Real time closed loop control of plasma assisted semiconductor manufacturing processes has received significant attention in recent years. Therefore we have developed and tested a customized optical sensor based on buffer gas (argon) actinometry which has been used to determine relative densities of atomic and molecular oxygen in an Ar/O2 radio–frequency ICP chamber. The operation and accuracy of our optical sensor compared favorably with a high resolution commercial spectrometer but at lower cost and exhibited improved actinometric performance over a low resolution commercial spectrometer. Furthermore, threshold tests have been performed on the validity of buffer gas based actinometry in Ar/O2 ICP plasmas where Ar is no longer a trace gas through Xe actinometry. The plasma conditions for which this customized optical sensor can be used for closed loop control have been established.
© 2007 Optical Society of America
The use of plasma processes is an important step in the fabrication of semiconductor devices. The semiconductor industry requires high yield, performance and throughput of its devices throughout the manufacturing process. Of critical importance to the industry is the elimination of process drift within the etch and deposition processes used to manufacture such devices, either within one chamber or across a range of identical chambers. Off–line metrology of selected semiconductor wafers which have passed through such chambers is commonly used to determine process drift along a manufacturing line but this is expensive and time consuming.
Real time control of plasma assisted semiconductor manufacturing processes is seen as a viable alternative to off–line metrology and could greatly improve process yield and performance. One possible control strategy, which is referred to as the plasma parameter control strategy, may reduce plasma process disturbances and drifts. This strategy aims to control parameters internal to the plasma itself e.g. electron density, ion flux to a surface or radical species density as opposed to external process variables such as gas flow, rf power, chamber pressure etc. This control methodology has yet to be demonstrated and as a first step, a proof of principle experiment was required to be performed on a relatively simple plasma process with an Ar/O2 mixture operated in a laboratory inductively coupled plasma chamber. In order to perform such control experiments, non–invasive plasma diagnostics that can respond quickly to changes in plasma state are required which can operate at high speeds and be compatible with the surrounding control infrastructure and protocols. Plasma diagnostics such as laser–induced fluorescence (LIF), optical absorption spectroscopy, optical emission spectroscopy and actinometry fulfill this control criteria to varying degrees.
Laser-induced fluorescence [1, 2, 3, 4] and optical absorption spectroscopy are very powerful techniques used to determine species density in a plasma. These techniques allow measurement of the density of ground state atoms of a chosen species present in a plasma. Both techniques are very powerful but their application in an industrial context is unlikely due to the complexity of the laser and optical systems required. This problem can be solved if we employ optical emission spectroscopy (OES), which is a relatively simple technique to implement in an industrial setting. However, OES measurements provide information about excited electronic states, which are not necessarily related to the density of atoms in the ground state [5, 6, 7].
Optical actinometry is an OES technique which is widely used for in–situ monitoring of spatial and temporal variations of atomic and molecular concentrations [8, 9, 10]. This technique uses the addition of a small amount of gas in the discharge e.g. argon where the intensities of its spectral lines are known to be representative of the excitation mechanism. Argon is frequently chosen as the actinometer gas because the transition 3s23p54s–3s23p54p from multiplet 2[1/2]o 1–2[1/2]0 at 750.387 nm is insensitive to two step excitation . The excitation of the 2[1/2]0 4p level in Ar I is 13.48 eV  which is higher than the excitation threshold for the 4So 3p level in O I of 10.74 eV. A well known condition for actinometry is that the excited state of the actinometer (in our case argon) should have nearly the same energy as an excited state of the species of interest (i.e. oxygen). However there is a 2.7 eV difference between oxygen and argon electronic thresholds and thus it is clear that this condition is not satisfied. Nevertheless, if the remaining conditions for actinometry are fulfilled , then a difference between the energies of the excited states of argon and oxygen is not so critical and therefore Ar I 750.387 nm and O I 777.417 nm can be used for actinometry purposes [14, 15]. Comparison of the emitted line intensity for the desired species under examination (O, O2) with the intensity of an emitted line of argon, allows one to eliminate the influence of line intensity changes due to excitation conditions and evaluate the real behavior of the emitted line intensity due to the changes in the species ground state concentration. The relative density of atomic oxygen is monitored by calculating the ratio of the 3s–3p line intensity at 777.417 nm for O I and the 4s–4p line intensity at 750.387 nm for Ar. Similarly the relative density of molecular oxygen is determined by the ratio of the intensity of the B3∑ → X3∑ molecular band head at 417 nm  and the 750.387 nm emission line for Ar. To this end, we have developed a customized optical sensor for oxygen in an Ar/O2 discharge which measures the emitted intensity at these specific wavelengths in order to determine the density of atomic and molecular oxygen. This diagnostic can easily be used for the implementation of closed loop control of atomic and molecular oxygen density in the discharge.
However, the use of a trace gas as an actinometer in an industrial setting is unlikely as the requirement for an additional gas line would be outweighed by cost considerations. Many plasma processes in the semiconductor industry dilute the chemically reactant gases by using a buffer gas such as Ar or He. The ability to use this buffer gas as the actinometric gas would negate the requirement for additional trace gases and therefore allow this technique to be applied in a manufacturing process. It must be stated, however, that the validity of actinometry  is somewhat controversial and the criteria for the utilization of the technique and its limits of validity must be verified in each case. Actinometry becomes invalid if the excited state from which emission is being detected has not been created by electron impact excitation from the ground state. The validity of actinometry using excited oxygen and argon atoms was investigated by Walkup et al . They found that the actinometric determination was well correlated with the variation of atom concentration up to 50 Pa in RF O2-CF4 plasmas, but discrepancies occurred in pure O2 plasmas. In addition, it was shown by Booth et al  that the ratio of the intensities of the oxygen line to the argon line, was poorly correlated to the oxygen atom concentration in Electron Cyclotron Resonance (ECR) plasmas operated at ultra-low-pressures (0.1–0.8 Pa).
It is because of this cautionary note that extensive validation measurements have been performed on our system which uses argon as the buffer gas. The validation procedure uses the actinometric technique itself as trace amounts of Xe gas have been added to the Ar/O2 discharge to confirm the applicability of the technique for a range of Ar/O2 discharges. The validity of using the buffer gas as the actinometer should be manifested as an agreement between the argon and xenon actinometry data under a range of plasma conditions. Once this validation procedure is complete, we no longer use trace gas Xe in the process and can revert back to actinometry measurements using the buffer gas of the process (argon).
2.1. Basic radio-frequency inductive source
The plasma chamber used in this campaign is the BARIS (BAsic Radio-frequency Inductive Source) system. The BARIS discharge chamber consists of a water-cooled helical antenna which is isolated from the discharge by a quartz dielectric tube and is centered co-axially inside a cylindrical stainless steel discharge chamber as shown in Fig. 1. The antenna is driven at 13.56 MHz using a radio-frequency generator at powers from 25–300W and uses a standard ’L’-type impedance matching unit in automatic matching mode to ensure maximum power transfer from the generator to the plasma.
The discharge chamber itself is a stainless steel cylindrical vacuum chamber of internal diameter 200 mm and length 860 mm and has several vacuum ports to allow for diagnostic access. The chamber is evacuated using a turbomolecular pump and a rotary vane pump which gives a base pressure of the order of 5× 10-5 Pa (5× 10-7). Chamber pressures from 1.33Pa (10 mTorr) to 13.3Pa (100 mTorr) are achieved using a stepper motor driven gate valve positioned above the turbomolecular pump. Gas flow into the chamber is controlled via digital mass flow controllers (MFCs) which precisely determine gas content in the chamber. At these pressures, electron densities of the order of 1010 cm-3-1011cm-3 are common in this system for rf input powers from 25–300 W. In fact the large parameter space required of this work meant that a Design of Experiment (DOE) had to be implemented i.e RF power: 25–300 W, Ar gas flow rate (7–350 sccm), O2 gas flow rate: (2–100 sccm) and chamber pressure: 10–100 mTorr. The application of this DOE reduced the number of experimental runs to 48.
2.2. The customized optical device
The requirement for a real time, practical and easy to use sensor for oxygen detection in a plasma processing environment resulted in the construction of the customized optical device detailed schematically in Fig. 2. This device may be applied to process chambers in the semi-conductor industry as it is much cheaper to construct and easier to use than many high resolution spectrometers available on the market. Relatively cheap low resolution spectrometers are available but, as is discussed later, such devices can have a limited range of operation for actinometry purposes.
The actinometric technique applied here requires the recording of the intensity of emission lines at three specific wavelengths i.e. 750.387 nm for Ar I, 777.417 nm for O I and 417 nm for O2. Our customized optical device operates on the principle that the polychromatic light emitted from the plasma enters the device and is split into the three wavelengths of interest which fall with approximately equal intensity onto their associated optical filters/photodiode detection channel. Each optical filter (OF), manufactured by LOT Oriel, has a very narrow bandwidth (FWHM: OF1=0.67 nm, OF2=0.71 nm, OF3=0.57 nm), centered at a specific wavelength as shown on Fig. 2 and is therefore only sensitive to its particular species’ emission line. The analog voltage from each photodiode detection channel is proportional to the emitted intensity at that wavelength which can be amplified via the photodiode’s amplifier circuitry. The plasma light was collected by this device from a region about 2 cm in diameter and its entrance port was positioned about 16.5 cm above the antenna itself.
To enable direct comparison between DC signals from different photodiodes and optical filters with different transmission coefficients (OF1=0.77, OF2=0.74, OF3=0.64), a calibration procedure was performed on this optical device. The calibration has been carried out using a Gigahertz–Optik BN–0102–1 Reference Standard source .
Calibration procedures have also been performed to compare the operation of our customized optical device with two commercial spectrometers. The first spectrometer is a Carl Zeiss PGS-2 2 m focal length spectrometer with a 1302 lines/mm grating which operates from 200 to 1400 nm and has an intensified–CCD (iCCD) camera system, which is sensitive from 190 nm to 900 nm, positioned at its focal plane. The PGS-2/iCCD optical system has very high resolution of 9.17 pm/pixel at λ=200 nm and 7.5 pm/pixel at λ=900 nm at the optimum entrance slit width of 15μm. Although the PGS-2/iCCD optical system has very high resolution, it is bulky, difficult to operate and align and has just 5 nm across the CCD detector at any one time. An alternative spectrometer is the Ocean Optics USB2000 fibre optic spectrometer which is a low resolution spectrometer with a focal length of 42 mm, a 600 lines/mm grating which operates from 200 to 875 nm. The resolution of the USB2000 is 0.555 nm/pixel at λ=200 nm and 0.536 nm/pixel at λ=875 nm. This compact spectrometer is easy to operate with no optical alignment required and can display the entire spectrum from 200 to 875 nm at any one time. Both spectrometers use the same multimode fibre with core diameter of 200 μm and length of 2 m positioned at its slit entrance. A collimator lens at the other end of the fibre is used to collect light from plasma.
Actinometry experiments were performed in Ar/O2 plasmas under the 48 different plasma operating conditions determined by the DOE calculation. Actinometry threshold tests were performed with xenon as a trace gas (2% of total pressure) in the Ar/O2 plasmas to identify the plasma conditions for which buffer gas actinometry with argon is no longer valid as the argon content in the plasma has exceeded a validity threshold.
3. Results and Discussion
3.1. Comparison of the low and high resolution spectrometers for actinometry
If the USB2000 low resolution spectrometer were to be used for actinometry measurements in an Ar/O2 plasma, it would need to exhibit the same behavior as the PGS-2 high resolution spectrometer after the appropriate calibration procedures . Fig. 3 shows a typical optical spectrum recorded by both spectrometers in the region of the actinometric lines for O I (777.417 nm) and Ar I (750.387 nm) for an Ar/O2 plasma. As can be seen in the plots on the left side of Fig. 3, there are additional lines around the actinometric wavelengths of interest as measured by the PGS-2 spectrometer which are summed and appear as one broad emission peak in the USB2000 spectrometer shown on the right side of Fig. 3.
For comparative purposes, we have selected a reduced range of plasma operating conditions from the original DOE dataset of 48. The reduced dataset of 18 encompasses the extremes of rf power/pressure in the BARIS system allowing us to investigate in more detail how the actinometry peak intensities at 777.417 nm and 750.387 nm behave as measured by both spectrometers. The intensity of each spectral peak was measured by determining the area under the line of interest as summed over a spectral region from -2.5 full width half maximum (FWHM) step size up to +2.5 FWHM step size from the spectral line peak central position. If the spectral line was clearly isolated from other emission lines, then the intensity summation was straightforward as it is just a summation over ±2.5 FWHM step sizes from the spectral line peak central position. However, if other lines existed in close proximity to the spectral line of interest, as was the case for the O I 777 nm triplet, then the intensity summation had to neglect the intrusive line and the summation was performed down to the background continuum that existed in that spectral region.
The intensity of each line of the triplet around 777 nm in O I measured with the PGS-2 spectrometer exhibits similar behavior with changing rf power/pressure over the plasma operating conditions tested here. A similar observation was made for each line of the Ar I “doublet” around 750 nm. This infers that although the actual actinometry spectral line is not resolved by the USB2000 spectrometer, the measured broad peak intensity can by multiplied by a certain factor to give the actual intensity of the actinometric emission line within the particular unresolved peak. This factor was calculated to be 0.31 for the O I line at 777.417 nm and 0.42 for the Ar I line at 750.387 nm.
The ratio of the measured and inferred intensity of the Ar I line determined from the PGS-2 and USB2000 spectrometers should therefore be constant over all the plasma operating conditions. Similarly a fixed but different ratio should also be observed for the O I line. However, as the left side of Fig. 4 indicates, the line intensity ratio in each case is not actually constant, even for the restricted set of plasma operating conditions used here. This infers that one needs to be very careful when using a low resolution spectrometer, such as the USB2000, for actinometry purposes as it works only under certain experimental conditions. This discrepancy between the low and high resolution spectrometers for certain plasma conditions results from the variation in the continuum emission measured by each spectrometer and the contribution of unwanted spectral lines to the peak profile recorded by the low resolution spectrometer. This continuum emission varies with pressure and rf power in the plasma leading to an inaccuracy in the factor used to determine the actual peak intensity of the desired actinometry line in the broad peak profile.
To quantify the data comparison of the low resolution spectrometer and customized optical device with the high resolution spectrometer, as shown in Fig. 4, the root mean square (RMS) value of the ratio has been calculated for each case. The RMS value for the USB2000/PGS-2 ratio for Ar I was 3.8 × 10-3 and for O I was 7.4 × 10-3. The RMS value for the customized optical device/PGS-2 ratio for Ar I was 4.7 × 10-4 and for O I was 5.1 × 10-4. This would indeed indicate the data associated with the customized optical device is more reliable than a low resolution spectrometer when compared with an optical spectrometer of high resolution.
In addition, the 4p–4d spectral line of Ar I at 751.041nm (see Fig. 3) makes a significant contribution to the broad unresolved profile at 751 nm in the argon spectrum at higher rf powers. Thus three lines are contributing to the unresolved peak profile instead of the two that were previously assumed to be involved when the factor was calculated. Other unwanted spectral lines are seen in the cluster of three lines around 778 nm in the top left pane of Fig. 3. These lines are not from the oxygen spectrum and are in fact impurity lines from the plasma discharge.
The reduced operating space of the low resolution spectrometer for actinometry measurement necessitated the development of a customized optical device which could be used for real time control of O I and O2 species density in an rf ICP plasma.
3.2. Comparison of the customized optical device with a high resolution spectrometer for actinometry
The limited range of use of the USB2000 spectrometer for actinometry purposes, resulted in the construction of the customized optical device, as depicted in Fig. 2. To test the reliability of this device, a comparison was again made of its measured intensity with that of the high resolution PGS-2 spectrometer. The ratio of the intensities of the Ar I and O I lines was calculated and is shown on the right side of Fig. 4.
The intensity ratio is much flatter for both the Ar I and O I lines than that observed previously with the low resolution spectrometer (see Fig. 4 - left side). Therefore the intensity measurements made with the customized optical device are similar to the PGS–2 and thus reliable actinometry measurements can be made without the need for bulky spectrometers that are difficult to align. Furthermore, the customized optical device delivers a signal that is directly proportional to intensity (area) of the spectral line. Thus no additional recording software, spectral “line” deconvolution procedures or theoretical line profile fitting techniques need to be applied to the measured emission line profiles. Improved actinometric performance may be achieved with a commercial compact spectrometer of moderate resolution which could be optimized for operation in a specific spectral region e.g. 750–778 nm. However, even in this scenario, the optimized compact spectrometer would still require additional and time consuming line profile fitting procedures to be employed when compared with our customized device. Fig. 4 also shows that the absolute value of ratio between our device and the PGS–2 spectrometer is higher than that of the USB2000/PGS–2 spectrometer of Fig. 4. This infers another advantage associated with our optical device over a low resolution spectrometer as it can be used to record spectral lines of low intensity. Obviously, spectrometers can record low intensity spectral lines by increasing integration time but this will also lead to an increase in the continuum measured which would also contribute substantially to the unresolved peak profile leading to a miscalculation of the peak intensity of the line under examination.
The customized optical device can be used for actinometric purposes to determine O I and O2 densities in an Ar/O2 plasma and so the next step was to investigate precisely where argon actinometry is valid in a system where Ar is no longer used in trace amounts as is the case in these experiments. The actinometry threshold tests are detailed in the following subsection.
3.3. Threshold tests for argon actinometry
Optical actinometry with argon as the actinometer was required to determine the relative atomic and molecular oxygen concentration in an Ar/O2 plasma where argon is no longer used in trace amounts. This is not the usual manner in which Ar actinometry is applied as it is normally introduced in trace amounts in the plasma under investigation. Thus another technique was required to determine the validity of Ar actinometry in an environment where the content of argon varies considerably. The technique chosen to test this principle was xenon actinometry where Xe is added in trace amounts (about 2% of the total pressure) to the Ar/O2 plasma.
It was essential that as little xenon as possible be introduced into the argon–oxygen plasma i.e. the requirement was to be able to record the xenon actinometry line at 834.7 nm whilst ensuring that the entire argon–oxygen emission spectrum was not altered. There are two xenon actinometry lines that are frequently used: (1) 834.682 nm from the 6s–6p transition and (2) 828.012 nm from the 6s–6p transition. We chose to use the Xe I line at 834.682 nm as the actinometry line as it does not overlap with any argon or oxygen spectral lines in that region.
The broadband spectral sensitivity of the USB2000 spectrometer meant that it could very easily be used to test for any disturbance in the argon–oxygen optical spectrum under all of the 48 plasma conditions required as part of our original design of experiment (DOE). Two sets of optical data were then recorded for each of the 48 experimental runs, the first set was in Ar/O2 plasma only and the second set with 2% of xenon added to the Ar/O2 plasma. Fig. 5 shows the recorded spectra for one such argon–oxygen and argon–oxygen–xenon plasma and reveals how similar the two spectra are with the only apparent difference being the Xe I line at 834.7 nm. Hence we can be assured that xenon can reliably be used as our actinometer reference for the threshold tests required for argon actinometry in Ar rich plasmas.
The actinometry results with argon as the actinometer were then compared with those with xenon as the actinometer. The validity for actinometry by argon in the Ar/O2 plasmas is studied by monitoring the point when the relative atomic oxygen and molecular oxygen determined by Ar and Xe actinometry begins to diverge. Fig. 6 shows a selection of such actinometry measurements as the percentage of argon in the argon–oxygen plasma and argon–oxygen–xenonplasma is increased from 0 to 100% of total flow. This plot shows some of the limiting cases where Ar actinometry is valid e.g. Ar actinometry is not valid when the total Ar flow is greater than 45% at low rf powers and pressures (25 W, 1.3 Pa). The validity of argon actinometry at higher rf powers and pressures (300 W, 13.3 Pa) is restricted to cases where Ar flow is less than 20% of the total flow into the chamber. In fact, Fig. 6 shows that for the range of rf powers and pressures used in this experiment, Ar actinometry is only valid in Ar/O2 plasmas when the Ar flow is less than 20% of the total flow.
The atomic oxygen and molecular oxygen densities were calculated to be of the order of 1018 m-3 – 1019 m-3 for the Ar/O2 plasma conditions where Ar actinometry was valid. These densities were determined using all the relevant actinometry data [13, 21] and assuming a gas temperature, according to previous investigations, which changes linearly with RF power from 300 K (the reactor-wall temperature) at 0 W up to 450 K at 300 W. The calculated value for the densities of atomic and molecular oxygen are within the expected limits for the powers and pressures used in the plasma chamber and are typical for many similar RF plasmas .
From the above discussion it is clear that a customized optical device can be used for buffer gas actinometry purposes in an Ar/O2 plasma when the Ar content is below a certain threshold. As a general rule validation of the use of a particular buffer gas for actinometry purposes is required. The customized optical device is a viable alternative to a high resolution spectrometer for the measurement of atomic and molecular oxygen densities in such plasmas. It is envisaged that this sensor will be used for real time measurement of species concentration which can be used to achieve active species control (O, O2) in a radio–frequency inductively coupled plasma.
A real time sensor for atomic and molecular oxygen in a radio–frequency inductively coupled plasma has been developed. This sensor uses buffer gas actinometry where argon is no longer a trace gas to determine relative atomic and molecular oxygen concentrations in an Ar/O2 plasma. The validity of using the buffer gas as the actinometer has been determined by a one–off set of actinometric measurements with xenon gas used in trace amounts. This sensor exhibits improved performance over a low resolution spectrometer which cannot always be used for precise actinometric measurements. This sensor is also seen as a practical alternative to more cumbersome, expensive and complex high resolution optical spectrometers for radical species density measurement.
This work is a part of the projects: No. 02/IN.1/I147 founded by Science Foundations Ireland, EURATOM contract FU06-CT-2004-00068 founded by European Union, “Measurement of atomic Oxygen in an industrial plasma etcher” and project “The use of non invasive OES as method for the determination of gas species concentration and the main plasma parameters in industrial plasma processing equipment” founded by Enterprise Ireland. V. Milosavljević is grateful to the Ministry of Science and Environment Protection of the Republic of Serbia under grand No.: OI141031 ”Nonlinear dynamical phenomena in photorefractive media, liquid crystals, plasmas and left–handed materials”.
References and links
1. D. Lee, G. Severn, L. Oksuz, and N. Hershkowitz, “Laser-induced fluorescence measurements of argon ion velocities near the sheath boundary of an argon-xenon plasma”, J. Phys. D: Appl. Phys. 395230–5235 (2006). [CrossRef]
2. T. Lee, W. G. Bessler, C. Schulz, M. Patel, J. B. Jeffries, and R. K. Hanson, “UV planar laser induced fluorescence imaging of hot carbon dioxide in a high-pressure flame”, Appl. Phys. B: Lasers & Optics 79/4427–430 (2004). [CrossRef]
3. W. Koban, J. D. Koch, R. K. Hanson, and C. Schulz, “Toluene LIF at elevated temperatures: implications for fuel-air ratio measurements”, Appl. Phys. B: Lasers & Optics 80/2,147–150 (2005). [CrossRef]
4. J. Amorim, G. Baravian, J. Jolly, and M. Touzeau, “Two-photon laser induced fluorescence and amplified spontaneous emission atom concentration measurements in O2 and H2 discharges”, J. Appl. Phys. 76/31487–1493 (1994). [CrossRef]
5. B. L. Preppernau, K. Pearce, A. Tserepi, E. Wurzberg, and T. A. Miller, “Angular momentum state mixing and quenching of n=3 atomic hydrogen fluorescence”, Chem Phys. 196, 371–381 (1995). [CrossRef]
6. J. C. Thomaz, J. Amorim, and C. F. Souza, “Validity of actinometry to measure N and H atom concentration in N2-H2 direct current glow discharges”, J. Phys. D: Appl. Phys. 323208–3214 (1999). [CrossRef]
7. N. G. Ferreira, E. J. Corata, V. J. Trava-Airoldia, and N. F. Leitea, “OES study of the plasma during CVD diamond growth using CCl4/H2/O2 mixtures”, Diamond and Related Materials 9/3–6368–372 (2000). [CrossRef]
8. J. W. Coburn and M. Chen, “Optical emission spectroscopy of reactive plasmas: A method for correlating emission intensities to reactive particle density”, J. Appl. Phys. 513134–3136 (1980). [CrossRef]
9. S. De Benedictis, A. Gicquel, and F. Cramarossa, Proc. 8th Int. Symp. Plasma Chem. ISPC’87, (Ed. K. Akashi and A. Kinbara), Tokyo (1987).
10. P. Macko, P. Veis, and G. Cernogora, “Study of oxygen atom recombination on a Pyrex surface at different wall temperatures by means of time-resolved actinometry in a double pulse discharge technique”, Plasma Sources Sci. Technol. 13251–262 (2004). [CrossRef]
11. T. Czerwiec, F. Greer, and D. B. Graves, “Nitrogen dissociation in a low pressure cylindrical ICP discharge studied by actinometry and mass spectrometry”, J. Phys. D: Appl. Phys. 38/244278–4289 (2005). [CrossRef]
12. NIST - Atomic Spectra Data Base Lines (wavelength order) 2007 - http://physics.nist.gov
13. M. Lieberman and A Lichtenberg, “Principles of Plasma Discharges and Materials Processing” (New York: Wiley), (1994).
14. S. Fujimura, K. Shinagawa, M. Nakamura, and H. Yano, “Additive Nitrogen Effects on Oxygen Plasma Downstream Ashing”, Jpn. J. Appl. Phys. 29/102165–2170 (1990). [CrossRef]
15. A. Granier, D. Chéreau, K. Henda, R. Safari, and P. Leprince, “Validity of actinometry to monitor oxygen atom concentration in microwave discharges created by surface wave in O2–N2 mixtures”, J. Appl. Phys. 75/1104–114 (1994). [CrossRef]
16. R. W. B. Pearse and A. G. Gaydon, “The identification of molecular spectra”, (Chapman & Hall LTD., London) (1941).
17. C. Guyon, S. Cavadias, and J. Amouroux, “Heat and mass transfer phenomenon from an oxygen plasma to a semiconductor surface”, Surf. Coat. Technol. 142–144959–963 (2001). [CrossRef]
18. R. E. Walkup, K. L. Saeneer, and G. S. Selwyn, “Studies of atomic oxygen in O2+CF4 rf discharges by two- photon laser-induced fluorescence and optical emission spectroscopy”, J. Chem. Phys. 842668–2674 (1986). [CrossRef]
19. J. P. Booth, O. Joubert, J. Pelletier, and N. J. Sadeghi, “Oxygen atom actinometry reinvestigated: Comparison with absolute measurements by resonance absorption at 130 nm”, J. Appl. Phys. 69618–626 (1991). [CrossRef]
20. V. Milosavljević and A R Ellingboe, “Quantum efficiency of Spectrometers”, PRL Internal report (Dublin: Dublin City University) (2004).
21. A. D. Richards, B. E. Thompson, K. D. Allen, and H. H. Sawin, “Atomic chlorine concentration measurements in a plasma etching reactor. I. A comparison of infrared absorption and optical emission actinometry”, J. Appl. Phys. , 62/3792–798 (1987). [CrossRef]
22. H. M. Katsch, A. Tewes, E. Quandt, A. Goehlich, T. Kawetzki, and H. F. Döbele, “Detection of atomic oxygen: Improvement of actinometry and comparison with laser spectroscopy”, J. Appl. Phys. 88/116232–6238 (2000). [CrossRef]