ARTICLE pubs.acs.org/ac
Evaluating Headspace Component VaporTime Profiles by Solid-Phase Microextraction with External Sampling of an Internal Standard William MacCrehan,* Stephanie Moore, and Michele Schantz Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States
bS Supporting Information ABSTRACT: The vaportime profiles of explosive materials are of valuable interest to Homeland Security, providing critical information that can aid in the detection of explosive-containing devices. An approach is described that achieves reproducible characterization of volatile components as a function of time based on comparison of the sample response to an externally sampled internal standard (ESIS). Utilizing nonequilibrium solid-phase microextraction (SPME) measurements, this SPMEESIS technique improves reproducibility (reported as percent relative standard deviation) of vaportime profiles by approximately an order of magnitude and allows for an equitable comparison of the target compound between diverse materials. Two odorants associated with canine detection of explosives, 2-ethyl-1-hexanol and 2,4-dinitrotoluene, are used to optimize parameters for the SPME-ESIS technique.
T
he stability of compounds, environmental impact conditions, and mechanical influences can all affect the rate of disappearance of chemical odors over time. One example is the detection of the impurity 2,4-dinitrotoluene (2,4-DNT) present in 2,4,6-trinitrotoluene (TNT). 2,4-DNT is a volatile component in headspace of the explosive TNT and is the target analyte used by instrumentation and canines for vapor detection. However, simulated TNT residues lose their 2,4-DNT content in a matter of hours when exposed to air on various substrates.1 Therefore, dynamically monitoring compounds such as 2,4-DNT can help determine which volatile compounds are present, for how long, and how external environmental factors can affect these parameters. This study focuses on two materials identified as canine target odorants (2-ethyl-1-hexanol in composition C-4 and 2,4-DNT in TNT2). In a previous investigation, the ATASS procedure, or automated training aid simulation using solid-phase microextraction (SPME), was developed to monitor the dynamic headspace above canine training aid samples of the improvised explosive triacetone triperoxide (TATP) over time.3 Although the overall vaportime patterns for TATP release from different aid formulations could be qualitatively reproduced, the absolute peak areas differed by a factor of 2 when the procedure was repeated with additional fibers of the same composition at a later date. It was clear that, in order to accurately compare between samples over time, further development was needed to properly normalize the results and increase instrumental and SPME fiber reproducibility. SPME, or solid-phase microextraction, employs an absorbentcoated glass fiber to preconcentrate sample analytes.4 The analytes This article not subject to U.S. Copyright. Published 2011 by the American Chemical Society
from the sample matrix are absorbed into a layer of polymeric organic phase, followed by desorption of analytes in a chromatographic inlet system.4,5 Typically, for headspace analysis the fiber is allowed to come to equilibrium with the volatile sample components, a process that can take from seconds to hours, providing a one-time characterization of the sample. However, a method for standardizing this approach to achieve reproducible and quantitative results via SPME for solid materials is not straightforward. Traditional SPME calibration methods such as external standardization (calibration curve), internal standardization, and standard addition have been used to quantify SPME results; however, each method has advantages and disadvantages depending on the sample type.6 Typically, solid compounds sampled by SPME are externally calibrated via liquid standards or standard addition through spiking.7,8 With either method it can be a challenge to obtain relative standard deviations (RSDs) as low as 5%, which are typical for SPME measurements of aqueous samples.9 Recently, inkjet microdrop printing directly onto the SPME fiber has been used to prepare analyte mass response curves.10 This technique is similar to the syringe-fiber transfer method11 where a solution is deposited directly onto the fiber and the solvent is allowed to evaporate. However, microdrop printers are not yet widespread technology in analytical laboratories10 and the syringe-fiber transfer approach was determined to be unsuitable for calibration of volatile compounds.11 Received: July 22, 2011 Accepted: October 3, 2011 Published: October 03, 2011 8560
dx.doi.org/10.1021/ac201888r | Anal. Chem. 2011, 83, 8560–8565
Analytical Chemistry
ARTICLE
One of the most powerful approaches to reduce uncertainty and provide a path to quantitation is to use an internal standard (IS). Ideally, the IS is a compound that is structurally similar to the analyte of interest, exhibits analogous behavior in the analysis scheme, and is not present in the sample. However, for neat solid samples, like many of the explosive-related materials that we are studying, the means to realistically incorporate an internal standard into the sample matrix is not straightforward.7,12 The concept of loading an internal standard onto a SPME fiber prior to sampling the analyte has been previously described with application in techniques such as stepwise SPME, in-fiber standardization, and kinetic calibration.6,11,13,14 Spiking a fiber with an internal standard prior to exposing the fiber to the analyte can provide many of the advantages of a conventional internal standard, such as correcting for measurement response drift, sample loss, and matrix effects. In addition, using an internal standard in SPME can provide an overall improvement in accuracy and precision.11 Stepwise SPME has been applied in forensic analysis to develop an air sampling device that mimics canine scent detection.15 However, previous stepwise SPME studies have focused on compounds with high vapor pressures that may be calibrated by use of a gas-generating device.14,15 Previously identified canine target odorants 2-ethyl-1-hexanol and 2,4-dinitrotoluene,2 studied here, have much lower reported vapor pressures16,17 than the compounds previously calibrated via gas-generating systems.14,15 In-fiber standardization requires measurements of samples at near equilibrium (t ≈ t95) or preequilibrium extraction of materials that exhibit isotropic absorption/desorption.6,11,13 Using an approach that requires SPME measurements near equilibrium can be time-consuming. For example, investigation of the headspace composition of explosives and explosive odorants would require several hours for some of the components of interest to reach equilibrium with the fiber. Since the odor profiles of training aid materials may change over the time scale of minutes,3 the long SPME absorption times associated with near-equilibrium sampling do not provide sufficiently rapid characterization of the dynamic training aid behavior. More rapid sampling may be achieved as dominant odorants, such as 2,4-DNT,2 adsorb sufficiently to the fiber within seconds for reliable detection. In this paper, a simple approach is presented to normalize the results obtained via headspace SPME by frequent sampling. By taking the ratio of the relative signal response of the analyte to that of a previously sampled internal standard, neat explosive compounds and canine training aid materials can be compared with high reproducibility over a time frame of many hours. All measurements are indexed to the saturated headspace of the analyte and internal standard. The standard is not a true “internal” standard since it is not present within the sample vial; however, it is not an “external” standard since it is present on the fiber when the sample is interrogated. Therefore, we introduce the terminology for this “externally sampled internal standard” as ESIS. For analysis, the ratio (A/E) of the peak areas for the analyte to that of the ESIS is calculated as A=E ¼
peak area of analyte peak area of ESIS
ð1Þ
This SPME-ESIS technique was applied to provide a quick and reproducible way to compare canine training aid materials to the respective neat explosive compounds. For this study, two canine odorants, 2-ethyl-1-hexanol (2-EH), a liquid associated with the
plastic explosive C-4,2,18 and 2,4-dinitrotoluene (2,4-DNT), a solid associated with TNT,2 were selected. Sample vials containing 2-EH and 2,4-DNT were studied in static, closed vial systems at vapor-saturated equilibrium. To investigate dynamic systems, 2-EH was also investigated in open vials. The dynamic ESIS ratio approach is useful for comparison of functional canine training aids, which are ultimately utilized in a variety of open systems with varying environmental conditions. To conclude this study, SPME-ESIS was used to monitor and compare the vaportime profiles for three vapor delivery mechanisms for the canine target odor 2-EH.
’ EXPERIMENTAL SECTION Sample Preparation. Note: Certain commercial equipment, instruments, or materials are identified in this paper to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. To investigate the potential of SPME-ESIS, both solid and liquid samples were tested. The liquid samples included 2-EH (g99.6%, manufactured by Eastman Chemical Co. and distributed by Aldrich Chemical Co.) and 1-octanol (99%, Chem Service). One milliliter of 1-octanol (1-Oct) was chosen as the ESIS for 2-EH, as it is a geometric isomer having a similar vapor pressure19 and is baseline-separated in the gas chromatograph (GC). For solid samples, three separate ≈1 mg samples of 2,4-DNT (military grade) were analyzed with ≈1 mg of 2,3dinitrotoluene (2,3-DNT, 99%, SigmaAldrich, St. Louis, MO) as the ESIS. We learned that it takes a considerable time (≈ 10 h) for solid samples to reach vapor-saturated equilibrium in the septum-capped 20 mL sample vial at 30 °C. Headspace equilibrium was indicated by achievement of a steady SPME/GC/mass spectrometry (MS) signal. All solid and liquid samples were evaluated within a 20 mL autosampler vial in a temperatureregulated tray set at 30 °C (Gerstel Inc., Linthicum, MD). Analyte samples were evaluated in both static systems (closed vials) and dynamic systems (open vials). The dynamic system samples were first allowed to equilibrate (≈0.5 h) with a septum cap in place; the septum was then removed and the samples were evaluated. Without the septum present, the aluminum screw-top caps afforded a means for the vapor to escape through the 8 mm circular hole to simulate field conditions. ESIS vials containing either 1-Oct or 2,3-DNT were septum-capped and remained closed to avoid loss of the internal standard over time. Once static and dynamic parameters were established (detection limits, optimum SPME absorption/desorption times, etc.) SPME-ESIS was used to compare between actual canine training aids at 30 °C containing 2-EH. Overall, four separate 2-EH vapor delivery mechanism systems were placed into sample vials. The control consisted of ≈1 mL of neat 2-EH liquid to provide a steady 2-EH release rate. Additional samples consisted of ≈1 μL of 2-EH, a permeation bag, and a particulate odor delivery. The permeation bag contained ≈2 mL of 2-EH heat sealed in fluorpolymer permeation bag. The particulate sample consisted of three samples (≈30 mg each) of a 1% 2-EH material. This canine training material consisted of a chromatographic support (C18 Vydac silica TP2030, Grace Discovery Sciences, Deerfield IL) coated with 1% (mass fraction) of 2-EH. As previously reported, explosives coated onto octadecylsilane-modified silica 8561
dx.doi.org/10.1021/ac201888r |Anal. Chem. 2011, 83, 8560–8565
Analytical Chemistry (C18) with a 2030 μm particle size provided a practical surrogate for real explosive residues.20 The procedure for rotoevaporatory coating of the 2-EH from 2-methylbutane onto the silica substrate is similar to the procedure described previously.20 Gas Chromatography/Mass Spectrometry Measurements. An Agilent 6890 gas chromatograph with a Hewlett-Packard 5973 mass-selective detector (Agilent Technologies, Santa Clara, CA) fitted with a nonpolar Agilent DB-5MS column (30 m, 0.25 mm i.d., 0.25 μm film) and a 5 m Restek Siltek guard column of 0.25 mm i.d. was used for analysis of all samples. The GC injection method was splitless with a constant flow rate of 1 mL/min helium. The mass spectrometer was operated in scan mode from 40 to 400 m/z, the source temperature was 230 °C, and analyzer temperature was 150 °C for all samples. The mass spectrometer was tuned weekly using perfluorotributylamine (PFTBA). Agilent Enhanced MSD ChemStation was used to control the GC/MS parameters. The enhanced data analysis feature was used in conjunction with the NIST Mass Spectral Search Program (NIST/ EPA/NIH Mass Spectral Library, version 2.0d) to analyze all spectra collected. The GC oven temperature program for the 2-EH/1-Oct separation began at 40 °C (5 min hold) and increased to 130 °C (0 min hold) at 10 °C/min. The injector temperature was held constant at 110 °C. For the DNT experiment, the injector was held constant at 220 °C and the temperature program began at 80 °C (0 min hold), increased initially to 115 °C (0 min hold) at 10 °C/min,and then was followed by an increase to 230 °C (3.83 min hold) at 15 °C/min. Solid-Phase Microextraction Parameters. Poly(dimethylsiloxane) (PDMS) fibers were reported as the most effective SPME fiber for rapid field headspace analysis of explosives2 and were used throughout. All fused nonpolar PDMS-coated silica fibers (Supelco, SigmaAldrich, St. Louis, MO, 100 μm nonbonded, 24 gauge) used in this study were conditioned at 250 °C for 0.5 h before use. A Gerstel MPS2 autosampler with automated SPME capabilities was used to monitor the headspace for all experiments. The set-up for static and dynamic ESIS sampling systems is presented in Figure S1 in Supporting Information. In the static experiments, the SPME fiber was sampled first from the closed ESIS vial, followed by sampling from a closed vial containing the analyte, with both vials providing saturated vapors in equilibrium with the liquid(s) or solid phase(s). For the dynamic experiments, the fiber sampled the sealed ESIS vial and was exposed for a predetermined amount of time. The fiber was then brought to the vial containing the analyte and the dynamic headspace was extracted for a second predetermined amount of time. Both types of sampling occurred in tandem and were immediately followed by thermal desorption in the GC injection port for 5 min. Three separate SPME fibers were used to analyze the 2-EH training aids. For all experiments reported here, sampling data were collected every hour. However, more frequent sampling is possible with the SPME-ESIS approach coupled to a more rapid GC separation. Solid-Phase Microextraction Calibration. Since the samples utilized are pure components, it is important to make sure that the amount of analyte absorbed onto the SPME fiber is within the GC/MS detection range. To avoid overloading the mass spectrometer, typical solutions should have a concentration
