Anal. Chem. 2009, 81, 1040–1048
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 Alan M. Piwowar,* Nick P. Lockyer, and John C. Vickerman Surface Analysis Research Centre, Manchester Interdisciplinary Biocentre, CEAS, The University of Manchester, Manchester M1 7DN, United Kingdom In secondary ion mass spectrometry, the molecular environment from which a sample is analyzed can influence ion formation, affecting the resulting data. With the recent surge in studies involving examination of biological specimens, a better understanding of constituents commonly found in biological matrixes is necessary. In this article we discuss results from an investigation directed at understanding the role of salts doped as alkali chlorides in a model biological environment, arginine. The data show that addition of salt to the model system causes ion suppression of all the major mass spectral peaks attributed to arginine, with KCl having the largest suppression effect. Potential causes for the suppression effects are briefly discussed in relation to collected data. These theories include sample degradation, formation of salt adduct peaks, and anion neutralization. Investigation of the arginine salt data in comparison with data collected from pure salt systems indicates that suppression of the positive secondary ions is likely caused by a neutralization process involving the salt counteranion, chloride. To address the suppression issue, various procedures were performed on the arginine films such as sample washing with a cleaning solution (ammonium formate, ethanol, water) and analysis of films in a frozen-hydrated state. We present data from the analysis of the frozen-hydrated samples that shows both an ion yield enhancement and a significant amelioration of the salt suppression effects when compared to the samples run under standard conditions, demonstrating that it is a helpful approach to dealing with salt suppression. Secondary ion mass spectrometry (SIMS) is a surface sensitive technique that provides spatially specific chemical information from a vast number of materials and surfaces.1 Recently the technique has been used in an increasing number of investigations of biological samples, due mainly to its chemical imaging capabilities and to the availability of polyatomic primary ion sources.2-6 This current drive for biomolecular imaging and analysis has even * To whom correspondence should be addressed. E-mail: alan.piwowar@ manchester.ac.uk. (1) ToF-SIMS - Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; SurfaceSpectra and IM Publications: Chichester, 2001. (2) Nygren, H.; Malmberg, P. Trends Biotechnol. 2007, 25, 499–504.
1040
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
led to the recent development of novel instrumentation.7 Typically, the analysis of biological samples involves the examination of tissue or cellular samples which consist of a large number of molecular and atomic species, including salts. Both the yield of secondary ions characteristic of an analyte and the SIMS spectrum of an analyte are strongly influenced by the molecular environment from which it is analyzed.8-10 This molecular environment or “matrix” can increase, decrease, or have no effect on the ionization process, depending on the analytes present. Matrix effects which suppress or enhance ion formation place severe limits on the quantitative and qualitative ability of SIMS analysis of biological samples. Although a fully comprehensive theory for this interaction has yet to be confirmed, the ion suppression/enhancement effects observed in SIMS analysis have been attributed in part to the competition for protons between sputtered analyte molecules, which is related to the gas phase basicity of the analyte.8 Other potential ion suppression/enhancement interactions involving analytes and biologically relevant salts have only recently been examined in detail despite their presence in the majority of biosamples. Jones et al. identified a correlation between the presence of native salt levels in rat brain tissue section with decreased analyte signal intensity.11 Data from a depth profile of that tissue section showed that with increasing dose density, the Na+ and K+ peaks and probable salt adduct peaks dominate the spectra while analyte signal intensity from common protein and lipid fragments decrease dramatically. When salt levels were reduced after washing with a solution of ammonium formate, peak intensities were vastly improved indicating that salts were at least partially responsible for suppressing ion formation. (3) Fletcher, J. S.; Lockyer, N. P.; Vaidyanathan, S.; Vickerman, J. C. Anal. Chem. 2007, 79, 2199–2206. (4) McDonnell, L. A.; Heeren, R. M. A. Mass Spectrom. Rev. 2007, 26, 606– 643. (5) Parry, S.; Winograd, N. Anal. Chem. 2005, 77, 7950–7957. (6) Griffiths, J. Anal. Chem. 2008, 80, 7194–7197. (7) Fletcher, J. S.; Rabbani, S.; Henderson, A.; Blenkinsopp, P.; Thompson, S.; Lockyer, N. P.; Vickerman, J. C. Anal. Chem. 2008, in press. (8) Jones, E. A.; Lockyer, N. P.; Kordys, J. J. Am. Soc. Mass Spectrom. 2007, 18, 1559–1567. (9) Jones, E. A.; Lockyer, N. P.; Vickerman, J. C. Appl. Surf. Sci. 2006, 252, 6727–6730. (10) Biddulph, G. X.; Piwowar, A. M.; Fletcher, J. S.; Lockyer, N. P.; Vickerman, J. C. Anal. Chem. 2007, 79, 7259–7266. (11) Jones, E. A.; Lockyer, N. P.; Vickerman, J. C. Anal. Chem. 2008, 80, 2125– 2132. 10.1021/ac8020888 CCC: $40.75 2009 American Chemical Society Published on Web 01/06/2009
The presence of salt in samples analyzed with mass spectrometric techniques has also been shown to provide some beneficial effects. Cationization of intact molecular species by the addition of alkali, alkali earth, or transition metals can aid in molecular weightdeterminationorchemicalidentificationinSIMSanalysis,12-18 while Matrix-Assisted Laser Desorption Ionization (MALDI) experiments routinely utilize the presence of cations to form high molecular weight polymeric ions.13,19,20 Although addition of salt to a known sample material to enhance cationization is easy to implement and a suitable method for some samples, the detrimental effects of its presence in a sample mixture is not often taken into account. Also, the presence of salts in biological samples can potentially spread signal intensity into many different spectral peaks via the formation of salt adducts, decreasing the intensity of characteristic protonated ions and complicating spectral interpretation.13,20 One procedure for addressing the negative effects of the presence of salts in tissue sections or cell cultures is to wash the sample with a salt removal solution such as ammonium formate, ethanol, or water.21-23 However, total removal of salts and salt cations may limit the validity of a biological investigation. The biological integrity of some samples is judged on the basis of the observed peak ratio of K+/Na+. A ratio ∼10:1 for intracellular regions suggests the structural integrity of a biological specimen has been maintained through sample preparation and introduction into ultrahigh vacuum.24 Additionally, washing with a salt removal solution may eliminate or relocate some soluble analytes from the biomaterial, altering the analysis area. Therefore, a means of addressing salt suppression issues without chemically removing the salts present, such as analysis of frozen-hydrated samples, may provide a more suitable means of addressing the issue. Currently our understanding of salt effects in Time-of-FlightSecondary Ion Mass Spectrometry (ToF-SIMS) is limited.11 Because of the presence of salts in almost all biological samples, a better understanding of their effects on analyte ion formation is crucial. In this article, we discuss data collected from experiments in which we examined the effect of doping a series of alkali chlorides into an amino acid matrix to observe the effects of the salts on the formation of characteristic ions from that biomaterial. In addition, a series of procedures for addressing the salt issue including washing the sample with a salt removal solution and (12) Hagenhoff, B. In ToF-SIMS Surface Analysis by Mass Spectrometry; Vickerman, J. C., Briggs, D., Eds.; SurfaceSpectra and IM Publications: Huddersfied, 2001, pp 285-308. (13) Fujii, T. Mass Spectrom. Rev. 2000, 19, 111–138. (14) Delcorte, A.; Bertrand, P. Anal. Chem. 2005, 77, 2107–2115. (15) Gusev, A. I.; Choi, B. K.; Hercules, D. M. J. Mass Spectrom. 1998, 33, 480–485. (16) Grade, H.; Cooks, R. G. J. Am. Chem. Soc. 1978, 100, 5615–5621. (17) Cornelio-Clark, P. A.; Gardella, J. A. Langmuir 1991, 7, 2279–2286. (18) Linton, R.; Mawn, M.; Belu, A.; Desimone, J.; Hunt, M.; Menceloglu, Y.; Cramer, H.; Benninghoven, A. Surf. Interface Anal. 1993, 20, 991–999. (19) Nielen, M. W. F. Mass Spectrom. Rev. 1999, 18, 309–344. (20) Zhang, J.; Zenobi, R. J. Mass Spectrom. 2004, 39, 808–816. (21) Fernandez-Segura, E.; Canizares, F. J.; Cubero, M. A.; Campos, A.; Warley, A. J. Microsc. 1999, 196, 19–25. (22) Sjovall, P.; Lausmaa, J.; Johansson, B. Anal. Chem. 2004, 76, 4271–4278. (23) Berman, E. S. F.; Fortson, S. L.; Checchi, K. D.; Wu, L.; Felton, J. S.; Wu, K. J. J.; Kulp, K. S. J. Am. Soc. Mass Spectrom. 2008, 19, 1230–1236. (24) Chandra, S.; Morrison, G. H. Int. J. Mass Spectrom. Ion Processes 1994, 143, 161–176.
analysis of samples in a frozen-hydrated state were performed in an effort to find an efficient means of eliminating the unwanted salt effects. EXPERIMENTAL SECTION Materials and Sample Preparation. L-Arginine (SigmaAldrich, UK) was dissolved in Chromasolv HPLC grade water (Sigma-Aldrich, U.K.) as a 0.3 M solution and run as a pure film, or with 1% mol/mol addition of one of following: LiCl, NaCl, KCl, or CsCl (Sigma-Aldrich, U.K.). Standard films were made by pipetting 7 µL of solution onto a cleaned (wash with HPLC grade water, analytical grade ethanol (Fisher Scientific, U.K.), and 0.15 M ammonium formate (Sigma-Aldrich, U.K.)) 0.5 × 0.5 cm2 silicon shard (Agar Scientific, Essex, U.K.) and left to evaporate in a fume hood for approximately 25 min before introduction into the sample analysis chamber. Reference salt samples were prepared by pressing a thick film of the various salts into indium foil. For the washing procedures, the dried films were dipped into one of the following cleaning solutions for 5 s: 0.15 M ammonium formate, analytical grade ethanol, or HPLC grade water. Frozen-hydrated films were made in an identical manner to the standard films with the exception that the films were not dried in a fume hood. Instead, they were immediately fixed to a sample analysis stub and plunged into liquid nitrogen until frozen. They were then transferred to a sample stage held at approximately -170 °C. ToF-SIMS Analysis. ToF-SIMS analysis was performed using a Bio-ToF SIMS instrument, the design of which has been described previously.25 Data was collected using a 20 kV C60+ ion gun (Ionoptika Ltd.) with a 30 ns pulse width and a beam current of approximately 0.2 nA rastered over a 200 × 200 µm2 area. The spectra consisted of 100,000 shots for each acquisition and had an average mass resolution of 650 at m/z 28. Secondary ion yields reported correspond to the number of secondary ions detected per primary ion impact. Data was collected from several regions on the film to provide an error assessment of the values presented. Low energy (25 eV) electrons were flooded onto the sample between primary ion pulses to limit any effects from sample charging. The sample stage was held at ground during ion impact, and the secondary ions were directed into a two-stage reflectron by applying a delayed extraction pulse of 2.5 kV to the stage. The ions were postaccelerated to 20 keV and detected using a dual microchannel plate assembly with the flight times being recorded on a 1 ns time-to-digital converter (Fast Comtec, GmbH). Depth profiling data was collected using a 20 kV C60+ beam current of approximately 0.3 nA. Etching of the film took place over an area of 620 × 620 µm2, and information was collected from a 200 × 200 µm2 area inside of the etch crater. RESULTS AND DISCUSSION The purpose of this study is to form a better understanding of peak suppression issues involving salts within biological samples. To that end, we set out to examine the two most common biological cations Na+ and K+ as alkali chlorides to determine (25) Braun, R. M.; Blenkinsopp, P.; Mullock, S. J.; Corlett, C.; Willey, K. F.; Vickerman, J. C.; Winograd, N. Rapid Commun. Mass Spectrom. 1998, 12, 1246–1252.
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
1041
Figure 1. Secondary ion yield comparison for m/z 175 (dots), m/z 87 (horizontal lines), m/z 70 (checkered pattern) in an undoped arginine film and in films in which arginine was doped with 1% (mol/ mol) of LiCl, NaCl, KCl, and CsCl, collected using a 20 keV C60+ primary ion source. The numbers located above the bars correspond to the percent of secondary ion yield observed relative to the undoped arginine film.
what effect they may have on ion suppression in ToF-SIMS analysis. In addition, samples doped with LiCl and CsCl were also investigated to provide a more thorough study of the role of cations. The counteranion used in this study, chloride, is an anion of physiological relevance in extracellular processes and is of significant importance in cellular ion transport, justifying its use in this study.26 Further to identifying the suppression of peaks, we set out to examine possible sample preparation protocols that may be utilized to lessen the salt suppression issue, such as analysis of samples in a frozen-hydrated environment or by washing the films with various salt removal solutions. Since the study is geared toward an understanding of salts in a biological sample, we chose to examine salt effects in the amino acid system arginine. In addition to its biological significance, arginine was chosen because it is hydrophilic, allowing both the salts and the amino acid to be dissolved in the same solution. Furthermore, analysis of a hydrophilic medium would facilitate our preliminary work with a frozen-hydrated matrix. An example of the suppression effects of the different salts on arginine ion formation is presented in Figure 1, which displays a bar chart of the average positive secondary ion yields for an undoped arginine film and for arginine films in which 1% (mol/ mol) of a salt was added. The data in Figure 1 corresponds to the protonated molecular ion [M+H]+ of arginine (m/z 175) and two of its main fragment peaks (m/z 70; C4NH8 and m/z 87; C4N2H11). The spectra were collected using a dose of ∼5.0 × 109 ions/cm2, and the values presented are an average of the yields collected from several spots on the film surface. Depending on the alkali halide added, the presence of the 1% salt in the arginine system reduces the protonated molecular ion yield to between 10% to 35% of the value observed from the undoped sample while m/z 87 and m/z 70 decrease in intensity to 20% to 60% of that observed in an undoped film. The values above the bars provide the average percent yields relative to the undoped sample. A pattern from the data that forms indicates the salt KCl suppresses the secondary ions to the greatest (26) Nelson, D. L.; Cox, M. M. In Lehninger Principles of Biochemistry, 3rd ed.; Nelson, D. L., Ed.; Worth Publishers: New York, 2000; pp 82-108.
1042
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
degree, followed by an almost equal suppression by NaCl and CsCl. LiCl appears to have the lowest suppression effect on the secondary ions for all of the salts examined. Salt Suppression Hypotheses. On the basis of the first results from the suppression data, three basic hypotheses were developed to explain the salt suppression in the arginine system and were tested against the data. The first theory postulated that the presence of the salt acted as a “catalyst” that enhanced the breakdown of larger molecules. For this model, the observed suppression of large molecular fragments and the protonated molecular ion is attributed to the salt acting in a way to increase the fragmentation of those analyte ions. If this model were to be observed, larger molecular fragments (greater than about 70 Da) should decrease in intensity with the presence of a salt while smaller less chemically characteristic fragments (less than about 50 Da) would show an increase in intensity because of the catalytic breakdown of the larger fragments.27 The values for “larger” and “smaller” fragments were arbitrarily assigned based on the arginine system and its known fragment ions. The formation of salt adduct peaks is a second possible explanation for the lowered intensity of analyte ions in the salt doped system. The formation of salt adduct peaks would involve some arginine molecules or fragments becoming ions by cationization instead of protonation, spreading the ions into a series of peaks instead of a single ion location. This may result in decreased intensity of characteristic peaks observed in the undoped arginine system. A third hypothesis proposed to explain the suppression of the analyte ions was that the addition of the salt anions to the arginine system were “neutralizing” the formation of characteristic positive secondary ions in the spectra. This neutralization would arise from addition of Cl- to the arginine system. The model would suggest that all of the fragment ions attributed to arginine should decrease in intensity regardless of mass range, although suppression may vary based on the location of ion formation and differences in positive ion interaction with Cl-. To provide data to enable these hypotheses to be discussed, both high and low mass fragments were examined in the undoped and the salt doped arginine systems, with results from that investigation presented in Figures 2 and 3. In the range of m/z e50, peaks at m/z 18, 27, 28, 29, 30, and 43 were examined. Peaks at m/z 12, 13, and 15 were omitted because of insufficient intensity. In this low mass region, the secondary ion yield of all the peaks examined from the salt doped systems decreased relative to the undoped arginine sample. Additionally, most of the peaks examined in this range are suppressed as a function of alkali halide doping in a similar pattern to that observed in Figure 1: KCl > NaCl, CsCl > LiCl. A few exceptions to the suppression trend are observed in Figure 2, some of which are attributed to the development of new salt adduct peaks. Comparable suppression effects are also observed in the mass range from m/z 50 to m/z 100 (see Figure 3) and from m/z 101 to m/z 176 (data not shown), although a few new peaks were observed that can be attributed to alkali cationized arginine molecules. In summary, the majority of peaks from all ranges of the examined spectra appear to be suppressed by the presence of salts to varying levels, with the pattern of suppression often (27) Leggett, G. J.; Vickerman, J. C. Annu. Rep. Prog. Chem. 1993, 88, 77–133.
Figure 2. Secondary ion yields of selected low mass fragments for an undoped arginine film (dots) and for LiCl/arginine (diagonal stripes), NaCl/arginine (vertical stripes), KCl/arginine (wave lines), and CsCl/arginine (checkered pattern) collected using a 20 keV C60+ primary ion source.
Figure 3. Secondary ion yields of select mass fragments in the range of 55 to 100 for an undoped arginine film (dots) and for LiCl/arginine (diagonal stripes), NaCl/arginine (vertical stripes), KCl/arginine (wave lines), and CsCl/arginine (checkered pattern) collected using a 20 keV C60+ primary ion source.
observed to be KCl > NaCl, CsCl > LiCl. This suggests that the suppression is not due to the salts acting to catalytically break down the molecules, but instead the suppression may be due to one of the other two mechanisms. To examine the role of cationization on the suppression of ion intensity (the second hypothesis), potential salt adduct peaks were divided into two categories and investigated. The first group corresponds to high mass adducts that form from cationization of an unfragmented arginine molecule while the second is related to adducts involving smaller fragment molecules. In the high mass range the protonated molecular ion, [M+H]+, forms from the association of a proton with an intact, neutral arginine molecule. Potential salt adducts that would compete with this process would be the substitution of the H+ by an alkali metal cation forming an [M+A]+ ion (where A ) Li, Na, K or Cs) or possibly an association of two cations with an arginine molecular ion
after the ejection of a proton [M-H+2A]+. Other possible combinations of the arginine-cation association are feasible, for example, [M-2H+3A]+, but no peaks corresponding to these ions were detected in this set of experiments. For the investigation into the high mass arginine salt adducts, the total counts from the two major adduct peaks identified in the previous paragraph were summed together and ratioed to the intensity loss of the [M+H]+ peak observed in the arginine/ salt systems. This analysis suggests that only 7 to 8% of the molecular ion intensity loss could potentially be attributed to the formation of salt adduct peaks. It should be pointed out that there is no conclusive evidence that the salt adduct peaks are “stealing” neutral arginine molecules from the [M+H]+ peak. The salt adduct peaks may form from an entirely different set of circumstances that does not interfere with formation of the [M+H]+. The data indicates that even if the salt adduct Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
1043
Figure 4. ToF-SIMS spectral data representative of a potassium chloride salt reference sample (A) positive ion, (C) negative ion; and an arginine film doped with 1% mol/mol of KCl (B) positive ion, (D) negative ion, collected with 20 keV C60+. From the spectra, the probable chemical structure of the identified peaks in the positive spectra are: m/z 39 - K+, m/z 104 - K2CN+, m/z 113 - K2Cl+, m/z 175 C6H15O2N4+ (arginine [M+H]+) and m/z 187 - K3Cl2+. The probable chemical structure of the identified peaks in the negative spectra are: m/z 26 - CN-, m/z 35 - 35Cl-, m/z 37 - 37Cl-/C3H-, m/z 109 - K35Cl2-, m/z 111 - K37Cl2-, m/z 131 - C5H11O2N2- (arginine [M-[C(NH)NH2]]), m/z 173 - C6H13O2N4(arginine [M-H]-), m/z 183 - K237Cl3- and m/z 185 - K237Cl3-.
peaks are contributing to the suppression of the [M+H]+ intensity, their influence is not substantial and that there must be another cause for the suppression. In the low mass range of the spectra, some peaks appear to increase in secondary ion yield that correspond to salt adducts of [A2CN]+, [A2OH]+, [A235Cl]+, and [A237Cl]+, similar to those proposed by Jones et al.11 In further studies it was observed that these adduct peaks increased with excess amounts of NaCl or KCl (20% mol/mol), confirming their origin as salt adduct peaks. In the 1% salt doped arginine systems, peaks corresponding to [A2CN]+ were observed in all of the systems with the exception of LiCl. Only the CsCl spectra contained peak intensity increases corresponding to the other proposed salt adduct peaks ([A2OH]+, [A235Cl]+, and [A237Cl]+). To explain the suppression of the arginine secondary ions in the presence of salt we are left with the possibility that it could be a consequence of the positive secondary ions becoming neutralized by the presence of the salts. This hypothesis follows 1044
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
from results observed previously by Garrison et al. 28,29 In a combined molecular dynamics modeling and experimental study, Garrison et al. indicated that neutralization of ion signal by oppositely charged particles is common in water ice films containing salts (preformed ions) and concluded that this type of interaction should be ubiquitous in all SIMS experiments. To investigate the potential role of the anion in the suppression effect, pure salt samples were examined, and their peak intensities were compared to the peak intensities observed in the arginine/salt system (see Figure 4). As a qualitative observation, the negative ion spectra for all the pure alkali chloride salts examined follow a similar pattern: The A35Cl2- ion peak is equal to, or of slightly greater intensity than the 35Cl- peak (see Figure 4C) and they are of the highest intensity. A similar pattern is observed for the positive ion data: (28) Wojciechowski, I. A.; Kutliev, U.; Sun, S.; Szakal, C.; Winograd, N.; Garrison, B. Appl. Surf. Sci. 2004, 231-232, 72–77. (29) Wojciechowski, I. A.; Sun, S.; Szakal, C.; Winograd, N.; Garrison, B. J. Phys. Chem. A 2004, 108, 2993–2998.
The A+ and the A235Cl+ ion peaks are similar in their intensities and represent the highest intensity peaks in the spectrum (see Figure 4A). When we examine the data for the mixed arginine/ salt systems, in contrast to pure salt the 35Cl- peak is observed to be a relatively insignificant peak in the negative ion spectrum (Figure 4D) while the cation signal again dominates the positive ion spectrum (see Figure 4B). The conspicuous reduction in the intensity of the chloride signal while the cation signal is of such high relative intensity is noteworthy. The correlation between the apparent reduction in the chloride signal and the concurrent decrease in positive secondary ions from the arginine matrix with the addition of the salt suggests that the two observations are correlated. This interpretation of the data suggests then that the chloride ion yield is reduced in the arginine/salt system because the chloride ions are combining with positive secondary ions, causing suppression of arginine signal. The observation that the cation peak appears relatively unaffected by the presence of the arginine matrix suggests that sputtering of the salt components is occurring and that the absence of the chloride signal is significant. It is also notable that peaks corresponding to A2Cl+ and ACl2- evident in the pure salt spectrum are also absent. This suggests that the salt ions have dispersed in the arginine matrix, and the added salt does not exist in a crystalline state. Therefore ion recombination into the A235Cl- and A35Cl2- is less probable. Furthermore, if the anion is active in the neutralization of positive secondary ions from the arginine matrix it would leave less of the chloride for recombination with the cation present to form these salt cluster peaks or to neutralize the cation signal. This discussion thus far has focused on the KCl data. However, similar observations have been obtained from the other alkali chloride systems. In each case, intense Cl- signals were obtained in the negative ion spectra from the pure salt, whereas in the arginine salt doped systems the chloride signal was very much reduced. In the positive ion spectra the cation signal was very intense and dominated the spectrum in both the pure salt systems and the arginine salt doped systems. Again, there appears to be a correlation between the reduction of the Cland the decrease in the arginine positive ion intensity. It is tempting to quantitatively compare the decrease in Cl- signal from the pure salt spectra to the changes in the arginine ion yields observed in the arginine doped systems. However, the yield of secondary ions is the product of ionization probability and sputter yield. To confidently compare the secondary ion yield between the pure salt and the arginine based samples we would need to know whether and to what extent the sputter yields differed. Currently we do not have sputter yield information for these systems, and it is not straightforward to obtain. Our observation then that the decrease in the raw yield data for the Cl- between the pure salt and the arginine system (corrected for salt concentration) does broadly correlate with the loss of the arginine ion yield across the four arginine/salt samples cannot provide quantitative and conclusive support for the neutralization hypothesis. However, Figure 4 and the totality of the information obtained in combination with the failure of the other proposed theories to adequately describe the suppression observed in the arginine salt system leads us to conclude that
the majority of signal intensity loss in this system is caused by the presence of chloride anions that combine with the arginine positive secondary ions. It should be noted that although chloride is not abundant within intracellular fluids (1-3 mM), it is significant in extracellular fluids (∼100 mM).30 Therefore, the chloride may be a source of biological ion suppression, particularly in tissue samples that are analyzed as freeze-dried specimens, since extra-cellular material will be part of the system analyzed and salt levels will be concentrated by the removal of water and other volatiles. There are many other significant intracellular anions such as carbonates, phosphates, and sulfates which may also play a role in positive ion suppression of biological specimens. The role that the cation plays in the suppression of positive ions is not clear. It appears from the data collected that the cations have some effect on the secondary ion yields of the positive ions but that the anion is the likely cause of the suppression. From the current data, we speculate that if the suppression is caused by the presence of Cl-, the variation observed from different cationic salts is related to the interaction of the cation with the chloride ion and not due to the cation having any direct effect on the positive ions. The majority of atomic and thermodynamic data follows a linear trend of increasing or decreasing values from lithium to cesium, depending on the property,31 suggesting that ionization energy, electronegativity, and lattice enthalpy are not of themselves likely causes for the changes. Salt Reduction Procedures. One simple procedure to address the salt suppression issue is to briefly dip the sample into a salt removal solution. 11,21 This procedure acts to chemically remove the salts from the surface of the film with the intent of lowering or eliminating their interference with analysis. Results presented in this section focus on the use of three commonly used solutions for salt removal: HPLC grade water, 0.15 M ammonium formate, and ethanol. The results from the salt washing procedures are presented in Figure 5. Data from the right side of the figure indicates that washing with HPLC grade water or 0.15 M ammonium formate may have some beneficial effects for the arginine/KCl system as evidenced by the increase in the protonated molecular ion yield in those systems compared to the unwashed arginine/KCl system. The water and the ammonium formate washes increased the protonated molecular ion yield (m/z 175) by factors of 2.1 and 3.6, respectively, when compared to the unwashed arginine/KCl sample. No enhancements of molecular ion yield were observed for the ethanol wash for the KCl/arginine system. Despite the observed benefits of the wash, a noteworthy drawback is the effect the washing procedure has on the undoped arginine sample. From the left side of Figure 5, the protonated molecular ion yield for the arginine molecule for the washed undoped samples are observed to decrease dramatically compared to the film that was not washed. Additionally, the protonated molecular ion yield for the washed arginine/KCl samples are still lower than that observed in the unwashed arginine film. Although the intensities are an improvement over the unwashed arginine/ KCl sample, the wash does not entirely reduce the effects of the presence of KCl in the arginine system. It should be noted that (30) Di Stasio, E. Biophys. Chem. 2004, 112, 245–252. (31) Stark, J. G.; Wallace, H. G. Chemistry Data Book, 2nd ed.; John Murray Publishers, Ltd: London, 1984.
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
1045
Figure 5. Secondary ion yields of the protonated molecular ion from unwashed and washed (HPLC grade water, 0.15 M ammonium formate or reagent grade ethanol) films of arginine (left) and arginine doped with 1% KCl (mol/mol) (right) collected with a 20 keV C60+ primary ion source.
Figure 6. Secondary ion yields for m/z 175 (M+H+) for films of arginine analyzed at room temperature (empty squares); arginine/KCl analyzed at room temperature (empty triangles); arginine as a frozen-hydrated film (filled-in squares); and arginine/KCl as a frozen-hydrated film (filled-in triangles) as a function of primary ion dose density (ions/cm2). Point A marks the approximate point at which the ambient condensed water is etched from the frozen-hydrated film surface, and data was collected using a 20 keV C60+ primary ion source.
some sample removal is likely during the washes; however, while no thickness measurements were taken, the films were etched up to a dose of 5.0 × 1013 ions/cm2 after the wash process with no evidence of the substrate appearing in the spectrum. Clearly these films were of multilayer thickness even after washing; thus, signal loss cannot be attributed to removal of the films from the substrate. Frozen-Hydrated Analysis. Another procedure investigated with the intent of removing unwanted salt effects for this study was the analysis of a frozen-hydrated sample on a cold-stage. In this procedure, a sample is frozen in its aqueous state based on the idea that the formation of the solid-state water matrix may prevent salt migration, salt crystallization, and/or provide an ideal environment with excess hydrogen for protonation from the water matrix. Further, when the sample is analyzed in an aqueous environment, the water can stabilize the isolated cations and anions by weakening the electrostatic interactions between them thereby reducing their tendency to associate in a crystalline lattice.26 Furthermore, if the anions can be trapped inside a water matrix in hydrogen bonded networks, they may not associate as 1046
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
readily with exiting positive ions thereby reducing their unwanted effects of signal suppression.29 The freezing process is of vital interest if this method is to be applied to the analysis of biological systems. Our choice of liquid nitrogen for the cooling procedure is not appropriate for more complex biological specimens. For a biologically relevant system, a quicker freezing procedure (such as the use of liquid propane as the freezing medium) would be required to prevent the formation of crystalline water, which may damage the sensitive biological material.32 The yield of the molecular ion (m/z 175) as a function of primary ion dose density (ions/cm2) for the frozen-hydrated samples is presented in Figure 6, along with a comparison to results collected from similar films run under standard conditions (i.e., no sample freezing, sample analysis at room temperature). The lack of signal for the frozen-hydrated films at the beginning of the depth profile was due to an ambient ice layer acquired (32) Low Temperature Methods in Biological Electron Microscopy (Practical Methods in Electron Microscopy); Robards, A. W., Sleytr, U. B., Eds.; Elsevier: Amsterdam, The Netherlands, 1985.
during sample transfer, which required removal by sputter etching before analysis of the analyte could take place. The data shows that analyzing the aqueous arginine and aqueous arginine/KCl samples in a frozen-hydrated environment has a favorable effect on the protonated molecular ion yield. As reported earlier in the article, the intensity of the molecular ion peak decreases with the addition of 1% KCl in films examined under standard room temperature conditions. However, when frozen-hydrated samples are examined and compared to samples run under standard conditions, both an enhancement in yield and an amelioration of the salt effect for the protonated molecular ion is observed. The enhancement observed when the frozen-hydrated films reach their steady-state plateau (point A in Figure 6) is an average factor of 1.8 and 4.2 over the arginine and arginine/KCl films run under standard conditions, respectively. The frozen-hydrated arginine/ KCl system also is enhanced by an average factor of 1.8 over the arginine system run under standard conditions, ameliorating the salt suppression effects explained earlier in the article. Interestingly, when the frozen-hydrated films are compared to each other, there is no evidence of a salt suppression effect for the protonated molecular ion peak. The enhancement observed from the analysis of the frozenhydrated films over the dehydrated films run at room temperature is a consequence of at least two potential causes. The first is that, as suggested by earlier work on amino acids trapped in ice matrixes, the presence of the ice matrix provides an alternative source of protons, which increases the chance of protonation of ions, particularly for the protonated molecular ion.33 This is an area of interest within our research group and will be explored in greater detail in subsequent studies. A second cause is related to the formation of the ice matrix. Garrison et al. have argued that in frozen ice films containing salts, in contrast to the cations, anions do not disrupt the hydrogen-bonding network in the water and therefore become “trapped” inside the matrix and consequently may not eject as readily as the cations.29 This does not imply that anions do not eject from the water matrix; however, the long-range interaction with water should make desorption of the chloride and its solvation sphere a more energetic process than for an anion with no long-range interaction with water molecules. The increase in energy required to remove an anion may therefore reduce the probability of some of the chloride anions sputtering and interacting with the arginine positive secondary ions. However, a more significant effect is likely related to the fact that in the frozen hydrated sample the chloride anions and arginine molecules are all surrounded by a large number of water molecules, greatly reducing the probability that the anions and the positive arginine secondary ion would collide and neutralize each in the emission zone. A possible complication in this argument is that under the conditions used for freezing, the rate of cooling would not be great enough to enable vitrification with the formation of a homogeneous solid. Quench freezing in liquid nitrogen can result in the solutions phase separating to form crystallites of pure water ice and either amorphous dispersions or crystallites of the solutes.34,35 (33) Xavier, A.; Lockyer, N. P.; Vickerman, J. C. Rapid Commun. Mass Spectrom. 2006, 20, 1327–1334. (34) Franks, F. Water and aqueous solutions at sub zero temperatures; Plenum Press: New York, London, 1982.
It is possible then that the arginine could be segregated to either the ice surface or the supporting substrate ice interface. The data reported in Figure 6 suggests that neither of these extreme effects obtains. If the arginine segregated to the ice surface, the observed ion yield would be expected to be similar to that obtained from a pure arginine film at room temperature, and the yield might be expected to decay with ion fluence as the room temperature film does. In fact what is observed is that arginine molecular ion intensity is higher from the frozen hydrated sample than from a film comprised entirely of arginine, is not suppressed by the presence of salt, and does not decay with ion fluence. This provides some support for our proposal that analysis in the frozen hydrated state for such samples is beneficial because positive ion neutralization by alkali chloride ions is inhibited in some way by the water molecules, and the arginine is dispersed with the ice matrix such that the protonation process is enhanced. CONCLUSION From the data presented in this article, it is clear that salts doped into the arginine system adversely affect the formation of secondary ions. Indeed, the protonated molecular ion intensity of arginine is observed to decrease by as much as 85% of the value observed in an undoped sample from the presence of just 1% (mol/ mol) of salt added. Results from the salt doped arginine samples indicate that salt adduct peak formation may play a role in the suppression of peaks; however, that effect is minor and accounts for at most 7 to 8% of the peak suppression in this sample set. Therefore, another more significant mechanism must be taking place to cause the suppression. Lower than expected peak intensities for chloride in the salt doped systems in conjunction with the signal suppression observed in the arginine/salt doped system suggest that a process of neutralization of positive ions may be a cause. This is in agreement with the observation that all of the arginine-related peaks in the spectrum are negatively affected by the presence of salts, ruling out the possibility that the salts are acting as catalysts to break down the larger fragment ions. Possible solutions for addressing the salt issues in the arginine/salt systems analyzed in this study yielded some interesting results. Chemical washing of the salt doped samples in HPLC grade water or ammonium formate appeared to counteract some of the salt suppression by increasing the protonated molecular ion yield of the arginine/KCl sample when compared to a similar unwashed sample. However, the washing procedure was also observed to decrease the secondary ion yield of the undoped arginine film indicating some deleterious effects of this procedure. Additionally, the washing procedure did not raise the yield for the washed salt-doped samples above 50% of the value observed in the unwashed undoped sample. Analysis of frozen-hydrated samples in which the analyte solution was frozen in an aqueous environment before introduction into the analysis chamber provided the most beneficial results. The arginine protonated molecular ion yield for the arginine/KCl film was 4.2 times greater in intensity than that of the same sample run as a dehydrated film at room temperature apparently providing a full reversal of the salt effect for that peak. The protonated molecular ion of the undoped frozen-hydrated arginine film was also enhanced by a (35) Pyne, A.; Suryanarayanan, R. Pharm. Res. 2001, 18, 1448.
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
1047
factor of 1.8 over a film run under standard conditions, suggesting that analysis in the frozen-hydrated state may also provide a means for increasing secondary ion yields. This indicates analysis of frozen-hydrated arginine films provides both a means for signal enhancement and a process for lessening the effects of the salts. ACKNOWLEDGMENT The authors thank the Life Sciences Interface program of EPSRC for their financial contribution (EP/C008251/1) to this work and also thank Dr. John Fletcher. In addition, we thank
1048
Analytical Chemistry, Vol. 81, No. 3, February 1, 2009
Professors Garrison and Winograd of Penn State University for their constructive comments and help with this article during a visit to Manchester funded by the Collaborating for Success through People initiative of EPSRC (EP/F012985/1).
Received for review October 3, 2008. Accepted December 9, 2008. AC8020888
