Improving Quantification of tabun, sarin, soman, cyclosarin, and sulfur mustard by focusing agents: A field portable gas chromatography-mass spectrometry study

Journal of Chromatography A 1636 (2021) 461784

Contents lists available at ScienceDirect

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

Improving Quantification of tabun, sarin, soman, cyclosarin, and sulfur
mustard by focusing agents: A field portable gas
chromatography-mass spectrometry study
John T. Kelly a,∗, Anthony Qualley a, Geoffrey T. Hughes a, Mitchell H. Rubenstein b,∗,
Thomas A. Malloy c, Tedeusz Piatkowski c
a

UES, Inc., Air Force Research Laboratory, 711th Human Performance Wing/RHMO, 2510 Fifth Street, Area B, Building 840, Wright-Patterson AFB, OH 45433,
USA
b
United States Air Force 711th Wing – Air Force Research Laboratory, 711th Human Performance Wing/RHMO, 2510 Fifth Street, Area B, Building 840,
Wright-Patterson AFB, OH 45433, USA
c
Hazardous Materials Research Center (HMRC), Battelle Columbus Laboratories, Battelle Memorial Institute, Columbus, OH, USA

a r t i c l e

i n f o

Article history:
Received 27 July 2020
Revised 30 November 2020
Accepted 2 December 2020
Available online 13 December 2020

a b s t r a c t
Commercial gas chromatograph-mass spectrometers, one of which being Inficon’s HAPSITE® ER, have
demonstrated chemical detection and identification of nerve agents (G-series) and blistering agents (mustard gas) in the field; however most analyses relies on self-contained or external calibration that inherently drifts over time. We describe an analytical approach that uses target-based thermal desorption standards, called focusing agents, to accurately calculate concentrations of chemical warfare agents that are
analyzed by gas chromatograph-mass spectrometry. Here, we provide relative response factors of focusing agents (2-chloroethyl ethyl sulfide, diisopropyl fluorophosphate, diethyl methylphosphonate, diethyl
malonate, methyl salicylate, and dichlorvos) that are used to quantify concentrations of tabun, sarin, soman, cyclosarin and sulfur mustard loaded on thermal desorption tubes (Tenax® TA). Aging effects of
focusing agents are evaluated by monitoring deviations in quantification as thermal desorption tubes age
in storage at room temperature and relative humidity. The addition of focusing agents improves the quantification of tabun, sarin, soman, cyclosarin and sulfur mustard that is analyzed within the same day as
well as a 14-day period. Among the six focusing agents studied here, diisopropyl fluorophosphate has the
best performance for nerve agents (G-series) and blistering agents (mustard gas) compared to other focusing agents in this work and is recommended for field use for quantification. The use of focusing agent
in the field leads to more accurate and reliable quantification of Tabun (GA), Sarin (GB), Soman (GD),
Cyclosarin (GF) and Sulfur Mustard (HD) than the traditional internal standard. Future improvements on
the detection of chemical, biological, radiological, nuclear, and explosive materials (CBRNE) can be safely
demonstrated with standards calibrated for harmful agents.
© 2020 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license
( />
1. Introduction
The need for detecting chemical warfare agents (CWAs) is
ubiquitous and is considered of the highest importance for protecting soldiers, sailors, airmen, and Marines. The military community supports current methodology for detecting nerve agents
with significant gaps in reliability and transferability in fieldportable chemical identification methods. The organophosphorus

∗

Corresponding authors.
E-mail address: (J.T. Kelly).

nerve agents, tabun (GA), sarin (GB), soman (GD), cyclosarin (GF)
and the blistering agent sulfur mustard (HD) are recognized as
some of the most lethal CWAs with respect to persistency and toxicity [1–3]. A common point-sensing approach to detecting nerve

agents in the field is by mass spectrometry; which has the reputation of being the “gold standard” of chemical identification
with high sensitivity and high selectivity. Gas chromatographymass spectrometry (GC-MS) is commonly used in the identification of CWAs, as well as chemical precursors and decomposition
products however inter- and intra-instrument reproducibility puts
into question absolute quantification. The chemical structures of
the CWAs investigated in this work are shown in Fig. 1.

/>0021-9673/© 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
( />

J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 1. Structures of the chemical warfare agents: (a) GA, (b) GB, (c) GD, (d) GF and (e) HD. Carbon is black, oxygen is red, nitrogen is blue, fluorine is olive, chlorine is lime,
sulfur is mustard, and phosphorus is orange. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

For a number of years, response factors have been of interest for developing novel mass spectrometer motifs and fundamental aspects: mass analyzers type [4–8] (magnetic sector, time-of flight, quadrupole, ion trap, or ion cyclotron resonance), ionization
method [9–11] (electron impact, chemical, and electrospray) and
sample introduction [12–15] (liquid injection, solid phase microextraction and thermal desorption). Seto and coworkers have reported the response factors for G-series nerve agents for two fieldportable GC-MS spectrometers, the HAPSITE® (Hazardous Air Pollutants on Site) and the HAPSITE® ER (Hazardous Air Pollutants
on Site Extended Range) [16,17]. The results of earlier experiments
concluded with significant residual carryover effects from adsorption to the air sampling probe and transfer line to the concentrator. The HAPSITE® ER has an onboard internal standard, bromopentafluorobenzene (BPFB), that is commonly used for quantification however is known to fluctuate between days (26.3 %RSD)
and even more within a single day (32.9%RSD) [17].
While GC-MS has been shown to be capable of deconvoluting
complicated chemical mixtures in solution, there has been significant work in coupling sorbent traps to automated thermal desorption to improve detection limits. Combining thermal desorption with gas chromatography has been demonstrated for the application of CWAs in the laboratory setting [18–22] however literature for field-portable techniques is more sparse. KanamoriKataoka and Seto reported a thorough study on comparing CWAs
performance on Tenax® TA, Tenax® GR, and Carboxen® 1016 thermal desorption tubes. The thermal desorption GC-MS analysis revealed Tenax® TA as the most favorable of the three sorbents in
dry conditions as well as humid conditions up to 50% relative humidity [23]. The use of thermal desorption improves the detection
limits substantially however a number of considerations must be
evaluated (e.g. thermal desorption tube stability, capacity, carryover, and sampling rates.) [24]
The primary aim of this work is to provide relative response
factors [25] (RRFs) as target-based standards, referred to in this

study as focusing agents, for G-series agents and sulfur mustard
quantification after analyzed by two different HAPSITE® ER systems. Our previous work shows system-level improvement in the
data quality when switching from external and internal standards
to in-situ calibration with isotopic analogues [26]. Here, we demonstrate the quality control in quantification of by using the following focusing agents: 2-chloroethyl ethyl sulfide (2-CEES), diisopropyl fluorophosphate (DIFP), diethyl methylphosphonate (DEMP),
diethyl malonate (DEM), methyl salicylate (MES), and dichlorvos
(DCV) as an alternative focusing agent to demonstrate a transferable RRF for all field-portable GC-MS systems. Fig. 2 shows structures of the focusing agents used in this work. Thermal desorption
tubes are spiked with focusing agent and exposed to select con-

centrations of nerve gases (G-series) and blistering agents (mustard gas) to establish RRFs and are monitored over 14 days evaluate storage at room temperature and relative humidity ranging
from 24% to 34%. This transferring of RRFs across instruments that
would otherwise rely on external calibration alone is referred to as
calibration transportability.
2. Experimental
2.1. Reagents and material
A total set of 6 focusing agents were purchased from at Sigma–
Aldrich: 2-CEES, DIFP, DEMP, DEM, MES, and DCV. Isopropyl alcohol and acetonitrile (American Chemical Society grade or equivalent) were used as solvents for liquid dilution. GB, GD, GA, GF, and
HD are not commercially available; however, independent standard
solutions were prepared by two different analysts. CWA purity was
determined by preparing high concentration stock solutions in acetonitrile and analyzing the solution by gas chromatography-flame
ionization detection (GC-FID) and followed standard operating procedure Hazardous Materials Research Center [HMRC] IV-056.
2.2. Instrumentation
This investigation examined the performance of the HAPSITE®
ER (Inficon) with respect to five chemical warfare agents: sarin,
tabun, soman, cyclosarin, and distilled sulfur mustard by thermal
desorption from Supelco Tenax® TA (35/60) thermal desorption
tubes. The temperature of the thermal desorber sampling system
(TDSS) was set to 310 °C during which nitrogen carrier gas transferred the desorbed sample to a tri-bed concentrator that is held
at 45 °C for 12:00, where time in mm:ss. A new tri-bed concentrator was installed in each instrument at the start of testing and was
not replaced throughout the duration of the present work. The tribed concentrator is then heated to 280 °C in 11 seconds and then
introduced to a DB-1ms GC column (15 m, 0.25 mm ID, 1.0 μm

df ). The column temperature, membrane and valve oven were set
at 60 °C, 120 °C, and 120 °C respectively. The temperature profile
for all measurements held the GC column at 60 °C for 01:15, followed by a thermal ramp at a rate of 8 °C min−1 for 03:45. The
temperature was then ramped at a rate of 25 °C min−1 for 04:24
after reaching 90 °C. The GC column temperature was limited to
200 °C for the remaining time of the 15:30 experiment. The carrier gas (nitrogen) flow rates through the GC column were dependent on a constant inlet pressure of 88 kPa and the column temperature. The spatial and temporally separated eluates then pass
through a hydrophobic membrane that interfaced with the electron ionization (EI, positive) quadrupole mass spectrometer. The
2


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 2. Structures of focusing agents: (a) 2-CEES, (b) DIFP, (c) DEMP, (d) DEM (e) MES, and (f) DCV. Carbon is black, oxygen is red, hydrogen is grey, fluorine is olive, chlorine
is lime, sulfur is mustard and phosphorus is orange. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

employed source settings were previously established for consistency with Automated Mass Spectral Deconvolution and Identification System [27,28] or AMDIS (Version 2.72, 2014, National Institute of Standards and Technology), using NIST mass spectral libraries. Mass spectra were acquired for the mass-to-charge range
from 45 to 300 (dwell time is 300 μs dwell time, 0.765 scans per
second). The internal standard for this work was bromopentafluorobenzene (BPFB, 5.5 ppm) and has previously shown to have poor
reproducibility (R2 ranging from 0.88 to 0.982).

Sample preparation and thermal desorption tube conditioning
was previously described [26]. In brief, thermal desorption tubes
were procured from Supelco and conditioned using a Markes International TC-20 with dry-purging for multiple sorbent tubes. Flow
rates did not exceed 50 mL min−1 and conditioning temperatures
were set to 280 °C for 120 mins at Wright-Patterson Air Force Base
(WPAFB). Thermal desorption tubes were loaded with a Markes
International Calibration Solution Loading Rig (CSLRTM ) with flow

rates not exceeding 50 mL min−1 of ultra-high purity nitrogen gas.
Single and dual regulator pneumatics controllers (Markes International Gas01 and Gas03) were used for setting backing pressures.

2.3. Thermal desorption tube preparation

2.4. RRFs

For baseline testing and calibration, Supelco Tenax® TA (35/60)
thermal desorption tubes were spiked with 1, 2, 5, 10, and 50 ng of
each focusing agent. thermal desorption tubes were characterized
for stability: aging, carry-over and residual focusing agent on the
thermal desorption tube were evaluated by two HAPSITE® ER (instruments 112 and 121) analysis as previously described [26,29].
In brief, carryover was determined by a subsequent blank tube
thermally desorbed after a spiked tube analysis. Spiked thermal
desorption tubes were analyzed in repeat desorptions to measure
residual focusing agent on the thermal desorption tube. Aging effects of focusing agents spiked on thermal desorption tubes were
measured by comparing mid-point signal response of 5 ng of focusing agent (capped and stored at ambient temperatures and relative humidity levels) on days 0, 3, 7, and 14.
External calibrations for CWAs (GA, GB, GD, GF, and HD) were
created in triplicate by a six-point curve (1, 2, 5, 10, 20, and 50 ng)
and internal standard values reported in the literature (26.3 %RSD)
were consistent with the values reported here (24.2 %RSD) [17].
The Stability and thermal desorption tube aging effects were established for CWAs following the same approach as for the focusing
agents. Relative response factors (RRFs) are compared through internal standard and focusing agents in the quantification of CWAs.
Area determinations were made for the internal standard and focusing agents using the HAPSITE® ER IQ software package (v. 2.32,
Inficon) and AMDIS protocol.

RRFs are calculated for GA, GB, GD, GF, and HD using the areas
of the CWA (ACWA ) and focusing agent (AFA ) and the ratio of the
mass of focusing agent (mFA ) and mass of CWA (mCWA ) loaded on
the thermal desorption tube as shown in Eq. 1. The derivative of

the relative response function accounts for the area response ratio
(ACWA /AFA ) and mCWA therefore the RRF is a product of the derivative and mFA .

RRF =

ACWA mFA
AFA mCWA

(1)

RRFs are calculated by fitting a six-point analysis at 1, 2, 5, 10,
20 and 50 ng of CWAs by the area response ratio of the CWA
and focusing agent. Using a forced-origin linear analysis (OriginPro 2020b), values for relative response functions including slopes,
standard errors, R2 and adjusted-R2 are reported. Tabulated values
can be found in the Supplemental Materials.
2.5. Safety considerations
The CWAs investigated here are highly toxic and were handled with personal protective equipment (PPE). Full-face respiratory masks, self-contained breathing apparatuses, and liquid-proof
and vapor-impermeable suits are highly recommended for handling CWAs [1]. Safety protocol established at Battelle followed
3


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 3. Mass spectra of (a) GA, (b) GB, (c) GD, (d) GF, and (e) HD and asterisk denoting ions used in quantification and confirmation. Total ion chromatogram of (f) CWAs via
thermal desorption GC-MS (HAPSITE® ER) including focusing agents. Combines extracted ion chromatogram of (g) CWAs and (h) focusing agents. Area response of 50 ng of
GB, GD, GA, GF, and HD loaded on to Tenax® TA (35/60) thermal desorption tubes.

Table 1

CWA retention times (RT) and mass-tocharge ratio for ions used in quantification
and confirmation (m/z) compared to precious values [17].

standard operating procedure Hazardous Materials Research Center [HMRC] IV-056.

3. Results and discussion

This work

3.1. Determination of CWAs
GB
GD
GA
HD
GF

We examined the capability to use focusing agents on thermal desorption tubes as a method of improving quantification of
GA, GB, GD, GF and HD in the field. Six different focusing agents
were nominated and characterized by performance, chemical stability and reproducibility. The NIST library mass spectra of CWAs
is shown in Fig. 3a-e and a representation of the total ion chromatogram (TIC) containing CWAs and focusing agents is shown in
Fig. 3f. Extracted ion chromatograms (EIC) of all CWAs is shown
in Fig. 3g and the EIC for all focusing agents is shown in Fig. 3h.
The retention time (mm:ss) and quantification ions (m/z) for CWAs
are presented in Table 1 and are compared to previously reported
values [17]. The four-step AMDIS approach was used in the identification of CWAs (noise characterization, perception of CWAs, extraction of CWA spectrum, and library comparison with a target
library for mass spectrum and retention time) [28]. Fig. 3(f-h) are
reduced representations of experimental data by the HAPSITE® ER

[Ref. 17]

RT

m/z

RT

m/z

02:48
06:13
06:58
07:27
07:44

99
126
133
111
99

02:47
05:57
06:47
07:05
07:52

99
99
70
109

99

by adapting retention times and response areas to a summation of
a gaussian functions with the peak areas reflecting the observed
values. The experimental TICs and EICs can be found in Fig. S1 in
the Supplemental Materials.
Carryover and residual CWAs remaining on a thermal desorbed
thermal desorption tube was measured by comparing area response from a 50 ng spike of CWAs along with that of the internal standard. Carryover for GA, GD, GF and HD were 0.26%, 0.04%,
0.03% and 0.02% respectively. The thermal desorption tube residual for GA, GD, GF and HD were 0.26%, 0.04%, 0.01% and 0.02%
4


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 4. Mass spectra of (a) 2-CEES, (b) DIFP, (c) DEMP, (d) DEM, (e) MES and (f) DCV and asterisks denoting ions used in quantification and confirmation.

respectively. The nominally larger variations in internal standard
for BPFB range from 58% to 78% for carryover and 48% to 87% for
residual BPFB. This lack of stability and reproducibility mandates
an alternative quantitative methodology, and we employ focusing
agents to circumvent this discrepancy. Carryover and residual focusing agent on thermal desorption tubes were less than 0.1% for
both instruments and solvents (isopropyl alcohol and acetonitrile).
There is no significant influence of carryover on the quantification of the analysis. Neither solvent showed favorable chromatographic response or instrument response leading to the inference
that either could be used in standard preparation. The results for
all carry-over and residual experiments for four HAPSITE® ER systems are in the Supporting Information (Table S1).
Influence of carryover on the quantification of the following
sample Influence of carryover on the quantification of the following sample Influence of carryover on the quantification of the following sample Influence of carryover on the quantification of the
following sample Influence of carryover on the quantification of

the following sample

torically a significant methodology for determining concentration
of species in an analytical setting ranging for gas chromatography.
In 1976, Pacer states “The use of relative response factors in an experiment illustrates the fact that components of a mixture, present in
equal amounts, need not respond equally in order for that response
to be useful for quantitative purposes. As long as the response is reproducible, a quantitative method is feasible” and later states that
RRFs can provide as little as 1% error by a GC approach [30]. Pardue et al. [31] and Karasek et al. [32] demonstrate the peak area
methodology for determining RRFs by GC. Here, focusing agent
area responses are used for obtaining the RRFs for CWAs.
Here, we report the RRFs for GA, GB, GD, GF, and HD for focusing agents (2-CEES, DIFP, DEMP, DEM, MES, and DCV). This quantitative relationship provides a reliable and reproducible approach
for quantifying G-series CWAs by thermal desorption GC-MS. All
CWAs and focusing agents were analyzed in a single thermal desorption tube and were done in triplicate across four HAPSITE® ER
systems. The data that is used for discussion is selected with nearly
average results in attempt to reflect the overall system performance. The 5 ng of each 2-CEES, DIFP, DEMP, DEM, MES, and DCV
reflected area responses of 5612366, 9170962, 8204334, 7931053,
391090 03, and 2730 0138. These values are used for the RRFs for
all CWAs that are used to calculate each relative response function.

3.2. Characterization of focusing agents
Previous evaluation of thermal desorption analysis on a
portable GC–MS systems compares performance to that of a standard bench-top instrument with one of the most problematic inconsistencies is the response area of the internal standards, 1,3,5tris(trifluoromethyl)benzene and bromopentafluorobenzene [29].
We have demonstrated the improvement of data quality for fieldportable GC-MS through the use of focusing agents and employ them for determining reliable RRFs and demonstrating interinstrument transportability [26]. The retention time (mm:ss) and
quantification ions (m/z) for the selected focusing agents are 2CEES – 04:38 (75), DIFP – 05:08 (101), DEMP – 05:52 (79), DEM –
06:31 (115), MES – 07:48 (120) and DCV – 08:08 (109) as seen in
the extract ion chromatogram in Fig. 3f and mass spectra in Fig. 4.
External calibrations for focusing agents, as seen for DIFP in Fig. 3,
are consistent and reliably linear within the data set. RRFs are his-

3.2.1. Tabun (GA)

RRFs for GA are by a six-point quantitative analysis at 1, 2,
5, 10, 20 and 50 ng of GA loaded on thermal desorption tubes
and analyzed by thermal desorption GC-MS. The linear relative response functions have slopes of 0.241, 0.105, 0.129, 0.092, 0.0192,
and 0.0303 for 2-CEES, DIFP, DEMP, DEM, MES and DCV respectively. The derivative of the relative response function provides the
calculated RRFs without accounting for the focusing agent mass
loaded on the thermal desorption tubes. The RRFs are 1.203, 0.523,
0.647, 0.458, 0.0962, and 0.1515 for 2-CEES, DIFP, DEMP, DEM, MES
and DCV respectively with 5 ng of focusing agent. All relative response functions have a fixed y-intercept of 0 for the linear analysis. The reliability of focusing agents determined by looking at
5


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 5. RRFs for GA for (a, red) 2-CEES, (b, orange) DIFP, (c, yellow) DEMP, (d, green) DEM, (e, blue) MES and (f, purple) DCV for day 0. The six-point quantitative analysis at 1,
2, 5, 10, 20 and 50 ng are denoted by dots, the relative response function are black traces and the colored shadows represent the standard error in the fit. (For interpretation
of the references to color in this figure legend, the reader is referred to the web version of this article.)

the aging effects of the focusing agents have on the quantification of GA after 3, 7 and 14 days after being capped and stored
at ambient temperatures and relative humidity levels ranging from
24% to 34%. The percent relative standard deviation (%RSD) for 2CEES, DIFP, DEMP, DEM, MES and DCV are 45%, 14%, 12%, 22%, 30%,
and 20% respectively over 14 days. In the case of GA, the focusing agents DIFP, DEMP, DEM, MES, and DCV are an improvement
on reliability beyond the BPFB internal standard (32.9%) [17]. Fig. 5
shows the relative response functions of GA and different focusing
agents (a-f) along with the standard error in the slope. The coefficient of determination (R2 ) and adjusted-R2 values for relative
response functions of GA are 0.994 or better. See Table S1 in the
Supplemental Materials for the slope, standard error in the slope,
R2 and adjusted-R2 values.

(25%), DEM (9%), MES (2%) and DCV (9%) are more reliable than
the internal standard, BPFB (32.9%). The only focusing agent that
performed worse than BPFB is 2-CEES (37%). GD has a retention
time of 06:13 and there are two neighboring focusing agents in the
TIC for DEMP (05:52) and DEM (06:31). Detection techniques other
than mass spectrometry can encounter problems in quantification
in the presence of poor peak shapes or exceedingly high concentrations of GD. Fig. 7 shows the relative response functions of GD
and different focusing agents (a-f) along with the standard error
in the slope. The coefficient of determination (R2 ) and adjustedR2 values for relative response functions of GD are 0.997 or better.
See Table S3 in the Supplemental Materials for the slope, standard
error in the slope, R2 and adjusted-R2 values.
3.2.4. GF
RRFs for GF are calculated through the relative response functions for day 0 for 2-CEES, DIFP, DEMP, DEM, MES and DCV. The
slopes of these functions are 2.1, 0.93, 1.15, 0.82, 0.171 and 0.27.
The RRFs are 10.7, 4.66, 5.75, 4.08, 0.856, and 1.35. The %RSD over a
14-day aging study reveals that DIFP (12%), DEMP (18%), DEM (2%),
MES (9%) and DCV (3%) are more reliable than the internal standard, BPFB (32.9%). The only focusing agent that performed worse
than BPFB is 2-CEES (41%). We report the retention time for GF is
07:44 and the retention time for the focusing agent MES is 07:48.
While this method does not resolve individual peaks in the TIC,
quantification of GF is performed by using unique features in mass
spectra where GF and MES can be separated. Detection techniques
such as FID or PID can encounter problems in quantification for GF
if using MES as a focusing agent. Fig. 8 shows the relative response
functions of GF and different focusing agents (a-f) along with the
standard error in the slope. The coefficient of determination (R2 )
and adjusted-R2 values for relative response functions of GF are
0.983 or better. See Table S4 in the Supplemental Materials for the
slope, standard error in the slope, R2 and adjusted-R2 values.

3.2.2. Sarin (GB)
The linear relative response functions for GB have slopes of
0.179, 0.078, 0.096, 0.068, 0.0143, and 0.0225 and RRFs 0.894,
0.389, 0.481, 0.340, 0.0716, and 0.1126 for 2-CEES, DIFP, DEMP,
DEM, MES and DCV respectively. The RRFs are calculated for 5 ng
of focusing agent. The reliability of all focusing agents, with the
exception of 2-CEES, was significantly better for GB than GA with
%RSDs of 51%, 3%, 8%, 10%, 21%, and 9% for 2-CEES, DIFP, DEMP,
DEM, MES and DCV respectively over 14 days. In the case of GB,
the focusing agents DIFP, DEMP, DEM, MES, and DCV are an improvement on reliability beyond the BPFB internal standard (32.9%)
[17]. Fig. 6 shows the relative response functions of GB and different focusing agents (a-f) along with the standard error in the
slope. The coefficient of determination (R2 ) and adjusted-R2 values
for relative response functions of GB are 0.997 or better. See Table
S2 in the Supplemental Materials for the slope, standard error in
the slope, R2 and adjusted-R2 values.
3.2.3. GD
RRFs for GD are calculated for 2-CEES, DIFP, DEMP, DEM, MES
and DCV by the relative response functions for day 0. The slopes
of these functions are 0.40, 0.172, 0.213, 0.151, 0.0317, and 0.0499.
The RRFs are 1.983, 0.861, 1.067, 0.754, 0.1587, and 0.2497. The
%RSD over a 14-day aging study reveals that DIFP (18%), DEMP

3.2.5. HD
RRFs for HD are calculated through the relative response functions for day 0. The slopes of these functions are 0.74, 0.32, 0.40,
0.28, 0.059, and 0.093 and the RRFs are 3.71, 1.31, 1.99, 1.41, 0.296,
6


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 6. RRFs for GB for (a, red) 2-CEES, (b, orange) DIFP, (c, yellow) DEMP, (d, green) DEM, (e, blue) MES and (f, purple) DCV for day 0. The six-point quantitative analysis at 1,
2, 5, 10, 20 and 50 ng are denoted by dots, the relative response function are black traces and the colored shadows represent the standard error in the fit. (For interpretation
of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 7. RRFs for GD for (a, red) 2-CEES, (b, orange) DIFP, (c, yellow) DEMP, (d, green) DEM, (e, blue) MES and (f, purple) DCV for day 0. The six-point quantitative analysis at 1,
2, 5, 10, 20 and 50 ng are denoted by dots, the relative response function are black traces and the colored shadows represent the standard error in the fit. (For interpretation
of the references to color in this figure legend, the reader is referred to the web version of this article.)

and 0.466 for 2-CEES, DIFP, DEMP, DEM, MES and DCV respectively.
The %RSD over a 14-day aging study reveals that 2-CEES (29%),
DIFP (32%), DEM (22%), MES (12%) and DCV (23%) are more reliable
than the internal standard, BPFB (32.9%). The only focusing agent
that performed worse than BPFB was DEMP (37%). The linear fit
analysis shows the largest error in the slope for the relative response function for HD and the focusing agents in this work, however this is still an improvement from the average BPFB internal
standard quantification method. Fig. 9 shows the relative response
functions of HD and different focusing agents (a-f) along with the
standard error in the slope. The coefficient of determination (R2 )
and adjusted-R2 values for relative response functions of HD are

0.983 or better. See Table S4 in the Supplemental Materials for the
slope, standard error in the slope, R2 and adjusted-R2 values.
3.3. RRFs and transferability
The acquisition of data on a single instrument provides reliability and reproducibility, however it is often difficult to compare
inter-instrument data with respect to a number of propagating errors [33]. Here, we define calibration transferability as an analytical
method that lacks the strict protocols applied to laboratory-based
analyses through the use of focusing agents and analysis by fieldportable thermal desorption GC-MS. Table 2 shows the RRFs for the
7

J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

Fig. 8. RRFs for GF for (a, red) 2-CEES, (b, orange) DIFP, (c, yellow) DEMP, (d, green) DEM, (e, blue) MES and (f, purple) DCV for day 0. The six-point quantitative analysis at 1,
2, 5, 10, 20 and 50 ng are denoted by dots, the relative response function are black traces and the colored shadows represent the standard error in the fit. (For interpretation
of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 9. RRFs for HD for (a, red) 2-CEES, (b, orange) DIFP, (c, yellow) DEMP, (d, green) DEM, (e, blue) MES and (f, purple) DCV for day 0. The six-point quantitative analysis at 1,
2, 5, 10, 20 and 50 ng are denoted by dots, the relative response function are black traces and the colored shadows represent the standard error in the fit. (For interpretation
of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 2
RRFs for GA, GB, GD, GF, and HD when compared to 5 ng of focusing agent. RRFs were calculated using
Eq. 1 and the slopes of the relative response function for two different HAPSITE® ER systems (Instrument A and Instrument B). The %RSD is calculated [33] from the difference in area responses by each
instrument for all CWAs and focusing agents. Similar RRFs indicate a stable focusing agent across systems.
Instrument A

2-CEES
DIFP
DEMP
DEM
MES
DCV

Instrument B

GA

GB

GD

GF

HD

GA

GB

GD

GF

HD

%RSD

1.20
0.52
0.65
0.46
0.10
0.15

0.89
0.39
0.48
0.34

0.07
0.11

1.98
0.86
1.07
0.75
0.16
0.25

10.73
4.66
5.75
4.08
0.86
1.35

3.71
1.61
1.99
1.41
0.30
0.47

0.90
0.62
0.66
0.71
0.14
0.19

0.66
0.46
0.49
0.52
0.11
0.14

1.12
0.77
0.83
0.88
0.18
0.24

6.24
4.30
4.64
4.92
0.99
1.33

2.53
1.75
1.87
2.00
0.40
0.54

21%

6%
6%
16%
13%
7%

8


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

data provided by a HAPSITE® ER (Instrument A) and an additional
system to demonstrate the independence of this method of calibration, i.e. calibration transferability. The data from Sections 3.2.15 suggests that the most reliable focusing agents for GA, GB, GD,
GF, and HD are DIFP, DEMP and DEM on a single instrument, but
Table 2 shows that the DIFP and DEMP are better focusing agents
for comparing inter-instrument quantified values.

Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at doi:10.1016/j.chroma.2020.461784.
References
[1] R.J. Brennan, J.F. Waeckerle, T.W. Sharp, S.R. Lillibridge, Chemical warfare
agents: emergency medical and emergency public health issues, Ann. Emerg.
Med. 34 (2) (1999) 191–204.
[2] G. Hurst, Medical Management of Chemical Casualties Handbook, Government
Printing Office, 2015.
[3] E. Raber, T.M. Carlsen, K.J. Folks, R.D. Kirvel, J.I. Daniels, K.T. Bogen, How clean
is clean enough? recent developments in response to threats posed by chemical and biological warfare agents, Int. J. Environ. Health Res. 14 (1) (2004)

31–41.
[4] M. Sinha, A. Tomassian, Development of a miniaturized, light-weight magnetic
sector for a field-portable mass spectrograph, Rev. Sci. Instrum. 62 (11) (1991)
2618–2620.
[5] S.A. Ecelberger, T.J. Cornish, B.F. Collins, D.L. Lewis, W.A. Bryden, Suitcase TOF:
a man-portable time-of-flight mass spectrometer, Johns Hopkins APL Tech. Dig.
25 (1) (2004) 14–19.
[6] T. Spaeder, R. Walton, Hapsite GC/MS analysis of low level chemical warfare
agents, Toxic Industrial Compounds and Toxic Industrial Materials in the Field
by First Responders, Abstracts of Papers of the American Chemical Society,
2003 U128-U128.
[7] D.T. Snyder, C.J. Pulliam, Z. Ouyang, R.G. Cooks, Miniature and fieldable mass
spectrometers: recent advances, Anal. Chem. 88 (1) (2016) 2–29.
[8] Z. Ouyang, R.G. Cooks, Miniature mass spectrometers, Annu. Rev. Anal. Chem.
2 (2009) 187–214.
[9] J. Liigand, A. Kruve, P. Liigand, A. Laaniste, M. Girod, R. Antoine, I. Leito, Transferability of the electrospray ionization efficiency scale between different instruments, J. Am. Soc. Mass Spectrom. 26 (11) (2015) 1923–1930.
[10] W. Lotz, An empirical formula for the electron-impact ionization cross-section,
Z. Physik 206 (2) (1967) 205–211.
[11] W.L. Fitch, A.D. Sauter, Calculation of relative electron impact total ionization
cross sections for organic molecules, Anal. Chem. 55 (6) (1983) 832–835.
[12] K. Demeestere, J. Dewulf, K. De Roo, P. De Wispelaere, H. Van Langenhove,
Quality control in quantification of volatile organic compounds analysed by
thermal desorption–gas chromatography–mass spectrometry, J. Chromatogr. A
1186 (1-2) (2008) 348–357.
[13] S. Popiel, M. Sankowska, Determination of chemical warfare agents and related
compounds in environmental samples by solid-phase microextraction with gas
chromatography, J. Chromatogr. A 1218 (47) (2011) 8457–8479.
[14] D. Patterson Jr, L. Alexander, L. Gelbaum, R. O’Connor, V. Maggio, L. Needham,
Synthesis and relative response factors for the 22 tetrachlorodibenzo-p-dioxins (TCDDs) by electron-impact ionization mass spectrometry, Chemosphere 15
(9-12) (1986) 1601–1604.

[15] K. Lundgren, C. Rappe, M. Tysklind, Low-resolution mass spectrometric relative
response factors (RRFs) and relative retention times (RRTs) on two common
gas chromatographic stationary phases for 87 polychlorinated dibenzofurans,
Chemosphere 55 (7) (2004) 983–995.
[16] H. Sekiguchi, K. Matsushita, S. Yamashiro, Y. Sano, Y. Seto, T. Okuda, A. Sato,
On-site determination of nerve and mustard gases using a field-portable gas
chromatograph-mass spectrometer, Forensic Toxicol. 24 (1) (2006) 17–22.
[17] H. Nagashima, T. Kondo, T. Nagoya, T. Ikeda, N. Kurimata, S. Unoke, Y. Seto,
Identification of chemical warfare agents from vapor samples using a field-portable capillary gas chromatography-membrane-interfaced electron ionization quadrupole mass spectrometry instrument with tri-bed concentrator, J.
Chromatogr. A (2015) (1406) 279–290.
[18] R.M. Black, R.J. Clarke, D.B. Cooper, R.W. Read, D. Utley, Application of
headspace analysis, solvent extraction, thermal desorption and gas chromatography—mass spectrometry to the analysis of chemical warfare samples containing sulphur mustard and related compounds, J. Chromatogr. A 637 (1)
(1993) 71–80.
[19] W.A. Carrick, D.B. Cooper, B. Muir, Retrospective identification of chemical
warfare agents by high-temperature automatic thermal desorption–gas chromatography–mass spectrometry", J. Chromatogr. A 925 (1-2) (2001) 241–249.
[20] J. Oostdijk, C. Degenhardt, H. Trap, J. Langenberg, Selective and sensitive
trace analysis of sulfur mustard with thermal desorption and two-dimensional
gas chromatography–mass spectrometry", J. Chromatogr. A 1150 (1-2) (2007)
62–69.
[21] O. Terzic, I. Swahn, G. Cretu, M. Palit, G. Mallard, Gas chromatography–full scan
mass spectrometry determination of traces of chemical warfare agents and
their impurities in air samples by inlet based thermal desorption of sorbent
tubes, J. Chromatogr. A 1225 (2012) 182–192.
[22] Y. Juillet, C. Dubois, F. Bintein, J. Dissard, A. Bossee, Development and validation of a sensitive thermal desorption–gas chromatography–mass spectrometry
(TD-GC-MS) method for the determination of phosgene in air samples", Anal.
Bioanal. Chem. 406 (21) (2014) 5137–5145.
[23] M. Kanamori-Kataoka, Y. Seto, Measurement of breakthrough volumes of
volatile chemical warfare agents on a poly(2, 6-diphenylphenylene oxide)-based adsorbent and application to thermal desorption–gas chromatography–mass spectrometric analysis, J. Chromatogr. A 1410 (2015) 19–27.

4. Conclusion

The conclusive remarks that can be made upon the completion
of this work are the overall improvement of thermal desorption capabilities on the HAPSITE® ER for CWAs. Focusing agents not only
provide an accurate methodology for intra-instrument calibration
but also provides calibration transportability for inter-instrument
comparison. The inconsistencies in the performance that were previously identified (i.e. variations in internal standard peak areas
and daily response) [29] have been improved by the implementation of focusing agents. Seto previously compares the HAPSITE®
ER to other point sensor technologies which compares most closely
with electrochemical sensors (limits of alarms in the sub mg/m3
range) [34]. The use of focusing agent in the field leads to more
accurate and reliable quantification of GA, GB, GD, GF and HD than
the traditional internal standard (BPFB) within a single day as well
as across a multiple day timespan. BPFB has shown to have as high
as 32.9 %RSD within a single day and 26.3 %RSD between days
[17]. Aging effects are tabulated in the Supplemental Material (Table S6) in term of %RSD with respect to time. Future work will establish measurement detection limits (MDLs) for the HAPSITE® ER
and relative response factors for VX and Russian VX using focusing agents. While focusing agents have been demonstrated against
chemical warfare agents in our laboratory settings, they have yet to
be reported for field use. Future work expands upon this work by
extending experiments beyond laboratories and into the intended,
more harsh environments.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.
CRediT authorship contribution statement
John T. Kelly: Writing - original draft, Writing - review & editing, Data curation, Visualization. Anthony Qualley: Writing - original draft, Writing - review & editing, Data curation, Visualization.
Geoffrey T. Hughes: Writing - original draft, Writing - review &
editing, Data curation, Visualization. Mitchell H. Rubenstein: Conceptualization, Methodology, Writing - original draft, Writing - review & editing, Project administration. Thomas A. Malloy: Conceptualization, Formal analysis. Tedeusz Piatkowski: Conceptualization, Formal analysis.
Acknowledgements
We would like to thank the United States Air Force for funding as well as Drs. Darrin Ott and Claude C. Grigsby for their support and encouragement. We recognize Mr. Will Bell and Dr. R.
Craig Murdoch who provided program and financial management.
The views expressed in this article are those of the author and do

not reflect the official policy or position of the United States Air
Force, Department of Defense, or the U.S. Government. This work
has been approved for public distribution (Distribution-A, Public
88ABW-2020-2344).
9


J.T. Kelly, A. Qualley, G.T. Hughes et al.

Journal of Chromatography A 1636 (2021) 461784

[24] S.W. Harshman, V.L. Dershem, M. Fan, B.S. Watts, G.M. Slusher, L.E. Flory,
C.C. Grigsby, D.K. Ott, The stability of tenax TA thermal sesorption tubes in
simulated field conditions on the HAPSITE® ER, Int. J. Environ. Anal. Chem. 95
(11) (2015) 1014–1029.
[25] L. Ettre, Nomenclature for Chromatography (IUPAC Recommendations 1993),
Pure Appl. Chem. 65 (4) (1993) 819–872.
[26] A. Qualley, G.T. Hughes, M.H. Rubenstein, Data quality improvement for field-portable gas chromatography-mass spectrometry through the use of isotopic analogues for in-situ calibration", Environ. Chem. 17 (1) (2020) 28–
38.
[27] S.E. Stein, An integrated method for spectrum extraction and compound identification from gas chromatography-mass spectrometry data, J. Am. Soc. Mass
Spectrom. 10 (8) (1999) 770–781.
[28] W.G. Mallard, AMDIS in the chemical weapons convention, Anal. Bioanal.
Chem. 406 (21) (2014) 5075–5086.

[29] S.W. Harshman, M.H. Rubenstein, A.V. Qualley, M. Fan, B.A. Geier, R.L. Pitsch,
G.M. Slusher, G.T. Hughes, V.L. Dershem, C.C. Grigsby, Evaluation of thermal
desorption analysis on a portable GC–MS system", Int. J. Environ. Anal. Chem.
97 (3) (2017) 247–263.
[30] R.A. Pacer, Quantitative gas chromatography using peak heights and relative
response factors. an undergraduate student experiment, J. Chem. Educ. 53 (9)

(1976) 592.
[31] H.L. Pardue, M.F. Burke, J.R. Barnes, Quantitative gas chromatography, J. Chem.
Educ. 44 (11) (1967) 695.
[32] F. Karasek, E. De Decker, J. Tiernay, Qualitative and quantitative gas chromatography for the undergraduate, J. Chem. Educ. 51 (12) (1974) 816.
[33] G.L. Long, J.D. Winefordner, Limit of detection. A closer look at the IUPAC definition, Anal. Chem. 55 (7) (1983) 712A–724A.
[34] Y. Seto, On-site detection of chemical warfare agents, in: Handbook of Toxicology of Chemical Warfare Agents, Elsevier, 2015, pp. 897–914.

10