Preparation by low-temperature nonthermal plasma of graphite fiber and its characteristics for solid-phase microextraction

Analytica Chimica Acta 631 (2009) 62–68

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Preparation by low-temperature nonthermal plasma of graphite fiber and its characteristics for solid-phase microextraction Fan Luo a , Zucheng Wu a,∗ , Ping Tao b , Yanqing Cong c a Department of Environmental Engineering, State Key Laboratory of Clean Energy Utilization, Key Laboratory of Polluted Environment Remediation and Ecological Health, MOE, Zhejiang University, Hangzhou 310027, PR China b Institute of Structural Mechanics, China Academy of Engineering Physics, Mianyang 621900, PR China c College of Environmental Science and Engineering, Zhejiang Gongshang University, Hangzhou 310012, PR China

a r t i c l e

i n f o

Article history: Received 17 July 2008 Received in revised form 3 October 2008 Accepted 6 October 2008 Available online 22 October 2008 Keywords: Solid-phase microextraction Graphite fiber Nonthermal plasma Adsorption

a b s t r a c t Low-temperature nonthermal plasma has been used to prepare solid-phase microextraction (SPME) fibers with high adsorbability, long-term serviceability, and high reproducibility. Graphite rods serving as fiber precursors were treated by an air plasma discharged at 15.2–15.5 kV for a duration of 8 min. Sampling results revealed that the adsorptive capacity of the homemade fiber was 2.5–34.6 times that of a polyacrylate (PA) fiber for alcohols (methanol, ethanol, isopropyl alcohol, n-butyl alcohol), and about 1.4–1.6 times and 2.5–5.1 times that of an activated carbon fiber (ACF) for alcohols and BTEX (benzene, toluene, ethylbenzene, and xylenes), respectively. It is confirmed from FTIR (Fourier transform infrared spectrophotometer) and SEM (scanning electron microscope) analyses that the improvement in the adsorptive performance attributed to increased surface energy and roughness of the graphite fiber. Using gas chromatography (GC)-flame-ionization detector (FID), the limits of detection (LODs) of the alcohols and BTEX ranged between 0.19 and 3.75 ␮g L−1 , the linear ranges were between 0.6 and 35619 ␮g L−1 with good linearity (R2 = 0.9964–0.9997). It was demonstrated that nonthermal plasma offers a fast and simple method for preparing an efficient graphite SPME fiber, and that SPME using the homemade fiber represents a sensitive and selective extraction method for the analysis of a wide range of organic compounds. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Pawliszyn and co-workers introduced solid-phase microextraction (SPME) nearly twenty years ago [1]. The sampling method is efficient in that it integrates extraction, concentration, and sample introduction steps into a single process [2–4]. Since its introduction, several related techniques have stemmed from SPME, such as hollow fiber-protected liquid-phase microextraction [5], thin-film microextraction [6], membrane SPME [7], stir bar sorption [8], intube SPME [9,10], and cold-fiber SPME [11]. Although it has evolved into a family of methods, the traditional SPME method remains the most popular. Undoubtedly, an efficient fiber is the key element of the traditional SPME technique. At present, among seven kinds, 30 types of commercial fibers are available. Most of these fibers are made by simultaneously preparing the film and drawing silica rods [12]. Therefore, they have inherent shortcomings, such as limited coat-

∗ Corresponding author at: Box 1707, 38 Zheda Road, Yuquan Campus, Zhejiang University, Hangzhou 310027, PR China. Tel.: +86 571 85660839. E-mail address: [email protected] (Z. Wu). 0003-2670/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2008.10.025

ings, fragile supports, thermal and chemical instabilities, and high cost. Furthermore, some coated fibers also suffer from significant variations in coating volumes, which result in poor fiber-to-fiber reproducibility [13]. Many researchers have motivated to seek alternative fibers with lower cost, longer lifetime and better extraction performance. Some of them have focused on carbon-based adsorbents, which have displayed non-selective high adsorbability, thermal and chemical stability, cost efficiency, and long-term reusability [14–18]. Preparation processes have included the adhering of carbon granules [19], or the bending [20], evaporation [14], or carbonizing [21] carbon coatings; curing precursors through carbonization, immersion, activation, and post-treatment [22,23]; and combining conglutination of porous silica particles, heat-curing, and sol–gel processing [24]. These methods have mostly aimed at increasing the specific surface areas of the fibers to improve the adsorbability of the material. Though these fabrication procedures have yielded good results, they are time-consuming and labor-intensive, which may result in limited reproducibility, low availability, and restricted applicability of the fibers. In the present study, an innovative plasma surface treatment method is proposed for preparing carbon fibers. Low-temperature

F. Luo et al. / Analytica Chimica Acta 631 (2009) 62–68

63

nonthermal plasma (NTP) can activate surfaces furthermore impart material with specific properties by grafting chemical functions (plasma-active species) and varying its surface energy [25]. The technology has been used to improve the surface properties of metal [26], glass, fibers, polymers and cotton fabrics [27] by enhancing their adhesion, roughness and hydrophilicity. The aim of this paper is to investigate the feasibility of NTP to prepare carbon SPME fibers. The precursor of the fiber is graphite rods which have shown to be an effective sorbent phase for organic compounds [15,28]. The performance of the prepared fibers was evaluated by determining BTEX (benzene, toluene, ethylbenzene, and xylenes) and alcohols (methanol, ethanol, isopropyl alcohol, and n-butyl alcohol) from gas samples. 2. Experimental 2.1. Reagents and materials The reagents used in the experiments were all of analytical reagent grade. BTEX, alcohols, hydrochloric acid (HCl), and sodium chloride (NaCl) were purchased from a local reagent company (Huipu, Hangzhou, China). The precursors of the fibers were pieces of graphite rods (length 360 mm; diameter 0.3 mm) produced by Sakura (Tokyo, Japan). The SPME holder for manual sampling and commercially available fibers, including 100 ␮m polydimethylsiloxane (PDMS) fibers and polyacrylate (PA) 85 ␮m fiber, were obtained from Supelco (Bellefonte, PA, USA). The activated carbon fibers (ACF) were provided by Shanghai Jiao Tong University (Shanghai, China). 2.2. Apparatus Low-temperature NTP was generated in a glass plasmatron [29] of length 29 cm and inner diameter 10 cm. A circular stainless steel net served as the cathode electrode (diameter 8 cm), the planes of which were parallel to the axis of the chamber. The anode electrode was an aciculate copper rod, which stood in the middle of the chamber surrounded by the cathode electrode. The electrodes were connected to a high-tension direct current power supply (12.31–21.45 kV). The analytes were separated in a gas chromatograph with a flame-ionization detector (GC-FID) (model GC-2010, Shimadzu (Tokyo, Japan), or model GC-6820, Agilent (Santa Clara, CA, USA)). A capillary GC column of 30 m × 0.53 mm i.d. coated with 0.25 ␮m film thickness (DB-624, J & W Scientific (Folsom, CA, USA)) was used for all investigations. A NEXUS-670 Fourier-transform infrared (FTIR) spectrometer (Nicolet (Madison, WI, USA)) was used to investigate the fiber surfaces. A SIRION-100 scanning electron microscope (SEM) (FEI (Amsterdam, Holland)) was used to study the surface structural features of the fiber. 2.3. Fiber preparation Before the treatment, the graphite rods were immersed in 1% (v/v) aqueous HCl at 50 ◦ C for 1 h and then washed with distilled water and dried in air. The graphite rods were then placed flat on the cathode between the pair of parallel electrodes during surface modification. By adjusting the applied potential to 15.2–15.5 kV, a uniform plasma was generated to erode the graphite rods. Atmospheric air was used as the plasma-forming gas and low-temperature NTP was produced at atmosphere pressure and room temperature. The substrates were subsequently rinsed three times in distilled water. Prior to use, the fibers were cut into

Fig. 1. Setup for comparison of different fibers with SPME: (1) compared fiber (PDMS, PA, or ACF); (2) the homemade fiber; (3) gas sample; (4) water bath.

1.2 cm lengths and mounted on a commercial SPME assembly. The usable length of the fiber during extraction was 1.0 cm. Prior to their first usage, the fibers were conditioned for 10 min at 220 ◦ C with a 5 mL min−1 flow of nitrogen in the GC injection port. 2.4. Gas sampling system Gaseous samples containing BTEX and alcohols were prepared by two different installations. Samples used in comparative experiments were produced by the technique shown in Fig. 1. Certain quantities of analytes (5 ␮L of each BTEX and 20 ␮L of each alcohol) were injected and volatilized in a 600 mL volume glass vessel, which was airtight and could be vibrated for mixing. After a certain adsorption time, two SPME with the fibers to be compared were simultaneously conducted under the same conditions. A Teflon gasket was placed under the capsule to minimize loss of analytes. The standard gas-generating device shown in Fig. 2 was designed to provide a wide range of concentrations of the target compounds by mixing the gaseous compounds volatilized from liquid compounds with zero air at a certain flow rate, under ambient pressure at 30 ◦ C. To obtain better precision in determining the concentrations, a dilution step was employed to stabilize the airflow. SPME was performed from a glass vessel of about 500 mL volume (5.0 cm i.d.) with one stopcock at each end. A third orifice (Fig. 2, item 10), situated at the center of the cap, fitted with a silicone septum to enable airtight operations, was used to introduce the SPME fiber. The concentrations of the gas samples were calculated from the decreased weights of the reagents in the vials and the volume of gas that had flowed through the system. After adsorption, the fiber was retracted back into the needle and immediately transferred to the injector of the GC. The extraction process was repeated three times for each sample.

64

F. Luo et al. / Analytica Chimica Acta 631 (2009) 62–68

Fig. 2. System for the generation of gaseous standard mixtures: (1) air pump; (2) mass flow controller; (3) humidity controller; (4) zero air chamber, dilution step; (5) mixing chamber; (6) sampling chamber, 2nd dilution step; (7) target compounds; (8) Teflon tube; (9) decontaminator; (10) sampling point; (11) water bath.

2.5. GC conditions In order to optimize the chromatographic conditions, 0.2 ␮L samples of the headspace from above each of the pure BTEX and alcohols and their mixtures were injected separately into the GC injector port and analyzed by the capillary GC-FID method. Separated peaks were obtained when the column temperature was maintained at 40 ◦ C (hold 5 min) and then ramped to 120 ◦ C at a rate of 10 ◦ C min−1 (hold 5 min). The temperatures of the FID system and injector were held at 250 and 220 ◦ C, respectively. Nitrogen was used as carrier gas, at a flow rate of 5 mL min−1 . Splitless mode was used for all the investigations. 2.6. FTIR and SEM FTIR is a very useful technique for assessing the chemical changes induced by plasma treatment. The fibers were conditioned under nitrogen protection for 2 h, and dried before the FTIR analyses. A section of the coating was harvested from the surface of the fiber with a razor blade, then ground and blended with potassium bromide (KBr). The KBr pellet spectra of the coatings were acquired with air as background, with a 10-scan data accumulation at a resolution of 4 cm−1 over the full mid-IR range (4000–400 cm−1 ). The surface morphologies of the fibers before and after treatment with NTP were analyzed by SEM (at 10000× magnification).

Fig. 3. Comparison of SPME fibers before and after treatment (for 8 min) by nonthermal plasma and direct injection of 0.5 mL samples. The concentrations of ethanol, n-butyl alcohol, benzene and ethylbenzene were 26310, 27030, 7300 and 7220 ␮g L−1 , respectively. The extraction and desorption time was 10 and 5 min, respectively.

cess. The NTP was clearly effective for treating the graphite fiber to improve its adsorptive performance. The properties of the homemade fiber were further investigated by comparisons with other SPME fibers and direct injection (DI, 0.5 mL gas sample by syringe) under the same conditions. Two kinds of fibers were employed as extracting phases: commercial fibers, including PDMS and PA, as examples of organic coatings, and a laboratory-prepared carbon fiber, ACF. PDMS and PA fibers are considered to be appropriate for extracting nonpolar and polar compounds, respectively, while ACF shows high adsorptive efficiency for a wide range of organic compounds [11]. It can be seen from Fig. 4(a) that although the PDMS fiber displayed better extraction efficiency for BTEX, the homemade fiber also showed enrichment capability compared to direct injection. Furthermore, as shown in Fig. 4(b), the fiber exhibited an especially high affinity for alcohols compared to the PA fiber, and the latter is known

3. Results and discussion 3.1. Evaluation of the fiber The performance of the fiber was evaluated by extracting ethanol, n-butyl alcohol, benzene, and ethylbenzene from gas samples. SPME with the NTP-treated graphite fiber was compared to SPME with the original graphite fiber. The results of these comparisons are shown in Fig. 3. The adsorptive capacity of the homemade fiber (THF, the fiber after NTP treatment) was 1.0–15.3 times higher than that of the original fiber for the four compounds. The high adsorbability of the homemade fiber must be due to the physical and chemical changes on its surface as a result of the treatment proTable 1 Comparisons of different fibers and direct injection (peak area/peak area). Methanol THFa /PDMS THF/PA THF/ACF THF/DIb THF/DI THF/DI a b

Ethanol

Isopropyl alcohol

n-Butyl alcohol

34.6 1.5

8.0 1.4

11.8 1.5

2.5 1.6

1.1 0.2

1.7 0.4

2.0 0.5

7.0 4.0

THF: the homemade fiber;. DI: direct injection, average data of three samplings.

B

T

E

m, p-X

o-X

0.3

0.3

0.2

0.3

0.2

4.6 0.3

4.8 1.0

5.1 2.5

4.7 3.5

2.5 2.2

0.5

1.8

5.2

6.3

3.5

F. Luo et al. / Analytica Chimica Acta 631 (2009) 62–68

65

adsorbing the eight compounds. The contrasting data are shown in Table 1. The results of the comparisons indicated that the homemade fiber showed relative better affinity for alcohols, which was probably due to the polar functional groups on the surface of the fiber. Moreover, it was also found that the enrichment of the fiber for high-molecular-mass analytes (such as n-butyl alcohol, ethylbenzene, and xylene with adsorption ratios as 4.0, 5.2, 6.3 and 3.5) was higher than that of low-molecular-mass analytes (such as methanol, ethanol, and isopropyl alcohol, with adsorption ratios as 0.2, 0.4 and 0.5). The reason for this phenomenon may possibly be that the size of the cavities on the surface was almost the same as that of the high-molecular-mass compounds. From Fig. 4, it can also be seen that the adsorption of BTEX increased while that of alcohols decreased when sampling mixtures of the two groups of compounds. This observable fact may be ascribed to the limited number of adsorption sites on the surface of the fiber and competitive adsorption between the compounds. 3.2. FTIR and SEM surface analysis The improved adsorptive performance of the treated fiber was evidently due to surface reactions induced by the NTP. In a discharge, free electrons gain more energy from an imposed electric field. Energy transfers to the molecules lead to the formation of a variety of new species, such as metastable free radicals and ions, along with UV radiation and other photons. Furthermore, these new species interact with the surface of the material and cause some physical and chemical changes, such as roughness, oxidation, polymerization, cross-linking, and etching [30]. Evidence of these effects was obtained by FTIR and SEM. Fig. 5 shows FTIR spectra of the fiber surface before and after NTP treatment. The irradiation time of the plasma was 0, 8, and 15 min, respectively, while the other operational parameters were kept constant. It is clear that the plasma treatment considerably enhanced the formation of polar species on the fiber surface. The absorbance peaks of the fiber at 3319, 1647, 1308, and 1021 cm−1 correspond to OH, C=O, C–O stretching, and C–H bending vibrations, respectively. Organic groups on the surface have adhesive properties, so that the fiber after treatment showed greater adsorbability for organic compounds. The spectra obtained from the plasma-treated fibers were very similar, but the absorption intensity of the short-term-treated fiber

Fig. 4. GC-FID chromatogram of organic compounds obtained from different SPME fibers: (a) BTEX samples; (b) alcohols samples; (c) mixed BTEX and alcohols samples. Peaks: (1) methanol, (2) ethanol, (3) isopropyl alcohol, (4) benzene, (5) n-butyl alcohol, (6) toluene; (7) ethylbenzene, (8) m, p-xylene, (9) o-xylene. The homemade fiber was treated by plasma for 8 min. The concentrations of alcohols and BTEX were 26380, 26310, 26180, 27030, 7300, 7220, 7230 and 7300 ␮g L−1 , respectively. The adsorption and desorption time was 40 and 5 min, respectively.

to be effective in the extraction of polar compounds. The result of the comparison between the homemade fiber and ACF, which is presented in Fig. 4(c), also indicates good adsorbability of the homemade fiber (1.4–5.1 times higher). Compared to direct injection, it is obvious that the homemade fiber was all-purpose for

Fig. 5. FTIR spectra, from the top down, of homemade fibers exposed to plasma treatment for 8, 15, and 0 min, respectively.

66

F. Luo et al. / Analytica Chimica Acta 631 (2009) 62–68

was higher than that of the long-term-treated fiber. Two reasons for this may be cited. First, the etching process was predominant on the amorphous region of the surface as opposed to the crystalline region. Therefore, it is possible that the initial rate of etching may have been more rapid. Once all the amorphous materials on the surface had been removed, the remaining crystalline and tightlybound amorphous regions might not have been removed so easily [30]. Hence, a prolonged treatment time would not have led to any great improvement in the adsorbability of the fiber. Second, a long irradiation time could lead to collapse of the cavities on the surface or the ablation of functional groups, thereby decreasing the adsorbability. Usually, enhanced roughness of a surface increases the surface area and hence increases the adsorptive efficiency of analytes. The surface morphologies of the fibers treated for 0, 8, and 15 min were investigated by SEM and the images obtained are shown in Fig. 6. The SEM images clearly reveal the cleavage of the topmost layers of the fiber surface. The degrees of micro-roughness of the plasmatreated fiber shown in Fig. 6(b) are higher than that of the untreated fiber shown in Fig. 6(a). Some cracks and holes are distributed over the surface after NTP treatment. These surface changes, due to the irradiation of the plasma, can provide a new pathway for organic compounds to enter the fiber and hence increase its adsorbability. However, when the plasma irradiation time was extended to 15 min, the surface started to degenerate, which hampered the adsorption or desorption of organic compounds (Fig. 6(c)). Surface modification, including the immobilization of functional groups and physical etching by NTP, adequately accounts for the observed improvement in the affinity of the materials. In addition, materials with special functional groups and high surface area can be expected to display excellent adsorptive performance. 3.3. Properties of the homemade fiber 3.3.1. Adsorption equilibrium time To carry out an SPME procedure, the distribution and adsorption equilibrium of the analytes between the gaseous and solid phases needs to be established. In general, the equilibrium is dependent on the sorbent used and can vary significantly among different compounds. The optimal adsorption time was assessed using gas samples, whereby only the adsorption time was varied while the rest of the experimental conditions were kept constant. Fig. 7 shows that a 30 min adsorption time was sufficient to achieve equilibrium for most of the eight compounds. Because SPME is a non-exhaustive approach, the adsorption time was set at 5 min considering the total throughput time and minimizing the effects of competitive adsorption. 3.3.2. Desorption temperature and time The temperature of the GC injector and the appropriate desorption time were important parameters in ensuring that analytes were efficiently desorbed from the stable fibers. For the homemade fiber, five desorption temperatures ranging from 200 to 320 ◦ C in 20 ◦ C increments were selected to test the thermal stability of the fiber and to determine the optimal desorption temperature. The desorption time was 5 min in the experiments. The results showed that the fiber remained stable up to 320 ◦ C, and that the peak areas were hardly enhanced on increasing the temperature from 200 to 320 ◦ C. It was thus concluded that 200 ◦ C was high enough for desorption of the compounds adsorbed on the fiber. Nevertheless, to ensure sensitivity and repeatability, a temperature of 220 ◦ C was selected for both groups of compounds in further investigations. The desorption time should be sufficient for quantitative unloading of the adsorbed analytes from the surface of the SPME

Fig. 6. Micrographs of the graphite fibers: (a) without treatment, (b) after treatment for 8 min, (c) after treatment for 15 min.

fiber. Desorption profiles of the compounds were obtained by plotting GC-FID response against desorption time (0.5, 1, 2, 3, 4, and 5 min). From the results, it can be seen that the peak areas of the analytes increased slightly as the desorption time was increased from 0.5 to 5 min. This indicates that a short time is enough for desorption of the compounds adsorbed on the fiber. A desorption time of 3 min was used in the subsequent experiments, to achieve the highest sensitivity while avoiding carry-over.

F. Luo et al. / Analytica Chimica Acta 631 (2009) 62–68

67

Fig. 8. The performance of the homemade fiber used to adsorb n-butyl alcohol and benzene from gaseous samples with the same concentrations as in Fig. 4. The adsorption and desorption time was 5 and 3 min, respectively.

Fig. 7. Adsorption time profiles of the homemade fibers; the desorption time and the concentrations of the compounds were the same as in Fig. 4.

3.3.3. Durability of the fibers The lifetime of the homemade fiber under the proposed operation conditions was investigated (Fig. 8). On average, a mounted homemade fiber could be used over 400 times. In addition, no significant deterioration of the adsorbability of the fiber was found. This high operational stability may be ascribed not only to the heat-resistance and chemical stability of the material, but also to the novel method of preparation. Surface activation of atmospheric plasma can remain stable over quite a long period [32].

3.3.4. Analytical performance The validation of SPME method with the homemade fiber was conducted by determining the analytes spiked in gas samples at different concentration levels. The linear range (LR), limits of detections (LODs), repeatability (for one fiber) and reproducibility (fiber-to-fiber) of the method were all investigated and the results are listed in Table 2. The linear range of the method was tested by extracting different gas standards with increasing concentrations. All compounds exhibited wide linear ranges of five orders of magnitude with squared regression coefficients (R2 ) ranging between 0.9964 and 0.9997. This allowed the quantification of these compounds by the method of external standardization. The LODs defined as three times the baseline noise, were in the range 0.19 to 3.75 ␮g L−1 . It is clear that the LOD values for BTEX were a little higher than those for the alcohols. There are two possible reasons for this: (1) alcohols are more volatile than the BTEX; and (2) the affinity of BTEX for the fiber is stronger than that of alcohols. The data indicate that the concentration range of linear peakarea-to-concentration ratio is large, with a good coefficient. The simple fiber preparation processes should ensure good reproducibility of its performance. In order to check the reproducibility of the homemade fiber, four fibers were prepared in the same batch and four repeated extractions were performed with each fiber. The repeatability for each single fiber was evaluated by consecutive extracting gas samples spiked of each compound (six replicates). The relative standard deviations (R.S.D.s) were calculated to be from 1.6% to 6.2%. Fiber-to-fiber reproducibility

Table 2 Linear range (LR), linearity (R2 ), limits of detection (LOD), repeatability (R.S.D.%), and reproducibility (R.S.D.%) (Some analytical data) obtained for SPME of alcohols and BTEXs using the homemade fiber and GC-FID. Analyte

Regression equationa

LR (␮g L−1 )

R2

LOD (␮g L−1 )

Repeatabilityb (%, n = 6)

Reproducibilityb (%, n = 4)

Methanol Ethanol Isopropyl alcohol n-Butyl alcohol Benzene Toluene Ethyl benzene m, p- Xylene o-Xylene

A = 0.165C + 56.236 A = 0.342C − 22.323 A = 0.690C − 22.936 A = 1.740C + 47.669 A = 1.254C + 8.146 A = 2.330C + 85.379 A = 4.229C + 0.181 A = 1.740C − 70.587 A = 0.357C + 29.779

5.6–35619 4.2–35529 1.9–35358 1.6–12443 1.4–11240 1.5–10631 0.6–11090 3.1–10997 6.1–10997

0.9994 0.9997 0.9973 0.9991 0.9964 0.9997 0.9967 0.9996 0.9985

3.75 1.82 1.30 0.60 0.54 0.27 0.19 0.70 3.29

1.6 6.2 5.6 2.5 2.8 2.0 4.3 4.3 4.5

5.4 7.7 3.0 1.5 7.5 5.9 6.5 5.4 2.9

a b

A is the peak area and C is the analyte concentration in ␮g L−1 . The concentrations of the compounds were the same as in Fig. 4.

68

F. Luo et al. / Analytica Chimica Acta 631 (2009) 62–68

method proposed here is suitable for investigating complex mixtures of volatile organic compounds. 4. Conclusions

Fig. 9. Gas chromatogram of an extract from exhaust gas by means of the homemade fiber. Peaks: (1) methanol, (2) formaldehyde, (3) acetone, (4) isopropyl alcohol, (5) carbon disulfide, (6) ethyl acetate, (7) cyclohexane, (8) benzene, (9) n-heptane, (10) n-butyl alcohol, (11) toluene, (12) tetrachloroethylene, (13) chlorobenzene, (14) ethylbenzene, (15) m, p-xylenes, (16) o-xylene, (17) cyclohexanone, (18) benzoic aldehyde, (19) benzyl alcohol. The adsorption and desorption time was 30 and 3 min, respectively.

performed using four fibers was in the range 1.5–7.7%. The results demonstrate that the fiber preparation method has the advantage to manufacture steady-performance fibers. 3.4. Applicability The SPME method, optimized by means of the experiments described above, was successfully applied to the analysis of gas emissions containing BTEX and alcohols. Gas samples from a paint plant were trapped in gas bags and then extracted by the SPME method with the homemade fiber under the optimal conditions. From Fig. 9, it can be seen that the target peaks were well separated, despite the complexity of the sample. Eight compounds were found at concentrations in the analytical linear ranges listed in Table 2. The recovery of the SPME method, investigated by analysis of two samples spiked with standard gas at appropriate concentrations, was found to be 92.7–108.3% (Table 3). This indicated that the precision was acceptable when the homemade SPME fiber was applied to the analysis of real samples. The Table 3 Results from the analysis of polluted gasous samples by SPME–GC. Analyte

Nativea (␮g L−1 )

Addition (␮g L−1 )

Apparent (␮g L−1 )

Methanol

200

3210 25904

3215 25899

94.3 ± 3.1 99.2 ± 2.8

Ethanol

1680

3202 25839

5054 25568

103.5 ± 1.6 92.9 ± 3.6

Isopropyl alcohol

2170

3214 25715

5435 24471

100.9 ± 1.9 98.5 ± 2.5

n-Butyl alcohol

1720

7194 9442

8829 10547

99.1 ± 4.6 94.5 ± 3.8

Benzene

3140

6637 9125

9339 11364

95.5 ± 0.9 92.7 ± 1.7

Toluene

1100

7087 9745

7927 10854

96.8 ± 1.5 100.1 ± 4.4

Ethyl benzene

90

7097 9315

6923 9708

96.3 ± 3.7 103.2 ± 2.6

m, p-Xylene

2930

7038 9238

10799 11500

108.3 ± 3.3 94.5 ± 1.8

o-Xylene

190

7918 9238

8021 9222

98.9 ± 2.8 97.8 ± 3.6

a

Exhaust gas from Hangzhou Paint Plant, Zhejiang, China.

Recovery (%)

The studies reported herein provide an alternative method for the preparation of SPME fibers by treatment with low-temperature NTP. The preparative procedures are simple, convenient, and efficient. Besides high adsorptive efficiency, the fiber displayed a wide applicability for polar and nonpolar compounds, long-term stability of over 400 usages, and high reproducibility among four fibers with R.S.D. below 7.7%. It can be concluded that the increase in surface energy and roughness due to the plasma treatment were responsible for the improvement in the adsorbability of the fibers. Although the preliminary work has achieved amazing results, comprehensive optimization of the treatment conditions and variation of the materials deserve further investigation. The results of this study have demonstrated that NTP is effective for preparing graphite fibers with versatile performance in SPME. New materials and techniques might be applied in order to develop excellent SPME fibers and to offer new analysis methods. Acknowledgments The authors would like to thank the Natural Science Association Foundation of China (NSAF, 10376032), the National Science Foundation of China (20706048), the National High-tech Foundation of China (2002AA529182) and the Natural Science Foundation of Zhejiang province (Z505060) for financial support of this study. Dr. Ying Kang who offered the plasmatron and Professor Jinping Jia in Shanghai Jiao Tong University who provided the ACF are also acknowledged. References [1] S.B. Hawthorne, D.J. Miller, J. Pawliszyn, C.L. Arthur, J. Chromatogr. 603 (1992) 185. [2] J.M. Liu, Q.F. Zhou, G.B. Jiang, J.F. Liu, J.Y. Liu, M.J. Wen, J. AOAC Int. 86 (2003) 461. [3] H. Bagheri, E. Babanezhad, F. Khalilian, Anal. Chim. Acta 616 (2008) 49. [4] C.H. Deng, N. Li, X.M. Zhang, J. Chromatogr. B 813 (2004) 47. [5] H.S.N. Lee, M.T. Sng, C. Basheer, H.K. Lee, J. Chromatogr. A 1196 (2008) 125. [6] I. Bruheim, X.C. Liu, J. Pawliszyn, Anal. Chem. 75 (2003 1002). [7] R.Q. Yang, W.L. Xie, Forensic Sci. Int. 162 (2006) 135. [8] J.H. Loughrin, J. Agric. Food Chem. 54 (2006) 3237. [9] M.E.C. Queiroz, E.B. Oliveira, F. Breton, J. Pawliszyn, J. Chromatogr. A 1174 (2007) 72. [10] B. Lin, M.M. Zheng, S.C. Ng, Y.Q. Feng, Electrophoresis 28 (2007 2771). [11] X.L. Chai, J.P. Jia, T.H. Sun, Y.L. Wang, Chromatographia 67 (2008) 309. [12] D. Panavaite, A. Padarauskas, V. Vickackaite, Anal. Chim. Acta 571 (2006) 45. [13] C. Bicchi, C. Cordero, E. Liberto, B. Sgorbini, P. Rubiolo, J. Chromatogr. A 1152 (2007) 138. [14] J.X. Wang, D.Q. Jiang, Z.Y. Gu, X.P. Yan, J. Chromatogr. A 1137 (2006) 8. [15] R. Aranda, P. Kruus, R.C. Burk, J. Chromatogr. A 888 (2000) 35. [16] H.B. Wan, H. Chi, M.K. Wong, C.Y. Mok, Anal. Chim. Acta 298 (1994) 219. [17] D. Djozan, T. Baheri, R. Farshbaf, S. Azhari, Anal. Chim. Acta 554 (2005) 197. [18] D. Djozan, Y. Assadi, Chromatographia 60 (2004) 313. [19] S.T. Wang, Y. Wang, H. You, J. Yao, Chromatographia 63 (2006) 365. [20] D. Djozan, Y. Assadi, Chromatographia 45 (1997) 183. [21] A. Gierak, M. Seredych, A. Bartnicki, Talanta 69 (2006 1079). [22] T.H. Sun, J.P. Jia, N.H. Fang, Y.L. Wang, Anal. Chim. Acta 530 (2005) 33. [23] C.L. Mangun, Z.R. Yue, J. Economy, Chem. Mater. 13 (2001) 2356. [24] M. Giardina, S.V. Olesik, Anal. Chem. 73 (2001) 5841. [25] C. Tendero, C. Tixier, P. Tristant, J. Desmaison, P. Leprince, Spectrochim. Acta Part B 61 (2006) 2. [26] D.H. Shin, C.U. Bang, J.H. Kim, K.H. Han, Y.C. Hong, H.S. Uhm, D.K. Park, K.H. Kim, Surf. Coat. Technol. 201 (2007) 4939. [27] M. Okubo, M. Tahara, N. Saeki, T. Yamamoto, Thin Solid Films 516 (2008) 6592. [28] L. Xu, H.K. Lee, J. Chromatogr. A 1192 (2008) 203. [29] Y. Kang, Z.C. Wu, Chin. Sci. Bull. 53 (2008) 2248. [30] K.N. Pandiyaraj, V. Selvarajan, J. Mater. Process. Technol. 199 (2008) 130.