The influence of polyatomic primary ion chemistry on matrix effects in secondary ion mass spectrometry analysis

A.M. Alnajeebi1,2, J.C. Vickerman1,3, N.P. Lockyer1,2*

1Manchester Institute of Biotechnology, The University of Manchester, 131 Princess St., Manchester M1 7DN, U.K2School of Chemistry, The University of Manchester, Oxford Rd, Manchester M13 9PL, U.K3School of Chemical Engineering and Analytical Science, The University of Manchester, Oxford Rd, Manchester M13 9PL, U.K

*Correspondence to: N.P. Lockyer, School of Chemistry, The University of Manchester, Oxford Rd, Manchester M13 9PL, U.K.E-mail:Nick.Lockyer@manchester.ac.uk

ABSTRACTRationale: The application of mass spectrometry imaging techniques to determine 2D and 3D chemical distribution ideally provides uniform, high sensitivity to multiple components and reliable quantification. These criteria are typically not met due to variations in sensitivity due to the chemistry of the analyte and surrounding surface chemistry. Here we explore the influence of projectile beam chemistry and sample chemistry in time-of-flight secondary ion mass spectrometry (ToF-SIMS). To the authors’ knowledge this is the first time the combined effects of projectile chemistry and sample environment on the quantitative determination of mixed samples have been systematically studied.

Methods: Secondary ion yields of lipid and amino acid mixtures were measured under 20 keV C60, Arn, and (H2O)n

cluster ion bombardment (n=2000 or4000) using ToF-SIMS. Ion suppression/enhancement effects were studied in dry sample films and in trehalose and water ice matrices.

Results: The extent of the matrix effects as well as the secondary ion yield was found to depend on the chemistry of the primary ion beam and (for C60, Arn) on the nature of the sample matrix. Under (H2O)n bombardment the sample matrix had negligible effect on the analysis.

Conclusions: Compared to C60 and Arn, water-containing cluster projectiles enhanced the sensitivity of ToF-SIMS determination of the chosen analytes and reduced the effect of signal suppression/enhancement in multicomponent samples and in different sample matrices. One possible explanation for this is that the (H2O)4000 projectile initiates on impact a nanoscale matrix environment that is very similar to that in frozen hydrated samples in terms of the resulting ionisation effects. The competition between analytes for protons and the effect of the sample matrix is reduced with water-containing cluster projectiles. These chemically reactive projectile beams have improved characteristics for quantitative chemical imaging by ToF-SIMS compared to their non-reactive counterparts.

1

Introduction

Secondary ion mass spectrometry (SIMS) is a surface analysis technique which provides detailed chemical characterisation of complex samples using energetic 'primary' ions to sputter surface molecules for mass spectrometric determination [1]. The primary ions can be focused to <1 μm enabling high-resolution microprobe chemical imaging. The introduction of cluster ion beams including SF5, Au3, Bi3, C60 and massive gas clusters extends the sensitivity and mass range of detected analytes and, for the polyatomic projectiles, reduces the chemical damage left in the sputter crater, enabling 3D molecular imaging and depth profiling. These novel ion beams have revolutionized the application of SIMS in the Life Sciences and Biomedicine [1] [2] [3] [4] [5] [6].

A significant challenge faced by SIMS analysts is determining chemical composition in micron-sized pixels, particularly in biological applications (cells and tissues etc) where analytes are chemically diverse and present at trace levels. The power of mass spectrometry imaging techniques, including SIMS, is in measuring subtle chemical heterogeneity, which relies on quantitative and highly sensitive methodology [7]. In SIMS and other techniques, which rely on the ionisation of analytes from condensed phases, in the absence of closely matched standards ion suppression limits quantitative data interpretation. In SIMS this 'matrix effect' arises because the measured secondary ion (SI) yield of an analyte is not only proportional to its surface concentration, but also its ionisation probability, which itself is strongly influenced by the surrounding chemical environment [8]. Jones et al. analysed atropine in cholesterol and in DPPC mixtures with SIMS and laser post-ionisation, simulating the situation of the pharmaceutical distribution between white and grey matter in brain tissue. A significant (~×10) difference in the atropine intensities between the two samples was observed in SIMS, which was negated using laser post-ionisation by decoupling the ionisation from the lipid matrix [9]. Nevertheless, SIMS is increasingly applied in (bio)imaging applications where a straightforward quantitative relationship between signal level and analyte concentration across different sample matrices is implicitly assumed in data interpretation. This raises concerns over potential misinterpretation of analyte distribution.

Sample matrix modification can be applied in SIMS to address matrix effects and increase the ionisation probability. Methods include dispersing the analyte in organic matrices (matrix-enhanced SIMS, ME-SIMS) [10] [11] or deposition of metals such as gold onto a sample (metal-assisted SIMS, MetA-SIMS) [12] [13] [14]. Such approaches can be very successful for surface analysis, but not readily transferred to a depth-profiling experiment or 3D imaging, where the sample is sputter etched in-vacuo during the analysis. An attractive option is to prepare samples for analysis in a frozen-hydrated state where the ice matrix constitutes a potential proton donor. Such methodology lends itself to the analysis of biological samples where rapid freezing fixes hydrated samples for in-vacuo analysis, maintaining chemical distributions in a 'life-like' state [15] [16]. A study conducted on the analysis of frozen samples of pure water and aqueous solutions of alanine, arginine and adenine showed that the yield ratios of protonated water cluster ions to non-protonated water cluster ion were higher in pure water than metabolite solutions suggesting that the protonated water clusters act as a source

2

of protons that enhance [M+H]+ yields of metabolite molecules [17]. Metabolite [M+H]+

yields were shown to be particularly sensitive to the chemical environment in comparison with non-protonated fragment ions [18].

It is now firmly established in the SIMS literature that polyatomic primary ion beams significantly enhance secondary molecular ion yield relative to fragments ions [19] [20] [21]. The distribution of the projectile impact energy amongst constituent species is an important parameter in determining the sputtering characteristics of different sized cluster projectiles. Frequently, sputtered ion/neutral yields are compared to total projectile energy E per constituent atom/molecule in the projectile n [22]. Empirical equations describe the effect of E/n on molecular sputtering [21]. Sputter yield is determined primarily by the total projectile energy whereas the secondary ion yield is found to decrease with increasingly cluster nuclearity. The threshold for secondary ion formation under Arn bombardment appears to be in the region of several eV/atom. Whilst the details of C60 and Arn sputtering have been quite widely reported in the past decade, other polyatomic projectiles have been less well studied. In particular the role that the chemical composition of the polyatomic projectile, and its distinctive sputtering dynamics play in matrix ionisation effects has yet to be determined. Cluster ion beams such as C24H12 [23], (H2O)n

[24], HCl-doped Arn[25], CH4-doped Arn [26] or

water-doped Arn [27] provide potentially reactive species such as protons, which in addition to increasing overall ionisation yields, might influence the suppression/enhancement effects in mixed samples. Biddulph et al. showed that switching from Au to C60 or C24H12 primary ions ameliorated the matrix effects in the analysis of mixtures of 2,4,6 trihydroxyacetophenone (THAP) with cytosine and THAP with barbituric acid (BA) [23]. A recent study of a binary amino acid mixture showed that matrix effects were ameliorated using water-containing polyatomic clusters [28]. There is evidence too from the study of tissue samples that projectile properties play a role in determining matrix effects. In SIMS imaging of brain tissues with metal clusters or C60 beams, the cholesterol intensity is stronger in the white matter in comparison to the grey matter (see for example [29]). With water-containing ion beams i.e. (H2O)6000 and (Ar/H2O)2000, the cholesterol intensity was shown to be similar in the two regions, more closely representing the actual in-vivo concentrations, whereas in the analysis with Ar2000, cholesterol was hardly detected in the grey matter [27]. In another study on brain tissue, Tian et al. applied a dynamic reactive ionisation (DRI) approach in which HCl was doped into an Arn cluster ion beam accompanied by supplying the sample with H2O/D2O vapour [25]. The water-mediated surface acid-dissociation resulted in increased protonated lipid signals and reduced matrix effects compared with Arn projectiles.

Given the current interest in using chemically reactive polyatomic projectiles to enhance SIMS sensitivity and the strong evidence that protonation effects involving the sample matrix and/or the primary projectile influence the SI yields of different analytes to a different degree, this paper explores the implications for quantitative SIMS analysis of biosystems. The study was carried out using three different ion beams C60, Arn and (H2O)n and three different sample preparation methods in order to investigate the effect of the different ion beams and the sample state on the ionisation of standard samples.

3

Experimental

Sample PreparationThe biomolecules used in this study were lipids (dipalmitoylphosphatidylcholine (DPPC) and cholesterol) and amino acids, representing analytes commonly detected in SIMS analysis of biological samples. All materials were purchased from Sigma Aldrich (Gillingham, UK) unless otherwise stated. Stock solutions of DPPC and cholesterol were prepared at 0.30 M in chloroform:methanol (10:1). Uniform films of 1:1 lipid mixtures were made by spin coating 200 μL of equal volume DPPC/cholesterol solutions onto 5 × 5 mm2 Si wafers (Agar Scientific, Essex, UK). Prior to spin coating, wafers were cleaned by sonication for 30 min in hexane, methanol and HPLC-grade water.

A standard amino acid mixture (Standard H, Thermo Fisher Scientific Ltd) was also analysed. Standard H comprises of 18 amino acids; alanine, arginine, aspartic acid, cysteine, glutamic acid, glycine, histidine, isoleucine, leucine, lysine, methionine, phenylalanine, proline, serine, threonine, tyrosine and valine at an identical concentration of 2.5 μM in 0.1 M HCl, except cysteine at 1.25 μM. For the amino mixture, three different sample preparation methods were used to represent typical SIMS analysis conditions for biological samples and to determine the effects of a disaccharide matrix (trehalose) and a more biologically relevant matrix (water ice) which can also act as a proton source. Dry film - 15 μL of sample solution was deposited on a silicon wafer and spin-coated at 800 rpm until dry. Trehalose matrix - the sample was combined in 1:1 ratio in 0.2 M trehalose dissolved in HPLC-grade water and a dry film prepared by spin coating 15 μL at 6500 rpm onto a silicon wafer. Frozen-hydrated - the sample was mixed in 1:1 ratio with HPLC water then 15 μL placed onto silicon wafers and exposed immediately to liquid nitrogen within the glovebox of the SIMS instrument to be frozen and analysed at cryogenic temperature.

SIMS CharacterisationThe SIMS data presented in this paper were acquired on a J105-3D Chemical Imager ToF-SIMS instrument (Ionoptika Ltd, Chandler’s Ford, U.K.). The instrument has been described in detail elsewhere [30]. The primary ion beams can operate in direct current (DC) mode providing secondary ions (SI) continuously, allowing acquisition with a high duty cycle and high lateral resolution. A linear buncher produces a time-focus of the continuous SI fluence at the entrance to a harmonic-field reflectron ToF-analyser. The primary ion beams used in this study were 20 keV C60, Arn and (H2O)n (n = 2000 or 4000). The Arn and (H2O)n beams are formed via adiabatic expansion of Ar (10-20 bar back-pressure) or steam (from a heater water reservoir) through a 30 m aperture into a differentially pumped expansion chamber. The resulting cluster beam is collimated through a conical skimmer and enters a further chamber where electron ionisation occurs via a heated filament. Mass-filtering of the ionized clusters is achieved with a Wien filter providing a resolving power (m/m) of ~10. The operation of the cluster beam source has been described previously [27]. The analysis was carried out in both positive and negative ion mode to access both protonation and deprotonation and its relation to the matrix effect. The C60

+ analysis is generally excluded from the discussion that follows as it has a vastly different E/n and therefore sputter yield and

4

hence signal levels would not be directly comparable with the massive cluster beams in terms of sputtering dynamics. There are however some interesting observations in the C60 data which we will comment on further below.

Data Analysis.SIMS data were processed using Image Analyser software (Ionoptika Ltd) to determine peak areas of selected m/z values. Further data analysis was performed using Microsoft Office Excel. To assess the results quantitatively it was important to avoid any surface contamination or component segregation effects in the samples. Depth profiles were carried out with each ion beam and data taken from regions where the variation of signal intensity with ion dose was minimal. For multi-component samples it was assumed that in this region the sample composition is the same as the preparation composition. Data were averaged over 10 mass spectra after an ion dose 1.5 × 1013 ions/cm2 which established a relatively invariant ion signal, with each subsequent data acquisition corresponding to a primary ion dose of 5 × 1011 ions/cm2 per spectrum. Secondary ion (SI) yields were calculated by determining the number of detected ions at a given m/z per incident primary ion. C60

+ analysis was used to check for any drift in instrument sensitivity between measurements as it was not possible to operate the cluster source with all necessary primary ions in a single analysis session.

To compare total ionisation efficiencies it would be necessary to measure the signal of all fragments resulting from each component and also the component sputter yields. Due to isobaric interferences and the requirement to use mixed samples in this quantification study this is not possible. Our approach therefore is to measure diagnostic ions of each component and determine changes in the relative signal intensity of these ions as ion beam chemistry and sample matrix is changed. Published data from pure samples of arginine and irganox show that the sputter yields (nm3/ion) of Arn and (H2O)n (n=500 and 2000) are within experimental error [24]. Therefore any significant change in the signal intensity we interpret as an ionisation effect.

RESULTS AND DISCUSSION

Lipid mixture analysis

DPPC and cholesterol are lipids which are readily detected during the SIMS analysis of biological specimens such as cells and tissues. DPPC is a polar lipid accounting for 60 mol% of total plasma membrane lipids, and contains a basic phosphocholine headgroup with pKa ~1 [31] . Cholesterol is a neutral, non-polar lipid accounting for 40 mol% of total membrane lipids [32]. Previous SIMS studies have used DPPC and cholesterol as model cell membrane systems and it is known that cholesterol ion signals are suppressed in the presence of DPPC as a result of the matrix effect [33] [34]. When these analytes were segregated by the incorporation of the membrane protein glycophorin A, lipid sensitivity was enhanced [35].

Depth profile analysis of a binary mixture of DPPC and cholesterol showed no sign of significant chemical segregation in the resulting film (Fig S1). Figure 1 shows yields for the

5

DPPC [M+H]+ m/z 734 ion and the cholesterol [M+H-H2O]+ m/z 369 ion from the pure lipids and a (1:1) equimolar mixture, using different primary ion beams, after a pre-etch dose of 1.5 × 1013 ions/cm2. We make a number of observations from these data concerning relative ion yields in different sample matrices and with different primary ion beams.

In comparing the properties of atom clusters and molecular clusters we express the cluster energy in terms of energy per nucleon to account for the different masses of constituent atoms. This approach compares yields on the basis of projectile velocity, rather than kinetic energy per constituent atom/molecule. In this experiment our E/n values are 0.25 eV/nucleon and 0.13 eV/nucleon for the Ar2000 and Ar4000 beams respectively, and 0.56 eV/nucleon and 0.28 eV/nucleon for the (H2O)2000 and (H2O)4000 beams respectively. Although the Ar2000 and (H2O)4000 beams have very similar eV/nucleon, the ion yield from the water beam is enhanced a factor of ×4 for DPPC (m/z 734) and ×18 for cholesterol (m/z 369). The ion yield from the less basic analyte is enhanced to a greater extend by the proton-containing projectile. Of course there are various factors including the sputter yield and fragmentation probability which influence the yield of the secondary ions in question. For the molecular water cluster a fraction of the collision energy is expected to be dissipated into intra/intermolecular bond breaking and vibrational excitation in the projectile, and less is transferred to the sample for fragmentation and sputtering [36]. The corresponding yield enhancement with (H2O)4000 for the m/z 184 fragment is only a factor of ×2 (data not shown). As n increases from 2000 to 4000 the ion yields decrease from both pure samples.

Figure 1: The effect of different 20 keV primary ion beams on SI yields obtained from analysis of DPPC and cholesterol samples, in pure form and in (1:1) mixture.

Ionisation matrix effects become apparent through the change of secondary ion yields on moving from single-component to mixed samples. In our simple model lipid system, the following statement holds true for all projectile beams reported here – comparing ion yields in pure and mixed samples, cholesterol [M-H2O+H]+ is greatly suppressed whilst DPPC [M+H]+ is enhanced in the mixture. The more basic DPPC molecule suppresses the protonation of cholesterol under these conditions. Figure 1 shows the ion yield ratio m/z 734:m/z 369 which provides a quantitative measure of the matrix effects in this lipid mixture sample. Although in all cases the DPPC yield increases from the mixed sample, and the

6

cholesterol yield decreases, the extent of the change is dependent on the primary beam used. Switching from Ar2000 to Ar4000 brings about a small reduction in the ratio – from the pure samples we note that the larger DPPC ion is reduced to a greater extent than cholesterol ion as the projectile E/n decreases. However, a much more significant change in ion ratio (up to an order of magnitude) is observed on switching to the (H2O)n beams. The (H2O)4000 beam delivers an ion yields ratio close to unity for these selected ions.

Multi-component amino acid mixture

This study assessed combinations of different ion beams and sample preparation methods to determine the effects of ion beam chemistry and sample matrix, and the combination thereof. Secondary ion yields were determined from components of the standard H mixture samples prepared as dry films, trehalose-embedded and frozen-hydrated as described above. Depth profile analysis of samples showed relatively constant SI signals after a dose of 4 × 1012

ions/cm2 (Fig S2). This is consistent with previous studies on trehalose using water clusters [24]. As with the DPPC/cholesterol sample, yield data was acquired after a pre-etch dose of 1.5 × 1013 ions/cm2. Component secondary ion yields from samples prepared and analysed under different conditions are summarized in Fig 2. Positive ion data is plotted in order of gas phase basicity (GPB) [37] of the component amino acids, to test the hypothesis that this is a parameter influencing the protonated ion yield response of an analyte under different analysis conditions. The general observation for all ion beams tested is that the high GPB analytes demonstrate the highest ion yield, as would be expected purely on a thermodynamic consideration of this parameter alone. Although the components of the sample are all amino acids, some variation in their sputter yield and stability to ion beam induced fragmentation is to be expected.

Considering the dry film and trehalose-embedded samples Fig. 2 (a & b), the highest ion yields are from the (H2O)4000 ion beam for all analytes except Cys and Gly. In order to focus on the ionisation component of any ion beam effects, it is instructive to examine the extent of SI yield enhancement for (H2O)4000

compared with Ar4000 as a function of analyte GPB, which ranges from ×5 to ×90 in the dry sample (Fig. S3(a)). We assume that the SI yield enhancement is dominated by ionisation effects due to the cluster chemistry rather than energetic effects as these beams have relatively similar E/n and sputter yield. We consider data only for the components with the most significant ion yields (Arg through Val in descending GPB order), which provide more confidence in the calculated ion ratio. The less basic components are enhanced to a greater extent by the (H2O)4000 beam in this sample (Fig S3 (a-b)).

For the frozen-hydrated sample the situation is very different (Fig 2 (c)). Component secondary ion yields are greatest for C60 bombardment (as would be expected due to the high E/n of 20 keV C60), and are very similar between Ar4000 and (H2O)4000 across the range of GPB (see also Fig S3(c)). This suggests that the presence of the water ice matrix mediates to a large extent a similar protonation enhancement to that of the water cluster beam.

7

Examination of the mass spectra from the frozen-hydrated samples confirms that a series of water cluster secondary ions are detected, as reported previously [38], and that the yields of these species, which may themselves provide additional routes to analyte protonation, are greatest for C60 bombardment (Fig S4).

8

Figure 2: Positive secondary ion yield of amino acids obtained by analysis of standard H mixture (a) dry film sample, (b) trehalose embedded sample and (c) frozen-hydrated sample. Analysis performed with 20 keV C60

+, 20 keV Ar4000+ and 20 keV (H2O)4000

+.

9

The influence of the sample state on the component ion yields is more clearly described by replotting the data from each ion beam separately (Fig 3). It is apparent from the general form of these plots that the relative secondary ion yields of different components of the sample is remarkably consistent, regardless of the ion beam or sample state. The relative yields of the selected components are not predicted entirely by GPB values but may be attributed to the other ion formation/stability parameters as noted above. Switching from a pure dry film to a trehalose matrix has little effect on the yields under C 60 bombardment but provides some enhancement (less than ×10) under Ar4000 (Fig S5 (a)). Different sputtering dynamics could result in these analytes benefitting more from matrix-derived protons resulting from the Ar4000 impact, or from the change in analyte-analyte interactions which results from dispersion of the sample in a matrix. Ion yield enhancement from analytes embedded in trehalose have been reported previously [39].

Under C60 or Ar4000 bombardment, the ice matrix provides a very significant ion yield enhancement over dry samples. The enhancement is greater for Ar4000 (up to ×100) than C60

(up to ×20) and generally greatest for the lower GPB components (Fig S5 (b)). Previous studies of cryogenic sample analysis reported that yield is enhanced for a number of molecules analysed in frozen-hydrated and/or low temperature conditions compared to room temperature analysis [17] [40] [41] [42]. This has been attributed to proton donation from sputtered ice. It is noteworthy that for the more basic amino acids (Arg and His) the ion yields from the frozen-hydrated sample using the C60 beam are approximately a factor of 2 greater than the corresponding yields from (H2O)4000 bombardment of dry or trehalose-embedded samples. For the less basic components this factor is increased to ~4.

Importantly, the sample matrix has very little effect on the component ion yields under (H2O)4000 bombardment (Fig 3 (c) and Fig S5), which are within a factor of ×2 between different samples for any given component.

10

11

Figure 3: Positive secondary ion yield of amino acids obtained by analysis of standard H mixture in a dry film, trehalose-embedded, and frozen-hydrated samples. Analysis performed (a) 20 keV C60

+ (b) 20 keV Ar4000+ and (c) 20 keV (H2O)4000

+.

Finally, we compare the [M-H]- negative SI yields from the dry films of Standard H, under different ion beams. These data are plotted against predicted gas phase acidity (GPA) [43] (Fig. 4). No systematic dependence on GPA is observed, and compared with the positive ion data (Fig 2) the SI yields are more consistent across the range of components. The yields from Ar4000 and (H2O)4000 bombardment are within a factor of 2, with the Ar4000 yield the larger of the two. This is consistent with our previously published conclusion [44] that the water beam promotes protonation (and not it would appear deprotonation at E/n = 5 eV/atom). In considering the formation probability and relative stability of positive or negative reactive species from water cluster beams, it may also be relevant than on sputtering ice with C60 the yield of positive ion clusters exceeds that of negative ion clusters by an order of magnitude [38]. Similarly, there may be lower concentrations of ions of the form (H2O)nOH-, (H2O)nO- or (H2O)nOH3

- resulting from shattered water clusters compared to (H2O)nH+ etc. Interestingly, C60 negative SI yields are lower than those of Ar4000 projectiles, which is the opposite of what is seen in the positive ion data. Although both C60 and Arn clusters can be considered chemically unreactive projectiles (unlike water clusters), it would appear for these samples at least, C60 also favours protonation over proton abstraction in comparison to Arn clusters.

12

Figure 4. [M-H]- secondary ion yield of amino acids obtained by analysis of a dry film sample of standard H mixture using 20 keV C60

+, 20 keV Ar4000+ and 20 keV (H2O)4000

+.

Summary and conclusions

In this paper we report on the relative secondary ion yield enhancement of exemplar biomolecules in different matrices using reactive and non-reactive cluster ion beams.

In lipid/lipid binary mixtures water-containing projectiles disproportionately enhance the yield of less basic analytes. The enhancement is greater for larger projectile clusters with a higher water-content. This represents a reduction in the ionisation suppression effect of the more basic analyte. No such effects are evident from bombardment with Arn, which can be considered a non-reactive projectile. The delivery of hydrogen-containing species in the primary ion beam appears to mitigate the ionisation suppression effects between these biologically important lipid molecules, as has been observed directly in the SIMS imaging of brain tissues using these primary ions [45].

In a multi-component amino acid sample prepared as a dry film or embedded in a trehalose matrix, the (H2O)4000

projectile produced the greatest positive secondary ion yields compared to C60 and Ar4000. The yield enhancement over Ar4000

projectiles was greatest for components with the lower gas phase basicity. For the same sample in a frozen-hydrated state the ion yields from the (H2O)4000

and Ar4000 projectile were mostly within experimental error. For C60

and Ar4000 projectiles the sample matrix (dry film, trehalose or water-ice) has a very

significant effect on the positive ion yields, particularly for the less basic analytes, which are greatly enhanced in the ice matrix. However, for (H2O)4000

projectiles the sample matrix has

13

no significant effect on the corresponding ion yields. One possible explanation for this is that the (H2O)4000 projectile initiates on impact a nanoscale matrix environment in the sputter crater that is very similar to that in frozen hydrated samples in terms of the resulting ionisation effects on sputtered analytes.

This study has highlighted the significance of both the sample matrix and the primary beam chemistry in the relative ion yields of mixed biomolecular systems. Reactive projectiles, for example those containing water, provide opportunities not only to increase the sensitivity of the SIMS measurement, but also to mitigate ion suppression effects which currently limit quantitative interpretation of SIMS data. Further work is needed to explore other physical and chemical properties of projectiles including cluster size, energy and composition, in order to extend the capabilities for molecular SIMS analysis.

ACKNOWLEDGEMENTS

The instrumentation used in this work was funded by the UK Engineering and Physical Sciences Research Council, EPSRC, under grants EP/K01353X/1 and EP/C008251/1. The authors thank Sadia Sheraz and Irma Berrueta Razo for assistance with the ToF-SIMS system.

14

Supplementary figures

Fig S1: Depth profile yield of DPPC [M+H] + m/z 734 and fragment [C5H15NO4P]+ m/z 184.07 and cholesterol [M-H2O+H]+ m/z 369 obtained from the analysis of single-component samples of DPPC and cholesterol with different ion beams 20 keV Ar2000

+ and 20 keV H2O2000

+.

15

Fig S2: Depth profile of arginine m/z 175 [M+H] + obtained by analysis of standard H mixture with C60

+, Ar4000+ and (H2O)4000

+ from (a) dry film sample, (b) trehalose matrix embedded sample and (c) ice matrix embedded sample.

16

17

Fig S3: Ratio of secondary ion yield of amino acids obtained by analysis of standard H mixture with (H2O)4000

+ and Ar4000+ from (a) dry film sample, (b) trehalose matrix embedded

sample and (c) ice matrix embedded sample.

Fig S4: Secondary ion yield of water cluster molecules obtained from the analysis of standard H embedded in ice matrix and analysed at cryogenic temperature of -196°C with 20 keV C60

+, 20 keV Ar4000

+ and 20 keV (H2O)4000+. Yields were generated at the steady state region after an

accumulated ion dose of 5×1012 ion/cm2.

18

Fig S5: Secondary ion yield ratio from the standard H mixture (a) embedded in trehalose matrix and (b) ice matrix sample to dry film samples analysed with 20 keV C60

+, 20 keV Ar4000

+ and 20 keV (H2O)4000+.

19

Amino acidAtomic mass(u)

Gas-phase basicity(kJ/mol)

[37]

Gas-phase acidity

(kJ/mol) [43]

Positive ion(m/z) [46]

Negative ion

[M-H]-

(m/z)

Arginine (Arg)C6H14N4O2

174.20 991.60 331.8175.20 [M+H] +

158.17 [M+H-NH3]+

130.19 [M+H-HCO2]+173.19

Histidine (His)C6H9N3O2

155.15 935.96 334.6 156.16 [M+H] +

110.13 [M-HCO2]+ 154.14

Lysine (Lys)C6H14N2O2

146.18 928.01 329.9147.19 [M+H] +

84.13 C5H10N56.08 C3H6N

145.17

Glutamate (Glu)C5H9NO4

147.12 902.07 316.4 148.13[M+H]102.11 [M-HCO2]+ 146.12

Proline (Pro)C5H9NO2

115.13 896.63 340.5 116.13 [M+H] +

70.11 [M-HCO2]+ 114.12

Methionine (Met)C5H11NO2S

149.21 892.44 331.1 150.22 [M+H] +

104.19 [M-HCO2]+ 148.20

Tyrosine (Tyr)C9H11NO3

181.18 891.61 330.4 182.19 [M+H] +

136.17 [M-HCO2]+ 180.18

Phenylalanine (Phe)C9H11NO2

165.18 887.42 330.4 120.17 [M-HCO2]+ 164.18

Threonine (Thr)C4H9NO3

119.11 885.75 324.8 120.12 [M+H] +

74.10 [M-HCO2]+ 118.11

Isoleucine(Ile)C6H13NO2

131.17 881.98 333.5 132.18 [M+H] +

86.15 [M-HCO2]+ 130.16

Leucine (Leu)C6H13NO2

131.17 876.96 333.5 132.18 [M+H] +

86.15 [M-HCO2]+ 130.16

Valine (Val)C5H11NO2

117.14 873.20 333.2 118.15 [M+H]+

72.12 [M-HCO2]+ 116.13

Aspartate (Asp)C4H7NO4

133.10 872.78 315.4 134.11 [M+H] +

116.09 [M+H-H2O]+ 132.09

Serine (Ser)C3H7NO3

105.09 868.59 325.7 106.10 [M+] +

60.07 [M-HCO2]+ 104.08

Alanine (Ala)C3H7NO2

89.09 863.57 224.6 90.10 [M+H]+ 88.08

Cysteine (Cys)C3H7NO2S

121.15 862.74 327.1 122.16 [M+H]+ 120.15

Glycine (Gly)C2H5NO2

75.07 848.09 335.3 76.07 [M+H]+ 74.05

20

Table S1: Positive and negative molecular ions and fragments peaks observed in analysis of amino acids (standard H) and their corresponding gas-phase basicity [37] and acidity [43].

21

Fig S6: Positive ion SIMS spectra acquired from analysis of the dry film of standard H mixture analysed 20 keV C60

+, 20 keV Ar4000+ and 20 keV (H2O)4000

+. (f-* fragment of the amino acid).

References

[1] J. C. Vickerman, D. Briggs. ToF-SIMS: Materials Analysis by Mass Spectrometry, SurfaceSpectra Ltd. 2013.

[2] J. S. Fletcher, N. P. Lockyer, J. C. Vickerman. Developments in molecular SIMS depth profiling and 3D imaging of biological systems using polyatomic primary ions. Mass Spectrom Rev 2011, 30, 142.

[3] M. K. Passarelli, N. Winograd. Lipid imaging with time-of-flight secondary ion mass spectrometry (ToF-SIMS). Biochimica et biophysica acta 2011, 1811, 976.

[4] S. Fearn. Characterisation of biological material with ToF-SIMS: a review. Materials Science and Technology 2015, 31, 148.

[5] N. P. Lockyer. Secondary ion mass spectrometry imaging of biological cells and tissues. Methods in molecular biology 2014, 1117, 707.

[6] C. Bich, D. Touboul, A. Brunelle. Biomedical studies by TOF-SIMS imaging. Biointerphases 2015, 10, 018901.

[7] S. R. Ellis, A. L. Bruinen, R. M. A. Heeren. A critical evaluation of the current state-of-the-art in quantitative imaging mass spectrometry. Analytical and bioanalytical chemistry 2014, 406, 1275.

[8] E. A. Jones, N. P. Lockyer, J. Kordys, J. C. Vickerman. Suppression and enhancement of secondary ion formation due to the chemical environment in static-secondary ion mass spectrometry. Journal of the American Society for Mass Spectrometry 2007, 18, 1559.

[9] E. A. Jones, N. P. Lockyer, J. C. Vickerman. Mass spectral analysis and imaging of tissue by ToF-SIMS—The role of buckminsterfullerene, C60+, primary ions. International Journal of Mass Spectrometry 2007, 260, 146.

[10] K. J. Wu, R. W. Odom. Matrix-enhanced secondary ion mass spectrometry:  a method for molecular analysis of solid surfaces. Analytical chemistry 1996, 68, 873.

[11] F.N. Svara, A. Kiss, T.W. Jaskolla, M. Karas and R.M.A. Heeren. High-Reactivity Matrices Increase the Sensitivity of Matrix Enhanced Secondary Ion Mass Spectrometry. Anal. Chem. 2011, 83, 8308..

[12] L. Adriaensen, F. Vangaever, R. Gijbels. Metal-assisted secondary ion mass spectrometry: influence of Ag and Au deposition on molecular ion yields. Analytical chemistry 2004, 76, 6777.

[13] A. Heile, D. Lipinsky, N. Wehbe, A. Delcorte, P. Bertrand, A. Felten, L. Houssiau, J.-J. Pireaux, R. De Mondt, L. Van Vaeck, H. F. Arlinghaus. Metal-assisted SIMS and cluster ion bombardment for ion yield enhancement. Applied Surface Science 2008, 255, 941.

[14] M. Inoue, A. Murase. Reduction of matrix effects in TOF-SIMS analysis by metal-assisted SIMS (MetA-SIMS). Surface and Interface Analysis 2005, 37, 1111.

[15] S. Chandra, M. T. Bernius, G. H. Morrison, Ion Microanalysis of Frozen-Hydrated Cultured Cells. Springer, Berlin, Heidelberg, 1986, pp. 429–431.

[16] Thomas L. Colliver, Christopher L. Brummel, Michaeleen L. Pacholski, Franklin D. Swanek, Andrew G. Ewing and N. Winograd. Atomic and Molecular Imaging at the Single-Cell Level with TOF-SIMS. Anal. Chem. 1997, 96, 2225.

22

[17] X. A. Conlan, N. P. Lockyer, J. C. Vickerman. Is proton cationization promoted by polyatomic primary ion bombardment during time-of-flight secondary ion mass spectrometry analysis of frozen aqueous solutions? Rapid communications in mass spectrometry, 2006, 20, 1327.

[18] J. S. Fletcher, H. L. Kotze, E. G. Armitage, N. P. Lockyer, J. C. Vickerman. Evaluating the challenges associated with time-of-fight secondary ion mass spectrometry for metabolomics using pure and mixed metabolites. Metabolomics 2013, 9, 535.

[19] C. M. Mahoney. Cluster secondary ion mass spectrometry of polymers and related materials. Mass Spectrometry Reviews 2009, 29

[20] S. Rabbani, A. M. Barber, J. S. Fletcher, N. P. Lockyer, J. C. Vickerman. TOF-SIMS with argon gas cluster ion beams: a comparison with C60+. Anal. Chem 2011, 83, 3793.

[21] M. P. Seah. Universal Equation for Argon Gas Cluster Sputtering Yields. The Journal of Physical Chemistry C 2013, 117, 12622.

[22] A. Delcorte, B. J. Garrison, K. Hamraoui. Dynamics of Molecular Impacts on Soft Materials: From Fullerenes to Organic Nanodrops. Anal. Chem 2009, 81, 6676.

[23] G. X. Biddulph, A. M. Piwowar, J. S. Fletcher, N. P. Lockyer, J. C. Vickerman. Properties of C84 and C24H12 molecular ion sources for routine TOF-SIMS analysis. Anal. Chem. 2007, 79, 7259.

[24] S. Sheraz née Rabbani, I. B. Razo, T. Kohn, N. P. Lockyer, J. C. Vickerman. Enhancing Ion Yields in Time-of-Flight-Secondary Ion Mass Spectrometry: A Comparative Study of Argon and Water Cluster Primary Beams. Anal. Chem. 2015, 87, 2367.

[25] H. Tian, A. Wucher, N. Winograd. Reduce the matrix effect in biological tissue imaging using dynamic reactive ionization and gas cluster ion beams. Biointerphases 2016, 11, 02A320.

[26] A. Wucher, H. Tian, N. Winograd. A mixed cluster ion beam to enhance the ionization efficiency in molecular secondary ion mass spectrometry. Rapid communications in mass spectrometry 2014, 28, 396.

[27] I. B. Razo, S. Sheraz née Rabbani, A. Henderson, N. P. Lockyer, J. C. Vickerman. Mass spectrometric imaging of brain tissue by time-of-flight secondary ion mass spectrometry--How do polyatomic primary beams C₆₀+, Ar₂₀₀₀+, water-doped Ar₂₀₀₀+

and (H₂O)₆₀₀₀+ compare? Rapid communications in mass spectrometry 2015, 29, 1851.[28] A. M. Alnajeebi, J. C. Vickerman, N. P. Lockyer. Matrix effects in biological SIMS

using cluster ion beams of different chemical composition. Biointerphases 2016, 11, 02A317.

[29] P. Sjövall, B. Johansson, J. Lausmaa. Localization of lipids in freeze-dried mouse brain sections by imaging TOF-SIMS. Applied Surface Science 2006, 252, 6966.

[30] J. S. Fletcher, S. Rabbani, A. Henderson, P. Blenkinsopp, S. P. Thompson, N. P. Lockyer, J. C. Vickerman. A new dynamic in mass spectral imaging of single biological cells. Anal. Chem. 2008, 80, 9058.

[31] M. R. Moncelli, L. Becucci, R. Guidelli. The intrinsic pKa values for phosphatidylcholine, phosphatidylethanolamine, and phosphatidylserine in monolayers deposited on mercury electrodes. Biophysical journal 1994, 66, 1969.

[32] X. Han, R. W. Gross. Shotgun lipidomics: Electrospray ionization mass spectrometric analysis and quantitation of cellular lipidomes directly from crude extracts of biological samples. Mass Spectrometry Reviews 2005, 24, 367.

[33] A. G. Sostarecz, D. M. Cannon, C. M. McQuaw, S. Sun, A. G. Ewing, N. Winograd. Influence of molecular environment on the analysis of phospholipids by time-of-flight

23

secondary ion mass spectrometry. Langmuir 2004, 20, 4926.[34] C. Bich, R. Havelund, R. Moellers, D. Touboul, F. Kollmer, E. Niehuis, I. S. Gilmore,

A. Brunelle. Argon Cluster Ion Source Evaluation on Lipid Standards and Rat Brain Tissue Samples. Anal. Chem. 2013, 85, 7745.

[35] M. J. Baker, L. Zheng, N. Winograd, N. P. Lockyer, J. C. Vickerman. Mass spectral imaging of glycophospholipids, cholesterol, and glycophorin a in model cell membranes. Langmuir  2008, 24, 11803.

[36] A. Delcorte, M. Debongnie. Macromolecular Sample Sputtering by Large Ar and CH4

Clusters: Elucidating Chain Size and Projectile Effects with Molecular Dynamics. The Journal of Physical Chemistry C 2015, 119, 25868.

[37] A. G. Harrison. The gas-phase basicities and proton affinities of amino acids and peptides. Mass Spectrometry Reviews 1997, 16, 201.

[38] X. A. Conlan, J. S. Fletcher, N. P. Lockyer, J. C. Vickerman. A Comparative Study of Secondary Ion Emission from Water Ice under Ion Bombardment by Au+ , Au3

+ , and C60

+ . The Journal of Physical Chemistry C 2010, 114, 5468.[39] J. Cheng, N. Winograd. Depth profiling of peptide films with TOF-SIMS and a C60

probe. Anal. Chem. 2005, 77, 3651.[40] T. P. Roddy, J. Donald M. Cannon, Sara G. Ostrowski, A. Nicholas Winograd, A. G.

Ewing. Identification of Cellular Sections with Imaging Mass Spectrometry Following Freeze Fracture. Anal Chem 2002, 74, 4020.

[41] Christine M. Mahoney, Albert J. Fahey, Greg Gillen, A. Chang Xu, J. D. Batteas. Temperature-Controlled Depth Profiling of Poly(methyl methacrylate) Using Cluster Secondary Ion Mass Spectrometry. 2. Investigation of Sputter-Induced Topography, Chemical Damage, and Depolymerization Effects. Anal. Chem. 2006, 79, 837.

[42] A. M. Piwowar, N. P. Lockyer, J. C. Vickerman. Salt Effects on Ion Formation in Desorption Mass Spectrometry: An Investigation into the Role of Alkali Chlorides on Peak Suppression in Time-of-Flight-Secondary Ion Mass Spectrometry. Anal. Chem. 2009, 81, 1040.

[43] M. L. Stover, V. E. Jackson, M. H. Matus, M. A. Adams, C. J. Cassady, D. A. Dixon. Fundamental Thermochemical Properties of Amino Acids: Gas-Phase and Aqueous Acidities and Gas-Phase Heats of Formation. The Journal of Physical Chemistry B 2012, 116, 2905.

[44] S. Sheraz née Rabbani, A. Barber, J. S. Fletcher, N. P. Lockyer, J. C. Vickerman. Enhancing Secondary Ion Yields in Time of Flight-Secondary Ion Mass Spectrometry Using Water Cluster Primary Beams. Anal. Chem. 2013, 85, 5654.

[45] I. Berrueta Razo. Molecular imaging of mouse brain tissue using Cluster Time-of-Flight Secondary Ion Mass Spectrometry. The University of Manchester, Manchester, UK, 2015.

[46] N. T. Samuel, M. S. Wagner, K. D. Dornfeld, D. G. Castner. Analysis of Poly(amino acids) by Static Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS). Surface Science Spectra 2001, 8, 163.

24