Kitmin Chen1, Alexander S Edgar1, Camille H Wong1, Dali Yang1. 1. MST-7: Engineered Materials Group, Materials Science and Technology Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States.
Abstract
As an antioxidant, N-phenyl-β-naphthylamine (PBNA) inhibits the activity of oxidants, such as NO x , to prevent the degradation of energetic materials. In the presence of NO x , nitrated products can be generated in the process potentially. To characterize nitrated PBNA in a nontargeted analysis of complex samples as such, liquid chromatography tandem quadrupole time-of-flight (LC-QTOF), as an excellent analytic technique, is used due to its high resolution and sensitivity. However, a systematic approach of instrumentation optimization, data interpretation, and quantitative determination of products is needed. Through a step-by-step evaluation of the instrumental parameters used in the Q0, Q1, and Q2 compartments of LC-QTOF, optimal ion yields of precursor ions and high-resolution MS2 fragmentation spectra at low mass defects were obtained in both negative and positive electrospray ionization modes. Through rationalization of the fragmentation pathways and verification using theoretical masses, the mononitro derivative of PBNA was accurately identified as N-(4-nitrophenyl)-naphthalen-2-amine and further confirmed using a reference standard. Using strict criteria provided by the analytical guidelines (e.g., SANTE), limit of quantitation, limit of detection, and calibration were established for the quantitation of PBNA and nitrated PBNA. From optimization to characterization and subsequent quantification of the mononitro-PBNA derivative, for the first time, the applicability of this strategy is demonstrated in the aged energetic binders.
As an antioxidant, N-phenyl-β-naphthylamine (PBNA) inhibits the activity of oxidants, such as NO x , to prevent the degradation of energetic materials. In the presence of NO x , nitrated products can be generated in the process potentially. To characterize nitrated PBNA in a nontargeted analysis of complex samples as such, liquid chromatography tandem quadrupole time-of-flight (LC-QTOF), as an excellent analytic technique, is used due to its high resolution and sensitivity. However, a systematic approach of instrumentation optimization, data interpretation, and quantitative determination of products is needed. Through a step-by-step evaluation of the instrumental parameters used in the Q0, Q1, and Q2 compartments of LC-QTOF, optimal ion yields of precursor ions and high-resolution MS2 fragmentation spectra at low mass defects were obtained in both negative and positive electrospray ionization modes. Through rationalization of the fragmentation pathways and verification using theoretical masses, the mononitro derivative of PBNA was accurately identified as N-(4-nitrophenyl)-naphthalen-2-amine and further confirmed using a reference standard. Using strict criteria provided by the analytical guidelines (e.g., SANTE), limit of quantitation, limit of detection, and calibration were established for the quantitation of PBNA and nitrated PBNA. From optimization to characterization and subsequent quantification of the mononitro-PBNA derivative, for the first time, the applicability of this strategy is demonstrated in the aged energetic binders.
N-Phenyl-β-naphthylamine
(PBNA) is a widely
used antioxidant in numerous industrial applications, such as the
manufacture of rubber, plastic, dyes, various greases, lubricating
and transformer oils, and energetic materials.[1−9] In the presence of oxidative or hydrolyzing agents such as oxygen,
nitrogen oxides (NO), and nitrous and
nitric acids (HNO), this secondary arylamine
stabilizer can effectively protect energetic materials from oxidation
or hydrolysis by scavenging the free radicals (e.g., NO2 radical) and hence reducing the acidity.[1,2,5,7,10] In a similar class of antioxidants (e.g., diphenylamine
and triphenylamine), nitration of arylamines generally forms nitro
derivatives by reacting with nitrogen oxide radicals[6−9,11−14] and can be characterized using
mass spectrometry (MS) through rationalization of fragmentation.[15] Accordingly, PBNA can also form mononitro derivatives
when it undergoes similar reactions. In studies of phosphite antioxidants,
MS ionization is successfully employed as the primary identification
method.[16,17] Likewise, PBNA and its mononitro derivative
can be analyzed using the electrospray ionization (ESI). However,
there are three major challenges: (1) because a liquid chromatography
quadrupole time-of-flight (LC-QTOF) mass spectrometer is capable of
detecting analytes at sub-ppb levels, the accuracy of identification
and quantitation in nontargeted analysis can be significantly hindered
in sampling systems containing a diverse population of decomposed
species (e.g., energetic binders); (2) the concentration of PBNA is
extremely low,[18,19] and extraction may destroy valuable
unknown products in the aged energetic binders. (3) There is a lack
of literature concerning the LC-QTOF analysis of PBNA activity in
energetic binders. Hence, the development of an optimal analytical
method is critical to accurately determine the concentration of PBNA,
identify its nitro derivatives, and evaluate the quality of the energetic
binders. In this investigation, we explored various instrumental parameters,
established limits of quantitation (LOQ) for compounds of interest,
and developed a characterization strategy using the mononitro derivative
of PBNA in the aged energetic binders as a case study.
Results and Discussion
pH Effect
Due to the inductive and
resonance effects, substitution of hydrogen by NO2 can
decrease the acidity constant (pKa) of the polycyclic compounds.[20] Since the retention time and ionic strength
of an analyte is correlated to the pKa value and can be significantly
altered under the pH effect,[21] the pH of
the mobile phases was investigated to optimize chromatographic separation.
In unbuffered mobile phases, both mono- and dinitro-PBNA derivatives
are deprotonated, and therefore, optimal separation of retention times
can be achieved. However, poor solubility of ammonium acetate in 95:5
(v/v) ACN:MeOH poses a major concern, which increases the risk of
column clogging and inconsistency of retention time. Therefore, the
pH is adjusted to 6.0 to maintain the balance between the solubility
of ammonium acetate salt and chromatographic resolution of analytes.
As a result, the retention times of 2,2-dinitropropanol (DNPOH), PBNA,
mononitro-PBNA, and dinitro-PBNA were measured at 3.300, 6.400, 6.129,
and 6.103 min, respectively, with a standard deviation of ±0.030
min.
Optimization of the MS Experiment
The optimization targets four components of the LC-QTOF ion path:
(1) droplet ionization, (2) the ion guide at Q0, (3) mass filtration
at the Q1 analyzer, (4) fragmentation at the Q2 collision cell, and
separation at the TOF analyzer. Optimization of MS parameters for
PBNA and its derivatives was heavily focused on ESI+ mode due their
poor sensitivity and selectivity in positive ionization, as shown
in Figure . (1) To
maximize the ionization efficiency, the flow rate (FR) was reduced
from 0.40 mL/min to 0.35 mL/min, the spray voltage (SV) was increased
from 4500 to 5500 V, the ion source gas 1 (GS1) pressure was increased
from 30 to 50 psi, and the source temperature (TEM) was increased
from 100 to 250 °C. These four parameters specifically improve
the desolvation rate and therefore the droplet size is reduced before
entering the mass analyzer. Also, the addition of acetic acid in the
mobile phases promotes ionization efficiency, and hence increases
the signal responses of the targeted analytes in ESI+.[22] As a complimentary effect of the decreased FR,
the risk of column overpressure is also reduced: the column pressure
dropped from initial 6500 to 4750 psi. (2) At the orifice plate, although
the declustering potential (DP) can be adjusted to prevent ions from
clustering, the loss of analyte signals was observed when a stronger
DP was applied, perhaps due to unwanted in-source fragmentation of
the precursor ions. Hence, the DP was settled at 80 V. (3) By injecting
a sample of 1:1 mobile phases A:B using the information-dependent
acquisition (IDA) experiment, the exclusion lists of interfering ions
were obtained for both ESI+ and ESI– (Table S2). To minimize these contaminants or impurities from reaching
the collision cells, the exclusion lists were applied to the IDA method
with a mass tolerance of 10 mDa and an intensity threshold of 500
cps. (4) While increasing the collision energy (CE) on the MS acquisition
lowers the ion yields of the precursors, increasing the CE of the
MS2 acquisition induces higher order of fragmentation in
the MS2 spectra. Therefore, the CE of the MS1 acquisition remains unchanged (10 V) in ESI+. Furthermore, since
minimal fragmentation is desired to distinguish aromatic compounds
from the aliphatic compounds, only a narrow range of collision energies
in the MS2 acquisition was explored, and the resultant
CE of 25 V is obtained as the optimal condition for characterization
of MS2 spectra in ESI+ mode (e.g., minimal presence of
low intensity or insignificant fragments).
Figure 1
Evaluation of MS parameters
in ESI+ using the signal responses
of PBNA, mononitro and dinitro-PBNA derivatives. The same aged energetic
sample is used across all tested conditions. The intensities on the
left indicate PBNA and on the right indicate nitro derivatives.
Evaluation of MS parameters
in ESI+ using the signal responses
of PBNA, mononitro and dinitro-PBNA derivatives. The same aged energetic
sample is used across all tested conditions. The intensities on the
left indicate PBNA and on the right indicate nitro derivatives.The MS parameters in ESI– were only evaluated
using the
nitro derivatives due to the lack of PBNA signals (Figure ). Since the ESI– signals
were two to three times stronger than ESI+ signals to begin with,
minimal changes in the instrumental parameters were required for further
signal enhancement of nitro derivatives: increased GS1 from 30 to
50 psi and the decreased DP in MS acquisition from −50 to −80
V. Finally, the dwell times or accumulation times were optimized to
100 and 50 ms in MS1 and MS2 acquisitions, respectively,
to obtain a desirable signal-to-noise ratio (S/N). As a result, the
estimated scan cycles were increased from 969 to 1759, which enables
at least 12-point measurements across each analyte peak.
Figure 2
Evaluation
of MS parameters in ESI– at different conditions.
Evaluation
of MS parameters in ESI– at different conditions.
Characterization of Mononitro-PBNA
Using the high-resolution spectral data attained by the optimized
method, the mononitro derivative of PBNA was characterized through
the fragmentation patterns, mass verification, retention time, and
reference comparison. Although MS fragmentation generally obeys the
even-electron rule, losses of OH and NO radicals are common fragmentation
behaviors in nitroaromatic compounds.[15,23] Based on the
consecutive cleavages of radicals and the subsequent step of H abstraction,
the three stable fragment ions (m/z 217.0895, 218.0966, and 248.0947) are logically explained in Figure . In addition to
the fragment below the mass calibration range (e.g., NO2–), the accuracy of all observed m/z (e.g., precursor, fragment, and isotopic masses)
are within 10 ppm of the theoretical m/z, and good agreement is demonstrated between the mass profiles of
the aged sample and the reference standard (Tables and 2). As demonstrated
in the mass verification above, the standard addition method was also
utilized to further verify the mononitro derivative of PBNA. In both
ESI– and ESI+, the extracted ion chromatograms (XIC) of the
standard addition (Figure , right column) illustrated uniform symmetrical peaks at the
same retention time and there is no significant indication of difference
such as the shoulder peak or peak broadening. Although a minor coeluent
is observed at the base of the targeted peak in ESI+ (Figure , bottom left), the peak alignment
between the aged sample and the reference standard is not impacted.
The retention time of mononitro-PBNA in the aged samples is averaged
to 6.127 ± 0.013 min across 55 measurements, matching the retention
time obtained in the calibration standards (6.115 ± 0.016 min
across 105 measurements). Furthermore, the fragmentation profile of
mononitro-PBNA in the aged sample displays the same fragment fingerprints
when compared to the MS2 spectra of the reference standard
and the standard addition (Figure ). Since the coeluents or the nearby compounds (Figure , bottom row) can
potentially influence the fragmentation efficiency and contaminate
the MS2 spectra, the varying intensities of the ion peaks
at m/z 265, 248, and 218 in Figure (bottom row) are
therefore explained.
Figure 3
Rationalization of fragmentation pathway for mononitro-PBNA
using
the ESI+ MS2 spectrum.
Verification of the Isotopic Pattern
in the MS1 Spectra
isotope,
ESI+
isotope,
ESI–
observed m/z (Da)
266.1002
264.0855
predicted m/z (Da)
266.1003
264.0857
mass error (ppm)
–0.38
–0.76
observed abundance (%)
19.44
17.31
predicted abundance (%)
18.26
18.24
abundance error (%)
6.46
–5.10
Figure 4
XICs of the aged sample (left), N-(4-nitrophenyl)
naphthalen-2-amine (middle), and standard addition of the aged sample
(right) obtained from ESI– (top row) and ESI+ (bottom row).
The isomers of mononitro derivatives are detected in ESI+ (6.500 min)
but not examined in detail because the positions of NO2 addition is unknown and the reference standards are not available.
Figure 5
MS2 spectra of the aged sample (left), N-(4-nitrophenyl) naphthalen-2-amine (middle), and standard
addition
of the aged sample (right) obtained from ESI– (top) and ESI+
(bottom).
Rationalization of fragmentation pathway for mononitro-PBNA
using
the ESI+ MS2 spectrum.XICs of the aged sample (left), N-(4-nitrophenyl)
naphthalen-2-amine (middle), and standard addition of the aged sample
(right) obtained from ESI– (top row) and ESI+ (bottom row).
The isomers of mononitro derivatives are detected in ESI+ (6.500 min)
but not examined in detail because the positions of NO2 addition is unknown and the reference standards are not available.MS2 spectra of the aged sample (left), N-(4-nitrophenyl) naphthalen-2-amine (middle), and standard
addition
of the aged sample (right) obtained from ESI– (top) and ESI+
(bottom).
Evaluation of Quantitative Criteria
The LOQ, limits of detection (LODs), and calibration curves were
established following various analytical guidelines.[24,25] The average height of the matrix blank (Figure , bottom) is measured across a 1 min retention
time window as baseline noise. Using the measured height of the reference
standard CAL1 (Figure ) as the signal, S/N ratios are calculated at 124, 6, and 25 for
DNPOH, PBNA, and mononitro-PBNA, respectively. Using the known concentrations
of the reference standards and the calculated S/N ratios, the LOQs
and LODs are estimated given by the commonly accepted tolerances of
S/N ratios at ≥10 and ≥3:[24,25] respectively,
162.00 and 49.00 ppb for DNPOH, 6.00 and 2.00 ppb for PBNA, and 1.00
and 0.30 ppb for mononitro-PBNA. As recommended by the guideline,[25] weighted (1/x) linear regressions
were applied to the calibration curves of the mixed standards (Figures , S1, and S2). The linearities of all calibration
curves are measured at a R2 value of greater than 0.98
(Table ). In addition
to DNPOH, the accuracies or recoveries of all standards in PBNA and
the mononitro derivative are within the ±20% tolerance of the
calibration curve.[25] Due to poor sensitivity
in the precursor ion of DNPOH (m/z 149.0203), the in-source fragment ion, deprotonated 1,1-dinitroethane
(m/z 119.0098, C2H3N2O4–), was used as
the quantifier. Perhaps associated to the effect of fragmentation
efficiency, most of the recovery measurements in DNPOH standards are
deviated beyond the tolerance, as depicted by the subtle quadratic
characteristic in Figure (left). However, accuracy can be improved either by using
a targeted approach (e.g., multiple reaction monitoring or MRM) and/or
through internal standard correction (e.g., using compounds that exhibit
similar retention time and fragmentation behavior or deuterated DNPOH),
as demonstrated by the example in Figure (right) and the improved R2 value
of 0.997 in Table .
Figure 6
Average height of the baseline noises: 782 cps for DNPOH (left),
131 cps for PBNA (middle), and 200 cps for mononitro-PBNA (right)
in the corresponding retention time windows.
Figure 7
Peak heights of the CAL1 reference standards: 96,700 cps
for DNPOH
at 2.00 ppm in ESI– (left), 739 cps for PBNA at 3.00 ppb (middle)
in ESI–, and 4991 cps for mononitro-PBNA at 2.50 ppb in ESI+
(right).
Figure 8
Calibration curves of DNPOH without ISTD correction (left)
and
with ISTD correction (right) in ESI–.
Table 3
Linear Regressions of PBNA, Mononitro-PBNA,
and DNPOH
analytes, ionization
mode
linear regression
equation (1× weighting)
coefficient
of determination, r2
PBNA, ESI+
y = 1079.04394 x + 200.18303
0.995
PBNA-NO2, ESI-
y = 8774.94923 x + −3417.05100
0.999
PBNA-NO2, ESI+
y = 834.19067 x + 169.35237
1.000
DNPOH, ESI-
y = 1.64960e5 x + 3.12804e5
0.981
DNPOH (with
ISTD correction), ESI-
y = 3.22355 x + −0.14011
0.997
Average height of the baseline noises: 782 cps for DNPOH (left),
131 cps for PBNA (middle), and 200 cps for mononitro-PBNA (right)
in the corresponding retention time windows.Peak heights of the CAL1 reference standards: 96,700 cps
for DNPOH
at 2.00 ppm in ESI– (left), 739 cps for PBNA at 3.00 ppb (middle)
in ESI–, and 4991 cps for mononitro-PBNA at 2.50 ppb in ESI+
(right).Calibration curves of DNPOH without ISTD correction (left)
and
with ISTD correction (right) in ESI–.
Conclusions
A nontargeted LC-QTOF method
was developed to analyze PBNA and
its nitro derivatives in energetic binder materials for the first
time. The challenges of selectivity and sensitivity, due to the complexity
of the matrix and the low concentration of PBNA, respectively, were
overcame through a comprehensive assessment of instrumental parameters.
Based on the demonstration of characterizing mononitro-PBNA, the quality
of the spectrometric information obtained (XIC, MS1, and
MS2 spectra) is very promising (mass defect <10 ppm).
Using the reference material of PBNA, the determination of PBNA concentration
in energetic binders is achievable. Using this experimental design,
the degradation study of aged energetic binders[26−29] will be revisited: to search
for and/or verify the identities of other PBNA nitro derivatives and
to uncover the mechanism of antioxidants in the thermal aging process
of energetic binders.
Experimental Section
Reference Chemicals
DNPOH was synthesized
by David Langlois at Los Alamos National Laboratory.[30] The PBNA standard was purchased from Sigma-Aldrich. N-(4-nitrophenyl) naphthalen-2-amine, the reference standard
of the mononitro-PBNA derivative, was purchased from Aurora Fine Chemicals
(San Diego, CA). All chemicals were used as received and/or as synthesized.
Sample and Calibration Preparation
To evaluate the performance (e.g., peak intensity) of the changes
in LC-QTOF parameters and for the sake of brevity, only a selected
few of the aged energetic binders from the thermal aging experiment[26−29] that contains PBNA and PBNA nitro derivatives were tested, which
were prepared by dissolving 3.0 ± 0.5 mg of the aged samples
in 10 mL of ACN. Stock solutions of DNPOH, PBNA, and N-(4-nitrophenyl)-naphthalen-2-amine were prepared in ACN at 4244,
239, and 1843 ppm, respectively. By diluting the stock solutions in
a 10 mL volumetric flask with appropriate volume of ACN, nine calibration
standards were prepared in the concentrations as described in Table S1.
LC-QTOF Parameters
Chromatographic
separation was achieved on a reversed-phase (RP) LC column: Phenomenex
Kinetex 2.6 μm C8 100 Å, 150 × 2.1 mm. The aqueous
and organic mobile phases were 13 mM ammonium acetate at pH 6.0 in
water and in 95:5 (v/v) ACN/methanol, respectively. An InfinityLab
quick change inline filter (Agilent, Part No. 5067-1603) with a stainless-steel
filter disk (2.1 mm I.D., 0.2 μm porosity) was used to prevent
column clogging from potential particles in the mobile phases. A SecurityLINK
PEEKsil UHPLC finger-tight fitting (Phenomenex, Part No. AJ1-2441)
was used to reduce the risk of over- or undertightening and minimize
the impact of dead volume to the peak shape. The autosampler was operated
at room temperature. The column oven temperature was held at 40 °C.
The sample injection volumes were set to 3 μL in ESI–
and 5 μL in ESI+, and a rinse method was employed with an injection
volume of 10 μL of acetone after each injection to prevent carryover
contamination from the needle. The details of the optimized HPLC programs
are summarized in Table .
Table 4
LC Parameters (ExionLC AC)
acquisition gradient: time (min), % organic, flow rate (mL/min)
ESI and IDA parameters in positive and negative modes
were optimized
as described in Table . Attributes of precursor ions (e.g., isotopic masses and relative
abundances) were extracted from the spectra in the first stage of
the tandem mass spectrometry (MS1). Fragment ion spectra
(MS2) were generated using collisionally activated dissociation
(CAD) techniques at low CE. Frequency of external calibration was
set to every two injections.
Authors: Karsten Levsen; Hans-Martin Schiebel; Johan K Terlouw; Karl J Jobst; Manfred Elend; Alfred Preiss; Herbert Thiele; Arnd Ingendoh Journal: J Mass Spectrom Date: 2007-08 Impact factor: 1.982
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8