Hard X-ray Photoelectron Spectroscopy as a Key Tool for Energy and Quantum Materials

Resistive random access memory (ReRAM) has been proposed as a new application for oxide materials and advanced to the commercial manufacturing stage. An oxide sandwiched between two metal electrodes shows reversible electric field–induced resistance switching behaviors. These are many resistivity changing mechanisms for the oxide based ReRAM structures such as the insulator–metal transition in perovskite oxides and conductive-bridging (CBRAM). In our research, we focus on the CBRAM with nano-electrolyte reaction, whose advantage is the low forming voltage. The CBRAM with nano-electrolyte reaction comprises the generation and rupture of a metal filament using a metal such as Ag and Cu acting as a fast mobile ion in oxides. Hafnium oxide (HfO2), which is used as a high-k gate insulator for advanced complementary metal-oxide-semiconductor (CMOS) technologies, has shown resistance switching phenomena and been increased interest in the use of HfO2 and related oxides as potential ReRAM materials.1) To put the oxide based ReRAM on practical applications, understanding on controls of metal/oxide interface is essentially important. Here, we employed hard x-ray photoelectron spectroscopy (HX-PES) under bias operation. HX-PES is a powerful tool for investigating the electronic structure and chemical state of the surface/interface of stacking structures for nanoelectronics devices without any degradation because it has a longer photoelectron mean free path than conventional x-ray photoelectron spectroscopy using Al Kα radiation (hν= 1486.6 eV). The detection depth of HX-PES with an energy of 6 keV is approximately three times deeper than that of conventional XPS, so the photoelectron from a metal/oxide interface, which works as an electrical device, can be detected by HX-PES. With this method, bias-induced compositional changes around the metal/oxide interface during device operation have been directly observed. HX-PES was performed at the SPring-8 BL15XU undulator beamline. The incident X-ray energy and the total energy resolution were 5.95 keV and 240 meV, respectively. We have demonstrated resistance switching using HfO2 film with a Cu top electrode for nonvolatile memory applications, and revealed the Cu diffusion into the HfO2 layer during the conductive filament formation process. Resistive switching was clearly observed in the Cu/HfO2/Pt structure by performing current-voltage measurements. The current step from a high resistive state to a low resistive state was of the order of 103-104, which provided a sufficient on/off ratio for use as a switching device. The filament formation process was investigated by employing HX-PES under bias operation. The application of a bias to the structure reduced the Cu2O state at the interface and the intensity ratio of Cu 2p3/2/Hf 3d5/2, providing evidence of Cu2O reduction and Cu diffusion into the HfO2 layer. These results also provide evidence that the resistance switching of the Cu/HfO2/Pt structure originates in a solid electrolyte reaction containing Cu ions. HX-PES also revealed the top or bottom electrode dependences of interface reduction or oxidization, and ion migration behaviors.2-5) In the presentation, we will show the details of the correlation between the switching mechanism and the interface reaction in the electrode/high-k dielectrics based ReRAM structure.6) We are grateful to HiSOR, Hiroshima Univ. and JAEA/SPring-8 for the development of HX-PES at BL15XU of SPring-8. The HX-PES measurements were performed under the approval of the NIMS Beamline Station (Proposal Nos. 2009A 4600, 2010B 4600, 2011A 4611, 2011B 4613, and 2012A 4613).

X-ray Photoelectron Spectroscopy (XPS) is a widely used surface analysis technique with many well established industrial and research applications. The surface sensitivity (top 5-10 nm) of XPS and its ability to provide short-range chemical bonding information make the technique extremely popular in materials characterization and failure analysis laboratories. While its surface sensitivity is an important attribute, in some cases, the depth of analysis of XPS is not sufficient to analyze buried interfaces without first sputter etching the sample surface. However, sputter etching can often lead to alterations of the true surface chemistry. An alternative to sputter etching the sample is Hard X-ray Photoelectron Spectroscopy (HAXPES), available at some synchrotron facilities. HAXPES utilizes X-rays typically defined as having energies greater than 5 keV. By increasing the photon energy of the X-ray source, the mean free path of photoelectrons is increased, resulting in an increased information depth obtained from the sample. Depending on the energy used, these hard X-rays can provide depths of analysis three or more times than that of soft x-rays used on conventional XPS systems. HAXPES measurements are, therefore, more sensitive to the bulk, and contributions from the surface are minimized.1,2 This presentation will describe a laboratory-based instrument, the PHI Quantes, equipped with two scanning microprobe monochromated X-ray sources, Al Kα (1486.6 eV) and Cr Kα (5414.9 eV), thus enabling both traditional XPS and HAXPES experiments in the same instrument. Combining both soft and hard X-ray analyses, we can gain an even better understanding of composition with depth and information at buried interfaces. References Kobayashi, K. Hard X-ray photoemission spectroscopy, Nucl. Instr. Meth. Phys. Res. A 2009, 601, 32-47.Fadley, C.S. Hard X-ray Photoemission: An Overview and Future Perspective. In Hard X-ray Photoelectron Spectroscopy (HAXPS); Woicik, J. C., Ed; Springer: Switzerland 2016.

Resistive random access memory (ReRAM) has been proposed as a new application for oxide materials. An oxide sandwiched between two metal electrodes shows reversible electric field–induced resistance switching behaviors. Hafnium oxide (HfO2), which is used as a high-k gate insulator, has also shown resistance switching phenomena and been increased interest in the use of HfO2 and related oxides as potential ReRAM materials [1]. For the oxide based ReRAM with the conductive filament formation model, two mechanisms of resistance switching have been proposed. One is the nanoionics model, which comprises the generation and rupture of a metal filament using a metal such as Cu acting as a fast mobile ion in oxides. The other model is that of oxygen vacancy nucleation at the metal/oxide interface. To put the oxide based ReRAM on practical applications, understanding on controls of metal/oxide interface is essentially important.Here, we employed hard x-ray photoelectron spectroscopy (HX-PES) under bias operation to examine the electronic structure of Pt or Cu/HfO2 ReRAM structures in an operating device. HX-PES is a powerful tool for investigating the electronic structure and chemical state of the surface/interface of stacking structures for nanoelectronics devices without any degradation because it has a longer photoelectron mean free path than conventional x-ray photoelectron spectroscopy using Al Kα radiation (hν= 1486.6 eV). The detection depth of HX-PES with an energy of 6 keV is approximately three times deeper than that of conventional XPS, so the photoelectron from a metal/oxide interface, which works as an electrical device, can be detected by HX-PES. With this method, bias-induced compositional changes around the metal/oxide interface during device operation have been directly observed. The interface electronic states of 10-nm-thick Pt or Cu top electrodes/HfO2/Pt structures were measured with HX-PES in the SPring-8 BL15XU undulator beamline. The incident X-ray energy was 5.95 keV. In the case of the Pt/HfO2 interface, applying a forward bias increased the Pt–O bonding peak as shown in Fig 1, indicating evidence of Pt electrode oxidization and oxygen vacancy formation around the interface [2]. In contrast, the application of a bias to the Cu/HfO2 interface reduced the copper oxide bonding state, providing evidence of oxygen reduction and Cu diffusion into the HfO2 layer [3,4]. We achieved direct observation of oxygen migration at the metal/HfO2 interface under device operation, which is the key to controlling the electrical properties of oxide based ReRAM. Based on these results, we demonstrated the control of the switching voltage and the initial conductive filament formation process of the Ta−Nb binary oxide ((TaxNb1−x)2O5) ReRAM structure using a combinatorial method [5]. The relationship between the interface structure and the electrical properties also will be discussed in detail at the presentation.

1. Introduction As a packing density in the large-scale integration (LSI) becomes higher, the devices exhibit three-dimensional complicated structure. The deep trench with high aspect ratio (AR), i.e. opening/depth, is one of the components to realize the structure [1]. The atomic layer deposition (ALD) is one of techniques suitable for trench structure, because which can conformally deposit a thin film along the high AR trench structure [2,3]. The ALD film theoretically consist of the same composition with the same thickness along the trench no matter how the AR is high. In this study, we evaluated the ALD deposited SiN films along the deep trench with high AR of 3 and 7.5. 2. Experiment 10 nm-thick SiN films were deposited using alternate supply of SiH2Cl2 and NH3 precursor with plasma on the plane and trench structure with 40 nm opening and 120 or 300 nm deep, AR of 3 and 7.5 at 350, 550, 630oC, respectively. The SiN film were conformally deposited along the trench as shown in Fig. 1. The fabricated samples were evaluated using, conventional X-ray photoelectron spectroscopy (XPS) and Hard X-ray photoelectron spectroscopy (HAXPES). For the XPS and HAXPES evaluation, we carried out angle-resolved spectroscopy. As the HAXPES equipment, a laboratory-based HAXPES (Lab. HAXPES) equipment product of Scienta Omicron Inc. was used [4]. The X-ray source energy was 9,251.74 eV from liquid GaKα. The photoemission angle (Take-off angle: TOA) was varied from 90 degree to 30, with the photoelectron energy resolution of approximately 0.5 eV. 3. Results and Discussion Figure 1 illustrates the trench sample schematics with the explanation of the angle-resolved measurements. One can understand by reducing the TOA from 90 degrees, the photoelectron from the lower part of the trench side walls cannot reach to the photoelectron energy analyzer, therefore the measurement becomes top layer sensitive. Figure 2 compares the Si 1s spectra from the samples at 350oC with (blue) and without (red) trenches. The film on the flat surface, without trenches, showed mostly composed of Si-N chemical bonds, while the spectrum from the trench sample showed more Si-O bonds in the film. The films on or in the trench structure seem to be different from the flat surface, more oxide than nitride, although the ALD was performed to deposit SiN film. Figure 3 shows the angle-resolved HAXPES results for the trench sample at 350oC. From Fig. 3, it can be recognized the Si-O component in the film decreased by reducing the TOA from 90 to 60, implying there are more oxide in the lower part of trench than the upper part. Plasma activated N precursors might lose their energy (activity) during the proceeding narrow high AR trench, resulting in the non-stoichiometric film remained close to the bottom part of trench, which might be oxidized after the film deposition, although the film thickness seems to be the same for all over the trench structures. The stoichiometry uniformity was achieved by elevating the deposition temperature up to 550oC. 4. Conclusion In conclusions, we have evaluated the ALD SiN film conformally formed physically in the high AR trench and found there were possible non-uniformity in the chemical structure along the deep trench. Acknowledgements We appreciate the SPring-8 BENTEN database for the Si 1s spectra assignment. Reference [1] T. Franza et al., ACS Appl. Mater. Interfaces 9, 1858 (2017).[2] T. Antonio et al., Materials Matters 13, 55 (2018).[3] X. Meng et al., Materials 9, 1007 (2016).[4] A. Regoutz et al., Rev. Sci. Instrum. 89, 073105 (2018). Figure 1

Determination of electronic and atomic properties of surface, bulk and buried interfaces: Simultaneous combination of hard X-ray photoelectron spectroscopy and X-ray diffraction

X-ray photoelectron spectroscopy for analysis of plasma–polymer interactions in Ar plasmas sustained via RF inductive coupling with low-inductance antenna units

The experimental determination of valence band offsets (VBOs) at interfaces in complex-oxide heterostructures using conventional soft x-ray photoelectron spectroscopy (SXPS, hν ≤ 1500 eV) and reference core-level binding energies can present challenges because of surface charging when photoelectrons are emitted and insufficient probing depth to clearly resolve the interfaces. In this paper, we compare VBOs measured with SXPS and its multi-keV hard x-ray analogue (HXPS, hν > 2000 eV). We demonstrate that the use of HXPS allows one to minimize charging effects and to probe more deeply buried interfaces in heterostructures such as SrTiO3/LaNiO3 and SrTiO3/GdTiO3. The VBO values obtained by HXPS for these interfaces are furthermore found to be close to those determined by first-principles calculations.

Hard x-ray photoelectron spectroscopy (HAXPES) is a powerful and novel emerging technique for the nondestructive determination of electronic properties and chemical composition of bulk, buried interfaces and surfaces. It benefits from the exceptionally large escape depth of high kinetic energy photoelectrons, increasing the information depth up to several tens of nanometers. Complementing HAXPES with an atomic structure sensitive technique (such as x-ray diffraction) opens a new research field with major applications for materials science. At SpLine, the Spanish CRG beamline at the European Synchrotron Radiation Facility, we have developed a novel experimental set-up that combines HAXPES and x-ray diffraction (x-ray reflectivity, surface x-ray diffraction, grazing incidence x-ray diffraction, and reciprocal space maps). Both techniques can be operated simultaneously on the same sample and using the same excitation source. The set-up includes a robust 2S + 3D diffractometer hosting a ultrahigh vacuum chamber equipped with a unique photoelectron spectrometer (few eV < electron kinetic energy < 15 keV), x-ray tube (Mg/Ti), 15 keV electron gun, and auxiliary standard surface facilities (molecular beam epitaxy evaporator, ion gun, low energy electron diffraction, sample heating/cooling system, leak valves, load-lock sample transfer, etc.). This end-station offers the unique possibility of performing simultaneous HAXPES + x-ray diffraction studies. In the present work, we describe the experimental set-up together with two experimental examples that emphasize its outstanding capabilities: (i) nondestructive characterization of the Si/Ge and HfO2/SiO2 interfaces on Ge-based CMOS devices, and (ii) strain study on La0.7Ca0.3MnO3 ultrathin films grown on SrTiO3(001) substrate.

Photoelectron spectroscopic study on band alignment of poly(3-hexylthiophene-2,5-diyl)/polar-ZnO heterointerface

1. Introduction GeSn alloy is an attractive material as a high hole mobility channel substituting Si and as a stressor for strained channels. Furthermore, modulation of band structure from indirect to direct gap by the increasing Sn composition up to 6-10 % is expected to improve the performance of optical devices [1, 2]. Then, a high-quality GeSn thin film with higher Sn concentration is a promising for higher performance and integrated devices on the Si platform. However, it has been difficult to increase Sn concentration, because the solubility limit of Sn within a GeSn film is considered to be approximately 1 atomic % [3]. Hence, a growth technique which enables the suppression of Sn segregation during GeSn growth and a characterization to analyze the surface segregation occurred with a thickness of several nm, are essential. For these purpose, we have developed a metal-organic chemical-vapor-deposition (MOCVD) method with new source gases that is safe, uniform, and industrially applicable [4]. In addition, we also investigated a hard X-ray photoelectron spectroscopy (HAXPES) and X-ray diffraction (XRD) on synchrotron technique to analyze thin GeSn films. In this presentation, we introduce high Sn concentration GeSn thin film exceeding 6% prepared by combining the new MOCVD growth and synchrotron analysis methods. 2. Experimental Thin GeSn films with the thickness of typically 30-100 nm were grown on (001) Ge substrates at low temperature (320-360 degrees) by the MOCVD method using Ge (t-C4H9GeH3) and specially prepared Sn ((C2H5)4Sn) source gases. The Sb was doped from the atmosphere with residual triisopropyl-antimony [(i-C3H7)3Sb]. The target compositions of Sn were 2%, 3%, 6%, respectively. To clarify the crystallinity of GeSn films, HAXPES and XRD were carried out at SPring-8. Since the inelastic mean free path (IMFP) is several times deeper than that for conventional X-ray photoelectron spectroscopy (XPS), HAXPES is expected to be useful to identify simultaneously the variation of chemical state at a surface part and the underlying bulk part of a GeSn film [5]. The Sn and Sb concentrations were calculated by the combination of Rutherford back-scattering spectroscopy (RBS) and secondary ion mass spectrometry (SIMS) measurements. Cross-sectional TEM was also carried out for the evaluation of the crystal quality. 3. Results and discussion Initially, we observed the splitting of the Sn3d5/2 spectrum into two peaks for 3% Sn composition GeSn film by the HAXPES measurement, whereas the spectrum of a low Sn composition (2%) GeSn film remained single (Fig.1). The binding energy of the newly split peak (M2) was lower than that of the Sn3d5/2 peak (M1) and the peak position of M1 approximately coincided with that of the 2% GeSn film. To clarify the newly split Sn3d5/2 peak (M2), total reflection mode HAXPES (TR-HAXPES) measurement was also carried out for the 3% GeSn film. As a result, only a single Sn3d5/2 spectrum was observed by the measurement. Since the position of the observed peak by the TR-HAXPES, closely coincided with the split peak (M2), the M2 was identified as a peak derived from the Sn segregation formed at the film surface. Hence, it is concluded that depth profile characterization of the Sn chemical state within a GeSn film possible by combining the HAXPES and TR-HAXPES measurements. In order to further increase the Sn concentration, we employed Sb as a surfactant [6]. As a result, neither extreme change of XRD spectrum due to strain relaxation or asymmetric structure of HAXPES by the Sn segregation was observed for the newly grown GeSn film of 6% Sn. A decrease in the broad peak intensity on the high binding energy side due to surface oxidation was also confirmed. Since the surface is more stabilized by Sb with lower interfacial free energy, it is expected that suppression of Sn segregation and uniform Sn concentration in depth direction are simultaneously achieved. The concentration distribution of Ge and Sb were confirmed by the RBS and SIMS, and it was revealed that the Sn concentration was uniform in the depth direction, and the composition of Sn and Sb was 6.6 at.% and 0.5 at.%, respectively. Hence, by combining newly developed MOCVD technique and synchrotron technique, it was confirmed that uniform GeSn thin layer with higher Sn composition was realized. Reference [1] S.Zaima, JJAP 52 (2013) 030001, [2] R.Cjhen et al., APL 99, (2011) 181125, [3] C. Thurmond et al., J.Chem.Phys. 25,(1956)799, [4] K.Suda et al., ECS J.SSST, 4 (2015) 152. [5] K.Usuda et al., MRS spring meeting, (2016), EP11.6.10. [6] X. Yang et al., IEEE Photon.Tech.Lett.12,(2000) 128. Figure 1

The spin-resolved electronic structure of buried magnetic layers is studied by hard X-ray photoelectron spectroscopy (HAXPES) using a spin polarimeter in combination with a high-energy hemispherical electron analyzer at the high-brilliance BL47XU beamline (SPring-8, Japan). Spin-resolved photoelectron spectra are analyzed in comparison with the results of magnetic linear and circular dichroism in photoelectron emission in the case of buried Co2FeAl0.5Si0.5 layers. The relatively large inelastic mean free path (up to 20 nm) of fast photoelectrons enables us to extend the HAXPES technique with electron-spin polarimetry and to develop spin analysis techniques for buried magnetic multilayers and interfaces.

The spin-resolved electronic structure of buried magnetic layers is studied by hard X-ray photoelectron spectroscopy (HAXPES) using a spin polarimeter in combination with a high-energy hemispherical electron analyzer at the high-brilliance BL47XU beamline (SPring-8, Japan). Spin-resolved photoelectron spectra are analyzed in comparison with the results of magnetic linear and circular dichroism in photoelectron emission in the case of buried Co2FeAl0.5Si0.5 layers. The relatively large inelastic mean free path (up to 20 nm) of fast photoelectrons enables us to extend the HAXPES technique with electron-spin polarimetry and to develop spin analysis techniques for buried magnetic multilayers and interfaces.

Hard x-ray photoelectron spectroscopy (HAXPES) is a powerful technique to characterize the chemical and electronic structures of materials. In energy conversion devices, often composed of a stack of thin layers and thus containing multiple buried interfaces, the increased probing depth of HAXPES, compared to conventional x-ray photoelectron spectroscopy, makes it a technique of choice to ultimately reveal a more comprehensive device-relevant picture. In this contribution, we provide a brief review of recent HAXPES experiments conducted at the High Kinetic Energy Photoelectron Spectrometer endstation located at the BESSY II KMC-1 beamline at Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, focusing on three different prominent sample material systems widely used in energy conversion devices: (1) Cu-chalcogenides, (2) metal oxides, and (3) halide perovskites. The findings revealed by these studies highlight the advantage of knowledge-based heterointerface design in energy conversion devices, building interface models based on direct measurements targeting the entire structure as only possible by HAXPES. We conclude by giving an update on the new and enhanced HAXPES experimental capabilities starting to be offered by the Energy Materials In-situ Laboratory Berlin facility.

To obtain depth-resolved magnetic information from Fe films grown on fcc Co∕Cu(001), we have used various signal sources for the detection of x-ray absorption spectroscopy. These include total electron yield (TEY) and partial electron yield (PEY) of inelastic electrons at various kinetic energies between 70 and 470eV as well as PEY using photoelectrons at a fixed binding energy (constant initial state: CIS) near the Fermi level. Inelastic electron yield at electron emission angles up to 87° from the surface normal was found to be as nonsurface sensitive as TEY, however, the CIS mode shows a shorter information depth, comparable to the inelastic mean free path of photoelectrons. No difference in the dichroic signal at the Fe L3 edge was found between the CIS and TEY modes for a 2-monolayer (ML) Fe∕Co film, but an 8-ML Fe∕Co film showed a much higher dichroic signal in the CIS mode than that in the TEY mode. This is consistent with a homogeneous magnetic film at an Fe thickness of 2 ML and a nonhomogeneous magnetic film with a live ferromagnetic layer on the surface with nonferromagnetic underlayers at an Fe thickness of 8 ML. Thus, it is possible to extract depth-resolved magnetic information from x-ray magnetic circular dichroism by combining the surface sensitive CIS mode with other detection modes with less surface sensitivity.

The interactions of ions and photons in ultraviolet (UV) and vacuum ultraviolet (VUV) regions from argon plasmas with polymer surfaces were investigated by of depth analysis of chemical bonding states in the nano-surface layer of poly(ethylene terephthalate) (PET) films via conventional X-ray photoelectron spectroscopy (XPS) and hard X-ray photoelectron spectroscopy (HXPES). The PET films were exposed to argon plasmas by covering the PET films with MgF2 and quartz windows as optical filters to compare the irradiation effects with ions and photons. The conventional XPS results indicated that oxygen functionalities (the C–O bond and the O=C–O bond) were degraded by ion bombardment in the shallower region up to about 10 nm from the surface, whereas the effect of photoirradiation in the UV and VUV regions was insignificant. The HXPES analysis showed that irradiation with ions and photons did not cause serious damage in chemical bonding states in the deeper region up to about 50 nm from the surface.