Journal of Chromatography A, 1343 (2014) 42–46

Contents lists available at ScienceDirect

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

Increasing the sensitivity of headspace analysis of low volatility

solutes through water removal by hydrate formation
Hui-Chao Hu a , Xin-Sheng Chai a,∗ , Chao-Hai Wei b , Donald Barnes b
a
State Key Laboratory of Pulp and Paper Engineering, South China University of Technology, Guangzhou, China
b
School of Environment and Energy, South China University of Technology, Guangzhou, China

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

Article history: This paper reports on the development of a new headspace analytical technique that is based on water
Received 7 February 2014 removal by hydrate formation (WRHF). By adding anhydrous salt, the liquid water in an aqueous sample
Received in revised form 24 March 2014 will be removed leaving behind volatile analytes that are fully vaporized at temperatures well below
Accepted 25 March 2014
their boiling points. With WRHF, the amount of sample in the headspace can be significantly increased,
Available online 3 April 2014
thereby dramatically improving the detection sensitivity. The technique reduces the risk of possible
column damage in gas chromatography (GC) systems. The technique was applied to the determination of
Keywords:
phenol at different stages of a coking wastewater treatment plant. The results showed that up to mL-levels
Phenol
Coking wastewater
of sample solution can be used in WRHF HS-GC analysis when 5 g of CaCl2 were used as the anhydrous
Water removal by hydrate formation salt. The detection sensitivity for phenol content was 500 times greater than that in earlier HS-GC work
Full evaporation that did not incorporate hydrate formation. The proposed WRHF headspace analysis technique is simple
Headspace analysis and practical, making it a useful tool for quantifying low concentrations of volatile analytes in aqueous
Gas chromatography samples.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction in FE HS-GC, however, is the low detection sensitivity associated

Headspace gas chromatography (HS-GC) is now playing an
important role in the quantification of volatile and semi-volatile
analytes in samples with complicated matrices; e.g., many types of
industrial samples [1–4]. Because the headspace-based analysis can
effectively prevent non-volatile species (e.g., salts and polymers)
from entering the GC system, it avoids many of the problems asso-
ciated with contamination, thereby reducing sample preparation
procedures. For example, the need for filtration and/or solvent-
extraction is eliminated in many cases [5–9].
Among the techniques used in HS-GC, the full evaporation (FE)
technique was developed as a means to sidestep the weakness
in other methods; i.e., the changing in sample matrices. Due to
the matrix-independent nature of FE, the calibration and analy-
sis of many difficult samples becomes much simpler. To meet the
requirement for the full evaporation [10–14], a very small aliquot
of liquid sample (␮L) is placed into a headspace vial and allowed to
equilibrate at an elevated temperature. One of the major problems
∗ Corresponding author. Tel.: +86 20 87113713; fax: +86 20 87113713.
E-mail address: xschai@gmail.com (X.-S. Chai).
with the small sample size.
Coupling a mass spectrometer (MS) to a GC system (i.e., GC–MS)
can significantly improve the system sensitivity, compared to
results with more conventional GC detectors, such as thermal con-
ductivity and flame ionization detectors. MS has the added benefit
of being a powerful technique for identifying the chemical struc-
ture of species involved. However, GC–MS is not suitable for use in
the FE analysis of aqueous samples, because the significant amount
of water vapor resulting from the full evaporation can damage the
MS detector, as well as deteriorate the separation performance of
GC column [15,16].
Liquid water combines with many inorganic salts to form solid
crystalline hydrates that retain the water even at elevated tem-
perature [17]. Thus conversion of the liquid water in an aqueous
sample into solid crystalline hydrate form using appropriate anhy-
drous salts could be a way to eliminate or minimize the negative
impact of water vapor in HS-GC measurements.
In the present work, we demonstrate a new technique based
on water removal by hydrate formation (WRHF) for use in HS-GC
analysis. The major focus is on optimizing conditions for WRHF,
including the selection of salts, the amount of salt added, and the
equilibration temperature and time. Phenol, a semi-volatile com-
pound with a high boiling point, was the model analyte in the study.

http://dx.doi.org/10.1016/j.chroma.2014.03.066
0021-9673/© 2014 Elsevier B.V. All rights reserved.
 H.-C. Hu et al. / J. Chromatogr. A 1343 (2014) 42–46 43

The results provide guidance for the application of the WRHF HS- 2.3. Sample preparation and measurement
GC method to the determination of other low volatility analytes in
aqueous solutions in future work. 2.3.1. Water vapor in headspace vial
1.0 g of anhydrous CaCl2 and differing amounts of water (0 mL,
0.1 mL, 0.2 mL) were placed and sealed into headspace vials. These
2. Experimental sample vials were equilibrated at three different equilibrium tem-
peratures (105, 130 and 150 ◦ C) for 60 min prior to HS-GC-TCD
2.1. Chemicals and samples analysis.

All chemicals used in the study, including anhydrous CaCl2 ,
anhydrous K2 CO3 , phenol and methanol were analytical grade and
obtained from commercial sources. A standard phenol solution
(4026 ppm) and methanol solutions (70.7 ppm, 2120 ppm) were
prepared respectively by adding a 4.0 g of phenol and 0.09 and
2.7 mL of methanol, respectively, to 1 L of distilled water. A set
of coking wastewater samples was obtained from Coke Plant of
Shaoguan Iron & Steel Group Co. Ltd. for use in a practical applica-
tion of the WRHF method.
2.2. Apparatus and operations
HS-GC measurements were carried out with an automatic
headspace sampler (DANI HS 86.50, Italy) and a GC system
(Agilent GC 7890A, US). The flame ionization detector (FID) mea-
surement system operating conditions were as follows: detector
temperature was 250 ◦ C; the flow rates of hydrogen and air
were 40 and 400 mL/min, respectively; a DB-5 capillary column
(30 m × 0.32 mm × 0.25 ␮m) operated with nitrogen carrier gas,
oven temperature at 60 ◦ C. The thermal conductivity detector
(TCD) measurement system operating conditions were as fol-
lows: detector temperature = 200 ◦ C, oven temperature = 120 ◦ C,
and a GS-Q capillary column (30 m × 0.53 mm, J&W Scientific,
US) operating with nitrogen carrier gas (flow rate = 10.3 mL/min).
Other operating conditions of the GC were as following: injector
temperature = 250 ◦ C; pressure of injection port = 20 psi; split-
less with purge time = 0.2 min. Headspace operating conditions
were as follows: equilibration temp. = 150 ◦ C; needle and samp-
ling coil temp. = 155 ◦ C; transfer temp. = 160 ◦ C; vial pressurization
time = 0.2 min; sample loop fill time = 0.15 min; and sample loop
volume = 0.5 mL.
2.3.2. Methanol measurement
6.0 g of anhydrous K2 CO3 were added to a headspace sample vial
containing 0.3 mL of 70.7 ppm methanol solution and was sealed
immediately with a PTFE/butyl septum and aluminum cap. 10 ␮L
of a 2120 ppm methanol solution was added to another headspace
sample vial (without salt) that was sealed with a PTFE/butyl sep-
tum and aluminum cap. The contents of these sample vials were
analyzed by HS-GC-FID after different equilibrium times at 105 ◦ C.
2.3.3. Determination of phenol in wastewater from a coking
operation
5.0 g of anhydrous CaCl2 and a small flask (2 mL) containing a
0.5 mL sample of wastewater were placed in a headspace sample
vial and sealed immediately with a PTFE/butyl septum and alu-
minum cap. The vial was placed in a laboratory oven at 130 ◦ C
for 6 h. Then, it was placed in the headspace sampler and allowed
to equilibrate at 150 ◦ C for 30 min with strong shaking. Finally, a
portion of the vapor in the headspace sample vial was withdrawn
and analyzed by GC-FID.
3. Results and discussion
3.1. Inorganic salts and their crystalline hydrates
Table 1 lists common inorganic salts, each with the chemical
formula and melting point of its salt hydrate [18]. It is clear that
these anhydrous salts can react with liquid water to form crystalline
hydrate(s) at temperatures below their melting points. Based on
the chemical formula, we can calculate the theoretical maximum
amount of water incorporated into each of the hydrates. Thus, we
can select the appropriate amount of salt to use. From the amount
of salt needed, the melting point of the resulting hydrate and the

Table 1
Properties of crystalline hydrates of common inorganic salts [18].

Anhydrous salt Hydrated salt Hydrates held in 1 g of anhydrous salt (g) Melting points (◦ C) Acid–base property*

K2 CO3 ·H2 O 0.130 273 b

K2 CO3
K2 CO3 ·1.5H2 O 0.196 235

Na2 CO3 ·H2 O 0.171 100 b

Na2 CO3
Na2 CO3 ·10H2 O 1.71 34

Na2 SO4 ·7H2 O 0.887 – n

Na2 SO4
Na2 SO4 ·10H2 O 1.27 32

CuSO4 ·H2 O 0.113 200 a

CuSO4
CuSO4 ·5H2 O 0.563 110

CaSO4 ·1.5H2 O 0.199 – n

CaSO4
CaSO4 ·2H2 O 0.265 150

MgSO4 ·H2 O 0.150 200 a

MgSO4
MgSO4 ·7H2 O 1.05 150

CaCl2 ·H2 O 0.162 260 n

CaCl2 CaCl2 ·2H2 O 0.324 175
CaCl2 ·6H2 O 0.973 30

LiCl LiCl·H2 O 0.429 98 b

CoCl2 ·2H2 O 0.277 – a

CoCl2
CoCl2 ·6H2 O 0.831 87
*
a: acidic; b: alkaline; n: neutral.
44 H.-C. Hu et al. / J. Chromatogr. A 1343 (2014) 42–46
Fig. 1. Diagram of sample preparation via the vapor contact mode.
volatility of the analyte, an appropriate salt can be selected for a
particular application. Note that, in general, highly hydrated salts
have lower melting points (just above room temperature) and are,
therefore, not suitable for use in headspace analysis. Further, the
acidity or alkalinity of the salts may affect the analysis significantly.
For example, phenol in an alkaline medium can form phenolate
salts which are unsuitable for HS analysis.
3.2. Water removal by hydrate formation
3.2.1. Modes of hydrate formation
It was observed that the liquid water in an aqueous sample can
be quickly incorporated into crystal hydrates by the mixing with an
excess of anhydrous salt. For most of salts, this reaction is exother-
mic, resulting in a rapid rise in temperature in the system which
may not be desirable. As an alternative, the hydrate can be formed
more slowly by reaction of the salt with water vapor, rather than
with liquid water, as shown in Fig. 1. In this case, the hydration
reaction is not accompanied by the rapid release of heat associated
with the reaction between the salt and liquid water. This mode is
particularly useful for analytes that decompose and/or otherwise
react at higher temperatures; e.g., the formation of phenolate salts.
3.2.2. Effects of equilibration time and temperature on WRHF
A certain period of time is required to remove all of the water
by hydrate formation. Fig. 2 shows the time-dependent release of
methanol from an aqueous sample (300 ␮L with 70.7 ppm) at 105 ◦ C
during hydrate formation. For comparison, the results of a full evap-
oration (FE) using the same amount of methanol but dissolved in
Fig. 2. Effect of equilibration time on the methanol release in WRHF and FE pro-
cesses.
only 10 ␮L of water (2120 ppm) are shown. It can be seen that
the complete release of methanol in the aqueous sample through
WRHF process can be achieved in 10 min, which is slower than that
through the FE process (6 min).
Since the equilibration time of the vapor contact mode for WRHF
was longer than the equilibration time of the liquid contact mode,
it is useful to conduct the equilibration at an elevated tempera-
ture (e.g., above the boiling point of water). Doing so will release
water vapor from the sample more rapidly, thereby reducing the
equilibration time for hydrate formation.
3.2.3. Water vapor in WRHF headspace system
Fig. 3 shows the amount of water vapor in headspace (quantified
by GC equipped with TCD) in presence of one gram of anhydrous
calcium chloride (using the liquid contact mode). Clearly, even if
the amount of liquid water introduced is below the salt’s theoreti-
cal maximum of water contact in the hydrates, detectable amounts
of water vapor remain in the headspace. This finding reflects the
vapor–solid phase equilibrium (VSE) between the water vapor and
solid hydrate. This VSE partitioning is also temperature-dependent.
Therefore, the selection of a lower headspace equilibration temper-
ature can minimize the water vapor that enters the GC system as a
result of sampling the vapor in the headspace.
Fig. 3 shows that using a 0.5-mL of headspace sample loop
with 2.0 bar of pressurized headspace sampling pressure, the water
introduced into GC system is less than 0.06 ␮L at 150 ◦ C when using
a 200 ␮L aqueous sample in the headspace analysis. This amount
of water vapor introduced into the GC system is lower than that
in the conventional headspace analysis (e.g., equilibrated at 80 ◦ C)
Fig. 3. The variation of water in vapor and water transferred to GC with sample size
at different equilibrium temperature.
 H.-C. Hu et al. / J. Chromatogr. A 1343 (2014) 42–46
Fig. 4. Saturated vapor pressure of analytes at various temperatures as derived from
[18].
and will have, therefore, a less negative impact on the GC column
and any MS detector that might be attached. [19–21].
3.3. Equilibration in the headspace vial
In headspace analysis, temperature is an important parameter in
sample equilibration. The equilibration temperature must be below
solvent boiling point (b.p.) in the conventional headspace analysis;
while it is usually above solvent’s b.p. in the FE headspace analysis.
The major difference between these two techniques is the liquid
sample size (volume) used in the testing. Full evaporation can only
be used with a small sample size, typically below 30 ␮L (depends
on the solid content) [10–14]. Otherwise, the increased pressure in
the vial will raise the boiling point of the contents to such an extent
that some of the material will remain in a liquid phase, including
some of the analytes.
In the present WRHF headspace analysis, the selection of equil-
ibration temperature depends mainly on the vapor pressure (a
function of temperature) of the analytes, because the water in the
test sample is almost completely eliminated by reaction with the
anhydrous salt. Fig. 4 shows the relationships of the vapor pressure
of methanol, phenol, and camphor with temperature [18]. Clearly,
higher temperatures lead to higher vapor pressures of the analytes,
resulting in higher concentrations in the vapor phase.
Fig. 5 shows the calculated percentage of methanol, phenol,
and camphor in the sample that appears in the vapor phase at
Fig. 5. Amounts of methanol, phenol and camphor in vapor at various temperatures.
45
Fig. 6. Effect of temperature on the time to reach equilibrium for phenol in a WRHF
system.
different temperatures. The calculation is based on the analyte’s
vapor pressure data at different temperatures (Fig. 4) when a 5 mg
sample of the analyte is placed in a 20-mL headspace vial. It can
be seen that the 100% vaporization of methanol can be achieved
at temperature that is a much lower than the greater-than-100 ◦ C
temperature required for methanol in an aqueous solution [13]. The
result also indicates that WRHF headspace technique can even be
used for camphor quantification analysis at a temperature of 110 ◦ C.
It should be pointed out that the water is being removed via
hydrate formation at the same time that the analyte is going into
the vapor state. As shown in Fig. 2, the analyte’s vaporization in FE
is much faster that the water removal process in WRHF. Therefore,
the time required for the WRHF analysis using the vapor contact
mode of hydrate formation is dominated by the time for hydrate
formation and equilibration.
3.4. Application to the determination of phenol content in coking
wastewater
Phenol is a toxic aromatic organic compound that is corrosive
to the eyes, the skin, and the respiratory tract in both liquid and
vapor form [22]. There are significant amounts of phenol in many
industrial wastewater streams, especially those associated with
the coking industry [23]. Therefore, phenol, a moderately volatile
impurity, was chosen as a model compound to evaluate the effec-
tiveness of a coking wastewater treatment plant in removing this
pollutant, and at the same time, test the utility of the WRHF method.
3.4.1. Selection of conditions for phenol analysis
In the experiments that follow, 5 g of CaCl2 were used along with
600 ␮L samples of wastewater from various stages of the waste-
water treatment process. As shown in Fig. 6, the time to complete
hydration formation (determined by the method shown in Fig. 1)
falls rapidly with increasing temperature. At temperatures above
the b.p. of water, the time to reach equilibrium is significantly
reduced.
In headspace analysis, we found that the phenol equilibrium in
the system can be achieved in 20 min at 150 ◦ C, much lower than
that of phenol’s b.p. (181.7 ◦ C), as shown in Fig. 7. Therefore, the
batch phenol analysis in the present method was first performed
at 130 ◦ C for 6 h in laboratory oven and then transferred to HS-GC
for analysis after an equilibration in headspace sampler at 150 ◦ C
for 20 min.
46 H.-C. Hu et al. / J. Chromatogr. A 1343 (2014) 42–46
Fig. 7. Time required for phenol equilibration in headspace sample vial at 150 ◦ C.
3.4.2. Method sensitivity in phenol detection by WRHF
A set of standard phenol aqueous solutions (40–400 ppm) were
prepared and a 500 ␮L sample of each solution was place a
headspace sample vial containing 5.0 g of anhydrous CaCl2 . After
hydrate formation was facilitated by 6 h in a 130 ◦ C oven, the vapors
in the vials were analyzed by HS-GC as described above. The results
of these measurements from a standard addition calibration were
found to fit the following linear equation:
A = 14.7(±0.1)C + 763(±7) (n = 7, r 2 = 0.999) (1)
where A is the GC signal response for phenol and C is the phenol
content (in ␮g) added in the spiked sample.
In this case, the limit of detection (LOD) was 1.5 ␮g. With the
WRHF technique, the sample size can be increased up to 2.5 mL
when using a 5.0 g of anhydrous CaCl2 , compared to only 5 ␮L in
the FE headspace analysis [14]. Thus, the detection limit for phe-
nol in the wastewater sample is reduced to 0.6 ppm at the given
conditions, which is much sensitive than 300 ppm when using FE
method.
3.4.3. Phenol contents at different stages in the treatment of
coking wastewater
At the coking wastewater treatment plant, there are seven
treatment stages, including flotation, adjusting, anaerobic diges-
tion, hydrolysis and sedimentation stages. Fig. 8 shows the phenol
Fig. 8. Phenol contents at the exit of each stage in the treatment of coking waste-
water.
concentrations at the exits of each of these stages, demonstrating
the significant reductions that occur in the adjusting and anaerobic
treatment stages. These data show that the water treatment plant
is effectively removing the phenol from the wastewater with major
reduction taking place at the adjusting stage and the anaerobic
stage. The data also show that WRHF method can yield timely and
useful information in a short time and at low concentrations.
4. Conclusions
A WRHF headspace analysis technique has been developed for
the determination of analytes at low concentrations in aqueous
samples. With WRHF, the sample size used in headspace analysis
can be significantly increased over previously used methods which
lead to a significant improvement in the detection sensitivity. By
minimizing the water content in headspace vapor, the WRHF tech-
nique could also used with a HS-GC system that is coupled to a mass
spectroscopic detector with minimal concern about damage to the
GC column or the MS detector.
The new method was successfully applied to the analysis of
phenol in the process streams of a coking wastewater treatment
plant. The results showed that the detection sensitivity in the phe-
nol quantification is much greater than that associated with the
earlier full evaporation HS-GC method. The present method is both
simple and practical, making it a useful tool for quantifying low
concentrations of analytes, especially for low volatility compounds
with high boiling point, in aqueous samples.
Acknowledgements
The authors acknowledge the Natural Science Foundation of
China (No. 21037001) and Research Fund for the Doctoral Program
of Higher Education of China (No. 20110172110026) for sponsoring
the research.
References
[1] B. Kolb, L.S. Ettre, Static Headspace-Gas Chromatography: Theory and Practice,
2nd ed., John Wiley & Sons, Hoboken, NJ, 2006.
[2] E. Pagliano, J. Meija, J.F. Ding, R.E. Sturgeon, A. D’Ulivo, Z. Mester, Anal. Chem.
85 (2013) 877.
[3] H.-C. Hu, Y.-X. Tian, H.-J. Jin, X.-S. Chai, D.G. Barnes, J. Agric. Food Chem. 61
(2013) 9362.
[4] E. Pagliano, J. Meija, R.E. Sturgeon, Z. Mester, A. D’Ulivo, Anal. Chem. 84 (2012)
2592.
[5] Q. Guan, H. Chen, P. Ning, S. Tian, J. Gu, H.-C. Hu, X.-S. Chai, Energy Fuels 27
(2013) 5135.
[6] H. Li, X.-S. Chai, M. Liu, Y. Deng, J. Agric, Food Chem. 60 (2012) 5307.
[7] H.-C. Hu, X.-S. Chai, J. Chromatogr. A 1320 (2013) 125.
[8] H.-C. Hu, H.-J. Jin, X.-S. Chai, J. Ind. Eng. Chem. 20 (2014) 13.
[9] X.-S. Chai, C.L. Verrill, Ind. Eng. Chem. Res. 50 (2011) 6413.
[10] J.-F. Zhong, X.-S. Chai, X.-L. Qin, S.-Y. Fu, Carbohydr. Polym. 86 (2011)
367.
[11] H. Li, X.-S. Chai, H. Zhan, S. Fu, J. Chromatogr. A 1217 (2010) 7616.
[12] H. Li, X.-S. Chai, Y. Deng, H. Zhan, S. Fu, J. Chromatogr. A 1216 (2009)
169.
[13] H. Li, H. Zhan, S. Fu, M. Liu, X.-S. Chai, J. Chromatogr. A 1175 (2007)
133.
[14] M. Markelov, J.P.J. Guzowski, Anal. Chim. Acta 276 (1993) 235.
[15] Agilent 5975 series MSD operation manual, G3170-90036, 4th ed., Agilent Tech-
nologies, Santa Clara, 2012.
[16] Maintaining your Agilent GC and GC/MS systems, 5989-7612EN, Agilent Tech-
nologies, Santa Clara, 2008.
[17] P. Patnaik, Handbook of Inorganic Chemicals, McGraw-Hill, New York, 2002.
[18] D.R. Lide, CRC Handbook of Chemistry and Physics, 86th ed., CRC Press, Boca
Raton, 2005.
[19] M. Serrano, M. Gallego, M. Silva, J. Chromatogr. A 1307 (2013) 158.
[20] I. Missaoui, L. Sayedi, B. Jamoussi, B.B. Hassine, J. Chromatogr. Sci. 47 (2009)
257.
[21] N. Sugaya, T. Nakagawa, K. Sakurai, M. Morita, S. Onodera, J. Health Sci. 47
(2001) 21.
[22] M.J. Neil, The Merck Index: An Encyclopedia of Chemicals, 15th ed., Drugs, and
Biologicals Merck and Co. Inc., Whitehouse Station, NJ, 2006.
[23] M. Ahmaruzzaman, Adv. Colloid Interface Sci. 143 (2008) 48.