Dee

ATa

Vb

c

a

ARR2AA

KCsCLPP

1

timlaTsf

mba

0d

Journal of Chromatography A, 1218 (2011) 8561– 8566

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A

jou rn al h om epage: www.elsev ier .com/ locat e/chroma

etermination of carbohydrate- and lignin-derived components in complexffluents from cellulose processing by capillary electrophoresis withlectrospray ionization-mass spectrometric detection

nna Bogolitsynaa, Manuel Beckera, Anne-Laurence Dupontb, Andrea Borgardsc,homas Rosenaua, Antje Potthasta,∗

Department of Chemistry and Christian Doppler Laboratory “Advanced cellulose chemistry and analytics”, University of Natural Resources and Life Sciences, Muthgasse 18, A-1190ienna, AustriaCentre de recherches sur la conservation des collections, Muséum national d’Histoire naturelle, CNRS USR 3224, 36, rue Geoffroy-Saint-Hilaire, F-75005 Paris, FranceProcess Innovation, Lenzing AG, A-4860 Lenzing, Austria

r t i c l e i n f o

rticle history:eceived 30 May 2011eceived in revised form2 September 2011ccepted 23 September 2011vailable online 29 September 2011

eywords:apillary electrophoresis-mass

a b s t r a c t

Degradation products from lignocellulosic materials receive increasing attention due to the continuouslygrowing interest in their utilization. The inherent structural variance of lignocellulosics combined withthe intricacy of lignocellulosic processing (e.g. pulping of wood and bleaching of cellulosic pulps) and thecomplexity of degradation processes occurring therein result in rather complex mixtures in the processstreams and effluents that contain a large quantity of structurally different degradation products. This istrue for most processing steps, but also for degradation reactions occurring during aging of lignocellulosicmaterials, such as paper, cellulosic tissue or textiles. In order to render such mixtures better analyticallyaccessible than hitherto possible a CE-ESI-MS method was established for the simultaneous determina-

pectrometryarboxylic acidsignocellulosesulp bleaching effluentaper aging

tion of aliphatic carboxylic acids from the degradation of (hemi)celluloses and aromatic compounds fromlignin degradation. CE and ESI-MS parameters have been optimized towards sensitivity and good repro-ducibility. The method was tested in two real-world scenarios: the determination of major componentsin effluents from bleaching stages in the pulp and paper industry, and the analysis of degradation prod-ucts in extracts of naturally aged papers. The advantages and drawbacks of this approach are criticallydiscussed.

© 2011 Elsevier B.V. All rights reserved.

. Introduction

The analysis of lignocellulose degradation and identification ofhe products formed from these materials are still posing analyt-cal challenges, especially due to the complexity of the reaction

ixtures. This applies to nearly all byproduct streams in lignocel-ulose processing, no matter whether related to “classical” pulpnd paper processing or to more recent biorefinery scenarios.

he degradation products in industrial effluents after differentteps of pulp bleaching have large similarities to the productsound in extracts of aged paper. The mixtures of degradation

Abbreviations: CE, capillary electrophoresis; ESI-MS, electrospray ionization-ass spectrometry; GC–MS, gas chromatography–mass spectrometry; BGE,

ackground electrolyte; LMM, low molecular mass; MT, migration time; PA, peakrea.∗ Corresponding author. Tel.: +43 1 47654 6071; fax: +43 1 47654 6059.

E-mail address: antje.potthast@boku.ac.at (A. Potthast).

021-9673/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2011.09.063

products of low-molecular weight carbohydrates under stronglyalkaline, alkaline-oxidative or acidic-oxidative conditions arecomplex, being formed by superposition of multiple fragmentation,rearrangement, and condensation reactions. These carbohydrate-derived products come along with fragmentation products of(residual) lignin. In the pulp and paper industries, such degrada-tion products contribute significantly to the spent liquor streams,and hence also to the corresponding organic effluent load (totalorganic carbon). Degradation reactions similar to those enforcedduring the pulp bleaching stages by drastic reaction environmentsare proceeding also under ambient conditions over longer timesupon natural or artificially accelerated paper aging. Also here, car-bohydrates and lignins are fragmented and degraded into verycomplex compound mixtures. Degradation processes in paper aredetermined by numerous factors including the material of the

paper (fiber type, sizing, fillers, etc.) as well as storage condi-tions (temperature, relative humidity, acids, pollutants, etc.). As aresult of such degradation, mechanical and optical properties ofpaper can suffer dramatically [1]. More detailed information on the

8 matog

dgad

aadrei(Tv[l[

eoaIamT[

slbvpc[imta

anmno(pdp

2

2

wS

ai((aatpt

562 A. Bogolitsyna et al. / J. Chro

egradation mechanisms and the reaction products formed is ofreat significance for a suitable conservation treatment of papernd in summary for the general understanding of lignocelluloseegradation.

For the separation of degradation products of (hemi)celluloses,n analytical methodology is required which shows high sensitivitynd robustness as well as the ability to simultaneously determineegradation products of both carbohydrates and lignin. The task isendered even more complicated by the complexity in case of efflu-nt mixtures, the matrix and the rather low concentrations of itsndividual components. Gas chromatography/mass spectrometryGC–MS) is often applied for the analysis of such complex mixtures.his method requires preliminary sample derivatization to produceolatile analytes, e.g. by esterification, etherification, silylation, etc.2–4]. This is not only a time-consuming procedure, but might effectosses during sample preparation, and cause discrimination effects5–7].

In the recent years capillary electrophoresis (CE) has beenstablished as a fast and suitable method for both the analysisf carbohydrates, their degradation products (aliphatic carboxyliccids) and the determination of phenolic and aromatic compounds.t provides short analysis times and, at the same time, high sep-ration efficiencies without a preliminary derivatization step. Inost cases UV-detection is applied after separation by CE [8–11].

his method was used to study effluents of pulping woody biomass12–14], which offers complex matrices.

Hyphenation to MS provides more information on the actualtructures of the analytes. CE-MS was applied for the analysis ofow molecular mass carboxylic acids in atmospheric particles [7],everages [5,15], biological fluids [16,17], urban atmosphere andehicle emissions [6]. CE-MS was used further for the analysis ofhenolic compounds (with similar structures to the lignin-derivedompounds) in biomass pyrolysis and burning [18], virgin olive oil19,20], walnut [21] and atmospheric aerosols [22]. Pulp bleach-ng effluents and celluloses aging extracts have been studied by CE

uch less extensively, since due to dilution effects of the processhe overall concentrations – at similarly complex compositions –re significantly lower.

In this study, we would like to communicate our efforts towards CE-MS-based analytical technique for the simultaneous determi-ation of carbohydrate-derived reaction products, such as aliphaticono- and di-carboxylic acids and hydroxy-acids, as well as phe-

olic lignin-fragments in effluents of pulp bleaching and extractsf aged paper artifacts. Together with our parallel study onhemi)cellulose and lignin degradation products in artificially agedapers [23], this is the first report on a MS-hyphenated method forirect analysis of the complex pulp bleaching effluents and agedaper extracts.

. Materials and methods

.1. Chemicals

All chemicals were of the highest purity available and were usedithout further purification. Ultrahigh quality water (HPLC grade,

igma–Aldrich) was used for all aqueous solutions.The chemicals were obtained from the following suppliers:

mmonium formate (97%), ammonium hydroxide solution (≥25%n water) and sodium hydroxide (≥98%) from Sigma–Aldrich–FlukaSchnelldorf, Germany), 2-propanol (99.9%) from Fisher ScientificGermany). All the model compounds (either as the free acids ors their sodium salts) were obtained from Sigma–Aldrich–Fluka,

t a purity of 97% or better except lactic acid (90%). Stock solu-ions of each model compound with a concentration of 1 g/l wererepared in deionized water and stored at 4 ◦C. Solid phase extrac-ion cartridges SupelTM-Select HLB SPE 1 g/20 ml were purchased

r. A 1218 (2011) 8561– 8566

from Supelco (Bellefonte, PA, USA). Phenex polytetrafluoroethy-lene (PTFE) syringe filters with pore size of 0.2 and 0.45 �m andvarious diameters were supplied by Phenomenex (Aschaffenburg,Germany).

2.2. Sample preparation

2.2.1. Pulp bleaching effluentsThe pulp bleaching effluent sample was obtained from com-

bined effluents of a totally chlorine free (TCF) bleaching ofhardwood sulfite pulp, prior to the biological effluent treatment,from Lenzing AG, Lenzing, Austria. The sample was stored at 4 ◦Cand was warmed to room temperature before sample preparation,which started with vacuum filtration through 0.4 �m membranefilters. As the concentration of analytes was too low for thedirect analysis, effluent samples were concentrated by solid-phaseextraction (SPE) according the following procedure [24]: the SPEcartridge was conditioned with 10 ml methanol and 10 ml ofdeionized water, 200 ml of sample was loaded with 1 drop persecond. Salts are not retained on the SPE cartridge, which thusallows the separation of organic analytes from salts in the load-ing step. Carboxylic acids were eluted from the cartridge with amethanol/acetonitrile mixture (v/v = 1:1). The organic solvent wasremoved in vacuum under exclusion of oxygen, and the samplewas redissolved in deionized water. This removes long-chain car-boxylic acids in amounts according to their solubility in water[25]: acids with up to 10–12 carbon atoms are still in the sam-ple, larger ones are completely removed. This step is necessaryto protect the separation capillary from clogging. Evaporation ofthe organic solvent did not cause changes in the analyte compo-sition which was proved by GC–MS analyses of silylated samplesbefore and after the treatment (recovery above 90%). Prior toinjection, the sample was filtered through a 0.2 �m membranefilter.

2.2.2. Aged paper extractsAqueous extracts from an old book and aged papers were pre-

pared as follows: 17 g of air dry book paper were extracted in 230 mlof deionized water containing 10% methanol for 45 h. The extractwas brought to a volume of about 25 ml by freeze-drying and wasfiltered through a 0.2 �m membrane filter prior to injection.

Characteristics of the book paper: publication year 1924, brit-tle paper, double fold number: 2. The weighted average molecularweight (Mw) of the cellulose as determined by GPC in DMAc/LiCl[26] was between 100 and 120 kg/mol.

2.3. CE-ESI-MS

CE-MS analysis was performed on a G1600 Agilent capil-lary electrophoresis system (Agilent Technologies, Waldbronn,Germany) in combination with an Agilent 6320 series ion trap massspectrometer equipped with an Agilent CE-ESI-MS sprayer (Agi-lent Technologies). For separation, a fused-silica capillary (AgilentTechnologies) with a total length of 60 cm and an inner diameterof 50 �m was used. For maintaining constant performance overtime, the capillary was flushed daily with 0.1 M sodium hydroxidefor 10 min, water for 10 min and background electrolyte (BGE) for5 min. Between the actual runs, the capillary was preconditionedby flushing 5 min with water and 5 min with BGE. BGE and sampleswere filtered through 0.2 �m membranes. Hydrodynamic injec-tion was used at 50 mbar for 10 s followed by injection of buffer

at 50 mbar for 5 s. The separation voltage was 20 kV, the resultingcurrent was 20 �A. The capillary was thermostated at 25 ◦C.

The sheath liquid was delivered by an Agilent 1200 series iso-cratic pump equipped with a 1:100 splitter. System control, data

matogr. A 1218 (2011) 8561– 8566 8563

as

3

3

tao1leflt

pActwsvtc

lTi0o

2Ftothw

rhtiop

tcfae

iwdt

lBtctB

Fig. 1. Electropherograms of standard LMM carboxylic acids with optimized CE-ESI-MS parameters: total ion electropherogram (a) and extracted ion electropherogramsfor glucuronate (b), decanoate (c), 8-hydroxyoctanoate (d), xylonite (e), threonate

A. Bogolitsyna et al. / J. Chro

cquisition and data analysis were performed using Agilent Chem-tation for CE and Agilent LC/MSD Trap software for MS.

. Results and discussion

.1. Method optimization

For method optimization a model mixture was used con-aining succinic (1,4-butanedioic acid), n-valeric (n-pentanoiccid), pelargonic (n-nonanoic acid) and sebacic acid (1,8-ctanedicarboxylic acid) dissolved in water at concentrations of0 mg/l each. The acids were chosen to represent various chain

engths as well as mono-acids and di-acids. Optimized MS param-ters were dry gas temperature and flow, nebulizing gas pressure,ow rate, content of ammonium hydroxide and organic solvent inhe sheath liquid.

As organic solvents for the sheath liquid, methanol and 2-ropanol were tested. In agreement with Hagberg [5] as well ashrer and Buchberger [27], the results were successful in thease of both alcohols, but 2-propanol has an additional advan-age of improving the ionization of acids. Hence, 2-propanolas used for further concentration optimization. Four different

heath liquid concentrations were tested, having an organic sol-ent content of 30, 50, 70 and 80%. The results showed thathe peak area and intensities increased with higher 2-propanolontent.

Optimal concentration of ammonium hydroxide in the sheathiquid was tested with solutions of 0.025, 0.05, 0.075 and 0.1%.his parameter did not have a strong influence on peak areas andntensities, and there was no significant difference in the range of.025–0.075%, whereas a slight increase was observed in the casef 0.1% ammonium hydroxide.

The sheath liquid flow rate was studied in the range of–6 �l/min. A flow rate of 2 �l/min was insufficient for separation.rom 3 to 5 �l/min, a slight increase of peak areas and intensi-ies was monitored for all tested standard acids. At a flow ratef 6 �l/min a decreased intensity was observed. Resulting fromhese tests, a sheath liquid with a concentration of 0.1% ammoniumydroxide and 80% 2-propanol content at a flow rate of 4 �l/minas used.

With regard to nebulizing gas pressure, an optimal signalesponse was found between 12 and 16 psi, the range of 5–20 psiaving been tested. Values below 12 psi did not produce a satisfac-ory signal at all, while pressures above 16 psi caused lower peakntensities. A fine tuning experiment in the range of 12–16 psi gaveptimal results for 14 psi, which then was chosen as the workingressure for the nebulizing gas.

Dry gas temperatures of 200 ◦C, 250 ◦C, 300 ◦C and 350 ◦C wereested during parameter optimization. The signal intensities forarboxylic acids increased with higher temperatures up to 300 ◦C,urther temperature increase caused the intensity to decreasegain. Thus, a dry gas temperature of 300 ◦C was used for furtherxperiments.

Dry gas flow rates between 3 and 7 l/min did not significantlynfluence the signal intensities. While between 3 and 5 l/min there

as no difference in peak areas and intensities at all, a minorecrease in intensity was observed in the range of 5–7 l/min. Fur-her experiments used a dry gas flow of 5 l/min.

Ammonium formate was used as the BGE for the separation ofow molecular mass (LMM) carboxylic acids. The concentration ofGE was varied between 10 and 40 mM. While for the model mix-

ure analysis a BGE concentration of 10 mM was sufficient, moreomplicated multi-component samples demanded higher concen-rations to maintain separation quality. On the other hand, highGE concentrations of 40 mM caused instabilities in the capillary

(f), glycerate (g), azelate (h), succinate (i), malate (j). Capillary 60 cm × 50 �m I.D.;electrolyte, 20 mM ammonium formate; pH 9; applied potential, 20 kV.

current and lower peak intensities. The pH of the BGE was stud-ied in the range between 9 and 11. pH values above 10 caused CEcurrent instability and therefore cannot be applied. A satisfactoryseparation was achieved with 20 mM ammonium formate buffer ofpH 9.

3.2. Analysis of LMM carboxylic acids

After optimization of the fundamental parameters, the methodwas tested on a model mixture containing nine LMM carboxylicacids (Fig. 1; Table 1), among them two sugar acids, representativefor effluents with wood processing and degradation products. Themixture was reliably separated within 15 min, detection was byESI-MS.

Limits of the method arise for the determination of carboxylicacids with molecular mass below 70. For example, acetic acid(60 g/mol) is outside the MS-detection limit. Also short diacids can-not be MS-detected due to their high ionization potential and thedouble charge (z = 2): injection of oxalic acid did not result in a sig-nificant peak, even at comparatively high concentration. Aliphatic

acids with long chains (>C12) were not analyzed as they had beenexcluded from the sample during sample pre-treatment to avoidproblems with capillary clogging.

8564 A. Bogolitsyna et al. / J. Chromatogr. A 1218 (2011) 8561– 8566

Table 1Model mixture content and method evaluation.

Peak ID Compound Mw m/z in(−)ESIa

Migration time(min)

Linear correlationcoefficient, R

LOD (mg/l) LOQ (mg/l) Repeatability,% (n = 7)

Time Area

Aliphatic carboxylic acidsb Glucuronic acid 194 193 4.5 0.9973 0.73 2.45 0.3 3c Decanoic acid 172 171 4.4 0.9677 1.10 3.67 0.5 8d 8-Hydroxyoctanoic acid 160 159 4.5 0.9894 0.63 2.11 0.6 12e Xylonic acid 166 165 4.8 0.9992 0.58 1.94 0.7 10f Threonic acid 136 135 5.1 0.9973 0.55 1.82 0.5 4g Glyceric acid 106 105 5.8 0.8532 0.69 2.31 0.8 16h Azelaic acid 188 187 6.8 0.9925 0.91 3.03 0.3 22i Succinic acid 118 117 12.1 0.9520 0.92 3.07 0.5 17j Malic acid 133 132 12.5 0.9921 0.25 0.85 0.6 16

Lignin-derived compoundsb Acetovanillone 166 165 5.7 0.9951 0.19 0.64 0.6 15c 4-Hydroxyacetophenone 136 135 5.9 0.9959 0.16 0.53 0.7 20d Vanillin 152 151 6.2 0.9981 0.31 1.03 0.7 22e Ferulic acid 194 193 6.5 0.9953 0.30 1.01 0.8 18f 4-Hydroxybenzaldehyde 122 121 6.6 0.9946 0.09 0.31 0.8 14g Vanillic acid 168 167 7.0 0.9965 0.86 2.87 0.4 12

0.9959 0.53 1.78 0.2 20

3

acma

abwHsnsis

3

martLrifptRaos

3

eta

Fig. 2. Electropherograms of standard lignin derivatives with optimized CE-ESI-MSparameters: total ion electropherogram (a) and extracted ion electropherograms for

h 4-Hydroxybenzoic acid 138 137 7.5

a Post-run single ion.

.3. Analysis of lignin fragments

The method optimized for the separation of LMM carboxyliccids was also applied to the separation of phenolic modelompounds (Fig. 2). The lignin model mixture contained lignin frag-ents characteristic for softwood, hardwood and herbaceous lignin

s well as typical lignin degradation products (Table 1) [28,29].Injecting a model mixture of seven representative lignin-related

romatic compounds under conditions optimal for the LMM car-oxylic acids produced good results, confirming that the methodas successful as well for the analysis of lignin-derived compounds.ence, the method can be used for the simultaneous separation and

ubsequent identification of both LMM carboxylic acids and phe-olic lignoid compounds. Although the electropherogram showedome overlapping, this does not cause any serious problem for peakdentification and integration, which is done based on the well-eparated ion traces extracted from the electropherogram.

.4. Evaluation of the method

A method evaluation has been carried out to compare theethod against other options for LMM acid and lignin fragment

nalysis. Limit of detection (LOD), limit of quantification (LOQ) andepeatability (relative standard deviations, RSD%) for the migra-ion time (MT) and the peak area (PA) are given in Table 1. LOD andOQ were quantified from the threefold and tenfold signal-to-noiseatio, respectively. The repeatability was checked by repetitivenjection of the pulp bleaching effluent sample (n = 7). RSD valuesor MT and PA were less than 0.8% and 3–22%, respectively. Com-arison with literature data showed acceptable repeatability forhe ESI-MS detection. Van Pinxteren and Herrmann [7] reportedSD values of 0.2–0.5% for MT and 4–21% for PA in the analysis ofliphatic and aromatic carboxylic acids. Also in comparison withther literature accounts [15,18,27] the parameter values in thistudy were similar or better.

.5. Analysis of pulp bleaching effluents

The developed method was applied to the analysis of efflu-nts from pulp bleaching stages of the TCF (totally chlorine free)ype. The samples contained a large number of aliphatic carboxyliccids, which comprise short-chain mono- and di-acids, short-chain

acetovanillone (b), 4-hydroxyacetophenone (c), vanillin (d), ferulic acid isomers (e),4-hydroxybenzaldehyde (f), vanillic acid (g), 4-hydroxybenzoic acid (h). Capillary60 cm × 50 �m I.D.; electrolyte, 20 mM ammonium formate; pH 9; applied potential,20 kV.

A. Bogolitsyna et al. / J. Chromatogr. A 1218 (2011) 8561– 8566 8565

Fig. 3. Base peak electropherogram of combined bleaching effluent prior biologicalt

hnateptwttspcm

cps

TC

Fig. 4. Base peak electropherogram of aqueous extract from naturally aged bookpaper. For peak labeling see Table 3.

Table 3Compounds identified in the aqueous paper extract.

Peak ID Compound Mw m/z in(−)ESIa

1 Xylonic acid 166 1652 Threonic acid 136 1353 Vanillic acid 168 1674 Glyceric acid 106 1055 Lactic acid 90 896 Glycolic acid 76 757 Unidentified sugar acid 196 1958 Succinic acid 118 1179 Malic acid 134 133

10 Tartaric acid 150 149

reatment. For peak labeling see Table 2.

ydroxyacids, and medium-chain monoacids; components origi-ating from lignin detected in the sample were vanillic, syringicnd cinnamic acids (Fig. 3; Table 2). Peak assignment and quan-ification has been performed using the post-run single-ion tracesxtracted from the electropherogram, and were verified by com-arison of the migration time with standard compounds. Due tohe lack of standard compounds, not every peak was verified in thisay. Identification of compounds by ion fragmentation was rather

ime consuming due to the complexity of sample. Combination ofhe CE-MS analysis with a GC–MS approach (after derivatization,uch as silylation) provided additional information on the sam-le composition as it offers access to a much broader database ofhemical compounds which was used to support the data for CE-MSethod development.Generally, difficulties in the analysis arise from the high con-

entration of inorganic salts in the sample, which might hamper aerfect separation of some compounds in the effluent sample andhift the migration times.

able 2ompounds identified in the bleaching effluent sample.

Peak ID Compound Mw m/z in(−)ESIa

1 Syringic acid 198 1972 8-Hydroxyoctanoic acid 160 1593 Threonic acid 136 1354 Decanoic acid 172 1715 2-Hydroxyhexanoic acid 132 1316 Octanoic acid 144 1437 Tricarboxylic acid (citric or isocitric) 192 1918 Cinnamic acid 148 1479 Xylonic acid 166 165

10 Vanillic acid 168 16711 Unidentified sugar acid 196 19512 Azelaic acid 188 18713 Hexanedioic acid 146 14514 2-Ketogluconic acid 211 21015 2,4,5-Trihydroxypentanoic acid 150 14916 Malonic acid 104 10317 Unidentified sugar acid 166 16518 Succinic acid 118 11719 Malic acid 134 13320 Maleic acid 116 115

a Post-run single ion.

11 Maleic acid 116 115

a Post-run single ion.

3.6. Analysis of extract from aged paper

The extract from aged paper (Fig. 4; Table 3) mainly containedLMM mono- and di-acids, originating from far-reaching carbohy-drate degradation, as well as sugar acids. As mentioned above,carboxylic acids with very short carbon chain length (e.g. formicacid and acetic acid) were not detected due to their low molecularmass beyond the grasp of the MS detector, but they can be deter-mined by other means, e.g. conventional CE with UV-detection.As to the phenolic compounds from lignin, only vanillic acid wasdetected in the sample. The concentrations of the other lignin-derived compounds in the sample were too low to be identified.A matrix effect has been observed during the analysis of the paperextract sample as well, but it was by far not as strong as comparedto the effluent sample and did not disturb the separation. This canmainly be attributed to the missing inorganic salt freight comparedto the bleaching effluent.

4. Conclusions

The present CE-MS method offers an approach to direct analysisof most individual organic components, especially polar ones suchas acids and phenols, in some overly complex mixtures of carbo-hydrate and lignin degradation products, as occurring in industrialpulp processing or upon aging of cellulosic materials. The directanalysis refers to the fact that no derivatization is necessary as

in GC analysis (for volatility) which renders the method simpleand rapid, and the loss of components during such pretreatmentsteps is avoided. For unknown components not being identifiableby comparison to standards, the ESI-MS-hyphenation additionally

8 matog

ptaicpavo

cuoCoprpiodlcs

A

Ntp

DCgMAD

[[[[

[

[[[[[

[

[

[

[[

[

[Biomacromolecules 3 (2002) 959.

[27] W. Ahrer, W. Buchberger, J. Chromatogr. A 854 (1999) 275.

566 A. Bogolitsyna et al. / J. Chro

rovides (albeit limited) indications as to the chemical nature ofhese compounds. A comparably fast and efficient separation ofliphatic carboxylic acids and phenolic compounds from the stud-ed analytes can be achieved in one single run without the need tohange solvents. Only sample pre-concentration might be requiredrior the injection in case of very dilute solutions. The method ispplicable to any aqueous or water soluble sample and requiresery small sample amounts as the injection volume is in the rangef nanoliters.

The method presented can be complemented by other analyti-al methods as for instance GC–MS, which on one hand offers thetilization of more comprehensive compound databases, but on thether hand entails a derivatization step. Compared to GC–MS, theE-MS method shows lower sensitivity, but presents informationn the unaltered sample composition as no derivatization or sam-le preparation is involved as in the case of GC, and it offers moreapid and robust analysis. Although method improvements couldossibly further reduce the negative matrix effects of inorganic salts

n the analyte samples, the present method, being the first reportf the simultaneous determination of (hemi)cellulose and ligninegradation products by CE-ESI-MS in very complex mixtures of

ignocellulosic degradation products, is expected to find wide appli-ation in the pulp and paper industries as well as in conservationcience of historic cellulosic objects.

cknowledgements

The authors would like to thank Kyujin Ahn, MSc, University ofatural Resources and Life Sciences, Vienna, for providing data on

he book sample and Dr. Sonja Kirschnerova, for the book samplereparations.

The financial support by COST Action FP0901, the Christianoppler Research Society (CD laboratory “Advanced Cellulosehemistry and Analytics”) and Lenzing AG, Lenzing, Austria, is

ratefully acknowledged. We appreciate the helpful advice of Dr.arkus Himmelsbach and Dr. Manuela Haunschmidt, Institute of

nalytical Chemistry, Johannes-Kepler-University Linz, Austria andr. G. Götzinger, Lenzing AG, Lenzing, Austria.

[[

r. A 1218 (2011) 8561– 8566

References

[1] M. Strlic, J. Kolar, Ageing and Stabilization of Paper, National and UniversityLibrary, Ljubljana, 2005.

[2] R. Malinen, E. Sjöström, Paperi ja Puu 11 (1975) 728.[3] K. Niemelä, R. Alen, in: E. Sjöström, R. Alen (Eds.), Analytical Methods in Wood

Chemistry, Pulping, and Papermaking, Springer, Heidelberg, Germany, 1999.[4] H. Hirschlag, R. Küster, Fresen. J. Anal. Chem. 263 (1998) 274.[5] J. Hagberg, J. Chromatogr. A 988 (2003) 127.[6] E. Dabek-Zlotorzynska, R. Aranda-Rodriguez, L. Graham, J. Sep. Sci. 28 (2005)

1520.[7] D. van Pinxteren, H. Herrmann, J. Chromatogr. A 1171 (2007) 112.[8] E. Sjoholm, N.-O. Nilvebrant, A. Colmsjo, J. Wood Chem. Technol. 13 (1993)

529.[9] A.-L. Dupont, C. Egasse, A. Morin, F. Vasseur, Carbohydr. Polym. 68 (2007) 1.10] T. Javor, W. Buchberger, O. Faix, Anal. Chim. Acta 484 (2003) 181.11] C. Klampfl, W. Buchberger, Trends Anal. Chem. 16 (1997) 221.12] D. Volgger, A. Zemann, G. Bonn, J. High Res. Chromatogr. 21 (1998) 3.13] D.T. Nguyen, H. Lerch, A. Zemann, G. Bonn, Chromatographia 46 (1997)

113.14] S.M. Masselter, A.J. Zemann, G.K. Bonn, J. High Res. Chromatogr. 19 (1996)

131.15] H. Sawada, C. Nogami, Analyt. Chim. Acta 507 (2004) 191.16] B. Baena, A. Cifuentes, C. Barbas, Electrophoresis 26 (2005) 2622.17] W.-C. Yang, F.E. Regnier, J. Adamec, Electrophoresis 29 (2008) 4549.18] Y. Iinuma, H. Herrmann, J. Chromatogr. A 1018 (2003) 105.19] A. Carrasco-Pancorbo, L. Cerretani, A. Bendini, A. Segura-Carretero, G. Lercker,

A. Fernandez-Gutierrez, J. Agric. Food Chem. 55 (2007) 4771.20] J.J.B. Nevado, G.C. Penalvo, V.R. Robledo, G.V. Martinez, Talanta 79 (2009)

1238.21] A.M. Gomez-Caravaca, V. Verardo, A. Segura-Carretero, M.F. Caboni, A.

Fernandez-Gutierrez, J. Chromatogr. A 1209 (2008) 238.22] M.M. Yassine, E. Dabek-Zlotorzynska, P. Schmitt-Kopplin, Electrophoresis 30

(2009) 1756.23] A.L. Dupont, A. Seeman, B. Lavedrine, Talanta (submitted for publication).24] Supel-Select HLB SPE Instruction/Data Sheet & Troubleshooting Guide;

Data Sheet No. T708013, SUPELCO, Bellefonte, PA, 2008. http://www.sigmaaldrich.com/etc/medialib/docs/Supelco/Datasheet/t708013.Par.0001.File.tmp/t708013.pdf (accessed 7.02.11).

25] J. Dee, H.S. Stoker, Organic and Biological Chemistry, Cengage Learning,Brooks/Cole, 2009.

26] J. Röhrling, A. Potthast, T. Rosenau, T. Lange, G. Ebner, H. Sixta, P. Kosma,

28] E. Adler, Wood Sci. Technol. 11 (1977) 169.29] D. Fengel, G. Wegener, Wood: Chemistry, Ultrastructure, Reactions, Walter de

Gruyter, Berlin/New York, 1989.