Improving the Molecular Ion Signal Intensity for In Situ Liquid SIMS Analysis
In situ liquid secondary ion mass spectrometry (SIMS) enabled by system for analysis at the liquid vacuum interface (SALVI) has proven to be a promising new tool to provide molecular information at solid–liquid and liquid–vacuum interfaces. However, the initial data showed that useful signals in positive ion spectra are too weak to be meaningful in most cases. In addition, it is difficult to obtain strong negative molecular ion signals when m/z>200. These two drawbacks have been the biggest obstacle towards practical use of this new analytical approach. In this study, we report that strong and reliable positive and negative molecular signals are achievable after optimizing the SIMS experimental conditions. Four model systems, including a 1,8-diazabicycloundec-7-ene (DBU)-base switchable ionic liquid, a live Shewanella oneidensis biofilm, a hydrated mammalian epithelia cell, and an electrolyte popularly used in Li ion batteries were studied. A signal enhancement of about two orders of magnitude was obtained in comparison with non-optimized conditions. Therefore, molecular ion signal intensity has become very acceptable for use of in situ liquid SIMS to study solid–liquid and liquid–vacuum interfaces.
KeywordsIn situ liquid SIMS Solid-liquid interface Molecular imaging Ionic liquid Biofilm Cell Electrolyte
Ionic and molecular distribution and transport at solid–liquid and gas–liquid interfaces are of great interest in many fields. However, such information has been difficult to obtain because of the lack of desirable in situ analysis tools. At biointerfaces, for instance, the distribution of small functional ions such as Na+, K+, Ca2+, or small organic molecules such as glucose and phospholipid molecules remains unclear. This is despite the use of many in situ techniques such as atomic force microscopy [1, 2], fluorescence [3, 4], Raman , infrared [6, 7], and surface plasmon resonance .
Secondary ion mass spectrometry (SIMS) is a very powerful surface analysis tool. It can provide elemental, isotopic, and molecular information with high sensitivity (ppb to ppm level) and good spatial resolution (down to sub-nm in depth resolution and ~50 nm in lateral resolution) [9, 10]. However, as a high vacuum technique, it has been traditionally used to analyze solid samples. In situ liquid SIMS was developed in the last several years and one of its unique capabilities is the analysis of liquid–solid and liquid–vacuum interfaces [11, 12, 13, 14, 15, 16, 17, 18]. To enable in situ liquid SIMS measurement, a vacuum-compatible device named system for analysis at the liquid–vacuum interface (SALVI) was developed [11, 12], where a liquid of interest can be sealed in a microfluidic channel under a thin (e.g., 100 nm) silicon nitride (SiN) membrane. The key to in situ liquid SIMS measurements is the use of a ~2 μm diameter aperture on the thin SiN membrane, through which the liquid interface and the liquid itself can be analyzed. Because the aperture size is small, surface tension can hold the liquid beneath the SiN membrane, and the evaporation of the liquid from the aperture does not cause any unacceptable problem to the vacuum system of the SIMS instrument . The SiN membrane surface can be modified and the liquid in the cell can be changed, allowing various liquid–solid interfaces to be studied. As a result, in situ liquid SIMS has been successfully applied to analysis of live biofilms [16, 17], protein-modified Au nanoparticles , electrochemical reactions [15, 19], and the formation of solid–electrolyte interface in a Li-ion battery .
However, initial studies reported detection of only negative molecular ions [11, 12, 13, 14, 15, 16, 17] because the signals of the positive ion spectra were too weak to be meaningful. In addition, although negative molecular ions could be successfully analyzed, their signal intensity was not strong [16, 17]. Therefore, the SIMS community has some genuine concern for using this new innovation for further practical applications.
In this work, we report an enhancement of two orders of magnitude in molecular signal intensity by using a cluster primary ion beam with reasonable currents. Both positive and negative molecular ion signals can be readily detected so that signal intensity will no longer be a major concern in in situ SIMS analysis. Four model systems, comprising a 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)-base switchable ionic liquid , a live Shewanella oneidensis biofilm , a hydrated mammalian cell, and a LiPF6 electrolyte  were studied.
The details of SALVI fabrication have been extensively described in our previous publications [11, 12, 16]. In brief, a microfluidic channel with a cross-section of 200 × 300 μm2 was prepared on a PDMS block. A 100 nm thick SiN membrane with a window size of 1.5 × 1.5 mm2 supported on a 7.5 × 7.5 mm2 silicon wafer was bonded to the PDMS block with the microfluidic channel located at the center of the SiN window. The liquid was introduced into the microfluidic channel through PTFE tubes, and the two ends of the PTFE tubes were sealed by a union . The sealed SALVI device was loaded on a sample holder prior to ToF-SIMS analysis.
The biofilm culture followed the description in our previous papers [16, 17, 21]. In brief, the Shewanella oneidensis MR-1 was grown overnight in the MM1 medium at 30 °C. After the SALVI microfluidic device was sterilized by 70% ethanol, the diluted culture was inoculated into the microchannel. After inoculation, the flow was stopped in 4 h to facilitate the bacteria cells’ initial attachment. Afterwards, MM1 medium was continuously supplied for biofilm growth. After ~7 d, sufficient biofilm was formed in the microchannel, which was confirmed by fluorescence analysis. The SALVI device was then disconnected from the supply of the MM1 medium and sealed for immediate ToF-SIMS analysis.
The mammalian cells used in this research were alveolar type II epithelial mouse lung C10 cells. The details of C10 cell culturing in SALVI have been described in our previous paper . In brief, after being thoroughly washed by 70% ethanol and sterilized DI water, the microchannel was treated with 10 mg mL–1 fibronectin solution overnight. Subsequently, 500 μL C10 cell suspension (the concentration is about 1 × 106 cells per mL) was injected into the microchannel, and then incubated at 37 °C for 1 d. Fresh medium was applied for the following 2-d cell culturing. Then the cultured cells were washed with PBS and fixed with 4% paraformaldehyde. DI water was introduced into the channel after the cells were fixed before ToF-SIMS analysis in order to reduce the interference signals from the PBS during the ToF-SIMS analysis.
A battery grade electrolyte, composed of 1.0 M LiPF6 in ethylene carbonate (EC) and dimethyl carbonate (DMC) (EC:DMC=1:2, volume ratio), was purchased from Sigma Aldrich (St. Louis, MO, USA) and directly used for ToF-SIMS analysis without any further treatment.
Beam Currents and Pulse Widths of Bi+, Bi3+, and Bi3++ Beams Used in This Study
Probe size (nm)
DC beam current (pA)
Pulsed current (pA) @ 10 kHz with maximum pulsed width
Pulsed current (pA) used in this research @ 10 kHz and corresponding pulsed width
Pulse width (ns)
Assignment of Several Representative Negative Ion Signals of the DBU-Base Ionic Liquid
Ionic liquid, characteristic fragment ion/SiN membrane
Ionic liquid, characteristic fragment ion
1-Hexanol, characteristic fragment ion
1-Hexanol, molecular ion
Ionic liquid, molecular ion
A high beam current can be obtained if compromising beam focus. The principle is shown in Supplementary Figure S1. One could use the Lens Source to focus the most emission current onto the sample surface; however; the focus was poor (Supplementary Figure S1a). Alternatively, one could defocus the beam using the Lens Source and only allow a small angle of the emission current to pass through Aperture 1 before focusing it onto the sample surface. In this case, the focus is improved (Supplementary Figure S1b) with a low current. Roughly, the beam current on the sample surface was proportional to the square of the optimum beam diameter value. In this research, three diameters were used: 200, 500, and 650 nm. In addition, it was also possible to adjust the beam current by varying the pulse widths. Owing to the time-of-flight mass selection principle of the Bi source (as shown in Flash S1 in Supplementary Materials), for each ionic species there was a maximum pulse width over which interference from other ion species appeared. For the Bi+, Bi3+, and Bi3++ beams (25 kV acceleration), the maximum pulse widths were ~280, ~160, and ~250 ns, respectively. It should be noted that the pulse widths used in this study were narrower than the maximum values. The beam sizes, pulse widths, as well as corresponding DC and pulsed beam currents, are listed in Table 1. A repeat rate of 10 kHz (cycle time = 100 μs) was used and the effective mass range was about m/z 0–800. For readers’ convenience, we have used beam size, pulsed current, and ionic species to describe a beam condition (e.g., a 500 nm 0.36 pA Bi3+ beam, or more briefly, a 500 nm Bi3+ beam).
An optical microscope (Olympus BX60, with a 100X objective (NA = 0.80), Center Valley, PA, USA) and a scanning electron microscope (SEM) (Hitachi TM-1000, Tarrytown, NY, USA) were used to determine the size of the apertures on SiN membranes. The optical images were measured as soon as the device was taken out of the SIMS instrument. A thin layer of Au (~10 nm) was deposited on the SiN membrane surface before SEM measurement.
Results and Discussion
Beam Damage Evaluation
In most of our previous studies, a 25 keV Bi+ beam was used as the primary ion beam, and the beam size was focused to ~200 nm diameter with a current of 0.2–1.0 pA [11, 13, 14, 15]. The Bi+ beam was chosen because it was the most abundant primary ion species, and it could provide the highest current with our instrument (IONTOF V) [29, 30]. In addition, it had been expected that the liquid interface should be self-renewable due to liquid diffusion, and the beam damage of Bi+ might not be problematic .
The positive secondary ion spectra showed similar results (i.e., Bi+ induces significant damage) but Bi3+ and Bi3++ can provide strong and stable molecular ion signals (Supplementary Figure S2). Apparently, using cluster primary ions is the key to in situ liquid SIMS analysis.
It should be noted that unlike C60 or Ar cluster ions, molecular depth profiling of organic thin films using metal cluster primary ions is not feasible because of accumulation of beam damage . Therefore, the self-renewable feature of liquid interfaces (attributable to liquid diffusion and evaporation in many cases) plays an important role to compensate beam damage in in situ liquid SIMS analysis.
Optimization of Beam Current
To further improve the signal intensity, a straightforward consideration is to increase the incident beam current. One option for increasing the beam current is to increase the pulse width. However, the pulse width cannot be too long (≤280 ns) because of the ion separation mechanism used in the liquid metal ion gun (See Flash S1 in Supplemental Materials). In addition, though only unit mass resolution is used in this study, a short pulse width is still very helpful for assigning high mass signals (m/z > 200). Therefore, increasing pulse width may not be an appropriate choice. It is also possible to reduce cycle time (i.e., increase pulse frequency) to increase the beam current. However, cycle time is proportional to the square root of the measurable mass range due to the time-of-flight principle. Since a reasonable mass range is necessary for in situ SIMS analysis, reducing cycle time may not be a desirable choice in many cases either. Another option to increase the beam current is compromising lateral resolution (increasing beam diameter) of the incident ion beam. Generally speaking, better lateral resolution is a plus for in situ liquid SIMS study because three-dimensional (3-D) distributions of molecular signals are interesting in many cases. However, better signal intensity is normally more desirable.
Increasing the primary ion beam current is helpful in improving signal intensity of the molecular ion, as shown in Figure 3c and d. An interesting observation is that the signal does not increase linearly with beam current. For example, a comparison of Figure 3c with d shows that the m/z 145 signal exhibits ~20 times enhancement for only a six times increase in beam current. This situation was commonly observed in all model systems for both negative and positive secondary ions in this research. A plausible explanation is that a stronger beam may create a larger aperture in the SiN membrane, even though the scanned area is kept the same (2 μm in diameter). This explanation is supported by our optical images and SEM images, in which the aperture made by the 200 nm beam is considerably smaller than that made by the 500 nm beam (Supplementary Figure S3).
Bi3+ or Bi3++
In agreement with previous reports [28, 30, 35, 36, 37], the performances of Bi3+ and Bi3++ were found to be similar (for example, Figure 3d and e). Different pulse widths were, however, required to obtain similar currents: 160 ns for the Bi3+ beam and 250 ns for the Bi3++ beam at the maximum current. Although only unit-mass resolution can be achieved in our in situ liquid SIMS measurement currently, a narrow pulse width is still desirable because the narrow pulse width can reduce peak overlap (as shown in Figure 3d and e). Therefore, Bi3+ is better than Bi3++ for in situ liquid SIMS study.
One of the unique capabilities of ToF-SIMS is that it can provide 3-D ion images. In acquiring 3-D data, a larger aperture is better because it can provide more spatial information. However, there is a clear tradeoff here that larger aperture size may lead to unwanted leaking of the liquid to the vacuum. The liquid is confined in the aperture because of the liquid’s surface tension. Smaller aperture size provides higher holding pressure to confine the liquid in the aperture . If the aperture size becomes larger, surface tension may not be able to confine the liquid within the aperture anymore. Our experimental results show that if we increased the scanned area to 3 μm in diameter, the chance of liquid leaking to the vacuum greatly increased. The practical limit to the scanned area is, therefore, approximately 2 μm in diameter for the ionic liquid sample in this study.
Signal Improvement of Biofilm and Cell Samples
A schematic illustration of the optimization of experimental conditions for in situ liquid SIMS analysis is shown in Figure 4. Based on the previous discussion, a 500 nm 0.36 pA Bi3+ beam scanned on an area 2 μm in diameter is recommended for general in situ liquid SIMS analysis. This condition has been successfully used in our recent biofilm and single cell research [21, 27].
LiPF6 EC/DMC Electrolyte
The optimized conditions can also be used to analyze other nonaqueous systems, for example the analysis of a 1.0 M LiPF6 EC/DMC electrolyte (Supplementary Figure S4). Very strong molecular ion signals are observed because of the ionic character of this sample.
Similar to the other model systems already discussed, beam damage induced by the Bi3+ beam was not a problem. In contrast, serious beam damage was observed when the Bi+ beam was used.
It should be noted that although the optimized conditions recommended here can be used in a variety of non-aqueous systems, there is room for further optimization in specific cases. For example, the liquid leaking through the aperture was more likely to occur when analyzing the LiPF6 electrolyte. This situation may be attributed to the low surface tension of the organic solvents (EC/DMC). Therefore, the SIMS conditions may need to be tuned accordingly for different liquids using the principles outlined here.
As mentioned before, only unit mass resolution can be achieved because high spatial resolution mode is used in in situ liquid SIMS analysis. The detailed reason why high mass resolution and high spatial resolution cannot be achieved simultaneously in a traditional ToF-SIMS instrument is described in Flash S2 in Supplementary Materials. The low mass resolution may be the next challenge for extending applications of in situ liquid SIMS. Recently, several new ToF-SIMS instrumentations have been developed that can simultaneously provide high mass resolution and high spatial resolution with reasonable signal intensity. Two examples are the J105 from IONOPTIKA  and the attachment of an Orbitrap on a ToF-SIMS . The combination of in situ liquid SIMS with these instrumental advances will be ideal for investigation of liquid interfaces at the molecular level.
In situ liquid SIMS is a promising new technique. However, the initial results suffered from the low signal intensity for molecular and characteristic fragment ion species, especially in the positive ion mode. In this study, we used four model systems to optimize the experimental conditions, and successfully improved signal intensities for both positive and negative characteristic secondary ions. A 500 nm 0.36 pA Bi3+ beam (25 kV) scanned on an area 2 μm in diameter is recommended for general in situ liquid SIMS analysis. Beam damage induced by the 25 keV Bi3+ beam is not problematic for in situ liquid SIMS measurement despite the high ion doses that are two to four orders of magnitude higher than the static limit. The signal intensities of molecular ions and characteristic fragment ions improve by about two orders of magnitude compared with our initial 200 nm Bi+ beam. The optimized condition can be used to study various aqueous or nonaqueous liquid samples, although there may be room for further improvement because of variations of surface tensions among different liquids. This study provides a reliable set of conditions to investigate liquid interfaces using in situ liquid SIMS.
This work was funded by Laboratory Directed Research and Development (LDRD) programs of the Pacific Northwest National Laboratory (PNNL), which include an FY16 Open Call LDRD, an EBSD FY16 Seed LDRD, and an MS3 LDRD. The work was performed at EMSL, a national scientific user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research located at PNNL. We also appreciate Mr. Decarle Jin for his English editing.
- 8.Abadian, P.N., Tandogan, N., Jamieson, J.J., Goluch, E.D.: Using surface plasmon resonance imaging to study bacterial biofilms. Biomicrofluidics 8, 021804 (2014)Google Scholar
- 10.Wirtz, T., Philipp, P., Audinot, J.N., Dowsett, D., Eswara, S.: High-resolution high-sensitivity elemental imaging by secondary ion mass spectrometry: from traditional 2-D and 3-D imaging to correlative microscopy. Nanotechnology 26, 434001 (2015)Google Scholar
- 12.Yang, L., Yu, X.Y., Zhu, Z.H., Thevuthasan, T., Cowin, J.P.: Making a hybrid microfluidic platform compatible for in situ imaging by vacuum-based techniques. J. Vac. Sci. Technol. A 29, 061101 (2011)Google Scholar
- 17.Hua, X., Marshall, M.J., Xiong, Y.J., Ma, X., Zhou, Y.F., Tucker, A.E., Zhu, Z.H., Liu, S.Q., Yu, X.Y.: Two-dimensional and three-dimensional dynamic imaging of live biofilms in a microchannel by time-of-flight secondary ion mass spectrometry. Biomicrofluidics 9, 031101 (2015)Google Scholar
- 18.Zhu, Z.H., Zhou, Y.F., Yan, P.F., Vemuri, R.S., Xu, W., Zhao, R., Wang, X.L., Thevuthasan, S., Baer, D.R., Wang, C.M.: In situ mass spectrometric determination of molecular structural evolution at the solid electrolyte interphase in lithium-ion batteries. Nano Lett. 15, 6170–6176 (2015)CrossRefGoogle Scholar
- 19.Yu, J., Zhou, Y., Hua, X., Liu, S.Q., Zhu, Z.H., Yu, X.Y.: Capturing the transient species at the electrode‐electrolyte interface by in situ dynamic molecular imaging. Chem. Commun 52, 10952–10955 (2016)Google Scholar
- 21.Ding, Y.Z., Zhou, Y.F., Yao, J., Szymanski, C., Shi, L., Cao, B., Zhu, Z.H., Yu, X.Y.: Imaging Shewanella oneidensis biofilm matrix by in situ liquid ToF-SIMS. Anal. Chem. Submitted, (2016)Google Scholar
- 22.Yu, J.C., Zhou, Y.F., Hua, X., Zhu, Z.H., Yu, X.Y.: In situ characterization of hydrated proteins in water by SALVI and ToF-SIMS. J. Vis. Exp. 108, e53708 (2016)Google Scholar
- 24.Liu, Y., Lu, C.Y., Twigg, S., Ghaffari, M., Lin, J.H., Winograd, N., Zhang, Q.M.: Direct observation of ion distributions near electrodes in ionic polymer actuators containing ionic liquids. Sci. Rep. UK 3, 973 (2013)Google Scholar
- 31.Szakal, C., Kozole, J., Russo, M.F., Garrison, B.J., Winograd, N.: Surface sensitivity in cluster-ion-induced sputtering. Phys. Rev. Lett. 96, 216104 (2006)Google Scholar
- 42.Available atL https://www.iontof.com/news-ion-tof-sims-events.html; asscessed March 11, (2016)