3:00 PM - 3:15 PM
[MZZ45-09] Development of a Stigmatic Ion Imaging Method of Secondary Ion Mass Spectrometry by Using a qCMOS Camera with Ultra-low Readout Noise
Keywords:SIMS, Ion imaging, qCMOS, Micro channel plate
In order to understand the origin and processes of terrestrial and extraterrestrial materials, many studies have been conducted using stable isotopic compositions as tracers. Since stable isotope ratios of solar-system materials vary by at most four orders of magnitude, a dynamic range of at least four orders is required. In addition, ion imaging on the µm-scale using in situ analysis is required for discussions that combine petrological information with isotopic spatial distributions. Secondary ion mass spectrometry (SIMS) is an analytical technique that enables this type of imaging analysis. SIMS provides the analysis of all nuclides from hydrogen to uranium. SIMS has two methods for ion imaging. The scanning imaging method, which uses a focused primary ion beam, is limited in spatial resolution to the µm-order. The stigmatic imaging method, which preserves the isotopic distribution of secondary ions in the range of several hundred µm, is limited in spatial resolution by the optical aberrations of the mass spectrometer. To achieve both spatial resolution and dynamic range in ion imaging on the µm-scale by in situ analysis, stigmatic ion imaging is expected to provide more precise isotopic analysis than scanning ion imaging.
Previous quantitative ion imaging techniques began with a microchannel plate (MCP), a fluorescent screen (FS), and a CCD camera to amplify signals and acquire image data. The MCP/FS method has been treated as a qualitative analysis method due to the nonlinear relationship between the ion counts before amplification, C (Counts Per Second; cps), and the digital values obtained from the camera, I (Analog to Digital Unit; ADU). On the other hand, Mantus and Morrison (1990) reported a correspondence using the characteristic curve I=pCq. Tokunaga et al. (2021) developed a quantitative imaging method that can correct for local degradation of MCP and differences in quantum efficiency by calculating this characteristic curve for each pixel. However, noise components originating from the camera and the influence of external light that could not be excluded limited the quantitative performance when signal of ion counts were weak less than 10-1cps/pixel, and the systematic evaluation of error factors was insufficient.
In this study, we attempted to develop an imaging method that reduces noise, which can be a systematic error component. A qCMOS camera, manufactured by Hamamatsu Photonics, which has less noise than in previous studies, was incorporated, and a darkroom covering the MCP/FS/qCMOS camera system was specially designed. Using this new system, we evaluated the noise components in the experimental environment. In addition, the characteristic curves were calculated by optimizing the experimental conditions using SIMS.
The readout noise and dark current noise of the qCMOS camera in the experimental environment were evaluated to be 0.41 electrons rms and 0.016 electrons/pixels/s, respectively, which are equivalent to the camera specs, and the effect of external light was negligible. By adjusting the SIMS secondary ion optics, a characteristic curve with a dynamic range of 4 orders from 10-3 cps/pixel to 101 cps/pixel was obtained by using a uniform secondary ion image of less than a few %.
In order to evaluate the possibility of statistical processing of the above obtained characteristic curves, we applied these curves to calculate the silicon isotope ratio measured with silicon wafers assumed with homogeneous Si isotope distribution . The isotopic distributions of δ29Si and δ30Si were calculated by applying the characteristic curve to the image intensity of each pixel and converting it to ion counts. The errors of δ29Si and δ30Si for each pixel calculated by propagating the errors of the constants of the characteristic curves were consistent with the isotope distributions calculated by 50 x 50 pixel without the influence of interference for 28SiH- and 29SiH-. This indicates that the errors in ion counts can be estimated by a statistical process.
Previous quantitative ion imaging techniques began with a microchannel plate (MCP), a fluorescent screen (FS), and a CCD camera to amplify signals and acquire image data. The MCP/FS method has been treated as a qualitative analysis method due to the nonlinear relationship between the ion counts before amplification, C (Counts Per Second; cps), and the digital values obtained from the camera, I (Analog to Digital Unit; ADU). On the other hand, Mantus and Morrison (1990) reported a correspondence using the characteristic curve I=pCq. Tokunaga et al. (2021) developed a quantitative imaging method that can correct for local degradation of MCP and differences in quantum efficiency by calculating this characteristic curve for each pixel. However, noise components originating from the camera and the influence of external light that could not be excluded limited the quantitative performance when signal of ion counts were weak less than 10-1cps/pixel, and the systematic evaluation of error factors was insufficient.
In this study, we attempted to develop an imaging method that reduces noise, which can be a systematic error component. A qCMOS camera, manufactured by Hamamatsu Photonics, which has less noise than in previous studies, was incorporated, and a darkroom covering the MCP/FS/qCMOS camera system was specially designed. Using this new system, we evaluated the noise components in the experimental environment. In addition, the characteristic curves were calculated by optimizing the experimental conditions using SIMS.
The readout noise and dark current noise of the qCMOS camera in the experimental environment were evaluated to be 0.41 electrons rms and 0.016 electrons/pixels/s, respectively, which are equivalent to the camera specs, and the effect of external light was negligible. By adjusting the SIMS secondary ion optics, a characteristic curve with a dynamic range of 4 orders from 10-3 cps/pixel to 101 cps/pixel was obtained by using a uniform secondary ion image of less than a few %.
In order to evaluate the possibility of statistical processing of the above obtained characteristic curves, we applied these curves to calculate the silicon isotope ratio measured with silicon wafers assumed with homogeneous Si isotope distribution . The isotopic distributions of δ29Si and δ30Si were calculated by applying the characteristic curve to the image intensity of each pixel and converting it to ion counts. The errors of δ29Si and δ30Si for each pixel calculated by propagating the errors of the constants of the characteristic curves were consistent with the isotope distributions calculated by 50 x 50 pixel without the influence of interference for 28SiH- and 29SiH-. This indicates that the errors in ion counts can be estimated by a statistical process.