Dh

TLa

b

c

a

ARRAA

KAFHV

1

uopauaTat

A2

0d

Journal of Chromatography A, 1216 (2009) 3819–3824

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

etermination of alkenes in cracking products by normal-phaseigh-performance liquid chromatography–diode array detection�

atjana Tomic a, Sandra Babic b,∗, Nada Uzorinac Nasipaka, Maja Fabulic Ruszkowskia,ivijana Skrobonjac, Marija Kastelan-Macanb

INA-Industrija nafte d.d., Corporate Processes Function, Research and Development Sector, Central Testing Laboratory, Lovinciceva bb, Zagreb, CroatiaDepartment of Analytical Chemistry, Faculty of Chemical Engineering and Technology, University of Zagreb, Marulicev trg 19, Zagreb, CroatiaINA-Industrija nafte d.d., Rijeka Oil Refinery, Urinj bb, Rijeka, Croatia

r t i c l e i n f o

rticle history:eceived 24 October 2008eceived in revised form 18 February 2009ccepted 20 February 2009vailable online 26 February 2009

eywords:lkeneCC liquid productsPLC–DADalidation

a b s t r a c t

Alkene content determinations in fluid catalytic cracking (FCC) liquid products were performed bymeans of normal-phase high-performance liquid chromatography (NP-HPLC) with diode array detection(UV/DAD). Separation of alkenes from aromatic hydrocarbons was performed on amino-modified silicagel column with n-heptane as mobile phase. The column has a little affinity to alkenes and saturatedhydrocarbons and a pronounced affinity to aromatic compounds. The problem of alkenes and saturatesco-elution on this column type was overcome with the detection system, UV/DAD, sensitive and selectiveto alkenes, while saturates are inactive in UV field. Total alkene content was determined as a sum of mono-and dialkene groups quantified by external standard method. Validation and verification of the developedmethod proved their applicability. The following criteria were used to validate the HPLC–DAD method:selectivity, linearity, precision, limits of detection and quantification. Alkene contents were quantifiedwith the external standard method of wide calibration range, so both low and high alkene contents can

be determined by the single calibration. Correlation coefficients were higher than 0.99. Precision was eval-uated as repeatability and intermediary precision with relative standard deviations less than 5%. Somestructural investigation of alkene groups was performed to confirm the assumption. Proposed methodwas compared with certified NMR method. Six commercial motor gasoline samples were analyzed bythese two methods. Obtained results indicate good agreement between alkene content determined byboth methods. The developed method was applied to the determination of alkene content in liquid FCC

nge f

products in the boiling ra

. Introduction

Common characteristic of crude oil and most of crude oil prod-cts is a complex hydrocarbon composition. Structure and presencef some specific hydrocarbon groups depend strongly on type ofroduct, its origin and type of petrochemical process in which theyre made from. Determination of alkene hydrocarbon in oil prod-cts is from special interest for oil refinery processes as they are

bsent in crude oil and occurs in secondary refinery processes.he determination of the alkene can help in clearly understandingnd optimizing of conversion process variables explaining reac-ion pathways, physico-chemical mechanisms and their kinetics.

� Presented at the 14th International Symposium on Separation Sciences “Newchievements in Chromatography”, Primosten, Croatia, 30 September–3 October008.∗ Corresponding author. Tel.: +385 1 45 97 208; fax: +385 1 45 97 250.

E-mail address: sandra.babic@fkit.hr (S. Babic).

021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2009.02.062

rom 70 ◦C to 190 ◦C.© 2009 Elsevier B.V. All rights reserved.

Then the application and ecological properties of final productwould be determined. Fluid catalytic cracking (FCC) is one of therefinery processes in which alkene occurs by conversion of theheavy oil fraction with high sulphur and nitrogen species contentin the lighter fractions on fluidized catalyst. Liquid FCC prod-ucts, gasoline and light cyclic oils, have been used as blendingcomponents for commercial motor gasoline and diesel fuels andalkene content defines their application and ecological character-istic of the commercial fuels. Alkenes significantly influence onthe application characteristics of commercial gasoline as they arehigh octane number component. Octane number is a measure foranti-detonation characteristic of gasoline and it has been definedas isooctane mass portion in mix with n-heptane. It is found thatFCC gasoline octane number significantly increases higher alkene

content of production as compared with other hydrocarbon groups[1].

Depending on physical and chemical properties of petroleumproduct in which alkene is determined, different analytical meth-ods have been used. Liquid FCC products consist of hydrocarbon

3 togr. A

cptaaswtnocphbm

dtfipctoiharata[d(Ifg(caat

amtafphhUaomtF1tlcosiwe

820 T. Tomic et al. / J. Chroma

ompound classes, and in addition to alkenes there are alsoresent saturated, aromatic and polar components. Complex struc-ure composition and wide boiling point ranges make qualitativend quantitative investigations very demanding. Use of differentnalytical techniques and detection systems is often necessary,o as to be able to determine a more detailed composition,hile developed methods may very frequently be applied only

o a particular FCC temperature fraction. Chromatographic tech-iques are widely used for group and detailed determinationf alkene hydrocarbon composition in petroleum products. Gashromatography is used for determination of detailed liquid FCCroduct composition [2,3,4]. As this is a very sensitive methodaving very powerful separation and sensitive detection systems,oth very low and higher alkene concentrations can be deter-ined.Adsorption liquid chromatography is used for alkene content

etermination in petroleum samples. The most common method ishe fluorescent indicator adsorption (FIA) method [5]. FIA is usedor alkane, alkene and aromatic compound content determinationn different petroleum products with boiling point under 315 ◦C (i.e.etrol, jet fuels, aviation fuels, diesel fuels, gas oils), and alkeneontent can be determined in the range of 0.3–55% (v/v). Limita-ions of this method lie in its frequently provided insufficient datan chemical composition, poor precision, and low reproducibil-ty of results. High-performance liquid chromatography (HPLC)as been widely used for analysis of crude oil, middle distillates,nd liquid oil products [6–8]. Use of this method includes sepa-ation and identification according to hydrocarbon composition,s well as a separation, identification, and sometimes quantifica-ion of target components in many petrochemical samples. HPLCpplication procedures have been described in numerous papers9,10], and the method of determining aromatic groups in mid-le petroleum fractions has been accepted as European StandardEN 12916) [11] based on the preceding standard IP 391/95 [12].t is normal-phase (NP) HPLC method where separation is per-ormed on polar columns (silica gel or NH2/ CN modified silicael), with non-polar mobile phase and detection by refractive indexRI) detector. As a result of column selectivity, aromatic hydro-arbons are separated from saturated and alkene hydrocarbonsnd form distinct peaks according to their ring number. Alkanesnd alkenes have the same retention behaviour on this columnype.

In this paper we present the HPLC method developed forlkene content determination in liquid FCC products in chro-atographic conditions equivalent to those for aromatic group

ype determination, but with a difference in detection system,s we used hyphenated HPLC system with UV/DAD in the fieldrom 190 to 400 nm. As mentioned above, polar column and non-olar mobile phases enable separation of alkene and saturatedydrocarbons from aromatic compounds. Alkene and saturatedydrocarbon co-eluting problem has been overcome by usingV/DAD detection, as saturated compounds cannot absorb radi-tion in UV field, while alkenes absorb it owing to the presencef double bond � electrons. Developed and validated HPLC–DADethod was compared with certified NMR method for the quanti-

ative analysis of alkene contents in commercial gasoline products.CC liquid products in the boiling range from 70 ◦C to 190 ◦C, in5 ◦C cuts, were analyzed. Alkene contents were quantified withhe external standard method of wide calibration range, so bothow and high alkene contents can be determined by the singlealibration. Some limited structural data for FCC fractions were

btained by comparing UV spectra of unknown sample with thetandard component spectra stored in spectral library. Compar-son is performed according to spectral (maximum absorptionavelength, �max) and chromatographic (retention time, tR) param-

ters.

1216 (2009) 3819–3824

2. Experimental

2.1. Chemicals and standards

As a solvent and mobile phase in the system, n-heptane (Kemika,Zagreb, Croatia) of chromatographic purity with <50 ppm watercontent has been used. Prior to initiating the analysis, helium 5.0,purity 99.999% (Messer Croatia Plin, Zapresic, Croatia) has gonethrough the mobile phase to remove dissolved gasses. Standards2,3-dimethyl-1 butene, purity ≥98% (GC) and 2,3-dimethyl-1,3-butadiene, monomer ≥97% (GC) obtained from Fluka (Buchs,Switzerland) were used as representatives of mono- and dialkenes,respectively.

2.2. Samples

FCC liquid samples analyzed in the work have been FCCpetrol distillation fraction, produced from the FCC feedstock ofhigh-sulphur oil origin. Origin characteristics are: sulphur con-tent S = 1.27%, average molecular weight: 408, ASTM D 1160:IBP = 211.3 ◦C. By n-d-M method (ASTM D 3238) carbon distributionin crude oil and oil fractions is calculated. From the experimentaldata (refraction index (n), density (d) and molecular weight (M)) byempirical calculations, structural-group type carbon in alkane (Cp),aromate (CA) and cycloalkane (CN) hydrocarbon forms is calculated.Results in form of mass fractions are: % CA = 13.65, % CN = 18.64% andCP = 67.91.

FCC raw material has catalytic crack on commercial octane cat-alyst (Grace-Davison), at 525 ◦C on MAT (micro-activity test) plant.The following products have been produced: FCC petrol tv < 190 ◦C,LCO (light cyclic oil) tv = 190–343 ◦C and HCO (high cyclic oil)tv > 343 ◦C. In the paper we have focused on the FCC petrol anal-ysis, fractions from 70 ◦C to 190 ◦C. Lower FCC petrol fractions havenot been analyzed, because of volatility and weight losses whichmake quantification unreliable. LCO and HCO have not been inves-tigated, because they are usually further processed on other plants,while light FCC petrol has been directly blended in motor gasoline.

Before the chromatographic analysis samples have been dilutedin n-heptane (0.4 g/5 mL) and filtrated through 0.45 �m membranefilter. A volume of 10 �L of prepared solution has been injectedusing a partial filling of the sample loop.

2.3. HPLC–DAD procedure

Liquid chromatography system used for analyses is a complexmodular system consisting of the following units: 9012Q HPLCpump (Varian Chromatography System, Walnut Creek, CA, USA),injection system (ProStar 410 Autosampler, Varian), column oven(Mistral column oven, Varian) and UV/DAD detector (PolychromDiode Array Detector).

Alkenes were separated from aromatic compounds on the Zor-bax polar amino bonded silica column, 250 mm × 4.6 mm I.D.,particle size 5 �m (Agilent, Santa Clara, CA, USA) and n-heptane wasused as mobile phase with 0.8 mL/min flow rate. The column tem-perature was set at 30 ◦C and injection volume was 10 �L. Sampleswere analyzed in 190–367 nm UV field.

UV/DAD detector works with a deuterium lamp in 190–367 nmwavelength range, with precision of ±1 nm. As a difference fromusual UV detectors which collect data in time and absorbancedimensions at one wavelength, DAD collects data in whole wave-length range and makes possible the third dimension measurement

(wavelength). Whole chromatographic system is controlled by theVarian programme STAR, version 5. Data collecting and comput-ing are also integrated in the STAR programme. Special application,PolyView Spectral Processing, enable a post-run analysis data col-lected from Polychrom Diode Array Detector. This application can

togr. A 1216 (2009) 3819–3824 3821

eattq

scs

3

3

ttamaphadfimm

wTbmwab

d

3

iid(

sm2

Fig. 2. Spectroscopic curve of 2,3-dimethyl-1,3-butadiene.

Table 1Operating and instrumental conditions for alkene determination.

Monoalkenes Dialkenes

Calibration standard 2,3-Dimethyl-1-butene 2,3-Dimethyl-1,3-butadiene� (nm) 210 244F (mL/min) 0.8 0.8T (◦C) 30 30Mobile phase n-Heptane, HPLC grade n-Heptane, HPLC gradeMobile phase preparation Helium purging Helium purgingStationary phase Amino-modified silica Amino-modified silica gel

T. Tomic et al. / J. Chroma

xtract spectra at any time point from the chromatogram or gener-ted chromatogram absorbance in any UV field wavelength. Underhis application, spectral library can be created, and by comparinghe spectra of unknown component with the known spectra, someualitative and structural data can be achieved.

In the STAR programme PolyView application, there is the alkenetandard spectral library, and alkene peaks from samples have beenompared with the known components peaks and certain limitedtructural information have been obtained.

. Results and discussion

.1. HPLC–DAD analysis

This paper aims at developing the method for determination ofotal alkene content in FCC liquid products in the way that separa-ion system matches the one from EN 12916 for determination ofromatic hydrocarbons in middle petroleum distillates. The amino-odified silica gel shows a minor affinity to saturated hydrocarbons

nd olefins, while a significantly large affinity to aromatic com-ounds. Co-elution problem of saturated and alkene hydrocarbonsas been solved by use of UV/DAD. Saturated hydrocarbons are notctive in UV field, and therefore they are not detected, while alkenes,ue to the presence of double bond � electrons get absorbed in UVeld. Radiation of minor wavelengths is necessary for exciting theore tightly bonded electrons, while the more weakly bonded onesay get excited by larger wavelength radiation.Prior to performing analyses, it is necessary to define the

avelengths on which olefin hydrocarbons will be monitored.he UV spectra of 2,3-dimethyl-1-butene and 2,3-dimethyl-1,3-utadiene and their maximum UV absorption were determined. Foronoalkenes the maximum has been found close to 210 nm (Fig. 1),hile for dialkenes almost 244 nm (Fig. 2), which, according to the

bove mentioned relation between wavelengths and tightness ofonds, has been expected.

Experimental conditions for determination of mono- andialkenes are shown in Table 1.

.2. Method validation

Validation and verification of the method have been performedn order to confirm its validity and applicability. Prior to perform-ng validation of the total alkene determination method, there areefined validation parameters and criteria of their acceptability

Table 2).

Selectivity was first tested on a mixture of mono- and dialkenetandards and mono-aromatic standard. Monoalkenes have beenonitored at 210 nm and retention time of 4.2 min; dialkene at

44 nm and 4.6 min, and mono-aromatic standard at 254 nm and

Fig. 1. Spectroscopic curve of 2,3-dimethyl-1-butene.

gelVinjection 10 �L 10 �LtR (min) 4.2 4.6

5.4 min. By testing the selectivity on an actual FCC product, thesedata were confirmed, thus proving that the matrix has no influenceon determination and that the method is selective.

Linearity, limit of detection (LOD), limit of quantification (LOQ),precision and recovery were evaluated for quantitative purpose.

The linearity was evaluated for both alkene groups with 2,3-dimethyl-1-butene and 2,3-dimethyl-1,3-butadiene as calibrationstandards. Calibration curves for groups were made by means ofthe external standard method. For monoalkenes, calibration is per-formed with six concentrations in a wider calibration range from0.1 g/100 mL to 10 g/100 mL. Calibration for dialkenes is performedwith five concentrations in range from 0.1 g/100 mL to 1 g/100 mL.Blanks were also prepared and included in regression analysis. Lin-ear calibration curves are obtained between the peak areas andstandard concentrations in the entire analyzed concentration range.Coefficients of correlation were higher than 0.99 thus confirmingthe linearity of the method (Table 3). By means of calibration curves,individual alkene groups are quantified as mass fractions (%, w/w).Total alkene content, as mass fraction, is calculated by summation

of mono- and dialkene contents.

Verification is procedure of confirming the accuracy of calibra-tion by analyzing the calibration standard of exact concentration asunknown sample. The resulting value of area is converted into con-

Table 2Validation parameters, acceptability criterion and validation results.

Validation parameter Acceptability criterion Validation results

Selectivity Information Acceptable

Linearity R2 ≥ 0.99 R2 = 0.994R2 = 0.998

Repeatability RSD ≤ 5% RSD = 2.64%Intermediary precision RSD ≤ 5% RSD = 4.83%LOD Information 0.02% (w/w)LOQ Information 0.06% (w/w)

3822 T. Tomic et al. / J. Chromatogr. A 1216 (2009) 3819–3824

Table 3Calibration equations and results of method verification for alkene groups.

Monoalkenes Dialkenes

CalibrationCalibration curve equation Y = 2.62 × 105x + 1.56 × 105 Y = 1.13 × 106x − 1.37 × 106

Correlation coefficient (r) 0.994 0.998

VerificationExpected value 0.2207 0.2019

Y

cepa

bmutmtsoi

oftqt

st

mflmaatdit

3

dpic

TI

G

123456

Calculated value 0.2183Deviation (%) 1.1

: peak area (pixel); x: sample mass concentration (g/mL).

entration via calibration parameters and gets compared with thexpected (nominal) value, while result is expressed as a deviationercentage of the expected value. Results of verification procedurere shown in Table 3.

Precision of the method is checked by determining the repeata-ility and intermediary precision. Repeatability is calculated byeasuring the alkene content in a corresponding sample six times

nder identical conditions and performed by the same operator inhe laboratory. Since the relative standard deviation is below 5%, the

ethod is considered repeatable. Intermediary precision expresseshe precision within laboratory variations. In the correspondingample, different operators have tested the alkene content six timesn different days. As it is shown in Table 2 the RSD was 4.83%, whichs acceptable for intermediary precision.

LOD and LOQ were experimentally estimated from the injectionf standard solutions serially diluted until the signal-to-noise ratioor analyte reached a value of ten for LOQ and three for LOD. Underhe given conditions, noise of base line has been measured. Theuantification limit of 0.06% and detection limit of 0.02% for theotal alkenes were obtained.

The results obtained by validation and verification procedureshow that the method is applicable for alkene analyses in the men-ioned concentration range.

Based on presented results the precision of developed HPLCethod is considerably improved compared to the precision of the

uorescent indicator adsorption method, which is used most com-only for class composition determination of gasoline. LOD, LOQ

nd precision is similar to precision obtained by HPLC with DADnd RID [10]. However, method presented enables single calibra-ion in wide boiling range and separate quantification of mono- andialkene. Information about structure of alkene present in sample

s very important because of determined application properties ofhe commercial fuels (e.g. octane number).

.3. Method comparison

Developed and validated HPLC–DAD method was applied toetermination of alkene contents in commercial motor gasolineroducts. Samples were taken from different gas stations originat-

ng from different refinery processes and processed from differentrude oils. The obtained results were compared with results

able 4nterlaboratory comparison results obtained by HPLC and NMR method.

asoline sample number HPLC

Monoalkenes, % � Dialkenes, % �

8.85 1.908.96 1.978.83 1.928.90 1.698.27 1.588.41 1.12

0.19891.5

obtained by NMR method [13] certified (ISO/IEC 17025) for thehydrocarbon contents in motor gasoline. Six commercial motorgasoline samples were analyzed by both methods and obtainedresults are given in Table 4.

The results of determination of alkenes content by the developedHPLC–DAD method are in good agreement with those obtained byNMR method. However, NMR could be applied only for quantitativealkene determination in light distillates.

3.4. Analysis of liquid FCC products

Liquid FCC product is fractionated into the nine temperaturefractions starting from 70 ◦C, all the way to the residue with boilingpoint above 190 ◦C. Fractions below 70 ◦C have not been analyzed,because a mass loss caused by the product volatility has been found.Light and middle fractions get blended into motor gasoline, whilehigher fractions are blended into diesel fuels. This method has beendeveloped in order to enable, with a single analysis, determinationof not only the content and distribution of aromatic hydrocarbons,but also the content of alkene hydrocarbons. In such a way, totaltime required for analyzing samples gets shortened, while con-sumption of power and chemicals is reduced as well. Fig. 3 showsthe FCC fraction (130–145 ◦C) chromatogram on the wavelengthsdefined for monitoring the particular alkene groups.

Three-dimensional presentation of FCC product chromatogramprovides a more complete picture of alkene hydrocarbons in thesample, because monitoring is performed simultaneously, accord-ing to the absorption intensity, wavelength, and retention time.Fig. 4 shows three-dimensional chromatogram of FCC fraction(130–145 ◦C). Two peaks are identified: the one that absorbs UVradiation from 190 to 220 nm corresponds to monoalkene group(1), while that with absorption from 220 to 260 nm correspondsto dialkene group (2). Considering that dialkenes have a strongerabsorption in UV field, which is visible from Figs. 1 and 2, dialkenefraction content in these samples is expected to be smaller, becausetheir absorption intensity on the chromatogram is considerably

weaker.

Although retention times of alkene groups are similar to eachother, resolution is satisfactory due to absorption on different wave-lengths and to the fact that dialkene content in the products ofcracking is considerably smaller.

NMR alkenes, % � � %

Total alkenes, % �

10.75 11.37 5.510.93 11.27 3.010.75 10.68 0.710.59 11.54 8.29.85 11.26 12.59.53 9.64 1.1

T. Tomic et al. / J. Chromatogr. A 1216 (2009) 3819–3824 3823

Fig. 3. HPLC–DAD chromatogram of FCC fraction (130–145 ◦C) on both 210 and 244 nm, each wavelength is corresponding for the other alkene group.

30–14

wi

ig

Fct

TA

L

FFFFFFFFF

Fig. 4. Three-dimensional chromatogram of FCC fraction (1

Alkene contents in liquid FCC products determined by HPLCith diode array detector expressed as mass fraction are shown

n Table 5.The obtained results confirm the literature information that by

ncreasing the fraction boiling point temperature, alkene contentets reduced.

After quantification of alkene hydrocarbon contents in liquidCC fractions, UV spectra of particular fractions were analyzed andompared to the spectra of standard components stored in the spec-ra library, in order to evaluate the data on dominant structures of

able 5lkene contents in liquid FCC products.

iquid FCC product fractions Fraction boiling point temperature, ◦C

r. 1 70–85r. 2 85–100r. 3 100–115r. 4 115–130r. 5 130–145r. 6 145–160r. 7 160–175r. 8 175–190r. 9 >190

5 ◦C); peak 1: monoalkene group; peak 2: dialkene group.

alkenes present in particular fractions. Similarity parameter, cal-culated by PolyView Spectral Processing software, larger than 0.98was used as a criterion. Similarity is the parameter for mathemati-cal comparison of two data sets. For very similar spectra, similarityparameter is close to 1. By this, assumption could be confirmedthat certain retention times and wavelengths correspond to the

appurtenant mono- and dialkenes structures. It was found that infractions of lower boiling point in the monoalkene retention time,majority of the present structures are those with shorter hydrocar-bon chains. As the fraction boiling point temperature rises, longer

Monoalkenes, % Dialkenes, % Total alkenes, %

53.56 6.03 59.5933.15 5.68 38.8331.77 7.42 39.1930.32 6.97 37.2919.58 5.34 24.9214.03 4.31 18.349.31 2.91 12.226.51 1.30 7.814.09 0.50 4.59

3 togr. A

cttdeb

4

mFcwefiesabtqngTpttd

[[

824 T. Tomic et al. / J. Chroma

hains become dominant. In the dialkene section, structure withwo double bonds has been confirmed. Analyses show that duringhe FCC process there appears unsaturated structures of variousegrees of unsaturation. The number of alkene structures with sev-ral double bonds is considerably smaller, which may be explainedy their higher reactivity, but also by a lower appearance degree.

. Conclusion

The paper has presented development of the method for deter-ination of total olefin hydrocarbon content in liquid products of

CC process by HPLC–UV/DAD. Separation of olefins from aromaticompounds is performed in the conditions defined by the EN 12916,hile separation of olefins and saturated hydrocarbons is not nec-

ssary, because saturated hydrocarbons are not active in the UVeld. Samples contain mono- and dialkene groups with consid-rably different retention times and absorption maximum pointso they are quantified separately by the external standard methodt 210 and 244 nm, respectively. Total alkene content is obtainedy summation of mass shares of both groups. The method valida-ion and verification have confirmed applicability of the method atuantification limit of 0.06%. A comparison of quantitative determi-ation of alkene contents in commercial gasoline products indicatesood agreement between results obtained by HPLC–DAD and NMR.

he method has been successfully applied in analysis of liquid FCCroducts in the boiling point range of 70–190 ◦C. The method selec-ivity is satisfactory, because, regardless of the vicinity of retentionimes (4.2 and 4.6 min), wavelengths of maximum absorptions areifferent, so quantification is possible without interference. In this

[

[

1216 (2009) 3819–3824

boiling point area, satisfactory quantification has been found bymeans of a single calibration curve for each alkene group. Presenceof somewhat longer chains was found in higher FCC fractions, whilealkene chains are shorter in the fractions of lower temperature.

Acknowledgements

Vlasta Srica and Jelena Parlov Vukovic (INA–Industrija nafte d.d.,Central Testing Laboratory) are gratefully acknowledged for theirassistance in interlaboratory testing with NMR.

References

[1] E.T. S Habib, Preprints Symposia, The Hydrocarbon Chemistry of FCC NaphtaFormation, vol. 34, Miami, American Chemical Society, 10–15 September 1989,1989, p. 674.

[2] L. Gustavo, S. Nora, G.L. Daniel, J. Chromatogr. 369 (1986) 419.[3] K.N. Patero, E. Wandzik, Chem. Anal. 15 (1970) 1977.[4] D.L. Schmitt, H.B. Jonassen, Anal. Chim. Acta 49 (1970) 580.[5] Method D 1319, Annual Book of ASTM Standards, American Society for Testing

and Materials, Philadelphia, PA, 2003.[6] J.C. Suatoni, H.R. Garber, J. Chromatogr. Sci. 14 (1976) 546.[7] J.C. Suatoni, R.E. Swab, J. Chromatogr. Sci. 14 (1976) 535.[8] T.V. Alfredson, J. Chromatogr. 218 (1981) 715.[9] S.L.S. Sarowha, K.S. Brajendra, S.D. Bhagat, Fuel 75 (1996) 1743.10] M. Kaminski, R. Kartanowicz, A. Przyjazny, J. Chromatogr. A 1029 (2004) 77.11] European Standard EN 12916, Petroleum Products – Determination of Aromatic

Hydrocarbon Types in Middle Distillates – High Performance Liquid Chromatog-raphy Method with Refractive Indeks Detection, CEN, Brussels, 2000.

12] IP 391, Standard Methods for Analyzing and Testing of Petroleum Related Prod-ucts, Aromatic Hydrocarbon Types in Diesel Fuels and Distillates, The Instituteof Petroleum, London, UK, 1997.

13] J. Burri, R. Crockett, R. Hany, D. Rentsch, Fuel 83 (2004) 187.