To improve properties of materials, the effect of additive elements was investigated using high-throughput (HT) experiments and materials informatics (MI) technique. In case of MI, a small number of datasets becomes a problem. Furthermore, the data in the literatures were not obtained under the same conditions, so the homogeneity of the data is questionable. The use of such datasets makes it difficult to search for materials by MI. As a countermeasure, we should collect data under the same conditions by our own HT experiments. Especially, we would like to suggest that a large number of homogeneous datasets is better than a small number of highly accurate datasets for MI. In this study, we focused on phosphorescent materials [1], oxide ion conductors [2] and electrode materials for lithium-ion battery [3].

At first, material libraries were synthesized by HT experiments. In case of oxide materials, the ink-jet technique as shown in Fig. 1 [2, 3] is one of the synthetic method suitable for HT experiments, therefore we have mainly used the instrument, which had 4 ink-jet heads. In case of the ink-jet method, the so-called “coffee stain effect” often becomes a problem to cause inhomogeneous film quality. Multi-target sputtering equipment [1] is another suitable method, especially searching for alloys and semiconductors. However, the expensive target can be often problem.

Next, the properties of material libraries were estimated using HT measurement system, for example X-ray diffraction (XRD), X-ray fluorescent analysis (XRF) and X-ray absorption near edge structure (XANES) spectrum measured at SPring-8. Alternatively, we made our own dedicated HT measuring device. In this case, it is important to screen the materials using properties that can measure at high-speed without laborious experiments. Data from such HT experiments are guaranteed to be homogeneous and suitable for material screening. Note that the purpose of this study is the screening for materials, not a prediction.

Appropriate materials were searched by machine learning models using composition-based explanatory variables and experimentally-obtained objective variables. In many cases, we trained a machine learning model using the dataset by leave-one-out cross validation (LOOCV) method because the dataset is not big data. The regression model selection methods are case-by-case and there is no universal selection method.

Finally, specimens proposed by MI were synthesized, often by solid-state reaction method, and then the properties of the materials were verified experimentally. In conclusion, the materials with high properties could be searched in a shorter time. The results suggest that the combination between the HT experiments and the MI technique is effective for searching additives under the limited conditions like this study.

However, MI did not show the reason for the selection of the materials. It is very difficult to explain the reason why the MI choose the materials because the machine learning system is “black box”. If there is big data about the materials, the problem will be clarified. In the next project, we will investigate the reason and the chemical interpretability of the regression results.

[1] H. Hazama et al., Inorg. Chem., 58 (2019) 10936-10943, [2] M. Matsubara et al., ACS Comb. Sci., 21(2019) 400-407. [3] S. Tajima et al., STAM method, (submitted).

X-ray photoelectron spectroscopy (XPS) is a non-destructive surface analysis technique that identifies the elements contained in a solid sample and their chemical states by measuring the energy spectrum of photoelectrons emitted by irradiating an X-ray on the sample. Simulation of Electron Spectra for Surface Analysis (SESSA) is a well-known simulator to facilitate quantitative interpretation of the spectra of XPS and Auger electron spectroscopy [1]. In principle, it is possible to solve an inverse problem using the simulator to estimate a sample structure from measurement data, but there is no automatic way to solve the inverse problem. Because users have to run the simulator and manually compare the results with measured data, there is a need to automate the process.

We have developed a framework for solving the inverse problem of XPS by incorporating the SESSA, into Bayesian estimation to obtain an overall picture of the distribution of plausible sample structures from the measured XPS data [2]. In this framework, the procedure for running the simulator, which originally required human trial and error, was fully automated. In addition, it is now possible to not only obtain the optimal values of the sample structure parameters to be estimated, but also to evaluate the accuracy of the parameters from the posterior distribution in Bayesian estimation. The drawback of SESSA, which is its high computational cost, was overcome by combining the replica-exchange Monte Carlo method with the straight-line approximation.

As a concrete example, an artificial sample with a four-layered structure of C₂O (10 Å)/HfO₂ (25 Å)/SiON (16 Å)/Si was designed, and the angle resolved XPS (ARXPS) intensity data were generated by superimposing statistical noise on the SESSA data that simulate the photoelectron intensities from the Hf4f, Si2p, C1s, N1s, and O1s orbitals at two emission angles (0° and 54.7°). Using the ARXPS intensity, we performed an inverse-problem solution in the framework to obtain the composition and thickness of each layer of the pre-designed sample, and we succeeded in estimating the sample structure, including the true model.

Figure 1 shows posterior distributions of the sample structure parameters. (a), (b), and (c) are compositional fractions of the first CON layer, second HfON layer, and third SiON layer, respectively. (d) shows posterior distributions of film thicknesses T₁, T₂, and T₃. The points indicated by open circles in Fig. 1(a)–(c) and the dotted lines in Fig. 1(d) correspond to the true values of the sample structure parameters. The posterior distributions of the compositions of the first CON and second HfON layers were very sharp. The distribution of the compositions of the third SiON layer was a linearly spreading distribution, indicating that the composition ratio of silicon to nitrogen was constant. The true values were included in those distribution. We prepared three different data sets with different measurement times, and confirmed that the estimation accuracies of sample structure parameters were higher when the time of measurement of the spectra was longer.

The advanced features and many adjustable parameters of SESSA make it difficult for users to manually search for the sample structure parameters that match the experimental data. The proposed method incorporates SESSA into Bayesian estimation to find the appropriate sample structures among many candidates, paving the way for future use of SESSA.

[1] W.S.M. Werner, W. Smekal, C.J. Powell, Simulation of Electron Spectra for Surface Analysis (SESSA) - 2.2.0, National Institute of Standards and Technology, Gaithersburg, MD, 2021.

View PDF for the rest of the abstract.

In quantitative XPS analysis, the peak area I_i of the narrow spectrum of each component i is measured and the atomic concentration C_x of component x is determined by taking the ratio

C_x=(I_x/S_x)/Σ_i(I_i/S_i)

S_i is the relative sensitivity factor, which depends on the characteristics of the measurement device and measurement conditions. For the same measurement conditions, S can be assumed to be a constant, which affects the accuracy of the measurements but not effective on the reproducibility of the XPS measurements. If the same instrument and recipe are used, and the changes in measurement environment are negligible, the precision depends on the reproducibility of the peak area I of the spectrum.

The reproducibility of the peak intensity can be easily improved by increasing the measurement time sufficiently long and improving the signal-to-noise ratio. However, in practical XPS analysis, it is often required to submit correct directional results at maximum speed even though the measurement time cannot be lengthened in many cases, such as when measurement is difficult, peak intensity is weak due to trace components, or sufficient measurement time cannot be taken because the sample is degraded. However, the correct directional results are often required to be submitted with the best possible speed. These issues of reproducibility of practical XPS analysis should have been dictated in the field, but is thought to be often forgotten and not to be proven.

In XPS quantitative analysis, the energy range of the peak is determined and the area of the peak is measured using an appropriate background function (e.g., Shirley method, linear, etc.). To extract error factors attributable to the analyst in this method, as simple as possible, consider the case of a Gaussian peak calculated with a linear background. As is field proven[1], the intensity at each measurement point (j_q) can assume a Poisson distribution except in the case of extremely low counts and extremely high count rates. Therefore, the measurement error in the XPS peak area can be estimated numerically when the noise (√(j_q)) is the standard deviation of the square root of the counts.

The peak area is obtained by subtracting the background (BG)-derived area from the peak-derived area in the determined energy range.

Using n=1 (background determined by 2n+1=3-point averaging) as an example commonly used in practical XPS analysis, the impact of background on quantitation is roughly 3 times larger than that of peak origin. When the peak-to-background ratio (SBR) becomes about 10, the influence of the approximation almost disappears, and when it becomes about 100, the background-derived noise becomes negligible, but there are not many cases where such time-consuming measurements are performed in practical XPS analysis. In other words, when quantification is performed under conditions with poor S/N, background noise can be considered to affect the accuracy of the peak intensity.

In conclusion, it was shown that it is better to calculate the peak area by "setting the peak range as narrow as possible and just at the edge of the peak" for peaks with poor signal-to-noise ratios such as practical XPS analysis. Another strategy is taking plenty of background points[2]. Optimal number of points depends on the level of reliability required for the data.

References

[1] S. Fukushima, J. Surf. Anal. 16, 70 (2009).

[2] K. Harrison, L. B. Hazell, Surf. Interface Anal. Mb>18, 368 (1992).

Surface treatment and coating process involves changing the properties of the surface in a demanded purpose. For getting effectively coating techniques involving the experience and knowledge of expert, the faster learning process when engaging coating is one of the most important key factors. We have been carrying out research and development of several coating databases. We would outline a term of ‘coating process informatics’ that allows. These approaches would be effective for coating design, doing trouble care, obtaining coating know-how, and accumulating well-skilled person’s knowledge.

1. Questionnaire results from job shop small and medium enterprises (SMEs)

we show a portion of the results obtained from questionnaires from job shop SMEs. The companies replied to our questionnaires were 70 (electro-plating), 32 (thermal spraying), and 17 (PVD/CVD). We asked for coating methods they engaged, the substrate they used, wear resistance coating materials, and applications, etc. Also, we asked for a few questions related to the human skills involved in the coating process; how skills are maintained in SME, how skills are transferred to younger generation, etc. The results of questionnaire would be helpful in future business, business operations and R&D.

2. Coating/surface treatment database We have been developed that coating/surface treatment databases which were WEB-based public service in Japan via the Internet. These databases were focused on surface treatment manufacturing field such as electrical and electroless plating, thermal spraying, PVD and CVD. One of the software functions of database is as follows. Searching the data using these databases, case data could be obtained from actual experiments classified by the function of film properties such as hardness, coefficient of friction, electrical resistivity, crystallinity, and so on. Using developed database software, persons engaged in coating would be able to find suitable coating condition, share information and obtain useful technological hints for the resolution of engineering problem.

3. Coating/Surface treatment process mapping using machine learning method We also apply support vector machines for making decision boundary of coating process map is chosen to be the one for which the margin is maximized. Multiple process maps of hardness, brightness, and internal stress of electrically plated Cr film, which are add in the same 2D axis of bath temperature and current density. As well-known, the overlay multiple coating properties of mechanical property and optical property has been mostly used in industrial application such as roll, plastic parts, etc. In this application, expert would be engaged to keep the plating condition, and would know the importance of the combination of film properties. Our approach would be succeeded in digitalizing plating expert’s knowledge.

4. Convertible PVD by Replacing Coating Components We also developed a methodology of PVD thin film deposition which allowed the process comparison between different PVD processes. We divided the PVD equipment by component based on coating phenomena and redesigned each component. By replacing components, one can realize desired PVD coating on same vacuum chamber such as vacuum deposition, ion plating, sputtering, etc. We call it ‘‘Convertible PVD’’ system. Thin film of aluminum oxide has been formed in three PVD processes of RF ion plating, RF sputtering and EB vacuum deposition. By comparing coating data of obtained film properties, the proper selection of coating process would be done when using convertible PVD system.

In summary, we would show our developed coating/surface treatment databases, and simulation related to coating field as ‘‘coating/surface treatment process informatics’’ tool. We believe that these studies would be able to be useful in faster coating prototype making and accumulating expert’s knowledge.

Machine-learning of interatomic potentials using data from first-principles calculations has been actively discussed in the field of data-driven materials science. This approach has been gaining attention as it offers prediction accuracy comparable to first-principles calculations while maintaining low computational costs. Such computational methods are crucial for analyzing phenomena observed within complex solid systems, either internally or at their surfaces. With this situation in mind, by employing the high-dimensional neural network potential (NNP) [1], one of the machine-learning interatomic potentials, we have been working on its applications and developments.

In this talk, we will first present the thorough exploration of stable phases in Au-Li binary alloy systems and the dynamical calculations of alloying processes of Au/Li interfaces using the NNP [2]. Subsequently, based on our preliminary results, we will discuss the applicability of NNPs for investigating ion dynamics in glass ceramic materials as well as for exploring high-entropy alloy catalysts.

In addition to the above, we proposed a different scheme of NNP to analyze the point defect behavior in multiple charge states [3]. Particularly, we will demonstrate the prediction performance of the proposed NNP using wurzite-GaN with a nitrogen vacancy with charge states of 0, 1+, 2+, and 3+. Furthermore, we developed a NN model to predict the Born effective charges of ions from atomic structures [4]. The proposed NN model was used to evaluate the external forces acting on ions from electric fields. Using Li₃PO4 as a prototype material, we will present the Li ion motion under electric fields in both crystalline and amorphous structures.

This work was supported by JST CREST (JPMJCR1996), e-ASIA JRP, and JSPS KAKENHI (19H02544, 20K15013, 22H04607, 23H04100).

References:

[1] J. Behler and M. Parrinello, Phys. Rev. Lett. 98, 146401 (2007).

[2] K. Shimizu, E.F. Arguelles, W. Li, Y. Ando, E. Minamitani, and S. Watanabe, Phys. Rev. B 103, 094112 (2021).

[3] K. Shimizu, Y. Dou, E.F. Arguelles, T. Moriya, E. Minamitani, S. Watanabe, Phys. Rev. B 106, 054108 (2022).

[4] K. Shimizu, R. Otsuka, M. Hara, E. Minamitani, and S. Watanabe, arXiv.2305.19546.

Spectroscopic measurements such as electron spectroscopy for chemical analysis (ESCA) and Raman are fundamental tools for obtaining information on the structure and properties of crystals. Recently, the sophistication, complication, and automation of measurement and the emergence of machine learning, which requires huge amounts of data and/or analysis, have led to an increase in the number of datasets and an explosion in the burden of its analysis quickly and accurately.

Transition metal dichalcogenides, such as MoS₂ and WS₂, are establishing themselves as the possible post-Si candidate as the next generation semiconductor. However, the issue of immaturity of the doping technique, which is the most fundamental requirement for the application of semiconductor devices, has not yet been fully solved. To solve the issue, we have been developing carrier doping technique for MoS₂ by elemental substitutional doping during the chemical vapor deposition growth. Through the study, we have successfully obtained partially-Nb-doped MoS₂ monolayers (i.e., MoS₂–MoS₂:Nb) [1]. In this talk, we will focus on machine learning analysis of spectroscopic data of MoS₂–MoS₂:Nb using ESCA.

The optical microscope image of the obtained sample is shown in Fig. 1a. Triangular-shape crystal can be found, which is a typical crystal shape of monolayer MoS₂. The contrast is different only at the outer edge, indicating that the optical properties of the obtained sample have changed. Raman spectra showed that the inner region was pristine MoS₂, while the outer one was doped MoS₂ (data not shown in the abstract). To visualize the planar distribution of Nb⁴⁺ 3d_5/2 of the crystal, we used a synchrotron soft X-ray scanning photoelectron microscopy system, which can obtain three-dimensional spatially resolved ESCA datasets, called “3D nano-ESCA” [2]. The lateral spatial resolution of “3D nano-ESCA” is ~100 nm. However, due to the peculiarities of the obtained results, it was challenging to analyze the obtained spectra by curve fitting with hands. Therefore, we performed spectral curve fitting using the spectral-adopted expectation-constrained maximization algorithm [3, 4]. The results are shown in Fig. 1b. The presence of Nb⁴⁺ was confirmed only at the outer edge of the crystal, confirming the selective Nb doping at the outer edge. Furthermore, we have made Mo⁴⁺ 3d_5/2 peak position mapping through the same process (Fig. 1c). We observed a low binding energy shift of Mo⁴⁺ 3d_5/2 peak, indicating the p-type nature of the Nb-doped region.

In summary, we have successively measured and analyzed spatially-resolved ESCA spectra of MoS₂–MoS₂:Nb with machine-learning-based spectral analysis. 3D nano-ESCA measurement and following analysis showed the Nb doping and corresponding p-type nature at the edge region of the crystal. Not only the ESCA spectra, the fast and automated curve fitting method would drive and save time for the analysis of various spectroscopic measurements.

References

[1] M. Okada et al., APL Mater., 9, 121115 (2021).

[2] N. Nagamura et al., Carbon, 152, 680 (2019).

[3] T. Matsumura et al., Sci. Technol. Adv. Mater., 20, 733 (2019).

[4] T. Matsumura et al., Sci. Technol. Adv. Mater.: Methods, 1, 45 (2021).

In light of the growing demand for rapid advancements in materials and manufacturing, computational materials science, especially data-centric methodologies, have seen marked growth. Historically, these methods centered on molecular descriptors, such as elemental species, bond types, and electronic interactions, aiding in the discovery of molecules, inorganic compounds, and crystals. Yet, their effectiveness diminishes for materials like plastics, metal alloys, and rubber with undefined structures due to the necessity for precise elemental coordinates. This paper presents a cutting-edge method in materials design: multimodal deep learning, which bridges the gaps inherent in traditional techniques. By mimicking human cognition, it synthesizes information across sensory channels (e.g., sight, sound) for a comprehensive grasp. Although multimodal deep learning's efficacy is evident in human-centric domains like emotion analysis, its application to materials design has not been achieved. In our approach, we emphasize the integration of extensive data on material structures. Our novel methodology includes: (1) generative deep learning models characterizing physical or chemical structures, and (2) a combined model integrating diverse data to predict material properties. Its efficacy was demonstrated in polymer composites across ten compositions, predicting eight distinct properties using diverse data sources such as optical microscope images, mid-infrared spectra, and Raman spectra, and compositional information. Based on our trained model, over 114,210 compositional scenarios were virtually examined, discerning optimal compositions and multi-property trade-offs. We assert that this innovative method holds vast potential for diverse materials, offering deep insights and shaping the trajectory of future materials design research.

References:

[1] Shun Muroga, Yasuaki Miki, Kenji Hata, Advanced Science, 2302508 (2023).

[2] Shun Muroga, Yasuaki Miki, Kenji Hata, arXiv:2303.16412 (2023).

[3] Shun Muroga, figshare (2023), available at https://doi.org/10.6084/m9.figshare.23358398.

Acknowledgements:

This work was supported by a project (JPNP16010) commissioned by the New Energy and Industrial Technology Development Organization (NEDO). Computational resource of AI Bridging Cloud Infrastructure (ABCI) provided by National Institute of Advanced Industrial Science and Technology (AIST) was used.

Fuel cells, which use the reaction of H₂ and O₂ as an energy source, are attracting attention as an essential device for a low-carbon society. In a polymer electrolyte fuel cell, platinum is used as a catalyst in cathode to promote oxygen reduction reaction. However, platinum needs to be replaced because of its high cost and limited reserves. Nitrogen-doped carbon is focused on an alternative catalyst for platinum owing to its high durability, low-cost, and high activity in basic media. However, the catalytic activity is very low in acidic media in cell assembly. Up to now, we have clarified pyridinic nitrogen (pyri-N) is the active site of nitrogen-doped carbon catalysts¹⁾ and the reduction reaction of pyri-NH⁺ (protonated pyridinic nitrogen) proceeds simultaneously with O₂ adsorption in acidic media²⁾. However, the reason of the decreased activity in acidic media has not been clarified yet. Therefore, to establish design principles for highly active catalysts, we aim to investigate the mechanism of the activity decrease based on the role of pyri-N in the elementary process.

For this purpose, we used a model molecular catalyst which has uniform active site because practical catalysts are difficult to analyze the mechanism due to the existence of several nitrogen species. The model catalyst was prepared by impregnating carbon black (CB) with 1,10-phenanthroline (1,10-phen) which includes pyri-N. Using this catalyst, the role of pyri-N was investigated from electrochemical measurement and electron density by a surface chemical method of XPS.

First, I measured 1,10-phen/CB model catalyst activity by electrochemical measurement in each pH. I plotted the onset potential which is the current density reached -0.1 mA/cm² as a function of pH as shown in Figure 1A. The slope of onset potential was changed at pH 5. This is due to a change in the acid-base equilibrium of pyri-N, thereby changing the reaction mechanism. Also, X-ray photoelectron spectroscopy was used to observe the N 1s spectra after applied potential in solution under the oxygen atmosphere. It was found that the formation of pyri-NH, a hydrogenation-reduction of pyri-N, took place simultaneously with oxygen adsorption at higher potential than the onset potential in any pH (Fig. 1B). This indicates that pyri-NH behaves like a radical so that it is easy to take place the adsorption of oxygen molecules on the nitrogen-doped carbon catalyst. Also, the fact that the formation potential of pyri-NH is higher than the onset potential in any pH indicates that the elementary process which determines the onset potential is the after pyri-NH formation reaction.Lastly, the formation potential of pyri-NH in acidic media was lower than the one in basic media. That means that hydration of pyridinium is stabilized in acidic media so that the catalytic activity decrease in acidic media. Thus, we obtained the guideline for increasing catalytic activity by importing hydrophobicity because of prevention the hydration.

1) D. Guo, T. Kondo, J. Nakamura, et al., Science, 351, 361-365 (2016).

2) K. Takeyasu, J. Nakamura et al., Angew. Chem. Int. Ed. 60, 5121 (2021).

The proton-doped SmNiO₃ (SNO) thin film, where protons are dissociated from hydrogen molecules by utilizing Pt catalytic effect and then doped into the film (hydrogenation), can lead to an ~10⁸ of resistance modulation at 300 K. Recently, the giant resistance modulation has been demonstrated to be directly related to the proton doping concentration within the perovskite lattice of RNiO₃. Since protons in SNO tend to aggregate at the disordered crystal structures such as grain boundaries and defects, the SNO films without crystal inhomogeneity are required to improve the protonation resistive switching property. One of the effective approaches is improve the surface condition (flatness, crystal structure) of the substrate, where the thin film growth starts. In this study, we attempted to improve protonation resistive switching property by virtue of on a atomically-ordered and flat LaAlO₃(LAO)(001) substrate. Commercially available single-crystal substrates (pristine-LAO) and smoothened LAO substrates (STEP-LAO) with an atomic step/terrace structure were used for the SNO film growth. The 30 nm-thick SNO films were deposited on these substrates and the change in resistivity during hydrogen doping was monitored. The time profile of the resistance of SNO films on pristine-LAO and on STEP-LAO showed a steep resistance increasement at about t~200 sec for both LAO films while the difference in the saturated resistance switching ratio was observed. The one-order higher resistance change in the SNO/STEP-LAO is thought to reflect the ordered and flat LAO substrate surface, which lead to an undistorted interface following the suppressed the crystal inhomogeneity of the SNO film. In this presentation, we will discuss in detail the crystallinity enhancement of the SNO thin films by virtue of atomically-ordered and flat substrate with an ideal crystal structure, and the hydrogen diffusion resistance switching characteristics derived from this enhancement.

Introduction

In recent years, as part of efforts to reduce carbon dioxide, research has been conducted on CCU, a technology to generate useful substances from CO₂. In the CO₂ conversion by using heterogeneous catalysts, C1 products such as methanol, CO, and CH₄ are produced. On the other hand, it has been reported that ethanol was produced from CO₂ without the application of potential using heterogeneous catalysts with two components. However, the activation energy for C-C bond formation is more than 1.5 eV. This means that the reaction hardly proceeds only by thermochemical heterogeneous catalysis. Therefore, we assume a contribution of electrochemical processes. An example of an electrochemical reaction that can proceed without the application of an external potential is the formation of a mixed potential. We have been attempting to demonstrate the mechanism of electrochemical reactions using a model reaction system, assuming that the reaction proceeds as an electrochemical reaction driven by the overvoltage caused by the formation of a mixed potential. In this study, we aim to prove the the reaction mechanism and to clarify the reaction pathway for the ethanol production.

Method

In this study, experiments were conducted using a model reaction system to detect the reaction current in a mixed potential-driven reaction. First, Cu nanoparticle catalysts and Pd/CB (carbon black) catalysts, which are considered to be responsible for CO₂ reduction and hydrogen oxidation, were coated on carbon paper and used as electrodes. The reaction current flowing between the electrodes was measured with an ammeter installed outside the reactor. The model reactor was filled with a CO₂/H₂ mixture at 40 atm and the reaction temperature was 150℃. The amount of product and the reaction current were measured.

Result

The reaction current was observed between the Cu nanoparticle and Pd/CB catalyst electrodes, and electrochemical reduction proceeded at the Cu nanoparticle electrode and electrochemical oxidation at the Pd/CB catalyst electrode. Methanol, ethanol, and CO were confirmed as products by gas chromatography. The reaction current was measured as a function of CO₂/H₂ pressure at room temperature, and it was found to increase and decrease reversibly with the introduced pressure. The amount of CO, the product of CO₂ hydrogenation, was measured by switching the electrical connections between catalysts, and a constant rate of CO was produced regardless of the electrical connections. The amount of product was measured while switching the electrical connections between the catalysts. This indicates that electron transfer is not involved in the CO formation process, which is a thermochemical reaction. Furthermore, the formation of formic acid as a liquid phase product was confirmed by liquid chromatography. As shown in the Fig.1, the selectivity of formic acid is 99.7%, which is very high, indicating that it is the main product of this model. Since there was no significant difference in the amount of formic acid formed between the short-circuited and non-short-circuited cases, we assume that it is formed by thermochemical reaction. We also conducted experiments using a CO₂ reduction catalyst other than Cu, considering that the reduction of oxides on the Cu catalyst proceeds during the reaction and affects the current value.

Reference

1) Bai et al. JACS 140, 524 (2018).

Chemical reactions at surfaces have been widely used for chemical processes such as catalytic synthesis, energy conversion, environmental cleanup, and sensor. Surface science techniques enable us to understand physicochemical fundamental processes on surfaces. However, one drawback of such surface science techniques is that the experiments are carried out under vacuum in many cases. Recent decades, some advanced experimental techniques have been developed to overcome such pressure gap problems, which are so called in situ and operando techniques.

We have developed some in situ / operando experimental techniques for observing surface reactions on liquid/solid and gas/solid interfaces in energy range from infrared to soft X-ray. Here, we present our recent results obtained by ambient pressure X-ray photoelectron spectroscopy (AP-XPS) and X-ray absorption spectroscopy (AP-XAS). The experiments were carried out at a soft X-ray beamline BL-13 at the Photon-Factory of High Energy Accelerator Research Organization (KEK-PF) in Tsukuba, Japan. The AP-XPS system is consisted of a high-pressure chamber, preparation chamber and load-lock chamber. The high-pressure chamber is equipped with an electron energy analyzer modified for high pressure experiments via the differential-pumping system. The XAS system is consisted two sections, high-vacuum front-end and ambient pressure reaction cell. A Si₃N₄ membrane is sandwiched between front-end and reaction cell. The reaction cell consists of CF 40 SUS flanges. A gold-coated cupper plate is used as a collector electrode. Further details of our experimental techniques are described elsewhere[1,2].

Figure 1 shows a result of operando AP-XPS measurement for H₂ sensing Pt-Rh thin-film sensor. The Pt-Rh sensor detects the atmospheric concentration of H₂ gas by changing in electric resistivity. Here, a 10 nm-thick Pt–Rh thin-film deposited on a SiO₂ substrate was used for the measurements. Figure 1(a) shows time evolution of relative electric resistivity (ΔR/R). The resistivity decreases with exposing H₂ gas to the sensor surface, whereas it increases with exposing O₂ gas. Figure 1(b) shows corresponding Rh 3d and Pt 4f XPS. Before the gas dosing, the surface was dominated by Rh oxide. When the surface was exposed to the H₂ gas, the chemical state clearly changed. The Rh oxide was completely reduced to the metallic state. When, the surface was exposed to the O₂ gas, the oxygen-induced species grew up again. Those findings indicate the surface chemical state strongly relate to the material functions (i.e. electric resistivity).

References

[1] Toyoshima, R. et al. J. Phys. Chem. Lett. 2022, 13, 8546–8552.

[2] Shimizu, H. et al. Phys. Chem. Chem. Phys. 2022, 24, 2988–2996.

Methanol synthesis by hydrogenation of CO₂ is industrially used with Cu/ZnO/Al₂O₃ catalysts, but the effect of Zn in the catalyst and other factors remain to be clarified. We have so far created an energy diagram of the methanol synthesis by CO₂ hydrogenation on the model surface of a Cu(111). By comparting the energy diagram for Zn/Cu(111), we are aiming to clarify the role of Zn for the methanol synthesis. In the experiments, the Cu(111) surface was cleaned by Ar⁺ sputtering and annealing at 773 K. Zn/Cu(111) was prepared by depositing Zn on the Cu(111) at room temperature. Formate was prepared by exposing the surface to formic acid also at room temperature. To observe the surface species and the kinetic characters, infrared-reflectance absorption spectroscopy (IRAS) and temperature programmed desorption (TPD) were performed. TPD was performed for Zn (~0.2 ML) /Cu(111) with adsorbed formate, which showed two peaks at 399 K and 472 K (Figure 1). The peak at 399 K is attributed to the decomposition of formate on Cu(111). On the other hand, the additional peak at 472 K was assigned to the decomposition of formate on Cu-Zn alloy sites. From this result, the activation energy for formate decomposition corresponding to the new peak was estimated to be 128 kJ/mol, which is 20 kJ/mol higher than that of formate on Cu(111). This means that the formate is stabilized by Zn. The roll of Zn was discussed based on an energy diagram of methanol synthesis for CO₂ hydrogenation.

Introduction Global warming caused by greenhouse gases such as carbon dioxide is currently an international issue. Carbon capture and sequestration (CCS) has become important for mitigating global warming and the method for CCS is actively studied. [1] Among the CCS methods, those using polymeric gas separation membranes have attracted particular attention as a separation method with a lower energy load than conventional methods. We are studying incorporation of ionic liquid into the porous membranes including polymers and covalent organic frameworks (COFs)[2]. In this presentation, we report MD simulations of the thin films of ionic liquids ([BMPyr][DCA], Fig. 1) and CO₂ and compare it with experimental results on the CO₂ gas separation membranes. Methods LAMMPS was used as the MD engine and the Generalized Amber Force Field (GAFF) was employed as the force field. The charge of each atom was calculated using Gaussian 16w with B3LYP/6-31G(d,p) level of theory. About 1nm thick film of the ionic liquid was prepared and one side was loosely fixed with a wall with van der Waal-type potential working only for the ionic liquid. Initially the ionic liquid and CO₂ gas were equilibrated separately and then the wall between the ionic liquid and the CO₂ was removed. The motion of CO₂ after moving out from the ionic liquid film beyond a certain distance was restricted to simulate pressure difference across the ionic liquid film. Part of the calculations were performed using the Okazaki Center for Computational Science. Results and Discussion Figure 2 (a) shows the initial configuration, where the ionic liquid film was separated from CO₂ gas. Figures 2(b) and (c) are snapshots 200 ps after starting interaction. At 300 K (Fig.2(b)), CO₂ agglomerates on the surface of the ionic liquid, and filamentary penetration behavior was observed. On the other hand, at 600 K, CO₂ dissolves and diffuses as a single molecule (Fig. 2(c)). The solubility and number of the permeated CO₂ molecules at different temperatures are shown in Figs. 2(d) and (e), respectively. Experimentally, we have observed non-monotonical dependence of various permeation parameters as a function of temperatures. Comparison with the MD results and experiments is under way. Figure 3 shows the powder x-ray diffraction pattern of the COF before and after immersing in the ionic liquid. The results show that the absolute intensity of COF decreased more when COF was impregnated with ionic liquids than when COF was impregnated with ethanol. The intensity ratio of in-plane related peaks (hk0) has also changed. MD simulations are being performed to investigate the factors that cause these changes in the X-ray diffraction peaks.

[1] A.Katare et al., ACS Omega 8, 17511 (2023).

[2] M. Kato et al., ACS Appl. Nano Mater. 5, 2367 (2022).

Ice provides one of the most important molecular crystals in our life. Because of its critical role in science and engineering, a number of experimental studies have been conducted on ice–vacuum [1] and also on ice–vapor [2] interfaces . However, molecular-scale knowledge of the ice–water interface is still quite limited. The problem is that an ice–water interface fluctuates in space, even when the interface is held exactly at the freezing point. In addition, the interface is lost when its temperature deviates from the freezing point.

To solve this problem and to analyze the interface at the molecular level, we considered using atomic force microscopy (AFM) at the interface between ice and organic solvents. For instance, ice in contact with liquid octanol (C₈H₁₇OH) is stable at temperatures below the freezing point of ice (0°C) and above the freezing point of octanol (-16°C), although the organic solvent is not equivalent to water, we expected that it is likely to mimic some features, hopefully important features, of the ice-water interface.

We investigated different ways to keep the ice–liquid interface below 0°C on a Dimension XR Icon NanoEC microscope (Bruker) being operated at the Institute for Molecular Science. The best way we found was to cool the entire microscope placed in an acoustic enclosure. The temperature inside the box was controlled at -10 ± 5°C stable enough to take topographic images of ice films in octanol with a spatial resolution of 0.1 nm (Fig. 1). In a nitrogen vapor atmosphere, ice films exhibited a more rippled topography than those observed in octanol. The finite dissolution of water to octanol and recrystallization on ice may have helped flatten the ice–octanol interfaces. In addition to topographic imaging, force curves were measured to follow the elastic response of the ice under the AFM tip.

References

[1] J. Peng, J. Guo, R. Ma, Y. Jiang, Surf. Sci. Rept. 77, 100549 (2022).

[2] F. Tang, T. Ohto, S. Sun, J. R. Rouxel, S. Imoto, E. H. G. Backus, S. Mukamel, M. Bonn, Y. Nagata, Chem. Rev. 120, 3633 (2020).

Introduction

Methanol is a valuable feedstock for various chemical products, including formaldehyde, formic acid, and acetic acid ^[1]. Copper is an essential metal in catalytic reactions involving methanol synthesis, methanol dehydrogenation, water gas shift reactions etc. Among these reactions, the dehydrogenation of methanol where formaldehyde is formed, is particularly important.

Methoxy species is an important intermediate for conversion of methanol to formaldehyde. However, the dissociation does not take place on the clean Cu flat surfaces [2]. In this study, we conducted the following experiments using the modified Cu surfaces to investigate the effects of steps and additives.

The first model catalyst is the Cu(977) step surface. It is known that the step sites exhibit different catalytic activity from those of the terrace sites [3]. The second one is the Pd deposited Cu(997) surface. Since Pd is known to have a strong interaction with hydrogen, we expect that a small amount of deposited Pd atoms on a Cu surface would facilitate the dehydrogenation of methanol with the advantage of retaining the catalytic properties of Cu metal [4].

In this work, we have investigated the adsorption, dissociation and desorption of methanol and its reaction products on the Cu(977) and Pd/Cu(977) surfaces by using TPD and infrared reflection absorption spectroscopy (IRAS).

Experimental

All experiments were performed in an ultra-high vacuum (UHV) chamber. The Cu(977) clean surface was prepared by repeated cycles of Ar+ sputtering and annealing at 645 K. The Pd/Cu(977) surface was prepared by depositing Pd atoms on the Cu(977) surface at 380 K in UHV. The prepared surfaces were cooled to 85 K with liquid nitrogen, and then gaseous methanol was introduced on the surfaces. We have conducted TPD and IRAS measurements on these surfaces.

Results and Discussion

From the TPD spectra of methanol adsorbed Cu(977) surfaces, three desorption maxima by methanol were observed. Based on the comparison with the previous study [5], the peaks at 141 K, 153 K, and 195 K can be attributed to the methanol desorption from multilayer, terrace and step, respectively.

On the clean Cu(977) surface, a desorption peak for m/e=30 was little observed, indicating most of methanol were molecularly desorbed without any reaction. On the other hand, on the Pd/Cu(977) surface, a significant desorption peak was observed at 350 K. This peak can be assigned to the desorption of formaldehyde formed via methoxy decomposition [3].

A part of methanol molecules adsorbed on Pd/Cu(977) led to methoxy species formation (Equation (1)), followed by decomposition-limited desorption as formaldehyde and hydrogen above 300 K (Equation (3)).

CH₃OH(ad) → CH₃O(ad) + H (ad) (1)

H(ad) → 1/2 H₂(g) (2)

CH₃O(ad) → CH₂O(g) + 1/2 H₂(g) (3)

The present experimental results indicate that the presence of Pd on Cu(977) lowers the activation barrier of methanol decomposition.

Figure shows the TPD spectra of methanol adsorbed Cu(977) and Pd/Cu(977) for m/e=2. The 250 K peak is due to hydrogen desorption by the dissociation of methanol into methoxy and hydrogen (equation(2)), while the peak at 350 K is associated with methoxy decomposition (equation(3)); these peak intensities were nearly equal. From these results, we conclude that the methoxy species selectively decomposes into formaldehyde and hydrogen with almost 100% efficiency.

In the presentation, the experimental results by IRAS measurements will be discussed including the intermediate species of methanol dissociation as a function of temperature.

View PDF for the rest of the abstract.

On photocatalytic hydrogen evolution, photogenerated electrons reduce proton derived from H-containing molecules, such as methane and water, to produce hydrogen; thus, the photogenerated electrons are key active species toward sustainable hydrogen supply. Briefly, electrons generated in the conduction band of semiconductor photocatalysts are immediately trapped at various in-gap electronic states in the timescale of less than 1 picosecond [1]. Then, only fractions of these stabilized electrons are transferred to surface active sites to induce reduction reactions with the timescale in the order of microseconds or milliseconds [2]. To prolong lifetime of the photogenerated electrons and enhance the photocatalytic performance, metal cocatalysts are often loaded on oxide semiconductor photocatalysts. The metal cocatalysts have traditionally been asserted to serve as sinks of the reactive electrons and to provide reactive sites for proton reduction [3]. However, microscopic understanding of the reactive long lived photogenerated electrons on metal-loaded oxide systems under reaction conditions is lacking, and frameworks for rational design of high-performance catalysts remain unclear.

Here, we conducted operando FT-IR spectroscopy to unveil the reactive electrons contributing to hydrogen evolution reactions. We focused on steam reforming of methane (CH₄ + 2H₂O → 4H₂ + CO₂) for Pt- or Pd-loaded β-Ga₂O₃ particulate catalysts as a model system of photocatalytic hydrogen evolution [4, 5]. The catalyst temperature inevitably rises owing to irradiation of excitation light, and the faint signal of the reactive species generally gets overwhelmed by the huge optical response derived from thermally excited electrons [6]. To suppress the rise of sample temperature and reveal the buried reactive electron species, we periodically modulated the intensity of excitation light in the appropriate reaction timescale of the order of milliseconds [2]. By synchronizing Michelson interferometer with the modulation period, we succeeded in removing most of the signal derived from sample heating and observing photogenerated electrons under reaction conditions. We demonstrated that the only band intensity derived from electrons trapped at the in-gap state ~0.26 eV below the conduction band minimum of Ga₂O₃ changed in clear response to the enhancement of hydrogen formation rate among the several observed absorption bands attributed to various photogenerated electron species. This indicates that these shallowly trapped electrons in Ga₂O₃ directly contribute to hydrogen evolution instead of electrons in metal cocatalysts. Furthermore, capacity of the reactive electrons was drastically changed between Pt/Ga₂O₃ and Pd/Ga₂O₃, suggesting that these reactive electronic states are localized at the periphery of metal cocatalysts, i.e., three-phase boundaries between oxide, metal cocatalyst and reactant gas. Precise design of the three-phase boundaries could be relevant for further efficient photocatalytic processes.

This work was supported by JST-PRESTO [JPMJPR16S7], JST-CREST [JPMJCR22L2]; JSPS KAKENHI Grant-in-Aid for Specially Promoted Research [JP17H06087], for Scientific Research (A) [JP22H00296], for JSPS Fellows [JP22J13055]; Joint Research by the National Institutes of Natural Sciences [01112104]; and Demonstration Project of Innovative Catalyst Technology for Decarbonization through Regional Resource Recycling, the Ministry of the Environment, Government of Japan.

References

[1] K. Shirai, G. Fazio, T. Sugimoto et al., J. Am. Chem. Soc., 140, 1415 (2018).

View PDF for the rest of the abstract.

Introduction

(La,Sr)MnO3 (LSMO), a perovskite manganese oxide has been drawing significant interest in its potential applications as a next generation spintronic material owing to its ferromagnetic and half-metallic properties at room temperature. Along this purpose, recent studies on (La,Sr)(Mn,Ru)O3 (LSMRO) thin films, where about 10% of Mn at the B site of LSMO is substituted with Ru, have been one of the attempts to further modulate the properties of LSMO. These films, different from the parent LSMO, exhibit perpendicular magnetic anisotropy while maintaining room-temperature ferromagnetism ^[1]. On the other hand, in the research field of magnetic metal alloys, compositionally-graded structures, where the composition continuously varies from the interior to surface of a material, have been extensively investigated to lead to the discovery of unique magnetic properties that had never been found in the corresponding homogeneous compositions ^[2]. However, there are few reports on the fabrication of such graded-composition structures for magnetic oxides such as LSMO. This study thus focuses on the fabrication of compositionally-graded epitaxial thin films of LSMRO and investigates possible influences of the graded structures on the magnetic properties.

Experimental section

All thin films, with their thickness of 30 nm, were fabricated on (LaAlO₃)_0.3-(SrAl_0.5Ta_0.5O₃)_0.7 (LSAT) (001) substrates, employing pulsed laser deposition with a rapid beam deflection (RBD-PLD: Pascal co.) system [3]. The Ru composition and its gradient structure were controlled by nanoscale alternating ablation of two sintered ceramics targets of LSMO and Ru15%: LSMRO. The oxygen partial pressure was maintained to be 200 mTorr throughout this process, while the substrate temperature was set to 675 °C. The laser conditions were set to a fluence of 0.6 J cm⁻² and a repetition rate of 2 Hz, respectively.

Results and discussion

A compositionally-graded Ru-LSMO epitaxial thin film, with its Ru-gradient ranging from 0% to 15% along the growing direction (referred to as “UP-graded film”), was confirmed to coherently grow in the (001) orientation on an LSAT (001) substrate. From reciprocal space mapping measurement (Fig. (a)), it exhibited that the peak shape extended over the peak positions of homogeneous Ru 0% and 15% epitaxial thin films, a distinct peak which is characteristic of the compositionally-graded structure ^[4]. SIMS measurements also confirmed successful implementation of the designed compositional gradient structure. M-H measurements at 100 K (Fig. (b)) revealed that the introduction of Ru substitution, irrespective of whether it was in the homogeneous or gradient mode, resulted in the disappearance of magnetic anisotropy and the enhancement in coercivity. However, there were also found some different magnetic behaviors between them. In the homogeneous films, the coercive force more increased with higher Ru%, while the saturation magnetization showed a monotonic decrease: 3.6 μB/u.c. for Ru 0%, 3.2 μB/u.c. for Ru 7.5%, and 2.7 μB/u.c. for Ru 15%. In contrast, UP-graded film still maintained a high saturation magnetization value of 3.6 μB/u.c., which was very close to that of the Ru 0% film, and its coercive force values were 150 Oe (in-plane) and 250 Oe (out-of-plane), as similarly observed in the Ru 7.5% homogeneous film (Note: the Ru composition was set to the same as the average one in UP-graded film for comparison). From these results, the emphasis should be placed on such unique magnetic properties of UP-graded film distinct from those of the homogeneous Ru-substituted films.

View PDF for the rest of the abstract.

1.Introduction

Deep ultraviolet (DUV) and vacuum ultraviolet (VUV) radiation, with wavelengths ranging from 250 to 200 nm and 200 to 100 nm respectively, have significant technological applications, such as in photochemical processing, sterilization, and surface treatment, including semiconductor substrate cleaning. To meet the demands of these applications, tremendous research has gone into the development of DUV and VUV light sources. Equally crucial is the development of detectors for this short wavelength region.

This work harnesses the wide band gap of Titanium dioxide (TiO₂) show that TiO₂ thin films can be used as photoconductive detectors of VUV radiation. Moreover, the study explores irradiating TiO₂ thin films with gamma rays for their potential use as photoconductive detectors for high-energy ionizing radiation and to investigate TiO₂'s radiation hardness for such applications.

2.Experiment and Result

TiO₂ thin films were fabricated on the surface of unheated soda lime glass substrates through magnetron sputtering. The deposited film had a thickness of 100 nm and was annealed at 450℃ for 8 hours after sputtering. To create the photoconductive detector, interdigitated aluminum electrodes whose thickness is 500nm were deposited on the thin film. Gamma ray irradiation was performed for 4 hours at room temperature.

Fig. 1 (a) shows the photoconductivity of the thin film before and after irradiation with gamma rays. After gamma-ray irradiation, the photo current decreased by about an order of magnitude, and the photosensitivity has likewise decreased by a factor of about 8.

Fig. 1 (b) displays transmission spectra of the gamma ray-irradiated thin films. The transmission edge is red shifted as the absorbed dose increased. And The Tauc plots in Fig. 1 (c), calculated from the deconvolved transmittance, reveal that the optical band gap decreased from 3.6 eV for the pristine thin film to 3.3 eV for the thin film with the highest absorbed dose. They are attributed to the formation of trivalent titanium (Ti³⁺) ions as defects induced by gamma rays. Oxygen vacancies (V_O) and Ti³⁺ ions are the most common defects formed in TiO₂. Fig. 1 (d) shows the photoluminescence spectra revealed a broad luminescence centered around 420 nm for the pristine thin film. The irradiated thin films showed an additional peak around 446 nm, and its intensity increased with the absorbed dose, indicating the presence of V_O and Ti³⁺ defects as trapping centers for photo-generated electrons. These traps reduce the effective charge carrier concentration and gamma-ray irradiation resulted to a decrease in the photo current.

Over a period of 720 hours, the irradiated thin films exhibited a recovery in transparency, and their photoconductivity improved. Fig. 1 (a) shows the photo current increased by an order of magnitude compared to the pristine detector, and the photosensitivity increased by a factor of about 16 and 124 post-recovery compared to its sensitivity before and immediately after irradiation, respectively.

3.Conclusion

In this study, TiO₂ thin films were investigated as photoconductive detectors for VUV radiation. The films were subjected to gamma-ray irradiation, resulting in the formation of defects, particularly V_O and Ti³⁺ defects, which were observed. These defects caused changes in the absorption spectrum and a narrowing of the band gap in the thin films. As a result of the gamma-ray irradiation, the photoconductivity of the thin films decreased due to the trapping of electrons by the V_O and Ti³⁺ defects.

View PDF for the rest of the abstract.

In thin film deposition by sputtering, it takes a lot of time to optimize conditions for obtaining desired film thickness distribution and film properties due to many condition parameters. Experimental time costs can be reduced by predicting film thickness distributions and film properties of deposited thin films. Film thickness distribution can be predicted by transport analysis simulation of sputtered particles, but there is no method for predicting film properties. On the other hand, there are methods for predicting film structures that are closely related to film properties, using the structural zone model (SZM). Using the SZM proposed by Thornton[1], the film structure of the deposited thin film can be predicted from the main parameters, sputtering pressure and substrate temperature. However, even if the film is formed under the same Ar gas pressure, the structure of the obtained film will be different if the apparatus is different. An improved Structural Zone Diagram (SZD) was proposed by Anders[2]. The axes of this SZD are the normalized incident energy of the sputtered particles and the substrate temperature. Therefore, even if a different apparatus is used, so far as the incident energy of the sputtered particles to the substrate is the same, the structure of the obtained film can be regarded as the same. However, Anders' SZD was proposed assuming a sputtering method with a high ionization rate, and it is not clear whether it can be applied to the conventional magnetron sputtering method.

In the previous study, a sputtered particle transport simulation program considering the thermal motion effect of the process gas[3] was created, and the film thickness distribution on the substrate was calculated and compared with the experimental results. Also, from the deposition conditions and the simulation results, the position on the Anders' SZD corresponding to the deposition conditions was obtained, and the conductivity distribution of the thin film was shown on the diagram[4]. As a result, it was found that the SZD is mostly applicable to conventional magnetron sputtering method, but there were partial differences in conductivity distribution due to differences in target-substrate distance conditions.

In this study, we report the validation results of this SZD for titanium films deposited by magnetron sputtering method. Using two general types of magnetron sputtering equipment, film samples were prepared with several combinations of sputtering pressure (0.3-4.0 Pa), target-substrate distance (50-100 mm), and substrate temperature (RT-450^oC). A DC power supply was used for sputtering, and the power was constant at 300 W. In addition, we observed the samples by Scanning Electron Microscope (SEM). It was confirmed whether the characteristics of the film structure indicated by the points on the SZD corresponding to the experimental conditions were consistent with the observed results of the samples prepared by either equipment. As a result, under the conditions corresponding to the film structures of ZONE 1 and ZONE 2 on the SZD, the SEM images of the film had the characteristics of each ZONE. On the other hand, under conditions corresponding to the film structure of the ZONE T, which is the transition region, it was found that the characteristics of the film structure did not show the characteristics of ZONE T so much, but strongly show the characteristics of the adjacent ZONE, especially in the area near the boundaries of the zones.

A part of this work was supported by Kyoto University Nano Technology Hub in "Nanotechnology Platform Project" and "Advanced Research Infrastructure for Materials and Nanotechnology Project" sponsored by MEXT, Japan.

References

[1] J. A. Thornton, Ann. Rev. Mater. Sci., 7, 239 (1977).

View PDF for the rest of the abstract.

[Introduction] Rare earth hydrides have been extensively studied due to their potential applications for electronic and energy devices [1,2]. Thus, thin film growth methods are being developed [3]. Recently, we synthesized LaH₂(111) epitaxial thin films by reactive magnetron sputtering for the first time [4]. In contrast with a metallic conduction in LaH₂ bulk polycrystal, the thinner LaH₂(111) film showed semiconducting conduction [5]. This different conduction behavior could be affected by the film orientation. In this study, we grew LaH_x(001) epitaxial thin films with different thickness, and evaluated the electrical properties in order to clarify the relation between electrical conduction and the film orientation.

[Experiment] LaH_x epitaxial thin films (31−123 nm) were grown on CaF₂(001) substrate by reactive rf magnetron sputtering of La metal target in Ar/H₂ mixed atmosphere (H₂ concentration: 3.5%). The range of growth temperature (T_g) was 250–500 °C. After the film growth, the temperature was quenched to room temperature followed by the deposition of Si₃N₄ cap layer (~ 5 nm). The crystal structure was evaluated by X-ray diffraction method. The electrical properties were measured by van der Pauw method.

[Results and discussion] Above T_g =350 °C, impurity phases such as LaF₃ were formed in the films in addition to LaH₂ phase. However, LaH_x phase was obtained in the films without such impurity phases with T_g = 300 °C (Fig. 1(a)). For the 31 nm-thick film, single phase LaH₂(001) epitaxial thin film was obtained. For the larger thicknesses, LaH₃(001) epitaxial thin film was included in the LaH₂(001) epitaxial thin film. These results suggest that hydrogen was desorbed near the surface of the LaH₃(001) film, in contrast with the presence of only LaH₂(111) epitaxial thin film without LaH₃ phase irrespective of the thickness [4], indicating easier hydrogen desorption for the (111)-orientation. Concerning the temperature dependence of electrical resistivity, the 31 nm-thick LaH₂(001) film showed semiconducting conduction similar to the 33 nm-thick LaH₂(111) film, whereas the thicker LaH_x (001) films showed much higher electrical resistivity (Fig. 1(b)), attributed to the inclusion of LaH₃ phase [4]. Taking these results into account, it is plausible that the hydration process in LaH_x depends significantly on the crystal plane. The difference in hydrogen desorption speed would result in the formation of only LaH₂ phase in the (111)-orientation while LaH₂ and LaH₃ mixed phase in the (001)-orientation. Accordingly, the film orientation is an important factor to control the amount of hydrogen in rare earth hydride thin films, leading to the control of electrical conduction.

[References] [1] J. N. Huiberts et al., Nature 380, 231–234 (1996). [2] K. Fukui et al., J. Am. Chem. Soc. 144, 1523–1527 (2022). [3] R. Shimizu et al., J. Phys. Soc. Jpn. 89, 051012 (2020). [4] S. Miyazaki et al., JSAP spring meeting, 15a-PB1-15 (2020). [5] S. Uramoto et al., JSAP spring meeting, 25a-F307-8 (2022).

In the previous study, it was confirmed that Anders' SZD is applicable to conventional magnetron sputtering for film structure[1]. Anders' structure zone diagram (SZD)[2] shows the film structure distribution on the normalized incident energy/normalized substrate temperature plane. Because the property of film was related to film structure, we assumed the film property distribution will be decided on this two-dimensional diagram.

We have already reported the conductivity distribution of Ti films on SZD[3]. However partial differences were observed in the conductivity distribution due to the difference in target-substrate distance conditions. Therefore, the ultimate pressure was lowered to improve the distribution. In addition, we found the reflectance distribution of Ti film and confirmed the relationship between these distributions and the film structure distribution on the Anders' SZD. Using two equipments, film samples were made in some combination of sputtering pressure (0.3-4.0 Pa), target-substrate distance (50-100 mm), and substrate temperature (RT-450^oC). A DC power supply was used for sputtering, and the power was constant at 300 W. In addition, we measured the surface conductivity and the reflectance of the samples. The obtained properties were plotted on a two-dimensional diagram with incident energy and substrate temperature. About the film property distribution, we confirmed no difference between equipments. As a result, about the distribution of the conductivity, it was confirmed that a contour line became parallel to ZONE border on SZD as shown in Fig. 1, and the film property that reflects the film structure is well explained by the SZD. On the other hand, about the distribution of the reflectance, the contour line did not become parallel to the ZONE border and was not able to confirm the relationship with the film structure. In addition, it was revealed that the high background pressure strongly influenced film properties distribution.

References

[1] K. Kuroshima et al., Annual Meeting of the Japan Society of Vacuum and Surface Science 2023, Nagoya, October 31-November 2, 2023 (to be presented).

[2] A. Anders, Thin Solid Films, 518, 4087 (2010).

[3] K. Kuroshima et al., The 22nd International Vacuum Congress, Sapporo, September 11-16, 2022, Tue-PO1D-13.

Amorphous carbon (a-C(:H)) thin films have been studied in a wide range of fields as protective films for automotive parts, hard masks for semiconductor device fabrication, and biocompatible films for medical devices due to their excellent characteristics. In particular, mechanical properties (film stress and fracture toughness) related to film delamination are important because they are related to the durability of the films, which in turn are related to film stress and adhesion strength. Recently, we have shown that the introduction of carbon nanoparticles (CNPs) between two layers of a-C:H thin films reduces film stress[1]. In this study, we evaluated other properties of the CNP-inserted sample and examined the effect of CNPs on the mechanical properties of the film toward the practical stage.

Sandwich structure films were fabricated using a capacitively coupled plasma-enhanced chemical vapor deposition (PECVD) system [1]. Ar and CH4 gases were introduced from the top at 19 sccm and 2.6 sccm, respectively. The thickness of the first and second layers was 154 nm. For the nanoindentation test, a nanoindentation tester (ENT-1100a) was employed and a Berkovich indenter was used.

The load-unloading curve by nano-indentation showed a typical curve at 5 mN, and a step in the curve occurred at over 8 mN, and SEM images of the indentation showed that the membrane peeled off in a circular shape when the step occurred. EDS analysis of the peel scar revealed that the peel occurred at the interface between the first and second layers. In addition, the fracture toughness of the film was determined from the SEM images of the delamination traces and the load-unloading curve at the time of step generation, and it decreased with increasing Cp in the region where the film stress was constant for the CNP coverage. These results suggest that CNP coverage has a negative correlation with fracture toughness and that there is an optimum value for improving mechanical properties. Other properties will be discussed in detail in the presentation.

[1] S.H. Hwang et al., Jpn. J. Appl. Phys. 59 100906, (2020).

[introduction] Water splitting is a promising way for mass production of hydrogen from renewable energy sources. To increase the efficiency of water splitting, it is important to reduce the overpotential of oxygen evolution reaction (OER). For this purpose, a use of OER catalysts with inexpensive and highly active materials is highly demanded. In this study, we focused on LiNi_0.5Mn_1.5O₄ (LNMO), which is composed of inexpensive transition metals. So far, the OER catalytic activity of LiNi_0.5Mn_1.5O₄ has been reported in powder[1]. However, there is no report on quantitative studies of catalytic activity using LNMO thin films with well-defined surface area. Here, we study the OER catalytic activity of LNMO thin films and report a recovery of catalytic activity after the application of low potential. [experiments] An Au(111) thin film (80 nm thick) was deposited as a current collector on an Al₂O₃(0001) substrate using a direct current magnetron sputtering at room temperature. A LiNi_0.5Mn_1.5O₄(111) thin film (80 nm thick) was deposited on the Au(111)/Al₂O₃(0001) substrate using radio-frequency magnetron sputtering (substrate temperature: 500°C, Ar partial pressure: 0.98 Pa, oxygen partial pressure: 0.02 Pa). X-ray diffraction and Raman spectroscopy were used for the structural characterization of the thin films. A three-electrode cell was prepared (0.1 M KOH, pH=13.0) with the LiNi_0.5Mn_1.5O₄ (111) film, Ag/AgCl, and Pt wire as the working, reference, and counter electrodes, respectively. The OER activity was evaluated by cyclic voltammetry (CV) as follows. First, a potential was swept eight times between +0.8 and +2.0 V (vs. reversible hydrogen potential (RHE)), followed by a low potential sweep in the range of -0.8 to +0.1 V (vs. RHE). The potential was then swept again between +0.8 and +2.0 V. [results] Figure 1 shows the CV curve observed in the OER. The current density at +2.0 V (vs. RHE) was 15.7 mAcm^-2 in the first cycle. Note that the current density of 0.07 mAcm^-2 at 1.70 V (vs. RHE) is 2.5 times larger than the previous report in bulk[1]. Also, we confirmed the activity decrease from 15.7 mAcm^-2(1st) to 6.70 mAcm^-2 (8th) (Fig. 1(a) blue line). When the OER was performed again after the application of low potential (-0.8 V vs RHE), surprisingly, we found that the current density was recovered to 16.5 mAcm^-2. (Fig. 1(b), red line). In this talk, we will discuss the mechanism of this activity recovery based on the valence states of transition metals characterized by X-ray photoelectron spectroscopy. References [1] Y. Ren et al., ACS Appl. Energy Mater. 2021, 4, 10731-10738. [2] L. Wang et al., Solid State Ionics 2011, 193, 32-38.

Alkaline-earth doped lanthanum trifluoride (La_1-xBa_xF_3-x, La_1-xSr_xF_3-x) has attracted much attention as a solid electrolyte in all-solid-state fluoride ion batteries [1]. Epitaxial thin films provide an ideal platform to quantitatively evaluate F^- conductivity due to well-defined size, surface area, and crystal orientation. We have previously reported on epitaxial thin film growth of La_1-xBa_xF_3-x and evaluation of their ionic conductivity [2]. However, since Ba is located next to La in the periodic table, it is difficult to analyze the composition of the thin films, hindering the exact amount of substitution. In this study, we focused on Sr instead of Ba, and reported on the fabrication of La_1-xSr_xF_3-x (x=0.05) epitaxial thin films by reactive magnetron sputtering using a custom-made target, and the characterization of their ionic conductivity.

LaF₃ and SrF₂ powders were mixed at 95:5 and milled using a planetary ball mill (800 rpm, 50 sets of 6 min (operation) and 2 min (break)). Subsequently, the powder was pressed into pellets (φ 20 mm × 3 mm), followed by sintering at 700^oC for 5 hours in vacuum. Thin films were deposited using magnetron sputtering with the sintered target on CaF₂(111) substrates heated to 200^oC. To compensate F deficiencies, CF₄ gas is introduced with Ar gas during sputtering at a constant total pressure of 1.0 Pa. X-ray diffraction (XRD) was used for the structural characterization. F^- conduction properties were evaluated by alternating current (AC) impedance spectroscopy after the deposition of Au comb electrodes.

Figure (a) shows CF₄ partial pressure (P_CF4) dependence of XRD patterns. Tysonite-type (001)-oriented La_0.95Sr_0.05F_2.95 thin films were fabricated at P_CF4 ≧ 0.02 Pa. Pole figure measurements confirmed the in-plane orientation relationship of [110]_LaF3 // [11-2]_CaF2 (not shown here). Figure (b) shows the Nyquist plot, indicating semicircles originating from F^- conduction. The ionic conductivity of ~6.6×10^-3 S/cm at 150^oC was obtained, which is consistent with that in bulk 1.0×10^-2 S/cm at 150^oC [3]. Thus, we successfully fabricated La_0.95Sr_0.05F_2.95 epitaxial thin films with F^- conduction.

References [1]: H. Bhatia et al., ACS Appl. Mater. Interfaces 9 23707 (2017), [2]: K. Fukatsu et al., The 90th ECSJ Annual Meeting, 3I09 (2023), [3]: N. I. Sorokin et al., Phys. Solid State 40 604 (1998).

[Introduction] Electrides are ionic compounds in which electrons are occupied in the anionic sites instead of the usual anions. Among them, Ca₂N electride has two-dimensional layered structures with alternating layers of Ca₂N⁺ and electrons (e^-)[1]. Due to the two-dimensional distribution of electrons, high in-plane electron mobility (200 cm²V^-1s^-1 @ room temperature) has been reported in a bulk single crystal. Although device applications of such Ca₂N electride thin films are expected, so far, there has been no report of epitaxial thin-film fabrication of Ca₂N. In this study, we report on the fabrication of Ca₂N electride thin films using reactive magnetron sputtering.

[Experimental] Ca₂N thin films (thickness: ~300 nm) were deposited on yttria-stabilized zirconia (YSZ)(111) substrates using reactive magnetron sputtering. A Ca metal target was used for the deposition. A total pressure is set to 1.0 Pa. In addition to Ar gas, N₂ gas was introduced for nitrogenation, and the partial pressure of N₂ (P_N2) is varied from 0.05 to 0.25 Pa. The substrate temperature (T_s) was 400 - 600°C. X-ray diffraction (XRD) and Raman spectroscopy were used for the structural characterization of the thin films. Since Ca₂N is unstable in air, all the characterization processes were performed using the Ar-filled cell sealed in a glove box. Figure (a) Substrate temperature dependence of X-ray diffraction patterns. (b) Raman spectra (red: our Ca₂N thin film, black: Ref [3] (bulk))

[Results and Discussion] Figure (a) shows the T_s dependence of XRD patterns. The diffraction peaks originating from Ca₂N 003, 006, and 009 reflections were observed, indicating 00n orientation at T_s ≥ 400°C. Since the peak intensity decreased with increasing T_s, we set the optimum T_s as 400°C. Next, we studied P_N2 dependence during the sputtering in the range of P_N2 = 0.05 - 0.25 Pa. As a result, the peak intensities in the XRD pattern and Raman spectrum (Fig. (b)) were highest at P_N2 = 0.13 Pa. We also found that the c-axis length of the thin film is 19.1 Å, ~1.6% longer than that reported for the bulk (18.8 Å). This suggests the possibility of non-stoichiometry or stacking faults in the thin films. Further optimization of synthesis conditions will be planned for the evaluation of physical properties in the future.

[1] Lee et al., Nature, 2013, 494, 336

[2] E. T. Keve et al., Inorg. Chem., 1968, 7, 1757

[3] M. Kitano et al., Chem. Sci., 2016, 7, 4036

Introduction

The era of "autonomous" experiments using machine learning and robotics has arrived [1,2]. To explore novel solid materials, we are constructing an automated and autonomous experimental system involving thin film synthesis and various characterization instruments such as X-ray diffraction (XRD), Raman spectroscopy, UV-visible spectroscopy, and scanning electron microscopy/energy dispersive X-ray spectroscopy (Fig. (a)). To realize the autonomous experimental cycle, it is necessary to automatically analyze the results successively acquired from each instrument. For XRD measurements, the automation of analyses has been reported in powders [3]. However, there is no report on the automatic analysis methods for thin films due to the limitation of a few diffraction peaks. Therefore, we are studying automatic XRD analysis methods for thin films. Here, using a LiCoO₂(001) epitaxial thin film as a model case, we report the automatic processes of peak detection in XRD patterns and successful peak extraction originating from LiCoO₂(001).

Methods

LiCoO₂(001) epitaxial thin films (40 nm thick) were fabricated on Al2O₃(0001) substrates using radio-frequency (RF) magnetron sputtering (substrate temperature: 700 °C, Ar pressure: 0.9 Pa, O₂ pressure: 0.1 Pa, RF power: 80 W). Out-of-plane XRD patterns were obtained on the LiCoO₂(001) thin film samples. BEADS (baseline estimation and denoising using sparsity) [4] and SSFP (scipy.signal.find_peaks) packages were used for the estimation of background and noise, and the diffraction peak detection, respectively. Indexing for the detected peaks was based on the value closest to 2θ in the reference powder X-ray diffraction.

Results and Discussion

Figures (b)-(i) and (b)-(ii) show the experimentally obtained out-of-plane XRD patterns (LiCoO₂(001) thin film on Al₂O₃(0001) substrate and only Al₂O₃(0001) substrate) and the estimated pattern after the background subtraction and denoising by BEADS, respectively. Subsequently, the peak detection was performed using SSFP on the estimated patterns. Compared with the peaks detected from the pattern on the Al₂O₃(0001) substrate (Fig. (b)-(ii), blue ×), we succeeded in the extraction of the peak components originating from the thin films (Fig. (b)-(ii), red ♦). Each diffraction peak from the thin film is assigned to 003, 006, 009, 0012, and 0015 reflection of LiCoO₂, which is consistent with the previous report [5]. Note that the detected positions slightly shift to lower angles than the reference values in powder, suggesting the possibility of the incorporation of Co₃O₄ impurities. To further improve the crystal quality in thin films, we next plan to combine Raman spectroscopy in our automated system.

References

[1] R. Shimizu et al., APL Mater. 8, 111110 (2020).

[2] N. Ishizuki et al., STAM-M. 3, 2197519 (2023).

[3] Y. Ozaki, et al., Npj Comput. Mater. 6, 75 (2020).

[4] X. Ning, et al., Chemometr Intell Lab Syst. 139, 156 (2014).

[5] T. Tsuruhama et al., Appl. Phys. Express 2, 085502 (2009).

Catalytic chemical vapor deposition (Cat-CVD) is a film deposition method in which gas molecules are decomposed through a catalytic reaction on the surface of a heated catalyzing wire, and can realize film formation without plasma damage on substrates. This feature is suitable particularly for the deposition of thin films on crystalline silicon for application to solar cells which needs good surface passivation to minimize the recombination of photo-generated minority carriers. In this talk, I review the fundamentals of Cat-CVD and its photovoltaic application. Cat-CVD can produce amorphous silicon (a-Si) films with good surface passivation ability on the surface of crystalline silicon wafers, and Cat-CVD a-Si films are thus widely used in silicon heterojunction (SHJ) solar cells. Cat-CVD silicon nitride (SiN_x) can be used as anti-reflection coating and passivation films for crystalline silicon solar cells. It has been recently demonstrated that ultra-thin Cat-CVD SiN_x films are available for the tunneling layer of a carrier selective contact. In addition, by taking advantage of the ability of Cat-CVD to deposit high-density thin films even at low temperatures, CVD SiN_x films are also used as gas barrier films protecting materials and devices with low temperature and humidity tolerance such as organic light-emitting diodes (OLEDs) and perovskite solar cells.

Introduction Thermoelectric (TE) power generation, which converts heat to electricity, is one of the sustainable energy suppliers. Specifically, TE film on Si substrate is drawing much attention as a power source for internet of things sensor. The TE performance is quantified by a dimensionless figure-of-merit ZT; ZT=S²σTκ^-1, where S is Seebeck coefficient, σ is electrical conductivity, T is absolute temperature, and κ is thermal conductivity. The crucial bottleneck for high ZT is the interdependent relationship among three TE physical parameters [1, 2]. Heavy element-based materials likely exhibit high ZT due to inherently-low κ. In recent years, GeTe is intensively studied because of low phonon group velocity and large band degeneracy, which lead to inherently-low κ and -high S²σ, respectively. However, the carrier concentration p is too high because of the generation of defect (Ge vacancy) working as carrier supplier. Although the amount of Ge vacancy is decreased by introducing other elements such as In, Sb, etc. for suppressing its generation [3, 4], such introduced elements can also work as carrier scattering centers, leading to mobility μ decrease. In this study, we demonstrate the p decrease in epitaxial GeTe film without introducing other elements. Therein, p decrease is brought by utilizing high volatility of Te atoms during the film growth in vacuum. This outstanding technique largely enhances TE properties.

Experimental Epitaxial GeTe films were grown using pulsed laser deposition. The GeTe target was fabricated by spark plasma sintering under 93 MPa at 723 K for 5 min. Chemically-cleaned Si(111) substrates were introduced into vacuum chamber. By irradiating the target with ArF laser, GeTe films were grown on Si(111) substrates at the growth temperature T_S=473-593 K in vacuum. Structural analyses were performed using transmission electron microscopy (TEM) and x-ray diffraction (XRD). The p, σ, and S were obtained by Hall effect measurement, van der Pauw method, ZEM-3 (Advance Riko Inc.), respectively. κ was measured by time domain thermoreflectance method.

Results and discussions XRD measurement revealed that T_S<523 K is a key to growing epitaxial GeTe films/Si. Furthermore, from TEM observation, it was found that the epitaxial GeTe films had nanoscale domain structures, the size of which was ~60 nm. These films had relatively low p (~3×10²⁰ cm^-3) compared with GeTe bulk reported by the preceding studies [5]. Thanks to p decrease, the films exhibited high μ and S, resulting in high S²σ. The κ of these films was measured to be ~2 Wm^-1K^-1, which was extremely lower than those of GeTe bulk. In this talk, we will discuss the detail of film growth and the TE properties.

References [1] Y. Nakamura, et al., Nano Energy 12, 845-851(2015). [2] T. Ishibe, et al., ACS Appl. Mater. Interfaces 10, 37709-37716 (2018). [3] J. Dong, et al., Energy Environ. Sci. 12, 1396-1403 (2019). [4] S. Imam, et al., Mater. Today Phys. 22, 100571 (2022). [5] T. Ishibe, et al., ACS Appl. Mater. Interfaces 15, 26104-26110 (2023).

Introduction Organic ionic plastic crystals (OIPCs) have attracted much attention for their possible applications as solid electrolytes and for other specific purposes owing to their negligible vapor pressure, nonflammability, and thermal and electrochemical stability. However, there have been quite few reports on the fabrication of OIPC thin films.^[1] We have proposed an ionic liquid (IL)-assisted vacuum deposition method, in which ILs are used as nonvolatile solvents in a vacuum for thin film growth.^[2],[3] This deposition technique is effective for controlling crystal polymorphism and growth form as demonstrated in the growth of organic semiconductor crystals,^[2] as well as in the epitaxial growth of KBr(111) on α-Al₂O₃(0001).^[3] In this study, we applied this method to a typical OIPC, 1-ethyl-1-methylpyrrolidinium bis(trifluoromethanesulfonyl)amide ([C₂mpyr][TFSA]), and investigated the solvent and growth temperature dependences on its growth orientation and morphology.

Experimental An α-Al₂O₃(0001) substrate was covered with IL 1-ethyl-1-methylimidazolium bis(trifluoromethanesulfonyl)amide ([C₂mim][TFSA]) and annealed in a vacuum at 300 °C for 20 minutes to evaporate the IL, whereby the wettability of subsequently deposited liquid on the substrate was greatly enhanced.^[3] On the treated substrate, a 100 nm-thick film of polar [C₂mim][TFSA] or non-polar bis(2-ethylhexyl) sebacate (B2EHS) was deposited at room temperature, and subsequently [C₂mpyr][TFSA] whose amount was 400 nm in thickness was deposited, by vacuum deposition using an infrared laser deposition technique.^[2],[3]

Results and Discussion Fig (a) shows the XRD patterns and optical microscope images of [C₂mpyr][TFSA] films grown with and without IL at different temperatures of 30 °C and -70 °C. The film grown without the IL at 30 °C showed a polycrystalline structure with droplet-like morphology, while a (010)-oriented uniform thin film was obtained for the deposition temperature of -70 °C. On the other hand, the films grown with the IL showed a (100) preferential orientation with an island-like morphology, a growth habit, which did not change regardless of the deposition temperature. Furthermore, a polycrystalline film was grown when using non-polar B2EHS instead of the IL, as in the case grown without the IL at 30 °C. These results suggest that the (100) plane of [C₂mpyr][TFSA] is polar and that the solvation effect by the polar nature of IL during the crystal growth process is responsible for the (100) orientation of [C₂mpyr][TFSA]. Fig (b) shows a series of in situ optical microscope images taken during the deposition process of [C₂mpyr][TFSA] onto [C₂mim][TFSA] at a substrate temperature of -70 °C and its subsequent process of increasing the sample temperature after the deposition. There were no significant morphological changes while increasing the deposition amount, but a morphological change, which seemingly results from the precipitation of crystallites, started at a sample temperature of around -10 °C during the temperature increase. Fig (c) shows the XRD intensities of the 200 and 020 peaks of [C₂mpyr][TFSA] in the films grown with the IL as a function of the growth temperature. The intensity of the 200 diffraction is generally stronger than that of the 020 diffraction, except for the growth at -100 °C, and increased with decreasing the growth temperature.

View PDF for the rest of the abstract.

[INTRODUCTION] Recent studies have highlighted the potential to improve the battery performance of LiCoO₂ (LCO), a layered rock-salt structured cathode material for lithium-ion secondary batteries, through partial substitution of Co³⁺ with heavy metal elements [1]. Also, in the field of lithium secondary batteries, a concerted effort has emerged to further enhance such a substitutional approach by developing compositionally-graded materials, wherein a continuous transition of the composition of a substituted metal element is made from the material's interior to its topmost surface [2]. However, prior investigations have primarily centered around bulk polycrystalline materials, thereby necessarily including considerations of grain boundaries and polymorphism. In contrast, epitaxial thin films offer a more ideal platform for unraveling the intrinsic correlation with battery performance. In this study, our focus centers on the substitution of La as a heavy metal element in LCO. We investigate the possible effects of the La substitution and its gradient structures on the battery properties of LCO by preparing both homogeneous and compositionally-graded La-doped LCO (La:LCO) epitaxial thin films.

[EXPERIMENTAL] Thin films were fabricated using a pulsed laser deposition system with rapid beam deflection (RBD-PLD) [3]. Three types of targets were used: Li_1.4CoO_x (La0%), Li_1.4La_0.05Co_0.95O_x (La5%), and Li_1.4La_0.10Co_0.90O_x (La10%). A bottom electrode layer of 50 nm-thick SrRuO₃ (SRO) was first deposited on a SrTiO₃ (STO)(100) substrate. Four 130 nm-thick, homogeneous composition films with different La compositions were prepared. The La compositions were determined as La0%, La4.4%, La4.7%, and La9.4% in terms of La/(La+Co) mol%, verified using inductively coupled plasma mass spectrometry (ICP-MS). For compositionally-graded films, UP-graded (with La10% at the top surface) and DOWN-graded (with La0% at the top surface) thin films were fabricated, featuring a concentration gradient in the film thickness direction spanning from La0% to La10%. This was achieved through high-speed alternative ablation of two types of targets, La0% and La10%:LCO. Secondary ion mass spectrometry (SIMS) depth profiling confirmed the presence of a linear La concentration gradient. The epitaxial growth of these fabricated samples was confirmed through X-ray diffraction (XRD). The battery performance evaluation involved using a CR2032 coin cell with 1M LiPF₆ (EC:DEC=3:7) as the electrolyte and Li metal as the anode, with charge-discharge measurements conducted at different charge rates ranging from 1 C to 20 C.

[RESULTS] Cycle characteristic evaluation of the homogeneous films (depicted in Fig. (a)) demonstrated that the La4.7% film, predominantly substituted with Co sites, exhibited improvements in rate performance in comparison to the La0% film. However, the La4.4% and La9.4% films, predominantly substituted with Li sites, showed a decline in performance. Considering the results of the density of states calculations utilizing the DFT+U method (shown in Fig. (b)), it is postulated that the Co site substitution facilitated Li ion deinsertion due to the enhanced electrical conductivity from a narrower bandgap, thereby resulting in the enhanced rate performance. On the other hand, the cycle characteristics evaluation for the compositionally-graded films (also shown in Fig. (a)) unveiled the improved rate performance in the UP-graded film with La10% at the top surface, relative to the homogeneous La0%. Conversely, the DOWN-graded film with La0% at the top surface exhibited degraded rate performance.

View PDF for the rest of the abstract.

Hexagonal boron nitride (hBN) is a similar property of graphene, such as strong covalent bonds of B and N atoms reserved for C atoms. Recently, hBN has been the most popular material due to its flexible physical and chemical properties and is used in a variety of applications, especially in electronics and biomedical fields. Still, a lack of study on hBN thin film and improving properties of hBN based heterostructures fabrication are essential for future electronic devices. In this work, the synthesis and functionalize hBN sheets are transferred on both SiO₂/Si substrate and reduced graphene oxide (RGO) thin film. Accordingly, hBN and RGO/hBN thin films are studied by X-ray diffraction (XRD), high-resolution transmission electron microscopy (HR-TEM), Raman spectroscopy, and atomic force microscopy (AFM). The h-BN sheets and h-BN/RGO heterostructures thin films are uniform with a large area on the SiO₂/Si substrate, which is observed by an optical microscope. And a single microscale hBN sheet (> 2 nm thickness) and hBN/RGO heterostructures were observed by AFM. Interestingly, the Kelvin probe force microscopy (KPFM) showed a higher surface potential of the hBN/RGO thin film than that of the h-BN thin film. These interesting surface potential values of single and heterostructure thin films are promising 2-dimensional (2D) materials for nanoelectronics applications.

Titanium dioxide (TiO₂) thin films have a wide range of applications, including photocatalytic degradation of pollutants, self-cleaning surfaces, and anti-reflective coatings. RF magnetron sputtering is a widely used method for preparing TiO₂ thin films due to its high deposition rates, good uniformity, and low substrate temperatures. However, the reproducibility and stability of TiO₂ thin films prepared by RF magnetron sputtering can be affected by a number of factors, such as the deposition parameters and the substrate material [1].

In this study, we investigated the reproducibility and stability of TiO₂ thin films prepared by RF magnetron sputtering on pre-treated quartz substrates. The quartz substrates were annealed in oxygen atmosphere before the deposition of TiO₂ films. Photoluminescence (PL) measurement was carried out to study the presence of defect states in the films and to confirm the reproducibility of the films. Photocatalytic degradation of the methylene blue (MB) was carried out to check the photocatalytic activity of the prepared films.

We found that the reproducibility of TiO₂ thin films prepared on pre-treated quartz substrate was significantly improved compared to films prepared without pre-treatment. The PL spectra of the films showed in fig. 1 (A), (B) and (C) indicate the temporal change in the films over a period of three weeks. No change in defect states was observed during this time which indicates good stability of the film. Fig. 1 (d) shows the calculated rate constant for the films A, B and C .The surface defects were responsible for not only enhancing the photocatalytic activity but also the good stability of the films. We also found that the pre-treatment of the quartz substrates can affect the uniformity of the surface state condition of the substrate, in turn giving reproducibility.

Reference:

[1] J. Daughtry et al., Nanoscale Adv. 3, 1077 (2021)

For crystalline Si solar cells theoretical limit [1], heterojunction, tunnel oxide passivated contact, and carrier selective contact (CSC) structure have been proposed. Especially in the CSC structure, the CSC layer is the key to conversion efficiency. Various CSC layer materials such as TiOx [2, 3] and MoOx [4, 5] have already been developed for CSC solar cells. We have been trying to investigate a new CSC layer for further improvement of efficiency. As a novel CSC layer, we focused on transition metal dichalcogenides (TMD), which are two-dimensional layered materials. TMD is a generic name for materials combining metallic elements of 4-6 group (Mo, W, Zr, etc.) and chalcogens of 16 group (S, Se, Te, etc.). TMDs have covalent and ionic bonds in the in-plane and van der Waals force bonds in the out-of-plane (between layers). In an ideal TMD, there are no dangling bonds in the out-of-plane direction [6], and thus no interface levels are formed. Furthermore, depending on the combination of metallic elements and chalcogens, it has a wide variety of physical properties such as band gap, work function, and electrical conductivity [7-9].

For applications in the CSC layer, we need to understand the band structure. In the band structure evaluation technique, we used hard X-ray photoelectron spectroscopy (HAXPES). HAXPES provides several times the detection depth by using hard X-rays as shown in Fig. 1 (a) [10-12], combined with angle resolution, enables the evaluation of band structure. For application to the CSC layer, we evaluated the band bending at the interface with the n-type c-Si using HAXPES. We deposited 7 nm MoSx film on an n-Si with mirror surface by RF sputtering. MoS₂ is a representative TMD material that has already been successfully used in transistor applications [13]. In this experiment, we used HAXPES equipment at BL46XU at SPring-8 in Japan. The basic synchrotron radiation source energy of HAXPES was set to be 7,939 eV. And we evaluated the TOA dependence of the Si 1s spectra. The Fig 1 (b) shows the Si-Si bond in Si 1s spectra obtained with various TOA. With decrease in TOA, Si-Si bonding peak shifts to 0.08 eV higher binding energy side. Since the binding energy in photoelectron spectroscopy corresponds to the energy difference between the Fermi level and the core level, the shift of the core level due to a change in TOA mainly represents band bending. Thus, the shift of the Si-Si bonding toward the higher binding energy side suggests a downward band bending. Figure 1 (c) shows the band diagram at the MoS_x/n-Si interface calculated from the Si-Si bonds obtained. From these results, we believe that MoSx has a smaller conduction band barrier to n-Si and is more likely to work as an electron transport layer.

Using HAXPES, we succeeded in nondestructively clarifying the band bending at MoSx/n-Si interface. This result suggests that the MoSx layer may work as a CSC layer, especially as an electron selective layer.

Part of this study was supported by NEDO. The authors thank Prof. Yoshio Ohshita and Mr. Tomohiko Hara for his cooperation in sample fabrication.

Reference

[1] A. Richter et al., IEEE J. Photovol. 3, 1184 (2013). [2] T. Matsui et al., ACS Appl. Mater. Interface 12, 49777 (2020). [3] T. Matsui et al., Sol. Energy Mater. Sol. Cells 209, 110461 (2020). [4] T. Kamioka et al., Jpn. J. Appl. Phys. 57, 076401 (2018). [5] C. Battaglia et al., Nano Lett. 14, 967 (2014). [6] Q.H. Wang et al., Nat. Nanotechnol. 7, 699 (2012). [7] X. Duan et al., Chem. Soc. Rev. 44, 8859 (2015). [8] Y. Hibino et al., MRS. Adv. 5, 1635 (2020).

View PDF for the rest of the abstract.

Introduction

Zirconium oxide (ZrO₂) has attracted attention as a catalyst material, which exhibits activity in various reactions because of its excellent chemical properties. On the other hand, it has a large band gap of about 5 eV, making it difficult to utilize sunlight as a photocatalyst. Defect introduction into semiconductor oxide is a useful technique to narrow its inherent band gap which leads to a broadening of the light absorption for photocatalytic applications.^{[1, 2]} Recently, this technique has been used to endow large band gap metal oxides, which have been difficult to be used as photocatalysts, with photocatalytic activity. Reduced ZrO_2−x, in which oxygen vacancies (V_O) or Zr³⁺ are formed, is one of the promising materials which exhibit photocatalytic activity by defect formation.^[3]

Crystal phases of ZrO₂ can be changed by different temperatures and pressures. Among them, Monoclinic and Tetragonal can exist stably under mild conditions, which makes them suitable for catalytic reaction. However, the effect of crystal phases of ZrO₂ on the photocatalytic activity has rarely been studied.

In this study, we performed hydrogen reduction treatments on two types of ZrO₂, Monoclinic and Tetragonal, to investigate the effect of crystal phases and defect formation on the photocatalytic activity.

Experimental

ZrO₂ with monoclinic (m-) or tetragonal (t-) crystal phase was synthesized in a previously reported method.^[4] Then, Pt was deposited on ZrO₂ at 1 wt%. The Pt/ZrO₂ was subsequently reduced at 200, 400, or 600 °C under H₂ flow to obtain Pt/ZrO_2−x. The prepared Pt/ZrO_2−x photocatalysts were evaluated by X-ray diffraction (XRD), electron spin resonance (ESR), and so on. The photocatalytic activity in H₂ production was investigated with aqueous methanol solution under UV light (λ = 365 nm) irradiation.

Results and Discussion

Fig. 1 (a, b) shows the XRD patterns of the Pt/ZrO_2−x after reduction under various conditions. No phase transition was observed for Pt/m-ZrO_2−x. For Pt/t-ZrO_2−x, the diffraction pattern did not change after reduction below 400 °C, but a peak attributed to Monoclinic appeared after 600 °C reduction.

Fig. 1 (c, d) shows the ESR spectra of Pt/ZrO_2−x after reduction under various conditions. In the Pt/m-ZrO_2−x spectrum, the signal attributed to Zr³⁺ was observed before reduction, and a new signal appeared at 200 °C reduction, which was attributed to oxygen vacancies. When the reduction temperature increased above 400 °C, the signal attributed to Zr³⁺ disappeared, while the intensity due to oxygen vacancies increased. On the other hand, in Pt/t-ZrO_2−x, oxygen vacancies were present even before reduction, and their amount increased significantly with increasing reduction temperature. In addition, the amount of oxygen vacancies was much larger in Pt/t-ZrO_2−x than in Pt/m-ZrO_2−x. Therefore, it is clear that, Zr³⁺ disappears and oxygen vacancies are formed by reduction treatment in Pt/m-ZrO_2−x, while the amount of oxygen vacancies increases by reduction treatment in the Pt/t-ZrO_2−x.

View PDF for the rest of the abstract.

The reaction of halogen atoms with metal surfaces is important in applications such as catalysis, dissolution, deposition, complex formation, and energy conversion [1-3]. Chloride ions on gold surfaces are a typical system for such applications, however, the elementary processes of the reaction are not fully understood. In this presentation, we report the results of our analysis of the reaction mechanism of chloride ions with gold surfaces in aqueous solution by scanning probe microscopy.

When the potential of the Au electrode is swept in the positive direction in an aqueous solution containing chloride ions, dissolution of Au is observed. This reaction is generally considered to proceed by the reaction of chloride ions with Au atoms on the surface. However, the atomic structure related to this reaction has not been clarified. In this study, we successfully observed atomic resolution image of the surface and found a structure different from that of chloride ions on the Au surface observed in ultrahigh vacuum. This structure is considered to be related to the progress of the dissolution reaction. In addition, a change in the mechanical properties of the surface at the dissolution zone was observed with the progress of the dissolution reaction. This change in physical properties is thought to trigger the dissolution reaction.

References

[1] Yuan Fang, Song-Yuan Ding, Meng Zhang, Stephan N. Steinmann, Ren Hu, Bing-Wei Mao, Juan M. Feliu, and Zhong-Qun Tian, J. Am. Chem. Soc. 142, 9439−9446 (2020).

[2] T. Tansel and O. M. Magnussen, Phys. Rev. Lett, 96, 026101 (2006).

[3] Hiroaki Konishi, Taketoshi Minato, Takeshi Abe, and Zempachi Ogumi, J. Phys. Chem. C 123, 10246−1025 (2019).

Photocatalytic dye degradation is an attractive method for “green” environmental remediation. The Cu₂O is a natural p-type semiconductor and the most promising photocathode for solar energy harvesting and conversion. However, the Cu₂O-based photocathodes still suffer severe self-photo-corrosion and fast surface electron-hole recombination issues. Earlier reports state that the carbon coating over the Cu₂O structure effectively suppresses the photo-corrosion and recombination [1,2]. Herein, a new hydrogen-substituted graphdiyne (HsGDY) is grown over the Cu₂O nanowire to protect it from photo corrosion. The HsGDY is grown by a Glaser coupling reaction for 1, 3, 6, 9 and 12 hours. In all the experiments 10 mM MB dye and 5 ml volume are fixed. The effect of dye degradation is monitored using a UV-visible spectrometer. Fig.1(a) shows the Raman spectrum of HsGDY@Cu₂O, which contains three peaks at 2215 cm^-1 1933 cm^-1 and 1578 cm^-1 corresponding to CºC bond, Cu-metallated CºC bond and C=C bond, respectively [3]. The combined HsGDY@Cu₂O heterostructure is used for photocatalytic degradation of methylene blue (MB) dye (Fig.1(b)) under 1 sun illumination. Over time, the intensity of MB dye at 664 nm decreases with continuous illumination. Around 60 % dye was degraded at 90 minutes of illumination.

References 1. Y. Li, et al. Catalysis Communications 66 (2015): 1-5. 2. W. Shi, et al. Applied Surface Science 358 (2015): 404-411. 3. X. Zhou, et al. Nature Communications 13.1 (2022): 5770.

Multifunctional materials have attained inevitable growth in recent years due to their unique properties and wide range of applications. Recently, emerging 2D layered materials possess excellent electrical, chemical, optical, and mechanical properties. Moreover, its tunable bandgap and ability to switch between metallic and semiconducting behavior allows for versatile electronic applications, including field-effect transistors, sensors, photodetectors, etc. Additionally, its unique surface and optical properties, like large surface area, strong light-matter interactions, and high light absorption, have led to its application in catalysis, photonics, and energy harvesting devices. This abstract aims to provide an overview of the key aspects and potential applications of MoS₂ as a 2D layered multifunctional material. We have investigated 2D layered materials for diverse applications and achieved state-of-art results by carefully tuning their properties as per the requirement. In our laboratory, thermoelectric energy conversion is the most explored application of MoS2. We have studied the thermoelectric properties of MoS₂ in all the available forms ranging from bulk to thin-film to flexible thick films to wearable fabrics, and have employed diverse strategies to improve its performance. We have formed MoS₂/MoO₃ hierarchical structures by hydrothermal method and studied the effect of the interface in its thermoelectric performance. With the simultaneous enhancement of electrical conductivity and Seebeck coefficient along with reduced thermal conductivity, we have achieved a record high zT of 1.18 at 600 K. This was attributed to enhanced phonon scattering and improved electrical conductivity by zero-barrier charge injection at MoS₂/MoO₃ interface [1]. Similarly, we have grown a few-layer MoS₂ on 290 nm-SiO₂/Si by a two-zone atmospheric pressure chemical vapor deposition technique and investigated its thermoelectric properties. Here, we observed the decoupling of electrical conductivity and Seebeck coefficient after 592 K resulting in an enhanced power factor of 116 nW/mK² at 734 K [2]. To overcome the challenges associated with using binders in the fabrication and enhance flexibility, we processed in-situ binder-free growth of MoS₂ nanosheets on conductive carbon fabric for wearable thermoelectric applications. We fabricated a wearable thermoelectric generator using MoS₂ and generated an output voltage of 1.2 mV at a temperature difference of 20 K. This paves a promising route for the future development of wearable thermoelectric. Further, we employed MoS₂ as the catalyst to degrade the environmental pollutants under light irradiation. MoS₂, when composited with ZnS, degraded the pollutants in 32 min (99.89%). This was attributed to the interfacial charge carriers being transferred between MoS₂ and TiO₂. We have constructed a MoS2-based gas sensor for the highly sensitive room temperature NO₂ sensing and have studied the effect of doping (Co, Ni) in its sensing performance [3]. Parallelly, we fabricated a solid-state asymmetric supercapacitor using Ni-doped MoS₂ for energy storage applications[4]. The DSSC device assembled with as-fabricated N-GQD@MoS₂@rGO possessed a superior power conversion efficiency of 4.65% due to the enhanced electrochemical active site and electrical conductivity property of rGO and MoS₂ [5].

Introduction

Vacuum ultraviolet (VUV) radiation spans the wavelength range from 100 to 200 nm. Its high photon energy renders VUV radiation indispensable in a wide range of applications, including semiconductor lithography, sterilization of surfaces, optical cleaning, surface modification, and spectroscopy, to name a few, so the development of VUV detectors is very important. In order to detect high-energy light, it is necessary to use materials with a large bandgap. Therefore, wide bandgap (WBG) semiconductors such as nitrides and oxides, as well as diamond, are being investigated. On the other hand, we focused on fluoride materials, which have higher durability against UV light irradiation than these, and carried out a similar approach using this material. Unlike silicon-based detectors developed by relatively mature production technologies and WBG semiconductor-based detectors, which are widely investigated in current literature, the development of fluoride photodetectors is at its infancy and thus not much is known regarding the correlation between fabrication parameters and photodetector performance. In this work, we investigated the influence of annealing on the interface between neodymium trifluoride (NdF₃) thin film and quartz (SiO₂) substrate. In semiconductor device fabrication, annealing is the process of heating the semiconductor to a predetermined temperature, keeping it at that temperature for a fixed time, and then cooling to room temperature. During annealing, the fabricated NdF₃ film on SiO₂ substrate is heated to a temperature of 200, 400, and 600 ^o C inside a furnace for 3 hours and then subsequently cooled to room temperature.

Experimental and Result

The effect of annealing on the interface between NdF₃ thin film and SiO₂ substrate and hence on the photoconductivity of NdF₃ thin film VUV photodetector, was investigated. Thin films were deposited using a pulsed laser deposition (PLD) under two substrate conditions: unheated substrate at R.T. and heated substrate at 600 ^o C. The deposited films were then annealed in vacuum and their structure and photoconductivity were compared to those of the unannealed (as-grown) films. The films deposited on the unheated substrate exhibited an increase in dark current resistance and hence a decrease in dark current after annealing. Consequently, our findings revealed a seven-fold increase in the SN ratio of the detectors following post deposition annealing. Resistance to dark current of the films deposited on the 600 ^o C-heated substrate decreased, resulting in similar values to those of the photocurrent at high bias voltages and suggesting poor photoconductivity. Investigation of the NdF₃/SiO₂ interface revealed that fluorine diffused from the film to the substrate during deposition. Diffusion can be attributed to the energy deposited by PLD ablation; however, diffusion is exacerbated when heating the substrate during deposition. Post deposition annealing did not seem to significantly contribute to fluorine diffusion. Diffusion of fluorine was found to degrade NdF₃ crystallinity near the interface, explaining the decrease in dark current resistance and increase in dark current. Deposition of NdF₃ on a MgF₂ substrate heated at 600 ^oC did not result in fluorine diffusion. Consequently, the SN ratio of the NdF₃/MgF₂ detector that was deposited on a 600 ^oC-heated substrate without post deposition annealing was similar to that of the NdF₃/SiO₂ detector that was deposited on an unheated substrate with post deposition annealing at 600 ^oC.

View PDF for the rest of the abstract.

Metal pattern formation by vacuum deposition is an important process from basic research to industrial mass production. Selective metal deposition utilizing desorption of metal atoms from organic surfaces is a promising method for forming metal patterns by maskless vacuum deposition. Maskless vapor deposition for metal pattern preparation is possible by changing the glass transition temperature (Tg) due to photoisomerization of photochromic diarylethene (DAE) [1,2]. The core phenomenon of selective metal deposition is metal atom desorption from the low Tg surface near room temperature. Metal atoms including Mg, Mn, and Pb desorb and the metal film is not formed on the colorless DAE surface. However, selective metal deposition of DAE could not be attained for such metal species having a high melting-point and low intrinsic-vapor-pressure as Ag, Cu, and Cr [3].

In this work, we demonstrate the formation of metal patterns by maskless deposition for various metal species using vacuum-depositable and printable perfluoropolyether (PFPE)-based materials [4]. A simple fluorinated alkyl chain film (a), which is silane-oupled to a glass substrate, exhibited low surface energy, but did not exhibit good desorption property of metal atoms. PFPE film (b), in which an ether moiety was incorporated in the alkyl chain, exhibits good metal atom desorption due to its low surface energy and surface molecular flexibility. It was found, furthermore, that the highest metal desorption property was attained for PFPE polymerized film (c) because of its large surface softness.

We adopted the antifouling/mold-release agent KY-1901 (Shin-Etsu Chemical Co., Ltd.) as a PFPE material, which can be used for both vacuum deposition and printing. Since this PFPE material was supplied as a solution, it was dropped into an evaporation boat and then annealed at 120°C for 1 hour to evaporate the solvent before being placed in a vacuum chamber as an evaporation source. In the PFPE film with a thickness of about 10 nm, a metal film was formed when the deposition amount of Ag and Cr was large. On the other hand, a 30-nm-thick PFPE film showed good desorption properties. PFPE-patterns were prepared on a glass substrate by vacuum deposition with a shadow mask and then metal evaporation was carried out without a shadow mask, resulting metal pattern formation for a variety of metal species including Ag, Cu, Cr, and Ni (d). This method provides complementary patterns for direct metal patterning using a shadow mask, including the pattern with isolated non-metallized areas. Even in the case of PFPE pattern formation by printing method, metal pattern formation was possible by maskless metal deposition. The resolution of the metal pattern obtained by this method depends on the fineness of the original PFPE pattern.

This method enables metal pattern formation by maskless vacuum deposition for various metals with high melting points and low intrinsic vapor pressures and can be applied to electrode pattern formation in electronics and other applications.

[1] T. Tsujioka, et al., J. Am. Chem. Soc., 130, 10740 (2008).

[2] T. Tsujioka, Chem. Rec., 16, 231 (2016).

[3] T. Tsujioka, S. Matsumoto, J. Mater. Chem. C, 6, 9786 (2018).

[4] T. Tsujioka, H. Kusaka, Adv. Mater. Interfaces, 9, 2201096 (2022).

The technological innovations are remarkable in artificial intelligence, internet of things, and automatic driving. These technologies are supported by semiconductor devices, that are required to be more highly integrated with a higher performance. Recent semiconductor devices have three-dimensional (3D) structures due to the scaling limits, and the ratio of vertical to horizontal sizes, or an aspect ratio, tends to increase. In order to manufacture those semiconductor devices for a practice use, cleaning processes are becoming much more important. While there are many reports on cleaning processes and their evaluations for a flat sample surface, research on 3D devices is limited.

Our final goal is to develop a non-destructive method to evaluate cleaning characteristics at the bottoms of 3D structures of semiconductor devices, which are probably the most difficult areas to be cleaned. As a test sample, in this study, we focused on 3D nanostructures on a Si surface. Our proposed method is as follows. (i) We fabricate 3D structures with different materials only at their bottoms. (ⅱ) In Angle-Resolved X-ray Photoelectron Spectroscopy (ARXPS) measurements, we control the take-off angles of photoelectrons strictly so that signal from the heterogeneous “landmark” element is detected. This guarantees the collection of photoelectrons not from the sidewalls but from the bottoms. (ⅲ) XPS spectra in (ii) also include signal from the top surface of the 3D structure, which needs to be removed. Thus, as a next step, we set a very shallow take-off angle to collect photoelectrons emitted only from the top surface. By subtracting this signal from that in (ii), we expect to extract only the “bottom condition”. The most important point in this scheme is to embed a heterogeneous element only at the bottoms of 3D nanostructures on a Si surface without introducing mechanical damages. In order to achieve this, we used Metal-Assisted Chemical Etching (MACE).

MACE is a solution process to fabricate various 3D nano- and micro-structures, including nanoporous layers, nanowires, 3D objects, micro-electromechanical systems, x-ray optics. [1]. In typical MACE of Si, a metal-loaded Si surface is immersed into a mixed solution of HF and H₂O₂ to cause local electrochemical reactions at the metal/Si interface. The metal serves as a catalyst for the reduction of H₂O₂, which injects holes into the Si. Because the hole concentration in Si becomes higher around the metal catalyst, a Si surface is readily oxidized underneath the metals. This is followed by the prompt dissolution of the SiO₂ layer in HF as a silicon fluoride compound, leaving nanostructures possessing a high aspect ratio with the metal catalyst at their bottoms. Figure 1(a) shows a schematic drawing for this etching mechanism. MACE is a low-cost and anisotropic etching method without leaving any crystallographic defects on the fabricated Si structures.

Figure 1(b) depicts the process flow of a trench formation. First, thin metal films composed of Au /Ti was deposited on a Si substrate by electron beam (EB) evaporation. Second, the photoresist was patterned using conventional photolithography. Au stripes were then formed by the use of an etchant and acetone to remove unnecessary Au and the resist, respectively. The Au/Ti/Si sample was immersed into a mixed solution of HF, H₂O₂, and H₂O. A resultant surface structure was imaged by Scanning Electron Microscope (SEM) as shown in Fig. 1 (c). It demonstrates the formation of Si nanotrenches with a width and a depth of around 200 nm and 1.2 µm, respectively. And their aspect ratio is approximately 6. It was also confirmed that Au catalysts reside at the bottoms of the trench structures in Figure 1(c) as we intended.

View PDF for the rest of the abstract.

The development of semiconductor devices, which opens technological innovation in mobile devices, automobiles, industrial equipment, etc. is accelerating year by year to achieve miniaturization with a higher integration. In this situation, a three-dimensional design, or a vertical integration, has emerged as a trend. For example, in Fin-Field-Effect Transistors (Fin-FETs), the aspect ratios (ARs) of the fin structure have reached as high as 10. Another example is a three-dimensional (3D) NAND flash memory of which ARs have exceeded 40 [1,2]. In the near future, we expect that the device structure will become much more complex and the ARs of 3D structures further increase. In order to fabricate those devices with a high yield, extremely strict cleanliness is required on the complexed 3D structures. In contrast to a large number of reports on cleaning and its evaluation on a flat surface, those on 3D structures are limited.

Based on this background, we aim at developing a novel non-destructive method to evaluate the cleaning performance of contaminants and oxides at the bottoms of 3D nanostructures, which are probably the most difficult areas to make clean.

We apply Angle-Resolved X-ray Photoelectron Spectroscopy (AR-XPS) for 3D structures with high ARs such as deep trenches and pores. In order to collect photoelectrons from their bottoms, we need to align the detector with the bottoms by adjusting the take-off angle of photoelectrons with a high accuracy. At this specific angle, we collect photoelectrons emitted from both the surface and the bottom. On the other hand, at a very shallow take-off angle, XPS spectra represent the condition of solely the surface. By subtracting the latter signal from the former one, we expect to extract information only from the bottoms. In this scheme, a key is to guarantee that we can detect signals from the bottoms, not from sidewalls, on the nanostructures. In order to achieve this, we propose to embed a heterogeneous “landmark” element at some bottoms.

To test the feasibility of the proposed method, we took XPS spectra of Au embedded at the bottoms of trenches with different ARs (1–6) on Si. Our concept is schematically depicted in Fig. 1(a). The sample was fabricated by metal-assisted chemical etching in solutions [3]. Figures 1(b) and 1(c) show the XPS spectra of Si2p and Au4f orbitals, respectively, at different take-off angles of photoelectrons from a Si surface consisting of trenches with an aspect ratio of approximately 6. Because Au resides only at the bottoms of the trenches, the Au intensity attenuates more significantly than that of Si as the take-off angle decreased. And it is barely detectable at angles below 60°. Figure 1(d) summarizes the area ratios of Au to Si signals as a function of take-off angles for samples with different ARs. It shows that the signal intensity of Au is strong only at an angle of 90°, and attenuates steeply at lower take-off angles. This trend is clearer for samples with higher ARs. These results demonstrate that the signal intensity from the “landmark” element, or Au in this experiment, serves as a guarantee to detect photoelectrons from the bottoms of 3D nanostructures. This will be more important when we analyze spectra on a complex structure with a higher AR. We expect that the combined method of AR-XPS measurements with our unique sample structure contributes to unveiling the cleaning performance at the bottoms of 3D structures.

References

[1] Yue-Gie Liaw et al, Solid-State Electronics 126(2016)46-50.

[2] Chong-Jhe Sun et al, IEEE journal of the Electron Devices Society 8(2020)1016-1020.

[3] Tomoki Higashi et al., Annual Meeting of JVSS 2023 [to be presented].

Flexible thermoelectric materials have been investigated for developing wearable power generators and self-powered wireless physiological sensors. One of ways to enhance the thermoelectric properties is the introduction of nanostructures. Therefore, we have investigated to form the oxide semiconductor nanomaterials on fabric substrates, as a flexible thermoelectric material, by solvothermal synthesis [1-7]. In the present study, with the aim of realizing oxide-nanostructure/fabric materials appropriate for wearable devices, we investigated the vertical growth of taller ZnO nanorods on conductive NiCu fabrics by microwave-assisted solvothermal synthesis.

ZnO nanostructures were formed on a NiCu fabric substrate using two-step microwave-assisted solvothermal synthesis. Based on solvothermal synthesis with microwave, some seeds of ZnO were fixed on the fabric in the first step, and ZnO nanorods grew up in the second step. In this study, we repeated the growth process (second step) for making the ZnO nanorod longer.

Figure shows scanning electron microscope (SEM) images of ZnO morphological structures grown on a NiCu fabric after (a) one-time and (b) four-times growth process, for microwave power of 100 W with an irradiation time of 10 min in the crystal-growth step. It is found that the fabric surface is fully covered with dense nanorods. In the case of conventional solvothermal synthesis with a furnace, it spent several hours for ZnO growth, and ZnO nanosheets and nanopillars were frequently observed [5,6]. Therefore, the microwave has effects of shortening the synthesis time and of forming ZnO nanorods. In addition, the nanorods seem decently vertical and become larger by repeating the growth process. Actually, it was confirmed that the diameter and height of nanorods evaluated from the SEM images increase linearly with the repeat count of the growth process, which indicates that the nanorod length can be controlled to obtain the desired length.

This work was financially supported in part by a Grant-in Aid for Challenging Exploratory Research (No.20K21886) from the Japan Society for the Promotion of Science and by the Cooperative Research Project on Research Centre for Biomedical Engineering.

[1] V. Pandiyarasan et al., Carbohydr. Polym. 157, pp. 1801-1808 (2017).

[2] V. Pandiyarasan, et al., J. Alloys Compounds 695, pp. 888-894 (2017).

[3] V. Pandiyarasan, et al., Mater. Lett. 188, pp. 123-126 (2017).

[4] V. Pandiyarasan, et al., Appl. Surf. Sci. 418, pp. 352-361 (2017).

[5] F. Khan, et al., IEICE Trans. Electron. E101-C, pp. 343-346 (2018).

[6] A.P. Kristy, et al., J. Mater. Sci. Mater. Electron. 33, pp. 9301-9311 (2022).

[7] V. Shalini, et al., J. Colloid Interface Sci. 633, pp. 120-131 (2023).

Anodizing is applied to the aluminum alloy of the structural material of chemical vapor deposition and etching equipment for semiconductor manufacturing to prevent degradation caused by reactive and corrosive gases. The anodized aluminum coating should be highly durable and have good vacuum properties. A new anodized aluminum coating with advanced durability and vacuum properties was developed, and its physical and vacuum-related characteristics were studied.

Conventional samples were prepared, including sulphate-anodized aluminium oxide with a thin sealing film (C-1), oxalic aluminium oxalate with a thin sealing film (C-2), and oxalic aluminium oxalate with a thick sealing film (C-3). To improve durability, an anodic aluminium oxalate oxalate coating was developed with no sealing treatment, a lower oxygen content and higher flexibility than conventional coatings. Surface SEM images of conventional and developed anodised aluminium samples are shown in Fig. 1. Conventional anodized aluminium samples display a foliated or needle-like surface structure, which can be attributed to the presence of the sealing film. The developed anodised aluminium has a comparatively smooth surface, but still has the pores characteristic of anodic oxidation.

The heat resistance of the prepared anodised aluminium coating samples was investigated by keeping them at a constant temperature for one hour and then counting the number of cracks that developed on the anodised aluminium surface. The heating temperatures were 70°C, 100°C, 150°C, 200°C, 250°C and 300°C. The results showed that sulphate-anodised alumina (C-1) with a thin sealing film and alumina oxalate (C-2) with a thin sealing film developed cracks in the sample heated to 100 °C, and the number of cracks in the sample increased at a higher heating temperature. The oxalic anodised aluminium oxalate sample (C-3) with a thick sealing film did not crack when heated to 100°C, but the sample heated to 150°C did crack. However, the developed anodised aluminium sample did not show any cracks when heated to 300°C. This suggests that the developed anodised aluminium oxide possesses high heat resistance.

Figure 2 shows the time dependence of the outgassing rate after vacuum pumping of the prepared anodised aluminium oxide samples from atmospheric pressure. The outgassing rate of the three conventional anodised alumina samples decreased with time but remained in the order of 10^-4 Pam³s^-1m^-2 up to 20 hours and in the order of 10^-5 Pam³s^-1m^-2 after 20 hours. The outgassing rates of the developed anodised alumina sample remained in the order of 10^-5 Pam³s^-1m^-2 after 4 hours and in the order of 10^-6 Pam³s^-1m^-2 after 30 hours. As a result, the outgassing rate of the developed anodised aluminium sample was an order of magnitude lower than that of the conventional anodised aluminium samples.

1. Introduction

It is necessary to reduce the outgassing from the materials for achieving the extreme high vacuum, and research on surface treatment and other methods. The component of temperature rises due to heat source such as the filament of the vacuum gauge and the mass spectrometer become to increase the outgassing [1]. BeCu material has characteristics high thermal conductivity (13 times more than stainless steel) and low thermal emissivity (1/7 of stainless steel). The special heat treatment is able to reduce the diffusion of gases from BeCu material. Furthermore, an oxide film (BeO) is formed on the surface of BeCu material to prevent gas permeation. Due to these characteristics, the use of BeCu material made it possible to reduce the outgassing [2]. So BeCu material with the special heat treatment used for extreme high vacuum equipment such as vacuum gauge, mass spectrometers, and vacuum piping [3]. The performance was confirmed by the build-up method.

2. X-ray Photoelectron Spectroscopy (XPS)

Mr. Watanabe reported that the BeO film thickness on the surface is about 10 nm, but no verification data remains. Therefore, the surface condition with and without heat treatment was observed by XPS analysis. The result of analysis shows Be (0%), C (49.9%), O (24.5%) and Cu (14.5%) without heat treatment, and Be (34.5%), C (9.8%), O (52.9%) and Cu (2.5%) with the heat treatment. The thickness of the surface oxide film was 1 nm (CuO) without the heat treatment and 7.4 nm (BeO) with heat treatment. The type of each oxide film was different. Details will be reported on the day.

3. Build up test

The build-up evaluation method is 25 hours baking at 200 degrees C. The chamber is isolated after the chamber pressure is stabilized after baking. The chamber pressure measures during 120 minutes after isolated. After the measurement was completed, the baking is restarted without purge to the atmosphere, and the same measurement was performed 5 times. For performance comparison, we also measured SUS304 chamber was pretreated by vacuum firing (10 hours baking at 850 degrees C) [4]. Both of the ultimate pressure and the buildup pressure without the baking were lower for SUS304 than for BeCu. This is because the BeCu chamber before test was stored in the atmosphere, and the SUS304 chamber before test was stored in a vacuum pack after vacuum firing. BeCu was lower in both the ultimate pressure and the buildup pressure in the first baking. As the number of baking times increased, both the ultimate pressure and the build-up pressure of BeCu decreased. For SUS304, the ultimate pressure remained almost unchanged, but the buildup pressure decreased. Details will be reported on the day.

Reference

[1] F. Watanabe: J. Vac. Soc. Jpn. Vol. 49, No.6, 2006

[2] F.Watanabe: J. Vac. Sci. Technol. A, Vol. 22, No. 1, Jan/Feb 2004

[3] F. Watanabe: J. Vac. Soc. Jpn. Vol. 56, No.6, 2013

[4] J. Kamiya: e-j Surf. Sci. Nanotechnol. 21, 144-153 (2023)

The adoption of 0.2% BeCu alloy and its related products inherit the technology of the late Dr. Fumio Watanabe of Vaclab Co., Ltd. who was involved in the development for many years. This was due to a strong desire to achieve E-13Pa. In order to achieve this, He started with the development of key points, the Q material, the S vacuum pump, and the P vacuum gauge (Fig. 1). Vacuum structure material to replace stainless steel, 0.2%BeCu alloy has been used as a vacuum structure material for XHV/UHV due to its high thermal conductivity and low thermal emissivity, and has been used for many years as a low outgassing technology. As a result, it is now used in vacuum gauges and mass spectrometers, and has successfully commercialized many products such as vacuum Nipple connections, vacuum chambers, and NEG pumps. However, due to the material's high thermal conductivity, electron beam or laser welding is difficult, so forged blocks are manufactured by mechanical cutting now. In addition, processing technology with 5-axis control may be required, but it will be extremely expensive. At this time, we will introduce the results of using the latest Hybrid Lasers and its welding technology, with the aim of improving manufacturing processes, reducing costs, and responding to various shape requirements. Here we will describe the characteristics of the 0.2% BeCu alloy and its manufacturing process. When this alloy is heat treated at 400℃ for 3 days using a vacuum furnace, almost 100% of the alloy surface is covered with Be atoms in the bulk. (Fig.2). During this process, the hydrogen dissolved in the bulk is sufficiently degassed. After cooling and purging with oxygen, the Be metal forms a stable oxide film of BeO, which absorbs water. This oxide film becomes a barrier film that prevents re-dissolution of the hydrogen generated by the hydrogen gas release rate of the 0.2% BeCu alloy treated with BeO is as low as 5E-13 Pa m/s, which is 1/100 of Stainless Steel and 1/10 Titanium alloy subjected to similar treatment. However, because the BeO film is thin, the part facing the atmosphere is oxidized during baking, and the CF edge of the flange cold-bonds with the copper gasket and adheres to it. Electroless nickel plating (NiP) was applied to the 0.2% BeCu product. NiP plating can form uniform plating on any complex shape, and increases surface hardness. CF flanges with this NiP plating reliability of the gasket is extremely high, and even after repeated baking at 250°C, no peeling of the plating or knife edge defects will occur. In addition, the coefficient of thermal expansion of the 0.2% BeCu alloy perfectly matches that of the pure copper gasket, so it is possible to connect a stainless-steel CF flange and a copper gasket (Fig.3). The manufacturing process consists of a forged block, mechanical cutting, NiP plating, removal of the NiP plating on the vacuum side (contact surface), chemical polishing of the copper surface, and BeO film formation in a vacuum furnace. The problem with laser welding of Cu that heat diffusion is fast and light absorption is low in the IR region, so stable heat input is not possible, resulting in unstable welding and vacuum leaks. As a method to solve this problem, Furukawa Electric Industry's Hybrid Laser has good results were obtained using BRACE®-X (Blue 1kw, IR 3kw), which combines Blue + IR lasers, so we will explain it at Oral Session. (Fig.4)

References

1) Fumio Watanabe: J. Vac. Sci. Technol. A22 (2004) 181 & 739. 2) Fumio Watanabe: J. Vac. Soc. Jpn, Vol. 56, No. 6, 2013 3) https://www.furukawa.co.jp/fiber-laser/product/lineup/hybrid_blue.html

Introduction

An ultra-high vacuum compatible ethanol liquid target is being developed to investigate the reaction products produced by the interaction of a high intensity laser with ethanol. One of the issues to be solved is the need to maintain an ultra-high vacuum in the main chamber for analysis within ~1 m of the liquid target generator, where the reaction chamber containing the generator is planned to be operated about 10 Pa or less. The constraint is to have a straight-through section without obstructions, so that reaction products can fly into the main chamber from the vicinity of the liquid target. In this development, this straight-through section is a cylinder with a diameter of 20 mm, as shown in Fig. 1. We attempted to solve the above problem by installing an in-line liquid nitrogen trap between the reaction chamber and the main chamber and improving the trap.

Trap evaluation experiment

An experimental system was prepared using a thimble-type in-line liquid nitrogen trap. The flanges A and B mating with the liquid target chamber and the main chamber were set to ISO NW25 with a face-to-face distance of 416 mm. The outer diameter and depth of the trap are 140 mm and 90 mm, respectively. A copper plate of 120 mm OD and 5 mm thick was attached to the bottom of the trap. This is for attaching the auxiliary plates and other materials that can be used in the following. Connect C₂H₅OH vapor introduction system to flange A. Ethanol vapor pressure is monitored by a Pirani vacuum gauge. Then an evaluation system is connected to flange B for flow evaluation and partial pressure measurement. Experiments were conducted at ethanol vapor pressures ranging from 0.01 Pa to 20 Pa.

Examples of the results obtained so far

(1) Trap efficiency with two additional copper plates

When the ethanol vapor was 4.93 Pa, the pressure in the trap was 1.73×10^-2 Pa. This indicates that the trap's pumping speed for ethanol vapor is about 2.3 m³/s. Since the area of the cooling surface is ~950 cm², the ethanol condensation coefficient is evaluated to be ~0.25.

(2) Trap efficiency with the two plates and an additional cooling pipe

When a cooling pipe with an inner diameter of 20 mm and a length of 100 mm was added at flange A to cover the straight pipe section, the pressure in the trap was 8.31 × 10^–4 Pa in response to ethanol vapor of 3.97 Pa. From this result, it can be seen that the ethanol vapor conductance between the introduction chamber and the trap was drastically reduced to about 5.5×10⁻⁴ m³/s by installing the cooling pipe. Also, if the same cooling pipes are installed not only at the flange A but also at the flange B, and if the main chamber is evacuated with a pump of ~1 m³/s, you can almost certainly keep the main chamber in UHV.

Details will be given at the lecture.

In J-PARC LINAC, the vacuum system is in place to maintain an ultra-high vacuum in the beam transport line (LINAC to 3GeV RCS beam transportation line: L3BT) between the LINAC to the 3-GeV synchrotron. The vacuum system is installed in the LINAC and L3BT buildings and consists of vacuum pumps, vacuum gauges, beam line gate valves (BLGVs), and other vacuum. Vacuum pumps and vacuum gauges are controlled by a Programmable Logic Controller (PLC) which provides the logic to safely transition from atmospheric to ultra-high vacuum conditions. On the other hand, the BLGV is controlled by the BLGV relay unit and VME controller using the Machine Protection System (MPS), which remotely opens and closes the BLGV and forcibly closes the BLGV when the beamline pressure exceeds a threshold value according to an interlock (ILK) logic. In the existing vacuum system, the vacuum equipments in each area between BLGV is independently controlled, and the vacuum system does not monitor the information on the equipments or vacuum pressure between nearby areas. This means that vacuum equipment can be operated regardless of the condition of adjacent areas, resulting in sudden deterioration of the vacuum in high vacuum areas, or injecting air into vacuum pumps that are in operation. In addition, when a vacuum deterioration occurs in the beam transport line, the vacuum deterioration ILK signal is transmitted to the BLGV relay unit via the MPS transmission signal, which causes the BLGVs to be forcibly closed. Because the ILK signal transmission range extends to all BLGVs in the L3BT, however, BLGVs in areas unaffected by vacuum deterioration are also forced to close. This could cause problems such as unnecessary open/close operations leading to more frequent maintenance cycles of the BLGVs. In addition, Since the BLGVs are operated using the MPS signal path, the open/close signal in the vacuum deteriorating ILK signal can only be sent uniformly to all BLGVs, and each individual control is not possible. Furthermore, maintenance of the vacuum control system also requires work involving the MPS signal path, making it difficult to maintain the vacuum control system alone and making the work complicated. To solve these problems, it is necessary to improve maintainability by separating the signal paths and automatically controlling BLGV separately. Therefore, the vacuum control system was modified and constructed with the aim of realizing a control system that takes into account the safety and efficient maintenance and operation of the L3BT vacuum system. This report summarizes the development and use of the L3BT vacuum system control system.

1. Introduction

In the Direct Simulation Monte Carlo (DSMC) method[1] which is generally used to analyze rarefied gas flows in, e.g., vaccum chambers and space, porous media have been modeled by randomly arranged solid cells as obstacles[2], packed cylinders or spheres[3], or using geometries imaged by computed tomography[4]. Unlike these conventional ways, a novel porous media model called virtual solid cell (VSC)[5] will be demonstrated in this abstract and at the meeting.

2. VSC model

In the VSC model, porous media is treated as a fluid domain, and the same effects as the conventional models are achieved by setting a stochastic limit corresponding to the porosity on the inter-cell transfer of gas molecules. The advantage of the VSC model as a stochastic approach enables one-dimensional (1D) simulation of a rarefied gas flow passing through porous media, contributing to a significant reduction in computation time[5]. In addition, it can be applied to complex porous media with various porosities and to rough surfaces as a new surface boundary model. These two application examples of the VSC model are described below.

Figure 1(a) shows the number density distribution of N₂ gas flow passing through a porous wall formed with multiple cross-shaped obstacles. The geometrical porosity of the porous wall is 0.445, and the cross-shaped obstacles are modeled using VSCs with a porosity of 0.1. The gas pressures P_in at the upstream boundary and P_out at the downstream boundary are 1 Pa and 0 Pa, respectively. The top and bottom boundary conditions are specified as symmetry. The N₂ gas mainly flows through the gap between the cross-shaped obstacles. In contrast, some obstacles with a higher N₂ number density than the main flow indicate that N₂ molecules are trapped inside the obstacles. As shown in this example, the VSC model can represent porous media with various porosities.

Figure 1(b) illustrates the temperature distribution of He gas enclosed between parallel plates. The Knudsen number Kn with reference to the distance between the plates is 4.59. The conditions of the wall temperatures and thermal accommodation coefficients are given in Fig. 1(b). In this example, two VSC layers are adopted to model surface roughness at the right side of the top wall. Thus, the simulated gas temperature is higher on the right side, especially in the vicinity of the VSC layers, than on the left side between the parallel plates. The gas temperature approaches the top wall temperature inside the VSC layers due to repeated molecular reflections as on a rough surface. This example depicts that the VSC model can also mimic surface roughness.

3. Summary

The VSC model to treat porous media in the DSMC method has been briefly demonstrated in this abstract. The conventional porous media models with complex geometry can be simplified by a stochastic procedure in the VSC model. Consequently, the VSC model enables 1D simulation of a rarefied gas flow passing through porous media, contributing to a significant reduction in computation time. It also facilitates the treatment of complex porous media with various porosities. Furthermore, the VSC model shows its potential as a new boundary model to describe surface roughness. Further details will be presented at the meeting.

References

[1] G. A. Bird, Molecular Gas Dynamics and the Direct Simulation of Gas Flows (Oxford Univ. Press, 1994).

[2] A. Saito et al., Trans. Jpn. Soc. Mech. Eng. B. 61, 248 (1995).

[3] Y. Kawagoe et al., Microfluid Nanofluid 20, 162 (2016).

[4] C. Christou and K. Dadzie, Society of Petroleum Engineers, Heriot-Watt Univ., SPE-173314-MS (2015).

[5] K. Denpoh, Vac. Surf. Sci. 66, 490 (2023).

Calculating the gas flow rate through a cylindrical tube of known geometry is actually complicated because the characteristics of the gas flow depend on pressure, gas species, temperature, tube diameter, and tube length. There are at least six flow regimes to explain the characteristics of the gas flow, such as molecular flow, viscous laminar flow, turbulent flow, critical flow, subcritical flow, and their intermediates including slip flow.

In recent, the author and coworkers have developed the modified Knudsen equation (MK equation) by combining equations of each gas flow, which is applicable to the whole flow regime for arbitrary length of the tubes [1,2]. This equation has two advantages; one is that this equation is used without considering Knudsen number (Kn), Reynolds number (Re), Mach number (Ma) and the length-to-diameter ratio of tube, and the other is that it can be solved in straight forward without an iterative procedure although the other equations sometimes need it. This equation is especially useful when one does not know which flow regime the gas flow is in. The MK equation typically produces results that agree to within 20 % to 30 % of previous results from experiments and computer simulations, but more study is still needed to confirm what range of Kn, Re, and Ma is applicable. Another aim is to improve the MK equation by examining the conditions under which it produces relatively large differences from existing results. As the results of comparison of MK equation with 82 literatures, it is found that the differences can be reduced mostly to below 20 % by introducing the effective length of turbulent flow into the MK equation. On the other hand, it is also found that considerable differences from 30 % to 70 % arise for the gas flow through an orifice when the pressure ratio of downstream to upstream is close to 1 and high-Re flow in short tubes.

Also, as an application of the MK equation, the pump-down time of vacuum chamber is tried to calculate by using the MK equation. Figure 1 shows the comparison of pump-down time calculated by the MK equation and the theory of Senda [3], and the experiments result Yokogawa et. al. [4], where the volume of vacuum chamber is 30 L, the pumping speed of the vacuum pump is 60 L/min and the diameter and the length of the connecting tube are 1.83 mm and 2 m, respectively. The results of MK equation are closer to the experimental results than Senda’s equation until 10³ s because the MK equation includes the effect of turbulent flow, but the Senda’s equation does not. The result of MK equation, however, underestimates the pump-down time similar to Senda’s equation at the time longer than 10³ s. This is because these calculations do not include the effects of outgassing from the chamber and reduction of pumping speed at lower pressure. Details will be discussed at the presentation.

[1] H. Yoshida, Y. Takei and K. Arai, Vacuum and Surface Science 63 (2020) 373.

[2] H. Yoshida, M. Hirata, T. Hara, Y. Higuchi, Packag Technol Sci. 34 (2021) 557.

[3] Y. Senda, SEI TECHNICAL REVIEW 176, 1 (2010).

[4] K. Yokogawa, K. Seto and T. Fukuda, Vacuum and Surface Science Vol. 61, No. 5 (2018).