The volatile organic compounds (VOCs) as important air pollutants have attracted more and more concerns in the world. Some studies presented that the chemical composition of VOCs emissions varied with different fuels, industries and living regions [1, 2]. Liquefied petroleum gas (LPG) is one of the commonly used fuel sources for heating appliances, vehicles and even cooking purposes. There are some organic residues especially aromatics in LPG besides the main component of propane [3-7]. The aromatic residues will produce more toxic sub-aromatics after high temperature firing and cause secondary environmental pollution. Some investigation results have been reported on the impacts of LPG on air pollution and the obtained results were anxious [8-10]. To monitor organic residue levels in LPG and guide the refining improvement to reduce secondary pollution is important. Up to now, few references are available for directly and accurately monitoring aromatics in LPG [11]. The American ASTM D2158-05 standard detection method of organic residues in LPG adopted solvent adsorption for sampling [12].
For monitoring trace aromatics from complex co-existing matrices in LPG, efficient sample enrichment procedure is required before chromatography or GC-MS analysis. Solid phase extraction (SPE) is a simple, convenient and cost-effective technique, which is widely used for liquid sample extraction and preconcentration. For solid and airborne particulate matrix, analytes always need to be devolved into solution firstly before SPE extraction [13-16]. SPE has been rarely used to extract analytes from gas samples directly [17-19], and the absorbents were essential for gas sampling [20-22]. Graphitized carbon black (GCB), produced by heating carbon black to 2 700-3 000 ℃ in an inert atmosphere, is a good carbon-based SPE sorbents [23-25]. The hexagonal structure of graphite surface makes it show a selective adsorption to aromatic compounds. Compared with other sorbent materials, two advantages of GCB adsorbents are outstanding for adsorption of aromatics from gas. Firstly, the excellent Van der Waals adhesion and non-covalent π -π stacking interactions improve adsorption capacity for both non-polar and weakly polar organic compounds bearing aromatic moieties [26]. Secondly, the dry GCB sorbent still presents excellent adsorption capability which is very important for direct gas sampling [17].
In this work, a GCB sorbent-based flow-through sampling method for preconcentration of trace aromatics from LPG is proposed. The collected aromatics were rapidly characterized and quantified by gas chromatography-mass spectrometry (GC-MS) analysis. The breakthrough volume, extraction ability and reprodu-cibility of the GCB sorbent for extraction of aromatics were examined. The accurate concentrations of aromatics given by this GCB method have significance for improvement industrial processes. The assembled GCB sampling cartridges with caps and stoppers were designed for convenient long-time storage and transportation. The developed flow-through GCB-based sorption technique has the potential to be applied to the analysis of other fuel gases.
Seven standard substances including benzene (99.5%), toluene (99.5%), o-xylene (99.8%), m-xylene (99.2%), p-xylene (99.0%), styrene (99.7%) and naphthalene (99.8%) dissolved in methanol were obtained from AccuStandard (USA). The mass concentration of each compound in the stock mixture solution was 2 000 mg/L. High-purity indene (98%) was also purchased from AccuStandard and used for chromatographic analysis directly. The standard solutions were stored at -20 ℃.
All organic solvents including methanol, dichloromethane, and n-hexane were of HPLC grade and obtained from J. T. Baker (USA). The graphitized carbon black sorbents (40-60 μ m in size) and C18 cartridges (250 mg, 6 mL) were all supplied by Agela Technologies (China). Poly (styrene-divinylbenzene) copolymer (PS-DVB) disks (47 mm diameter, ca. 0.50 mm thickness) were obtained from Empore (USA).
The LPG flow-through sampling setup is illustrated in Fig. 1. The two packed GCB cartridges were connected in series, with the 1# cartridge for sampling and the 2# cartridge for monitoring gas breakthrough. A glass buffer bottle was connected between the GCB cartridge and LPG tank to obtain a stable flow path. The GCB cartridges were prepared by packing 250 mg amount of GCB sorbents into empty polytetrafluoroethylene (PTFE) cartridges with holding spacer. Before sampling, the packed GCB cartridges were conditioned twice by 5 mL dichloromethane to remove organic contaminants. Then flow-through sampling was carried out in a passive mode and the flow rate was regulated by a flow meter. The total volume of 4 L LPG was sampled at 200 mL/min flow rate.
After sampling, the double ends of the GCB cartridges were sealed with custom-built stoppers and caps (silicone rubber) and the cartridges were stored in a refrigerator. Just before GC-MS analysis, the sampled GCB cartridges were eluted by 2 mL dichloromethane for analysis. A control GCB cartridge without sampling was used as blank, which was handled and prepared similarly to the sampling GCB cartridge.
The PS-DVB disk was placed on a sampler supported by a stainless steel net and two teflon rings. After sampling, the disk was transferred to a 10 mL centrifuge tube and 5 mL dichlorome-thane was added. Then the centrifuge tube was sealed, sonicated and centrifuged to remove particulate matters from the upper supernatant. The operations of C18 cartridge sampling and pretreatment were the same to the process of GCB cartridge extraction.
The aromatics determination was carried out by using the GC/MS QP 2010 instrument (Shimadzu, Japan). An RTX-50 fused silica capillary column (30 m×0.25 mm i. d. with 0.25 μ m thickness coating, Restek Corporation, USA) was used for GC separation. The optimized separation program was performed by starting at 40 ℃ for 3 min, increasing to 220 ℃ at 10 ℃/min speed and keeping for 2 min. The carrier gas was helium (99.999%) at 1.0 mL/min flow rate. The sample (2.0 μ L) was injected at the split ratio of 1 : 10. Detection of the targets was performed by the electron impact ionization (EI) with selected ion monitoring (SIM) mode.
Each aromatic standard (20.0 ng) was spiked to the blank GCB cartridge (18 cartridges in total). The aromatics collected in the three parallel cartridges were analyzed immediately after spiking as the control value at time zero. The remaining sampled cartridges were covered with assembling stoppers and caps, and stored at 4 ℃ until analysis. Every three samplers as one group were analyzed periodically in 1, 2, 3, 5 and 7 d.
Total ion chromatogram (TIC) of the standard solution containing eight aromatics is shown in Fig. 2. The separation was completed within 15 min. For the overlapping m-xylene and p-xylene, detection and quantification were not compromised by overlapping retention times since their compound-specific target and reference ions produced clearly consistent signals in the MS [27] as shown in Fig. 2.
The linearity for the eight aromatics is calculated by five mass concentration levels of standards covering the range of 0.03-2.0 mg/L, which were equivalent to the mass concentration of aromatics from 15 to 1 000 μ g/m3 in 4 L LPG. Good quantitative linearity of individual aromatic target could be obtained with correlation coefficients (r2) varying from 0.993 4 to 0.999 5.
In this work, GCB, C18 cartridges and the PS-DVB disks were tested for aromatics sampling. All the experiments were performed by spiking 20 μ L standard aromatics at 1.0 mg/L mass concentration into 4 L high-purity nitrogen. The adsorption efficiency was evaluated with the ratio of measured adsorption value of aromatics on sorbents and the spiked amount, i. e., adsorption recovery. The average recoveries of C18 cartridge and PS-DVB disk were all below 80% as shown in Fig. 3. The strong adsorption ability of PS-DVB to aromatics because that the π -π interaction makes it difficult to release the targets and restricts the extraction recoveries. The good recoveries of GCB sorbents demonstrated the proper ability to selectively adsorb and release aromatics during the sampling and eluting procedures [28].
It is well known that benzene solvent class, chlorinated, alcohols and hexane are good elution solvents for hydrophobic compounds in sample pretreatment. In this experiment, benzene solvent class was out of consideration since the preconcentration targets were aromatics. Dichlorome-thane is lower toxic than the three carbon tetrachloride and carbon tetrachloride. So dichloromethane, methanol and n-hexane were tried as elution solvents due to their low toxicity and different physical properties. The impacts on the adsorption recoveries are given in Fig. 4. It is observed that the best recoveries could be obtained with dichloromethane as eluent.
As to gas sampling, complete retention during sampling (no breakthrough or back-diffusion) is essential. The breakthrough experiment could be performed by purging the standard aromatics from solution in custom-built vials with nitrogen gas flow [29, 30]. An amount of 20 μ L of mixed aromatics solution containing 20.0 ng of each aromatic hydrocarbon was spiked into a vial which was designed with a low-level inlet and a high-level outlet for gas flow in and out. Two blank GCB cartridges were connected for sampling and monitoring breakthrough, respectively. The vial was connected between the nitrogen container and the sampling GCB cartridges. When nitrogen gas passed through the vial, the mixed aromatics were evaporated into the gas flow and adsorbed on the GCB cartridges.
In order to allow a slow and continuous releasing of aromatics, the inlet and outlet of the vial should be as small as possible. The spiked aromatics solution was totally evaporated with nitrogen purging for 20 min at a flow rate of 200 mL/min. The adsorbed extract in 2# monitoring cartridge (MS2) and in 1# sampling cartridge (MS1) was analyzed separately. It was reported that the 300 mg carbograph packed tube presented a breakthrough volume of 16 L/g for benzene [17].
Similarly, to correct the possible presence of the target aromatics in the purging gas, a same nitrogen purging program was taken without spiking aromatics in the vial. The eluent of the cartridge was also detected as the background value Mblank. The retention efficiency of sampling aromatics on 1# sampling cartridge was calculated by the following Equation (1).
where MS1 and Mblank are the means (n=3) of the detected aromatic concentrations, and Mspiked is the spiked concentration, respectively. As can be seen in Table 1, the retention efficiencies of the aromatics on the sampling GCB cartridge ranged from 96.4% to 102.1%. The breakthrough was estimated by the leakage amount of aromatics on the 2# monitoring cartridge. It was found that the spiked amount of 20.0 ng of each aromatic was not up to the breakthrough volume.
The retaining ability of extraction sorbents on adsorbents over long-time transportation is an important factor, because sampling sites are always far away from the analytical laboratory. Since the aromatics are volatile, the storage stability of GCB sampling cartridge needs to be testified. The GCB cartridges loaded with aromatics were wrapped with assembled stoppers and caps, then stored at 4 ℃. All of the aromatics had no significant loss over 7 days storage as listed in Table 2. Therefore, GCB sampling and storage is promising for remote sites.
Method detection limits (MDLs) were calculated as three times the standard deviation determined from three repetitive runs of the lowest aromatic concentration by nitrogen purging experiments. The MDL values were given by conversion the spiked aromatics amounts to the gas concentration in 4 L sampling volume. As in Table 3, the values range from 1.0 to 6.2 μ g/m3, and the MDL of toluene is the lowest.
The precision and accuracy of the flow-through sampling method were valuated by purging experiments, which were similar to the purging operation in Section 2.3 except that nitrogen gas was replaced by LPG. The mixed aromatics solution containing 20.0 ng amount of each of aromatic hydrocarbon was spiked into the vial and was purged for 20 min at 200 mL/min LPG flow rate. Accordingly, another control GCB cartridge was directly purged by LPG. The precision was evaluated by RSDs with replicate assays, and the accuracy was evaluated by the relative errors (REs) of the assayed samples to their spiked concentrations. In addition, the precision and accuracy for lower concentration of aromatics were also estimated by purging 5 μ L of the mixed aromatics stock solution containing 5.0 ng amount of each aromatic hydrocarbon. The precision and accuracy of GCB sampling method are listed in Table 3. For all the aromatics, the RSDs ranged from 0.6% to 5.8%, and the REs ranged from 0.8% to 8.2%. It proved that GCB sorbents were suitable for sampling aromatics from LPG.
Finally, the proposed dynamic GCB-based sampling method was applied to preconcentration of trace aromatics in LPG. A total volume of 4 L LPG was sampled at 200 mL/min flow rate at room temperature. The extract was eluted from GCB cartridge by 2 mL dichloromethane. As shown in Fig. 5, the aromatics (benzene, toluene, o/m/p-xylene, styrene and naphthalene) except for indene are observed in the LPG. The aromatics in LPG are just at trace levels ranging from 16 to 868 ng/m3 in Table 3.
In this paper, a GCB-based sampling system for preconcentration of trace aromatics from LPG has been demonstrated. The GCB-packed cartridge presented practical utility for sampling the aromatics in flow gas with high efficiency and good reproducibility. The aromatics at nanogram levels in LPG were found. Furthermore, the outstanding storage stability of GCB benefits remote sampling. This solid sorbents based direct adsorption strategy provides a feasible way to develop rapid and simple fuel gas sampling methodologies.
2017,
Vol. 35 Issue (1): 47-53
PDF
2017,
Vol. 35 Issue (1): 47-53
PDF