Sensitive Analysis of Nitroguanidine in Aqueous and Soil Matrices by LC-MS

Article abstract of DOI:10.1021/acs.analchem.7b02364

Nitroguanidine, a widely used nitramine explosive, is an environmental contaminant that is refractory, persistent, highly mobile in soils and aquifers, and yet under-researched. Nitroguanidine determination in water and soil poses an analytical challenge due its high hydrophilicity, low volatility, charge neutrality over a wide pH range, and low proton affinity which results in low electrospray interface (ESI)-MS sensitivity. A sensitive method for the determination of nitroguanidine in aqueous and soil matrices was developed. The method is based on reduction by zinc in acidic solution, hydrophobization by derivatization, preconcentration on C18 cartridge, and LC-MS quantification. The demonstrated limit of detection (LOD) reaches 5 ng/L and 22 ng/g in water and soil, respectively. Analysis of a contaminated site demonstrates that it is possible to map a contamination plume that extends over 1 km from the source of the contamination.

Full text of DOI:10.1021/acs.analchem.7b02364

Article
pubs.acs.org/ac
Sensitive Analysis of Nitroguanidine in Aqueous and Soil Matrices by
LC-MS
Anna Voloshenko Rossin,§ Sergey Sladkevich,§ Guy Gasser,§ Artem Melman,† and Ovadia Lev*,§
§The Casali Center, The Institute of Chemistry, The Hebrew University of Jerusalem, Edmond J. Safra Campus, Jerusalem 91904,
Israel
†Department of Chemistry and Biomolecular Science, Clarkson University, 8 Clarkson Ave., Potsdam, New York 13699-5810, United
States
S
* Supporting Information
ABSTRACT: Nitroguanidine, a widely used nitramine ex-
plosive, is an environmental contaminant that is refractory,
persistent, highly mobile in soils and aquifers, and yet under-
researched. Nitroguanidine determination in water and soil
poses an analytical challenge due its high hydrophilicity, low
volatility, charge neutrality over a wide pH range, and low
proton affinity which results in low electrospray interface (ESI)-
MS sensitivity. A sensitive method for the determination of
nitroguanidine in aqueous and soil matrices was developed. The
method is based on reduction by zinc in acidic solution,
hydrophobization by derivatization, preconcentration on C18
cartridge, and LC-MS quantification. The demonstrated limit of
detection (LOD) reaches 5 ng/L and 22 ng/g in water and soil,
respectively. Analysis of a contaminated site demonstrates that it is possible to map a contamination plume that extends over 1
km from the source of the contamination.
days.7,14−18 Thus, NQ can persist in the environment,
itroguanidine (NQ) is a nitramine explosive, which is
Noften used in combination with 1,3,5-trinitroperhydro- especially in groundwater sources, where the absence of
photolysis and anoxic conditions can hinder NQ degradation.
Thus, the large production rate and its high mobility make
this compound a potential risk to underground water resources.
The low retardation in the ground implies that it can also serve
as a potential indicator for propagation of aquifer pollution
plumes created by the military industry.
1,3,5-triazine (RDX), nitrocellulose, or nitroglycerin as a
double- or triple-based propellant. It was demonstrated that
NQ can be used as an ingredient in an insensitive munitions
formulation (IMX-101) aimed to replace 2,4,6-trinitrotoluene
(TNT) with safer energetic materials that are more stable
against detonation stimuli.1−3 Thus, NQ is an intensively used
material that can be found in the areas adjacent to military
facilities, where live-fire exercises are held.4 It is produced in
large quantities; the annual production of nitroguanidine was
estimated at 8600 metric tons in 2016.5
However, NQ cannot be easily extracted from water, and
thus, the sensitivity of all of the current analytical techniques is
rather limited (Table 1). Most of these methods are based on
HPLC-UV/vis analysis, which has low sensitivity with a limit of
detection (LOD) of 100 μg/L or higher for water samples, with
the exception of the work of Walsh19 reporting a LOD of 5 μg/
L. The specificity of HPLC with UV detection is inferior to
mass spectrometric analysis. A preconcentration by rotary
evaporation followed by electrochemical analysis was proposed
by Manning and Maskarinec,20 but despite the inherent
sensitivity of electrochemical techniques, the LOD for NQ by
this technique is as high as 40 μg/L. The most remarkable
performance reported to date was published in this journal by
Mu et al.21 Reverse phase UPLC (using methanol−water with 1
mM ammonium acetate) coupled with 4000Q TRAP triple
NQ is highly soluble in water (4.4 g/L at 25 °C)6 and has a
low soil−water partition coefficient of 0.15−0.43 L/kg.7,8 In
addition, NQ is not charged between pH −0.5 and 12.2,9,10 and
as such, it is highly mobile in aqueous systems and soils.
Although the reduction of NQ can yield nitrosoguanidine,
which was shown to be carcinogenic in rats,11 it is classified by
the USEPA as a group D substance, i.e., a compound that is not
classifiable as a human carcinogen.12
The research on NQ stability in water is rather limited. The
reported products of abiotic photocatalytic NQ degradation are
guanidine, hydroxyguanidine, cyanoguanidine, urea, ammonia,
and nitrosoguanidine.7,13 The biodegradation of NQ in aerobic
conditions results in formation of nitrourea,14,15 which can
further decompose to NH3, N2O, and CO2.15 The reported
degradation time scale varies from several hours to 85
Received: June 20, 2017
Accepted: July 19, 2017
Published: July 19, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Table 1. Reported Methods for Analysis of NQ
LOD
soil [μg/g]
method of analysis and
water
[μg/L]
detection
column type
comments
reference
HPLC-UV
NP: neutral alumina
RP: ODS
800
100
100
40
254 nm
263 nm
254 nm
22
23
18
20
HPLC-UV
HPLC-UV
HPLC-EC
RP: C8
RP: C18
preconcentration by rotary
evaporation
HPLC-UV
mixed mode: C18/cation exchange
RP: Hypercarb
5
0.5
263 nm
19
24
25
a
HPLC-UV
286 nm
LC-ESI (+)-MS
RP column 1: Synergi hydroRP; RP column 2: Pinnacle II
biphenyl
9
UFLC-ESI (−)-MS/MS
HPLC-UV
RP: Knietex C18
0.7
2000
21
26
Synergi 4 μm Polar-RP 80
0.95−1.74
263 nm two-stage soil
extraction
a
LOD was not established; the method was used for dissolution study only.
a
Scheme 1. Schematic Illustration of NQ Reduction Followed by Derivatization with 4-NBA
a
Below pH 3 only the trans-isomer is obtained.
quadrupole mass spectrometry was used to analyze NQ and
other nitro compounds. The authors reported a LOD of 0.7
μg/L for water samples. This, however, may be regarded as too
high since, for comparison, the allowable level of any single
pesticide in the EU drinking water is 0.1 μg/L. The situation is
even more complicated, since NQ is eluted very close to the
unretained peak, and thus, analysis of high ionic strength water
is expected to result in ion suppression.
In this research, we present a new methodology for
determination of NQ based on reduction of NQ to
aminoguanidine (AQ) by hydrogenation on zinc under acidic
conditions, followed by derivatization by 4-nitrobenzaldehyde
(4-NBA) to give a Schiff base 2-[(4-nitrophenyl)methylene]-
hydrazinecarboximidamide (AQ-4NBA, Scheme 1). The latter
can be preconcentrated on a C18 SPE cartridge prior to LC-MS
analysis.
etry. The synthesis of AQ-4NBA was carried out according to
ref 28. The identity of the product was confirmed using high
resolution mass spectrometry and 1H NMR spectroscopy.
Stock solutions (1 mg/mL) of NQ and 15N-labeled NQ were
prepared by dissolving the compounds in ethanol/water 80:20
(v/v). Stock standard solution of AQ (1 mg/mL) was prepared
by dissolving the compound in HPLC-grade methanol.
Working solutions of the standards were prepared by dilution
of the respective stock solutions with water. All solutions were
freshly prepared or kept refrigerated for less than 1 week prior
to the analysis.
Solid-phase extraction (SPE) cartridges (500 mg, 6 mL)
Strata C18-E (55 μm, 70A) were purchased from Phenomenex.
The nitrogen used for drying the solid-phase cartridges and for
evaporation of solvents was of 99.995% purity from Maxima,
Israel. Ultrapure water was obtained by a Millipore laboratory
water purification system.
EXPERIMENTAL SECTION
Instrumentation. LC-MS-MS quantifications were per-
formed using an Agilent 1200 high performance liquid
chromatograph (HPLC) coupled via an electrospray interface
(ESI) to an Agilent 6410A triple quadrupole MS (QQQ-MS).
Agilent 1100 HPLC coupled via an ESI to an Agilent 6520A
QTOF MS was used for conclusive identifications.
■
Materials. The following solvents and chemicals were
purchased and used without further purification. Methanol
(HPLC grade), acetonitrile (HPLC grade), formic acid, acetic
acid glacial, ethanol absolute (dehydrated) ULC/MS, and
sodium hydroxide were supplied by Bio-Lab LTD (Israel). 4-
Nitrobenzaldehyde 99% (4-NBA) was obtained from Alfa
Aesar; zinc powder was supplied by Fisher Scientific, and
aminoguanidine hydrochloride (≥98%), acridine (≥97%),
nitroguanidine, and semicarbazide hydrochloride (SEM)
(≥99%) were obtained from Sigma-Aldrich (Israel).
1H NMR spectra of samples in D2O were collected on a
Bruker Avance-400 (9.4 T) spectrometer. The measurements
were performed using a single pulse sequence with rf pulse
duration of 10 μs and recycling time of 9−60 s.
Single crystal X-ray analysis for the synthesized AQ-4NBA
was performed on a Bruker SMART APEX CCD diffractometer
using graphite monochromatized Mo Kα radiation (λ =
0.71073 Å). The structure was solved by direct methods and
15N-labeled nitroguanidine was synthesized in our laboratory
according to ref 27. The identity and purity (>99.5%) of the
product was confirmed using high resolution mass spectrom-
B
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Table 2. LC-MS/MS Parameters
product ions (m/z)
compound
AQ-4NBA
retention time (min)
precursor ion
208.3
quantifier
qualifier
collision energy (eV)
fragmentation voltage (V)
110
191.3
15
20
15
15
15
40
55
10
15
20
7.8
162.3
149.3
194.3
178
15N-labeled AQ-4NBA
acridine
7.8
7.4
212.3
180
120
130
150.2
151
192
SEM
8.7
209.2
149
103
120
refined by full matrix least-squares on F2. Details of single
crystal X-ray studies are given in Table S1.
mL/min until t = 25 min. Finally, the composition of the eluent
was changed gradually to initial conditions at t = 17 min and
remained until t = 25 min for stabilization of the column before
the next injection. 1.5 min was set for post run at initial
conditions. Injection volume was 15 μL, unless otherwise
specified. The MS parameters are presented in Table 2.
Quantification was carried out with isotopically labeled internal
standard and by multipoint calibrations with good linear
correlation (R2 = 0.9999 in the range of 1−500 μg/L).
Extraction of NQ from Soil. The analysis of NQ in soils
was carried out on the basis of the soil extraction protocol
reported in ref 19, using the calibration curve with 15N-labeled
NQ as internal standard. Briefly, 5 g of the examined soil was
weighed into a 50 mL centrifuge tube, and 2 mL of deionized
water was then added. The sample was shaken in the dark at
room temperature. After 20 h, the sample was transferred to a
round-bottom flask for lyophilization. After the sample was
completely dried, 10 mL of deionized water was added to each
sample and the mixture was transferred to an ultrasonic water
bath (GT Sonic, 40 kHz) maintained at less than 25 °C for 2 h.
The extracted sample was then centrifuged and filtered through
a 0.45 μm PVDF membrane filter. NQ content in obtained
water extracts was estimated according to the analytical
procedure of NQ analysis in water samples described above.
Soils were analyzed in triplicates.
Sampling and Storage. Water samples were collected in
amber glass bottles from several water wells in the vicinity of an
abandoned industrial site in Israel. The water samples were
transported and kept under refrigeration (2−4 °C). Extraction
was conducted within 120 h of sampling. Extracts were stored
for less than 1 week in amber glass vials in a refrigerator.
Soil samples were collected at different depths of a secondary
effluent recharge of Soil Aquifer Treatment (SAT) plant
managing domestic wastewater in the Tel-Aviv region (Israel).
Different depths of the SAT are characterized by different
lithology and water and organics content. In addition to the
analysis of soils prior to NQ enrichment, we have carried out
analysis of the effluent collected from the relevant wastewater
treatment facility to show that the water treatment system and
the soil do not contain any NQ or AQ.
Analytical Procedure of NQ Analysis in Water
Samples. Reduction of NQ. NQ was reduced to AQ by zinc
powder in acidic solution. Glacial acetic acid (0.3 mL), zinc
powder (200 mg), and 15N-labeled-NQ internal standard (10
μL of the 1 mg/L) were added to 10 mL of water sample. The
mixture was stirred for 20 min at room temperature and then
filtered through a 0.45 μm PVDF syringe filter (Millex-HV,
Merck-Millipore).
Derivatization of AQ. The AQ solution obtained in the
previous step was reacted with 0.2 mL of 4-NBA (10 mg/mL)
to form AQ-4NBA. SEM was added as a surrogate standard for
the derivatization process. The mixture was shaken for 3 h, and
then, the derivate was extracted by SPE using a C18 cartridge.
The cartridge was preconditioned with 6 mL of methanol and 6
mL of water consecutively. The sample was loaded into the
cartridge and left to percolate by gravity followed by the
cartridge drying step under vacuum for 2 min. Analyte (AQ-
4NBA) was subsequently eluted with 6 mL of methanol, and
the obtained extract was concentrated to a final volume of 0.5
mL by gentle nitrogen flow at 55 °C. The final sample was
filtered through a 0.22 μm PVDF syringe filter (Millex-GV,
Merck-Millipore). Five μL of a 5 mg/L acridine solution was
added as an instrument internal standard.
LC analyses were performed on an Agilent ZORBAX Eclipse
Plus C18 (2.1 mm ID, 100 mm length, and 3.5 μm particle
size). Column temperature was set at 25 °C. The mobile phase
consisted of 10% acetonitrile, 90% H2O, and 0.1% formic acid.
The eluent composition was as follows: initial conditions, 10%
aetonitrile fed at 0.2 mL/min for 1 min. At t = 10 min, the flow
rate was increased to 0.3 mL/min, and the composition was
ramped to 100% acetonitrile, maintaining a flow rate of 0.3
RESULTS AND DISCUSSION
■
Preliminary Studies. Preliminary studies were carried out
to examine the possibility of direct ESI-MS analysis of NQ. The
obtained LOD of NQ was higher than 5 μg/L. The attempts to
extract NQ from water in a wide pH range using various
commercial SPE cartridges did not provide reasonable recovery
(additional details are specified in the Supporting Information).
Then, an alternative approach involving functionalization of
the NQ in the aqueous solution was pursued to attain better
affinity to SPE cartridges (further details on the unsuccessful
protocols can be found in the Supporting Information). We
decided to reduce NQ to AQ which has higher MS visibility
and is more amenable for derivatization. Different reduction
approaches were tested (further details on the unsuccessful
reduction protocols can be found in the Supporting
Information). By far, the most successful approach involved
the reduction of NQ by Zn in acidic solution.29 Parametric
dependence of this method is presented below. Preconcentra-
tion of the protonated AQ on a cation exchange SPE cartridge
failed due to the presence of Zn2+, and a protocol involving
derivatization by a Schiff base reagent and SPE preconcentra-
C
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
tion for sensitive determination of the AQ product was
developed.
Method Optimization. The optimized method of analysis
is detailed in the Experimental Section, and only pertinent
parametric dependencies are discussed here.
Reduction Step. NQ reduction was carried out by addition
of 20 mg/mL zinc powder and 3% (v/v) acetic acid. Synthesis
of amino compounds by reduction of the respective nitro
compounds on zinc under acidic conditions was discussed by
Gowda et al.30 who showed that reduction time (using 90%
acetic acid and 78.5 mg/mL zinc) of 20 different nitro-
compounds ranged between 2 and 20 min, though NQ
reduction was not studied. The dependence of the observed
relative LC-MS signal of AQ-4NBA on reaction time is
depicted in Figure 1. A monotonous increase of the yield with
Figure 2. Dependence of the relative response of the trans- (dark
shade) and cis-isomers. Bottom frame depicts chromatograms of the
two isomers of AQ-4NBA at different pH values. The cis-isomer is the
first eluting peak.
The observed ratio between the two chromatographic peaks
was similar to the ratio between the trans- and cis-isomers of
AQ-4NBA in the 1H NMR study in D2O. The 1H NMR of the
trans-isomer showed two apparent doublets at 7.89 and 8.27
ppm and a singlet at 8.12 ppm, which were attributed to the
hydrogen atoms on C5, C3, C2, C6, and C7, respectively
(Figure S2). 1H NMR after the irradiation showed two
additional apparent doublets at 7.77 and 8.36 ppm and a
singlet at 7.82 ppm which were attributed to the hydrogens on
C5, C3, C2, C6, and C7 atoms, respectively (Figure S2). On
the basis of these observations, the first chromatographic peak
was attributed to the cis-isomer and the second peak to the
trans-isomer (Figure S3, Scheme 1). Single crystal X-ray
analysis was carried out in order to get additional information
about isomer structure of the synthesized AQ-4NBA crystals.
Figure S4 represents the structure of the molecule (as AQ-
4NBA acetate), confirming the formation of trans-isomer in the
reaction between AQ and 4-NBA. The presence of dissolved
zinc at pH 3 did not change the method performance, and
therefore, we could analyze the AQ-4NBA at pH 3 without zinc
removal.
Derivatization. The derivatization process of AQ was based
on the method described by Beaven et al.32 Schiff base
reactions are frequently used for coupling of amines and
aldehydes or ketones in aqueous solutions. We screened
different derivatization reagents including 3,4-dihydroxybenzal-
dehyde, heptaldehyde, 3-pyridinecarboxaldehyde, methyl ethyl
ketone, acetone, cyclohexanecarboxaldehyde, and 4-nitro-
benzaldehyde. The relative response levels of the different
products by direct analysis by LC-QTOF-MS are depicted in
Figure 3. The highest responses were obtained after
derivatization with either acetone or 4-NBA. We preferred
the latter to endow hydrophobicity, since SPE of the 1-
(propan-2-ylideneamino)guanidine (acetone derivatization
product) had lower recovery compared to the 4-NBA
derivatization product.
Figure 1. Dependence of the MS response level of AQ-4NBA on the
reduction time of NQ by zinc (3.5 μg/L NQ was used). Bars delineate
standard deviation of triplicates.
time of reaction was observed, and 20 min was selected as an
optimal reaction time for the reduction process. Note that the
initial reduction step is rather rapid and almost 50% of the
reduction is completed within less than 1 min, which may be
used when rapid analysis is desirable.
Zinc Cleanup. Comparative tests were made in order to
determine whether the production of Zn(II) during the
reduction step interferes with the analysis of NQ. The zinc
removal process (by increasing pH with sodium hydroxide
resulting in precipitation of zinc hydroxide31) did not change
the LOD or the recovery, but two chromatographic peaks (at
retention time tr = 5.7 and 7.8 min), instead of a single one at
7.8 min, were obtained. These two peaks with identical m/z
([M + H]+ = 208.0829) and MS/MS fragmentation were
assigned to cis- and trans-isomers of AQ-4NBA product. In
order to clarify the nature of these isomers, AQ-4NBA was
synthesized separately and LC-MS analysis of the product as a
function of pH was performed. At pH 3, we observed a peak (tr
= 7.8 min), constituting 100% of the material, and at pH higher
than 5, the two peaks were observed at the ratio of ca. 2:3
(Figure 2).
NMR studies were carried out in order to glean insight into
1
the identity of the two chromatographic peaks. However, H
NMR of the synthesized AQ-4NBA in D2O revealed only a
spectrum of the trans-isomer (Figure S1). To synthesize cis-
AQ-4NBA, we used UV irradiation (300 W Xe lamp, Oriel) of
a solution of the synthesized AQ-4NBA in D2O. The 1H NMR
spectrum of irradiated sample revealed a mixture of trans-
(major) and cis- (minor) isomers of AQ-4NBA in 1.0:0.43
ratio. The LC-MS of the solution soon after the irradiation
revealed two chromatographic peaks both having the same
mass and retention times that were observed in the pH study
(vide supra, Figure 2).
We optimized the reaction time for derivatization of AQ. For
this purpose, we analyzed 3.5 μg/L NQ according to the full
procedure, including reduction by zinc and derivatization with
200 mg/L 4-NBA, varying only the reaction time. Figure 4
shows product formation as a function of time. The tests were
conducted in triplicates, and their average and standard
deviations are presented. Optimal derivatization time under
these conditions was 3 h; further increase of the reaction time
D
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Recovery Tests. An analysis of the recovery loss at each of
the different analytical steps was carried out. The recovery loss
at the reduction step was evaluated by 3-points standard
addition of AQ to the vial after reduction of NQ (2.5 μg/L) by
the optimized zinc reduction protocol. Calculated recovery in
triplicate was 97 12%.
The recovery after the derivatization step was evaluated by a
5-point calibration curve for home synthesized AQ-4NBA
standard in the range of 25−150 μg/L prepared using blank
water, which was subjected beforehand to the full reduction−
derivatization process (in order to account for possible matrix
effects). To calculate derivatization recovery, 1.8 μg/L AQ
standard was passed through an optimized derivatization
protocol (with 4-NBA) and analyzed. Calculated recovery in
triplicate was 83 7%.
Figure 3. Relative response of AQ after derivatization with different
Schiff base reagents.
The recovery of the SPE extraction step was evaluated by
comparison of the response of 5 μg/L AQ-4NBA that was
spiked in distilled water that was previously subjected to the
zinc reduction and derivatization protocol. Then, SPE
extraction was conducted, and the response was compared to
a calibration curve describing the direct response of AQ-4NBA
standard in 60:40 methanol/water solution at pH 3. The
calibration curve was measured in the range of 25−150 μg/L.
Obtained recovery was 91 11%. The overall recovery of the
whole process starting with NQ spiked distilled water was 73
14%.
Figure 4. MS signal dependence on derivatization time (for 1.7 μg/L
NQ standard). Bars indicate standard deviation of triplicates.
Analysis of a Contaminated Site. In order to examine the
applicability of the new method for real water samples, we have
analyzed water wells in the vicinity of an industrial site that was
contaminated by NQ. Eight water wells at various distances and
directions from the focal point of contamination were examined
(Figure 5). The analysis was carried out by three different
modes: (i) direct injection, i.e., analysis of NQ by direct
injection of the sample to HPLC-QQQ-MS without reduction
and derivatization and quantification by the calibration curve;
(ii) optimized reduction−derivatization, i.e., quanitification of
NQ by the method reported here, using the calibration curve
did not improve the obtained signal. However, it was possible
to decrease the derivatization time to 30 min by carrying out
the derivatization reaction at 70 °C (relative response was 99
7% of the response obtained after 3 h of derivatization at room
temperature). This protocol is, however, more complicated and
requires an additional heating bath, and therefore, the
optimized protocol was based on 3 h of derivatization at
room temperature.
Increasing 4-NBA dose can reduce the derivatization time,
but the unreacted 4-NBA interferes with the extraction and MS
analysis. Therefore, 200 mg/L 4-NBA was chosen as a
compromise between lower reaction time and lower analytical
interferences of the unreacted reagent.
Method Validation. The performance of the proposed
method was established by employing assays with sample
blanks and spiked samples. Linearity, matrix effects, selectivity,
trueness, precision, and detection limit were studied.
Quantifications were carried out by multipoint calibration
with 15N-labeled NQ as internal standard. An 8-points
calibration curve (Figure S5) was prepared in the range of
1−500 μg/L by addition of NQ standards to deionized water
and carrying out the full process, including reduction,
derivatization, and extraction in triplicates. A good linear
correlation was observed with correlation coefficient R2
=
0.9999. The LOD of the analytical method, calculated at 3
times the noise level, was 10 ng/L for a 15 μL injection or 5
ng/L for a 30 μL injection. Since the injected sample contained
over 60% water, the peak maintained its nearly Gaussian shape
even for a 30 μL injection. Peak symmetry, computed as the
ratio between the front half-width and the back half-width,33
was 1.6 (the respective calibration curve and the chromato-
graphic peaks shape for a 30 μL injection are depicted in Figure
S5). These limits of detection are 70 and 140 times lower than
previously reported for 15 and 30 μL per injection, respectively
(Table 1).
Figure 5. Scheme of the relative location of the studied water wells
(designated by letters a−h) in the vicinity of the examined site.
Numbers in the scheme report NQ concentration in the
corresponding wells (in μg/L). The concentrations of NQ and other
contaminants are delineated in the table.
E
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
and 15N-labeled NQ as internal standard; (iii) standard
addition; i.e., quantification of NQ for the low concentration
range was also carried out by reduction−derivatization but with
4 point standard addition calibration. Table 3 summarizes the
Table 4. LOD and Spike Recovery of NQ in Different Soil
Matrices
soil A
sand
soil B
soil C
soil D
sand
type
clay
sandy−clay
93% sand, 19% sand, 14% 60% sand, 10% 89% sand, 7%
7% silt silt, 67% clay silt, 30% clay silt, 4% clay
Table 3. Comparison of NQ Analyses of Water Samples from
the Studied Wells by Direct Injection (Without Reduction
and Derivatization, Method (i)), by Optimized Reduction−
Derivatization (Method (ii)), and by Standard Addition
depth
3 m 9 m 16.5 m 18 m
organic
matter
0.05%
0.36%
0.2%
0.12%
a
(Method (iii))
average
SD
[ng/g]
4.9 1.0
5.3 0.8
4.9 0.7
4.9 0.2
NQ concentration [μg/L]
recovery % 97 20
105 16
21
99 14
20
98
22
5
point
name
(i) direct water sample
injection
(ii) calibration
curve
(iii) standard
addition
(n = 3)
LOD
[pg/g]
22
a
b
c
d
e
f
43.6 3.0
ND
42.7 2.7
0.010 0.004
ND
NA
0.010 0.002
ND
ND
b
3.7 0.4
8.1 0.4
0.15 0.05
0.31 0.02
ND
8.5 0.8
0.20 0.07
0.33 0.04
ND
various soil blanks. Table 4 depicts the test results for the
different samples. The recovery was in the range of 97.3−
105.2% with RSD of triplicates in the range of 5−20%. The
LOD of the method was in the range of 20−22 pg/g.
ND
ND
g
h
ND
81.8 6.5
126.2 3.5
NA
a
The point names (a−h) designate the sampling well location
b
CONCLUSIONS
depicted in Figure 5. ND: not detected; NA: not analyzed. The
averaged value is below the LOD.
■
A highly sensitive method for the determination of NQ, a
widely used nitramine explosive, was developed on the basis of
reduction, hydrophobization, and SPE preconcentration
followed by LC-MS analysis. The method exhibits high
recovery and low limit of detection and is applicable for the
analysis of water samples as well as soils.
results of the different analyses. NQ was found only in 3 water
wells by direct injection (method (i), above) and in 6 out of the
8 water wells that were analyzed by method (ii). Good fit was
observed between method (ii) and method (iii) (Figure S6).
The slope of the fit between the results of the two methods was
1.05 (instead of a perfect 1.0) with R2 = 0.99997. Even for the
point with the lowest detectable NQ level (10 ng/L, at well b),
the fit between the two methods was good (both methods gave
10 ng/L).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.anal-
chem.7b02364.
■
S
Analysis of the 6 different field samples (that tested positive
for NQ) without the reduction step but after derivatization with
4-NBA did not show any false positive, which could
theoretically occur due to the presence of AQ in the water
samples. It should be noted that, although AQ was tested as a
pharmaceutical compound,34 it did not pass phase III FDA
analysis and has never been in use. Our literature survey failed
to show any report on the presence of AQ in the environment.
To complete the description of the contamination in the field
test site, Figure 5 (and the table inset) depicts the observed
concentrations of perchlorate (CLO4), chloroform (TCMA),
tetrachloroethylene (TCE), and cis-1,2-dichloroethylene (1,2-
DCE) at the same points. The NQ plume was much larger than
the plumes of all other organic contaminants, though smaller
than the perchlorate plume. This illuminates the risks entailed
by NQ due to its high mobility compared to other organic
contaminants, as well as its possible application as an indicator
for pollution plumes around polluting military industry sites.
Analysis of NQ in Soils. Recovery and LOD of NQ were
examined using 4 different soils (sandy, clayey, and sandy
clayey) with different organic matter content (0.05−0.36%).
The samples were spiked at a level of 5 ng/g and analyzed
according to the soil extraction protocol detailed in the
Experimental Section, followed by NQ analysis of the water
extract by the optimized NQ protocol. Recovery was evaluated
by a comparison with the response factor of the calibration
curve for NQ in water (Figure S3). LOD for each soil type was
calculated at 3 times the noise level obtained by analysis of the
Preliminary study: detailed unsuccessful protocols for (i)
preconcentration, (ii) aqueous derivatization, and (iii)
reduction of NQ; H NMR spectra of AQ-4NBA before
and after irradiation under UV; structures of cis- and
trans-isomers of AQ-4NBA; single crystal X-ray structure
of trans-AQ-4NBA; crystal data, data collection, and
refinement parameters for trans-AQ-4NBA acetate;
calibration curve for NQ; comparison of NQ quantifica-
tion by standard addition method and by calibration
using 15N-labeled NQ as internal standard (PDF)
1
AUTHOR INFORMATION
Corresponding Author
*E-mail: ovadia@mail.huji.ac.il.
ORCID
Sergey Sladkevich: 0000-0002-2486-0078
Ovadia Lev: 0000-0002-3536-2277
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the Scientific Infrastructure Program of the
Ministry of Science, Technology and Space and the Ministry of
agriculture and rural development for financing this research.
The authors thank the Harvey M. Krueger Family Center for
Nanoscience and nanotechnology.
F
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
(22) Bissett, F. H.; Levasseur, L. A. Analytical Methods for
Nitroguanidine and Characterization of its Degradation Products;
Technical Report TR-76/47; U. S. Army, Natick Research and
Development Command: Natick, MA, 1976.
REFERENCES
■
(1) Lee, K. E.; Balas-Hummers, W. A.; Di Stasio, A. R.; Patel, C. H.;
Samuels, P. J.; Roos, B. D.; Fung, V. Qualification Testing of the
Insensitive TNT Replacement Explosive IMX-101. In Insensitive
Munitions and Energetic Materials Technology Symposium, Munich,
Germany, October 11−13, 2010; pp 1−13.
(2) Oxley, J. C.; Smith, J. L.; Donnelly, M. A.; Colizza, K.; Rayome, S.
Propellants, Explos., Pyrotech. 2016, 41, 98−113.
(3) Picatinny Public Affairs. Army approves safer explosive to replace
TNT; 2010; https://www.army.mil/article/43553/army- (accessed
July 10, 2017).
(4) Walsh, M. R.; Walsh, M. E.; Ampleman, G.; Thiboutot, S.;
Brochu, S.; Jenkins, T. F. Propellants, Explos., Pyrotech. 2012, 37, 393−
406.
(5) Nitroguanidine (CAS 556-88-7) Market Research Report 2016;
http://www.businesswire.com/news/home/20161109005957/en/
Global-Nitroguanidine-CAS-556-88-7-Market-Research-Report (ac-
cessed July 10, 2017).
(23) Kenyon, K. F. A Data Base Assessment of Environmental Fate
Aspects of Nitroguanidine; U. S. Army Medical Bioengineering Research
and Development Laboratory: Frederick, MD, 1982.
(24) Braida, W.; Wazne, M.; Pavlov, J.; Kalomoiri, A.; Koutsospyros,
A.; Balok, V. In Proceedings of Protection and Restoration of the
Environment XI, Thessaloniki, Greece, July 3−6, 2012; pp 1634−1642.
(25) Russell, A. L.; Seiter, J. M.; Coleman, J. G.; Winstead, B.;
Bednar, A. J. Talanta 2014, 128, 524−530.
(26) Felt, D.; Gurtowski, L.; Nestler, C. C.; Johnson, J.; Larson, S.
Chemosphere 2016, 165, 18−26.
(27) Bulusu, S.; Dudley, R. L.; Autera, J. R. Magn. Reson. Chem. 1987,
25, 234−238.
(28) Ring, J. R.; Zheng, F.; Haubner, A. J.; Littleton, J. M.; Crooks, P.
A. Bioorg. Med. Chem. 2013, 21, 1764−1774.
(29) Shriner, R. L.; Neumann, F. W. Org. Synth. 1946, 26, 7−10.
(30) Gowda, D. C.; Mahesh, B.; Gowda, S. Indian J. Chem., Sect. B
2001, 40B, 75−77.
(6) Yalkowsky, S. H.; He, Y.; Jain, P. Handbook of Aqueous Solubility
Data, 2nd ed.; CRC Press: Boca Raton, FL, 2010.
(7) Haag, W. R.; Spanggord, R.; Mill, T.; Podoll, R. T.; Chou, T. W.;
Tse, D. S.; Harper, J. C. Environ. Toxicol. Chem. 1990, 9, 1359−1367.
(8) Pennington, J. C.; Jenkins, T. F.; Ampelman, G.; Thiboutot, S.;
(31) Albrecht, T. W. J.; Addai-Mensah, J.; Fornasiero, D. In Chemeca
2011: Engineering a Better World, Sydney, Australia, September 18−21,
2011, pp 2100−2110.
(32) Beaven, M. A.; Gordon, J. W.; Jacobsen, S.; Severs, W. B. J.
Pharmacol. Exp. Ther. 1969, 165, 14−22.
(33) Dolan, J. W. LC GC Eur. 2003, 16, 610−613.
(34) Bolton, W. K.; Abdel-Rahman, E. Expert Opin. Invest. Drugs
2002, 11, 565−574.
́
Brannon, J.; Clausen, J.; Hewitt, A. D.; Brochu, S.; Dube, P.; Lewis, J.;
Ranney, T.; Faucher, D.; Gagnon, A.; Stark, J. A.; Brousseau, P.; Price,
C. B.; Lambert, D.; Marois, A.; Bouchard, M.; Walsh, M. E.; Yost, S. L.;
Perron, N. M.; Martel, R.; Jean, S.; Taylor, S.; Hayes, C. A; Ballard, J.-
M.; Walsh, M. R.; Mirecki, J. E; Downe, S.; Collins, N. H.; Porter, B.;
Karn, R. Distribution and Fate of Energetics on DoD Test and Training
Ranges: Interim Report; ERDC TR-04-4; U. S. Army Corps of
Engineers: Washington, DC, 2004.
(9) DeVries, J. E.; Gantz, E. S. C. J. Am. Chem. Soc. 1954, 76, 1008−
1010.
(10) Bonner, T. G.; Lockhart, J. C. J. Chem. Soc. 1958, 3858−3861.
(11) Bower, M. E.; Hartley, W. R.; Roberts, W. C. In Drinking Water
Health Advisory: Munitions; Roberts, W. C.; Hartley, W. R., Eds.; CRC
Press, Lewis Publishers: Boca Raton, FL, 1992; pp 201−246.
(12) U. S. Environmental Protection Agency. Nitroguanidine (CAS
556-88-7). Integrated Risk Information System (IRIS). Chemical
Assessment Summary. https://cfpub.epa.gov/ncea/iris/iris_
documents/documents/subst/0402_summary.pdf (accessed July 10,
2017).
(13) Dickinson Burrows, W.; Schmidt, M. O.; Chyrek, R. H.; Noss,
C. I. Photochemistry of Aqueous Nitroguanidine; Technical Report 8808;
U. S. Army, Toxic and Hazardous Materials Agency: Aberdeen, MD,
1988.
(14) Richard, T.; Weidhaas, J. J. Hazard. Mater. 2014, 280, 372−379.
(15) Perreault, N. N.; Halasz, A.; Manno, D.; Thiboutot, S.;
Ampleman, G.; Hawari, J. Environ. Sci. Technol. 2012, 46, 6035−6040.
(16) Mirecki, J. E.; Porter, B.; Weiss, C. A., Jr. Environmental
Transport and Fate Process Descriptors for Propellant Compounds;
ERDC/EL TR-06-7; U. S. Army Corps of Engineers: Washington,
DC, 2006.
(17) Mulherin, N. D.; Jenkins, T. F.; Walsh, M. E. Stability of
Nitroguanidine in Moist, Unsaturated Soils; ERDC/CRREL TR-05-2; U.
S. Army Corps of Engineers: Washington, DC, 2005.
(18) Kaplan, D. L.; Kaplan, A. M. Degradation of Nitroguanidine in
Soils; Technical Report TR-85/047; U. S. Army, Natick Research and
Development Command: Natick, MA, 1985.
(19) Walsh, M. E. Analytical Methods for Determining Nitroguanidine
in Soil and Water; Special Report 89-35; U. S. Army Corps of
Engineers: Washington, DC, 1989.
(20) Manning, D. L.; Maskarinec, M. Analysis of Nitroguanidine in
Aqueous Solutions by HPLC with Electrochemical Detection and
Voltammetry; ORNL/TM-10295; U. S. Army Toxic and Hazardous
Materials Agency: Aberdeen Proving Ground, MD, 1987.
(21) Mu, R.; Shi, H.; Yuan, Y.; Karnjanapiboonwong, A.; Burken, J.
G.; Ma, Y. Anal. Chem. 2012, 84, 3427−3432.
G
DOI: 10.1021/acs.analchem.7b02364
Anal. Chem. XXXX, XXX, XXX−XXX