Identification of inorganic ions in post-blast explosive residues using portable CE instrumentation and capacitively coupled contactless conductivity detection

Abstract

Novel CE methods have been developed on portable instrumentation adapted to accommodate a capacitively coupled contactless conductivity detector for the separation and sensitive detection of inorganic anions and cations in post-blast explosive residues from homemade inorganic explosive devices. The methods presented combine sensitivity and speed of analysis for the wide range of inorganic ions used in this study. Separate methods were employed for the separation of anions and cations. The anion separation method utilised a low conductivity 70 mM Tris/70 mM CHES aqueous electrolyte (pH 8.6) with a 90 cm capillary coated with hexadimethrine bromide to reverse the EOF. Fifteen anions could be baseline separated in 7 min with detection limits in the range 27–240 μg/L. A selection of ten anions deemed most important in this application could be separated in 45 s on a shorter capillary (30.6 cm) using the same electrolyte. The cation separation method was performed on a 73 cm length of fused-silica capillary using an electrolyte system composed of 10 mM histidine and 50 mM acetic acid, at pH 4.2. The addition of the complexants, 1 mM hydroxyisobutyric acid and 0.7 mM 18-crown-6 ether, enhanced selectivity and allowed the separation of eleven inorganic cations in under 7 min with detection limits in the range 31–240 μg/L. The developed methods were successfully field tested on post-blast residues obtained from the controlled detonation of homemade explosive devices. Results were verified using ion chromatographic analyses of the same samples.

1 Introduction

Inorganic ions are highly prevalent and measurement of their levels is important in a wide variety of sample matrices, including environmental 1–5, foods 6, industrial 7, 8 and forensic samples 9–11. Determination of these species can be performed by wet chemical methods 12, gravimetry 13, potentiometry 14, flow-injection analysis 15, atomic spectroscopy 16, 17, ion chromatography 2 and electrophoretic methods in capillary 5, 18, 19 and microchip 20 formats. Of increasing interest is the use of portable and field-deployable instrumentation for these analyses so that the distance between sample collection and analysis can be minimised. CE is especially amenable to miniaturisation 21–24.

Conductivity detection is the most universal form of detection for inorganic ions and, in the case of CE separations, capacitively coupled contactless conductivity detection (C⁴D) has been used with increasing frequency over recent years (for recent reviews see 18, 25–27. Detection limits are in the range 10⁻⁵–10⁻⁷ mol/L and are generally 1–2 orders of magnitude lower when compared with indirect photometric detection 18. Of the CE separations of inorganic and small organic ions in the literature using C⁴D detection, the work presented by Mayrhofer et al.28 shows the separation of eight anions in 2.8 min using a sorbic acid/arginine BGE. Furthermore, contactless conductivity detection provided LODs for anions and cations as low as 30 μg/L. Kubáò et al.29 were later able to separate eight anions and fourteen cations in less than 3 min in a single run using the technique of dual opposite end injection, with LODs approaching 7 μg/L. Tanyanyiwa et al.30 improved upon the electronics of their contactless conductivity detector by using a higher excitation voltage, which resulted in higher signal strength, better S/N ratio and improved stability. LODs of approximately 3 μg/L were obtained for inorganic ions using a 10 mM MES/His BGE. Further work 31 using this method allowed for the separation of fifteen cations in 5.5 min, although some cations exhibited tailing and were not completely resolved.

Contactless conductivity detection has also been integrated into portable CE instrumentation. Using portable instrumentation in the field has certain advantages over laboratory-based instrumentation, such as potentially reducing sample degradation and contamination and decreasing the time from sample collection to analysis. Early analytical results, obtained at or near a bombing scene (e.g. using instrumentation setup within a mobile laboratory facility), for example, can provide preliminary information on the explosive device to assist the investigation. The first portable CE instrumentation was developed by Kappes and Hauser 32 and the instrument was equipped with potentiometric 32, amperometric 33 and conductometric 34 detectors, weighed 7.5 kg and was housed in a poly(vinyl chloride) case. More recently, Hauser and co-workers 22 designed a new portable instrument of similar dimensions, which was better suited to C⁴D detection, providing LODs as low as 10 μg/L for the separation of ten inorganic cations and five inorganic anions. The only other body of work utilising portable CE instrumentation is that of Li and co-workers. Initially, potential gradient detection was employed for the analysis of DNA fragments 35 and the water-soluble antibiotic neomycin 36. The instrument was then equipped with a C⁴D for the detection of DNA fragments 37 and samples containing organic acids and chlorinated herbicides 21.

Previous work in our laboratory 11 has used a commercial portable CE instrument modified to include a LED indirect photometric detector for the universal detection of inorganic anions and cations in post-blast residues from homemade explosives. The present study extends the previous work by introducing novel methods for the separation of a large range of inorganic anions and cations using a C⁴D. C⁴D detection is shown to be significantly more sensitive than indirect photometric detection, and the new separations are considerably faster and use less hazardous electrolytes than the methods reported previously.

2 Materials and methods

2.1 Apparatus

A portable CE-P2 instrument (8 kg, 50 W and equipped with both a mains power adaptor and a battery capable of providing 3 h of use) from CE Resources (Ayer Rajah, Singapore) was used for all analyses. This instrument was modified in-house to mount the earth electrode and outlet electrolyte vial internally onto the instrument. The instrument was operated either under standard laboratory conditions or inside a purpose-built mobile laboratory housed in a standard small delivery van. The battery life of the CE-P2 was insufficient for performing the required field experiments; hence a portable generator was used as the power source for all equipment used.

A commercial C⁴D (TraceDec) was purchased from Innovative Sensor Technologies (Strasshof, Austria). This detector has previously been compared with several other contactless conductivity detectors in our laboratory and has been shown to provide the lowest LOD for inorganic ions 38. A high-sensitivity detector head was used and this was positioned 8 cm from the outlet end of the capillary.

2.2 Reagents

All reagents used were of analytical reagent grade, purchased from Sigma-Aldrich, (Milwaukee, WI, USA) and dissolved in Milli-Q water unless otherwise stated. Tris and CHES were used to prepare the BGE for the separation of anions. The final composition of this electrolyte was 70 mM Tris and 70 mM CHES prepared in Milli-Q water, buffered at pH 8.6. The BGE for separation of cations consisted of 10 mM L-histidine (His) and 50 mM acetic acid (Biolab, VIC, Australia) at pH 4.2. Hydroxyisobutyric acid (HIBA, 1 mM) was added to aid selectivity and 0.7 mM 18-crown-6 ether was used to separate potassium and ammonium ions. Analytical grade sodium hydroxide and hydrochloric acid were obtained from BDH (Kilsyth, VIC, Australia) and Ajax (Sydney, NSW, Australia), respectively, and 1 M aqueous solutions of these reagents were used for conditioning the capillary prior to use. An aqueous solution of hexadimethrine bromide (HDMB, 1% solution) was used as an EOF reversal agent to provide a co-EOF separation of anions, towards the detector.

Anion standard solutions (1000 mg/L) were prepared individually by volumetric dissolution in Milli-Q water of sodium or potassium salts obtained from Sigma-Aldrich unless otherwise stated. A 15 anion working standard (0.5 mg/L unless otherwise stated) was prepared daily as follows: acetate (1 mg/L, May and Baker, VIC, Australia), benzoate (1 mg/L), carbonate (BDH), chlorate, chloride (BDH), chlorite (1 mg/L, technical grade 80%), cyanate, nitrate, nitrite, fluoride, perchlorate, phosphate (1 mg/L), sulphate, thiocyanate and thiosulphate. 1-Hexanesulphonic acid, sodium salt, was added as an internal standard at a concentration of approximately 5 mg/L for the anion separations to indicate when the run was complete. A ten anion working standard containing only the essential ions for post-blast explosive residue identification was prepared for the fast screening method. This working standard was prepared as stated above, but with cyanate, nitrite, sulphate, thiocyanate and thiosulphate being omitted.

The following 1000 mg/L atomic absorption standard solutions were purchased from Sigma-Aldrich: barium, calcium, lead, manganese, magnesium, sodium, strontium and zinc. Standard cation solutions (1000 mg/L) of ammonium, methylammonium and ethylammonium were prepared by volumetric dissolution of ammonium nitrate (BDH, 98%), methylammonium chloride (Sigma, laboratory grade) and ethylammonium chloride (Fluka, Puriss grade) using Milli-Q water. A mixed working standard of these cations was prepared daily containing, 0.61 mg/L ammonium, 1.23 mg/L (methylammonium, ethylammonium, potassium, sodium, magnesium, calcium), 2.44 mg/L (strontium, manganese, zinc) and 4.89 mg/L barium. Benzylamine (∼5 mg/L) was used as an internal standard and added to all post-blast samples.

2.3 Electrophoretic procedures

2.3.1 Separation of anions

Separations were performed using 75 μm id fused-silica capillary purchased from Polymicro Technologies (Phoenix, AZ, USA). A 90 cm length of capillary was used for the separation of the 15 anion target set with the detector positioned 8.0 cm from the anode. A 30.5 cm capillary was used for the fast screening method. The capillary wall was permanently coated with HDMB for all anion separations, resulting in both anions and the EOF migrating towards the anode. The capillary was preconditioned prior to use by flushing 1 M NaOH at 20 psi for 2 min, water for 1 min, a 1% aqueous solution of HDMB for 5 min, then BGE for 2 min. Prior to each run, the capillary was conditioned by flushing the BGE at 20 psi for 0.8 min. Injection was performed hydrodynamically at 1 psi for 5 s, representing an injection volume of 33 nL (2.1% of capillary length). For the fast anion separation method, injection was performed electrokinetically for 5 s at −1 kV. Separation was performed by applying −20 or −25 kV across the capillary depending on the method chosen. The TraceDec detector was operated at a gain of 200% and −18 Db (approximately 10% of full output voltage).

2.3.2 Separation of cations

A 73 cm length of 75 μm id fused-silica capillary was used with the detector placed 8.0 cm from the cathode. This capillary was used uncoated, but prior to use it was preconditioned by flushing the BGE at 20 psi through the capillary for 5 min. Prior to each run, the capillary was conditioned at 20 psi for 0.7 min with BGE. Introduction of the sample was performed using hydrodynamic injection at 0.1 psi for 5 s, representing an injection volume of 4 nL (0.3% of capillary length). The voltage applied during the separation was 25 kV. The TraceDec detector was operated at −12 Db (approximately 25% of full output voltage) and a gain of 100%. The signal filter function on the TraceDec was turned on to reduce baseline drift and had the added benefit of reducing baseline noise.

2.3.3 Ion chromatography procedures

Two Dionex ICS-2000 (Sunnyvale, CA, USA) systems were used in conjunction with a single Dionex AS autosampler to enable simultaneous analysis of anions and cations. The methods employed were those outlined in our previous work 39, whereby optimised gradient elution profiles were coupled with conductivity detection to separate target anions and cations in explosive residues.

2.3.4 Sample collection and preparation

Witness plates were used to collect explosive residues from field tests. These plates were composed of galvanised steel sheets of rectangular dimensions, 34 cm×32 cm. For each inorganic homemade explosive, a witness plate was placed on the ground (base-plate) and the explosive device positioned in the centre of the plate. Four further witness plates were arranged laterally around the explosive at distances of 1 and 2 m. The explosive was detonated and the witness plates collected in polyethylene bags, which were subsequently heat-sealed to avoid contamination. Soil samples were also taken from directly under the base-plate both before and after the explosion.

The sampling of ions deposited onto the witness plates was performed as follows. A sterile MWE102 rayon swab (Imbros, Tasmania, Australia) was moistened in Milli-Q water and wiped over a 50 cm² portion of the witness plate and then placed in a Dionex (Sunnyvale, CA, USA) polypropylene 10 mL sample vial. The vial was filled with 8 mL of Milli-Q water, capped, shaken and sonicated for 5 min to aid extraction. The aqueous sample was then filtered through a 0.45 μm nylon syringe filter (Phenomenex, NSW, Australia) into a 2 mL snap-top glass sample-vial (Agilent Technologies, VIC, Australia) in order to remove any particulate material. The same extraction procedure was performed on an unused witness plate and this formed the sample blank. Benzylamine (∼5 mg/L) was used as an internal standard for the cation analyses and for the anion analyses, ∼5 mg/L of the sodium salt of 1-hexanesulphonic acid was added to each sample after filtration as an internal standard to indicate when the separation was complete. Residue samples known to contain metal powders, such as magnesium, were acidified with a 100 mM HCl solution to convert insoluble metals and their oxides into water-soluble species.

Soil samples were extracted as follows. Approximately 150 mg of soil was placed into a polypropylene 10 mL sample vial and 8 mL Milli-Q water added. This vial was then sonicated for 2 min to extract ions and the extract was filtered with a 0.45 μm nylon filter to remove the particulate material.

3 Results and discussion

3.1 Instrumental considerations

Compared with other techniques used for the analysis of inorganic ions, CE instrumentation is particularly amenable to miniaturisation and portability due to the minimal hardware requirements and small sample volumes inherent to the operation of the technique. A drawback of current commercially available portable CE instrumentation is that it does not provide temperature control during analysis. However, internal standards were used in this study to overcome temperature-related variations in the migration times of ions. The C⁴D and associated electronics could be easily miniaturised to a cell with the size of a few centimetres. The portable CE-P2 instrument was purchased with the earth electrode situated externally on a UV detector. The instrument was modified in-house to place the earth electrode and C⁴D housing on the body of the portable instrument, thus increasing its portability when taken for fieldwork. A photograph of the modified instrument can be seen in Fig. 1.

Figure 1. Portable CE-P2 instrument modified in house to mount the earth electrode onto the instrument and to fit the TraceDec C⁴D.

3.2 Choice of analyte ions

The goal of this study is to develop sensitive separation methods for a wide variety of inorganic anions and cations known to be present in the residues of homemade explosives. Moreover, it was necessary that these ions be separated from common ions likely to be present in the sampling environment. Inorganic homemade explosive devices are composed of fuels and inorganic oxidisers, which undergo chemical reaction during the explosion to produce ionic species present in post-blast residues. The 15 target anions (chloride, thiosulphate, nitrite, nitrate, sulphate, perchlorate, thiocyanate, chlorate, cyanate, fluoride, chlorite, phosphate, carbonate, acetate and benzoate) and 11 target cations (ammonium, methylammonium, potassium, sodium, magnesium, calcium, ethylammonium, strontium, manganese, zinc and barium) chosen in this study comprise the inorganic species that are prevalent in inorganic explosive residues, together with potential interfering species. The justification for the inclusion of each analyte has been outlined previously 11. The anion profile of a post-blast residue is particularly indicative as to the type of device used and the dominant peak in the separation is usually the major anion present initially or produced by the explosion. On the other hand, the cation profile is used primarily to identify the counter-ion and hence the specific salt used in the preparation of the explosive. This information is of great importance in the investigation of terrorist attacks.

3.3 Anion separation method

Conductivity detection involves the measurement of a difference in conductance between the migrating analyte zones and the BGE, with a low conductance BGE being the most common choice. The BGE must also have sufficiently high ionic strength to reduce the effects of electromigrational dispersion. Ampholytic buffers offer such characteristics and a combination of MES and His is commonly used. However, MES/His BGEs were found to be unsuitable because weak acid anions were partially protonated at the pH used (6.0), leading to reduced detection signals. A higher pH BGE formed by combining equal portions of Tris (pK_a=8.1 40) and CHES (pK_a=9.55 40) was therefore used and provided sensitive detection of all analytes. The effect of the concentration of the BGE (in the range 10–100 mM of each component) on peak resolution was investigated and it was found that increasing the concentration of the buffer components improved peak efficiency but reduced the detection sensitivity. Baseline resolution of all 15 target anions was obtained (Fig. 2) using 70 mM Tris and 70 mM CHES in the BGE, with LOD⩾27 μg/L (S/N=3) being obtained. LODs were approximately ten-fold lower than those obtained in our previous study using indirect photometric detection 11 and the separation was accomplished in a considerably shorter time. In addition, non-hazardous BGE components were used. Analytical figures of merit calculated for ten consecutive replicates of the separation are listed in Table 1.

Figure 2. Separation of target anion mixture using the portable CE-P2 with TraceDec C⁴D. Conditions: 75 μm id fused-silica capillary, L_total=90.0 cm. BGE was 70 mM Tris/70 mM CHES in aqueous solution at pH 8.6. Separation voltage: −25 kV. Detector 8.0 cm from cathode (L_detector=82.0 cm). Detector operational parameters: frequency: 2×high; voltage: −18 Db; gain: 200%; offset: 038. Hardware settings: ADC: 19.80 Hz; filter: fast; DAC: 18-bit; contrast: 50%. Samples injected using a pressure of 1 psi for 5 s. Sample concentrations: 0.5 mg/L of the following: chloride, thiosulphate, nitrite, nitrate, sulphate, perchlorate, thiocyanate, chlorate, cyanate, fluoride and carbonate, and 1 mg/L of the following: acetate, benzoate, chlorite and phosphate.

Table 1. Analytical figures of merit for anion separation system by CE-C⁴D

Analyte	Retention time (n=10)		Peak area n=10	Theoretical plates (N/m)	Resolution (calculated using the following peak)	LOD (S/N=3, mg/L)	Calibration R^2 (range 0–10 ppm)
	(min)	RSD (%)	(RSD %)
a) a) The LOD calculated for carbonate may be understated due to the adsorption of carbonate from the atmosphere.
1 Chloride	4.55	0.98	5.47	109 000	1.23	0.027	0.9943
2 Thiosulphate	4.61	0.99	3.99	216 000	1.36	0.052	0.9910
3 Nitrite	4.67	0.97	5.77	195 000	1.12	0.041	0.9953
4 Nitrate	4.72	0.94	5.80	213 000	2.02	0.052	0.9980
5 Sulphate	4.82	0.96	3.67	142 000	1.63	0.044	0.9948
6 Perchlorate	4.91	0.92	3.49	231 000	1.59	0.084	0.9983
7 Thiocyanate	4.97	0.92	5.85	331 000	1.04	0.066	0.9993
8 Chlorate	5.01	0.90	5.46	337 000	1.74	0.069	0.9950
9 Cyanate	5.09	0.89	6.78	207 000	6.31	0.053	0.9949
10 Fluoride	5.46	0.84	5.05	117 000	2.97	0.026	0.9965
11 Chlorite	5.66	0.78	3.90	186 000	2.48	0.092	0.9946
12 Phosphate	5.81	0.78	9.33	163 000	4.11	0.081	0.9958
13 Carbonate	6.12	0.81	9.89	105 000	3.52	0.018a)	0.9908
14 Acetate	6.36	0.69	11.46	289 000	13.1	0.11	0.9969
15 Benzoate	7.03	0.58	13.40	373 000		0.24	0.9955

3.4 Cation separation method

MES/His BGEs have also been used widely for the determination of cations in combination with conductivity detection, but these BGEs proved to be unsuccessful in separating all 11 of the target cations. An acetate/His (pH 4.1) BGE was found to provide good peak shapes but some peak overlaps occurred. However, addition of HIBA 25 and 18-crown-6 ether permitted all 11 target cations to be separated. The final BGE composition was 50 mM acetic acid, 10 mM His, 0.7 mM 18-crown-6 ether and 1 mM HIBA at pH 4.1 and the corresponding separation is shown in Fig. 3. Analytical figures of merit calculated for ten consecutive runs are given in Table 2. The cation separation method utilising C⁴D detection developed on the portable instrumentation was significantly faster and again provided 10-fold lower LOD than the methods developed previously 11.

Figure 3. Separation of target cation mixture using the portable CE-P2 with TraceDec C⁴D. Conditions: 75 μm id fused-silica capillary, L_total=73.0 cm. BGE was 10 mM His, 50 mM acetic acid, 1 mM HIBA and 0.7 mM 18-crown-6 ether in aqueous solution, pH 4.2. Voltage: 25 kV. Detector 8.0 cm from cathode (L_detector=65.0 cm). Detector operational parameters: frequency: high; voltage: −12 Db; gain: 100%; offset: 059. Hardware settings: ADC: 19.80 Hz; filter: fast; DAC: 18-bit; analogue filter: on. Samples injected using a pressure of 0.1 psi for 5 s. Sample concentrations: 1.1–4.4 mg/L mixture of 11 cations prepared in water. The following analytes were at a concentration of 1.1 mg/L: ammonium, methylammonium, potassium, sodium, magnesium, calcium; the following were at a concentration of 2.2 mg/L: ethylammonium, strontium, manganese, zinc; barium was at a concentration of 4.4 mg/L.

Table 2. Analytical figures of merit for cation separation system by CE-C⁴D

Analyte	Retention time (n=10)		Peakarea n=10	Theoretical plates (N/m)	Resolution (calculated using the following peak)	LOD (S/N=3, mg/L)	Calibration R^2 (range 0–10 ppm)
	(min)	RSD (%)	(RSD %)
1 Ammonium	4.07	0.77	9.00	57 400	2.15	0.031	0.9916
2 Potassium	4.25	0.82	5.71	83 400	9.48	0.053	0.9907
3 Methylammonium	5.01	0.86	8.05	83 600	4.13	0.058	0.9967
4 Calcium	5.37	0.92	7.84	92 900	3.45	0.048	0.9941
5 Sodium	5.67	0.93	6.50	107 000	1.39	0.052	0.9957
6 Strontium	5.79	0.97	8.16	128 000	1.16	0.12	0.9913
7 Magnesium	5.89	0.97	7.74	112 000	1.14	0.073	0.9996
8 Manganese	6.02	0.97	7.07	106 000	1.10	0.040	0.9925
9 Ethylammonium	6.11	0.95	5.08	153 000	1.84	0.129	0.9933
10 Zinc	6.27	0.92	7.25	108 000	4.02	0.095	0.9982
11 Barium	6.68	1.14	6.29	118 000		0.24	0.9912

3.5 Determination of background ions

Before post-blast residues can be identified, it is important to determine background levels of ions introduced by the sampling procedure. There are two important sources: the ions present in the soil at the test site blown onto the witness plate used for collecting residues and ions introduced during the swabbing procedure itself. Water-soluble ions were extracted from a representative soil sample taken directly underneath where the explosive device was to be detonated. A sample of soil (160 mg – determined from previous tests to be the approximate amount of soil deposited on a 50 cm² section of the witness plate) was extracted and the anion and cation profiles are shown in Fig. 4a and b, respectively. In an actual bombing incident, soil samples would be collected away from the blast site to determine the normal background levels of anions and cations of interest. Only trace amounts of inorganic anions and cations were extracted from the small amount of soil that is deposited onto the witness plate during detonation, with the main anions being chloride, nitrate and a small amount of sulphate. Carbonate was present in all electropherograms due to absorption of carbon dioxide from the atmosphere into the aqueous samples. The cations found in the aqueous soil extract were ammonium, potassium, calcium, sodium and magnesium. Background ions introduced by the sample collection procedure were assessed by swabbing a new, unused witness plate using the procedure outlined for the collection of real samples. Figure 4 indicates that the main ions extracted from the cotton swabs were sulphate and sodium, but the background levels of these ions could be reduced to below the detection limit if the swabs were pre-rinsed with Milli-Q water prior to use.

Figure 4. Analysis of background (A) anions and (B) cations extracted from a cotton swab, a pre-blast soil sample and residues from the detonation of two different organic explosive devices. Conditions: Aqueous samples were injected without dilution. CE methods as for Figs. 2 and 3 for anions and cations, respectively. Approximately 5 mg/L of hexanesulphonate and benzylamine were added as internal standards for the anion and cation analyses, respectively, and provided an indication to when the run had finished.

Organic explosive devices (which do not produce residues containing inorganic ions) were also detonated in the same location as a further means of investigating the levels of ions contributed from soil deposited onto the witness plate. Two separate organic explosives (250 g of PE4 (1,3,5-trinitro-1,3,5-triazacyclohexane (RDX))/pentaerythritoltetranit rate (PETN) and 600 g of 2,4,6-trinitrotoluene (TNT)/(RDX)) were detonated and the witness plates swabbed. The ions present in the samples were considered to result only from the soil and the swabbing procedure and the results shown in Fig. 4 confirm that levels of ions detected were very similar to those observed for the soil extract.

3.6 Analysis of post-blast residues from inorganic homemade explosives

The developed method and instrumentation were evaluated in the field by analysing residues from controlled detonations of homemade explosive devices, as listed in Table 3. Figure 5 shows electropherograms for the anionic and cationic components of the residues and the components identified for each device were in close agreement with those listed as major marker ions in Table 3. Of note was the fact that benzoate was not observed from device 2 and the presence of zinc in the acidified cationic sample from device 3. Acidification of the sample was required to convert insoluble metals and their compounds such as magnesium and its oxide into water-soluble ions for detection. Corrosion products of zinc are present on the surface of the galvanised witness plates, which were subsequently collected by the swabbing procedure and converted to zinc ions upon acidification. In all cases, the residues also showed the presence of ions not listed in Table 3, but these ions could be attributed to three sources, namely ions produced in the chemical reaction (especially chloride produced by reduction of perchlorate and chlorate), contaminants (from the swabs, soil and the galvanised coating on the witness plates), or to carbonate formed from absorption of atmospheric carbon dioxide. The presence of these ions did not prevent positive identification of each type of explosive using the “fingerprint” of indicator anions and cations.

Figure 5. Analysis of (A) anions and (B) cations extracted from post-blast residues resulting from the detonation of four inorganic homemade explosive devices. Conditions: Samples were injected without dilution. CE methods as for Figs. 2 and 3 for anions and cations, respectively. Approximately 5 mg/L of hexanesulphonate and benzylamine were added as internal standards for the anion and cation analyses, respectively, and provided an indication to when the run had finished. The cationic residue from device 3 was split into two portions and only the portion used for the identification of cations was acidified with 100 mM HCl to dissolve magnesium not soluble in water alone. Zn²⁺ was a by-product of acidifying the sample resulting from swabbing the galvanised witness plates.

Table 3. Summary of the inorganic homemade explosive devices detonated during this study and the associated marker ions expected in the resulting post-blast residues

	Type of inorganic homemade device	Composition	Marker ions expected in residue
			Anions	Cations
Device #1	Ammonium nitrate:fuel oil	300 g ammonium nitrate: diesel (ratio 16:1)	NO3^−	NH4^+
Device #2	Black powder	165 g potassium nitrate	NO3^−	K^+
		60 g barium nitrate	SO4^2−	Ba^2+
		30 g sulphur	C6H5COO^−	Na^+
		45 g charcoal
		25 g sodium benzoate
Device #3	Chlorate/sulphur/magnesium	150 g sodium chlorate	ClO3^−	Na^+
		50 g sulphur	SO4^2−	Mg^2+
		25 g magnesium powder
Device #4	Chlorate/perchlorate/nitrate/sulphur/charcoal	40 g barium nitrate	NO3^−	Na^+
		80 g calcium nitrate	ClO3^−	K^+
		30 g sodium chlorate	ClO4^−	Ca^2+
		40 g potassium perchlorate	SO4^2−	Ba^2+
		25 g sulphur
		35 g charcoal

A previous study in this laboratory was performed using ion chromatographic (IC) methods to determine inorganic ions in post-blast residues 39. The IC methods were applied to the same residue samples to provide complementary identification of the marker ions. Excellent agreement was obtained between the two techniques, with the LOD when using IC being approximately an order of magnitude lower than when using CE-C⁴D. The differences in detection limits can be explained by the larger volumes injected in IC and the nature of direct conductivity detection used in conjunction with electrolytic suppression of the eluent. However, the poorer sensitivity of CE-C⁴D detection was not a limitation because marker ions were present in the mg/L range, as shown in Fig. 5. If greater sensitivity was required, less water can be used in the swabbing/extraction process to reduce dilution factors and a larger portion of the plate can be swabbed.

3.7 Rapid separation of inorganic marker anions

The anion separation shown in Fig. 2 permits the identification of major types of homemade inorganic explosives using a single set of separation conditions. However, in some cases there is a requirement to perform faster, but less comprehensive, separations for the rapid preliminary screening of samples. Therefore, a fast separation method was developed to separate the ten marker and background anions (chloride, nitrate, perchlorate, chlorate, fluoride, chlorite, phosphate, carbonate, acetate and benzoate) deemed to be of greatest significance in the identification of inorganic post-blast residues. Figure 6 shows the separation of these species in 45 s performed using a 30.6 cm capillary (compared with the 90 cm capillary required for the separation of all 15 analytes). LODs (S/N=3) ranged between 12 and 160 μg/L. The disadvantage of the faster screening method was the co-migration of the following groups of analytes: chloride/thiosulphate, nitrite/nitrate, sulphate/perchlorate and thiocyanate/chlorate/cyanate. Of these, the most significant is sulphate/perchlorate because of the widespread occurrence of sulphate and the importance of perchlorate as a marker anion. The proposed rapid separation could therefore be used as a preliminary screen to suggest the presence of marker ions, but to eliminate false-positive results the comprehensive separation would need to be employed. Figure 7 provides an example of two real post-blast residues analysed using the fast separation method. One residue resulted from an ammonium nitrate:fuel oil device and the other device was known to contain nitrate, chlorate and perchlorate. Peaks were observed suggesting the presence of the relevant marker ions for both explosives; yet it is possible that this is false positive due to contaminant peaks co-migrating in the same position as the marker ions. Such a method is advantageous as a quick screening method; however, it also highlights the importance of developing a method capable of baseline resolving all analytes that may arise in real samples.

Figure 6. Fast CE separation of the ten most significant anions. Conditions: 75 μm id fused-silica capillary, L_total=30.6 cm. Detection using TraceDec contactless conductivity detector, 24.6 cm from cathode (L_detector=24.6 cm). Separation voltage: −20 kV. Other conditions were the same as those in Fig. 2. Samples injected electrokinetically of −1 kV for 5 s. Sample concentrations: 0.5 mg/L of the following anions: chloride, nitrate, perchlorate, chlorate, fluoride and carbonate, and 1 mg/L of the following anions: acetate, benzoate, chlorite and phosphate.

Figure 7. Fast CE analysis of a post-blast residue from two inorganic improvised homemade explosive devices. Conditions used were the same as those in Fig. 6. The samples were spiked with approximately 5 mg/L of the internal standard, hexanesulphonate, to indicate the end of the separation. (a) Thiosulphate potentially co-migrates with chloride. (b) Nitrite potentially co-migrates with nitrate. (c) Sulphate potentially co-migrates with perchlorate. (d) Thiocyanate and cyanate potentially co-migrate with chlorate.

4 Conclusions

C⁴D detection combined with portable CE separation methods have been developed for the sensitive analysis of inorganic anions and cations in residues from inorganic homemade explosives. C⁴D detection provided 10-fold lower detection limits and separations were completed more quickly compared with separations reported previously using the portable CE instrument with indirect photometric detection based on highly absorbing probes and LED light sources 11. In comparison with other methods on portable CE instrumentation, the detection limits achieved for inorganic ions in this study were quite similar to those achieved by Kubáò et al. 22 performing CE-C⁴D detection and those obtained by Kappes et al. 33, 34 using amperometric detection. Moreover, these detection limits are 1–3 orders of magnitude lower than those achieved using potentiometric or potential gradient detection 32, 33, 35. The comprehensive separation methods developed could be used to positively identify the major types of homemade explosives using a single set of separation conditions for anions or cations. In addition, a rapid separation of ten anions in 45 s was developed and is proposed for preliminary screening of samples to indicate the possible presence of marker anions. Although these methods have been developed specifically for inorganic post-blast residues, the large number of ions separated means that the methods will also have applicability for the sensitive, field-based determination of inorganic anions and cations in a wide range of other sample types.

Acknowledgements

This project was supported by a grant from the National Security Science & Technology Unit, Department of Prime Minister and Cabinet, undertaken with the support of the Australian Federal Police and the National Institute for Forensic Science. The project also forms part of an Australian Research Council Federation Fellowship awarded to PRH. The authors acknowledge the assistance of the Australian Bomb Data Centre for the preparation and detonation of the homemade explosive devices used to generate the post-blast residues analysed in this study.

The authors have declared no financial or commercial conflict of interest.