辉光放电光谱法在深度分析上的应用现状

本文简单地介绍了辉光放电光谱法（GD-OES）的基本原理。分别对深度分析的定量方式、放电方式、应用领域和相关标准进行了详细地阐述。重点描述了商品化仪器中使用的SIMR深度分析定量方法。分别对三种放电方式（如直流、射频和脉冲）在深度分析中的特点进行了介绍。综述了GD-OES在金属镀层、复杂涂镀层、纳米级薄膜和样品制备领域的具体应用。最后,介绍了GD-OES在深度分析方面的标准     

Radio frequency glow discharge optical emission spectroscopy: a new weapon in the depth profiling arsenal

While the array of analytical methods routinely applied for depth profile analysis was fairly static over the decades of the 1980s and 1990s, there appears to be an emerging technique that has a number of very positive and complementary attributes, and warrants serious consideration by the thin film community. Radio frequency glow discharge optical emission spectroscopy (rf-GD-OES) is a technique that provides depth-resolved elemental composition information on a wide variety of sample types. In a manner very much like most depth profiling methods, the rf-GD plasma utilizes an ion sputtering step to ablate sample material in a layer-by-layer fashion. Different from the more commonly applied methods, the device operates at elevated pressures [2-10 Torr Ar (266-1,330 Pa)] and has the inherent capability of sputtering electrically insulating materials directly, without any auxiliary means of charge compensation. In addition, sputtering rates on the order of 1 micro m/min provide rapid analysis, with depth resolving powers that are comparable to high-vacuum sputtering methods. Three examples of the use of the rf-GD-OES method are presented as examples of its analytical potential: (1) boron-implanted silicon wafer, (2) a barrier-type alumina film, and (3) a porous-type alumina film. It is believed that the method holds a great deal of promise as part of the arsenal of weapons in the thin films laboratory.     

Measuring metal substrate roughness with glow discharge optical emission spectrometry

Although there are many methods to evaluate metal surface roughness, it is difficult to detect the substrate roughness of coated samples. Depth profile analysis (DPA) is proposed as a new method for the substrate roughness measurement, and glow discharge optical emission spectrometry (GDOES) becomes a candidate in substrate roughness measuring instruments. With this method, the typical roughness parameters R_a (arithmetic mean of the absolute deviation values of the roughness profile) and R_y (maximum value of consecutive peak and valley heights of the roughness profile along the sampling length) can be easily obtained. The principle of this method is discussed and the formulasof the principle are deducted in this paper. Electrodeposited zinc coating on copper substrate specimens is used as an example to explain the measuring process. Copyright © 2007 John Wiley & Sons, Ltd.     

Quantitative depth profile analysis of metallic coatings by pulsed radiofrequency glow discharge optical emission spectrometry

In recent years particular effort is being devoted towards the development of pulsed GDs because this powering operation mode could offer important analytical advantages. However, the capabilities of radiofrequency (rf) powered glow discharge (GD) in pulsed mode coupled to optical emission spectrometry (OES) for real depth profile quantification has not been demonstrated yet. Therefore, the first part of this work is focussed on assessing the expected advantages of the pulsed GD mode, in comparison with its continuous mode counterpart, in terms of analytical emission intensities and emission yield parameters. Then, the capability of pulsed rf-GD-OES for determination of thickness and compositional depth profiles is demonstrated by resorting to a simple multi-matrix calibration procedure. A rf forward power of 50 W, a pressure of 600 Pa, 1000 Hz pulse frequency and 50% duty cycle were selected. The quantification procedure used was validated by analysing conductive layers of thicknesses ranging from a few tens of nanometer up to about 20 μm and varied compositions (hot-dipped zinc, galvanneal, back contact of thin film photovoltaic solar cells and tinplates).     

Wavelength table of chromium emission lines in argon glow discharge optical emission spectrometry

A wavelength table of chromium lines emitted from an argon glow discharge plasma, which comprises 2049 atomic and ionic emission lines in the wavelength range of 200–440 nm, is presented. The relative intensities are rather different from the data of published wavelength tables based on arc-excited and spark-excited spectra. Emission lines of Ar, Ti, V, Fe, Ni, and Cu in the neighborhood of the prominent Cr emission lines are also compiled as a table. These tables could be employed for the analytical applications in glow discharge optical emission spectrometry. All of the data are presented as Supplementary Electronic Material. Recieved: 22 December 1999 / Revised: 25 February 2000 / Accepted: 25 February 2000     

Wavelength table of chromium emission lines in argon glow discharge optical emission spectrometry

A wavelength table of chromium lines emitted from an argon glow discharge plasma, which comprises 2049 atomic and ionic emission lines in the wavelength range of 200–440 nm, is presented. The relative intensities are rather different from the data of published wavelength tables based on arc-excited and spark-excited spectra. Emission lines of Ar, Ti, V, Fe, Ni, and Cu in the neighborhood of the prominent Cr emission lines are also compiled as a table. These tables could be employed for the analytical applications in glow discharge optical emission spectrometry. All of the data are presented as Supplementary Electronic Material.     

Quantitative nitrogen analysis by Auger electron spectrometry and glow discharge optical emission spectrometry

Nitrides of refractory metals are investigated as diffusion barriers for Cu metallization. The composition, thermal stability and inter diffusion in layered systems are characterized by depth profile analysis. For the quantification of depth profiles determination of sensitivity factors is essential. For nitrogen and other light elements matrix specific standards are often not available and compound standards are used for calibration. We have investigated the systems Ta–N and Ta–Si–N and for comparison Cr–N by means of Auger electron spectrometry (AES) and glow discharge optical emission spectrometry (GDOES). A non-linear calibration curve for the N/Cr intensity ratio was observed with GDOES in the Cr–N-system, probably caused by self-absorption of the Cr line.     

Rotational temperature of nitrogen glow discharge obtained by optical emission spectroscopy

Measurements of rotational temperature as low as several hundred Kelvin have been measured using optical emission spectroscopy (OES) in nitrogen direct current (DC) glow discharge. The strongest band of the first negative system of nitrogen was chosen to deduce the rotational temperature at four different positions in nitrogen DC glow discharge, the back of cathode; cathode sheath; positive column; and anode glow. In positive column the rotational temperature increased apparently with the increasing discharge voltage from 500 to 1000 V when the pressure was 10 Pa. But with pressure of 20 Pa the rotational temperature in positive column increased slightly with the increase of discharge voltage. On the contrary, the rotational temperature in cathode sheath took reverse tendencies when the discharge voltage varies from 500 to 1000 V. As regard the anode glow, the rotational temperature at 10 Pa decreased with the increase of discharge voltage, but that at pressure of 20 Pa increased. We attribute the different tendencies of the rotational temperature to the different discharge statues at different pressures. When the discharge voltage varies from 500 to 1100 V, the discharge with pressure of 10 Pa is normal glow and that with 20 Pa is abnormal glow.     

Analytical performance of high-voltage neon plasma in glow discharge optical emission spectrometry

In a high-voltage Ne glow discharge plasma (Ne-GDP), calibration factors as well as the limit of determination were compared between atomic resonance lines and singly-ionized lines of copper and aluminium in optical emission spectrometry. These elements have intense ionic lines which are excited by resonance charge-transfer collisions of Ne ions. The ionic lines gave better detection sensitivity in the Ne-GDP, whereas the atomic resonance lines were commonly employed as analytical lines in the other plasma sources such as Ar-GDP and ICP. The limit of determination was 1.3 × 10^–3 mass % for the Cu II 248.58 nm line and 1.0 × 10^–3 mass % Al for the Al II 358.66 nm line at a discharge parameter of 1.60 kV/36 mA. Received: 22 January 1999 / Revised: 15 March 1999 / Accepted: 20 March 1999     

Analytical performance of high-voltage neon plasma in glow discharge optical emission spectrometry

In a high-voltage Ne glow discharge plasma (Ne-GDP), calibration factors as well as the limit of determination were compared between atomic resonance lines and singly-ionized lines of copper and aluminium in optical emission spectrometry. These elements have intense ionic lines which are excited by resonance charge-transfer collisions of Ne ions. The ionic lines gave better detection sensitivity in the Ne-GDP, whereas the atomic resonance lines were commonly employed as analytical lines in the other plasma sources such as Ar-GDP and ICP. The limit of determination was 1.3 × 10^–3 mass % for the Cu II 248.58 nm line and 1.0 × 10^–3 mass % Al for the Al II 358.66 nm line at a discharge parameter of 1.60 kV/36 mA.     

Application of glow discharge optical emission spectroscopy (GDOES) to the analysis of PVD- and CVD-layers

Summary PVD- and CVD-layers were analysed by GDOES using a homemade laboratory device and the commercial Spectrumat 1000. The presented examples demonstrate the many possibilities of GDOES and give information about the further development of this analytical technique.     

Multilayer thin-film coatings based on chromium nitride and aluminum nitride: Comparative depth profiling by secondary ion mass spectrometry and glow-discharge optical emission spectrometry

Round-robin characterization is reported on the sputter depth profiling of CrN/AlN multilayer thin-film coatings on nickel alloy by secondary ion mass spectrometry (SIMS) and glow-discharge optical emission spectrometry (GD-OES). It is demonstrated that a CAMECA SIMS 4550 Depth Profiler operated with 3 keV O ₂ ⁺ primary ions provides the best depth resolution and sensitivity. The key factor is sample rotation, which suppresses the negative influence of the surface topography (initial and ion-induced) on the depth profile characteristics.     

Modeling of glow discharge optical emission spectrometry: Calculation of the argon atomic optical emission spectrum

The optical emission spectrum of the argon atomic lines in a glow discharge is calculated, using a collisional–radiative model for argon, which was recently developed (A. Bogaerts et al., Collisional–radiative model for an argon glow discharge, J. Appl. Phys., vol. 84, No 1, 1998). It is shown that the lines corresponding to 4p→4s transitions clearly dominate the spectrum. They are, however, not responsible for the characteristic visible light in the glow discharge, because they are lying between 700 and 1000 nm, which is mainly in the near infrared. The characteristic blue light of the glow discharge is caused by the lines corresponding to 5p→4s transitions (lying in the blue–violet part of the spectrum). Beside these two most important line groups (the so-called `red' and `blue' lines) a large number of other lines are present, making the entire argon spectrum quite complex. The calculated spectrum is compared with experimental spectra from the literature, and excellent qualitative agreement is obtained.The calculated spatial distributions of optical emission lines originating from low excited levels (i.e., 4p, 3d, 5s, 5p, 4d, 6s) show a maximum in the cathode glow, caused by fast argon ion and atom impact excitation, to these levels, and a second maximum in the beginning of the negative glow, due to electron impact excitation. The maximum in the cathode glow is very pronounced for lines originating from the 4p levels, which is in agreement with experimental observations. The higher excited levels are not populated by fast argon ion and atom impact excitation but only by electron impact excitation; hence, lines originating from these levels exhibit only a maximum in the beginning of the negative glow.     

Hydrogen effects on copper,zinc and nickel atomic emission lines in argon radiofrequency glow discharge optical emission spectrometry

The effect of hydrogen (0.5%, 1% and 10% v/v) added to the argon plasma gas on the emission spectra of selected atomic lines for copper, zinc and nickel has been studied by radiofrequency glow discharge optical emission spectrometry (rf-GD-OES). Conductive homogeneous samples containing different concentrations of the elements under study in different matrices have been investigated. Results show different trends of the emission intensity lines with increasing hydrogen concentration in the rf-GD, depending on the line characteristics. In most cases, the emission yields of the lines under study did not change or increased when hydrogen was added to the discharge (no decreases were observed). The emission yields of certain lines showed much higher increases than other lines of the same element (for example, lines 213.86 nm of Zn and 231.10 nm of Ni). Our experiments indicate that such notorious increases could be related with the possible decrease of the self-absorption when hydrogen is added to the discharge. Overall, the results obtained for the emission yield changes of certain lines of a given element in different matrices (with different analyte content) showed that while for resonance emission lines very notorious increases are observed, the values for non-resonance lines do not change significantly (specially if the matrices employed are similar).     

Improved voltage transfer coefficients for nonconductive materials in radiofrequency glow discharge optical emission spectrometry

In radiofrequency glow discharge emission spectrometry (RF-GDOES), the excitation voltage used to create the plasma is applied to the back or front end of the sample to be analyzed. In this paper we focus on back-applied voltage systems (a configuration that represents about half of the instruments available on the market), and on applied voltage problems (the power coupling efficiency and materials analysis are beyond the scope of this study). In the RF-GDOES of nonconductive samples, a voltage drop develops inside the material. The voltage transfer coefficient is defined as the ratio between the peak voltage in front of the sample (facing the plasma) and the peak voltage applied to the back of the sample. In this work, we show that it is possible to increase the voltage transfer coefficient by increasing the capacitance of the sample. The capacitance of a given nonconductive material depends on its surface, its thickness and its permittivity. Increasing the voltage transfer coefficient permits higher power deposition in the plasma. This study is based on an electrical equivalent circuit for the discharge device, which takes into account the sample and reactor capacitances as well as the voltage probes used for the measurements. This circuit, when modeled by a commercial electrical circuit simulator, gives the voltage transfer coefficient as a function of the sample capacitance. Different approaches to increasing the sample capacitance and their influence on the voltage transfer coefficient are presented and related to the 750.4 nm argon line intensity, which is correlated to the electron density.     

Comparison of modeling calculations with experimental results for rf glow discharge optical emission spectrometry

A model developed recently for a radio-frequency glow discharge, is applied to experimental Grimm-type discharge conditions, to check the validity of the model calculations. The calculated electrical characteristics (rf voltage, d.c. bias voltage, electrical power, peak-to-peak current, as well as the time-profiles of voltage and current), and the calculated erosion rates are compared with the measured values for an rf frequency of 3.5 MHz. The electrical characteristics are found to be in fairly good agreement. The calculated and measured erosion rates show larger discrepancies. Compared to the d.c. Grimm-type glow discharge, where similar quantities were compared and were found in excellent agreement, the agreement is less satisfactory in the rf discharge. This illustrates that the rf discharge is much more complicated than a d.c. discharge, and that more fundamental studies are required.     

Emission characteristics of Cu in Dc voltage modulation applied glow discharge optical emission spectrometry.

In order to obtain the depth profile of a thin film, we investigated the emission characteristics of a voltage modulation glow discharge to optimize the modulation parameters (modulation voltage, offset voltage, and modulation frequency). In this study, a phase-sensitive detection method with a lock-in amplifier to the modulation technique led to a higher sensitivity and a larger signal-to-noise ratio in the emission analysis compared to the normal dc amplification method. Upon increasing the maximum voltage, the emission intensity of the Cu atomic line (CuI 239.34 nm) increased linearly at a modulation voltage of 400 V and an offset voltage of 300 V. On the other hand, the emission intensity was gradually reduced when a modulation frequency increased. It is advantageous for surface analysis that the voltage modulation technique gives a lower sputtering rate rather than the conventional dc discharge.     

Time-resolved measurement of emission profiles in pulsed radiofrequency glow discharge optical emission spectroscopy: Investigation of the pre-peak

Radiofrequency glow discharge coupled to optical emission spectroscopy has been used in pulsed mode in order to perform a detailed study of the measured temporal emission profiles for a wide range of copper transitions. Special attention has been paid to the early emission peak (or so-called pre-peak), observed at the beginning of the emission pulse profile. The effects of the important pulse parameters such as frequency, duty cycle, pulse width and power-off time, have been studied upon the Cu pulse emission profiles. The influence of discharge parameters, such as pressure and power, was studied as well.     

Comparison of modeling calculations with experimental results for direct current glow discharge optical emission spectrometry

A comparison is made between numerical modeling and experimental results for the electrical characteristics, the erosion rates and the optical emission intensities of various argon and copper lines in a direct current glow discharge, to verify the model calculations and to illustrate some features and limitations of the model. In order to reach good agreement with the current–voltage characteristics, the gas temperature, which was treated as an adjustable parameter, was assumed to increase slightly as a function of voltage and pressure. This assumption is in accordance with theoretical predictions and experimental observations in the literature. The erosion rates and optical emission intensities, calculated as a function of voltage and pressure, were also found to be in reasonable agreement with the experimental data. However, it appeared that still better agreement with the measured data could be reached when the gas temperature was assumed to be constant as a function of voltage. This illustrates that the effect of voltage cannot yet be completely correctly predicted for both the electrical current and the erosion rates and optical emission intensities at the same time, and that, therefore, the glow discharge behavior is not yet perfectly described in the model. This is not unexpected in view of the complexity of the model calculations and the uncertainties of some input data. However, in general, the agreement between model results and experimental data is satisfactory, so that it can be concluded that the model gives already a realistic picture of the direct current glow discharge.     

Lithium-fusion sample preparation method for radio frequency glow discharge optical emission spectroscopy (rf-GD-OES): analysis of coal ash specimens

The use of lithium-fusion sample preparation methods for the analysis of powder-form specimens by radio frequency glow discharge optical emission spectroscopy (rf-GD-OES) has been evaluated for the test case of coal ash samples. The development of an elemental analysis technique for chemically complex coal ashes presents a challenge in terms of having simple sample preparation, providing accurate results, and minimizing analysis time. The adopted sample preparation procedure for the coal specimens involved a standard lithium fusion methodology. This procedure circumvents many problems associated with the common compaction methods employed for GD analysis of inorganic powders. It was found that coal ashes prepared as glass disks using a mixture of lithium compounds as the host matrix and analyzed by rf-GD-OES provide good plasma stabilization, with analyte optical emission signals stabilizing in ∼1 min and exhibiting ∼2% RSD variations for sputtering times of up to 10 min thereafter. The evaluation of discharge operating parameters and the assessment of approaches to quantitative analysis were also investigated. Discharge parameters of an rf power of 30 W and an Ar discharge gas pressure of 4 Torr yielded rapid signal stabilization and optimized S/B ratios. Sample-to-sample precision of better than 7% RSD was achieved for repetitive samplings (in the same sample locations) for species present at the parts-per-million level in the sample. Limits of quantification could not be adequately evaluated as the levels of the target analytes in the fusion components (i.e. lithium compounds) were above the method detection limit. It is believed that the general methodology holds promise for rf-GD analysis of a range of inorganic solids in powder form.