|
Please click the images for a enlarged view.
|
Glow Discharge Optical Emission Spectroscopy (GDOES)
Introduction and Overview
|
|
1. General information
|
Back to top |
Glow Discharge Optical Emission Spectroscopy (GDOES) provides rapid analysis of the atomic composition of solids both conducting and non-conducting.
GDOES was introduced by W. Grimm in the late 1960s. Its first applications were the determination of the bulk atomic composition of metals. Only a few years later Berneron and his co-workers from the IRSID research laboratories in France demonstrated the surface analysis capabilities of the GDOES. One of the first centres of interest were the study of oxidation processes of steel. In its early years GDOES received much interest from a large group of analysts and much expectation were put on this technique. However, it did not find an important market immediately. In fact for most analytical problems concerning bulk analysis of metallic species it was outperformed by spark, not always in terms of analytical performance but often by price and ease of use. A group of industrial researchers and analysts kept their interest in this technique and developed and improved its capabilities. Performing quantitative depth profiling and the introduction of the RF excitation allowing the analysis of non-conducting coatings and films presented a major break through. The interest in this new/old technique has considerably increased in the last five years. This revived interest is partly due to the improvement of the technical performance, but mainly due to the increasing industrial need of characterising surface treatments and coatings. |
|
1.1. Neighbouring techniques
|
Back to top |
The field of application of the GDOES technique is extremely wide, ranging from elemental bulk analyses to analyses of thin films. The range of neighbouring techniques is consequently large. In terms of bulk analysis, GDOES has to compete with Spark Emission Spectroscopy. The market has already answered the question of superiority. In most cases, a well performing spark system will supply better detection limits and often accuracy, but GDOES often gives equivalent answers. Employing GDOES for bulk analyses only has some justification in cases where spark systems notoriously have had problems, grey cast iron for example. Although GDOES systems are generally more expensive to produce than spark systems, mainly because of the vacuum technology involved, GDOES sometimes is economically the first choice in particular when many different materials must be analysed with the same instrument. In this case, GDOES is advantageous because of the reduced number of spectral lines needed to cover the complete field of application.
Providing detection limits of generally 1 ppm to 10 ppm GDOES is an honest solution for bulk analysis applications in particular when depth profiling is requested as well.
Coating of bulk material, in particular steel sheets is a well established technique for corrosion protection. Inexpensive and non-destructive X-ray technology is often industry's first choice. In cases, when only the coating thickness needs to be determined, this is an appropriate choice, except when the coating contains the same elements as the substrate as in the case for Zn/Fe coating on steel sheets. The needs of the industry, however, goes more and more beyond a simple determination of the layer thickness, as coating structure and composition become more and more relevant to the quality of the final product. Quantitative depth profiles, as measured by GDOES, is the preferred method to ensure the quality of the coated product.
Surface analysis has long been a domain for physicists employing sophisticated tools as Secondary Ion Mass Spectrometers, Auger Electron Spectrometers and others. These techniques are certainly very powerful, but they do have several major drawbacks for their introduction into industry: price, sample throughput, ease of result interpretation etc. Glow Discharge can, in a majority of industrial surface analytical applications, replace these techniques. It goes beyond the scope of this text to prove and explain this, but it has been shown that GDOES can produce results comparable in depth resolution to SIMS, but at a definitely lower cost. If you look at GDOES from the spark point of view it may appear expensive, but from the surface analysis point of view it is a powerful and inexpensive tool, allowing high sample throughput and low cost of ownership. In order to avoid misunderstanding and disappointment, it should be stated that GDOES is not a microanalysis tool, the analysed area is of some mm² of size. |
|
2. Basic Processes in the Glow Discharge Source
|
Back to top |
2.1. Basic design
|
Back to top |
| A hollow anode glow discharge chamber serves as the light source for the spectrometer. The two electrodes of the discharge cavity are designed as a tubular grounded copper tube and the flat sample to be analysed. In fact, the sample closes the discharge cavity: Sample introduction is therefore extremely simple. The electrical power is supplied directly to the sample. In the case of a DC discharge a regulated negative voltage of some hundreds of Volts is supplied. In the case of an RF discharge a RF power of usually less then 30 Watts is supplied to the sample. The carrier gas within the discharge cavity is in most cases Argon. The Argon pressure in the cavity is maintained at about 5 hPascal. An Argon flow of about 0.2 l/min sweeps the cavity to ensure a clean Argon atmosphere during the analysis process. |
Figure 1: |
The analytical figures of merit of the GDOES technique are closely linked to the physics of the glow discharge.
A plasma is created within the tubular grounded anode. The electrical power to ignit and maintain the plasma is fed through the sample, which acts as the cathode. Argon ions and free electrons are generated in the plasma (1). The potential difference between the plasma and the sample accelerate the Argon ions towards the sample (2). The ions are bombarding the sample surface with an average kinetic energy of 100 eV (3). This bombardment creates a sputtering process, atoms from the sample surface leave the sample lattice. In addition to the atoms some free electrons, called secondary electrons, are generated during the sputtering process as well (4). The secondary electrons are accelerated towards the plasma and maintain the plasma activity. The atoms diffuse towards the plasma. Here they may be excited (7) or ionised through collisions with free electrons, Argon ions or meta-stable Argon atoms (5). |
Figure 2: |
The excited analyte atoms subsequently relax to the energetic ground state by emission of characteristic photons (6). The characteristic photon spectrum is observed through the observation window.
In fact, the analyte atoms, after being sputtered from the surface, are diluted in the Argon atmosphere of the plasma and excited as single atoms. It is this separation of sputtering and excitation as well as the dilution in the Argon plasma, that makes GDOES analyses widely independent of matrix effects. This finally allows the quantification of depth profiles, where analyses of different base materials, in the coating and the core, are to be performed with only one general calibration curve. |
|
|
3. Applications
|
Back to top |
| GDOES finds its application mainly in Quantitative Depth Profiling or Surface and Interface analysis tasks. The field of possible applications is extremely wide. Today, GDOES is well introduced in the steel manufacturing and the car manufacturing industry. Also the semiconductor industry is of upcoming importance. Different industrial sectors, research centres and universities discovered GDOES as a valuable tool for characterisation of surface and interfaces in order to optimise and to control the development and manufacturing processes. |
|
4. RF/DC Plasma Sources Introduction
|
Back to top |
| Radio Frequency discharges are frequently used for sputtering, etching, and coating processes as well as for analytical purposes. The semiconductor industry has long found the interest of RF driven discharges for etching and coating in particular of non-conducting materials. Processes and physics linked to RF discharges are described in detail in various textbooks. The GDA 750 uses this technology to drive a GDOES plasma in order to sputter and excite non-conducting as well as conducting material. |
|
5. DC Glow Discharges
|
Back to top |
In a Grimm type Glow discharge source the sample is used as the discharge cathode. In order to run the discharge the sample is set to a negative Voltage of some hundred Volts. A plasma is ignited and an electron and ion current flows between the anode and the cathode. The electrical circuit is closed and the voltage drop occurs mainly across the plasma.
If we now replace the conduction sample or cathode by some non-conducting material and we again set the sample to a negative potential of some hundred Volts, a plasma will again be ignited. Free electrons and Argon ions will be created in the negative glow area. The sample has been set to a negative voltage and its surface is consequently bombarded by positively charged Argon ions. These Argon ions recombine with electrons on the sample surface. The electrons now missing on the sample surface, however, cannot be replaced by electrons from the interior of the sample because it is not conductive. Therefore no current across the sample is possible. The sample surface is charged positively and the potential difference between the two electrodes will decrease. Once the potential difference between the two electrodes is too small to maintain a plasma, no free electrons and Argon ions are available to maintain a current and the above described process is interrupted. These processes are very fast and the plasma will exist for less then a millisecond. |
|
6. Pulsed Discharges
|
Back to top |
| We have seen, an insulating cathode in a DC discharge will be positively charged and the plasma will be interrupted as soon as the voltage drop between the two electrodes is not any more sufficient to sustain a discharge. In order to reduce the charge on the cathode surface we can change the sign of the applied voltage. The potential on the sample surface will be positive and the voltage drop across the discharge chamber high enough to ignit a plasma. But this time the sample will act as the anode. The sample attracts electrons. Again the same procedure will happen as described above. |
Figure 3: |
The sample, being a non-conductor, can not conduct any electrical current. It’s surface will be charged negatively this time, the potential difference between the two electrodes will decrease until it is not any more sufficient to maintain the plasma. We can, however, repeat the switching of polarity over and over. The net time averaged, current will or must be equal to zero. In fact this technique of switching polarity is used in SIMS when non-conducting samples are analysed. An ion gun producing a weak current of highly energetic ions replaces the plasma as a source of free ions and electrons. For GDOES this technique does have a major drawback, each time the sample is positively charged, the “anode” tube will act as cathode and attract the Argon ions. The copper tube will be sputtered as well and copper will be detected everywhere. |
|
|
7. RF Discharges
|
Back to top |
It has been explained that it is possible to run a discharge with a Grimm type source even with a non-conductive sample as a cathode, if the polarity of the excitation voltage is changed continuously. Using an RF generator to produce the alternating voltage is nothing else than changing the polarity very rapidly. Rapid in this context means: faster than the associated processes in the plasma. Typically RF generators for GDOES are operated at 13.56 MHz, whereas the plasma reaction times correspond to a frequency of 1 kHz or a few kHz. In fact as a consequence the plasma will not decay when the polarity is changed at this high frequencies, but it will glow continuously.
The sputtering of the anode is avoided by a simple particularity of the configuration. If we look at the two electrodes, the copper tube and the sample, the area exposed to the plasma, and sputtering is very different. The sputter area on the sample is restricted by the anode tube to a circular surface of 2.5 mm to 8 mm in diameter. The area exposed to electron or ion current on the anode tube is not restricted at all can be extended as much as requested by the plasma size. |
Figure 4:
|
Figure 4 shows the potential distribution across an asymmetric planar diode discharge lamp. The two electrodes are very different in size, the grounded electrode being much larger than the RF-powered electrode. The blocking capacitor between the RF source and the electrode is necessary to allow the electrode be on a floating potential.
Figure 4 shows the local distribution of the time averaged potential across the discharge for different power and pressure levels. The grounded anode is at zero potential.
The plasma potential is always slightly above the potential of the surrounding walls. The cathode is at a negative potential. The potential difference between the cathode and the plasma is called the auto bias voltage. Across the blocking capacitor the potential will again drop to the ground level. The large potential difference between anode and cathode in an asymmetric discharge lamp is built up to ensure the time average current is zero. |
| The plasma potential, which is due to the high mobility of the electrons in the plasma, is not strong enough to considerably accelerate the Argon ions towards the anode. No sputtering of the anode occurs. |
Figure 5:
|
When the RF Voltage is switched on and the Plasma is ignited the auto bias voltage is established within a few RF cycles. (Figure 5) The time averaged potential difference between the two electrodes, the anode tube and the cathode, is non-zero. In fact, only during a very short period the plasma polarity is inverted and the electrons are accelerated towards the sample and the positive charge accumulated at the sample surface is removed. Again, the difference in mobility of electrons and Argon ions is very important here. Due to the higher mobility of electrons, the electron current occurring during the inversion of polarity equals the Argon ion current. The time averaged net current equals zero. |
|
|
8. Impedance Matching
|
Back to top |
| DC glow discharges used for GDOES with a discharge area of only a few mm² are known to have a resistance of 10 kOhm to 100 kOhm. It is therefore surprising to find RF GDOES sources to show only low impedance of a few Ohms. In fact, the discharge itself still has a high real-impedance, but the source including the sample and cables supplying the power do have a high capacitive effect, thus lowering the impedance of the complete source. |
Figure 6 :Impedance matching
|
RF power supplies are generally delivered with exit impedance of 50 Ohms. In order to deliver the power emitted by the power supply to the load, the load must have an impedance of 50 Ohms as well. For RF discharges this is generally achieved using an impedance matching system, consisting of a set of capacitors and coils. The two capacitors C1 and C2 are variable. They are automatically adjusted to assure total load impedance, discharge chamber and matching circuit, of 50 Ohms. The impedance of RF discharges for GDOES is generally above 50 Ohms. The matching circuit is therefore designed as an L-circuit.
The impedance of the discharge chamber depends on its general design as well as on the nature of the sample, its size and material. The automatic matching system must therefore continuously adapt the values of C1 and C2 to obtain optimal and stable discharge conditions. |
|
|
Sources:
B. Chapman: Glow Discharge Processes 1980, John Wiley, New York. |
|
|
|
