TOF‐SIMS Applications - mikee9265/SIMS-Wiki GitHub Wiki

The geometry of an organic contaminant strongly affects the chance that an attempt to identify it using TOF-SIMS will be successful. The larger the surface area subtended by a contaminant, the better the chances of obtaining an adequate spectrum. As the surface area gets smaller, problems arise.

1. If you are using a conventional ion source (atomic or small cluster), the primary ion dose you can use to get your spectrum is limited. If the static SIMS limit is exceeded, you may still be able to obtain useful information (although you will no longer be able to trust relative peak intensities), but at some point you will stop getting any unique signals. The surface will have been reduced to an inorganic material bearing little resemblance to its original form. The surface area needed for an analysis is a function of the efficiency of which ions are formed from the contaminant; so success in obtaining an adequate spectrum may depend on what the material is, which is not known at the start of the analysis. The smaller the area covered by the contaminant, the less the chances for a useful result. Submicron contaminants will almost never yield enough signals to enable the analyst to identify the material. Above a micron in radius, chances improve. Unfortunately some materials that produce poor ion yields (well cross-linked polymers come to mind) may require very large areas for identification to be possible. The problem is the competing rate of data acquisition and the rate at which the sample is damaged, both of which are functions of the species being used for the analysis, and not the resolution of the instrument. The good news is that the contaminant’s thickness is not generally an issue for TOF-SIMS. If you have enough surface area, a monolayer is plenty.

2. If you have a large cluster ion source, you can analyze without facing the problem of damage described above. The damaged material is removed as quickly as it is analyzed. As long as you do not exhaust the material, you can keep analyzing. Now, you are limited by your spot size. If the area subtended by the contaminant is only a small fraction of the size of your spot, it may be hard to pick out the contaminant signals from those of the surrounding area. You can instead depth profile the contamination, using an ion beam with a small spot size to take the data, interspersed with the large cluster beam to remove the damage. For contaminants with small areas but significant thickness, this approach may be workable. If the damage produced by the small spot beam is too much, however, this approach may not work. In any case, this method demands that the contaminant be thick enough for enough signals to be obtained before the material is completely etched away.

Other problems that arise in the analysis of unknown organics result from the difficulties analysts have in directly interpreting the results and from the paucity of standard spectra. For some laboratories that work with limited sample sets and for whom the question may be not “what is this contaminant?” but rather, “which contaminant is this?” a private collection of standard spectra obtained locally can be quite sufficient. For other laboratories, an excellent analytical result may still not be sufficient to make the ID. If unique peaks are sufficiently intenese, MS/MS analysis can save the day.

Two other issues in the data analysis include “contamination on top of the contamination” and the problem presented by mixtures. The first issue is a problem exacerbated by the wide range of sensitivities TOF-SIMS has for different organic materials. A submonolayer of poly-dimethyl siloxane (PDMS) can completely dominate the spectrum. Even adventitious hydrocarbons can produce signals far more intense than many polymeric materials. A common example is polymethyl methacrylate (PMMA), a commonly used adhesive. PMMA when prepared as an adhesive, as it is present on many tapes, is a naturally tacky substance. It will collect a thin layer of whatever is in the environment. Unless efforts are made to keep the surface clean, and this is generally not possible for many industrially relevant samples, the SIMS analyst, without the use of large cluster sources, will collect the spectrum of whatever has stuck to the PMMA instead of the well-documented spectrum for PMMA. The second issue is due to the fact that unlike mass spectrometric methods, such as gas chromatograph mass spectrometry (GCMS), the TOF-SIMS has no separation stage. Mixtures at best will produce a combination of signals from the different materials present, requiring a daunting interpretation effort. At worst, one material in the mixture will be preferentially present at the surface, and the true nature of the mixture will be hidden under static SIMS conditions. It is for both these reasons that other analytical methods may be needed.

If there is sufficient material present for a vibrational spectroscopy method (Raman spectroscopy or Fourier transform infrared spectroscopy—FTIR), that method should be tried first (McCreery 1005; Smith 2011). Contamination that is optically visible is quite likely to be amenable to Raman spectroscopy. It generally takes a significant amount of material for FTIR, but the database for this method is even better than that for Raman, and more can be gleaned from the spectrum even if there is no direct match. In cases where TOF-SIMS and vibrational spectroscopy have been tried, but yield different results, the answer obtained via vibrational spectroscopy should be preferred. As noted above, mixtures, and contaminants on top of contaminants can confuse the TOF-SIMS analysis. The vibrational spectroscopy methods generate results from the bulk of the contamination present. If the area with the material of interest is small, AFM-IR should be considered as a complimentary or if necessary, alternative technique.

When vibrational spectroscopy will not work and TOF-SIMS has failed, and you have not damaged the sample too much (or if there are multiple samples, the sign of a true contamination crisis), and if the area subtended by the contaminant is sufficiently large (generally 10 µ in diameter or greater) then X-ray photoelectron spectroscopy (XPS may be of use. The method cannot specifically identify an organic compound, but it can tell a good deal about it.

The sensitivity for inorganic species in SIMS ranges over many orders of magnitude, as has been noted before. Inorganic clusters also do not always directly allow identification of the starting material. However, the TOF-SIMS is a very fast method for the analysis of most inorganic contaminants, and as such, it may be a good starting point for analysis, even if it is not ideal.

When looking at the SIMS ion images of a contaminant, one may find a number of inorganic species associated with the species of interest. The material detected may be sufficient to allow identification, especially in a situation where the potential for contamination is understood. However, strong signals for an element can be misleading if it is a species to which the SIMS is very sensitive. For example, alkali elements (Li, Na, K) may produce very strong images and still represent only a trace level mixed in with a material that is primarily not ionic in nature. The alkalis may give hints as to the origin of the contamination (similar levels of Na and K tend to suggest biological origins, Li is not a common environmental contaminant, samples with either pure Na or K contamination tend to be of industrial origin, etc.) but not to the nature of the bulk of the contaminant. Further, major elements may not be readily detected at all (Zn is an element, for example, which may produce only weak signals in the SIMS depending on the matrix).

Clusters also cannot give clear indications as to the exact nature of a contaminant. As an example, you can readily determine the difference between reduced S (which produces a strong S− signal and no SOx− peaks) and oxidized S (which will produce a series of SOx− peaks including SO2−, SO3−, and HSO4−), but all of these ions will be present in the spectrum of any of the oxidized S anions, and their relative intensities will be affected not only by the species in question, but by the matrix in which they are present and the analytical conditions (primary beam ion and energy) being used for the analysis. Further, the intensities of the entire series can vary widely with the nature of the anion. Sulfide and sulfate are detected with great sensitivity in the TOF-SIMS, but some mixed oxidation states (thiosulfates) produce much weaker spectra.

For small contaminants, Auger spectroscopy is often the method of choice. It may be needed not only in cases where the contaminant is too small for TOF-SIMS but also in cases where the TOF-SIMS analysis is somewhat ambiguous; the added semiquantitative nature of Auger spectroscopy can identify what is most important in the SIMS results, what represents a major element, and what is only present as a trace (Wolstenholm 2015). For contamination that subtends a larger surface area, XPS is generally the preferred method, because in addition to its ability to quantify elements, the “chemical state” information obtained can lead to direct identification of inorganic materials. It should be noted, however, that XPS and Auger have different sensitivities to different elements. The low level of C present in a sample will be blindingly obvious in the Auger while it will produce only a weak signal in the XPS. Whichever is used, the results can “calibrate” the TOF-SIMS analysis, even for qualitative work. Once a contaminant is identified (or at least better characterized) by Auger or XPS, the TOF-SIMS result obtained can be remembered for what it really represents, even if a clean library spectrum is not obtained.

It is often said that TOF-SIMS analyses cannot be quantified, but what is really true is that TOF-SIMS spectra cannot directly yield quantification. Because TOF-SIMS results are highly reproducible for a given sample type and analysis condition, the method can be quantified via the use of standards (Spool and Finney 2024). TOF-SIMS is generally faster and more sensitive than other techniques, so it is often worthwhile to run a few samples through multiple analytical methods or to produce samples with known concentrations in another manner, and then let the TOF-SIMS take over for ongoing routine measurements. This can apply for both organic and inorganic materials.

The caveat is that the “matrix” from which the secondary ions are obtained must be similar. For example, determining the amount of an inorganic contaminant on a silica surface can be readily quantified given a series of standards, as long as those contaminants are not present at so high a level as to change the nature of the surface and therefore, potentially, the ion yields. It also follows, that a given contaminant, present at low levels on the surface, must be present in chemically similar environments to be quantified. As an example, consider two silica surfaces with identical Fe concentrations. The Fe is evenly dispersed so that each Fe atom is surrounded by silica on the first surface. The Fe is present in clusters on the second. ,The ion yields of Fe from the two surfaces will be different. Fortunately, in many cases the source of Fe on a series of related samples will be similar, and therefore, the chemical environment of the Fe will be similar and thus the TOF-SIMS measurement will be reliable. If, for example, contamination is due to the inclusion of Fe in the silica from a sputter process, the chemical environment is likely to be quite consistent. The same standards developed for that analysis may not be as reliably used for samples with Fe contamination from a lapping process.

Auger and XPS are natural allies of TOF-SIMS in the business of producing quantifiable data. Both produce direct measurements of atomic percent. In the case of organic materials, XPS data can be used to produce measurements of layer thickness and monolayer coverage or molecules per unit area. Auger with its different sensitivities and better spot sizes can be ideal for other materials. While the results are “semiquantitative” without themselves being calibrated with standards, the level of accuracy they provide is sufficient to solve many problems. Care must be taken with these methods to avoid electron or X-ray induced damage that may alter the results, however these tools may be all you need to solve problems involving a slow stream of samples. For heavier streams of samples, it is often worthwhile to use Auger or XPS to calibrate the SIMS and let the nimbler technique handle the sample load. Care must be taken when using XPS and Auger to calibrate SIMS results that the relative depths of analysis for the techniques be kept in mind. Static SIMS has produces data from a shallower depth than does Auger, and that shallower than XPS.

In any endeavor, one comes across constructs that work well for the purpose for which they were prepared and those that do not. The laboratory analyst is many times asked to determine what are the significant differences between these “good” and ”bad” samples. In these cases, the samples are usually very similar (they are meant to be identical). Absolute quantification of the materials present at the sample surface is not as important as it is to determine the significant qualitative differences. In other cases, the samples are all believed to be good, but in the interest of quality control, analysis is performed to ensure that there are no significant differences.

Because TOF-SIMS is so reproducible and is highly sensitive to differences in sample surfaces, it is often an ideal method for this type of analysis. Its speed also makes it the first stop in a crisis. TOF-SIMS analysts have directly solved many problems of this type.

Both in the case where there are “good” and “bad” samples, and in the quality control case, a word that is key to the analysis is “significant.” TOF-SIMS can often see differences between samples that are quite real, but which have no technological significance whatsoever. In some cases, the analysis can produce more problems than it solves for the organization involved, because when differences are detected, no matter how trivial, they are difficult to ignore. In cases where low levels of materials to which the SIMS is sensitive are not important, it may be preferable to start the analysis with other methods. In cases where TOF-SIMS analysis finds differences that are likely to be at a low level, a level set by another method like Auger or XPS can be important to determine if the SIMS result is indeed “significant”. On the other hand, the finding of a trace can sometimes be a clue as to what is different in the history of the samples.

Making artificial constructs biocompatible is a major focus of many laboratories, and in this field, TOF-SIMS has been shown to be a powerful tool. As a recent example, the figure below shows the spectra of two contact lenses, one more biocompatible than the other (Clark et al. 2015). The phosphorylcholine moiety, a constituent of the Lens II polymer that is shown in the spectra to be present at the surface, is known to enhance hydration at the surface, and inhibit protein adsorption, a significant cause of discomfort and vision loss.

Source: The results are presented here courtesy of Tascon USA and Tascon Gmbh.

The next figure shows the relative surfaces of the less compatible lens type I when it has not and has been worn for four weeks. Found in the spectra were many changes due to use including the added presence of fatty acid peaks (from lipids), lauryl sulfate, dioctadecyl dimethyl ammonium cation from the disinfecting solution, and low mass N containing species (not shown) indicative of protein adsorption.

Source: The results are presented here courtesy of Tascon USA and Tascon Gmbh.

Beyond the identification of contaminants, there are three other modes of failure analysis that can be done by TOF-SIMS, which are noted here. Low levels of wear not detectable optically can be evident in TOF-SIMS ion images. Traces of materials in a scratch or on a surface can indicate contact, and sometimes what made the contact. Analysis of a surface created by a failed bond can provide important clues as to the reason the bond failed.

Wear is especially apparent in TOF-SIMS images when the surface being examined has a thin coating. In many cases, even partial removal of the coating will result in a large increase in the intensity of signals coming from the underlying material, and the scratch or wear will be clearly revealed. Low energy secondary electron microscopy (SEM) analysis can sometimes also image such wear, especially when the layer and the underlying material differ significantly in average atomic number, and so in these cases the better resolution and often faster analysis in the SEM may be preferred. A common analytical sequence is to try the SEM first and then follow up with TOF-SIMS as needed. Prior SEM analysis can make the TOF-SIMS analysis more difficult. The electron beam in the high (but not ultra-high) vacuum environment of the SEM will deposit inorganic carbon onto the surface. Care should be taken, if further analysis is desired, not to spend more time than is necessary acquiring the images. Energy dispersive X-ray (EDX) analysis, usually taken in point mode with high currents, will definitely place a significant layer of carbon on the surface, obscuring it in the TOF-SIMS. Such layers induced by prior SEM analysis can be removed with the use of large cluster beams, a method that is particularly selective if the materials present are otherwise inorganic.

Scratches, worn surfaces, or even suspected points of contact can be analyzed in the TOF-SIMS to search for clues as to what made the scratch or wear, or simply to confirm that contact was made. If there is a lubricant on a surface, analysis of a contacting surface by TOF-SIMS can reveal where the lubricant transferred, and therefore confirm the contact. An example of this is shown in the figure below. In this case, the contact damaged a magnetic recording disk, and the lubricant definitively showed what disk drive part had damaged the disk. In other cases, contamination that is no longer present can be partially identified by the traces it has left behind. When the scratch or point of wear is very small, Auger may be a preferred method for such an analysis. If the area of interest is large enough, though, the usually higher sensitivity of TOF-SIMS and its speed is preferred.

on a hard disk drive damper from a failed disk drive.

Bond failure may not always be due to adhesive failure. Cohesive failure in one or the other material being bonded often occurs very close to the bond, so that surface analysis methods are needed to distinguish adhesive failure from cohesive failure (Spool 1994). With adhesive failure, the two sides of the failed bond will present differences to a surface analytical method, or else a contaminant will be obviously present. With cohesive failure, despite the fact that the bond looks to have failed at an interface, the two surfaces will look similar to a surface sensitive technique. TOF-SIMS is often preferred for these measurements due to its relative speed. XPS and Auger also do well for this sort of analysis, although, if the failure is due to an organic contaminant, the TOF-SIMS will be more informative.

An interesting example of an adhesion investigation that is both a comparison of “good” versus “bad” and a story of failure analysis involves the study of a pressure sensitive adhesive (Pacholski and Donkus 2015). An additive has a propensity to both migrate to the surface of the adhesive and to oxidize, and this is readily apparent in the SIMS analyses. The figure below shows the spectra from three labels made from the same acrylic adhesive. Shown in the spectra are peaks attributable to the adhesive, to PDMS, and to sulfonate-based surfactants.

Source: The results are presented here courtesy of the Dow Chemical Corporation.

There are multiple causes for reduced adhesion. The presence of PDMS from a silicone-based transfer liner is one, while this figure shows that inhomogeneity of the surfactant is another.

Source: The results are presented here courtesy of the Dow Chemical Corporation.

There are many situations in which surface treatments lead to significant changes in surface properties. The changes may lead to the enhancement of adhesion, a change in surface energy and therefore wettability, and a different tendency to adsorb biomolecules, among many others. The actual changes to the surfaces that these treatments produce may not be obvious just from knowledge of the treatment processes themselves. Understanding how these processes work can be important in targeted development of improved processes.

XPS is often the first stop for a study of the changes a process may produce in a surface, but it is often not chemically specific enough. It is for this reason that TOF-SIMS is often used for these types of study. The combination of the two techniques is often far more powerful than the use of either alone, but, unfortunately, the interpretation of the SIMS results when the treatment does not involve the attachment of specific known chemical species, can be ambiguous.

Tissue samples and even individual cells can be analyzed by TOF-SIMS. Biological systems are chemically complex, and it is clear that TOF-SIMS is sensitive to many lipids, metabolites, small molecules of various types (cholesterol is a good example), and select drugs, but is less sensitive to proteins, DNA, and RNA. Thus with biological samples TOF-SIMS has proven to be useful for a subset of problems of interest to biologists (molecular and other) and the medical community. TOF-SIMS also appeals because of its reproducibility and small spot size holding out the promise of subcellular imaging, and it does not require extensive sample preparation as do matrix-assisted laser desorption ionization (MALDI) and fluorescence microscopy. There is, of course, still the issue of matrix effects, which may make the images that are produced less directly interpretable. The matrix effects in organic samples tend to be less extreme than in inorganic samples, or in the analysis of thin organic materials on varied inorganic substrates. These effects in most biological samples are less than an order of magnitude, but are nonetheless very significant, especially to attempts at quantification.

The use of large cluster beams allows analysis of the entire sample depth (necessary when the material of interest either has a less than ideal ion yield, or when it is present in low concentrations). Species to which the TOF-SIMS is sufficiently sensitive can be imaged in 3D. When a cluster ion gun is used to produce the analysis beam, spot size makes subcellular imaging more difficult, but sensitivity can be enhanced. When ion formation is occurring via protonation, the use of large water clusters, and even more significantly, the use of acidic water clusters (via the incorporation of CO₂ or HCl). When large clusters are used for sputtering and small clusters are used for analysis, a good deal of the sputtered material is lost, and therefore the sensitivity is compromised, but the smaller spot size gives the possibility of sharper images, and the smaller clusters can ionize species that cannot be protonated or produced via a heterogeneous bond scission. It does not help to have a small spot size, however, when the ion yield is poor. These tradeoffs contribute to the limitations that TOF-SIMS faces, especially for biological samples, even with the use of large cluster sources.