Improved mass resolution and mass accuracy in TOF-SIMS spectra and images using argon gas cluster ion beams

Several types of cluster ion beams have been examined and applied to TOF-SIMS analyses to measure large organic- and biomolecules with high mass because these beams inflict less damage to the sample surface. According to computer simulations¹ and experimental^2–6 reports, the ejection of high mass secondary ions is enhanced with the nonlinear effects of multiple collisions when large-sized cluster ion beams bombard the organic surfaces. There are some reports that a cluster ion beam as the sputtering ion source can successfully accomplish depth profiling of organic^7–10 and amino acid films¹¹ without ion damage. For example, Mochiji and Oshima et al. used argon gas cluster ion beams (Ar-GCIB) to measure several peptides and proteins with molecular weights of up to 15 kDa such as insulin, cytochrome C, lysozyme, and chymotrypsin, which were previously undetected in TOF-SIMS using atomic ion beams.^4,6 Recently, a number of studies have examined the performance of cluster ion beams with respect to tissue imaging on the brains of rodents.^12–16 In these studies, various biological molecules such as cholesterols, fatty acids, and sulfatides were detected in the brain tissue, with the intensity of the secondary ions boosted significantly when using Ar-GCIB than when using a bismuth cluster ion beam.

Understandably, GCIB is gaining in popularity as the primary ion source for measuring organic- and biomaterials in TOF-SIMS analyses. Of these cluster ion beams, Ar-GCIB is the preferred analysis beam because argon gas is relatively cheap and is easily formed into various sizes of clusters. Unfortunately, Ar-GCIB has poor lateral resolution as well as poor mass resolution caused by difficulties in both spatial and time focusing of large argon clusters. Poor lateral resolution results in low image quality for small sample areas such as cells or hair in the SIMS imaging analysis. One method to overcome this has been to increase the energy of the primary ion beam, which facilitates the focusing of the Ar-GCIB. This approach was utilized by Angerer et al. who demonstrated that a 40 keV Ar₄₀₀₀⁺ beam could be focused to approximately 2–3 μm at the sample surface to successfully produce a chemical image of sphingomyelin on human hair.¹² Ar-GCIB is also limited by poor mass resolution, which prevents the isotope peaks from separating, thereby causing them to become overlapped in the spectrum. The resultant broadness of the peaks is problematic when trying to assign accurate mass for mass calibration; by the same token, the isotope distribution cannot be utilized to determine completely unknown peaks.

In this study, we are concerned with increasing the mass resolution and mass accuracy of the spectrum and image in the Ar-GCIB-TOF-SIMS analysis by using delayed extraction and external mass calibration. Delayed extraction¹⁷ has been widely used in matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry^18,19 and TOF-SIMS (Refs. 20–28) to achieve high mass resolution. This technique achieves time-of-flight compensation in ion energy spread by applying additional accelerating energy onto the sample plate with a specific time delay after the laser irradiates the sample surface. Thus, ions having the same mass-to-charge ratio (m/z) can arrive at the detector simultaneously, to generate peaks with narrow widths. We demonstrate here that the isotope peaks can be separated enough to distinguish them from the lowest to the highest after applying the delayed extraction technique to TOF-SIMS. In TOF-SIMS, delayed extraction achieves time-of-flight compensation in ion energy spread by applying extracting pulse energy onto the extractor of the secondary ions with a specific time delay after the pulsed primary ion beam irradiates the sample surface. Unfortunately, the spectrum acquired by using the delayed extraction method has lost the peaks of hydrogen and carbon ions, which causes difficulties in internal mass calibration—the typical method of calibrating mass in TOF-SIMS analysis. Furthermore, the asymmetric peaks in the low mass range make it difficult to determine their specific m/z values, which often leads to inaccurate internal mass calibration. As a result, we have used the external mass calibration method together with delayed extraction in our TOF-SIMS analysis. As with the delayed extraction method, the external mass calibration method is also widely used in various mass spectrometry techniques.^29–32 Although these two methods, delayed extraction and external mass calibration, are not new, they will be useful in identifying unknown materials when newly applied to an Ar-GCIB-TOF-SIMS analysis.

Irganox^® 1010 was purchased from Sigma Aldrich (St. Louis, MO, USA) and dissolved in ethanol. Tryptic digest of bovine serum albumin (BSA) and peptide calibration standard II were purchased from Bruker Daltonics (Bremen, Germany). Tryptic digest of BSA and the peptide calibration standard II samples were each dissolved in 125 μl of distilled water, according to the instructions provided. The silicon wafers were cleaned in a super-piranha solution³³ and rinsed sufficiently with distilled water. Each sample of 10 μl were loaded onto a clean silicon wafer and air dried. After preparing the samples, Irganox^® 1010 was added adjacent to the dried sample spots. For the rat brain sampling, all experimental procedures were conducted using protocols approved by the Institutional Animal Care and Use Committee of Konkuk University (Seoul, Korea). Male Sprague–Dawley rats (aged 10 weeks and weighing 300 g) were purchased from Orient Co., Ltd., a branch of Charles River Laboratories (Seoul, Korea). The animals were anesthetized using xylazine/ketamine (30 mg/75 mg/kg, ip) and maintained at 37 ± 1 °C during all surgical procedures. The dissected brains were immediately transferred to liquid nitrogen and then transferred to −80 °C. For TOF-SIMS analysis, the frozen brain tissues were sectioned at 10 μm-thickness in −20 °C using a cryostat (Leica CM 3050S, Leica Microsystems Inc., IL). The tissue sections were deposited onto a stainless steel substrate and then stored at −80 °C before the TOF-SIMS analysis.

TOF-SIMS spectra were obtained using a TOF-SIMS V instrument (ION-TOF GmbH). Ar₁₀₀₀⁺ with 10 keV was used as the analysis ion beam and the primary ion current was approximately 0.01 pA. During data acquisition, a pulsed electron flood gun was not used since charge compensation was unnecessary for Ar-GCIB. For the spectrum analysis, the analysis area was 300 μm × 300 μm and the primary ion dose density was 2.2 × 10¹¹ ions/cm² to ensure static limits. Five spectra were acquired for each sample in the conventional static and delayed extraction modes, whose timing diagrams are shown in Scheme 1. For the imaging analysis, we generated images with a wider field of view of 17.4 × 10.8 or 15 × 10 mm² using a “stage-scan” technique in a saw-tooth mode that scanned the entire surface area with a spatial resolution of 10 μm. The primary ion current was 0.03 or 0.01 pA and the primary ion dose density was 8.4 × 10¹⁰ or 4.5 × 10¹⁰ ions/cm² to ensure static limits. In the stage-scan mode, the sample was rastered by moving the sample stage. A delayed extraction time of 1.95 μs was used in the delayed extraction mode. The positive ion spectra were internally mass calibrated using C₃H₅⁺, C₃H₇⁺, C₄H₇⁺, C₄H₉⁺, and C₅H₁₁⁺ peaks for the tryp-BSA and C₃H₅⁺, C₃H₇⁺, C₄H₉⁺, and C₅H₁₁⁺ peaks for the peptide standard calibration II and rat brain samples. For external calibration, four peaks of Irganox^® 1010 were selected and used as external calibrants, as explained in Sec. III.

The molecular ion peaks of Irganox^® 1010, recorded using Ar₁₀₀₀⁺ in the conventional extraction mode, are broad and show poor mass resolution, as shown in Fig. 1(a). With a mass resolution of approximately 260, the isotope peaks of the molecular ion are not separated and cannot be distinguished from each other. One attempt to improve mass resolution in TOF-SIMS using Ar-GCIB involves applying delayed extraction, an essential technique in TOF mass spectrometry.^17–28 

In the conventional static extraction mode (standard mode) of TOF-SIMS analysis, the width of the primary ion pulse directly affects the start accuracy of the TOF measurement, regardless of whether the analysis is performed with Ar-GCIB or with a liquid metal ion gun. This is because the pulsed primary ions desorb the secondary ions while the extraction field is already switched on, as shown in Scheme 1(a). The difficulty of Ar-GCIB is that the pulse duration, even in a bunched mode, is quite long (more than 20 ns) due to the extreme mass distribution of the large argon cluster primary ions. This results in a limited time and, hence, mass resolution. In the delayed extraction mode, however, the pulsed primary ions desorb the secondary ions while the extraction field is still switched off so that the secondary ions drift in field-free space above the sample surface. After a specific delay time, the extraction field is switched on shortly after the end of the primary ions pulse, so the start of the TOF measurement is no longer spread out over the primary ions pulse width. In other words, the delayed extraction mode separates the pulsing of the secondary ion extraction from that of the primary, and with the right delayed extraction time, the secondary ions that are located between the sample surface and the secondary ion extractor are able to enter the analyzer at the same time. As a result, mass resolution is dramatically improved from m/z ∼ 30 to ∼1200. Figure 1(b) shows the molecular peaks of Irganox^® 1010 obtained with Ar-GCIB in the delayed ion extraction mode. The isotopic peaks of Irganox^® 1010 are clearly separated, and mass resolution is visibly improved to 4972 calculated at m/z 1176.78.

Mass accuracy is contingent on accurate peak assignments that are reliant on mass resolution. Without these peak assignments, mass calibration is difficult. In a TOF-SIMS analysis, internal mass calibration is typically carried out by choosing peaks generated from small molecules that already exist in the sample. Thus, the mass error between the measured and theoretical values increases with molecular weight. Consequently, identifying unknown peaks obtained by Ar-GCIB is difficult when using the internal mass calibration method, and is further hampered when using the delayed extraction mode, as described in the Introduction.

Irganox^® 1010 was selected for external mass calibration^29–32 because this organic material is inexpensive and commonly used to evaluate the performance of cluster ion beams through depth profiling.^7–9 Further, positive and negative spectra that contain well-known multiple characteristic and molecular peaks from the low- to the high-mass spectral regions (up to m/z 1200) can be obtained from this material in a reasonable time (less than 10 s). Figures 2(a) and 2(b) are positive and negative spectra obtained by using Ar-GCIB in the delayed ion extraction mode. The peaks selected as the external mass calibrants are indicated with arrows in Figs. 2(a) and 2(b). The Irganox^® 1010 solution was deposited adjacent to the analytes, and a spectrum was acquired before the analytes were analyzed. Mass calibration was performed by using the four selected peaks shown in Fig. 2(a) for the positive spectrum and Fig. 2(b) for the negative spectrum, and the channel values or flight times of the selected peaks were measured from the mass calibrated spectrum. Then, the spectra of the analytes were externally calibrated using these channel values or flight times of the four specific m/zs. It is important to note that the analytes should be analyzed without z-movement.

To evaluate the effectiveness of the external mass calibration method, trypsin-digested bovine serum albumin (tryp-BSA) and peptide calibration standard II were used. Figure 3 shows positive spectra of tryp-BSA and peptide calibration standard II obtained with Ar-GCIB using the delayed extraction mode and external mass calibration. Internal mass calibration was first performed using the C₃H₅⁺, C₃H₇⁺, C₄H₇⁺, C₄H₉⁺, and C₅H₁₁⁺ peaks for the tryp-BSA, and the C₃H₅⁺, C₃H₇⁺, C₄H₉⁺ and C₅H₁₁⁺ peaks for the peptide standard calibration II. External mass calibration was then performed using the four peaks of Irganox^® 1010. Here, it should be noted that in the delayed extraction mode, the setting values of the delay time can be crucial—particularly for the flight time of the low-mass ions because the setting condition of the delayed extraction mode is typically optimized to separate the high-mass ion peaks (m/z ∼ 1000 in our study). This is because low-mass ions drift faster than ions with higher mass, meaning they experience a smaller fraction of the extraction field compared to the slower (high-mass) ions. As a consequence, the deviation of the mass calibration curve for low-mass ions is generally larger than that of the high mass ions. For internal calibration, then, it is advisable to use ions with higher mass—the higher the better—preferably above m/z 40 in order to minimize this effect.³⁴ 

Figure 4 shows the mass differences of the internally mass calibrated and externally mass calibrated spectra. The average values and deviations calculated from the five spectra for each sample are also indicated in the graph. For the internally mass calibrated spectra, the mass differences are in the range of 0.7–2.0 Da, with the values tending to increase as a function of m/z. Here, in the case of tryp-BSA, the average mass difference of the eight peaks is approximately 1.2 Da. However, after external mass calibration, the average mass difference rapidly decreases to 0.03 Da. Since these values can be also expressed by means of mass error in parts per million (ppm), they were converted to 827.14 ppm for the internal mass calibration and 23.64 ppm for external mass calibration. For the peptide calibration standard II, the average mass difference of the seven peaks is approximately 1.3 Da in the internally mass calibrated spectrum and is reduced to 0.002 Da in the externally mass calibrated spectrum. These values correspond to 925.86 and 1.65 ppm, respectively. From these results, we are confident that the external mass calibration method is a useful calibration method for Ar-GCIB-TOF-SIMS analysis.

Rather than using external mass calibration, well-known higher m/z peaks could be included to improve mass accuracy in the high mass range. For example, the average mass difference of the eight peaks can be reduced from 1.2 Da (827.14 ppm) to 0.013 Da (8.78 ppm) for tryp-BSA with the use of m/z 1283.71 and 1479.80 as internal calibrants. Similarly, the average mass difference of the seven peaks can be reduced from 1.3 Da (925.86 ppm) to 0.007 Da (6.49 ppm) for the peptide calibration standard II with the use of m/z 1347.74 and 1758.93 as internal calibrants. However, in general, it is difficult to determine the exact chemical formula of a high molecular peak for the internal calibrant unless peak identification with the MS/MS measurement can be successfully performed for each peak.

In addition to a spectrum analysis, the delayed extraction and external calibration methods were also applied to an image analysis of a rat brain sample obtained by using Ar-GCIB. Figure 5 shows the images and spectra of protonated cholesterol trimer¹⁵ [(C₂₇H₄₆O)₃H⁺, m/z (calc.) = 1160.07] obtained by conventional [(a) and (c)] and delayed extraction modes [(b) and (d)], as explained above. As with the spectrum analysis case, mass error was significantly reduced from 215.5 to 60.34 ppm by using delayed extraction and external mass calibration, although it was higher than in the spectrum analysis due to the large scanning area for image acquisition.

The use of Ar-GCIB as analysis beams in the TOF-SIMS technique has rapidly increased because they produce molecular ions of biomolecules with high mass. However, their poor mass resolution results in insufficient separation of the isotope peaks, thus making it difficult to distinguish them from each other. This affects mass calibration by aggravating the mass error up to several thousands of ppm. To circumvent these difficulties, two common techniques in MALDI-TOF mass spectrometry, delayed extraction and external mass calibration, were applied to improve mass resolution and mass accuracy, respectively. Mass resolution was improved to 4972 for the molecular peak (m/z 1176.78) of Irganox^® 1010, and mass accuracy was enhanced up to 1.65 ppm in the spectrum analysis and up to 60.34 ppm in the image analysis. When Ar-GCIB is used as the analysis beam in TOF-SIMS, the delayed extraction technique and external mass calibration can be valuable tools to identify unknown peaks.

This study was supported by the Development of Platform Technology for Innovative Medical Measurements Program (GP2016-0022) from the Korea Research Institute of Standards and Science, the Pioneer Research Center Program (NRF-2012-0009541), the Bio & Medical Technology Development Program (NRF-2015M3A9D7029894), and Global Frontier Project (H-GUARD_2013M3A6B2078962) of the National Research Foundation (NRF) funded by the Ministry of Science, ICT & Future Planning. The authors are also grateful to Derk Rading (ION-TOF GmbH) for his valuable comments and suggestions.