Laser Desorption for

Jan 19, 2010 - since the sample is introduced as a gas packet into the MS and the laser pulse ... laser desorption (online COLD).21 A capillary with a...
0 downloads 0 Views 250KB Size
Download PDF
Anal. Chem. 2010, 82, 1283–1287

Online Concentration by Analyte Adsorption/Laser Desorption for Application to Gas Chromatography/ Resonance-Enhanced Multiphoton Ionization Time-of-Flight Mass Spectrometry Yuji Sakoda,† Tomohiro Uchimura,*,†,‡ and Totaro Imasaka†,‡ Department of Applied Chemistry, Graduate School of Engineering, and Division of Translational Research, Center for Future Chemistry, Kyushu University, 744, Motooka, Nishi-ku, Fukuoka 819-0395, Japan A novel sample introduction technique, based on online concentration by analyte adsorption/laser desorption, was applied to resonance-enhanced multiphoton ionization time-of-flight mass spectrometry (REMPI-TOFMS). Signal enhancement and optical selectivity, based on supersonic jet spectrometry, were examined by measuring the REMPI spectra. This sample introduction technique was applied at the interface of a gas chromatograph (GC) and a mass spectrometer (MS). The signal intensity was enhanced ca. 80-fold, compared with that of a conventional technique based on continuous sample introduction. The analyte molecule was selectively measured using a tunable laser emitting at a wavelength of the 0-0 transition. With the use of a desorption laser, a linear calibration curve was extended over 4 orders of magnitude. The sensitivity, selectivity, and dynamic range were all superior for this sample introduction system, as were the inherent characteristics, e.g., a near-zero dead volume and thermal durability. This system would provide a useful means for practical trace analysis of aromatic hydrocarbons by GC/REMPI-TOFMS. Gas chromatography/resonance-enhanced multiphoton ionization time-of-flight mass spectrometry (GC/REMPI-TOFMS)1-4 has been developed as a sensitive and selective analytical means for the measurement of aromatic hydrocarbons including environmental pollutants such as polychlorinated dibenzo-p-dioxin/ dibenzofurans (PCDD/Fs)5,6 and polychlorinated biphenyls * To whom correspondence should be addressed. E-mail: uchimura@ cstf.kyushu-u.ac.jp. † Graduate School of Engineering. ‡ Center for Future Chemistry. (1) Rhodes, G.; Opsal, R. B.; Meek, J. T.; Reilly, J. P. Anal. Chem. 1983, 55, 280–286. (2) Dobson, R. L. M.; D’Silva, A. P.; Weeks, S. J.; Fassel, V. A. Anal. Chem. 1986, 58, 2129–2137. (3) Wilkerson, C. W., Jr.; Colby, S. M.; Reilly, J. P. Anal. Chem. 1989, 61, 2669–2673. (4) Zimmermann, R.; Rohwer, E. R.; Heger, H. J. Anal. Chem. 1999, 71, 4148– 4153. (5) Yamaguchi, S.; Kira, F.; Miyoshi, Y.; Uchimura, T.; Watanabe-Ezoe, Y.; Zaitsu, S.; Imasaka, T.; Imasaka, T. Anal. Chim. Acta 2009, 632, 229–233. (6) Imasaka, T.; Nakamura, N.; Sakoda, Y.; Yamaguchi, S.; Watanabe-Ezoe, Y.; Uchimura, T.; Imasaka, T. Analyst 2009, 134, 712–718. 10.1021/ac902273b  2010 American Chemical Society Published on Web 01/19/2010

(PCBs).7 Superior selectivity can be provided by a combination with supersonic jet spectrometry,8-14 in which high spectral resolution is derived by jet expansion and subsequent molecular cooling. In REMPI-TOFMS, there are two approaches for sample introduction: continuous and pulsed sample. A capillary with a restricted tip has been utilized for the former.15 Such a sample inlet system is suitable for use at the interface of the GC and MS, since it has a near-zero dead volume and can be heated to high temperature. However, the duty cycle, which is defined as the ratio of the amount of the sample probed by an ionization laser to the amount of the sample supplied, is very low because of the pulsed operation of the ionization laser. On the other hand, the latter sample introduction technique provides a higher duty cycle, since the sample is introduced as a gas packet into the MS and the laser pulse is synchronously fired, thus allowing superior sensitivity. This pulsed sample introduction technique uses a mechanically driven pulsed nozzle that cannot be effectively applied at the interface of the GC and MS because of a large dead volume and limitations in the maximum operation temperature, although several studies have reported the employment of a pulsed nozzle for this purpose.16-20 We have recently reported a versatile sample introduction technique based on online concentration by analyte adsorption/ laser desorption (online COLD).21 A capillary with a restricted (7) Imasaka, T.; Sakai, K.; Imasaka, T. Anal. Chem. 2004, 76, 5534–5538. (8) Smally, R. E.; Wharton, L.; Levy, D. H. Acc. Chem. Res. 1977, 10, 139– 145. (9) McClelland, G. M.; Saenger, K. L.; Valentini, J. J.; Herschbach, D. R. J. Phys. Chem. 1979, 83, 947–959. (10) Amirav, A.; Even, U.; Jortner, J. Chem. Phys. 1980, 51, 31–42. (11) Hayes, J. M.; Small, G. J. Anal. Chem. 1983, 55, 565A–574A. (12) Lubman, D. M. Anal. Chem. 1987, 59, 31A–40A. (13) Syage, J. A. Anal. Chem. 1990, 62, 505A–509A. (14) Imasaka, T.; Moore, D. S.; Vo-Dinh, T. Pure Appl. Chem. 2003, 75, 975– 998. (15) Hafner, K.; Zimmermann, R.; Rohwer, E. R.; Dorfner, R.; Kettrup, A. Anal. Chem. 2001, 73, 4171–4180. (16) Pepich, B. V.; Callis, J. B.; Burns, D. H.; Gouterman, M.; Kalman, D. Anal. Chem. 1986, 58, 2825–2830. (17) Imasaka, T.; Tashiro, K.; Ishibashi, N. Anal. Chem. 1986, 58, 3242–3244. (18) Ko ¨ster, C.; Grotemeyer, J.; Schlag, E. W. Z. Naturforsch. 1990, 45a, 1285– 1292. (19) Zimmermann, R.; Lermer, Ch.; Schramm, K. W.; Kettrup, A.; Boesl, U. Eur. Mass Spectrom. 1995, 1, 341–351. (20) Heger, H. J.; Dorfner, R.; Zimmermann, R.; Rohwer, E. R.; Boesl, U.; Kettrup, A. J. High Resolut. Chromatgr. 1999, 22, 391–394. (21) Uchimura, T.; Sakoda, Y.; Imasaka, T. Anal. Chem. 2008, 80, 3798–3802.

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1283

Figure 1. Experimental apparatus of GC/REMPI-TOFMS combined with an online COLD sample introduction technique (a) and configuration of the ion source (b).

tip was used for sample introduction into a vacuum to form a supersonic jet. The analyte was found to be adsorbed at the tip of the capillary by the cooling effect arising from adiabatic expansion of a carrier gas, and a pulsed laser was utilized for desorption of the analyte to inject it into a vacuum as a packet. The analyte molecule was concentrated in the jet and, then, ionized by the other laser. The duration of the sample pulse passing through an ionization region can be reduced to 300 °C. The enhancement factor, i.e., the relative signal increase by the use of a desorption laser, and the dynamic range of the analytical curve were examined for practical trace analysis. EXPERIMENTAL SECTION Figure 1 shows the experimental apparatus developed in this study. The operation principle for the online COLD technique has been described elsewhere.21 A sample of p-chlorophenol (KANTO Chemical, Tokyo), which is one of the precursors of PCDD/Fs, was employed. The second-harmonic emission of a Nd:YAG laser (Minilite 1, Continuum, 532 nm, 5 ns, 10 Hz, Excel Technology, Tokyo) was used as a desorption laser, and the second-harmonic (22) Uchimura, T. Bunseki Kagaku 2009, 58, 119–126.

1284

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

emission of a dye laser (Sirah dye laser, 5 ns, 10 Hz, Spectra Physics) was used as an ionization laser. Rhodamine 590 (Exciton, Ohio) dissolved in ethanol was utilized as a laser dye. In order to measure the mass and REMPI spectrum, the analyte was mixed with helium and was stored in a sampling bag (Tedlar bag, 2 L, Omi Odor-Air Service, Shiga). The gas was introduced into a mass spectrometer, and the flow rate was typically adjusted to ca. 5 mL/min. A 1 m deactivated fused-silica capillary (0.32 mm i.d., GL Sciences, Tokyo), the tip of which was restricted to ca. 50 µm, was used. The position of the tip was adjusted to about 1 mm above the surface of the repeller electrode and was supported by a metal jacket, as shown in Figure 1b. The desorption laser was used with no focus of the beam, and the power density was adjusted to 1 MW/cm2. After 8-12 µs following laser desorption, the ionization laser was focused into the sample in a packet by a planoconvex lens with a focal length of 300 mm. The timing between the two lasers was adjusted using a delay/pulse generator (DG535, Stanford Research Systems, California). The time duration for the sample passing through the ionization region was typically 6.5 µs. The energy of the ionization laser was attenuated to 80-300 µJ, and the laser wavelength was changed from 284 to 288 nm at a scan rate of 0.01 nm/s. The induced ion was detected by means of an assembly of microchannel plates (F4655-11, Hamamatsu, Shizuoka) after passing through a linear-type time-of-flight tube (ca. 60 cm, homemade). The mass spectrum was recorded using a digital oscilloscope (TDS5104, Tektronix Japan, Tokyo). From the data stored in a computer, the mass-selected REMPI spectrum was extracted with a homemade software program that used LabVIEW (National Instruments Japan, Tokyo). All experiments for measuring the mass and REMPI spectra were performed at room temperature. In order to measure the mass chromatogram, the end of the capillary opposite the restricted tip was connected to a capillary column (Agilent Technologies Japan, HP-5, 30 m long, 0.32 mm i.d., Tokyo) for sample separation in the GC (Agilent Technologies Japan, 6890N, Tokyo). An autosampler (Agilent Technologies Japan, 7683B, Tokyo) was used to inject 1 µL of the sample solution dissolved in hexane. The sample was prepared at concentrations from 10 pg/µL to 100 ng/µL. The temperature of the GC oven was programmed as follows: the initial temperature of 50 °C was ramped at 20 °C/min until it reached 250 °C, which was then held for 3 min. The temperatures of the injection port and the transfer line were maintained at 270 and 250 °C, respectively. The temperature of the metal support holding the jacket of the capillary tip was kept at either 130 or 320 °C to measure the mass chromatogram. The flow rate of helium was 5 mL/min, and the eluted analyte was introduced into the TOFMS. A mass chromatogram was obtained by extracting a series of recorded data using the homemade software program mentioned previously. RESULTS AND DISCUSSION Mass Spectra. Figure 2 shows the mass spectra for pchlorophenol without and with a desorption laser. Four isotope peaks of p-chlorophenol were observed in the mass spectrum obtained using the desorption laser. The peak height was increased 130-fold at m/z ) 128 and 190-fold at m/z ) 130. The mass resolution was 1200, as shown in Figure 2a, whereas it was

Figure 2. Mass spectra for p-chlorophenol. Desorption laser: (a) 0 MW/cm2 (no desorption laser); (b) 0.8 MW/cm2. The timing sequence in each experiment is shown in the figure.

Figure 3. REMPI spectra for p-chlorophenol at different power densities of the desorption laser: (a) 0 MW/cm2 (no desorption laser), (b) 0.5 MW/cm2, and (c) 0.9 MW/cm2.

reduced to 780 for the peak at m/z ) 128 in Figure 2b. Since the mass resolution at m/z ) 130 was as high as 1200, as shown in Figure 2b, the broadening of the peak was not caused by an introduction of the desorption laser. Thus, the peak broadening arose from the saturation of the detector response. Therefore, the online COLD technique could be used for sensitive detection of the analyte without degradation of the mass resolution. REMPI Spectra. Figure 3 shows the REMPI spectra for p-chlorophenol measured at different power densities of the desorption laser. As shown in Figure 3a, several sharp peaks were clearly observed without using the desorption laser. The peak at 287.17 nm can be assigned to the 0-0 transition of p-chlorophe-

Figure 4. Mass chromatograms for p-chlorophenol. Temperature of the capillary holder: (a-d) 130 °C; (e and f) 320 °C. Desorption laser: (a and e) OFF; (b-d and f) ON. Concentration: (a-c and e and f) 1 ng/µL; (d) 0.01 ng/µL. Wavelength of the ionization laser: (a and b and d-f) 278.17 nm; (c) 278.27 nm.

nol.23 The spectral line width was 0.05 nm, suggesting sufficient temperature cooling by supersonic jet expansion. We were able to increase the signal intensity by introducing the desorption laser (0.5 MW/cm2, 532 nm), as shown in Figure 3b. An enhancement factor, which is defined as a ratio of the signal intensities at the 0-0 transition obtained both with and without the desorption laser, could be calculated to 130. No difference was observed in the spectral feature, as shown in Figure 3, parts a and b. These results suggest that a sufficient amount of sample was reliably adsorbed at the tip of the capillary and desorbed by the laser. Accordingly, sensitive detection, as well as optically selective ionization, can be achieved with an online COLD technique. However, the baseline of the signal was unfavorably increased when the power density was increased to 0.9 MW/cm2 as shown in Figure 3c, although the enhancement factor was increased to 320. In this case, the tip of the capillary could be substantially heated, probably through a multiphoton absorption process using an intense desorption laser. The increase in the tip temperature would increase the temperature of the supersonic jet, resulting in an increase of the baseline. Thus, optimization in the energy of a desorption laser is necessary for sensitive and selective detection of the analyte molecule. Mass Chromatograms. This COLD sample introduction technique was applied to the interface of the GC and MS. Figure 4 shows the mass chromatograms obtained for p-chlorophenol. The result shown in Figure 4a was obtained without use of the desorption laser. In this case, some portion of the sample was adsorbed at the tip of the capillary, and the other portion was continuously introduced into MS. The retention time of p(23) Tembreull, R.; Sin, C. H.; Li, P.; Pang, H. M.; Lubman, D. M. Anal. Chem. 1985, 57, 1186–1192.

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1285

chlorophenol was 4.1 min, and the full width at half-maximum (fwhm) of the peak was 2 s. The signal intensity was increased 82-fold by introducing the desorption laser, as shown in Figure 4b. It should be mentioned that no signal broadening was observed in the mass chromatogram shown in Figure 4b. This could be due to a near-zero dead volume for the interface. Therefore, this online COLD technique is useful for the enhancement of the signal in GC/REMPI-TOFMS. The signal enhancement obtained (82-fold) was slightly lower than the value of 130fold obtained in the previous section. This is probably due to the elevated temperature of the tip, since the metal support was heated to 130 °C. When the wavelength of the ionization laser was slightly shifted to 287.27 nm, no peaks were observed in the mass chromatogram, as shown in Figure 4c. This suggests superior spectral selectivity of supersonic jet spectrometry achieved by a combination of online COLD and GC/MS. Figure 4d shows the mass chromatogram for a 10 pg sample of p-chlorophenol. The signal-to-noise ratio (S/N) of the observed peak was 6.8, and the detection limit (S/N ) 3) was calculated to be 4.4 pg. On the other hand, the detection limit obtained by using a continuous sample introduction technique was calculated to be 320 pg from the data shown in Figure 4a. Therefore, the detection limit could be improved ca. 73-fold using the online COLD technique. The mass chromatogram was measured for p-chlorophenol at an elevated temperature (320 °C) with a metal jacket holder. The signal intensity obtained using a desorption laser was enhanced only 2.2-fold (see Figure 4, parts e and f). These results suggest that fewer analyte molecules were adsorbed at the higher temperature of the tip. In fact, the amount of the analyte adsorbed at 320 °C was reduced to 2.7% (2.2/82) of the analyte adsorbed at 130 °C. However, the detection limit (S/N ) 3) of p-chlorophenol measured without using the desorption laser at 320 °C (330 pg) was nearly identical to that obtained at 130 °C (320 pg) (see Figure 4, parts a and e). This result suggests that only a small percentage of the analyte in the gas was adsorbed at the tip for both experiments at 130 and 320 °C and that the percentage of the analyte molecule continuously flowing through the tip was nearly identical. However, the enhancement factor was substantially decreased in the experiment performed at 320 °C, since the absolute amount of the analyte molecule adsorbed at the tip was significantly reduced by increasing the temperature of the tip from 130 to 320 °C. It should be mentioned that the enhancement factor generally increased at lower tip temperatures. The tip would be, however, plugged for a sample prepared at extremely high concentrations. Therefore, optimization of the tip temperature is required for optimal use of the online COLD technique. Calibration Curves. Figure 5 shows the calibration curves for p-chlorophenol. The calibration curve provided a correlation coefficient of R ) 0.9999 in the concentration range from 1 to 100 ng/µL and R ) 0.9991 in the range from 10 pg/µL to 100 ng/µL, without and with a desorption laser, respectively, allowing quantitative analysis in a wide dynamic range of more than 4 orders of magnitude. However, the slope of the curve (0.94) shown in Figure 5a was slightly smaller than 1.00, which could be explained by the space-charge effect in the ionization region. In Figure 5b, the same value of the slope (0.94) was obtained in a concentration range from 10 pg/µL to 10 ng/µL, but the signal intensity at a concentration of 100 ng/µL was slightly lower than 1286

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

Figure 5. Calibration curves for p-chlorophenol. Desorption laser: (a) OFF; (b) ON. The error bar in curve b shows the standard deviation of the data (n ) 3).

the expected value from a slope of 0.94. A possible explanation is a larger effect of space charge at the higher concentration of 100 ng/µL, resulting in a spread of ions such that fewer ions reach the detector than without space charge. Another explanation could be a lower percentage of the analyte adsorbed at higher concentrations by a multilayer coverage of the analyte at the capillary tip surface, as described in the following section. Model for Analyte Adsorption. The coverage ratio, θ ) V/Vm, is defined as the ratio of the analyte weight adsorbed on the tip surface (V) and the maximum weight calculated by assuming a full coverage of the tip surface as a monolayer (Vm). The value of Vm is given by Vm )

SM σNA

where S is the area at which the analyte can be adsorbed in the capillary, M is the molecular weight of the analyte, σ is the cross section of the adsorbed analyte, and NA is Avogadro’s number. The surface area can be calculated from the equation S ) 2πrL, where L is the length of the analyte zone in the capillary and r is the inner radius of the tip (50/2 µm). The parameter of L can be assumed to be ca. 400 µm from ref 18, although the flow rate was 1 mL/min in ref 18 and 5 mL/min in the present study. Thus, S can be calculated to be 6 × 1010 nm2. The cross section, σ, can be calculated as follows:24

(

σ ) 2√3

M 4√2NAdt

)

2/3

where dt is the density of the solid or liquid analyte. Then, the value of σ is calculated to be 0.33 nm2 for p-chlorophenol. Thus, the value of Vm becomes 40 pg. The weight of the analyte passing through the tip during the interval of the laser pulses (Va) is given by

Va )

Vi f∆T

(24) Young, D. M.; Crowell, A. D. In Physical Adsorption of Gases; Niessner, R., Ed.; Butterworths: London, 1962; p 226.

where Vi is the weight of the analyte molecule injected into GC, f is the repetition rate of the laser, and ∆T is the fwhm of the chromatographic peak. In the present study, the value of Va was calculated to be 50 pg under the assumption that Vi ) 1 ng and ∆T ) 2 s. The value of V can be estimated using the following equation: V ) VaAf∆t where ∆t is the pulse duration of the sample packet and A is the enhancement factor. Using ∆t ) 6.5 µs and A ) 82, the value of V was calculated to be 0.3 pg. Therefore, the coverage ratio of θ was calculated to be 0.8% when Vi ) 1 ng. Generally, the amount of the analyte adsorbed on the tip (V) should obey Henry’s absorption isotherm, and it should be proportional to the partial vapor pressure (or concentration) when the analyte molecules are adsorbed as a monolayer, and then, the coverage ratio of θ would be much less than a few percentage points. Multilayer coverage of the analyte should be taken into account at higher concentrations. Under such conditions, the molecules were insufficiently adsorbed on the surface of the capillary. This could be the reason the calibration curve was slightly saturated at high concentrations (100 ng/µL), as discussed in the previous section. CONCLUSIONS In this study, an online COLD sample introduction technique was utilized as a sensitive and selective means for the detection

of an analyte based on REMPI-TOFMS. In fact, the signal intensity was enhanced by 2 orders of magnitude, and superior spectral selectivity was achieved by supersonic jet spectrometry. These advantages were also examined and confirmed in a study using the same combination with a GC. A straight calibration curve was obtained in the dynamic range that exceeded 4 orders of magnitude. In addition, the nozzle has near-zero dead volume and can be heated to >300 °C. Thus, the nozzle has a distinct advantage at the interface of a GC and MS. Therefore, this novel sample introduction technique has the potential for use in practical trace analysis of pollutants in the environment, such as PCDD/Fs and PCBs, using GC/REMPI-TOFMS. ACKNOWLEDGMENT This research was supported by a Grant-in-Aid for the Global COE program “Science for Future Molecular Systems” from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This work was also supported by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) and the Creation and Support Program for StartUps from Universities from the Japan Science and Technology Agency.

Received for review October 8, 2009. Accepted January 5, 2010. AC902273B

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1287