Determination of benzene, toluene, ethylbenzene and xylene in field and laboratory by means of cold fiber SPME equipped with thermoelectric cooler and GC/FID method

Abstract A simple and effective cooling device based on a thermoelectric cooler was applied to cool the SPME fiber. The device was used for quantitative extraction of aromatic hydrocarbons in the air. Several factors such as coating temperature, extraction temperature and relative humidity in the laboratory setting were optimized. Comparison of the results between the cold fiber SPME (CF-SPME) and NIOSH 1501 method on standard test atmosphere indicated a satisfactory agreement. The CF-SPME and SPME method were also compared. The results revealed that CF-SPME has the most appropriate outcome for the extraction of aromatic hydrocarbons from the ambient air. The cold fiber SPME technique showed good results for several validation parameters. Under the optimized conditions, the limits of detection (LOD) and the limits of quantification (LOQ) ranged from 0.00019 to 0.00033 and 0.0006 to 0.001 ng ml−1, respectively. The intra-day relative standard deviation (RSD) showed ranging from 4.8 to 10.5%.


INTRODUCTION
BTEX (benzene, toluene, ethylbenzene, and xylene) are common hazardous volatile organic compounds (VOCs) and toxic air pollutants emitted into the atmosphere from natural and artifi cial sources 1 . These compounds are widely used in industries, such as paint, printing, synthetic rubber and resin, detergent, ink and pesticides. Vehicle fuel combustion and industrial processes are the major sources of emission of these pollutants in the outdoor air 2, 3 . Exposure to the BTEX can cause adverse health effects such as cancer, neurological disorders and damage to respiratory system, liver and kidneys 4, 5 .
Due to the potential health problems and environmental impacts that VOCs may cause, several sampling and analysis methods have been developed for air monitoring of these pollutants, of which the passive or active sampling by sorbent tubes, gravimetric fi lters, impingers and canisters followed by solvent or thermal desorption using gas chromatography for detection purposes are the most predominant ones 6 . Despite their fair reliability, many of these methods have several serious drawbacks, such as the need for considerable sampling expertise, complex, laborious, multistep and lengthy sample collection and preparation along with sophisticated equipment and complicated and costly extraction procedures 7, 8 . In many cases, conventional air-sampling methods are not applicable to indoor and outdoor air sampling, particularly in cases where very low LOD are required 9 . Recently, extensive efforts towards modernization of analytical instruments and elimination of multistep sample-preparation techniques have led to the widespread application of solvent-free approaches in environmental and occupational exposure assessments 10, 11 .
Solid phase microextraction (SPME) is an innovative technique with several major advantages, including simplicity of use, faster implementation, low solvent consumption, higher time and cost-effi ciency and automation capability 12, 13 . Since its development, is has been shown to be convenient for fi eld and laboratory analysis and has been successfully applied to the sampling and analysis of various contaminants in the air, water, and soil 14-16 . Nonetheless, due to several limitations particularly in terms of selectivity, sensitivity and extraction capability of SPME, researchers proposed new confi gurations for this technique in order to minimize its exothermic effect and improve the extraction effi ciency 17, 18 .
The cold fi ber SPME (CF-SPME) was introduced in 1995 to signifi cantly overcome some of these drawbacks 19 . The cold fi ber method involves simultaneous increase in sample temperature and decrease in fi ber temperature 20 . Heating the sample to the elevated temperatures usually provides the necessary energy for the target analytes to overcome the barriers that bind them to the sample matrix, and therefore improves the mass transfer process and maximizes the vapor pressure of the analyte. On the other hand, due to the exothermic nature of the adsorption, application of high temperature can adversely affect the partition coeffi cient of the analytes and reduce the extraction capabilities. In order to tackle this problem, the coating temperature is therefore lowered, which in turn, signifi cantly improves the extraction process 21 .
The fi rst cold fi ber SPME device was introduced in 1995 by Zhang and Pawliszyn for the extraction of BTEX compounds from clay soil and sand samples 19 . Such device was then miniaturized and automated by Chen et al. 22 . Since then several authors such as Ghiasvand et al. A review of the previous studies reveals that CO 2 and N 2 gases have been widely used as coolant agents for sampling and determination of several volatile compounds in solid and aqueous samples 21, 27 . As the use of such techniques usually entails high voltage and heavy equipment, the portability and cost-effectiveness characteristics of them along with the instability in controlling the fi ber temperature are the most considerable problems associated with the utilization of these gas cooling techniques.
Referring to the previous studies conducted in this fi eld it is almost evident that CF-SPME has been mostly used for sampling and determination of VOCs compounds in solid and aqueous samples. Air sampling with CF-SPME, however, is the novel application of this technique that has been less investigated than the other techniques in this fi eld. The purpose of the current study is to investigate the applicability of CF-SPME as a monitoring tool in air sampling particularly in occupational and/or environmental assessment studies. In this research a CF-SPME device was constructed based on thermoelectric cooler (TEC) for cooling the SPME fi ber in order to improve extraction effi ciency and resolve the remaining problems with ordinary SPME fi bers. The cooling device described in this study has proven to be an effective, inexpensive, fast, environmentally safe and reliable technique for the determination of volatile organic compounds in ambient air. However, to the best of our knowledge there is no similar study based on cooling techniques for quantification of these compounds in the air.
Two stock mixtures of BTEX were prepared in methanol and carbon disulfi de with a concentration of 1000 μg ml -1 for each compound. They were stored in a refrigerator at 4 o C. Standard working solutions of the analytes were prepared daily from the stock solution in methanol.

Instruments
Gas chromatographic analyses of the air samples were performed by a Shimadzu GC-2010 system (Kyoto, Japan) equipped with a split/splitless injector and fl ame ionization detector (FID). Analytes were separated using an Rtx®-5 (30 m*0.25 mm*0.25 μm) fused silica column from Restek (Bellefonte, PA, USA). Ultra pure N 2 at a fl ow rate of 1.1 and 40 ml min -1 was used as a carrier and make-up gas, respectively. The column temperature was initially held at 35 o C for 1 min and gradually increased to 100 o C at a rate of 4 o C min -1 and then to 200 o C at a rate of 20 o C min -1 . The temperatures associated with the injector and detector were set and kept at 290 and 250 o C, respectively.
An Alfa hs-810 model hot plate-stirrer (Tehran, Iran) was used in a humidity generation system for adjusting the relative humidity inside the standard chamber. Temperature controller (Busan, South Korea), the thermoelectric cooler and related heat sink and fan (Busan, South Korea) and other electronic components were purchased from the electronic stores. The SPME holder and carboxen/polydimethylsiloxane (CAR/PDMS, 75 μm) fi ber were provided by Supelco (Bellefonte, PA, USA). When new, the fi ber was conditioned in a GC injector according to manufacturer instructions.
A syringe pump, JMS SP-510 (Hiroshima, Japan), was used for preparing the standard concentration and thereby determined the calculated amount of target analytes injected into the sampling chamber. A high volume sampling pump SKC (PA, USA) was used for drawing air through the chamber. A low volume sampling pump (SKC 222 series, PA, USA), with a sampling fl ow rate of 1-200 mL min -1 and 150 mg charcoal sorbent tubes (SKC Inc., PA) were used for the performance evaluation of CF-SPME device and accurate drawing of air through the sorbent bed.

Cold fi ber-SPME device
A simple, inexpensive and effective cooling device based on the use of a thermoelectric cooler (TEC) as the cooling source was applied to cool the SPME fi ber via SPME needle. The schematic illustration of the cold fi ber SPME device is shown in Figure 1. A heat sink, fan and silicon-oil based thermal grease were attached to the hot side of the thermoelectric cooler in order to dissipate the generated heat and make it cool. To achieve more effi cient cooling and improve the contact between the SPME needle and the cold side of TEC, A copper plate was mounted on the cold surface of the TEC. A thermocouple was embedded in this copper plate to monitor the temperature of the cold side of the TEC. A temperature controller was also used to adjust and precisely control the fi ber temperature.

Selection of SPME fi ber
Choosing of an appropriate coating is the most important step in the sampling and analysis of the analytes in the air. The selected coating should have a good sensitivity toward the target analytes. CAR/PDMS is the mixed coating which is mainly composed of microporous structure that make it appropriate for the analysis of small molecular and non-polar compounds 11, 28 . In the current study, CAR/PDMS was chosen as a proper coating for simultaneous extraction of BTEX, because it has better sensitivity and higher selectivity for extraction of aromatic compounds than other types of available fi bers and also it is recommended by the Supelco fi ber selection guide for gases and low molecular weight compounds 4, 29, 30 .

Standard gas generating system and Sampling by CF--SPME device
A dynamic gas generation system was applied to prepare certain concentrations of gaseous standards in the range of interest for laboratory testing. Predetermined amount of each analyte was injected using a syringe pump into a fl ow direction line connected to the sampling chamber in order to provide dynamic standard concentration of BTEX inside a chamber. The sampling temperature was adjusted to 20, 30 and 40 o C using a thermostated lighting radiation lamp inside a heat control chamber, located upstream of the standard chamber. This temperature controller system managed to adjust the temperature inside the chamber in a defi ned range. For adjusting the relative humidity inside the chamber, a humidity generation system consisted of an impinger, heater and control valve for changing air fl ow rate and temperature of water in the impinger was used, and relative humidity was also successfully set at two levels of 30% and 70%. A hygrometer was applied to monitor relative humidity. A high volume sampling pump was used to draw air through the sampling chamber. A dry test meter (Elster--Handel, Germany) that had been calibrated according to the primary standard device was employed to continuously check the diluent gas fl ow rate in the system. Figure 2 shows the sampling set used in this study. Sampling and adsorption of the analytes were performed by inserting the SPME needle into the sampling chamber and exposing the fi ber to the analytes. After sampling, analysis was performed by thermal desorption of adsorbed analyte in GC injection port (by GC/FID system).

Active sampling
Active air sampling was performed according to NIOSH 1501 method. According to this method, a personal pump calibrated at 100 ml/min fl ow rate, drew a specifi ed volume of air to the charcoal tubes. Samples were extracted with 1 ml of carbon disulphide, agitated with a rotary shaker for 30 minutes and then analyzed by GC-FID 31 .

Parameters related to the laboratory phase of the study
The sampling rate of SPME could be affected by environmental parameters. In order to achieve the best sampling effi ciency, several factors such as coating temperature, extraction temperature and relative humidity inside the standard chamber, were investigated and optimized.

Eff ect of coating temperature on sampling effi ciency
To determine the ideal fi ber temperature for CF-SPME technique, the effect of this parameter on the sampling rate was investigated in the range 5-20 o C. As it is illustrated in Figure 3, the extraction capabilities increased by decreasing the coating temperature to 5 o C. Regarding the exothermic nature of the adsorption process of volatile compounds onto SPME fi ber, these compounds tend to retain in the fi ber at lower temperatures 32 . Therefore, 5 o C was fi nally chosen as the optimum coating temperature for the subsequent evaluations. The effects of temperature, relative humidity and coating temperature on the sampling phase were examined. The type of fi ber coating and GC desorption time and temperature parameters were also determined from the other studies. Having optimized the performance parameters of the SPME for use as a passive sampler, the device was applied to the determination of BTEX concentrations in several gas stations and car painting workshops.

Eff ect of temperature on sampling effi ciency
To use SPME and CF-SPME in the fi eld, the effect of the fi eld related parameters on the sampling effi ciency and the performance of the new device was investigated. The air temperature in the sampling chamber was varied from 20 to 40 o C, to study its effect on sampling effi ciency.
For SPME, an increase in extraction temperature caused a decrease in extraction rate and peak area responses of the GC analysis 33 . The reason for such phenomenon is adverse effect of temperature on the distribution coeffi cient. In other words, the increase in temperature usually enhances headspace extraction and simultaneously reduces the fi ber coating/sample distribution constant. Since the adsorption is an exothermic process, reduction in extracted mass at equilibrium is usually expected in high extraction temperature. Thus, an extraction temperature of 20 o C was found to be optimal.
For CF-SPME, the amount of adsorbed analytes increased almost linearly with the increase in air temperature from 20 to 40 o C. In CF-SPME extraction mode, adverse effect of the high temperature on the distribution constant and extraction effi ciency is resolved by lowering the coating temperature. As a result, sampling in higher temperature is possible. With this regard, 40 o C was identifi ed as an optimal extraction temperature (Fig. 4).
tained with the CF-SPME and NIOSH method under the laboratory conditions. Table 1 also revealed that the reproducibility of the CF-SPME technique was to some extent better than the conventional charcoal tube method for air sampling.

Comparison and validation of the optimized method
In order to demonstrate the effect of cold fi ber on the extracted amount, SPME and CF-SPME approach were compared under the optimum sampling conditions. As can be seen from Figure 6, the extraction rate in CF-SPME is more effi cient than SPME. This high sampling rate can be attributed to the large distribution coeffi cient of target analytes at lower coating temperature and high sample temperature. Therefore, to obtain high sensitivity, CF-SPME was applied for further analysis.

Eff ect of humidity on sampling effi ciency
The effect of 30 and 70% relative humidity on sampling effi ciency was investigated. In general, relative humidity had a negative effect on the extracted amount of both techniques; i.e., less mass was adsorbed at high level of RH (Fig. 5). The fi ndings of the current study are consistent with those of  and Zare Sakhvidi (2012) who found that an increase in the relative humidity caused a decrease in extraction rate 14, 34 . It seems that, the competition between water molecules and target analytes for occupying active adsorption sites is the reason of such decrease in the amount of adsorbed analytes. In fact, condensation of water molecules at the surface of coating alters the coating properties and fi ber selectivity and thereby reduces the sampling rates 34, 35 .

Validation with NIOSH method
NIOSH 1501 method and the CF-SPME method were compared, in terms of accuracy, under laboratory-controlled conditions (Table 1). Laboratory results indicated that there is a fair agreement between the results ob-In order to verify the CF-SPME method for the analysis of the air samples, analytical characteristics of this technique such as limit of detection (LOD), limit of quantifi cation (LOQ) and repeatability were investigated. The quantifi cation of the BTEX using CF-SPME device was based on the calibration curves obtained under non-equilibrium conditions and 15 min sampling time. Calibration curves were plotted using fi ve concentration levels ranging from 0.03 to 160 ng ml -1 , with three extraction replicates for each level. The LOD and LOQ were estimated according to the recommendations of Eurachem Guide 36 using ten measurements of the blank.
The LOD and LOQ for all analytes ranged from 0.00019 to 0.00033 and 0.0006 to 0.001 ng ml -1 , respectively. Repeatability of the method was examined by calculating the relative standard deviation (RSD) of peak response of fi ve samples at the concentration level of 2 and 50 ng ml -1 on the same day. The RSDs also demonstrated a reasonable repeatability for the proposed method (see Table 2). The calculated relative standard deviation indicated a range of 8.2% to 10.5% at the low level and from 4.8% to 9.4% for the high level.

Field sampling by CF-SPME device
Having optimized all aspects of SPME sampling in the laboratory, the device was applied to determine the concentration of BTEX in several petrol stations and car painting shop to test the method. Four calibrated CAR/PDMS fi bers were used for air sampling purpose in each environment. Sampling was performed using SPME and NIOSH 1501 techniques in the respiration zone (about 1.70 m above the ground level). Immediately after sampling, all fi bers were withdrawn into the barrel and capped with a new and clean silicone septum and placed on a dry ice bed. As the analytes were highly volatile, all SPME fi bers were analyzed during maximum 2 h after sampling in order to avoid volatilization. The analytical results for determination of these analytes in air samples are given in Table 3. At the petrol station capability of the method for trace level analysis in different environments.  Table 3. Concentrations (mg/m 3 ) of the analytes in outdoor and indoor air samples by CF-SPME and painting shop the concentrations of toluene, ethyl benzene and xylenes were lower than those recommended by NIOSH recommended exposure limits (RELs) but benzene concentration in both environments was higher than the recommended value (0.32 mg/m 3 ).

CONCLUSION
The current study detailed the development of an alternative approach for the analysis of aromatic compounds in ambient air with cold fi ber solid phase microextraction (CF-SPME). It was concluded that the proposed method was relatively precise, fast and sensitive for quantitative extraction of volatile compounds in air samples. The method also demonstrated good sensitivity and precision for the range of environmental interest. Comparison of the CF-SPME and NIOSH method showed that the proposed technique is a powerful alternative for NIOSH--based fi eld sampling. The applicability of the method for outdoor and indoor air samples demonstrated the