tese doutoramento sofia catarino.pdf

160
UNIVERSIDADE TÉCNICA DE LISBOA INSTITUTO SUPERIOR DE AGRONOMIA Metais contaminantes nos vinhos. Ocorrência por influência das bentonites Sofia Cristina Gomes Catarino Orientador: Engenheiro António Sérgio Curvelo Garcia Co-orientador: Doutor Raul Filipe Xisto Bruno de Sousa Júri: Presidente: Reitor da Universidade Técnica de Lisboa Vogais: Doutora Maria Teresa Sá Dias de Vasconcelos - Professora Catedrática da Faculdade de Ciências da Universidade do Porto Doutora Maria Arlete Mendes Faia - Professora Catedrática da Universidade de Trás-os- Montes e Alto Douro Doutor Raul Filipe Xisto Bruno de Sousa - Professor Catedrático do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Doutor Manuel Armando Valeriano Madeira - Professor Catedrático do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Engenheiro António Sérgio Curvelo Garcia - Investigador Coordenador da Estação Vitivinícola Nacional do Instituto Nacional de Investigação Agrária e das Pescas Doutor Jorge Manuel Rodrigues Ricardo da Silva - Professor Associado do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Doutor Miguel Pedro de Freitas Barbosa Mourato - Professor Auxiliar do Instituto Superior de Agronomia da Universidade Técnica de Lisboa DOUTORAMENTO EM ENGENHARIA AGRO-INDUSTRIAL Lisboa 2006

Upload: vankhue

Post on 08-Jan-2017

256 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Tese doutoramento Sofia Catarino.pdf

UNIVERSIDADE TÉCNICA DE LISBOA

INSTITUTO SUPERIOR DE AGRONOMIA

Metais contaminantes nos vinhos.

Ocorrência por influência das bentonites

Sofia Cristina Gomes Catarino

Orientador: Engenheiro António Sérgio Curvelo Garcia Co-orientador: Doutor Raul Filipe Xisto Bruno de Sousa Júri: Presidente: Reitor da Universidade Técnica de Lisboa Vogais: Doutora Maria Teresa Sá Dias de Vasconcelos - Professora Catedrática da Faculdade de

Ciências da Universidade do Porto Doutora Maria Arlete Mendes Faia - Professora Catedrática da Universidade de Trás-os-Montes e Alto Douro Doutor Raul Filipe Xisto Bruno de Sousa - Professor Catedrático do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Doutor Manuel Armando Valeriano Madeira - Professor Catedrático do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Engenheiro António Sérgio Curvelo Garcia - Investigador Coordenador da Estação Vitivinícola Nacional do Instituto Nacional de Investigação Agrária e das Pescas Doutor Jorge Manuel Rodrigues Ricardo da Silva - Professor Associado do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Doutor Miguel Pedro de Freitas Barbosa Mourato - Professor Auxiliar do Instituto Superior de Agronomia da Universidade Técnica de Lisboa

DOUTORAMENTO EM ENGENHARIA AGRO-INDUSTRIAL

Lisboa 2006

Page 2: Tese doutoramento Sofia Catarino.pdf

UNIVERSIDADE TÉCNICA DE LISBOA

INSTITUTO SUPERIOR DE AGRONOMIA

Metais contaminantes nos vinhos.

Ocorrência por influência das bentonites

Dissertação apresentada neste Instituto para a obtenção do grau de Doutor

Sofia Cristina Gomes Catarino

Orientador: Engenheiro António Sérgio Curvelo Garcia Co-orientador: Doutor Raul Filipe Xisto Bruno de Sousa Júri: Presidente: Reitor da Universidade Técnica de Lisboa Vogais: Doutora Maria Teresa Sá Dias de Vasconcelos - Professora Catedrática da Faculdade de

Ciências da Universidade do Porto Doutora Maria Arlete Mendes Faia - Professora Catedrática da Universidade de Trás-os-Montes e Alto Douro Doutor Raul Filipe Xisto Bruno de Sousa - Professor Catedrático do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Doutor Manuel Armando Valeriano Madeira - Professor Catedrático do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Engenheiro António Sérgio Curvelo Garcia - Investigador Coordenador da Estação Vitivinícola Nacional do Instituto Nacional de Investigação Agrária e das Pescas Doutor Jorge Manuel Rodrigues Ricardo da Silva - Professor Associado do Instituto Superior de Agronomia da Universidade Técnica de Lisboa Doutor Miguel Pedro de Freitas Barbosa Mourato - Professor Auxiliar do Instituto Superior de Agronomia da Universidade Técnica de Lisboa

DOUTORAMENTO EM ENGENHARIA AGRO-INDUSTRIAL

Lisboa 2006

Page 3: Tese doutoramento Sofia Catarino.pdf

O presente trabalho foi realizado na Estação Vitivinícola Nacional (EVN) e financiado

pelos projectos PARLE - Projecto A e PIDDAC 720/02. Complementarmente obteve também

financiamento da Fundação para a Ciência e a Tecnologia, através de uma bolsa de doutoramento

(POCI 2010, medida IV.3, BD/17237/2004).

Page 4: Tese doutoramento Sofia Catarino.pdf

Agradecimentos

Fruto da intensidade com que foram vividos, os aproximadamente cinco anos que este trabalho

demorou a preparar passaram muito rapidamente. Os primeiros passos datam contudo de 1997,

ano em que sob a orientação do Investigador Coordenador A.S. Curvelo-Garcia comecei a estudar

a composição mineral dos vinhos, e mais tarde traduzidos sob a forma de tese de mestrado. Ao

longo destes anos colocaram-se vários desafios maiores: a entrada em funcionamento de um

equipamento de ETAAS, a acreditação do laboratório, mais recentemente a selecção, instalação e

entrada em funcionamento de um ICP-MS, equipamento raro no nosso país.

Nesta etapa final desejo expressar os meus agradecimentos a todos os que ajudaram a tornar

realidade este trabalho.

À Estação Vitivinícola Nacional, verdadeira entidade de referência na investigação em viticultura

e enologia em Portugal e a nível internacional, à qual se encontra estreitamente ligado o meu

percurso científico. Agradecimento extensivo a todos os seus trabalhadores.

Ao seu Director, e meu orientador, Investigador Coordenador A.S. Curvelo-Garcia, cuja

sensibilidade e trabalho promovem a permanente construção da EVN, por ter proporcionado

condições para o crescimento e afirmação do sector de análise mineral. Obrigada pela partilha de

conhecimentos e experiências, pela capacidade de visão, pelos “2 minutos” sempre disponíveis. A

confiança depositada, a liberdade concedida, o rigor exigido. Sobretudo, a amizade.

Ao Professor R. Bruno de Sousa, um agradecimento muito especial por ter aceitado a co-

orientação deste trabalho. Sempre presente nos momentos essenciais, com críticas e sugestões

preciosas, apontando pistas e ajudando a vencer as dificuldades. Obrigada por tudo!

À responsável pelo Departamento de Enologia da EVN, Investigadora Principal Isabel Spranger,

por ter tornado possível a concretização deste trabalho.

A todas as pessoas que me ajudaram no “arranque” do equipamento de ICP-MS, Professora

Maria Teresa Vasconcelos e Doutora Cristina Almeida (LAQUIPAI – Faculdade de Ciências –

UP), Engª Mª José Machado (Instituto Geológico e Mineiro - Porto), Dr. Pedro Brito (IPIMAR),

Doutor Bernard Médina e Drª Annie-Claude Ladrat (DGCCRF, Bordéus).

Aos investigadores de várias instituições pelas preciosas colaborações iniciadas neste trabalho,

Doutor José-Luís Capelo (Departamento de Engenharia Química - Instituto Superior Técnico –

UTL, REQUIMTE – Faculdade de Ciências e Tecnologia – UNL), Professor Manuel Madeira e

Professor Fernando Monteiro (Departamento de Ciências do Ambiente - Instituto Superior de

Agronomia – UTL), Professor Fernando Rocha (Departamento de Geociências - UA).

Page 5: Tese doutoramento Sofia Catarino.pdf

À Professora Ana Mota, pela disponibilização de condições para o tratamento de mostos nos

laboratórios do Departamento de Engenharia Química do Instituto Superior Técnico.

À Professora Manuela Abreu (Departamento de Ciências do Ambiente - Instituto Superior de

Agronomia – UTL), pelo interesse e sugestões relativas à linha de trabalho das bentonites.

À Doutora Margarida Baleiras Couto, pela valiosíssima revisão dos textos em inglês.

Às Investigadoras Auxiliares Ilda Caldeira e Sara Canas, pelas sugestões relativas ao tratamento

estatístico de dados. À Ilda também pela leitura crítica de alguns textos. Apetece dizer “Ilda, foste

muito bem!”

Ao Investigador Auxiliar José Silvestre, pelo apoio prestado ao nível da informática.

Aos Estagiários sob minha orientação durante estes anos, Daniela Pinto, José Soares, Inês

Pimentel, Vera Silva e Ricardo Braz, pelo seu contributo para este trabalho. Sobretudo por terem

possibilitado a realização de actividades que me dão bastante prazer: a orientação e o ensino.

À Mestre Nilza Eiriz, minha companheira de “aquário”.

Ao Ministério da Ciência, Tecnologia e Ensino Superior – Fundação para a Ciência e a

Tecnologia, pelo financiamento do PARLE - Projecto A, no qual o presente trabalho se inseriu e

pela bolsa de doutoramento atribuída em 2004 (BD/17237/2004).

A todos os que, pelo apoio e amizade, me ajudaram a realizar este objectivo.

À minha família, no princípio de tudo.

Page 6: Tese doutoramento Sofia Catarino.pdf

Filhos e versos, como os dás ao mundo?

Como na praia te conversam sombras de corais?

Como de angústia anoitecer profundo?

Como quem se reparte?

Como quem pode matar-te?

Ou como quem a ti não volta mais?

Jorge de Sena

Ao meu filho João

Page 7: Tese doutoramento Sofia Catarino.pdf

ÍNDICE

Resumo i

Abstract ii

Capítulo 1. INTRODUÇÃO GERAL

1.1. Sobre a composição mineral do vinho 3

1.2. Metais contaminantes do vinho 5

1.2.1. Origens da sua presença, teores, influência dos factores tecnológicos e definição de

limites

6

1.2.2. Determinação analítica 23

1.3. Influência das bentonites na composição mineral do vinho 23

1.4. Objectivos 25

1.5. Referências bibliográficas 27

Capítulo 2. DESENVOLVIMENTO E VALIDAÇÃO DE MÉTODOS DE ANÁLISE

2.1. DETERMINATION OF ALUMINIUM IN WINE BY GRAPHITE FURNACE AAS: VALIDATION

OF ANALYTICAL METHOD

39

2.1.1. Introduction 40

2.1.2. Experimental 41

2.1.2.1. Instrumentation 41

2.1.2.2. Labware 41

2.1.2.3. Reagents and calibration 41

2.1.3. Results and discussion 42

2.1.3.1. Selection of aluminium furnace program 42

2.1.3.2. Algorithm calibration curve 43

2.1.3.3. Selectivity 44

2.1.3.4. Accuracy 44

2.1.3.5. Analytical limits 44

2.1.3.6. Precision 45

2.1.4. Conclusion 45

2.1.5. References 45

2.2. DETERMINATION OF COPPER IN WINE BY ETAAS USING CONVENTIONAL AND FAST

THERMAL PROGRAMS: VALIDATION OF ANALYTICAL METHOD

49

2.2.1. Introduction 50

2.2.2. Experimental 51

Page 8: Tese doutoramento Sofia Catarino.pdf

2.2.2.1. Instrumentation and analytical conditions 51

2.2.2.2. Materials 52

2.2.2.3. Reagents and calibration 52

2.2.3. Results and discussion 53

2.2.3.1. Thermal programs 53

2.2.3.2. Validation 56

2.2.3.2.1. Algorithm calibration curve, selectivity and accuracy 56

2.2.3.2.2. Analytical limits 58

2.2.3.2.3. Precision 58

2.2.4. Conclusion 58

2.2.5. References 58

2.3. FOCUSED ULTRASOUND VERSUS MICROWAVE DIGESTION FOR THE DETERMINATION

OF LEAD IN MUST BY ELECROTHERMAL-ATOMIC ABSORPTION SPECTROMETRY

61

2.3.1. Introduction 62

2.3.2. Experimental 63

2.3.2.1. Apparatus 63

2.3.2.2. Reagents 64

2.3.2.3. Sample collection 65

2.3.2.4. Sample treatment 66

2.3.3. Results and discussion 66

2.3.3.1. Thermal program 66

2.3.3.2. Ultrasonic treatment 69

2.3.3.3. Analytical figures of merit 70

2.3.3.4. Determination of Pb in must samples 71

2.3.4. Conclusions 72

2.3.5. References 72

2.4. MEASUREMENTS OF CONTAMINANT ELEMENTS IN WINE BY INDUCTIVELY COUPLED

PLASMA-MASS SPECTROMETRY: A COMPARISON OF TWO CALIBRATION APPROACHES

75

2.4.1. Introduction 76

2.4.2. Experimental 77

2.4.2.1. Apparatus 77

2.4.2.2. Material and reagents 78

2.4.2.3. Samples 78

2.4.2.4. ICP-MS determinations 79

2.4.2.5. Determination of Cl and Ca species interference ratios – an approach to improve

accuracy

79

Page 9: Tese doutoramento Sofia Catarino.pdf

2.4.2.6. Quantitative approach 79

2.4.2.7. Semi-quantitative approach 80

2.4.2.8. Validation of ICP-MS procedures 80

2.4.3. Results and discussion 81

2.4.3.1. Cl and Ca species interference ratios and mathematical corrections 81

2.4.3.2. ICP-MS quantitative and semi-quantitative approaches figures of merit 83

2.4.3.3. Comparison of analytical results obtained by ICP-MS quantitative and semi-

quantitative approaches

85

2.4.3.4. Validation of ICP-MS procedures 87

2.4.4. Conclusions 89

2.4.5. References 90

Capítulo 3. AVALIAÇÃO DA COMPOSIÇÃO DE MOSTOS E VINHOS EM METAIS

CONTAMINANTES

3.1. EVALUATION OF CONTAMINANT ELEMENTS IN PORTUGUESE WINES AND ORIGIN

MUSTS BY HIGH INTENSITY FOCUSED ULTRASOUND COMBINED WITH INDUCTIVELY

COUPLED PLASMA MASS SPECTROMETRY

95

3.1.1. Introduction 96

3.1.2. Materials and methods 97

3.1.2.1. Apparatus 97

3.1.2.2. Material and reagents 97

3.1.2.3. ICP-MS determinations 98

3.1.2.4. Musts and wines 99

3.1.2.5. Microvinification processes 99

3.1.2.6. Must sample collection 99

3.1.2.7. Must sample treatment by HIFU 99

3.1.3. Results and discussion 100

3.1.3.1. Ultrasonic treatment 100

3.1.3.2. Must ICP-MS analysis 100

3.1.3.3. Contents of elements/changes from must to wine 101

3.1.4. Conclusions 106

3.1.5. References 106

Capítulo 4. CEDÊNCIA AO VINHO DE METAIS CONTAMINANTES DE BENTONITES

4.1. RELEASE OF CONTAMINANT ELEMENTS FROM BENTONITES TO WINE: EFFECT OF

BENTONITE CHARACTERISTICS AND WINE PH

111

4.1.1. Introduction 112

Page 10: Tese doutoramento Sofia Catarino.pdf

4.1.2. Materials and methods 113

4.1.2.1. Bentonite characteristics 113

4.1.2.2. Extraction essays 115

4.1.3. Results and discussion 116

4.1.3.1. Physical and chemical characteristics of bentonites 116

4.1.3.2. Mineral contents in bentonite treated wines 119

4.1.3.3. Physical and chemical characteristics of bentonites and mineral release 124

4.1.3.4. Interaction between wine pH and bentonites 125

4.1.4. Conclusions 126

4.1.5. References 127

4.2. RELEASE OF CONTAMINANT ELEMENTS FROM BENTONITES TO WINE: A CONTRIBUTION

TO ACHIEVE A TEST SOLUTION

129

4.2.1. Introduction 130

4.2.2. Material and methods 131

4.2.2.1. Test solutions 131

4.2.2.2. Extraction essays 131

4.2.2.3. Mineral composition analysis 132

4.2.2.4. Statistical analysis 133

4.2.3. Results and discussion 133

4.2.4. Conclusions 141

4.2.5. References 142

Capítulo 5 – CONSIDERAÇÕES FINAIS

5.1. Considerações finais 147

5.2. Referências bibliográficas 155

Divulgação do conhecimento 157

Page 11: Tese doutoramento Sofia Catarino.pdf

i

RESUMO

Foram desenvolvidos e validados métodos de análise para a determinação de metais

contaminantes dos vinhos, por ETAAS e ICP-MS, e uma metodologia inovadora de preparação de

mostos por ultra-sons de alta intensidade e focalização.

Avaliou-se a composição de vinhos e mostos correspondentes em metais contaminantes,

verificando-se que a variação do mosto para o vinho depende do elemento e da sua concentração.

Num vasto conjunto de metais contaminantes ocorreram decréscimos significativos, revelando uma

importante capacidade natural de eliminação.

No estudo da cedência ao vinho de metais contaminantes de bentonites, observaram-se

enriquecimentos para a generalidade dos elementos, fortemente relacionados com as características

físico-químicas das bentonites, sendo maiores a pH mais baixo. A caracterização físico-química das

bentonites, pela primeira vez relacionada com a sua cedência mineral, permitiu constatar não

conformidades com as especificações da OIV, revelando a necessidade de controlo efectivo. Os

resultados sugerem que a percentagem de montmorilonite não é um indicador satisfatório da

reactividade da bentonite. Foram ensaiadas soluções de extracção para controlo da qualidade das

bentonites, verificando-se que a natureza da proteína é um factor importante. A solução de

extracção com proteína de vinho simulou melhor o vinho que a solução teste estabelecida pela

OIV.

Palavras-chave: metais contaminantes, vinho, mosto, bentonite, ETAAS, ICP-MS.

Page 12: Tese doutoramento Sofia Catarino.pdf

ii

Occurrence of contaminant metals in wines by bentonite influence

ABSTRACT

A new methodology based on focused ultrasound for must pre-treatment and analytical

methods for the determination of contaminant metals of wines, by ETAAS and ICP-MS techniques,

have been developed and validated.

The composition of wines and their corresponding must in contaminant metals were

evaluated. The variation from must to wine depended on the element and its concentration. For a

large group of contaminant metals decreases were verified, suggesting an important capacity of

elimination.

The study of the release of contaminant elements from bentonites to wine showed

enrichment for the generality of the elements, strongly related to bentonites characteristics, and

highest at the lowest pH. Physical, chemical and mineralogical characteristics of bentonites, for the

first time connected to its mineral release, revealed several non-conformances with OIV

specifications, demonstrating the need for an effective control. The results suggested that

nontmorillonite content is not a satisfactory indicator of bentonite reactivity. For bentonite quality

control purposes extraction solutions were essayed. The protein nature is an important factor on

extraction. The extraction solution with wine protein was a better simulator of wine than the test

solution indicated by OIV.

Key-words: contaminant metals, wine, must, bentonite, ETAAS, ICP-MS.

Page 13: Tese doutoramento Sofia Catarino.pdf

Capítulo 1

INTRODUÇÃO GERAL

Page 14: Tese doutoramento Sofia Catarino.pdf

Introdução geral

3

1.1. SOBRE A COMPOSIÇÃO MINERAL DO VINHO

A composição mineral do vinho reflecte a sua origem e percurso particulares, sendo por

isso singular e identificadora, contribuindo de forma substancial para as características sensoriais,

com influência na cor, limpidez, gosto e aroma.

Representando cerca de 1,5 a 3 g/L no seu conjunto (Ribéreau-Gayon et al., 1982), os

elementos minerais presentes no vinho provêm, em grande parte, da absorção radicular,

verificando-se um constante enriquecimento durante a formação e maturação do bago. À

semelhança de grande parte das plantas, a videira tem a capacidade de absorver quantidades

relativamente elevadas de elementos tóxicos, sem que se verifique manifestação de toxicidade, uma

vez que os metais são complexados e integrados em moléculas biológicas num mecanismo de

autodefesa da planta (Prasad, 1998).

Um critério frequentemente utilizado na classificação dos elementos minerais do vinho

consiste na sua expressão quantitativa. De acordo com esta abordagem clássica, como exemplos de

elementos maioritários, em concentrações de 10 mg/L até 1 g/L, incluem-se alguns metais alcalinos

e alcalino-terrosos, tais como o Na, K, Mg e Ca, principais responsáveis pela “estrutura metálica”

dos vinhos e pela sua capacidade tampão ácido-base. Outros elementos químicos que integram este

grupo são o Si sob a forma de ácido silícico, o P, presente no vinho essencialmente sob a forma

mineral (fosfatos) mas também orgânica, o S sob a forma de sulfatos, sulfitos e outras espécies, e o

Cl sob a forma de cloretos.

A um nível de concentração inferior, geralmente entre 0,1 mg/L e 10 mg/L, são exemplo o

B sob a forma de ácido bórico, o Al, Mn, Fe, Cu, Zn, Rb, Sr e o Mo, elementos minoritários, e na

sua grande maioria oligo-elementos (indispensáveis aos seres vivos em pequenas quantidades). Em

teores normalmente inferiores a 100 µg/L, o Li, V, Cr, Co, Ni, Ga, As, Se, Cs, Ba, Pb, Br, I e F, são

elementos vestigiais. A concentração desce para níveis inferiores a 1 µg/L para elementos (sub-

vestigiais) como o Be, Cd, Sb, W, Hg, Tl, Bi, U e lantanídeos (terras raras). Naturalmente, os

intervalos de variação entre vinhos são, para a maioria dos elementos, bastante consideráveis, pelo

que as fronteiras entre os diferentes grupos de classificação são apenas indicativas. Por outro lado,

não existe unanimidade quanto ao número de classes, por vezes distinguindo-se apenas entre

elementos maioritários e elementos vestigiais (Eschnauer, 1982a). Há ainda autores que

diferenciam entre elementos maioritários, oligoelementos e elementos vestigiais, incluindo neste

último grupo apenas aqueles que apresentam características nocivas (Flanzy, 1998).

O K, Ca, Fe e Cu, por estarem relacionados com fenómenos de instabilidade físico-química

dos vinhos, e o Na, por poder indiciar práticas enológicas não autorizadas, são desde há muito

tempo objecto de especial atenção no âmbito da Química Enológica. A carência de elementos

essenciais nos mostos e seus efeitos na actividade fermentativa de leveduras e bactérias lácticas têm

Page 15: Tese doutoramento Sofia Catarino.pdf

Introdução geral

4

sido também investigados. Numa perspectiva global, e apesar da sua não negligenciável

importância nutricional e do seu contributo fundamental para as características do vinho, verifica-

se que as abordagens sobre os elementos minerais do vinho têm incidido, fundamentalmente, sobre

os seus potenciais efeitos negativos.

Os avanços no conhecimento científico estão associados ao desenvolvimento de novas

tecnologias e vice-versa. No caso do estudo da composição mineral dos vinhos, a aplicação de

técnicas de análise multi-elementar e isotópica, com progressivo abaixamento dos limites analíticos

dos métodos, permitindo quantificar cada vez mais elementos no vinho, abriu novas perspectivas e

áreas de investigação.

Nos últimos vinte anos, com base no pressuposto de que a composição mineral do vinho

reflecte a composição mineral do solo de origem, diversos estudos têm sido realizados, com o

objectivo de garantir a origem geográfica e autenticidade do vinho (Gonzales-Larraina et al., 1987;

Herrero-Latorre e Médina, 1990; Day et al., 1995; Médina, 1996; Baxter et al., 1997; Greenough et

al., 1997; Rizzon et al., 1997; Martin et al., 1999; Almeida e Vasconcelos, 2003a; Frias et al.,

2003a; Taylor et al., 2003; Gómez et al., 2004a; Gremaud et al., 2004; Jos et al., 2004; Thiel et al.,

2004; Coetzee et al., 2005). A razão de isótopos estáveis de elementos, cuja composição isotópica

varie naturalmente com a região geográfica, como são os casos do Sr e do Pb, constitui um outro

potencial marcador da proveniência do vinho (Dean et al., 1990; Almeida e Vasconcelos, 2001;

Almeida e Vasconcelos, 2003b).

Com recurso a técnicas acopladas, especial atenção tem sido também dedicada ao estudo de

fenómenos de complexação de catiões metálicos, alguns deles com potencial tóxico, com

polifenóis, polissacáridos e proteínas (Pellerin et al., 1997; Szpunar et al., 1998; Esparza et al.,

2004; Salinas et al., 2005). Os elementos minerais encontram-se solubilizados no mosto e no vinho

sob a forma de sais orgânicos, tais como tartaratos, malatos, succinatos, acetatos, etc, de sais

minerais, como cloretos, sulfatos, fosfatos e de complexos orgânicos com outras espécies químicas.

A forma química em que o catião se encontra no vinho condiciona enormemente a sua

biodisponibilidade e toxicidade.

O enriquecimento do vinho em alguns metais pode originar fenómenos depreciativos da

sua qualidade tais como turvações, precipitações e oxidações. Para além disso, e dependendo da

concentração do elemento, podem substancializar problemas de ordem toxicológica e legal. A

investigação sobre a presença de metais contaminantes no vinho, nomeadamente no que respeita

aos níveis de ocorrência, evolução ao longo do processo tecnológico, identificação de fontes de

contaminação e ao próprio desenvolvimento de métodos de análise apropriados ao seu doseamento,

tem beneficiado enormemente com a disponibilidade de técnicas analíticas cada vez mais

poderosas. Estes estudos, interpretação e significado dos seus resultados, apresentam, para além do

Page 16: Tese doutoramento Sofia Catarino.pdf

Introdução geral

5

inegável interesse académico, relevância sob o ponto de vista tecnológico, nutricional e

especialmente de segurança alimentar.

De resto, é objectivo da Subcomissão de Métodos de Análise da Organização Internacional

da Vinha e do Vinho (OIV), a constituição de um “banco de dados mundial” sobre os teores em

contaminantes metálicos dos vinhos, em associação com estudos para identificação da sua origem e

minimização dos seus teores, visando o estabelecimento de limites com maior fundamento

científico (OIV, 1995a).

1.2. METAIS CONTAMINANTES DO VINHO

Por metais contaminantes dos vinhos entende-se normalmente o conjunto dos metais

pesados, associando-lhe o Al, e ainda alguns não metais tais como o As e o Se, e mais

recentemente o Be, com exclusão dos metais alcalinos e restantes alcalino-terrosos. É nesta

perspectiva, em sentido lato, que esta temática será abordada.

A designação de metal pesado, frequentemente utilizada na classificação dos elementos

minerais, não possui uma definição única, variando de acordo com o ramo da ciência que o aborda.

A ideia comum às diferentes áreas é a de serem metais ou não metais que apresentam uma

densidade relativamente alta, associados com poluição e toxicidade, ainda que alguns elementos

sejam essenciais para os seres vivos em baixas concentrações (V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As,

Se, Mo, entre outros) (Hurrel et al., 2001). De facto, a essencialidade ou toxicidade do elemento

depende do nível de concentração em que este se apresenta. A título de exemplo, o Mn, Fe e o Cu

são indispensáveis ao desenvolvimento de leveduras e bactérias lácticas (Ribéreau-Gayon et al.,

1998), intervindo como co-factores na actividade de enzimas, tais como as oxireductases e as

quinases. Pelo contrário, o Cd e o Pb são potencialmente tóxicos mesmo em pequenas quantidades,

e a sua essencialidade está ainda por esclarecer.

A presença de metais pesados no vinho está directamente relacionada com o

desenvolvimento da actividade industrial e com a poluição gerada. De acordo com Ribéreau-Gayon

et al. (1998), todos os catiões minerais estão presentes naturalmente nos mostos e vinhos em teores

não tóxicos (presença endógena). Ao longo do processo tecnológico, podem ocorrer contaminações

de origem diversa: atmosférica, práticas culturais, aditivos e auxiliares tecnológicos, equipamentos

utilizados na vinificação, estabilização e conservação (tendo em conta a sua acidez, o vinho é

susceptível de, por corrosão dos materiais, dissolver alguns catiões metálicos).

No sentido oposto, como fenómeno natural de estabilização do vinho, no decorrer da

fermentação alcoólica e da fermentação maloláctica, ocorre uma eliminação parcial destes metais,

Page 17: Tese doutoramento Sofia Catarino.pdf

Introdução geral

6

por precipitação sob a forma de sais orgânicos e/ou sulfuretos, e por absorção e adsorção por

leveduras e bactérias (Blackwell et al., 1995; Ruzic, 2000; Brandolini et al., 2002).

1.2.1. Origens da sua presença, teores, influência dos factores tecnológicos e definição

de limites

Aspectos como a origem, níveis de ocorrência, evolução ao longo do processo

tecnológicos, fontes de contaminação, estabilidade fisico-química, distribuição sob diferentes

formas químicas, toxicidade e métodos de análise, entre outros, variam consideravelmente com o

metal em consideração. Alguns metais contaminantes adquiriram uma prioridade de estudo

superior à de outros, eventualmente por estes se encontrarem em quantidades extremamente

reduzidas (tendo em conta a importância toxicológica dessa presença) ou ainda devido à

dificuldade do seu doseamento, com evidentes reflexos na extensão da bibliografia existente.

Segue-se a descrição pormenorizada dos principais metais contaminantes do vinho, apresentados

por ordem crescente de número atómico.

Berílio

Não se inclui no grupo de metais pesados, embora seja um elemento potencialmente tóxico.

De acordo com a revisão bibliográfica, o Be surge nos vinhos normalmente em teores inferiores a 5

µg/L (Eschnauer, 1982a; Greenough et al., 1997; Taylor et al., 2003; Thiel et al., 2004). A

bentonite é referida como fonte de contaminação especialmente relevante, podendo triplicar ou

quadriplicar a sua concentração (Gómez et al., 2004b; Nicolini et al., 2004).

A determinação analítica deste elemento nos vinhos, mais generalizada, é certamente a que

recorre à técnica de espectrometria de massa com plasma acoplado por indução (Inductively

Coupled Plasma - Mass Spectrometry, ICP-MS) (Baxter et al., 1997; Pérez-Jordán et al., 1998;

Almeida e Vasconcelos, 2002; Gómez et al., 2004a; Thiel et al., 2004).

Alumínio

O envolvimento do Al em diversas disfunções neurológicas, particularmente na doença de

Alzheimer, tem sido objecto de numerosos estudos e de alguma controvérsia (Aikoh e Nishio,

1996; Seruga et al., 1998). Este elemento apresenta como especificidade enológica a capacidade de

se combinar com ácidos orgânicos, o que explica a sua biodisponibilidade. A sua presença no vinho

é de natureza essencialmente exógena, associada à utilização de pesticidas (com Al), partículas de

terra, contacto com superfícies de alumínio e alguns produtos enológicos, entre os quais as

Page 18: Tese doutoramento Sofia Catarino.pdf

Introdução geral

7

bentonites, taninos e adjuvantes de filtração (Ribéreau-Gayon et al., 1982; McKinnon et al., 1992;

Scollary, 1997; Kelly et al., 2005). A bentonite é, aliás, considerada a mais importante fonte de

contaminação de Al do vinho, tendo sido observados aumentos de aproximadamente 100%

(McKinnon et al., 1992; Nicolini et al., 2004).

Os dados publicados indicam que o seu teor nos vinhos é normalmente inferior a 2 mg/L

(Larroque et al., 1994; Lopez et al., 1998; Seruga et al., 1998; Taylor et al., 2003; Jos et al., 2004).

Quando em teores superiores a 10 mg/L pode estar na origem de acidentes de estabilidade físico-

química e gostos e aroma depreciativos (Lay e Meyer, 1989; Scollary, 1997; Seruga et al., 1998).

O doseamento pode ser realizado com recurso a diversas técnicas analíticas, sendo as mais

vulgarizadas a espectrofotometria de absorção atómica com atomização electrotérmica

(Electrothermal Atomic Absorption Spectrometry, ETAAS) (Almeida et al., 1992; McKinnon et al.,

1992; Larroque et al., 1994; Canuto et al., 2004), espectrometria de emissão com plasma acoplado

por indução (Inductively Coupled Plasma – Optical Emission Spectroscopy, ICP-OES) (Jos et al.,

2004; Nicolini et al., 2004) e ICP-MS (Aikoh e Nishio, 1996; Rodushkin et al., 1999; Castiñeira et

al., 2001; Almeida e Vasconcelos, 2002; Coetzee et al., 2005). Para estudos de especiação é

referida a voltametria de redissolução anódica diferencial com impulsos (Salinas et al., 2005).

Vanádio

Trata-se de um metal essencial que nos últimos anos foi objecto de especial interesse

farmacológico devido às suas propriedades anti-diabéticas, entre outras (Teissèdre et al., 1998a;

Hurrel et al., 2001). Os níveis da sua presença são especialmente elevados nos combustíveis

fósseis, pelo que em zonas industriais ou de elevado tráfego rodoviário, a contaminação

atmosférica é bastante relevante. Este metal é também utilizado em algumas ligas metálicas,

especialmente no aço inoxidável (Teissèdre et al., 1998a).

De acordo com a revisão bibliográfica realizada, o intervalo de variação dos teores de V

nos vinhos, normalmente superiores aos dos mostos, é bastante largo (1 a 447 µg/L), com um valor

médio da ordem de 50 µg/L (Teissèdre et al., 1998a; Taylor et al., 2003; Gómez et al., 2004a),

ainda que alguns autores refiram valores ligeiramente mais baixos (Greenough et al., 1997). Para

além da sua presença endógena, a contaminação pelo aço inoxidável e pelos óxidos metálicos

utilizados na pigmentação das garrafas podem aumentar o seu teor.

A determinação analítica deste metal em vinhos pode ser realizada por ETAAS (Teissèdre

et al., 1998a), por ICP-OES (Nicolini et al., 2004) e por ICP-MS (Stroh et al., 1994; Baxter et al.,

1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Almeida e Vasconcelos, 2002; Almeida et

al., 2002; Taylor et al., 2003).

Page 19: Tese doutoramento Sofia Catarino.pdf

Introdução geral

8

Crómio

O Cr é um elemento essencial ao metabolismo da glucose, colesterol e proteínas (Hurrel et

al., 2001). A sua toxicidade depende da forma química, sendo os compostos de Cr(VI)

extremamente tóxicos, mutagénicos e carcinogénicos. A forma Cr(III), a mais frequente no vinho,

apresenta toxicidade reduzida, sendo a ingestão a via principal de introdução de Cr no organismo,

não devendo exceder 200 µg por dia (Lendinez et al., 1998).

À semelhança da maioria dos metais, factores de ordem antropogénica podem contribuir

para o aumento do seu conteúdo no ambiente. A contaminação dos solos contribui para a absorção

radicular. A dissolução do Cr a partir do aço inoxidável é provavelmente a fonte de contaminação

principal do vinho (Médina e Sudraud, 1980; Eschnauer, 1982b; Cabrera-Vique et al., 1997),

responsável pela tendência para o aumento do teor de Cr nos vinhos observada nas últimas décadas.

Também a conservação em garrafa de vidro pode conduzir a um enriquecimento em Cr (Médina e

Sudraud, 1980). De acordo com a revisão bibliográfica, surge normalmente em concentrações

inferiores a 60 µg/L nos vinhos (Ribéreau-Gayon et al., 1982; Cabrera-Vique et al., 1997;

Greenough et al., 1997; Lendinez et al., 1998).

O seu doseamento no vinho pode ser realizado por ETAAS (Médina e Sudraud, 1980;

Cabrera-Vique et al., 1997; Lendinez et al., 1998; Kristl et al., 2002), ICP-OES (Nicolini et al.,

2004) e por ICP-MS (Baxter et al., 1997; Rodushkin et al., 1999; Almeida e Vasconcelos, 2002).

Manganês

O Mn é um oligoelemento com intervenção a vários níveis no organismo humano. A sua

deficiência, bem como a sua toxicidade quando em excesso, podem afectar o metabolismo cerebral

(Cabrera-Vique et al., 2000).

O seu teor nos vinhos é característico dos solos de onde são provenientes, e normalmente

inferior a 3 mg/L, verificando-se no entanto uma enorme variabilidadede dos valores extremos e

médio (Gonzales-Larraina et al., 1987; Herrero-Latorre e Médina, 1990; Latorre et al., 1992; Day

et al., 1995; Rizzon et al., 1997; Cabrera-Vique et al., 2000; Costa et al., 2000; Taylor et al., 2003;

Jos et al., 2004). A natureza exógena deste elemento é associada à utilização de produtos

fitossanitários contendo sais de Mn, conservação em alguns tipos de inox, bentonites e enzimas

pectolíticas (Cabrera-Vique et al., 2000). O envolvimento do Mn em processos de oxidação do

vinho é conhecido (Cacho et al., 1995).

O doseamento nos vinhos pode ser realizado por espectrofotometria de absorção atómica

com chama (Flame Atomic Absorption Spectrometry, FAAS) (Gonzalez-Larraina et al., 1987;

Latorre et al., 1992; Rizzon et al., 1997; Frías et al., 2003), por Sequential Injection Analysis –

Page 20: Tese doutoramento Sofia Catarino.pdf

Introdução geral

9

FAAS (SAI-FAAS) (Costa et al., 2000), ETAAS (Latorre et al., 1992; Cabrera-Vique et al., 2000),

ICP-OES (Gremaud et al., 2004; Nicolini et al., 2004) e ICP-MS (Stroh et al., 1994; Baxter et al.,

1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Castiñeira et al., 2001; Almeida e

Vasconcelos, 2002).

Ferro

O vinho contém normalmente alguns mg/L de Fe (2 a 20 mg/L), em parte de natureza

endógena (2 a 5 mg/L) (Curvelo-Garcia, 1988; Latorre et al., 1992; Martin de la Hinojosa et al.,

1994; Sudraud et al., 1994; Escobal et al., 1995; Costa et al., 2000; González e Peña-Méndez,

2000; Taylor et al., 2003; Gremaud et al., 2004). O restante é veiculado pelas partículas de terra

que acompanham a uva, diferentes equipamentos utilizados e materiais utilizados na vindima e nos

processos tecnológicos de vinificação, estabilização e conservação dos vinhos, incluindo alguns

aditivos e auxiliares tecnológicos (Curvelo-Garcia e Ghira, 1979; Curvelo-Garcia e Catarino,

1998). Parte do Fe presente no mosto precipita em quantidade variável, dependendo das condições

de oxidação-redução durante e após a fermentação.

Quando em teores de 10 a 20 mg/L pode provocar turvações ou alterações da cor. Em

vinhos conservados ao abrigo do ar, o Fe encontra-se sob a forma Fe(II), solúvel. Quando ocorre

arejamento, o Fe(II) pode ser oxidado a Fe(III), capaz de precipitar a matéria corante dos vinhos

tintos (casse azul) ou o ácido fosfórico (casse branca), essencialmente nos vinhos brancos. A

generalização do emprego de aço inoxidável contribuiu fortemente para a redução do risco de

excesso de Fe e de ocorrência de casse férrica. O Fe desempenha também um papel importante nos

fenómenos de oxidação, como catalisador, e nos fenómenos de envelhecimento (Cacho et al., 1995;

Ribéreau-Gayon et al., 1998). Aos níveis em que se encontra nos vinhos, não é susceptível de

provocar problemas de ordem toxicológica.

Os teores em que se apresenta nos vinhos permitem geralmente o seu doseamento por

FAAS (Escobal et al., 1995; Catarino et al., 2003; Frias et al., 2003a). Outros métodos têm sido

descritos e aplicados: SAI-FAAS (Costa et al., 2000), ICP-OES (Jos et al., 2004; Nicolini et al.,

2004), ICP-MS (Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Castiñeira et

al., 2001; Almeida e Vasconcelos, 2002; Taylor et al., 2003). No âmbito da OIV (OIV, 2005) e da

União Europeia (CEE, 1990) são descritos dois métodos, um método colorimétrico (usual) e um

método por FAAS (referência).

Page 21: Tese doutoramento Sofia Catarino.pdf

Introdução geral

10

Cobalto

A toxicidade deste elemento revela-se através de manifestações alérgicas (Berg e Licht,

2002). Os teores de Co em vinhos referidos na literatura são normalmente inferiores a 10 µg/L

(Soares et al., 1995; Greenough et al., 1997; Gómez et al., 2004a; Thiel et al., 2004). A sua

presença exógena no vinho resulta principalmente do ataque a algumas ligas metálicas com este

elemento (Curvelo-Garcia, 1988) e do tratamento de mostos e vinhos com bentonites (Nicolini et

al., 2004).

Nos últimos anos a determinação analítica do Co em vinhos tem sido realizada por ICP-MS

(Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Almeida e Vasconcelos,

2002; Almeida et al., 2002; Taylor et al., 2003; Thiel et al., 2004). É possível ainda encontrar

referências ao seu doseamento por FAAS, ETAAS, ICP-OES e voltametria (Soares et al., 1995).

Níquel

Nos últimos anos a importância e interesse do Ni cresceram marcadamente, fruto de

descobertas no âmbito das suas implicações toxicológicas e fisiológicas (Hurrel et al., 2001).

Essencial em baixas concentrações, tóxico em elevadas concentrações, este metal pesado surge

associado a dermatites, podendo ter um efeito cancerígeno (por ingestão), embora os estudos

epidemiológicos em humanos sejam ainda escassos (Teissèdre et al., 1998b; Berg e Licht, 2002).

A presença endógena de Ni nos vinhos deve-se à absorção radicular, sendo contudo a

utilização de equipamentos em aço inoxidável apontada como a principal origem deste metal nos

vinhos (Eschnauer, 1982b; Teissèdre et al., 1998b). Outras fontes de contaminação possíveis são os

produtos fitossanitários, contaminação atmosférica, bentonites e a conservação em garrafa devido

aos pigmentos (contendo Ni como impureza) utilizados no seu fabrico. À semelhança de outros

metais pesados presentes nas emissões de veículos rodoviários, a sua concentração desce de forma

bastante acentuada com a distância à via rodoviária (Teissèdre et al., 1998b).

Nas últimas décadas parece ter ocorrido um ligeiro aumento do teor de Ni nos vinhos, na

ordem de 20 a 30 µg/L, provavelmente associado à introdução de materiais e equipamentos em aço

inoxidável (Ribéreau-Gayon et al., 1982; LaTorre et al., 1992; Teissèdre et al., 1998b; Jos et al.,

2004; Thiel et al., 2004).

A técnica de ETAAS é perfeitamente adequada à sua determinação analítica (Teissèdre et

al., 1998b; Jos et al., 2004), sendo também referidas a FAAS (LaTorre et al., 1992), ICP-OES

(Nicolini et al., 2004) e ICP-MS (Stroh et al., 1994; Baxter et al., 1997; Pérez-Jordán et al., 1998;

Rodushkin et al., 1999; Castiñeira et al., 2001; Almeida e Vasconcelos, 2002; Almeida et al., 2002;

Thiel et al., 2004).

Page 22: Tese doutoramento Sofia Catarino.pdf

Introdução geral

11

Cobre

O Cu é um elemento indispensável para o funcionamento normal dos tecidos vegetais,

sendo co-factor de numerosas reacções enzimáticas.

A origem do Cu nos vinhos deve-se à própria constituição das uvas (Ribéreau-Gayon et al.,

1982; Magalhães et al., 1985; Ribéreau-Gayon et al., 1998), mas essencialmente a tratamentos

cúpricos anticriptogâmicos, reflectindo-se na predominância do elemento sob a forma iónica

(Scollary, 1997; Fournier et al., 1998a; Salvo et al., 2003; Pietrzak e McPhail, 2004; Garcia-

Esparza et al., 2006), que poderá explicar a elevada capacidade de ser assimilado (Azenha e

Vasconcelos, 2000; Vasconcelos e Azenha, 2001).

Os mostos podem apresentar teores na ordem dos 10 a 20 mg/L, precipitando em grande

parte no decorrer da fermentação, sob a forma de sulfuretos extremamente insolúveis, juntamente

com leveduras, originando vinhos com teores relativamente pequenos, da ordem de 0,1 a 0,2 mg/L

(Ribéreau-Gayon et al, 1982; Curvelo-Garcia, 1988; LaTorre et al., 1992; Martin de la Hinojosa et

al., 1994; Sudraud et al., 1994; Escobal et al., 1995; Fournier et al., 1998b; Costa et al., 2000;

Taylor et al., 2003; Jos et al., 2004). Essa eliminação é favorecida pela presença de enxofre e

nitidamente acelerada quando o pH é elevado, sendo no entanto essencial a existência de proteínas

(Curvelo-Garcia, 1988). Em alguns casos, a adsorção de Cu pelas leveduras e bactérias lácticas

pode afectar a sua actividade (Vidal et al., 2001). Durante a conservação, o seu teor pode aumentar

por contacto do vinho com materiais de cobre, latão e bronze (Ribéreau-Gayon et al., 1982), e, nos

vinhos licorosos, através da aguardente adicionada (Almeida et al., 1994). No âmbito da OIV é

permitida a utilização de sulfato de cobre (1 g/hL) para eliminação de espécies químicas portadoras

de odores indesejáveis, o que poderá contribuir para o enriquecimento do vinho neste metal.

Quando em teores normalmente superiores a 1 mg/L, na presença de proteínas e em

ambiente redutor, pode estar na origem de uma turvação (casse cúprica) ou da formação de um

precipitado, fenómenos acelerados pela luz e temperatura elevada. No entanto mais do que o seu

teor total, a distribuição sob diferentes formas químicas parece ser um melhor indicador do

potencial de ocorrência destes fenómenos (Scollary, 1997). Pode actuar também como catalisador

da oxidação do ferro, originando a ocorrência de casse férrica (Ribéreau-Gayon et al., 1982). Por

outro lado, mesmo em baixos teores, o Cu actua como catalisador de oxidação dos vinhos (Cacho

et al., 1995; Scollary, 1997; Clark et al., 2003).

É de considerar a sua importância sob o ponto de vista toxicológico (Hurrel et al., 2001). O

limite máximo admissível para vinhos estabelecido pela OIV é de 1 mg/L de Cu (OIV, 2005).

O doseamento do Cu em vinhos pode ser realizado por FAAS (LaTorre et al., 1992; Costa

et al., 2000; Catarino et al., 2003), embora esta técnica apresente algumas dificuldades face aos

baixos teores em que este elemento se apresenta na generalidade dos vinhos. No âmbito da OIV

Page 23: Tese doutoramento Sofia Catarino.pdf

Introdução geral

12

(OIV, 2005) e da União Europeia (CEE, 1990), o único método descrito baseia-se exactamente

nesta técnica analítica. São várias as referências à aplicação da ETAAS (Almeida et al., 1994;

Escobal et al., 1995), ICP-OES (Jos et al., 2004; Nicolini et al., 2004) e ICP-MS (Stroh et al.,

1994; Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Castiñeira et al., 2001;

Almeida e Vasconcelos, 2002; Almeida et al., 2002; Taylor et al., 2003). A medição do teor total

pode ser realizada com recurso a técnicas electroquímicas, tais como a potenciometria de

redissolução e a voltametria de redissolução (Clark e Scollary, 2000; Brainina et al., 2004; Misiego

et al., 2004). A determinação analítica dos iões livres pode ser realizada por polarografia (Fournier

et al., 1998a).

Zinco

O Zn encontra-se na natureza sob a forma de minerais combinados com enxofre, numa

mistura de sulfuretos de Zn e de Pb (Ribéreau-Gayon et al., 1998), sendo um elemento essencial no

desenvolvimento e crescimento vegetal. Ao contrário do que se verifica para o Cu, parece não

haver acumulação significativa deste elemento nos solos (Magalhães et al., 1985).

Para além da presença endógena, como resultado da sua assimilação pela videira, é também

de referir, como significativa, a resultante da aplicação de fungicidas, sendo nomeados os

pertencentes à família dos ditiocarbamatos (Salvo et al., 2003). O contacto com materiais à base de

ligas metálicas contendo Zn (ferro galvanizado, latão) e da utilização de produtos enológicos

(Curvelo-Garcia, 1988; Fournier et al., 1998a) podem constituir formas de introdução de Zn no

vinho. Por um processo idêntico ao que se passa com o Cu (reduzida solubilidade do sulfureto

respectivo), é por outro lado conhecida a eliminação natural de uma parte significativa deste metal,

durante a fermentação alcoólica (Curvelo-Garcia, 1988).

À semelhança de outros metais, o teor total de Zn nos vinhos depende da intensidade dos

fenómenos de maceração, extracção e solubilização ocorridos durante a fermentação, já que se

localiza preferencialmente nas películas e graínhas da uva. Fermentações decorridas a temperaturas

mais elevadas, em igualdade das restantes condições, originam vinhos com maior teor de Zn

(Fournier et al., 1998a). A distribuição do Zn no vinho sob formas orgânicas e minerais é também

influenciada por este parâmetro, que favorece a solubilização de complexos organometálicos. Por

outro lado, a presença simultânea de formas orgânicas e minerais de Zn nos vinhos reflecte a sua

origem, as primeiras relacionadas com os fungicidas, as últimas por contacto com materiais

metálicos (Fournier et al., 1998a).

A generalidade dos valores encontrados na bibliografia é inferior 5 mg/L (Curvelo-Garcia e

Ghira, 1981; Martin de la Hinojosa et al., 1994; Sudraud et al., 1994; Soares et al., 1995;

Greenough et al., 1997; Rizzon et al., 1997; Costa et al., 2000; González e Peña-Méndez, 2000;

Page 24: Tese doutoramento Sofia Catarino.pdf

Introdução geral

13

Frias et al., 2003a; Taylor et al., 2003; Gómez et al., 2004a; Gremaud et al., 2004; Jos et al., 2004).

Não obstante os teores normalmente encontrados no vinho serem demasiado reduzidos para

provocar intoxicações, este metal pesado é reconhecido pela sua potencial acção tóxica sobre o

organismo humano. A OIV (2005) indica como limite máximo admissível em vinhos 5 mg/L, o que

se julga um valor relativamente elevado face às actuais possibilidades tecnológicas, tanto mais que

elevados teores de Zn poderão ser sede de acidentes de estabilidade fisico-química, e afectar as

características sápidas dos vinhos (foi relatada a existência de relação entre a sensação de

adstringência e a concentração de Zn) (Fournier et al., 1998a).

Os teores em que surge geralmente nos vinhos permitem o seu doseamento, após

desalcoolização, por FAAS, técnica na qual se baseia o método (único) indicado pela OIV (2005) e

pela União Europeia (CEE, 1990), sendo numerosas as referências à sua aplicação (LaTorre et al.,

1992; Day et al., 1995; Catarino et al., 2002; Frias et al., 2003a). O doseamento pode ser também

realizado por SIA-FAAS (Costa et al., 2000), ETAAS (Soares et al., 1995), ICP-OES (Jos et al.,

2004; Nicolini et al., 2004), ICP-MS (Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et

al., 1999; Castiñeira et al., 2001; Almeida e Vasconcelos, 2002; Almeida et al., 2002; Taylor et al.,

2003), por técnicas electroquímicas tais como a potenciometria de redissolução (Chen et al., 1994)

e a voltametria de redissolução (Mikkelsen e Schroder, 2002; Brainina et al., 2004). A

concentração de iões livres pode ser determinada por polarografia (Fournier et al., 1998a).

Arsénio

Em conjunto com o Pb, o Cd e o Hg é um dos elementos cuja presença está relacionada

com um elevado número de intoxicações ao longo da história da humanidade. As formas químicas

inorgânicas de As podem causar intoxicação aguda ou crónica, encontrando-se associados a vários

tipos de cancro, particularmente ao cancro de pele (por ingestão) (Herce-Pagliai et al., 2002;

Larsen et al., 2002; Mandal e Suzuki, 2002). A forma química mais tóxica é As(III), seguindo-se

As(V), mono-metil-arseniato, e finalmente, o di-metil-arseniato. O As está presente em quase todos

os sulfuretos metálicos naturais, em particular com o Cu, Sn e Ni. O As presente nos vinhos pode

ter origem no solo, ou em herbicidas e fungicidas utilizados na vinha (principalmente como

arseniato de sódio) (Jaganathan, 2001; Herce-Pagliai et al., 2002).

O teor no vinho é normalmente inferior a 0,02 mg/L (Martin de la Hinojosa et al., 1994;

Sudraud et al., 1994; Greenough et al., 1997; Jaganathan, 2001; Herce-Pagliai et al., 2002;

Barbaste et al., 2003; Taylor et al., 2003; Jos et al., 2004; Thiel et al., 2004). No entanto, vinhos

com origem em vinhas tratadas com sais arseniosos podem apresentar teores mais elevados

(Eschnauer, 1982b; Ribéreau-Gayon et al., 1998). Normalmente as formas orgânicas de As são

predominantes nos vinhos, em resultado da sua biometilação pelas leveduras durante a fermentação

Page 25: Tese doutoramento Sofia Catarino.pdf

Introdução geral

14

alcoólica. Nos mostos, as formas inorgânicas são mais significativas do que nos vinhos, tendo sido

registados nestas formas cerca de 25% do As total (Herce-Pagliai et al., 2002). No âmbito da OIV o

limite para a sua concentração no vinho é de 0,2 mg/L (OIV, 2005). Quando em concentrações

superiores a 1 mg/L constitui perigo para a saúde humana.

O teor de As total em vinhos pode ser determinado por AAS com gerador de hidretos

(Segura et al., 1999; Tasev et al., 2005), espectrometria de fluorescência atómica com gerador de

hidretos (Segura et al., 1999; Karadjova et al., 2005), ETAAS (Bruno et al., 1994; Kildahl e Lund,

1996; Jaganathan, 2001) e ICP-MS (Stroh et al., 1994; Baxter et al., 1997; Pérez-Jordán et al.,

1998; Rodushkin et al., 1999; Wangkarn e Pergantis, 1999; Castiñeira et al., 2001; Almeida et al.,

2002; Barbaste et al., 2003). A especiação do As tem sido realizada com recurso à cromatografia

de troca iónica, cromatografia líquida de alta resolução-AAS (HPLC-AAS) com gerador de

hidretos, HPLC/ICP-MS e HPLC/ICP-OES (Jaganathan, 2001; Herce-Pagliai et al., 2002).

Selénio

Os efeitos do Se enquanto elemento essencial são bem conhecidos, subsistindo alguma

polémica sobre os contornos da sua potencial toxicidade (Boulyga et al., 2000; Hurrel et al., 2001;

Larsen et al., 2002).

O teor deste elemento no vinho é normalmente inferior a 2 µg/L (Eschnauer, 1982a;

Eschnauer et al., 1989a; Bellanger et al., 1992, Frias et al., 2003b), havendo no entanto registos de

concentrações até 26 µg/L (Greenough et al., 1997; Jaganathan e Dugar, 1998). À semelhança de

grande parte dos metais pesados o Se é eliminado em grande quantidade durante a fermentação

alcoólica. Como fontes de contaminação são referidos pesticidas e a bentonite.

No que respeita ao seu doseamento, encontra-se descrito um método por ETAAS

(Jaganathan e Dugar, 1998), apresentando contudo um limite de detecção (9 µg/L) superior ao teor

de Se apresentado pela generalidade dos vinhos, um método por espectrometria de fluorescência

atómica (Bellanger et al., 1992), encontrando-se na bibliografia numerosas referências à técnica de

ICP-MS (Stroh et al., 1994; Baxter et al., 1997; Pérez-Jordán et al., 1998; Boulyga et al., 2000;

Almeida e Vasconcelos, 2002).

Molibdénio

Este elemento essencial (Hurrel et al., 2001) surge geralmente nos vinhos em teores até 15

µg/L, havendo contudo observações até 64 µg/L. Na escassa bibliografia existente a bentonite é

referida como potencial fonte de contaminação (Gómez et al., 2004b). A determinação analítica do

Page 26: Tese doutoramento Sofia Catarino.pdf

Introdução geral

15

Mo pode ser realizada por ICP-MS (Stroh et al., 1994; Baxter et al., 1997; Pérez-Jordán et al.,

1998; Rodushkin et al., 1999; Almeida e Vasconcelos, 2002; Thiel et al., 2004).

Cádmio

O Cd é um elemento mineral extremamente tóxico, bioacumulativo e cancerígeno,

susceptível de ser introduzido no organismo humano por via respiratória, oral e cutânea (Cabanis et

al., 1996; Larsen et al., 2002). Paralelamente ao Pb, é um dos metais contaminantes dos vinhos

mais investigados.

Apresenta-se naturalmente sob a forma de sulfureto e associado a minérios de Zn, sendo

absorvido pelas raízes da videira. A sua origem, à semelhança de outros metais pesados, encontra-

se fortemente relacionada com a sua aplicação industrial. A sua presença exógena está sobretudo

associada à poluição atmosférica, a produtos fitossanitários, ao contacto com materiais com Zn, e a

aços inoxidáveis (Curvelo-Garcia, 1988; Sudraud et al., 1994; Catarino, 2000). No decorrer da

fermentação alcoólica ocorre uma acentuada eliminação de Cd (Teissèdre et al., 1996; Catarino,

2000).

Os dados divulgados nos últimos 15 anos revelam teores quase sempre inferiores a 5 µg/L

em vinhos (Chen et al., 1994; Martin de la Hinojosa et al., 1994; Sudraud et al., 1994; Simões et

al., 1995; Teissèdre et al., 1996; Greenough et al., 1997; Catarino, 2000; Barbaste et al., 2003;

Taylor et al., 2003; Lima et al., 2004; Thiel et al., 2004), evidenciando uma tendência para a sua

diminuição, atribuída à melhoria das condições de elaboração (Eschnauer, 1982a; Ribéreau-Gayon

et al., 1982). O teor médio por nós encontrado em 95 vinhos portugueses foi de 0,28 µg/L, e o teor

máximo de 1,24 µg/L (Catarino e Curvelo-Garcia, 1999; Catarino, 2000). Com base nesses

resultados, foi apresentada uma proposta, junto da Sub-Comissão de Métodos de Análise da OIV,

com vista à diminuição do limite máximo em vigor, 10 µg/L (OIV, 2005), o qual nos parece

exagerado face à situação actual. A sua diminuição, justificada por razões de ordem toxicológica,

não implicaria a alteração de práticas tecnológicas por parte do sector produtivo.

A determinação analítica deste elemento pode ser realizada por ETAAS (Simões et al.,

1995; Catarino, 2000; Kristl et al., 2002; Canuto et al., 2004), técnica à qual recorre o método

indicado pela OIV (OIV, 2005) e pela União Europeia (CEE, 1990). A técnica de ICP-MS tem sido

também utilizada por vários autores (Stroh et al., 1994; Baxter et al., 1997; Pérez-Jordán et al.,

1998; Rodushkin et al., 1999; Castiñeira et al., 2001; Almeida e Vasconcelos, 2002; Barbaste et al.,

2003; Thiel et al., 2004). Em alternativa, o doseamento pode realizar-se com recurso à

potenciometria de redissolução (Chen et al., 1994) e à voltametria de redissolução (Brainina et al.,

2004).

Page 27: Tese doutoramento Sofia Catarino.pdf

Introdução geral

16

Estanho

É bastante reduzida a informação sobre a presença de Sn em vinhos e ainda mais sobre a

sua origem, essencialmente exógena resultante do contacto com materiais. A bibliografia refere

teores até 700 µg/L, embora normalmente sejam bastante mais baixos, na ordem dos 10 µg/L

(Martin de la Hinojosa et al., 1994; Greenough et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et

al., 1999; Nicolini et al., 2004).

As referências à sua determinação analítica indicam essencialmente o recurso a técnicas

espectrométricas: ETAAS (Martin de la Hinojosa et al., 1994), ICP-OES (Nicolini et al., 2004) e

ICP-MS (Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Castiñeira et al.,

2001; Almeida e Vasconcelos, 2002).

Antimónio

O Sb e os seus compostos apresentam uma toxicidade similar à dos compostos de arsénio,

surgindo na natureza sob diferentes formas químicas de toxicidade variável, sendo o Sb(III)

bastante mais tóxico do que o Sb(V) (Feng et al., 2000).

Os vinhos apresentam normalmente teores na ordem de 1 µg/L, contudo há registo de

observações até 122 µg/L (Eschnauer, 1982a; Kildahl e Lund, 1996; Greenough et al., 1997;

Rodushkin et al., 1999; Taylor et al., 2003; Thiel et al., 2004). De acordo com a escassa

bibliografia existente, a presença exógena deste elemento no vinho está relacionada com a

utilização de tubagens de borracha (em mau estado de conservação), materiais com zinco e

cápsulas de chumbo (Eschnauer, 1982a).

As técnicas de ICP-MS (Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al.,

1999; Almeida e Vasconcelos, 2002; Taylor et al., 2003; Thiel et al., 2004) e de ICP-OES com

gerador de hidretos são apropriadas ao seu doseamento nos vinhos. O seu doseamento por ETAAS

encontra-se também descrito (Kildahl e Lund, 1996). A determinação das espécies Sb(III) e Sb(V)

pode ser realizada por ICP-MS de alta resolução com gerador de hidretos (Feng et al., 2000).

Mercúrio

Os vapores deste metal bioacumulativo são extremamente tóxicos. O Hg armazenado em

sedimentos volta às cadeias alimentares por actividade microbiana que transforma o Hg metálico

em Hg orgânico, sob a forma de metilmercúrio, extremamente tóxico e volátil (Ribéreau-Gayon et

al., 1998). Os efeitos adversos do Hg inorgânico verificam-se ao nível dos rins enquanto as formas

orgânicas afectam o sistema nervoso central (Larsen et al., 2002).

Page 28: Tese doutoramento Sofia Catarino.pdf

Introdução geral

17

Apresenta-se nos vinhos normalmente em teores inferiores a 5 µg/L (Eschnauer, 1982a;

Martin de la Hinojosa et al., 1994; Sudraud et al., 1994; Ribereau-Gayon et al., 1998; Frias et al.,

2003b), verificando-se um acentuado decréscimo no decorrer da fermentação alcoólica.

No âmbito da OIV (OIV, 2005, Resolução OIV Oeno 15/2002) o doseamento de Hg em

vinhos é realizado por espectrometria de fluorescência atómica com gerador de hidretos. A técnica

de AAS com gerador de hidretos (Frias et al., 2003b), após mineralização da amostra, e a de ICP-

MS (Baxter et al., 1997; Pérez-Jordán et al., 1998; Almeida e Vasconcelos, 2002) são também

apropriadas.

Tálio

O Tl é um elemento tóxico sem acção biológica benéfica conhecida (Cvetkovic et al.,

2002). Os teores em vinhos não contaminados são normalmente inferiores a 1 µg/L, embora

contaminações entre 2 e 8 µg/L possam ter origem nos solos, adubos, ou nas emissões atmosféricas

de cimenteiras (Eschnauer, 1982a ; Eschnauer et al., 1984; Greenough et al., 1997; Cvetkovic et

al., 2002; Taylor et al., 2003; Nicolini et al., 2004; Thiel et al., 2004).

O doseamento do Tl nos vinhos apresenta alguma dificuldade devido aos níveis de

concentração baixos em que normalmente se apresenta, inferiores aos limites analíticos da AAS

com com gerador de hidretos e ETAAS, implicando a preconcentração da amostra (Cvetkovic et

al., 2002). É possível encontrar na bibliografia várias referências ao seu doseamento directo no

vinho por ICP-MS (Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999; Almeida

e Vasconcelos, 2002; Taylor et al., 2003; Thiel et al., 2004).

Chumbo

O Pb tem sido objecto de numerosos estudos e de especial vigilância, facto para o qual

contribui certamente a sua elevada toxicidade e níveis de ocorrência que permitem a quantificação

na generalidade dos vinhos. Este metal contaminante tem sido, inclusivamente, utilizado como

elemento modelo de outros metais, com algumas semelhanças químicas, mas que apresentam maior

dificuldade de determinação analítica. A problemática da presença de Pb nos vinhos foi aliás

objecto de um volume publicado pela OIV em 1995, reunindo artigos de especialistas em

segurança alimentar da Subcomissão de Nutrição e Saúde, em colaboração com a Subcomissão de

Métodos de Análise (OIV, 1995b).

A absorção do Pb ocorre essencialmente por ingestão, sendo o contributo da via

respiratória e da via cutânea pouco significativo. Apenas uma pequena parte do Pb ingerido é de

facto absorvido na barreira intestinal e transferido através do sangue para os tecidos, sendo

Page 29: Tese doutoramento Sofia Catarino.pdf

Introdução geral

18

susceptível de se acumular sob a forma mineral, principalmente nos tecidos duros (esqueleto,

dentes, cabelo). Os efeitos patológicos do Pb fazem-se sentir ao nível do sistema sanguíneo, por

inibição da síntese de hemoglobina, ao nível do sistema nervoso provocando encefalopatia crónica,

problemas neurológicos e psicomotores, ao nível do sistema renal (nefropatia e alteração

progressiva da função renal) e ao nível do sistema cardio-vascular (Cármen de la Torre, 1997).

A utilização do Pb em diversos sectores na indústria, e particularmente a sua incorporação

nos combustíveis, foi responsável pela introdução de grandes quantidades deste metal no ambiente.

O Pb encontra-se naturalmente nos solos maioritariamente sob a forma de sulfuretos, carbonatos e

sulfatos, apresentando quatro isótopos naturais: 204Pb, 206Pb, 207Pb e 208Pb.

Numerosos trabalhos têm sido desenvolvidos com vista à identificação de fontes de

contaminação (Sudraud et al., 1989; Terwell, 1990; Harding, 1991; Gulson et al., 1992; Cantagrel

et al., 1992; Henick-Kling e Stoewsand, 1993; Soares et al., 1993; Stockley e Lee, 1995; Minguez

et al., 1997; Kaufmann, 1998; Catarino, 2000; Almeida, 2002; Salvo et al., 2003; Stockley et al.,

2003). O estudo da composição isotópica do Pb no vinho permitiu comprovar a poluição

atmosférica como uma das principais origens deste elemento no vinho (Augagneur et al., 1997a;

Rosman et al., 1998; Médina et al., 2000). Também com base na composição isotópica, o

contributo das cápsulas de chumbo-estanho para o conteúdo do vinho em Pb foi considerado

insignificante (Gulson et al., 1990; Gulson et al., 1992), contrariando resultados de estudos

anteriores (Terwell, 1990; Harding 1991). Apesar disso, a utilização de cápsulas desta liga metálica

é interdita desde 1994 (Resolução OIV Oeno-Eco 1/90).

Foi comprovada a relação entre o teor de Pb das uvas com a sua proximidade a vias

rodoviárias (Teissèdre et al., 1993a). Foi objecto de discussão a eventual biometilação do Pb

durante a fermentação alcoólica e maloláctica: Teissèdre et al. (1994b) não encontraram evidências

da formação de compostos orgânicos de Pb, atribuindo a origem desses compostos aos

combustíveis. Estima-se que os combustíveis sejam responsáveis por cerca de 90 a 95% da

poluição atmosférica em Pb, por utilização de aditivos anti-detonantes organometálicos voláteis e

liposolúveis [Pb(CH3)4 e Pb(C2H5)4].

A localização e distribuição do Pb nos orgãos da videira, aspectos de especial relevância na

optimização de parâmetros tecnológicos químicos e físicos com influência na extracção e

solubilização do Pb, são hoje relativamente bem conhecidas (Teissèdre et al., 1993a; Teissèdre et

al., 1993b; Teissèdre et al., 1994a). Durante a fermentação alcoólica e nas etapas seguintes de

vinificação parte do Pb é removido, por precipitação sob a forma de sulfuretos, por adsorção e

absorção das leveduras e por precipitação com macromoléculas do vinho.

Relativamente aos níveis de ocorrência deste metal pesado nos vinhos, os teores entre 0,1 e

0,4 mg/L indicados por Ribéreau-Gayon et al. (1982) encontram-se francamente desactualizados,

Page 30: Tese doutoramento Sofia Catarino.pdf

Introdução geral

19

observando-se actualmente teores normalmente inferiores a 100 µg/L (Catarino, 2000; Almeida,

2002; Larsen et al., 2002; Stockley et al., 2003; Taylor et al., 2003; Gómez et al., 2004a; Jos et al.,

2004; Lima et al., 2004).

O decréscimo dos teores de Pb nos vinhos nas últimas décadas está relacionado com a

eliminação e controlo de fontes de contaminação importantes, nomeadamente pela substituição de

materiais e equipamentos utilizados na vinificação e conservação dos vinhos, e pela redução

significativa da utilização de combustíveis com chumbo. Este decréscimo, indicativo da

inexistência de limitações tecnológicas por parte do sector produtivo, encontra-se relacionado com

a redução progressiva do limite máximo do teor de Pb no vinho, no âmbito da OIV: 0,6 mg/L em

1953; 0,5 mg/L em 1975; 0,3 mg/L em 1987; 0,25 mg/L em 1993; 0,20 mg/L em 1996 e, muito

recentemente, 0,15 mg/L, limite aplicado a vinhos obtidos a partir da vindima de 2007 (Resolução

OIV Oeno 13/2006).

Estudos pioneiros, de enorme relevância, iniciados por Pellerin et al. (1997) abriram novas

perspectivas e relançaram a questão de ordem toxicológica da presença de Pb no vinho. De acordo

com estes autores, o Pb encontra-se maioritariamente (cerca de 85%) sob a forma de complexos

estáveis com ramnogalacturona II (RG-II), um polissacárido péctico das paredes celulares da uva.

A RG-II encontra-se nos vinhos sob a forma de dímeros (d RG-II-B) ligados por um diéster de

ácido bórico (Ishii e Matsunaga, 1996; O’Neill et al., 1996; Vidal et al., 1999). Para além do Pb, os

dímeros de RG-II podem formar complexos de coordenação com determinados catiões divalentes

ou trivalentes. Com base nos teores médios de RG-II nos vinhos, maior nos tintos do que nos

brancos, a sua capacidade de complexação supera os teores normalmente encontrados nos vinhos, o

que indica que a capacidade total de complexação pela RG-II está longe da saturação. Os

polissacáridos pécticos pertencem ao grupo das fibras alimentares, não sendo degradados no

estômago ou no intestino. Por outro lado, o RG-II ingerido com o vinho possui a capacidade de

fixar outros metais tóxicos veiculados pela restante alimentação, contribuindo para redução da sua

absorção intestinal (Tahiri et al., 2000). Diversos autores referem a ligação do Pb a

macromoléculas do vinho, capazes de complexar o Pb adicionado (Arcos et al., 1993; Lobinski et

al., 1993; Scollary, 1997; Fournier et al., 1998a; Pellerin e O’Neill, 1998; Green e Scollary, 2000),

com evidentes reflexos na sua assimilação (Azenha e Vasconcelos, 2000).

No âmbito da OIV e da União Europeia, a determinação analítica do Pb total em vinhos é

realizada por ETAAS (CEE, 1990; OIV, 2005; Resolução OIV Oeno 3/94) encontrando-se na

bibliografia várias referências (Almeida et al., 1992; Matthews e Parsons, 1993; Bruno et al., 1994;

Jorhem e Sundström, 1995; Catarino, 2000; Kristl et al., 2002; Canuto et al., 2004). O doseamento

é também possível por FAAS, após pré-concentração da amostra (Lemos et al., 2002; de Peña et

al., 2004). A técnica de ICP-MS tem sido vastamente utilizada (Goossens et al., 1993; Goossens et

al., 1994; Stroh et al., 1994; Baxter et al., 1997; Pérez-Jordán et al., 1998; Rodushkin et al., 1999;

Page 31: Tese doutoramento Sofia Catarino.pdf

Introdução geral

20

Castiñeira et al., 2001; Almeida e Vasconcelos, 2002; Almeida et al., 2002; Barbaste et al., 2003).

As técnicas electroquímicas de voltametria de redissolução (Mikkelsen e Schroder, 2002) e de

potenciometria de redissolução (Chen et al., 1994) permitem o doseamento do Pb total em vinhos,

podendo a potenciometria e a polarografia ser aplicadas para estudar a sua distribuição sob

diferentes formas químicas (Fournier et al., 1998a; Green e Scollary, 2000). Outras técnicas

utilizadas na especiação do Pb têm sido a cromatografia em fase gasosa / ICP-OES e HPLC de

exclusão /ICP-MS (Szpunar et al., 1998; Szpunar et al., 1999). Por último, a composição isotópica

do Pb em vinhos tem sido determinada por ICP-MS (Dean et al., 1990; Gulson et al., 1992;

Augagneur et al., 1997b; Almeida e Vasconcelos, 1999; Médina et al., 2000; Larcher et al., 2003).

No Quadro 1 apresenta-se informação sistematizada sobre os níveis de ocorrência, fontes

de contaminação identificadas e limites máximos admissíveis estabelecidos pela OIV para os

metais contaminantes do vinho anteriormente caracterizados e outros de menor prioridade de

estudo, por ordem crescente de número atómico.

Pelo que anteriormente se referiu conclui-se sobre o desajustamento dos limites legais em

vigor, definidos em termos de teor total do elemento, independentemente da(s) forma(s) química(s)

em que este se encontra no vinho, sendo importante distinguir entre as formas químicas

assimiláveis pelo organismo, eventualmente susceptíveis de serem acumuladas, e as formas

químicas não assimiláveis, sendo por isso rejeitadas. Avanços nesta área do conhecimento serão,

certamente, de fulcral importância para um efectivo controlo dos metais contaminantes com

implicações toxicológicas nos vinhos.

Apesar de actualmente apenas alguns metais contaminantes do vinho serem alvo de

controlo legal, a preocupação crescente com sua segurança alimentar provavelmente resultará no

alargamento da lista de elementos tóxicos a controlar. Futuros desenvolvimentos na toxicologia,

tecnologias analíticas, assim como no comércio internacional vão certamente implicar alterações na

regulamentação actual. Por outro lado, com a tendência para o estabelecimento de regulamentação

internacional, surge a necessidade de serem desenvolvidos e validados métodos de análise

reconhecidos internacionalmente.

Page 32: Tese doutoramento Sofia Catarino.pdf

Introdução geral

21

Quadro 1 – Níveis de ocorrência, limites máximos admissíveis (OIV) e fontes de contaminação de metais contaminantes do vinho

Concentrações expressas em mg/L para Al, Mn, Fe, Cu, Zn, e em µg/L para os restantes elementos

Metal

contaminante

Níveis de

ocorrência

Fontes de contaminação

(identificadas)

Nº da referência

bibliográfica

Berílio (Be) < 0,01 – 13,1 Bentonite 8,54,69,76,125,126,144, 166,175

Alumínio (Al) 0,18 - 8,6

Produtos fitossanitários, contacto com superfíces de alumínio, bentonite, cartuchos de filtração, tanino enológico

8,48,50,69,76,77,90,93,98, 107,117,125,126,134,140, 144,149, 151,166

Escândio (Sc) 0,091 - 64,8 Bentonite

8,69,76,77,134,144

Titânio (Ti) 7,1 – 300 Bentonite

8,69,76,77,134,166,175

Vanádio (V) 1 - 447,0

Contaminação atmosférica, aço inoxidável, conservação em garrafa de vidro (óxidos metálicos utilizados na sua pigmentação), bentonite

8,68,69,76,126,134,144, 158,166,172

Crómio (Cr) 3,94 – 139 Aço inoxidável, conservação em garrafa

8,27,55,76,95,103,119,121,126,131,140,144

Manganês (Mn) 0,150 - 7,836 Produtos fitossanitários, bentonites

8,28,43,44,50,65,68,69,70, 76,77,83,90,100,114,121, 126,134,143,144,155,158, 160,166

Ferro (Fe) 0,24 - 19,40 Partículas de terra, contacto com superfícies de aço corroídas, bentonite

8,44,50,59,65,68,69,70,71, 76,83,90,100,113,121,126, 134,142,143,144,149,155, 160,166

Cobalto (Co) 1,0 - 20,0 Bentonite, materiais de ligas metálicas com cobalto

8,68,69,76,126,134,144, 154,166,175

Níquel (Ni) 1 – 510

Poluição atmosférica, produtos fitossanitários, bentonite, aço inoxidável, conservação em garrafa de vidro (pigmentos com níquel)

8,69,76,77,90,100,119,126,131,134,140,144,158,166, 173,175

Cobre (Cu) 0,02 - 3,0 1

Fungicidas cúpricos (calda bordalesa), utensílios em bronze e latão, sulfato de cobre (aplicado no vinho para eliminar odores indesejáveis)

5,8,44,50,59,63,64,65,67, 69,70,71,76,83,90,100,113,121,124,125,126,134,136, 143,144,148,149,155,158, 160,166

Zinco (Zn) 0,11 - 9,40 5

Produtos fitossanitários, materiais em ferro galvanizado e latão

8,40,44,50,63,64,65,68,69, 71,76,77,90,100,113,121, 126,134,140,141,143,144, 148,154,155,160

Gálio (Ga) 0,76 – 23,50 Bentonite

8,69,114,126,134,175

Arsénio (As) 0,4 - 597,7 200

Produtos fitossanitários (com sais arseniosos)

16,25,55,69,76,82,86,94, 113,114,131,134,140,158, 160,166,175,179

Selénio (Se) 0,29 – 26 Produtos fitossanitários, bentonite

18,54,57,66,76,87,158,166

Page 33: Tese doutoramento Sofia Catarino.pdf

Introdução geral

22

Ítrio (Y) 0,07 - 6,89 Bentonite

8,69,126,134,144,175

Zircónio (Zr) 0,3 – 85 Bentonite

8,69,134,144,175

Nióbio (Nb) < 0,08 – 5,0 Bentonite

8,69,134,144,175

Molibdénio (Mo) 0,7 – 64 Bentonite 8,69,76,134,144,158,166, 175

Prata (Ag) 0,013 - 270 Cloreto de prata (aplicado no vinho para eliminar odores indesejáveis)

140,144,154,166

Cádmio (Cd) 0,1 - 8,1 10

Poluição atmosférica, contacto com materiais de zinco, bentonite

8,16,34,35,40,54,55,76,90, 95,104,113,121,134,144, 158,160,166,171,175

Estanho (Sn) 0,99 - 700 Contacto com materiais metálicos

76,113,126,134,144

Antimónio (Sb) 0,39 - 122 Tubagens de borracha com antimónio, materiais com zinco, cápsulas de chumbo

54,76,94,126,134,144,166, 175

Césio (Cs) 0,04 – 24 8,54,69,76,126,134,144, 158,166,175

Háfnio (Hf) 0,04 - 5,5

54,134,144

Tantálio (Ta) 0,01 - 3,8

54,144

Tungsténio (W) 0,09 - 10,5 Aço inoxidável

8,54,134,144,175

Platina (Pt) <0,0002

144

Ouro (Au) 0,0017 - 0,76

54,144

Mercúrio (Hg) 0,01 – 6 Contaminação ambiental 54,55,66,113,134,140,141, 149,160

Tálio (Tl) 0,032 - 4,2

Poluição atmosférica (proximidade a minas abandonadas e fábricas de cimento), bentonite

49,54,56,76,126,134,144, 166,175

Chumbo (Pb) 1 – 1125 150

Poluição atmosférica (proximidade a vias rodoviárias, fundições e incineradoras), partículas de terra, arseniato de chumbo e determinados fungicidas, contacto com materiais de ligas metálicas contendo chumbo, materiais em bronze e latão, bombas com corpo em bronze, revestimentos cerâmicos, tintas, presença de soldaduras e estanhagens; cápsulas de Pb-Sn, garrafas de cristal, bentonite, operação de aguardentação

8,16,25,30,34,40,55,63,64, 68,69,71,76,80,81,89,90,92,99,104,106,113,115,121, 123,126,128,131,134,140, 144,148,156,157,158,159, 160,166,167,174

Bismuto (Bi) 0,04 - 2,8

76,144,158,166

Page 34: Tese doutoramento Sofia Catarino.pdf

Introdução geral

23

1.2.2. Determinação analítica

Do que foi exposto anteriormente verifica-se que diversas técnicas analíticas têm sido

utilizadas no doseamento dos metais contaminantes dos vinhos, com especial destaque para as

técnicas espectrométricas. Os métodos oficiais para a determinação de metais pesados em vinhos

indicados pela OIV, e adoptados pela União Europeia, baseiam-se essencialmente na AAS, com

chama (FAAS) e com atomização electrotérmica (ETAAS). A técnica de FAAS é uma das técnicas

mais vulgarizadas nos laboratórios de Enologia, apresentando contudo limites analíticos

relativamente altos para grande parte dos metais contaminantes, entre os quais o As, Cr, Hg ou Pb,

excepto para amostras muito concentradas ou sujeitas a pré-concentração. É método de referência

para doseamento de Fe, e método único para o doseamento de Cu, Zn e Ag (OIV, 2005), embora

evidencie manifestas dificuldades no doseamento do Cu na generalidade dos vinhos. A AAS com

gerador de hidretos é método de referência para o doseamento do As, podendo ser igualmente

utilizada no doseamento de outros elementos, por exemplo o Se, Sn, Sb, Pb e Bi. A ETAAS é

vastamente utilizada na determinação de elementos vestigiais, sendo método oficial único para a

determinação do Cd e do Pb.

A aplicação da técnica multi-elementar de ICP-OES é referida na literatura (Eschnauer et

al., 1989b; Fournier et al., 1998b; Szentmihalyi et al., 2000; Aceto et al., 2002; Lara et al., 2005),

apresentando contudo dificuldades no doseamento de alguns metais contaminantes devido aos seus

limites analíticos (entre 1 e 10 µg/L, para a generalidade dos elementos). Trata-se de uma técnica

analítica pouco vulgar nos laboratórios de Enologia.

A técnica de ICP-MS apresenta como grandes vantagens a análise multielementar e limites

analíticos bastante baixos, sendo aplicada cada vez com mais frequência no estudo da composição

mineral dos vinhos, embora seja ainda rara nos laboratórios de Enologia devido aos custos

envolvidos (Pyrzynska, 2004).

Embora de utilização menos vulgarizada, diversas técnicas electroquímicas são referidas na

bibliografia: voltametria de redissolução e potenciometria de redissolução, potenciometria com

eléctrodo selectivo e polarografia (Fournier et al., 1998a; Clark e Scollary, 2000). Para além do

teor total do elemento, algumas destas técnicas fornecem ainda informação sob a sua distribuição

por diferentes formas químicas (Chen et al., 1994; Scollary et al., 1997; Fournier et al., 1998; Clark

e Scollary, 2000; Green e Scollary, 2000; Brainina et al., 2004).

1.3. INFLUÊNCIA DAS BENTONITES NA COMPOSIÇÃO MINERAL DO VINHO

A bentonite é um auxiliar tecnológico com vasta aplicação na clarificação e estabilização

proteica de mostos e vinhos brancos, sendo referida como uma das suas principais fontes de

Page 35: Tese doutoramento Sofia Catarino.pdf

Introdução geral

24

contaminação mineral, como aliás se verifica pelo que anteriormente se referiu (Postel, 1986;

Enkelmann, 1988; Wurzinger et al., 1994; Gössinger et al., 1997; Machado-Nunes et al., 1998;

Jakubowski et al., 1999; Catarino et al., 2004; Gómez et al., 2004b; Nicolini et al., 2004).

Bentonite é a designação comercial de um material argiloso (silicatos de alumínio

hidratados) composto principalmente por montmorilonite, uma esmectite dioctaédrica 2:1,

representada pela fórmula geral (Maujean, 1993; Resolução OIV Oeno 11/2003): Si4 (Al(2-x)Rx)

(O10, H2O) (Cex, nH2O) ou Si4(Al(2-x)Rx) (H2O)n, onde: R = Mg, Fe, Mn, Zn, Ni e Ce (catiões de

troca) = Ca2+, Na+, Mg+. Em adição à montmorilonite pode ainda conter quartzo, feldspatos, calcite,

pirite, caulino, entre outros, como materiais geológicos acessórios. A estrutura da montmorilonite

consiste em duas camadas tetraédricas de sílica e uma camada octaédrica de alumina, combinadas

numa unidade estrutural cristalina (Figura 1). Parte do alumínio das posições octaédricas encontra-

se substituído por Mg2+, Fe2+, Fe3+ ou por outros catiões bivalentes. As cargas negativas resultantes

destas substituições são neutralizadas por catiões que se alojam nos espaços interfoliares e na

superfície externa das partículas argilosas (Maujean, 1993). Esses catiões são de natureza diversa:

Ca2+, Na+, Mg2+ e, com menor importância, Cu+, Fe2+, K+ (Marchal et al., 1995), variando

significativamente entre bentonites.

Figura 1 – Estrutura dioactaédrica da montmorilonite.

nH2O

catiões de troca

oxigénios

Al, Fe, Mg

Si, Fe

Page 36: Tese doutoramento Sofia Catarino.pdf

Introdução geral

25

Actua como um permutador de iões, embora de capacidade de troca relativamente fraca.

Assim, quando é adicionada a um mosto ou vinho, assiste-se a uma troca entre os seus catiões

compensadores e as proteínas do vinho (Marchal et al., 2002), tendo ainda a capacidade de fixar

matéria corante e, de modo indirecto, diminuir o risco de ocorrência de casse cúprica (Peynaud,

1996). A adsorção das proteínas pode ser afectada pela competição de catiões em solução,

nomeadamente K+, Ca2+, Na+ e Mg2+, bem como de outras espécies químicas com carga positiva

(Blade e Boulton, 1988). O pH da matriz, ao alterar a conformação espacial e a carga eléctrica das

proteínas, interfere no processo de adsorção, podendo igualmente influenciar a capacidade de troca

da bentonite (Blade e Boulton, 1988).

A investigação sobre a utilização de bentonites no vinho tem incidido essencialmente sobre

os seus efeitos na eliminação de proteínas, amino-ácidos, polifenóis e compostos do aroma. Na

literatura são também referidos alguns estudos sobre o efeito na composição mineral,

especialmente sobre elementos maioritários, tendo sido verificados aumentos nos teores de

numerosos elementos (Postel et al., 1986; Enkelmann, 1988; Wurzinger et al., 1994; Gössinger et

al., 1997; Machado-Nunes et al., 1998; Gómez et al., 2004b; Nicolini et al., 2004) e decréscimos

nos teores de K, Cu, Zn e Rb (Machado-Nunes et al., 1998; Nicolini et al., 2004).

A diversidade de bentonites no mercado, no que respeita à sua origem, tipo e rotulagem é

considerável, não sendo normalmente fornecida informação sobre a potencial cedência mineral ao

vinho. Com base na revisão bibliográfica realizada, verifica-se que a caracterização físico-química

das bentonites não foi anteriormente estudada na perspectiva da cedência de elementos minerais ao

vinho, concretamente de metais contaminantes. Actualmente, a OIV estabelece um teor mínimo de

montmorilonite e concentrações máximas para alguns elementos, a determinar analiticamente numa

solução de extracção composta por ácido tartárico (Resolução OIV Oeno 11/2003). No entanto,

considerando a complexidade química natural da matriz do vinho, e sendo conhecidos os

fenómenos de competição entre diferentes espécies químicas na adsorção pelas bentonites, é de

admitir resultados distintos quando a extracção é conduzida com a solução teste ou com vinho.

1.4. OBJECTIVOS

Nas últimas décadas a grande transformação verificada pelas adegas ao nível da qualidade

dos equipamentos e materiais utilizados no transporte, vinificação, estabilização e conservação dos

vinhos, aliada à criação de conhecimento sobre as potencias fontes de contaminação, permitiu

reduzir os níveis de ocorrência de grande parte dos metais contaminantes. Apesar disso, o

conhecimento e controlo de produtos enológicos de utilização generalizada, nos quais se incluem as

bentonites, exaustivamente apontadas como veículo de contaminação mineral, é bastante

Page 37: Tese doutoramento Sofia Catarino.pdf

Introdução geral

26

deficitário. Os aspectos relacionados com a cedência de metais contaminantes em ligação com as

características físico-químicas das bentonites, e o seu controlo da qualidade, necessitam de ser

clarificados de modo a prevenir a ocorrência de contaminações.

Assim, na génese do presente trabalho encontra-se o objectivo geral de contribuir para o

aumento do conhecimento sobre a ocorrência de metais contaminantes nos vinhos por influência

das bentonites. Para tal pretende-se:

- Estudar a evolução do teor em metais contaminantes ao longo do processo tecnológico do

vinho;

- Aprofundar o conhecimento sobre a cedência de metais contaminantes de bentonites para

o meio, bem como dos factores com influência nesses fenómenos (nomeadamente o pH do

vinho);

- Estudar a influência das características físico-químicas das bentonites na cedência de

elementos minerais ao vinho;

- Avaliar a capacidade da solução teste indicada pela OIV para fins de controlo da

qualidade das bentonites;

- Optimizar uma solução teste (modelo) de extracção, representativa dos vinhos, para

controlo da qualidade das bentonites.

De modo complementar, como instrumentos de suporte ao trabalho experimental

necessário para atingir os objectivos anteriormente referidos, pretende-se:

- Desenvolver e validar métodos de análise para doseamento dos metais contaminantes em

vinhos por ETAAS (programas térmicos convencionais e rápidos) e ICP-MS (método

quantitativo e semiquantitativo).

- Desenvolver e validar uma metodologia inovadora de preparação de amostras de mosto

por processos de oxidação avançada.

Page 38: Tese doutoramento Sofia Catarino.pdf

Introdução geral

27

1.5. REFERÊNCIAS BIBLIOGRÁFICAS

Aceto M., Abollino O., Bruzzoniti M.C., Mentasti E., Sarzanini C., Malandrino M., 2002. Determination of metals in wine with atomic spectroscopy (flame-AAS, GF-AAS and ICP-AES); a review. Food Additives and Contaminants, 2, 126-133. [1]

Aikoh H., Nishio M.R., 1996. Aluminum content of various canned and bottled beverages. Bulletin of Environment Contamination and Toxicology, 56, 1-7. [2]

Almeida C.M.R., 2002. Isotopic and multi-element characterisation of wine for identification of lead contamination sources and of the provenance region. Tese de Doutoramento, Faculdade de Ciências da Universidade do Porto, Porto. [3]

Almeida A.A., Bastos M.L., Cardoso M.I., Ferreira M., Lima J.L.F.C., Soares M.E., 1992. Determination of lead and aluminium in Port wine by electrothermal atomic absorption spectrometry. Journal of Analytical Atomic Spectrometry, 7, 1281-1285. [4]

Almeida A.A., Cardoso M.I., Lima J.L.F.C., 1994. Determination of copper in Port wine and Madeira by electrothermal atomization AAS, Atomic Spectroscopy, 2, 73-77. [5]

Almeida C.M.R., Vasconcelos M.T.S.D., 1999. Determination of lead isotope ratio in port wine by inductively coupled plasma mass spectrometry after pré-treatment by UV-irradiation. Analytica Chimica Acta, 396, 45-53. [6]

Almeida C.M., Vasconcelos M.T.S.D., 2001. ICP-MS determination of strontium isosotope ratio in wine in order to be used as a fingerprint of its regional origin. Journal of Analytical Atomic Spectrometry, 16, 607-611. [7]

Almeida C.M., Vasconcelos M.T.S.D., 2002. Advantages and limitations of the semi-quantitative operation mode of an inductively coupled plasma-mass spectrometer for multi-element analysis of wines. Analytica Chimica Acta, 463, 165-175. [8]

Almeida C.M.R., Vasconcelos M.T.S.D., Barbaste M., Médina B., 2002. ICP-MS multi-element analysis if wine samples – a comparative study of the methodologies used in two laboratories. Anaytical and Bioanalytycal Chemistry, 374, 314-322. [9]

Almeida C.M.R., Vasconcelos M.T.S.D., 2003a. Multielement composition of wines and their precursors including provenance soils and their potentialities as fingerprints of wine origin. Journal of Agricultural and Food Chemistry, 51, 4788-4798. [10]

Almeida C.M.R., Vasconcelos M.T.S.D., 2003b. Multi-element composition and 87Sr/86Sr of wines and their potentialities as fingerprints of wine provenance. Ciência e Técnica Vitivinícola, 18 (1), 15-27. [11]

Arcos M.T., Ancín M.C., Echeverría J.C., González A., Garrido J.J., 1993. Study of lability of aging through differential pulse anodic stripping voltammetry. Journal of Agricultural and Food Chemistry, 41, 2333-2339. [12]

Augagneur S., Médina B., Grousset F., 1997a. Mise en evidence de la pollution par le plomb d’origine atmospherique dans les vins. Feuillet Vert de l’OIV, 1053. [13]

Augagneur S., Médina B., Grousset F., 1997b. Measurement of lead isotope ratios in wine by ICP-MS and its applications to the determination of lead concentration by isotope dilution. Fresenius Journal of Analytical Chemistry, 357, 1149-1152. [14]

Azenha M.A.G.O., Vasconcelos M.T.S.D., 2000. Assessment of the Pb and Cu in vitro availability in wines by means of speciation procedures. Food and Chemical Toxicology, 38, 899-912. [15]

Barbaste M., Médina B., Perez-Trujillo J.P., 2003. Analysis of arsenic, lead and cadmium in wines from the Canary Islands, Spain, by ICP-MS. Food Additives and Contaminants, 20 (2), 141-148. [16]

Baxter M.J., Crews H.M., Dennis M.J., Goodall I., Anderson D., 1997. The determination of the authenticity of wine from its trace element composition. Food Chemistry, 60 (3), 443-450. [17]

Page 39: Tese doutoramento Sofia Catarino.pdf

Introdução geral

28

Bellanger J., Bouvier J.C., Lab C., 1992. Determination of selenium in wine. Application to several French wines. Journal International des Sciences de la Vigne et du Vin, 26, 287-293, 297-298. [18]

Berg T., Licht D., 2002. International legislation on trace elements as contaminants in food: a review. Food Additives and Contaminants, 19 (10), 916-927. [19]

Blackwell K.J., Singleton I., Tobin J.M., 1995. Metal cation uptake by yeast: a review. Applied Microbiology and Biotechnology, 43, 579-584. [20]

Blade W.H., Boulton R., 1988. Adsorption of protein by bentonite in a model wine solution. American Journal of Enology and Viticulture, 39 (3), 193-199. [21]

Boulyga S., Dombovári J., Becker J.S., Dietze H.J., 2000. Determination of selenium in biological samples using ICP-QMS. Atomic Spectroscopy, 21 (5), 149-164. [22]

Brainina Kh.Z., Stozhko N.Yu., Belysheva G.M., Inzhevatova O.V., Kolyadina L.I., Cremisini C., Galletti M., 2004. Determination of heavy metals in wines by anodic stripping voltammetry with thick-film modified electrode. Analytica Chimica Acta, 514, 227-234. [23]

Brandolini V., Tedeschi P., Capece A., Maietti A., Mazzota D., Salzano G., Paparella A., Romano P., 2002. Saccharomyces cerevisiae wine strains differing in copper resistence exhibit different capability to reduce copper content in wine. World Journal of Microbiology & Biotechnology, 18, 499-503. [24]

Bruno S., Campos R.C., Curtius A.J., 1994. Determination of lead and arsenic in wines by electrothermal atomic absorption spectrometry. Journal of Analytical Atomic Spectrometry, 9, 341-344. [25]

Cabanis M.T., Teissèdre P.L., Cabanis J.C., 1996. Toxicite du cadmium. Feuillet Bleu de l’OIV, 18. [26]

Cabrera-Vique C., Teissèdre P-L., Cabanis M-T., Cabanis J-C., 1997. Determination and levels of chromium in French wine and grapes by graphite furnace atomic absorption spectrometry. Journal of Agricultural and Food Chemistry, 45, 1808-1811. [27]

Cabrera-Vique C., Teissèdre P.L., Cabanis M.T., Cabanis J.C., 2000. Manganese determination in grapes and wines from different regions of France. American Journal of Enology and Viticulture, 51 (2), 103-107. [28]

Cacho J., Castells J.E., Esteban A., Laguna B., Sagristá N., 1995. Iron, copper and manganese influence on wine oxidation. American Journal of Enology and Viticulture, 46 (3), 380-384. [29]

Cantagrel R., Vidal J.P., Chetaneau B., 1992. Evolution de la teneur en plomb dans du vin ou du cognac logés en bouteille de verre ou en carafe cristal. Feuillet Vert de l’OIV, 938. [30]

Canuto M.H., Siebald H.G.L., Franco M.B., Silva J.B.B., 2004. Determination of Al, Cd, and Pb in Brazilian sugar cane spirit, cachaça, by ETAAS using matrix-matched calibration. Atomic Spectroscopy, 25 (3), 140-144. [31]

Carmen de la Torre M., 1997. Les contaminants du vin, aspects toxicologiques et de sécurité alimentaire. Analusis, 25 (3), M21-M26. [32]

Castiñeira M.M., Brandt R., von Bohlen A., Jakubowski N., 2001. Development of a procedure for the multi-element determination of trace elements in wine by ICP-MS. Fresenius Journal of Analytical Chemistry, 370, 553-558. [33]

Catarino S.C.G., 2000. A ocorrência de chumbo e cádmio em vinhos. Dissertação de Mestrado em Viticultura e Enologia. Faculdade de Ciências da Universidade do Porto, Porto. [34]

Catarino S., Curvelo-Garcia A.S., 1999. Les teneurs en plomb et en cadmium de quelques vins portugais. Feuillet Vert de l’OIV, 1081. [35]

Catarino S., Curvelo-Garcia A.S., Bruno de Sousa R., 2002. Determinação do zinco por espectrofotometria de absorção atómica com chama. Validação do método de análise. Ciência e Técnica Vitivinícola, 17 (1), 15-26. [36]

Catarino S., Pinto D., Curvelo-Garcia A.S., 2003. Validação e comparação de métodos de análise em espectrofotometria de absorção atómica com chama para doseamento de cobre e ferro em vinhos e aguardentes. Ciência e Técnica Vitivinícola, 18 (2), 65-76. [37]

Page 40: Tese doutoramento Sofia Catarino.pdf

Introdução geral

29

Catarino S., Soares J., Curvelo-Garcia A.S., Bruno de Sousa R., 2004. Implicações da utilização de bentonites sobre a fracção mineral de vinhos: potássio, sódio, cálcio, alumínio e chumbo. Efeito do pH. Ciência e Técnica Vitivinícola, 19 (1), 29-45. [38]

CEE, 1990. Regulamento (CEE) nº 2676/90 de 1990-09-17, Jornal Oficial das Comunidades Europeias, 3 de Outubro de 1990. [39]

Chen G.N., Scollary G.R., Vicente-Beckett V.A., 1994. Potentiometric stripping determination of lead, cadmium and zinc in wine. American Journal of Enology and Viticulture, 45 (3), 305-311. [40]

Clark A., Scollary G.R., 2000. Determination of total copper in white wine by stripping potentiometry utilising medium exchange. Analytica Chimica Acta, 413, 25-32. [41]

Clark A.C., Prenzler P.D., Scollary G.R., 2003. The role of copper(II) in the bridging reactions of (+)-catechin by glyoxylic acid in a model white wine. Journal of Agricultural and Food Chemistry, 51 (21), 6204-6210. [42]

Coetzee P.P., Steffens F., Eiselen R., Augustyn O.P., Balcaen L., Vanhaecke F., 2005. Multi-element analysis of South African wines by ICP-MS and their classification according to geographical origin. Journal of Agricultural and Food Chemistry, 53, 5060-5066. [43]

Costa R.C.C., Cardoso M.I., Araújo A.N., 2000. Metals determination in wines by sequential injection analysis with flame atomic absorption spectrometry. American Journal of Enology and Viticulture, 51 (2), 131-136. [44]

Curvelo-Garcia A.S., 1988. Controlo da Qualidade dos Vinhos. Química Enológica. Métodos analíticos. Instituto da Vinha e do Vinho, Lisboa. [45]

Curvelo-Garcia A.S., Ghira J.J., 1979. Uma causa da presença de ferro em vinhos brancos. A incorporação de terra no mosto. De Vinea et Vino Portugaliae Documenta, II, 9, 1-8. [46]

Curvelo-Garcia A.S., Ghira J.J., 1981. Considerações sobre a presença de zinco e chumbo em vinhos, policop., Estação Vitivinícola Nacional, Dois Portos. [47]

Curvelo-Garcia A.S., Catarino S., 1998. Os metais contaminantes dos vinhos: origens da sua presença, teores, influência dos factores tecnológicos e definição de limites (revisão bibliográfica crítica). Ciência e Técnica Vitivinícola, 13 (1-2), 49-70. [48]

Cvetkovic J., Arpadjan S., Karadjova I., Stafilov T., 2002. Determination of thallium in wine by electrothermal atomic absorption spectrometry after extraction preconcentration. Spectrochimica Acta Part B, 57, 1101-1106. [49]

Day M.P., Zhang B., Martin G.J., 1995. Determination of the geographical origin of wine using joint analysis of elemental and isotopic composition. II – Diferentiation of the principal production zones in France for the 1990 vintage. Journal of the Science of Food and Agriculture, 67, 113-123. [50]

Dean J.R., Ebdon L., Massey R.C., 1990. Isotope ratio and isotope dilution analysis of lead in wine by inductively coupled plasma-mass spectrometry. Food Additives and Contaminants, 7 (1), 109-116. [51]

de Peña Y.P., Paredes B., Rondón W., Burguera M., Burguera J.L., Rondón C., Carrero P., Capote T., 2004. Continuous flow system for lead determination by FAAS in spirituous beverages with solid phase extraction and on-line copper removal. Talanta, 64, 1351-1358. [52]

Enkelmann R., 1988. Migration of heavy metals from bentonites to wine. Bentonites. Deutsche Lebensmittel-Rundschau, 84, 243-247. [53]

Eschnauer H., 1982a. Éléments ultra-traces dans le vin. Bulletin de L’OIV, 617/618, 592-597. [54]

Eschnauer H., 1982b. Trace elements in must and wine: primary and secondary contents. American Journal of Enology and Viticulture, 33 (4), 226-230. [55]

Eschnauer H., Gemmer-Colos V., Neeb R., 1984. Thallium in wine – trace element vinogram of thallium. Z. Lebensm. Unters. Forsh, 178 (6), 453-460. [56]

Page 41: Tese doutoramento Sofia Catarino.pdf

Introdução geral

30

Eschnauer H.R., Alt F., Messerschmidt J., Toelg G., 1989a. A contribution to the enology and ecology of selenium. Trace element vinogram. Fresenius Zeitschrift fur analytische Chemie, 332, 874-879. [57]

Eschnauer H., Jakob L., Meierer H., Neeb R., 1989b. Use and limitations of ICP-OES in wine analysis. Microchimica Acta, 99, 291-298. [58]

Escobal A., Iriondo C., Laborra C., Ulibarrena E., 1995. Application of FAAS and GFAAS to the control of cloudiness in Txakoli wine. Atomic Spectroscopy, 4, 162-164. [59]

Esparza I., Salinas I., Caballero I., Santamaría C., Calvo I., García-Mina J.M., Fernández J.M., 2004. Evolution of metal and polyphenol content over a 1-year period of vinification: simple fractionation and correlation between metals and anthocyanins. Analytica Chimica Acta, 524, 215-224. [60]

Feng Y.-L., Narasaki H., Tian L.-C., Chen H.-Y., 2000. Speciation of Sb(III) and Sb(V) by hydride generation high-resolution ICP-MS combined with prereduction of Sb(V) with L-cysteine. Atomic Spectroscopy, 21 (1), 30-36. [61]

Flanzy C., 1998. Œnologie. Fondements scientifiques et technologiques. Lavoisier Tec.& Doc., Paris. [62]

Fournier J.-B., Hirsch O., Martin G.J., 1998a. Analyse du zinc, du cuivre et du plomb dans le vin. Rapport entre les concentrations totales analysées para spectrométrie d’absorption atomique et les concentrations des ions libres déterminées par polarographie. Journal International des Sciences de la Vigne et du Vin, 32 (1), 45-50. [63]

Fournier J.-B., Hirsch O., Martin G.J., 1998b. Intérêt de l’analyse élémentaire des produits viticoles: dosage de vingt-cinq éléments par spectrométrie d’émission atomique dans un plasma à couplage induit. Analusis, 26, 28-32. [64]

Frías S., Conde J.E., Rodríguez-Bencomo J.J., García-Montelongo F., Pérez-Trujillo J.P., 2003a. Classification of comercial wines from the Canary Islands (Spain) by chemometrics techniques using metallic contents. Talanta, 59, 335-344. [65]

Frías S., Diaz C., Conde J.E., Perez-Trujillo J.P., 2003b. Selenium and mercury concentrations in sweet and dry bottled wines from the Canary Islands, Spain. Food Additives and Contaminants, 20 (3), 237-240. [66]

García-Esparza M.A., Capri E., Pirzadeh P., Trevisan M., 2006. Copper content of grape and wine from Italian farms. Food Additives and Contaminants, 23 (3), 274-280. [67]

Gómez M.M.C., Feldmann I., Jakubowski N., Andersson J.T., 2004a. Classification of german white wines with certified brand of origin by multielement quantitation and pattern recognition techniques. Journal of Agricultural and Food Chemistry, 52, 2962-2974. [68]

Gómez M.M.C., Brandt R., Jakubowski N., Andersson J.T., 2004b. Changes of the metal composition in german white wines through the winemaking process. A study of 63 elements by inductively coupled plasma-mass spectrometry. Journal of Agricultural and Food Chemistry, 52, 2953-2961. [69]

Gonzales-Larraina M., Gonzales A., Médina B., 1987. Les ions métalliques dans la différentiation des vins rouges des trois régions d’appellation d’origine Rioja. Connaissance de la Vigne et du Vin, 21 (2), 127-140. [70]

González G., Peña-Méndez E.M., 2000. Multivariate data analysis in classification of must and wine from chemical measurements. European Food Research and Technology, 212, 100-107. [71]

Goossens J., De Smaele T., Moens L., Dams R., 1993. Accurate determination of lead in wines by inductively coupled plasma mass spectrometry. Fresenius Journal of Analytical Chemistry, 347, 119-125. [72]

Goosens J., Moens L., Dams R., 1994. Determination of lead by flow-injection inductively coupled plasma mass spectrometry comparing several calibration techniques. Analytica Chimica Acta, 293, 171-181. [73]

Gossinger M., Schodl H., Steidl R., Meier W., 1997. Comparison of commercial must and wine bentonites. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Fruechteverwertung, 47 (1-2), 1-7. [74]

Green A.M., Scollary G.R., 2000. Influence of metal ions on lead complexation in wine. Australian Journal of Grape and Wine Research, 6 (3), 197-202. [75]

Page 42: Tese doutoramento Sofia Catarino.pdf

Introdução geral

31

Greenough J.D., Longerich H.P., Jackson S.E., 1997. Element fingerprinting of Okanagan Valley wines using ICP-MS: Relationships between wine composition, vineyard and wine colour. Australian Journal of Grape and Wine Research, 3, 75-83. [76]

Gremaud G., Quaile S., Piantini U., Pfammater E., Corvi C., 2004. Characterization of Swiss vineyards using isotopic data in combination with trace elements and classical parameters. European Food Research and Technology, 219, 97-104. [77]

Gulson B.L., Mizon K.J., Korsch M.J., Eschnauer H.R., Lee T.H., 1990. Are tin-lead capsules a source of lead in wine? Wine Industry Journal, 368, 274-276. [78]

Gulson B.L., Lee T.H., Mizon K.J., Korsch M.J., Eschnauer H.R., 1992. The application of lead isotope ratios to determine the contribution of the tin-lead to the lead content of wine. American Journal of Enology and Viticulture, 43 (2), 180-190. [79]

Harding R., 1991. Origine du plomb dans le vin. Feuillet Vert de l’OIV, 899. [80]

Henick-Kling T., Stoewsand G.S., 1993. Lead in wine. American Journal of Enology and Viticulture, 44, 459-463. [81]

Herce-Pagliai C., Moreno I., González G., Repetto M., Cameán A.M., 2002. Determination of total arsenic, inorganic and organic species in wine. Food Additives and Contaminants, 19 (6), 542-546. [82]

Herrero-Latorre C., Médina B., 1990. Utilisation de quelques éléments minéraux dans la différenciation des vins de Galice de ceux d’autres régions d’Espagne. Journal International des Sciences de la Vigne et du Vin, 24 (4), 147-156. [83]

Hurrel R., Moser R.U., Staehelin H.B., Walter W., Zimmermann M., 2001. Micronutrients. Science des Aliments, 21, 375-413. [84]

Ishii T., Matsunaga T., 1996. Isolation and characterisation of a boron-rhamnogalacturonan II complex from cell walls of sugar beet pulp. Carbohydrate Research, 284, 1-9. [85]

Jaganathan J., 2001. A random testing of table wines for arsenic using electrothermal atomic absorption spectrometry. Atomic Spectroscopy, 22 (2), 280-283. [86]

Jaganathan J., Dugar S.M., 1998. Determination of selenium in wines usign elctrothermal atomic absorption spectrometry with Zeeman background correction. American Journal of Enology and Viticulture, 49 (2), 115-118. [87]

Jakubowski N., Brandt R., Stuewer D., Eschnauer H.R., Görtges S., 1999. Analysis of wines by ICP-MS: is the pattern of the rare earth elements a reliable fingerprint for the provenance? Fresenius Journal of Analytical Chemistry, 364, 424-428. [88]

Jorhem L., Sundström B., 1995. Direct determination of lead in wine using graphite furnace AAS. Atomic Spectroscopy, 5, 226-228. [89]

Jos A., Moreno I., González A.G., Repetto G., Cameán A.M., 2004. Differentiation of sparkling wines (cava and champagne) according to their mineral content. Talanta, 63, 377-382. [90]

Karadjova I.B., Lampugnani L., Onor M., D’Ulivo A., Tsalev D., 2005. Continuous flow hydride generation-atomic fluorescence spectrometric determination and speciation of arsenic in wine. Spectrochimica Acta Part B, 60, 816-823. [91]

Kaufmann A., 1998. Lead in wine. Food Additives and Contaminants, 15 (4), 437-445. [92]

Kelly M.T., Gilis M., Blaise A., 2005. Utilisation des bentonites: effect sur les teneurs en aluminium des vins. Revue Française d’Oenologie, 210, 15-18. [93]

Kildahl B.T., Lund W., 1996. Determination of arsenic and antimony in wine by electrothermal atomic absorption spectrometry. Fresenius Journal of Analytical Chemistry, 354, 93-96. [94]

Kristl J., Veber M., Slekovec M., 2002. The application of ETAAS to the determination of Cr, Pb and Cd in samples taken during different stages of the winemaking process. Analytical and Bioanalytical Chemistry, 373, 200-204. [95]

Page 43: Tese doutoramento Sofia Catarino.pdf

Introdução geral

32

Lara R., Cerutti S., Salonia J.A., Olsina R.A., Martine L.D., 2005. Trace element determination of Argentine wines using ETAAS and USN-ICP-OES. Food and Chemical Toxicology, 43, 293-297. [96]

Larcher R., Nicolini G., Pangrazzi P., 2003. Isotope ratios of lead in Italian wines by inductively coupled plasma mass spectrometry. Journal of Agricultural and Food Chemistry, 51, 5956-5961. [97]

Larroque M., Cabanis J.C., Vian L., 1994. Determination of aluminium in wines by direct graphite furnace atomic absorption spectrometry. Journal of AOAC International, 77, 463-466. [98]

Larsen E.H., Anderson N.L., Moller A., Peterson A., Mortensen G.K., Peterson J., 2002. Monitoring the content and intake of trace elements from food in Denmark. Food Additives and Contaminants, 19 (1), 33-46. [99]

Latorre M.J., Herrero C., Médina M., 1992. Utilisation de quelques éléments minéraux dans la différenciation des vins de Galice. Journal International des Sciences de la Vigne et du Vin, 26 (3), 185-193. [100]

Lay H., Meyer L., 1989. The distribution of aluminium in marc, yeast, wine and deposit. Wein-Wissenschaft, Wiesbaden, 44, 173-176. [101]

Lemos V.A., de la Guardia M., Ferreira S.L.C., 2002. An on-line system for preconcentration and determination of lead in wine samples by FAAS. Talanta, 58, 475-480. [102]

Lendinez E., Lopez C., Cabrera C., Lorenzo M.L., 1998. Determination of chromium in wine and other alcoholic beverages consumed in Spain by electrothermal atomic absorption spectrometry. Journal of AOAC international, 8 (5), 1043-1047. [103]

Lima M.T.R., Cabanis M.T., Matos L., Cassanas G., Kelly M.T., Blaise A., 2004. Determination of lead and cadmium in vineyard soils, grapes and wines of the Azores. Journal International des Sciences de la Vigne et du Vin, 38 (3), 163-170. [104]

Lobinski R., Szpunar-Lobinska J., Adams F., 1993. Speciation analisis of organolead compounds in wine by capillary gas chromatography/microwave-induced-plasma atomic emission spectrometry. Journal of AOAC International, 76 (6), 1262-1267. [105]

Lopes J.M.P., Christo M.A.H., Barros P., Moreira M.S.R.B., Simões J.T., 1991. Sondage des valeurs de plomb dans les vins portuguais. Feuillet Vert de l’OIV, 890. [106]

Lopez F.F., Cabrera C., Lorenzo M.L., Lopez M.C., 1998. Aluminium levels in wine, beer and other alcoholic beverages consumed in Spain. The Science of the Total Environment, 220, 1-9. [107]

Machado-Nunes M., Laureano O., Ricardo-da-Silva J.M., 1998. Influência do tipo de cola (caseína e bentonite) e da metodologia de aplicação nas características físico-químicas e sensoriais do vinho branco. Ciência e Técnica Vitivinícola, 13 (1-2), 7-28. [108]

Magalhães M.J., Sequeira E.M., Lucas M.D., 1985. Copper and Zinc in Vineyards of Central Portugal. Water, Air, and Soil Pollution, 26, 1-17. [109]

Mandal B.K., Suzuki K.T., 2002. Arsenic round the world: a review. Talanta, 58, 201-235. [110]

Marchal R., Barret J., Maujean A., 1995. Relations entre les caractéristiques physico-chimiques d’une bentonite et son pouvoir d’adsorption. Journal International des Sciences de la Vigne et du Vin, 29 (1), 27-42. [111]

Marchal R., Weingartner S., Voisin C., Jeandet Ph., Chatelain F., 2002. Utilisation de modèles mathématiques pour optimiser des doses de bentonite gonflée et non gonflée lors du collage des vins blancs. Partie I : Clarification et stabilisation colloïdale. Journal International des Sciences de la Vigne et du Vin, 36 (3), 169-176. [112]

Martin de la Hinojosa, Dapena P., Cerezo M.J., Hitos P., 1994. Etude de la présence de certains contaminants dans les vins espagnols. Feuillet Vert de l’OIV, 965. [113]

Martin G.J., Mazure M., Jouitteau C., Martin Y.L., Aguile L., Allain P., 1999. Characterization of the geographic origin of Bordeaux wines by a combined use of isotopic and trace element measurements. American Journal of Enology and Viticulture, 50 (4), 409-417. [114]

Page 44: Tese doutoramento Sofia Catarino.pdf

Introdução geral

33

Matthews M.R., Parsons P., 1993. A simple method for the determination of lead in wine using Zeeman electrothermal atomization atomic absorption spectrometry. Atomic Spectroscopy, 14, 41-46. [115]

Maujean A., 1993. Propriétés physico-chimiques des bentonites: applications oenologiques. Revue Française d’œnologie, 143, 43-53. [116]

McKinnon A.J., Cattrall R.W., Scollary G.R., 1992. Aluminum in wine – its measurement and identification of major sources. American Journal of Enology and Viticulture, 43 (82), 166-170. [117]

Médina B., 1996. Wine authenticity. In: Food authentication. Edited by P.R. Ashurst and M.J. Dennis, Blackie Academic & Professional (Chapman and Hall). [118]

Médina B., Sudraud P., 1980. Teneur des vins en chrome et en nickel. Causes d’enrichissement. Connaissance de la Vigne et du Vin, 14 (2), 79-96. [119]

Médina B., Augagneur S., Barbaste M., Grousset F.E., Buat-Ménard P., 2000. Influence of atmospheric pollution on the lead content of wines. Food Additives and Contaminants, 17 (6), 435-445. [120]

Mer H., 2004. Uruguayan wines by atomic absorption spectroscopy. Atomic Spectroscopy, 25 (4), 177-184. [121]

Mikkelsen O., Schroder K.H., 2002. Voltammetry using a dental amalgam electrode for heavy metal monitoring of wines and spirits. Analytica Chimica Acta, 458, 249-256. [122]

Minguez S., Hernandez P., Garcia J., Médina B., Bartra E., Vilavella M., Espinás E., 1997. Presence du plomb dans les vins et efficacite des techniques soustractives, Doc OIV (Groupe d’Experts “Technologie du vin”). [123]

Misiego A.S., Carra R.M.G.-M., Carracedo M.P.A., Sánchez-Simón M.T.G., 2004. Electroanalytical determination and fractionation of copper in wine. Journal of Agricultural and Food Chemistry, 52, 5316-5321. [124]

Nicolini G., Larcher R., Mattivi F., 2001. ICP-OES and ICP-MS measurement of mineral elements in grapes and wines: analytical, compositional and enological aspects. Actas do VIII Viticulture and Enology Latin-American Congress, Montevideo, Uruguai. [125]

Nicolini G., Larcher R., Pangrazzi P., Bontempo L., 2004. Changes in the contents of micro- and trace-elements in wine due to winemaking treatments. Vitis, 43 (1), 41-45. [126]

OIV, 1995a. Compte rendu des travaux de la Sous-Comission des Methods d’Analise, 2291/95. Office International de la Vigne et du Vin, Paris. [127]

OIV, 1995b. Le plomb. Chiers Scientifiques et techniques. Office International de la Vigne et du Vin, Paris. [128]

OIV, 2005. Recueil des méthodes internationales d’analyse des vins et des môuts. Organisation International de la Vigne et du Vin, Paris. [129]

O’Neill M.A., Warentfeltz D., Kates K., Pellerin P., Doco T., Darvill A., Albersheim P., 1996. Rhamnogalacturonan II, a pectic polysaccharide in the walls of growing plant cells, forms a dimer that is covalently cross-linked by a borate ester. The Journal of Biological Chemistry, 271, 22923-22930. [130]

Pederson G.A.L., Mortensen G.K., Larsen E.H., 1994. Beverages as a source of toxic trace elements intake. Food Additives and Contaminants, 11 (3), 351-363. [131]

Pellerin P., O’Neill M.A., Pierre C., Cabanis M.-T., Darvil A.G., Albersheim P., Moutounet M., 1997. Complexation du plomb dans les vins par les dimères de rhamnogalacturonane II, un polysaccharide pectique du raisin. Journal International des Sciences de la Vigne et du Vin, 31 (1), 33-41. [132]

Pellerin P., O’Neill M.A., 1998. The interaction of the pectic polysaccharide rhamnogalacturonan II with heavy metals and lanthanides in wines and fruit juices. Analusis Magazine, 26 (6), M32-M36. [133]

Pérez-Jordán M.Y., Soldevila J., Salvador A., Pastor A., de la Guardia M., 1998. Inductively coupled plasma mass spectrometry analysis of wines. Journal of Analytical Atomic Spectrometry, 13, 33-39. [134]

Page 45: Tese doutoramento Sofia Catarino.pdf

Introdução geral

34

Peynaud E., 1996. Enologia pratica. Conocimiento y elaboracion del vino. Ediciones Mundi-Prensa, Bilbao. [135]

Pietrzac U., McPhail D.C., 2004. Copper accumulation, distribution and fractionation in vineyard soils of Victoria, Australia. Geoderma, 122 (2-4), 151-166. [136]

Postel W., Meier B., Markert R., 1986. Influence on processing aids on the content of mineral compounds and trace elements in wine. Bentonite. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria, 36 (1), 20-27. [137]

Prasad M.N.V., 1998. Metal-biomolecule complexes in plants: occurrence, functions and applications. Analusis Magazine, 26, 25-28. [138]

Pyrzynska K., 2004. Analytical methods for the determination of trace metals in wine. Critical Reviews in Analytical Chemistry, 34, 69-83. [139]

Ribéreau-Gayon J., Peynaud E., Sudraud P., Ribéreau-Gayon P., 1982. Sciences et techniques du vin. Tome 1 – Analyse et contrôle des vins. 2e édition. Dunod, Paris. [140]

Ribéreau-Gayon P., Glories Y., Maujean A., Dubourdieu D., 1998. Traité d’Oenologie 2. Chimie du vin. Stabilisation et traitements. Dunod, Paris. [141]

Riganakos K.A., Veltsistas P.G., 2003. Comparative spectrophotometric determination of the total iron content in various white and red Greek wines. Food Chemistry, 82, 637-643. [142]

Rizzon L.A., Miele A., Rosier J.P., 1997. Discrimination of wines from the mercosul countries according to their mineral composition. Journal International des Sciences de la Vigne et du Vin, 31 (1), 43-47. [143]

Rodushkin I., Ödman F., Appelblad P.K., 1999. Multielement determination and lead isotope ratio measurement in alcoholic beverages by high-resolution inductively coupled plasma mass spectrometry. Journal of Food Composition and Analysis, 12, 243-257. [144]

Rosman K.J.R., Chisholm W., Jimi S., Candelone J-P., Boutron C.F., Teissèdre P-L., Adams F.C., 1998. Lead concentrations and isotopic signatures in vintages of french wine between 1950 and 1991. Environmental Research, Section A, 78, 161-167. [145]

Ruzic N., 2000. Effect of copper and lead on the fermentation of grape juice. Lozarstvo i Vinarstvo, 48 (2), 57-59. [146]

Salinas I., Esparza I., Gómez S., Santamaría C., Fernández J.M., 2005. A study of heavy metal complexation in grape juice. Electroanalysis, 17 (5-6), 469-475. [147]

Salvo F., La Pera L., Di Bella G., Nicotina M., Dugo G., 2003. Influence of different mineral and organic pesticida treatments on Cd(II), Cu(II), Pb(II), and Zn(II) contents determined by derivative potentiometric stripping analysis in Italian white and red wines. Journal of Agricultural and Food Chemistry, 51, 1090-1094. [148]

Scollary G., 1997. Metals in wine: contamination, spoilage and toxicity. Analusis, 25 (3), M26-M30. [149]

Segura M., Madrid Y., Camara C., 1999. Evaluation of atomic fluorescence and atomic absorption spectrometric techniques for the determination of arsenic in wine and beer by direct hydride generation sample introduction. Journal of Analytical Atomic Spectrometry, 14 (2), 131-135. [150]

Seruga M., Grgic J., Zdravko G., Seruga B., 1998. Aluminium content of some croation wines. Deutsche Lebensmittel-Rundschau, 94 (19), 336-340. [151]

Simões M.M.T., Granhão S., Simões J.T, 1995. Étude du cadmium dans les vins de liqueur portugais. Feuillet Vert de l’OIV, 1004. [152]

Soares M.E., Bastos M.L., Ferreira M.A., 1993. Determination of copper, iron, aluminium, lead and cadmium in cork stoppers by electrothermal atomic absorption spectrometry. Journal of Analytical Atomic Spectrometry, 8, 655-657. [153]

Soares M.E., Bastos M.L., Ferreira M.A., 1995. Quantification of Ag, Co, Si, and Zn in Port Wine by atomic absorption spectrometry. Atomic Spectroscopy, 6, 256-260. [154]

Page 46: Tese doutoramento Sofia Catarino.pdf

Introdução geral

35

Soulis Th., Arvanitoyannis I., Kavlentis E., 1989. Iron, copper, manganese and zinc contents of some bottled and non-bottled Greek wines. Science des Aliments, 9, 799-803. [155]

Stockley C.S., Lee T.H., 1995. Much ado about lead in wine? An Australian review. Journal of Wine Research, 6 (1), 5-17. [156]

Stockley C.S., Smith L.H., Tiller K.G., Gulson B.L., Osborn C.D., Lee T.H., 2003. Lead in wine: a case study on two varieties at two wineries in South Australia. Australian Journal of Grape and Wine Research, 9 (1), 47-55. [157]

Stroh A., Brückner P., Völlkopf U., 1994. Multielement analysis of wine using ICP-MS. Atomic Spectroscopy, 2, 100-106. [158]

Sudraud P., Médina B., Grenon J.P., 1989. Enrichissement des vins en plomb par les capsules de surbouchage plomb-etain. Essai de capsules a proteccion interne. Feuillet Vert de l’OIV, 842. [159]

Sudraud P., Médina B., Grenon J.P., 1994. Teneurs en éléments minéraux des vins. Journal International des Sciences de la Vigne et du Vin, 28 (1), 69-75. [160]

Szentmihalyi K., Csiktusnadi-Kiss G.A., Keszler A., Kotai L., Candeias M., Bronze M.R., Vilas-Boas L., Spauger I., Forgacs E., 2000. Method development for measurement of elements in Hungarian red wines by inductively coupled plasma optical emission spectrometry (ICP-OES). Acta Alimentaria, 29 (2), 105-121. [161]

Szpunar J., Pellerin P., Makarov A., Doco T., Williams P., Médina B., Lobinski R., 1998. Speciation analysis for biomolecular complexes of lead in wine by size-exclusion high-performance liquid chromatography-inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry, 13, 749-754. [162]

Szpunar J., Pellerin P., Makarov A., Doco T., Williams P., Lobinski R., 1999. Speciation of metal-carbohydrate complexes in fruit and vegetable samples by size exclusion HPLC-ICP-MS. Journal of Analytical Atomic Spectrometry, 14, 639-644. [163]

Tahiri M., Pellerin P., Tressol J.C., Doco T., Pépin D., Rayssiguier Y., Coudray C., 2000. The rhamnogalacturonan-II dimer decreases intestinal absorption and tissue accumulation of lead in rats. The Journal of Nutrition, 130 (2), 249-253. [164]

Tasev K., Karadjova I., Stafilov T., 2005. Determination of inorganic and total arsenic in wines by hidryde generation atomic absorption spectrometry. Microchimica Acta, 149, 55-60. [165]

Taylor V.T., Longerich H.P., Greenough J.D., 2003. Multielement analysis of Canadian wines by inductively coupled plasma mass spectrometry (ICP-MS) and multivariate statistics. Journal of Agricultural and Food Chemistry, 51, 856-860. [166]

Teissèdre P.L., Cabanis M.T., Daumas F., Cabanis J.C, 1993a. Evolution de la teneur en plomb au cours de l’élaboration des vins des Côtes du Rhône et de la Vallée du Rhône. Revue Française d’Oenologie. Cahiers Scientifiques, 14, 6-18. [167]

Teissèdre P.L., Cabanis M.T., Champagnol F., Cabanis J.C., 1993b. Étude de la teneur en plomb des feuilles et de quelques organes du cep de vigne. Bulletin de l’OIV, 66, 843-854. [168]

Teissèdre P.L., Cabanis M.T., Champagnol F., Cabanis J.C., 1994a. Lead distribution in grape berries. American Journal of Enology and Viticulture, 45 (2), 220-228. [169]

Teissèdre P.L., Lobinski R., Cabanis M-T., Lobinski J.S., Cabanis J-C., Adams F.C., 1994b. On the origin of organolead compounds in wine. The Science of the Total Environment, 153, 247-252. [170]

Teissèdre P.L., Cabanis M.T., Daumas F., Cabanis J.C., 1996. Evolution des teneurs en cadmium au cours de l’elaboration des vins des cotes-du-rhone et de la vallee du rhone. Feuillet Bleu de l’OIV, 19. [171]

Teissèdre P.L., Krosniak M., Portet K., Gasc F., Waterhouse A.L.L., Serrano J.J., Cabanis J.C., Cros G., 1998a. Vanadium levels in French and Californian Wines: influence on vanadium dietary intake. Food Additives and Contaminants, 15 (5), 585-591. [172]

Page 47: Tese doutoramento Sofia Catarino.pdf

Introdução geral

36

Teissèdre P.L., Cabrera Vique C., Cabanis M.T., Cabanis J.C., 1998b. Determination of nickel in French wines and grapes. American Journal of Enology and Viticulture, 49 (2), 205-210. [173]

Terwell, 1990. Résumé d’une etude sur le plomb dans le vin au Danemark. Feuillet Vert de l’OIV, 874. [174]

Thiel G., Geisler G., Bleschschmidt I., Danzer K., 2004. Determination of trace elements in wines and classification according to their provenance. Analytical and Bioanalytical Chemistry, 378, 1630-1636. [175]

Vasconcelos M.T., Azenha M., 2001. In vitro assessment of the bioavailabitlity of Pb and Cu in different Portuguese wines. Feuillet Vert de l’OIV, 1135. [176]

Vidal S., Doco T., Moutounet M., Pellerin P., 1999. Le rhamnogalacturonane II, un polysacharide complexe du vin aux propriétés remarquables. Revue Française d’Oenologie, 178, 12-17. [177]

Vidal M.T., Poblet M., Constantí M., Bordons A., 2001. Inhibitory effect of copper and dichlofluanid on Oenococcus oeni and malolactic fermentation. American Journal of Enology and Viticulture, 52 (3), 223-229. [178]

Wangkarn S., Pergantis S.A., 1999. Determination of arsenic in organic solvents and wines usign microscale flow injection inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry, 14, 657-662. [179]

Wurzinger A., Netzer M., Heilli K., Bandion F., 1994. Migration of components of bentonites during the fermentation of must. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria, 44, 218-221. [180]

Page 48: Tese doutoramento Sofia Catarino.pdf

Capítulo 2

DESENVOLVIMENTO E VALIDAÇÃO DE MÉTODOS DE ANÁLISE

Page 49: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

39

2.1. DETERMINATION OF ALUMINIUM IN WINE BY GRAPHITE FURNACE AAS:

VALIDATION OF ANALYTICAL METHOD

S. Catarino1,2, A.S. Curvelo-Garcia1, R. Bruno de Sousa2

1INIA. Estação Vitivinícola Nacional. 2565-191 Dois Portos. Portugal 2Instituto Superior de Agronomia – Departamento de Química Agrícola e Ambiental. Tapada da Ajuda. 1349-017 Lisboa. Portugal

Abstract

A reliable and rapid method is presented for the determination of the total concentration of Al in wine by graphite furnace atomic absorption spectrometry (GFAAS). The method was validated and its analytical characteristics were checked (algorithm calibration curve, selectivity, analytical limits, accuracy, and precision). White, red, and fortified wines were used for the analysis. The method involves reduced risk of sample contamination using a one-step pre-treatment procedure. The calibration does not require matrix simulation since it is obtained with aqueous solutions of Al. The matrix effect is minimized by the high dilution of the sample (1:40). Calibration curves from aqueous standards were used to calculate the detection and quantification limits of the analyte, resulting in 1 µg.L-1 and 2 µg.L-1, respectively. The determination was performed in the linear range of 2-50 µg.L-1. Spiking the wine samples with aqueous Al resulted in recoveries between 96% and 105%. Replicate wine samples at several Al concentration levels resulted in relative standard deviations better than 3.0% (n = 10), indicating that the analytical method is of high precision. The advantages of the method (practicability, specificity, precision, and accuracy) make it suitable for routine Al determinations in wine.

Atomic Spectroscopy (2002) 23 (6), 196-200

Page 50: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

40

2.1.1. INTRODUCTION

The determination of metals in wines is important because some metals are toxic and thus a

health hazard, while others can lead to spoilage of the wine. Information about the metal content is

also indicative of the geographic origin of the wines (1-13).

From an enological point of view, Al is a trace element and its determination is important

because it is toxic and can also lead to spoilage through haze formation as well as the generation of

undesirable tastes and aromas (14,15). With regard to the enological specificities of Al, it is

possible that Al combines with tartaric acid and other organic acids (16-18).

The major sources of Al contamination are pesticides, soil, tanks, and some enological

products such as bentonite and red tannin (16-21). Other products used as technological aids such

as infusorial earths, perlite and carbonate of lime, are also potential sources of Al contamination in

wines (22-24). In bottle-aged wines, another possible source of contamination are the aluminum

capsules (16,25). Contamination with this metal, abundant and omnipresent, is difficult to control

during wine processing and even during analytical monitoring (Al is the most abundant metal and

the third most widespread element in the earth’s crust). In every wine, natural and soil-related Al

can be found. Published data indicate that the content of total Al in wines is lower than 2 mg.L-1

(16,17,21,26-28), which suggests that there has been a reduction in Al found in wines in recent

years (15,19). Currently, no maximum limit for Al in wines has been established by the

International Organisation of Vine and Wine (OIV, Paris, France).

The determination of Al in wines presents problems such as poor sensitivity and selectivity

of the method and additional contamination risks during the analytical processes. The preferred

method used in wine laboratories is electrothermal AAS (ETAAS) (6,16,20,25,26,28) because of

its high sensitivity, precision, and selectivity, and it requires no sample preconcentration in

comparison to flame AAS (FAAS). Other methods used include inductively coupled plasma optical

emission spectrometry (ICP-OES) (7,11,29,30,31) and inductively coupled plasma mass

spectrometry (ICP-MS) (8,9,12,27,32,33).

The aim of the present work was to develop and validate a reliable and rapid analytical

method for the GFAAS determination of Al, with high sensitivity, selectivity, accuracy and

precision, applicable to different wines (table and fortified wines). The instrumental and analytical

conditions for pre-treatment of the samples were optimized. Since only a simple dilution of the

wine is required as pre-treatment, it reduces the risk of sample contamination. The method

presented is improved over previous methods and its complete validation (including algorithm

calibration curve, selectivity, accuracy, analytical limits and precision) is new.

Page 51: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

41

2.1.2. EXPERIMENTAL

2.1.2.1. Instrumentation

A PerkinElmer Model 4110 ZL graphite furnace atomic absorption spectrometer

(PerkinElmer, Shelton, CT USA) was used, equipped with Zeeman-effect background correction

and a Model AS-72 autosampler, and AAWinLab, version 2.5, spectrometer software. Argon

N50 (purity > 99,999%) was used to protect and purge the graphite tubes, with an internal flow rate

of 250 mL.min-1. A PerkinElmer Lumina hollow cathode lamp was used. Table 1 lists the

spectrometer settings and furnace programs for the pyrolytically coated graphite tubes with end

caps and L’vov platforms. The measurement mode used for all elements was integrated

absorbance. The autosampler was programmed to pipette and dispense 20 µL of the sample

(standard or wine) onto the platform. Deionized water using a Seralpur Pro 90 CN purifier was

used throughout.

2.1.2.2. Labware

To eliminate possible contamination, all glassware and polyethylene materials (volumetric

flasks, micropipette tips, and autosampler cups) were immersed for 24 hours in freshly prepared

20% (v/v) HNO3 and then rinsed thoroughly with deionized water before use. Thorough cleaning

of the glassware and the use of plastic materials are very important for this type of analysis.

2.1.2.3. Reagents and calibration

All standards (10, 20, 30, 40, and 50 µg.L-1) were prepared daily from a 1000 mg.L-1

solution (CertiPUR, Merck) of Al(III) in HNO3 (0.2% v/v) and C2H6O (0.3% v/v) with deionized

water (conductivity < 0.1 µS.cm-1). The wine samples were diluted (1:40) with HNO3 (0.2% v/v).

Immediately after preparation, the standards and sample solutions were transferred into

polyethylene labware. All standards and samples were analyzed in triplicate. The high dilution of

the samples suggests that the viscosity and matrix effects can be controlled without the use of a

surfactant. However, the efficiency of using both palladium nitrate and magnesium nitrate [0.005

mg Pd(NO3)2 and 0.003 mg Mg(NO3)2], and magnesium nitrate [0.015 mg Mg(NO3)2 as

recommended by PerkinElmer] as the chemical modifiers was tested for the wine samples and

standards.

Page 52: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

42

2.1.3. RESULTS AND DISCUSSION

2.1.3.1. Selection of aluminium furnace program

Optimization of the furnace program was carried out with a solution of Al at a

concentration of about 30 µg.L-1, in parallel with a diluted sample of red wine. These solutions

were also used for the optimization of the overall furnace program, which is summarized in Table

1. This study was directed not only at obtaining a more sensitive and reproducible method, but also

at reducing analysis time and cost.

Table 1 - Furnace programme

Step Temperature (ºC) Ramp time (s) Hold time (s)

Dry 1 110 1 30 Dry 2 140 15 30 Pyrolysis 1600 10 20 Atomizationa 2400 0 5 Clean-out 2500 1 3

a stop argon flow Notes: wavelength (λ) = 309.3 nm; slit-width = 0.7 nm; hollow cathode lamp current = 25 mA; background correction = Zeeman effect; integration time = 5 s; injection volume = 20 µL; total time = 115 s.

The dry step was established by visual control of the sample behaviour during the heating

stage and by the consistency of instrumental output. A two-stage procedure for the dry step proved

convenient. During the first stage (110ºC), the alcohol volatilizes without spattering and drying of

the sample is completed during the second stage (140ºC).

For the ashing and atomization steps, the change in integrated absorbance with a change in

temperature was studied. The influence of different temperatures, ramp times, and hold times was

also assessed. This study was carried out in three ways: with two matrix modifiers (magnesium

nitrate and palladium nitrate), with only one matrix modifier (magnesium nitrate), and without

matrix modifiers. The optimum ashing and atomization temperatures using these three options

were: 1600ºC and 2350ºC, 1700ºC and 2500ºC, 1600ºC and 2400ºC, respectively.

Ashing and atomization temperatures of 1600ºC and 2400ºC, respectively, proved to be

suitable for the adequate elimination of the matrix without loss of the element, thereby rendering

chemical modifiers unnecessary. The effect of the matrix is reduced because the relatively high

concentrations of Al in wine require a substantial sample dilution (1:40). The possibility of greatly

diluting the samples also allows the viscosity of the samples and standards to be controlled without

the use of a surfactant. Thus, no chemical modifiers, which would constitute an additional source of

contamination, were used.

Page 53: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

43

2.1.3.2. Algorithm calibration curve

Calibration against acidified standard solutions was carried out and the standard additions

technique used to investigate the linearity (y=a+bx) of the calibration curve. The wines were spiked

with five concentration levels of Al, never exceeding the total content of 50 µg.L-1.

The variance homogeneity was investigated and confirmed by the Cochran test. The results

presented in Table 2 confirm that for each curve a high percentage of the total variance is explained

by the regression, which has a calculated F higher than the Fisher’s F at 99.5% (10.6). The residual

variance due to the adjustment error (lack of fit) is significant in all cases, except for Red Wine 2.

This may result from the small variability within each group. Although the residual variability is

very small (since the pure and adjustment errors are not of the same magnitude), the test is

significant. This explanation is supported by the variance analysis of curve 2 (Table 2), which was

plotted from an independent set of standard solutions and presents an F value near theoretic F.

Another explanation for curves obtained from standard solutions may be for the group of zero

concentration. This is the only group where the variance is zero. By exclusion of this group, it was

found to be true for curve 2.

Table 2 - Characteristic parameters of the calibration curves (variance analysis of the regressions)

Curve 1 Curve 2 White wine Red wine 1 Red wine 2 Fortified

wine

y = a + bx y = -0.00198 +

0.00327x

y = -0.00283 +

0.00346x

y = 0.01803 +

0.00377x

y = 0.02173 +

0.00338x

y = 0.02680 +

0.00349x

y = 0.08437 +

0.00381x

r 0.9993 0.9985 0.9968 0.9983 0.9979 0.9976

a 0.00198 ±

0.00203

-0.00283 ±

0.00305

0.01803 ±

0.00488

0.02173 ±

0.00321

0.02680 ±

0.00442

0.08437 ±

0.00213

b 0.00327 ±

0.00006

0.00346 ±

0.00010

0.00377 ±

0.00016

0.00338 ±

0.00011

0.00349 ±

0.00014

0.00381 ±

0.00014

∆b/b 1.8% 2.9% 4.2% 3.3% 3.9% 3.7%

(F1)

F(0.005; 1,16) = 10.6

10711.0 (s) 5275.5 (s) 2454.6 (s) 4574.8 (s) 3135.2 (s) 3288.6 (s)

(F2)

F(0.005; 4,12) = 6.5

F(0.005; 3,10) = 8,1

26.1 (s)

10.6* (s)

9.6 (s)

6.9* (ns)

43.2 (s) 77.5 (s) 6.6 (ns) 43.7 (s)

Cochran (C)

C(0.05; 6,3) = 0.72

0.42

(ns)

0.40

(ns)

0.34

(ns)

0.49

(ns)

0.45

(ns)

0.50

(ns)

(s): significant difference; (ns): without significant difference; the calculate values should be lower than theoretic values, except for F1; *F2 calculated by exclusion of zero concentration group.

Page 54: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

44

2.1.3.3. Selectivity

When there are no interferences, the slope of the calibration curve from aqueous standards

should be the same as the slope of the curve plotted by wine standard addition. Except for the

fortified wine, there are no significant differences (α = 5%) (Table 3). However, if α = 1%, there

are no significant differences for all wines. The high dilution of the samples reduces the matrix

effect, allowing calibration with aqueous standards.

Table 3 - Selectivity: comparison of slopes

Calibration curves

(b)

Wines

(b)

Student t

(calculated value)

0.00359 0.00377 (White wine) 1.9 (ns)

0.00327 0.00338 (Red wine) 1.9 (ns)

0.00359 0.00349 (Red wine) 1.2 (ns)

0.00359 0.00381 (Fortified wine) 2.6 (ns, α = 0.01)

t (0.05; 32) = 2.0 ; t (0.01; 32) = 2.7; (ns): without significant difference.

2.1.3.4. Accuracy

Accuracy was determined by the method of standard additions, and used to determine the

recovery of the spiked analyte. This approach was used because it was not possible to prepare a

blank sample without the presence of the analyte. The recovery may change with the analyte level

and sample matrix. The results of these experiments confirmed linearity. Spiking wine samples

with aqueous aluminum gave recoveries between 96% and 105%, demonstrating the accuracy of

the method (Table 4).

Table 4 - Recovery estimates

5 addition levels (5 to 40 µg.L-1)

White wines – 101.7 ± 3.0%; 98.3 ± 1.6% (96.8 to 105%)

Red wines – 99.5 ± 3.1%; 99.8 ± 1.6 (96.1 to 102%)

Fortified wine – 99.9 ± 2.6% (97.3 to 102.5%)

2.1.3.5. Analytical limits

The detection limit was calculated as the mean concentration, plus three standard

deviations of the blank, obtained for 20 determinations. The quantification limit was calculated as

Page 55: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

45

the mean concentration, plus 10 standard deviations of the blank. The experimental conditions used

enabled the determination of 1 µg.L-1 and the quantification of 2 µg.L-1. The detection limit of 1

µg.L-1 permits the analytical determination of Al in undiluted wines with 40 µg.L-1. In fact, the Al

concentration in wines is generally higher than 40 µg.L-1.

2.1.3.6. Precision

Replicates of three white wines, three red wines, and one fortified wine, at several Al

concentration levels, gave relative standards deviations between 2% and 3% (n = 10), indicating

that the analytical method is of high within-run precision (repeatability) (Table 5). The between-run

precision (reproducibility) was also determined in three red wines and one fortified wine. The

relative standard deviations were from 1% to 8% (N = 3). The higher value was found in the wine

with lower Al concentrations.

Table 5 - Repeatability (within-run precision) and reproducibility (between-run precision)

Repeatability White wine 34 ≤ r ≤ 92 µg.L-1 (1.54% ≤ RSD ≤ 2.54%)

Red wine 29 ≤ r ≤ 108 µg.L-1 (2.11% ≤ RSD ≤ 2.64%)

Fortified wine r = 83 µg.L-1 RSD = 2.08%

Reproducibility White wine 17 ≤ R ≤ 100 µg.L-1 (0.88% ≤ RSD ≤ 8.31%)

Red wine 56 ≤ R ≤ 140 µg.L-1 (3.17% ≤ RSD ≤ 4.31%)

Fortified wine R = 58 µg.L-1 RSD = 1.43%

2.1.4. CONCLUSION

The method described can be used to determine total Al concentration in table and fortified

wines on a routine basis. The analytical limits for Al in samples prior to dilution are satisfactory for

the control of this element in wines. The analytical procedure requires the dilution of the sample to

reduce the Al concentration, which has the advantage of overcoming matrix effects. The main

difficulty of the method is to keep the labware free from contamination during the analytical

process.

2.1.5. REFERENCES 1. M. Gonzales-Larraina, A. Gonzales, and B. Médina, Connaissance de la Vigne et du Vin, 21 (2), 127

(1987).

2. C. Herrero-Latorre and B. Médina, Journal International des Sciences de la Vigne et du Vin, 24 (4), 147 (1990).

Page 56: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

46

3. M.J. Latorre, C. Herrero, and B. Médina, Journal International des Sciences de la Vigne et du Vin, 26, 185 (1992).

4. M.P. Day, B.L. Zhang, and G.J. Martin, American Journal of Enology and Viticulture, 45 (1), 79 (1994).

5. M.P. Day, B.L. Zhang, G.J. Martin, C. Asselin, and R. Morlat, Journal International des Sciences de la Vigne et du Vin, 29 (2), 75 (1995a).

6. M.P. Day, B. Zhang, and G.J. Martin, Journal of Science and Food Agriculture, 67, 113 (1995b).

7. J.B. Fournier and O. Hirsch, Actes Symp. In Vino Analytica Scientia, 532 (1997).

8. M.J. Baxter, H.M. Crews, M.J. Dennis, I. Goodall, and D. Anderson, Food Chemistry, 60 (3), 443 (1997).

9. J.D. Greenough, H.P. Longerich, and S.E. Jackson, Australian Journal of Grape and Wine Research, 3 (2), 75 (1997).

10. L.A. Rizzon, A. Miele, and J.P. Rosier, Journal International des Sciences de la Vigne et du Vin, 31 (1), 43 (1997).

11. G. Thiel and K. Danzer, Fresenius Journal of Analytical Chemistry, 357 (5), 553 (1997).

12. N. Jakubowski, R. Brandt, D. Stuewer, H.R. Eschnauer, and S. Gortges, Fresenius Journal of Analytical Chemistry, 364 (5), 424 (1999).

13. G.J. Martin, M. Mazure, C. Jouitteau, Y.L. Martin, L. Aguile, and P. Allain, American Journal of Enology and Viticulture, 50 (4), 409 (1999).

14. B. Rankine, Australian Grapegrower Winemaker, 234, 18 (1983).

15. H. Lay and L. Meyer, Wein-Wissenschaft, 44, 173 (1989).

16. M. Larroque and J.C. Cabanis, Journal of AOAC International, 77, 463 (1994).

17. H.R. Eschnauer and G.R. Scollary, Die Wein-Wissenschaft, 1, 24 (1995).

18. G.R. Scollary, Analusis, 25 (3), M26 (1997).

19. F. Testa, Vini d’Italia, 4, 7 (1989).

20. A.J. McKinnon, R.W. Cattrall, and G.R. Scollary, American Journal of Enology and Viticulture, 43 (2), 166 (1992).

21. K. Heili, M. Netzer, and F. Bandion, Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Fruchteverwertung Getranken, 47 (5), 159 (1997).

22. R. Enkelmann, Deutsche Lebensmittel-Rundschau, 85, 216 (1989).

23. R. Enkelmann, Deutsche Lebensmittel-Rundschau, 86, 314 (1990).

24. R. Enkelmann, Deutsche Lebensmittel-Rundschau, 88, 217 (1992).

25. A.A. Almeida, M.L. Bastos, M.I. Cardoso, M.A. Ferreira, J.F.L.C. Lima, and M.E. Soares, Journal of Analytical Atomic Spectrometry, 7, 1281 (1992).

26. F.F. Lopez, C. Cabrera, M.L. Lorenzo, and M.C. Lopez, The Science of the Total Environment, 220, 1(1998).

27. I. Rodushkin, F. Odman, and P.K. Appelblad, Journal of Food Composition and Analysis, 12 (4), 243 (1999).

28. A.M. Cameán, I.M. Moreno, M. López-Artíguez, M. Repetto, and A.G. González, Sciences des Aliments, 20, 433 (2000).

29. H. Eschnauer, L. Jakob, H. Meierer, and R. Neeb, Mikrochimica Acta, 3, 291 (1989).

30. K.H. Bauer, S. Hinkel, R. Neeb, P. Eichler, and H.R. Eschnauer, Weinwissenschaft, Wiesbaden, 49, 209 (1994).

31. J.B. Fournier, O. Hirsch, and G.J. Martin, Analusis, 26 (1), 28 (1998).

Page 57: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

47

32. M.Y. Pérez-Jordán, J. Soldevila, A. Salvador, A. Pastor, and M. Guardia, Journal of Analytical Atomic Spectrometry, 13, 33 (1998)

33. M.M. Castineira, R. Brandt, A.V. Bohlen, and N. Jakubowski, Fresenius Journal of Analytical Chemistry, 370 (5), 553 (2001).

Page 58: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

49

2.2. DETERMINATION OF COPPER IN WINE BY ETAAS USING CONVENTIONAL AND FAST

THERMAL PROGRAMS: VALIDATION OF ANALYTICAL METHOD

S. Catarino, Inês Pimentel, A.S. Curvelo-Garcia INIAP. Estação Vitivinícola Nacional. 2565-191 Dois Portos. Portugal

Abstract

A study involving conventional and fast thermal programs, with and without matrix modifiers, for the determination of Cu in wine by ETAAS is presented. With the conventional thermal program, better results were obtained when using matrix modifiers [Pd(NO3)2 and Mg(NO3)2]. This program presents better sensitivity and consequently it was the program selected for method validation. The fast thermal program allows a higher throughput, and can be a useful alternative to the conventional program. The method involves reduced risk of sample contamination eliminating prior treatment other than dilution (1:5). White and red wines with several Cu concentrations were used for method validation. The determination was performed in the linear range 1-50 µg.L-1; the detection limit in undiluted samples was 5 µg.L-1; the recoveries were between 93% and 100%; the repeatability (n=10), expressed as % RSD, was lower than 3%. Comparable values of concentrations were obtained by ETAAS and ICP-MS (differences<15%). The advantages of the method (practicability, sensitivity, precision, and accuracy) make it useful for routine determination of this metal in wine, when compared to flame AAS, which inevitably requires either sample pre-concentration or the use of standard additions method.

Atomic Spectroscopy (2005) 26 (2), 73-78

Page 59: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

50

2.2.1. INTRODUCTION

The content of metals at different stages of winemaking process is of great concern because

of legal and wine quality requirements (1,2). From an enological point of view, copper is a trace

element and its determination is important because of its toxicity, which can lead to spoilage

through haze formation (2-5). In addition, Cu2+ presents inhibitory effects on malolactic

fermentation, enhanced by higher concentrations of alcohol and other factors such as low pH and

higher contents of SO2 (6).

Part of the Cu found in wine is of endogenous origin due to the nature of the grapes

themselves. The major sources of Cu contamination are pesticides, namely copper sulphate applied

to prevent mildew, blending, and contact with copper, tin, or bronze materials during winemaking

and storage processes (2,3,7).

For musts, a range of acceptable Cu levels (1 - 7 mg.L-1) has been referenced (8). During

the fermentation process, most of the Cu is eliminated since it precipates mainly as copper

sulphide. Some published data indicate Cu wine concentrations of 4 mg.L-1, but most of wines

show concentrations below 0.2 mg.L-1 (1,3,8-12). In reduction conditions, wine containing more

than 0.2-0.4 mg.L-1 may suffer from copper cloudiness, a phenomenon well known which affects

stability and commercial acceptability. Nowadays, the Office International Organisation of Vine

and Wine (OIV, Paris, France), prescribes a maximum limit of 1 mg.L-1 for Cu (13).

The most frequently used method for Cu determination in wine is flame atomic absorption

spectrometry (FAAS) (1,11,14,15). This method is employed by the OIV and it is also the official

Portuguese and European Union method for Cu analysis (13,16). However, these procedures

present problems due to the complexity of the matrix in wine samples and the sensitivity of FAAS,

which is too low for the Cu content in the majority of wines. The standard additions method, used

in these situations, is not a good enough alternative since it presents low precision for the lowest Cu

concentrations (15). On the other hand, the alternative of sample pre-concentration decreases the

practicability of the method.

The determination of low Cu concentrations is feasible with electrothermal atomization

atomic spectrometry (ETAAS) (5,17), potentiometric stripping analysis (18), polarography (19),

inductively coupled plasma optical emission spectrometry (ICP-OES) (20,21) and inductively

coupled plasma mass spectrometry (ICP-MS) (12,22,23).

The conventional thermal program of ETAAS technique usually includes the following

steps: drying, pyrolysis, atomization, and cleanout. The pyrolysis step is used to remove as many

matrix components as possible. A matrix modifier can be used to stabilize the analyte or aid in

removing matrix components (24,25,26).

The most time consuming steps of thermal programs are drying and pyrolysis. Analysts

have long thought some means to reduce or eliminate the time required for these pre-treatment

Page 60: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

51

stages. The combination of two techniques, Zeeman-effect background correction and Stabilized

Thermal Platform Furnace (STPF), was the key to providing faster analyses.

In fast furnace analysis, the pyrolysis step and matrix modification are usually eliminated.

Using these procedures, considerable time saving is achieved without sacrificing analytical

precision or accuracy. This technique may not be compatible with all instrumentation (the

instrumentation used must provide Zeeman background correction) and with all sample types. Very

complex matrices may still require at least a short pyrolysis step and the use of a matrix modifier

for optimum results (25,27).

The main aim of the present work was to compare conventional and fast thermal programs,

with and without matrix modifiers, in order to improve the analytical performance of Cu

determination in wine, in terms of higher throughput and lower cost, without sacrificing analytical

sensitivity, precision, or accuracy. In addition, a method using the selected thermal program was

validated.

2.2.2. EXPERIMENTAL

2.2.2.1. Instrumentation and analytical conditions

The ETAAS instrument used was a PerkinElmer model 4110 ZL graphite furnace atomic

absorption spectrometer (PerkinElmer Life and Analytical Sciences, Shelton, CT, USA), using

Zeeman-effect background correction with a model AS-72 autosampler, and the PerkinElmer

AAWinLabTM software, version 2.5. Argon N50 (purity > 99.999%) was used to protect and purge

the graphite tubes, with an internal flow rate of 250 mL.min-1. A PerkinElmer LuminaTM hollow

cathode lamp was used. The spectrometer settings for pyrolytically coated graphite tubes with end

caps and L’vov platforms are given in Table 1. The measurement mode was integrated absorbance.

The auto sampler was programmed to pipette 20 µL of the sample (Cu standard solution or wine)

and 5 µL of each matrix modifier onto the platform.

ICP-MS measurements were carried out with an ELAN 9000 ICP-MS (PerkinElmer

SCIEX, Concord, Ontario, Canada), equipped with a crossflow nebulizer, a Scott-type spray

chamber made of Ryton material, nickel cones, and a peristaltic sample delivery pump. A

PerkinElmer autosampler AS-93Plus, protected by a laminar-flow-laminar clean room class 100,

(Reinraumtechnik Max Petek, Germany) was used. The water was deionized with a Seralpur Pro 90

CN purifier (Seral, Ransbach-Baumbach, Germany).

Page 61: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

52

Table 1 - Furnace thermal programs

Step Temperature (ºC) Ramp time (s) Hold time (s)

Dry 1 110 1 30 Dry 2 140 10 25 Pyrolysisa 1100 10 30 Atomizationb 2300 0 5 Clean-out 2450 2 5

Wavelength (λ) 324.8 nm Slit width 0.7 nm Hollow cathode lamp current 15 mA Background correction Zeeman-effect Integration time 5 s Injection volume 20 µL Total time 118 s (conventional thermal program)

78 s (fast thermal program) a step eliminated at fast thermal program; b stop argon flow.

2.2.2.2. Materials

To eliminate possible contamination, all glassware and polyethylene material (volumetric

flasks, micropipette tips and auto sampler cups) were immersed for 24 hours in freshly prepared

20% (v/v) HNO3 and then rinsed thoroughly with deionized water before use. An exhaustive

cleaning of the glassware and the use of plastic materials are very important. For ICP-MS

determinations, only polyethylene material was used.

2.2.2.3. Reagents and calibration

Standard solutions (10, 20, 30, 40, and 50 µg.L-1) were prepared daily from a 1000 mg.L-1

solution (CertiPUR, Merck) of Cu in HNO3 (0.2% v/v) and C2H6O (2% v/v) with deionized water

(conductivity<0.1 µS.cm-1). Palladium nitrate and magnesium nitrate, both from Merck, were tested

as chemical modifiers in standard solution and wine samples [5 µg Pd(NO3)2 and 3 µg Mg(NO3)2].

The wine samples were diluted (1:5) with HNO3 (0.2% v/v). Immediately after preparation, the

standard and sample solutions were transferred to polyethylene material. Standards and samples

were analysed in triplicate. For ICP-MS determinations a multi-element standard solution of 10

µg.L-1 (PerkinElmer) was used for external calibration; Rh and Re were used as the internal

standards in 10 µg.L-1 concentrations (Merck) and ultrapure HNO3 (J.T. Baker).

Page 62: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

53

2.2.3. RESULTS AND DISCUSSION

2.2.3.1. Thermal programs

Different studies involving conventional and fast thermal programs with and without

matrix modifiers were developed. The optimized thermal program was directed not only at

obtaining a more sensitive and reproducible method, but also for optimizing the respective costs

and analytical time in order to allow its routine use. This work was carried out with a solution of

Cu at a concentration of about 20 µg.L-1 in parallel with diluted samples of a white and a red wine.

The dry step was established by the visual control of the sample behaviour during the

heating stage and by the consistency of instrumental output. A two-stage procedure for the dry step

proved convenient. During the first stage (110ºC), the alcohol volatilizes without spattering and the

drying of the sample is completed during the second stage (140ºC). For the pyrolysis and

atomization steps, the variation of the integrated absorbance with temperature was studied for the

determination of Cu. The influence of different temperatures, ramp, and hold times were also

assessed.

The results using the conventional program with and without matrix modifiers are shown in

Figures 1A and 1B, respectively. The pyrolysis curves (atomization temperature 2000ºC) show that

the combination of Pd(NO3)2 and Mg(NO3)2 matrix modifiers was able to thermally stabilize Cu up

to 1100 ºC. Although high absorbance signals were obtained at 2000ºC, with and without matrix

modifiers, it was observed that some wines presented unsatisfactory absorption profiles: the analyte

peak shows a “tail” (no return to baseline). The increment of the temperature did produce

appreciable effects; at 2300ºC the absorption profiles were acceptable. The use of matrix modifiers

led to a higher absorbance signal (about 20% higher), except for the standard solution.

An additional study was developed varying the amount of both matrix modifiers deposited

in the graphite tube. Better results were obtained with the initial conditions: 5 µg Pd(NO3)2 and 3

µg Mg(NO3)2.

The results using the fast program with and without matrix modifiers are shown in Figures

2A and 2B, respectively. For both modalities, the optimum atomization temperature is 2100ºC. The

white wine shows a particular behaviour: when using matrix modifiers, the atomization

temperature could be 2000ºC. However, at 2300ºC the absorption profiles were preferable.

Page 63: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

54

0

0,02

0,04

0,06

0,08

0,1

0,12

500 1000 1500 2000 2500

Temperature (ºC)

Abso

rban

ce

Cu standard solution

white wine

red wine

Figure 1A - Conventional thermal program: pyrolysis and atomization curves in white and red wines (diluted 1:5) and Cu standard solution (20 µg.L-1), with matrix modifiers [5 µg Pd(NO3)2 and 3 µg Mg(NO3)2].

0

0,02

0,04

0,06

0,08

0,1

0,12

500 700 900 1100 1300 1500 1700 1900 2100 2300 2500

Temperature (ºC)

Abso

rban

ce

Cu standard solutionwithe winered wine

Figure 1B – Conventional thermal program: pyrolysis and atomization curves in white and red wines (diluted 1:5) and Cu standard solution (20 µg.L-1), without matrix modifiers.

Page 64: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

55

0,00

0,02

0,04

0,06

0,08

0,10

0,12

1400 1600 1800 2000 2200 2400

Temperature (ºC)

Abs

orba

nce

Cu standard solutionwhite winered wine

Figure 2A - Fast thermal program: pyrolysis and atomization curves in white and red wines (diluted 1:5) and Cu standard solution (20 µg.L-1), with matrix modifiers [5 µg Pd(NO3)2 and 3 µg Mg(NO3)2].

Figure 2B - Fast thermal program: pyrolysis and atomization curves in samples in white and red wines (diluted 1:5) and Cu standard solution (20 µg.L-1), without matrix modifiers.

The use of matrix modifiers did not lead to a significant variation in the absorbance signal.

The overall optimized furnace programs are summarized in Table 1.

In conclusion, fast and conventional programs (with or without matrix modifiers) can be

used for determining Cu in wines. However, the conventional program with matrix modifiers

allowed a higher sensitivity and consequently it was the program selected for the method

validation.

0,00

0,02

0,04

0,06

0,08

0,10

0,12

1400 1600 1800 2000 2200 2400

Temperature (ºC)

Abs

orba

nce

Cu standard solution

w hite w ine

red w ine

Page 65: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

56

2.2.3.2. Validation

2.2.3.2.1. Algorithm calibration curve, selectivity and accuracy

Calibration against acidified standard solutions was carried out. Two white wines and two

red wines were spiked with four concentration levels of Cu, never exceeding the total concentration

of 50 µg.L-1. The linearity (y=a+bx) of the response curves was investigated.

The results presented in Table 2 confirm that ∆b/b coefficients are lower than 5%, meaning

that curves are satisfactory. The variance homogeneity was investigated and confirmed by the

Cochran test. For each curve, a high percentage of the total variance is explained by the regression,

which has a calculated F (F1) higher than the Fisher’s F at 99.5%. The residual variance due to the

adjustment error is significant in all cases (F2), except for the Red Wine 2. Although the residual

variability is very low, since the pure error and adjustment error are not of the same magnitude, the

test is significant. This may result from the small variability within each group because the

replicates are not independent.

Table 2 - Calibration and wines response curves, characteristic parameters (Variance analysis of the linear regressions)

Calibration Curve White Wine 1 White Wine 2 Red Wine 1 Red Wine 2

y = a + bx y = (0.007 ± 0.003) +

(0.0078 ± 0.0001)x

y = (16.7 ± 0,.8) +

(1.02 ± 0.03)x

y = (32.5 ± 0.5) +

(1.02 ± 0.04)x

y = (10.6 ± 0.3) +

(0.95 ± 0.01)x

y = (27.9 ± 0.3) +

(1.02 ± 0.02)x

r2 0.999 0.997 0.995 0.999 0.999

∆b/b 1.5% 3.2% 4.3% 1.5% 2.1%

Cochran (C)

C(0.05; 6,2) = 0.62

C(0.05; 5,2) = 0.68

0.44 (ns)

0.47 (ns)

0.31 (ns)

0.43 (ns)

0.53 (ns)

(F1)

F(0.005; 1,16) =10.6

F(0.005; 1,13) =11.4

18864.9 (s)

4445.5 (s)

2563.2 (s)

21419.1 (s)

10460.7 (s)

(F2)

F(0.005; 4,12) =6.52

F(0.005; 3,10) =8.08

110.2 (s)

46.9 (s)

12.1 (s)

21.9 (s)

5.8 (ns)

(s): significant difference; (ns): without significant difference; the calculate values should be lower than theoretic values, except for F1.

Page 66: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

57

The slope of the curves plotted by wine standard additions (wine Cu concentration as

function of Cu addition) are not different of 1, except for Red Wine 1, indicating that there are no

matrix effects (Figure 3). Thus, the method presents a very satisfactory selectivity.

Accuracy was determined by the technique of standard additions to determine recovery of

spiked analyte. This approach was used because we could not find reference materials for this

parameter. Spiking wine samples with aqueous Cu gave recoveries between 93% and 100%:

94±7%; 97±6%; 93±3%; 100±2%, for White Wine 1, White Wine 2, Red Wine 1 and Red Wine 2,

respectively. The accuracy of the method is satisfactory and in agreement with the required quality

of the analytical results.

One white wine and one red wine were also analyzed by ICP-MS (semi-quantitative

method). The analytical results obtained using the semi-quantitative mode of ICP-MS technique

presented an accuracy of ±20%. Thus, the results obtained by the two techniques are comparable

(differences lower than 15%).

Figure 3 – Curves obtained for white and red wines by standard additions method.

0

10

20

30

40

50

60

0 5 10 15 20 25 30 35 40 45

Cu addition (µg.L-1)

Cu

conc

entr

atio

n (

g L-1

)

w hite w ine 1w hite w ine 2red w ine 1red w ine 2Linear (w hite w ine 2)Linear (red w ine 2)Linear (w hite w ine 1)Linear (red w ine 1)

y = (32.5± 0.5) + (1.02 ± 0.04)x

y = (27.9 ± 0.3) + (1.02± 0.02)x

y = (16.7 ± 0.8) + (1.02 ± 0.03)x

y = (10.6 ± 0.3) + (0.95 ± 0.01)x

Page 67: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

58

2.2.3.2.2. Analytical limits

The detection limit was calculated as the mean concentration, plus three standard deviation

of the blank, obtained with 20 determinations. The quantification limit was calculated as the mean

concentration, plus ten standard deviation of the blank. The experimental conditions used enabled

to detect 1 µg.L-1 and to quantify 3 µg.L-1. The quantification limit permitted the determination of

Cu in undiluted wines with 15 µg.L-1. In fact, the Cu concentration of wines is generally higher

than 15 µg.L-1.

2.2.3.2.3. Precision

Replicates of four white wines and four red wines, at several concentration levels, gave relative

standards deviations between 1% and 3% (n = 10), indicating that the analytical method is of high

within-run precision (repeatability). The relative standards deviations of between-run precision

(reproducibility) are from 5% to 8% (N = 3). The higher values were found in wines with lower

concentrations.

2.2.4. CONCLUSION

The method presented for direct analysis by ETAAS can be used to measure the copper

content in wines on a routine basis, especially in samples containing very low concentrations of this

metal. The characteristics and figures of merit of the method described (practicability, reduced risk

of sample contamination, sensitivity, selectivity, analytical limits, accuracy, and precision) make it

useful, in contrast to flame AAS which inevitably requires sample pre-concentration or the use of

method of standard additions. Both fast and conventional thermal programs (either with or without

matrix modifiers) can be used; fast program allowed to a higher throughput; however, the

conventional program with matrix modifiers allowed a higher sensitivity and was consequently

validated.

2.2.5. REFERENCES 1. A.S. Curvelo-Garcia, Controlo de Qualidade dos Vinhos, Química Enológica, Métodos Analíticos, 420

p, Instituto da Vinha e do Vinho, Lisboa (1988).

2. A.S. Curvelo-Garcia and S. Catarino, Ciência e Técnica Vitivinícola, 13 (1-2), 49 (1998).

3. J. Ribéreau-Gayon, E. Peynaud, P. Sudraud, and P. Ribéreau-Gayon, Sciences et Techniques du Vin. Tome I – Analyse et contrôle des vins, 645 p., Dunod, Paris (1982).

4. M.T. Vidal, P. Poblet Montserrat, M. Constanti, and A. Bordons, American Journal of Enology and Viticulture, 52 (3), 223 (2001).

5. M.T. Vasconcelos and M. Azenha, Feuillet Vert de l’OIV, 1135 (2001).

Page 68: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

59

6. M.T. Vidal, M. Constanti, and A. Bordons, Vignevini Bologna, 26 (7-8), 50 (1999).

7. J-P. Quinche, Revue Suisse de Viticulture, Arboriculture, Horticulture, Nyon, 17, 341 (1985).

8. H. Lay and W. Lieb, Wein-Wissenschaft, 43, 107 (1988).

9. C.S. Ough and M.A. Amerine, Methods for Analysis of Musts and Wines, 377p, John Wiley & Sons, New York (1988).

10. H. Lay, Deutsches Weinbau-Jahrbuch, 44, 255 (1993).

11. P. Sudraud, B. Médina, and J.P. Grenon, Feuillet Vert de l’OIV, 984 (1995).

12. I. Rodushkin, F. Odman, and P.K. Appelblad, Journal of Food Composition and Analysis, 12, 243 (1999).

13. OIV, Recueil des Méthodes Internationales d’Analyse des Vins, Office International de la Vigne et du Vin, Paris (1990).

14. M. Netzer and F. Bandion, Feuillet Vert de l’OIV, 1024 (1996).

15. S. Catarino, D. Pinto, and A.S. Curvelo-Garcia, Ciência e Técnica Vitivinícola, 18 (2), 65 (2003).

16. CT83, Norma Portuguesa NP 2442, Instituto Português da Qualidade, Lisboa (1988).

17. A.A. Almeida, M.I. Cardoso, and J.F.C. Lima, Atomic Spectroscopy, 2, 73 (1994).

18. A.M. Green, A.C. Clark, and G.R. Scollary, Fresenius’ Journal of Analytical Chemistry, 358 (6), 711 (1997).

19. J.B. Fournier, M. El Hourch, and G.J. Martin, Journal International des Sciences de la Vigne et du Vin, 32 (1), 45 (1998).

20. G. Thiel and K. Danzer, Fresenius’ Journal of Analytical Chemistry, 357 (5), 553 (1997).

21. G. Nicoli, R. Larcher, P. Pangrazzi, and L. Bontempo, Vitis, 43 (1), 41 (2004).

22. J.P. Greenough, H.P. Longerich, and S.E. Jackson, Australian Journal of Grape and Wine Research, 3 (2), 75 (1997).

23. S. Catarino, J. Soares, A.S. Curvelo-Garcia, and R. Bruno de Sousa, Ciência e Técnica Vitivinícola, 19 (1), 29 (2004).

24. PerkinElmer, The THGA Graphite Furnace, Techniques and Recommended Conditions, PerkinElmer (1995).

25. R. Beaty and J.D. Kerber, Concepts, Instrumentation and Techniques in Atomic Absorption Spectrometry, PerkinElmer (1993).

26. G. Schlemmer and B. Radziuk, Analytical Graphite Absorption Spectrometry – A Laboratory Guide, Birkhauser Verlag, Basel (1999).

27. S. Catarino, J.L.Capelo, A.S. Curvelo-Garcia, and M. Vaião, Journal of AOAC International (in press).

Page 69: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

61

2.3. FOCUSED ULTRASOUND (FU) VERSUS MICROWAVE DIGESTION FOR THE DETERMINATION

OF LEAD IN MUST BY ELECTROTHERMAL ATOMIC ABSORPTION SPECTROMETRY

J.L. Capelo1

Centro de Química Estrutural, Instituto Superior Técnico de Lisboa, Avda Rovisco Pais, s/n, 1049-001, Lisboa, Portugal

S. Catarino1, A. S. Curvelo-Garcia

INIAP, Estação Vitivinícola Nacional, 2565-191, Dois Portos, Portugal

M. Vaião Laboratório de Análises do IST, Instituto Superior Técnico de Lisboa, Avda Rovisco Pais, s/n, 1049-001, Lisboa, Portugal

Abstract

The content of metals at the different stages of the winemaking process is of great concern because of legal and wine quality reasons. In the present work, we have developed a new fast procedure for metal extraction from must based on focused ultrasound with HNO3 and H2O2, with which microwave digestion of the must is not necessary. Lead was used as a model element to achieve the optimum combination of reagents in order to achieve total extraction. Results were compared with those obtained after microwave digestion of the samples. In addition, conventional and fast programs in conjunction with different matrix modifiers were studied. The method requires few sample mass (ca 0.5 g) and few reagents in low volume (HNO3 50 µL + H2O2 100 µL). The total sample treatment time is only 60 s.

Journal of AOAC International (2005) 88 (2), 585-591

Page 70: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

62

2.3.1. INTRODUCTION

The content of metals at the different stages of the winemaking process is of great concern

because of legal and wine quality reasons. Maximum acceptable values for some toxic metals in

wine are defined by wine regulations in the European Union (EU) and in the United States (USA).

In addition, must quality, and subsequently the wines, depends on the metallic elements present and

on their concentration (1,2). Must is one of the most important stages of the winemaking process

because, once fermented, it will give the wine (must is formed when the grape pomace has been

removed).

Must is a complex substance with a high organic matrix content, mainly sugars and large

agglomerations of organic matter. Hence, a sample pre-treatment is required in order to destroy the

organic matter that could interfere with the analytical determination by electrothermal-atomic

absorption spectrometry (ET-AAS) or inductively coupled plasma-mass spectrometry (ICP-MS),

which are widely used in the analysis of the metal content of many products with high organic

content, such as must (3-6). Early sample pretreatments described in literature for metal

determination in must included conventional wet or dry ashing (7,8).

Today, microwave digestion offers better performance in the destruction of organic matter

in samples with high sugar content (3), compared to conventional approaches. Many applications

of microwave energy assisted-acid digestion procedures using both domestic or special constructed

microwave ovens can be found in literature, as reviewed by Chakraborty et al. for ET-AAS (9).

These procedures completely destroy the organic matter content of the sample by using

concentrated acids, and the subsequent solution may be used for both ET-AAS and ICP-MS.

New trends in sample pretreatments based on advanced oxidation processes (AOPs) are

emerging in routine analysis (10). The AOPs are characterized by the in situ generation of highly

potent chemical oxidants, such as the hydroxyl radical (OH). This alternative methodology avoids

the complete sample digestion and, when working with ET-AAS, offers as an advantage the

possibility of using aqueous calibration, because matrix effects are not important. These procedures

involve the breakdown of the chemical bonds between the trace elements and matrix sample

constituents. In addition, naturally occurring sequestering agents are destroyed or inactivated.

Furthermore, the requisite equipment for AOP application is inexpensive and simple to assemble.

Focused ultrasound is an AOP that, in conjunction with different chemicals, has demonstrated to be

an elegant way to overcome many problems dealing with the use of microwave digestion, such as

long manipulation time, the use of concentrated acids, or safety problems (3,11-15). To the best of

our knowledge, no attempts to use focused ultrasound have been cited in literature as sample

treatment for the determination of metals in must.

The main aim of the present work was to assess the feasibility of probe sonication in

conjunction with different chemicals for the determination of lead in must, as a fast alternative to

microwave digestion for sample pretreatment. The study was conducted by monitoring the levels of

Page 71: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

63

lead, as a model element in must. In addition, different chemicals modifiers were studied in order

to improve the analytical performance of the lead ET-AAS determination in terms of suppressing

the interferences associated with complex sample matrixes and to enhance Pb absorption signal.

Results were validated against spiked samples, by comparing probe sonication with microwave

digestion, and by analysing the samples by ET-AAS and ICP-MS. In addition, 3 different

laboratories reported results for the same sample after microwave digestion and probe sonication.

2.3.2. EXPERIMENTAL

2.3.2.1. Apparatus

(a) Microwave oven - MLS 1200 mega (Milestone Inc., Shelton, CT); used for microwave

digestion.

(b) Ultrasonic cell disruptor-homogenizer - Branson Sonifier 150, 63 W, 22.5 kHz,

(Branson Ultrasonics Corp., Danbury, CT); equipped with a 3 mm titanium microtip. Ultrasonic

energy irradiation was fixed at 10% level with the 3 mm microtip. The Sonifier 150 has a digital

LCD display that provides a continuous readout of the watts delivered to the end of the probe

(range 5-6 W in this work).

(c) Eppendorf cups - 2 mL.

(d) Centrifuge - Wifug (London, UK) labor 50M.

(e) Atomic absorption instruments - Atomic absorption measurements [at the Estação

Vitivinícola Nacional, Instituto Nacional de Investigação Agrária e das Pescas (INIAP), and

Laboratório de Análises do Instituto Superior Técnico (LAIST) laboratories] were performed with

a Perkin-Elmer Model 4110 ZL graphite furnace atomic absorption spectrometer (Perkin-Elmer,

Norwalk, CT), equipped with Zeeman-effect background correction, Model AS-72 autosampler,

and a transverse heated graphite atomizer (THGA). A Varian (Cambridge, UK) atomic absorption

spectrometer Model SpectrAA-300 Plus equipped with a graphite furnace, an autosampler, and a

Zeeman background correction was used at the Instituto Superior Técnico (IST). Pyrolytic

graphite-coated graphite tubes were used with the L´Vov platform. The electrothermal parameters

and thermal programmes are presented in Table 1.

(f) ICP-MS instruments – ICP-MS measurements were performed at the INIAP with a

Perkin-Elmer Model ELAN 9000 equipped with a crossflow nebulizer, a Scott-type spray chamber

made of Ryton, and nickel cones. A peristaltic sample delivery pump (a 4-channel Model Gilson)

was used. The Perkin-Elmer (AS-93 Plus) autosampler was protected by a laminar-flow-chamber

clean room, Class 100 (Max Petek, Radolfzell, Germany). The application software was ELAN-

6100/Windows NT (Version 2.4). The operational parameters of the ICP-MS are presented in

Table 2.

Page 72: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

64

Table 1- Instrumental parameters and the thermal programs for lead used with the Perkin-Elmer and Varian atomic absorption spectrometers

Instrumental parameters Perkin-Elmer INIAP (LAIST) Varian

Background correction Zeeman effect Zeeman effect

Lamp current, mA 12 10

Atomizer type Transversely heated graphite tubes Pyrolytic graphite-coated graphite

Wavelength, nm 283.3 283.3

Spectral bandpass, nm 0.7 0.7

Thermal programs

Stage Temperature(ºC)/time(s)/ramp(ºC/s) b

Perkin-Elmer instrument INIAPc LAIST

Dry 1 110/30/10 110/50/5

Dry 2 140/30/15 130/30/5

Ashing 900/20/10 900/20/10

Atomization 1600/5/0 1500/4/0

Cleaning 2500/3/1 2400/2/1

Varian instrumentd IST Dry 1 85/5/17 Dry 2 95/40/0.25 Dry 3 120/40/0.63 Ashing 1 850/40/18.2 Ashing 2 850/60/5 Atomization 2250/5.1/0 Cleaning 2600/2.0/75

aLAIST = Laboratório de Análises do Instituto Superior Técnico; INIAP = Instituto Nacional de Investigação Agrária e das Pescas; IST = Instituto Superior Técnico; bThe ramp is given in seconds for the Perkin-Elmer instrument; cThe purge gas flow rate (300 mL/min) was stopped; fast program: (Perkin-Elmer instrument) Dry 2 and ashing stages were replaced by a single drying stage at 400ºC, time 20 s, ramp 15 s; dTime needed to achieve the corresponding temperature with the indicated ramp.

2.3.2.2. Reagents

Special care was taken in order to choose the highest pure reagents available in the market.

Milli-Q ultrapure water was used throughout.

(a) KMnO4 (No. 105084), K2S2O8 (No. 5092), and HNO3 (N0. 30709) - Merck (Darmstadt,

Germany).

(b) HCl ACS (No. 30721) - Riedel-de Häen (Seelze, Germany).

(c) H2O2 (No. 31642) - Sigma-Aldrich (St. Louis, MO).

(d) Palladium nitrate atomic absorption modifier solution – Perkin-Elmer (No.

BO190635).

(e) Stock standard solutions - Stored in a refrigerator at 4ºC and protected from light.

Page 73: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

65

(f) Working standard solutions - Prepared just before use by appropriate dilution of the

stock standard solution.

For ICP-MS determinations:

(g) Multi-element standard solution - Perkin-Elmer; used for external calibration (10 µg/L

Ag, Al, As, Ba, Be, Bi, Cd, Co, Cr, Cs, Cu, Fe, Ga, In, Li, Mn, Ni, Pb, Rb, Se, Sr, Tl, U, V, and

Zn).

(h) Rh and Re - Used as internal standards in 10 µg/L concentrations (Merck)

(i) HNO3 - Ultrapure (Mallinckrodt Baker, Phillipsburg, NJ).

(j) Deionized water - Obtained by using a Seralpur Pro 90 CN (Seral, Ranbach-Baumbach,

Germany); used throughout.

Table 2 - Operating conditions and mass spectrometer settings of the ELAN 9000 ICP-MS semi-quantitative mode of analysis

ICP RF power, W 1200 Nebulizer gas flow rate, L/min 0.91

AutoLens DAC values: 6.0 (Be); 6.5 (Co); 7.5 (In); 9.0 (U)

Sample uptake flow, mL/min 0.85

MS acquisition settings

Dwell time, ms 50

Sweeps/reading 6

Number of replicates 1

Time per run, s 65

Isotopes monitored Range from 6 to 240

Scan mode Peak hopping

MCA channels for peak 1

Detector Dual (pulse and analogic)

Dead time, ns 60

2.3.2.3. Sample collection

Exogenous contamination was avoided by cleaning all the plastic bottles used for sample

collection with 10% HNO3. The bottles were then rinsed gently with ultrapure water and dried at

room temperature. The must samples were obtained immediately after the grapes were pressed at

the INIAP, and 5 mL of HNO3 was added to a sample volume of ca 0.5 L. The samples were stored

at -10ºC until they were analized.

Page 74: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

66

2.3.2.4. Sample treatment

Procedure 1: Focused ultrasound - To 0.5 mL must sample placed in eppendorf cups were

added the following reagents: (1) HNO3 + H2O2 (50 µL concentrated + 100 µL concentrated), (2)

KMnO4 + HCl (150 µL 3 × 10-3 M + 150 µL 2 M), (3) K2S2O8 (150 µL 7 x 10-2 M), and (4) HNO3

(50 µL concentrated). The reagents, their mixtures and the quantities added were chosen based on

their recognised oxidizing properties and on our own experience (12,13,15). Then, focused

ultrasound was applied for 60 s at 10% amplitude (5-6 watts delivered at the end of the 3 mm

probe). Finally, the sample was centrifuged for 5 min at 5000 rpm, and the supernatant was used

for measurements.

Procedure 2: Microwave digestion - To 2.5 mL must sample placed in the microwave

reactor was added 2 mL concentrated HNO3 and 0.5 mL 30% H2O2. Once the reactor was cupped,

it was placed in the microwave oven and the following program was run: (1) 1 min at 250 W; (2) 3

min at 0 W; (3) 5 min at 250 W; (4) 5 min at 400 W; (5) 5 min at 600 W. After cooling to ambient

temperature, the reactor was opened and the resultant solution was transferred into a 10 mL

calibrated flask. The reactor was carefully cleaned with 2 mL Milli-Q water using a Pasteur pipette,

and the water was added to the calibration flask. Finally, the solution was made up to volume with

Milli-Q water.

2.3.3. RESULTS AND DISCUSSION

2.3.3.1. Thermal program

The thermal program was optimized at the INIAP for determination of lead in sample

solutions and aqueous standard. Different studies involving conventional and fast programs along

with different matrix modifiers were performed.

Pd was chosen as matrix modifier because of its recognised thermal stabilization

characteristics for lead. The KMnO4 and the combination KMnO4 plus HCl were chosen for

different reasons. KMnO4 has been recently cited as the matrix modifier in the determination by

ET-AAS (15) and electrothermal vaporization (ETV)-ICP-MS (16) of highly volatile Hg. KMnO4

+ HCl has been successfully used as a powerful oxidant for organic matter (human urine;12,15) in

conjunction with focused ultrasound. Nevertheless, the presence of the chloride ion in the

determination of lead by ET-AAS has been reported to lead to the formation of volatile ClPb

complexes (17). To the best of our knowledge, no data have been reported dealing with KMnO4 +

HCl as a matrix modifier for lead determination in ET-AAS.

Results using the conventional program after must pre-treatment with probe sonication and

HNO3 + H2O2 are shown in Figure 1A.

Page 75: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

67

Figure 1 - Ashing and atomization curves. Chemical modifiers: ο, 1:1 (v/v) KMnO4 (0.5 g/L)/HCl (2 M); □, Pd(NO3)2 , 200 µg/g; ∆, KMnO4 (0.5 g/L); ◊, no modifier. Absorbances obtained after ultrasound treatment with HNO3 + H2O2 in must with (A) conventional program; (B) fast program. Absorbances obtained after ultrasound treatment in must with (C) KMnO4 + HCl. Absorbances obtained with aqueous standards: (D) conventional program; (E) fast program.

Page 76: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

68

The pyrolysis curves (atomization temperature, 1800ºC) show that there is no difference in

the thermal stabilisation of Pb caused by either KMnO4 or the combination KMnO4 plus HCl. Both

chemicals thermally stabilised the Pb up to 500ºC, hence the presence of HCl did not have

noticeable effects. The Pd matrix modifier was able to thermally stabilise Pb up to 1100ºC. Without

the modifier, Pb losses by volatilization began at 350ºC (Figure 1A). Shapes of the absorption

profiles were acceptable from 500ºC for Pd, from 400ºC for KMnO4, and from 500ºC for KMnO4 +

HCl. Lower calcinations temperatures led to unacceptable absorption profiles (e.g. irreproducible

shapes and double peaks).

Figure 1A also shows the atomization curves obtained in the range 1100º-1900ºC

(calcination temperature: Pd, 900ºC; KMnO4, 300ºC; KMnO4 + HCl, 400ºC, no modifier, 400ºC).

Neither KMnO4 nor KMnO4 + HCl was able to stabilise the Pb and, at 1100ºC atomization

temperature, the absorptions were 75 and 58% of the expected signal, respectively. Finally, and

additional study was performed by varying the amount of Pd deposited in the tube in the range 1-6

µg. No significant variation was observed in the Pb signal (n=3, t-test, P = 0.05, data not shown).

One of the drawbacks of conventional determinations by ET-AAS is the time required to

produce a result. Three approaches (18) have been tried to reduce the total analysis time:

minimization of the time in conventional drying, injection of the sample onto a preheated tube, and

high-temperature drying. Our attempts were focused on reducing the total time per cycle by

reducing the time in the drying step along with an increase in the drying temperature, and

suppressing the ashing stage.

Figure 1B shows the results for the fast program developed (see Table 1 for details). The

worst results were obtained in the presence of HCl. In this case, the standard deviation (SD),

background signal, and peak shape were unacceptable. This result is in contrast with the one

previously cited for the conventional program, in which no effect was observed when the chloride

was present.

Although high background signals were obtained with the fast program, either with Pd or

KMnO4, the correction provided by the Zeeman effect was enough to guarantee reasonable results,

with relative standard deviation (RSD) values around 10%.

Results using the conventional program for sample pretreatment with KMnO4 + HCl and

probe sonication are shown in Figure 1C. As can be seen, the trends are similar to those obtained

for probe sonication with HNO3 + H2O2. The increment in the quantities of KMnO4 and HCl

present in the graphite tube, as a result of the utilization of both reagents in the sample

pretreatment, did not produce appreciable effects in the absorption profiles or in the absorbance

signal but, when compared with the pretreatment with HNO3 + H2O2, backgrounds signals were at

least twice as high. A recovery test was performed using spiked samples treated with the

combination of KMnO4 + HCl, with Pd as the matrix modifier (calcination temperature, 700ºC and

Page 77: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

69

atomization temperature, 1500ºC); recoveries were 104 ± 5% (n=3) and 102.4 ± 6% (n=3) for

concentrations of ca 20 and 30 ppb, respectively.

The trends showen by the Pb aqueous standards when KMnO4 and KMnO4 + HCl were

studied as potential modifiers were different from those observed when the same reagents were

used in the presence of sample matrix. Thus, when working with aqueous standards and the

conventional program, no lead stabilization was observed (Figure 1 D) with HCl + KMnO4 or

KMnO4 alone because Pb losses began at 600ºC, the same temperature at which Pb losses were

noted when no modification was used. The same was later observed with the fast program, with

which similar peaks profiles and absorbance signals were obtained, with or without modification.

In conclusion, fast (with Pd) and conventional programs (with or without modification) can

be used for determining Pb in must extracts when using HNO3 + H2O2 as the liquid medium.

However, the fast program allows a higher throughput and, consequently, was the program

selected.

2.3.3.2. Ultrasonic treatment

Because must has large agglomerations of organic matter, Eppendorf tips were modified in

order to make weighing of the sample as precise as possible. The tip was cut 1 cm above its end in

order to make the hole larger. The result for weighing eight 0.5 mL samples was 0.559±0.005 g

(RSD = 0.9%). Samples were stirred to ensure that they were homogeneous.

Previous research showed that quantitative recoveries of Pb could be obtained from

biological samples prepared with an acidic diluent in conjunction with ultrasonic processors, even

when large particle sizes were present in the slurry (19). This process can be attributed to acoustic

cavitation and the consequent disruptive action caused by ultrasound. Sonication with probes is

able to dissipate much higher powers than ultrasonic cleaner baths and is more convenient for

metal extraction in order to achieve quantitative recovery (20). Another important observation

concerning the effect of ultrasound was the formation of OH radicals yielding hydrogen peroxide

(21), which has oxidant properties when used with an acidic medium. This process might

contribute to the partial oxidation of the biological matrix, thus aiding metal extraction. The values

of the variables influencing extraction with the probe sonicator, i.e., sonication time, sonication

amplitude, and reagent concentrations, were chosen based on our previous experience with these

procedures (10,13). Microwave digestion of the must sample yielded a Pb concentration of 13.4 ±

2.4 ng/g for 4 different digestions. In addition, the same sample was analysed in the INIAP and IST

laboratories, and the data are reported in Table 3. Furthermore, results of INIAP analyses of the

digested samples by ICP-MS are given in Table 3.

Concerning ultrasonic extractions, Table 3 shows that the best results were obtained with

the mixture HNO3 + H2O2. Pb concentration obtained after sonication with the former reagents

Page 78: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

70

does not differ statistically from the values obtained after microwave digestion (t-test, P = 0.05).

Although the F-test reveals no difference between the RSD of both analysis (5 vs 11%), the lower

value achieved with the probe may be due to the few steps and reagent volumes involved with this

procedure, which increases precision. Sonication with KMnO4 + HCl or K2S2O8 gave lower

extraction efficiencies (66 and 73%, respectively) compared to the concentration value attained

after complete dissolution of the sample by microwave digestion. This poor performance was also

observed in the high RSD values obtained when using the above-mentioned reagents. Interestingly,

the use of sonication without any additional chemical (only the initial acidification with acid nitric,

when the sample was taken) led to Pb concentrations of the same order as the ones found with

microwave digestion (no significant difference by the t-test, P= 0.05). However, it should be

stressed that this result is due to the poor precision obtained when the extraction was performed

without the aid of any chemical (RSD = 18%).

Table 3 - Pb content in must (ng/g)

Sample treatment

Microwave

digestion (MW) HNO3 + H2O2

a KMnO4 + HCla K2S2O8a

Ultrasonic

extraction

(US)a

x ± ts/√n (n = 4) 13.4 ± 2.4b 13.6 ± 1.1b 8.9 ±1.7 9.8 ± 1.4 10.8 ± 3.1

13.6 ± 1.9c 13.5 ± 0.9c

14.1± 2.1d 14.3 ± 1.7d

RSD, % 11 5 11 9 18

Fexpe - 2.94 2.25 2.77 1.77

texpf - 0.24 5.01 4.13 2.08

ICP-MS analysisg: Pb, 14 ± 2; Al, 1.9 ± 0.2; V, 13.0 ± 0.2; Mn, 450 ± 50; Co, 2.0 ± 0.3; Ni, 92 ±

0.2; Zn, 54 ± 1; As, 4.0 ± 0.3; Cd, 3.0 ± 0.3 aResults with ultrasonication; bResults from IST; cResults from LAIS; dResults from INIAP; eF-test for comparing variances between ultrasonic procedures and microwave digestion; Ftab = 9.277 (for n1 = n2 = 4); ft-test for comparing concentration values between ultrasonic procedures and microwave digestion; ttab = 2.45 (for n = 6); gICP-MS analysis of the microwave digests (n = 4), ng/g (µg/g for Al).

2.3.3.3. Analytical figures of merit

Analytical figures of merit are presented in Table 4. Y is the integrated absorbance and Pb

is the concentration reported in pg. Characteristic masses for Pb are given based on integrated

absorbance. The limit of detection (LOD) was defined as 3 SD/m, where SD is the standard

deviation corresponding to 10 blank injections and m is the slope of the calibration graph. When

the analyte extraction efficiency is ca 100%, the between-batch precision is mainly determined by

the homogeneity of the whole mass used for preparation.

Page 79: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

71

Homogeneity of the must samples used in this work was assessed by measuring the Pb

content in 4 separate sample preparations with both probe sonication and microwave digestion. A

volume of 0.5 mL was chosen to perform the homogeneity assessment with the sonication probe,

whereas 2.5 mL was used with microwave digestion. As can be observed in Table 4, between-batch

precision was slightly higher in all cases than within-batch precision. When variances for between-

and within-batch measurements were compared, no significant differences were found for P = 0.05,

thus indicating good homogeneity at the 0.5 mL volume level.

Table 4 - Analytical figures of merit

Laboratory (ETAAS)

Figure of merit INIAP LAIST IST

Calibration grapha y = -10 × 10-5 + 1.5 ×

10-4 (Pb)

y = 16 × 10-4 + 0.86 ×

10-4 (Pb)

y = 4 × 10-4 + 2 × 10-4

(Pb)

Linear range, pg, up to 1000 1200 1500

LOQ, pg 60 58 35

LOD, pg 20 17.5 10.5

Characteristic mass, pg 29 51 22

Between-batch precision

(RSD, %, n = 4) 10.7b - 5c

Within-batch precision

(RSD, %, n = 4) 3.0b - 2.4c

Fexp 12.72 - 4.3d

Matrix modifier, µg Pd(NO3)2e

(NH4)H2PO4 +

Mg(NO3)2f

Pd(NO3)2g

a Pb in pg; b Microwave digestion; c Probe sonication with HNO3 + H2O2; d Ftab = 15.44 for n = 4, P = 0.05; e 3µg; f 250 pg

+ 75 pg; g 5 µg.

2.3.3.4. Determination of Pb in must samples

Analytical results for ultrasonic extraction of Pb from 3 different must samples are shown

in Table 5. Three replicates from each were performed. Optimum extraction conditions found for

Pb were used. Calibration with simple aqueous Pd standards was performed. In order to compare

the proposed extraction method, microwave digestion was performed in triplicate for each must

sample. No significant differences were observed for P = 0.05 when comparing the values obtained

by both methods (unpaired t-test).

Page 80: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

72

Table 5 - Analytical results obtained for Pb in musta

Sample MW valueb

(x ± ts/√n)

US valuec

(x ± ts/√n) Recovery, %

texp

(tteo = 2.72)

Must A 15 ± 2 14 ± 2 93 1.22

Must B 21 ± 5 23 ± 5 109 2.44

Must C 28 ± 5 27 ± 2 96 1.22 aAverage value (ng/g) ± confidence interval (n = 3) for P = 0.05; bMW = Microwave digestion; cUS = Ultrasonic extraction. Conditions: HNO3 + H2O2 (50 µL concentrated + 100 µL 30%, v/v), 30 s sonication time, 10% amplitude (4 – 5 W output).

2.3.4. CONCLUSIONS

Sonication with probe in conjunction with HNO3 and H2O2 is a useful sample treatment

procedure for the extraction of metals in the organic matter present in must. In conjunction with

HNO3 and H2O2, it was statistically demonstrated that lead concentration values obtained with

USLE did not differ from those provided by the common sample treatment method, microwave

digestion. The proposed methodology is fast, 60 s/sample, which allows high sample throughput.

Another advantage is the low reagent consumption, in the order of some hundreds of µL or below.

In addition, the low sample handling allows better performance in terms of reproducibility and

safety. The present work was focused on lead, but from a preliminary screening, it can be

anticipated that USLE could be also useful for other metals in must samples.

Acknowledgements José L. Capelo acknowledges the postdoctoral grant SFRH/BDP/9481/2002 of FCT (Science and

Technical Foundation) from Portugal. Sofia Catarino appreciates the financial support given by the Programa

de Apoio à Reforma das Instituições Públicas ou de Interesse Público de Investigação.

2.3.5. REFERENCES

(1) Curvelo-Garcia, A.S., Catarino, S. (1998) Ciência Téc. Vitiv. 13, 49-70.

(2) Curvelo-Garcia, A.S. (1988) Quality Control of Wines: Analytical methods, Instituto da Vinha e do Vinho, Lisbon, Portugal, p. 420.

(3) Garcia-Rey, R.M., Quiles-Zafra, R., Luque de Castro, M.D. (2003) Anal. Bioanal. Chem. 377, 316-321.

(4) Kristl, J., Veber, M., Slekovec, M. (2002) Anal. Bioanal. Chem. 373, 200-204.

(5) Catarino, S., Curvelo-Garcia, A.S., de Sousa, B. (2002) Atom. Spectros. 23, 196-200.

(6) Catarino, S. (2000) Levels of Lead and Cadmium in Wines. MSc Thesis, University of Porto, Porto, Portugal.

(7) Puig-Deu, M., Lamuela-Raventos, R.M., Buxaderas, S., Torre-Boronat, C. (1994) Am. J. Enol. Vitic. 45, 25-28.

Page 81: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

73

(8) Ough, C.S., Crowell, E.A., Benz, J. (1984) J. Food Sci. 47, 825-828.

(9) Chakraborty, R., Das, A.K., Cervera, M.L., de la Guardiã, M. (1996) Fresenius’ J. Anal. Chem. 355, 99-111.

(10) Capelo-Martínez, J.L., Ximenez-Embún, P., Madrid, Y., Cámara, C. (2004) Trends Anal. Chem. 23, 331-340.

(11) Capelo-Martínez, J.L., Ximenez-Embún, P., Madrid, Y., Cámara, C. (2004) Anal. Chem. 76, 233-237.

(12) Capelo, J.L., Maduro, C., Mota, A. (2004) J. Anal. At. Spectrom. 19, 414-416.

(13) Capelo, J.L., Maduro, C., Vilhena, C. (2005) Ultrason. Sonochem. 12, 225-232.

(14) Capelo, J.L., Lavilla, I., Bendicho, C. (2001) Anal. Chem. 73, 3733-3376.

(15) Capelo, J.L., dos Reis, C.D., Maduro, C., Mota, A. (2004) Talanta 64, 217-223.

(16) da Silva, A.F., Dias, L.F., Saint-Pierr, T.D., Curtius, A.J., Welz, B. (2003) J. Anal. At. Spectrom. 18, 344-347.

(17) Slavin, W. (1991) in Graphite furnace AAS: A Source Book, The Perkin-Elmer Corp., 2nd Ed., Norwalk, CT, pp 109-111.

(18) Halls, D.J. (1995) J. Anal. At. Spectrom. 10, 169-175.

(19) Amoedo, L., Capelo, J.L., Lavilla, I., Bendicho, C. (1999) J. Anal. At. Spectrom. 14, 1221-1226.

(20) Mamba, S., Kratochvil, B. (1995) Int. J. Environ. Anal. Chem. 60, 295-299.

(21) Mason, J.T., Lorimer, J.P. (1988) Sonochemistry: Theory, Applications and Uses of Ultrasound in Chemistry, Ellis Horwood, Chichester, UK, pp. 44-45.

Page 82: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

75

2.4. MEASUREMENTS OF CONTAMINANT ELEMENTS OF WINES BY INDUCTIVELY COUPLED

PLASMA MASS SPECTROMETRY: A COMPARISON OF TWO CALIBRATION APPROACHES

S. Catarino1,2, A.S. Curvelo-Garcia1, R. Bruno de Sousa2

1INIAP, Estação Vitivinícola Nacional, 2565-191 Dois Portos, Portugal 2Instituto Superior de Agronomia, Departamento de Química Agrícola e Ambiental, 1349-017 Lisboa, Portugal

Abstract

The aim of the present work was to develop and validate an accurate method by ICP-MS focalized to the measurement of contaminant elements in wines, in special those with legal importance. In addition, we intended to evaluate the suitability of ICP-MS semi-quantitative methodology in order to reduce the time and cost of analysis. Twenty-six contaminant elements of wine (Li, Be, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Ag, Cd, In, Cs, Ba, Hg, Tl, Pb, Bi and U) were measured using quantitative and semi-quantitative calibration approaches, in diluted white and red wines. In an early step potential interferences caused by Cl and Ca species were evaluated, in order to establish suitable mathematical corrections. For validation of ICP-MS procedures a few elements were determined by flame and electrothermal AAS. Reference wines from 1992 year, with provisional values, were analyzed and the results showed satisfactory agreement. The semi-quantitative calibration provided slightly higher limits of detection than those obtained by the quantitative calibration, and always lower than 0.1 µgl-1, except for Fe and Zn. For most elements the recovery percentages (between 90 and 100%) and precision of the results (R.S.D. (%) < 4%) were similar for both modes. Differences lower than 20% of concentration was obtained for most elements. Both methodologies offer valuable alternatives to wine characterization and comparison purposes. For legal requirements control purposes, with reference to the importance of accurate results, quantitative approach is the most suitable alternative. Keywords: wine; contaminant elements; ICP-MS; quantitative and semi-quantitative calibration

Talanta (2006) 70, 1073-1080

Page 83: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

76

2.4.1. INTRODUCTION

Mineral content of wines depend on several factors, including soil, variety of grape,

environmental conditions and viticultural and enological practices. The determination of some

elements is of interest due to their toxicological or physiological proprieties, while others can lead

to wine spoilage [1-5]. It should be noted that the levels of certain contaminant elements, such as

Cu, Zn, As, Cd and Pb, at different stages of the winemaking process are of great concern because

of legal requirements [6].

Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is a multi-element technique

applied to a wide range of applications in analytical chemistry, with high selectivity and sensitivity

and low analytical limits, so that it is an excellent tool for isotopic analysis [7-15] and for detailed

characterization of elemental composition of wine [16-27].

The application field of ICP-MS technique to wine is large. Different methodologies have

been used in preliminary studies to classify wines according to their geographic origin or grape

variety, from its trace composition, as a fingerprint [17,18,24,27,28]. A number of papers have

been published reporting the use of ICP-MS methods for determine trace metals and rare earth

elements in wine fingerprints [16,18,21-23,25,27,29]. In the last years this technique has been a

very important tool to evaluate the changes in the contents of trace elements through the

winemaking process. Nicolini et al. [25] have applied the ICP-MS technique to evaluate the

changes in the contents of micro and trace elements in wine due to winemaking treatments. The

changes of the metal composition in German white wines through the winemaking process were

evaluated by Gómez et al. [26].

ICP-MS offers different quantification procedures depending on the accuracy and precision

required. Isotope dilution mode of analysis presents the highest quality of results. However, when a

high quality of results is needed quantitative approach may be a suitable strategy as well. This

procedure requires external calibration with standards of each element to be determined. This

strategy is time consuming and it is not easy to have a complete set of the multi-element standards

required for the calibration. The application of this mode of quantification in wines has already

been reported [12,16-19,22,23,25,27,30].

The semi-quantitative approach is a versatile application of ICP-MS, which is claimed to

allow the determination of 81 elements with accuracy errors lower than 20% for most elements.

The semi-quantitative analysis software available for ICP-MS instrumentation (Total Quant III

from Perkin-Elmer – ELAN software, for instance) automatically corrects for isotopic interferences

and interfering molecular species and produces a comprehensive report, listing each element

present in the sample along with its concentration. Unlike quantitative analysis methods, calibration

is achieved using just a few elements distributed throughout the mass range of interest. The

calibration process is used to update internal response data that correlates measured ion intensities

to the concentrations of elements in a solution. During calibration, the response data is adjusted to

Page 84: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

77

account for changes in the instrument’s sensitivity due to variations in the sample matrix. This

methodology has already been applied to wine [15,18,20,21,25,29,30]. Pérez-Jordán et al. [19] and

Sagrado et al. [30] have compared two ICP-MS methods for analyzing wine samples (quantitative

and semi-quantitative methodologies). For the majority of elements determined by both

methodologies, comparable results were found. The advantages and limitations of the semi-

quantitative calibration of an inductively coupled plasma-mass spectrometer for multi-element

characterization of fortified and table wines were tested by Almeida et al. [22]. The wine samples

were submitted to an UV-irradiation treatment prior to analysis. Pruszkowski [31] have applied a

semi-quantitative methodology of ICP-MS technique to wine, supporting its suitability for final

sample characterization, such as total metal content and fingerprint.

Although the concentrations for only few elements in wines are currently under regulation,

the increasing concern for health effects and environmental exposure will probably result in a

longer list of potentially toxic elements, the concentrations of which have to be monitored in a near

future. As far as we know, ICP-MS methodologies are not yet used in routine wine analysis by

control laboratories. Nevertheless, the establishment of an ICP-MS methodology for contaminant

elements of wine will carry great advantages to these control laboratories.

The purpose of the present study was to develop and validate an accurate method by ICP-

MS focalized to the determination of contaminant elements in wines, in particular those which

concentration is, or could be in a near future legally controlled. In addition, we intended to evaluate

the suitability of ICP-MS semi-quantitative calibration, by comparison of the results obtained by

both calibration approaches, in order to reduce the time and cost of analysis relatively to those of

quantitative calibration.

2.4.2. EXPERIMENTAL

2.4.2.1. Apparatus

The analytical measurements were carried out with a Perkin Elmer SCIEX Elan 9000 ICP-

MS (Perkin-Elmer SCIEX, Norwalk, CT, USA) apparatus, equipped with a crossflow nebulizer, a

Scott-type spray chamber made of Ryton and nickel cones. A peristaltic sample delivery pump with

four channels, model Gilson, was used. Autosampler Perkin-Elmer AS-93 Plus was protected by a

laminar-flow-chamber clean room class 100 (Reinraumtechnik Max Petek). Application software

Elan – 6100 Windows NT (Version 2.4), was used.

For validation of the ICP-MS results, some elements were also determined by Flame

Atomic Absorption Spectrometry (FAAS) (AAnalyst 100, Perkin-Elmer, Norwalk, CT, USA) and

Electrothermal Atomic Absorption Spectrometry (ETAAS) (4110 ZL, Perkin-Elmer, Norwalk, CT,

USA).

Page 85: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

78

2.4.2.2. Material and reagents

Monoelement standard solutions of Be, Co, In (1000 mgl-1), from Merck, and a solution

with Mg, Cu, Rh, Cd, In, Ba, Ce, Pb and U (10 µgl-1), from Perkin-Elmer were used for ICP-MS

optimization procedures. Ultrapure concentrated HNO3 (J.T.Baker), C2H5OH (Lichrosolv, Merck)

and Au (Merck, 1000 mgl-1) for wash, blank, and standard solutions were used. ICP-MS calibration

was established with a multielement standard solution with 30 elements: Li, Be, Na, Mg, Al, K, Ca,

V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Ag, Cd, In, Cs, Ba, Hg, Tl, Pb, Bi, U (Perkin-

Elmer, 10 mgl-1). For ICP-MS internal standardisation standard solutions of Rh and Re (Merck,

1000 mgl-1), were used. For the determination of Cl and Ca interference free ranges and

interference ratios, a monoelement standard solution of Ca (1000 mgl-1) and Suprapur HCl (30%),

both from Merck, were used.

For AAS analysis, monoelement standard solutions of Al, Pb, Cd (Merck, 1000 mgl-1) and

Cu and Fe (Perkin-Elmer, 1000 mgl-1), Pd(NO3)2 and Mg(NO3)2 as matrix modifiers, both from

Merck (10 gl-1), were used.

Purified water was produced using a Seralpur Pro 90CN apparatus (Seral, Ransbach-

Baumbach, Germany). For solution preparation plastic material was only used. All the material was

soaked in 20% HNO3 (v/v) for at least 24 h and rinsed several times with purified water, before

use.

2.4.2.3. Samples

White and red table wines were used for the study and comparison of quantitative and

semi-quantitative approaches. A Portuguese red wine was used for AAS analysis comparison.

Three provisionally certified wines from a BCR intercomparison trial (BCR C white, BCR

D liquor and BCR E red) were analyzed for method validation. These wines are from 1992 year

and showed some precipitation material. Our attempt to get younger reference wine materials had

no success, so we decided to work with these wines, in spite of their state. To the best of our

knowledge, no recent reference wine materials can be found for contaminant elements analysis in

wines.

A 10-fold final dilution of wine samples, without any further sample preparation, was used

for ICP-MS analysis.

HNO3, a solution of Rh and Re (1 mgl-1) and a solution of Au (1 mgl-1) were added to

samples in order to provide a final concentration of 1% (v/v), 10 µgl-1 and 200 µgl-1, respectively.

Blank solution and standards contained 1% HNO3 (v/v), 1% C2H5OH (v/v), 10 µgl-1 of Rh and Re,

and 200 µgl-1 of Au. The Au was used to eliminate the Hg and U memory effect at nebulizer

chamber. The wine samples, blank and standard solutions were prepared daily, in polyethylene

tubes.

Page 86: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

79

2.4.2.4. ICP-MS determinations

The operating conditions were optimized daily, by using an aqueous solution containing 10

µgl-1 of Mg, Ba, Ce, Cu, Cd, Rh, In and Pb, and monitoring the intensities of the isotopes 24Mg, 103Rh, 114In, 208Pb, 138Ba and 140Ce as well as the intensities at mass 69, 156 and 220 (corresponding

to species 138Ba2+, 140Ce16O+ and background, respectively).

The chosen conditions were a compromise between the highest 103Rh ion signal and the

lowest percentage of doubly charge ions (obtained by the intensities ratio Ba2+/Ba+; always ≤ 3%)

and of oxide ions (obtained by the intensities ratio CeO+/Ce+, always ≤ 3%), precision better than

2% and background < 30 cps. The Autolens system was optimised with a 10 µgl-1 Be, Co and In

solution, and Dual Detector calibration with a 200 µgl-1 solution with twenty-nine elements.

Operating conditions used were as follows: RF power of 1200W; sample uptake rate of

0.85 mlmin-1; nebulizer argon flow between 0.85 and 0.95 lmin-1. For both modes of analysis, in

order to get signal stabilization, a sample read delay of 75 s was chosen. Between samples or

standards, the sampling system was rinsed with a 2% HNO3 (v/v), 1% C2H5OH (v/v), and 200 µgl-1

of Au solution for 75 s.

Rh and Re were used as internal standards for elements in the mass range (m/z) 7-138 and

205-238, respectively.

2.4.2.5. Determination of Cl and Ca species interference ratios – an approach to

improve accuracy

In an early step, potential polyatomic interferences caused by Cl and Ca, two main

elements in wine, their interference free ranges and interference ratios were evaluated, using 43Ca

and 35Cl as references. The effects of several concentrations of Cl (20 to 1000 mgl-1) and Ca (10 to

250 mgl-1) on 40Ar35Cl, 35Cl16O, 35Cl18O, 43Ca16O and 44Ca16O species signal were studied. The

stability of interference ratios over time was monitorized, at 20 mgl-1 and 10 mgl-1 Cl and Ca

concentrations, respectively. These concentrations have been used taking into account both the

usual levels of Cl and Ca in wines and the sample dilution factor.

2.4.2.6. Quantitative approach

External calibration was used and the appropriate interpolation was carried out for each

element to determine its concentration in the corresponding calibration line (Table 1).

A selection of isotopes of the elements to be determined was performed, except for

monoisotopic elements. The selected isotopes were those free from isobaric or important matrix-

induced interferences, when possible. Otherwise, suitable elemental equations, established after

preliminary tests, which are presented in the Results and Discussion section, were applied to

correct isobaric and matrix-induced interferences. Since Pb isotope ratios may change from sample

Page 87: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

80

to sample, 206Pb, 207Pb and 208Pb isotopes were measured. The 201Hg and 202Hg isotopes were added

in order to increase the signal intensity.

The experimental conditions used for the measurements were: dwell time = 50.0 ms;

sweeps/reading = 30; reading/replicates = 1; replicates = 4; time per run = 235 s.

Table 1 - Isotopes monitored and concentration range for external calibration in the quantitative approach of analysis

Isotopes monitored Concentration range (µgl-1)

for external calibration 201Hg; 202Hg 0.1; 0.2; 0.5 9Be; 59Co; 71Ga; 75As; 82Se; 107Ag; 111Cd; 115In; 133Cs; 205Tl; 209Bi; 238U 0.05; 0.25; 0.5; 2.5 7Li; 51V; 53Cr; 60Ni 0.25; 0.5; 2.5; 10 65Cu; 138Ba; 206Pb; 207Pb; 208Pb 0.5; 2.5; 10; 50 27Al; 55Mn; 57Fe; 66Zn; 85Rb; 88Sr 2.5; 10; 50; 200

2.4.2.7. Semi-quantitative approach

A full mass spectrum (m/z = 6-240, omitting the mass ranges 16-18, 40, 41, 211-229) was

obtained by full mass range scanning. A reference response table (Perkin-Elmer TotalQuant III)

was updated with the multielement standard solution at 10 µgl-1. The software of the instrument

performs automatic corrections of isobaric interferences.

The experimental conditions used for the measurements were: dwell time = 50 ms;

sweeps/reading = 6; reading/replicates = 1; replicates = 1; time per run = 67 s.

2.4.2.8. Validation of ICP-MS procedures The limits of detection (LOD) of the elements of interest were determined, for both

calibration approaches. The LODs were calculated as being the concentrations corresponding to

signals equal to three-fold the standard deviation of a blank solution signal (3 replicates).

Recovery tests were carried out for a white and a red wine samples spiked with 0.2 to 500

µgl-1 standard solution (depending of the element). Three different spikes of each element were

performed and the solutions obtained were analyzed. The mean and the respective standard

deviation were calculated from the three recovery values obtained for the three spikes.

The contents of Al, Cu, Cd and Pb, were determined by ETAAS [3-5], and those of Fe and

Zn by FAAS. A Portuguese red wine was used in this study.

Page 88: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

81

For validation of the ICP-MS results, three provisionally certified wines from a BCR

intercomparison trial (BCR C white, BCR D liquor and BCR E red), for Al, Fe, Cu, Zn, Cd and Pb,

were analyzed by quantitative analysis.

2.4.3. RESULTS AND DISCUSSION As with all other atomic spectrometric analytical techniques, interferences of various types

can occur during ICP-MS determinations. The most commonly observed result of matrix

interference is the ion intensity if an analyte element becomes dependent upon the total

composition of the sample. Several methods can be used to compensate for matrix-induced signal

suppression such as matrix dilution and standard sample matrix simulation. Wine is a complex

matrix that contains many inorganic and organic substances, which can affect signal intensity in

ICP-MS. So, in order to reduce the matrix effect, initial studies of the method development

included the establishment of the sample minimal preparation procedure. In an early step a five-

fold dilution of samples was tested. Problems with signal suppression and signal instability were

encountered probably due to progressive blocking of the sampling and skimmer cones. Thus a ten-

fold dilution of samples was undertaken. This procedure enabled to reduce matrix effect and still

yet allowed the trace elements quantification.

2.4.3.1. Cl and Ca species interference ratios and mathematical corrections

Spectral interferences are the result of other chemical species, which are present at the

same atomic mass as the analyte of interest. A careful selection of isotopes was made in order to

overcome isobaric overlaps. Sample constituents, usually in combination with oxygen, may

generate polyatomic ions. In most cases, these interferences can be compensated by knowledge of

the intensities of the oxide and the parent ion. A prior knowledge of polyatomic interferences cited

in literature for a particular analyte may be helpful to the analyst for selecting reagents and

conditions that would preclude or reduce [the possibility of] their formation. A good perspective of

know polyatomic interferences was given by May and Wiedmeyer [32]. The fact that a spectral

interference occurs at the isotope mass of an analyte does not necessarily mean that all

determinations of this particular element are interfered. Whether interference occurs or not depends

on the abundance of interfering species, intensities of parent ions of the interference, and on the

ratio at which the interference occurs [33].

Figure 1 shows the stability of Cl and Ca species interference ratios. 40Ar35Cl interference

ratio seems to be stable all over the concentration range in study, and it was always lower than

0.0005, while 35Cl16O and 35Cl18O interference ratios increase with Cl concentration. 43Ca16O and 44Ca16O species interference ratios seem stable in the 10 to 1000 mgl-1 range of concentration.

Page 89: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

82

Stability of CaO/Ca interference ratios (43Ca as reference)

0

0.002

0.004

0.006

0 50 100 150 200 250 300

Ca concentration (mgl-1)

Inte

rfer

ence

ratio

CaO(59)/Ca

CaO(60)/Ca

Figure 1 – Stability of Cl and Ca species interferences ratios with Cl and Ca concentrations.

Cl level in wines is usually lower than 50 mgl-1 nerveless in some typified and rare

situations it can achieve 1 gl-1. Ca level in wines is usually lower than 150 mgl-1. The

stability of calculated interference ratios over time were evaluated, in order to investigate

their suitability to be used as interference correction factor. The interference correction

Stability of ClO/Cl and ArCl/Cl interference rates (35Cl as reference)

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0 200 400 600 800 1000 1200

Cl concentration (mgl-1)

inte

rfer

ence

ratio

ClO(51)/ClArCl(75)/ClClO(53)/Cl

Page 90: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

83

factors were monitorised during three months, showing to be fairly stable over time (20

mgl-1 and 10 mgl-1 concentrations of Cl and Ca, respectively, were used in these studies).

The interference ratios were used in elemental equations to correct for the effect of

polyatomic Cl and Ca species on 51V, 53Cr, 57Fe, 59Co, 60Ni, 75As and 107Ag isotopes. The

software of the instrument performed an automatic correction for isobaric interferences.

The correction equations for both types of interferences are shown at Table 2.

Table 2 - Elemental equations applied for compensation of isobaric and matrix-induced polyatomic interferences in ICP-MS measurements

Analyte /Isotope Interfering species Correction equations

51V 35Cl16O V = I (51V) - 0.0015× I(35Cl)

53Cr 40Ar13C, 35Cl18O Cr = I (53Cr) – 0.002 × I(13C) – 0.0015 × I(35Cl)

57Fe 40Ca16O1H Fe = I (57Fe) - [0.0001 × A(40Ca) / A(43Ca) × I(43Ca)]

59Co 43Ca16O Co = I (59Co) – 0.0009 × I(43Ca)

60Ni 44Ca16O Ni = I (60Ni) - [0.0002 × A(44Ca) / A(43Ca) × I(43Ca)]

75As 40Ar35Cl As = I (75As) – 0.0002 × I(35Cl)

82Se 82Kr Se = I (82Se) - [A(82Kr) / A(83Kr) × I(83Kr)]

107Ag 91Zr16O Ag = I(107Ag) – 0.025 × I( 91Zr)

115In 115Sn In = I(115In) - [A(115Sn) / A(118Sn) × I(118Sn)]

138Ba 138Ce, 138La Ba = I(138Ba) - [A(138La) / A(139La) × I(139La)] –

[A(138Ce) / A(140Ce) × I(140Ce)]

187Re 187Os Re = I(187Re) - [A(187Os) / A(189Os) × I(189Os)]

2.4.3.2. ICP-MS quantitative and semi-quantitative approaches figures of merit

Table 3 shows the detection limits (LODs) for both calibration approaches. The values are

10 to 100 fold lower than those reported by Baxter et al. [18], Pérez-Jordan et al. [19], Almeida et

al. [22], Taylor et al. [24] and Gómez et al. [26], all of them obtained using ICP-Quadropole MS

apparatus. In general, major and minor elements in wine, such as Al, Cr and Fe, presents the

highest LODs, nevertheless always lower than 0.1 µgl-1. The values are low enough to enable

quality and legal control of the contamination elements in wine.

It should be referred that the International Organisation of Vine and Wine (OIV) defines

maximum acceptable values for the concentration of the elements Cu, Zn, As, Cd and Pb in wine,

of 1mgl-1, 5 mgl-1, 0.2 mgl-1, 0.01mgl-1 and 0.20 mgl-1, respectively [6]. The semi-quantitative

Page 91: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

84

approach presents LODs, for most of the studied elements, similar or slightly higher than those

obtained by the quantitative approach.

Results concerning the recovery tests are also present in Table 3. It can be seen that if Cr is

excluded, recoveries between 80 and 120% were obtained for white and red wines. These recovery

percentages are similar to those obtained by several authors [18,19,22] for table wines, using

sample pre-preparation other than dilution or sample nebulizers other than crossflow. For Cr, an

average recovery higher than 120% was obtained. This indicates that the correction of the matrix-

induced polyatomic of 40Ar13C and 35Cl18O species was not effective for this element. Therefore, it

was concluded that Cr could not be measured in the wine samples with the present methodology.

The mathematical corrections used for V, Fe, Co, Ni and As interferences seem to be appropriate.

Table 3 - Limits of detection (LODs)a for both quantitative (Q) and semi-quantitative (SQ) approaches of analysis (µgl-1). Recoveries percentagesb obtained using the quantitative (Q)

approach

Recovery (Q) Recovery (Q)

Element LOD

(Q)

LOD

(SQ) White wine Red wine Element LOD

(Q)

LOD

(SQ) White wine Red wine

Li 0.003 0.003 97 ± 1 88 ± 1 Se 0.01 0.02 97.7 ± 0.9 88.9 ± 2

Be 0.001 0.003 98 ± 2 89 ± 1 Rb 0.001 0.001 96 ± 2 103 ± 3

Al 0.1 0.08 90 ± 2 99 ± 5 Sr 0.002 0.004 98 ± 18 98 ± 2

V 0.001 0.003 102 ± 1 101 ± 2 Ag 0.001 0.006 79 ± 3 81 ± 2

Cr 0.2 0.3 124 ± 4 136 ± 14 Cd 0.001 0.001 82 ± 1 83.0 ± 0.1

Mn 0.001 0.003 104 ± 4 105 ± 8 In 0.0002 0.0005 84 ± 2 83.0 ± 0.3

Fe 0.2 0.2 96 ± 4 110 ± 2 Cs 0.0003 0.0004 89 ± 1 98 ± 1

Co 0.0004 0.001 93 ± 2 90.8 ± 0.5 Ba 0.002 0.010 109 ± 2 81 ± 3

Ni 0.006 0.02 99 ± 3 102 ± 4 Hg 0.01 0.008 99 ± 3 100 ± 3

Cu 0.01 0.03 97 ± 3 91 ± 2 Tl 0.0003 0.0006 91 ± 3 93.7 ± 0.9

Zn 0.02 0.2 95 ± 3 97 ± 3 Pb 0.001 0.006 99 ± 1 92 ± 1

Ga 0.002 0.004 87.6 ± 0.6 84.8 ± 0.6 Bi 0.001 0.02 88 ± 6 93 ± 6

As 0.002 0.005 98 ± 3 95.0 ± 0.8 U 0.003 0.5 104 ± 5 106 ± 2

aLODs were calculated as being the concentrations corresponding to signals equal to three-fold the standard deviation of a blank solution signal (3 replicates); bmean of recovery values of three different spikes ± SD.

Page 92: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

85

Some elements of medium mass (Ag, Cd, In) show recoveries close to 80%, despite the use

of an internal standard with a similar mass (Rh). Re seems to be an appropriate internal standard for

the elements with higher mass. It should be referred that in an initial step of our study, an attempt

was made to use Sc as internal standard for elements of low mass, such as Li, Be and Al. However,

the results were not satisfactory probably due to interferences affecting Sc. Once sample pre-

treatment consists in a dilution, without destruction of the organic material, the results obtained are

satisfactory. Preliminary tests in order to determine internal calibration recoveries showed similar

values to those obtained by external calibration.

2.4.3.3. Comparison of analytical results obtained by ICP-MS quantitative and semi-

quantitative approaches

In order to compare the results of quantitative and semi-quantitative approaches, white and

red table wines were analysed, in parallel, by both modes of analysis (Cr was excluded). For each

wine sample three independent replicates (external calibration) were carried out and the mean and

the respective RSD (%) calculated for each element.

The values of RSD (%) changed between 0.2 and 4% for most elements (Table 4), both for

quantitative and semi-quantitative approach. For Ga, Se, In and Bi RSD (%) values were observed

in white and red wine as follows: 12% and 10%, 11% and 33%, 8% and 60%, 8% and 50%,

respectively. These results could probably be explained by their relatively low concentration in

wines. Nevertheless, these precisions are slightly better than those obtained by Pérez-Jordán et al.

[19] and similar to those reported by Almeida et al. [22].

As shown in Table 4, comparable results (differences lower than 20%) were obtained for

all of the elements with exception of Fe, Zn, Se, Cd, In, Tl and U (white wine) and Al, Fe, Se, Cd,

In, Bi, U (red wine). In most cases the differences were lower than 6% (white wine) and 10% (red

wine). For Be and Ga the differences were between 10 and 20% in both wines. Se was not detected

by semiquantitative mode of analysis, neither in white nor red wine. For In and Tl the main reason

for the differences in white wine, 80% and 60%, respectively, may be the fact that the respective

concentrations are near the analytical limits of both modes of analysis, which precluded their

accurate determination. For Cd, differences of 42% and 30% in white and red wine were observed,

probably due to their low concentration.

The differences observed for U, near to 200%, may be explained by the unsuitability of the

semi-quantitative mode of analysis for such a high mass element. Differences between 42 and 45%

were observed for Fe results. The results of quantitative mode were confirmed by AAS analysis,

leading to the conclusion that either the automatic corrections made by the software for semi-

quantitative mode were not successful or the calibration was not fit. As like for Fe, the differences

observed for Al (21% in red wine) and Zn (23% in white wine and 18% in red wine) may be related

Page 93: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

86

to the unsuitability of the calibration used in semi-quantitative mode of analysis for major

elements. In fact, it was used a standard solution with 10 µgl-1 of major, minor, trace and ultra-trace

elements.

Table 4 - Concentrationsa of the contaminant elements obtained (µgl-1), using both the quantitative (Q) and semi-quantitative methods (SQ), for wine samples

White wine Red wine Element

Quantitative mode Semi-quantitative mode Quantitative mode Semi-quantitative mode

Li 4.4 (0.2) 4.3 (0.2) 7.8 (0.1) 7.55 (0.06)

Be 1.09 (0.04) 0.96 (0.06) 0.127 (0.006) 0.15 (0.07)

Al 1437 (32) 1364 (9) 671 (3) 528 (5)

V 59 (2) 58.0 (0.8) 24.2 (0.2) 23.17 (0.07)

Mn 1036 (28) 1009 (3) 879 (24) 828 (2)

Fe 2147 (21) 1168 (5) 4490 (12) 2587 (14)

Co 4.3 (0.1) 4.47 (0.09) 2.87 (0.02) 2.76 (0.03)

Ni 24.5 (0.9) 27.5 (0.8) 18.2 (0.4) 19.2 (0.5)

Cu 49 (2) 44.2 (0.7) 237 (3) 226 (4)

Zn 661 (10) 819 (6) 642 (19) 758 (6)

Ga 1.7 (0.2) 1.6 (0.1) 2.1 (0.2) 1.83 (0.07)

As 10.6 (0.6) 11.2 (0.3) 3.73 (0.09) 3.76 (0.09)

Se 0.9 (0.1) NDb 0.6 (0.2) NDb

Rb 453 (10) 429.9 (0.8) 674 (18) 627 (3)

Sr 243 (6) 239 (3) 326 (10) 303 (11)

Ag NDb NDb NDb NDb

Cd 0.48 (0.02) 0.28 (0.04) 0.30 (0.02) 0.21 (0.01)

In 0.12 (0.01) 0.02 (0.01) 0.005 (0.003) NDb

Cs 3.6 (0.01) 3.80 (0.01) 3.59 (0.03) 3.80 (0.1)

Ba 89 (2) 89.1 (0.3) 79 (1) 77.3 (0.03)

Hg NDb NDb NDb NDb

Tl 0.24 (0.01) 0.15 (0.03) 0.147 (0.006) 0.145 (0.002)

Pb 14.43 (0.02) 14.0 (0.3) 19.2 (0.3) 19.0 (0.1)

Bi 1.3 (0.1) 1.3 (0.7) 1.2 (0.6) 1.7 (0.1)

U 0.69 (0.01) 2.04 (0.08) 0.47 (0.02) 1.67 (0.04) aConcentration values correspond to the mean of three independent replicates (quantitative mode) and three dependent replicates (semi-quantitative mode) corresponding standard deviation (in brackets); bND – not detected.

As previously referred in the Introduction section, since sensitivity depends on

experimental factors, it is important to update the set of response factors for each element in order

Page 94: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

87

to achieve accurate results. With reference to the importance of accurate results, such update should

be performed with a multielement standard solution containing, for each element, the approximate

concentration of those observed in wine.

In reference to the elements under legal control, it should be noted that for Zn and Cd, the

comparison study revealed large differences between the results obtained by the two calibration

approaches (higher than 10%) and, hence, demonstrate the unsuitability of semi-quantitative

approach in the analysis of these elements in wine. A linear least-squares adjustment of the results

obtained by both modes of analysis was performed in order to evaluate the potential risk of error

(the 95% confidence limits of the linear regression parameters are given). Non comparable results

(differences > 20%) were not considered in the adjustment. A deviation of the slope from unity

indicates proportional discrepancies between the two methods. A non-zero intercept is diagnosed as

a constant discrepancy. The regression line for white wine yielded the equation: y = (0.957 ±

0.008)x + (1 ± 4), R = 0.9999 (n = 16), (95% confidence level). For red wine the equation was y =

(0.937 ± 0.005)x + (1 ± 2), R = 0.9999 (n = 15), (95% confidence level). The regression line of the

global results (including white and red wine samples) yielded the equation y = (0.952 ± 0.006)x ±

2, R = 0.9999 (n = 31), (95% confidence level). In both cases, white and red sample wines, high

correlations were found, being very close to unity. However, the semi-quantitative mode analytical

results are slightly lower than quantitative mode results, suggesting the existence of a slight bias.

These observations are similar to those register by Almeida et al. [21], obtained either with table

and fortified wines.

2.4.3.4. Validation of ICP-MS procedures

As referred, for validation of ICP-MS quantitative approach, six elements of a Portuguese

red table wine were measured both by ICP-MS and AAS (Figure 2). The differences between the

results obtained by the two techniques were lower than 5% for all elements, with exception of Al

(18%), demonstrating the accuracy of ICP-MS method. Linear least squares adjustment was

applied to the six elements determined by AAS (x- axis) and ICP-MS (y-axis). The regression line

of the global results yielded the equation: y = (1.01 ± 0.04)x + (3 ± 40) (95% confidence level),

with a correlation coefficient of R = 0.9995 (n = 6). Evidence of relative or fixed bias was not

observed.

The results concerning the analysis of provisionally certified wines by quantitative ICP-MS

approach are shown in Figure 3. Acceptable agreement was obtained within the tolerances supplied

for all samples and analytes with the exceptions of Fe (BCR D and BCR E wines), Cd and Cu

(BCR E wine). A linear least squares adjustment was applied to all elements determined,

provisional data (x-axis) and our data (y-axis).

Page 95: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

88

Figure 2 - Comparison between AAS and ICP-MS results (µgl-1) obtained for the elements Al, Fe, Cu, Zn, Cd and Pb measured in a Portuguese red wine. The y-axis represents the logarithm of concentrations.

For BCR C wine the equation y = (0.79 ± 0.09)x + (14 ± 55) (95% confidence level), with

a correlation coefficient of R = 0.99 (n = 9), was obtained. For BCR D wine the equation y = (0.80

± 0.03)x + (40 ± 75) (95% confidence level), with a correlation coefficient of R = 0.999 (n = 9),

was obtained. If we exclude Fe data, for BCR E wine the equation y = (0.8 ± 0.2)x - (7 ± 75) (95%

confidence level), with a correlation coefficient of R = 0.97 (n = 8) was observed. The regression

line of the global results (including the three wines), yielded the equation: y = (0.80 ± 0.03)x + (14

± 55) (95% confidence level), with a correlation coefficient of R = 0.99 (n = 26). The differences

between our data ant the provisional values indicate a systematic bias (our results were lower than

the provisional values, that is, a slope significantly < 1). The reason for this may related to the

precipitation phenomena observed in all the wines, but more intense in red wine (BCR E wine). As

previously mentioned the wines are from 1992 year and showed some precipitation material. It is

know that metals interact with polyphenols and pectic polysaccharides species, which, in part,

precipitate during the ageing of wines.

0,1

1

10

100

1000

10000

Al Fe Cu Zn Cd Pb

Element

Con

cent

ratio

n (u

gl-1

)

AAS results ICP-MS results

Page 96: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

89

Figure 3 - Data for reference wines (BCR C, BCR D and BCR E, in the graph referred as C, D and E, respectively), showing values obtained by the authors and provisional values. The y- axis represents the logarithm of concentrations (µgl-1) in order to obtain a better scatter of the data. Information about Li and Mn (provisional) concentration standard deviation of the three wines was not available.

2.4.4. CONCLUSIONS

In the present work it was demonstrated that ICP-MS quantitative and semi-quantitative

approaches offer valuable alternatives to table wine characterization and comparison purposes.

However, as a tool for legal control, quantitative approach is the most suitable alternative. For

some elements with legal importance, namely for Fe, Zn and Cd, the comparison study reveals

large differences between the results obtained using quantitative and semi-quantitative calibration.

In spite of that, analytical results of both methods are comparable in terms of precision and

accuracy (relative differences < 20%) for most of the elements. In contrast with quantitative

methodology, semi-quantitative methodology is faster, 67 s per sample, which allows high sample

throughput. Another advantage is the lower reagent consumption.

Acknowledgements To “Programa de Apoio à Reforma das Instituições Públicas e de Interesse Público de Investigação”

a research scholarship and to “Fundação para a Ciência e Tecnologia”, Lisbon, Portugal, a PhD scholarship

(PRAXIS XXI/BD/17237/2004).

0,1

1

10

100

1000

10000

C D E C D E C D E C D E C D E C D E C D E C D E C D E

Li Al Mn Fe Cu Zn As Cd Pb

Element / Wine

Ele

men

t con

cent

ratio

n (u

g.l-1

)

provisional values authors values

Page 97: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

90

2.4.5. REFERENCES

[1] A.S. Curvelo-Garcia, Controlo de Qualidade dos Vinhos, Química Enológica, Métodos Analíticos, Instituto da Vinha e do Vinho, Lisboa, 1988 (Chapter VI).

[2] A.S. Curvelo-Garcia, S. Catarino, Ciência Téc. Vitiv., 13 (1) (1998) 49.

[3] S. Catarino, Ocorrência de chumbo e cádmio em vinhos, Msc. Thesis, University of Porto, Porto, 2000.

[4] S. Catarino, A.S. Curvelo-Garcia, R. Bruno de Sousa, At. Spectrosc., 23 (6) (2002) 196.

[5] S. Catarino, I. Pimentel, A.S. Curvelo-Garcia, At. Spectrosc., 26 (2) (2005) 73.

[6] OIV, Recueil des methods internationals d’analyse des vins et des mouts, Organisation International de la Vigne et du Vin, Paris, 2005.

[7] J.R. Dean, L. Ebdon, C. Massey, Food Addit. Contam., 7 (1) (1990) 109.

[8] B.L. Gulson, T.H. Lee, K.J. Mizon, M.J. Korsch, H.R. Eschnauer, Am. J. Enol. Vitic., 43 (2) (1992) 180.

[9] S. Augagneur, B. Médina, F. Grousset, Fresenius J. Anal. Chem., 357 (1997) 1149.

[10] K.J.R. Rosman, W. Chisholm, S. Jimi, J-P. Candelone, C.F. Boutron, P-L. Teissèdre, F.C. Adams, Environ. Res., 78 (1998) 161.

[11] C.M.R. Almeida, M.T.S.D. Vasconcelos, J. Anal. At. Spectrom., 14 (1999) 1815.

[12] I. Rodushkin, F. Odman, P.K. Appelblad, J. Food Comp. Anal., 12 (1999) 243.

[13] B. Médina, S. Augagneur, M. Barbaste, F.E. Grousset, P. Buat-Ménard, Food Addit. Contam., 17 (6) (2000) 435.

[14] M. Barbaste, L. Halicz, A. Galy, B. Médina, H. Emteborg, F.C. Adams, R. Lobinski, Talanta, 54 (2001) 307.

[15] C.M. Almeida, M.T.S.D. Vasconcelos, J. Anal. At. Spectrom., 16 (2001) 607.

[16] A. Stroh, P. Bruckner, U. Vollkopf, At. Spectrosc., 2 (1994) 100.

[17] J.D. Greennough, H.P. Longerich, S.E. Jackson, Austral. J. Grape Wine Res., 3 (1997) 75.

[18] M.J. Baxter, H.C. Crews, M.J. Dennis, I. Goodall, D. Anderson, Food Chem., 60 (3) (1997) 443.

[19] M-Y. Pérez-Jordán, J. Soldevila, A. Salvador, A. Pastor, M. de la Guardia, J. Anal. At. Spectrom., 13 (1998) 33.

[20] N. Jakubowski, R. Brandt, D. Stuewer, H.R. Eschnauer, Gortges, Fresenius J. Anal. Chem., 364 (1999) 424.

[21] M.M. Castiñeira, R. Brandt, A. Von Bohlen, N. Jakubowski, Fresenius J. Anal. Chem., 370 (2001) 553.

[22] C.M. Almeida, M.T.S.D. Vasconcelos, Anal. Chim. Acta, 463 (2002) 165.

[23] C.M.R. Almeida, M.T.S.D. Vasconcelos, M. Barbaste, B. Médina, Anal. Bioanal. Chem., 374 (2002) 314.

[24] V.F. Taylor, H.P. Longerich, J.D. Greenough, J. Agric. Food Chem., 51 (2003) 856.

[25] G. Nicolini, R. Larcher, P. Pangrazzi, L. Bontempo, Vitis, 43 (1) (2004) 41.

[26] M.D.M.C. Gómez, R. Brandt, N. Jakubowski, J.T. Andersson, J. Agric. Food Chem., 52 (2004) 2953.

[27] G. Thiel, G. Geisler, I. Blechschmidt, K. Danzer, Anal. Bioanal. Chem., 378 (2004) 1630.

[28] G.J. Martin, M. Mazure, C. Jouitteau, Y.L. Martin, L. Aguile, P. Allain, Am. J. Enol. Vitic., 50 (4) (1999) 409.

[29] M. Barbaste, Recherches sur l’origine geographique et le millesime des vins, PhD Thesis, Universite de Pau et Des Pays de l’Adour, 2001.

[30] S. Sagrado, M.Y. Pérez-Jordan, A. Pastor, A. Salvador, M. De la Guardia, J. Chemometrics, 16 (2002) 305.

[31] E. Pruszkowski, Field Application Report, Perkin-Elmer, 2004.

Page 98: Tese doutoramento Sofia Catarino.pdf

Desenvolvimento e validação de métodos de análise

91

[32] T.W. May, R.H. Wiedmeyer, At. Spectrosc., 19 (5) (1998) 150.

[33] Software Guide, Elan Version 2.4., Perkin-Elmer Sciex, Canada, 2001.

Page 99: Tese doutoramento Sofia Catarino.pdf

Capítulo 3

AVALIAÇÃO DA COMPOSIÇÃO DE MOSTOS E VINHOS EM

METAIS CONTAMINANTES

Page 100: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

95

3.1. EVALUATION OF CONTAMINANT ELEMENTS IN PORTUGUESE WINES AND ORIGIN MUST BY

HIGH INTENSITY FOCUSED ULTRASOUND COMBINED WITH INDUCTIVELY COUPLED

PLASMA MASS SPECTROMETRY

S. Catarino1,2, J.L. Capelo3, A.S. Curvelo-Garcia1, R. Bruno de Sousa2 1INIAP, Estação Vitivinícola Nacional, 2565-191 Dois Portos, Portugal 2Instituto Superior de Agronomia, Departamento de Química Agrícola e Ambiental, 1349-017 Lisboa, Portugal 3REQUIMTE, Departamento de química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Monte da Caparica, Portugal

Abstract

Portuguese must and wines from several varieties and different vineyards were studied to evaluate their contents in contaminant elements, in two winemaking steps. HIFU (High Intensity Focused Ultrasound) combined with ICP-MS (Inductively Coupled Plasma Mass Spectrometry) was the methodology used. The variation from must to wines was a function of the element and its concentration. Be, Al, Mn, Co, Ni, Cu, Ga, Rb, Cd, In, Ba, Tl and U concentration in wines was lower than in must. Cu depletion achieved 99% of the initial value. For Li, Fe, Zn, Sr, Cs, Pb and Bi, white and red samples presented different trends, suggesting the importance winemaking process. The levels of Cu, Zn, As, Cd and Pb were lower than the acceptable maximal limits established by the International Organisation of Vine and Wine and in general could be considered relatively low values. Keywords: must, Portuguese wines, contaminant elements, HIFU, ICP-MS

Journal International des Sciences de la Vigne et du Vin (2006) 40 (2), 91-100

Page 101: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

96

3.1.1. INTRODUCTION

The elemental composition of wines is influenced by environmental factors, such as

geological and atmospheric phenomena, soil and climate, as well as by technological processing of

the grapes, must and wine (ESCHNAUER, 1982; CURVELO-GARCIA, 1988; STOCKLEY and

LEE, 1995; CURVELO-GARCIA and CATARINO, 1998; RIBÉREAU-GAYON et al., 1998).

Several sources contributing to the metal composition of a finished wine are known. An important

part of the total concentration comes from the vineyard soil via the roots of the grapevine. The

other part, considered as contamination, is introduced during the different processing stages, from

the grape culture to the finished wine. Several research studies have been developed in the last

decade in order to evaluate the potential sources of wine contamination (MCKINNON et al., 1992;

TEISSÈDRE et al., 1993; TEISSÈDRE et al., 1998a; GÓMEZ et al., 2004; LIMA et al., 2004;

NICOLINI et al., 2004). As an example, MÉDINA et al. (2000) have shown the influence of

atmospheric pollution on the lead content of wines. Recently, SALVO et al. (2003) have studied

the influence of different mineral and organic pesticide treatments on Cu, Zn, Cd and Pb contents

in italian wines.

The determination of some elements is of interest due to their toxicological or

physiological properties, while others can lead to wine spoilage. Hence, the presence of metallic

elements and their concentration may affect must and wine quality. Although the concentrations of

only few elements in wines are currently under regulation, the increasing concern for health effects

and environmental exposure will probably result in a longer list of potentially toxic elements to be

monitored. To date, the International Organisation of Vine and Wine (OIV) define maximum

concentration values for the elements Cu, Zn, As, Cd and Pb in wine (OIV, 2005) as follows: 1

mg/L, 5 mg/L, 0.2 mg/L, 0.01 mg/L and 0.2 mg/L, respectively.

Metal determination in must and wine can be performed only if matrix influences are

eliminated or at least diminished. Must is a complex sample with a high organic content, mainly

sugars and large agglomerations of organic matter. Therefore, a sample pre-treatment is required in

order to destroy the organic matter which could interfere with the analytical determination by

spectrometric techniques such as Inductively Coupled Plasma – Mass Spectrometry. Nowadays,

microwave digestion (MWD) is widely used in the destruction of organic matter in samples with

high sugar content (TEISSÈDRE et al., 1993; CABRERA-VIQUE et al., 1997; KRISTL et al.,

2002; GÓMEZ et al., 2004). However, High Intensity Focused Ultrasound (HIFU) is also a useful

technique for the extraction of metals associated with organic matter present in must. Recently, we

have evaluated the suitability of this methodology in grape must (CATARINO et al., 2005). It was

demonstrated that lead concentration values were not statistically different when HIFU or MWD

were used for sample treatment. Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is an

excellent tool for metal analysis because it is a multielemental technique, with high selectivity and

sensitivity and low analytical detection limits.

Page 102: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

97

The main purpose of this work was to evaluate the metal content in must and

corresponding wines from different grape varieties and regions from Portugal, by using HIFU

sample treatment in conjunction with ICP-MS for metal quantification.

3.1.2. MATERIALS AND METHODS

3.1.2.1. Apparatus

A Branson Sonifier 150 ultrasonic cell disruptor-homogeniser (100 W, 22.5 kHz, Branson

Ultrasonics Corporation, USA) equipped with a 3-mm titanium micro-tip was used. Ultrasonic

energy irradiation was fixed at 10% level with the 3-mm micro-tip. The Sonifier 150 has a digital

LCD display which provides a continuous read-out of the watts delivered to the end of the probe

(range 5-6 W in this work). Eppendorf cups (2 ml) were used throughout this work. A Wifug

(London, UK) labor-50M centrifuge was used.

The analytical measurements were carried out with a Perkin Elmer SCIEX Elan 9000 ICP-

MS (Perkin-Elmer SCIEX, Norwalk, CT, USA) apparatus, equipped with a crossflow nebulizer, a

Scott-type spray chamber made of Ryton and nickel cones. A peristaltic sample delivery pump with

four channels, model Gilson, was used. Autosampler Perkin-Elmer AS-93 Plus was protected by a

laminar-flow-chamber clean room class 100 (Reinraumtechnik Max Petek). Application software

Elan – 6100 Windows NT (version 2.4), was used.

3.1.2.2. Material and reagents

Monoelement standard solutions of Be, Co, In (1000 mg/L), from Merck (Darmstadt,

Germany), and a solution of Mg, Cu, Rh, Cd, In, Ba, Ce, Pb and U (10 µg/L), from Perkin-Elmer

(Connecticut, USA) were used for ICP-MS optimization procedures. Ultrapure concentrated HNO3

(J.T.Baker, Phillipsburg, USA), C2H5OH (Lichrosolv, Merck) and Au (Merck, 1000 mg/L) for

washing, blank, and standard solutions were used. ICP-MS external calibration was established

using a multielement standard solution with 30 elements (Perkin-Elmer, 10 mg/L). For ICP-MS

internal standardisation standard solutions of Rh and Re (Merck, 1000 mg/L), were used. H2O2

(Puriss. p.a.) from Fluka (Steinheim, Germany) was used for must treatment. Purified water

(conductivity < 0.1 µS/cm) was produced using a Seralpur Pro 90CN apparatus (Seral, Ransbach-

Baumbach, Germany). For sample and solution preparation only plastic material was used. All the

material was soaked in 20% HNO3 (v/v) for at least 24 h and rinsed several times with purified

water before use.

Page 103: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

98

3.1.2.3. ICP-MS determinations

The multielemental analysis of must and wines was performed by ICP-MS quantitative

approach (CATARINO et al., in press). The operating conditions were optimized daily, by using an

aqueous solution containing 10 µg/L of Mg, Ba, Ce, Cu, Cd, Rh, In and Pb, and monitoring the

intensities of the isotopes 24Mg, 103Rh, 114In, 208Pb, 138Ba and 140Ce as well as the intensities at mass

69, 156 and 220 (corresponding to species 138Ba2+, 140Ce16O+ and background, respectively). The

chosen conditions were a compromise between the highest 103Rh ion signal and the lowest

percentage of doubly charge ions (obtained by the intensities ratio Ba2+/Ba+; always ≤ 3%) and of

oxide ions (obtained by the intensities ratio CeO+/Ce+, always ≤ 3%), precision better than 2% and

background < 30 cps. The Autolens system was optimised with a 10 µg/L Be, Co and In solution,

and Dual Detector calibration with a 200 µg/L solution with twenty-nine elements. Operating

conditions were as follows: RF power of 1200W; sample uptake rate of 0.85 mL/min; nebulizer

argon flow between 0.85 and 0.95 L/min. The wash solution contained 2% HNO3 (v/v), 1%

C2H5OH (v/v) and Au (200 µg/L). The Au was used to eliminate the Hg and U memory effect at

nebulizer chamber. As internal standards Rh and Re, both at 10 µg/L concentrations, were used for

blank solution, standards and samples preparation. Blank solution and standards contained 1%

HNO3 (v/v), 1% C2H5OH (v/v), 10 µg/L of Rh and Re, and 200 µg/L of Au. Since must do not

contain C2H5OH, blank, standard and wash solutions used to analyse must samples were prepared

without this compound. The must and wine samples, blank and standard solutions were prepared

daily, in polyethylene tubes. In order to get signal stabilization, a sample read delay of 75 s was

chosen. Between samples or standards, the sampling system was rinsed with a 2% HNO3 (v/v), 1%

C2H5OH (v/v), and 200 µg/L of Au solution for 75 s. External calibration was used and the

appropriate interpolation was carried out for each element to determine its concentration in the

corresponding calibration line. Isotopes monitored and concentration range (µg/L) for external

calibration are the following: 201Hg; 202Hg (0.1; 0.2; 0.5); 9Be; 59Co; 71Ga; 75As; 82Se; 107Ag; 111Cd; 115In; 133Cs; 205Tl; 209Bi; 238U (0.05; 0.25; 0.5; 2.5); 7Li; 51V; 53Cr; 60Ni (0.25; 0.5; 2.5; 10); 65Cu; 138Ba; 206Pb; 207Pb; 208Pb (0.5; 2.5; 10; 50); 27Al; 55Mn; 57Fe; 66Zn; 85Rb; 88Sr (2.5; 10; 50; 200). The

selected isotopes were those free from isobaric or important matrix-induced interferences, when

possible. Otherwise, suitable elemental equations were applied to correct isobaric and matrix-

induced interferences. Since Pb isotope ratios may change from sample to sample, 206Pb, 207Pb and 208Pb isotopes were measured. The 201Hg and 202Hg isotopes were added in order to increase the

signal intensity. The experimental conditions used for the measurements were: dwell time = 50.0

ms; sweeps/reading = 30; reading/replicates = 1; replicates = 4; time per run = 235 s.

Page 104: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

99

3.1.2.4. Must and wines

Seven wines and origin must originated from 2003 and 2004 vintage years were used.

Several grape varieties (Vitis vinifera L.) from different regions of Portugal, were used in this

study. Samples A and B, both from white grapes, were treated with distinct pesticides. Samples C

and D are different clones of Fernão Pires, a Portuguese white variety, growing in Óbidos and

Caldas da Raínha vineyards. Samples E, F and G are from red varieties, the first two representing

different clones of Aragonez from Alentejo, while G sample is from Syrah variety (Dois Portos

vineyards).

3.1.2.5. Microvinification processes

Microvinifications were performed at the winery of Estação Vitivinícola Nacional (Dois

Portos, Portugal). The operating procedures were as follows: grapes from each white variety were

destemmed/crushed, and immediately pressed for skin and must separation and sulfited with

sulphur dioxide (100 mg/L). Must were decanted at 4ºC during 48 hours and fermentation

continued for 11 to 12 days in 60 L stainless steel tank at 19ºC. At the end of fermentation the wine

was decanted from lees and stored in 20 L glass carboys at room temperature. After 9 to 10 months,

the wines were racked and bottled in 750 mL glass bottles. Sulphur dioxide (40 mg/L) was added at

bottling. Grapes from red varieties were processed, also separately, into wine according to the

classical red wine technology. Grapes were destemmed/crushed and collected in 60 L stainless steel

tanks, sulfited with sulphur dioxide (60 mg/L) prior skin fermentation at 25ºC. The cap was

punched down twice daily until it remained submerged. When alcoholic fermentation was nearly

finished, after 5 days of maceration, the mash was pressed. The wine obtained was stored at 20 l

carboys at room temperature until bottling. Sulphur dioxide (20 mg/L) was added at bottling. With

exception of sulphur dioxide addition, in order to minimize potential contamination sources,

namely those represented by enological additives, no other winemaking treatment was made.

3.1.2.6. Must sample collection

Exogenous contamination was avoided cleaning all the plastic bottles used for must sample

collection with HNO3 10% v/v. The bottles were then rinsed gently with ultrapure water and dried.

The must samples were obtained in duplicate immediately after the grapes were pressed (before

sulphur dioxide addition), and 5 mL of HNO3 (70% v/v) were added to a sample volume of 0.1 L.

The samples were stored at -10 ºC until they were analyzed.

3.1.2.7. Must sample treatment by HIFU Following the procedure previously described elsewhere (CATARINO et al., 2005), to 0.5

mL must sample placed in eppendorf cups were added the following reagents: HNO3 (70% v/v) +

Page 105: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

100

H2O2 (30% v/v) (50 µl + 100 µL). Then, focused ultrasound was applied for 60 s at 10% amplitude

(5-6 watts delivered at the end of the 3 mm probe). Finally the sample was centrifuged for 5 min at

10000 rpm and the supernatant was used for measurements. For each must sample, this treatment

procedure was repeated several times in order to obtain the required volume to ICP-MS analysis.

3.1.3. RESULTS AND DISCUSSION

3.1.3.1. Ultrasonic treatment

Preliminary assays were performed to ensure the homogeneity and representativeness of

the must samples investigated. Since must samples have large agglomerations of organic matter,

eppendorf tips were modified in order to make weighting of the sample as precise as possible. In

order to do so, the tip was cut 1 cm above its end, the hole of the tip being modified to a higher

size. For each must, eight 0.5 mL samples were taken while stirring to ensure homogenization. The

precision (RSD, n=8) was always lower than 2% (Must A: 0.590 ± 0.007 g); (Must B: 0.60 ± 0.01

g); (Must C: 0.58 ± 0.01 g); (Must D: 0.576 ± 0.006 g); (Must E: 0.57 ± 0.01 g); (Must F: 0.576 ±

0.005 g); (Must G: 0.59 ± 0.01 g). In order to control any contamination, as a quality control tool of

the process, water was submitted to complete treatment procedure. Periodically, and between every

10 must samples, this procedure was repeated. Blank of HIFU sample treatment was subtracted to

ICP-MS analytical results. Blanks of HIFU sample treatment had high levels of aluminium (nearly

300 µg/L) and vanadium (about 200 µg/L), probably due to ultrasonic probe. The low levels of

vanadium in must make the measurement of this metal impossible. Aluminium content in must and

wine samples are high enough to allow its quantification.

3.1.3.2. Must ICP-MS analysis

In an initial step of our work, the feasibility of ICP-MS to analyse must samples was

assessed through recovery tests. White and a red must samples were spiked with 5 to 5000 µg/L

standard solution. Three different spikes of each element were performed. Table 1 shows the results

which could be considered satisfactory except for As and Ag. Excluding afore mentioned elements

the recovery averages were 98±5% (n=22) for white must and 100±6% (n=21) for red must. The

high recovery percentage observed for As, approximately 200%, probably means that severe

polyatomic interferences (namely 40Ar35Cl species) were not correctly eliminated (MAY and

WIEDMEYER, 1998). Taking this into account, the quantification of this element in must samples

was not carried out. Ag recovery was about 80%, in agreement with previous results observed in

our laboratory.

Page 106: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

101

Table 1 - Recoveries obtained for contaminant elements in must by ICP-MS analysis

Recovery Recovery Element

White must Red must Element

White must Red must

7Li 95.1 ± 0.1 91 ± 2 85Rb 96 97 ± 5

9Be 95.6 ± 0.5 95 ± 2 88Sr 100 ± 3 100 ± 2

27Al 105 ± 4 112 ± 2 107Ag 88 80

51V 112 ± 2 111 111Cd 90.1 ± 0.1 92 ± 2

55Mn 100 ± 4 107 ± 2 115In 94.01 ± 0.07 95 ± 2

57Fe 99 ± 4 109 ± 4 133Cs 91.6 ± 0.3 95 ± 2

59Co 100.2 ± 0.1 102 ± 2 138Ba 93 ± 1 97.1 ± 0.9

60Ni 100 ± 1 99 201Hg 105.5 ± 0.1 ND

65Cu 103 ± 3 108.9 ± 0.5 205Tl 90.9 ± 0.1 94 ± 3

66Zn 101 ± 5 102 208Pb 89.8 ± 0.7 92 ± 3

71Ga 98.4 ± 0.4 100 ± 2 209Bi 96 ± 2 97.1 ± 0.6

75As 208.3 ± 0.3 218 ± 3 238U 94.6 ± 0.3 97 ± 1

ND: not determined.

3.1.3.3. Contents of elements / changes from must to wine

The levels of elements (mean of three independent replicates ± standard deviation) in must

and wines are shown in table 2. The results in both must and wine are expressed in µg/L in order to

allow better must-wine comparison. Analytical results for V and As in must are not shown because

of the reasons explained previously. In addition, levels of Se in must and Hg in must and wine were

below our quantification limits. To afford better observation and analysis of the results, Figure 1

shows the changes of concentrations for each pair must-wine, expressed in percentage.

Important losses (expressed as percentage) of Be, Al, Mn, Co, Ni, Cu, Zn, Ga, Rb, Cd, Ba,

Tl and U occurred from must to wines during alcoholic fermentation, probably due to precipitation

as insoluble salts, namely sulphides. Furthermore, these phenomena are favoured both by sulphur

dioxide and pH value, which, in part, could explain the differences between samples along the

winemaking process (wines pH values are shown at table 2). The tendencies for higher decreases of

Mn and Co in red samples may be related to pH value. It should also be emphasized the almost

complete depletion of Cu from must to wine, that could be explained by the low solubility of CuS.

Page 107: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

102

Table 2 – Levels of contaminant elements in must and wines (µg/L)

A B C D E F G Element

/Isotope Must Wine

pH 3.12 Must

Wine

pH 3.14 Must

Wine

pH 3.06 Must

Wine

pH 3.02 Must

Wine

pH 3.73 Must

Wine

pH 3.83 Must

Wine

pH 3.37

7Li 4.6 (0.5) 4.0 (0.1) 5.021

(0.005) 5.7 (0.1) 24 (2) 25.2 (0.4) 13 (1) 13.0 (0.8) 6.8 (0.2) 11.7 (0.2) 5.5 (0.2) 9.0 (0.3) 24.8 (0.3) 28.1 (0.5)

9Be 0.058

(0.003)

0.013

(0.003)

0.04

(0.009)

0.011

(0.003)

0.187

(0.006)

0.080

(0.007)

0.1330

(0.0007)

0.059

(0.002)

0.154

(0.002) 0.06 (0.01)

0.19

(0.02) 0.07 (0.03)

0.114

(0.003) 0.12 (0.05)

27Al 301 (17) 134 (4) 526 (35) 119 (6) 671 (3) 201 (40) 685 (176) 116 (4) 1249 (20) 180 (39) 1501

(251) 236 (13) 527 (10) 273 (21)

51V ND 0.42 (0.04) ND 0.38 (0.03) ND 1.34 (0.03) ND 0.90 (0.07) ND 0.06 (0.02) ND 0.16 (0.09) ND 0.32 (0.06)

55Mn 553 (4) 482 (12) 2076 (25) 1343 (32) 471 (26) 407 (10) 602 (30) 546 (25) 1738 (30) 382 (21) 1627 (51) 329 (6) 433 (2) 498 (12)

57Fe 1212 (46) 352 (10) 1586 (98) 430 (12) 1144 (41) 540 (11) 1035 (22) 455 (24) 830 (20) 1526 (33) 892 (38) 1363 (27) 1192 (20) 1359 (41)

59Co 1.7 (0.1) 0.81 (0.02) 2.1 (0.01) 0.90 (0.04) 2.4 (0.1) 1.22 (0.04) 2.47

(0.08) 1.36 (0.07)

2.39

(0.02) 0.32 (0.02) 2.2 (0.2) 0.15 (0.03)

3.35

(0.02) 2.5 (0.1)

60Ni 13.8 (0.4) 9.7 (0.3) 13 (1) 9.8 (0.2) 14.1 (0.7) 7.5 (0.2) 16.5 (0.3) 7.6 (0.4) 17 (1) 7.1 (0.3) 13 (3) 7.7 (0.2) 14.1 (0.5) 16 (1)

65Cu 372 (3) 14 (1) 361.96

(0.02) 368 (9)

9571

(211) 45.3 (0.8)

5577

(245) 23 (1) 1050 (18) 5.8 (0.2) 1067 (28) 6 (2) 7912 (75) 20 (5)

66Zn 250.5

(0.7) 301 (34) 456 (18) 488 (40) 364 (10) 422 (13) 584 (63) 467 (27) 578 (57) 26 (5) 301 (34) 6 (2) 1092 (19) 901 (230)

71Ga 3.6 (0.1) 1.73 (0.04) 3.60

(0.07) 1.84 (0.05)

3.76

(0.05) 1.73 (0.04) 3.3 (0.5) 1.7 (0.2) 3.0 (0.2) 1.17 (0.03)

3.51

(0.08) 1.31 (0.05) 4.4 (0.4) 4.1 (0.2)

Page 108: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

103

75As ND 0.57 (0.03) ND 0.59 (0.04) ND 0.18 (0.01) ND 0.30 (0.02) ND 0.56 (0.02) ND 0.43 (0.06) 7.141

(0.008) 0.81 (0.09)

82Se NQ 0.37 (0.06) NQ 0.26 (0.06) NQ 0.33 (0.09) NQ 0.37 (0.04) NQ 0.27 (0.06) NQ 0.297 NQ 0.48 (0.07)

85Rb 1283 (18) 823 (18) 1708 (34) 1158 (29) 2735 (60) 1533 (28) 2196 (98) 1560 (65) 2696 (47) 2022 (32) 2375 (67) 704 (13) 2503 (5) 1839 (35)

88Sr 113 (2) 98 (2) 118 (2) 100 (3) 152.9

(0.6) 136 (3) 102 (6) 95 (4) 60.2 (0.8) 101 (1) 36.7 (0.9) 126 (3) 383 (3) 474 (9)

111Cd 0.95

(0.01)

0.34 (5×10-

9) 1.3 (0.1) 1.26 (0.09)

1.06

(0.01) 0.22 (0.01) 1.1 (0.2) 0.24 (0.02)

1.10

(0.04) 0.3 (0.5)

0.95

(0.04) 0.07 (0.02)

1.30

(0.01) 0.27 (0.01)

115In 0.045

(0.008)

0.006

(0.005)

0.021

(0.004)

0.009

(0.002)

0.0175

(0.0007)

0.001

(0.002)

0.017

(0.004)

0.001

(0.002)

0.0145

(0.0007)

0.011

(0.008)

0.0165

(0.0007) 0.04 (0.03) 0.018 (0) 0.03 (0.03)

133Cs 3.8 (0.1) 3.27 (0.09) 6.5 (0.1) 6.1 (0.1) 18.2 (0.4) 16.7 (0.3) 19.9 (0.9) 19.0 (0.7) 6.5 (0.1) 18.9 (0.4) 4.5 (0.1) 2.69 (0.05) 7.92

(0.06) 9.5 (0.2)

138Ba 19.2 (0.5) 13.3 (0.2) 17.53

(0.05) 10.8 (0.3) 80 (2) 50.4 (0.8) 71 (2) 50 (2)

111.3

(0.8) 11.4 (0.3) 99 (3) 14.9 (0.3)

121.1

(0.2) 104 (2)

201Hg NQ NQ NQ NQ NQ NQ NQ NQ NQ NQ NQ NQ NQ NQ

205Tl 0.36

(0.02)

0.198

(0.004)

0.516

(0.009)

0.278

(0.004)

0.41

(0.01)

0.0487

(0.0006)

0.54

(0.03)

0.096

(0.005)

0.603

(0.009)

0.358

(0.002)

0.40

(0.01) 0.09 (0.03)

1.0025

(0.0007) 0.47 (0.04)

208Pb 15.04

(0.09) 28 (0.4)

12.76

(0.09)

27.91

(0.09)

5.54

(0.02) 1.82 (0.02) 6.1 (0.3) 1.87 (0.08) 8.8 (0.2) 3.5 (0.4)

4.612

(0.008) 1.81 (0.05) 6.6 (0.1) 3.21 (0.04)

209Bi 0.94

(0.05) 0.5 (0.3) 0.5 (0.1) 0.13

0.27

(0.03) 0.07 (0.01)

0.14

(0.01) 0.13 (0.02)

0.007

(0.008) 0.6 (0.2)

0.01

(0.02) 2 (1)

0.07

(0.04) 0.6 (0.3)

238U 0.16

(0.04)

0.009

(0.005)

0.150

(0.008)

0.004

(0.002)

0.614

(0.004)

0.0143

(0.0006)

0.46

(0.06)

0.0057

(0.0006)

0.08

(0.01) 0.03 (0.01)

0.062

(0.003) 0.06 (0.05)

0.115

(0.001) 0.07 (0.06)

Concentration values correspond to the mean of three independent replicates. Corresponding standard deviation are given in brackets. ND: not determined, NQ: not quantified.

Page 109: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

104

Must/Wine A

-120

-100

-80

-60

-40

-20

0

20

40

60

80

100

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

209B

i

238U

Isotope/Element

Con

cent

ratio

n ch

ange

(%)

Must/Wine B

-120

-70

-20

30

80

130

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

209B

i

238U

Isotope/Element

Con

cent

ratio

n ch

ange

(%)

Must/Wine C

-120

-100

-80

-60

-40

-20

0

20

40

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

209B

i

238U

Isotope/Element

Con

cent

ratio

n ch

ange

(%)

Must/Wine D

-120

-100

-80

-60

-40

-20

0

20

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

209B

i

238U

Isotope/Element

Con

cent

ratio

n ch

ange

(%)

Must/Wine E

-120

-70

-20

30

80

130

180

230

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

238U

Isotope/Element

Con

cent

ratio

n ch

ange

(%)

Must/Wine F

-120

-70

-20

30

80

130

180

230

280

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

238U

Isotope/Element

Con

cenr

atio

n ch

ange

(%)

Must/Wine G

-120

-100

-80

-60

-40

-20

0

20

40

60

80

7Li

9Be

27Al

55M

n

57Fe

59C

o

60Ni

65C

u

66Zn

71G

a

75As

85R

b

88Sr

111C

d

115I

n

133C

s

138B

a

205T

l

208P

b

238U

Isotope/Element

Con

cent

ratio

n ch

ange

(%)

Page 110: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

105

Figure 1 – Concentration change (%) of contaminant elements from must to wine.

It is interesting to note that some elements showed different behaviour (throughout

winemaking process) depending on sample type: thus, red wines showed considerable higher

concentrations of Li, Fe, Sr, In and Bi than origin must; this trend was not observed for white

wines. An explanation for this could be related to the maceration of solid parts of red grapes.

Another hypothesis is the grape contamination with soil, a well know phenomenon (CURVELO-

GARCIA and CATARINO, 1998). In white samples, Li concentrations remain quite constant from

must to wines, accordingly to results reported by GÓMEZ et al. (2004). In the present study, Pb

was one of the few elements that increased concentration from must to wine (samples A and B),

suggesting the existence of a source of contamination, although not identified. However, the final

concentration of Pb is lower than the maximum accepted values allowed by the OIV (0.2 mg/L),

thus the contamination introduced should be of no concern. For some samples (C, D, E, F and G) a

main part of the Pb present in the must was removed, during alcoholic fermentation and

thefollowing vinification steps (Figure 1). Such a decrease has been previously reported by various

authors (TEISSÈDRE et al., 1993; HENICK-KLING and STOEWSAND, 1993). This elimination

may be attributed to the reaction of Pb with hydrogen sulphide formed during yeast fermentation:

Pb sulphide precipitate, is adsorbed to the yeast, and removed with the yeast lees. In addition, Pb

may also be removed after insoluble complexes with proteins and polysaccharides are formed. It is

also known that in wine a main part of the Pb is complexed with rhamnogalacturonan II

(PELLERIN et al., 1997).

Regarding the contents of elements both in must and wines, the ones could be considered at

low concentration, moreover when confronted with published data. Be, In and U occurred in sub-

trace levels. The following elements were found in lower concentrations than those reported on

literature: Be (THIEL et al., 2004), Al (MCKINNON et al., 1992; ESCHNAUER and

SCOLLARY, 1995; LOPEZ et al., 1998), Ni (TEISSÈDRE et al., 1998a), Sr (GREENOUGH et

al., 1997) and V (TEISSÈDRE et al., 1998b). Likewise other elements, levels of As and Cd were

very low in all wines, moreover very distant from the OIV maximum acceptable values (2005), of

0.2 mg/L and 0.01 mg/L, respectively). HERCE-PAGLIAI et al. (2002) have reported levels of As

in must and wines in the range of 2.10-14.60 µg/L, while JAGANATHAN (2001) and THIEL et al.

(2004) have found higher amounts (up to 22 µg/L). Recently, LIMA et al. (2004) observed 3-4

µg/kg and 0.58 µg/L of Cd in Azores must and wines, respectively. Pb levels were also low,

confirming a general tendency of decreasing in the levels of these elements observed in the last

years. In 1994, SUDRAUD et al. related values between 14 and 450 µg/L. Recently, LIMA et al.

(2004) reported an average concentration of 28.3 µg/L in Azores wines.

Page 111: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

106

The results obtained for the following elements are comparable to the amounts reported by

different authors: Mn (SUDRAUD et al., 1994; GREENOUGH et al., 1997; TAYLOR et al.,

2003), Co and Ga (TAYLOR et al., 2003; THIEL et al., 2004), Ba (TAYLOR et al., 2003), Tl

(GREENOUGH et al., 1997; TAYLOR et al., 2003), Bi (GREENOUGH et al. 1997; TAYLOR et

al., 2003) and Zn (SUDRAUD et al., 1994; GREENOUGH et al., 1997; SALVO et al., 2003). It

should be referred that Zn levels are far from the OIV maximum acceptable value (5 mg/L). In the

case of Rb, the levels are similar to those reported by SUDRAUD et al. (1994), in french wines, but

higher than those founded by TAYLOR et al. (2003) or THIEL et al. (2004). Must C, D and G

showed high levels of Cu, probably due to grape treatment with copper sulphate. Sample B should

be noted because it presents a peculiar behaviour: Cu concentration remains unchanged, perhaps

due to the pesticide type used in vineyard treatment. The OIV maximum acceptable value for this

element is of 1 mg/L.

The results obtained for Cs are of particular interest as they could be considered high

values (ESCHNAUER, 1982; GREENOUGH et al., 1997; TAYLOR et al., 2003; NICOLINI et al.,

2004). Finally, for some elements, namely Li, Mn, Zn, Sr, Cs and Ba, concentrations in must

samples seem to be a reflex of provenance soil.

3.1.4. CONCLUSIONS

The change in the element concentration from must to wine depends on the element, its

concentration and the sample type. Lower concentrations of the elements studied were found in

wines. The major part of these decreases can probably be explained by precipitation of the

elements, namely the heavy metals, as sulphured forms. Indeed, this emphasises the role played by

alcoholic fermentation in the elimination of toxic elements contained in the grapes (must). In some

cases, white and red samples presented different behaviour, suggesting the importance of

winemaking process. Nevertheless, some elements remain constant from must to wine, independent

of the winemaking process. Finally, the levels of all different elements were under the acceptable

maximal limits established by the OIV and in general can be considered low both for must and

wines.

Acknowledgments The authors are grateful to M. Cristina Clímaco, P. Clímaco and Goreti Botelho for providing the samples, to

“PARLE – Project A” for financial support and to “FCT” for a PhD scholarship (POCI 2010, medida IV.3.,

BD/17237/2004).

Page 112: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

107

3.1.5. REFERENCES

CABRERA-VIQUE C., Teissèdre P-L., Cabanis M-T. and Cabanis J-C., 1997. Determination and levels of chromium in French wine and grapes by graphite furnace atomic absorption spectrometry. J. Agric. Food Chem., 45, 1808-1811.

CATARINO S.C.G., 2000. Ocorrência de chumbo e cádmio em vinhos. MSc Thesis, Universidade do Porto, Portugal.

CATARINO S., Capelo J.L., Curvelo-Garcia A.S. and Vaião M., 2005. Focused ultrasound versus microwave digestion for the determination of lead in must by electrothermal-atomic absorption spectrometry. J. AOAC Int., 88 (2), 585-591.

CATARINO S., Curvelo-Garcia A.S., Bruno de Sousa R. (in press). Measurements of contaminant elements of wines by inductively coupled plasma mass spectrometry: a comparison of two calibration approaches. Talanta.

CURVELO-GARCIA A.S., 1988. Controlo de Qualidade dos Vinhos. Química Enológica. Métodos Analíticos. Instituto da Vinha e do Vinho, Lisboa, Portugal.

CURVELO-GARCIA A.S. and Catarino S., 1998. Os metais contaminantes dos vinhos. Origens da sua presença, teores, influência dos factores tecnológicos e definição de limites. Ciência Téc. Vitiv., 13 (1-2), 49-70.

ESCHNAUER H., 1982. Trace elements in must and wine: primary and secondary contents. Am. J. Enol. Vitic., 33 (4), 226-230.

ESCHNAUER H.R. and Scollary G.R., 1995. Aluminium in wein. Wein-Wiss., 1, 24-30.

FRANCESCO S., La Pera L., Di Bella G., Nicotina M. and Dugo G., 2003. Influence of different mineral and organic pesticide treatments on Cd(II), Cu(II), Pb(II), and Zn(II) contents by derivate potentiometric stripping analysis in Italian white and red wines. J. Agric. Food Chem., 51, 1090-1094.

GREENOUGH J.D., Longerich H.P. and Jackson S.E., 1997. Element fingerprint of Okanagan Valley wines using ICP-MS: relationships between wine composition, vineyard and wine colour. Austral. J. Grape Wine Res., 3, 75-83.

GÓMEZ M.D.L.C., Brandt R., Jakubowski N. and Anderson J.T., 2004. Changes of the metal composition in German white wines through the winemaking process. A study of 63 elements by inductively coupled plasma-mass spectrometry. J. Agric. Food Chem., 52, 2953-2961.

HENICK-KLING T. and Stoewsand G.S., 1993. Lead in wine. Am. J. Enol. Vitic. (Technical Brief), 44 (4), 459-463.

HERCE-PAGLIAI C., Moreno I., González G., Repetto M. and Camèan A.M., 2002. Determination of total arsenic, inorganic and organic arsenic species in wine. Food Addit. Contamin., 19 (6), 542-546.

JAGANATHAN J., 2001. A random testing of table wines for arsenic using electrothermal atomic absorption spectrometry. Atom. Spectrosc., 22 (2), 280-283.

LIMA M.T.R., Cabanis M.T., Matos L., Cassanas G., Kelly M. and Blaise A., 2004. Determination of lead and cadmium in vineyard soils, grapes and wines of the Azores. J. Int. Sci. Vigne Vin, 38, 163-170.

LOPEZ F.F., Cabrera C., Lorenzo M.L. and Lopez M.C., 1998. Aluminium levels in wine, beer and other alcoholic beverages consumed in Spain. Sci. Total Environ., 220, 1-9.

KRISTL J., Veber M. and Slekovec M., 2002. The application of ETAAS to the determination of Cr, Pb and Cd in samples taken during different stages of the winemaking process. Anal. Bioanal. Chem., 373, 200-204.

MAY T.W. and Wiedmeiyer R.H., 1998. A table of polyatomic interferences in ICP-MS. Atom. Spectrosc., 19 (5), 150-155.

MCKINNON A.J., Cattrall R.W. and Scollary G.R., 1992. Aluminum in wine – its measurement and identification of major sources. Am. J. Enol. Vitic., 43 (2), 166-170.

MÉDINA B., Augagneur S., Barbaste M., Grousset F.E. and Buat-Ménard P., 2000. Influence of atmospheric pollution on the lead content of wines. Food Addit. Contamin., 17 (6), 435-445.

NICOLINI G., Larcher R., Pangrazzi P. and Bontempo L., 2004. Changes in the contents of micro and trace elements in wine due to winemaking treatments, Vitis, 43 (1), 41-45.

Page 113: Tese doutoramento Sofia Catarino.pdf

Avaliação da composição de mostos e vinhos em metais contaminantes

108

OIV, 2005. Recueil des methodes internationals d’analyse des vins et des mouts. Organisation Internationale de la Vigne et du Vin, Paris.

PELLERIN P., O’Neill M.A., Pierre C., Cabanis M-T., Darvill A.G., Albersheim P. and Moutounet M., 1997. Complexation du plomb dans les vins par les dimères de rhamnogalacturonane II, un polysaccharide pectique du raisin. J. Int. Sci. Vigne Vin, 31 (1): 33-41.

RIBEREAU-GAYON P., Glories Y., Maujean A. and Dubourdieu D., 1998. Traité d’Oenologie. Chimie du vin, stabilisation et traitements. Dunod, Paris.

SALVO F., La Pera L., Di Bella G., Nicotina M. and Dugo G., 2003. Influence of different mineral and organic pesticide treatments on Cd(II), Cu(II), Pb(II), and Zn(II) contents determined by derivate potentiometric stripping analysis in Italian white and red wines. J. Agric. Food Chem., 51, 1090-1094.

STOCKLEY C.S. and Lee T.H., 1995. Much ado about lead in wine? An Australian review. J. Wine Research, 6 (1), 5-17.

SUDRAUD P., Médina B. and Grenon J.P., 1994. Teneurs en éléments minéraux des vins. J. Int. Sci. Vigne Vin, 28 (1), 69-75.

TAYLOR V.F., Longerich H.P. and Greenough J.D., 2003. Multielement analysis of Canadian wines by inductively coupled plasma mass spectrometry (ICP-MS) and multivariate statistics. J. Agric. Food Chem., 51, 856-860.

TEISSÈDRE P.L., Cabanis M.T., Daumas F. and Cabanis J.C., 1993. Evolution de la teneur en plomb au cours de l’elaboration dês vins dês cotes du rhone et de la vallee du rhone. R. F. OE. (cahier scientifique), 140 (Mars/Avril), 6-18.

TEISSÈDRE P.L., Cabrera Vique C., Cabanis M.T. and Cabanis J.C., 1998a. Determination of nickel in French wines and grapes. Am. J. Enol. Vitic., 49 (2), 205-210.

TEISSÈDRE P.L., Krosniak M., Portet K., Gasc F., Waterhouse A.L., Serrano J.J., Cabanis J.C. and Cros G., 1998b. Vanadium levels in French and Californian wines: influence on vanadium dietary intake. Food Addit. Contamin., 15 (5), 585-591.

THIEL G., Geisler G., Blechschmidt I. and Danzer K., 2004. Determination of trace elements in wines and classification according to their provenance. Anal. Bioanal. Chem., 378, 1630-1636.

Page 114: Tese doutoramento Sofia Catarino.pdf

Capítulo 4

CEDÊNCIA AO VINHO DE METAIS CONTAMINANTES DE

BENTONITES

Page 115: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

111

4.1. RELEASE OF CONTAMINANT ELEMENTS FROM BENTONITES TO WINE: EFFECT OF BENTONITE

CHARACTERISTICS AND WINE PH

S. Catarino1,2, M. Madeira3, F. Monteiro3, F. Rocha4, A.S. Curvelo-Garcia1, R. Bruno de

Sousa2 1INIAP, Estação Vitivinícola Nacional, 2565-191 Dois Portos. Portugal 2Instituto Superior de Agronomia, Departamento de Química Agrícola e Ambiental, 1349-017 Lisboa. Portugal 3Instituto Superior de Agronomia, Departamento de Ciências do Ambiente. Tapada da Ajuda, 1349-017 Lisboa. Portugal 4Universidade de Aveiro. Departamento de Geociências. Campus Universitário de Santiago, 3810-193 Aveiro. Portugal

Abstract

The aim of this study was to assess physical and chemical characteristics of several bentonites and measure their contaminant elements release to wine. Physical, chemical and mineralogical characteristics of six bentonites were assessed. Extraction essays using six bentonites in wine at three pH levels were carried out. The multilemental analysis of bentonites and extraction wines was performed by FAAS, ETAAS and ICP-MS. Bentonite addition resulted in significant higher concentrations of Li, Be, Na, Mg, Al, Ca, Sc, V, Mn, Fe, Co, Ni, Ga, Ge, As, Sr, Y, Zr, Nb, Mo, Cd, Sn, Sb, Ba, W, Tl, Bi and W. In contrast, concentrations of B, K, Cu, Zn and Rb in wine significantly decreased after bentonite addition. Distinct effects were observed for Pb concentrations in wines. In general, highest mineral enrichments were observed at pH 2.94. A strong correlation between Na concentrations of treated wines and its content in bentonite exchange complex was observed. Al and Fe contents reflected bentonite extractable aluminous and ferruginous constituents, while Be, Mg, Ca, V, Mn, Ni, Ge, Zr, Nb, Mo, Sn, Sb, Tl, Pb and U concentrations reflected the mineral composition of bentonites. None of the studied bentonites had the montmorillonite minimum content established by the OIV. Extractable Fe content of one bentonite sample largely surpassed the OIV maximum limit. The maximum soluble Na content defined by the OIV was exceeded by all of the natural bentonites. Results strongly suggest that montmorillonite content is not a satisfactory indicator of bentonite reactivity. Keywords: contaminant elements, wine, bentonite physical and chemical characteristics, pH, ICP-MS

Journal of Agricultural and Food Chemistry (em preparação para submissão)

Page 116: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

112

4.1.1. INTRODUCTION

Bentonite is a technological aid widely used in winemaking to remove or reduce the

concentration of undesirable constituents. In particular, it acts as a settling aid to remove proteins, thus

reducing the risk of protein haze in wine, whose occurrence could compromise its acceptance by the

consumer. Due to its mineral nature, bentonite treatment agent highly influences the final wine

concentration of mineral contaminants (Postel 1986, Enkelmann 1988, Wurzinger et al. 1994,

Gössinger et al. 1997, Machado-Nunes et al. 1998, Jakubowski et al. 1999, Catarino et al. 2004,

Nicolini et al. 2004, Catarino et al. 2006).

Bentonite is the commercial designation of an expansive clay material mainly composed by

montmorillonite (Lipka 1974, Maujean 1993, OIV 2003). In addition to montmorillonite, bentonite

could contain quartz, chalcedony, feldspars, calcite, dolomite, analcime, pyrite, among others, as

accessory minerals. Montmorillonite is a 2:1 dioctahedral smectite whose structure consists in two

tetrahedral silicon oxide sheets and one octahedral aluminium hidroxide sheet, combined as a

crystalline structural layer unit. Some Al3+ in octahedral positions is displaced by Mg2+, Fe2+, Fe3+ as

well as by other divalent cations, leading to charge imbalances in the structure (Maujean 1993). The

negative surface charge is mostly permanent and ranges from -0.25 to -0.6 (Z) according to smectite

type (Bailey 1980). In addition, bentonite structure also shows variable surface charge located in the

OH groups, which signal and intensity is a function of solution pH value (Greenland and Mott 1978).

The negative charge is balanced by exchangeable cations localized within the interlayer space or on

the external surface of the clay particles. These cations are mainly Ca2+, Na+ and Mg2+, but other

cations such as K+, Fe2+ and Cu+ are present in minor extent (Marchal et al. 1995). Both major and

minor cations ratio can significantly vary from one bentonite to another.

Bentonite nature can be changed through activation treatment, often employed on natural Ca

bentonites (high Ca2+/Na+ ratio). The activation process consists of a treatment of the wet mud by solid

Na2CO3 at a temperature of 80ºC, in order to obtain similar properties to natural Na bentonites (high

Na+/Ca2+ ratio), which have enhanced protein binding ability (Blade and Boulton 1988, Gougeon et al.

2003). For this reason, bentonites are classified according to the function of exchangeable cations into

three categories: natural Na bentonites, natural Ca bentonites and activated Ca bentonites.

The adsorption of positively charged proteins and other soluble cationic constituents by

bentonites in the wine is primarily due to the cationic exchange properties of these clays. The

adsorption of protein is affected through competition from other cations in the solution matrix, and

also affected by pH and ethanol content of solution (Blade and Boulton 1988). Competition would be

from K+, Ca2+, Mg2+, Na+, and H+, most amino acids, some peptides, and other cationic fractions. A

study developed by Blade and Boulton (1988), concerning several potential affecting factors, showed

that protein adsorption was independent of temperature, but varied with protein and ethanol contents,

and pH value.

Page 117: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

113

The effects of bentonite addition have been mostly studied in order to assess the depletion of

proteins, amino acids, bioamines, polyphenols and aroma compounds (Lipka 1974, Danilatos 1979,

Gorinstein et al. 1984, Vernhet and Bellon-Fontaine 1995, Barón et al. 1997, Machado-Nunes et al.

1998, Marchal et al. 2002, Balík 2003). Some studies concerning the variation of major elements

concentrations (Na, Ca, Mg), and also Al, Fe, Mn, Cr and Pb, when bentonites are used as fining

agents, have also been developed (Lipka 1974, Danilatos 1979, Ribereau-Gayon et al. 2000, Postel et

al. 1986, Enkelmann 1988, Wurzinger et al. 1994, Gössinger et al. 1997, Machado-Nunes et al. 1998).

Additionally, decreases of K, Cu, Zn and Rb concentrations in wine were reported by Lemperle et al.

(1988), Machado-Nunes et al. (1998) and Nicolini et al. (2004). Meanwhile, there is not complete

agreement among reported data, probably owing to different experimental conditions. Recently,

Gómez et al. (2004) and Nicolini et al. (2004) reported significant increases of micro and trace

elements in treated wine. Moreover, Jakubowski et al. (1999) reported that the concentration pattern of

the rare earth elements can be strongly affected by bentonite treatment and therefore is not suitable as

a fingerprint for the provenance of wines.

Although origin, type and label variety of bentonites on the market are considerable, a general

chemical and structural characterization and application instructions are usually supplied. However, no

information on the potential mineral release to wine is given. To date, the International Organisation

of Vine and Wine (OIV) defines a minimum content of montmorillonite and maximum contents for

extractable Pb, Hg, As, Fe, Al, Na and combined Ca and Mg (OIV 2003).

From an oenological point of view, the composition and structure of bentonites have been

studied and related to their protein adsorption capacities (Colagrande et al. 1973, Sudraud et al. 1985,

Maujean 1993, Marchal et al. 1995, Poinsaut and Hardy 1995). Nevertheless, similar studies regarding

mineral release to wine, namely micro and trace elements, have not been done so far. Thus, a study

was carried out to evaluate the release of contaminant elements from several bentonites to wine, and to

understand the effects of their physical and chemical characteristics. In addition, the effect of pH on

the release process was also studied.

4.1.2. MATERIALS AND METHODS

4.1.2.1. Bentonite characteristics

Six different bentonites (B1, B3, B4, B5, B8 and B9) supplied by four companies (in Portugal)

were used in the extraction essays (Table 1). Previously, some physical and chemical characteristics

were assessed to obtain more precise information on the studied bentonites. For that, each material

was forced to pass through a sieve of 2 mm. Particle size distribution was determined as described by

Póvoas and Barral (1992). Organic carbon content was determined by wet oxidation, by the Springer

and Klee method, following the methodology modified and reported by Póvoas and Barral (1992).

Page 118: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

114

Values of pH were measured in water (bentonite to water ratio 1:2.5) using a pH meter Metrohm 632.

Cation exchange capacity and exchangeable base cations (Ca, Mg, K and Na) were assessed by the

ammonium acetate (1M) adjusted to pH 7. The content of non-crystalline ferruginous and aluminous

constituents was evaluated by differential solubility of Fe and Al forms, by using the oxalic acid–

ammonium oxalate buffer (pH 3.2) (Schwertmann 1964, cited by Póvoas and Barral 1992). Base

cations and extractable Fe and Al were determined by Flame Atomic Absorption Spectrometry

(FAAS). Results were expressed on a dry basis (105ºC).

Table 1 - Characteristics of bentonites

Code Colour Form Type

B1 Beige Fine powder Activated Ca

B3 White ivory Granule Na

B4 Pale grey Granule Na

B5 Grey Powder Activated Ca

B8 White Elongated filament Na

B9 Grey Stony granule Na

For mineral constitution assessment, bentonites samples were grinded to a size <50 µm,

oriented on glass slides and then scanned, air-dried, after ethylene glycol solvation and after heating to

550 ºC. Scannings were performed using Cu Kα radiation (40 KV, 20 mA) and a 0.02º 2θ s-1 scanning

speed, 2 s counting time. X-ray diffractograms were recorded with a Philips X-ray diffractometer (a

PW 1732 generator, a PW 1050 goniometer, and a PV 1710 control unit, using the PC-APD software

version 3.6B).

External specific surface area was determined by the indirect method BET (Brunauer-

Emmett-Teller 1938) using a Gemini II 2370 apparatus. A continuous stream of N2 and He was passed

through a dry atmosphere previously cooled nearly to N2 boiling point. Then, a N2 single molecular

layer was adsorbed on bentonite particles, and its volume was quantified (N2 molecular surface is a

known characteristic). Subsequently, the temperature of bentonite samples was increased to room

temperature causing N2 release, and its volume was measured. The external specific surface area was

calculated applying the two N2 measured volumes and the BET equation. Due to their different

particle size, bentonite samples were fragmented using a mortar before analysis.

For chemical analysis, bentonite samples were grinded to a size <50 µm, dried at 105ºC and

separate in two parts. One part was used to determine the ignition loss (1000ºC, 30 min), and the other

for mineral composition analysis. Grinded bentonite samples were dissolved with a HNO3 (65% v/v)

Page 119: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

115

and H2O2 (30% v/v) solution, using a microwave apparatus CEM-MDS 2000. The dissolved material

was evaporated and once more dissolved with HCl 3M (Póvoas and Barral 1992). Ca, Mg, Na, K, Mn,

Fe and Al elements were determined by FAAS, and P by the molibdate method. All these elements

were expressed as oxides. The percentage of Si was estimated by the difference between 100 and the

sum of all other elements. Other elements (Li, Be, B, Sc, V, Mn, Co, Ni, Cu, Zn, Ga, Ge, As, Rb, Y,

Zr, Nb, Mo, Cd, In, Sn, Sb, Cs, Ba, W, Tl, Pb, Bi and U) were determined by ICP-MS (Inductively

Coupled Plasma Mass Spectrometry) semi-quantitative approach (Catarino et al. in press), using a

Perkin Elmer SCIEX Elan 9000 apparatus, equipped with a crossflow nebulizer, a Ryton Scott-type

spray chamber, nickel cones and a peristaltic sample delivery pump with four channels (Gilson)

(application software Elan – 6100 Windows NT - version 2.4). An autosampler Perkin-Elmer AS-93

Plus was protected by a laminar-flow-chamber clean room class 100 (Reinraumtechnik Max Petek).

Operating conditions used were as follows: RF power of 1200W; sample uptake rate of 0.85 mL/min;

nebulizer argon flow between 0.85 and 0.95 L/min. A full mass spectrum (m/z = 6-240, omitting the

mass ranges 16-18, 40, 41, 211-229) was obtained by full mass range scanning. A reference response

table (Perkin-Elmer TotalQuant III) was updated with a multi-element standard solution (Li, Be, Na,

Mg, Al, K, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Ag, Cd, In, Cs, Ba, Hg, Tl, Pb, Bi,

U) at 10 µg/L, from Perkin-Elmer. Rhodium (Rh) and rhenium (Re) were used as internal standards

for elements in the mass range (m/z) 7-138 and 205-238, respectively. The software of the instrument

performed automatic corrections of isobaric interferences. In order to get signal stabilisation, a sample

read delay of 75 s was chosen. Between samples or standards, the sampling system was rinsed with a

2% HNO3 (v/v), 1% C2H5OH (v/v), and 200 µg/L of Au solution for 75 s. The experimental conditions

used for the measurements were: dwell time = 50 ms; sweeps/reading = 6; reading/replicates = 1;

replicates = 1; time per run = 67 s. Only high purity reagents and deionised water (conductivity < 0.1

µS/cm) were used. All the material (polypropylene and Teflon PFA) was soaked in 20% HNO3 (v/v)

for at least 24 h and rinsed several times with deionised water, before use.

4.1.2.2. Extraction essays

Extraction essays were based in a factorial experiment, with two factors: bentonite (control,

B1, B3, B4, B5, B8 and B9) and wine pH value (2.94, 3.32 and 3.58). Each extraction was replicated

three times.

To simulate the natural pH range occurring in wine, three lots with different pH levels (2.94,

3.32 and 3.58) were obtained by adding tartaric acid (3.20 g/L and 0.55 g/L) and calcium carbonate

(0.80 g/L) to a white wine from 2001 vintage. Wine pH was controlled along the time and after

stabilisation extraction essays were carried out.

The dose of bentonite used in this study (25 g/L) was higher than that usually used in

winemaking, but lower than the concentration established by the OIV (2003) for bentonite quality

Page 120: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

116

control (50 g/L). This was to amplify the mineral release extent and consequently to detect eventual

effects of bentonites.

A sample of 5 g bentonite (105ºC drying) was shaken with 200 mL of wine in an Erlenmeyer

vessel for 15 min. After 24 hours of decantation, the suspension was centrifuged at 12000 rpm during

30 min, and the supernatant was used for measurements. Control samples (wine without bentonite

addition) were simultaneous and similarly prepared at the same pH, stirred and centrifuged. Before

use, the prolypropylene material was soaked in 20% HNO3 (v/v) for at least 24 hours and rinsed

several times with purified water. Measurements of Na, Mg, K, Ca and Fe were made by FAAS

(Catarino et al. 2003) and Al by ETAAS (Electrothermal Atomic Absorption Spectrometry) (Catarino

et al. 2002), while Li, Be, B, Sc, V, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Rb, Sr, Y, Zr, Nb, Mo, Cd,

In, Sn, Sb, Ba, W, Tl, Pb, Bi and U were analysed by the ICP-MS semi-quantitative approach as

described before (Catarino et al. in press).

Statistical analysis The variance analysis was performed using Statistica vs’ 98 edition (Statsoft Inc., E.U.A.).

4.1.3. RESULTS AND DISCUSSION

4.1.3.1. Physical and chemical characteristics of bentonites

External specific surface area, particle size distribution, organic carbon content, pH values and

exchange complex characteristics of studied betonites are shown in the Table 2.

Table 2 - Physical and chemical characteristics of bentonites: external specific surface, particle size distribution, organic carbon level, pH and exchange complex characteristics. CS – coarse sand; FS –

fine sand; SL – silt; CL– clay; SB –sum of bases; CEC – cationic exchange capacity; SB/CEC – cationic exchange proportion on exchange complex

Particle size distribution pH Exchange complex

Bent.

Specific

surface CS FS SL CL

C

Org H2O Ca Mg K Na SB CEC

m2/g --------------------cmolc/kg-----------------------

SB/CEC

% -------------------g/kg -------------------

B1 16.0591 0.1 199.7 239.7 560.5 1.9 9.70 43.0 17.0 0.93 61.3 122.3 80.4 152.1

B3 26.9062 0.1 79.4 270.2 650.3 0.7 10.21 12.6 13.5 1.56 69.1 96.8 59.9 161.7

B4 20.4110 0.1 65.7 307.4 626.8 1.4 10.31 16.8 9.9 1.68 74.4 102.8 57.0 180.3

B5 38.3820 2.1 160.6 179.9 657.4 3.1 10.06 29.7 15.8 1.62 73.4 120.4 72.5 166.0

B8 31.1781 0.2 63.7 308.7 627.4 2.3 10.22 25.8 14.7 1.77 90.4 132.6 58.4 226.2

B9 13.4071 0.1 69.5 139.4 791.0 2.5 9.67 20.9 8.4 1.13 51.1 81.5 72.5 111.9

Page 121: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

117

Concerning particle size distribution, studied bentonites showed negligible content of coarse

sand; the B5 and specially the B1 samples showed high contents of fine sand, while the B9 contained

low contents of fine sand and silt, and the highest content of clay, suggesting a high reactivity. It is

noteworthy that none of the studied bentonites contained a montmorillonite minimum content (80%)

established by the OIV (2003). Admitting that the clay fraction was exclusively constituted by

montmorillonite, only the B9 showed a close content (79%).

A precise evaluation of all mineral constituents of bentonite is required to assess their

characteristics and behaviour. Diffractograms of bentonites after ethylene glycol solvation are shown

in Figure 1. As expected, minerals of the smectite group were predominant in all bentonite samples,

specially montmorilonite, and probably also beidelite in the B1 and B5 samples. In addition, other

minerals in variable proportions were present, such as quartz and probably opal in the samples B3, B4,

B5 and B8, calcite in the B1, B3 and specially B5, plagioclase in the B1 and B3, and halite in the B5.

The B1 differed from the other samples due to its high proportion of inert constituents (approximately

20%). On the contrary, the B9 showed a smectite high proportion, with inert constituents being less

than 5%. Moreover, the bentonites may also contain amorphous materials (not detected by the applied

method), susceptible to release Ca, Mg or Na, and also minor and trace elements.

Figure 1 - RX diagram of bentonites (after ethylene glycol solvation), E-Esmectite; O-Opal; Q-Quartz; P-Plagioclase; C-Calcite; H-Halite.

Page 122: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

118

The values of external specific surface area ranged from 13.4 to 38.3 m2/g, respectively, in the B9

and B5 samples. Differences between samples are not relevant, and therefore can not explain distinct

mineral release capacity by itself. The B9 sample showed the highest montmorillonite content and the

lowest value of external specific surface area, suggesting that clay fraction characteristics are more

important than its content. Information on the internal specific surface area would be necessary to get

deeper understanding on bentonite reactivity.

It is understood that each clay mineral presents a characteristic range of cationic exchange

capacity values. Since bentonites are complexe combinations of various minerals the results are

difficult to interpret. For example, the B1 showed the highest cationic exchange capacity (80.4

cmolc/kg), followed by the B5 and B9 (72.5 cmolc/kg), whereas the B3, B4 and B8 samples presented

considerable lower cationic exchange capacity (57.0-59.9 cmolc/kg). Bearing in mind the particle size

distribution of bentonites these results are interesting. The bentonite with the lowest clay content (B1)

showed the highest cation exchange capacity (CEC), while the B9 with the highest clay content (79%),

suggesting a high reactivity, exhibited a low CEC value. These results are probably related to

montmorillonite structural particularities and suggest that its content is not a satisfactory indicator of

bentonite reactivity. Thus, montmorillonite content is not sufficient for bentonite characterization.

Concerning exchangeable bases (Table 2), Ca content in the exchange complex seemed to be

related to the bentonite nature. In fact, activated bentonites showed the highest concentrations, mainly

in the B1. On the other hand, the highest content of Na was observed in the B8, a natural Na bentonite.

As expected, low values of organic carbon were observed (Table 2). Nevertheless, the B5 sample

showed the highest content (3.1 g/kg), which was nearly four times the content determined for the B3

sample. As well known, organic material contains mineral elements by complexation phenomena,

which are susceptible to react in solution.

Bentonite elemental composition (Ca, Mg, Na, K, Mn, Fe, Al and P elements), expressed as

oxides, is shown in Table 3. High CaO contents in the B1 and B5 samples (17.3% and 16.3%,

respectively) are in agreement with the presence of calcite (Figure 1), and the high NaO content in the

B1 can be related to the presence of plagioclase (albite). Considering the mineralogical composition of

bentonites, there is no explanation either for the high content of MgO in the B3 and B4 samples

(17.0% and 13.1%, respectively) or for the level of NaO in the B4 (11%). A possible explanation for

these contents is that the B3 and B4 samples probably contain trioctahedral smectites (magnesian

smectites).

In relation to extractable Fe contents (Table 3), strong differences were observed between the B1

and the B5 samples, showing the minimum and maximum values, respectively. It should be

emphasized that the content of extractable Fe of the B5 largely exceeded the threshold of 600 mg/kg

established by the OIV (OIV 2003). Extractable Al contents ranged from 0.58 g/kg (B4) to 1.11 g/kg

(B8), being lower than the limit of 2.5 g/kg established by the OIV (2003). Contrarily to extractable

Al, extractable Fe was correlated with the total Fe in the bentonite sample (r = 0.740, P = 0.05).

Page 123: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

119

Table 3 - Ca, Mg, Na, K, Mn, Fe, Al, P contents in bentonites, expressed as oxides, Si was obtained by difference between 100% and the sum of the other elements percentage, ignition loss and non-

crystalline Fe and Al

Non-

crystalline

Fe and Al Bent. CaO MgO NaO K2O MnO Fe2O3 Al2O3 P2O5

Ignition

loss Partial SiO2

Fe Al

% % g/kg

B1 17.31 0.71 3.79 0.27 0.05 2.85 32.32 0.04 10.57 67.91 32.09 0.19 0.89

B3 6.01 16.98 11.57 0.36 0.03 1.87 18.72 0.05 6.46 62.05 37.95 0.28 0.75

B4 5.95 13.08 10.95 0.71 0.03 1.56 17.60 0.03 10.84 60.72 39.28 0.25 0.58

B5 16.29 1.49 3.88 1.02 0.12 3.85 28.89 0.11 8.67 64.32 35.68 1.39 0.93

B8 2.96 1.34 3.19 0.16 0.02 1.46 13.52 0.46 6.41 29.52 70.48 0.36 1.11

B9 3.06 0.72 2.80 0.34 0.02 3.35 16.89 0.05 6.91 34.14 65.86 0.58 0.86

The results regarding some minor and trace elements are reported in Table 4. For the most part of

the studied elements, including Al and Fe, the highest contents were observed in the B1, B5 and B9

samples.

4.1.3.2. Mineral contents in bentonite treated wines

The composition of control and bentonite treated wines are shown in Table 5. The

interpretation of the results should take into account that they were obtained under experimental

conditions, with a bentonite dose higher than that commonly used in winemaking.

Enrichment of mineral elements Wines treated with bentonites showed significantly different concentrations of all the minerals

elements studied (P = 0.05) compared to the controls (Table 5). Bentonite treatment resulted in

significantly higher concentration of a large group of elements (Li, Be, Na, Mg, Al, Ca, Sc, V, Mn, Fe,

Co, Ni, Ga, Ge, As, Sr, Y, Zr, Nb, Mo, Cd, Sn, Sb, Ba, W, Tl, Bi and W), some of them contaminants

legally controlled. Calcium concentration in wines increased up to 220 µg/L (B1), from 54% to 237%,

which is very interesting as Ca concentration affects the tartaric stability of wine. Mg concentrations

increased up to 126%. As expected, Na concentrations in wines showed increases from 306 to 435

µg/L (corresponding to 339% and 482%). These increments are important because Na enrichment in

wine could carry legal implications (OIV 2005).

Page 124: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

120

Table 4 - Multielemental composition of bentonites expressed as µg/g (average of three replicates)

B1 B3 B4 B5 B8 B9

B1 B3 B4 B5 B8 B9

Li 27.9±0.3 7.7±0.6 6.02±0.06 12.3±0.3 6.1±0.4 16.4±0.1 Zr 142±2 90±9 82±1 122±6 84±3 116.1±0.4

Be 2.5±0.4 1.7±0.1 1.2±0.2 1.6±0.2 0.95±0.04 1.2±0.2 Nb 13.9±0.3 7.4±0.8 5.83±0.02 6.0±0.3 5.0±0.2 19.4±0.1

B 156±4 333±38 127.5±0.5 483±31 236±9 153.8±0.6 Mo 0.51±0.04 0.9±0.1 0.84±0.03 1.63±0.04 0.67±0.05 3.3±0.2

Sc 1.30±0.04 0.77±0.08 0.63±0.01 1.7±0.1 0.54±0.04 0.72±0.03 Cd 0.196±0.005 0.06±0.06 0.08±0.01 0.09±0.02 0.09±0.05 0.109±0.008

V 25.7±0.2 31±3 29.4±0.2 114±7 28.6±0.5 14.5±0.2 In 0.058±0.007 0.018±0.006 0.010±0.001 0.0305±0.0007 0.02±0.02 0.062±0.004

Mn 278±4 177±19 178.8±0.2 562±21 122±4 56.9±0.4 Sn 3.7±0.2 5.2±0.9 3.0±0.2 1.5±0.2 6.4±0.2 7.72±0.09

Co 2.65±0.05 3.6±0.5 2.89±0.06 8.9±0.4 2.8±0.2 0.88±0.04 Sb 1.36 0.9±0.1 1.38±0.09 2.7±0.2 0.80±0.04 2.74±0.03

Ni 3.09±0.09 18±2 17.3±0.3 15.0±0.7 17.3±0.6 5.8±0.2 Cs 0.86±0.04 10±1 12.7±0.2 11.4±0.3 13.9±0.4 0.38±0.01

Cu 9.1±0.5 23±2 14.5±0.3 20±1 9.8±0.5 6.1±0.3 Ba 1518±41 640±95 1163±8 706±24 357±17 931±5

Zn 113±27 66.0±0.7 56±10 87±6 73±14 127±4 W 0.80±0.01 0.7±0.2 0.72±0.04 0.793±0.004 0.54±0.01 0.24±0.02

Ga 16.1±0.2 10±1 9.4±0.1 13.4±0.6 9.3±0.4 19.41±0.09 Tl 0.25±0.01 0.8±0.1 1.21±0.02 1.11±0.03 1.33±0.07 0.447±0.007

Ge 0.41±0.01 0.159±0.006 0.16±0.01 0.25±0.01 0.112±0.008 0.134±0.006 Pb 46.9±0.1 19.2±0.4 30.9±0.4 18.7±0.3 25±1 70±1

As 39.1±0.9 8.5±0.6 32.1±0.6 9.1±0.3 4.5±0.2 38.2±0.6 Bi 2±2 0.46±0.06 0.7±0.2 0.42±0.08 2±2 1.7±0.1

Rb 8.2±0.1 35±4 44.4±0.3 47±2 26±1 9.09±0.03 U 5.0±0.2 34.9 26.57 52.262 46±2 -

Y 109±45 13.8±0.8 103±16 - 91±47 101±1

Page 125: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

121

It should be noted that none of the natural bentonites (B3, B8 and B9) respected the maximum

soluble sodium content defined by OIV of 10 g/kg (OIV 2003). As regards to some minor elements in

wine with relevant technological interest, remarkable Al enrichments up to 19.77 mg/L (16 times)

were observed. Fe enrichments varied between 0.46 and 6.43 mg/L (25% and 350%). Bearing in mind

the potential Fe precipitation phenomena, these increments are relevant. Extractable Al and Fe

contents from bentonites were in accordance with the limits defined by the OIV (2.5g/kg and 600

mg/kg, respectively). The introduction of 2400 µg/L of Mn caused by the B5, representing an increase

of 523%, should be pointed out.

As regards to the following elements, including several with toxicological interest, the content

increases were up to: 39 µg/L of Li (corresponding to an enrichment of 366%, caused by the B9); 15

µg/L of Be (B1); 41 µg/L of V (763%, B5); 24 µg/L of Co (B3); 77 µg/L of Ni (331%, B4); 1211

µg/L of Sr (763%, B9); 8 µg/L of Y (B9); 391 µg/L of Zr (B1); 30 µg/L of Nb (B9); 24 µg/L of Mo

(B9); 9 µg/L of Sn, (B9); 8 µg/L of Sb (500%, B9); 239 µg/L of Ba (608%, B5); 6 µg/L of Tl (B4) and

111 µg/L of U (B9). The concentration of Be, Co, Y, Zr, Nb, Mo, Sn, Tl and U increased by more than

1000%. Significant but few relevant increases of Sc, Ga, Ge, In, W and Bi in wines were observed.

Concerning elements under legal control, particular attention must be paid to As remarkable

enrichments, from 2.1 to 46.71 µg/L (163 to 3621%), specially from the B9. Bearing in mind the OIV

maximum contents for extractable As (2 mg/kg), the B9 should be signalised as being responsible for

a As release of 1.87 mg/kg. The maximum acceptable value for As in wine is of 0.2 mg/L (OIV 2005).

Cd introductions were technologically irrelevant as final concentrations were very distant from the

legal limit (0.01 mg/L). All bentonites were in conformity with OIV maximum contents for extractable

Pb (5 mg/kg). Distinct bentonite effects were observed on Pb wine contents: enrichments were

promoted by the B1 (5 µg/L) and mainly by B9 (43 µg/L). The maximum acceptable value for Pb in

wine is of 0.20 mg/L (OIV 2005). In contrast, the B4, B5 and B8 leaded to depletion of 4, 14 and 5

µg/L, respectively. From a technological point of view, these observations are exciting. The most

reasonable explanation for distinct trends is that Pb release from B4, B5, and B8 bentonites to wine

was lower than its depletion due to the formation of proteins and polysaccharides complexes. In fact,

the B4, B5 and B8 samples showed low amounts of Pb in their composition, as compared with the

others. On a distinct application field but somehow related with these observations, bentonite ability to

remove certain heavy metals from liquid waste have been observed (Kayabali and Kezer 1998).

Page 126: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

122

Table 5 - Effect of the bentonite and pH value on the multielemental compositon of wine (the results are given in µg/L, except for Na, Mg, Al, K, Ca and Fe which are expressed as mg/L)

Element pH effect pH level (n=7) Bent.

effect Bentonite (n=3)

2.94 3.32 3.58 Control B1 B3 B4 B5 B8 B9

pH –bentonite interation

Li ** 20.31b 16.50a 16.53a ** 10.71a 19.53b 11.39a 10.99a 11.02a 10.90a 49.90c **

Be ** 11.31c 8.17b 5.53a ** 0.09a 15.18f 9.12d 6.78b 8.92d 7.98c 10.28e **

B ** 1342.9c 1234.6b 1133.9a ** 1500.5e 1314.2d 1234.5c,d 1272.4d 1121.2a,b 1163.6b,c 1053.6a **

Na n.s. 421.9 419.9 428.2 ** 90.2a 419.8c 490.8d 517.1e 525.1e 524.5e 395.8b n.s.

Mg n.s. 112.19 112.48 113.05 ** 69.11a 98.89b 156.22f 152e 131.78d 109.33c 70.67a **

Al ** 14.383c 13.163b 12.229a ** 1.223a 15.107c 14.394c 12.017b 12.725b 20.988e 16.355d **

K ** 491.4a 762.0b 783.9c ** 865.3f 563.6a 694.7e 682.3d,e 627.3b 672.9d 647.6c **

Ca ** 157.5a 154.7a 252.4b ** 92.9a 313.2f 175.8d 163.1c 268.9e 143.3b 159.8c **

Sc * 1.40b 1.39b 1.17a ** 1.09b 2.08d 1.51c 1.25b,c 1.45c 0.79a 1.08b **

V ** 28.89c 22.04b 16.61a ** 5.32a 32.70d 25.14c 24.85c 45.91e 17.63b 5.76a **

Mn ** 1675.61c 1429.67b 1101.64a ** 460.34a 1519.10c 1814.92d 1947.65e 2868.29f 750.29b 455.56a **

Fe ** 5.11b 6.02c 3.93a ** 1.84a 2.30b 4.56c 4.95d 8.27b 6.12e 7.10f **

Co ** 19.04c 15.80b 13.00a ** 1.92a 6.79b 26.43e 24.31d 25.88e 20.48c 5.81b **

Ni ** 64.28c 55.55b 48.05a ** 23.18a 25.89a 94.91e 99.95f 49.28c 66.00d 32.52b **

Cu ** 23.45c 18.79b 11.48a ** 57.23e 13.45c 11.64b,c 12.35c 9.22b 17.44d 4.01a **

Zn ** 580.19c 477.67b 377.64a ** 623.22e 390.11a 489.27d 487.37c,d 413.25a,b 446.40b,c 499.86d **

Page 127: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

123

Ga ** 2.19c 1.80b 1.50a ** 1.01a 2.03e 1.98d,e 1.86c.d 1.63b 2.35f 1.83c **

Ge ** 0.056b 0.041a 0.040a ** 0.014a 0.066d 0.037b 0.031b 0.046c 0.030b 0.097e **

As ** 22.30b 15.90a 15.13a ** 1.29a 3.39b 12.84c 17.22d 25.24e 16.40d 48.05f **

Rb ** 382.89a 433.04b 389.76a ** 794.37d 266.71a 408.92c 422.63c 285.70a 349.17b 285.78a **

Sr ** 700.79b 568.29a 714.24b ** 158.55a 885.85d 560.64b,c 543.20b 605.13c 505.13b 1369.24d **

Y ** 4.58b 1.25a 5.64b ** 0.04a 3.95c,d 4.67d 3.71c 2.10b 3.88c.d 8.39e **

Zr ** 231.76c 160.98b 98.22a ** 1.47a 392.25g 119.55c 76.55b 138.69d 150.51e 266.54f **

Nb ** 9.89c 6.75b 4.80a ** 0.03a 16.86b 0.56a 0.29a 0.11a 0.13a 30.01c **

Mo ** 8.01b 5.08a 4.96a ** 1.33a 2.09b 3.29c,d 3.25c,d 3.76d 2.76c 25.63e **

Cd ** 0.691b 0.584a 0.712b ** 0.218a 0.998e 0.563c 0.750d 0.455b 0.985e 0.667d **

In n.s. 0.07a 0.04a 0.06a ** 0.01a 0.02a 0.01a 0.01a 0.01a 0.01a 0.33b n.s.

Sn ** 3.42c 1.69b 0.90a ** 0.70b 0.71b 0.88c,d 0.75b,c 0.54a 0.94d 9.50e **

Sb ** 4.47c 3.07b 2.65a ** 1.69a 2.74b 1.87a 1.86a 3.67c 1.81a 10.13d **

Ba ** 102.59a 109.24b 122.93c ** 39.31a 76.90c 184.45e 278.50f 60.34b 97.79d 43.83a **

W ** 0.91b 0.89b 0.70a ** 0.25a 1.55f 1.33e 1.61g 0.46d 0.35c 0.31b **

Tl ** 2.21a 2.33b 2.25a ** 0.12a 0.59b 2.86e 3.11f 2.12d 5.59g 1.42c **

Pb ** 33.55c 25.24b 15.79a ** 21.49c 26.25d 20.27c 17.35b 8.01a 16.55b 64.08e **

Bi ** 1.44b 1.41b 1.14a ** 1.02a,b 2.08e 1.51d 1.33c.d 1.45d 0.79a 1.27b,c **

U ** 43.96a 52.06b 57.35c ** 0.58a 8.09b 22.62c 35.73d 108.10f 71.33e 111.41f **

Means followed by the same letter in a line are not significantly different at the 0.05* or 0.01** level of significance; n.s. = without significant difference.

Page 128: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

124

Depletion of mineral elements

Concentrations of B, K, Cu, Zn and Rb in wine significantly decreased after bentonite

treatment (Table 5). Variations of K (-20% to -35%), Zn (-20% to -37%) and mainly Cu (-70% to -

93%) were very interesting, as they contribute to wine tartaric stabilization and to limit Cu wine haze.

It should be mentioned that Cu and Zn concentrations in wine are legally controlled with maximum

acceptable values of 1 and 5 mg/L, respectively (OIV 2005).

Probably, K is removed from wine through an exchange with Na and Ca in exchange positions

(among other cations). This hypothesis is supported both by its natural high concentration in wine and

by its high affinity to aluminium silicates structure. Similar explanation is credible for Rb decrease,

likewise K an element of low ionic potential. Proteins and K competition to bentonite adsorption have

been previously reported (Blade and Boulton 1998).

Depletion of B, Cu and Zn may be related to their complexation with condensed tannins,

which are in part bound to proteins. Additionally, Zn could be complexed with dimmers of

rhamnogalacturonan II, a grape pectic polysaccharide (Pellerin et al. 1997). Bentonites capacity to

indirectly prevent wine Cu haze was already reported (Peynaud 1996, Ribéreau-Gayon et al. 2000).

4.1.3.3. Physical and chemical characteristics of bentonites and mineral release

Concerning the studied elements, significant differences among bentonites were found,

reinforcing the importance of bentonite chemical and structural characteristics on extraction

phenomena. The B1 sample showed a high release of Li, Be, Ca, Sc, V, Zr, Nb, Cd and Bi, and a

depletion of K, Zn and Rb. These results are explained by its relative high cationic exchange capacity,

calcite presence and by its high proportion of inert constituents. The B3 and B4 should be noticed by

the high introduction of Mg in wine, related to the probable presence of magnesian smectites. Ni and

Ba releases were also relevant for those bentonites. The B3 and B4 were responsible for a low removal

of K, Zn and Rb, which is in agreement with their low cationic exchange capacity. The high Ca release

from the B5 is ascribed to the presence of calcite. The remarkable introduction of V, Mn, Fe, Co, As

and U is probably due to the presence of non crystalline materials. Moreover, K concentration in wine

showed a strong decrease reflecting an important cationic exchange capacity. A remarkable addition of

Al was observed for the B8. Ga and Tl introductions were also highest in this bentonite. Finally, the

B9 was characterized by Be, Fe, Y, Zr release and in particular Li, As, Sr, Nb, Mo, Sn, Sb, Pb and U

introductions. This bentonite showed a high organic C content, a smecticte high proportion, an

important CEC and the lowest Ca, Mg, K and Na proportion in the exchange complex. Relevant

amounts of B, Cu and Rb were removed.

The correlation between the concentration of each element in treated wine and its content in

bentonites was assessed (P = 0.05). Na concentrations in wines were positively correlated with its

proportion in bentonite exchange complex (r = 0.881), but not with its total content in bentonites. In

fact, part of the Na is included in plagioclase mineral, thus being unattainable. Increments of Al and Fe

Page 129: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

125

in wines were positively correlated with extractable Al (r = 0.784) and Fe (r = 0.808) constituents in

bentonites. Regarding bentonite multielemental composition, correlation values higher than 0.800

were found for: Be (r = 0.857); Mg (r = 0.820); Ca (r = 0.981); V (r = 0.812); Mn (r = 0.814); Ni (r =

0.867); Ge (r = 0.988, if B9 is excluded); Zr (r = 0.836); Nb (r = 0.992); Mo (r = 0.948); Sn (r = 0.967,

if B9 is excluded); Sb (r = 0.920, if B9 is excluded); Tl (r = 0.839); Pb (r = 0.933) and U (r = 0.872).

Correlation values between 0.600 and 0.700 were observed for Sc, Co, In and W. Finally, considering

Li, Ga, As, Y, Cd with B1 exclusion, and Ba and Bi elements, correlation values lower than 0.420

were found.

4.1.3.4. Interaction between wine pH and bentonites

The cationic exchange capacity of bentonites is not significantly affected by the pH range

studied. As previously mentioned, the negative charge resultant from isomorphic substitutions in

smectites is constant and not affected by pH variation. However, the surface charge is slightly variable

in the studied pH range (Abreu 1994). Thus, at low pH values the development of some positive

charge begins, while at high pH values the negative charge is more intense. In fact, the protein

adsorption process is influenced by the pH value, which can also change the structural conformation

and the electric charge of proteins. A significant increase in adsorption capacity at pH 3.8 was

observed by Blade and Boulton (1988).

It is known that at lower pH values more protons are available in solution. Additionally, as already

stated, there is a strong preference for hydrogen adsorption. Therefore, at low pH values the adsorption

extent of competitor cations increases.

Mineral contents in wines treated with bentonite was significantly affected by the pH value

(2.94 – 3.58 range) (Table 5), which is in disagreement with observations reported by Nicolini et al.

(2004). Most of the elements presented the highest concentration at pH 2.94. Concerning some

elements contained in the crystalline structure, such as Al, this tendency could be in part related to

structure degradation, which could occur at these low pH levels. On the other hand, at highest pH

value some losses by insolubleness and precipitation may occur.

The following exceptions were observed: Ca, Ba and U concentrations increased with pH

value; a peculiar behaviour was shown by Sr, Zr, Cd and Tl with the highest concentration both at pH

2.94 and 3.58, but with little expressive differences comparatively to pH 3.32; Fe highest

concentration was observed at pH 3.32 probably as a consequence of precipitation phenomena at pH

3.5. On the other hand, no significant effect of pH was observed for Na, Mg and In contents. In some

cases, important differences were observed: Be, Zr, Nb and Pb average amounts at pH 2.94 are twice

as much than at pH 3.58; V, Co, Ni, Zn, As, Mo, Sb contents are also strongly affected by wine pH. It

is woth noting the drastic pH effect on Sn content in wines: at pH 2.94 the concentration was fourfold

at pH 3.58.

Page 130: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

126

With respect to B, K, Cu, Zn and Rb elements, whose concentrations decreased after bentonite

application, differences of each element level were observed among the control wines, probably due to

precipitation at pH 3.58. So, in order to enable a rigorous data interpretation, variance analysis of the

differences between each modality and respective control sample was carried out (results not shown).

With regard to B, Cu, Zn and Rb, lower contents were observed at pH 3.32, while for K the lower

contents were observed at pH 3.58, which could be explained by its precipitation with tartaric acid.

It should be stressed that the previous statements are based on average values obtained from

seven modalities (control wine and bentonite treated wines) at the same pH. In fact, as shown in Table

5, interaction between bentonite and pH was observed. Not all the bentonites showed the highest

release at the same pH value. For instance, at pH 3.32, the B9 and B1 samples showed the lowest and

the highest Al release, respectively. On the other hand, in some cases, the pH value had no effect on

element release, such as Al release from the B3. Consequently, our data show that the pH effect on

bentonite mineral release depends on its chemical and structural characteristics.

4.1.4. CONCLUSIONS

This study proves that bentonite affects the mineral composition of wines. Bentonite

application resulted in significantly higher concentrations of Li, Be, Na, Mg, Al, Ca, Sc, V, Mn, Fe,

Co, Ni, Ga, Ge, As, Sr, Y, Zr, Nb, Mo, Cd, Sn, Sb, Ba, W, Tl, Bi and W. In contrast, the

concentrations of B, K, Cu, Zn and Rb in wine have significantly decreased. Distinct effects were

observed for Pb wine contents, suggesting the bentonite ability to its removal, which is remarkable

from a technological point of view. As a general trend, the highest transfer of elements from bentonite

to wine was observed at the lowest pH value. This information should be taken into account to

minimize the risk of contamination during the winemaking process.

Distinct effects among bentonites were observed, indicating the importance of their physical

and chemical characteristics on extraction phenomena, thus suggesting the need for more information

about the trade mark products. Sodium concentration of wine was strongly correlated to its content in

the bentonite exchange complex, while the increment of Al and Fe is associated with bentonite

extractable Al and Fe constituents. Concentrations of Be, Mg, Ca, V, Mn, Ni, Ge, Zr, Nb, Mo, Sn, Sb,

Tl, Pb and U in treated wines were related to their content in bentonites.

Some bentonite characteristics were in disagreement with OIV specifications. Indeed, none of

the bentonites contained the required minimum content of montmorillonite. All studied bentonites

exceeded the allowed maximum content of soluble Na, while that of extractable Fe was exceeded only

in one of them. The B9 was responsible for a As release of 1.87 mg/kg, very near from the defined

maximum content of extractable As. On the other hand, assuming that clay fraction properties were

more important than its content, the montmorillonite content did not seem to be a satisfactory indicator

of bentonite reactivity.

Page 131: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

127

Finally, the wide diversity in the composition of wines (specially with respect to mineral,

protein and polysaccharide contents) is an additional influent factor on bentonite mineral release.

Acknowledgements The authors are grateful to AEB Bioquímica, Ecofiltra, Meller & Essink, Proenol and Soeno for providing the

bentonite samples, “PARLE – Project A” for financial support and “Fundação para a Ciência e a Tecnologia” for

a PhD scholarship (POCI 2010, medida IV.3, BD/17237/2004).

4.1.5. REFERENCES Bailey, S.W. (1980) Summary and recommendations of AIPEA nomenclature committee. Clays and Clay Minerals 28, 73-78.

Balík, J. (2003) Effect of bentonite clarification on concentration of anthocyanins and colour intensity of red and rosé wines. Journal of the American Society for Horticultural Science 30 (4), 135-141.

Barón, R., Mayen, M., Merida, J. and Médina, M. (1997) Changes in phenolic compounds and colour in pale Sherry wines subjected to fining treatments. Zeitschrift für Lebensmitteluntersuchung und -Forschung A 205, 474-478.

Blade, W.R. and Boulton, R. (1988) Adsorption of protein by bentonite in a model wine solution. American Journal of Enology and Viticulture 39 (3), 193-199.

Catarino, S., Curvelo-Garcia, A.S. and Bruno de Sousa, R. (2002) Determination of aluminum in wine by graphite furnace AAS: validation of analytical method. Atomic Spectroscopy 23 (6), 196-200.

Catarino, S., Pinto, D. and Curvelo-Garcia A.S. (2003) Validation and comparison of analytical methods by flame atomic absorption spectrometry for determination of copper and iron in wines and brandies. Ciência e Técnica Vitivinícola 18 (2), 65-76.

Catarino, S., Soares, J., Curvelo-Garcia, A.S. and Bruno de Sousa, R. (2004) Influence of bentonites on mineral composition of wine: potassium, sodium, calcium, aluminium and lead. pH effect. Ciência e Técnica Vitivinícola 19 (1), 29-45.

Catarino, S., Madeira, M., Monteiro, F., Curvelo-Garcia, A.S. and Bruno de Sousa, R. (2006) Release of contaminant elements from bentonites to wine: a contribution to achieve a test solution. Ciência e Técnica Vitivinícola 21 (1), 17-31.

Catarino, S., Curvelo-Garcia, A.S. and Bruno de Sousa, R. (2006). Measurements of contaminant elements of wines by inductively coupled plasma mass spectrometry: a comparison of two calibration approaches. Talanta 70, 1073-1080.

Colagrande, O., Griselli, F. and Del Re, A.A. (1973) Étude des phenomenes déchange lors de l’emploi des bentonites oenologiques. I – Les echanges de sodium et de proline. Connaissance de la Vigne et du Vin 2, 93-106.

Danilatos, N. (1979) Données récentes sur l’emploi des bentonites. Bulletin de l’OIV 580, 456-481.

Enkelmann, R. (1988) Migration of heavy metals from bentonites to wine.1. Note: Bentonites. Deutsche Lebensmittel-Rundschau 84, 243-247.

Gómez, M.M.C., Brandt, R., Jakubowski, N. and Anderson, J.T. (2004) Changes of the metal composition in german white wines through the winemaking process. A study of 63 elements by inductively coupled plasma-mass spectrometry. Journal of Agricultural and Food Chemistry 52, 2953-2961.

Gorinstein, S., Goldblum, A., Kitov, S., Deutsch, J., Loinger, C., Cohen, S., Tabakman, H., Stiller, A. and Zykerman, A. (1984) The relationship between metals, polyphenols, nitrogenous substances and treatment of red and white wines. American Journal of Enology and Viticulture 35, 9-15.

Gössinger, M., Schödl, H., Steidl, R. and Meier, W. (1997) Comparison of commercial must and wine bentonites. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria 47 (1-2), 1-7.

Page 132: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

128

Gougeon, R.D., Soulard, M., Miehé-Brendlé, J., Chézeau, J-M., Le Dred, R., Jeandet P. and Marchal, R. (2003) Analysis of two bentonites of enological interest before and after commercial activation by solid Na2CO3. Journal of Agricultural and Food Chemistry 51, 4096-4100.

Greenland, D. J., Mott, C. J. B. (1978) Surfaces of soil particles. In: “The Chemistry of Soil Constituents”. Eds. D. J. Greenland& M. H. B. Hayes (John Wiley & Sons, Chichester) pp. 321-353.

Jakubowski, N., Brandt, R., Stuewer, D., Eschnauer, H.R. and Görtges S. (1999) Analysis of wines by ICP-MS: Is the pattern of the rare earth elements a reliable fingerprint for the provenance?. Fresenius Journal of Analytical Chemistry 364, 424-428.

Lemperle, E. and Kerner, E. (1988) Influence of fermentation and finishing on the ash content of wines. Weinwirtschaft Technik 124 (8), 19-21.

Lipka, Z. (1974) Qualité des bentonites utilisées pour le traitement des moûts et des vins. Revue Suisse de Viticulture, Arboriculture et Horticulture VI (5), 147-155.

Kayabali, K. and Kezer, H. (1998) Testing the ability of bentonite-amended natural zeolite (clinoptinolite) to Remove heavy metals from liquid waste. Environmental Geology 34 (2/3), 95-102.

Machado-Nunes, M., Laureano, O. and Ricardo-da-Silva, J.M. (1998) Influência do tipo de cola (caseína e bentonite) e da metodologia de aplicação nas características físico-químicas e sensoriais do vinho branco. Ciência e Técnica Vitivinícola 13 (1-2), 7-28.

Marchal, R., Barret, J. and Maujean A. (1995) Relations entre les caractéristiques physico-chimiques d’une bentonite et son pouvoir d’adsorption. Journal International des Sciences de la Vigne et du Vin 29 (1), 27-42.

Marchal, R., Weingartner, S., Voisin, C., Jeandet, Ph. and Chatelain, F. (2002) Use of mathematical laws for optimizing the doses of swelled and dry bentonite during the fining of white wines. Part I: Clarification and colloidal stability. Journal International des Sciences de la Vigne et du Vin 36 (3), 169-176.

Maujean, A. (1993) Proprieties physico-chimiques des bentonites: applications œnologiques. Revue Française d’Oenologie 33 (143), 43-53.

Nicolini, G., Larcher, R., Pangrazzi, P. and Bontempo, L. (2004) Changes in the contents of micro- and trace-elements in wine due to winemaking treatments. Vitis 43 (1), 41-45.

OIV (2003) Bentonites. Resolution OENO 11/2003. Office International de la Vigne et du Vin.

Pellerin, P., O’Neill, M.A., Pierre, C., Cabanis, M-T., Darvill, A.G., Albersheim, P. and Moutounet, M. (1997) Complexation du plomb dans les vins par les dimères de rhamnogalacturonane II, un polysaccharide pectique du raisin. Journal International des Sciences de la Vigne et du Vin 31 (1), 33-41.

Peynaud, E. (1996) “Enologia pratica. Conocimiento y elaboracion del vino” (Ediciones Mundi-Prensa: Bilbao).

Poinsaut, P. and Hardy, G. (1995) Bentonites. Analyses et comportements des bentonites. Revue des Œnologues 21 (76), 17-21.

Postel, W., Meier, B. and Markert, R. (1986) Influence on processing aids on the content of mineral compounds and trace elements in wine. I. Bentonite. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria 36 (1), 20-27.

Póvoas, I. and Barral, M.F. (1992) Métodos de análise de solos. Série de Ciências Agrárias, Instituto de Investigação Científica Tropical. Secretaria de Estado da Ciência e Tecnologia, Lisboa.

Ribéreau-Gayon, P., Glories, Y., Maujean, A. and Dubourdieu, D. (2000). “Handbook of Enology” Vol. 2: The Chemistry of Wine Stabilization and Treatments (John Wiley & Sons Lta: Paris).

Sparks, D.L. (1995) “Environmental Soil Chemistry” (Academic Press: London).

Sudraud, P., Sudraud, F. and Gaye, J. (1985) Observations sur la composition et le contrôle d’efficacité des bentonites. Revue Française d’ Oenologie 25 (97), 21-24.

Vernhet, A. and Bellon-Fontaine, M.N. (1995) Role of bentonites in the prevention of Saccharomyces cerevisiae adhesion to solid surfaces. Colloids and Surfaces B: Biointerfaces 3, 255-262.

Wurzinger, A., Netzer, M., Heili, K. and Bandion, F. (1994) Migration of components of bentonites during the fermentation of must. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria 44, 218-221

Page 133: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

129

4.2. RELEASE OF CONTAMINANT ELEMENTS FROM BENTONITES TO WINE:

A CONTRIBUTION TO ACHIEVE A TEST SOLUTION

S. Catarino1,3, M. Madeira2, F. Monteiro2, A.S. Curvelo-Garcia1, R. Bruno de Sousa3

1INIAP. Estação Vitivinícola Nacional. 2565-191 Dois Portos. Portugal. 2Instituto Superior de Agronomia. Departamento de Ciências do Ambiente. Tapada da Ajuda, 1349-017 Lisboa. Portugal. 3Instituto Superior de Agronomia. Departamento de Química Agrícola e Ambiental. Tapada da Ajuda, 1349-017 Lisboa. Portugal.

Abstract

The release of mineral elements from six bentonites to wine and to the test solution established by the OIV was evaluated in order to compare the extraction performance of both solutions. Significant differences between wine and tartaric solution results for thirty-four mineral elements analysed by AAS, ETAAS and ICP-MS were observed, suggesting that the extraction solution proposed by the OIV is not suitable for bentonite quality control purposes. Taking into account bentonite maximum extractable amounts defined by the OIV, some samples showed higher concentrations for Na (B3 and B8), Al (B8), Fe (B5) and As (B9). An additional extraction essay involving two bentonites, wine and two complex test solutions containing the major mineral elements of wine (K, Ca, Na and Mg) and protein (wine protein and BSA) was carried out. Significant differences were observed for all elements with exception of Sc, Mn, Co, W and Bi. For several elements, such as Na, Mg (test solution with BSA) V, Ni, Ga, Zr, Cd, In, Sb, Tl and U, the similarity of wine and test solutions content variations was satisfactory. The protein nature of test solutions seemed to have a decisive role in some element changes, probably related to wine protein and BSA distinct volume. In general, the test solution containing wine protein was a slightly better simulator of wine performance. However, the above mentioned test solutions did not simulated wine performance in a total satisfactory way, and further work should be performed on extraction solution composition. Keywords: contaminant elements, bentonite, wine, test solution, ICP-MS

Ciência e Técnica Vitivinícola (2006) 21 (1), 17-31

Page 134: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

130

4.2.1. INTRODUCTION

Bentonite is a technological additive with a high potential influence on mineral

composition of the wine and one of its main sources of contamination. Over the last twenty years,

several authors have reported important concentration increases of metals in wines due to bentonite

addition (Postel et al., 1986; Enkelmann, 1988; Wurzinger et al., 1994; Gössinger et al., 1997;

Machado-Nunes et al., 1998; Catarino et al., 2004; Gómez et al., 2004; Nicolini et al., 2004). Wine

enrichment in some mineral elements may affect its sensory quality, namely appearance, its legal

quality as well as an impact on consumer health (Curvelo-Garcia and Catarino, 1998).

Usually, a general physical and chemical characterization as well application instructions

are provided with the commercial bentonite, without information about its potential mineral release

to wine. Considering that bentonite origin and type on the market is considerably diverse, each

product may cause a particular effect on mineral composition of wine. This statement is supported

by the results of a previous study, in which we have studied the release of contaminant elements

from bentonites to wine, in relation with their physical and chemical characteristics (submitted for

publication).

Assuming that quality control of bentonites is essential to assure its safe use as a

technological additive, the International Organisation of Vine and Wine (OIV, 2003) defines limit

values for a set of elements to be determined in a bentonite extraction solution. According to the

OIV, the extraction trial should be carried out with a tartaric acid solution. Nevertheless,

considering the wine natural complexity, it should be admitted that the results obtained using this

test solution could be very different from wine results.

Due to the high cationic exchange capacity of bentonites, cations are released when

positively charged proteins and other soluble cationic constituents in wine are adsorbed. Blade and

Boulton (1988) have stated that adsorption of proteins suffers competition from other cations in the

solution matrix and the solvent properties. These authors suggested that competition would be from

K+, Ca2+, Mg2+, Na+ and H+, most amino acids, some peptides and other cationic fractions.

Concerning protein adsorption by bentonites, few studies have been carried out using wine model

solutions (Blade and Boulton, 1988; Achaerandio et al., 2001). However, as far as we know there

are no published studies developed in view to attain a representative test solution of wine, in what

concerns to its extraction effect on bentonites. Furthermore the natural diversity of wine

composition represents an additional obstacle to a test solution establishment.

The main purpose of our work was to evaluate the release of mineral elements from

bentonites employing: (1) the test solution indicated by the OIV, and (2) wine, in order to compare

their performance. In addition, complex solutions were also tested in order to improve a test

solution fairly representative of wine.

Page 135: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

131

4.2.2. MATERIAL AND METHODS

Six different bentonites follow designated by B1, B3, B4, B5, B8 and B9, obtained from

four companies in Portugal, were used for extraction essays. Tartaric acid was added (3.20 g/L) to a

white wine from 2001 vintage in order to obtain a pH 3.0. Wine pH value was controlled through

the time until stability was reached then extraction trials with bentonites were carried out.

4.2.2.1. Test solutions

Test solutions were prepared using deionized water (conductivity < 0.1 µS/cm) obtained

through a Seralpur Pro 90 CN (Seral, Ransbach-Baumbach, Germany).

Test solution A (OIV) - a L(+) tartaric acid solution of 5 g/L (Merck) and pH level of 3

(adjusted with sodium hidroxide 10 N).

Test solution B and test solution C - 5 L of a solution with the follow composition was

prepared: tartaric acid from Merck (5 g/L), potassium chloride from Merck (1.90 g/L); calcium

carbonate from Riedel (250 mg/L) and hexahydrated magnesium chloride from Fluka (400 mg/L),

in order to simulate K, Ca and Mg wine contents. The pH value was adjusted to 3 with sodium

hydroxide 10 N. Afterwards, this solution was separated in two flasks and 120 mg/L of wine

protein (test solution B) or standard protein BSA (bovinum serum albumin from Sigma) (test

solution C) was added. Wine protein was obtained as described by Mesquita et al. (2001). Chosen

concentration is considered typical in Portuguese white wines (Mesquita et al., 2001).

4.2.2.2. Extraction essays

An amount of 5 g anhydrous bentonite was shaken together with 200 mL of wine (or test

solution) in an Erlenmeyer vessel for 15 min. After 24 hours of decantation, the liquid obtained was

centrifuged at 12000 rpm during 30 min and the supernatant was used for measurements. Samples

were prepared in parallel, in triplicate, with control samples (without bentonite addition), similarly

stirred and centrifuged. Only polypropylene material was used. All the material was soaked in 20%

HNO3 (v/v) for at least 24 hours and rinsed several times with deionized water before use.

Trial 1 – Extraction essay using wine and test solution A (OIV)

This study was based in a factorial experiment, with two factors: bentonite (B1, B3, B4,

B5, B8 and B9) and extraction solution (wine, test solution A), with three replications (Figure 1).

Trial 2 – Extraction essay using wine, test solution B and C

This study was also based in a factorial experiment, with two factors: bentonite (B1 and

B9) and extraction solution (wine, test solution B, test solution C), with three replications (Figure

2).

Page 136: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

132

Figure 1 - Trial 1 scheme

Figure 2 – Trial 2 scheme

4.2.2.3. Mineral composition analysis

Na, Mg, K, Ca and Fe were analysed by flame atomic absorption spectrometry (FAAS)

(Catarino et al., 2003), Al by electrothermal atomic absorption spectrometry (ETAAS) (Catarino et

al., 2002), while Li, Be, B, Sc, V, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Rb, Sr, Y, Zr, Nb, Mo, Cd,

In, Sn, Sb, Ba, W, Tl, Pb, Bi and U were analysed by inductively coupled plasma mass

spectrometry (ICP-MS) semi-quantitative approach (Catarino et al., in press), using a Perkin Elmer

SCIEX Elan 9000 apparatus, equipped with a crossflow nebulizer, a Ryton Scott-type spray

chamber, nickel cones and a peristaltic sample delivery pump with four channels (Gilson)

(application software Elan – 6100 Windows NT - version 2.4). An autosampler Perkin-Elmer AS-

93 Plus was protected by a laminar-flow-chamber clean room class 100 (Reinraumtechnik Max

Petek). Operating conditions used were as follows: RF power of 1200W; sample uptake rate of

Bentonite

(B1, B3, B4, B5, B8, B9)

Bentonite

(B1, B3, B4, B5, B8, B9)

Control; W-B1; W-B3;

W-B4; W-B5; W-B8; W-B9

Wine Test solution A

Control; A-B1; A-B3;

A-B4; A-B5; A-B8; A-B9

Bentonite

(B1, B3, B4, B5, B8, B9)

Bentonite

(B1, B3, B4, B5, B8, B9)

Control; W-B1; W-B3;

W-B4; W-B5; W-B8; W-B9

Wine Test solution A

Control; A-B1; A-B3;

A-B4; A-B5; A-B8; A-B9

Bentonite(B1, B9)

Bentonite(B1, B9)

Control; W-B1; W-B9

Wine Test solution B

Control; B-B1; B-B9

Test solution C

Bentonite(B1, B9)

Control; C-B1; C-B9

Bentonite(B1, B9)

Bentonite(B1, B9)

Control; W-B1; W-B9

Wine Test solution B

Control; B-B1; B-B9

Test solution C

Bentonite(B1, B9)

Control; C-B1; C-B9

Page 137: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

133

0.85 mL/min; nebulizer argon flow between 0.85 and 0.95 L/min. A full mass spectrum (m/z = 6-

240, omitting the mass ranges 16-18, 40, 41, 211-229) was obtained by full mass range scanning. A

reference response table (Perkin-Elmer TotalQuant III) was updated with a multielement standard

solution (Li, Be, Na, Mg, Al, K, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Ag, Cd, In,

Cs, Ba, Hg, Tl, Pb, Bi, U) at 10 µg/L, from Perkin-Elmer. Rh and Re were used as internal

standards for elements in the mass range (m/z) 7-138 and 205-238, respectively. The software of

the instrument performed automatic corrections of isobaric interferences. In order to get signal

stabilization, a sample read delay of 75 s was chosen. Between samples or standards, the sampling

system was rinsed with a 2% HNO3 (v/v), 1% C2H5OH (v/v), and 200 µg/L of Au solution for 75 s.

The experimental conditions used for the measurements were: dwell time = 50 ms; sweeps/reading

= 6; reading/replicates = 1; replicates = 1; time per run = 67 s. Only high purity reagents and

deionized water (conductivity < 0.1 µS/cm) were used. All the material (polypropylene and Teflon

PFA) was soaked in 20% HNO3 (v/v) for at least 24 h and rinsed several times with deionized

water, before use.

4.2.2.4. Statistical analysis The variance analysis was performed using Statistica version 98 edition (Statsoft Inc.,

E.U.A.). As control samples of wine and test solutions showed different mineral contents,

concentration variations between control and bentonite treated samples were considered for

statistical analysis.

4.2.3. RESULTS AND DISCUSSION

Table 1 shows the average variations in the concentration of several elements, both in wine

and test solution A, caused by bentonite addition. For each element, the variation of concentration

between each modality (extraction solution and bentonite association) and the respective control

was calculated. Afterwards, for each extraction solution (wine and test solution A) the mean of the

corresponding six modalities, considering the three independent replicates, was calculated.

The interpretation of the results should take into account that those were obtained under

experimental conditions, using a bentonite dose higher than usually applied in winemaking. Thus,

concerning wine enrichment or loss in different elements, the results could not be directly

extrapolated to production scope. On the other hand, for some elements, it should be noted that

significant differences between extraction solutions were not relevant, bearing in mind the low

amounts involved.

Considering the whole mineral elements in study, with exception of Pb, the average

concentration variation of wine and test solution A were significantly different (99% confidence

level). At once, the results concerning to B, K, Cu and Zn should be pointed out. In consequence of

Page 138: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

134

bentonite addition, their contents in wine suffered important decreases while in test solution A

increases occurred. The different trends between the two extraction solutions were certainly related

to their distinct composition, namely mineral and protein contents. The highest weight cation of

wine, K, was present in test solution A as a trace element. Probably this element replaced bentonite

Na and Ca positions being consequently partially removed from wine. In fact, proteins and K

competition to bentonite adsorption was previously reported (Blade and Boulton, 1998). Similar

explanation is credible for Rb decrease, since it is an element of low ionic potential, likewise K.

Table 1 - Average changes in the concentrations of several elements in wine and test solution A caused by bentonites addition (Trial 1). The results are expressed in µg/L, except for Na, Mg, Al,

K, Ca and Fe (mg/L)

Element Extraction solution effect

Wine

mean

(n=18)

Test sol. A

mean

(n=18)

Element Extraction solution effect

Wine

mean

(n=18)

Test sol. A

mean

(n=18)

Li ** 9.80a 20.31b As ** 24.34b 31.27a

Be ** 13.03b 4.03a Rb ** 409.48b 28.89a

B ** (-)331.31b 13.75a Sr ** 665.95b 127.01a

Na ** 358.28b 247.67a Y ** 4.597a 5.338b

Mg ** 50.00b 28.89a Zr ** 269.30a 412.68b

Al ** 15.416a 25.804b Nb ** 11.50b 8.68a

K ** (-)193.6b 5.4a Mo ** 7.55a (-)46.55b

Ca ** 77.22b 2.27a Cd ** 0.55b 0.20a

Sc ** 0.76a 2.11b In ** 0.06b (-)0.04a

V ** 27.10a 44.88b Sn ** 2.60a 7.30b

Mn ** 1406.54b 964a Sb ** 4.37b 2.88a

Fe ** 3.69a 8.02b Ba ** 83.03a 218.46b

Co ** 19.77b 18.29a W ** 0.736a 0.998b

Ni ** 45.85b 23.53a Tl ** 2.41b 0.91a

Cu ** (-)44.44b 16.10a Pb n.s. 13.13 13.33

Zn ** (-)109.85a 125.83b Bi ** 0.847a 66.749b

Ga ** 1.44a 5.88b U ** 43.80a 64.34b

Ge ** 0.046a 0.250b

Averages followed by the same letter are not significantly different at α = 0.01**; ns = without significant difference.

The results concerning Mo element are probably explained by its distinct concentration in

control samples. The control sample of test solution A presented a higher concentration of Mo,

probably introduced as an impurity by tartaric acid. Likewise K and Rb, Mo probably replaced

bentonite Na and Ca positions. The explanation for B, Cu and Zn depletions may be associated to

Page 139: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

135

their complexation with condensed tannins, which are in part bind to proteins. Moreover, the

extraction solution A has no protein or tannins in its composition.

Regarding to bentonite major exchangeable cations, Na, Mg and Ca, important enrichments

were verified in both extraction solutions, specially in wine. This observation suggests that cationic

exchange phenomena were more intense in wine somehow as expected, since the test solution A

contained no protein.

For the following elements, Al, Sc, V, Fe, Ga, Ge, Y, Zr, Sn, Ba, Bi and U, and in less

extent Y, higher increases were observed in test solution A. Probably, the partial removal of these

elements from wine by precipitation with other chemical species such as organic acids, proteins,

tannins or even polysaccharides occurred. On the other hand, the release of the above mentioned

elements from smectite structure or from other minerals and impurities could have been enhanced

by the tartaric acid. Maujean (1993) stated that clays could be partial or totally destroyed by tartaric

acid.

It should be remembered that the previous statements are based on average values. In fact,

interaction between bentonite and extraction solution was observed for all the elements.

Concerning Pb results, the similar average value observed in both extraction solutions were

obtained by chance as for none of the individual studied bentonites the results were similar.

Additionally, for each element a linear least-squares adjustment of the results obtained

using wine and test solution A was performed in order to evaluate their correlation (r). The most

satisfactory results obtained, corresponding to Mg, Co, Fe, As, Zr, Sn, Sb, W, Tl and U elements,

are shown in Figure 3 (the 95% confidence limits of the linear regression parameters are given). A

deviation of the slope from unity indicates proportional discrepancies between wine and test

solution A results. A non-zero intercept is diagnosed as a constant discrepancy.

The previous results confirmed that tartaric acid solution was not able to simulate wine

extraction performance. Thus, concerning bentonite potential mineral contamination effect on

wine, it could be stated that test solution A is not a suitable extraction solution for bentonite quality

control purposes.

Taking into account bentonite maximum extractable contents defined by the OIV (OIV,

2003) some observations should be emphasized: the extractable Al content of B8 in tartaric acid

solution was of 2.6 mg/L, a slightly higher value than the established 2.5 mg/kg; the B9 soluble As

content, of 2.5 mg/kg, exceed the defined 2 mg/kg; B5 soluble Fe content was of 865 mg/kg (the

established content is of 600 mg/kg). Finally, B3 and B8 samples both showing 13 g/kg exceeded

the soluble Na content (less than 10 g/kg) defined for natural bentonites. These results are of major

importance as to our knowledge the quality control of bentonites is not a regular practise.

Page 140: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

136

Figure 3 - Relationship between the changes in the concentrations of Mg, Co, Fe, As, Zr, Sn, Sb, W, Tl and U in wine and test solution A, caused by bentonite addition. The linear least squares adjustment and correlation coeficient are shown (the 95% confidence limits of the linear regression parameters are given).

Sb

0

5

10

15

20

0 5 10 15

Wine

Tes

t sol

utio

n A

Mg

-10

0

10

20

30

40

50

60

-20 0 20 40 60 80 100

Wine

Tes

t sol

utio

n A

Co

0

10

20

30

40

50

0 10 20 30 40

Wine

Tes

t sol

utio

n A

Fe

0

5

10

15

20

25

0 2 4 6 8

Wine

Tes

t sol

utio

n A

As

-200

20406080

100120

0 20 40 60 80

Wine

Tes

t sol

utio

n A

Zr

0

200

400

600

800

1000

0 200 400 600 800

Wine

Tes

t sol

utio

n A

Sn

-1

0

1

2

3

4

-0,8 -0,6 -0,4 -0,2 0 0,2 0,4

Wine

Test

sol

utio

n A

W

0

0,5

1

1,5

2

2,5

0 0,5 1 1,5 2

Wine

Tes

t sol

utio

n A

Tl

0

0,5

1

1,5

2

0 1 2 3 4 5 6

Wine

Tes

t sol

utio

n A

U

0

50

100

150

200

-20 0 20 40 60 80 100 120

Wine

Tes

t sol

utio

n A

r = 0.9464 (n = 18) y = (0.6 ± 0.1) x + (1 ± 6)

r = 0.9163 (n = 18) y = (1.2 ± 0.3) x + (-5 ± 6)

r = 0.7830 (n = 18) y = (2 ± 1) x + (-1 ± 4)

r = 0.9686 (n = 17) y = (1.5 ± 0.2) x + (-5 ± 6)

r = 0.9732 (n = 18) y = (1.3 ± 0.2) x + (68 ± 50)

r = 0.8416 (n = 15) y = (4 ± 1) x + (1.7 ± 0.4)

r = 0.9921 (n = 18) y = (1.39 ± 0.09) x + (0.4 ± 0.5)

r = 0.9705 (n = 18) y = (1.5 ± 0.9) x + (91 ± 88)

r = 0.9461 (n = 18) y = (0.2 ± 0.2) x + (0.1 ± 0.2)

r = 0.8601 (n = 17) y = (1.1 ± 0.3) x + (19 ± 20)

Page 141: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

137

The next step in our work involved essays carried out with two extraction solutions

containing the major mineral elements of wine and also protein: (1) test solution B with wine

protein and (2) test solution C with standard protein (BSA), both solutions tested in parallel with

wine (Trial 2). Table 2 shows the variations in the contents of several elements in wine, test solution B and

test solution C, caused by bentonite addition. The multielemental composition of control samples is

shown at Table 3. For each modality (extraction solution and bentonite association) and element,

the difference between its content and the respective control was calculated. Afterwards, for each

type of extraction solution the mean of the corresponding two modalities, considering the three

independent replicates, was calculated.

As previously stated cations release from bentonites depends on protein adsorption extent.

Adsorption of proteins with distinct molecular weight by bentonite in a white wine model solution

was studied by Achaerandio et al. (2001). These authors proposed that protein volume is the factor

that most affects the adsorption process. The molecular weight of the large majority of wine

proteins ranges from 20 to 30 kDa, while Bovine Serum Albumin (BSA), the standard protein also

used in our study, has a considerable higher molecular weight, of approximately 66 kDa.

Concerning to isoelectric points (pI), wine proteins and BSA show similar values, from 4.1 to 5.8

and of 4.7, respectively (Hsu and Heatherbell, 1987; Achaerandio et al., 2001; Monteiro, 2001).

The protein nature of test solutions seemed to have a decisive role in some element

changes, such as Mg, K, Zn and Pb. Mg and Zn and probably Pb are exchangeable cations of

bentonite, suggesting that their release depend on the extent of cation adsorption. On the other

hand, as shown on table 3, the control solution of test solution B contained a high concentration of

Zn, probably due to wine protein. K is a protein competitor, so its final content was strongly related

to protein adsorption. In fact, higher K depletions were observed in test solution C modalities,

probably because protein adsorption occurred at less extent due to BSA larger volume. Significant

different variations (95% confidence level) were observed between extraction solutions for all

elements with exception of Sc, Mn, Co, W and Bi. For several elements, such as Na, Mg (test

solution C) V, Ni, Ga (B9), Zr (in special B9), Cd, In, Sb, Tl and U the similarity between wine and

test solutions variations was satisfactory, in spite of those variations being statistically different.

In regard to Fe, Y and Sn contents, both test solutions showed similar variations with

bentonite addition, higher than wine variations but nevertheless indicative. It should be pointed out

that both extraction solutions showed similar variations in the contents of Fe, probably limited by

its solubility.

Observing the elements which concentration decreased in wine with bentonite addition, B,

K, Cu, Zn and Rb, the distinct concentrations of control samples could explain, in part, the results.

Page 142: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

138

Table 2 - Changes in the concentrations of several mineral elements in wine, test solution B and test solution C, caused by bentonite addition (Trial 2). The results are expressed in µg/L, except for Na, Mg, K, Ca and Fe (mg/L)

Extraction

solution effect

Wine

mean

(n=6)

Test sol. B

mean

(n=6)

Test sol. C

mean

(n=6)

W-B1

(n=3)

B-B1

(n=3)

C-B1

(n=3)

W-B9

(n=3)

B-B9

(n=3)

C-B9

(n=3)

Li ** 76.17c 50.25b 47.79a 19.57b 20.68b 11.69a 132.77e 79.83c 83.88d

Be ** 33.04b 7.97a 8.69a 35.71e 5.48a 4.72a 30.38d 10.45b 12.67c

B ** (-)193.06c 1.90a 12.59b (-)388.93b 4.93a 8.26a 2.82a (-)1.14a 16.92a

Na ** 276b 215a 313c 283d 220a 234b 269c 210a 391e

Mg ** 13.95b 7.38a 13.60b 23.77c 14.3b 23.87c 4.13a 0.48a 3.33a

Al * 5169a 6809b 5283a 6060b 10465c 7709b 4280a 3152a 2856a

K ** (-)75b (-)4a (-)258c (-)121c 16a (-)247d (-)28b (-)24b (-)269d

Ca ** 98c 18a 28b 115e (-)6a 12b 80d 42c 44c

Sc n.s. 1.44 1.57 1.73 1.24a 1.92b 1.66a,b 1.63a,b 1.21a 1.81b

V * 21.79b 19.69a,b 15.70a 39.71c 37.50c 28.34b 3.86a 1.89a 3.07a

Mn n.s. 851.58 960.30 926.85 1424.60c 1521.24c 1371.54c 278.57a 399.26a,b 482.17b

Fe ** 3.94a 6.08b 6.20b 0.39a 1.47b 1.61b 7.49c 10.69d 10.79d

Co n.s. 7.35 8.28 8.30 6.14a 7.01a 6.11a 8.57b 9.55b,c 10.48c

Ni ** 15.33a 17.01b 18.77c (-)1.49b 4.59c 0.97a 32.15d 32.94d 33.06d

Cu ** (-)96.38b -52.72a -54.13a (-)85.02d (-)68.50b (-)72.68c (-)107.74e (-)36.94a (-)35.58a

Zn ** (-)176.75b (-)8.98a 243.75c (-)592.82e (-)138.51b 17.99a 239.31c 120.56a 469.51d

Ga

** 1.97b 1.31a 1.34a 2.22e 1.22a,b 1.12a 1.74d 1.40b,c 1.57c,d

Page 143: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

139

As ** 160.21b 51.38a 49.27a 10.02a 2.07a 1.94a 310.40c 100.69b 96.59b

Rb ** (-)673.53b 18.04a (-)16.37a (-)784.62d 17.08a,b 11.79a (-)562.44c 19.00b 20.94b

Sr n.s. 1647.00 1633.73 1705.88 819.97a 778.03a 700.78a 2474.03b 2489.42b 2710.97c

Y ** 18.74a 33.77b 38.29c 4.41a 5.36a 6.90a 33.07b 62.18c 69.68d

Zr ** 660.89b 540.60a 523.60a 673.43d 495.69b 418.19a 648.34d 585.50c 629.01c,d

Nb ** 49.70c 19.04a 20.77b 28.89b 5.54a 4.82a 70.52e 32.55c 36.71d

Mo ** 27.45a 33.24b 36.12c 0.70a 1.32a 1.04a 54.20b 65.16c 71.20d

Cd * 0.89a 1.03a,b 1.12b 0.78a 0.83a 0.84a 1.00a,b 1.22b,c 1.40c

In ** 0.330b 0.215a 0.22a 0.042b 0.001a (-)0.002a 0.618d 0.429c 0.444c

Sn ** 10.80a 15.80b 17.12c (-)0.18a 0.89b 0.75a,b 21.78c 30.70d 33.49e

Sb * 10.20b 8.99a 9.90b 1.6a 1.55a 1.49a 18.78c 16.43b 18.32c

Ba ** 30.06a 599.90b 654.18b 41.52a 1062.50c 1136.90c 18.60a 137.31b 171.47b

W n.s. 0.628 0.639 0.650 1.157b 1.185b 1.164b 0.099a 0.092a 0.136a

Tl ** 0.81b 0.50a 0.50a 0.44b (-)0.05a (-)0.05a 1.19c 1.05c 1.03c

Pb ** 29.18a 91.43b 170.50c 5.52a 10.97a 22.56b 52.85c 171.89d 286.75e

Bi n.s. 4.22 4.02 4.05 (-)0.93b (-)0.16a (-)0.06a 9.37d 8.21c 8.16c

U ** 57.18b 54.28b 47.41a (-)0.47a (-)0.35a 2.30a 114.84d 108.91c 92.52b

Averages followed by the same letter are not significantly different at α = 0.01** or α = 0.05*; ns = without significant difference.

Page 144: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

140

Table 3 - Mineral compositiona of wine and test solutions B and C control samples. The results are expressed in µg/L, except for Na, Mg, K, Ca and Fe (mg/L)

Wine

control

(n = 3)

Test sol. B control

(n = 3)

Test sol. C control

(n = 3)

Wine

control

(n = 3)

Test sol. B control

(n = 3)

Test sol. C control

(n = 3)

Li 31.1 (0.1) NDb NDb As 5.4 (0.3) 5.0 (0.3) 4.2 (0.9)

Be 0.21 (0.07) NDb NDb Rb 1102 (21) 19 (1) 18.5 (0.6)

B 2921 (64) 37 (6) 9 (1) Sr 207 (6) 114.1 (0.5) 112 (2)

Na 103.8 (0.7) 310 (3) 320 (3) Y 0.287 0.41 (0.03) 0.45 (0.02)

Mg 78.6 (0.2) 64 (1) 62.6 (0.6) Zr 3.8 (0.3) 4.1 (0.5) 5.2 (0.9)

Al 459 (30) 211 (35) 144 (8) Nb 0.04 (0.01) 0.09 (0.02) 0.03 (0.01)

K 539 (15) 671 (28) 881 (3) Mo 2.31 (0.09) 0.5 (0.2) 0.43 (0.08)

Ca 112 (2) 115.2 (0.4) 115.6 (0.7) Cd 0.44 (0.03) 0.32 (0.09) 0.15 (0.04)

Sc 2.4 (0.2) 0.44 (0.04) 0.19 (0.04) In 0.02 (0.01) NDb NDb

V 9.1 (0.3) 8.1 (0.4) 7.8 (0.4) Sn 2.2 (0.2) 0.66 (0.01) 0.66 (0.02)

Mn 769 (16) 17.3 (0.6) 16.0 (0.8) Sb 2.7 (0.1) 0.5 (0.3) NDb

Fe 1.93 (0.02) 0.05 (0.01) 0.04 (0.01) Ba 35 (1) 16.1 (0.9) 12.9 (0.1)

Co 3.4 (0.2) 0.26 (0.02) 0.24 (0.03) W 0.26 (0.01) 0.06 (0.01) NDb

Ni 40 (1) 5.2 (0.6) 4.2 (0.2) Tl 0.13 (0.01) 1.5 (0.1) 1.56 (0.03)

Cu 121 (5) 79 (2) 78 (1) Pb 23 (1) 0.9 (0.2) 0.67 (0.06)

Zn 1552 (46) 468 (20) 34 (1) Bi 1.2 (0.9) NDb NDb

Ga 1.7 (0.2) 0.05 (0.04) 0.035 U 0.47 (0.05) 0.4 (0.2) 0.29 (0.03)

aConcentration values correspond to the mean of three independent replicates and corresponding standard deviation; bND – not detected

As a general trend, the solution presenting the initial higher concentration of the element

showed the highest depletion with bentonite addition. An exception of this was observed for B in

samples treated with B9. Wine control, with the highest concentration of Cu, Zn and Rb, suffered

the major depletions. It could be concluded that in future studies the test solution should be

prepared in order to present similar amounts of these elements to wine, in particular K amounts,

since this element plays an important role in cationic exchange phenomena competing with

proteins to bentonites complex exchange. Though the mineral profile of wines is widely variable it

is difficult to establish a standard composition.

For Ca element the results were very unsatisfactory, with higher enrichment in wine than in

test solutions. A probable explanation could be its reaction with tartaric acid originating calcium

tartrate that could precipitate. This explanation is also valid for K element. Concerning Li element

and bentonite B1, its content variation in test solution B was similar to wine enrichment. Both test

Page 145: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

141

solutions, either with B1 or B9, verified low Be enrichments by comparison to wine. The results

concerning As and Nb elements were unsatisfactory, with wine presenting the highest enrichments,

in spite of the similar values shown by both test solutions. On the contrary, Ba enrichments were

much higher in test solutions than in wine.

Additionally, linear least-squares adjustment of the results (considering only the elements

that have increased with bentonite treatment) obtained using wine and test solutions, were

performed in order to evaluate their correlation (95% confidence level). For B1, the correlation

value between wine and test solution B results was higher (r = 0.840) than wine and test solution C

correlation (r = 0.791). Similar results were obtained for B9, with correlation values of r = 0.903

and r = 0.860 between wine and test solutions B and C, respectively. Thus, in general it seems that

test solution B, containing wine protein, was a better simulator of wine performance.

Finally, the previous results suggested that neither test solution B nor test solution C

simulated wine performance in a total satisfactory way, and additional studies should be done in

view to improve solution performance.

4.2.4. CONCLUSIONS

When using extraction solutions containing wine protein or BSA and some major mineral

elements, similar average variations in the contents of Sc, Mn, Co, W and Bi in wine and test

solutions were observed. In spite of significantly different, for Na, Mg, V, Ni, Ga, Zr, Cd, In, Sb, Tl

and U similar results were observed. The protein nature of test solutions seemed to have an

important role in some element variations, perhaps due to wine protein and BSA different volumes

that could affect the adsorption process. In fact, it was possible to observe that, globally, the test

solution containing wine protein was a better simulator of wine performance. In spite of that,

neither test solution B nor test solution C simulated wine performance in a total satisfactory way.

Thus, further work should be carried out on test solution composition in view to improve its

performance. Taking into account the active role played by the major mineral elements of wine in

cationic exchange phenomena, it seems important that the test solution nearly reflects the wine

composition. As the mineral profile of wines is widely variable it is difficult to settle a standard

composition. Bearing in mind the different strength of acids could be advantageous to simulate the

wine acidic composition.

It should be emphasized that the tartaric acid solution indicated by OIV was not able to

simulate wine performance on bentonite extraction essays, leading to significant different average

variations in the contents of thirty-four elements. Thus, as a principal conclusion, this work

evidences the unsuitability of the test solution recommended by the OIV for bentonite quality

control purposes, concerning its potential effect on wine mineral contamination.

Page 146: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

142

At last, it was possible to verify that some bentonites were in disagreement with OIV

resolutions, as maximum extractable contents of Na, Al, Fe and As were exceeded. These

observations are of major importance as to our knowledge the quality control of bentonites by

control laboratories or by winemakers is not a regular practise in Portugal.

Aknowledgements The authors are grateful to Doctor Sara Monteiro from Instituto Superior de Agronomia for providing the

wine protein and Bovine Serum Albumin; to AEB Bioquímica, Ecofiltra, Meller & Essink, Proenol and

Soeno, for providing the bentonite samples; to “Programa de Apoio à Reforma dos Laboratórios de Estado

(PARLE)” for financial support and to “Fundação para a Ciência e a Tecnologia” for a PhD scholarship

(POCI 2010, medida IV.3, SFRH/BD/17237/2004).

4.2.5. REFERENCES Achaerandio I., Pachova V., Güell C., López F., 2001. Protein adsorption by bentonite in a white wine model solution: effect of protein molecular weight and ethanol concentration. American Journal of Enology and Viticulture, 52 (2), 122–126.

Blade W.R., Boulton R., 1988. Adsorption of protein by bentonite in a model wine solution. American Journal of Enology and Viticulture, 39 (3), 193-199.

Catarino S., Curvelo-Garcia A.S., Bruno de Sousa R., 2002. Determination of aluminum in wine by graphite furnace AAS: Validation of Analytical Method. Atomic Spectroscopy, 23 (6), 196-200.

Catarino S., Pinto D., Curvelo-Garcia A.S., 2003. Validação e comparação de métodos de análise em espectrofotometria de absorção atómica com chama para doseamento de cobre e ferro em vinhos e aguardentes. Ciência e Técnica Vitívinícola, 18 (2), 65-76.

Catarino S., Soares J., Curvelo-Garcia A.S., Bruno de Sousa R., 2004. Implicações da utilização de bentonites sobre a fracção mineral de vinhos: potássio, sódio, cálcio, alumínio e chumbo. Efeito do pH. Ciência e Técnica Vitívinícola, 19 (1), 29-45.

Catarino S., A.S. Curvelo-Garcia, R. Bruno de Sousa. Measurements of contaminant elements of wines by inductively coupled plasma-mass spectrometry: a comparison of two calibration approaches. Talanta (in press).

Curvelo-Garcia A.S., Catarino S., 1998. Os metais contaminantes dos vinhos. Origens da sua presença, teores, influência dos factores tecnológicos e definição de limites. Ciência e Técnica Vitivinícola, 13 (1-2), 49-70.

Enkelmann R., 1988. Migration of heavy metals from bentonites to wine.1. Note: Bentonites. Deutsche Lebensmittel-Rundschau, 84, 243-247.

Gómez M.M.C., Brandt R., Jakubowski N., Anderson J.T., 2004. Changes of the metal composition in german white wines through the winemaking process. A study of 63 elements by inductively coupled plasma-mass spectrometry. Journal of Food Agriculture and Food Chemistry, 52, 2953-2961.

Gössinger M., Schödl H., Steidl R., Meier W., 1997. Comparison of commercial must and wine bentonites. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria, 47 (1-2), 1-7.

Hsu J.C., Heatherbell D.A., 1987. Isolation and characterization of soluble proteins in grapes, grape juice and wine. American Journal of Enology and Viticulture, 38, 6-10.

Machado-nunes M., Laureano O., Ricardo-da-Silva J.M., 1998. Influência do tipo de cola (caseína e bentonite) e da metodologia de aplicação nas características físico-químicas e sensoriais do vinho branco. Ciência e Técnica Vitivinícola, 13 (1-2), 7-28.

Page 147: Tese doutoramento Sofia Catarino.pdf

Cedência ao vinho de metais contaminantes de bentonites

143

Maujean A., 1993. Propriétés physico-chimiques des bentonites: applications oenologiques. Revue Française d’Oenologie, 143, 43-53.

Mesquita P., Piçarra-Pereira M.A., Monteiro S., Loureiro V., Teixeira A., Ferreira R.B., 2001. Effect of wine composition on protein stability. American Journal of Enology and Viticulture, 52 (4), 324–330.

Monteiro S.A.V.S., 2001. Análise molecular das proteínas responsáveis pela turvação dos vinhos. PhD Thesis. Instituto Superior de Agronomia, Universidade Técnica de Lisboa, Lisboa.

Nicolini G., Larcher R., Pangrazzi P., Bontempo L., 2004. Changes in the contents of micro- and trace-elements in wine due to winemaking treatments. Vitis, 43 (1), 41-45.

OIV, 2003. Bentonites. Resolution OENO 11/2003. Office International de la Vigne et du Vin.

Postel W., Meier B., Markert R., 1986. Influence on processing aids on the content of mineral compounds and trace elements in wine. I. Bentonite. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria, 36 (1), 20-27.

Wurzinger A., Netzer M., Heili K., Bandion F., 1994. Migration of components of bentonites during the fermentation of must. Mitteilungen Klosterneuburg, Rebe und Wein, Obstbau und Früchteverwertung Austria, 44, 218-221.

Page 148: Tese doutoramento Sofia Catarino.pdf

Capítulo 5

CONSIDERAÇÕES FINAIS

Page 149: Tese doutoramento Sofia Catarino.pdf

Considerações finais

147

5.1. CONSIDERAÇÕES FINAIS A investigação sobre a presença de metais contaminantes no vinho, no que respeita aos

níveis de ocorrência, evolução ao longo do processo tecnológico, fontes de contaminação e ao

desenvolvimento de métodos de análise adequados ao seu doseamento, é de enorme relevância e

interesse, particularmente no âmbito tecnológico e da segurança alimentar.

Numa primeira fase do presente trabalho foram desenvolvidos e validados métodos para a

determinação analítica de metais contaminantes dos vinhos, com recurso às técnicas analíticas de

ETAAS e ICP-MS, e uma metodologia inovadora de preparação de mostos por ultra-sons de alta

intensidade e focalização. Os métodos de análise desenvolvidos foram posteriormente aplicados

nas etapas seguintes do trabalho, nomeadamente no estudo da cedência ao vinho de metais

contaminantes das bentonites.

As técnicas analíticas de ETAAS e de ICP-MS apresentam limites analíticos, selectividade

e precisão bastante apropriados ao doseamento dos metais contaminantes dos vinhos. A

possibilidade de dosear vários elementos em simultâneo torna a técnica de ICP-MS especialmente

interessante. No capítulo 2 deste trabalho apresentou-se uma metodologia inovadora de preparação

de mosto e também o desenvolvimento e validação de métodos de análise para determinação do

teor de Al e de Cu por ETAAS, e de um vasto conjunto de elementos por ICP-MS.

O controlo analítico do Al no vinho não é realizado por rotina em laboratórios de controlo

da qualidade, não obstante os efeitos importantes e diversificados na sua qualidade. Acresce que a

presença exógena deste elemento no vinho é preponderante, estando actualmente identificadas

numerosas fontes de contaminação. O método analítico desenvolvido para doseamento do Al no

vinho por ETAAS (secção 2.1 do capítulo 2) apresenta elevada selectividade, precisão e exactidão,

permitindo a quantificação em amostras que contenham mais de 80 µg/L, teor largamente excedido

pela generalidade dos vinhos. Os limites analíticos, precisão e taxas de recuperação observados são

similares aos valores indicados por outros autores (Almeida et al., 1992; Lopez et al., 1998). A

elevada diluição da amostra (1:40), factor potencial de erro associado ao resultado analítico,

minimiza os efeitos de matriz tornando desnecessária a adição de modificadores desta. Uma das

principais dificuldades do método proposto consiste no controlo de potenciais fontes de

contaminação durante o processo analítico, dada a presença importante deste elemento também no

ambiente laboratorial. Contudo, a simplicidade da preparação da amostra e a dispensa de utilização

de modificadores de matriz, contribuem para a redução do risco de contaminação.

Em consequência da sua participação em fenómenos de instabilidade físico-química, o Cu

é um dos metais contaminantes do vinho com maior interesse enológico e cujo teor é alvo de

controlo legal (OIV, 2005). A técnica analítica de FAAS apresenta sensibilidade insuficiente para a

determinação analítica do teor de Cu da generalidade dos vinhos, tornando necessária a pré-

concentração da amostra ou o recurso ao método das adições de padrão (Catarino et al., 2003). Na

Page 150: Tese doutoramento Sofia Catarino.pdf

Considerações finais

148

secção 2.2 do capítulo 2 apresentou-se o estudo de diferentes metodologias para o doseamento do

Cu em vinhos, por ETAAS, orientado para a obtenção da sensibilidade necessária e para a

minimização do custo económico da análise.

Uma das desvantagens das determinações convencionais por ETAAS é o tempo necessário

para produzir um resultado analítico. Numa tentativa de minimizar essa desvantagem,

desenvolveram-se programas térmicos alternativos, um programa convencional e um programa sem

a etapa de pirólise (rápido), com ou sem adição de modificadores de matriz, embora o programa

convencional conduza a maior sensibilidade do método. Quando aplicado o programa

convencional, a utilização de modificadores de matriz [Mg(NO3)2 e Pd(NO3)2] para estabilização

térmica do Cu resultou no aumento da sensibilidade (cerca de 20%). Contrariamente, quando

aplicado o programa sem etapa de pirólise, a utilização de modificadores de matriz não afectou de

modo significativo a sensibilidade. Este método apresenta assim como principais vantagens a

dispensa da utilização de modificadores de matriz, maior rapidez no fornecimento de resultados

analíticos e consequentemente menor custo económico. A sensibilidade analítica do método com

aplicação do programa convencional constitui a sua vantagem comparativa.

O método com recurso ao programa térmico convencional (seleccionado para validação)

apresenta características de selectividade, precisão e exactidão elevadas, de acordo com o nível de

qualidade exigido aos resultados analíticos. Por último, o limite de detecção do método em

amostras não diluídas, 5 µg/L, é perfeitamente adequado para fins de controlo legal permitindo

igualmente o doseamento na generalidade dos vinhos, sendo ligeiramente inferior ao valor de 6

µg/L, referido por Almeida et al. (1994).

A técnica multielementar de ICP-MS constitui um excelente instrumento para a

caracterização da composição mineral do vinho. Na secção 2.4 do capítulo 2 apresentou-se o

desenvolvimento de um método quantitativo e de um método semi-quantitativo para a

determinação de 26 metais contaminantes do vinho. Numa fase inicial as potenciais interferências

espectrais causadas por iões poliatómicos com origem nos elementos Ca e Cl foram estudadas e

estabelecidas correcções matemáticas sobre os isótopos 51V, 53Cr, 57Fe, 59Co, 60Ni, 75As e 107Ag. Os

limites de detecção de ambos os métodos são suficientemente baixos para permitir o controlo da

qualidade do vinho, e 10 a 100 vezes inferiores aos limites de detecção encontrados por outros

autores (Baxter et al., 1997; Pérez-Jordan et al., 1998; Almeida e Vasconcelos, 2002; Taylor et al.,

2003; Gómez et al., 2004), utilizando equipamentos de ICP-MS com dispositivo para separação de

massas semelhante (quadropolo). Os limites de detecção do método quantitativo para o Cu, Zn, As,

Cd e Pb, elementos para os quais se encontram definidos limites máximos aceitáveis (OIV, 2005),

são de 0,01; 0,02; 0,002; 0,001 e 0,001 µg/L, respectivamente. Para a maioria dos elementos o

método semi-quantitativo apresenta limites de detecção similares ou ligeiramente superiores aos do

método quantitativo, como seria de esperar.

Page 151: Tese doutoramento Sofia Catarino.pdf

Considerações finais

149

O estudo das taxas de recuperação, conduzido para avaliação da exactidão do método,

permitiu observar valores entre os 80 e 120% para todos os elementos, com excepção do Cr. As

interferências poliatómicas que afectam o isótopo 53Cr, não foram eficazmente corrigidas pelo que

se considerou não ser possível quantificar o Cr pelo método proposto. As taxas de recuperação

referidas são similares às taxas de recuperação encontradas por outros autores (Baxter et al., 1997;

Pérez-Jordan et al., 1998) para vinhos de mesa, utilizando tratamento da amostra similar e com

recurso a nebulizadores tipo crossflow. O método quantitativo e o método semi-quantitativo são

comparáveis em termos de precisão. As diferenças entre os resultados analíticos obtidos por

aplicação dos dois métodos são inferiores a 20% (erro do método semiquantitativo) para todos os

elementos estudados, com excepção do Fe, Zn, Se, Cd, In, Tl e U (em vinho branco) e do Al, Fe,

Se, Cd, In, Bi e U (em vinho tinto). A calibração realizada no método semi-quantitativo, utilizando

uma concentração única para todos os elementos independentemente do seu nível de concentração

no vinho, poderá em parte explicar as diferenças observadas para os elementos situados nos

extremos da gama de trabalho, nomeadamente para o Se, Cd, In, Tl e U (extremo inferior), e para o

Fe, Zn e Al (extremo superior).

De um modo geral, o método semi-quantitativo conduziu a resultados analíticos

ligeiramente inferiores aos resultados analíticos fornecidos pelo método quantitativo, o que está de

acordo com observações de Almeida e Vasconcelos (2002). Para alguns elementos, Fe, Cu, Zn, Cd

e Pb, os resultados analíticos obtidos pelo método quantitativo foram validados por comparação

com os resultados obtidos por AAS. Para fins de controlo legal do Zn e Cd, o método semi-

quantitativo desenvolvido não é adequado, uma vez que não apresenta a exactidão necessária.

Finalmente, vinhos provisoriamente certificados foram analisados pelo método

quantitativo, verificando-se a ocorrência de erro sistemático, com as concentrações obtidas sempre

inferiores às concentrações certificadas, provavelmente explicado pela diminuição do conteúdo

mineral do vinho por precipitação. Em conclusão, ambos os métodos representam alternativas

valiosas com vista à caracterização mineral do vinho, nomeadamente para fins de comparação de

amostras, sendo que o método semi-quantitativo fornece resultados analíticos mais rapidamente e

com menor custo económico. Contudo, como instrumento para controlo legal o método

quantitativo é sem dúvida o mais adequado.

O mosto contém uma quantidade elevada de matéria orgânica que se torna necessário

eliminar de modo a evitar interferências na análise por ETAAS ou ICP-MS. Os métodos clássicos

de digestão apresentam um risco importante de introdução de contaminação mineral inerente às

manipulações da amostra e introdução de reagentes, não sendo por isso adequados para análise de

metais vestigiais. Actualmente, a metodologia mais vulgarizada consiste na digestão por

microondas, o que implica obviamente a utilização de equipamento apropriado, nem sempre

disponível nos laboratórios de enologia. Por outro lado, os processos de oxidação avançada

Page 152: Tese doutoramento Sofia Catarino.pdf

Considerações finais

150

possuem numerosas vantagens em relação à digestão por microondas (Capelo-Martinez et al.,

2004).

Na secção 2.3 do capítulo 2 apresentou-se o desenvolvimento e validação de uma

metodologia inovadora para extracção dos metais ligados à matéria orgânica do mosto, consistindo

na aplicação de ultra-sons de alta intensidade e focalização, conjugados com HNO3 e H2O2. Após a

determinação do teor de Pb (elemento modelo) por ETAAS, em três laboratórios diferentes, os

resultados do tratamento proposto foram comparados com os resultados obtidos por digestão por

microondas, não tendo sido observadas diferenças estatisticamente significativas. No âmbito da

análise do Pb por ETAAS foram desenvolvidos diferentes programas térmicos, convencional e sem

etapa de pirólise, em associação com diferentes modificadores de matriz [Pd(NO3)2; KMnO4;

KMnO4 + HCl; Pd(NO3)2]. A determinação do Pb no mosto por ETAAS, sem etapa de pirólise,

revelou-se possível na presença de Pd(NO3)2 como modificador de matriz. Por outro lado, o

programa térmico convencional, com ou sem modificador de matriz, permite o doseamento em

amostras de mosto tratadas por ultra-sons na presença de HNO3 e de H2O2. Das vantagens desta

metodologia, comparativamente às metodologias clássicas e à digestão por microondas, salienta-se

desde logo a necessidade de um volume de amostra e de reagentes reduzido e a rapidez do

tratamento, para além do custo reduzido do equipamento de ultra-sons.

Um aspecto fulcral e comum às metodologias referidas anteriormente para a análise de

elementos minoritários, vestigiais e sub-vestigiais no vinho consiste nos cuidados que

obrigatoriamente devem ser colocados ao longo do processo analítico, de modo a minimizar a

ocorrência de contaminações.

Em conclusão, os resultados constantes do capítulo 2 demonstram a disponibilidade de

métodos de análise, com recurso às técnicas de ETAAS e ICP-MS, adequados ao doseamento de

metais contaminantes nos vinhos, apresentando requisitos de qualidade que permitem depositar

confiança nos seus resultados analíticos, e eventualmente aplicáveis para fins de controlo legal.

Considerando a sua aplicabilidade em laboratórios de enologia e/ou de controlo, a técnica de

ETAAS apresenta como principais vantagens comparativas o pequeno volume de amostra

necessário, ausência de interferências espectrais e custo moderado. Por sua vez, a técnica de ICP-

MS destaca-se por permitir a análise multi-elementar, e elevada produtividade. Contudo, é afectada

por importantes interferências espectrais e de matriz, sendo particularmente exigente ao nível de

desenvolvimento de métodos, e apresenta custos bastante elevados relacionados com a aquisição de

equipamento e com custos operacionais, o que explica a sua ainda escassa implantação nos

laboratórios de enologia (actualmente única em Portugal).

A composição mineral do vinho é influenciada por factores ambientais, tais como

fenómenos geológicos e atmosféricos, e por factores antrópicos. Com vista a estudar a evolução do

Page 153: Tese doutoramento Sofia Catarino.pdf

Considerações finais

151

teor em metais contaminantes ao longo do processo tecnológico, avaliou-se a composição de sete

vinhos (obtidos por microvinificação) e dos respectivos mostos de origem em Li, Be, Al, V, Mn,

Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Cd, In, Cs, Ba, Hg, Tl, Pb, Bi e U (capítulo 3). As amostras

de mosto foram preparadas para análise por aplicação da metodologia com recurso a ultra-sons de

alta intensidade e focalização para preparação de mostos para análise (secção 2.3 do capítulo 2). O

conteúdo mineral de mostos e vinhos foi analisado pelo método quantitativo por ICP-MS (secção

2.4 do capítulo 2).

Os resultados comprovam uma importante capacidade natural de eliminação de metais

contaminantes por parte dos mostos, sendo esta a principal conclusão a destacar. Para um

importante conjunto de elementos, Be, Al, Mn, Co, Ni, Cu, Zn, Ga, Rb, Cd, Ba, Tl and U,

observaram-se decréscimos importantes do mosto para o vinho, provavelmente devido a

precipitações sob a forma de sulfuretos e a fenómenos de adsorção e absorção por leveduras e

bactérias. Essas perdas, como seria de esperar, foram mais acentuadas nos vinhos com pH mais

elevado. O decréscimo do teor de Cu foi particularmente relevante, representando em média cerca

de 90% do teor inicial, atingindo cerca de 8 mg/L. Verificou-se nos vinhos tintos um

enriquecimento em Li, Fe, Sr, In e Bi, provavelmente relacionado com a tecnologia de vinificação

aplicada, concretamente com a maceração das partes sólidas. Em duas amostras verificou-se um

aumento ligeiro do teor de Pb (15 µg/L, aproximadamente) do mosto para o vinho, evidenciando a

ocorrência de contaminação de origem não identificada. Para as restantes amostras foram

observadas reduções importantes no teor deste elemento, corroborando observações de outros

autores (Teissèdre et al., 1993; Pellerin et al., 1997).

As concentrações dos elementos analisados são inferiores aos limites máximos

estabelecidos e de um modo geral podem considerar-se reduzidas, especialmente quando

confrontadas com a literatura (McKinnon et al., 1992; Sudraud et al., 1994; Eschnauer e Scollary,

1995; Greenough et al., 1997; Teissèdre et al., 1998a; Teissèdre et al., 1998b; Lima et al., 2004;

Thiel et al., 2004). No entanto, na interpretação e discussão destes resultados deverá ser incluído o

processo tecnológico aplicado na obtenção dos vinhos, nomeadamente a não utilização de

auxiliares tecnológicos, à excepção de dióxido de enxofre. De facto, na produção de vinhos à escala

industrial, e também à escala artesanal, são utilizados diversos produtos susceptíveis de introduzir

contaminação mineral, sendo também maior o número de operações tecnológicas que implicam o

contacto do vinho com materiais.

Do conjunto de auxiliares tecnológicos autorizados pela OIV, a bentonite é certamente um

dos produtos de utilização mais generalizada na produção de vinho. A sua aplicação visa

essencialmente a clarificação e estabilização proteica dos vinhos, representando contudo um risco

bastante importante de contaminação mineral. O conhecimento dos fenómenos de extracção e

cedência de elementos minerais para o meio, e dos factores com influência nesses fenómenos,

Page 154: Tese doutoramento Sofia Catarino.pdf

Considerações finais

152

sendo fundamental para prevenir a ocorrência dessas contaminações, é ainda manifestamente

insuficiente. Acresce que a oferta deste produto no mercado é bastante diversificada e a informação

sobre a sua potencial cedência mineral para o vinho muito escassa.

A utilização de bentonites no âmbito da sua influência na composição mineral do vinho

apresenta diferentes vertentes de estudo. No capítulo 5, numa fase inicial desta linha de trabalho,

realizaram-se ensaios de extracção com vinho e soluções teste, utilizando bentonites adquiridas no

mercado. A análise da composição mineral das bentonites e das soluções de extracção foi realizada

por AAS e ICP-MS (métodos descritos no capítulo 2). O efeito de contaminação mineral

provocado pela adição de bentonites ao vinho foi confirmado para um largo conjunto de elementos

(Li, Be, Na, Mg, Al, Ca, Sc, V, Mn, Fe, Co, Ni, Ga, Ge, As, Sr, Y, Zr, Nb, Mo, Cd, Sn, Sb, Ba, W,

Tl, Bi e W). Em oposição, e de acordo com observações de outros autores (Enkelmann, 1988;

Jakubowski et al., 1999; Gómez et al., 2004; Nicolini et al., 2004), a adição de bentonite provocou

um decréscimo nos níveis de B, K, Cu, Zn e Rb. Importa salientar que a dose de bentonite utilizada

nos ensaios de extracção foi bastante superior às doses normalmente utilizadas na prática

enológica, pelo que os resultados não podem ser directamente transpostos. Estes resultados devem

ser considerados nos trabalhos que visam estabelecer uma relação directa entre a composição

mineral de um vinho com a sua origem geográfica. Esta é ainda uma área de trabalho envolvida em

alguma polémica (Jakubowski et al., 1999; Gómez et al., 2004; Thiel et al., 2004), à

semelhança de outros trabalhos que visam a determinação da origem do vinho em função da

presença e/ou conteúdo em determinada espécie química.

Verificaram-se diferenças significativas entre as bentonites no que respeita ao seu efeito

sobre a composição mineral do vinho, revelando a importância das suas características físico-

químicas nos fenómenos de extracção. À diversidade natural das bentonites acresce a diversidade

natural da composição dos vinhos, o que torna ainda mais complexa a previsão do efeito de

determinada bentonite.

Uma importante e inovadora vertente desta linha de trabalho consistiu no estudo da relação

das características físico-químicas da bentonite com a cedência mineral ao vinho. Verificou-se a

existência de uma correlação forte entre a cedência de Na ao vinho e a proporção deste elemento no

complexo de troca de cada bentonite (r = 0,881, P = 95%). Foram ainda constatadas correlações

importantes entre a cedência de Al e de Fe e o conteúdo das bentonites em compostos não

cristalinos de Al e Fe (r = 0,784 e r = 0,808, respectivamente; P = 95%). Para o Be, Mg, Ca, V, Mn,

Ni, Ge, Zr, Nb, Mo, Sn, Sb, Tl, Pb e U observaram-se boas correlações com os resultados da

composição multi-elementar das bentonites.

Actualmente, a OIV estabelece um conjunto de especificações relativas a características

físico-químicas que as bentonites deverão respeitar com vista à sua utilização enológica (Resolução

OIV Oeno 11/2003). De salientar que algumas bentonites estudadas apresentaram características

não conformes com essas especificações: o teor mínimo de montmorilonite estabelecido não foi

Page 155: Tese doutoramento Sofia Catarino.pdf

Considerações finais

153

respeitado por nenhuma das bentonites; uma das bentonites apresentou um teor em Fe extraível

superior ao limite máximo; o conteúdo em Na solúvel das bentonites naturais (não activadas) foi

superior ao teor máximo estabelecido. O controlo da qualidade das bentonites não é realizado por

rotina por nenhum laboratório nacional de controlo oficial, o que permite supor o desconhecimento

sobre a qualidade dos produtos e a circulação de produtos não conformes.

Outra conclusão importante retirada deste trabalho foi que a percentagem de

montmorilonite não é um indicador satisfatório da reactividade da bentonite, dado que as

características da fracção argilosa revelaram ser mais importantes do que a sua quantidade. Um

outro aspecto observado, com especial importância tecnológica, diz respeito ao efeito significativo

do pH na cedência mineral, contrariando estudos realizados anteriormente (Nicolini et al., 2004).

De um modo geral, foram observados maiores enriquecimentos ao nível de pH mais baixo. Esta

informação deverá ser tida em conta de modo a minimizar o risco de contaminação ao longo do

processo tecnológico.

No âmbito da OIV, para controlo da qualidade das bentonites no que respeita ao seu efeito

potencial de contaminação mineral, é estabelecida a realização de ensaios de extracção com uma

solução teste cuja composição consiste em uma solução de ácido tartárico a 5 g/L (Resolução OIV

Oeno 11/2003).

A comparação entre o enriquecimento verificado pelo vinho e o enriquecimento verificado

pela solução teste indicada pela OIV permitiu observar resultados muito diferentes, pelo que se

conclui que a solução de ácido tartárico não é indicada para fins de controlo da qualidade das

bentonites, uma vez que não simula o efeito de extracção do vinho. Verificou-se que em algumas

amostras de bentonite os teores extraídos de Na, Al, Fe e As superaram os limites máximos

estabelecidos (Resolução OIV Oeno 11/2003), o que, mais uma vez, permite colocar algumas

questões sobre a qualidade destes produtos e o seu controlo oficial.

Constatada a inadequação da solução teste preconizada pela OIV, ensaiaram-se soluções de

extracção com composições alternativas em paralelo com o vinho. A diversidade dos vinhos no que

respeita à sua composição química (constituindo um forte obstáculo à sua padronização) é mais um

factor com influência na cedência mineral, e que acresce à diversidade natural das bentonites.

Elementos com acção directa nos fenómenos de troca catiónica (K, Mg, Na, Mg) e proteína

(proteína extraída de vinho e BSA) foram incluídos nas soluções de extracção, de modo a simular a

composição do vinho. A natureza da proteína mostrou ser um factor importante, provavelmente

relacionado com o efeito do seu volume no processo de adsorção, o que está de acordo com

conclusões de Achaerandio et al. (2001). De um modo geral a solução com proteína de vinho foi

capaz de simular ligeiramente melhor o vinho do que a solução com BSA. Embora nenhuma das

duas soluções tenha simulado o vinho de modo totalmente satisfatório foram retiradas indicações

de grande utilidade para o prosseguimento desta linha de trabalho.

Page 156: Tese doutoramento Sofia Catarino.pdf

Considerações finais

154

Em conclusão, os resultados alcançados com este trabalho confirmaram as bentonites como

uma importante fonte de contaminação mineral dos vinhos, tendo sido detectadas amostras deste

auxiliar tecnológico não conformes com as especificações para uso enológico. O controlo da

qualidade das bentonites, que urge intensificar, beneficiará com o estabelecimento de uma solução

teste representativa do vinho, objectivo para o qual se contribuiu.

No seguimento natural deste trabalho, e para além do estabelecimento de uma solução teste

para controlo da qualidade das bentonites, diversas linhas de investigação que consideramos de

interesse relevante poderão ser desenvolvidas, representando um importante contributo para o

aumento do conhecimento sobre a ocorrência de metais contaminantes nos vinhos. A título de

exemplo, apresenta-se uma descrição necessariamente breve de três áreas de trabalho:

1. O estudo da distribuição de metais contaminantes nos vinhos sob formas químicas

minerais e orgânicas, e dos fenómenos de complexação com outras espécies químicas. Tal

como referido no capítulo 1, a forma química em que o metal se apresenta influencia

decisivamente a sua biodisponibilidade e toxicidade, aspecto que actualmente não é

contemplado na definição dos limites legais. Por outro lado, a distribuição do elemento sob

diferentes formas químicas pode ser em alguns casos um bom indicador do potencial de

ocorrência de fenómenos de instabilidade físico-química.

2. A análise da composição isotópica de determinados metais contaminantes dos vinhos

poderá ser uma ferramenta bastante útil na identificação de fontes de contaminação.

3. Por fim, a caracterização multi-elementar, por si só ou aliada ao estudo das razões

isotópicas de alguns elementos (ambas com recurso à técnica de ICP-MS), não obstante a

reserva natural face às sucessivas contaminações minerais susceptíveis de ocorrer ao longo

do processo tecnológico, poderão ser uma útil fonte de informação para o estabelecimento

de padrões regionais com vista à identificação da proveniência geográfica de um mosto ou

vinho. Esta linha de trabalho, actualmente em desenvolvimento em outros países

vitivinícolas, representaria um forte contributo para o conhecimento sobre os vinhos

portugueses, e nesse sentido para o sector vitivinícola português.

Page 157: Tese doutoramento Sofia Catarino.pdf

Considerações finais

155

5.2. REFERÊNCIAS BIBLIOGRÁFICAS Achaerandio I., Pachova V., Güell C., López F., 2001. Protein adsorption by bentonite in a white wine model solution: effect of protein molecular weight and ethanol concentration. American Journal of Enology andViticulture, 52 (2), 122-126.

Almeida A.A., Bastos M.L., Cardoso M.I., Ferreira M., Lima J.L.F.C., Soares M.E., 1992. Determination of lead and aluminium in port wine by electrothermal atomic absorption spectrometry. Journal of Analytical Atomic Spectrometry, 7, 1281-1285.

Almeida A.A., Cardoso M.I., Lima J.L.F.C., 1994. Determination of copper in Port wine and Madeira wine by electrothermal atomization AAS. Atomic Spectroscopy, 15 (2), 73-77.

Almeida C.M., Vasconcelos M.T.S.D., 2002. Advantages and limitations of the semi-quantitative operation mode of an inductively coupled plasma-mas spectrometer for multi-element analysis of wine. Analytica Chimica Acta, 463, 165-175.

Baxter M.J., Crews H.M., Dennis M.J., Goodall I., Anderson D., 1997. The determination of the authenticity of wine from its trace element composition. Food Chemistry, 60 (3), 443-450.

Capelo-Martinez J.L., Ximenez-Embún P., Madrid Y., Cámara C., 2004. Advanced oxidation processes for sample treatment in atomic spectrometry. Trends in Analytical Chemistry, 23 (4), 331-340.

Catarino S., Pinto D., Curvelo-Garcia A.S., 2003. Validação e comparação de métodos de análise em espectrofotometria de absorção atómica com chama para doseamento de cobre e ferro em vinhos e aguardentes. Ciência e Técnica Vitívinícola, 18 (2), 65-76.

Enkelmann R., 1988. Migration of heavy metals from bentonites to wine. 1. Note: Bentonites. Deutsche Lebensmittel-Rundschau, 84, 243-247.

Eschnauer H.R., Scollary G.R., 1995. Aluminium in wein. Wein-Wissenschaft, 1, 24-30.

Gómez M.D.L.C., Brandt R., Jakubowski N., Andersson J.T., 2004. Changes of the metal composition in German white wines through the winemaking process. A study of 63 elements by inductively coupled plasma – mass spectrometry. Journal of Agriculture and Food Chemistry, 52, 2953-2961.

Greenough J.D., Longerich H.P., Jackson S.E., 1997. Element fingerprint of Okanagan Valley wines using ICP-MS: relationships between wine composition, vineyard and wine colour. Australian Journal of Grape and Wine Research, 3, 75-83.

Jakubowski N., Brandt R., Stuewer D., Eschnauer H.R., Görtges S., 1999. Analysis of wines by ICP-MS: Is the pattern of the rare earth elements a reliable fingerprint for the provenance?. Fresenius Journal of Analytical Chemistry, 364, 424-428.

Lima M.T.R., Cabanis M.T., Matos L., Cassanas G., Kelly M., Blaise A., 2004. Determination of lead and cadmium in vineyard soils, grapes and wines of the Azores. Journal International des Sciences de la Vigne et du Vin, 38, 163-170.

Lopez F.F., Cabrera C., Lorenzo M.L., Lopez M.C., 1998. Aluminium levels in wine, beer and other alcoholic beverages consumed in Spain. The Science of the Total Environment, 220, 1-9.

McKinnon A.J., Cattrall R.W., Scollary G.R., 1992. Aluminum in wine – its measurement and identification of major sources. American Journal of Enology and Viticulture, 43 (2), 166-170.

Nicolini G., Larcher R., Pangrazzi P., Bontempo L., 2004. Changes in the contents of micro- and trace-elements in wine due to winemaking treatments. Vitis, 43 (1), 41-45.

OIV, 2005. Recueil des methodes internationals d’analyse des vins et des mouts. Organisation Internationale de la Vigne et du Vin, Paris.

Pellerin P., O’Neill M.A., Pierre C., Cabanis M-T., Darvill A.G., Albersheim P., Moutounet M., 1997. Complexation du plomb dans les vins par les dimères de rhamnogalacturonane II, un polysaccharide pectique du raisin. Journal International des Sciences de la Vigne et du Vin, 31 (1), 33-41.

Pérez-Jordán M.Y., Soldevila J., Salvador A., Pastor A., de la Guardia M., 1998. Inductively coupled plasma mass spectrometry análisis of wines. Journal of Analytical Atomic Spectrometry, 13, 33-39.

Page 158: Tese doutoramento Sofia Catarino.pdf

Considerações finais

156

Sudraud P., Médina B., Grenon J.P., 1994. Teneurs en éléments minéraux des vins. Journal International des Sciences de la Vigne et du Vin, 28 (1), 69-75.

Taylor V.F., Longerich H.P., Greenough J.D., 2003. Multielement analysis of Canadian wines by inductively coupled plasma mass spectrometry (ICP-MS) and multivariate statistics. Journal of Agriculture and Food Chemistry, 51, 856-860.

Teissèdre P.L., Cabrera Vique C., Cabanis M.T., Cabanis J.C., 1998a. Determination of nickel in French wines and grapes. American Journal of Enology and Viticulture, 49 (2), 205-210.

Teissèdre P.L., Krosniak M., Portet K., Gasc F., Waterhouse A.L., Serrano J.J., Cabanis J.C., Cros G., 1998b. Vanadium levels in French and Californian wines: influence on vanadium dietary intake. Food Additives and Contaminants, 15 (5), 585-591.

Thiel G., Geisler G., Blechschmidt I., Danzer K., 2004. Determination of trace elements in wines and classification according to their provenance. Analytical and Bioanalytical Chemistry, 378, 1630-1636.

Page 159: Tese doutoramento Sofia Catarino.pdf

157

Divulgação do conhecimento

PUBLICAÇÕES EM REVISTAS CIENTÍFICAS INTERNACIONAIS COM REFEREES

Catarino S., Curvelo-Garcia A.S., Bruno de Sousa R., 2002. Determinação analítica do zinco em vinhos por espectrofotometria de absorção atómica com chama. Validação do método de análise. Ciência e Técnica Vitívinícola, 17 (1), 15-26.

Catarino S., Curvelo-Garcia A.S., Bruno de Sousa R., 2002. Determination of aluminum in wine by graphite furnace AAS: Validation of Analytical Method. Atomic Spectroscopy, 23 (6), 196-200.

Catarino S., Pinto D., Curvelo-Garcia A.S., 2003. Validação e comparação de métodos de análise em espectrofotometria de absorção atómica com chama para doseamento de cobre e ferro em vinhos e aguardentes. Ciência e Técnica Vitívinícola, 18 (2), 65-76.

Catarino S., Soares J., Curvelo-Garcia A.S., Bruno de Sousa R., 2004. Implicações da utilização de bentonites sobre a fracção mineral de vinhos: potássio, sódio, cálcio, alumínio e chumbo. Efeito do pH. Ciência e Técnica Vitívinícola, 19 (1), 29-45.

Catarino S., Pimentel I., Curvelo-Garcia A.S., 2005. Determination of copper in wine by ETAAS using conventional and fast thermal programs: validation of analytical method. Atomic Spectroscopy, 26 (2), 73-78.

Catarino S., Capelo J.L., Curvelo-Garcia A.S., Vaiâo M., 2005. Focused Ultrasound (FU) vs Microwave Digestion for the Determination of Lead in Must by Electrothermal-Atomic Absorption Spectrometry. Journal of AOAC International, 88 (2), 585 – 591.

Maduro C., Vale G., Alves S., Galesio M., Gomes da Silva M.D.R., Fernandez C., Catarino S., Rivas M.G., Mota A.M., Capelo J.L., 2006. Determination of Cd and Pb in biological reference materials by electrothermal atomic absorption spectrometry: a comparison of three ultrasonic-based sample treatment procedures. Talanta, 68, 1156-1161.

Catarino S., Capelo J-L., Curvelo-Garcia A.S., Bruno de Sousa R., 2006. Evaluation of contaminant elements in Portuguese wines and origin musts by inductively coupled plasma mass spectrometry. Journal International des Sciences de la Vigne et du Vin, 40 (2), 91-100.

Catarino S., Madeira M., Monteiro F., Curvelo-Garcia A.S., Bruno de Sousa R., 2006. Release of contaminant elements from bentonites to wine: a contribution to achieve a test solution. Ciência e Técnica Vitivinícola, 21 (1), 17-31.

Catarino S., Curvelo-Garcia A.S., Bruno de Sousa R., 2006. Measurements of contaminant elements of wines by inductively coupled plasma mass spectrometry: a comparison of two calibration approaches. Talanta, 70, 1073-1080.

Catarino S., Madeira M., Monteiro F., Rocha F., Curvelo-Garcia A.S., Bruno de Sousa R. Release of contaminant elements from bentonites to wine: effect of bentonite characteristics and wine pH. Journal of Agricultural and Food Chemistry (a submeter para publicação).

Mira H., Leite P., Catarino S., Ricardo-da-Silva J., Curvelo-Garcia A.S. PVI-PVP copolymer use for wine metal reduction. Effects on wine characteristics. Vitis (submetido para publicação).

COMUNICAÇÕES EM LIVROS DE ACTAS DE ENCONTROS CIENTÍFICOS

Comunicações orais

Catarino S., Soares J., Curvelo-Garcia A.S., Bruno de Sousa R., 2004. Implicações da utilização de bentonites sobre a fracção mineral de vinhos. In: 6º Simpósio de Vitivinicultura do Alentejo. vol. 2, pp.927-933. Évora.

Page 160: Tese doutoramento Sofia Catarino.pdf

158

Comunicações em painel

Catarino S., Curvelo-Garcia A.S., Sousa R.B., 2003. Controlo analítico do alumínio em vinhos. Optimização e validação de um método de análise por espectrofotometria de absorção atómica com atomização electrotérmica. In: 6º Encontro de Química dos Alimentos. pp. 927-933. Lisboa.

Mira H., Leite P., Ricardo da Silva J.M., Catarino S., Curvelo-Garcia A.S., 2004. Influência do copolímero PVI-PVP nas características do vinho. In: 6 Simpósio de Vitivinicultura do Alentejo. vol.2, pp. 272-280, Évora.

Capelo J.L., Galésio M., Catarino S., Curvelo-Garcia A.S., 2005. Focused ultrasound (FU) vs. microwave digestion for the determination of lead in must by electrothermal-atomic absorption spectrometry. In: 7º Encontro de Química dos Alimentos. pp.108. Viseu.

Catarino S., Soares J., Curvelo-Garcia A.S., Bruno de Sousa R., 2005. Measurements of contamination elements of wines by inductively coupled plasma-mass spectrometry (ICP-MS): a comparison of the quantitative and semi-quantitative operation modes. In: In Vino Analytica Scientia. pp. 87. Montpellier, França. Trabalho distinguido com 3º prémio.

AULAS LECCIONADAS

Segurança alimentar - Metais pesados, no I Curso avançado sobre Vinho - Qualidade e Segurança Alimentar, Estação Vitivinícola Nacional, 4 de Novembro de 2002.

Metais pesados em vinhos, no âmbito da disciplina de Química Enológica, Instituto Superior de Agronomia, Novembro de 2002.

Metais pesados em vinhos, colaboração no Curso de Pós-Graduação “Especialista en Elaboración y Análisis de Vinos” (Programa Leonardo Da Vinci), módulo de Análise de Metais Pesados, Estação Vitivinícola Nacional, 27 de Novembro de 2002.

Metais pesados em vinhos, no âmbito da disciplina de Química Enológica, Instituto Superior de Agronomia, Novembro de 2003.

Segurança alimentar - metais pesados, no colóquio XXXIII “Valorização do património vitivinícola português pela qualidade, diversidade e segurança alimentar dos seus produtos”, Estação Vitivinícola Nacional, 23 de Novembro de 2004.

Metais pesados em vinhos, no âmbito da disciplina de Química Enológica, Instituto Superior de Agronomia, Dezembro de 2004.

Fenómenos de instabilidade físico-química do vinho, no Curso de Iniciação à Prova de Vinhos, Estação Vitivinícola Nacional, 15 de Abril de 2005.

Metais pesados, no curso “Segurança Alimentar de Vinhos”, Estação Vitivinícola Nacional, 7 de Março de 2006.

Acreditação de laboratórios de ensaio, no 4º Curso de Mestrado em Viticultura e Enologia, módulo de Controlo da Qualidade, Estação Vitivinícola Nacional, 17 de Maio de 2006.

Metais pesados, no 4º Curso de Mestrado em Viticultura e Enologia, módulo de Controlo da Qualidade, Estação Vitivinícola Nacional, 18 de Maio de 2006.

Alterações físico-químicas, no Curso de Iniciação à Prova de Vinhos, Estação Vitivinícola Nacional, 29 de Junho de 2006.

Metais pesados em vinhos, no âmbito da disciplina de Química Enológica, Instituto Superior de Agronomia, 8 de Novembro de 2006.