Article pubs.acs.org/ac
Dynamic Vapor Generator That Simulates Transient Odor Emissions of Victims Entrapped in the Voids of Collapsed Buildings M. Statheropoulos,*,† G. C. Pallis,† K. Mikedi,† S. Giannoukos,†,‡ A. Agapiou,† A. Pappa,† A. Cole,§ W. Vautz,¶ and C. L. Paul Thomas∥ †
School of Chemical Engineering, National Technical University of Athens, Field Analytical Chemistry and Technology Unit, 9 Iroon Polytechniou Street, Athens, 157 73, Greece ‡ Department of Electrical Engineering and Electronics, University of Liverpool, Brownlow Hill, Liverpool, L69 3GJ, United Kingdom § Markes International Ltd, Gwaun Elai Medi Science Campus, Llantrisant, Rhondda Cynon Taf CF72 8XL, United Kingdom ¶ Leibniz-Institut für Analytische Wissenschaften − ISAS − e.V., Bunsen-Kirchhoff-Str. 11, Dortmund, North Rhine-Westphalia 44139, Germany ∥ Department of Chemistry, Centre for Analytical Science, Loughborough University, Loughborough, Leicestershire LE11 3TU, United Kingdom ABSTRACT: The design, development, and validation of a dynamic vapor generator are presented. The generator simulates human scent (odor) emissions from trapped victims in the voids of collapsed buildings. The validation of the device was carried out using a reference detector: a quadrupole mass spectrometer equipped with a pulsed sampling (PS-MS) system. A series of experiments were conducted for evaluating the simulator’s performance, defining types and weights of different factors, and proposing further optimization of the device. The developed device enabled the production of stable and transient odor profiles in a controllable and reproducible way (relative standard deviation, RSD < 11%) at ppbv to low ppmv concentrations and allowed emission durations up to 30 min. Moreover, the factors affecting its optimum performance (i.e., evaporation chamber temperature, air flow rate through the mixing chamber, air flow rate through the evaporation chamber, and type of compound) were evaluated through an analysis of variance (ANOVA) tool revealing the next steps toward optimizing the generator. The developed simulator, potentially, can also serve the need for calibrating and evaluating the performance of analytical devices (e.g., gas chromatographers, ion mobility spectrometers, mass spectrometers, sensors, e-noses) in the field. Furthermore, it can contribute in better training of urban search and rescue (USaR) canines.
A
dynamic, periodic, and even variable process, due to a number of independent or nonindependent factors. The factors that mostly contribute to a dynamic concentration profile are the wind velocity which affects the motion of the plume and influences the concentration through dilution effects, the rubbles structure (plume accumulation or dispersion), and the interactions of the plume with the construction/building materials (e.g., absorption, chemical reaction).4 In the same context, the chemical environment in the debris is affected by a number of random, nonreproducible, and uncontrollable factors, such as the dynamic emission of VOCs and gases from the various human biological sources (e.g., expired air, urine, blood, sweat, etc.), the temperature, the humidity, the dust, and the interactions between the emitted compounds and with the background environment.5 Therefore, the evolved
s part of a research project that investigates chemical human signatures and uses combined chemical, audio, and video sensors for detecting entrapped victims under the ruins of collapsed buildings (EC project “SGL for USaR”, www.sgl-eu. org), a vapor generator was developed for simulating transient odor emissions. Volatile organic compounds (VOCs) released from tissues (skin, lungs) and/or biological fluids (sweat, urine, blood), contributing to human scent, were targeted as potential markers of human presence.1,2 During victims’ entrapment, the concentrations of endogenous volatiles inside the voids vary over time. The released chemical plume will be entrained by adducted air currents and will equilibrate with the surfaces it passes across. The general concentration profile may be envisaged to follow a breakthrough profile with concentration transients superimposed over the underlying trend. The transients are caused by wind vector fluctuations and metabolic events; urination, for example, is such an event.3 The release of a plume of VOCs from the ruins of a collapsed building is not a continuous and stable process but a rather © 2014 American Chemical Society
Received: December 23, 2013 Accepted: March 24, 2014 Published: March 24, 2014 3887
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
Table 1. Commercial Atmosphere Gas Generator Systems GEN-SYS by OWLSTONE (www.owlstonenanotech. com) technology
permeation tube
portability number and type of analytes number of independent permeation ovens total flow (mL/min) output concentration humidity range
AUTOBLEND by KINTEK (www.eco-scientific. co.uk)
491M modular by KINTEK (www.eco-scientific. co.uk)
HOVACAL by IAS (www. hovacal.de)
permeation tube
yes 1
permeation tube and dynamic headspace saturation no 48 and 2
yes 2−5
evaporation of a volumetrically controlled liquid flow no 3 (at least) + humidity
1
6
2−5
0
up to 500 ppbv−ppmv 40−90% or 1−90%
250 to 5000 ppbv to 1000 ppmv
ppbv to 1000 ppmv optional
50−5000 pptv−ppmv up to 100%
LENGEN by US ARMY (www.ecbc. army.mil) sorbent tubes yes 2, low volatility chemicals 2
pptv−ppbv
Figure 1. General setup of the fully fledged vapor generator.
mixtures is indispensable. Some of these commercial generators are examined and presented in Table 1. Indicatively, Hovacal requires around 0.1−1 h for a concentration change. It is designed in a modular concept with a basic module including one evaporator, one dilution stage, and flow control, as well as software and hardware for control and data acquisition (∼20 k€). For each particular analyte, a syringe system module (∼10 k€) provides a continuous analyte flow towards the evaporator for the generation of concentrations in the upper ppbV range. Due to the modular design, a further dilution stage (∼10 k€) is available for shifting the concentration range down to pptV, and additional syringe systems enable the generation of mixtures of different analytes. Furthermore, by operating a second evaporator (∼10 k€) coupled with an additional syringe system, humid calibration gases in the range of 0−100% relative humidity can be provided. The majority of commercial vapor generators is based on permeation tubes; however, this principle of operation requires long-term operation of each source to produce stable concentrations. According to our knowledge, there is not any commercial field transportable system for generating reliable concentration transients at the time-scale of seconds or even minutes. The aim of the present work is to describe the design, development, and validation of a dynamic vapor generator for producing transient odors of interest evolved in the ruins of collapsed buildings after natural or man-made disasters. The dynamic module can be part of a more complicated system that includes a stable concentration module and a humidifier. These kind of devices are mainly developed for calibrating and evaluating the performance of chemical instruments used in the field: gas sensors, portable gas chromatographers, mass spectrometers (MS), and ion mobility spectrometers (IMS) for safety, security, and situation awareness applications. Although, there has been some research in the development of generating gaseous standard mixtures,6 as well as in emphasizing the importance of humidity in vapor generators,14
concentrations will vary over time with randomized concentration transients. Finally, concentrations of VOCs in the debris of collapsed buildings are usually in the range of pptv to ppbv and potentially can increase to the ppmv level after hours or even days of entrapment. Thus, there is a great need to reproduce and reliably test and validate the chemical sensors/ devices that are used in such complicated and harsh environments. This can be succeeded through the use of VOCs generating systems (chemical environment simulators). In general, the generation of test gases is performed mainly through static and dynamic methods; however, other methods are also possible (e.g., exponential dilution).6 Although static methods are simple and inexpensive, they present adsorption and condensation problems.7 Moreover, they are affected by leakages and pressure changes, and an amount of test gas can be generated only once. On the other hand, dynamic methods overcome the above drawbacks, but they are more complex and expensive.8 There are different categories of dynamic methods including injection, permeation, diffusion, and evaporation methods;9 the accuracy and characteristics of each method is described and discussed elsewhere.10 Examples of in-house dynamic test gas generation devices based on the evaporation principle include the impinger bubbler system,10 the microwave-assisted method,11 and the thermal decomposition of suitable surface compounds.12 A variety of commercial gas atmosphere generators is also available in the market; Gen-Sys by Owlstone, Autoblend and 491 M modular by Kin-Tek, and Hovacal13 by IAS. Another interesting gas generator is LenGen, developed by Edgewood Chemical Biological Center (US Army). These systems can be mainly categorized on the basis of their technology (principle of operation), portability, number or type of analytes, number of independent permeation ovens, total flow, and output concentration. As humidity influences the reliability of many analytical devices and biological detectors as well, for the mentioned application, the generation of humid calibration 3888
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
Figure 2. The main parts of the developed dynamic simulator prototype: (1) evaporation chamber, (2) mixing chamber, (3) adsorptive tissue, (4a) mass flow controller, (4b) steering for mass flow controller, (5a) heating bin, and (5b) temperature controller for heating the bin.
this is to the best of our knowledge, the first study focusing on the development of a vapor generator that enables the simultaneous production of both stable and transient odors of interest.
■
monitoring by sampling at atmospheric pressure: the pulsed sampling-mass spectrometer (PS-MS). Pulsed Sampling-Mass Spectrometer (PS-MS). A HewlettPackard quadrupole MS detector equipped with an in-house developed sampling probe, the PS system, was used for monitoring the evolved volatiles. The PS minimizes the impact of oxygen or contamination on the MS ion source and analyzer, when sampling from oxidative or highly contaminated environments. Therefore, it is capable to perform near real-time monitoring of rapid dynamic changes in various concentrations, enabling the fast sampling and transfer of a wide range of VOCs (boiling point < 214 °C) to the MS. Further details about the PS-MS can be found in the literature.15−18 The capillary column, used for transferring the gaseous products into the MS, was a 2.15 m long uncoated fused silica of 0.15 mm id (Varian, USA). The transfer line and the pulsed sampling switch of PSMS were heated at 210 and 200 °C, respectively. The PS-MS was operated in continuous sampling mode. Method Development. PS-MS linearity range and detection limit for online monitoring of VOCs was determined using benzene gas standards in synthetic air. Gas standards of 0.5 and 1 ppmv, as well as, a blank sample were prepared by evaporating appropriate quantities of benzene methanolic solutions in 2.8 L clean, synthetic air purged glass bottles according to the procedure described by McClennen et al.19 The same procedure was followed to prepare blank synthetic air bottles. The dilution bottles were prepared on selected days, prior running the experiments. The single ion monitoring (SIM) mode was applied for monitoring the following chemical components: acetone (m/z 58), 2-butanone (m/z 72), benzene (m/z 78), n-hexane (m/z 86), toluene (m/z 92), dimethyl disulfide (DMDS, m/z 94), xylene (m/z 106), and dimethyl trisulfide (DMTS, m/z 126). The ions 40 and 78, attributed to Ar and benzene, respectively, were also monitored in SIM mode with dwell time of 200 ms. Mass 40 for Ar was monitored as an indicator of stable sampling conditions. HP Chemstation was used for data acquisition. The benzene gas standards were analyzed sampling directly from the bottles through a Teflon tube, in the following sequence: (a) 1st blank synthetic air sample, (b) 2nd blank
EXPERIMENTAL SECTION
Design. The main parts of the fully fledged odor simulator (stable and dynamic modules) are shown in Figure 1. A stable concentration module generates mixtures of stable ultra low level concentrations, and a dynamic concentration module generates dynamic concentration profiles for simulating transient phenomena. The latter allows one to add specific vapors to the mixture produced by the stable concentration module. A humidifier introduces certain levels of humidity in the produced mixtures, and a packed construction material tube (PCMT) allows the mixture to interact with any type of material (e.g., construction materials). A reference detector monitors the performance of the simulator while zero air provides extra pure air. In the present work, only the development and validation of the dynamic module is presented. Development. The dynamic module of the vapor generator consists of the following components: an evaporation chamber, a mixing chamber, an adsorptive tissue, a mass flow controller (Model PR 4000-S2 V1N, SN:13605G, MKS Instruments Deutschland, GmbH), a heating bin, a temperature controller for heating the bin, and a flow meter (Dwyer Series RM RateMaster Flow meters, CAT. NO. RMA-21-SSV, Dwyer Instruments, INC., USA). The main parts of the simulator in the first laboratory setup are presented in Figure 2. The primary pressure on the mass flow controller has to be in the range of 2−3 bar. The tissues for absorption of the analyte could be of any absorptive material, in the best case without odorants or other additives. The diameter of the tissue is 12−13 mm. Validation. The validation procedure of the dynamic vapor generator included: (a) the validation of the reference detector, (b) the method development, (c) the experimental design, (d) the synthesis of odors of interest, and (e) the statistical analysis. Reference Detector. The vapor generator was validated using a sensitive reference detector able to perform online 3889
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
Experimental Design. Five sets of experiments were carried out in total; details of the experimental cycles are given in Table 2. Two of them, were used for evaluating the simulator performance, defining types and weights of variances of different factors, and proposing optimization of the simulator. The other three sets were carried out to investigate its operation using complex mixtures (including compounds of different volatility and chemical class), the interaction with building materials (through PCMT), and the influence of humidity. Three replicates were performed per experimental cycle. When data acquisition ended, the evaporation chamber was purged with zero air at a flow rate of 10 mL/min for 2−3 min. The PCMT was constructed as follows: the cement samples were grinned up to a diameter of 1−1.68 mm using special sieves from Test-Sieve, type ASTM E11-61, 1000 and 1680 μm. An empty glass sampling sorbent tube (6 mm i.d., 115 mm length, Supelco, USA) was filled with 0.9 g of construction material with a bulk density of 1.1 g/cm3. Glass wool was used at both sides of the tube. The PCMT material was left to dry for 30 min at 120 °C before use. In the first two sets, the production of two different evolution profiles was examined, a “stepwise” increase profile and an “increase−decrease” profile based on the following variables: evaporation chamber temperature (X1), air flow rates in mixing (X2) and evaporation (X3) chambers, and type of chemical compound (X4). Stepwise increase profile corresponds to pulses of odors in which the levels of concentration increase at prefixed steps. The level of concentrations (or level of abundances) is controlled by the flow through the evaporation and the mixing chamber. With a constant flow rate through the mixing chamber, the flow rate through the evaporation chamber determines the final output concentration or the output relative abundance. Synthesis of Odors of Interest. In the environment of building collapse, odors of interest refer to VOCs released from a variety of sources. For the needs of the present work, the odors of urine3,20−22 and human decay (early stages)23−26 were synthesized. These particular odors are considered complex and rich; therefore, a simplified version of them was developed on the basis of the most abundant VOCs, and their produced average responses were monitored using a TD-GC-TOF-MS system; thermal desorption unit (Series 2 Unity, Markes International) combined with a gas chromatographer (2010
synthetic air sample, replicate, (c) 1st 0.5 ppmv benzene standard sample, (d) 2nd 0.5 ppmv benzene standard sample, replicate, (e) 1st 1 ppmv benzene standard sample, and (f) 2nd 1 ppmv benzene standard sample, replicate, as shown in Figure 3.
Figure 3. PS-MS response for ion m/z 78 sampled directly from synthetic air bottles (served as blank) and from benzene gas standards at 0.5 and 1 ppmv: (a) 1st blank synthetic air sample, (b) 2nd blank synthetic air sample, replicate, (c) 1st 0.5 ppmv benzene standard sample, (d) 2nd 0.5 ppmv benzene standard sample, replicate, (e) 1st 1 ppmv benzene standard sample, and (f) 2nd 1 ppmv benzene standard sample, replicate. The sparks and/or the decrease of plateaus are attributed to the changing of the dilution bottles.
The MS response of ion mass 78 (benzene) in the validation experiment is presented in Figure 3. The abundance recorded at each concentration level in terms of height was determined. An average abundance value was estimated for the blank and the various standards. A linear calibration curve was estimated for benzene, described by y = 357.25x + 535.36 (R2 = 0.9984), where y is the average abundance and x is the concentration of the gas standards. The MS detection limit (cL) for online gas monitoring was estimated at 210 ppbv for benzene using the calibration curve and eq 1:18 3SDBLANK (1) a where SDBLANK is the standard deviation of the MS signal for the blank and a is the slope of the calibration curve. cL =
Table 2. Overall Experimental Design for Testing the Performance of the Dynamic Odor Simulator experimental design “stepwise increase” profile evaporation chamber temperature (o C) air flow rate through the mixing chamber (L/min) evaporation chamber flow rate (mL/min) volume of analyte in the evaporation chamber (μL) compound(s)
“increase− decrease” profile
response to synthetic mixtures
interaction with construction material
water phase impact
40, 60
40, 60
60
40
40
4, 6
4, 6
4
4
4
0, 1, 2, 3, 4, 5
3, 0, 5, 0, 3, 0, 5, 0, 3, 0, 5, 0 40, 60
0, 1, 2, 3, 4, 5
0, 1, 2, 3, 4, 5
0, 1, 2, 3, 4, 5
40
40
40
acetone, 2-butanone, toluene, xylene, DMDS, DMTS
acetone, 2-butanone, benzene, toluene, xylene,
acetone, 2-butanone, benzene, toluene, xylene, ammonia
40 acetone, hexane, benzene, 2-butanone, DMDS
benzene
3890
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
Figure 4. (a) Stepwise increase profile of mass 78 for 40 μL of benzene at 40 °C. Each step relates to flow of air through the evaporation chamber at 1, 2, 3, 4, and 5 mL/min, respectively. The air flow through the mixing chamber was set at 4 L/min; (b) increase−decrease profile of mass 78 for 40 μL of benzene at 40 °C. The pattern relates to flow of air through the evaporation chamber at 3, 0, 5, 0, 3, 0, 5, 0, 3, 0, 5, and 0 mL/min, respectively. 3891
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
Figure 4. continued The air flow through the mixing chamber was set at 4 L/min; (c) characteristic masses m/z 58, 94, 72, 106, 92, and 126 attributed to acetone, DMDS, 2-butanone, xylene, toluene, and DMTS, respectively, in the isomolecular early decay body synthetic odor mixture; (d) mass profiles of m/z 72, 92, 78, and 106 attributed to 2-butanone, toluene, benzene, and xylene, respectively, in the urine synthetic odor mixture; (e) mass profiles of m/z 72, 92, 78, and 106 attributed to 2-butanone, toluene, benzene, and xylene, respectively, in the urine synthetic odor mixture after interaction with construction materials; (f) mass profiles of m/z 72, 92, 78, and 106 attributed to 2-butanone, toluene, benzene, and xylene, respectively, in the urine synthetic odor mixture after interaction with water.
Shimadzu) and a time of flight-MS (Bench TOF-dx, ALMSCO International). Chemicals. Acetone (99.8%) and DMDS (99%) were obtained from Sigma-Aldrich (Germany). Benzene was provided from BDH Chemicals Ltd. (England), whereas nhexane (p.a.) was purchased from Riedel-de Haen (Germany). 2-Butanone (99.7%) was provided from Chromasolv, DMTS (>98%) from SAFC Supply Solutions, toluene (p.a.) from FERAK (Berlin, Gemany), and xylene from Mallinckrodt (USA). Finally, ammonia was provided from Carlo Erba (Italy), and the gas cylinders with air-zero air were supplied from Air Liquid (Greece). Statistical Analysis. The mathematical software MATLAB 7.5.0 (R2007b) was applied for the analysis of variance (ANOVA). The aim of the factorial design was to reveal the significant parameters that heavily impact the performance of the simulator. The abundance values were preprocessed before applying ANOVA. First, the background (average abundance of the respective compound for air flow rate through the evaporation chamber equal to 0 mL/min) was subtracted for omitting variance due to background noise. The ratio of this difference to the average abundance of m/z 40 of the particular experiment was determined. This ratio was used as input for further calculations (normalized abundance). Calculations for determining the mean relative abundance for all combinations in the different experimental sets were performed. The normalized abundance of characteristic masses (m/z) related to each individual compound was used as dependent variable. The relative standard deviation (RSD) of these normalized abundances over time was also used as a dependent variable.
72), toluene (m/z 92), DMDS (m/z 94), xylene (m/z 106), and DMTS (m/z 126). In addition, a proportional mixture of compounds found in urine odor was also prepared using acetone (m/z 58), 2-butanone (m/z 72), benzene (m/z 78), toluene (m/z 92), and xylene (m/z 106). The parameters considered for the selection of the characteristic chemicals were: the relevant literature,3,20−26 the abundance of chemicals in the TD-GC-TOF-MS chromatograms (data not shown), the volatility, the uniqueness (if possible) in relation to the source, the expected concentration range in specific time frame in the debris, the PS-MS limit of detection, and the interactions with the building background.4 The stepwise increase profile was followed in both cases. In Figure 4c,d, representative profiles of specific ions of chemical mixtures of interest are being presented. As shown, the profiles of the selected chemicals in the mixtures presented a constant slope or a decrease. This trend is the result of the dominating phenomena inside the evaporation chamber, the chemical interactions, and the flash vaporization of chemicals. Overall, the performance of the vapor generator is strongly affected by the combination of vapor pressure (acetone > 2-butanone > benzene > toluene > xylene > DMDS > DMTS), liquid quantity, and selected air flow rate. Exemplary, compounds with low vapor pressure and at low/medium evaporation chamber flow rate used in high liquid quantity followed the increased stepwise profile, whereas volatiles with high vapor pressure and at high evaporation chamber flow rate used in low liquid quantity were quickly vaporized (decreased profile). However, compounds combined with high evaporation chamber flow rate and low/medium vapor pressure but low liquid quantity presented a constant slope. Note that acetone (highest vapor pressure) presented high abundances compared to the rest of the VOCs and therefore is not shown at all in Figure 4d−f. The main aim of the next experimental cycle was to test the interaction of the device with construction materials that can be found in the debris of a collapsed building. The outlet of the odor simulator was connected to the inlet of the PCMT, while its outlet was connected to the inlet of the PS-MS reference detector. The mixture of analytes resembling urine odor was further employed to study the interaction of the construction materials with the produced odors. The stepwise increase profile was followed for the evaporation chamber air flow rate. Figure 4e presents the interaction of analytes of interest with a PCMT; a minor impact on the synthetic mixture was observed as can be seen from the similar slopes in (e) compared to (d). In the last experiment, the water phase impact on odor generation was examined; practically, this resembles water vapor impact on odor production. An aqueous ammonia solution was selected with the mixture of components found in urine odor for simulating original samples. This provides both the ammonia content of urine and water vapors, simulating the increased humidity in building surroundings. The importance of humidity in confined spaces was highlighted during the trapped human experiment.1 Ten μL of ammonia 25% v/v was
■
RESULTS AND DISCUSSIONS The first experimental set was carried out to examine the response of the vapor generator to the stepwise increase profile of the evaporation chamber flow. The total data acquisition time was approximately 16 min. In Figure 4a, the MS response at m/z 78 is being presented. The stepwise increase of the evaporation chamber flow resulted in a linear increased response. The second experimental set aimed at examining the response of the vapor generator to the increase−decrease profile of the evaporation chamber flow. The total data acquisition time was approximately 26 min. In Figure 4b, the MS response at m/z 78 is being presented. The increase− decrease profile of the flow rates resulted in a relevant pattern of the MS response. In both cases, the produced signal was relatively stable showing a fast decrease to the baseline. The third experimental cycle included measurements carried out on synthetic mixtures of different analytes. The aim of these experiments was to test (a) the simulator performance when using mixtures of compounds and (b) the capability of generating odors of interest. An isomolecular mixture of 6 analytes that can be found in human decay was prepared using the following components: acetone (m/z 58), 2-butanone (m/z 3892
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
added to 40 μL of the mixture. The stepwise increase profile was selected for the evaporation chamber air flow rate. Figure 4f presents the interaction of the analytes that compose the urine synthetic odor mixture along with the ammonia aqueous solution; a minor impact was noticed on the generated vapor as can be seen from the similar slopes in (f) compared to (d). The overall performance of the vapor generator in the presence of synthetic urine mixture or synthetic urine mixture with PCMT or synthetic urine mixture with NH3 aqueous solution is presented in Figure 5. The stepwise profile is clearly shown but
Table 3. Overall Characteristics of the Dynamic Odor Simulator Technology: Portability: Number and type of analytes: Total flow (mL/min): Output concentration: Odor profile: Duration of odor emissions: RSD of odor emissions’ pulse: Affecting factors:
Single compounds and complex mixtures: Interaction:
Reference detector:
controlled evaporation of a liquid quantity yes not limited up to 6 L/min ppbv to low ppmv transient odor profile 30 min 11% evaporation chamber temperature air flow rate through the mixing chamber air flow rate through the evaporation chamber type of compound (different volatility) yesa minimum impact of constructions materials minor impact of humidity pulsed sampling-mass selective detector
a
The number of the single compounds depends on the number of evaporation chambers used.
■
CONCLUSIONS In this work, the construction (design, development, and validation) of a new dynamic gas generator with significant advantages over existing systems (i.e., inexpensive, simple, convenient for field applications) is presented; it generates odors to simulate analysis in harsh environments (e.g., human and nonhuman vapor plume) and enables on-site instrument testing. The developed device can simulate in a controllable and reproducible way vapor phase mixtures, such as the different odors evolving from trapped victims in the voids of a collapsed building (especially the odor of human decay and urine). The synthetic used odors were mixtures of a few characteristic analytes that resemble the real odors emitted from various sources frequently found in the debris of collapsed buildings. The simulator can produce ppbv to low ppmv concentrations that can be detected by the reference detector (online MS), and it also has the potential of producing dynamic concentrations with various levels of humidity. The odor simulator met the requirements that were set during the design phase. This was quite clear for independent compounds of interest, as well as for mixtures of 5 or 6 different analytes. The duration of odor generation was as long as 30 min. Average stability of odor pulse was 11%, as measured by RSD. A minor impact was observed on the vapor generation in the presence of construction materials and humidity. The ANOVA revealed that the independent parameters X2 (air flow rate through the mixing chamber), X3 (air flow rate through the evaporation chamber) and X4 (type of compound) had significant impact on both dependent variables (abundance of m/z and its RSD). However, parameter X1 (evaporation chamber temperature) seems to affect only the abundance of m/z. The controlling and monitoring of these factors strongly affects the overall performance of the generated vapor. The developed simulator potentially can also serve the need for calibrating and evaluating the performance of chemical analytical devices (e.g., gas chromatographers, ion mobility spectrometers, mass spectrometers, sensors, e-noses) in the field. Moreover, it can be used for
Figure 5. Total ion current (TIC) of synthetic urine mixture, synthetic urine mixture with PCMT, and synthetic urine mixture with NH3 aqueous solution. The MS response was identical in the different experimental settings, and the observed delay might be attributed to humidity impact and the interactions with the construction materials.
with lower abundances though in the last two scenarios, because of humidity impact and the interactions with the construction materials. The overall performance of the developed dynamic vapor generator related to the human scent is presented in Table 3. The developed device can simulate in a controllable and reproducible way odors generated from entrapped victims in collapsed building voids. Additionally, the vapor generator can have applications in instruments calibration and testing, in air quality control and environmental pollution systems, in chemical vapor deposition, and in adsorption studies.10 Moreover, the device can be used for enhancing the training of search and rescue canines. The results of statistical analysis (ANOVA) showed that the four independent factors X1, X2, X3, and X4 have significant impact on the dependent variable (normalized abundance) for a confidence interval of 95%. The resulting p-values of the corresponding factors were less than 0.05; therefore, the statistical hypothesis H0 for which there is no significant impact is rejected. Additionally, it was revealed that only the combinations of factors (X1, X2) and (XI, X3) have no significant impact on this dependent variable. Moreover, the statistical hypothesis H0 for factors X2, X3, and X4 and the combinations (X1, X3), (X1, X4), (X2, X4), and (X3, X4) with p-value less than 0.05 are rejected and thus they have significant impact on the other dependent variable (RSD of normalized abundances). 3893
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
Analytical Chemistry
Article
(20) Statheropoulos, M.; Sianos, E.; Agapiou, A.; Georgiadou, A.; Pappa, A.; Tzamtzis, N.; et al. J. Chromatogr., B 2005, 822, 112−117. (21) Agapiou, A.; Mochalski, P.; Schmid, A.; Amann, A. In Volatile Biomarkers: Non-Invasive Diagnosis in Physiology and Medicine; Amann, A., Smith, D., Ed.; Elsevier: New York, 2013; pp 514−558. (22) Mochalski, P.; Buszewska, M.; Agapiou, A.; Statheropoulos, M.; Buszewski, B.; Amann, A. Chromatographia 2012, 75 (1−2), 41−46. (23) Statheropoulos, M.; Spiliopoulou, C.; Agapiou, A. Forensic Sci. Int. 2005, 153 (2−3), 147−155. (24) Statheropoulos, M.; Agapiou, A.; Zorba, E.; Mikedi, K.; Karma, S.; Pallis, G. C.; Eliopoulos, C.; Spiliopoulou, C. Forensic Sci. Int. 2011, 210 (1−3), 154−163. (25) Vass, A. A.; Smith, R. R.; Thompson, C. V.; Burnett, M. N.; Wolf, D. A.; Synstelien, J. A.; et al. J. Forensic Sci. 2004, 49 (4), 760− 769. (26) Stadler, S.; Stefanuto, P.-H.; Brokl, M.; Forbes, S. L.; Focant, J.F. Anal. Chem. 2013, 85 (2), 998−1005. (27) Giannoukos, S.; Brkić, B.; Taylor, S.; France, N. Anal. Chem. 2014, 86, 1106−1114.
improving the training of search and rescue canines, as well as in the identification of illegal human immigration.27
■
AUTHOR INFORMATION
Corresponding Author
*Tel.: +30-210-7723109. Fax: +30-210-7723188. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
■
REFERENCES
The work presented was funded by the European Community’s Seventh Framework Programme (FP7/2007-13) under grant agreement No. 217967; “SGL for USaR” project (Second Generation Locator for Urban Search and Rescue operations, www.sgl-eu.org). Furthermore, financial support of the Bundesministerium f ür Bildung und Forschung and the Ministerium f ür Innovation, Wissenschaf t und Forschung des Landes Nordrhein-Westfalen is gratefully acknowledged.
(1) Huo, R.; Agapiou, A.; Bocos-Bintintan, V.; Brown, L. J.; Burns, C.; Creaser, C. S.; Davenport, N. A.; Gao-Lau, B.; Guallar-Hoyas, C.; Hildebrand, L.; Malkar, A.; Martin, H. J.; Moll, V. H.; Patel, P.; Ratiu, A.; Reynolds, J. C.; Sielemann, S.; Slodzynski, R.; Statheropoulos, M.; Turner, M. A.; Vautz, W.; Wright, V. E; Thomas, C. L. P. J. Breath Res. 2011, 5, 1−12. (2) Vautz, W.; Slodzynski, R.; Hariharan, C.; Seifert, L.; Nolte, J.; Fobbe, R.; Sielemann, S.; Lao, B. C.; Hildebrand, L. Anal. Chem. 2013, 85 (4), 2135−2142. (3) Rudnicka, J.; Mochalski, P.; Agapiou, A.; Statheropoulos, M.; Amann, A.; Buszewski, B. Anal. Bioanal. Chem. 2010, 398 (5), 2031− 2038. (4) Mochalski, P.; Agapiou, A.; Statheropoulos, M.; Amann, A. Analyst 2012, 137 (14), 3278−3285. (5) Statheropoulos, M.; Mikedi, K.; Agapiou, A.; Georgiadou, A.; Karma, S. Anal. Chim. Acta 2006, 566 (2), 207−216. (6) Naganowska-Nowak, A.; Konieczka, P.; Przyjazny, A.; Namieśnik, J. Crit. Rev. Anal. Chem. 2005, 35 (1), 31−55. (7) Namiesnik, J. J. Chromatogr., A 1984, 300, 79−108. (8) Koziel, J. A.; Martos, P. A.; Pawliszyn, J. J. Chromatogr., A 2004, 1025 (1), 3−9. (9) Barratt, R. S. Analyst 1981, 106 (1265), 817−849. (10) Li, Y.; Taffner, T.; Bischoff, M.; Niemeyer, B. Int. J. Chem. Eng. 2012, article ID 417029, 6 pages (DOI:10.1155/2012/417029). (11) Xiong, G.; Pawliszyn, J. Anal. Chem. 2002, 74 (10), 2446−2449. (12) Przyk, E.; Konieczka, P.; Szczygelska-Tao, J.; Teschner, R.; Biernat, J. F.; Namiesnik, J. J. Chromatogr., A 2001, 928 (1), 99−108. (13) Vautz, W.; Schmäh, M. Int. J. Ion Mobility Spectrom. 2009, 12, 139−147. (14) Thomas, C. L. P.; Alder, J. F.; Fielden, P. R.; Przybylko, A. R. M.; Snook, R. D.; Watson, A. F. R. Anal. Chim. Acta 1993, 214, 179− 183. (15) Statheropoulos, M.; Kyriakou, S.; Pappa, A. Thermochim. Acta 1999, 329 (1), 83−88. (16) Statheropoulos, M.; Kyriakou, S.; Tzamtzis, N. Thermochim. Acta 1998, 322 (2), 167−173. (17) Pappa, A.; Kyriakou, S.; Mikedi, K.; Tzamtzis, N.; Statheropoulos, M. J. Therm. Anal. Calorim. 2004, 78 (2), 415−426. (18) Statheropoulos, M.; Mikedi, K.; Stavrakakis, P.; Agapiou, A.; Karma, S.; Pallis, G. C.; Pappa, A. Sens. Actuators, B: Chem. 2011, 159 (1), 193−200. (19) McClennen, W. H.; Vaughn, C. L.; Cole, P. A.; Sheya, S. N.; Wager, D. J.; Mott, T. J.; Dworzanski, J. P.; Arnold, N. S.; Meuzelaar, H. L. C. Field Anal. Chem. Technol. 1996, 1 (2), 109−116. 3894
dx.doi.org/10.1021/ac404175e | Anal. Chem. 2014, 86, 3887−3894
