Anal. Chem. 2010, 82, 3454–3456
Letters to Analytical Chemistry Multiphoton Ionization Spectroscopy as a Diagnostic Technique of Surfaces Under Ambient Conditions Yuheng Chen, Valery Bulatov, Natalia Vinerot, and Israel Schechter* Schulich Faculty of Chemistry Technion-Israel Institute of Technology, Haifa 32000, Israel Direct detection of solid substances is an important yet challenging issue in analytical chemistry. Laser multiphoton ionization spectroscopy has been applied for the first time for direct analysis of solids under ambient conditions. In this method, a solid powder/film is placed on a conductive surface and is irradiated by a pulsed tunable laser while an electrical field of ∼2 kV cm-1 is applied across this conductive surface and another electrode. The resulting photoelectrons and negative ions are measured by recording the current as a function of wavelength to produce a multiphoton ionization spectrum that is characteristic of the surface. Results indicate rich spectral features that can be used for compound identification. The present sensitivity is in the low picomole range. This method has been successfully tested for direct detection of various organic molecules, including explosives, narcotic drugs, and polycyclic aromatic compounds. Multiphoton ionization (MPI) is a nonlinear process where an atom or a molecule can simultaneously absorb more than one photon and get ionized. The process can readily be monitored based on either the photoelectrons or the produced ions. Current applications of MPI are mainly in mass spectrometry, since the resulting molecular ions do not undergo extensive fragmentation. Under vacuum conditions, tuning the irradiation wavelength may reach resonant transitions, thus considerably enhancing the ionization probability. MPI has also been tested under ambient conditions.1-15 It has been found that the resulting photocurrent can be used for quantitative estimation of the concentration of an easily ionized compound in a hardly ionizable matrix.1-13 In its resonant mode, the method can be used for detection of gaseous organic compounds.14,15 * To whom correspondence should be addressed. E-mail: israel@ techunix.technion.ac.il. (1) Yamada, S.; Sato, N.; Kawazumi, H.; Ogawa, T. Anal. Chem. 1987, 59, 2719–2721. (2) Schechter, I.; Schro ¨der, H.; Kompa, K. L. Anal. Chem. 1992, 64, 2787– 2796. (3) Gridin, V. V.; Korol, A.; Bulatov, V.; Schechter, I. Anal. Chem. 1996, 68, 3359–3363. (4) Tanaka, J. Bull. Chem. Soc. Jpn. 1965, 38, 86–103.
3454
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
We succeeded, for the first time, to utilize MPI spectroscopy for direct chemical analysis of solids under ambient conditions. In our experimental setup, various organic solids have been irradiated with a laser and the photoelectrons were collected using a positive electrode placed nearby. The laser wavelength has been tuned in a wide range, and the time-integrated photocurrent was registered as a function of wavelength. EXPERIMENTAL SECTION The experimental setup is shown in Figure 1. It consists of a sample holder, a tunable laser irradiation source, and a detection unit. The solid samples were placed between two electrodes connected to a high voltage supply (PS325, Stanford Research Systems). The gap between the electrodes was 10 mm, and the applied voltage was 2 kV. Scanning of the laser irradiation point was possible using an XY-stage, which was connected to the negative electrode. Local ionization has been induced by a solid state optical parametric oscillator (OPO) pumped by the third harmonic of a Nd:YAG laser (355 nm) and frequency doubled. It provided 5 ns pulses in the wavelength range 220-355 nm (5.64-3.54 eV), at 0.1 nm steps. The laser power is measured using a beam splitter and a pyroelectric sensor. The irradiation power is set at the desired value using an attenuator. (5) Litani-Barzilai, I.; Bulatov, V.; Gridin, V. V.; Schechter, I. Anal. Chim. Acta 2004, 501, 151–156. (6) Gridin, V. V.; Bulatov, V.; Korol, A.; Schechter, I. Anal. Chem. 1997, 69, 478–484. (7) Gridin, V. V.; Litani-Barzilai, I.; Kadosh, M.; Schechter, I. Anal. Chem. 1997, 69, 2098. (8) Gridin, V. V.; Bulatov, V.; Korol, A.; Schechter, I. Instrum. Sci. Technol. 1997, 25, 321–333. (9) Gridin, V. V.; Litani-Barzilai, I.; Kadosh, M.; Schechter, I. Anal. Chem. 1998, 70, 2685–2692. (10) Inoue, T.; Gridin, V. V.; Ogawa, T.; Schechter, I. Anal. Chem. 1998, 70, 4333–4338. (11) Gridin, V. V.; Inoue, T.; Ogawa, T.; Schechter, I. Instrum. Sci. Technol 2000, 28, 131–141. (12) Litani-Barzilai, I.; Fisher, M.; Gridin, V. V.; Schechter, I. Anal. Chim. Acta 2001, 439, 1–8. (13) Kadosh, M.; Gridin, V. V.; Schechter, I. Israel J. Chem. 2001, 41, 99–104. (14) Cullum, B. M.; Shealy, S. K.; Angel, S. M. Appl. Spectrosc. 1999, 53, 1646– 1650. (15) Chen, K.; Pender, J. E.; Ferry, J.l.; Angel, S. M. Appl. Opt. 2004, 43, 6207– 6212. 10.1021/ac100539x 2010 American Chemical Society Published on Web 04/09/2010
Figure 1. Experimental setup.
The resulted photocurrent was amplified (Keithley 428, gain, 107 V/A) and measured using a digital oscilloscope (Tektronix, TDS 220). A typical temporal profile of the photocurrent is shown in the inset. Under ambient conditions, the signal spans over ∼100 µs and is composed of two parts: a sharp peak due to displacement charges and due to fast photoelectrons reaching the anode and a broad peak due to photoelectrons entrapped by O2 molecules on their way to the anode. MPI spectra were obtained by scanning the laser wavelength and integrating the photocurrent over time for each wavelength. The MPI signals were normalized to the laser power produced at each wavelength. The tested organic substrates included some explosives, drugs, and polycyclic aromatic compounds. They were placed on the conductive surface in the form of powder, of ∼0.5 mm thickness. For quantification, solutions of various concentrations were prepared and applied on the condutive surface. Upon vaporization of the solvent, known surface concentrations were obtained. RESULTS AND DISCUSSION The results indicate a spectral fingerprint of the various compounds, which resembles the corresponding absorption spectra in solution and in thin solid films. An example is presented in Figure 2a. As far as we know, this is the first report of MPI spectra of solid samples. Note the detailed information obtained from the MPI spectra, which is much richer than that obtained from absorption measurements. This is attributed to differences in process probabilities of multiphoton ionization and single photon absorption. It seems that MPI better reflects the molecular energy levels of the sample. Some of the known absorption transition bands are also provided in this figure. These results indicate that MPI performed under ambient conditions and monitored via the emitted photoelectrons is a new kind of spectroscopy that provides useful analytical information on solids. The signals obtained in this operational mode depend on electrical conductivity, which is essential for preventing severe
Figure 2. (a) MPI spectrum of solid pyrene under ambient conditions. The solid and solution absorption spectra are provided for comparison. (b) The integrated signal as a function of the quantity of material in the laser spot. The 95% confidence intervals used for estimating the LOD are also shown. The inset shows the time-dependent signal under ambient conditions.
space charge effects. However, as previously shown, solid materials under ambient conditions possess a certain level of conductivity, due a persistent water layer.6 This native conductivity is sufficient for ensuring reproducible waveforms at low laser repetition rates. Nevertheless, it has also been shown that when the thickness of an insulating sample reaches a certain value, the native conductivity fails to overcome space-charges and the Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
3455
Figure 3. MPI spectra of some solid explosives, directly obtained under ambient conditions.
resulted waveform might be affected.13 Since the thickness that was sufficient for discharging the space-charges was of the order of 0.5 mm, and since linear calibration plots were obtained for rather thick layers, we can assume that the MPI process might involve not only the upper molecular monolayer. This should depend on the light penetration, on one hand, and on the ability to extract electrons from the inner layers, on the other hand. More quantitative conclusions require additional investigation. Quantification is possible, since the time-integrated signals are proportional to the number of molecules in the laser interaction spot. Note that at such a high voltage, the collection efficiency is close to 1. An example of a calibration plot for pyrene is shown in Figure 2b. The inset shows the time dependent signal resulting from a single 5 ns laser pulse. Quantification was based on averaging over 16 such pulses. The limits of detection (LOD) depend on the material and on the laser wavelength. For pyrene (at 270 nm), the 95% confidence limit based LOD was 2 pmol. Similar spectra were also obtained for a variety of other solid compounds, including HMX, RDX, TATP, TEN, TNT (explosives), MDA, THC (drugs), coronene, anthracene, chrysene, and melamine (PAHs). (The latter has been of considerable interest due to its fatal addition to milk powder in China and due to its difficult detection16). In most cases the spectra were detailed and allowed for chemical characterization. An example is shown in Figure 3. The major advantage of the proposed detection method is that it provides more detailed information than regular absorption spectroscopy, and it can be carried out directly on the sample surface, without special preparations. (Optical absorption measurements require preparing a thin film of the material.) (16) Chinese milk powder contaminated with melamine sickens 1,253 babies. The Times, September 16, 2008.
3456
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
The method provides information on molecules in the irradiated spot of a slightly focused laser. Therefore, the method can be used for scanning a surface, spot by spot. Better spatial resolution can be obtained by further focusing the laser. However, the power density should be kept below the breakdown threshold. This imposes a practical resolution of about 20 µm. Air humidity has only a small effect upon the MPI signals. Changes due to regular day to day variations were insignificant. Testing the effect of sample water content was carried out using dry, moist, and wet soil samples. The MPI signals were higher by ∼20% in the wet samples. However, such high variations are not relevant to hydrophobic organic compounds. The method is technologically simple. The only rather expensive component is the laser, but it will probably become less expensive in the future. The major drawback of this method is that it requires electrical contact with the sample. Nonconducting materials can also be analyzed, provided that the electrical contact is not too far away from the irradiation spot (conductivity takes place via natural humidity and/or the thin layer of surface water persistent on all materials under ambient conditions. Another drawback is related to the fact that an efficient collection of the photoelectrons requires application of high electrical fields (order of 1000 V cm-1). Moreover, the results presented here can only be considered as an indication of feasibility, and further studies are required for establishing a new analytical method. CONCLUSIONS We conclude that the proposed method introduces a new kind of spectroscopy that has the potential to be commonly used for direct detection of solids. For example, the method could be utilized for detection of explosives, for quality assurance in food and in pharmaceutical industries, and even for detection of pesticides on fruits and vegetables (confirming organic products). The method can also become useful in numerous applications where nondestructive analysis is required. Quantification capability has been demonstrated; however, further investigation is needed for understanding various effects, such as sample conductivity (affected by humidity and thickness), mixtures, and matrix effects. ACKNOWLEDGMENT This research was supported, in part, by the James Frank Program in Laser Matter Interaction. V.B. thanks the Israel Ministry of Absorption for financial support in the framework of the Kamea program. Received for review February 27, 2010. Accepted April 8, 2010. AC100539X
