ARTICLE pubs.acs.org/ac
TOF-SIMS with Argon Gas Cluster Ion Beams: A Comparison with C60þ Sadia Rabbani,† Andrew M. Barber,‡ John S. Fletcher,*,† Nicholas P. Lockyer,† and John C. Vickerman† †
Manchester Interdisciplinary Biocentre, School of Chemical Engineering and Analytical Science, University of Manchester, 131 Princess Street, Manchester, M13 9PL, U.K. ‡ Ionoptika Ltd., Southampton. U.K. ABSTRACT: Time-of-flight secondary ion mass spectrometry (TOF-SIMS) is an established technique for the characterization of solid sample surfaces. The introduction of polyatomic ion beams, such as C60, has provided the associated ability to perform molecular depth-profiling and 3D molecular imaging. However, not all samples perform equally under C60 bombardment, and it is probably naïve to think that there will be an ion beam that will be applicable in all situations. It is therefore important to explore the potential of other candidates. A systematic study of the suitability of argon gas cluster ion beams (Ar-GCIBs) of general composition Arnþ, where n = 603000, as primary particles in TOF-SIMS analysis has been performed. We have assessed the potential of the Ar-GCIBs for molecular depth-profiling in terms of damage accumulation and sputter rate and also as analysis beams where spectral quality and secondary ion yields are considered. We present results with direct comparison with C60 ions on the same sample in the same instrument on polymer, polymer additive, and biomolecular samples, including lipids and small peptides. Large argon clusters show reduced damage accumulation compared with C60 with an approximately constant sputter rate as a function of Ar cluster size. Further, on some samples, large argon clusters produce changes in the mass spectra indicative of a more gentle ejection mechanism. However, there also appears to be a reduction in the ionization of secondary species as the size of the Ar cluster increases.
S
econdary ion mass spectrometry has become established as a powerful tool for surface analysis and is routinely used for depth-profiling atomic species in the semiconductor industry.1 Static SIMS, in which the primary ion beam dose is restricted so that only 1% of the surface is impacted by a primary ion, allows molecular information to be extracted from the sample that is representative of the chemistry of the sample.2 Frequently coupled to a time-of-flight (TOF) mass analyzer, the technique has been used in a wide range of studies from atmospheric chemistry3 to the imaging of zebra finch brains4 and single cells.5 In the last 10 years, the field has been transformed by introduction of novel ion sources. Initially, heavy atomic ion beams610 were introduced, followed by liquid metal cluster ions of gold11 and bismuth12 for routine SIMS analysis. These ions offered the advantage of a nonlinear increase in the secondary ion yields, particularly at higher mass range.11 Perhaps more significantly, the implementation of the polyatomic ion projectiles, particularly SF513,14and C60,15 have enabled molecular depthprofiling. This allows the concentration of specific “molecular” components to be determined as a function of depth due to the low damage cross section and damage accumulation associated with these beams. As a result, analysis at an ion fluence higher than the static limit could be performed while maintaining molecular information, allowing the prospect of organic molecular depth-profiling and 3D molecular imaging to be realized. Differences in the sputtering mechanism of these beams have been explained through molecular dynamics simulation studies.16 Upon impact, the monatomic and the cluster ions such as Aunnþ r 2011 American Chemical Society
and Binnþ penetrate deep into the sample and deposit energy many layers below the surface. This causes subsurface mixing, causing damage deep into the sample. In comparison, the C60 disintegrates into its constituent atoms as it strikes the sample, and each subsequent carbon, having 1/60th of the particle kinetic energy, creates its own cascade of moving particles. Since the carbon atoms deposit energy very close to the surface, it results in an increase in the amount of material removed by a single ion impact. In addition, this damage, which is limited to the surface region, is substantially removed by subsequent impacts, allowing molecular information to be maintained as a function of primary ion dose. Gillen and co-workers demonstrated molecular depth-profiling in 1998 using a polyatomic, SF5 ion beam,17 and since then, it has been used extensively to depth-profile biopolymers, particularly for characterization of drug eluting medical devices.18 Since its introduction, C60 has also been successfully and widely utilized to depth-profile a series of polypeptide films in trehalose,19 LangmuirBlodgett (LB) multilayer films20 and organic delta layers.21 Similarly, studies completed in the Vickerman lab demonstrated the potential and benefits of the C60 projectile to depth-profile a range of biological systems, including cellulose, which previously had not been possible with SF5þ.22 Received: February 2, 2011 Accepted: April 4, 2011 Published: April 04, 2011 3793
dx.doi.org/10.1021/ac200288v | Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry Although C60 has been recognized as a very effective beam for molecular depth-profiling of numerous organic samples, the analysis of cross-linked polymers and organic electronic devices such as organic light-emitting diodes (OLEDS) are still proving to be difficult with this beam. This is due to the significant accumulation of damage caused when these samples are impacted with the C60 ion beam, which results in rapid loss of the secondary ion signal and a decrease in sputter yields. Various parameters, such as cooling the sample, varying the incident angle, and energy of the primary ion beam as well as rotation of the sample have been explored to improve performance.23 Gas cluster ion beams (GCIBs) were originally developed for sample cleaning and polishing in the semiconductor industry and subsequently modified for application in SIMS by Matsuo and co-workers.24 Ion beams were generated with an average Ar cluster size of 2000 atoms. In early work, the authors used simulation studies to show the potential benefits of utilizing Ar clusters as primary ion particles.25 It was suggested that these benefits included high sputter yields, higher secondary ion signal, and minimal development of ion-beam-induced topography or subsurface chemical damage. Ar-GCIBs have shown promising results in the depth-profiling of organic/biological material, particularly samples that have proved difficult under C60 bombardment. Nimomiya et al. have reported the use of large argon cluster beams for the depthprofiling of organic semiconductor materials.26 Spectra were produced for Alq3 analyzed using monatomic Arþ at 7 keV and Ar700þ at 5.5 keV. The atomic Arþ spectrum is dominated by the aluminum signal with a small peak corresponding to the Alq2 moiety, whereas in the Ar700þ spectrum, it is the Alq2þ (m/z 315) peak that dominates, and the Alþ signal is barely present (S/N ∼ 2). In contrast to the results reported by Shard et al., who performed a similar experiment using C60 primary ions, the Alq2þ signal is stable throughout the depth profile with a flat, steady state signal observed until the substrate is reached when using the large argon clusters, and the erosion of the sample is linearly dependent on the primary ion fluence. Ar-GCIBs have also been successfully used to depth-profile polycarbonate and polystyrene by monitoring diagnostic fragment ions.27 Polystyrene has also proven difficult to profile using C60 because the sample is thought to cross-link under ion beam radiation, leading to a dramatically reduced sputter yield.28 Another interesting observation is that for a range of molecules, the fragmentation pattern changes. Particularly, signals from the low mass fragments seem to be significantly reduced. This “cleaning up” of the spectrum has been shown by Ninomiya and co-workers on small biomolecular species on arginine and the tripeptide Gly-Gly-Gly. In each case, the mass spectrum showed the molecular ion with very little fragmentation when using clusters with energy per Ar atom (Eatom) of 5 and 1.7 eV, respectively.29 We have recently reported on the development of a novel time-of-flight SIMS instrument that allows the analysis to be performed using continuous primary ion beams, the J105 3D Chemical Imager (Ionoptika Ltd., U.K.).30 A linear buncher is used to sample the continuous stream of secondary ions that are ejected from the surface, thus overcoming many of the fundamental limitations of conventional TOF-SIMS instrumentation. A large part of the impetus for the development of this instrument was to fully realize the potential of polyatomic ion beams, such as C60 primary ions for SIMS. The instrument is therefore ideally suited for testing the capabilities of large argon clusters for SIMS analysis. Specifically, it is almost impossible to generate sufficiently
ARTICLE
short pulses of pure Arnþ species to produce useful mass spectrometric data on a pulsed TOF instrument, and as a result, the interest in the conventional TOF-SIMS field is focused primarily on applying these beams for sputter etching. An additional issue is that, to date, the SIMS results that have emerged from measurements using Ar clusters as primary ions have resulted from experiments performed on instruments designed primarily for ion beam development in which the mass spectrometric capabilities have been secondary and not of high quality. By no means does this reduce the importance of the measurements made and the observations that have been reported, but it means that the picture is still far from complete. We are in a unique position in that we can accommodate an Ar cluster gun on our instrument along with a C60 ion gun for comparison and are able to operate both beams without compromising the mass spectrometry. Here, we report on initial studies that focus on the use of Ar cluster beams as primary beams for SIMS and therefore focus on two key claims/hypotheses relating to the application of these beams in SIMS: 1 Damage accumulation during organic depth-profiling is reduced to a level below that observed with C60. This implies that more signal may be available from some samples and that more materials may be amenable to molecular depth-profiling. 2 The spectra are “cleaner”; that is, fragmentation during the secondary ion generation process is reduced, and molecular ions at higher mass are therefore more readily detected.
’ EXPERIMENTAL SECTION TOF-SIMS analysis was performed on a J105 3D Chemical Imager (Ionoptika Ltd. U.K.) mentioned above and described in detail elsewhere.30 Briefly, the instrument uses a continuous primary ion beam to generate a stream of secondary ions that are then sampled by a buncher that produces a tight packet of ions at the entrance to a harmonic field reflectron. We have also recently developed a 20 kV Ar-GCIB ion source in collaboration with Ionoptika Ltd. (Southampton, U.K.). The resulting ion beam system comprises three main chambers: an expansion chamber, an ionization chamber, and the ion optical column, all of which are pumped independently. Formation of neutral clusters occurs in the expansion chamber, which consists of the Laval-style nozzle with a 30 μm aperture. Ar gas is introduced at high pressure and expands adiabatically into vacuum through the nozzle and subsequently cools to generate clusters ranging from 1 to >3000 atoms in size. The nozzle can be adjusted in X, Y, and Z to optimize signal levels for a range of cluster sizes. The beam is collimated via a conical skimmer having an exit aperture diameter of 0.5 mm. The beam then enters the ionization chamber, where it is ionized by electron bombardment and mass-filtered using a Wien filter arrangement, which can be tuned to the species of interest, providing a resolving power (m/Δm) of ∼10. Obviously, for larger cluster sizes (e.g., Ar2000 with a mass of ∼80 kDa), a single species cannot be isolated, but mass resolution is sufficient to easily investigate the change in the interaction of the ion beam with the sample as a function of particle nuclearity. The operating pressure in the ionization chamber is 5 105 Torr. Once the beam of interest has been selected, it is passed through the ion optical column where it is focused and can be rastered over an area of the sample. The beam can be accelerated at different energies ranging from 5 to 20 keV, and presets of different cluster sizes with different energies can be saved and loaded in software to be specified as required. 3794
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry
ARTICLE
Figure 1. Plots of the [M þ H H2O]þ ion from cholesterol (m/z 369.3) as a function of ion dose. Signals have been normalized to their steady state.
The GCIB gun is mounted on the instrument at 45° to the sample, similar to the C60 beam. Our C60 ion gun, which is nominally a 40 kV ion source, was run at 20 kV acceleration voltage to match the Arn beams because the aim of this study was to investigate spectral changes and depth-profiling under different ion beams. All the beams used were defocused to provide a standard beam diameter of ∼30 μm with 80 pA of dc current at the sample (measured in a Faraday cup with 27 V stage bias to suppress secondary electron emission). Hence, the experimental setup remained constant on each sample, with only ion gun selection and corresponding preset (for each Arn species) needing to be changed in the software. Depth-profiling experiments were performed using a single beam (as opposed to the dual beam approach adopted on conventional TOF-SIMS instruments), in which a series of images of the sample (600 600 μm2, 32 32 pixels unless specifically stated otherwise) were acquired using a primary ion fluence 2 1012 ions/cm2 per image. Depth profiles were then generated using data from the central 16 16 pixels in each image, thus avoiding the crater edge effects. Cholesterol films were made by vacuum deposition onto clean silicon wafers in a custom-built vacuum evaporator and depthprofiled using C60 and a range of argon cluster beams. Films appeared uniform and smooth and of a homogeneous blue color. Thin films of angiotensin III were prepared by spin-casting 10 μL of 1 mg/250 μL in methanol at 1000 rpm for 1 min. Irganox 1010 films were prepared by vacuum evaporation/deposition onto Si wafer substrates.
’ RESULTS AND DISCUSSION In terms of biological sample analysis, there have been many studies of different cell and tissue types using SIMS and also numerous studies of cellular mimics, such as lipid films31 and liposomes.32 There are some species that SIMS can detect quite readily, particularly lipids, whereas some species are seldom observed using SIMS, particularly protein and peptide fragments.
Figure 2. Secondary ion yields for [M H]þ (m/z 385) and [M þ H H2O]þ (m/z 369) at steady state relative to C60.
We have selected three compounds of increasing mass as example targets for SIMS analysis of biological samples: cholesterol, dipalmitoylphosphatidylcholine (DPPC), and angiotensin III. We also report on the analysis of Irganox 1010, a polymer additive often used as a “standard” for TOF-SIMS analysis, and a low-molecular-weight polystyrene (PS1000). Cholesterol has been studied by a number of groups as a target for molecular depth-profiling in biological systems and, therefore, makes a good test sample.33 The data presented here were acquired on the same film for all beams. Figure 1 shows depth profiles acquired from cholesterol films using C60, Ar60, Ar200, Ar1000, and Ar2000. In all cases, a steady state was observed following an initial transient region for both [M H]þ and [M þ H H2O]þ ions and those corresponding to the dimer. Further, the fluence required to erode the sample was within 20% for all beams (Figure 1), despite the large degree of variation in particle mass and energy/atom (Eatom). Lee et al.34 have reported on the use of large argon cluster beams for depthprofiling an Irganox multilayer sample and also observed a 3795
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry
ARTICLE
Figure 3. TOF-SIMS spectra of DPPC normalized to the m/z 184.07 ion signal. The initial spectra from a depth profile and spectra from the steady state for experiments using Ar1000þ (a, b) and C60þ (c, d). The spectral region between 400 and 800 has been multiplied by 20.
sputter yield dependence on total kinetic energy of the primary particle with no dependence on the energy per constituent Ar atom. Figure 1 shows that under all the argon cluster beams used in this study, the steady state signal was equal to or greater than the initial signal. Where n > 60, there was considerable increase in signal from the initial bombardment to the steady state. This is in contrast to the situation using C60, where there is an initial decrease in signal intensity before the steady state is reached. A comparison of the cholesterol secondary ion yields for each of the different clusters (relative to C60) is given in Figure 2. C60 delivers overall a higher yield than the heavier argon clusters, although Ar60 and Ar200 produce more of the [M þ H H2O]þ (m/z 369), but less of the [M H]þ (m/z 385) compared with C60. Indeed, C60 delivers a higher yield of [M H]þ than any of the argon clusters and the yields of both secondary ions decrease with increasing argon cluster nuclearity. The C60 beam provides a 100% higher initial ion yield while producing some ion beam accumulated damage; hence, a steady state below the initial intensity. The Ar cluster beams seem to produce little if any chemical damage, although the increase in signal to the steady state suggests that an altered layer is formed and that ionization is more favorable once this layer is formed. It has been shown that bombardment of water ice films using cluster and polyatomic ion beams results in an increase in the protonated (H2O)nHþ species.35 Such a phenomenon may explain what is occurring in the cholesterol films using the large gas clusters, particularly as the enhancement at the steady state is greatest for the species requiring a proton to form. It is noticeable that there is a steady fall in ion yield as argon cluster size increases. The steady state signals during etching indicate a constant erosion rate through the depth profile. Because the silicon substrate is reached following a similar fluence, (20%, with all beams, this erosion rate is very similar for all the ion beams in this study. The reduction in detected secondary ion signal therefore indicates
that ionization efficiency falls as cluster size increases and is probably attributable to the concomitant fall in the energy per impacting atom. Phosphatidylcholine-containing lipids are commonly observed in SIMS analysis of tissue, particularly rodent brain sections, where [M þ H]þ ions are usually detected along with a very strong signal from the lipid headgroup at m/z 184.07. Dipalmitoylphosphatidylcholine, DPPC, has been shown to be quite stable under continued C60þ bombardment, where although there is loss of signal from the characteristic peaks, a steady state is reached. The m/z 184.07 fragment can be routinely depth-profiled and imaged in 3D in single biological cells.5 Figure 3 shows the initial spectra and the spectra from the steady state from a depth profile of a dried droplet sample of DPPC under C60þ and Ar1000þ bombardment. Although the m/z 184.07 moiety is a useful lipid marker, it only identifies the class of phospholipids; high-mass fragments are required to determine the fatty acid chains associated with the intact lipid. Therefore, increasing the signal from the [M þ H]þ ion relative to this would be very useful. Both ion beams produce “good” mass spectra of the lipid, with the [M þ H]þ ion clearly identifiable along with characteristic, structurally significant fragments; however, the Ar1000þ analysis shows an increase in the intensity of the higher-mass ions relative to the phosphocholine headgroup. Analysis with either beam produces a steady state, where the m/z 184.07 peak has reduced in intensity to ∼60% of its starting value. The steady state level for the molecular ion is reduced relative to the m/z 184.07, but is stable and clearly visible in all spectra. Angiotensin III is an oligopeptide consisting of seven amino acids [HArg-Val-Tyr-Ile-His-Pro-PheOH] arranged in a linear chain and plays a significant role in the renninangiotensin system for regulating blood pressure in humans. Large peptides and proteins are not easily detected in SIMS, but small peptides, such as angiotensin or protein fragments produced in the sputtering 3796
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry
ARTICLE
Figure 4. Comparison of standard spectra of angiotensin III acquired with C60, Ar60, Ar500, and Ar1000 at ion fluence of 2 1012 ions/cm2.
process or by enzymatic digestion, are viable candidates for imaging SIMS. Standard spectra of angiotensin III acquired with different beams and beam energies are shown in Figure 4. The fragmentation of oligopeptides can be rationalized using the scheme devised by Roepstorff and Fohlman in 1984 and later modified by Johnson and co-workers in 1987.36 Comparing the spectra acquired with different beams at different energies, it can be seen that the larger clusters, n = 500 and 2000, generate a somewhat “cleaner” spectrum. There appears to be lower fragmentation and reduced detection of background chemical species as compared with analysis with C60þ and Ar60þ. Interestingly, lowering the Ar2000þ kinetic energy to 15 and then 5 keV and, hence, the energy/atom, Eatom, from 10 through 7.5 to 2.5 eV seems to have no significant effect on the degree of fragmentation. Although C60þ and Ar60þ clearly deliver more fragmentation, it is not dramatically greater. The ratio of the fragment m/z 402, the b3-NH3 ion, as a result of cleavage of the peptide bond between
tyrosine (Tyr) and isoleucine (Ile) with the loss of ammonia to [M þ H]þ is less than twice that observed from the Ar2000þ spectrum. On the other hand, upon depth-profiling these films (see Figure 5), it is clear that using C60þ and Ar60þ at 20 keV results in significant chemical damage and the almost complete loss of the signal for [M þ H]þ m/z 931 before the substrate is reached, whereas with Ar2000þ at 20, 15, and 5 keV, the signal decreases significantly but reaches and maintains a quasi-steady state for longer as a function of primary ion dose. The sputter rate falls with the kinetic energy of the beam, but the degree of damage generated also falls; for example, by a fluence of 4 1013 cm2, the signal is 27% of the initial level at 5 keV, whereas it is only 9% at 20 keV. At this stage, it is useful to briefly review the observations on the three biomolecules. Under argon cluster bombardment, in contrast to C60þ, the ion yields from cholesterol rose from the 3797
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry
ARTICLE
Figure 5. Secondary ion intensity of angiotensin III [M þ H]þ m/z 931 plotted as a function of increasing primary ion dose acquired using different beams at different energies. The signal intensity has been normalized to initial signal intensity.
Figure 6. Relative steady state signals for m/z 1177 and m/z 899 from Irganox 1010 and m/z 385 and 369 from cholesterol under 20 keV Arn cluster bombardment as a function of cluster size.
initial spectrum to the steady state level as a function of ion fluence, even for Ar60þ, whose behavior might have been expected to be similar to C60þ. Cholesterol is a small molecule (384 Da), and during argon cluster sputtering, it is clearly relatively easily lifted out of its molecular solid environment with little damage to its molecular structure. Sputter yield appears to be independent of cluster size, although ion formation is sensitive to the chemistry of the sputtering species and the Eatom. When we move on to the larger DPPC molecule (734 Da), while the yield of the molecular ion at steady state using Ar1000þ is greater than with C60þ or Ar60þ, there is still a significant reduction in yield relative to the initial spectrum. Even with the larger cluster and lower energy per atom, it does not seem to be possible to lift the whole molecule out without causing some damage. This is even more evident with angiotensin III (930 Da). Both C60þ and Ar60þ generate significant fragmentation and chemical damage. However, increasing the cluster size to Ar2000 and lowering the energy per atom to 2.5 eV, while beneficial within the range we have presently studied, does not enable the whole molecule to be lifted off with high yield undamaged.
Figure 7. Comparison of the oligomer region of PS1000 spectrum obtained under 20 keV C60, Ar60, and Ar1000 bombardment.
Irganox 1010 films show a trend similar to that observed with cholesterol. A steady state signal level is observed for the molecular ion [M þ H]þ m/z 1177 for n = 500, 1000, and 2000 following an initial increase in secondary ion signal. Ar60þ, however, shows an initial drop in signal, followed by a pseudosteady state with a gradual decline in signal until a sharp drop, when the substrate is reached. As reported by Lee et al.,34 we observe that the sputter rate is again approximately constant under each of the Arn species used at 20 keV energy. As with the 3798
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry cholesterol sample, the signal level at the steady state decreases as a function of particle nuclearity, despite the initial damage observed in the Ar60 profile. Again, due to the similar sputter rate with each beam, this is attributed to a decrease in ionization. Figure 6 shows the yield at the steady state for the molecular ion and the m/z 899 fragment for the different argon beams as a function of cluster size (therefore, proportional to Eatom). The cholesterol data ([M H]þ and [M þ H H2O]þ) is also included to highlight a clear decrease in secondary ion signal as a function of increasing particle nuclearity. Whether ionization occurs via protonation, deprotonation, or fragmentation, the energetics in the impact region is clearly important. A lower Eatom in the sputtering particle may help preserve chemical structure; however, it also seems to reduce ion formation. Moritani et al. 37 have shown that the fragmentation pattern from polystyrene varies as a function of cluster size under Ar n bombardment. Three trends in ion yields were observed as a function of energy/constituent atom (E atom ). Fragment ions from main chain cleavage without fragmentation of the phenyl ring show little change in intensity between the E atom of 22 to 3 eV, but a rapid decrease below 3 eV per Ar atom was observed. Side chain fragmentation including fragmentation within the phenyl ring decreased exponentially with E atom , and small aliphatic fragments, possibly from the terminal groups of the polymer, showed no change as a function of E atom . Here, we show that the oligomer type ions from a sample of PS1000 also change as the impinging ion beam is varied at 20 kV acceleration potential (Eatom = 333 eV for C60 and Ar60 to 20 eV for Ar1000). Figure 7 shows the oligomer region of the PS1000 spectrum under C60, Ar60, and Ar1000 in the first layer of a depth profile (accumulated ion beam fluence of 2 1012 ions/cm2) and the third layer of the depth profile (i.e., after the sample has experienced a fluence of 6 1012 ions/cm2). Characteristic ions are observed under all three beams, with a regular spacing of 104 mass units corresponding to the repeating monomer. However, the specific species observed change both as a function of ion beam (chemistry and energy) and also as a function of fluence. The most prominent peaks in the spectra are the [M þ H]þ ions of the oligomers (labeled in Figure 7 with filled circles) and a second series of peaks 91 mass units below these corresponding to the loss of C7H7 from the molecule (labeled in Figure 7 with filled squares). Of the three beams, Ar1000 shows the least change as a function of ion beam fluence and also shows less fragmentation. The C 60 analysis, in contrast, produces only the ions that have lost the C 7 H 7 group, and these species are quickly lost in the chemical background under continued bombardment. Ar 60 produces an initial spectrum that is of intermediate character between the C 60 and Ar 1000 spectrum in that the most intense peaks are those corresponding to the loss of C 7 H7 from the pseudo-molecular ion, whereas the intact molecular species are still visible. Upon continued bombardment with Ar60, the spectrum changes, and by 6 1012 ions/cm2 resembles the initial C60 spectrum. As we have seen in the case of the biomolecules, the results here demonstrate not only a change in the process leading to secondary ion formation as a function of primary ion beam characteristics but also that these differences lead to further changes to the sample and, hence, to the secondary ion spectrum under increasing primary ion fluence.
ARTICLE
’ CONCLUSIONS A number of clear trends are evident from this initial set of data: 1 On samples for which C60 has previously been shown to be “successful” for depth-profiling (cholesterol, DPPC), there is little change in the secondary ion spectrum when Ar clusters are used, and a steady state was also reached as a function of ion beam fluence. 2 Relative to C60, the degree of damage accumulated at the start of the profile before the steady state is reached decreases under Arn bombardment as the argon cluster size is increased. 3 At constant beam energy, the sputter rate for the different Arn beams on these organic samples appears to be approximately constant as a function of cluster size and, interestingly, is also similar to C60. 4 At constant beam energy, there is an approximately linear reduction in ionization efficiency as the argon cluster size is increased. 5 The mechanisms of secondary ion formation appear to be influenced by the chemistry, energy, and nuclearity of the primary ion. 6 For the biomolecular samples, as molecular mass increases, the ability of C60þ and the argon clusters to sputter the molecular species without damage decreases. While the quality of depth profiles increases with argon cluster size, and there is some evidence of decreased fragmentation, argon clusters cannot be said to universally improve the SIMS analysis capability. At this stage, what are the indications with regard to the potential benefits of using argon cluster ions as primary ions in SIMS analysis? In many cases, hypothesis 1 (see above) is supported, and chemical damage is reduced significantly compared with the smaller polyatomic ions, such as C60þ. This theory has been previously suggested by Matsuo et al. The authors successfully depth-profiled polymers using large gas cluster ions and observed that the accumulation of damage during sputtering was substantially reduced.26 However, the hoped-for benefits of higher yield of larger chemically significant ions may not be realized because the ionization efficiency also falls as cluster size increases and Eatom falls. Surprisingly, the jury is still out on hypothesis 2. Although earlier work by Matsuo et al. on very small molecules, arginine, and the tripeptide Gly-Gly-Gly did report a significant reduction in fragmentation,29 the present work on larger molecules does not suggest that large argon clusters or low Eatom, even down to 2.5 eV, results in a dramatic change in fragmentation compared with smaller polyatomic primary ions, such as C60þ. It is possible that we have not taken Eatom low enough; however, ion beams of very large clusters or very low beam energies are difficult to focus and control and deliver rather low yields, so again, the benefits may be difficult to realize. We are carrying out further studies including imaging to continue our exploration of the benefits of argon cluster beams; however, at this stage, it is difficult to see a clear case that argon cluster beams will replace other primary beams. Clearly, they have a place in dual beam depth-profile studies of organic systems, in which they will be valuable as low-damage, although not zero-damage, sputtering beams. A downside of this approach is that most of the material sputtered is not available for analysis. 3799
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800
Analytical Chemistry
’ AUTHOR INFORMATION Corresponding Author
*john.fl
[email protected].
’ ACKNOWLEDGMENT The authors gratefully acknowledge the financial support of the Engineering and Physical Sciences Research Council, EPSRC, UK under grants EP/C008251 and EP/G045623/1 and a University of Manchester EPSRC Impact Pathfinder Award. The authors also acknowledge Gareth Hampton for producing vapor-deposited films for the study. ’ REFERENCES
ARTICLE
(30) Fletcher, J. S.; Rabbani, S.; Henderson, A.; Blenkinsopp, P.; Thompson, S. P.; Lockyer, N. P.; Vickerman, J. C. Anal. Chem. 2008, 80, 9058–9064. (31) McQuaw, C. M.; Sostarecz, A. G.; Zheng, L.; Ewing, A. G.; Winograd, N. Langmuir 2005, 21, 807–813. (32) Baker, M. J.; Fletcher, J. S.; Jungnickel, H.; Lockyer, N. P.; Vickerman, J. C. App. Surf. Sci. 2006, 252, 6731–6733. (33) Jones, E. A.; Lockyer, N. P.; Vickerman, J. C. Int. J. Mass Spectrom. 2007, 260, 146–157. (34) Lee, J. L. S.; Ninomiya, S.; Matsuo, J.; Gilmore, I. S.; Seah, M. P.; Shard, A. G. Anal. Chem. 2010, 82 (1), 98–105. (35) Conlan, X. A.; Lockyer, N. P.; Vickerman, J. C. Rapid Commun. Mass Spectrom. 2006, 20, 1327–1334. (36) Roepstorff, P.; Fohlman, J. Biomed. Mass Spectrom. 1984, 11, 601–601. (37) Moritani, K.; Mukai, G.; Hashinokuchi, M.; Mochiji, K. Appl. Phys. Expr. 2 2009, 046001.
(1) McPhail, D; Dowsett, M. In Surface Analysis The Principal Techniques, 2nd ed.; Vickerman, J.C., Gilmore, I.S., Eds.; John Wiley and Sons Ltd.: Chichester 2009, Chapter 5. (2) Benninghoven, A. Z. Physik. 1970, 230 (5), 403–417. (3) Fletcher, J. S.; Henderson, A.; Horn, A. B.; Vickerman, J. C. J. Phys. Chem. B 2004, 108, 5960–5966. (4) Amaya, K. R.; Monroe, E. B.; Sweedler, J. V.; Clayton, D. F. Int. J. Mass Spectrom. 2007, 260 (23), 121. (5) Fletcher, J. S. Analyst 2009, 134, 2204–2215. (6) Andersen, H. H.; Bay, H. L. J. Appl. Phys. 1974, 45 (2), 953–954. (7) Wittmaack, K. Surf. Sci. 1979, 90, 557–563. (8) Johar, S. S.; Thompson, D. A. Surf. Sci. 1979, 90, 319–330. (9) Blain, M. G.; Della-Negra, S.; Joret, H.; Le Beyec, Y.; Schweikert, E. A. Phys. Rev. Lett. 1989, 63, 1625–1628. (10) Blain, M. G.; Della-Negra, S.; Joret, H.; Le Beyec, Y.; Schweikert, E. A. J. Vac. Sci. Technol. 1990, A8, 2265–2268. (11) Davies, N.; Weibel, D. E.; Blenkinsopp, P.; Lockyer, N.; Hill, R.; Vickerman, J. C. Appl. Surf. Sci. 2003, 203204, 223–227. (12) Kollmer, F. Appl. Surf. Sci. 2004, 231232, 153–158. (13) Appelhans, A. D.; Delmore, J. E. Anal. Chem. 1989, 61 (10), 1087–1093. (14) Kotter, F.; Benninghoven, A. Appl. Surf. Sci. 1998, 133, 47–57. (15) Weibel, D. E.; Wong, S. C. C.; Lockyer, N. P.; Hill, R.; Blenkinsopp, P.; Vickerman, J. C. Anal. Chem. 2003, 75, 1754–1764. (16) Garrison, B. J.; Postawa, Z. Mass Spectrom. Rev. 2008, 27, 289–315. (17) Gillen, G.; Roberson, S. Rapid Commun. Mass Spectrom. 1998, 12, 1303–1312. (18) Mahoney, C. M. Mass Spectrom. Rev. 2009, 29 (2), 247–293. (19) Cheng, J.; Winograd, N. Anal. Chem. 2005, 77, 3651–3659. (20) Zheng, L. L.; Wucher, A.; Winograd, N. J. Am. Soc. Mass Spectrom. 2008, 19 (1), 96–102. (21) Shard, A. G.; Green, F. M.; Brewer, P. J.; Seah, M. P.; Gilmore, I. S. J. Phys. Chem. B. 2008, 112 (9), 2596–2605. (22) Fletcher, J. S.; Conlan, X. A.; Jones, E. A.; Biddulph, G.; Lockyer, N. P.; Vickerman, J. C. Anal. Chem. 2006, 78, 1827–1831. (23) Sjovall, P.; Rading, D.; Ray, S.; Yang, L.; Shard, A. G. J. Phys. Chem. B. 2010, 114 (2), 769–774. (24) Yamada, I.; Matsuo, J.; Toyoda, N.; Kirkpatrick, A. Mater. Sci. Eng. 2001, 34 (6), 231–295. (25) Toyoda, N.; Matsuo, J.; Aoki, T.; Yamada, H.; Fenner, D. B. Nucl. Instrum. Methods Phys. Res., Sect. B 2002, 190, 860–864. (26) Ninomiya, S.; Ichiki, K.; Yamada, H.; Nakata, Y.; Seki, T.; Aoki, T.; Matsuo, J. J. Rapid Commun. Mass Spectrom. 2009, 23, 3264–3268. (27) Ninomiya, S.; Ichiki, K.; Yamada, H.; Nakata, Y.; Seki, T.; Aoki, T.; Matsuo, J. Rapid Commun. Mass Spectrom. 2009, 23, 1601–1606. (28) Lida, S.; Miyayama, T.; Sanada, N.; Suzuki, M.; Fisher, G. L.; Bryan, S. R. App. Surf. Sci. 2011, 43 (12), 214–216. (29) Ninomiya, S.; Nakata, Y.; Honda, Y.; Ichiki, K.; Seki, T.; Aoki, T.; Matsuo, J. App. Surf. Sci. 2008, 255, 1588–1590. 3800
dx.doi.org/10.1021/ac200288v |Anal. Chem. 2011, 83, 3793–3800