Deconvolution in X-ray photoelectron spectroscopy

Published Mg and Al Kα X-ray lineshapes have been used to study the removal of the effects of X-ray broadening in XPS spectra by deconvolution. These results are compared with spectra obtained for the same specimens in the same instrument using a monochromatic X-ray source. The use of deconvolution to remove analyzer broadening from spectra has also been examined, and it has been verified experimentally that the maximum improvement in the full width at half-maximum of a peak that can be achieved is 30%. Deconvolution has also been examined as a means to remove backgrounds from XPS spectra over wide energy ranges, up to 100 eV.     

The application of deconvolution methods in electron spectroscopy — a review

In this review of the application of deconvolution in electron spectroscopy the operations of convolution and deconvolution are first defined. The theoretical background to direct methods, Fourier Transform techniques and iterative approaches to deconvolution is given. The use of smoothing techniques, the removal of secondary electron background and the determination of broadening functions are then considered. Examples of deconvolution in electron spectroscopy are discussed. These include valence and core levels in XPS, gas-phase UPS studies and resolution enhancement in Auger spectroscopy. Deconvolution of Auger spectra to yield density of states curves is examined. The likely future status of deconvolution in electron spectroscopy is considered.     

Iterative deconvolution of X-ray photoelectron spectra generated by an achromatic source

Three cases are presented where X-ray photoelectron spectra generated by an achromatic X-ray source have been deconvoluted by the iterative, Van Cittert method. These are (i the gold 4 f doublet, (ii) the gold valence band, and (iii) the aluminium 2 p spectrum of an oxidized aluminium foil. All prominent features in the deconvolute are shown to be real by comparison with high resolution studies, and the noise levels are low. Resolution after deconvolution is similar to that obtained with a monochromatized X-ray source. The spectra are also similar to those obtained by Fourier transform deconvolution.     

High resolution XPS study of the large-band-gap semiconductor stibnite (Sb2S3): Structural contributions and surface reconstruction

Conventional X-ray photoelectron spectroscopy (XPS) and synchrotron radiation XPS (SRXPS) were used to probe the chemical state properties of stibnite (Sb₂S₃), a large-band-gap semiconductor of complex structure. The conventional spectra were obtained with a Kratos Axis Ultra XPS with magnetic confinement charge neutralization, which is very effective in minimizing both uniform charging and differential charging on this large-band-gap semiconductor. The narrow linewidths (much narrower than previously obtained) for single doublet fits (e.g. Sb 4d_5/2 of 0.57 eV and S 2p_3/2 of 0.63 eV) enabled the observation of a small peak on the low binding energy side of the Sb 3d and Sb 4d lines. With the aid of the very surface-sensitive Sb 4d SRXPS spectra, these low energy peaks are assigned to small Sb metal clusters at the surface after cleavage; the signal for these clusters increases with X-ray dose on the sample.A detailed analysis of the Sb 4d and S 2p linewidths concludes that the Sb 4d_5/2 linewidth is larger than expected based on the inherent linewidth of the instrument and the Sb 4d lifetime width, and on comparison with the As 3d linewidth (0.52 eV) for the analogous As₂S₃. Also, the S 2p_3/2 linewidth is substantially broader than the Sb 4d_5/2 linewidth. These larger than expected linewidths are due to two structurally distinct Sb atoms and three structurally distinct S atoms in the Sb₂S₃ crystal structure. Accordingly, the Sb 4d and S 2p spectra have been fitted to two and three doublets respectively, and the linewidth for all peaks is 0.53 eV. Using recent molecular orbital calculations, the doublets have been assigned to the different structural Sb and S sites.     

Maximum-entropy deconvolution applied to electron energy-loss spectroscopy

In the present paper, we investigate the performance of maximum-entropy deconvolution, in removing the instrumental response function from electron energy-loss spectra. To this end we make use of spectra acquired from the carbon K-edge in graphite for a range of signal-to-noise ratios. The zero-loss peak is used as the instrumental profile. The resolution improvement obtained through the application of the deconvolution algorithm as a function of the signal-to-noise ratio is well described by a logarithmic dependency. The claimed resolution improvement is further substantiated by demonstrating the consistency between improvement obtained for the width of the instrumental response function, the width of the π¹ peak and the splitting of the σ¹ peaks for a range of signal-to-noise ratios.     

Relation between 2p X-ray photoelectron and Kα X-ray emission spectra of manganese and iron oxides

The high resolution Mn and Fe Kα X-ray emission spectra (XES), and Mn and Fe 2p X-ray photoelectron spectra (XPS) for manganese and iron oxides were measured. The spectra were compared with those of [MnO₄]⁻, [Fe(CN)₆]⁴⁻ and [Fe(CN)₆]³⁻ ions. As the electronic structure of the latter compounds do not change with electron hole creation in the core levels, satellite peaks due to charge transfer are not observed in the 2p XPS spectra, and the peak profiles of metal 2p XPS and Kα XES are governed by the exchange splitting between 2p and valence electrons. The metal 2p XPS spectra of the oxides had satellite peaks, but the XES spectra had no satellites. FWHMs of the metal 2p_3/2 main peaks of the compounds being low spin states are smaller than those of metal Kα₁ XES spectra. However, FWHMs of Mn 2p_3/2 of the manganese oxide were nearly equal to those of Mn Kα₁ XES spectra, and those of Fe 2p_3/2 XPS spectra of the iron oxides are greater than those of Fe Kα₁ XES spectra.     

X-ray photoelectron spectroscopic study of nitrogen incorporated amorphous carbon films embedded with nanoparticles

The effect of substrate bias on X-ray photoelectron spectroscopy (XPS) study of nitrogen incorporated amorphous carbon (a-C:N) films embedded with nanoparticles deposited by filtered cathodic jet carbon arc technique is discussed. High resolution transmission electron microscope exhibited initially the amorphous structure but on closer examination the film was constituted of amorphous phase with the nanoparticle embedded in the amorphous matrix. X-ray diffraction study reveals dominantly an amorphous nature of the film. A straight forward method of deconvolution of XPS spectra has been used to evaluate the sp³ and sp² contents present in these a-C:N films. The carbon (C 1s) peaks have been deconvoluted into four different peaks and nitrogen (N 1s) peaks have been deconvoluted into three different peaks which attribute to different bonding state between C, N and O. The full width at half maxima (FWHM) of C 1s peak, sp³ content and sp³/sp² ratio of a-C:N films increase up to −150 V substrate bias and beyond −150 V substrate bias these parameters are found to decrease. Thus, the parameters evaluated are found to be dependent on the substrate bias which peaks at −150 V substrate bias.     

High energy xps using a monochromated Ag Lα source: resolution,sensitivity and photoelectric cross sections

Resolution, sensitivity and calibration data are presented for a novel high energy XPS source, monochromated Ag Lα radiation (hv = 2984.3 eV). Adequate resolution is attainable for good signal/noise spectra, whilst values for experimental sensitivity factors agree well with theoretical cross section values calculated by Nefedov. This allows an evaluation of ESCA 3 Mk. II transmission function up to 3000 eV, which appears to obey an approximate E^{$\text{?1}\text{2}$} dependence. Monochromated Ag Lα (linewidth 1.3 eV) overcomes the problem of broad natural linewidths for high energy sources, such that chemical state information can be gained. Various new core level and Auger peaks are developed, a notable feature being the 1s core level and KLL Auger transition capability from Al through to Cl. Improved sensitivity is experienced for elements whose major peaks occur in the 1500–3000 eV BE range, whilst there is no serious reduction of sensitivity in the conventional XPS energy range.     

Fourier-wavelet regularized deconvolution (ForWaRD) for lidar systems based on TEA-CO2 laser

Lidar is an efficient tool for remotely measuring physical quantities or detecting targets. To improve the range resolution in long pulse lidars, such as lidar systems based on TEA-CO₂ lasers, deconvolution methods were used by previous investigators. Deconvolution is a noise sensitive process. In order to avoid noise amplification during the deconvolution process, the Fourier-wavelet regularized deconvolution method is used to deconvolve and denoise the back-scattered lidar signal simultaneously. This method is applied to lidar systems based on the TEA-CO₂ laser and the results are compared to nitrogen tail clipping method. Numerical simulation shows, in comparison to the clipping nitrogen tail technique, by using the ForWaRD method; the range resolution and working distance of the lidar is improved and the clipping tail apparatus is also eliminated.     

A Structural Study of Some Metal Diethyldithiocarbamate Chelates by X-ray Photoelectron Spectroscopy

The diethyldithiocarbamate chelates of copper (II), lead (II) and tin (IV) have been prepared and characterized. Pure or nearly pure metal diethyldithiocarbamate (DEDTC) chelates could be prepared through recrystallization. The x-ray photoelectron spectra (XPS) for the recrystallized samples were obtained. Three different carbons (C Is) in the diethyldithiocarbamate molecule can be recognized through a computer deconvolution system. Sulfer 2p showed two peaks of spin doublet S 2p_1/2 and S 2p_3/2. Based on the XPS results we may conclude that the diethyldithiocarbamate ligand forms an unusual 4-membered chelate ring structure with the metal ion, both sulfur having bonding with the metal. The N in the DEDTC may also bind with the metal.     

Interference of O Kα ghost features in X-ray-excited Auger spectra

We have noted the interference of O Kα ghost peaks in spectra obtained using unmonochromatized X-ray sources, particularly in C(KVV) and O(KVV) X-ray-excited Auger spectra. The source of the undesirable O Kα photons is a surface oxide on the Mg or Al anode of the polychromatic X-ray sources used. A calculation of the oxide thickness required to give the observed ghost intensity in measurements on a gold sample using a heavily oxidized Mg anode is given and the estimated oxide thickness found plausible. Since it may be impossible to operate the anode with no oxide present, the use of a standard Au sample is suggested as an excellent test for the presence of O Kα X-radiation from unmonochromatized X-ray sources.     

Comparison of site-specific valence band densities of states determined from Auger spectra and XPS-determined valence band spectra in GeS (001) and GeSe (001)

Auger lineshapes of the Ge M₁M_4,5V and M₃M_4,5V and Se _M1M_4,5V transitions in GeS (001) and GeSe (001) are measured and compared to XPS valence band spectra. Distortions in both types of spectra due to inelastic scattering, analyzer and source broadening, and core level lifetime broadening are removed by deconvolution techniques. The valence band consists of three main peaks at ?2 eV, ?8 eV, and ?13 eV. There is excellent agreement of peak positions in AES and XPS spectra. The Auger lineshapes can be interpreted in terms of site-specific densities of states. They indicate that the states at ～?8 eV and at ～?13 eV are associated with the cation and anion sites respectively. The bonding p-like states at the top of the valence band have both cation and anion character. The Auger lineshapes indicate that the states closest to the valence band maximum are preferentially associated with Ge.     

Thermoluminescence glow curves of CaF2: Dy crystals irradiated by soft X-rays

Applying a deconvolution of the thermoluminescence glow curves, parameters of single glow peaks of CaF₂: Dy TLD 200 dosemeters irradiated by soft X-rays were determined. A dependence of the height ratio of low temperature (T393, 413 and 473 K) single peaks on energy of absorbed photons was measured in a region of 1–22.2 keV. Standard radionuclides¹⁰⁹Cd,²³⁸Pu,⁵⁵Fe and iodine laser produced aluminium plasma (T _e 500 eV) were used as soft X-ray sources. The ratios of the heights of different single peaks are discussed with respect to high local doses. The decreasing ratio of the heights of the first and third and/or second and third peak with increasing photon energy allows to determine reversely a mean photon energy of absorbed soft X-ray radiation.     

盲目反卷积光谱图超分辨复原算法

反卷积是实现光谱图超分辨复原的重要手段,与常规反卷积相比,盲目反卷积具有不需要预先准确获取卷积核函数的优势。着眼于充分利用光谱信号的特点和已有的光谱图反卷积成果,详细讨论了空域迭代盲目反卷积方法用于光谱图反卷积时的算法实现问题,并在分析光谱图卷积退化过程的基础上,针对光谱图反卷积算法特点,提出了光谱图卷积退化简化计算模型和最小二乘高斯拟合模型,以解决算法中相应的计算问题。基于Matlab平台的仿真表明,对于所用的高斯型谱线和点扩散函数,空域迭代盲目反卷积算法效果良好,在信噪比为50 dB时,分辨率提高约30％。     

A complete and self-consistent evaluation of XPS spectra of TiN

The electron configuration in single crystalline (sc-)titanium nitride (TiN) has been quantitatively studied using angle resolved X-ray photoelectron spectroscopy (AR-XPS). All samples were fabricated and transferred in situ so that only minimal surface contaminations were observed. The residual oxygen contamination was separated from the bulk information by extrapolating angle resolved measurements. Special attention is given to the quantitative evaluation of the spectra based on basic principles. Shake-up features are observed on core level lines and appear due two final ionized states. The ratio of this shake-up and the main Ti 2p energy line are investigated in this paper. In order to quantify this shake-up a precise evaluation is required. Here we present an approach to evaluate the XPS spectra of the Ti 2p photoemission line in TiN in a self-consistent manner that accounts for all features observable in an energy window of 80 eV. The evaluation considers the appropriate Tougaard background correction, shake-up features as well as surface and bulk plasmons. The ratio of the Ti 2p_1/2 and Ti 2p_3/2 and the corresponding peaks in the energy loss features fulfill the requirements given by quantum mechanics. The energy loss ΔE due to the shake-up process and the shake-up ratio have been determined quantitatively for oxygen-free bulk titanium nitride. The origin of the shake-up, its intensity and energy difference ΔE are explained by a two electron excitation process.     

增量维纳滤波法在波前探测解卷积中的应用

一个常规的自适应光学系统通常包含三个重要环节：波前探测、波前校正和波前重构。因此对系统的技术要求非常高,造成系统复杂,成本昂贵。基于哈特曼-夏克波前探测的图像解卷积处理就是“事后”处理的一种,它省去了波前校正环节,使常规的自适应光学系统得到简化,降低了系统成本。其基本原理为对瞬时波面进行短时间曝光探测,同时记录相应的短时间曝光图像,来进行解卷积处理。将增量维纳滤波法应用于基于哈特曼-夏克波前探测的解卷积中,并对室内模拟点源情况下的三组畸变光斑图像数据进行了解卷积恢复处理。结果表明,将增量维纳滤波法应用于基于波前探测的解卷积是完全可行的,在室内模拟点源情况下,恢复的图像可以达到衍射极限分辨力。与维纳滤波相比,它扩展了噪声抑制因子的选取范围,在噪声抑制因子选取不正确的情况下,仍能得到比维纳滤波更好的结果。     

An exploratory study of the reactive Ni-GaAs (1 1 0) interface

Soft X-ray photoemission spectroscopy measurements have been carried out on cleaved n-type GaAs (1 1 0) surfaces covered with Ni overlayers ranging in thickness from 0.05 to 53 Å. The results of these room temperature measurements show that we have band bending effects occurring in conjunction with strong interfacial chemical reactions. Deconvolution of the Ga 3d core line into substrate and metallic components shows dissolution of the substrate at the interface with Ga diffusing into the surface of the metal overlayer for the intermediate coverages (1–15 Å). Observation of the As 3d core level shows out-diffusion of As to the surface over the entire Ni coverage range. Using this deconvolution scheme we are able to follow the band bending of the Schottky barrier formed here up to the 8 Å coverage. The Schottky barrier height is 1.0 ± 0.1 eV for this overlayer thickness.     

Analysis of XPS spectra of Fe and Fe ions in oxide materials

Samples of the iron oxides Fe_0.94O, Fe₃O₄, Fe₂O₃, and Fe₂SiO₄ were prepared by high temperature equilibration in controlled gas atmospheres. The samples were fractured in vacuum and high resolution XPS spectra of the fractured surfaces were measured. The peak positions and peak shape parameters of Fe 3p for Fe²⁺ and Fe³⁺ were derived from the Fe 3p XPS spectra of the standard samples of 2FeO·SiO₂ and Fe₂O₃, respectively. Using these parameters, the Fe 3p peaks of Fe₃O₄ and Fe_1−yO are analysed. The results indicate that high resolution XPS techniques can be used to determine the Fe²⁺/Fe³⁺ ratios in metal oxides. The technique has the potential for application to other transition metal oxide systems.     

An X-ray photoelectron spectroscopic study of some nickel (II) amine complexes

The X-ray photoelectron (XPS) spectra of several Ni(II) diamine complexes, a Ni(II) triamine complex and several Ni(II) diimine complexes are reported in terms of the Ni 2p, Cl 2p and N 1s core level binding energies. Theoretical models are presented to account for the quadratic relationship observed between XPS Ni 2p_{$\text{3}\text{2}$}binding energy shifts and the N-H infrared rocking frequency as related to these complexes. On the basis of this correlation, it is possible to discriminate between free counter-ions coordinated to the metal or involved in cluster-ion formation. Individual XPS data for these complexes are interpreted in terms of the intrinsic nature of the metal environment and an attempt is made to calculate the metal and donor nitrogen atomic charges for selected complexes.     

Further performance tests of a picosecond X-ray and laser induced streak camera system with fast scintillation materials

A further performance test of a fluorescence lifetime spectroscopy system with 2D-photon counting based on a streak camera was carried out. This system offers three different excitation light sources. It includes a pulsable X-ray tube with 40 kV and a laser. A pulsed UV-LED has been installed for sample excitation. The table top measurement device has a temporal resolution down to 20 ps after deconvolution and fluorescence/scintillation decay times can be evaluated in a wavelength selectable range between 200 and 850 nm. Cesium fluoride, cesium chloride and cesium bromide as fast core-valence luminescence emitters with a relatively low light yield were used in the presented performance test. The easily manageable system is able to detect low light levels and to determine the decay times in agreement with literature in most instances. Beside mentioned scintillators PbCO₃, SrI₂ and Lu₂SiO₅:Ce as a high light yield material were also measured.