· Table of Contents
· Methods & Instrumentation
Depth-Profile Analysis of Elements by Glow Discharge Optical Emmission SpecrometryJohn M. Long
Lecturer, School of Engineering and Technology
Deakin University, Victoria, Australia
Glow-discharge optical emission spectrometry (GD-OES) is a powerful tool for the rapid analysis of elements in the surface of solids. One may employ GD-OES to determine quantitatively the bulk concentration of elements in a sample. With further calibration, one may also obtain elemental concentrations as a function of depth into the sample. This allows depth profiling on a host of advanced materials: treated metals, coated metals and other materials, multi-layers, painted surfaces, hard samples coated with polymers, thin films, and many others.
A consortium of institutions in Victoria, led by Deakin University, has purchased a new glow-discharge optical emission spectrometer. This instrument has the ability to perform elemental depth profiling on a wide range of materials. This technique, the first of its kind in Australia, is of particular interest to those working on metals, ceramics, glasses, coatings, semi-conductors, and multi-layers. We present here an overview of depth profiling by GD-OES and some examples of its use.
Keywords: GD-OES, glow discharge, depth profiling, surface analysis, emission spectrometry
Surface engineering in solids has become an increasingly important field in materials science and engineering. Surface treatments have advanced from classic treatments such as galvanising and case hardening, to diffusion heat treatment and chemical vapour deposition of hard coatings. New technologies have made it possible to modify the chemistry of a material at the extreme outer surface, while leaving the interior unchanged. Surface treatments and coatings have a wide variety of applications, such as increasing wear resistance in cutting tools and other components, providing protection from corrosion and other environmental conditions, producing sophisticated electronic devices, and improving the appearance of common products, like metal watch bands. Common products that contain thin coatings or multi-layers on the surface vary widely from drill bits to painted automobile panels to semiconductor devices to concrete roofing tiles to glass bottles. In such applications, knowledge of coating or layer compositions, thicknesses, and sharpness of transition from one layer to another and to the substrate is critical. All these factors influence the overall performance of the material.
For all these applications, it is important to be able to identify and quantify the distribution of elements in the surface. Such data is necessary for process development, research, and routine quality control. Until recently, elemental determination on surfaces has been difficult. Techniques have been indirect, non-quantitative, or highly time-consuming; and often destructive to the test samples. For instance, one technique used in the steel industry is cutting a thin slice off a surface, dissolving the sample in acid, and employing techniques such as atomic absorption or ICP spectroscopy to quantify the elements present.
Glow-discharge optical emission spectrometry (GD-OES) had proven to be a powerful tool for the rapid analysis of elements in the surface of solids (Payling et al. 1997, Payling 1998). With appropriate calibration one may employ GD-OES to quickly determine quantitatively the bulk concentration of elements in a sample. More recent advances allow the calibration of sputter rate. Once the sputter rate is calibrated, one may not only determine elemental concentrations in the whole sample, but also elemental concentrations as a function of depth into the sample.
Emission spectrometry is a well-known technique for analysing an unknown material for the elements present in the sample. When a material is heated sufficiently, it will emit visible light in a discrete spectrum characteristic of the elements in the sample. Each element has its unique atomic emission spectrum (both for visible light and for x-rays). Payling and Larkins (2000) have recently published a new catalogue of atomic spectra for all the elements. Thus if one obtains an atomic spectrum from an unknown sample, he may identify the elements in the sample by comparing the spectrum with those of the elements in the spectral database.
There are three common means of heating the sample to produce the optical emission. The first is to apply a high electrical voltage across the sample, which heats the sample in a spark discharge (arc/spark instruments). Another is to dissolve the sample in acid, and "burn" the solution in an argon plasma (ICP). The third is to sputter the sample surface with argon, and excite the sputtered atoms in the argon plasma (glow discharge). The latter is the principle behind the new Deakin spectrometer.
|Fig 1: Schematic of the glow-discharge lamp and associated process.|
Figure 1 shows a schematic of the glow-discharge lamp. The sample forms the cathode and a thin (2-8 mm diameter) metal tube forms the anode. A small O-ring separates the anode from the cathode. High-purity argon is pumped into the anode chamber. A high voltage (DC or RF) between sample and anode ionises the argon to produce a glow discharge. The excited argon ions bombard the sample and cause uniform sputtering of the sample surface. The sputtered atoms from the sample enter the plasma, and they are excited by collisions with electrons or other argon atoms. Those excited atoms from the sample eventually have energy decays to stable states, and the optical emission forms the atomic spectrum from the sample.
The light emitted by the sample, which is characteristic of the atoms sputtered from the surface, passes through a window into a standard polychromator to be dispersed off a curved, holographic diffraction grating. A series of fixed photomultiplier tubes detect and record specific wavelengths of visible light, corresponding to selected elements.
The applied voltage may be either DC or RF. DC lamps work on conductive samples only. RF lamps are a more recent development in GD-OES (Marcus 1996), and enable the analysis of non-conductive materials. A typical bias voltage is -1000 volts from anode to cathode.
With appropriate calibration one may employ GD-OES to determine quantitatively the bulk concentration of elements in a sample. This technique is very common. The signals from the detectors are integrated over the analysis time (1 minute or less for a bulk metal) to yield bulk elemental concentrations. More recent advances in techniques and instrumentation allow the calibration of sputter rate. The glow discharge sputters the sample uniformly, producing a shallow crater in the sample. The plasma sputters a measurable thickness of sample in a given time, on the order of microns per minute in metals. The detectors acquire data very fast, up to 2000 measurements per second for each photomultiplier. Thus a plot of detector intensity versus time can be transformed through sputter-rate calibrations to a depth profile of elemental concentration versus depth into the sample.
Depending on the standards used and the quality of the calibration, depth may be resolved to 100 nm or less. This allows depth profiling (detailed surface analysis) on a host of advanced materials: heat-treated metals, diffusion-treated metals, coated metals and other materials, multi-layers, painted surfaces, hard samples coated with polymers, thin films, and many others. If an RF glow-discharge source is used, quantitative depth profiling (QDP) on ceramics, polymers, and non-conductive coatings on conductors is possible. QDP analysis by means of an RF source is the recent innovation in GD-OES, and is still under development. QDP by a DC source is more established and better understood.
Applications of depth profiling by GD-OES include the analysis of galvanised steel, coated tool steels, anodised aluminium, and painted surfaces (Payling 1998, Shimizu et al. 1999, Weiss and Marshall 1997).
The Deakin instrument is a Leco GDS-850A GD-OES spectrometer, equipped with both a Grimm-type DC lamp for conductive samples, and an RF lamp for the analysis of non-conductors. The anode has a diameter of 4 mm, for a sampling area of 12.5 mm2. The spectrometer is equipped with dual Rowland circles, having curved diffraction gratings of 1800 lines/mm and 3600 lines/mm, respectively, for a spectral range of 120-800 nm. There are 32 detectors, with the ability to detect 29 elements, including most standard metals, oxygen, nitrogen, boron, carbon, hydrogen, chlorine, some rare earths, and others.
The instrument was calibrated prior to shipping against a large number of certified and non-certified reference materials. Sputter rates are calibrated against a database developed and provided by Leco.
The spectrometer is corrected daily for drift in the optics by means of a smaller series of standards - a high-concentration standard and a low-concentration standard for each calibration curve stored in the particular test method. The optical alignment is also checked daily against two iron emission lines, one for each circle.
Three samples were analysed by GD-OES:
Sample No. 2 was prepared by nitriding in a fluidised bed furnace by standard methods (Reynoldson 1993). Sample No. 3 was prepared by coating by means of cathodic arc vapour deposition (Doyle et al. 1998). The first sample was analysed for bulk elemental concentration, with an applied voltage of 1050 volts, and a plasma current of 45 mA. The data were collected six times from various parts of the sample. This sample required minimal preparation, only some light grinding on 220-grit SiC paper to smooth out the surface. The QDP results from samples 2 and 3 were typical of their respective surfaces.
Table 1 shows the results of the bulk analysis of sample No. 1. After the instrument was corrected for drift, the total analysis time for this sample was around 5 minutes. In the bulk analysis, prior to collecting the data, the instrument sputtered on the surface for 1 minute to remove any surface impurities or fingerprints. The consistency in the results is excellent. The test also measured several other elements whose concentrations were 0.007 wt% or less.
|Table 1: Bulk Quantitative Elemental Concentrations (wt%) for Sample No. 1|
Figures 2 (qualitative raw data) and 3 (quantitative) show the depth-profiling results for sample No. 2. The elemental concentrations for iron, chromium, carbon, and oxygen are shown, although the test itself measured many other elements whose concentrations were very low. In 28 minutes of analysis time, the lamp sputtered about 60 microns into the sample. The results show that the surface layer affected by the diffusion treatment was 10 microns thick. The nitrogen penetration was non-uniform, reaching a maximum concentration 4 microns below the surface, decreasing very gradually to a depth of 5 microns, and then decreasing more rapidly. The iron concentration followed the nitrogen behaviour in a reverse manner. Also detected was significant oxygen diffusion in the first half micron of the sample. QDP also detected a very thin layer of carbon on the surface, presumably from handling the specimen. The elemental concentrations below a depth of 25 microns were characteristic of a normal tool steel.
|Fig 2: Qualitative depth-profiling results (raw data) for sample No. 2.|
|Fig 3: Quantitative depth-profiling results for sample No. 2.|
The QDP results for sample No. 3 are in figure 4. The instrument sputtered to a depth of 11 microns in 5 minutes of data collection. Again more elements were analysed than the 6 shown in the figure. The overall coating is 1.1 microns thick, in two distinct layers. The very first 0.3 microns is a non-uniform layer of TiAlN. The aluminium concentration drops to zero by 0.5 microns. The remainder of the coating is layer of non-uniform TiN. The transition between coating and substrate occurs between 1 micron and 1.8 microns. The composition of the substrate matches that of a high-alloy tool steel.
|Fig 4: Quantitative depth-profiling results for sample No. 3.|
The results indicate that useful depth profiling information can be obtained in as little as a few minutes in a metallic sample. Sputtering into a ceramic, such as glass, would be expected to take much longer. Qualitative depth profiling is also possible without the need to run numerous calibration standards and produce calibration curves. Qualitative analysis is still useful, especially in industry because of the consistency of the test methods. As long as the tests are performed with the same operating parameters, useful comparisons can be made among groups of similar samples. In many cases this is sufficient for quality control or for an initial indication of surface characteristics.
Deakin's initial applications for this instrument include surface analysis of heat-treated and coated tool steels, galvanised and similar sheet steels, analysis of corrosion, ceramic-coated aluminium alloys, thermally-sprayed ceramic coatings, multi-layers on semiconductors, and of course general bulk analysis of metal alloys. In the future more analytical detectors will be added to increase the number of elements it can analyse, and its applications. This technique also complements other surface analytical techniques such as Auger and X-ray photoelectron spectroscopies.
In partnership with several Victorian universities and industry, Deakin University has installed a new Leco glow-discharge optical emission spectrometer. It is believed that this instrument is the first GD-OES instrument in Australia capable of performing quantitative, elemental depth profiling. It has proved useful in the bulk analysis of steel, and in depth profile analysis on nitrided and hard-coated tool steels. This instrument shows much promise in the surface analysis of many different and industrially important materials.
The Deakin spectrometer was funded in part by the Research Infrastructure Equipment and Facilities Scheme of the Australian Research Council. The author thanks R. Reynoldson and E.D. Doyle for providing the QDP test samples. He is also grateful to R. Payling, P. Hunault, and F. Schafer for several useful discussions.
|© AINDT , created by NDT.net|||Home| |Top||