Sensitivity enhancement using chemically reactive gas cluster ion beams in secondary ion mass spectrometry (SIMS)

We report for the first time on significant molecular secondary ion yield increases by modifying the chemistry of a water cluster primary ion beam. This was demonstrated using 70‐keV ion beams of 0.15 eV/amu. For the neutral drug Bezafibrate, secondary ion yield enhancements ×5–10 were observed when replacing the Ar carrier gas in a water gas cluster ion beam (GCIB) source with a mixture containing 12% CO2 and 2% O2 in Ar. For the cationic drug Ranitidine, the ion yield enhancements using the CO2‐containing carrier gas were up to ×20–50 in positive mode and ×2–4 in negative mode. The extent of molecular fragmentation was very similar from both cluster beams. We conclude that additional chemically reactive species are present in the impact zone using the (H2O/CO2)n projectile, which promote the formation of secondary ions of both polarity through projectile impact‐induced chemical reactions. This methodology can be applied to further extend the capabilities of high‐resolution 3‐dimensional mass spectral imaging using reactive GCIB‐SIMS.


| INTRODUCTION
The capabilities of secondary ion mass spectrometry (SIMS) for (bio) organic analysis have been greatly extended by the introduction of massive gas cluster primary ion beams (GCIBs). [1][2][3][4][5][6][7] For a recent review on the subject of cluster beam interactions with surfaces, see Delcorte et al. 8 Sputtering of organic species by keV projectiles containing thousands of constituent atoms is characterised by the emission of intact molecules and low residual damage to the sample. This is thought to result from the localised energy deposition from the cluster projectile and the low internal energy of sputtered molecules. 9 Importantly, this facilitates molecular depth profiling and 3D imaging, whereby underlying molecular layers are revealed in a stepwise manner as a function of primary ion fluence. 10 Seah and co-workers developed a generic description of molecular sputter yields under GCIB projectiles including angulardependence and the importance of projectile energy (E) and cluster size (n). [11][12][13] For SIMS analysis, sputtered secondary ions (SI) are required, and their internal energy and probability of ionisation is dependent on the velocity of the primary cluster. 14 To compare ion velocity of clusters with constituents of different chemistry (mass), it is convenient to refer to projectile energy per nucleon or E/m rather than E/n. For atomic cluster beams of a given E, a more massive cluster with lower E/m results in reduced damage but also lower SI yield. The first GCIBs applied in SIMS, and still the most prevalent, consist of Ar n clusters. 15 However, the current density of Ar n GCIBs remains low in comparison to liquid metal beams or C 60 and 3D molecular imaging at (sub)micrometer resolution often employs a dual-beam approach 16 where the low-damage GCIB sputtering is interleaved with a high-resolution/high-damage analytical beam, for example, Bi 3 To optimise efficient 3D molecular SIMS imaging applications using a single beam approach, a number of groups have investigated methods to increase the SI yield under GCIB bombardment. A strategy which shows particular promise, is to change the chemistry of the GCIB such that ionisation is promoted within the impact crater or the sputter plume. This methodology minimises sample modification and provides enhancement specifically at the impact site throughout a depth-profile or 3D image. An analogous approach is used in atomic SIMS analysis using Cs + or O 2 +/À projectiles, but these beams are destructive of molecular samples. The principle of surface-projectile chemical reactions with GCIBs has been established in the field of inorganic materials processing-a methodology termed reactive accelerated cluster erosion replaces Ar n projectiles with (SF 6 ) n , (CO 2 ) n , and so forth to facilitate chemical reaction with the surface and the formation of volatile products. 17 Combining the ionisation benefits of H 2 O in sample environments [18][19][20]  2 | METHODS

| Sample preparation
SIMS is increasingly being applied to determine the localisation of pharmaceuticals within biological systems including single cells and tissues, where the limit of detection determines the level of spatial resolution that can be obtained. We therefore chose two drug compounds for this study, with a significant difference in physico-chemical properties.

| ToF-SIMS characterisation
ToF-SIMS measurements were performed on a J105 3D Chemical Imager (Ionoptika Ltd, UK) described previously. 30 A 70-keV GCIB system was employed, equipped with a source which is capable of producing primary cluster projectiles of Ar n , (Ar/CO 2 ) n , or (H 2 O) n with n ≤ 50 k. Clusters are formed by adiabatic expansion from a highpressure region into vacuum and ionised by electron bombardment.  We did not measure sputter yields specifically in this study, but previous work has indicated that E and E/n are the principal parameters determining sputter yield 11 and these values are identical for both GCIBs in our study. The Ranitidine depth profiles ( Figure 1B)   Molecular Dynamics simulations have provided many insights into the mechanisms of the interaction of (poly)atomic projectiles with surfaces (see for example a recent review on GCIB impacts by Delcorte et al. 8 ). An important parameter is the ratio of the projectile energy impact energy E is greater (70 keV vs. 40 keV), which will increase the energy imparted to the surface and available for chemical reactions.
The details of the energy deposition mechanism remain uncertain and require sputter yield measurements in tandem with molecular dynamics simulations 35 to further understand. predicting similar sputter yields for these projectiles, we infer that impact-induced chemical reactions involving CO 2 are responsible for an increased ionisation efficiency.
Our observations suggest the charge state of the sample prior to analysis affects the magnitude of this effect which may differ in each polarity. This work demonstrates the potential for further significant SIMS performance gains using reactive GCIBs. SI yield enhancements over an order of magnitude have been observed in this study, while changes in the relative fragmentation of secondary ions are within a factor of two. Further studies are needed to better understand the relationship between analyte/cluster beam physico-chemical properties and SIMS performance using novel GCIB projectiles. Although we limit our discussion in this paper to Ranitidine and Bezafibrate, similar SI yield enhancement have also been observed with Acetaminophen and Diclofenac and these will be reported as part of a wider study into matrix and primary beam effects on quantification in GCIB-SIMS.