Electrospray Mass

This automation system enables both “hands-free” formation of the liquid microjunction used to sample material from the surface and hands-free reo...
0 downloads 0 Views 352KB Size
Anal. Chem. 2005, 77, 7183-7189

Automation of a Surface Sampling Probe/ Electrospray Mass Spectrometry System Vilmos Kertesz,* Michael J. Ford,† and Gary J. Van Berkel*

Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6131

An image analysis automation concept and the associated software (HandsFree TLC/MS) were developed to control the surface sampling probe-to-surface distance during operation of a surface sampling electrospray system. This automation system enables both “hands-free” formation of the liquid microjunction used to sample material from the surface and hands-free reoptimization of the microjunction thickness during a surface scan to achieve a fully automated surface sampling system. The image analysis concept and the practical implementation of the monitoring and automated adjustment of the sampling probe-tosurface distance (i.e., liquid microjunction thickness) are presented. The added capabilities for the preexisting surface sampling electrospray system afforded through this software control are illustrated by an example of automated scanning of multiple development lanes on a reversed-phase C8 TLC plate and by imaging inked lettering on a paper surface. The post data acquisition processing and data display aspects of the software package are also discussed. Analysis of organic and biological analytes in and on surfaces such as tissue1 and affinity materials2 using mass spectrometric techniques is an expanding area of research and application. More mature ionization/sampling techniques for this purpose include secondary ion mass spectrometry used mainly for elemental and small-molecule analysis and matrix-assisted laser desorption/ ionization used for higher molecular mass analytes such as peptides and proteins.1 Other recently developed surface analysis techniques include desorption electrospray ionization3,4 and direct analysis in real time,5 a type of atmospheric pressure chemical ionization (APCI). Each of these techniques is performed under ambient conditions and appears to be useful largely for the same * Corresponding authors. E-mail: [email protected]. Phone: 865-574-4878. Fax: 865-576-8559. E-mail: [email protected]. Phone: 865-574-1922. Fax: 865576-8559. † Present address: Pharmaceutical Research Institute, Bristol Myers Squibb, Research Parkway, Wallingford, CT 06492. (1) Todd, P. J.; Schaaff, T. G.; Charurand, P.; Caprioli, R. M. J. Mass Spectrom. 2001, 36, 355-369. (2) Tang, N.; Tornatore, P.; Weinberger, S. R. Mass Spectrom. Rev. 2004, 23, 34-44. (3) Taka´ts, Z.; Wiseman, J. M.; Golagan, B.; Cooks, R. G. Science 2004, 306, 471-473. (4) Van Berkel, G. J.; Ford, M. J.; Deibel, M. A. Anal. Chem. 2005, 77, 12071215. (5) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 22972302. 10.1021/ac0510742 CCC: $30.25 Published on Web 10/08/2005

© 2005 American Chemical Society

molecules that typically might be analyzed by electrospray (ES) ionization6 and APCI,7 respectively. Surface sampling probes used in conjunction with ES ionization8-10 and APCI11 have also been demonstrated. In all these cases, complete automation of the surface analysis process is either a critical part of the functionality of the system or a desired capability. In several reports, we have demonstrated the use of a combined surface sampling probe/ES emitter as the interface for coupling thin-layer chromatography (TLC) and mass spectrometry (MS).12-15 This device exploits a surface sampling probe-to-surface liquid microjunction and a self-aspirating ES emitter to sample material from the stationary phase of developed TLC plates for analysis by ES-MS. The analytical utility of this TLC/ES-MS coupling has been demonstrated by the qualitative12-14 and quantitative15 analysis of a variety of analytes separated on commercially available reversed-phase (RP) C8 and C18 TLC plates. We also have recently shown that the same surface sampling concept can be used for direct sampling of analytes on surfaces (including TLC plates) and in liquid solutions with a selfaspirating APCI source.11 Computer-controlled scanning of one development lane at a time on a TLC plate has been accomplished in our prior reports.12-15 However, initial formation of the liquid microjunction and maintenance of the optimum microjunction thickness during the course of an experiment have required manual adjustments by a skilled operator. Analysis of additional development lanes or “spot sampling” required further manual positioning of the surface relative to the probe. The present report describes an image analysis-based concept and the implementation of computercontrolled adjustment of the probe-to-surface distance to automate the surface sampling system. The software package is also used for processing of the acquired data for display. The added (6) Cech, N. B.; Enke, C. G. Mass Spectrom. Rev. 2001, 20, 362-387. (7) Bruins, A. P. Mass Spectrom. Rev. 1991, 10, 53-77. (8) Wachs, T.; Henion, J. Anal. Chem. 2001, 73, 632-638. (9) Modestov, A. D.; Srebnik, S.; Lev, O.; Gun, J. Anal. Chem. 2001, 73, 42294240. (10) Luftmann, H. Anal. Bioanal. Chem. 2004, 378, 964-968. (11) Asano, K. G.; Ford, M. J.; Tomkins, B. A.; Van Berkel, G. J. Rapid Commun. Mass Spectrom. 2005, 19, 2305-2312. (12) Van Berkel, G. J.; Sanchez, A. D.; Quirke, J. M. E. Anal. Chem. 2002, 74, 6216-6223. (13) Ford, M. J.; Van Berkel, G. J. Rapid. Commun. Mass Spectrom. 2004, 18, 1303-1309. (14) Ford, M. J.; Kertesz, V.; Van Berkel, G. J. J. Mass Spectrom. 2005, 40, 866875. (15) Ford, M. J.; Deibel, M. X.; Tomkins, B. A.; Van Berkel, G. J. Anal. Chem. 2005, 77, 4385-4389.

Analytical Chemistry, Vol. 77, No. 22, November 15, 2005 7183

Figure 1. Structure and mass-to-charge ratio for the rhodamine dyes.

capabilities for our surface sampling electrospray system afforded through this software are illustrated by an example of automated scanning of multiple development lanes of a reversed-phase C8 TLC plate containing separated bands of three rhodamine dyes (compounds 1-3 in Figure 1) and by imaging inked lettering (containing 2) on a paper surface. EXPERIMENTAL SECTION Chemicals. HPLC grade methanol and water were purchased from Burdick & Jackson (Muskegon, MI). Ammonium acetate (99.999%) and acetic acid (PPB/Teflon Grade) were obtained from Sigma Aldrich (Milwaukee, WI). Rhodamines 6G (1) and B (2) were purchased from Eastman Kodak Co. (Rochester, NY) and rhodamine 123 (3) was obtained from Sigma Aldrich (see structures in Figure 1). Stock solutions were prepared by dissolving the dyes in methanol at a concentration of 200 ng/µL (0.45 nmol/µL for 1 and 2; 0.57 nmol/µL for 3). Analytical standards were prepared from these stock solutions by dilution with methanol for TLC spotting and 60/40 (v/v) methanol/water with 0.1% acetic acid by volume for ES-MS optimization. Thin-Layer Chromatography. TLC was carried out using hydrophobic Merck RP C8 plates (P/N 13725/5, EM Science, Gibbstown, NJ). Dye standards were spotted on the plates in 0.5-µL aliquots. Plates were developed with 80/20 (v/v) methanol/water containing ∼200 mM ammonium acetate. Developed plates were dried in an oven (110 °C) for 30 min just prior to mass spectrometric analysis. Photographs of the developed TLC plate were taken with a Coolpix 990 digital camera (Nikon, Tokyo, Japan) using long-wavelength UV illumination. Inked Lettering on Paper. A red ink Stamp-Ever “COPY” logo stamp (U.S. Stamp and Sign, Cookeville, TN) was used to impress the word COPY onto hp LaserJet Tough Paper (HewlettPackard Co., Palo Alto, CA). The red dye component of the stamp ink was 2 determined by ES-MS and MS/MS analysis (data not shown). The Tough Paper (cut to 100 × 100 mm) was fixed to a glass backing plate (100 × 100 mm) with UHUstic glue adhesive (Manco Inc., Avon, OH) and then stamped with the red COPY logo stamp. The Tough Paper was sufficiently hydrophobic to allow surface sampling using 60/40 (v/v) methanol/water (0.1% acetic acid by volume) as the eluting solvent system without wetting of the paper surface.12 Photographs of the impressed COPY logo were taken in white light. 7184

Analytical Chemistry, Vol. 77, No. 22, November 15, 2005

Surface Sampling Probe/ES-MS System. The major portion of the experimental setup shown in Figure 2 has been previously described.14,15 The mass spectrometer was a 4000 QTrap (MDS Sciex, Concord ON, Canada) operated using Analyst software version 1.4 on computer PC 1. Data were collected in selected reaction monitoring (SRM) mode with the dissociation conditions optimized separately for each of the rhodamine dyes. Fixed atmospheric sampling interface conditions were as follows: curtain gas (CUR) ) 15, ion spray voltage (IS) ) 4500 V, interface heater (IH) ) 120 °C, and entrance potential (EP) ) 10 V. Compounddependent conditions were as follows: 1, declustering potential (DP) ) 150 V, collision energy (CE) ) 50 eV, collision exit potential (CXP) ) 10 V; 2, DP ) 185 V, CE ) 58 eV, CXP ) 5 V; and 3, DP ) 185 V, CE ) 60 eV, CXP ) 10 V. The SRM transitions monitored were m/z 443 f 415 (1), m/z 443 f 399 (2), and m/z 345 f 285 (3) with a 500-ms dwell time for each transition. The eluting solvent used to sample material from the TLC plates was pumped (solvent pump, SP) from a 1.0-mL glass syringe with a syringe pump (Harvard Apparatus, Inc., Cambridge, MA). An Agilent 1100 LC pump (Agilent Technologies, Inc., Palo Alto, CA) was used to deliver the eluting solvent for the analysis of the ink stamp on the Tough Paper. The aspiration rate of the probe/ emitter was matched to the pumped flow rate by adjustment of the nebulizing gas (nitrogen) flow rate. Analyte standard solutions used to optimize mass spectrometer detection conditions were infused through the probe/emitter via the eluting solvent line. An MS2000 XYZ stage (Applied Scientific Instrumentation Inc., Eugene, OR) was used to move the surface to be sampled relative to the stationary surface sampling probe/emitter during an experiment as described previously.14 The surface was held in the vertical XY plane, perpendicular to the Z axis of the sampling probe. The XYZ stage (ST) was connected to a stage control unit (SCU), which in turn was attached to a laptop computer (PC 2 in Figure 2) via a USB port. The platform was operated under computer control using HandsFree TLC/MS software version 1.2 developed in-house. The position of the TLC plate relative to the stationary sampling probe and the liquid microjunction thickness were monitored with a Panasonic GP-KR222 closed circuit camera (C, Panasonic Matsushita Electric Corporation of America, Secaucus, NJ) with Optem 70 XL zoom lens (Thales Optem Inc., Fairport, NY). The camera image was output (V/I) to a black and white monitor. The video signal from the monitor (V/O) was connected to a Belkin USB VideoBus II (Belkin Corp., Compton, CA) video capture device (VCD) coupled to PC 2 via a USB port to capture and analyze the images of the sampling probe. A Creative Notebook Webcam (W, Creative Labs Inc., Milpitas, CA) connected to PC 2 via USB port was used to obtain a wide-angle view of the surface for initial probe positioning. Safety Considerations. The surface sampling probe/emitter (SPE) floats at the high ES voltage and appropriate shields and interlocks should be used to avoid accidental contact with this device. RESULTS AND DISCUSSION Automation Concept and Implementation. Full computercontrolled operation of the surface sampling probe ES system required the capability, without operator intervention, to initiate formation of a liquid microjunction between the sampling end of the probe and the surface to be sampled. Full automation also

Figure 2. Schematic illustration of the automated TLC/ES-MS experimental setup (not to scale): LS, light source; SP, syringe pump; W, webcam; TV, TV monitor; V/I, video in; V/O, video out; MS, mass spectrometer; C, camera; PH, plate holder; SPE, sampling probe emitter; ST, XYZ stage; SCU, stage control unit; XYC, XY axis control; ZC, Z axis control; VCD, video capture device; PC1, computer running Analyst software for control of the mass spectrometer; PC2, computer running HandsFree TLC/MS software for stage control, for data acquisition from the stage control unit, and for image acquisition from the webcam and the camera.

required the capability to maintain the optimum microjunction thickness (Z axis control) as the surface was moved in the XY plane. This task required precise, real-time measurement of the probe-to-surface distance. That was accomplished here using image analysis, where a predetermined area of the captured image showing the sampling probe tip and the surrounding environment was analyzed. The image analysis distance measurement is illustrated in Figure 3a with a schematic of 9-pixel wide and 19-pixel high captured image of the probe tip and surface to be sampled. The area of the image analyzed was between two vertical lines, L1

and L2. By applying the proper lighting, the sampling probe and the surface being analyzed were brighter than the sampling probe tip where the eluting solvent protruded slightly from the tip (D, Figure 3a). The brightness of the pixels along the horizontal lines between L1 and L2 was summed (3 pixels in every line, marked by circles in the example in Figure 3a). This calculated number represented the average brightness of the horizontal line (LAB), which was plotted versus the Z axis position (probe-to-surface direction) (Figure 3b). The plotted LABs were normalized relative to the brightest and the darkest LAB value in the examined range. Also, seen in this illustration is this shadow (E) of the probe on Analytical Chemistry, Vol. 77, No. 22, November 15, 2005

7185

Figure 3. (a) Schematic representation of a theoretical 9-pixel by 19-pixel image to show the concept of the image analysis: A, background; B, nonexamined parts of the surface probe; C, examined parts of the surface probe; D, liquid/probe interface; E, probe shadow. L1 and L2 vertical lines show the limits of the “image-analyzed” area of the picture. (b) Calculated, normalized horizontal line average brightness (LAB) along the Z axis for the illustration in (a).

the surface. This shadow was used to advantage to estimate/ calculate the probe-to-surface distance (see below). Figure 4 shows actual captured images of the surface approaching the probe and the corresponding LAB versus Z axis position plots. The image in Figure 4a shows the probe ∼500 µm

from the surface with the resulting brightness versus Z axis plot indicating a peak corresponding to the sampling probe tip. As the distance between the probe tip and the surface decreased, the shadow of the probe entered the analyzed part of the image resulting in a second peak on the brightness plot (Figure 4b,c). The image in Figure 4d shows the optimum liquid microjunction. In this case, the brightness versus Z axis plot exhibited only one, relatively wide peak because there no longer was a gap between the probe and surface. These image data presented two alternatives to automate formation of the liquid microjunction and to maintain the optimum microjunction thickness. The first possibility was to allow the surface to approach the probe until the two peaks corresponding to the probe tip and the probe shadow appeared in the analyzed picture and then to track the z coordinates and merging of the two peaks. The calculation of the probe-to-surface distance in that case could be based on the peak separations and widths. However, experiments showed that dark spots derived from samples on the surface interfered with the detection of the second peak. Furthermore, when the surface was not uniform, the detection of the second peak was unreliable. The second possibility to automate control of the liquid microjunction was to follow the full width at half-maximum (fwhm) of the first peak. The fwhm was relatively constant as the surface approached the probe until a sudden rise when the probe tip and surface shadow peak began to merge followed by a linear decrease in the fwhm value. This method was improved by setting a line at the start of the experiment that represented the edge of the probe tip (L3 in Figure 4). The distance between this line and the halfpeak width on the surface side of the Z axis LAB peak (Wp,1/2) was then monitored to judge the probe-to-surface distance. This

Figure 4. Captured images showing different stages while the surface approached the probe: (a) surface relatively far from probe tip (∼500 µm) horizontal line average brightness (LAB) versus Z axis plot exhibits only one peak; (b) surface approaches closer to the probesthe shadow of the probe causes appearance of a second peak in the LAB versus Z axis plot; (c) surface nearing contact with probestwo peaks in the LAB versus Z axis plot merge increasing Wp,1/2; and (d) stable liquid microjunction formed. L1 and L2 vertical lines show the limits of the imageanalyzed area of the picture. L3 indicates the horizontal line that was manually set before the experiment to measure the distance Wp,1/2 (see text for details). 7186 Analytical Chemistry, Vol. 77, No. 22, November 15, 2005

adjustment eliminated unreliable detection of the edge of the probe tip. Successful long period automated surface sampling experiments proved that monitoring Wp,1/2 was a good approach to monitor the liquid microjunction thickness as demonstrated by the data shown below. Thus, in an actual automated surface sampling experiment there were four stages, with software variables for optimization of each, to form and maintain a stable liquid microjunction. In stage 1, the surface approached the probe until Wp,1/2 reached a preset value corresponding to the situation presented in Figure 4c. In stage 2, the surface was moved ∼5-10 µm closer to the probe than the optimal thickness of the liquid microjunction to initiate the liquid microjunction formation. In stage 3, the surface was kept there for a predetermined time (usually 3 s) to form a stable liquid microjunction and to provide time for the start of the mass spectrometry data acquisition. In stage 4, the surface was moved back from the probe to establish the predetermined optimal liquid microjunction thickness. This stage was followed by continuous monitoring and adjustment of the probe-to-surface distance between preset limits to maintain an optimal liquid microjunction during acquisition of the mass spectral data as the surface was moved in the XY plane. Data Processing with the HandsFree TLC/MS Software. In addition to controlling the liquid microjunction, the software was also used to synchronize mass spectrometric signals (time, intensity) with the probe location data (time, x and y locations) to produce x-y-intensity data sets from which three-dimensional (3D) plots or two-dimensional (2D) contour plots could be produced in external graphics programs. Visualization of the data as 2D contour plots was also provided within the software. Automated Analysis of Multiple Development Lanes on a TLC plate. Presented in Figure 5 are the results of an automated scan of four development lanes (Y ) 0, -10, -20, and -30 mm) on a RP C8 TLC plate. A spotted mixture of the three rhodamine dyes (1-3, Figure 1) was separated in each lane. Figure 5a is a photograph of the plate showing the separated components on the plate and an overlay of the scanning route through the development lanes. The plate was mounted in the XYZ stage and the length and distance of the lanes to be scanned set in the software. The surface was manually moved to position the probe above the starting point. The probe-to-surface distance was set at ∼300-500 µm. At this point, the liquid microjunction formation and surface scan was begun. Following software-controlled formation of the liquid microjunction, the actual surface scan began (44 µm/s) in synchronization with mass spectral data acquisition employing SRM detection. At the end of the multiple-lane scan, the software moved the surface back from the probe by 200 µm severing the liquid microjunction and the mass spectral data acquisition was stopped. The data files were recorded by the two computers into separate files, i.e., a file containing the time-x-y-z data sets (file A) stored in the computer that controlled the XYZ stage (PC 2 in Figure 2) and three files (files B-D) containing the time-SRM signals collected for 1-3 stored on the computer attached to the mass spectrometer (PC 1 in Figure 2), respectively. Files A and B were loaded into the software on PC 2 and by synchronizing the time axis in the files the program output one file containing the corresponding x-y-SRM ion current signal for 1. Similarly,

Figure 5. (a) Photograph of a RP C8 TLC plate showing the separated components of a spotted mixture (0.5 µL, 65 ng each component) of rhodamine 6G (1), rhodamine B (2), and rhodamine 123 (3) showing the route of the probe during the automated surface sampling. (b) SRM ion current profiles for 1 (black line, m/z 443 f 415), 2 (red line, m/z 443 f 399), and 3 (blue line, m/z 345 f 285) obtained during the automated multiple-lane scan. Development lanes were scanned at 44 µm/s using an eluting solvent (60/40 methanol/ water, 0.1 vol % acetic acid) flow rate of 15 µL/min and a 500-ms dwell time for each transition.

files including x-y-SRM ion current signal data sets for 2 and 3 were generated by loading files A and C and files A and D into the software, respectively. Figure 5b shows the SRM ion current profiles for 1-3 in a 3D plot (perspective angle is 0° for better viewing). The peak heights (and areas) for the respective dyes were obviously not equivalent in the four lanes even though the same amount of material was spotted in each lane. The most likely explanation for this observation was off-center sampling of the bands. We have previously observed this effect when manually analyzing replicate development lanes on a plate. In this current case, lane 1 (Y ) 0 Analytical Chemistry, Vol. 77, No. 22, November 15, 2005

7187

Figure 6. Representation of the continuous reoptimization of the probe-to-surface distance illustrating the cause for different ALJT values observed in opposite surface scan directions (indicated by the arrows).

mm) was most accurately lined up to start the experiment, and the positions of the three other lanes were calculated and the scan parameters entered into the program. A possible but experimentally unexplored contributing factor to differences in intensity between lanes scanned in different directions may be the less than perfectly perpendicular position of the probe and the surface. This condition resulted in small differences in the average liquid microjunction thickness (ALJT, calculated by averaging the probe/surface distance during a lane scan) when scanning a flat surface in opposite directions as indicated in Figure 6. The software reoptimizes the liquid microjunction thickness when it reaches the lower or the upper limit of the required range by setting the optimal probe-to-surface distance (i.e., the center value of that range). For the experiments shown in Figure 5, the ALJT for both lanes 1 (Y ) 0) and 3 (Y ) -20) was 36 µm, while the ALJT for both lanes 2 (Y ) -10) and 4 (Y ) -30), which were scanned in the opposite direction, was 40 µm. The actual effect of this small difference in ALJT on signal intensity will need further investigation. In any case, for qualitative work (i.e., compound identification), these changes in intensity would not compromise an analysis. Using isotopically labeled internal standards developed along with the analytes has been shown to eliminate this issue in quantitative analysis with our surface sampling system.15 Imaging of Inked Lettering on Paper. The lettering COPY was transferred to a sheet of Tough Paper using a stamp with rhodamine B containing red ink. The lettering measured ap7188 Analytical Chemistry, Vol. 77, No. 22, November 15, 2005

Figure 7. (a) Schematic representation of the impressed COPY logo showing the route of the surface sampler during the lane scanning. (b) Picture before lane scanning of the impressed COPY logo on Tough Paper applied by a stamp containing 2. (c) Picture of the impressed COPY logo on tough paper after the lane scanning. (d) Normalized SRM ion current profile for 2 (m/z 443 f 399) obtained during the automated multiple-lane scan. Darker red color represents higher SRM ion signal at that surface location. The lanes were scanned at 88 µm/s using an eluting solvent (60/40 methanol/water, 0.1 vol % acetic acid) flow rate of 15 µL/min and a 500-ms dwell time.

proximately 1.0 cm × 3.7 cm as illustrated in Figure 7a. This illustration also indicates the sampling path along the surface. In this experiment, 13 lanes were scanned and the distance between the lanes was selected as 1.0 mm. The Tough Paper was affixed to a glass plate, and the plate was mounted in the XYZ stage. As with the TLC plate readout, the surface was manually moved to position the probe ∼300-500 µm above the starting point and then the automated, preprogrammed scan pattern was begun (88 µm/s). Mass spectral data acquisition again employed SRM detection. Figure 7b and Figure 7c are photographs of the lettering taken before and after the surface sampling, respectively. The high efficiency of sampling the ink from the surface was indicated by the white tracks through the letters in Figure 7c. The data file containing time and x, y, z position and the other file containing time and SRM signal for 2 were processed by the software in a fashion similar to the TLC lane scan data discussed above. Figure 7d shows the image of the inked letters based on the normalized SRM ion current profile for 2 (m/z 443 f 399) along the 13 scanned lanes. The darker red color in this figure represents a higher SRM ion signal. There was a direct correlation between the photograph of the scanned lettering (Figure 7c) and the scanned image (Figure 7d). Some “tailing” of the signal in the scan direction was observed as the probe traveled across the lettering. This was largely a function of the high concentration of

the rhodamine B in the ink and the finite time to wash out the probe. At the scan rate employed, the probe was not completely washed out until the probe had traveled some distance (∼2 mm) beyond the ink sampled. The data in Figure 7d took 94 min to acquire. During this total time, the surface sampling system was under complete computer control; no operator intervention was required. The data in Figure 7d also show that the readout resolution in these experiments was sufficient to create a readable image of the inked letters. This resolution might not be suitable for other imaging applications (e.g., smaller font lettering). With the current sampling probe (635-µm outer diameter), readout resolution might be improved from 1.0-mm separated lane scans to ∼650-µm separated scans. A smaller diameter probe could be used to further improve resolution by decreasing the necessary distance between lane scans. However, as the probe diameter shrinks, less material will be sampled from the surface and signal levels may be reduced. CONCLUSIONS In this paper, software to control a surface sampling ES-MS system was described and demonstrated. The software automated formation and real-time reoptimization of the sampling probe-tosurface liquid microjunction using image analysis. The image analysis included periodic (1 Hz) capture of still images from a video camera monitoring the region at the tip of the sampling probe followed by analysis of the captured images to determine the sampling probe/surface distance. By determining this distance, the software first automatically formed an optimal liquid microjunction and continuously reoptimized it during the experiment by adjusting the probe-to-surface distance. The software then provided for scanning of multiple parallel lanes along the surface with equal or customized distance between the lanes. The current version of the program supported only constant scan speed during an experiment, but implementation of customized scan speed for each lane is planned in future software versions. With this capability, we demonstrated automated multiple-lane scans along a surface of a TLC plate and over inked lettering on paper. The software synchronized the mass spectrometric data collected during the surface scan with the surface location where those data were measured and produced data files containing x, y coordinates

of the surface and the corresponding mass spectral signal intensities. From these data, it was possible to produce a 3D plot or 2D contour plot, where the X and Y axis of the plot corresponded to the horizontal and vertical range of the scanned surface, and the Z axis or the pixel color represented the corresponding mass spectral signal intensity, respectively. Future work will implement the ability to spot sample rather than just lane scan. This feature will allow the probe to sample from predefined (grid style or customized) locations on the surface by automatically forming a liquid microjunction, accomplishing an analysis, and automatically moving to the next location to be examined. Further study and enhancement of the surface sampling part of the system is also planned. Fabrication of an emitter with a smaller (inner and outer) diameter should increase the resolution of the surface imaging. With the ability to precisely control the liquid microjunction thickness, we will explore sampling efficiency and mass spectral signal intensity dependence on liquid microjunction thickness to better understand the sampling process. ACKNOWLEDGMENT M.J.F. and V.K. acknowledge ORNL appointments through the ORNL Postdoctoral Research Associates Program. The microionspray head was supplied by MDS Sciex through a Cooperative Research and Development Agreement (CRADA ORNL02-0662). Research on the self-aspirating ES emitter using the microionspray platform was sponsored by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences, U.S. Department of Energy. Initial development of the TLC/ ES-MS readout system was sponsored by the Laboratory Directed Research and Development Program of ORNL. Development of the HandsFree TLC/MS software was supported by ORNL Technology Transfer and Economic Development (TTED) Royalty Funds. ORNL is managed by UT-Battelle, LLC for the U.S. Department of Energy. under contract DE-AC05-00OR22725.

Received for review June 17, 2005. Accepted September 8, 2005. AC0510742

Analytical Chemistry, Vol. 77, No. 22, November 15, 2005

7189