MS with Post-Column Switching for Large Volume Injection of

Jan 26, 2011 - introduction of target compounds into the MS ionization chamber. Our HS-GC/MS system enabled injection of as large as 5 mL of HS gas ...
0 downloads 0 Views 941KB Size
Download PDF
TECHNICAL NOTE pubs.acs.org/ac

GC/MS with Post-Column Switching for Large Volume Injection of Headspace Samples: Sensitive Determination of Volatile Organic Compounds in Human Whole Blood and Urine Kanako Watanabe,* Hiroki Fujita, Koutaro Hasegawa, Kunio Gonmori, and Osamu Suzuki Department of Legal Medicine, Hamamatsu University School of Medicine, Hamamatsu, 431-3192, Japan ABSTRACT: When volatile or semivolatile compounds are measured by headspace (HS) gas chromatography (GC)/ mass spectrometry (MS), the maximum gas volume to be injected is usually 0.5-1.0 mL; over the volume, the MS detector automatically shuts down due to impairment of the vacuum rate of the MS ionization chamber. To overcome the problem, we modified the gas flow routes of a new type of GC/MS instrument to create a postcolumn switching system, which can eliminate the large volume of gas before introduction of target compounds into the MS ionization chamber. Our HS-GC/MS system enabled injection of as large as 5 mL of HS gas without any disturbance. As the first example analysis, we tried to establish the analysis of naphthalene and p-dichlorobenzene in human whole blood and urine by this method with large volume injection. The limits of detection for both compounds in whole blood and urine were as low as about 10 and 5 pg/mL, respectively. The validation data and actual measurements were also demonstrated. The new GC/MS system has great potential to analyze any type of volatile or semivolatile organic compounds in biological matrixes with very high sensitivity and full automation.

P

urge-and-trap sample concentration followed by capillary gas chromatography (GC)/mass spectrometry (MS) is probably the most sensitive technique for detection of volatile organic compounds (VOCs) from a relatively large volume of water.1 This dynamic headspace (HS) technique is, however, not suitable for biological samples, such as blood and tissue homogenates, because it causes serious foaming and clogging of the gas flow routes. Another sensitive technique for VOCs is cryofocusing GC/MS;2 however, this may not be suitable for biological matrixes with high protein contents due to the same problems, because no report has been published on the use of cryofocusing GC/MS for analyses of VOCs in biological samples. The most popular method being recently used for VOCs in biological samples is HS-solid-phase microextraction (SPME) connected to GC/MS,3-6 because of its simple procedure for extraction, condensation, and injection into a GC port with an SPME fiber and easiness of automated analysis. However, the disadvantage of the SPME method is low efficiency rates (largely less than 5%) due to the small volumes of stationary phase attached to a thin SPME fiber.7 In our laboratories, we have tried injection of large volumes (5 mL or more) of HS samples into GC port to gain high sensitivities; it was successful, when flame ionization detection or nitrogen-phosphorus detection was used for GC analyses.8,9 However, when the MS detection is used for the large volume r 2011 American Chemical Society

injection GC, the MS detector automatically shuts down due to impaired vacuum degree of the ionization chamber; the maximum volumes admitted to GC/MS injection are 0.5-1.0 mL.10 To enable the large volume injection of HS vapor for GC/MS analysis, we have created a postcolumn switching system by modifying the gas flow routes of a new type of GC/MS instrument; the details of which are described in this report. As target compounds, we analyzed naphthalene and p-dichlorobenzene in human whole blood and urine by the new postcolumn switching GC/MS system, because in October 2008, there were strange incidents in Kanagawa, Japan, in which criminal tainting of instant cup-noodle products with either naphthalene or pdichlorobenzene was suspected and we were in the position to analyze them as forensic toxicologists.10

’ EXPERIMENTAL SECTION Chemicals. Pure crystals of naphthalene and p-dichlorobenzene were purchased from Wako (Osaka, Japan) and Aldrich (Milwukee, WI), respectively; both deuterated internal standards (ISs) naphthalene-d8 and p-dichlorobenzene-d4 were from Received: October 5, 2010 Accepted: December 8, 2010 Published: January 26, 2011 1475

dx.doi.org/10.1021/ac1026258 | Anal. Chem. 2011, 83, 1475–1479

Analytical Chemistry Kanto chemicals (Tokyo, Japan). Other common chemicals used were of the highest purity commercially available. Headspace Method. A 0.5 mL volume of whole blood or urine was mixed with 0.5 mL of distrilled water in a 10 mL HS glass vial (external diameter, 22 mm; height, 45 mm; Agilent, Santa Clara, CA, USA). To the mixture, a 10 μL aliquot of a methanol solution containing 200 ng each of both ISs was added and mixed briefly. When validation experiments with blank whole blood or urine were performed, a 10 μL aliquot of methanol solution containing a desired concentration of naphthalene and p-dichlorobenzene was also added to the vial together with the 10 μL of the above IS solution. Immediately after the additions, the HS vial was sealed with a 22 mm HS silicone polytetrafluoroethylene septum cap (Agilent) and put on a tray of an autosampler (G1888 Network Headspace Sampler; Agilent.). The headspace vapor was brought to a sampling loop (1 or 5 mL) with heating at 150 °C and then introduced into an injector of the following GC instrument via a transfer line heated at 200 °C. Modification of Gas Routes of the GC Instrument. The GC instrument to be connected to MS was an Agilent 7890 A system, which enabled “two dimensional GC” analysis with two different columns by their “capillary flow technology”. Both columns could be connected by a Deans switch. As the first modification of the system, we closed one of the openings of the Deans switch. The second modification was the separation of the solenoid valve from the pneumatic control module; this modification enables us to abandon the HS vapor after trapping the target compounds in the GC column by opening the solenoid valve. GC/MS Conditions. The above 7890A GC system connected with an MS detector (5975C inert MSD with Triple-Axis Detector; Agilent) and with the autosampler was used. The whole system was controlled by G1701 ChemStation (Agilent) to realize the full automatic analysis. The GC conditions were as follows: column, Rtx-1 fused silica capillary column (30 m  0.32 mm i.d., film thickness 0.25 μm; Bellefonte, PA); column temperature, 60 °C (1 min hold) to 120 °C at 10 °C/min and 120 to 280 °C at 60 °C/min; injector temperature, 250 °C; helium flow rate, 1.2 mL/min; injection mode at the injector, pulsed splitless at the pressure of 50 psi for 1 min ; solenoid valve, open upon injection and closed 0.2 s after the injection. The MS conditions were as follows: ionization mode, positive electron ionization (EI); ion source temperature, 250 °C ; ionization energy, 70 eV; emission current, 34.6 μA ; acceleration voltage, 3.0 kV; interface (restrictor) prior to the MS detector, deactivated fused-silica capillary (27 cm  0.1 mm i.d.) maintained at 200 °C; full-scan range, m/z 40-350; quantitation, selected ion monitoring (SIM) mode with peak area measurements; SIM ions used, m/z 128, 136, 146, and 152 for naphthalene, naphthalened8, p-dichlorobenzene, and p-dichlorobenzene-d4, respectively.10 Human Experiments. The human experiments were approved by the Institutional Review Board of Hamamatsu University School of Medicine. The chief scientist (male, 64-yearold, medical doctor) of this research project volunteered to undertake the experiments. He was healthy and well aware of the toxicities of naphthalene and p-dichlorobenzene and that the experiments were not hazardous to his health. Five grams of naphthalene or p-dichlorobenzene crystals were put at the bottom of a plastic bag (size: 52  60 cm). A sufficient amount of air was also introduced into the bag, and the bag was closed and shaken vigorously for 1 min to evaporate the crystals into the air sufficiently. The subject inhaled the naphthalene-containing air in the bag for 10 min. Immediately after his inhalation, several

TECHNICAL NOTE

Figure 1. Simplified schematic illustrations of gas flow routes for postcolumn switching GC/MS system created for enabling injection of large volume headspace gas samples. (A) Gas flow routes upon injection; (B) those upon introduction of target compounds into the mass detector.

milliliters of blood were collected from his vein; a small amount of blood was also taken 10 min later. His urine was also obtained 1 h after the end of his inhalation. For the experiments with pdichlorobenzene, all conditions for the inhalation and samplings of his blood and urine were the same as those with the above naphthalene, except that the inhalation duration was 5 min.

’ RESULTS AND DISCUSSION Automated Postcolumn Gas Flow Switching. Using the new type of GC instrument (Agilent 7890A), we modified the gas flow routes for enabling large volume injection as shown in Figure 1. A 5 mL volume of HS vapor was stored in the sampling loop and injected into GC inlet in the pulsed splitless mode; the target compounds (naphthalene, p-dichlorobenzene, and their ISs) were trapped in the stationary phase of the inlet end of a capillary GC column under cryogenic or not cryogenic oven temperature. The carrying gas (5 mL) passed through the column and was abandoned via the open solenoid valve. For naphthalene and p-dichlorobenzene, it was not necessary to use the cryogenic oven system; both compounds and their deuterated ISs were entirely trapped at 60 °C of column temperature10 1476

dx.doi.org/10.1021/ac1026258 |Anal. Chem. 2011, 83, 1475–1479

Analytical Chemistry

TECHNICAL NOTE

Figure 2. Relative abundance of peaks for naphthalene and p-dichlorobenzene spiked into human whole blood (WB) and urine as a function of the headspace gas sample volumes (1 and 5 mL).

(Figure 1A). Then, the solenoid valve was closed, and the column temperature was increased, which resulted in the transportation of the target compounds into the MS detector via the restrictor (Figure 1B). The whole procedure was automatically handled and controlled by the G1701 ChemStation system. Effect of HS Vapor Volumes. In our previous report10 on GC/MS analysis of naphthalene and p-dichlorobenzene in whole blood and urine, we used 1 mL of HS vapor for injection in the splitless mode. One milliliter of the HS vapor was the maximum limit volume, which did not result in the disturbance of the MS detection. When the volumes of HS vapor larger than 1 mL were introduced into the ionization chamber of the MS detector, it resulted in decreased sensitivity or shutdown by impairment of the MS function. Thus, we compared the results obtained from 5 mL of HS vapor with those obtained from 1 mL of HS vapor for naphthalene and p-dichlorobenzene spiked into whole blood and urine using the presently established method as shown in Figure 2. The 5 mL HS vapor samples at both concentrations (500 and 1000 pg/mL) in whole blood and urine gave abundances 5 times higher or even more than the 1 mL HS vapor samples. These results show that the large volume injection in the splitless mode has been realized by this postcolumn switching system successfully. Reliability of the Method. Figure 3 shows SIM chromatograms for naphthalene and p-dichlorobenzene spiked into human

whole blood and urine at the concentration of 100 pg/mL for both compounds; the concentrations of both ISs were 200 ng/mL for both whole blood and urine. The retention times of the peaks of naphthalene and p-dichlorobenzene were 8.90 and 7.35 min, respectively. There were almost no impurity peaks interfering with the test peaks. Table 1 shows regression equations for naphthalene or pdichlorobenzene spiked into human whole blood or urine. For whole blood, we constructed the calibration curves for both low and high concentration ranges. All of them showed good linearity with correlation coefficients not lower than 0.99. The detection limits (signal-to-noise ratio g3) for both naphthalene and p-dichlorobenzene were about 10 pg/mL for whole blood and 5 pg/mL for urine. The quantitation limits (signal-to-noise ratio = 10) for both compounds were about 20 pg/mL for whole blood and 10 pg/mL for urine. The efficiency, accuracy, and precision data are shown in Table 2. The data were obtained at 200, 500, and 1000 pg/mL for each compounds and matrix. The efficiency values, as a function of the shift rate from aqueous phase (1.0 mL) to the 5 mL of HS vapor used, were 7.94-38.1%; the efficiency rates were higher for p-dichlorobenzene than for naphthalene. The intra- and interday accuracies were 82.0-118%, and intra-and interday coefficients of variation were not greater than 11.7%, showing that the reproducibility of this method was generally satisfactory. 1477

dx.doi.org/10.1021/ac1026258 |Anal. Chem. 2011, 83, 1475–1479

Analytical Chemistry

TECHNICAL NOTE

Table 1. Regression Equations for Naphthalene or p-Dichlorobenzene Spiked into Human Whole Blood or Urine compound naphthalene

p-dichlorobenzene

range matrixa (pg/mL) WB WB urine WB WB urine

20-1500b 1000100000c 10-1500d 20-1500b 1000100000c 10-1500d

equation

correlation coefficient (r)

y = 0.00000208x þ 0.000880 y = 0.00000243x þ 0.00126

0.997 0.990

y = 0.00000239x þ 0.000354 y = 0.00000379x þ 0.00163 y = 0.00000371x þ 0.00730

0.999 0.998 0.998

y = 0.00000513x þ 0.000665

0.999

a

WB: human whole blood. b Nine plots with different concentrations. c Seven plots with different concentrations. d Ten plots with different concentrations. Table 2. Validation Data for Headspace Gas Chromatography/Mass Spectrometry (GC/MS) Analysis of Naphthalene or p-Dichlorobenzene Spiked into Human Whole Blood or Urine intraday (n = 5)

interday (n = 5)

efficiency concentration (n = 10; accuracy precision accuracy precision (%) (%CV) (%) (%CV) compound matrix (pg/mL) % ( SD)a 200 11.6 ( 1.64 99.6 7.11 113 9.24 500 11.7 ( 1.47 108 4.23 98.2 3.72 1000 7.94 ( 1.18 82.0 2.02 107 3.83 urine 200 14.0 ( 2.94 94.1 4.97 102 9.61 500 12.4 ( 4.17 111 3.65 110 4.68 1000 10.8 ( 2.90 107 5.22 99.4 4.00 p-dichloro- WB 200 38.1 ( 4.48 109 5.40 101 9.60 benzene 500 31.8 ( 4.65 108 4.17 107 6.17 1000 26.2 ( 3.56 101 1.28 98.8 6.63 urine 200 22.7 ( 5.76 82.6 5.16 117 11.7 500 17.5 ( 4.98 101 3.08 104 5.92 1000 18.0 ( 4.04 118 5.78 92.2 8.11 a Data expressed as mean ( standard deviation (SD); they were calculated by comparing peak areas obtained from the headspace gas of the spiked whole blood or urine samples with those obtained from nonextracted methanolic solution of each compound directly injected into GC/MS. The volume of the headspace vapor was taken as 5 mL out of 10 mL headspace. The values show the total shift rates of each compound from the aqueous phase (1.0 mL) to the 5 mL headspace vapor used. naphthalene WB

Figure 3. Selected ion monitoring (SIM) for naphthalene, p-dichlorobenzene, and their internal standards (ISs) spiked into whole blood (A) or urine (B). The concentrations of both compounds spiked were 100 pg/mL, and those of both ISs were 200 ng/mL.

The most prominent advantage of the HS extraction is that it gives very low background and few impurity peaks with a very simple procedure. The cleanliness of the HS samples seems to be good for less pollution of the ionization chamber of an MS instrument. However, the recovery rates are generally low, especially for compounds present in various matrixes containing large amounts of lipids and/or proteins, such as whole blood and tissue homogenates.6,11 This is probably the reason why most of the previous studies12-14 on the analysis of polycyclic aromatic hydrocarbons or VOCs including naphthalene and/or p-dichlorobenzene did not deal with whole blood samples. However, in postmortem forensic toxicological analysis, hemolyzed whole

blood should be frequently analyzed. The previous forensic toxicological report dealing with the GC/MS analysis of dichlorobenzene isomers in human whole blood employed HS extraction followed by SPME.15 According to their results, the limit of quantitation was 20 ng/mL of whole blood; that of our conventional HS-GC/MS for naphthalene and p-dichlorobenzene for the whole blood method was about 2.5 ng/mL10. In the present method, the limit of quantitation for both compounds was as low as 20 pg/mL of whole blood; the high sensitivity is probably due to large volume injection in the splitless mode giving much higher efficiency rates (Table 2) and also due to higher sensitivity of the new type of MS detector used. Application of the Present Method. To ascertain the applicability of this method to real human samples, a 64-yearold volunteer inhaled naphthalene or p-dichlorobenzene-containing air for 10 and 5 min, respectively on different days. The blood samples were collected immediately after and 10 min after the end of the exposure. As shown in Figure 4, intense peaks of both compounds could be detected from his blood; the urine samples obtained 1 h after inhalation gave too small a peak and also interfered with by impurities for naphthalene but gave a 1478

dx.doi.org/10.1021/ac1026258 |Anal. Chem. 2011, 83, 1475–1479

Analytical Chemistry

TECHNICAL NOTE

The system enabled injection of an HS gas volume as large as 5 mL into a GC/MS instrument, and we could present a highly sensitive method for analysis of naphthalene and p-dichlorobenzene in human whole blood and urine. To our knowledge, such a postcolumn switching system enabling large volume injection of gaseous samples for GC/MS has not been reported. All of the procedure was fully automated and controlled by a workstation; neither manual extraction of target compounds nor injection is required. In this study, we did not use the cryogenic oven system,8 because naphthalene and p-dichlorobenzene could be completely trapped by the capillary GC column at 60 °C. Nowadays, every new type of GC instrument is equipped with the cryogenic oven system, which enables the trapping of highly volatile organics of small molecules by a GC column. Therefore, the present new postcolumn switching GC/MS system has great potential to analyze any type of volatile and semivolatile organic compound present in biological matrixes with very high sensitivity and full automation.

’ AUTHOR INFORMATION Corresponding Author

*Address: 1-20-1 Handayama, Higashi-ku, Hamamatsu 431-3192, Japan. Tel: þ81-53-435-2239. E-mail: [email protected].

’ REFERENCES

Figure 4. SIM for naphthalene, p-dichlorobenzene, and their ISs extracted from whole blood and urine of a volunteer, who inhaled naphthalene or p-dichlorobenzene-containing air for 10 and 5 min, respectively. (A) SIM chromatograms for WB sampled immediately after the exposure; (B) those for urine sampled 1 h after the exposure. N. D.: not detectable.

clean peak for p-dichlorobenzene. The whole blood levels of naphthalene at 0 and 10 min after exposure were thus calculated to be 12.5 and 3.84 ng/mL, respectively; those of p-dichlorobenzene were 99.2 and 43.4 ng/mL, respectively. Naphthalene in urine 1 h after exposure was not detectable (Figure 4), but the pdichlorobenzene level in urine was calculated to be 390 pg/mL.

(1) Zoccolillo, L.; Amendola, L.; Cafaro, C.; Insogna, S. J. Chromatogr., A 2005, 1077, 181–187. (2) Chen, T.-C.; Her, G.-R. J. Chromatogr., A 2001, 927, 229–235. (3) Waidyanatha, S.; Zheng, Y.; Rappaport, S. M. Chem. Biol. Interact. 2003, 145, 165–174. (4) Campo, L.; Addario, L.; Buratti, M.; Scibetta, L.; Longhi, O.; Valla, C.; Cirla, P. E.; Martinotti, I.; Foa., V.; Fustinoni, S. Toxicol. Lett. 2006, 162, 132–138. (5) Campo, L.; Mercadante, R.; Rossella, F.; Fustinoni, S. Anal. Chim. Acta 2009, 631, 196–205. (6) Hara, K.; Kashiwagi, M.; Kageura, M.; Matsusue, A.; Kubo, S. Forensic Toxicol. 2009, 27, 52–53. (7) Kumazawa, T.; Lee, X. P.; Sato, K.; Suzuki, O. In Applications of Solid Phase Microextraction; Powliszyn, J., Ed.; The Royal Society of Chemistry: Cambridge, U.K., 1999; pp 470-485. (8) Watanabe-Suzuki, K.; Ishii, A.; Suzuki, O. Anal. Bioanal. Chem. 2002, 373, 75–80. (9) Watanabe, K.; Hasegawa, K.; Yamagishi, I.; Nozawa, H.; Suzuki, O. Forensic Toxicol. 2009, 27, 98–102. (10) Watanabe, K.; Hasegawa, K.; Yamagishi, I.; Nozawa, H.; Takaba, M.; Suzuki, O. Anal. Sci. 2009, 25, 1301–1305. (11) Wise, J.; Danielson, T.; Mozayani, A.; Li, R. Forensic Toxicol. 2008, 26, 66–70. (12) Hino, T.; Nakanishi, S.; Hobo, T. J. Chromatogr., A 1996, 746, 83–90. (13) Sakata, S. K.; Taniguchi, S.; Rodriguez, D. F.; Urano, M. E.; Wanderm€uren, M. N.; Pellizari, V. H.; Comasseto, J. V. J.Chromatogr., A 2004, 1048, 67–71. (14) Safarova, V. I.; Sapelnikova, S. V.; Djazhenko, E. V.; Teplova, G. I.; Shajdulina, G. F.; Kudasheva, F. K. J. Chromatogr., B 2004, 800, 325–330. (15) Liu, J.; Hara, K.; Kashimura, S.; Hamanaka, T.; Tomojiri, S.; Tanaka, K. J. Chromatogr., B 1999, 731, 217–221.

’ CONCLUSIONS In this study, we modified the gas flow routes of a new type of GC/MS instrument to create a postcolumn switching system. 1479

dx.doi.org/10.1021/ac1026258 |Anal. Chem. 2011, 83, 1475–1479