Accepted Manuscript

Title: Effect of halide ions on secondary electrospray

ionization–ion mobility spectrometry for the determination of
TNT extracted by dispersive liquid–liquid microextraction

Authors: Ahmad R. Jafari Horestani, Mohammad T. Jafari,

Elham Jazan, Mehdi Mossaddegh

PII: S1387-3806(18)30114-3
DOI: https://doi.org/10.1016/j.ijms.2018.08.006
Reference: MASPEC 15990

To appear in: International Journal of Mass Spectrometry

Received date: 28-4-2018

Revised date: 14-8-2018
Accepted date: 14-8-2018

Please cite this article as: Jafari Horestani AR, Jafari MT, Jazan E,
Mossaddegh M, Effect of halide ions on secondary electrospray ionization–ion
mobility spectrometry for the determination of TNT extracted by dispersive
liquid–liquid microextraction, International Journal of Mass Spectrometry (2018),
https://doi.org/10.1016/j.ijms.2018.08.006

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Effect of halide ions on secondary electrospray ionization–ion mobility spectrometry
for the determination of TNT extracted by dispersive liquid–liquid microextraction

Ahmad R. Jafari Horestania, Mohammad T. Jafaria,, Elham Jazanb, Mehdi Mossaddegha

a
Department of Chemistry, Isfahan University of Technology, Isfahan 84156–83111, Iran

T
b
Department of Chemistry, Islamic Azad University, Shahreza Branch, Shahreza, Isfahan 311-86145, Iran

IP

Corresponding author: Tel.: +98-311-391-2351, Fax: +98-311-391-2350

R
E-mail: jafari@cc.iut.ac.ir.

SC
Graphical Abstract

U
N
A
M

Highlights
ED

 Secondary electrospray ionization-ion mobility spectrometry was utilized for the

determination of TNT.

PT

KCl was selected as the most effective dopant in electrospray solution for signal
enhancement.
 The proposed method was used for determination of TNT, extracted from water samples by
E

DLLME.
CC

Abstract
In this study, secondary electrospray ionization source (SESI) was coupled with ion mobility
A

spectrometer (IMS) and utilized for the determination of 2,4,6-trinitrotoluene (TNT) in water and
wastewater samples. The sensitivity of SESI–IMS for TNT detection was compared with that
obtained by ESI–IMS. The effect of different halide ions (as dopants) on the signal enhancement

1
was evaluated and eventually KCl was selected as the most effective dopant. Sample preparation
step for the pre-concentration of TNT was performed by dispersive liquid-liquid microextraction
using an extraction solvent with the lower density than water. Methanol and toluene were selected
as the disperser and extraction solvents, respectively. Some other affected parameters on the
extraction efficiency; such as salt addition and sample acidity were also investigated. The linearity
was obtained 4.0–175.0 µg L-1 and the limit of detection and quantification were calculated 1.0 and

T
4.0 µg L-1, respectively. Tap, River, and laboratory wastewater samples were selected to evaluate

IP
the method capability for determining the analyte in real matrices, and satisfactory analytical results
were obtained.

R
SC
Keywords: Secondary electrospray ionization; Ion mobility spectrometry; Halide ions dopant;
Dispersive liquid-liquid microextraction; Trinitrotoluene

1. Introduction
U
N
Ion mobility spectrometry (IMS) is an efficient analytical method which is currently utilized
for detecting explosives, warfare agents, drugs, pesticides and some other chemical pollutants. The
A
main advantages of IMS are very low response time, high sensitivity, simplicity and low-cost [1,2].
M

The key part of IMS instrument is ionization source which provides the ionized compounds. In the
drift region (other section of instrument) ionized compounds can be separated based on their
ED

63
charge, size and shape. The most currently ionization sources have been used in IMS are Ni,
63
corona discharge (CD) and electrospray [3]. Ni is a radioactive source, so the regular leak test
may be required, before applying the instrument. CD and 63Ni are generally used for the analysis of
PT

samples in the gas phase. Although CD is known as highly sensitive ionization source, however,
solid or liquid samples should be firstly evaporated as a gas phase, prior to entering the ionization
E

region [4,5]. Therefore, non-volatile and/or low-thermal stability compounds cannot be analyzed by
CC

these ionization sources. Electrospray ionization (ESI) as an efficient ionization source has been
coupled with IMS and mass spectrometry (MS), for the analysis of explosives, drugs, etc [6–8].
A

ESI–IMS is a good alternative for the analysis of water samples and detection of non-volatile or
thermally unstable compounds. The main limitation in utilizing of ESI–IMS is incompetent
desolvation process, especially when the aqueous samples are sprayed. In contrast, the desolvation
process can be completely occurred, under the vacuum conditions in MS [9].
The most effective advantage of ESI is direct introduction of liquid samples including high
molecular weight compounds such as biomolecules for analyzing by IMS [10,11]. For

2
determination of samples in the liquid phase, different nonvolatile dopants can be used for
increasing the ionization efficiency. So far, some salts have been used as a dopant for the analysis
of explosives ionized by electrospray ionization [12–14]. Whereas, utilizing the gas-phase
63
ionization sources like Ni and CD restricts the advantage of using various types of chemical
dopants.
On the other hand, one of the main problems maybe occurred by ESI is non-sufficient

T
spraying for some organic solvents. This drawback decreases the sensitivity or even causes

IP
impossibility of analysis. In this case, direct and accurate analysis of target compound would be
impossible, due to existence of inappropriate solvents or other contaminants extracted from

R
complex matrices. To expand the ESI applications, Hill and co-workers in 2004 developed

SC
secondary electrospray ionization (SESI) as a strategy for the explosives detection [15]. In SESI,
the reactant ions are produced by spraying a solvent and subsequently react with the analyte

U
molecules to generate charged species. In that work, Hill’s group expected promotion of the
N
ionization efficiency using this strategy. In this regards, Wu et al. compared the sensitivity of SESI
and ESI sources to investigate some illicit drugs such as heroin, morphine and cocaine [16]. They
A
showed that SESI has more ionization efficient than ESI in analysis of test compounds. Up to now,
M

SESI has also been applied for analyzing some chemical warfare agent simulants [17] as well as
some explosives auch as pentaerythritol tetranitrate (PETN), nitroglycerin (NG) and cyclo-1,3,5-
ED

trimethylene-2,4,6-trinitramine (RDX) [15]. In addition to IMS, SESI has been used as the
ionization source in the mass spectrometry technique. In 2009, Martínez-Lozano et al. utilized SESI
in triple quadrupole MS for ambient vapor determination of PETN and TNT, at the concentrations
PT

below parts per trillion [18].

On the other hand, different microextraction methods have been developed in recent three
E

decades, for preconcentration and extraction of vast variety of compounds. Dispersive liquid-liquid
CC

microextraction (DLLME) is one of the most interesting methods of sample preparation has some
advantages of low consumption of solvent, short analysis time and high pre-concentration factor
A

[19]. In this method, two organic solvents including the disperser and extraction solvents are mixed
and injected into an aqueous sample solution. The disperser solvent, which should be miscible in
both aqueous and organic phases, increases the mass transfer rate, and so, decreases the analysis
time, significantly.
In this work, the main objective of this work is investigation of the effect of different halide
ions (as dopants) on the enhancement of the IMS signal using secondary electrospray ionization

3
compared with that obtained by conventional electrospray ionization. Potassium chloride was used
as the best dopant in electrospray solvent to enhance the ionization efficiency. In addition, this
paper makes a novel contribution in the combination of dispersive liquid-liquid microextraction
with SESI-IMS, has not been reported previously. In this regard, 2,4,6-trinitrotoluene (TNT) was
analyzed with SESI–IMS and the sensitivity of SESI was compared with that obtained by ESI.

T
IP
2. Experimental
2.1. Instrumentation

R
The ESI–IMS apparatus used in this work was constructed at Chemistry Department of

SC
Isfahan University of Technology which has been described, previously [20,21]. The injection port
based on thermal desorption was used for sample vaporization in SESI strategy. This introduction

U
system was also designed and constructed at Isfahan University of Technology [5]. The instrument
N
was equipped with a Rheodyne 7725 manual injector with a 10-µL loop for injecting the liquid
samples. The main parts of the instrument included an IMS cell, an electrospray needle, two power
A
supplies for supporting the high-voltages, a pulse generator, an analogue to digital converter and a
M

computer. The drift tube was made by 16 aluminum rings, separated from each other by a ring of
PTFE. In addition to electrospray region (3 cm), two other regions of cell were the desolvation
ED

region (4 cm in length, 2.1 cm i.d.) and the drift region (11cm in length, 4 cm i.d). The electrospray
needle (P/N 7768-01, Hamilton, Reno, NV, USA) was positioned towards the entrance of IMS cell.
To remove the interfering of corona discharge effect on the instrument, hollow needle was fixed
PT

inside a Teflon tube. The Bradbury-Nielsen shutter grid was applied to introduce the originated ion
swarms into the drift region. The grid potential was eliminated for about 300 µs, using a pulse
E

generator to obtain the sufficient resolution. Preheated nitrogen with the flow rates of 400 and 1000
CC

mL min-1 was used as the carrier and drift gases, respectively. The carrier gas passed through a
channel in the injection port, transferring the analyte vapor into the reaction region of IMS cell. The
A

temperature of injection port was adjusted by using a digital temperature controller (Hanyoung,
ED6, Korea). In this work, sample solution was injected into the glass liner (i.d. 1 mm) using a
microsyringe and through the septum located in head of the injection port. When the analyte vapor
reached to the reaction region, ion-molecule reactions would be occurred between the analyte
molecules and the reactant ion clusters originated by electrospray ionization. This is known as the
secondary electrospray ionization mechanism. To produce the reactant ion clusters, the electrospray

4
solution (methanol including dopant) was delivered by a programmable syringe pump (New Era
Pump System Inc., U.S.A.). Faraday plate was utilized to detect the current produced by ionized
molecules. In this work, data acquisition software based on LabView was used to collect the
spectra. The instrumental conditions of SESI–IMS were summarized in Table 1.

2.2. Chemicals

T
HPLC-grade toluene, n-hexane, methanol, cyclohexane, acetone, and acetonitrile were

IP
purchased from Merck (Darmstadt, Germany). Potassium chloride, potassium iodide and potassium

R
bromide were purchased from Sigma–Aldrich (St. Louis, USA). Commercial-grade TNT was

SC
obtained from Zarrin-Shahr Chemical Industries (Isfahan, Iran). The stock standard solution of
TNT in methanol was prepared at the concentration of 1200 mg L-1. The other working standard
solutions were daily prepared by stepwise dilution using pure water.

U
N
2.3. Extraction procedure
A
To perform the DLLME with low-density solvent, 5 mL aqueous solution of TNT was
M

transferred into the vessel of the extraction which was designed for DLLME with low-density
solvent and reported previously by Farajzadeh et al. [22]. 1 mL methanol and 40 μL toluene as
ED

dispersive and extraction solvents were rapidly injected into the aqueous solution to produce the
cloudy solution. After that, in order to separate the organic and aqueous phases, the mixture was
centrifuged at 3000 rpm for 5 min. The organic phase containing the concentrated TNT was
PT

collected at the top of the aqueous phase due to using the extraction solvent with low density
(toluene). By mildly injection of desired values of pure water into the DLLME vessel (through the
E

septum at the bottom of vessel), the organic phase was collected at the narrow neck of the vessel.
CC

Afterward, 5 μL of collected organic phase was withdrawn by a 25–μL microsyringe and injected
into the SESI–IMS, through the thermal desorption injection port.
A

2.4 Real samples

River water sample was collected from Zayandeh-rood (Isfahan, Iran) river. Tap water
sample was taken from Isfahan city (Iran) and laboratory wastewater sample was collected from a
laboratory in Zarrin-Shahr defense industry (Isfahan, Iran). River and tap water samples were
analyzed without filtration and dilution, but the laboratory wastewater sample was diluted 500

5
times prior to extraction, due to the high concentration of TNT. To deal with the matrix
interference, multiple standard addition method was performed and the calibration curves were
obtained. To that end, different volumes (0, 2.5, 5.0, 10, and 20 mL) of a TNT standard solution,
200 µg L-1, were added to 30 mL real samples in volumetric flask and diluted to 50 mL.

3. Results and discussion

T
3.1. Comparison of ESI and SESI ionization sources

IP
Since the designed analytical instrument in this research can operate with both ionization
sources of ESI and SESI, therefore the ionization efficiency and sensitivity of two sources were

R
investigated for TNT in negative mode. The ion mobility spectra obtained by the injection of 5 µL

SC
TNT solution (1.0 mg L-1) using ESI and SESI are depicted in Fig. 1. To calculate the reduced
mobility value (Ko) of produced ion, chloride ion with the Ko=2.94 cm2V-1s-1 [23] was used. The Ko

U
value for negative ion originated from TNT was calculated as 1.60 cm2V-1s-1 which is in conformity
N
with 1.54 cm2V-1s-1, reported by Hill [24] for MH-1 ion. As can be observed in this figure, the signal
intensity of the generated ion peak by SESI was about two times higher than that obtained by ESI.
A
This higher sensitivity might be due to the nature of the ionization mechanism of SESI occurred
M

during the gas-phase charge-transfer reactions. In addition to the signal improvement, two other
advantages of SESI must be noted here. Firstly, analysis time in the SESI is much shorter than that
ED

of the ESI because the liquid samples have to be delivered by a syringe pump (flow rate ~ 6 µL
min-1) in the ESI. Secondly, existence of some contaminants extracted from complex matrices may
disturb or even eliminate the signal of the ESI. Consequently, sample analysis by SESI increases
PT

our chance in selection of appropriate extraction solvents as well as simpler methods of

pretreatment.
E
CC

3.2. The effect of dopant type

To evaluate the dopant effect on the signal intensity, some halide ions were added to the
A

electrospray solvent. To that end, KCl, KBr and KI were individually added to the methanol solvent
with the concentration of 1 mM. The ion mobility spectra were obtained after injection of 5 µL
TNT solution at the concentration of 1 mg L-1 in the presence and absence of each halide ions. The
comparison between the data obtained by ESI and SESI revealed that KCl was the best dopant for
increasing the signal in both ionization sources (Fig. 2-A). This observation might be due to the
lower effective diameter of chloride relative to the other halide ions, make an easier attachment of

6
chloride ions to the TNT with three NO2 groups in its molecule. Moreover, the results show that the
dopant effect is more effective in SESI relative to ESI. Fig. 2-B shows the comparison of sensitivity
of ESI–IMS and SESI–IMS at the presence of Cl- dopant. Consequently, it can be concluded that
by using SESI instead of ESI, the sensitivity will be enhanced and also more reinforced by using
the Cl- dopant. Tam and Hill [15] analyzed some explosives by SESI-IMS in the presence of
NaNO2, NaNO3, and NaCl, as dopant added to electrospray solvent and showed the same results.

T
IP
3.3. Study of extraction conditions
The most important parameters affected the extraction efficiency of TNT were evaluated

R
and optimized. To that end, different parameters such as type and volume of the extraction solvent,

SC
type and volume of the disperser solvent, salt addition and pH were exhaustively investigated. In
this regard, all experiments were replicated at three times.

U
N
3.3.1. The effect of the type of extraction solvent
In this work, the DLLME procedure was performed with low density solvent for extraction.
A
The extraction solvent should be able to extract the target analyte form a cloudy solution in the
M

presence of the disperser solvent [25]. Accordingly, n-hexane, toluene, and cyclohexane were
tested for obtaining the best efficiency of extraction. To collect a constant volume of organic phase
ED

(20 µL), 1000 µL disperser solvent (methanol) and the different volumes of mentioned extraction
solvents (40, 40, and 50 µL of n-hexane, toluene, and cyclohexane, respectively) were injected into
the aqueous sample. The results (data not shown) indicated that the most efficiency of extraction
PT

was obtained using toluene. No ion peak originated from toluene in the negative operation mode of
SESI–IMS, helping us to determine the analyte, without interference. So, toluene was selected as
E

the extraction solvent in the DLLME procedure.

CC

3.3.2. The effect of the volume of extraction solvent

A

To optimize the volume of the extraction solvent in the presence of 1000 μL methanol,
different volumes of toluene in the range of 30 to 70 μL were studied for DLLME performance.
Based on the obtained results (data not shown), by increasing the volume of toluene up to 40 μL,
the extraction efficiency was increased and then declined with further volume of the extraction
solvent. In fact, the pre-concentration factor is decreased at high volumes of extraction solvent,

7
leads to decreasing the extraction efficiency. Therefore, 40 μL of toluene was chosen as the
extraction solvent volume for more experiments.

3.3.3. The effect of the type of disperser solvent

The disperser solvent in DLLME procedure should be soluble in both organic and aqueous
phases and also form a cloudy solution as the fine droplets in aqueous phase [26]. To optimize the

T
type of disperser solvent, methanol, acetonitrile and acetone were chosen. The experiments were

IP
performed with 40 μL of toluene and 1000 μL of disperser solvents. Based on obtained results (data
not shown), the highest signal was obtained using methanol as a disperser solvent. Therefore, it was

R
selected as the disperser solvent in this study.

SC
3.3.4. The effect of the volume of disperser solvent

U
To form an effective cloudy solution, the volume of disperser solvent should be optimized.
N
For investigation of this effect, different volumes of methanol in the range of 0.5 to 2 mL were
checked. Based on obtained results (data not shown), the maximum extraction efficiency was found
A
using 1 mL methanol. Therefore, the following experiments were performed with 1 mL methanol
M

as the disperser solvent.

ED

3.3.5. The effect of salt addition

The extraction efficiency in DLLME method may be changed by salting out effect. The
high ionic strength in aqueous sample can decrease the extraction efficiency due to creation the
PT

electrostatic media and increase the viscosity of sample [27]. Therefore, to investigate the salt
addition effect on the extraction efficiency of TNT, the different amounts of NaCl were added to
E

water samples between 0 and 20% (w/v). Obtained results (data not shown) indicated that 15%
CC

(w/v) NaCl was found as the optimum concentration of salt in water sample.
A

3.3.6. The effect of sample acidity

Since a non-polar solvent was used for the extraction of TNT, the acidity of the solution
should be adjusted to remain the TNT molecules in their neutral form. The compound of TNT has
an acidic nature due to high electron affinity of NO2 groups in its molecular structure. At very high
acidity value, the hydrogen ions react with TNT anions and result a neutral form of the analyte. To
study the sample acidity, pH of sample solution was studied in the range of 2 to 12. Based on the

8
obtained results (data not shown), the pH 3 was selected as the optimized acidity of the sample
solutions.

3.4. Analytical parameters

To evaluate the method capability and reliability of DLLME–SESI–IMS for the
determination of TNT, some analytical figures of merit were obtained. The optimum parameters

T
applied to perform these experiments were including 40 µL toluene; as the extraction solvent, 1000

IP
µL methanol; as the dispersive solvent, 15% (w/v) NaCl concentration and finally the pH of 3.0.
The investigated analytical parameters were limit of detection (LOD), limit of quantitation (LOQ),

R
linear dynamic range (LDR), repeatability, and pre-concentration factor. The precision was

SC
investigated as intra- and inter-day repeatability values. The intra-day repeatability values were
obtained in the range of 5–6%, by analyzing the TNT spiked water samples at the concentration of

U
10.0, 20.0 and 50.0 μg L-1 within one day. Also, the intra-day repeatability by analyzing the TNT
spiked water samples at the concentration of 10.0 μg L-1 over a period of three days was obtained
N
7%. The LDR was investigated at 6 point concentration levels in the range of 4.0–175.0 µg L-1 with
A
a good linearity (r2= 0.993). The equation of the obtained calibration plot is S=20.8 C+ 255, where,
M

S and C are the IMS signal and the analyte concentration, respectively. Based on the signal-to-noise
ratios of 3 and 10, the LOD and LOQ values were calculated as 1.0 μg L-1 and 4.0 μg L-1,
ED

respectively. The pre-concentration factor is the ratio of the analyte concentration in the extracted
phase to its concentration in the aqueous sample. Based on this, the pre-concentration factor was
calculated 92 for three replicated extractions of water sample which was spiked with 10.0 μg L-1 of
PT

TNT.
E

3.5. Real sample analysis

CC

To evaluate the method efficiency for the analysis of TNT in real samples, analysis of
River and tap waters, as well as laboratory wastewater sample collected from a company produces
A

explosives. TNT was detected in the laboratory wastewater sample; however two other water
samples were free from the TNT. The SESI–IMS spectra obtained for the spiked (at the
concentration of 10.0 µg L-1) and non-spiked laboratory wastewater sample were depicted in Fig. 3.
Considering the high concentration level of TNT in the wastewater sample and also its high
viscosity, the wastewater sample was diluted 500 times (with pure water), before applying the
extraction process. Based on the multiple standard addition method, the amount of TNT in

9
wastewater sample was calculated 9.5 mg L-1. Some samples were spiked at four concentration
levels (10.0, 20.0, 40.0 and 80.0 µg L-1 of TNT) and the recovery value was calculated using the
following equation:

𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑟𝑒𝑐𝑜𝑣𝑒𝑟𝑦 (%) = (𝐶𝑓𝑜𝑢𝑛𝑑 − 𝐶𝑟𝑒𝑎𝑙 )/𝐶𝑎𝑑𝑑𝑒𝑑

T
where, Creal and Cfound are the concentration of analyte calculated before and after addition of

IP
standard solution to the real sample, respectively. Cadded is also the concentration of standard which

R
was spiked to the real sample. The relative recovery values were calculated for all the studied
concentration levels and the results are in the range of 64–104% for tree real samples. As indicated

SC
in Table 2, at the concentration level of 10.0 µg L-1 of TNT the recovery values are 93, 91 and 64%
for tap, river and laboratory wastewater samples, respectively. The precision as RSD values for all

U
samples were between 6 and 8% (n = 3). Moreover, the pre-concentration factors of the extraction
N
method were obtained 89, 86 and 72 for tap, river and laboratory wastewater samples, respectively.
The results were tabulated in Table 2 revealed the performance and reliability of the method in
A
analyzing the real samples.
M

3.6. Comparison of the DLLME–SESI–IMS with other reported methods

ED

Some analytical parameters including LDR, LOD, RSD and recoveries obtained by the
proposed method were compared with those obtained by some other methods reported previously
for analysis of TNT. According to the results summarized in Table 3, it can be concluded that the
PT

detection limits were almost better than that of other methods. However, the RSD values were
comparable with that obtained by the other reported methods. Garofolo et al. [34] reported an
E

application of SPE-HPLC-IMS for extraction and determination of trace levels of explosives. They
CC

could analyze explosives in complex matrices due to application of both SPE and HPLC as
powerful extraction and separation techniques, respectively. However, the method of HPLC needs
A

a long run time in addition to expensive solvents for the gradient elution. In another reference [35],
63
Lu et al. used molecularly imprinted polymer as the extraction method before Ni-IMS and
analyzed nitrobenzene compounds in surface water. Although, they used MIP as a selective
technique in the sample preparation, however, they reported a lower sensitivity than this work, may
be due to application of 63Ni as ionization source. The recovery and the relative standard deviation
values reported by Lu et al. are the same as those obtained by this method. It is notable that the

10
method has a very shorter analysis time in comparison with the other ones which is very important
to employ the analytical method for routine analysis of different matrices.

4. Conclusion
In this work, secondary electrospray ionization (SESI–IMS) was applied for analyzing the
extracted TNT from water and wastewater samples. In the view of sensitivity, ESI and SESI

T
ionization sources were compared in the presence and absence of different halide ions. Dispersive

IP
liquid-liquid microextraction (DLLME) having a fast procedure time and low consumption of
organic solvent was used for the sample preparation. The results obtained from real samples

R
analysis confirmed the capability and reliability of the DLLME–SESI–IMS method to determine

SC
TNT. Some advantages of the SESI–IMS are the short analysis time, the capability of analysis of
gaseous and liquid samples, the feasibility of coupling the instrument with separation techniques

U
like gas-chromatography, as well as more stability, simplicity, and efficiency related to ESI–IMS.
N
These make SESI–IMS as a worthy method for the analysis of TNT in different matrices.
A
Acknowledgments
M

Authors gratefully acknowledge the financial support of the Research Council of Isfahan University
of Technology (IUT) and Center of Excellence in Sensor and Green Chemistry.
ED
E PT
CC
A

11
References
[1] M.J. Cohen, F.W. Karasek, Plasma chromatography™—a new dimension for gas
chromatography and mass spectrometry, J. Chromatogr. Sci. 8 (1970) 330−337.
[2] G.A. Eiceman, Z. Karpas, Ion mobility spectrometry, second ed., CRC Press, Boca Raton,
2005.
[3] R.G. Ewing, D.A. Atkinson, G.A. Eiceman, G.J. Ewing, A critical review of ion mobility

T
spectrometry for the detection of explosives and explosive related compounds, Talanta 54 (2001)

IP
515–529.
[4] T. Khayamian, M. Tabrizchi, M. T. Jafari, Analysis of 2,4,6-trinitrotoluene, pentaerythritol

R
tetranitrate and cyclo-1,3,5-trimethylene-2,4,6-trinitramine using negative corona discharge ion

SC
mobility spectrometry, Talanta 59 (2003) 327–33.
[5] M. T. Jafari, F. Riahi, Feasibility of corona discharge ion mobility spectrometry for direct

U
analysis of samples extracted by dispersive liquid–liquid microextraction, J. Chromatogr. A 1343
N
(2014) 63–68.
[6] G. R. Asbury, J. Klasmeier, H. H. Hill Jr, Analysis of explosives using electrospray
A
ionization:ion mobility spectrometry (ESI:IMS), Talanta 50 (2000) 1291–1298.
M

[7] W. E. Steiner, B. H. Clowers, K. Fuhrer, M. Gonin, L. M. Matz, W. F. Siems, A. J. Schultz, H.

H. Hill, Jr, Electrospray ionization with the ambient pressure ion mobility separation and mass
ED

analysis by orthogonal time-of-flight mass spectrometry, Rapid Commun. Mass Spectrom. 15

(2001) 2221–2226.
[8] M. T. Jafari, M. Javaheri, Selective Method Based on Negative Electrospray Ionization Ion
PT

Mobility Spectrometry for Direct Analysis of Salivary Thiocyanate, Anal. Chem. 82 (2010) 6721–
6725.
E

[9] M. L. Gieniec, J. Cox Jr., D. Teer, M. Dole, 20th Annual conference on mass spectrometry and
CC

allied topics, Dallas, TX, June 4–9, 1972.

[10] Y. H. Chen, Electrospray ionization-ion mobility spectrometry. Ph.D. Dissertation,
A

Washington State University, Pullman, WA, 1994.

[11] C. Wu, Development and application of electrospray ionization/high resolution ion mobility
spectrometry - mass spectrometry. Ph.D. Dissertation, Washington State University, Pullman, WA,
1997.

12
[12] Z. Wu, C. L. Hendrickson, R. P. Rodgers, A. G. Marshall, Composition of explosives by
electrospray ionization fourier transform ion cyclotron resonance mass spectrometry, Anal. Chem.
74 (2002) 1879–1883.
[13] B. Casetta, F. Garofolo, Characterization of explosives by liquid chromatography/mass
spectrometry and liquid chromatography/tandem mass spectrometry using electrospray ionization
and parent-ion scanning techniques, Org. Mass Spectrom. 29 (1994) 517–25.

T
[14] A. Gapeev, M. Sigman, J. Yinon, Liquid chromatography/mass spectrometric analysis of

IP
explosives: RDX adduct ions, Rapid Commun. Mass Spectrom. 17 (2003) 943–948.
[15] M. Tam, H. H. Hill, Jr, Secondary electrospray ionization-ion mobility spectrometry for

R
explosive vapor detection, Anal. Chem. 76 (2004) 2741–2747.

SC
[16] C. Wu, W. F. Siems, H. H. Hill, Jr. Secondary electrospray ionization ion mobility
spectrometry/mass spectrometry of illicit drugs, Anal. Chem. 72 (2000) 396–403.

U
[17] W. E. Steiner, B. H. Clowers, P. E. Haigh, H. H. Hill, Jr. Secondary ionization of chemical
N
warfare agent simulants: atmospheric pressure ion mobility time-of-flight mass spectrometry, Anal.
Chem. 75 (2003) 6068–6076.
A
[18] P. Martínez-Lozano, J. Rus, G. Fernández de la Mora,M. Hernández,J. Fernández de la Mora,
M

Secondary electrospray ionization (SESI) of ambient vapors for explosive detection at

concentrations below parts per trillion, J. Am. Soc. Mass Spectrom. 20 (2009) 287–294.
ED

[19] M. T. Jafari, M. Saraji, M. Mossaddegh, Combination of dispersive liquid–liquid

microextraction and solid–phase microextraction: An efficient hyphenated sample preparation
method, J. Chromatogr. A 1466 (2016) 50–58.
PT

[20] T. Khayamian, M. T. Jafari, Design for electrospray ionization-ion mobility spectrometry,

Anal. Chem. 79 (2007) 3199–3205.
E

[21] M. T. Jafari, Improved design for high resolution electrospray ionization ion mobility
CC

spectrometry, Talanta 77 (2009) 1632–1639.

[22] M. A. Farajzadeh, S. E. Seyedi, M. S. Shalamzari, M. Bamorowat, Dispersive liquid–liquid
A

microextraction using extraction solvent lighter than water, J. Sep. Sci. 32(2009) 3191–3200.
[23] E. Glenn, I. Charles, Reactant ions in negative ion plasma chromatography, Anal. Chem. 47
(1975) 393–402.
[24] G. Reid Asbury, J. Klasmeier, H. H. Hill, Analysis of explosives using electrospray ionization

ion mobility spectrometry, Talanta, 50 (2000) 1291–1298.

13
[25] K. Pytlakowska, R. Sitko, Energy-dispersive X-ray fluorescence spectrometry combined with
dispersive liquid–liquid microextraction for simultaneous determination of zinc and copper in water
samples, Anal. Method 5 (2013) 6192–6199.
[26] M. Uslu, H. Ulutürk, A. Yartaşı, S. Döke, A sensitive method for selective determination of
vanadium species by dispersive liquid–liquid microextraction (DLLME) with spectrophotometric
detection, Toxicol. Environ. Chem. 95 (2013) 1638–1649.

T
[27] L. Zhao, L. Zhu, H. K. Lee, Analysis of aromatic amines in water samples by liquid–liquid–

IP
liquid microextraction with hollow fibers and high-performance liquid chromatography, J.
Chromatogr. A, 963 (2002) 239–248.

R
[28] E. Ercag, A. emuzer, R. Apak, Selective spectrophotometric determination of TNT using a

SC
dicyclohexylamine-based colorimetric sensor, Talanta 78 (2009) 772–780.
[29] A. Uzer, E. Ercag, R. Apak, Selective spectrophotometric determination of trinitrotoluene,

U
trinitrophenol, dinitrophenol and mononitrophenol, Anal. Chem. Acta 505 (2004) 83–93.
N
[30] D. Nie, D. Jiang, D. Zhang, Y. Liang, Y. Xue, T. Zhou, L. Jin, G. Shi, Two-dimensional
A
molecular imprinting approach for the electrochemical detection of trinitrotoluene, Sens. Actuators
M

B 156 (2011) 43–49.

[31] A. Larkia, M. Rahimi Nasrabadi, N. Pourreza, UV- Vis spectrophotometric determination of

ED

trinitrotoluene (TNT) with trioctylmethylammonium chloride as ion pair assisted and disperser

agent after dispersive liquid–liquid microextraction, Forensic Sci. Inl. 251 (2015) 77–82.
PT

[32] A. Uzer, E. Ercag, R. Apak, Selective colorimetric determination of TNT partitioned between
an alkaline solution and a strongly basic Dowex 1-X8 anion exchanger, Forensic Sci. Int. 174
E

(2008) 239–243.
CC

[33] C. Cortada, L. Vidal, A. Canals, Determination of nitroaromatic explosives in water samples

by direct ultrasound-assisted dispersive liquid–liquid microextraction followed by gas
A

chromatography mass spectrometry, Talanta 85 (2011) 2546–2552.

[34] F. Garofolo, V. Migliozzi, B. Roio, Application of ion mobility spectrometry to the
identification of trace levels of explosives in the presence of complex matrices, Rapid Commun.
Mass Spectrom. 8 (1994) 527–532.

14
[35] W. Lu, H. Lib, Z. Meng, X. Liang, M. Xue, Q. Wang, X. Dong, Detection of nitrobenzene
compounds in surface water by ion mobility spectrometry coupled with molecularly imprinted
polymers, J. Hazard. Mat. 280 (2014) 588–594.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

15
Figure Captions

Fig. 1. The ion mobility spectra obtained by SESI (above) and ESI (below) for the 5-µL injection

of TNT in methanol (1.0 mg L-1).

Fig. 2. A) The effect of dopant halide ions on the signal intensity compared with pure methanol and

T
B) the comparison of instrument sensitivity in the presence of chloride dopant obtained by ESI–

IP
IMS and SESI–IMS.

R
Fig. 3. The ion mobility spectra obtained by SESI-IMS for the extracted samples from diluted

SC
laboratory wastewater sample before and after spiking with TNT (10.0 µg L-1).

U
N
A
M
ED
E PT
CC
A

16
 Fig. 1
A
CC
E PT
ED

17
M

TNT
TNT

A
N
U
SC
RIP
T
 Fig. 2
A
CC
EPT
ED

18
M
A
N
U
SC
RIP
T
 T
IP
TNT

R
SC
U
N
after spiking
A
M

before spiking
ED
E PT
CC

Fig. 3
A

19
Table 1

Instrumental parameters for SESI–IMS.

Parameter Setting
Needle voltage -10.5 kV
Target electrode voltage -8.0 kV
Temperature of injector 200 ºC

T
Temperature of injector 150 ºC

IP
Drift field 300 V cm-1
Drift gas flow (N2, 99.999%) 1000 mL min-1

R
Carrier gas flow (N2, 99.999%) 400 mL min-1

SC
Drift tube length 11 cm
Shutter grid pulse 0.3 ms
Number of IMS averages 25
Number of points per ion mobility spectrum 500
U
N
A
M
ED
E PT
CC
A

20
Table 2

Analysis of TNT in real samples using DLLME–SESI–IMS.

Sample TNT measured LDRc PFd LODe RSD%f Relative
in real samplea recovery (%)g
Laboratory wastewater 9500 10.0–100.0 72 3 8 64
Tap water NDb 10.0–100.0 89 3 7 93

T
River water ND 10.0–100.0 86 3 6 91
a
Concentration (µg L−1).

IP
b
Not detected.
c
Linear dynamic range (µg L-1).
d
Preconcentration factor.

R
e
Limit of detection (µg L-1).
f
Relative standard deviation (10.0 µg L-1 for water and wastewater samples).
g
Relative recovery was calculated by analyzing the real samples spiked with 10.0 µg L-1.

SC
U
N
A
M
ED
E PT
CC
A

21
Table 3

Comparison of the analytical parameters of the DLLME–SESI–IMS method with other methods used for the determination of TNT in real
samples.
Method Sample type Analysis time Dynamic range LODa RSDb (%) Recovery (%) Referenc
(min) (µg L−1) (µg L−1) e
Polymer Sensor– UV–vis Soil and groundwater 30 5000–50000 3000 5 97–112 [28]
LLEc–UV–vis Soil and laboratory wastewater 30 4000–40000 1200 – – [29]
MIPd– Voltammetry Soil and river water 10 9.1–730 2.9 2–5 97–98 [30]
DLLME– UV–vis Soil and river water 8 3–200 0.9 1.5–3.7 99–104 [31]

T
Anion exchanger e resin–UV– Soil and laboratory wastewater 50 400–40000 100 – 93–97 [32]
vis

IP
DLLME–GC–MS Well water, effluent wastewater 40 1–10 0.2 5 71–79 [33]
Direct analysis–ESI–IMS Environmental samples <3 – 15 – – [24]
20–105 – –

R
Direct analysis–CD–IMS Standard samples <2 7 [4]
SPE–HPLC–IMS Samples collected at the scene ~50 104-105 – – – [34]
of real detonations

SC
MIP–63Ni–IMS Industrial waste water and 35 500–50000 100 6.8 48–106 [35]
surface water
DLLME–SESI–IMS Laboratory wastewater, tap and 5 4.0–175.0 1.0 6–8 64–104 This
river waters study
a
Limit of detection.
b
c
d
Relative standard deviation.
Liquid-liquid extraction.
Molecularly imprinted polymer. U
N
A
M
ED
E PT
CC
A

22