A

ppmtwwmedrri©

K

1

at[fdo

Vf

0d

Journal of Chromatography A, 1157 (2007) 454–461

Analysis of diterpenoic compounds in natural resins applied as bindersin museum objects by capillary electrophoresis

Anna Findeisen a, Viliam Kolivoska a, Isabella Kaml a,b,Wolfgang Baatz b, Ernst Kenndler a,∗

a Institute for Analytical Chemistry, University of Vienna, Austriab Academy of Fine Arts, Vienna, Austria

Received 15 March 2007; received in revised form 30 April 2007; accepted 2 May 2007Available online 6 May 2007

bstract

The exudates of conifers consist mainly of diterpenoic acids of the abietane and pimarane type (abietic, neoabietic, dehydroabietic, palustric,imaric, isopimaric, levopimaric and sandaracopimaric acid) and larixol acetate. These natural resins were used as adhesives, coatings, varnishes orlasticizers in artistic and historic works since ancient times. For the purpose of conservation and restoration and for art historic examination of suchuseum objects the identification of the binding media used is undoubtedly of paramount importance. In the present paper, the characterization of

hese resins based on the pattern of their diterpenoid constituents is carried out by capillary electrophoresis. For separation a background electrolytehich has been initially introduced for the analysis of chlorinated and natural resin acids in waste water was modified and the experimental conditionsere adjusted in terms of resolution and analysis time. Separation was carried out in borate buffer at pH 9.25 (ionic strength 20 mmol L−1) withethyl-�-cyclodextrin and sulfobutylether-�-cyclodextrin as additives to increase selectivity and enhance the solubility of the analytes. With this

lectrophoretic system the resin acids of interest and larixol acetate – all as anionic cyclodextrin complexes – were separated within 5 min andetected at 200, 250 and 270 nm with a diode array detector. The electrophoretic patterns served for the characterisation of the relevant diterpenoic

esins, balsams and copals. Sample pre-treatment was limited to sonication in methanol at 55 ◦C for 30 min. This enables the identification of theesins in mixtures with other binders like plant gums, animal glues or drying oils, even when these media are present in excess. Colophony wasdentified as resinous constituent of a modelling mass for gilded frames originating from the 19th century.

2007 Elsevier B.V. All rights reserved.

ophor

nwatih

eywords: Museum objects; Natural binding media; Analysis; Capillary electr

. Introduction

Natural resins have been used since ancient times in manyreas for medical, pharmaceutical or cosmetic purposes. In artis-ic works they have been used as glues, varnishes or plasticizers1,2]. The resins (except shellac) are all products of certain treesrom which they exude, the higher molecular constituents being

issolved in the mono- and sesquiterpenoid essential oils. Theils evaporate and the resinous material remains and is collected.

∗ Corresponding author at: Institute for Analytical Chemistry, University ofienna, Wahringerstr. 38, Vienna, Austria. Tel.: +43 1427752305;

ax: +43 142779523.E-mail address: ernst.kenndler@univie.ac.at (E. Kenndler).

macAattho

021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2007.05.010

esis; Diterpenoid; Resins; Larixol acetate

From the chemical point of view we can differentiate theatural resins according to the number of carbon atoms fromhich their molecules are built up: diterpenoids possess 20 C-

toms, triterpenoids possess 30 C-atoms in their skeleton. Notehat the members of these two groups do not occur togethern a certain plant. Diterpenoid resins, the group we considerere, are produced mainly from Coniferae (conifers) and Legu-inosae (pea family); tropical Leguminosae form the copals

nd copaibas. The Coniferae, growing in the temperate regions,an be divided into the subfamilies Pinaceae, Cupressaceae andraucariaceae. They produce mainly compounds with pimarane

nd abietane skeleton, both consisting of a phenanthrene-like

hree-ring system. The major resin constituents are acids withhe COOH group at position 4 in the ring system, and alco-ols. Pinaceae consist of Pinus species (the source of rosinr colophony), Picea species (from which Burgundy pitch is

A. Findeisen et al. / J. Chromatogr. A 1157 (2007) 454–461 455

Table 1Structures of the main diterpenoic acids in natural resins and of larixol acetate

oaa

ttbpdctpOamc

d[laoot

ttmgwaacp[wGtib

(wo[

btained), Abies species (the source of Strasbourg turpentinend Canada balsam), and Larix species (from L. decidue larchnd Venice turpentine originates) [1,2].

The main diterpenoid constituents of the Pinaceae resins andheir structures are listed in Table 1. Pine resins consist mostly ofhe resin acids shown there, Strasbourg turpentine and Canadaalsam contain in addition a C-20 alcohol, abienol. Venice tur-entine is characterised by its content of larixol acetate, whichoes not occur in any other resin. It can thus be taken as an indi-ator for this resin. It is important to mention in this context thathe quantitative composition of the resins varies due to chemicalrocesses during the production and the storage of the material.xidation e.g. leads in many cases to a conversion of abietic

nd neoabietic acid into dehydroabietic acid, which is finally aain constituent of particular resins independent of their initial

omposition.Several methods have been applied for the analysis of the

ifferent resins (for a recent review on analysis of binders see3]). Early attempts used gas chromatography (GC) [1], thinayer chromatography [4,5], differential thermal analysis [6]

nd pyrolysis mass spectrometry (MS) [7]. Spectrometric meth-ds like FTIR [8,9] and MS supported by chemometry [10,11],r FT-Raman [12] and micro-Raman [13] were used. Directemperature-resolved MS [14] and pyrolysis GC in combina-

lbtm

ion with MS was applied to analyse the oxidation products ofhe resin components [15]. GC/MS is certainly the most com-

on technique nowadays for the purpose of binder analysis ineneral. We have characterised resinous additives in anatomicax models of the 18th century using GC hyphenated with MS

nd FTIR [16]. Many recent investigations use GC/MS for thenalysis of diterpenoic and triterpenoic [17] [9,11,18–24] resinomponents. This method served to investigate their oxidationroducts in presence of pigments [9]. The group in Valencia20–22] and others [18,25] used controlled pyrolysis (combinedith in-situ derivatisation) for inserting the sample into the GC.C/MS was applied in a series of papers from Colombini’s group

o analyse different binders [11,17,26–35], finally culminatingn a sophisticated procedure that allowed the identification of allinders in a single sample [23].

In the last years the potential of capillary electrophoresisCE) for the analysis of the binding media of museum objectsas explored in our laboratory. This technique was devel-ped and successfully applied for the analysis of drying oils36,37], animal glues [38,39] and plant gums [40,41]. Fol-

owing this objective we were interested to close the gap ofinder characterisation by CE by insertion of resinous mat-er. Starting point for the development of an appropriate CE

ethod was the outstanding work of Luong et al. [42], describ-

4 matog

ifuc(tlpia

2

2

mRiGwcKbT

2

cG4Lfb1t

(MtStBNnoa11Sip

2

s

awmtipCfir

3

rdws(amoaat

3

t2atueeiBorflpcpatfwutmrsi

56 A. Findeisen et al. / J. Chro

ng the separation of 11 resin acids including 3 chlorinated onesor the analysis of waste water from pulp mills. The authorssed methyl-�-cyclodextrin (MECD) and sulfobutylether-�-yclodextrin (SBCD) as modifiers of the background electrolyteBGE). However, the objective of this work was different fromhat relevant for binders of museum objects, concerning the ana-ytes and the sample matrix as well. It was thus the goal of theresent investigation to develop a CE method for the character-sation and identification of the natural diterpenoic resins useds binders in artistic and historic works.

. Experimental

.1. Materials

Resin acids were obtained from Helix Biotech (Rich-ond, British Columbia, Canada), larixol acetate from Sequoiaesearch (Pangbourne, UK). Boric acid and sodium hydrox-

de for the BGE and methanol was from E.Merck (Darmstadt,ermany; all analytical grade). Methyl-�-cyclodextrin (MECD)as from Fluka (Steinheim, Germany), sulfobutylether-�-

yclodextrin (Advasept 7, SBCD) was from CyDex (Lenexa,S, USA). Reference material was taken from the collection ofinding media at the Academy of Fine Arts, Vienna, Austria.he BGE was prepared with bidistilled water.

.2. Instrumentation

Electrophoretic measurements were carried out with an air-ooled CE instrument (HP3D, Agilent Technologies, Waldbronn,ermany) using uncoated fused-silica capillaries (total length8.5 cm, length to detector 40.0 cm, ID 50 �m I.D.; Microsolv,ong Branch, NJ, USA). UV-absorbance was measured at dif-

erent wavelengths (at 200, 250 and 270 nm, bandwidth 10 nm)y the aid of a diode array detector (DAD). Injection was at00 mbar s, separation voltage was 25 kV. The temperature ofhe capillary cassettes was set to 20 ◦C.

Separation buffer was boric acid/borate adjusted to pH 9.25ionic strength 20 mmol L−1) with NaOH. Solid SBCD and

ECD were dissolved in the BGE at different concentra-ion (increasing the ionic strength accordingly), all with anBCD/MECD ratio of 3/2. The solutions were sonicated and fil-

ered through 0.22 �m Corning® Spin-X® centrifuge tube filters.efore use, new capillaries were conditioned with 1 mol L−1

aOH for about 30 min and then washed with water and the run-ing buffer, respectively, for another 30 min each. Before dailyperation, the capillary was rinsed with 0.1 mol L−1 NaOH forbout 10 min, and then with water and the BGE for another0 min each. The capillary was further conditioned by applying5 kV voltage for approximately 10 min before the first injection.tock solutions of the standards (10 mmol L−1) were prepared

n methanol, and diluted to 1 mmol L−1with the respective BGErior to injection.

.3. Sample preparation

All samples were treated with methanol at 55 ◦C in an ultra-onic bath for 30 min. Samples were resin acids and larixol

e

if

r. A 1157 (2007) 454–461

cetate standards, the different natural resins and their mixturesith either linseed oil, gum arabic or bone glue, and the agedaterial taken from a modelling mass for gilded frames from

he 19th century. Typically 1 mg solid sample was suspendedn 500 �L methanol, after heating and ultrasonication the sus-ensions where then centrifuged and filtered through 0.22 �morning® Spin-X® centrifuge tube filters. The solutions werenally diluted to the according concentration (in the millimolarange) with the BGE prior to injection.

. Results and discussion

The BGE initially used for the present analysis of the coniferesins by CE was that developed by Luong et al. [42] for theetermination of natural and chlorinated resin acids in wasteater of pulp mills (see Introduction). However, there are two

ignificant differences to the objective of the present research:i) larixol acetate is an additional potential sample constituent;nd (ii) the matrix is a solid material, it is a natural resin or/and aixture with other binding media like animal glues, plant gums

r drying oils. Therefore, the CE conditions were accordinglydjusted by varying the parameters pH, additive concentrationnd separation voltage, with the goal to enable full resolution ofhe analytes in an appropriate analysis time.

.1. Separation of standard compounds

Luong et al. [42] used several BGEs for the separation ofheir analytes. One BGE was at intermediate pH (4.5) with0 mmol L−1 MECD and 30 mmol L−1 SBCD, and 20% (v/v)cetonitrile as buffer modifiers. This BGE was explored first inhe present research, because the resin acids are not dissociatednder these conditions, which makes their electrophoretic prop-rties more similar to larixol acetate: the latter compound is anster with only a very weak acidic alcoholic OH as potentialonisable functional group (see Table 1). It was found that thisGE was not suited for our purposes; it led to low repeatabilityf the runs often accompanied by the breakdown of the cur-ent, blocked capillaries, and strongly varying electroosmoticow (EOF). We have thus modified the BGE (boric acid/borate,H 9.25, ionic strength 20 mmol L−1) by varying the additiveoncentration, but applying both cyclodextrins in the same pro-ortions (3:2 molar ratio of SBCD to MECD). Note that uponddition of SBCD the ionic strength is increased. The elec-rophoretic mobilities (of the anionic cyclodextrin complexes) asunction of the SBCD concentration are given in Fig. 1, togetherith that of the EOF (the EOF is directed towards the cathode assual). In order to better illustrate the effect on separation selec-ivity, the total mobilities are plotted rather than the effectiveobilities. Because the mobilities are very similar, the loga-

ithmic scale is used to better visualise their differences. In theame figure, the migration time of the last eluting compounds also depicted, which directly reflects the analysis time of the

lectrophoretic runs.

It can be seen that the total mobilities decrease with increas-ng additive concentration, the curves for the analytes – exceptor LAXA – almost running in parallel. The mobility of LAXA

A. Findeisen et al. / J. Chromatogr. A 1157 (2007) 454–461 457

Fig. 1. Total mobilities,μtot, of the analyte–cyclodextrin complexes and the EOF(left ordinate) and analysis time (right ordinate) as function of the concentrationof sulfobutylated cyclodextrin (SBCD) in the BGE (borate buffer, pH 9.25). Co-additive was methyl-�-cyclodextrin at a concentration 2/3 of that of SBCD. Asmt

(dlcaaoacastnaSsi

aaaattsm

mdtciTti

Fig. 2. Electropherogram of a standard mixture of the natural resin acids andlarixol acetate. BGE: borate buffer, pH 9.25, with 10 mmol L−1 SBCD and6.6 mmol L−1 MECD as buffer modifiers. Symbols as in Table 1. Analyte con-centration: 1 mmol L−1 each. CE conditions: uncoated fused-silica capillary,tds

awaarhaLc

arvdfdrhgDntaatc

3

gundy pitch, Canada balsam, Congo and Manila copal, and

easure for the analysis time the migration time of the last eluting analyte wasaken. For symbols see Table 1.

for abbreviations of the analytes see Table 1) varies in aifferent manner: it exhibits the smallest mobility from all ana-ytes (and thus the longest migration time) at low cyclodextrinoncentration, but changes the elution order at concentrationsbove 20 mmol L−1 SBCD, finally migrating between DABAnd PAA. From these plots it can be concluded that in termsf selectivity and analysis time low cyclodextrin concentrationsre preferable, conditions which are favourable also for a lowonsumption of the (costly) additives. At low concentrations andditional advantage is the applicability of higher electric fieldtrength due to the lower current (SBCD is a highly charged addi-ive). Too low cyclodextrin concentrations, on the other hand, areot so supportive for the solubilisation of the sparingly solublenalytes. Therefore, as a compromise, a BGE with 10 mmol L−1

BCD and 6.6 mmol L−1 MECD was selected for further analy-is. This concentration allows the application of 25 kV resultingn an acceptable electric current of 65 �A.

The resulting electropherogram obtained from a mixture ofbietic acid, dehydroabietic acid, isopimaric acid, levopimariccid, neoabietic acid, palustric acid, pimaric acid, and larixolcetate is shown in Fig. 2. It can be seen that indeed all analytesre baseline separated. The combination of the high mobility andhe high field strength enabled the reduction of the analysis timeo less than 5 min. It should be mentioned that according to [42]andaracopimaric acid, which was not included in the sampleixture (it was not available) co-elutes with pimaric acid.It is obvious that the identification of the compounds by their

igration time or mobility is significantly supported by spectralata, given that there is a sufficient difference between the par-icular compounds. Although the analyte molecules lack stronghromophores (see Table 1), discrimination is at least possiblen part, as can be seen from the UV-vis spectra shown in Fig. 3.

hey were measured by the aid of the diode array detector from

he analyte peaks (better said from their cyclodextrin complexes)n the electropherogram shown in Fig. 2. It can be seen that ABA

Ctt

otal length 48.5 cm, length to detector 40.0 cm, ID 50 �m. UV-absorbanceetected at 200, 250 and 270 nm by the aid of a DAD. Injection at 100 mbar s,eparation voltage +25 kV. Temperature of the capillary cassettes set to 20 ◦C.

nd NABA – they possess two conjugated C C double bondshereas not in the same ring – have a clear maximum at 240

nd 250 nm, respectively, differently to all other analytes. LPIAnd PAA – with two conjugated C C double bonds in the sameing – have maxima at 270 nm. DABA with the aromatic ringas strong absorbance with a local maximum at 220 nm, PIAnd IPIA absorb considerably at 200 to 210 nm. In contrast,AXA has low absorbance over the entire spectral range underonsideration, similar to the BGE, which is also shown in Fig. 3.

These different spectral properties – on the one hand a dis-dvantage for the easy quantitative analysis as the individualesponse factors of the analytes differ – are, on the other hand,aluable for additional identification purposes based on theetector record at different wavelength. Indeed it can be seenrom Fig. 2 that ABA, which has a low response at 200 nm,elivers a nearly four times higher peak at 250 nm and does notespond at 270 nm. The peak of NABA at 250 nm is five timesigher than at 200 nm, and is also detectable at 270 nm. PAAives peaks with equal height at all three recorded wavelengths.ABA, with the highest response of all analytes at 200 nm, giveso peak at 250 and 270 nm, and can therefore clearly be differen-iated from the closely eluting ABA peak. LPIA responds bettert 200 and 270 nm than at 250 nm, in contrast to PIA and IPIA,nd LAXA is detectable only at 200 nm. However, its migra-ion time is far from the next eluting compound under the givenonditions, so that a misinterpretation can be avoided.

.2. CE of natural resins

Colophony, Venetian, larch and Strasbourg turpentine, Bur-

opaiba balsam samples were subjected to CE; here we showhe results from a few typical examples. It was already men-ioned that the natural variation of the composition of the resins

458 A. Findeisen et al. / J. Chromatogr. A 1157 (2007) 454–461

easu

ct

uNAiwjet

TlbbLoaogec

pcNe

Icotiatjo

3

yncatafcu

c

Fig. 3. UV-vis spectra of the diterpenoic compounds (and the BGE) m

an be enormous, dependent on the source of the material, orhe transport and storage conditions.

A typical electropherogram of colophony is shown in Fig. 4,pper panel. In the sequence of their migration, the peaks ofABA, PIA, IPIA, PAA, DABA (the main peak at 200 nm) andBA (the main peak at 250 nm) can be differentiated. It is not our

ntention to quantify these compounds in the resin by calibrationith the individual response factors due to the reasons mentioned

ust before. However, the pattern of the resin components in thelectropherogram agrees with the description of colophony inhe literature: all the detected acids are constituents of this resin.

A second example is larch turpentine (Fig. 4, middle panel).he peak of DABA is much smaller than in colophony, and a

arge peak is recorded at the position of IPIA, its identificationeing is supported by the spectra obtained by the DAD (andy standard addition). Most indicative is the large peak fromAXA, a finding which is in agreement with the compositionf this resin given in the literature. LAXA can indeed be useds the marker for larch and Venice turpentine. It was pointedut above that this compound possesses an alcohol and an esterroup, but not a carboxylate. Its electrophoretic migration isxclusively the result of the formation of a complex with theharged cyclodextrin added to the BGE.

A third example is Strasbourg turpentine (Fig. 4, bottom

anel). According to the literature [1,2], Strasbourg turpentineontains LPIA and PAA, together at around 10%, ABA andABA at about 20% each, and as major compound (47%) abi-nol, a resin alcohol with two conjugated C C double bonds.

isbi

red with the DAD from the peaks in Fig. 2. For symbols see Table 1.

n the electropherogram ABA and NABA can clearly be indi-ated, together with PAA. Some peaks migrate after ABA, onef them possibly originating from abienol. We cannot confirmhis assumption by comparison with a standard, because abienols not commercially available. We have nevertheless some doubtbout the assignment of one of these peaks to abienol becausehere is no response at 250 nm, which should be due to the con-ugated �-electron system of this compound (cf. with the spectraf the according acids).

.3. Resin components in binder mixtures

The suitability of the CE system was tested for further anal-sis of the resin constituents in mixtures with other binders,amely with plant gums, animal glues and drying oils. Chemi-ally, plant gums are polysaccharides, animal glues are proteins,nd drying oils are triglycerides. For the present explorationhe binders were mixed in a 10-fold excess with the resin,nd the samples were then subjected to sonication at 55 ◦Cor 30 min after addition of methanol; then the samples wereentrifuged and the supernatant injected into the CE instr-ment.

The electropherogram of an extract from a mixture ofolophony with dried linseed oil obtained in this way is shown

n Fig. 5, bottom panel. Compared to colophony (Fig. 2) somemall additional peaks arise, but they appear at a migration timeefore the first resin acid. Therefore, identification of colophonys clearly possible.

A. Findeisen et al. / J. Chromatogr. A 1157 (2007) 454–461 459

Fig. 4. Electropherogram of colophony, larch turpentine (from Larix Euro-pia

wsg(Tfip

Fig. 5. Electropherograms of a mixture of gum arabic and colophony, bone glueand colophony and linseed oil and colophony. The samples contained the secondbinder in a tenfold excess over colophony. Separation and sample pre-treatmentc

gbn

ea), and Strasbourg turpentine (from Abies picea). Samples were sonicatedn methanol for 30 min at 55 ◦C, then centrifuged through filter vials, dilutednd injected. Separation conditions as in Fig. 2. For symbols see Table 1.

Bone glue, an example for animal glues of the collagen type,as treated with methanol as usual, and after centrifugation the

upernatant was analysed by CE. The resulting electrophero-ram did not show any peak larger than the detection limitthree times the standard deviation of the background noise).

he according analysis of a mixture of colophony with the 10-

old excess of bone glue results in an electropherogram in whichndeed only the peaks of this resin are detected (Fig. 5, middleanel).

Tfai

onditions as in Fig. 2. For symbols see Table 1.

Finally, potential interferences originating from vegetableums were explored analysing a methanolic extract of gum ara-ic in the usual way. For pure gum arabic no peaks above theoise appeared at the migration time of interest (not shown).

he according electropherogram (Fig. 5, upper panel) obtained

rom a 1:10 (w/w) mixture of colophony with gum arabic againllows the clear identification of the resin due to the absence ofnterfering peaks.

460 A. Findeisen et al. / J. Chromatog

Ffa

3

amTmicwwtag

amgpasccp

4

dairbotbdl

tCt

tdvdpltsaisi

ctctdtestttmcofa

ltiwe

A

fi0

R

ig. 6. Electropherograms of a sample taken from a modelling mass for gildedrames from the 19th century. Separation and sample pre-treatment conditionss in Fig. 2. For symbols see Table 1.

.4. Museum objects

One example for the identification of the resinous material inhistoric object is given for a material which was widely used forodelling reliefs on gilded frames mainly in the 19th century.hese reliefs were applied instead of those made by the muchore time consuming and elaborate carving of the ornaments

nto the wooden frames. For modelling the reliefs a pasty massonsisting of animal glue, resin, and drying oil mixed with wateras prepared, heated and the resulting paste was pressed intoooden forms with the desired ornaments. Then the mass was

aken out from the form and fixed onto the frame. After dryingnd hardening different kinds of techniques were applied for theilding of the object.

We have taken a material from a frame originating fromround 1850, and analysed a small sample (about 1 mg) afterethanolic treatment as described. The resulting electrophero-

ram is shown in Fig. 6. The typical pattern from the acidsresent in colophony are found: NABA, LPIA, PAA, DABAnd ABA are recorded. Peak identification was supported bytandard addition. No interference from other, non-resinousomponents of the complex sample takes place. It can thus beoncluded that the resin used for the preparation of the modellingaste was colophony.

. Conclusions

It was found that CE with charged additives – to intro-uce separation selectivity even for uncharged analytes – isn interesting alternative to spectroscopic and chromatographicdentification and characterization methods for diterpenoicesins applied in historic and artistic works. Such resins haveeen used for many purposes in museum objects, e.g. as coatings,r additives for glues or binders. The analytes under considera-

ion are resin acids and a resin alcohol with a skeleton formedy a bi- or tricyclic C20-hydrocarbon: abietic, neoabietic, dehy-roabietic, palustric, pimaric, isopimaric, levopimaric acid andarixol acetate. Charged cylodextrins are applied to solubilise

r. A 1157 (2007) 454–461

he analytes in aqueous solution, to enable their transport in theE system under the influence of the applied electric field, and

o modify their electrophoretic properties.In comparison to chromatography, there are several advan-

ages and disadvantages. For GC the resin acids are normallyerivatised to decrease polarity and adsorptivity and increaseolatility. Derivatisation can be laborious; moreover, the con-itions for derivatisation can transform other potential analytesresent in the sample to products, which might thus possibly beost for other analytical means. The present CE method has cer-ainly the advantage that the sample pre-treatment is extremelyimple, and that it leaves the analytes unchanged. The samplesre only sonicated at elevated temperature (55 ◦C) for 30 minn methanol, then centrifuged, and subsequently the supernatantolution is diluted with BGE and directly injected into the CEnstrument.

The poor spectral properties caused by the lack of intensehromophors, on the other hand, result in a low performance inerms of detectability by UV-vis absorbance. For analysis theoncentrations of a single analyte by the present method are inhe millimolar range. This means that around 1 mg analyte isissolved in the 500 �L volume of the methanolic sample solu-ion. This high concentration needed is not only due to the lowxtinction coefficients of the analytes, but also due to the veryhort detection path length (50 �m) of the capillary. Therefore,he analyte concentrations necessary for detection are one orwo orders of magnitude higher than for GC. However, takinghe absolute injected amount of analyte into account, the two

ethods are quite comparable. As in the present method theoncentration is several hundred ng/�L, the absolute amount isnly several hundred pg per injection (the injection volume is aew nanoliters). This does not differ from the demands on GCnalysis, where the injected volume is in the microliter range.

Although the largely differing spectral properties of the ana-ytes can be advantageously used for identification purposes,hey make quantitation cumbersome, because each analyte needsts own calibration curve. This is a drawback compared to GC,here the most common detectors (FID or MS) respond nearly

qually to all resin compounds under consideration.

cknowledgement

The authors thank the Austrian Exchange Service fornancial support for V.K. (CEEPUS Network CII-HU-0010-01-607).

eferences

[1] J.S. Mills, R. White, Studies Conserv. 22 (1977) 12.[2] J.S. Mills, R. White, The Organic Chemistry of Museum Objects, Butter-

worth Heinemann, London, 1994.[3] I. Kaml, E. Kenndler, Curr. Anal. Chem. 3 (2007) 33.[4] L. Masschelein-Kleiner, PACT (Athens, Greece) 13 (1986) 185.

[5] B.V. Kharbade, N. Srivastava, G.P. Joshi, O.P. Agrawal, J. Chromatogr. 439

(1988) 430.[6] M. Odlyha, A. Burmester, J. Therm. Anal. 33 (1988) 1041.[7] M.M. Wright, B.B. Wheals, J. Anal. Appl. Pyrol. 11 (1987) 195.[8] M. Odlyha, Thermochim. Acta 269/270 (1995) 705.

matog

[

[

[

[

[

[

[[

[[[

[

[

[

[

[

[

[

[

[

[[

[[[

[

[[

[[

A. Findeisen et al. / J. Chro

[9] M.T. Domenech-Carbo, S. Kuckova, J. de la Cruz-Canizares, L. Osete-Cortina, J. Chromatogr. A 1121 (2006) 248.

10] M. Bacci, M. Fabbri, M. Picollo, S. Porcinai, Anal. Chim. Acta 446 (2001)15.

11] F. Modugno, E. Ribechini, M.P. Colombini, Rapid. Commun. Mass Spec-trom. 20 (2006) 1787.

12] H.G.M. Edwards, D.W. Farwell, L. Daffner, Spectrochim. Acta, Part A:Molecular and Biomolecular Spectroscopy 52A (1996) 1639.

13] P. Vandenabeele, B. Wehling, L. Moens, H. Edwards, M. De Reu, G. VanHooydonk, Anal. Chim. Acta 407 (2000) 261.

14] D. Scalarone, J. van der Horst, J.J. Boon, O. Chiantore, J. Mass Spectrom.38 (2003) 607.

15] K.J. van den Berg, J.J. Boon, I.I. Pastorova, L.F. Spetter, J. Mass Spectrom.35 (2000) 512.

16] E. Kenndler, F. Mairinger, Fresenius J. Anal. Chem. 338 (1990) 635.17] M.P. Colombini, F. Modugno, S. Giannarelli, R. Fuoco, M. Matteini,

Microchem. J. 67 (2000) 385.18] G. Chiavari, D. Fabbri, S. Prati, Chromatographia 55 (2002) 611.19] S. Watts, E.R. de la Rie, Studies Conserv. 47 (2002) 257.20] L. Osete-Cortina, M.T. Domenech-Carbo, J. Chromatogr. A 1065 (2005)

265.21] L. Osete-Cortina, M.T. Domenech-Carbo, R. Mateo-Castro, J.V. Gimeno-

Adelantado, F. Bosch-Reig, J. Chromatogr. A 1024 (2004) 187.22] J. de la Cruz-Canizares, M.T. Domenech-Carbo, J.V. Gimeno-Adelantado,

R. Mateo-Castro, F. Bosch-Reig, J. Chromatogr. A 1093 (2005)

177.

23] A. Andreotti, I. Bonaduce, M.P. Colombini, G. Gautier, F. Modugno, E.Ribechini, Anal. Chem. 78 (2006) 4490.

24] V. Pitthard, M. Griesser, S. Stanek, T. Bayerova, Macromol. Symp. 238(2006) 37.

[

[[

r. A 1157 (2007) 454–461 461

25] G. Chiavari, G.C. Galletti, G. Lanterna, R. Mazzeo, J. Anal. Appl. Pyrolysis24 (1993) 227.

26] M.P. Colombini, R. Fuoco, A. Giacomelli, B. Muscatello, Studies Conserv.43 (1998) 33.

27] M.P. Colombini, F. Modugno, M. Giacomelli, S. Francesconi, J. Chro-matogr. A 846 (1999) 113.

28] M.P. Colombini, F. Modugno, R. Fuoco, A. Tognazzi, Microchem. J. 73(2002) 175.

29] I. Bonaduce, M.P. Colombini, Rapid Commun. Mass Spectrom. 17 (2003)2523.

30] M.P. Colombini, I. Bonaduce, G. Gautier, Chromatographia 58 (2003) 1.31] R. Lleti, L.A. Sarabia, M.C. Ortiz, R. Todeschini, M.P. Colombini, Analyst

128 (2003) 281.32] I. Bonaduce, M.P. Colombini, J. Chromatogr. A 1028 (2004) 297.33] M.P. Colombini, F. Modugno, J. Sep. Sci. 27 (2004) 147.34] M.P. Colombini, G. Giachi, F. Modugno, E. Ribechini, Microchem. J. 79

(2005) 83.35] I. Bonaduce, M.P. Colombini, S. Diring, J. Chromatogr. A 1107 (2006)

226.36] I. Surowiec, I. Kaml, E. Kenndler, J. Chromatogr. A 1024 (2004) 245.37] S. Harrison, I. Kaml, F. Rainer, E. Kenndler, J. Sep. Sci. 28 (2005)

1587–1594.38] I. Kaml, K. Vcelakova, E. Kenndler, J. Sep. Sci. 27 (2004) 161.39] S. Harrison, I. Kaml, V. Prokoratova, M. Mazanek, E. Kenndler, Anal.

Bioanal. Chem. 382 (2005) 1520–1526.

40] M. Großl, S. Harrison, I. Kaml, E. Kenndler, J. Chromatogr. A 1077 (2005)

80.41] M. Mazanek, I. Kaml, E. Kenndler, Studies Conserv. 51 (2006) 130.42] J.H. Luong, T. Rigby, K.B. Male, P. Bouvrette, Electrophoresis 20 (1999)

1546.