Journal of Chromatography A, 1194 (2008) 139–142

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Short communication

Determination of phenols and chlorophenols as trimethylsilyl derivatives

using gas chromatography–mass spectrometry
Ágnes Kovács a,∗ , Anikó Kende a , Mária Mörtl b , Gábor Volk c , Tamás Rikker c , Kornél Torkos a
a
Wessling International Research and Education Centre, Laboratory of Separation Techniques, Eötvös University, Pázmány Péter sétány 1/A, H-1117 Budapest, Hungary
b
Institute of Chemistry, Eötvös University, Pázmány Péter sétány 1/A, H-1117 Budapest, Hungary
c
Dr. E. Wessling Chemical Laboratory Ltd., Fóti út 56, H-1047 Budapest, Hungary

a r t i c l e i n f o a b s t r a c t

Article history: A rapid and simple method is described for the simultaneous determination of 6 phenols (phenol, o-,
Received 11 December 2007 m-, p-cresol, catechol and resorcinol) and 19 chlorophenols (all mono-, di-, tri-, and tetrachlorophenol
Received in revised form 16 April 2008 isomers and pentachlorophenol) present in aqueous samples. The method is based on derivatization with
Accepted 21 April 2008
trimethylsilyl-N,N-dimethylcarbamate (TMSDMC). In contrast to other derivatization agents, TMSDMC
Available online 23 April 2008
instantaneously reacts with the phenolic compounds at room temperature and no further sample pro-
cessing is necessary prior to instrumental analysis. The determination of the derivatives was performed by
Keywords:
capillary gas chromatography–mass spectrometry (GC–MS). The stability of the most instable trimethylsi-
Phenols
Chlorophenols
lyl derivative (pentachlorophenol) was studied using different excess levels of the derivatization reagent.
Silylation The derivatization method was tested on spiked water samples preconcentrated by solid phase extraction
Derivatization of phenols on Isolute ENV+ cartridge. The overall method gave detection limits of 0.01–0.25 ␮g/L for all compounds
Derivatization of chlorophenols and <0.05 ␮g/L for 17 of them.
Gas chromatography–mass spectrometry © 2008 Elsevier B.V. All rights reserved.

1. Introduction butyldimethylsilyl)-N-methyltrifluoracetamide [18]. The reaction

Due to their toxicity, persistence, and unpleasant organolep-
tic properties [1,2] both the US Environmental Protection Agency
(EPA) and the European Union (EU) has classified several phenols
and chlorophenols as priority pollutants [3]. Phenolic com-
pounds are most often determined by gas chromatography with
various detectors [4] among which mass spectrometry offers
unrivalled resolution and selectivity. If underivatized, phenolic
compounds are prone to irreversible adsorption on active centers
of the separation system, resulting in peak tailing and substance
loss [5,6]. Although several derivatization methods have been
proposed, including acetic anhydride [7–12], diazomethane [1],
benzoyl-chloride [13], and pentafluorobenzoyl-chloride [2,14] or
pentafluorobenzyl-bromide [15] for electron capture detection
(ECD), the derivatization of phenols and chlorophenols is not
straightforward. Acylation requires alkaline pH, the optimum being
between 9 [7] and 11 [16], but catechol and resorcinol are prone
to oxidative degradation at pH above 8 [17]. Diazomethane is
highly explosive and carcinogenic [3] and needs to be generated
in situ. The most widely applicable method is silylation with N-(t-
requires 80 ◦ C and 1 h to take place.
Thus there is a need for new, less cumbersome methods appli-
cable to all phenols and chlorophenols relevant in environmental
analysis. Knausz et al. efficiently silylated alcohols, phenols, and
carboxylic acids with trimethylsilyl-N,N-dimethylcarbamate (TMS-
DMC) [19]:
R-OH + (CH3 )2 NC(O)OSi(CH3 )3 → ROSi(CH3 )3 + (CH3 )2 NH + CO2
Eluting early in the chromatogram, the reaction byproducts and
excess derivatization reagent do not interfere with analyte detec-
tion. The lack of steric hindrance, as well as the autocatalytic [20]
and non-equilibrium character of the reaction provide for short
reaction time. With a mild excess (10%) of TMSDMC, the silylation
of unsubstituted phenol took place at room temperature within
10 min [19]. Taking advantage of these characteristics, in this study a
fast and simple method was developed for the simultaneous deter-
mination of 6 phenols and 19 chlorophenols from aqueous samples.
2. Experimental

2.1. Chemicals

∗ Corresponding author. The investigated phenols and chlorophenols were purchased

E-mail address: agnes1062@freemail.hu (Á. Kovács). from Supelco (Phenols Kit 27, Gland, Switzerland), except for

0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2008.04.043
140 Á. Kovács et al. / J. Chromatogr. A 1194 (2008) 139–142
catechol (Aldrich, Milwaukee, WI, USA) and resorcinol (Merck,
Darmstadt, Germany). Stock solutions for each individual phe-
nol (ca. 7.5 mg/mL) were prepared in dichloromethane. 2,4,6-
Tribromophenol (Fluka Chemika, Buchs, Switzerland) was dis-
solved in n-hexane (5.892 ␮g/mL) and used as surrogate standard
(SSTD). A 400 ␮g/mL solution of 1,2,3,5-tetrachloro-4,6-dimethyl-
benzene (Aldrich), prepared in acetonitrile, was used as injection
standard. TMSDMC was prepared by one of us (MM), but the same
can be purchased from Fluka or Aldrich. All solvents and anhy-
drous sodium sulfate (pro analysi) were purchased from Merck.
37% (m/m) hydrochloric acid (puriss.) was obtained from Reanal
(Budapest, Hungary). For solid phase extraction, Isolute ENV+,
500 mg styrene–divinylbenzene copolymer cartridges were used
(IST, Mid Glamorgan, UK).
2.2. Instrumentation
Solid phase extractions (SPEs) were carried out on an IST Vac-
Master manifold. Vacuum through extraction and drying of the
sorbent was provided by a water jet pump. For elution, a KNF Neu-
berger rotary vacuum pump (model N022AN.18) was used.
GC–MS analysis was carried out on an Agilent 6890/Agi-
lent 5973 system, equipped with an Agilent 7683 Autosampler
and Agilent 7683B Injector. The instrument was operated in the
pulsed splitless mode (injection pulse pressure 103.4 kPa until
1.00 min) and the injection port temperature was maintained at
280 ◦ C. The injection volume was 2 ␮L. The column was a HP-5ms,
30 m × 0.25 mm × 0.25 ␮m. The oven temperature was held at 45 ◦ C
for 1.5 min, increased to 100 ◦ C at 10 ◦ C/min, increased to 180 ◦ C at
4 ◦ C/min, increased to 300 ◦ C at 40 ◦ C/min and held at 300 ◦ C for
5 min. The carrier gas was helium of 99.999% purity at a constant
flow rate of 1.1 mL/min. Standard electron impact conditions (70 eV)
were used, with transfer line, source and quadrupole temperatures
of 300, 230, and 150 ◦ C, respectively. The solvent delay was set to
8.3 min.
For exploratory work concerning derivatization, a Chrompack
CP 9000 gas chromatograph with flame ionization detection (FID)
was used. The column was an HP-ultra1, 25 m × 0.2 mm × 0.33 ␮m.
The temperature of the detector was 290 ◦ C, and the head pressure
was 125 kPa. 2 ␮L was injected manually. The temperature program
started at 60 ◦ C, increased to 260 ◦ C at 10 ◦ C/min, and then held at
260 ◦ C for 5 min.
2.3. Analytical procedures
2.3.1. Complete procedure
Solid phase extraction was performed as described by Reitzel
and Ledin [21]. Given the sensitivity of the silylating reagent and
the silyl derivatives to moisture, we deemed it necessary to insert a
1-mL polypropylene syringe filled with anhydrous sodium sulfate
under the cartridge, so that the eluent flowed through it before
dropping into the collector vial. To control for differences in extract
Fig. 1. TIC chromatogram of standard solution of the trimethylsilyl derivatives of 6 phenols and 19 chlorophenols (25 ␮g/mL).
volume, 25 ␮L injection standard was added to the extract before
derivatization. 100 ␮L of the extract was transferred to a crimp neck
vial equipped with a 0.2-mL micro-insert and 100 ␮L of TMSDMC
was added. The vial was immediately sealed with PTFE-lined cap
and shaken vigorously on Vortex mixer for 5 s, then analyzed on
GC–MS within 12 h.
2.3.2. Gas chromatography–mass spectrometry
Mass spectral data and retention time of the trimethylsilyl
derivatives were obtained in the full scan mode. For quantitative
purposes, selected ion monitoring (SIM) mode was used: three
characteristic ions per compound were selected and scanned using
time windows (Table 1). The separation of the trimethylsilyl deriva-
tives of the 25 compounds studied was complete in 28 min, except
for 2,5- and 2,6-dichlorophenol which yielded a common peak
(Fig. 1). GC–MS chromatograms were evaluated using relative peak
areas, i.e., area of analyte peak/area of ISTD peak.
2.3.3. Calibration and limits of detection (LODs)
Instrumental calibration curves were established with stan-
dard solutions of seven concentration levels between 0.001 and
25 ␮g/mL, two replicates each, prepared in ethyl acetate. The cali-
bration curve was linear in the entire range (R2 = 0.993–1.000) and
zero was included in the 95% confidence intervals of the intercepts.
Instrumental detection limits were determined by injecting stan-
dard solutions that were serially diluted down to 0.001 ng/mL. The
LOD represents the lowest concentration level at which the tar-
get ion could be definitely seen with a signal-to-noise ratio 3:1 or
greater. Detection limits for the complete procedure were calcu-
lated using:
(instrumental LOD) × 2 × 2.5 mL
LOD =
100 mL
The average volume of the extract was 2.5 mL, and a twofold dilu-
tion occurred when adding the derivatization agent to the aliquot.
2.3.4. Recoveries
Six replicates of deionized water (100 mL each) and 1 mL ethyl
acetate were spiked with 20 ␮L stock solution (250 ␮g/mL for each
phenol) and 10 ␮L SSTD solution. Solid phase extraction was per-
formed on the spiked water samples; the spiked ethyl acetate was
used as reference standard. Recovery for analytes and SSTD was cal-
culated by comparing the relative peak area measured in the solid
phase extraction extract to that measured in the reference solution.
Since the surrogate standard undergoes derivatization, monitoring
its recovery provided a simple way to detect both gross sample
processing errors and derivatization mishaps.
 Á. Kovács et al. / J. Chromatogr. A 1194 (2008) 139–142 141

Table 1
Retention times on HP-5ms column, molecular weights and mass spectral data of the trimethylsilyl derivatives of phenolic compounds

Compound Retention time (min) MW (g mol−1 ) M+ (relative intensity) [M−15]+ (relative intensity) Selected fragments (relative intensity) (m/z)

Phenol–TMS 8.50 166 166(25) 151(100) 151(100) 166(25) 152 (14)

o-Cresol–TMS 9.49 180 180(63) 165(100) 165(100) 135(55) 180(63)
m-Cresol–TMS 9.64 180 180(33) 165(100) 165(100) 180(33) 91(19)
p-Cresol–TMS 9.80 180 180(37) 165(100) 165(100) 180(37) 91(18)
2-Chlorophenol–TMS 10.85 200 200(28) 185(100) 185(100) 149(98) 200(28)
3-Chlorophenol–TMS 11.13 200 200(27) 185(100) 185(100) 187(36) 200(27)
4-Chlorophenol–TMS 11.46 200 200(33) 185(100) 185(100) 200(33) 187(36)
Catechol–TMS 13.43 254 254(12) 239(4) 73(100) 254(12) 151(5)
2,5/2,6-Dichlorophenol–TMS 14.11 234 234(15/22) 219(58/100) 219(58/100) 93(100/64) 234(15/22)
3,5-Dichlorophenol–TMS 14.25 234 234(29) 219(100) 219(100) 221(70) 234(29)
2,4-Dichlorophenol–TMS 14.56 234 234(20) 219(63) 219(63) 221(43) 234(20)
Resorcinol–TMS 15.27 254 254(67) 239(100) 239(100) 254(67) 73(52)
2,3-Dichlorophenol–TMS 15.27 234 234(13) 219(60) 219(60) 221(40) 234(13)
3,4-Dichlorophenol–TMS 15.55 234 234(24) 219(100) 219(100) 221(67) 234(35)
2,4,6-Trichlorophenol–TMS 17.75 268 268(25) 253(100) 255(97) 253(100) 268(25)
2,3,5-Trichlorophenol–TMS 18.32 268 268(11) 253(49) 255(49) 253(49) 268(11)
2,4,5-Trichlorophenol–TMS 18.50 268 268(13) 253(45) 255(44) 253(45) 268(13)
2,3,6-Trichlorophenol–TMS 18.68 268 268(20) 253(99) 255(99) 253(99) 268(20)
3,4,5-Trichlorophenol–TMS 19.62 268 268(38) 253(100) 255(98) 253(100) 268(38)
2,3,4-Trichlorophenol–TMS 20.02 268 268(14) 253(49) 255(49) 253(49) 268(14)
2,3,5,6-Tetrachlorophenol–TMS 22.62 302 – 287(66) 289(86) 287(66) 93(100)
2,3,4,6-Tetrachlorophenol–TMS 22.94 302 – 287(78) 289(100) 287(78) 93(98)
2,3,4,5-Tetrachlorophenol–TMS 23.65 302 – 287(35) 289(47) 287(35) 93(100)
Pentachlorophenol–TMS 27.53 336 – 321(46) 323(72) 325(47) 93(100)
1,2,3,5-Tetrachloroxylene (ISTD) 21.46 – – – 207 209 244
2,4,6-Tribromophenol–TMS (SSTD) 25.29 400 – 385(35) 387(100) 389(99) 391(35)

3. Results and discussion
3.1. Derivatization method development
Standard solutions containing 7.5 × 10−4 mol dm−3 of each phe-
nol were prepared in dichloromethane, and injected into the
GC-FID. For derivatization, 100 ␮L of this solution was transferred
to a ReactiTherm reaction vial and 20 ␮L TMSDMC was added
(i.e., the silylating reagent was in 64-fold excess). The reaction
vial was hermetically sealed and kept at 50 ◦ C for 3 h. 2 ␮L of this
solution was subjected to GC-FID analysis. The chromatograms
of the silylated phenols and chlorophenols contained no peaks
of underivatized compounds, indicating 100% efficiency of the
silylation. The silylation reaction was carried out at room tem-
perature (25 ◦ C) as well, yielding the same chromatogram as the
heated sample: no heating was necessary for the derivatization
reaction to take place. The derivatized compounds were analyzed
on GC–MS in the full scan mode. The most important fragments
and their relative intensities in the EI mass spectra are shown in
Table 1.
The derivatization was successfully carried out in ethyl acetate
and acetonitrile as well. We then performed solid phase extrac-
tion with different eluents and silylated the extracts using 100-fold
reagent excess. The inclusion of the SPE step resulted in a dramatic
decrease of silyl derivatives when ethyl acetate was used as elu-
ent but not when dichloromethane was used. Since there had been
no difference between these solvents with respect to conversion,
we hypothesized that ethyl acetate absorbed significant amount
of humidity from the atmosphere during the extraction, which in
turn reacted instantaneously with the silylating agent. Underiva-
tized compounds were detected on GC–MS in the ethyl acetate
extract, corroborating this hypothesis. To reduce uncertainty intro-
duced into the method by moisture, we effected two changes in
the analytical scheme. First, elution was performed with the help
of rotary vacuum pump instead of water jet pump, which had been
used throughout the SPE process initially. Second, we decided to
perform the derivatization reaction with a 1:1 extract:reagent vol-
ume ratio. This enormous amount of silylating agent made the
method robust and provided an increased stability of 24 h for all
trimethylsilyl derivatives including PCP–TMS.

3.3. Overall detection limits

3.2. Ruggedness of the derivatization method

To assess the usefulness and applicability of the derivatization

A couple of days later partial hydrolysis of 2,3,5,6-
method, it was integrated into a complete analytical scheme and
tetrachlorophenol trimethylsilyl ether and almost complete hydrol-
tested for the trace-level analysis of the 25 compounds included
ysis of pentachlorophenol trimethylsilyl ether (PCP–TMS) was
observed, while all the other trimethylsilyl ethers had remained
stable. Therefore, stability studies were carried out on PCP–TMS. Table 2
To the solution of pentachlorophenol (7.5 × 10−4 mol dm−3 in Percent of silylated PCP in time, using different amounts of the silylating agent

dichloromethane), silylating reagent was added in stochiometric Time Stochiometric Amount of silylating agent
amount, and in 1:10 and 1:100 excess, respectively. Sampling
10× excess 100× excess
occurred in every 30 min and conversion was calculated from
absolute peak areas in the GC-FID chromatograms (Table 2). At 0.5 min 5.8 90.4 98.5
30 min 12.3 64.1 98.5
stochiometric amount of the reagent resulted in a very low reaction 60 min 18.5 31.0 98.5
rate, i.e., only 25% of PCP reacted after 90 min. At 10-fold reagent 90 min 25.4 12.6 98.3
excess, the reaction completed faster but all the trimethylsilyl- 120 min – 0.0 98.6
PCP hydrolyzed within 2 h, while 100× reagent excess provided 5h – – 96.0
24 h 0.0 – 95.2
instantaneous reaction and minimum 2 h stability.
142 Á. Kovács et al. / J. Chromatogr. A 1194 (2008) 139–142

Table 3 method has two distinguished features: TMSDMC instantaneously

Recoveries (n = 6) and limits of detection in standard solutions
silylates phenolic compounds at room temperature, and no
Compound Instrumental Recovery ± SD (%) LOD (␮g/L) further sample preparation step is necessary prior to gas chro-
LOD (ng/mL) matographic analysis. The trimethylsilyl derivatives are stable
Phenol 0.1 97 (±5) 0.005 for at least 24 h when applying 1:1 (v/v) extract:reagent vol-
o-Cresol 0.25 87 (±4) 0.0125 ume ratio. Preconcentrating 100 mL sample with SPE, detection
m-Cresol 0.25 76 (±16) 0.0125 limits were 0.01–0.25 ␮g/L for all compounds and <0.05 ␮g/L
p-Cresol 0.25 89 (±4) 0.0125
for 17 of them. Further reduction of the detection limits can
Catechol 0.0025 73 (±6) 0.0001
Resorcinol 0.0025 0.34 (±0.02) – be achieved by concentrating the SPE extract prior to deriva-
2-Chlorophenol 2.5 98 (±4) 0.125 tization, or by increasing the sample volume processed on
3-Chlorophenol 5 95 (±3) 0.25 SPE.
4-Chlorophenol 0.5 98 (±4) 0.025
2,3-Dichlorophenol 0.5 95 (±6) 0.025
2,4-Dichlorophenol 0.5 92 (±6) 0.025 References
2,5/2,6-Dichlorophenol 0.5 94 (±6) 0.025
3,4-Dichlorophenol 0.5 93 (±6) 0.025 [1] A. Geißler, H.F. Schöler, Water Res. 28 (1994) 2047.
3,5-Dichlorophenol 0.5 95 (±8) 0.025 [2] F. Bianchi, M. Careri, C. Mucchino, M. Musci, Chromatographia 55 (2002)
2,3,4-Trichlorophenol 0.5 96 (±7) 0.025 595.
[3] D. Puı́g, D. Barceló, Trends Anal. Chem. 15 (1996) 362.
2,3,5-Trichlorophenol 0.5 94 (±7) 0.025
[4] I. Rodrı́guez, M.I. Turnes, M.H. Bollaı́n, M.C. Mejuto, R. Cela, J. Chromatogr. A
2,3,6-Trichlorophenol 1.5 94 (±7) 0.075
778 (1997) 279.
2,4,5-Trichlorophenol 1 97 (±8) 0.05
[5] J. Drozd, Chemical Derivatization in Gas Chromatography., Elsevier, Amsterdam,
2,4,6-Trichlorophenol 1 95 (±9) 0.05 1981, p. 2.
3,4,5-Trichlorophenol 1 95 (±9) 0.05 [6] K.D. Buchholz, J. Pawliszyn, Anal. Chem. 66 (1994) 160.
2,3,4,5-Tetrachlorophenol 0.2 99 (±11) 0.01 [7] K. Abrahamsson, T.M. Xie, J. Chromatogr. 279 (1983) 199.
2,3,4,6-Tetrachlorophenol 0.2 99 (±11) 0.01 [8] T.J. Boyd, J. Chromatogr. A 662 (1994) 281.
2,3,5,6-Tetrachlorophenol 0.2 90 (±9) 0.01 [9] I. Rodrı́guez, M.I. Turnes, M.C. Cejuto, R. Cela, J. Chromatogr. A 721 (1996)
Pentachlorophenol 1 102 (±18) 0.05 197.
[10] K. Bielicka-Daszkiewicz, A. Völkel, M. Szejner, J. Osypiuk, Chemosphere 62
(2006) 890.
in our study. The analytical scheme consisted of three steps: (1) [11] N. Fattahi, Y. Assadi, M.R.M. Hossadi, E.Z. Jahromi, J. Chromatogr. A 1157 (2007)
23.
SPE; (2) derivatization with TMSDMC; (3) analysis of the deriva- [12] L. Farina, E. Boido, F. Carrau, E. Dellacassa, J. Chromatogr. A 1157 (2007)
tives on GC–MS. Table 3 compiles the SPE recoveries and overall 46.
detection limits. Processing 100 mL spiked water, recoveries were [13] A. Kunugi, K. Tabei, J. Chromatogr. 398 (1987) 320.
[14] S. Boitsov, S. Meier, J. Klungsøyr, A. Svardal, J. Chromatogr. A 1059 (2004)
higher than 75% with standard deviation of less than 20% for 24 ana- 131.
lytes. Resorcinol recovery was close to zero, which can be explained [15] H. Jobst, Acta Hydrochim. Hydrobiol. 23 (1995) 20.
by its extremely high water solubility (717 g/L [22]). Preconcentrat- [16] D.R. Knapp, Handbook of Analytical Derivatization Reactions, John Wiley &
Sons, New York, 1979, pp. 8–10.
ing 100 mL sample, phenols can be detected at the 0.1–10 ng/L level [17] K. Kim, H. Kim, J. Chromatogr. A 866 (2000) 87.
and chlorophenols at 10–75 ng/L, except for 2- and 3-chlorophenol [18] T. Heberer, H.-J. Stan, Anal. Chim. Acta 341 (1997) 21.
(125 and 250 ng/L). [19] D. Knausz, A. Meszticky, L. Szakács, B. Csákvári, K. Újszászy, J. Organomet. Chem.
256 (1983) 11.
[20] D. Knausz, A. Meszticky, L. Szakács, B. Csákvári, K. Újszászy, J. Organomet. Chem.
4. Conclusion 268 (1984) 207.
[21] L.A. Reitzel, A. Ledin, J. Chromatogr. A 972 (2002) 175.
A rapid and simple derivatization method was developed for 6 [22] Y.C. Fiamegos, C.G. Nanos, G.A. Pilidis, C.D. Stalikas, J. Chromatogr. A 983 (2003)
phenols and 19 chlorophenols using TMSDMC. The derivatization 215.