Journal of Environmental Sciences 19(2007) 879–884
Formation of multiple trimethylsilyl derivatives in the derivatization of
17α-ethinylestradiol with BSTFA or MSTFA followed by gas
chromatography-mass spectrometry determination
ZHOU Yi-qi1, WANG Zi-jian1,∗, JIA Ning2
1. Sate Key Laboratory of Environmental Aquatic Chemistry, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences,
Beijing 100085, China. E-mail: email@example.com
2. College of Environment and Energy Engineering, Beijing University of Technology, Beijing 100022, China
Received 25 August 2006; revised 15 September 2006; accepted 20 October 2006
N,O-bis(trimethylsily)trifluoroacetamide (BSTFA) and N-methyl-N(trimethylsily) trifluoroacetamide (MSTFA) are common deriva-
tization reagents used in the GC-MS analysis of estrogen steroids such as estrone (E1) and 17α-ethinylestradiol (EE2). In this study,
three trimethylsilyl (TMS) steroid derivatives, mono- and di-trimethylsilyl EE2 and mono-trimethylsilyl E1, were observed during the
derivatization of EE2 with BSTFA or MSTFA and/or GC separation. Factors influencing the production of multiple TMS derivatives
and their relative abundance were examined. It was found that both methanol and bisphenol A competed with estrogenic esteroids when
reacting with silylation reagents, and thus affected the formation of TMS derivatives and their relative abundance in the derivatization
products. Methanol was found to be more reactive than bisphenol A with the BSTFA reagent. None of the three solvents tested in
this study could prevent the generation of multiple TMS derivatives during the derivatization of EE2 with BSTFA, followed by GC
analysis. A similar result was observed using MSTFA as the derivative reagent followed by GC analysis. Thus, the suitability of
BSTFA or MSTFA as the derivatization reagent for the determination of E1 and EE2 by GC-MS, under the conditions reported here,
is questionable. This problem can be solved by adding trimethylsilylimidaz (TMSI) in the BSTFA reagent as recommended, and the
performance of the method has been proved in this study.
Key words: estrogens; water sample; GC/MS determination; trimethylsilyl steroid derivatives
Steroid estrogens of natural and anthropogenic origin
disrupting activities in both sewage effluent and surface
water (Lai et al., 2000; Gomes et al., 2004; Zuo et al.,
2004;Zhangand Zuo,2005). Thereforeitisveryimportant
to determine a low concentration (ng/L) of estrogens in
environmental water. Considering the low volatility of
steroid estrogens, LC/MS (Hu and Cheng, 2003; Hu et
al., 2005) and LC/MS/MS methods have recently been
developed for the determination of these estrogenic com-
pounds. However, at present, the application of LC/MS
and LC/MS/MS in environmental analyses of estrogenic
steroids appears to be limited because of their capital costs.
Traditionally, low concentrations (ng/L) of estrogenic
steroids in environmental water samples are determined by
gas chromatography-mass spectrometry (GC-MS), follow-
ing extraction and derivatization (Montserrat and Dami` a,
1997; Mol et al., 2000; Jones et al., 2000; Lee Ferguson et
(No. 2007CB407301) and Beijing Municipal Natural Science Foundation
(No. 8061004). *Corresponding author. E-mail: firstname.lastname@example.org.
al., 2001; Xiao et al., 2001; Helaleh et al., 2001; Orwa
et al., 2002; Kelly, 2002; Jeannot et al., 2002; Brossa
et al., 2002; Regan et al., 2002; Rodr´ ıguez et al., 2003;
Miao and Metcalfe, 2003; Shareef et al., 2004; Quintana
et al., 2004; Zuo et al., 2004, 2005; Zhang and Zuo,
2005). N,O-bis(trimethylsily)trifluoroacetamide (BATFA)
and N-methyl-N(trimethylsily) trifluoroacetamide (MST-
FA) have been widely used in the derivatization of
estrogenic steroids by reason of their facile and low
cost, which leads to the formation of the trimethylsilyl
(TMS) derivative. Catalysts such as trimethylchlorosi-
lane (TMCS), trimethyksilyimidazole (TMSI) or t-
butyldimethylsilylchlorosilane are usually added to en-
hance derivatization efficiency (Shareef et al., 2004). It
has been recently reported that di-trimethylsilyl (TMS)
derivative of 17α-ethinylestradiol (EE2) was formed by the
silylation reaction at both 3-OH and 17-OH, and partially
converted to its respective estrone (E1) derivative when N,
O-bis (trimethylsily)trifluoroacetamide (BSTFA) was used
to derivatize the synthetic estrogen EE2. Quintana et al.
(2004). observed the BSTFA reagent was able to react with
the aromatic hydroxyl group, but not with the aliphatic
hydroxyl group of EE2. Thus, a controversy arose about
880ZHOU Yi-qi et al. Vol. 19
the feasibility of the simultaneous determination of the
estrogens E1 and EE2 by BSTFA prior to column deriva-
tization and GC-MS. Recently, Zuo et al. (2004), reported
that using pyridine as a solvent can obtain di-trimethylsilyl
EE2 derivative as a sole trimethylsilyl derivatized product,
for the silylation of EE2 with the BSTFA reagent (Zuo
et al., 2004, 2005; Zhang and Zuo, 2005). However, in
the previous studies, various solvents and matrices were
involved in the analytical derivatization of EE2 and other
steroid compounds with MSTFA followed by GC-MS
measurement. In this study, a systemic investigation was
performed to examine the effects of solvent and matrices
on the derivatization of EE2 with BSTFA and MSTFA
reagents, including the formation and relative ratio of TMS
derivatives of EE2 and E1, which could be used in the
evaluation of the analytical results of EE2 and E1 obtained
in the previous studies, and to avoid potential problems in
the GC-MS analysis of estrogenic steroids. This research
attested that adding TMSI properly could produce the
di-trimethylsilyl EE2 derivative as a sole trimethylsilyl
derivatized product, for the silylation of EE2 with the
BSTFA or MSTFA reagent, using hexane as the solvent.
This finding provides an efficient analytical tool for the
future study of estrogenic steroids in the environment.
Estrone (E1), deuterated bisphenol A, TMSI, MSTFA,
and EE2 standards were purchased from Aldrich Chem-
istry Corporation (Milwaukee, WI). Anhydrous methanol,
acetone, ethylacetate, and hexane were supplied by Dima
(Richmond Hill, ON., Canada). BSTFA + TMCA (99:1,
v/v) was obtained from Supelco (Supelco Park, PA).
1.2 Standard solution
An individual standard solution of E1 and EE2 was
prepared at 5.0 × 10 mg/L and 1.00 × 102mg/L in
anhydrous methanol, from which appropriate dilutions
were made according to need. A stock solution of internal
standard, deuterated bisphenol A, was made at 2.00 × 102
mg/L in anhydrous methanol.
TMS derivatives of E1 and EE2 standards were respec-
tively prepared by the addition of anhydrous ethyl acetate
(100 µl) and BSTFA + 1% TMCA (100 µl) to a 2-ml amber
reaction vial containing 50.0 µg of standard, obtained by
evaporating 1.0 ml of the standard solution to dryness
under a low nitrogen flow. Then the vials were capped and
heated in an air bath at 65°C for 30 min. After cooling, the
products of derivative were analyzed directly by GC-MS
employing the SCAN mode.
1.3 GC-MS analysis
Derivatized samples were analyzed using an Agilent
GC-6890 Gas Chromatograph and 5973 Quadrapole Mass
Spectrometer equipped with a nonpolar HP-5MS 30 m ×
0.25 mm capillary column with 0.25 µm film (Agilent,
USA). The injector was set at 300°C, and the oven
temperature was programmed at 80°C for 1 min, ramped
at 20°C/min to 200°C, and then ramped at 10°C/min to
300°C and maintained at this temperature for 10 min.
The carrier gas was helium with a constant flow rate of
1 ml/min. The mass spectrometer was operated in the
electron impact ionization mode at 70 eV with an interface
temperature of 280°C and a source temperature of 230°C.
Positive fragment ions were analyzed over 50–500 m/z
mass range in the SCAN mode. All control of GC and
MS parameters and analysis of data were performed by the
MSD Productivity Chemstation Software Rev. D.00.00.
2 Results and discussion
2.1 TMS derivatives of E1 and EE2 resulted from
silylation with BSTFA + TMCA
Total ion chromatograms (TICs) of the TMS derivatives
of E1 and EE2 are given in Fig.1, and the mass spectra
from individual peaks are displayed in Fig.2. Peak 1 in the
TIC of E1 (Fig.1b) and Peak 1 of EE2 (Fig.1a) correspond
to the same retention time of 12.25 min and have identical
mass spectra as shown in Fig.2a. The mass spectra of peaks
2 and 3 in the TIC of EE2 are shown in Figs.2b and 2c.
The major ions for TMS E1 (Fig.2a) were the molecular
Fig. 1 GC-MS TICs of the derivatization products of (a) 17α-ethinylestradiol (EE2) without internal standard, (b) TMS estrone (E1) and the internal
standard (IS), deuterated bisphenol A.
No. 7 Formation of multiple trimethylsilyl derivatives in the derivatization of 17α-ethinylestradiol with BSTFA or MSTFA······
Fig. 2 Mass spectra. (a) mono-trimethylsilyl (TMS) estrone (E1),
(b) mono-TMS 17α-ethinylestradiol (EE2), and (c) di-TMS 17α-
ion at m/z 342 [M]+(base peak), 257 [M-85]+, and 327,
because of the loss of a methyl group from the derivative,
and an ion with m/z 218. Peak 2 in Fig.1a resulted from a
mono-trimethylsilyl derivative of EE2, which was formed
via silylation with the 3-OH group of EE2. The mass
spectrum of the mono-trimethylsilyl derivative of EE2
contained the molecular ion with m/z 368 [M]+, and an ion
with m/z 285 [M–83]+, because of the loss of [C3H5-OH]
and ethynyl group from [M]+on the D ring. A compound
of TMS reaction with both 3-OH and 17-OH of EE2
(Fig.2c), contained the molecular ion with m/z 440 [M]+,
and an ion with m/z 425 [M–15]+because of the loss of a
methyl group from the derivative, and an ion with m/z 285
[M–155]+, because of the loss of [C3H5–O–Si–(CH3)3]
and ethynyl group from [M]+on the D ring (Helaleh et
al., 2001; Shareef et al., 2004).
The identical retention times and mass spectra, together
with fragmentation patterns that match those expected for
E1 and EE2, indicate that TMS derivatives of EE2 include
two peaks, which are products of the TMS reaction, with
only 3-OH (Peak 2) and with both 3-OH and 17-OH (Peak
3) of EE2. EE2 derivatives are also partially broken down
into E1 derivative (Peak 1), during the derivatization with
BSTFA, or during chromatographic separation, or both
(Helaleh et al., 2001).
Previously, Quintana et al. (2004) observed that the
BSTFA reagent derivatized only the aromatic hydroxyl
group of EE2. Neither the di-TMS derivative of EE2
(Shareef et al., 2004) nor the derivative of E1 converted
from the EE2 derivative (Quintana et al., 2004), was
reported by them.
2.2 Effect of active hydroxyl on the silylation of EE2
Compounds containing active hydroxyl groups, such as,
methanol and bisphenol A, are frequently involved in the
determination of estrogens E1 and EE2 by BATFA pre-
column derivatization and GC-MS. To test the effect of
the active hydroxyl group in methanol and bisphenol A, on
silylation of EE2, 100 µg of EE2 standard, in the presence
of 200 µg deuterated bisphenol A or 20 µl methanol was
derivatized as described in Section 1.1. The influence of
active hydroxyl on the stability of TMS derivatives of
EE2 was also examined, by adding 200 µg deuterated
bisphenol A or 20 µl methanol into the reaction vial
after the derivatization of EE2, with BSTFA + 1% TMCS
reagent. As shown in Fig.3a, the addition of deuterated
the TMS derivatives formed or their relative peak areas,
when the chromatographic results were compared with
those obtained in the absence of deuterated bisphenol A
or methanol (Fig.1a). Three TMS derivatives, mono-TMS
E1, mono- and di-TMS EE2, were generated. However,
when EE2 was derivatized in the presence of deuterated
bisphenol A, only two derivatives, mono-TMS E1 and
mono-TMS EE2, were observed (Fig.3b). The stability
of these two TMS derivatives is shown in Fig.4. The
ratio of peak areas of mono-TMS E1 to internal standard
increases with the reaction time, but the ratio of peak areas
between mono-TMS EE2 and internal standard decreases
with reaction time. The relative abundance of mono-TMS
EE2 to mono-TMS E1 decreases rapidly. In a previous
study, Shareef et al. (2004) reported that mono-TMS E1
and di-TMS EE2 were the derivatization products because
less amount of bisphenol A was used in their experiment.
No TMS derivative of EE2 was found when the BSTFA
+ 1% TMCS reagent was added into the mixture of EE2
and methanol, because of the reaction of methanol with the
silylation reagent. Even though methanol was added after
the derivatization of EE2, the relative ratio of mono-TMS
E1, and mono- and di-TMS EE2 derivatives was altered.
This indicated that different types of hydroxyl groups had
different types of reactivity with the BSTFA + 1% TMCS
silylation reagent and affected the silylation of estrogenic
882 ZHOU Yi-qi et al. Vol. 19
Fig. 3 GC-MS TICs of trimethylsilyl derivative products of 17α-ethinylestradiol (EE2). (a) deuterated bisphenol A was added after TMS reaction with
EE2, and (b) deuterated bisphenol A was added at the same time with EE2.
Fig. 4 Stability of the TMS derivative products of EE2 in ethyl acetate
2.3 Effect of solvent on conversion of EE2 to the E1
Table 1 shows the relative peak areas (against the
internal standard) of mono-TMS E1 and mono-TMS EE2
in three solvents. Peak areas of mono-TMS EE2 versus
those of mono-TMS E1 are 0.23:0.16, 0.18:0.20, 0.30:0.18
in hexane, acetone, and ethyl acetate, respectively. The ten-
dency of conversion of the mono-TMS derivative of EE2
to the respective derivatives of E1 in the solvents studied
follows the order: acetone > hexane > ethyl acetate.
2.4 Solution for formation of multiple trimethylsilyl
Taking MSTFA as the derivative reagent, a result similar
to the above is observed. To solve the problem formation
of multiple trimethylsilyl derivatives, addition of several
kinds of reagents to the derivative reagent, such as triethy-
lamine and TMSI, has been tried. The result shows that
the reaction of triethylamine with BSTFA+1% TMCS or
MSTFA produces a white sediment. It can be seen from
Table 2, however, that adding 0.5% TMSI to BSTFA+1%
TMCS or MSTFA reagent produces di-trimethylsilyl EE2
derivative as a sole trimethylsilyl derivatized product for
the silylation of EE2, with the BSTFA reagent. The sily-
lation reaction is viewed as a nucleophilic attack upon the
silicon atom of the silyl donor, producing a bimolecular
transition state (Knapp, 1979). As a weak base, TMSI
can activate the hydroxyl groups and also serve as an
acid scavenger by removing the acidic product resulting
from the TMS derivatives, which make the derivatization
reaction of EE2 complete. Comparison of the sensitivity
of E1 and EE2, silylated by 0.5% TMSI in BSTFA +
1% TMCS and MSTFA in Table 3, demonstrates that it
is a good choice to take BSTFA + 1% TMCS + 0.5%
TMSI as the derivative reagent. Ratio of the peak area of
the derivative product of E1 with BSTFA + 1% TMCS +
Table 1 Relative peak areas (normalized against the internal standard) for E1 and EE2 after derivatization with BSTFA in different solvents
AnalyteHexane solvent Acetone solventEthyl acetate solvent
Peak No. in Fig.1 Relative peak
area ±SD (n)
Peak No. in Fig.1 Relative peak
area ±SD (n)
Peak No. in Fig.1Relative peak
Table 2 Percentage of mono-trimethylsilyl E1, mono-trimethylsilyl EE2, and di-trimethylsilyl EE2 using different derivatization reagents
Derivative reagentmono-Trimethylsilyl E1 (%)mono-Trimethylsilyl EE2 (%) di-Trimethylsilyl EE2 (%)
MSTFA + 0.5%TMSI
BSTFA + 1%TMCS
BSTFA + 1%TMC + 0.5%TMSI
No. 7Formation of multiple trimethylsilyl derivatives in the derivatization of 17α-ethinylestradiol with BSTFA or MSTFA······
Table 3 Comparison of the sensitivity for E1 and EE2 silylated by
different derivatization reagents (hexane solvent)
Derivative reagentE1 EE2
BSTFA + 1%TMCS + 0.5%TMSI
MSTFA + 0.5%TMSI
aRatio of derivative product of specific compound with BSTFA +
1%TMCS + 0.5%TMSI and MSTFA + 0.5%TMSI.
Table 4 Calibration curve for E1 and EE2 using BSTFA + 1%TMCS
+ 0.5%TMSI as derivative reagent (hexane solvent)
Y = 1885.7Xc
Y = 5226.8X
aAll data are mean of five in dependent assays;bIDL is instrumental
detection limit;cX is the concentration of specific compound.
0.5%TMSI, and MSTFA + 0.5%TMSI is 1.23, and that for
EE2 with BSTFA + 1% TMCS + 0.5%TMSI, and MSTFA
+ 0.5%TMSI is 1.20. When taking BSTFA + 1% TMCS
+ 0.5%TMSI as the derivative reagent, it can be seen from
Table 4 that a good linearity (R2> 0.99) is achieved in
the hexane solvent for the E1 and EE2 mixture standard.
Derivative reaction was performed under 65°C for 30 min.
Addition of TMSI in the derivative reagent prevents the
effective formation of multiple trimethylsilyl derivatives
and ensures availability of analysis result of EE2, by
using the similar pre-column derivative tandem GC-MS
method. Furthermore, the results in Table 4 have proved
that the approved method can be applied to simultaneous
determination of E1 and EE2.
Three trimethylsilyl steroid derivatives, mono-TMS E1
and mono- and di-TMS EE2, were generated in the deriva-
tization with BSTFA or/and subsequent GC separation.
Both compounds containing active hydroxyl groups, such
as methanol and bisphenol A, and derivatization solvents,
affect the formation of multiple TMS steroid derivatives
and their relative abundance in the silylation of EE2 with
the BSTFA reagent. Methanol is much more reactive with
BSTFA than bisphenol A. With each of the three solvents
tested in this study, more than one TMS steroid derivative
was produced in the derivatization of EE2; mono-TMS
E1 derivative was observed in every case studied. Taking
MSTFA as the derivative reagent, a similar result was
In the present study, addition of 0.5% TMSI effec-
tively prevents the formation of multiple trimethylsilyl
derivatives in the derivatization of EE2 with BSTFA or
MSTFA, followed by GC-MS determination. Developed
methods have testified that it can be applied to simultane-
ous determination of E1 and EE2. The linearity relativity
coefficients of the calibration curve for E1 and EE2 are
0.9979 and 0.9991 respectively. The result of this research
provides an efficient analytical tool for a further study of
estrogenic steroids in the environment.
Brossa L, Marc´ e R M, Borrull F et al., 2002. Application
of on-line solid-phase extraction-gas chromatography-mass
spectrometry to the determination of endocrine disruptors
in water samples[J]. J Chromatogr A, 963: 287–294.
Gomes R L, Avcioglu E, Scrimshaw M D et al., 2004. Steroid
estrogen determination in sediment and sewage sludge: a
critique of sample preparation and chromatographic/mass
spectrometry considerations, incorporating a case study in
method development[J]. Trends Analyt Chem, 23: 737–
Helaleh M I H, Takabayashi Y, Fujii S et al., 2001. Gas
chromatographic-mass spectrometric method for separation
and detection of endocrine disruptors from environmental
water samples[J]. Analytica Chimica Acta, 428: 227–234.
Hu J, Cheng S, 2003. Products of aqueous chlorination of
17α-estradiol and their estrogenic activities[J]. Environ Sci
Technol, 37: 5665–5670.
Hu J, Zhang H, Chang H, 2005. An improved method for
analyzing estrogens in water by liquid chromatography-
electrospray mass spectrometry[J]. J Chromatogr A, 1070:
Jeannot R, Sabik H, Sauvard E et al., 2002. Determination of
endocrine-disrupting compounds in environmental samples
using gas and liquid chromatography with mass spectrome-
try[J]. J Chromatogr A, 974: 143–159.
Jones L T, Gerlach C L, Cooter E J, 2000. The power of analyt-
ical methods for measuring suspected endocrine disrupting
compounds: a pilot field study[J]. Trends Analyt Chem, 19:
Kelly C, 2002. Analysis of steroids in environmental water sam-
mass spectrometry[J]. J Chromatogr A, 872: 309–314.
Knapp D R, 1979. Handbook of analytical derivatization reac-
tions[M]. New York: John References Wiley and Sons.
Lai K M, Johnson K L, Scrimshaw M D et al., 2000. Binding
of waterborne steroid estrogens to solid phases in river and
estuarine systems[J]. Environ Sci Technol, 34: 3890–3894.
Lee F P, Iden C R, McElroy A E et al., 2001. Determination of
steroid estrogens in wastewater by immunoaffinity extrac-
tion coupled with HPLC-Electrospray-MS[J]. Anal Chem,
Miao X S, Metcalfe C D, 2003. Determination of carba-
mazepine and its metabolites in aqueous samples using
liquid chromatography-electrospray tandem mass spec-
trometry[J]. Anal Chem, 75: 3731–3738.
Mol H G J, Sunarto S, Steijger O M, 2000. Determination of
endocrine disruptors in water after derivatization with N-
methyl-N-(tert.-butyldimethyltrifluoroacetamide) using gas
chromatography with mass spectrometric detection[J]. J
Chromatogr A, 879: 97–112.
Montserrat C, Dami B, 1997. Analysis of industrial effluents to
determine endocrine-disrupting chemicals[J]. Trends Ana-
lyt Chem, 16: 574–583.
of polymyxins B1, E1 and E2 in aqueous solution using
liquid chromatography and mass spectrometry[J]. Pharm
Biomed Anal, 29: 203–212.
Quintana J B, Carpinteiro J, Rodr´ ıguez I et al., 2004. Deter-
mination of natural and synthetic estrogens in water by
gas chromatography with mass spectrometric detection[J].
J Chromatogr A, 1024: 177–185.
884ZHOU Yi-qi et al. Vol. 19 Download full-text
Regan F, Moran A, Fogarty B et al., 2002. Development
of comparative methods using gas chromatography-mass
spectrometry and capillary electrophoresis for determina-
tion of endocrine disrupting chemicals in bio-solids[J]. J
Chromatogr B, 770: 243–253.
Rodr´ ıguez I, Quintana J B, Carpinteiro J et al., 2003. De-
termination of acidic drugs in sewage water by gas
chromatography-mass spectrometry as tert.-butyldime-
thylsilyl derivatives[J]. J Chromatogr A, 985: 265–274.
Shareef A, Parnis C J, Angove M J et al., 2004. Suitability
of N,O-bis(trimethylsilyl) trifluoroacetamidend N-(tert.-
butyldimethylsilyl)-N-methyltrifluoroacetamide as deriva-
tization reagents for the determination of the estrogens
estrone and 17α-ethinylestradiol by gas chromatography-
mass spectrometry[J]. J Chromatogr A, 1026: 295–300.
Xiao X Y, McCalley D V, McEvoy J, 2001. Analysis of estrogens
in river water and effluents using solid-phase extraction
and gas chromatography-negative chemical ionisation mass
spectrometry of the pentafluorobenzoyl derivatives[J]. J
Chromatogr A, 923: 195–204.
Zhang K, Zuo Y, 2005. Pitfalls and solution for simulta-
neous determination of estrone and 17α-ethinylestradiol
by gas chromatography-mass spectrometry after deriva-
tization with N,O-bis(trimethylsilyl)trifluoroacetamide[J].
Anal Chim Acta, 554: 190–196.
Zuo Y, Zhang K, Deng Y, 2004. Occurrence and photochemical
fate of hormone steroids in New Bedford Harbor sea-
water[C]. The 227th American Chemical Society Annual
Meeting. Anaheim, CA, March 28–April 1.
Zuo Y, Zhang K, Deng Yl, 2005. Occurrence and photochemical
degradation of 17α-ethinylestradiol in Acushnet river estu-
ary[J]. Chemosphere, 63: 1583–1590.