Environ Monit Assess (2012) 184:1657–1669
DOI 10.1007/s10661-011-2068-9
Effect of process conditions on the analysis of free
and conjugated estrogen hormones by solid-phase
extraction–gas chromatography/mass
spectrometry (SPE–GC/MS)
Rominder P. S. Suri · Tony Sarvinder Singh ·
Robert F. Chimchirian
Received: 23 June 2010 / Accepted: 11 April 2011 / Published online: 5 May 2011
© Springer Science+Business Media B.V. 2011
Abstract Simultaneous analysis of 11 free estro-
gen hormones and five conjugated estrogens in
water and municipal wastewater was studied. The
analytical method was developed and tested for
different types of solid-phase extraction adsor-
bents, eluents, sample containers and storage con-
ditions, derivatization, and matrix effects. Varian
Bond Elut C-18 solid-phase extraction adsorbent
cartridge was selected based on its high recov-
eries for both free and conjugated estrogens.
Sample storage conditions, as well as selection
and pretreatment of sample container materials,
can affect the trace level analysis of estrogens.
Silanization of glassware is observed to provide
low relative standard deviation (RSD) in the
analysis and less percentage loss due to contacting
with sample container materials. Light exposure
during the test can significantly impact the results.
The derivatized samples stored at −20◦ C for at
least 6 days showed less than 10.5% average RSD
in the analysis. The recovery efficiency in clean
water varies from 72% to 101% for free estrogens
and 78% to 82% for conjugated estrogens. The
method detection limits (MDL) for most of the
compounds range from 30 to 870 ng/L using a
R. P. S. Suri (B) · T. S. Singh · R. F. Chimchirian
National Science Foundation - Water
and Environmental Technology (WET) Center,
Temple University, Philadelphia, PA 19122, USA
e-mail: rominder.suri@temple.edu
sample volume of 200 mL. With a sample volume
of 3 L, the most sensitive compound produces a
MDL of 0.03 ng/L. Dilute methanol is used to
wash the loaded cartridge as a cleanup step in
order to remove interfering species during analy-
sis of wastewater samples. Using the optimized
analytical methods, the concentration level of free
estrogens in the influent and effluent municipal
wastewaters is tested.
Keywords Estrogens · Free estrogens ·
Conjugated estrogens · Solid-phase extraction
(SPE) · Steroid hormones ·
Emerging contaminants
Introduction
The presence of steroid hormones in the environ-
ment is an issue of growing concern in the sci-
entific community (Halling-Sorensen et al. 1998;
Kolpin et al. 2002; Khanal et al. 2006). These
are endocrine-disrupting compounds and are ei-
ther produced naturally in the mammal body or
are synthetically created for use in birth control
pills and hormone replacement therapy (Halling-
Sorensen et al. 1998; Kolpin et al. 2002). The
primary sources of steroid hormones in the envi-
ronment include municipal wastewater treatment
plants (Johnson and Sumpter 2001; Ternes et al.
2004; Shore and Shemesh 2003), septic systems, as
1658
well as animal manure. Estrogens are largely ex-
creted from mammal bodies in conjugated forms
(as sulfate or glucuronide complexes). After en-
tering the wastewater treatment systems or nat-
ural ecosystems, these estrogen conjugates get
de-conjugated by bacteria and form free estro-
gens (Andersen et al. 2003; D’Ascenzo et al.
2003). In the environment, estrogens come in
contact with various forms of wildlife and could
produce unnatural changes. In studies on male
trout (Oncorhynchus mykiss), it has been shown
that exposure to estrogenic substances cause these
fish, and a variety of other species, to produce
a female-specific egg yolk protein (vitellogenin)
(Rodgers-Gray et al. 2000; Purdom et al. 1994).
17α-ethinyl estradiol (a synthetic estrogen) has
been shown to induce feminization in fish at levels
as low as 0.1 ng/L (Routledge et al. 1998; Larsson
et al. 1999). Sanderson et al. (2004) reported
that sex hormones show higher relative toxicity
than three other classes of pharmaceuticals includ-
ing antibiotics, antineoplastics, and cardiovascular
compounds.
There is a need to develop analytical methods
that can detect trace levels of estrogen hormones,
as well as their conjugates present in water and
wastewater to carry out studies on the fate, trans-
port, effects, and removal methods for these chem-
icals. The estrogen hormones have mostly been
analyzed by using solid-phase extraction (SPE)
followed with gas chromatography/mass spectro-
metry (GC/MS) or liquid chromatography/mass
spectrometry (LC/MS) (Belfroid et al. 1999; Lee
Ferguson et al. 2001; Lopez de Alda and Barcelo
2001; Xiao et al. 2001; Richardson and Ternes 2005;
Petrovic et al. 2002; Yamamoto et al. 2006;
Gabet et al. 2007). The recently published
analytical method (Method 1698) by the US
Environmental Protection Agency (USEPA)
involves liquid–liquid extraction, deivatization,
and high-resolution gas chromatograph/mass
spectrometer analysis (USEPA 2007).
The estrogen compounds are hydrophobic and
have strong sorption affinity. Many estrogen hor-
mones are light sensitive. There can be sorption
on the surface of the containers. Therefore, selec-
tion and pretreatment of sample container mate-
rials, as well as sample storage conditions, could
potentially affect the trace level analysis of estro-
Environ Monit Assess (2012) 184:1657–1669
gens. However, not much has been reported in
the literature regarding the analytical conditions
such as type of SPE adsorbents, effects of con-
tainer materials, sample storage, and glassware
preparation for steroid hormones. The affect of
background matrix in most environmental sam-
ples further complicates the analysis. The extrac-
tion efficiency and instrument chromatography
for hormones in wastewater samples may be
different from that in clean waters and may vary
significantly from sample to sample. For example,
the extraction efficiency in the influent waste-
water might be different from the effluent waste-
water sample due to different matrix effects.
Therefore, the calibration based on any specific
matrix (organic phase, clean water, or wastewater)
may not fit for different types of environmental
samples.
In this study, the effect of different type of SPE
adsorbents, sample container materials and their
treatment, and sample storage conditions were ex-
amined for a number of steroid hormones. The
eleven free and five conjugated steroid hormones
examined were: 17α-estradiol, 17β-estradiol, 17α-
dihydroequilin, 17α-ethinyl estradiol, estriol, estrone,
equilin, medrogestone, levonorgestrel, norgestrel,
gestodene, sodium α-estradiol 3-sulfate, sodium
β-estradiol 3-sulfate, sodium α-dihydroequilin 3-
sulfate, sodium estriol 3-sulfate, and sodium es-
trone 3-sulfate. The analytical method was tested
for analysis of steroid hormones in municipal
wastewater.
Experimental materials and methods
Chemicals and reagents
Free and conjugated hormones were obtained
from Sigma Aldrich, Steriloids Inc., and phar-
maceutical companies. The hormones (min-
imum purities) investigated were: 17α-estradiol
(98%), 17β-estradiol (97.1%), 17α-dihydroequilin
(99.4%), 17α-ethinyl estradiol (99.1%), estriol
(100%), estrone (100%), equilin (99.9%), medro-
gestone (99.8%), norgestrel (100%), levonorges-
trel (100%), gestodene (99.3%), and 3-O-methyl
estrone (internal standard, 98%). The conjugated
hormones, sodium α-estradiol 3-sulfate, sodium
Environ Monit Assess (2012) 184:1657–1669
β-estradiol 3-sulfate, sodium α-dihydroequilin
3-sulfate, sodium estriol 3-sulfate, and sodium es-
trone 3-sulfate were obtained from pharmaceuti-
cal company (purity data for conjugates hormones
was not available). All the solvents (HPLC grade),
chemicals, amber glass bottles, and Millipore ni-
trocellulose filters (47 mm white) were purchased
from Fisher Scientific. All the glassware such as
sample bottles, beakers, test tubes, GC inserts,
pipettes, etc. were silanized using a three-step
process described elsewhere (Suri et al. 2010). The
glassware was washed with soap water, rinsed with
hot water, and then with Milli-Q water. The glass-
ware was oven dried at 250◦ C for 3 h and rinsed
with a 5% dimethyldichlorosilane solution made
in toluene. It was then rinsed twice with toluene,
three times with methanol, and was air-dried at
250◦ C for 2 h. Mixed stock solutions of hormones
were prepared in methanol at 10 mg L−1 us-
ing a silanized amber glass volumetric flask and
stored at 4◦ C. From this stock solution, calibration
standards were prepared by spiking appropriate
amounts of stock solution to 200 ml Milli-Q wa-
ter (18.2 M cm, filtered to contain particles no
larger than 0.20 μm). All the experiments were
conducted with mixture of estrogen at initial con-
centration of 50 μg/L.
Municipal wastewater sampling
The influent and effluent wastewaters were col-
lected in triplicate in 4-L silanized amber glass
bottles with Teflon lined caps from a munici-
pal wastewater treatment plant in Southeastern
Pennsylvania, USA. Samples were stored at 4◦ C
until extraction was performed, for no more than
24 h.
Analytical method development
Solid-phase extraction
Free estrogens SPE experiments were performed
using six different adsorbents to test their ex-
traction efficiencies for free estrogens. The SPE
adsorbents tested were: Spec C-18 (Spec), Varian
Bond Elut C-18 (Varian), Waters Sep-pack C-18
(Waters), Phenomenex Strata-X 33 μm (Strata),
Supelco DSC-18 (DSC-18), and Supelco DSC-
1659
18LT (DSC-18LT). The adsorbent cartridges were
evaluated under identical conditions. Certain vol-
ume (0.2, 1, and 3 L) of sample containing 11 free
steroid hormones and the internal standard (3-O-
methyl estrone) was prepared in Milli-Q water for
SPE testing. The SPE cartridges were activated
using 3 mL methanol and rinsed with 3 mL Milli-
Q water before loading the sample. The sample
was then passed through the SPE cartridges at a
flow rate of 5 mL/min followed by 3 mL Milli-
Q water to rinse the cartridge. The adsorbates
were eluted from the cartridges with 3 mL eluent
into a clean test tube. The eluent was completely
dried using a Genevac EZ-2 vacuum evaporator
and then derivatized (described later) for GC/MS
analysis.
Conjugated estrogens The SPE adsorbents tested
for conjugated estrogens were: Spec C-18 (Spec),
Varian Bond Elut C-18 (Varian), Phenomenex
Strata-X 33 μm (Strata), Supelco DSC-18 (DSC-
18), Supelco DSC-PH (DSC-PH), and Supelco
DSC-CN (DSC-CN). The test conditions for con-
jugated estrogens were identical to that of free
estrogens (a mixed stock of conjugated estrogens
and a sample size of 200 mL). After loading with
the sample, the SPE cartridges were eluted, and
the eluent was transferred to a 35-mL centrifuge
tube for deconjugation.
Deconjugation procedure For deconjugation, the
eluent from the SPE was dried down to about
10 μL, and 15 mL acetate buffer (pH 5.2) and sul-
fatase enzyme equivalent to 2,500 U were added.
The sample was shaken in a rotary water bath
shaker for 20 min at 50◦ C. After agitation, 1 g of
barium chloride and 7 mL of toluene were added.
The tubes were capped and tumbled for 30 min.
The samples were then centrifuged, and the or-
ganic layer was placed in a separate, clean test
tube, and evaporated to dryness. The dried sam-
ple was then derivatized. The description of the
SPE procedure is also presented in Fig. 1.
Derivatization procedure
Pyridine (15 μL) and bis(trimethylsilyl)trifluoroa-
cetamide (65 μL) containing 1% trimethylchloro-
silane were added to the dried sample obtained
1660 Environ Monit Assess (2012) 184:1657–1669

Fig. 1 Flow diagram Water Sample (surface water, ground water,

showing the methods industrial, pharmaceutical or municipal
used for detection of free wastewater)

and conjugated estrogens

Filtration (0.45µm)

SPE extraction (Varian C18, 500 mg

cartridge)

Elution using 3 ml methanol

Free Estrogen Method Conjugated Estrogen Method

Evaporate to dryness Transfer eluent to a centrifuge tube and

evaporate

Derivatization using 15 µl of pyridine and De-conjugation with sulfatase enzyme in

65 µl of bis(trimethylsilyl)trifluoroacetamide acetate buffer

Reconstitute in 0.3 ml Toluene Liquid-Liquid extraction using toluene

Collect 3 ml toluene and evaporate to

dryness

Derivatization using 15 µl of pyridine and

65 µl of bis(trimethylsilyl)trifluoroacetamide

Reconstitute with 0.3 ml Toluene

from the SPE step. The sample was allowed to re-
act in a capped test tube for 15 min at 26◦ C. To the
derivatized sample, 0.3 mL of toluene was added.
The samples were vortexed and transferred into
an amber glass silanized GC vial.
Gas chromatography/mass spectrometry analysis
Sample analysis was performed using an Agilent
6890N GC coupled with a 5973N MS in SIM
mode. The splitless injections onto a Pursuit DB-
225MS (30 m × 0.25 mm × 0.25 μm, J & W Sci-
entific) capillary column was used with helium
(4.5 mL/min) as the carrier gas. The tempera-
tures of the injection port and ion source were
set at 240◦ C. The temperature of column was
programmed from 50◦ C (1 min) to 200◦ C at rate
of 50◦ C/min and held for 95 min. The oven tem-
perature was ramped to 220◦ C at 10◦ C/min and
held for 27 min. The post-run was held at 240◦ C
for 10 min. Initially the run started at 200◦ C
(column temperature) which produced good sep-
aration but the peak shape was not ideal. Decreas-
ing the initial column temperature from 200◦ C to
50◦ C did not affect the peak separation. The inlet
and source temperature was 240◦ C with a relative
source voltage of 1,447 V. The quad was set at
150◦ C. The method detection limit (MDL) was
defined as the lowest detectable and reproducible
analyte concentration for three successive sample
Environ Monit Assess (2012) 184:1657–1669 1661

Table 1 Summary of the Estrogen compound Retention Quantitation Confirming

retention times,
time (min) ion ions
quantitation and
confirming ions of steroid 17α-estradiol 25.0 416 285, 129
hormones 17β-estradiol 28.3 416 285, 129
17α-dihydroequilin 28.7 309 324, 310
17α-ethinyl estradiol 41.9 425 285, 300
Estriol 44.4 412 309, 505
Estrone 70.8 342 257, 343
Equilin 73.1 341 284, 342
Medrogestone 104.1 412 413, 397
Norgestrel/levonorgestrel 107.1 355 356, 73
Gestodene 108.6 341 117, 75

injections. No analytes were detected in proce- Results and discussion

dural blanks. Reported concentrations of analytes
were not corrected for recovery efficiencies. The
r2 values of calibration curves for all the com-
pounds were >0.99. Table 1 shows the quan-
tification and confirming ions of the analytes, as
well as the retention times under the optimized
GC/MS conditions. As shown in Table 1, lev-
onorgestrel and norgestrel are reported together
because they co-elute from the GC column and
have exactly the same mass to charge ratios. These
two compounds are structural isomers, and any
concentration data reported in this paper is the
sum of the two chemicals.
Certain samples were analyzed with and with-
out deconjugation steps. The analysis without
deconjugation step provided free estrogens, and
the analysis with deconjugation step provided
the total estrogen concentration. The difference
between total estrogens and free estrogens was
calculated and reported as the concentration
of conjugated estrogens. All experiments were
conducted in triplicates and their average values
are reported.
Selection of SPE adsorbent
The type of bonded hydrocarbon, type of deriva-
tization (mono or trifunctional), and the extent of
endcapping of the adsorbents can influence the
performance of SPE adsorbent on the extraction
of estrogens. The particle size and type of com-
mercially obtained SPE adsorbents are listed in
Table 2. Six out of eight SPE adsorbent cartridges
examined i.e. Spec, Varian, Strata, Waters, and
Supelco (DSC18 and DSC 18-LT), had C-18 sil-
ica adsorbent materials with different percentage
of carbon. DSC-PH and DSC-CN (Supelco) had
phenyl, florisil and cyanopropyl, alumina adsor-
bents, respectively, with 7% carbon.
Conjugated estrogens are polar in nature and
have basic sites and, hence, monomeric non end-
capped adsorbents, such as Spec C-18 and Strata
C-18, were expected to have higher capacity. On
the other hand, these adsorbates may be difficult
to extract due to their interaction with free silanols
on the adsorbent surface. Free estrogens are very

Table 2 Select Type of SPE cartridge Adsorbent Particle size (μm) End capping
characteristics of the SPE
cartridges and adsorbents Spec C-18; Silica 20 No
Varian C-18; Silica 47–60 Yes
Waters C-18; Silica 55–105 Yes
Strata C-18; Silica 33 No
DSC-18 (Supelco) C-18; Silica (18% C) 50 Yes
DSC-18LT (Supelco) C-18; Silica (11% C) 50 Yes
DSC-PH (Supelco) Phenyl; Florisil (7% C) 50 Yes
DSC-CN (Supelco) Cyanopropyl; Alumina (7%C) 50 Yes
1662
hydrophobic (Lai et al. 2000; Xia et al. 2005);
hence, polymeric coverage with endcapped ad-
sorbents, such as Varian C-18, Waters C-18, and
DSC-18 C-18, were expected to be favorable as
these would provide the maximum organic car-
bon phase on the adsorbent for the analyte to
partition.
The SPE cartridges were evaluated by deter-
mining the average recovery efficiency for estro-
gens, as shown in Fig. 2. In these tests, methanol
was used as the eluent. It may be observed from
Fig. 2 that the recovery of free estrogens was
highest with DSC-18 (92%), Varian (87%), Wa-
ters (77%), and Spec (67%) adsorbents. The re-
covery of conjugated estrogens was highest with
Varian (82%) and DSC-18 (62%) adsorbents.
Although the Varian cartridge provided slightly
lower recovery than DSC-18 for free estrogens,
overall, the Varian adsorbent may be considered
as the best with high extraction efficiencies for
free and conjugated estrogens. The phenyl- and
cyanopropyl-based adsorbents were not suitable
for extraction of conjugated estrogens. In the case
of Strata and DSC-18LT C-18 adsorbents, either
the adsorption of free estrogens and/or the sol-
vent extraction of free estrogens was poor. The
non-end capped adsorbents, Spec C-18 and Strata
C-18, provided low recoveries for free and conju-
gated hormones. This may be due to the strong
interaction of hormones with the free silanols on
the adsorbent surface and subsequent low extrac-
tion of adsorbates by the solvent.
100
80
Average recovery %
60
40
20
0 fering species were extracted when tested with
Spec Varian Waters Strata DSC-18
Fig. 2 Average recovery of various SPE cartridges tested
for free and conjugated estrogen extraction. DSC-18LT
was only tested for free estrogens and DSC-PH and DSC-
CN were tested with only conjugated estrogens
Environ Monit Assess (2012) 184:1657–1669
The Varian Bond Elut C-18 adsorbent was
selected for further study for SPE of free and
conjugated estrogen hormones. High estrogen re-
coveries of 87% and 82% were observed for free
and conjugated estrogens, respectively, and can
be attributed to its physical chemical character-
istics and surface functional groups. The Varian
Bond Elut C-18 is a non polar endcapped adsor-
bent that contains irregularly shaped acid-washed
mesoporous silica (47–60 μm) as base material
with 60 Å mean porosity. It contains 17.4% carbon
with trifunctional octadecyl function group. The
potential for polar interactions between analytes
and sorbent is less significant with C18 because of
the predominant effect of the long hydrocarbon
chain. This Varian Bond Elut C-18 has direct
Si-C linkages between the surface and the organic
trifunctional octadecyl function group. In order
to minimize the Si–OH group, a secondary reac-
tion with a smaller silane type molecule follows
after the bonding of the larger primary organic
group and which is called endcaping. In end-
caping, smaller entities reach some of the unre-
acted silanols between the larger bonded groups.
Free silanols can strongly retain compounds with
polar functional groups but makes elution more
difficult. However, with the presence of organic
functional groups attached, analytes elute more
easily using solvents.
Effect of eluent
Different eluents were evaluated for analyte re-
covery from the Varian Bond C-18 SPE adsor-
bent. The five eluents examined were: toluene,
methanol, hexanol, acetonitrile, and acetone. An
Free
efficient eluent would produce high recoveries
and have relatively short evaporation time to fur-
Conjugated
ther concentrate it. Among the tested eluents,
methanol and toluene showed high extraction
efficiencies for all the analytes (data not shown
here). Methanol provided the highest recovery for
Milli-Q water sample. However, significant inter-
DSC- DSC-PH DSC-CN
18LT the wastewater samples. The interfering species
caused poor reproducibility and loss of peak clar-
ity. In order to solve this problem, the cartridge
was washed with 3 mL Milli-Q water/methanol
(60/40, v/v) solution two times as a cleanup step
Environ Monit Assess (2012) 184:1657–1669 1663

to remove the interfering species. Loss of target
analytes during the pre-washing step was investi-
gated in separate experiments, and was observed
to be negligible. After washing, the cartridge
was eluted with 3 mL of pure methanol to ex-
tract the analytes. Removal of interfering species
also allows a more complete derivatization. Good
chromatographic peaks with higher peak areas
and reproducibility were observed on the GC/MS
chromatograms for the samples using the washing
steps.
In order to examine whether the size of one
SPE adsorbent column was sufficient, two SPE
cartridges were used in series to determine the loss
of analytes from the first column. The second car-
tridge would adsorb the analytes from the effluent
of the first cartridge. Each cartridge was eluted
separately and then extraction was performed for
estrogen hormone analysis. It was observed that
there was less than 1% analytes adsorbed on the
second column. More than 99% of analytes were
adsorbed by the first column. Hence, one SPE
column was sufficient to adsorb the estrogens.
Effects of sample container material: silanization
and sample storage
Due to the low solubility and high Log Kow (Lai
et al. 2000; Xia et al. 2005; Chimchirian et al.
2007), the estrogens hormones will have high
sorption tendency for the container surface. Since
estrogens are typically present at very low concen-
trations (nanograms per liter) in natural waters,
their loss due to sorption on the container material
during sampling, storage, transport, and analytical
procedures needs to be evaluated. The effect of
different sample containers (materials) and the
effect of glassware silanization on the analytical
reproducibility were studied by determining the
relative standard deviation (RSD) and percentage
loss. The container materials examined were glass
(G), high density polyethylene plastic (HDPE),
low-density polyethylene (LDPE), polypropylene
(PP), polypropylene light (PP light), and silanized
glass (SG). Silanization was performed to bond
the free silanols on the glass surface and prevent
the sorption of the analytes. It may be observed
from Table 3 that samples from SG container had
RSD of about 0.3% to 5.6% with an average of
3.2% and had the best overall performance. This
could be due to the coating of the silanization
agent on the glassware creating a non-polar film
which covers the free silanols and does not allow
estrogens to bind to the glassware. The HDPE
containers had RSD ranging from about 2% to
80%, for different compounds with an average
RSD of 16.7%. PP container performance was
comparable with G, and the average RSD value
observed was about 13%. The LDPE bottle had
the average RSD of about 15%.

Table 3 RSD of free estrogen hormones with different container materials (n = 3)

Estrogen compound RSD %
SG G HDPE LDPE PP
17α-estradiol 4.2 3.2 2.5 8.2 3.0
17β-estradiol 4.4 4.2 2.1 10.1 4.9
17α-dihydroequilin 5.2 30.1 6.1 17.4 26.8
17α-ethinyl estradiol 3.9 3.4 17.1 10.9 5.8
Estriol 0.5 3.3 3.5 8.2 7.2
Estrone 4.1 3.2 15.8 9.0 7.1
Equilin 5.6 24.3 5.3 28.9 19.3
Medrogestone 0.3 9.0 NA NA NA
Norgestrel/levonorgestrel 2.5 27.8 80.0 25.1 27.2
Gestodene 1.6 13.2 18.0 17.3 15.5
Average RSD 3.2 12.2 16.7 15.0 13.0
These chemicals were not detected in these samples and the RSD could not be determined
SG silanized glass; G glass; HDPE high-density polyethylene plastic; LDPE low-density polyethylene; PP polypropylene;
NA not available
1664 Environ Monit Assess (2012) 184:1657–1669

Table 4 Percentage loss of free estrogen hormones from different container materials as compared to the silanized glass
bottle
Compound SG G HDPE LDPE PP PP light GS
17α-estradiol 0 4.8 −0.3 15.1 1.8 6.8 14.3
17β-estradiol 0 4.8 −1.2 16.7 −0.8 9.6 16.0
17α-dihydroequilin 0 19.8 −20.5 44.4 −7.3 21.2 17.8
17α-ethinyl estradiol 0 1.3 14.9 32.0 11.8 7.5 −13.3
Estriol 0 5.1 0.4 19.7 2.0 10.3 5.3
Estrone 0 10.6 3.0 4.2 −1.3 6.5 33.1
Equilin 0 24.1 4.4 50.4 9.2 23.2 27.4
medrogestone 0 −15.4 100 100 100 N.A. −8.1
norgestrel/levonorgestrel 0 10.4 −18.9 39.6 19.4 100 −3.5
gestodene 0 27.8 9.6 47.9 18.1 35.3 20.6
Average loss 0 9.3 9.2 37.0 15.3 24.5 11.0
The data reported is average of three tests. The negative data in this table is because the analyte concentration was higher
than the control (Silanized glass bottle)
SG silanized glass; G glass; HDPE high-density polyethylene plastic; LDPE low-density polyethylene; PP polypropylene;
PP light polypropylene light; Gs glass scratched (internal surface); N.A. not available

Estrogen solution was placed in different, new wear on the surface. Regardless of the degree of
sample bottles and stored in the dark at 4◦ C for scratches or the age of glass bottle, the silanization
a period of 1 week. After 1 week, the percentage agent will bind the nonpolar layer to the glass
loss of estrogens for different bottle materials was surface and create a new homogeneous and con-
determined, and the data are shown in Table 4. sistent surface every time.
LDPE showed the highest average loss of about
37% followed by PP and PP-light bottles with Effect of storage time of derivatized samples
average loss of 15% and 25%, respectively. The
G and HDPE bottles showed comparable average A 6-day test was conducted to examine whether
loss of about 9%. The HDPE bottles had four an- the derivatized samples would be stable when
alytes which produced a better response i.e. lower stored for analysis. One derivatized sample was
loss than SG bottle, and they were: 17α-estradiol divided into seven vials, and one of these samples
(0.3%), 17β-estradiol (1.2%), 17α-dihydroequilin was analyzed on the GC–MS every day for the 6-
(20.5%), and levonorgestrel/norgestrel (18.9%). day test period. It may be observed from Table 5
However, medrogestone showed a 100% loss. The that the RSD (n = 7) for all analytes ranged
worst performance was observed from the LDPE
bottle with losses ranging from 4% to 100%.
Table 4 also shows the percentage loss of es- Table 5 Effect of storage time of derivatized samples;
trogen hormones from a new glass bottle (GS ) observed percent RSD (n = 7) of a six day stability test
that was scratched (internal surface) with a sand Estrogen compound RSD (%)
paper. This was done to simulate used glass bot- 17α-estradiol 7.5
tle which would get scratched over time during 17β-estradiol 7.3
normal working i.e. washing and cleaning using a 17α-dihydroequilin 4.5
17α-ethinyl estradiol 10.1
brush. The hypothesis was that scratching the glass
Estriol 7.7
may expose some active sites on the surface for Estrone 13.3
sorption of estrogens. The losses on the scratched Equilin 7.7
glass when compared to the silanized glass ranged Medrogestone 13.7
from a gain of 13% to a loss of 33%. Overall, the Norgestrel/levonorgestrel 16.8
silanization process has advantages when consid- Gestodene 15.8
ering the length of use of sample bottles and the Average 10.5
Environ Monit Assess (2012) 184:1657–1669 1665

Table 6 Loss of free estrogen hormones due to exposure

to ambient light were covered with aluminum foil to prevent light
exposure. As shown in Table 6, the losses due to
Estrogen compound Analyte loss (%)
light exposure ranged from 6% to 100%. Medro-
17α-estradiol 15.7
17β-estradiol 20.4
gestone was not detected in the bottle that was
17α-dihydroequilin 35.0 kept under the light. This study shows that the
17α-ethinyl estradiol 6.9 estrogen chemicals are sensitive to prolonged light
Estriol 18.7 exposure. Ultraviolet photo-induced degradation
Estrone 18.0 of estrogens has been widely reported (Coleman
Equilin 25.0 et al. 2004). Hence, exposure to ambient light
Medrogestone 100.0 could affect the analysis of estrogen hormones and
Norgestrel/levonorgestrel 29.0
should be avoided.
Gestodene 34.4
Average loss 30.3

Percent recoveries, detection limits, and RSD

from 4% to 17%. The derivatized analytes of
17α-estradiol, 17β-estradiol, 17α-dihydroequilin,
estriol, and equilin were more stable with RSD of
less than 10%, for the storage period of 6 days.
Effect of light exposure
The percentage loss of estrogens due to light
exposure during storage was examined in the lab-
oratory. The estrogen hormone solution was pre-
pared using sterilized water and placed in SG
bottles. The first set of three bottles was stored
in the dark at 4◦ C and the second set of three
bottles was placed on the laboratory windowsill
(under ambient temperature of 20◦ C and natural
light irradiation through the window glass) for
a period of 1 week. The bottles stored at 4◦ C
Table 7 reports recovery efficiencies, percent
RSD, and MDL for analyte compounds using
Varian bond C-18 SPE adsorbent cartridge and
methanol as eluent. The test was performed using
SG bottles. Recoveries ranged from 81% to 101%
for all of the free analytes, with the exception
of medrogestone that had a recovery of 72%.
Recoveries of conjugated estrogens ranged from
78% to 82%. Lower recoveries for conjugated
estrogens could be attributed to their higher sol-
ubility in water. Percent RSD for all the estrogen
compounds vary between 0.3 and 2.5. As may
be observed from Table 7, the MDL for all free
estrogen compounds ranged from 30 to 7,900 ng/L
using a sample size of 200 mL. When the sam-
ple volume was increased to 3,000 mL the most
sensitive compound produces a GC–MS response

Table 7 Observed accuracy (recovery), precision (%RSD). and method detection limits (MDL) for different solution
volumes
Estrogen compound Recovery Percent MDL MDL MDL
(%) (%) RSD 200 mL 1000 mL 3000 mL
sample (ng/L) sample (ng/L) sample (ng/L)
17α-estradiol 97.9; 81.3a 1.1 30; 20a 0.10 0.03
17β-estradiol 97.1; 82.1a 0.4 30; 20a 0.10 0.03
17α-dihydroequilin 99.5; 79.7a 0.4 30; 200a 1.00 0.28
17α-ethinyl estradiol 86.3 2.2 30 0.10 0.03
Estriol 95.1; 78.2a 0.9 30; 7900a 0.10 0.28
Estrone 101.1; 78.0a 1.4 30; 20a 0.10 0.03
Equilin 99.8 0.6 3990 10.0 3.21
Medrogestone 71.9 0.3 7900 10.0 0.28
Norgestrel/levonorgestrel 80.9 2.5 870 10.0 0.28
Gestodene 95.4 1.6 870 10.0 1.28
The RSD was calculated based on three sample analysis
a Conjugated forms
1666
at 0.03 ng/L. The conjugated estrogens showed a
detection limit of 20 ng/L with 200 mL sample
size, except that sodium estriol 3-sulfate had a rel-
atively higher MDL value of 7,900 ng/L. Kuch and
Ballschmiter (2001) reported LODs for estrone
(E1), 17α-estradiol (αE2), 17β-estradiol (βE2),
and ethynyl estradiol (EE2) using sample volume
of 1,000–5,000 mL (drinking water) in the range
of 0.05–0.10 ng/L. Quintana et al. (2004) reported
MDL for diethylstilbestrol, estrone, 17β-estradiol,
mestranol, 17α-ethinylestradiol, and estriol in the
range of 2 to 6 ng/L using a sample (Milli-Q water)
volume of 2,000 mL.
Wastewater sample analysis
Significant interfering species were extracted from
the wastewater samples when methanol was used
as the eluent for the SPE process. The interfering
species caused poor reproducibility and loss of
peak clarity on the chromatogram. This problem
was solved by washing the cartridge first with
dilute methanol solution (40:60), and then eluting
it with pure methanol. The municipal wastewater
samples were pre-filtered in a step-down method
using 8.0, 0.8, and 0.45 μm nitrocellulose filter to
prevent clogging of the SPE cartridges. Control
experiments conducted with estrogen solution in
Milli-Q water showed no significant loss of the
analytes by the filter paper. In the control test,
estrogen solution of 500 ng/L in Milli-Q water
was passed though the filter. The filtrate did not
show any significant difference in estrogen con-
centration when compared to the sample without
filtration.
The sample matrix could affect the extraction
and analysis of estrogen hormones. Therefore, a
modified analytical procedure was developed for
quantification of analytes in the wastewater. Oth-
ers have reported the matrix effects during the
analysis of pharmaceuticals in environmental sam-
ples (Vanderford and Snyder 2006; Vanderford
et al. 2003). In their approach, a solution of iso-
topically labeled standard that contained a stable
isotope of each analyte was spiked in the sam-
ple (Vanderford and Snyder 2006). This method
provides an accurate recovery and instrument re-
sponse track by using labeled surrogates or in-
ternal standards for each analyte. However, for a
Environ Monit Assess (2012) 184:1657–1669
large number of analytes in the solution, the iso-
tope dilution method can be very cost-intensive,
and moreover, the commercial availability of the
labeled standards is also limited at present. In this
study, the well-known standard addition method
was used. For each sample, the estrogen concen-
tration (C) was determined by spiking a known
amount of estrogens into the solution volume
(V). The extraction efficiency from spiked sample
would be similar to the unspiked samples. The
mass balance on analyte extraction of spiked and
un-spiked samples could be written as shown in
Eqs. 1 and 2.
(C × V × R) = m0 = k1 N0a (1)
(C + Cs ) × V × R = ms = k1 Nsa (2)
In Eqs. 1 and 2, C is the concentrations of analyte
in un-spiked sample whereas Cs is the correspond-
ing concentration in the solution due to the spiked
analyte amount (total concentration = C + Cs ).
The volume change due to the addition of the
analyte standard was negligible (0.01%). R is the
recovery efficiency and m0 and ms are the mass
of estrogens extracted from the un-spiked and
spiked samples, respectively. The values of m0 and
ms were obtained from the instrument standard
calibration curve using the normalized peak areas
(N0 for unspiked sample; Ns for spiked sample);
and kl and a are the correlation constants from
the instrument standard calibration. From Eqs. 1
and 2, the concentration of estrogens in the sam-
ple can be calculated using Eq. 3. The recovery
efficiency of the estrogens in the samples can be
obtained from Eq. 4.
m0 C s
C= (3)
ms − m0
m0 ms
R= = (4)
CV (Cs + C) V
Applications of the developed method
The developed SPE–GC/MS method for free
steroid hormones was applied for their determi-
nation in municipal wastewater. Tables 8 show
the analytical data for the estrogen hormones de-
tected in municipal wastewater. The samples were
spiked with known amount of analyte and recov-
Environ Monit Assess (2012) 184:1657–1669
Table 8 Estrogen hormone analysis for wastewater from a municipal wastewater treatment plant
Estrogen compounds Influent sample
Recovery (%) Concentration (ng/L)
17α-estradiol 94 ND
17β-estradiol 87 198.3 ± 30.2
17α-dihydroequilin 98 20.4 ± 1.1
17a-Ethinyl estradiol 117 16.4 ± 4.1
Estriol 72 709.8 ± 247.3
Estrone 88 49.2 ± 4.8
Equilin 65 ND
Spiked concentration was 50 ng/L for each analyte for the recovery data. All data are average of 3 (n = 3) samples. The
concentration data reported is with 90% confidence intervals
ND not detected
ery and efficiency was determined as explained
earlier. The recovery efficiencies and concentra-
tions of estrogens in the influent and effluent of
municipal wastewater are shown in Table 8. The
recovery efficiencies of estrogens in the influent
wastewater were in the range of 65% (equilin)
to 117% (17α-ethinyl estradiol). The recovery
efficiency for the effluent wastewater ranged from
97% (equilin) to 132% (17α-estradiol). It may
be observed from Tables 7 and 8 that the re-
covery efficiency for each estrogen compound in
the wastewater (effluent and influent) and Milli-Q
water were different. This variability is due to the
impact of background sample matrix. For exam-
ple, in case of 17α-ethinyl estradiol, the recovery
efficiencies were 86%, 117%, and 58% for Milli-
Q, influent wastewater, and effluent wastewater,
respectively.
Conclusions
Different process conditions were optimized to
develop an analytical method for extraction and
analysis of trace level of free and conjugated
estrogens in water and wastewater. Overall, the
Varian Bond C-18 Elut adsorbent was observed
to be the best with high extraction efficiencies
for free estrogens (average 87%) and conjugated
estrogens (average 82%). Among the tested elu-
ents, methanol and toluene showed high extrac-
tion efficiencies for all the analytes. Methanol
provided the highest recovery for Milli-Q water
sample. However, significant interfering species
were extracted when tested with the wastewa-
1667
Effluent sample
Recovery (%) Concentration(ng/L)
132 ND
116 ND
108 11.9 ± 0.5
58 ND
104 ND
123 ND
97 ND
ter samples. This problem was eliminated by a
cleanup step using Milli-Q water methanol (60/
40, v/v) solution to wash the SPE column prior to
extraction. The sample container material affects
the analytical reproducibility significantly. Sam-
ples from SG container showed the best overall
performance with an average RSD of 3.2%. Ex-
posure of the sample to ambient light results in
loss of analytes and should be avoided; derivatized
samples can be stored for certain period of time,
type of sample containers to be used, type of
SPE adsorbent, and type of eluent to be used.
Silanization of the glass container and storage in
the absence of light (dark) are recommended to
achieve high quality control. Derivatized analytes
(17α-estradiol, 17β-estradiol, 17α-dihydroequilin,
estriol, and equilin) were determined to be more
stable with RSD of less than 10%, for the storage
period of 6 days. Recoveries ranged from 81%
to 101% for all of the free analytes, with the ex-
ception of medrogestone (72%). Lower recoveries
of 78% to 82% were observed for conjugated es-
trogens, and which could be due to their higher
aqueous solubility. The MDL for all estrogen
compounds ranged from 30 to 870 ng/L using a
sample size of 200 mL with the exception of equi-
lin and medrogestone. With 1L sample volume,
the MDL for all the estrogen compounds ranged
from 0.1 to 10 ng/L. With 3 L sample volume,
corresponding MDL for all the estrogen com-
pounds ranged from 0.03 to 1.28 ng/L. Wastewater
samples showed a slightly higher RSD (2.7% to
11.1%) for the analytes. The analytical method
could be adapted for different types of water
and wastewater by appropriate selection of the
1668
method parameters such as sample size, cleanup
steps, pre-filtration, and spiking of analyte stan-
dard in each sample.
Acknowledgements Partial funding for this research was
provided by Wyeth Pharmaceuticals. Opinions, findings
and conclusions expressed in this paper are those of the
authors and do not necessarily reflect the views of Wyeth
Pharmaceuticals, Temple University Water and Environ-
mental Technology (WET) Center. Assistance provided
by Hongxiang Fu (post doc researcher) in the analysis is
acknowledged.
References
Andersen, H., Siegrist, H., Halling-SøRensen, B., &
Ternes, T. A. (2003). Scrutinizing pharmaceuticals and
personal care products in wastewater treatment. Envi-
ronmental Science Technology, 37, 4021–4026.
Belfroid, A. C., Van der Horst, A., Vethaak, A. D.,
Schafer, A. J., Rijs, G. B., Wegner, J., et al. (1999).
Analysis and occurrence of estrogenic hormones and
their glucuronides in surface water and waste water
in The Netherlands. Science of the Total Environment,
225, 101–108.
Chimchirian, R. F., Suri, R. P. S., & Fu, H. (2007). Free
synthetic and natural estrogen hormones in influent
and effluent of three municipal wastewater treatment
plants. Water Environmental Research, 79(9), 969–
974.
Coleman, H. M., Routledge, E. J., Sumpter, J. P., Eggins,
B. R., & Byrne, J. A. (2004). Rapid loss of estro-
genicity of steroid estrogens by UVA photolysis and
photocatalysis over an immobilised titanium dioxide
catalyst. Water Research, 38, 3233–3240.
D’Ascenzo, G., Di Corcia, A., Gentili, A., Mancini, R.,
Mastropasqua, R., Nazzari, M., et al. (2003). Fate of
natural estrogen conjugates in municipal sewage trans-
port and treatment facilities. Science of the Total En-
vironment, 320, 199–209.
Gabet, V., Miege, C., Bados, P., & Coquery, M. (2007).
Analysis of estrogens in environmental matrices.
Trends in Analytical Chemistry, 26, 1113–1131.
Halling-Sorensen, B., Nielsen, S. N., Lanzky P. F.,
Ingerslev, F., Lützhøft, H. C. H., & Jørgensen, S. E.
(1998). Occurrence, fate and effects of pharmaceu-
tical substances in the environment - A review.
Chemosphere, 36(2), 357–393.
Johnson, A. C., & Sumpter, J. P. (2001). Removal of
endocrine-disrupting chemicals in activated sludge
treatment works. Environmental Science Technology,
35, 4697–4703.
Khanal, S. K., Xie, B., Thompson, M. L., Sung, S., Ong,
S., & Van Leeuwen, J. (2006). Fate, transport, and
biodegradation of natural estrogens in the environ-
ment and engineered systems. Environmental Science
Technology, 40, 6537–6546.
Environ Monit Assess (2012) 184:1657–1669
Kolpin, D. W., Furlong, E. T., Meyer, M. T., Thurman,
E. M., Zaugg, S. D., Barber, L. B., et al. (2002). Phar-
maceuticals, hormones, and other organic wastewater
contaminants in U.S. streams, 1999–2000: A national
reconnaissance. Environmental Science Technology,
36, 1202–1211.
Kuch, H. M., & Ballschmiter, K. (2001). Determination of
endocrine-disrupting phenolic compounds and estro-
gens in surface and drinking water by HRGC-(NCI)-
MS in the Picogram per liter range. Environmental
Science Technology, 35, 3201–3206.
Lai, K. M., Johnson, K. L., Scrimshaw, M. D., & Lester,
J. N. (2000). Binding of waterborne steroid estrogens
to solid phases in river and estuarine systems. Envi-
ronmental Science Technology, 34, 3890–3894.
Larsson, D. G. J., Adolfsson-Erici, M., Parkkonen, J.,
Pettersson, M., Berg, A. H., Olsson, P. E., et al. (1999).
Ethinyloestradiol - an undesired fish contraceptive.
Aquatic Toxicology, 45, 91–97.
Lee Ferguson, P., Iden, C. R., McElroy A. E., &
Brownawell, B. J. (2001). Determination of steroid
estrogens in wastewater by immunoaffinity extrac-
tion coupled with HPLC-Electrospray-MS. Analytical
Chemistry, 73, 3890–3895.
Lopez de Alda, M. J., & Barcelo, D. (2001). Use of solid-
phase extraction in various of its modalities for sam-
ple preparation in the determination of estrogens and
progestogens in sediment and water. Journal of Chro-
matography A, 938, 145–153.
Petrovic, M., Eljarrat, E., Lopez de Alda, M. J., & Barcelo,
D. (2002). Recent advances in the mass spectrometric
analysis related to endocrine disrupting compounds
in aquatic environmental samples. Journal of Chro-
matography A, 974, 23–51.
Purdom, C. E., Hardiman, P. A., Bye, V. J., Eno, N. C.,
Tyler, C. R., & Sumpter, J. P. (1994). Estrogenic
effects of effluents from sewage treatment works.
Journal of Chemistry and Ecology, 8, 275–285.
Quintana, J. B., Carpinteiro, J., Rodríguez, I., Lorenzo,
R. A., Carro, A. M., & Cela, R. (2004). Determi-
nation of natural and synthetic estrogens in water
by gas chromatography with mass spectrometric de-
tection. Journal of Chromatography - A, 1024, 177–
185.
Richardson, S. D., & Ternes, T. A. (2005). Water analysis:
Emerging contaminants and current issues. Analytical
Chemistry, 77, 3807–3838.
Rodgers-Gray, T. P., Jobling, S., Steven, M., Kelly, C.,
Kirby, S., Janbakhsh, A., et al. (2000). Long-term
temporal changes in the estrogenic composition of
treated sewage effluent and its biological effects on
fish. Environmental Science Technology, 34, 1521–
1528.
Routledge, E. J., Sheahan, D., Desbrow, C., Brighty, G. C.,
Waldock, M., & Sumter, J. P. (1998). Identification
of estrogenic chemicals in STW effluent. 2. In vivo
responses in trout and roach. Environmental Science
Technology, 32, 1559–1565.
Sanderson, H., Brain, R. A., Johnson, D. J., Wilson,
C. J., & Solomon, K. R. (2004). Toxicity classification
and evaluation of four pharmaceuticals classes: An-
Environ Monit Assess (2012) 184:1657–1669
tibiotics, antineoplastics, cardiovascular, and sex hor-
mones. Toxicology, 203, 27–40.
Shore L. S., & Shemesh, M (2003). Naturally produced
steroid hormones and their release into the environ-
ment. Pure and Applied Chemistry, 75, 1859–1871.
Suri, R. P. S., Singh, T. S., & Abburi, S. (2010).
Influence of alkalinity and salinity on the sonochem-
ical degradation of estrogen hormones in aqueous
solution. Environmental Science Technology, 44,
1373–1379.
Ternes, T. A., Joss, A., & Siegrist, H. (2004). Scrutinizing
pharmaceuticals and personal care products in waste-
water treatment. Environmental Science Technology,
38, 393A–399A.
USEPA (2007) Method 1698: Of f ice of science and tech-
nology, engineering and analysis division. December
2007.
Vanderford, B. J., & Snyder, S. A. (2006). Analysis of
pharmaceuticals in water by isotope dilution liquid
chromatography/tandem mass spectrometry. Environ-
mental Science Technology, 40, 7312–7320.
1669
Vanderford, B. J., Pearson, R. A., Rexing, D. J., &
Snyder, S. A. (2003). Analysis of endocrine disruptors,
pharmaceuticals, and personal care products in water
using liquid chromatography/tandem mass spectrome-
try. Analytical Chemistry, 75, 6265–6274.
Xia, K., Bhandarim, K., Das, A., & Pillar, G. (2005). Occur-
rence and fate of pharmaceuticals and personal care
products (PPCPs) in biosolids. Journal of Environ-
mental Quality, 34, 91–104.
Xiao, X., 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 penta-
fluorobenzoyl derivatives. Journal of Chromatography
A, 923, 195–204.
Yamamoto, A., Kakutani, N., Yamamoto, K., Kamiura,
T., & Miyakoda, H. (2006). Steroid hormone profiles
of urban and tidal rivers using LC/MS/MS equipped
with electrospray ionization and atmospheric pressure
photoionization sources. Environmental Science Tech-
nology, 40, 4132–4137.