1/8

Evaluation of available pre-mixed replacement external renderings Luís MATEUS, Maria DO ROSÁRIO VEIGA, Jorge DE BRITO

Article disponible dans les actes du colloque Terra 2016: JOFFROY, Thierry, GUILLAUD, Hubert, SADOZAÏ, Chamsia (dir.) 2018, Terra Lyon 2016: Articles sélectionnés pour publication en ligne / articles selected for on-line publication / artículos seleccionados para publicación en línea. Villefontaine : CRAterre. ISBN 979-10-96446-12-4.

2/8

SUMMARY The Algarve region in southern Portugal has one of the most significant rammed earth con-struction legacies in the country. Some modern architects are adapting this kind of material to new construction. This construction method is characterized by its high susceptibility to water damage. The main objective of this characterization study of the renderings is to iden-tify composition and behaviour patterns in order to understand the natural degradation phe-nomena and how they are associated with the background characterization. Another pur-pose is the search for pre-mixed renderings compositions, potentially compatible with this kind of substrate and suitable for the repair of damaged external renderings. Fifteen earth buildings were selected to represent this heritage and its composition diversity. Rendering and substrate samples were extracted from those constructions. Hydric, physical, mechani-cal and chemical properties of ancient renderings are evaluated through laboratory tests. In situ tests are carried out to evaluate hydric and mechanical properties of some pre-mixed rendering solutions available in the market. Laboratory and in situ results are discussed. The conclusions are presented regarding the performance of ancient and pre-mixed renderings. None of the pre-mixed renderings show an ideal behaviour, regarding the features tested.

1. INTRODUCTION Within a PhD study at Instituto Superior Técnico, characterization tests were performed by the first author at the National Laboratory for Civil Engineering (LNEC) and on site, allow-ing the establishment of analysis criteria aiming at understanding the degradation phe-nomena of ancient renderings and the potential use of pre-mixed renderings available in the market for ancient buildings. The use of protective renderings is common in Portugal, unlike in other regions of the world. These constructions have external renderings made of coats of mortar with lime, clay and sand, made on site. Renderings are the first physical barrier inhibiting the direct contact of water with rammed earth substrates, preventing leaching of the fine aggregates and loss of load bearing cross-section (Magalhães and Veiga 2009). The renderings must also have high permeability, allowing the exchange of water vapour between the earth substrate and the exterior, preventing large periods of wa-ter retention in the substrate and rendering. The main degradation phenomena are also related to the poor resistance of the earth substrate to horizontal actions. The main objec-tive of the characterization process of the renderings is to understand the natural degrada-tion phenomena and how they are associated with the background characterization (Ma-teus et al. 2014). A set of ancient rendering samples was assembled for their hydric, phys-ical, mechanical and chemical characterization in laboratory. In addition, some pre-mixed mortars were applied in a rammed earth construction, in order to evaluate some of their hydric and mechanical features using in situ procedures. According to the manufacturers information, CBR1, CBR3 and CBR5 pre-mixed renderings are hydraulic lime based mixes with synthetic fibres, CBR2 is an air lime based mix, and CBR4 is a hydraulic-based mix with cork aggregates.

2. CHARACTERIZATION METHODOLOGY The laboratory campaign concerned ancient renderings and substrates. The collection of samples of ancient rammed earth buildings focused the following aspects: i) type of sub-strate: rammed earth; ii) type of coating: exterior; iii) geographical area: western Algarve;

3/8

iv) type of construction: current dwelling 1 to 2 storeys above ground; v) construction peri-od: from 1850 to 1950. Fifteen ancient buildings were selected to test their renderings. Rendering samples from all the fifteen ancient buildings were submitted to water absorp-tion and drying tests; elasticity modulus; open porosity; insoluble residue; and mineralogy. The capillary absorption and drying tests aimed at defining the level and rate of water ab-sorption of applied renderings and correlating them with the drying rate. The 45 specimens - 3 for each of 15 buildings selected were tested according to LNEC test procedures (test sheet Pa40), developed for irregular sized samples of traditional renderings within previ-ous projects (Veiga et al. 2004). The compressive strength tests aimed at establishing direct relationships between the strength capacity of existing coatings in the buildings selected and strength capacity of the rammed earth substrate (Veiga et al. 2001). For each of the fifteen buildings, 3 rendering samples and 5 substrate samples were tested. These rendering samples have irregular sizes and were extracted using low intrusive tools in order to preserve their integrity. Methods developed at LNEC within earlier research projects on old renderings were used (Valek and Veiga 2005), described in LNEC test procedure (test procedure Pa42), using a loading rate of 100N/s. The elasticity modulus tests aimed at evaluating the renderings stiffness, through the deter-mination of the ultra-sound pulse velocity using the indirect method, suitable for small sam-ples with variable thickness. This methodology, usually applied at LNEC (test sheet Pa43), is based on the European Norm EN 12504-4. For that purpose, one sample was selected for each of the 15 constructions. The open porosity tests purpose is to determine the rendering volume occupied by open pores. This is related with the type of pores that can be filled with water, pore radius distri-bution and connectivity. Therefore, the open porosity is related with the potential infiltration of soluble salts. The methodology used can be applied to regular or irregular samples, ac-cording to the RILEM I.1 and RILEM I.2 procedures. The insoluble residue test aimed at determining the ratio aggregate: binder. The most common binder of ancient renderings is aerial lime. In that kind of rendering, the insoluble residue can be, approximately, associated with the non-calcareous part of it, and mainly composed by the non-calcareous aggregate. Methods developed at LNEC within earlier research projects on old renderings were used (Silva 2003; Veiga et al. 2001). In addition, mineralogy tests were carried out, to identify the main chemical components of these 15 renderings. So, XRD (x-ray diffractometry) and TGA (thermogravimetric analysis) tests were performed (Silva 2003). Mass losses in TGA were determined for the following temperature limits: ambient temperature - 200 ºC, 200-380 ºC, 380-550 ºC, 550-850 ºC and 850-1000 ºC. These tests were performed in the Materials Department at LNEC. Five pre-mixed commercial brand rendering samples were applied in one of the ancient buildings. Each rendering sample is 1.80 x 1.00m wide. In situ tests were performed to evaluate the water absorption and surface hardness, using specific test procedures. These tests were carried out 90 days after in situ application. The Karsten tubes method was also applied to the rammed earth substrate of that specific building. These in situ tests deliv-ered different outputs that can be related with the laboratory ones, although they cannot be directly compared. The water absorption test using Karsten tubes aimed at establishing the commercial renderings response to water contact. The Karsten tubes simulate the real

4/8

scenario of rain exposition. This methodology follows the EN16302 and LNEC Pa39 test procedures. The water absorption was measured every minute in the first 5 minutes, and every 5 minutes for an hour period. For comparison purposes, Karsten tubes tests were also performed in the original rendering (Fig. 1).

Figure 1. Karsten tubes test in commercial renderings and ancient renderings; pendulum hammer test.

To analyse the mechanical compatibility of the pre-mixed renderings, mechanical strength tests were carried out using the Schmidt pendulum hammer model PM (Fig. 1). These tests were carried out 90 days after in situ application, according to RILEM TC 127 MS procedures. This non-intrusive method measures the rendering surface hardness. The output is the rebound of the pendulum mass, measured using the equipment degree scale. Since these renderings are available in the market, compression strength reference values provided by the rendering manufacturers can be taken to evaluate the reliability of this test. Twenty tests were performed per pre-mixed rendering, in a total of 100 tests. Twenty tests were also performed in the original ancient rendering of that specific building where the pre-mixed renderings were applied.

3. RESULTS AND DISCUSSION The highest absorption rate in old renderings occurs in the initial water absorption period (first 5 minutes) and overall water absorption occurs during the first 25 minutes of expo-sure, resulting in high amount of absorbed water by capillarity (Fig. 2); drying periods are significantly larger.This implies a water retention period in the interface between the coat-ing and the rammed earth substrate, and the contact of infiltrated water with the rammed earth substrates and their progressive degradation. The water absorption rates of the tested new commercial renderings have a wide range of values: two of these renderings absorb less water than the building existing rendering. Two of the pre-mixed renderings have much lower absorption rates then the ancient renderings and two others have much higher absorption rates then the ancient renderings (Fig. 3). The results of porosity of existing renderings are consistent with the ones usually obtained in traditional lime renderings; the average value is 26.3%. These porosity results reveal high porous renderings, with high potential of water retention and salt crystallization. In most ancient buildings tested, there is a significant difference between the average compressive strength of old renderings and substrates; in average, the old rendering’s strength is up to 2 times higher than that of the substrate (Fig 4); this generates excessive stresses in the interface with the rammed earth substrate (Binda et al. 2000), promoting loss of adherence between these materials. Pre-mixed renderings CBR1, CBR3 and CBR5 have similar levels of compressive strength and higher than most of the ancient

5/8

renderings tested, meaning low mechanical compatibility of those pre-mixed renderings with the rammed earth substrates. The CBR2 rendering has lower compressive strength and, in spite of that, a higher pendulum rebound, which can be due to its higher adherence to the substrate. It was not possible to perform the pendulum test in the CBR4 rendering because of its very low mechanical resistance (Fig. 4). The pendulum test rebound is not directly related with the level of compressive strength for all the renderings. The results of the elasticity modulus of existing renderings are higher than the ones ob-tained in traditional lime renderings using the same test method (Velosa and Veiga 2012) showing low deformation capacity of the existing renderings; these ancient renderings show mechanical incompatibility with the rammed earth substrates because they have high levels of compressive strength and, in some cases, high values of the modulus of elasticity. The ancient renderings have high contents of insoluble residue (average value is 88.3%) - meaning a low percentage of lime binder. The XRD and TGA tests show that ancient ren-derings have similar chemical composition: sand aggregates, lime and clay (illite and kaolin-ite) (Fig. 5); in most renderings, clay and lime have similar contents in the mixes; the volu-metric ratio aggregates: binder is 6.2, a very high value when compared with traditional lime renderings.

Figure 2. Capillarity water absorption of ancient renderings - laboratory test.

Figure 3. Water absorption of commercial renderings - in situ test (Karsten tubes).

6/8

Figure 4. Mechanical strength evaluation of ancient renderings, substrates and commercial renderings.

Figure 5. XRD (x-ray diffractometry) output for fifteen ancient renderings - laboratory test.

4. CONCLUSIONS The present rebirth of rammed earth construction has various reasons, linked to the acknowl-edgement of its advantages in terms of traditional construction identity to be maintained. How-ever, it also reveals some drawbacks, namely concerning its low resistance to earthquakes and susceptibility to water action. Tests performed in laboratory in the scope of this study revealed the existence of response patterns common to all the ancient renderings tested: they show mechanical incompatibility with the rammed earth substrates because of their higher values of compressive strength and elasticity modulus; they also reveal great sus-ceptibility to water due to their high water absorption rates, lower drying processes, and high open porosity levels, causing higher water retention periods in contact with the earth substrate surface, promoting their degradation. Some commercial brand renderings - CBR 1 and CBR3 - show better characteristics then the ancient renderings regarding water ab-sorption; however, these same two commercial renderings have higher compressive strength than most of the ancient renderings and much higher than the rammed earth sub-strate; these two renderings were also the hardest to apply on the rammed earth sub-strate, because they dry too fast during application, making it difficult to apply manually in large areas of wall surface. Three months after application, there were no relevant signs of

7/8

superficial cracking in pre-mixed commercial renderings. In the scope of this campaign, none of the pre-mixed renderings show an ideal performance, regarding all the features tested. Therefore, other tests must be carried out in laboratory to test and improve some of those features, particularly regarding water drying rate, deformation capacity and chemical composition. In that scope, pre-mixed lime and clay-based mixes should be tested. The Karsten tubes test revealed to be adequate and reliable for the purpose of this study. However, the pendulum test did not allow a safe correlation between the rebound level and the compressive strength for the five commercial pre-mixed renderings tested.

5. ACKNOWLEDGEMENTS The authors acknowledge the support of LNEC - National Laboratory of Civil Engineering, especially the team from the Renderings and Insulations Group of the Buildings Depart-ment (Project PRESERVe - Preservation of the Cultural Heritage Coatings) and the team from the Metallic Materials Group of the Materials Department (Project PTDC/EPH-PAT/4684/2014 DB-HERITAGE - Database of building materials with historical and heritage interest). The authors also acknowledge the support of ICIST-CERIS (Instituto de Engenharia de Estruturas, Território e Construção), IST (Instituto Superior Técnico), and the support of FCT - Foundation for Science and Technology, Ministry of Education, for provid-ing the PhD scholarship, ref. SFRH/BD/102052/2014.

BIBLIOGRAPHY Binda, L. et al (2000). Investigation procedures for the diagnosis of historic masonries, Construction and Building Materials 14, p 199-233. Magalhães, A. et Veiga, M. R. (2009). Physical and mechanical characterization of ancient mortars. Applica-tion to the evaluation of the state of conservation, Materiales de Construcción 59 (295), p 61-77. Mateus L. et al (2014). Characterization of external renderings of rammed earth constructions in the Algarve, International Conference on Vernacular Heritage, Sustainability and Earthen Architecture, Valencia, Spain, p 245-249. RILEM Test No. I.1 - Porosity accessible to water, RILEM 25-PEM - Recommandations provisoires. Essais recommandés pour mesurer l’altération des pierres et évaluer l’efficacité des méthodes de traitement. Maté-riaux et Construction, 1980, Vol. 13, Nº 75. RILEM Test No. I.2 - Bulk densities and real densities, RILEM 25-PEM - Recommandations provisoires. Essais recommandés pour mesurer l’altération des 102 pierres et évaluer l’efficacité des méthodes de traitement. Ma-tériaux et Construction, 1980, Vol. 13, Nº 75. Silva, A. (2003). Caracterização de argamassas antigas, 3º Encore - Encontro sobre conservação e reabili-tação de edifícios, LNEC, Portugal. Valek, J. et Veiga, M. R. (2005). Characterization of mechanical properties of historic mortars, STREMAH 2005 - 9th International Conference on Structural Studies, Repairs and Maintenance of Heritage Architecture, 22 / 24 June 2005, Malta. Veiga, M. R. et al (2001). Methodologies for characterization and repair of mortars of ancient buildings, 3rd International Seminar Historical Constructions, Guimarães, Universidade do Minho, Portugal. Veiga, M. R. et al (2004). Capillarity tests on historic mortar samples extracted from site. Methodology and compared results, 13th International Masonry Conference, Amsterdam, Holland. Velosa, A. L. et Veiga, M. R. (2012). Argamassas do Património Histórico: conhecer para conservar e reabili-tar, CIREA 2012 - Conferência Internacional sobre Reabilitação de Estruturas Antigas de Alvenaria, Univer-sidade Nova de Lisboa, Portugal, p 17-27.

8/8

AUTHORS BIOGRAPHICAL NOTICE Luís Mateus: Civil Engineer (IST- Instituto Superior Técnico, Lisbon University). MSc thesis on characterization of external renderings of rammed earth construction in western Algarve (IST). PhD student at IST, thesis “Development of pre-mixed external renderings for rammed earth constructions”. Partner of Monumenta Ltd. Considerable expertise in rehabilitating architectural heritage and ancient buildings. Maria do Rosário Veiga: Civil Engineer (IST - Instituto Superior Técnico, Lisbon University), PhD (FEUP - Faculdade de Engenharia, Porto University), Senior Researcher at LNEC - National Laboratory of Civil Engineering, Coordinator of the Coatings Laboratory of LNEC. She is the author or co-author of many scientific papers and technical publica-tions. Jorge de Brito: Civil Engineer (IST - Instituto Superior Técnico, Lisbon University); MSc in Structural Engineering (IST) and PhD in Civil Engineering (IST). Full Professor at IST, in the area of Construction Technology, comprising earth construction. He has supervised many MSc and PhD students and has authored several hundred sci-entific papers.