Int. J. Ion Mobil. Spec. DOI 10.1007/s12127-017-0221-z
ORIGINAL RESEARCH
Method validation parameters for drugs and explosives in ambient pressure ion mobility spectrometry Victoria Sedwick 1 & Monique Massey 1 & TeAsia Codio 1 & A Bakarr Kanu 1
Received: 5 July 2017 / Revised: 17 July 2017 / Accepted: 18 July 2017 # Springer-Verlag GmbH Germany 2017
Abstract An approach using method validation (MV) parameters, otherwise known as analytical figures of merit was combined with electrospray ionization high performance ion mobility spectrometry (ESI-HPIMS) to describe an approach for evaluating drugs and explosives analysis in the field. MV parameters such as reduced mobility (Ko), conditional reduced mobility (Kc), resolving power (Rp), theoretical plates (N), linearity, accuracy, precision, limit of detection (LOD), limit of quantitation (LOQ), repeatability, range, and reporting limit were investigated and developed for eleven drugs and six explosives. Our investigation estimated resolving power at 66 ± 0.64 for the ESI-HPIMS used. The LOD’s calculated ranged from 0.45–2.97 ng of material electrosprayed into the ESI-HPIMS. The LOQ’s calculated falls in the range 4.11– 8.63 ng of material electrosprayed into the ESI-HPIMS. The key findings from this investigation were the following: Kc proves to be a measure of the identity of an explosive or drug ion; a parameter that may be applied to help aid IMS devices when detecting drugs and explosives. MV parameters, especially, Kc, introduced in this study is an effective parameter for establishing a unique identity of a drug or explosive. A control chart is an effective way to monitor the performance of an instrument and may be a useful tool for establishing reliability of confirmatory data in forensic investigations. MV parameters may be a reliable, accurate and unique identification marker for target drugs and explosives capable of differentiating these substances from false positive responses.
* A Bakarr Kanu
[email protected] 1
Department of Chemistry, Winston-Salem State University, Winston-Salem, NC 27110, USA
Keywords Method validation parameters . Forensic analysis . Analytical figures of merit . False positive responses . Explosives . Drugs . Conditional reduced mobility
Introduction The separation detection capabilities of ambient pressure ion mobility spectrometry (IMS) has been documented for well over 50 years for the following applications; explosives [1–9], drugs [8, 10–16], chemical warfare agents [17–19], environmental contaminants including volatile organic compounds [20–23], food industry [24], biological compounds [25, 26], and medical applications [27, 28], to name a few. Several advantages have been reported for IMS, in addition, to its multiple detection capabilities. IMS is a fast technique, it is sensitive, usually more reproducible [29], and ion responses in an IMS can be collected and recorded efficiently. IMS is less expensive, requires less bench space, is mechanically more robust, and is much easier to operate, especially when compared to techniques like mass spectrometry. Despite five decades of development, standalone IMS instruments when applied for detecting samples in the field still face problems with false positive responses [30–32]. The problem has mainly been due to the low resolving power experienced in commercial IMS systems. With such limitations, it is possible for a false positive ion to drift similar to an actual explosive or drug ion and be falsely identified. Falsely identifying a drug or explosive could mean a field instrument indicating the presence of an explosive or drug ion when in fact none was present. In such instances, sophisticated tests must be performed to ensure reliable results costing both time and money. The development of method validation parameters that can serve as confirmatory tests and aid
Int. J. Ion Mobil. Spec.
the identification of explosives and drugs in the field is the subject of this article. The goal of any analytical measurement is to obtain reproducible, reliable and accurate data. Method validation (MV), which is a process of proving that an analytical method is acceptable for its intended purpose [33] plays a significant role in achieving this goal. Data generated from method validation can be used to evaluate the quality, reliability and consistency of analytical results, which is an important part of any good analytical practice. In the pharmaceutical industry, for example, MV requirement includes studies of multiple figures of merit, including linearity, accuracy, precision, limit of detection, limit of quantitation, method specificity, repeatability, confidence level, and robustness, to name a few. The hypothesis in this article is that the determination of key MV parameters may prove to be an effective strategy for establishing a unique identity of a drug or explosive ion. False positive responses are major issue in standalone IMS for field measurements. When drugs, explosives or other substances of forensic interests are analyzed in the field with standalone IMS, additional tests may be required in order to ensure the proper identification of the target. In this experiment, MV parameters, otherwise known as analytical figures of merit, were developed for specific target drugs and explosives using an electrospray ionization high performance ion mobility spectrometry (ESI-HPIMS) in order to implement a reliable, accurate and unique identification marker for target drugs and explosives that is capable of differentiating these substances from false positive responses.
Experimental section Materials and reagent The following drugs and explosives amphetamine (AMPHET), cocaine (COC), cannabidiol (CBD), codeine (COD), heroin (HER), methamphetamine (METHAMP), morphine (MOR), phentermine (PHENT), L-phenylephrine (L-PHERN), trinitrotoluene (TNT), 2,6-dinitrotoluene (2,6-DNT), 4-amino-2,6dinitrotoluene (4-A-2,6-DNT), Octahydro-1,3,5,7-tetranitro1,3,5,7-tetrazocine (HMX), 1,3,5-Trinitroperhydro-1,3,5-triazine (RDX), and tetryl were purchased as 1 mg mL−1 certified reference materials (CRM) from Cerilliant (Austin, TX). CRM of drugs and explosive samples were prepared at different concentrations and analyzed with a standalone electrospray ionization high performance ion mobility spectrometry (ESI-HPIMS) to determine method validation parameters for this study. Two other drug samples, pioglitazone (PIOGZ) and rosiglitazone (ROSGZ) were purchased from Sigma Aldrich (St. Louis, MO). The solid sample was then used to prepare standard
solutions at specific concentrations. All solvents and reference standards were obtained from Sigma Aldrich & J.T. Baker (Phillipsburg, NJ). Instrumentation The atmospheric pressure instrument used in this investigation was the commercial RTK2100 standalone ESI-HPIMS system (Excellims Corporation, Acton, MA). The complete RTK2100 consisted of an ESI source, HPIMS drift tube, universal ion mobility spectrometry controller (UIMSC), and a VisIon™ software. The very first step in IMS operation is to generate analyte ions in the gas-phase. Electrospray ionization (ESI) has been one of the ionization methods coupled with IMS [32, 34]. This ionization method has proved very useful for IMS because it allowed for the detection of high molecular weight biological compounds significantly extending the applications developed for IMS. The ESI solvent in the positive ion mode (90% Methanol/ 9% Water/1% HOAc) [35, 36] and negative ion mode (90% Methanol/10% Water) [37] was delivered with a syringe pump (Analytical West Inc., Corona, CA). The flow rates in the positive and negative ion modes was 3.00 μL min−1 and 6.00 μL min−1, respectively, and was delivered from a 500 μL gas-tight syringe (Fischer Scientific, Hanover Park, IL) for the entire experiment. The syringe from the pump was connected to a 150 μm o.d. × 10 μm i.d. × 17.4 cm long fused silica capillary (Polymicro Technologies, Phoenix, AZ) using a zero dead volume needle to capillary connector (Upchurch Scientific, Oak Harbor, WA). A second silica capillary (150 μm o.d. × 10 μm i.d. × 7.5 cm long) was used to transfer the ESI solution through another zero dead volume capillary to an electrospray needle connector. The electrospray needle was held at 2.90 kV in the positive ion mode above the drift potential of 8.00 kV. In the negative ion mode, the electrospray needle was held at 2.60 kV above the drift potential of 8.00 kV. Drugs and explosive analyte ions were continuously infused through the electrospray needle by applying a potential of 10.5 kV at a current limit of 1 μA to maintain a stable spray [36–39]. This resulted in ESI solvent being removed from the ionized droplets in the ~6.00 cm long desolvation region. Drugs and explosives analyte ions were subsequently introduced into the drift region held at a constant temperature of 150 °C. A Bradbury-Nielson ion gate operating under the influence of a pulsed gate width was used to admit ions from the desolvation region into 10.5 cm long drift region. The upper potential of the desolvation region was held at 8.00 kVand the gate reference voltage was held at approximately 7.91 kVover the 10.5 cm long drift cell creating an electric field of approximately 753 V/cm in the drift cell. Ions were separated according to their size-to-charge as they moved under the influence of the drift field through a 1.4 L/min counter flowing drift gas
Int. J. Ion Mobil. Spec.
in the drift region. Each mobility spectra represented a sum of 10 spectrum ranging in length of 40 ms which were sampled by a Faraday plate detector. Data was acquired using Excellims VisIon control and acquisition software and were exported as text files for processing using Microsoft Excel. Table 1 summarizes all the operating conditions and the drift gas used in this investigation for both the positive and negative ion modes. Figure 1 shows a cross-sectional schematic view of the ESI-HPIMS.
Calculations An alternative strategy to determine reduced mobility for an unknown ion in an ion mobility spectrometry is to relate the observed drift time of the unknown (td [unknown]) to the observed drift time of a standard (td [standard]) and the reduced mobility of the standard (Ko [standard]) [18, 21]. This relationship is shown in Eq. 1. K o ½unknown ¼
K o ½s tan dard t d ½s tan dard t d ½unknown
ð1Þ
This investigation used two standards; 1.95 ± 0.01 cm2 V−1 s−1 was the reduced mobility for 2,4-lutidine standard in the positive ion mode. The negative ion mode used 1.43 ± 0.02 cm2 V−1 s−1 as the reduced mobility of citric acid standard [40–42]. Ion mobility separation now has a capability of generating unique identity of a compound at specific gate pulse width. The parameter used to generate unique identity of an ion in an IMS separation was previously defined as Table 1 Summary of ESIHPIMS instrument conditions
conditional reduced mobility (CRM) [32]. The CRM (Kc) is defined by Eq. 2. Kc ¼
Ko Δt 0:5
ð2Þ
where Kc has a unit of 103 cm2 V−1 s−2 and Δt0.5 is the width-at-half-height. The detection limit was estimated as follows [22, 43, 44]. We assumed a linear response, the minimum concentration closer to the detection limit whose signal was closer to the instrument noise was electrosprayed into the instrument. The detection limits at 3σ were estimated using statistics according to Eq. 3. analyteLOD
¼
t α;n RSDx ðamount electrosprayed Þ 100%
ð3Þ
where tα,n = t-table value for n - 1 degrees of freedom; α = confidence interval (99%); n = number of replicate measurements; RSD = relative standard deviation [44]. Analyte limit of quantification (LOQ) was calculated as [45] analyteLOQ ¼
10σ B1
ð4Þ
where σ is the standard deviation from the lowest detectable concentration (n = 5), and B1 is the sensitivity/slope from the calibration curve. Ion separation capabilities in an IMS are quantified using resolving power. IMS measured resolving power is defined as the drift time (td) divided by the width-at-half-height (Δt0.5) of
Parameter
Settings (+ mode)
Settings (− mode)
Unit
Desolvation region length Drift region length
3.50 10.5
3.50 10.5
cm cm
ESI Voltage Source Voltage Drift Voltage Gate reference voltage ESI flow rate Drift flow Drift gas Temperature Pressure IMS gate pulse width Sample rate Spectral time IMS Summing Acquisition time Src. Supply Current
10.5 2.90 8.00 7.91 3 1.4 Air 150 742–760 10–100 200,000 40 10 30 6.12
10.2 2.60 8.00 −7.91 6 1.4 Air 180 742–760 10–100 200,000 40 10 30 −0.24
kV kV kV kV μL min−1 L min−1 o
C mm Hg μs samples/s ms s μA
Int. J. Ion Mobil. Spec. Fig. 1 Schematic cross-sectional view of electrospray ionization high performance ambient pressure ion mobility spectrometry
the ion mobility peak [22, 32, 46–48]. This relationship is given by Eq. 5. td Δt 0:5
Rp ¼
ð5Þ
where Rp is the measured resolving power.
In IMS, the theoretical plates, N, can be used for comparing separating capabilities of an IMS to chromatographic methods [18, 49]. Eq. 6 shows the relationship between Rp and the number of theoretical plates. 2 N ¼ 5:55 Rp ð6Þ
5
5 Ko
4
1.34 (m)
Rp
SNR
63
17
N
4
13143
3
Ko
Rp
SNR
1.43
39
25
N 8350
3
Int. (au)
Int. (au)
2
10 µg cm-3 2,6-DNT (m)
2
10 µg cm-3 4-A-2,6-DNT (m) 1
1
10 µg cm-3 4-A-2,6-DNT (d)
0
0 0
10
20
30
40
0
10
20
Drift Time (ms)
30
40
30
40
Drift Time (ms)
5
5 Ko
4
1.24 (m)
Rp
SNR
65
22
N
4
23622
3
Ko
Rp
SNR
1.31
62
29
N 21334
3
Int. (au)
10 µg cm-3 TNT
Int. (au)
2
2
10 µg cm-3 RDX (m) 1
1
10 µg cm-3 RDX (d) 0
0 0
10
20
30
40
0
10
Drift Time (ms)
20 Drift Time (ms)
5
5
4
Ko
Rp
SNR
1.15
48
21
N
4
Ko
Rp
SNR
1.11
63
17
N 22028
12260
3
3
Int. (au)
10 µg
cm-3
Int. (au)
cocaine
10 µg cm-3 PIOGZ
2
2
1
1
0
0
10
20 Drift Time (ms)
30
40
0 0
10
20
30
40
Drift Time (ms)
Fig. 2 Representative ESI-HPIMS for target drugs and explosives studied in this investigation. Monomer and dimer are observed for 4-A-2,6-DNT and RDX. The drift gas was air, and the gate pulse widths was 100 and 50 μs for positive and negative ions modes, respectively
Int. J. Ion Mobil. Spec.
where N is the number of theoretical plates and Rp is the resolving power. Signal-to-noise ratio (SNR) was determine by zooming in at a region of the spectra with the highest noise spikes and free of signal peaks. Two tangent lines (one at the top of the tallest spike and the other at the bottom) were then drawn to the noise peaks. The vertical distance between these two tangent lines was measured as the noise. The signal was measured from the top of the response to the baseline (mid-point of noise peaks). Dividing the value obtained for signal with that of the noise gives us the SNR reported in this work. The accuracy of the method was estimated using Eq. 7. Accuracy ¼
jmv − t v j 100% tv
ð7Þ
where mv is the measured peak area value and tv is the true peak area value. To estimate repeatability Eq. 8 was used. repeatibility ¼ ½hv − l v 100%
ð8Þ
where hv and lv were the highest peak area and lowest peak area spread values for five replicate measurements. Experimental procedure 1. Reduced Mobility, Resoling Power, Theoretical Plates, and Signal-to-Noise Ratio Determinations: Certified reference material standards purchased from either Supelco (Bellafonte, PA) or Cerilliant (Austin, TX) were prepared using a micro-pipette. To achieve a required concentration, the appropriate volume was withdrawn from the 1 mg mL−1 stock solution and transferred to 2.0 mL natural microcentrifuge tube (Fischer Scientific). ESI solvent for the appropriate IMS mode was then added to 1 mL total volume. All standards were prepared to 10 μg cm−3. PIOGZ and ROSGZ were prepared by dissolving 7 × 10−4 g of the solid sample in 1 mL ESI solvent. The resulting solution was diluted to 10 μg cm−3 for a total volume of 1 mL with the ESI solvent. Five replicate measurements were taken for each standard on the ESI-HPIMS instrument. The gate pulse width used in the positive and negative ion modes were 100 μs and 50 μs, respectively. All solutions were stored in the refrigerator at 4 °C when not in use. 2. Calibration Studies: Certified reference material standards were prepared for 1, 5, 10, 15, and 30 μg cm−3 for each compound studied in a similar procedure as described in (1) above. Five replicate measurements were taken for each concentration on the ESI-HPIMS. The data was used to construct a calibration curve and Eq. 4 used to determine LOQ. To determine LOD, Eq. 3 was used. The relative standard deviation was determined using Eq. 9. RSDx ¼
Sx mean
ð9Þ
where Sx is the standard deviation at 3σ. For five replicate measurements (n–1 = 4), α = 0.01 (i.e. 99% confidence that analyteLOD was greater than zero, tα,n = 4.60) from the student t-table. Using this approach all detection limits for drugs and explosives were estimated. The gate pulse width used in the positive and negative ion modes were 100 μs and 50 μs, respectively. 3. Control Chart Studies: A control chart was created for morphine. The measurement was conducted to monitor the performance of the ESI-HPIMS. n = 40 replicate for 10 μg cm−3 morphine quality control standards were measured each week over a 10-week period. The gate pulse width used in the positive ion mode was 100 μs. 4. Determining Accuracy, Repeatability, Range, and Reporting Limit: n = 5 replicate for 10 μg cm−3 standard for each compound was spiked into the ESI solvent and analyze on the ESI-HPIMS instrument. Standards from two different manufacturers (Sigma Aldrich and Cerilliant) was used to determine accuracy. The Cerilliant standard was considered the true value. The data was used to calculate accuracy and repeatability according to Eqs. 7 and 8. The range and reporting limit were estimated from the calibration studies data. The gate
Table 2 Reduced mobility (Ko), resolving power (Rp), theoretical plates (N) and signal-to-noise ratio (SNR) for compounds investigated with ESI-HPIMS in positive and negative ion modes Analyte
*Ko (Lit. Ko)
Positive ion modea AMPHET 1.55 ± 0.01 (1.55) CBD m: 1.07 ± 0.02 d: 0.93 ± 0.02 COC 1.15 ± 0.02 (1.16) COD 1.13 ± 0.01 (1.12) HER 1.05 ± 0.03 (1.04) METHAMP 1.61 ± 0.02 (1.63) MOR 1.24 ± 0.01 (1.25) PHENT 1.58 ± 0.01 L-PHENE 1.57 ± 0.02 PIOGZ 1.11 ± 0.01 ROSGZ 1.14 ± 0.01 Negative ion modeb TNT 1.31 ± 0.02 Tetryl 1.22 ± 0.03 RDX m: 1.24 ± 0.02 d: 1.09 ± 0.01 HMX m: 1.11 ± 0.04 d: 0.99 ± 0.03 c 1.73, 1.54 2,6-DNT 1.43 ± 0.01 4-A-2,6-DNT m: 1.34 ± 0.02 d: 1.16 ± 0.03
Rp
N
SNR
48 ± 0.12 60 ± 0.22 62 ± 0.39 48 ± 0.34 58 ± 0.15 59 ± 0.35 60 ± 0.31 54 ± 0.45 57 ± 0.12 49 ± 0.13 63 ± 0.06 61 ± 0.08
12,787 ± 0.21 20,304 ± 0.85 21,115 ± 1.25 12,260 ± 0.72 18,670 ± 0.99 19,320 ± 1.08 19,980 ± 0.82 16,134 ± 0.54 18,032 ± 0.66 13,326 ± 0.44 22,028 ± 0.26 20,652 ± 0.84
21 ± 0.17 31 ± 0.14 27 ± 0.24 21 ± 0.31 38 ± 0.39 21 ± 0.09 24 ± 0.07 20 ± 0.16 18 ± 0.19 17 ± 0.21 17 ± 0.16 15 ± 0.02
62 ± 0.19 55 ± 0.34 65 ± 0.44 66 ± 0.64 64 ± 0.19 45 ± 0.19
21,334 ± 0.73 16,789 ± 0.61 23,622 ± 0.92 24,108 ± 0.77 22,427 ± 0.41 11,481 ± 0.12
35 ± 0.26 23 ± 0.17 22 ± 0.08 16 ± 0.18 21 ± 0.39 14 ± 0.07
39 ± 0.25 8350 ± 0.16 25 ± 0.28 49 ± 0.39 13,143 ± 0.55 22 ± 0.11 56 ± 0.21 17,486 ± 0.79 18 ± 0.27
*air, cm2 V−1 s−1 ; m = monomer; d = dimer a
Calibrated using literature Ko for 2,4-lutidine (1.95 cm2 V−1 s−1 )
b
Calibrated using literature Ko for citric acid (1.43 cm2 V−1 s−1 )
c
Extra peaks for HMX
Int. J. Ion Mobil. Spec.
pulse width used in the positive and negative ion modes were 100 μs and 50 μs, respectively. 5. Conditional Reduced Mobility Studies: 10 μg cm−3 for each compound was prepared and studied at gate pulse widths ranging from 10 to 100 μs. The data was used to calculate conditional reduced mobility for each ion according to Eq. 2.
Safety consideration The ESI-HPIMS instrument was operated at high voltages; only trained personnel were given the privilege to operate the instrument.
Results and discussions Reduced mobility, resolving power, theoretical plates, and signal-to-noise ratio A typical mobility spectra obtained with the ESI-HPIMS instrument for both the positive and negative ion modes is shown in Fig. 2. Experiments conducted in this investigation allowed us to determine reduced mobility (Ko), resolving power (Rp), theoretical plates (N), and signal-to-noise ratio (SNR). Table 2 lists Ko, Rp, N, and SNR for fifteen compounds studied in the positive and negative ion modes. All Ko values were determined by comparison to known reduced mobility values for 2,4-lutidine (1.95 ± 0.01 cm2 V−1 s−1) in the positive ion mode
Fig. 3 Example calibration spectra for CBD (a) and tetryl (b). The spectra of tetryl showed two extra peaks at Ko of 1.33 and 1.16 cm2 V−1 s−1, respectively
Int. J. Ion Mobil. Spec.
[40, 41] and citric acid (1.43 ± 0.02 cm2 V−1 s−1) in the negative ion mode [37] and their experimentally determined drift times. The Ko values reported compared well to literature values and matched already reported Ko values to 1–2%. Because Ko is the drugs or explosives ion’s adjusted for temperature and pressure, our results demonstrated K o values for drugs and explosives ion’s based on size, shape, and charge. As reported previously [32], CBD showed a monomer and a dimer at the 100 μs gate pulse width used for the measurement of the Ko values. The monomer and dimer peaks of CBD had K o of 1.07 ± 0.02 and 0.93 ± 0.02 cm2 V−1 s−1, respectively. The explosives 4-A2,6-DNT, HMX, and RDX also showed dimer peaks. Two other peaks that differ from the reactant ion peak were observed for HMX at Ko of 1.54 and 1.73 cm2 V−1 s−1, respectively (see Table 2 for all Ko values). The spectra for tetryl also showed two extra peaks at K o of 1.33 and 1.16 cm2 V−1 s−1, respectively. Theoretical resolving power (Rp), a measure of the separating efficiency for the ESI-HPIMS was estimated at 66 ± 0.64, using the operating parameters from the ESI-HPIMS. This value is acceptable for the commercial ESI-HPIMS used in this investigation. Rp values could be much improved with a drift tube construction that incorporates greater field uniformity for the ions along the drift axis and increased accuracy in the time base from better gate control electronics that may
Table 3 Summary of calibration studies parameters for drugs and explosives investigated using the ESI-HPIMS
provide a narrower and more precise gate pulses. Eq. 5 describes this separation efficiency in terms of resolving power. In Table 2, the Rp values for all drugs and explosives studied are listed, where it can be seen that the values ranged from 41 to 63. These values are reported for a 100 μs gate pulse width. It was reported previously that decreasing the gate pulse width will result in an increased Rp value [22, 32]. Another approach that was used to enable IMS to be compared to other chromatographic methods and more effectively demonstrates its separation ability is to recast the resolving power equation in Eq. 5 in terms of number of theoretical plates, N, as shown in Eq. 6 [18, 49]. Theoretical plates reported for the standalone ESIHPIMS in Table 2 approached 23,000 for a single charge ion in less than 30 ms. This was achieved despite the fact that the standalone ESI-HPIMS used very modest voltages in this investigation. It is a further demonstration that standalone ESI-HPIMS has separation capability approaching that of HPLC [31, 49, 50]. High resolving power is directly related to the peak capacity of an analytical method. The signal-to-noise (SNR) ratio measured for the compounds studied at 100 μs gate pulse width are reported in Table 2. This investigation only analyze SNR at 100 μs gate pulse width. As reported previously [22, 32], we do not expect increasing or decreasing the gate pulse width will equate to improving SNR in an IMS system.
Analyte
B1/ppm
Bo
R2
LOD/ng
LOQ/ng
Positive ion modea AMPHET CBD
5.134 ± 0.004 12.959 ± 0.006
0.108 ± 0.017 0.159 ± 0.025
0.9987 0.9968
0.45 ± 0.01 2.58 ± 0.02
5.27 ± 0.02 7.28 ± 0.01
COC COD HER METHAMP MOR PHENT L-PHENE PIOGZ ROSGZ Negative ion modeb
4.175 ± 0.001 7.076 ± 0.004 0.003 ± 0.001 4.178 ± 0.002 5.304 ± 0.001 10.100 ± 0.006 4.254 ± 0.007 0.069 ± 0.003 0.857 ± 0.012
0.004 ± 0.002 0.002 ± 0.001 0.002 ± 0.001 0.037 ± 0.006 0.011 ± 0.001 0.032 ± 0.018 0.055 ± 0.002 0.174 ± 0.006 0.156 ± 0.008
0.9989 0.9991 0.9979 0.9988 0.9962 0.9975 0.9958 0.9974 0.9992
2.91 ± 0.03 2.81 ± 0.02 2.93 ± 0.07 1.28 ± 0.02 2.97 ± 0.06 0.54 ± 0.01 1.72 ± 0.03 1.11 ± 0.02 1.17 ± 0.03
7.27 ± 0.04 8.46 ± 0.13 8.52 ± 0.04 7.14 ± 0.01 8.63 ± 0.03 4.95 ± 0.01 4.63 ± 0.03 5.77 ± 0.05 5.86 ± 0.01
TNT Tetryl RDX HMX 2,6-DNT 4-A-2,6-DNT
0.059 ± 0.002 0.002 ± 0.001 0.006 ± 0.002 0.002 ± 0.001 0.003 ± 0.001 0.002 ± 0.001
0.003 ± 0.001 0.032 ± 0.002 0.030 ± 0.001 0.048 ± 0.002 0.018 ± 0.001 0.072 ± 0.003
0.9986 0.9959 0.9921 0.9963 0.9947 0.9899
0.64 ± 0.07 0.96 ± 0.04 1.25 ± 0.03 1.18 ± 0.06 0.57 ± 0.02 0.48 ± 0.01
5.16 ± 0.01 5.34 ± 0.02 7.28 ± 0.11 6.24 ± 0.06 4.78 ± 0.04 4.11 ± 0.03
The calibration summary for each response is given by the following equation: Peak area response = Bo + B1 (ppm−1 ) × [concentration] (ppm) where Bo is the intercept and B1 is the sensitivity or slope
Int. J. Ion Mobil. Spec.
Calibration studies Figure 3 shows example spectra for calibration studies conducted with the ESI-HPIMS. Table 3 summarizes the calibration responses for drugs and explosives studied. All sensitivities (slope) and intercepts are reported in Table 3. The limit of detection (LOD) commonly described as the smallest quantity of analyte significantly different from the blank was estimated by injecting 0.5 ppm concentration for each compound. This concentration was discernable from the instrument noise for each of the compounds studied. Table 3 also summarizes the detection limits estimated using Eqs. 3 and 9. The lowest LOD was 0.45 ng material recorded for AMPHET with a sensitivity of 5.134 ± 0.004, and the highest LOD was 2.97 ng material recorded for morphine with a sensitivity of 5.304 ± 0.001. A signal that is 10 times greater than the noise or the smallest amount that can be measured with reasonable accuracy is often referred to as the limit of quantitation (LOQ). This study estimated LOQs for drugs and explosives using Eq. 4. All data for LOQ determination are summarized in Table 3. Based on our calculations, the lowest LOQ was 4.11 ng material recorded for 4-A-2, 6DNT, and the highest LOQ was 8.63 ng material recorded for morphine. All LODs and LOQs parts per million concentrations were converted to mass using the reported ESI flow rates for both positive and negative ion modes and the sample run time of 30 s (see Table 1 for ESI flow rates). The correlation coefficient often described as R2 measures the observed variance that can be attributed to a straight line plot. For a major unknown component, a value of R2 above 0.995 is deemed a good fit for many purposes [45, 51]. The data reported in Table 3 indicated that most of the drugs and explosives studied produced linear responses with R2 values >0.99. In most chemical analysis, the response of a procedure has to be evaluated from known standards (the so called calibration curves) so that response to an unknown quantity can be evaluated. One good approach to get the best straight line through experimental data points that have some scatter points and do not perfectly fit on a straight line is the method of least squares. The approach to conducting method of least squares was described elsewhere [33]. The calibration data collected for drugs and explosives for this investigation was subjected to the method of least square approach. Figure 4 shows example method of least square calibration plots generated after the data was subjected to the approach. All R2 values in Fig. 4 equals 1 indicating a perfect fit. This approach is sometimes used in method validation studies to demonstrate that a straight line model accounts for all sources of variance during the calibration studies.
Fig. 4 Drugs (a) and explosives (b) example calibration curves generated with the method of least square. R2 = 1 for each calibration line indicating a perfect fit
Control chart studies To illustrate a visual representation of confidence intervals for a Gaussian distribution, a control chart is normally used. This chart warns an analyst when a property being measured strayed dangerously away from an intended target value. This property was tested in this investigation by studying MOR over a 10-week period. The control chart generated for MOR using the ESI-HPIMS instrument is shown in Fig. 5. n = 40 replicate quality control standards were measured each week over a 10-week period. The data collected in weeks 3 & 8 falls above the action lines. This is normally a pointer that a system or instrument must be totally shut down for troubleshooting. This approach might find application in forensic investigations especially when confirmatory tests [52] are conducted. Such measurements are necessary when crime samples are analyzed to increase on the reliability of the data. Thus, a control chart is an effective way to either monitor the performance of an instrument or may be a useful tool for establishing reliability of confirmatory tests in forensic investigations.
Int. J. Ion Mobil. Spec.
Fig. 5 Control chart for morphine. The measurement conducted to monitor the performance of the ESI-HPIMS. n = 40 replicate quality control standards were measured each week over a 10-week period. The data collected in weeks 3 & 8 falls above the action lines. This is normally a pointer to a system shut down for troubleshooting. Such measurements are necessary to increase on the reliability of the data when confirmatory tests are required for crime samples
showed the lowest accuracy of 1.78 ± 0.04%. Repeatability describes the spread in results when one person uses one procedure to analyze the same sample by the same procedure multiple times. This study estimated repeatability for n = 5 replicate measurements. Table 4 shows that METHAMP gave the highest precision of 0.34 ± 0.11% and COC shows the lowest precision of 4.93 ± 0.09%. It should be noted that repeatability is just one approach to express precision of an analytical method. Several other approaches exist for expressing precision of an analytical method. The concentration interval over which linearity, accuracy, and repeatability are acceptable [33] is known as the range. As an example of specifying range for CBD; 0.94 to 37.42 μg cm−3 is the concentration range that provided a correlation coefficient of R2 = 0.9968, accuracy of 1.64 ± 0.02%, and repeatability of 1.47 ± 0.14% for CBD. This same generalization can be drawn for all the other drugs and explosives studied in this investigation. In most regulatory work, such as food analysis,
Determining accuracy, repeatability, range, and reporting limit Table 4 summarizes accuracy, repeatability, range and reporting limit investigated in this study. The accuracy sometimes referred to as nearness to a true value was estimated through spiking of the same standards from two different chemical manufacturers. According to Table 4, PIOGZ showed the highest accuracy of 0.76 ± 0.03% and HER Table 4 Accuracy, repeatability, range and reporting limit for compounds investigated with ESIHPIMS in positive and negative ion modes Analyte
a
b
1.92 ± 0.11 1.47 ± 0.14 4.93 ± 0.09 0.62 ± 0.05 1.25 ± 0.11 0.34 ± 0.11 1.09 ± 0.08 0.80 ± 0.07 0.77 ± 0.02 0.89 ± 0.02 0.87 ± 0.02
0.94–30.0 0.91–37.9 0.94–37.4 0.89–38.0 1.11–29.9 0.90–30.0 1.14–30.0 1.04–30.0 1.02–35.2 1.57–38.5 1.07–37.5
2.25 12.9 14.6 14.1 14.7 6.40 14.9 2.70 8.60 5.55 5.85
1.06 ± 0.12 1.27 ± 0.09 0.31 ± 0.02 0.56 ± 0.04 0.45 ± 0.02 0.76 ± 0.03
0.91–30.0 0.86–33.9 0.96–28.0 0.98–25.5 0.77–32.6 0.74–34.0
3.20 4.80 6.25 5.90 2.85 2.40
Accuracy (%) Repeatability (%)
Positive ion modea AMPHET 1.11 ± 0.05 CBD 1.64 ± 0.02 COC 1.46 ± 0.07 COD 0.91 ± 0.01 HER 1.78 ± 0.04 METHAMP 1.18 ± 0.02 MOR 1.65 ± 0.02 PHENT 0.86 ± 0.02 L-PHENE 1.10 ± 0.01 PIOGZ 0.76 ± 0.03 ROSGZ 0.92 ± 0.05 Negative ion modeb TNT 1.23 ± 0.06 Tetryl 1.32 ± 0.07 RDX 0.96 ± 0.05 HMX 1.06 ± 0.04 2,6-DNT 0.72 ± 0.02 4-A-2,6-DNT 0.89 ± 0.01 a
Range in μg cm−3 concentration
b
Reporting limit in ng material
Range
Reporting Limit
Fig. 6 Conditional reduced mobility (Kc), for drugs (a) and explosives (b) calculated using reduced mobility and width-at-half-height of each peak. The intensity on the y-axis is × 103 and the unit of K c is cm2 V−1 s−2. The result demonstrate that Kc can be used to differentiate or separate drugs and explosives from each other
Int. J. Ion Mobil. Spec.
it is sometimes necessary to establish a concentration (reporting limit) below which the concentration for a particular analyte is reported as Bnot detected^. Reporting limits determined for this investigation using a concentration 5 times
the LOD are summarized in Table 4. Note that the reporting limit does not mean that a particular analyte is not detected. It simply means that a particular analyte is below the prescribed level for reporting.
Fig. 7 Representative ESI-HPIMS spectra for explosive 4-A-2,6-DNT at gate pulse width of 10 to 100 μs. At responses below 20 μs gate pulse width 4A-2,6-DNT disappeared into the background
Int. J. Ion Mobil. Spec.
Conditional reduced mobility studies An application for the conditional reduced mobility (Kc) was reported previously [32]. As discussed previously, Kc uses the hypothesis that if two peaks have similar Ko values, it is possible to use Kc to differentiate between the two peaks because the peaks will most certainly not have the same Δt0.5. Thus, it is possible to use Kc to differentiate between two peaks with matched or slightly different Ko values. All Kc values were calculated using Eq. 2. The effect of Kc on drugs and explosives studied in this investigation is demonstrated in Fig. 6. In Fig. 6a, plot of Kc versus gate pulse width for AMPHET, COC, CBD, HER, L-PHERN, PHENT, PIOGZ, and ROSGZ are shown. In a routine operation, IMS could not separate all of these seven drug ions from each other. When Kc is calculated, however, it could be seen from Fig. 6a that all seven peaks could be differentiated at 30, 40, 60, and 70 μs gate pulse widths. Several other gate pulse widths showed that a combination of any of the seven drug ions could be differentiated from each other using Kc. For example, at 50 μs gate pulse width, CBD, HER, PHENT, L-PHERN, and AMPHET could be differentiated from each other. The Kc values for CBD, HER, PHENT, L-PHERN, and AMPHET at 50 μs gate pulse width was 3.01, 4.28, 7.29, 8.14, and 8.67 × 103 cm2 V−1 s−2, respectively. At 10 μs gate pulse width, only three drug ions (PIOGZ, PHENT and ROSGZ) showed responses. Kc for ROSGZ, PHENT and PIOGZ at 10 μs gate pulse width occurred at 5.45, 8.91, and 9.55 × 103 cm2 V−1 s−2, respectively. Figure 6b shows a plot of Kc versus gate pulse width for six explosives (TNT, HMX, RDX, 2,6-DNT, 4-A-2,6-DNT, and tetryl), and the standard citric acid. From this plot, it can be seen that all six explosives and the citric acid can be separated at 30 and 50 μs gate pulse widths. Several other gate pulse widths demonstrated that any combination of the explosive ions can be separated. Example spectra for the explosive 4-A-2,6-DNT at different gate pulse widths is shown in Fig. 7. Kc values for both positive and negative mode IMS demonstrated that the drifting pattern in an IMS can be taken advantage of to differentiate between ions of similar Ko values. The Kc is thus another approach to determine the identity of an ion; a parameter that may be applied to help aid identification of drugs and explosives ions with IMS devices.
figures of merit must be used when validating an analytical method, especially for cases where the method will be transferred between laboratories. In this investigation, separation efficiency normally reported as Rp for an IMS was estimated at 66 ± 0.64 for the ESI-HPIMS. LOD, known as the minimum amount of analyte that can be detected falls in the range 0.45–2.97 ng of material electrosprayed into the ESI-HPIMS. LOQ, known as the minimum amount of analyte that can be detected quantitatively falls in the range 4.11–8.63 ng of material electrosprayed into the ESI-HPIMS. Kc is a measure of the identity of an explosive or drug ion; a parameter that may be applied to help aid IMS devices for drugs and explosives detection. Analytical figures of merits especially Kc introduced in this study is an effective parameter for establishing a unique identity of a drug or explosive. A control chart is an effective way to monitor the performance of an instrument and may be a useful tool for establishing reliability with confirmatory data generation in forensic investigations. Acknowledgements We gratefully acknowledge the support of the Professional Development Committee (PDC) at Winston-Salem State University for their support of this work. The authors also thank Mr. David Pollard for his support during this investigation.
References 1.
2. 3.
4.
5.
6. 7.
8.
Conclusion 9.
Validating an analytical method is proving that the method is suitable for its intended purpose. One goal in analytical measurement is to obtain reproducible, reliable and accurate data. Method validation plays a major role in achieving this goal. Data generated from method validation can be used to evaluate the quality, reliability and consistency of analytical results, which is an important part of any good analytical laboratory practice. Multiple
10.
11.
Asbury G. R., Klasmeier J., and Hill H. H. Jr. (2000). Analysis of explosives using electrospray ionization/ion mobility spectrometry (ESI/IMS). Talanta 50: 1291–1298. Neves J. L., Haigh P. E., Wu C., and McGann W. J. (2003). ITMSMS analysis of smokeless powder. Int J Ion Mobil Spectrom 6: 1–3. Marr A. J., and Groves D. M. (2003). Ion mobility spectrometry of peroxide explosives TATP and HMTD. Int J Ion Mobil Spectrom 6: 59–62. Perr J. M., Furton K. G., and Almirall J. R. (2005). Solid phase microextraction ion mobility spectrometer interface for explosive and taggant detection. J Sep Sci 28: 177–183. Gruznov VM, Baldin MN, Filonenko VG (2004) High-speed gas analysis for explosives detection, NATO Science Series, II: Mathematics, Physics and Chemistry 167 (Vapor and Trace Detection of Explosives for Anti-Terrorism Purposes), 87 Baumbach J. I. (2006). Process analysis using ion mobility spectrometry. Anal Bioanal Chem 384: 1059–1070. Buryakov I. A. (2006). Solution of the equation of continuity for ions moving in a gas under the action of a periodic asymmetric alternating waveform. Tech Phys Lett 32: 67–69. Luong J., Gras R., Van Meulebroeck R., Sutherland F., and Cortes H. (2006). Gas Chromatography with State-of-the-Art micromachined differential mobility detection: operation and industrial applications. J Chromatogr Sci. 44: 276–282. Buryakov IA (2004) NATO Science Series, II: The analytical characteristics of ion mobility increment spectrometer during the detection of explosive vapors and products of their degradation. Mathematics, Physics and Chemistry 167 (Vapor and Trace Detection of Explosives for Anti-Terrorism Purposes), 113 Wu C., Siems W. F., and Hill H. H. Jr. (2000). Secondary electrospray ionization ion mobility spectrometry/mass spectrometry of illicit drugs. Anal Chem 72: 396–403. Lawrence A. H. (1989). Characterization of benzodiapine drugs by ion mobility spectrometry. Anal Chem 61: 343–349.
Int. J. Ion Mobil. Spec. 12.
13. 14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Miki A., Tatsuno M., Katagi M., Nishikawa M., and Tsuchihashi H. (1997). Analysis of illicit drugs by ion mobility spectrometry. Jpn J Forensic Toxicol 15: 142–148. Carroll J. J., Le Tri, and DeBono R. (2004). An ultrafast alternative to HPLC for cleaning validation. Am Lab 36(4): 32–34. Debono R., Stefanou S., Davis M., and Walia G. (April 2002). Using Ion Mobility Spectrometry for Cleaning Verification in Pharmaceutical Manufacturing. Pharm Technol N Am 26(4): 72–78. Keller T., Keller A., Tutsch-Bauer E., and Monticelli F. (2006). Application of ion mobility spectrometry in cases of forensic interest. Forensic Sci Int 161: 130–140. Gillen G., Mahoney C., Wight S., and Lareau R. (2006). Characterization of high explosive particles using cluster secondary ion mass spectrometry. Rapid Commun Mass Spectrom 20: 1949– 1953. Tarver E. E. (2004). External second gate, fourier transform ion mobility spectrometry: parametric optimization for detection of weapons of mass destruction. Sensors 4: 1–13. Kanu A. B., Haigh P. E., and Hill H. H. Jr. (2005). Surface detection of chemical warfare agent simulants and degradation products. Anal Chim Acta 553: 148–159. Steiner W. E., Haigh P. H., Clowers B. H., and Hill H. H. Jr. (2003). Secondary ionization of chemical warfare agent simulants: atmospheric pressure ion mobility time-of-flight mass spectrometry. Anal Chem 75: 6068–6076. Kanu A. B., Hill H. H. Jr., Gribb M., and Walters R. (2007). A small subsurface ion mobility spectrometer sensor for detecting environmental soil-gas contaminants. J Environ Monit 9: 51–60. Clowers B. H., Steiner W. E., Dion H. M., Matz L. M., Tam M., Tarver E. E., and Hill H. H. Jr. (2001). Evaluation of sulfonylurea herbicides using high resolution electrospray ionization ion mobility quadrupole mass spectrometry. Field Anal Chem Techno 5: 302–312. Kanu A. B., Gribb M. M., and Hill H. H. Jr. (2008). Predicting optimal resolving power for ambient pressure ion mobility spectrometry. Anal Chem 80: 6610–6619. Palmer P. T., and Limero T. F. (2001). Mass spectrometry in the U.S. Space Program: past present and future. J Am Soc Mass Spectrom 12: 656–675. Vautz W., Zimmermann D., Hartmann M., Baumbach J. I., Nolte J., and Jung J. (2006). Ion mobility spectrometry for food quality and safety. Food Addit Contam 23(11): 1064–1073. Hoaglund-Hyzer C. S., Lee Y. J., Counterman A. E., and Clemmer D. E. (2002). Coupling ion mobility separations, collision activation techniques, and multiple stages of MS for analysis of complex peptide mixtures. Anal Chem 74: 992–1006. Sheibani A., and Haghpazir N. (2014). Application of ion mobility spectrometry for the determination of tramadol in biological samples. J Food Drug Anal 22: 500–504. Jünger M., Vautz W., Kuhns L., Hofmann S., Ulbricht S., Baumbach J. I., Quintel M., and Perl T. (2012). Ion mobility spectrometry for microbial volatile organic compounds: a new identification tool for human pathogenic bacteria. Appl Microbiol Biotechnol 93(6): 2603–2614. Vautz W., Baumbach J. I., Westhoff M., Züchner K., Carstens E. T. H., and Perl T. (2010). Breath sampling control for medical application. Int J Mass Spectrom 13(1): 41–46. Crawford C. L., Hauck B. C., Tufariello J. A., Harden C. S., McHugh V., Siems W. F., and Hill H. H. Jr. (2012). Accurate and reproducible ion mobility measurements for chemical standard evaluation. Talanta 101: 161–170. Kanu A. B., and Hill H. H. Jr. (2007). Identity confirmation of drugs and explosives in ion mobility spectrometry using a secondary drift gas. Talanta 73: 692–699. Kanu A. B., Wu C., and Hill H. H. Jr. (2008). Rapid preseparation of interferences for ion mobility spectrometry. Anal Chim Acta 610: 125–134.
32.
Kanu A. B., and Leal A. (2016). Identity efficiency for high-performance ambient pressure ion mobility spectrometry. Anal Chem 80: 6610–6619. 33. Harris D. C., and Lucy C. A. (2016). Quantitative Chemical Analysis, 9th edn., W.H. Freeman & Company, New York. 34. Wittmer D., Chen Y. H., Luckenbill B. K., and Hill H. H. Jr. (1994). Electrospray ionization ion mobility spectrometry. Anal Chem 66: 2348–2355. 35. Midey A. J., Camacho A., Sampathkumaran J., Krueger C. A., Osgood M., and Wu C. (2013). Highperformance ion mobility spectrometry with direct electrospray ionization (ESI-HPIMS) for the detection of additives and contaminants in food. Anal Chim Acta 804: 197–206. 36. Midey A. J., Patel A., Moraff C., Krueger C. A., and Wu C. (2013). Improved detection of drugs of abuse using high-performance ion mobility spectrometry with electrospray ionization (ESIHPIMS) for urine matrices. Talanta 116: 77–83. 37. Hilton C. K., Krueger C. A., Midey A. J., Osgood M., Wu J., and Wu C. (2010). Improved analysis of explosives samples with electrospray ionization-high resolution ion mobility spectrometry (ESIHRIMS). Int J Mass Spectrom 298: 64–71. 38. Krueger C. A., Hilton C. K., Osgood M., Wu J., and Wu C. (2009). High resolution electrospray ionization ion mobility spectrometry. Int J Ion Mobil Spectrom 12: 33–37. 39. Krueger C. A., Midey A. J., Kim T., Osgood M., Wu J., and Wu C. (2011). High performance ion mobility spectrometry as a fast and low cost green analytical technology part I: analysis of nutritional supplements. Int J Ion Mobil Spectrom 14: 71–79. 40. Thomas C. L. P., Rezgui N. D., Kanu A. B., and Munro W. A. (2000). Measuring the temperature of the drift gas in an ion mobility spectrometer: A technical note. Int J Ion Mobil Spectrom 5(1): 31–36. 41. Shumate C., St. Louis R. H., and Hill H. H. Jr. (1986). Table of reduced mobility values from ambient pressure ion mobility spectrometry. J Chromatogr A 373: 141–173. 42. Dwivedi P., Wu P., Kplosh S. J., Puzon G. J., Xun L., and Hill H. H. Jr. (2008). Metabolic profiling by ion mobility mass spectrometry (IMMS). Metabolomics 4: 63–80. 43. Kanu A. B., Hampikian G., Brandt S. D., and Hill H. H. Jr. (2010). Ribonucleotide and ribonucleoside determination by ambient pressure ion mobility spectrometry. Anal Chim Acta 658: 91–97. 44. Kanu A. B., Brandt S. D., Williams M. D., Zhang N., and Hill H. H. Jr. (2013). Analysis of psychoactive cathinones and tryptamines by electrospray ionization atmospheric pressure ion mobility time-offlight mass spectrometry. Anal Chem 85: 8535–8542. 45. Miller J. C., and Miller J. N. (1993). Statistics for Analytical Chemistry, 3rd edn., Ellis Horwood, Hemstead, Herts. 46. Spangler G. E. (2002). Expanded theory for the resolving power of a linear ion mobility spectrometer. Int J Mass Spectrom 220: 399– 418. 47. Siems W. F., Wu C., Tarver E. E., Hill H. H. Jr., Larsen P. R., and McMinn D. G. (1994). Measuring the resolving power of ion mobility spectrometers. Anal Chem 66: 4195–4201. 48. Rokushika S., Hatano H., Baim M. A., and Hill H. H. Jr. (1985). Resolution measurement for ion mobility spectrometry. Anal Chem 57: 1902–1907. 49. Ashbury G. R., and Hill H. H. Jr. (2000). Evaluation of Ultrahigh Resolution Ion Mobility Spectrometryas an Analytical Separation Device in Chromatographic Terms. J Microcolumn Sep 12: 172– 178. 50. Kanu A. B., and Hill H. H. Jr. (2008). Ion mobility detection for gas chromatography. J Chromatogr A 1177: 12–27. 51. de Levie R. (2003). Two linear correlation coefficients. J Chem Ed 80: 1030–1031. 52. Kanu A. B., and Kaplan L. J. (2016). The quest for confirmatory data in crime scene investigations. Chem Educator 21: 231–239.