Article pubs.acs.org/ac
Plasma Ionization Source for Atmospheric Pressure Mass Spectrometry Imaging Using Near-Field Optical Laser Ablation Maryia M. Nudnova, Jérôme Sigg, Pascal Wallimann, and Renato Zenobi* Department of Chemistry and Applied Biosciences, ETH Zurich, CH-8093 Zurich, Switzerland S Supporting Information *
ABSTRACT: Mass spectrometry imaging (MSI) at ambient pressures with submicrometer resolution is challenging, due to the very low amount of material available for mass spectrometric analysis. In this work, we present the development and characterization of a method for MSI based on pulsed laser ablation via a scanning near-field optical microscopy (SNOM) aperture tip. SNOM allows laser ablation of material from surfaces with submicrometer spatial resolution, which can be ionized for further chemical analysis with MS. Efficient ionization is realized here with a custom-built capillary plasma ionization source. We show the applicability of this setup for mass spectrometric analysis of three common MALDI matrices, α-4-hydroxycyanocinnamic acid, 3-aminobenzoic acid, and 2,5dihydroxybenzoic acid. Although the ultimate goal has been to optimize sensitivity for detecting material ablated from submicrometer diameter craters, the effective lateral resolution is currently limited by the sensitivity of the MS detection system. In our case, the sensitivity of the MS was about 1 fmol, which allowed us to achieve a spatial resolution of 2 μm. We also characterize the analytical figures of merit of our method. In particular, we demonstrate good reproducibility, a repetition rate in the range of only a few seconds, and we determined the amount of substance required to achieve optimal resolution and sensitivity. Moreover, the sample topography is available from SNOM scans, a parameter that is missing in common MSI methods.
M
topographic image and hence, the combination of SMOM with mass spectrometry offers interesting perspectives for MSI. With SNOM tips for laser ablation, ablation of material from an area of about 100 nm has been demonstrated,11 which is close to the average size of the sharpened end of a SNOM probe. The ablation products can then be transported into the ion source of a mass spectrometer for further MS analysis. The proof of principle of combining SNOM-based laser ablation at ambient conditions with submicrometer lateral resolution followed by mass spectrometric analysis was given by Stöckle et al. in 2001.12 In that work, the ablated material was transported through a heated sampling capillary into a quadrupole mass spectrometer, where it was ionized by electron impact. In that work, the authors presented signals only for a fixed m/z ratio. A combination of near-field laser ablation sampling and inductively coupled plasma ionization MS for elemental analysis was later developed by Becker and co-workers.13,14 Yet another alternative was proposed by Van Berkel and co-workers:15−17 a heated atomic force microscopy (AFM) probe was used to locally desorb analyte molecules
ass spectrometry imaging (MSI) at atmospheric conditions is of great importance for materials science and biological research. This type of chemical imaging, preferably using an ambient ionization technique, often does not require sample preparation and hence, atmospheric MSI is advantageous in terms of throughput and robustness.1−3 With the exception of secondary ion mass spectrometry (SIMS),4 most of the MSI techniques utilize laser ablation to volatilize material from the sample surface. Typically, far-field laser ablation techniques for chemical imaging at atmospheric pressures have a resolution of 30−100 μm5 and, in some special cases, down to 5−10 μm.6,7 Even better local analysis of surfaces at ambient pressures via far-field laser ablation has been demonstrated,8 using a home-built atmospheric pressure scanning microprobe matrix-assisted laser desorption/ionization (MALDI) ion source for MSI of peptides with a ≈1 μm diameter. Near-field laser ablation potentially allows one to improve the spatial resolution for surface analysis. The main advantage of near-field laser ablation is that it provides the possibility to break the optical diffraction limit when focusing the laser beam. Like corresponding far-field techniques, it also works at ambient pressures. Near-field ablation can be achieved by guiding a laser beam through a tapered optical fiber tip, e.g., a tip for scanning near-field optical microscopy (SNOM).9,10 At the same time, SNOM can be utilized for acquiring a © XXXX American Chemical Society
Received: October 29, 2014 Accepted: December 17, 2014
A
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry from a sample surface at atmospheric pressure, which was followed by postionization by an electrospray plume. In the case of MSI with micrometer or submicrometer resolution, only a very small amount of material is available for mass spectrometric analysis. Moreover, under ambient conditions, ablation itself produces not only the ions ready for further MS analysis, but also all other sorts of particles: clusters, neutral particles and ionized clusters. Neutral clusters and molecules may stick to the SNOM tip or to the sample surface surrounding the ablation crater, which significantly decreases the sampling efficiency. This was the subject of an earlier, detailed study of our group.18 It was shown that significantly less than half of the ablated material is available for MS analysis; over 50% is redeposited around the ablation crater or on the shaft of the SNOM tip. The main driving force for transport of the remaining material into the mass spectrometer is the pressure difference between atmosphere and the vacuum inside the mass spectrometer. The efficiency of the transportation thought a heated sampling capillary and differentially pumped pressure reduction stages was investigated by Zhu et al.,19 who found that only a few % of the substance finally end up in the mass spectrometer. These phenomena reduce the amount of analyte entering the mass spectrometer, and therefore, the material transport is the most critical part for MSI. We have recently proposed an active sampling capillary based on a plasma source, which has the potential to increase the sensitivity of this particular MSI setup.20 Another aspect that is of great importance for MSI is the reproducibility of the laser ablation. This includes the stability and reproducibility of the laser energy output from the SNOM tip, which in turn influences the reproducibility of the size and shape of the ablation craters. The shape and size of the SNOM tip itself is the key factor that determines the shape and size of the ablation craters, which governs the amount of substance available for MS analysis. None of these factors were ever the subject of prior investigations, and their influence on the stability and the reproducibility of the MS data have never been studied. The topic of this work is to show the influence these parameters, including the laser pulse parameters and the SNOM tip shape on the amount of the ablated sample in SNOM-MS experiments. To demonstrate the robustness of high-resolution laser sampling at atmospheric conditions, we performed SNOM laser ablation with a resolution in the low μm range of sinapic acid, 2,5-DHB and α-HCCA acid surfaces, followed by MS signal detection. Mass spectrometric analysis from craters down to 2 μm diameter could be achieved. Amounts in the range of 1−50 fmol were efficiently detected by a mass spectrometer equipped with the active plasma source capillary, which was used to simultaneously transport and ionize the ablated material.
Figure 1. Schematic of the SNOM-MS instrument. Ablation takes place in the optical near field on the sample at ambient conditions. The ablated neutral species are sampled with an active plasma sampling capillary and are transported into the mass spectrometer for chemical analysis.
information of the analyzed surface. In the current setup, SNOM was used for the two main purposes: (i) to acquire the topographic information on the analyzed surfaces, and (ii) to deliver the laser pulses to the surface via the SNOM tip. The tip “focuses” the laser beam to a spot on the surface whose size can be significantly smaller than that allowed by the diffraction limit. In fact, using this technique, one can produce ablation craters as small as tens of nanometers.11 The details of the laser ablation process, its limitations and requirements are discussed in detail below. The SNOM instrument was coupled to the mass spectrometer by a specially designed plasma ionization source (see Figure 1). In the following, we refer to this ionization source as “active sampling capillary”, to emphasize that ionization takes place during transport of the material from the ambient environment into the mass spectrometer. The details of the operation of the active sampling capillary operation and its analytical figures of merit were described in detail in our previous work.20 Here we modified the construction of the plasma source to virtually eliminate electrical interference between the capillary and the analyzed surface. This was achieved by the removing the electrode from the tip of the plasma capillary source, which allowed us to bring the capillary as close as at 20 μm to the analyzed surface. The active sampling capillary consisted of a quartz tube (o.d. 1.5 mm and i.d. 1.0 mm) and two copper pieces that served as electrodes. An inner copper capillary with 0.5 mm i.d. served as the ground electrode. The outer electrode was a copper tape (length 5 mm) surrounding the outside of the quartz tube. The copper tape and the inner tube overlapped by 2 mm. The wall of the glass tube between the copper capillary and the copper tape forms a dielectric barrier. A glass extension of the capillary allows separating the high voltage area from the SNOM head. Here, we further modified the capillary to suppress RF noise, which was found to perturb the feedback of the SNOM head: a metal screen was placed between the high voltage electrodes and the SNOM head to avoid electrical noise (Figure 1). The transport of the analyte was due to the pressure gradient between the ambient pressure region and the first low-pressure region of the mass spectrometer. This pressure gradient generates a gas flow with a rate of 0.8 L/min, as was measured by a gas flow meter (MR3000; Key Instruments, Trevose). SNOM Probes and Near-Field Laser Ablation. For both ablation and optical scanning experiments, we used tips that were “tube etched” according to the procedure described by Stöckle et al.21 A multimode silica fiber with good UV
■
EXPERIMENTAL SECTION Scanning Near-Field Optical Microscopy. To perform MSI at the atmospheric pressure, we used a custom-built experimental setup. A schematic is shown in Figure 1. The MS analysis was performed using an ion trap mass spectrometer (LCQ Deca XP, Thermo, San Jose). Laser ablation was performed with the help of a scanning near-field optical microscope (Lumina, TopoMetrix/Veeco, Santa Barbara, CA). This SNOM instrument was equipped with a noncontact shearforce scan head based on a tuning fork, which was controlled by a special software (SPMLab 6.02) to obtain topographic B
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 2. SEM images of etched tips fabricated by etching in one batch. The exit diameters of the tips is estimated to be (a) 400, (b) 450, and (c) ∼50 nm.
Switzerland) as crystalline powder. All chemicals were used without further purification. For analysis, we pressed pellets of the substances using a hydraulic press with a force of 1 × 104 N, applied for 10 min. To obtain a very smooth surface of the pellets, the material was covered by mica foil that served as a template during the pressing.
transmission (attenuation 70 dB/km at 350 nm) was used as raw material for all SNOM tips produced (Superguide G fiber, High-OH, type SFS50/125Y, multimode UV−vis, 0.22 NA, 50 μm core diameter, 125 μm cladding diameter, 250 μm acrylate jacket, from Fiberguide Industries, Stirling). No metallization of the outside of the tip was done. The tips were glued onto the tuning fork as shown in Figure 1. A Nd:YLF laser (Triton, Spectra Physics, Mountain View, CA) was used for near-field ablation. The laser beam was focused onto the blunt end of the optical fiber. The laser operation was synchronized with the SPMLab software by a pulse generator (model 555, Berkley Nucleonics Corp., San Rafael, CA). Unless otherwise noted, single-shot ablation experiments were carried out in all cases. The SNOM tips were also imaged with a scanning electron microscope. Images of the three tips are shown in Figure 2. Although all of these tips were etched in the same batch and under the same conditions, the tip dimensions vary quite a lot, from ≈50 nm diameter to ≈400 nm. The tip diameter influences the ablation laser power that arrives at the sample surface, which, in turn, defines the spatial resolution and the reproducibility of the experimental data. We did not attempt to achieve a better reproducibility of the tip etching. Instead, we simply took the actual tip shapes into account when evaluating the data. This requires an additional calibration step. Such a calibration was carried out to correlate the particular tip diameter and other parameters such as laser power and SNOM topographic resolution with the parameters of the crater produced by the laser ablation. The optical throughput of the tips was characterized by measuring the output laser power as a function of input laser power. For this, we employed a laser power meter (PE9, Ophir, Jerusalem, Israel) with a maximum allowed energy of 1 mJ/pulse, as well as a more sensitive version when necessary (PD10, Ophir Jerusalem, Israel) with a maximum allowed energy of 200 nJ/pulse. Auxiliary measurements of the crater sizes were also made with an atomic force microscope (NT-MDT, Zelenograd, Russia) Sample Preparation. The near-field laser ablation-MS technique presented here was applied to a range of substances that are normally used as matrices for matrix-assisted laser desorption ionization (MALDI). This choice of substances provides us with the possibility to directly compare the proposed technique with MALDI, which is the leading technique for MSI at ambient conditions with good spatial resolution. The following substances were used: α-4-hydroxycyanocinnamic acid (α-HCCA) and 3-aminobenzoic acid (3AB) were purchased from Sigma-Aldrich (St. Louis, USA) as crystalline powders. 2,5-Dihydroxybenzoic (2,5-DHB) acid was purchased from Acros Organics (Geel, Belgium) as crystalline powder. Sinapic acid was purchased from Fluka (Buchs,
■
RESULTS AND DISCUSSION A key problem in MSI is that a better spatial resolution results in less substance available for MS analysis. Therefore, the sensitivity of the mass spectrometer and the transportation efficiency of the material inside the mass spectrometer can become the limiting factor in performing MSI. In the following, we delineate the quantitative interdependence between the spatial resolution and the sensitivity of SNOM-laser ablation. Based on our findings, we determine a set of the parameters that represent the best trade-off between resolution and sensitivity. We also determine properties that are important for SNOM-MSI under the given set of the parameters. We address the issue of reproducibility, using a correlation between macroscopic parameters (such as the laser power) and microscopic parameters, the most important of which is the shape and size of the ablation crater. Calibration. As mentioned before, we could not fabricate etched SNOM tips with reproducible sizes. The tip size and tip apex shape are the crucial parameters that determine the output laser power, which, in turn, defines the pulse energy delivered to the sample surface and hence the amount of ablated material. Therefore, the optical throughput of SNOM tips was calibrated, because it is a key parameter that determines the ablation characteristics. For each tip, we measured the dependence between the input and output laser energies. Figure 3 shows the results for three different tips used in this work. The optical throughput was found to vary linearly with input laser energy, and lies between 0.3% for the lowest throughput (smallest) tip and 0.9% for the highest throughput tip (i.e., the tip with the largest aperture). Note that there is no single linear dependence on the apex area of the tip and the optical throughput; other factors such as the taper angle near the apex or the precise shape of the end of the tip come into play and influence the optical throughput. Crater Size Determination and Parameter Optimization for SNOM-MS Analysis. The tip shape directly influences the size of the laser spot on the surface, which defines the power reaching the sample and hence the amount of the ablated material. Therefore, an additional calibration parameter that depends on the tip shape, the amount of material ablated by a laser shot, needs to be determined. A reasonable estimate for the quantity of material ablated (in C
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
illustrated in Figure 4. The shot-to-shot reproducibility of the amount ablated, another important methodological parameter, also depends directly on the crater volume. The crater size also gives information about the energy distribution and yields a direct connection between experimental variables (tip shape, input energy) and the ablation (crater size, amount of substance liberated). Here we show results for α-HCCA and the three different SNOM tips whose calibration was shown above, in Figure 3 (Figure 5a). From Figure 5a, one can immediately conclude how important the tip shape is near-field laser ablation using the exact same experimental conditions but different tips results in substantially different amounts of ablated material, because tips with different diameters at the apex produce craters of different volume. The maximum difference is about a factor of 5 for tips 1 and 2 when the same energy was applied (Figure 5a). Another observation is that the amount ablated depends almost linearly on the output laser energy (within the region of interest), which validates our calibration procedure. Error bars were calculated from repeat measurements for the data points shown. From the data in Figure 5, we can also estimate the power density that is necessary for ablation, assuming that the area illuminated by the laser light is equal to the crater fwhm; it amounts to (6 ± 1) × 107 W/cm2. In principle, calibration curves similar to ones shown in Figure 3 would suffice for estimating the amount of substance ablated per shot, if the tip shape is known, for example from electron microscopy images. However, MSI relies on high spatial resolution, thus rendering the crater size, or more precisely, the fwhm more important than the laser output energy. Recalling that both the output energy and the crater size depend on the tip shape, we present results about the dependence of the amount of available material on the crater size. The first fact worth mentioning here is the dependence between the amount of the material and both crater depth and crater fwhm. Near-field laser ablation produces relatively shallow craters, with r ≫ d (r = crater radius, d = crater depth). In the region of interest, the r/d ratio changes from 3 to approximately 10 with increasing laser power. The crater fwhm showed a weaker dependence on laser power than the depth resulting in an increase of the overall r/d ratio. However, in the region of interest, the dependence of the ablated material of different substances increases close to linearly with the fwhm of the crater, as shown in Figure 5b. All three data sets were obtained with tip #2, shown in Figure 2b. Predicting the amount of substance produced in a SNOM ablation event and available for further MS analysis is of great
Figure 3. Optical throughput measured for the three different SNOM tips shown in Figure 2.
fmol) is to use the volume of the crater produced (in μm3), multiplied with the density (g/cm3) and divided by the molecular weight (g/mol). In the next step, we thus determined the ablation crater sizes. To illustrate how this was done, we show an ablation experiment done on a 50 × 50 μm area of a sample, onto which a series of 25 near-field laser ablation shots was fired, with a crater-to-crater distance of 10 μm. This procedure was carried out for all three different MALDI matrices; Figure 4 shows the data for sinapic acid, which was irradiated with a pulse laser energy 85 nJ. To evaluate the data, the craters were approximated as cones, and we measured the full-width at half-maximum (fwhm) and the depth of each crater. The distribution of the crater profiles is presented in Figure 4. We found that the tip shape does not influence the measurements of the craters size in experiments with crater diameters of 1.5 μm and higher. The ablation craters were measured both with the SNOM tips and independently by atomic force microscopy in semicontact mode, using silicon cantilevers. Measurements made by the two methods showed the same crater size (data not shown), indicating that tip convolution effects when using the SNOM to determine the crater topography were negligible for craters larger than 1.5 μm. Under normal experimental conditions, the fwhm varied by no more than 20%. The deviation from the mean of the crater depth was often larger, but never exceeded 50%. This, we postulate, is due to the roughness of the sample and the difficulty to distinguish between the border of the initial surface and hills formed from the redeposited material. This is
Figure 4. 50 × 50 μm area of the surface of a sinapic acid pellet scanned after 25 near-field laser ablation events with 85 nJ pulse laser energy each. (a) 3D view, (b) 2D view, and (c) depth profiles of the uppermost series of craters. The zero line in panel c represents the average height of the sinapic acid surface over the entire 50 × 50 μm scan, after flattening the image. D
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 5. (a) Amount of ablated substance of α-4-hydroxycyanocinnamic acid vs output laser energy for the three tips shown in Figure 3. (b) Amount of ablated substance vs fwhm of ablated craters. Tip #2 was used for all materials in Figure 5 (b). Representative error bars are indicated for the largest craters. These were calculated from the data, using the experimentally derived standard deviation, which was less than 20%, from 10 different measurements.
Figure 6. SNOM image of a four ablation craters generated in an experiment on a sinapic pellet: average depth = 200 nm, fwhm = 2 μm, estimated amount of ablated material ∼ 1.1 fmol.
Figure 7. (a) Selected ion current in the range m/z = 225.2−225.4. The ion current represents the signal intensity as detected by the LCQ mass spectrometer. The peaks correspond to the detection of material produced by single-shot laser ablation events. Laser ablation was done with a repetition rate of 3.5 s−1. Each shot was fired at a new position on the sample surface (see Figure 6). (b) Mass spectrum of the material ablated from position #3 in panel a. The crater of this particular ablation is shown in Figure 6. m/z = 225 corresponds to protonated sinapic acid, m/z = 149 is a background signal.
reproducibility and the limits of applicability of the technique. With the optimization, we aim at reaching two goals at the same time. First, using the plot in Figure 5a, we are able to
methodological importance. First, it is required to determine the macroscopic parameters for achieving the desired resolution and sensitivity. Second, it allows one to estimate the E
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
the speed. The mass spectrum of sinapic acid ablated in the shot #3 is shown in Figure 7). The signal at m/z = 225 Da corresponds to protonated sinapic acid. The main background signal is m/z = 149, which was identified as 4-ethylacetophenone by comparison with reference data.22 This signal was always present in mass spectra of ambient air. The sensitivity is clearly high enough to generate a stable signal from one laser shot: the S/N of signals obtained from all ablation events was about 3, which is postulated to be enough for chemical imaging of surfaces.
estimate the maximal amount of substance available in the SNOM-MS experiment and hence we obtain the minimal sensitivity, which must have a MS instrument to work successfully in the combination with SNOM. Second, we can estimate the amount of material and the spatial resolution of the SNOM-MS combination basing directly on the macroscopic parameters, which are easy to measure and control. SNOM-MS Results. To demonstrate the possibility of robust laser ablation sampling at atmospheric conditions, we performed SNOM-MS with a resolution in the low micrometer range on sinapic acid, 2,5-DHB, and α-HCCA acids pellets for laser ablation followed by the MS signal detection. The material for MS analysis was either produced by single shot event or by a 2 × 2 mesh of shots onto the analyte surface. In all experiments, we measured a topographic image of the surface after acquiring the SNOM-MS signal, to obtain the crater dimensions. For relatively large craters (a few micrometers fwhm), stable and reproducible MS signals with high signal-tonoise ratio were observed. An example of a 6 μm diameter nearfield ablation crater on a 2,5-DHB pellet and the resulting mass spectrum from an estimated amount of ablated 2,5-DHB of about 50 fmol is shown in Figures S1−S3 (Supporting Information). To demonstrate the capability of the SNOM-MS approach to perform high-resolution MSI, data from several ablation events on a sinapic acid sample is presented in Figures 6 and 7. The laser ablation produced craters of about 2 μm fwhm, and depth of about 200 nm. The amount of the ablated material from this crater was estimated to be ≈1.1 fmol (Figure 6). Note that a considerable fraction of the ablated material will be lost in the interface between the SNOM tip and the high vacuum side of the MS. For instance, we have previously shown18 that >50% of the ablated material was deposited on the near-field tip’s shaft. Moreover, part of the ablated material will stick to the walls of the capillary. We can roughly estimate that the maximum amount of the material accessible for MS analysis is ≈0.3 fmol. This small amount of substance was enough to produce reliable mass spectra from a single shot (Figure 7). Laser ablation was done with a repetition rate of 3.5 s−1. The selected ion current in the range m/z = 225.2−225.4 has a direct correlation with the laser shots: roughly, the times at which the laser shots were fired are marked with vertical lines in Figure 7a. We did not synchronize the laser ablation with the acquisition of the mass spectrometer, but the correlation is obvious, e.g., shot #2 corresponds to the peak at 0.12 min. (small peak), shot #3 to the peak at 0.18 min, etc. The mass spectrum is shown for shot #3, with the ion current representing the signal intensity as detected by the LCQ MS. The signal-to-noise ratio for the total ion current is much lower compared to data obtained with larger craters (cf. Figure S2, Supporting Information). This is due to the smaller amount of the ablated material, which scales with the third power of the linear crater dimensions. The ion current peak width is ∼3 s. This width is formed due to substance adsorbed on the walls of the active sampling capillary immediately after the ablation, followed by slow desorption from the capillary walls. Currently, this presents a limitation in the speed that can be achieved for MSI. Substance absorption and following desorption also can cause a shift of a peak maximum with respect to the laser ablation events. However, this issue can be addressed by heating the active sampling capillary, which was not attempted here. Heating should eliminate most of the analyte adsorption, and increase bot the overall sensitivity of the setup as well as
■
CONCLUSIONS In this paper, we presented a technological development of a method to perform high-resolution laser ablation of surfaces under ambient conditions, followed by mass spectrometric analysis of the ablation products. The method is based on the combination of laser ablation via a SNOM tip, and ion production via a custom-built plasma ionization source. We investigated the dependence between the experimental parameters (e.g., tip dimensions, applied laser power) and the analytical characteristics of the method (e.g., size of the craters amount of ablated material) and, in particular, the influence of these parameters on the sensitivity and spatial resolution. We showed that the crater size and therefore the method reproducibility are within 50%. The tip etching procedure used resulted in variety of tip sizes and shapes. This required a tip calibration to be carried out, in order to determine the parameters of the SNOM setup. We showed the applicability of the method for spatially resolved chemical analysis of surfaces of three common MALDI matrices. We achieved a sensitivity of ≈1 fmol, and resolution as low as 2 μm, with a signal-to-noise ratio of ≈3. The mass spectrometer that we used in this work is not state of the art; using a more sensitive machine will certainly better characteristics. We also outlined options to improve the temporal resolution of our method. Finally, scanning sample surfaces with a SNOM tip gives direct access to the sample topography, which is missing in common MSI methods. The technology described here could bring the analysis of surfaces in biology, microelectronics, and related fields to a new level.
■
ASSOCIATED CONTENT
S Supporting Information *
SNOM image of a single crater ablated in 2,5 DHB pellet, mass spectrum before ablation, mass spectrum of ~50 fmol of ablated material, total ion current registered during the single shot ablation of 2.5 DHB pellet, and selected ion current in the range of m/z = 155−155.2. This material is available free of charge via the Internet at http://pubs.acs.org/.
■
AUTHOR INFORMATION
Corresponding Author
*Prof. Renato Zenobi: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Dr. Frank Krumeich (ETH Zurich) for help with recording the electron microscopy images. F
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
■
REFERENCES
(1) Chughtai, K.; Heeren, R. M. A. Chem. Rev. 2010, 110, 3237− 3277. (2) Römpp, A.; et al. Anal. Bioanal. Chem. 2013, 405, 6959−6968. (3) Römpp, A.; Spengler, B. Histochem. Cell. Biol. 2013, 139, 759− 783. (4) Winograd, N. Appl. Surf. Sci. 2003, 203−204, 13−19. (5) Laiko, V. V.; Baldwin, M. A.; Burlingame, A. L. Anal. Chem. 2000, 72, 652. (6) Trimpin, S.; Herath, T. N.; Inutan, E. D.; Wager-Miller, J.; Kowalski, P.; Claude, E.; Walker, J. M.; Mackie, K. Anal. Chem. 2010, 82, 359. (7) Römpp, A.; Günther, S.; Schober, Y.; Schulz, O.; Takats, Z.; Kummer, W.; Spengler, B. Angew. Chem., Int. Ed. 2010, 49, 3834− 3838. (8) Koestler, M.; Kirsch, D.; Hester, A.; Leisner, A.; Guenther, S.; Spengler, B. Rapid Commun. Mass Spectrom. 2008, 22, 3275−3285. (9) Hecht, B.; Sick, B.; Wild, U. P.; Deckert, V.; Zenobi, R.; et al. J. Chem. Phys. 2000, 112, 7761−7774. (10) Shresta, B.; Vertes, A. Anal. Chem. 2009, 81, 8265−8271. (11) Zeisel, D.; Nettesheim, S.; Dutoit, B.; Zenobi, R. Appl. Phys. Lett. 1996, 68, 2491−2492. (12) Stöckle, R.; Setz, P.; Deckert, V.; Lippert, T.; Wokaun, A.; Zenobi, R. Anal. Chem. 2001, 73, 1399−1402. (13) Zoriy, M. V.; Kayser, M.; Becker, J. S. Int. J. Mass. Spectrom. 2008, 273, 151. (14) Zoriy, M. V.; Becker, J. S. Rapid Commun. Mass Spectrom. 2009, 23, 23−30. (15) Ovchinnikova, O. S.; Kertesz, V.; Van Berkel, G. J. Anal. Chem. 2011, 83, 598−603. (16) Ovchinnikova, O. S.; Nikiforov, M. P.; Bradshaw, J. A.; Jesse, S.; Van Berkel, G. J. ACS Nano 2011, 5, 5526. (17) Ovchinnikova, O. S.; Kjoller, K.; Hurst, G. B.; Pelletier, D. A.; Van Berkel, G. J. Anal. Chem. 2014, 86, 1083−1090. (18) Zhu, L.; Gamez, G.; Schmitz, T. A.; Krumeich, F.; Zenobi, R. Anal. Bioanal. Chem. 2010, 396, 163−172. (19) Zhu, L.; Stadler, J.; Schmitz, T. A.; Krumeich, F.; Zenobi, R. J. Phys. Chem. C 2011, 115, 1006−1013. (20) Nudnova, M. M.; Zhu, L.; Zenobi, R. Rapid Commun. Mass Spectrom. 2012, 26, 1447−1452. (21) Stöckle, R.; Deckert, V.; Fokas, C.; Zenobi, R.; Hecht, B.; Sick, B.; Wild, U. P. Appl. Phys. Lett. 1999, 75, 160−162. (22) https://www.waters.com/webassets/cms/support/docs/ bkgrnd_ion_mstr_list.pdf? locale=en_US.
G
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Plasma Ionization Source for Atmospheric Pressure Mass Spectrometry Imaging Using Near-Field Optical Laser Ablation Maryia M. Nudnova, Jérôme Sigg, Pascal Wallimann, and Renato Zenobi* Department of Chemistry and Applied Biosciences, ETH Zurich, CH-8093 Zurich, Switzerland S Supporting Information *
ABSTRACT: Mass spectrometry imaging (MSI) at ambient pressures with submicrometer resolution is challenging, due to the very low amount of material available for mass spectrometric analysis. In this work, we present the development and characterization of a method for MSI based on pulsed laser ablation via a scanning near-field optical microscopy (SNOM) aperture tip. SNOM allows laser ablation of material from surfaces with submicrometer spatial resolution, which can be ionized for further chemical analysis with MS. Efficient ionization is realized here with a custom-built capillary plasma ionization source. We show the applicability of this setup for mass spectrometric analysis of three common MALDI matrices, α-4-hydroxycyanocinnamic acid, 3-aminobenzoic acid, and 2,5dihydroxybenzoic acid. Although the ultimate goal has been to optimize sensitivity for detecting material ablated from submicrometer diameter craters, the effective lateral resolution is currently limited by the sensitivity of the MS detection system. In our case, the sensitivity of the MS was about 1 fmol, which allowed us to achieve a spatial resolution of 2 μm. We also characterize the analytical figures of merit of our method. In particular, we demonstrate good reproducibility, a repetition rate in the range of only a few seconds, and we determined the amount of substance required to achieve optimal resolution and sensitivity. Moreover, the sample topography is available from SNOM scans, a parameter that is missing in common MSI methods.
M
topographic image and hence, the combination of SMOM with mass spectrometry offers interesting perspectives for MSI. With SNOM tips for laser ablation, ablation of material from an area of about 100 nm has been demonstrated,11 which is close to the average size of the sharpened end of a SNOM probe. The ablation products can then be transported into the ion source of a mass spectrometer for further MS analysis. The proof of principle of combining SNOM-based laser ablation at ambient conditions with submicrometer lateral resolution followed by mass spectrometric analysis was given by Stöckle et al. in 2001.12 In that work, the ablated material was transported through a heated sampling capillary into a quadrupole mass spectrometer, where it was ionized by electron impact. In that work, the authors presented signals only for a fixed m/z ratio. A combination of near-field laser ablation sampling and inductively coupled plasma ionization MS for elemental analysis was later developed by Becker and co-workers.13,14 Yet another alternative was proposed by Van Berkel and co-workers:15−17 a heated atomic force microscopy (AFM) probe was used to locally desorb analyte molecules
ass spectrometry imaging (MSI) at atmospheric conditions is of great importance for materials science and biological research. This type of chemical imaging, preferably using an ambient ionization technique, often does not require sample preparation and hence, atmospheric MSI is advantageous in terms of throughput and robustness.1−3 With the exception of secondary ion mass spectrometry (SIMS),4 most of the MSI techniques utilize laser ablation to volatilize material from the sample surface. Typically, far-field laser ablation techniques for chemical imaging at atmospheric pressures have a resolution of 30−100 μm5 and, in some special cases, down to 5−10 μm.6,7 Even better local analysis of surfaces at ambient pressures via far-field laser ablation has been demonstrated,8 using a home-built atmospheric pressure scanning microprobe matrix-assisted laser desorption/ionization (MALDI) ion source for MSI of peptides with a ≈1 μm diameter. Near-field laser ablation potentially allows one to improve the spatial resolution for surface analysis. The main advantage of near-field laser ablation is that it provides the possibility to break the optical diffraction limit when focusing the laser beam. Like corresponding far-field techniques, it also works at ambient pressures. Near-field ablation can be achieved by guiding a laser beam through a tapered optical fiber tip, e.g., a tip for scanning near-field optical microscopy (SNOM).9,10 At the same time, SNOM can be utilized for acquiring a © XXXX American Chemical Society
Received: October 29, 2014 Accepted: December 17, 2014
A
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry from a sample surface at atmospheric pressure, which was followed by postionization by an electrospray plume. In the case of MSI with micrometer or submicrometer resolution, only a very small amount of material is available for mass spectrometric analysis. Moreover, under ambient conditions, ablation itself produces not only the ions ready for further MS analysis, but also all other sorts of particles: clusters, neutral particles and ionized clusters. Neutral clusters and molecules may stick to the SNOM tip or to the sample surface surrounding the ablation crater, which significantly decreases the sampling efficiency. This was the subject of an earlier, detailed study of our group.18 It was shown that significantly less than half of the ablated material is available for MS analysis; over 50% is redeposited around the ablation crater or on the shaft of the SNOM tip. The main driving force for transport of the remaining material into the mass spectrometer is the pressure difference between atmosphere and the vacuum inside the mass spectrometer. The efficiency of the transportation thought a heated sampling capillary and differentially pumped pressure reduction stages was investigated by Zhu et al.,19 who found that only a few % of the substance finally end up in the mass spectrometer. These phenomena reduce the amount of analyte entering the mass spectrometer, and therefore, the material transport is the most critical part for MSI. We have recently proposed an active sampling capillary based on a plasma source, which has the potential to increase the sensitivity of this particular MSI setup.20 Another aspect that is of great importance for MSI is the reproducibility of the laser ablation. This includes the stability and reproducibility of the laser energy output from the SNOM tip, which in turn influences the reproducibility of the size and shape of the ablation craters. The shape and size of the SNOM tip itself is the key factor that determines the shape and size of the ablation craters, which governs the amount of substance available for MS analysis. None of these factors were ever the subject of prior investigations, and their influence on the stability and the reproducibility of the MS data have never been studied. The topic of this work is to show the influence these parameters, including the laser pulse parameters and the SNOM tip shape on the amount of the ablated sample in SNOM-MS experiments. To demonstrate the robustness of high-resolution laser sampling at atmospheric conditions, we performed SNOM laser ablation with a resolution in the low μm range of sinapic acid, 2,5-DHB and α-HCCA acid surfaces, followed by MS signal detection. Mass spectrometric analysis from craters down to 2 μm diameter could be achieved. Amounts in the range of 1−50 fmol were efficiently detected by a mass spectrometer equipped with the active plasma source capillary, which was used to simultaneously transport and ionize the ablated material.
Figure 1. Schematic of the SNOM-MS instrument. Ablation takes place in the optical near field on the sample at ambient conditions. The ablated neutral species are sampled with an active plasma sampling capillary and are transported into the mass spectrometer for chemical analysis.
information of the analyzed surface. In the current setup, SNOM was used for the two main purposes: (i) to acquire the topographic information on the analyzed surfaces, and (ii) to deliver the laser pulses to the surface via the SNOM tip. The tip “focuses” the laser beam to a spot on the surface whose size can be significantly smaller than that allowed by the diffraction limit. In fact, using this technique, one can produce ablation craters as small as tens of nanometers.11 The details of the laser ablation process, its limitations and requirements are discussed in detail below. The SNOM instrument was coupled to the mass spectrometer by a specially designed plasma ionization source (see Figure 1). In the following, we refer to this ionization source as “active sampling capillary”, to emphasize that ionization takes place during transport of the material from the ambient environment into the mass spectrometer. The details of the operation of the active sampling capillary operation and its analytical figures of merit were described in detail in our previous work.20 Here we modified the construction of the plasma source to virtually eliminate electrical interference between the capillary and the analyzed surface. This was achieved by the removing the electrode from the tip of the plasma capillary source, which allowed us to bring the capillary as close as at 20 μm to the analyzed surface. The active sampling capillary consisted of a quartz tube (o.d. 1.5 mm and i.d. 1.0 mm) and two copper pieces that served as electrodes. An inner copper capillary with 0.5 mm i.d. served as the ground electrode. The outer electrode was a copper tape (length 5 mm) surrounding the outside of the quartz tube. The copper tape and the inner tube overlapped by 2 mm. The wall of the glass tube between the copper capillary and the copper tape forms a dielectric barrier. A glass extension of the capillary allows separating the high voltage area from the SNOM head. Here, we further modified the capillary to suppress RF noise, which was found to perturb the feedback of the SNOM head: a metal screen was placed between the high voltage electrodes and the SNOM head to avoid electrical noise (Figure 1). The transport of the analyte was due to the pressure gradient between the ambient pressure region and the first low-pressure region of the mass spectrometer. This pressure gradient generates a gas flow with a rate of 0.8 L/min, as was measured by a gas flow meter (MR3000; Key Instruments, Trevose). SNOM Probes and Near-Field Laser Ablation. For both ablation and optical scanning experiments, we used tips that were “tube etched” according to the procedure described by Stöckle et al.21 A multimode silica fiber with good UV
■
EXPERIMENTAL SECTION Scanning Near-Field Optical Microscopy. To perform MSI at the atmospheric pressure, we used a custom-built experimental setup. A schematic is shown in Figure 1. The MS analysis was performed using an ion trap mass spectrometer (LCQ Deca XP, Thermo, San Jose). Laser ablation was performed with the help of a scanning near-field optical microscope (Lumina, TopoMetrix/Veeco, Santa Barbara, CA). This SNOM instrument was equipped with a noncontact shearforce scan head based on a tuning fork, which was controlled by a special software (SPMLab 6.02) to obtain topographic B
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 2. SEM images of etched tips fabricated by etching in one batch. The exit diameters of the tips is estimated to be (a) 400, (b) 450, and (c) ∼50 nm.
Switzerland) as crystalline powder. All chemicals were used without further purification. For analysis, we pressed pellets of the substances using a hydraulic press with a force of 1 × 104 N, applied for 10 min. To obtain a very smooth surface of the pellets, the material was covered by mica foil that served as a template during the pressing.
transmission (attenuation 70 dB/km at 350 nm) was used as raw material for all SNOM tips produced (Superguide G fiber, High-OH, type SFS50/125Y, multimode UV−vis, 0.22 NA, 50 μm core diameter, 125 μm cladding diameter, 250 μm acrylate jacket, from Fiberguide Industries, Stirling). No metallization of the outside of the tip was done. The tips were glued onto the tuning fork as shown in Figure 1. A Nd:YLF laser (Triton, Spectra Physics, Mountain View, CA) was used for near-field ablation. The laser beam was focused onto the blunt end of the optical fiber. The laser operation was synchronized with the SPMLab software by a pulse generator (model 555, Berkley Nucleonics Corp., San Rafael, CA). Unless otherwise noted, single-shot ablation experiments were carried out in all cases. The SNOM tips were also imaged with a scanning electron microscope. Images of the three tips are shown in Figure 2. Although all of these tips were etched in the same batch and under the same conditions, the tip dimensions vary quite a lot, from ≈50 nm diameter to ≈400 nm. The tip diameter influences the ablation laser power that arrives at the sample surface, which, in turn, defines the spatial resolution and the reproducibility of the experimental data. We did not attempt to achieve a better reproducibility of the tip etching. Instead, we simply took the actual tip shapes into account when evaluating the data. This requires an additional calibration step. Such a calibration was carried out to correlate the particular tip diameter and other parameters such as laser power and SNOM topographic resolution with the parameters of the crater produced by the laser ablation. The optical throughput of the tips was characterized by measuring the output laser power as a function of input laser power. For this, we employed a laser power meter (PE9, Ophir, Jerusalem, Israel) with a maximum allowed energy of 1 mJ/pulse, as well as a more sensitive version when necessary (PD10, Ophir Jerusalem, Israel) with a maximum allowed energy of 200 nJ/pulse. Auxiliary measurements of the crater sizes were also made with an atomic force microscope (NT-MDT, Zelenograd, Russia) Sample Preparation. The near-field laser ablation-MS technique presented here was applied to a range of substances that are normally used as matrices for matrix-assisted laser desorption ionization (MALDI). This choice of substances provides us with the possibility to directly compare the proposed technique with MALDI, which is the leading technique for MSI at ambient conditions with good spatial resolution. The following substances were used: α-4-hydroxycyanocinnamic acid (α-HCCA) and 3-aminobenzoic acid (3AB) were purchased from Sigma-Aldrich (St. Louis, USA) as crystalline powders. 2,5-Dihydroxybenzoic (2,5-DHB) acid was purchased from Acros Organics (Geel, Belgium) as crystalline powder. Sinapic acid was purchased from Fluka (Buchs,
■
RESULTS AND DISCUSSION A key problem in MSI is that a better spatial resolution results in less substance available for MS analysis. Therefore, the sensitivity of the mass spectrometer and the transportation efficiency of the material inside the mass spectrometer can become the limiting factor in performing MSI. In the following, we delineate the quantitative interdependence between the spatial resolution and the sensitivity of SNOM-laser ablation. Based on our findings, we determine a set of the parameters that represent the best trade-off between resolution and sensitivity. We also determine properties that are important for SNOM-MSI under the given set of the parameters. We address the issue of reproducibility, using a correlation between macroscopic parameters (such as the laser power) and microscopic parameters, the most important of which is the shape and size of the ablation crater. Calibration. As mentioned before, we could not fabricate etched SNOM tips with reproducible sizes. The tip size and tip apex shape are the crucial parameters that determine the output laser power, which, in turn, defines the pulse energy delivered to the sample surface and hence the amount of ablated material. Therefore, the optical throughput of SNOM tips was calibrated, because it is a key parameter that determines the ablation characteristics. For each tip, we measured the dependence between the input and output laser energies. Figure 3 shows the results for three different tips used in this work. The optical throughput was found to vary linearly with input laser energy, and lies between 0.3% for the lowest throughput (smallest) tip and 0.9% for the highest throughput tip (i.e., the tip with the largest aperture). Note that there is no single linear dependence on the apex area of the tip and the optical throughput; other factors such as the taper angle near the apex or the precise shape of the end of the tip come into play and influence the optical throughput. Crater Size Determination and Parameter Optimization for SNOM-MS Analysis. The tip shape directly influences the size of the laser spot on the surface, which defines the power reaching the sample and hence the amount of the ablated material. Therefore, an additional calibration parameter that depends on the tip shape, the amount of material ablated by a laser shot, needs to be determined. A reasonable estimate for the quantity of material ablated (in C
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
illustrated in Figure 4. The shot-to-shot reproducibility of the amount ablated, another important methodological parameter, also depends directly on the crater volume. The crater size also gives information about the energy distribution and yields a direct connection between experimental variables (tip shape, input energy) and the ablation (crater size, amount of substance liberated). Here we show results for α-HCCA and the three different SNOM tips whose calibration was shown above, in Figure 3 (Figure 5a). From Figure 5a, one can immediately conclude how important the tip shape is near-field laser ablation using the exact same experimental conditions but different tips results in substantially different amounts of ablated material, because tips with different diameters at the apex produce craters of different volume. The maximum difference is about a factor of 5 for tips 1 and 2 when the same energy was applied (Figure 5a). Another observation is that the amount ablated depends almost linearly on the output laser energy (within the region of interest), which validates our calibration procedure. Error bars were calculated from repeat measurements for the data points shown. From the data in Figure 5, we can also estimate the power density that is necessary for ablation, assuming that the area illuminated by the laser light is equal to the crater fwhm; it amounts to (6 ± 1) × 107 W/cm2. In principle, calibration curves similar to ones shown in Figure 3 would suffice for estimating the amount of substance ablated per shot, if the tip shape is known, for example from electron microscopy images. However, MSI relies on high spatial resolution, thus rendering the crater size, or more precisely, the fwhm more important than the laser output energy. Recalling that both the output energy and the crater size depend on the tip shape, we present results about the dependence of the amount of available material on the crater size. The first fact worth mentioning here is the dependence between the amount of the material and both crater depth and crater fwhm. Near-field laser ablation produces relatively shallow craters, with r ≫ d (r = crater radius, d = crater depth). In the region of interest, the r/d ratio changes from 3 to approximately 10 with increasing laser power. The crater fwhm showed a weaker dependence on laser power than the depth resulting in an increase of the overall r/d ratio. However, in the region of interest, the dependence of the ablated material of different substances increases close to linearly with the fwhm of the crater, as shown in Figure 5b. All three data sets were obtained with tip #2, shown in Figure 2b. Predicting the amount of substance produced in a SNOM ablation event and available for further MS analysis is of great
Figure 3. Optical throughput measured for the three different SNOM tips shown in Figure 2.
fmol) is to use the volume of the crater produced (in μm3), multiplied with the density (g/cm3) and divided by the molecular weight (g/mol). In the next step, we thus determined the ablation crater sizes. To illustrate how this was done, we show an ablation experiment done on a 50 × 50 μm area of a sample, onto which a series of 25 near-field laser ablation shots was fired, with a crater-to-crater distance of 10 μm. This procedure was carried out for all three different MALDI matrices; Figure 4 shows the data for sinapic acid, which was irradiated with a pulse laser energy 85 nJ. To evaluate the data, the craters were approximated as cones, and we measured the full-width at half-maximum (fwhm) and the depth of each crater. The distribution of the crater profiles is presented in Figure 4. We found that the tip shape does not influence the measurements of the craters size in experiments with crater diameters of 1.5 μm and higher. The ablation craters were measured both with the SNOM tips and independently by atomic force microscopy in semicontact mode, using silicon cantilevers. Measurements made by the two methods showed the same crater size (data not shown), indicating that tip convolution effects when using the SNOM to determine the crater topography were negligible for craters larger than 1.5 μm. Under normal experimental conditions, the fwhm varied by no more than 20%. The deviation from the mean of the crater depth was often larger, but never exceeded 50%. This, we postulate, is due to the roughness of the sample and the difficulty to distinguish between the border of the initial surface and hills formed from the redeposited material. This is
Figure 4. 50 × 50 μm area of the surface of a sinapic acid pellet scanned after 25 near-field laser ablation events with 85 nJ pulse laser energy each. (a) 3D view, (b) 2D view, and (c) depth profiles of the uppermost series of craters. The zero line in panel c represents the average height of the sinapic acid surface over the entire 50 × 50 μm scan, after flattening the image. D
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 5. (a) Amount of ablated substance of α-4-hydroxycyanocinnamic acid vs output laser energy for the three tips shown in Figure 3. (b) Amount of ablated substance vs fwhm of ablated craters. Tip #2 was used for all materials in Figure 5 (b). Representative error bars are indicated for the largest craters. These were calculated from the data, using the experimentally derived standard deviation, which was less than 20%, from 10 different measurements.
Figure 6. SNOM image of a four ablation craters generated in an experiment on a sinapic pellet: average depth = 200 nm, fwhm = 2 μm, estimated amount of ablated material ∼ 1.1 fmol.
Figure 7. (a) Selected ion current in the range m/z = 225.2−225.4. The ion current represents the signal intensity as detected by the LCQ mass spectrometer. The peaks correspond to the detection of material produced by single-shot laser ablation events. Laser ablation was done with a repetition rate of 3.5 s−1. Each shot was fired at a new position on the sample surface (see Figure 6). (b) Mass spectrum of the material ablated from position #3 in panel a. The crater of this particular ablation is shown in Figure 6. m/z = 225 corresponds to protonated sinapic acid, m/z = 149 is a background signal.
reproducibility and the limits of applicability of the technique. With the optimization, we aim at reaching two goals at the same time. First, using the plot in Figure 5a, we are able to
methodological importance. First, it is required to determine the macroscopic parameters for achieving the desired resolution and sensitivity. Second, it allows one to estimate the E
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
the speed. The mass spectrum of sinapic acid ablated in the shot #3 is shown in Figure 7). The signal at m/z = 225 Da corresponds to protonated sinapic acid. The main background signal is m/z = 149, which was identified as 4-ethylacetophenone by comparison with reference data.22 This signal was always present in mass spectra of ambient air. The sensitivity is clearly high enough to generate a stable signal from one laser shot: the S/N of signals obtained from all ablation events was about 3, which is postulated to be enough for chemical imaging of surfaces.
estimate the maximal amount of substance available in the SNOM-MS experiment and hence we obtain the minimal sensitivity, which must have a MS instrument to work successfully in the combination with SNOM. Second, we can estimate the amount of material and the spatial resolution of the SNOM-MS combination basing directly on the macroscopic parameters, which are easy to measure and control. SNOM-MS Results. To demonstrate the possibility of robust laser ablation sampling at atmospheric conditions, we performed SNOM-MS with a resolution in the low micrometer range on sinapic acid, 2,5-DHB, and α-HCCA acids pellets for laser ablation followed by the MS signal detection. The material for MS analysis was either produced by single shot event or by a 2 × 2 mesh of shots onto the analyte surface. In all experiments, we measured a topographic image of the surface after acquiring the SNOM-MS signal, to obtain the crater dimensions. For relatively large craters (a few micrometers fwhm), stable and reproducible MS signals with high signal-tonoise ratio were observed. An example of a 6 μm diameter nearfield ablation crater on a 2,5-DHB pellet and the resulting mass spectrum from an estimated amount of ablated 2,5-DHB of about 50 fmol is shown in Figures S1−S3 (Supporting Information). To demonstrate the capability of the SNOM-MS approach to perform high-resolution MSI, data from several ablation events on a sinapic acid sample is presented in Figures 6 and 7. The laser ablation produced craters of about 2 μm fwhm, and depth of about 200 nm. The amount of the ablated material from this crater was estimated to be ≈1.1 fmol (Figure 6). Note that a considerable fraction of the ablated material will be lost in the interface between the SNOM tip and the high vacuum side of the MS. For instance, we have previously shown18 that >50% of the ablated material was deposited on the near-field tip’s shaft. Moreover, part of the ablated material will stick to the walls of the capillary. We can roughly estimate that the maximum amount of the material accessible for MS analysis is ≈0.3 fmol. This small amount of substance was enough to produce reliable mass spectra from a single shot (Figure 7). Laser ablation was done with a repetition rate of 3.5 s−1. The selected ion current in the range m/z = 225.2−225.4 has a direct correlation with the laser shots: roughly, the times at which the laser shots were fired are marked with vertical lines in Figure 7a. We did not synchronize the laser ablation with the acquisition of the mass spectrometer, but the correlation is obvious, e.g., shot #2 corresponds to the peak at 0.12 min. (small peak), shot #3 to the peak at 0.18 min, etc. The mass spectrum is shown for shot #3, with the ion current representing the signal intensity as detected by the LCQ MS. The signal-to-noise ratio for the total ion current is much lower compared to data obtained with larger craters (cf. Figure S2, Supporting Information). This is due to the smaller amount of the ablated material, which scales with the third power of the linear crater dimensions. The ion current peak width is ∼3 s. This width is formed due to substance adsorbed on the walls of the active sampling capillary immediately after the ablation, followed by slow desorption from the capillary walls. Currently, this presents a limitation in the speed that can be achieved for MSI. Substance absorption and following desorption also can cause a shift of a peak maximum with respect to the laser ablation events. However, this issue can be addressed by heating the active sampling capillary, which was not attempted here. Heating should eliminate most of the analyte adsorption, and increase bot the overall sensitivity of the setup as well as
■
CONCLUSIONS In this paper, we presented a technological development of a method to perform high-resolution laser ablation of surfaces under ambient conditions, followed by mass spectrometric analysis of the ablation products. The method is based on the combination of laser ablation via a SNOM tip, and ion production via a custom-built plasma ionization source. We investigated the dependence between the experimental parameters (e.g., tip dimensions, applied laser power) and the analytical characteristics of the method (e.g., size of the craters amount of ablated material) and, in particular, the influence of these parameters on the sensitivity and spatial resolution. We showed that the crater size and therefore the method reproducibility are within 50%. The tip etching procedure used resulted in variety of tip sizes and shapes. This required a tip calibration to be carried out, in order to determine the parameters of the SNOM setup. We showed the applicability of the method for spatially resolved chemical analysis of surfaces of three common MALDI matrices. We achieved a sensitivity of ≈1 fmol, and resolution as low as 2 μm, with a signal-to-noise ratio of ≈3. The mass spectrometer that we used in this work is not state of the art; using a more sensitive machine will certainly better characteristics. We also outlined options to improve the temporal resolution of our method. Finally, scanning sample surfaces with a SNOM tip gives direct access to the sample topography, which is missing in common MSI methods. The technology described here could bring the analysis of surfaces in biology, microelectronics, and related fields to a new level.
■
ASSOCIATED CONTENT
S Supporting Information *
SNOM image of a single crater ablated in 2,5 DHB pellet, mass spectrum before ablation, mass spectrum of ~50 fmol of ablated material, total ion current registered during the single shot ablation of 2.5 DHB pellet, and selected ion current in the range of m/z = 155−155.2. This material is available free of charge via the Internet at http://pubs.acs.org/.
■
AUTHOR INFORMATION
Corresponding Author
*Prof. Renato Zenobi: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Dr. Frank Krumeich (ETH Zurich) for help with recording the electron microscopy images. F
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
■
REFERENCES
(1) Chughtai, K.; Heeren, R. M. A. Chem. Rev. 2010, 110, 3237− 3277. (2) Römpp, A.; et al. Anal. Bioanal. Chem. 2013, 405, 6959−6968. (3) Römpp, A.; Spengler, B. Histochem. Cell. Biol. 2013, 139, 759− 783. (4) Winograd, N. Appl. Surf. Sci. 2003, 203−204, 13−19. (5) Laiko, V. V.; Baldwin, M. A.; Burlingame, A. L. Anal. Chem. 2000, 72, 652. (6) Trimpin, S.; Herath, T. N.; Inutan, E. D.; Wager-Miller, J.; Kowalski, P.; Claude, E.; Walker, J. M.; Mackie, K. Anal. Chem. 2010, 82, 359. (7) Römpp, A.; Günther, S.; Schober, Y.; Schulz, O.; Takats, Z.; Kummer, W.; Spengler, B. Angew. Chem., Int. Ed. 2010, 49, 3834− 3838. (8) Koestler, M.; Kirsch, D.; Hester, A.; Leisner, A.; Guenther, S.; Spengler, B. Rapid Commun. Mass Spectrom. 2008, 22, 3275−3285. (9) Hecht, B.; Sick, B.; Wild, U. P.; Deckert, V.; Zenobi, R.; et al. J. Chem. Phys. 2000, 112, 7761−7774. (10) Shresta, B.; Vertes, A. Anal. Chem. 2009, 81, 8265−8271. (11) Zeisel, D.; Nettesheim, S.; Dutoit, B.; Zenobi, R. Appl. Phys. Lett. 1996, 68, 2491−2492. (12) Stöckle, R.; Setz, P.; Deckert, V.; Lippert, T.; Wokaun, A.; Zenobi, R. Anal. Chem. 2001, 73, 1399−1402. (13) Zoriy, M. V.; Kayser, M.; Becker, J. S. Int. J. Mass. Spectrom. 2008, 273, 151. (14) Zoriy, M. V.; Becker, J. S. Rapid Commun. Mass Spectrom. 2009, 23, 23−30. (15) Ovchinnikova, O. S.; Kertesz, V.; Van Berkel, G. J. Anal. Chem. 2011, 83, 598−603. (16) Ovchinnikova, O. S.; Nikiforov, M. P.; Bradshaw, J. A.; Jesse, S.; Van Berkel, G. J. ACS Nano 2011, 5, 5526. (17) Ovchinnikova, O. S.; Kjoller, K.; Hurst, G. B.; Pelletier, D. A.; Van Berkel, G. J. Anal. Chem. 2014, 86, 1083−1090. (18) Zhu, L.; Gamez, G.; Schmitz, T. A.; Krumeich, F.; Zenobi, R. Anal. Bioanal. Chem. 2010, 396, 163−172. (19) Zhu, L.; Stadler, J.; Schmitz, T. A.; Krumeich, F.; Zenobi, R. J. Phys. Chem. C 2011, 115, 1006−1013. (20) Nudnova, M. M.; Zhu, L.; Zenobi, R. Rapid Commun. Mass Spectrom. 2012, 26, 1447−1452. (21) Stöckle, R.; Deckert, V.; Fokas, C.; Zenobi, R.; Hecht, B.; Sick, B.; Wild, U. P. Appl. Phys. Lett. 1999, 75, 160−162. (22) https://www.waters.com/webassets/cms/support/docs/ bkgrnd_ion_mstr_list.pdf? locale=en_US.
G
DOI: 10.1021/ac504039w Anal. Chem. XXXX, XXX, XXX−XXX
