A Comparison of Real Versus Simulated Contraband VOCs for Reliable Detector Dog Training Utilizing SPME-GC-MS

Odor detection has become a focused area of research because of its importance to the forensic, law enforcement, and legal communities. Despite the abundance of methods for the detection of these characteristic chemical odors, the use of trained canines as biological detectors remains widely accepted. Thus, detector dog response is one of the major applications involved in odor detection studies, both to determine the chemical signature of individual odors to which the canines are actually alerting, and to discover if there is a common element within different items to support the use of contraband mimics.

Previous research has demonstrated that a compilation of chemical odors can be detected in individual contraband samples, including several common odors as well as uncommon odors.1,2 Current commercially available pseudo aids contain different amounts of either the actual explosive/narcotic or the chemical compound of suspected interest by canine detectors. The main chemical compound in a substance is not always the dominant volatile compound due to the low vapor pressure or limited olfactory receptor response.3 In addition, it has been shown that only specific odors are used by canines to detect the various forms of contraband.

Narcotics detection

Detector dog teams are trained to detect most commonly found illicit narcotic substances (i.e., marijuana, heroin, cocaine, and methamphetamines). It has been shown that canines respond to volatile organic compounds (VOCs) in the headspace above the drug instead of the parent compound itself. One case in which this has been proven is for cocaine. Field tests have shown that the trained law enforcement detector dogs respond to the compound methyl benzoate.3–5 This implies that methyl benzoate must be the dominant odor chemical signature for cocaine.

The headspace above marijuana has been repeatedly sampled and shown to possess a complex array of organic compounds. This list includes α-pinene; β-pinene; myrcene; limonene; and, in many cases, β-caryophyllene. These compounds have been shown to dominate the headspace of marijuana samples (upwards of 85%6–8). In a similar manner, it has been conjectured that acetic acid is the dominant odor compound in heroin samples.

Explosives detection

Classification by chemical groups is a common method, especially for research purposes. This is done by classifying explosives into one of the following five groups: organic nitrates (which include aliphatic nitro and aromatic nitro), nitrate esters, nitramines, acid salts, and peroxides.1,2

As with drugs, canines respond to VOCs in the headspace above explosives instead of the parent compound itself. An example is single-based smokeless powder, in which the main compound is nitrocellulose, an involatile compound. However, certain common odors, including 2,4-dinitrotoluene (2,4-DNT), 2,6-dinitrotoluene (2,6-DNT), and diphenylamine (DPA—a stabilizer), and uncommon odors, such as ethyl centralite (a stabilizer), have been found to be present in these smokeless powders.1,2,9,10 The compound identified as the canine active odor for single-based smokeless powders is 2,4-DNT.


All solvents were purchased from Fisher Scientific (Pittsburgh, PA). The compounds used to create the field samples were purchased from Sigma-Aldrich (St. Louis, MO). A 70-μm StableFlex Carbowax®/divinylbenzene (CW/DVB) solid-phase microextraction (SPME) fiber was purchased from Supelco (Bellefonte, PA). The samples were presented using Sigma Pseudo scent cages. The commercial narcotic pseudo scents were Sigma Pseudo Narcotic scents. Nonhazardous explosives for security training and testing (NESTT) samples utilizing the silica form were purchased from Ray Allen Manufacturing Co. (Colorado Springs, CO). Headspace vials (40 mL) fitted with phenolic plastic caps and a PTFE/silicon septum were purchased from Supelco. Sterile 5.1 × 5.1 cm gauze sponges were obtained from IMCO (Independent Medical Co-op, Inc., Daytona Beach, FL).

The GC-MS analysis was performed on an Agilent 6890 series with HP 5973 quadrupole mass spectrometer (Agilent Technologies, Wilmington, DE). An HP5 30 m × 0.25 mm capillary column (Agilent) with 25 μm film thickness and a helium flow rate of 1 mL/min was used. The injection port was held at 235 °C with 5-min SPME desorption. The MS temperature was 230 °C. The oven program began with a 40 °C hold for 5 min, followed by a 10 °C/min ramp to 290 °C, and ending with a 1-min hold at 290 °C. A 1.2-min solvent delay was used. Two standard solutions were made for each compound at 100 ppm and 50 ppm with dichloromethane.

For the Sigma Pseudo scents, 5 g was weighed and heat-sealed within 3-mm low-density polyethylene (LDPE) bags. Two spiked gauzes were prepared (mixture A and mixture B) comprised of 50 μL each of five compounds: α-pinene, β-pinene, myrcene, limonene, and β-caryophyllene. Both types of samples were placed in the Sigma Pseudo scent cages on site for testing. For the NESTT canine field trials odor delivery system, 5 g of each sample contained within an open vial was placed inside a sterile quart can.

For the dissipation study, 50 μL of a compound (each in turn) was spiked onto a single piece of gauze. The mass of the gauze was recorded prior to spiking and immediately afterward. The mass was then recorded over a period of 30 min at 0.5-min, 1-min, and 5-min intervals. The process was repeated in triplicate. Statistical analysis was performed, and the results were plotted.


Table 1 - Field testing of heroin mimics

Field experiments were conducted with certified law enforcement drug detector dogs. The scent cages were arranged in a line such that each sample was approximately 1 m apart. Each detector dog team was allowed to walk the line and sample the odor emanating from each cage. The handlers were not informed of the contents of each sample prior to the run. In addition to the two samples, blank gauze sealed in a 3-mm LDPE bag was included as a negative control. As shown in Table 1, no canine alerted to the 5-g sample of the Sigma Pseudo Heroin scent. Canine 112 showed interest in the sample. Only one canine alerted (8%) to the acetic acid sample.

Table 2 - Field testing of marijuana mimics

A similar setup was used for the marijuana pseudo sampling. Each detector dog team was allowed to walk the line and sample the odor emanating from each cage. In addition to the target odors, empty scent cages, blank gauze sealed in a 3-mm LDPE bag, and a dichloromethane sample were included as negative controls. The results of the marijuana pseudo testing are shown in Table 2. None of the detection teams alerted to the Sigma Pseudo Marijuana scent. The mixtures were composed of previously reported headspace components of marijuana samples. None of the canines alerted to either of the mixed spikes. In addition, no false positives can be reported.