RU2587468C2 - Method of measuring density of electrons in plasma by optical spectroscopy - Google Patents

Abstract

FIELD: measuring equipment.

SUBSTANCE: invention relates to measurements by optical methods of electrophysical parameters of plasma, including electron density and electric field intensity and their distribution. Method of measuring spatial distribution of electron density of plasma includes measurement of intensity of plasma radiation from different coordinate areas of spark gap at wavelength, corresponding to spectral line of atomic or molecular band, which is selected so that radiation intensity of said line or band is mainly determined by excitation of radiating state by direct electron impact or fast compared with a period of high-frequency field cascade processes, with subsequent determination of spatial distribution of electron density of plasma by numerical simulation of plasma.

EFFECT: measurement by optical methods of electrophysical parameters of plasma.

1 cl, 5 dwg

Description

Technical field

The invention relates to the field of optical measurements of electrophysical plasma parameters, including electron density and electric field strength and their distributions.

State of the art

A distinctive feature of measuring the electrophysical parameters of plasma, including electron density, is that these parameters are not measured directly, but are calculated using a rather complicated algorithm based on the physical model of the plasma and its interaction with the detector. Therefore, in the General case, the method of measuring them includes recording the sensor signal, its digitization and processing in order to calculate the value of the required parameter. By the type of sensor, two types can be distinguished: a probe, which is placed in direct contact with the plasma, or an electromagnetic radiation detector that detects radio-frequency or optical radiation from the plasma and is located outside the plasma region. In the first case, the solution to the problem of determining the plasma parameters is complicated by the fact that the probe can introduce distortions into the plasma, and in the second, it is necessary to solve a complex inverse problem and it is not always possible to ensure sufficient measurement accuracy. Despite the fact that theories of probe and optical measurements are quite well developed, the creation and application of new methods for measuring the electron density of plasma is actively continuing, since plasma processes play a key role in many modern industrial technologies and, therefore, new design options for probes of various types and methods of measuring them help are subject to intensive development.

Known, for example, is a method of measuring electron density from measured signals of the radio frequency range, described in US patent N5471115 "Method and apparatus for measuring electron density of plasma", publ. November 28, 1995, according to which plasma oscillations in the 13.56 MHz RF discharge are excited in the reactor with an auxiliary RF, microwave potential or an alternating magnetic field superimposed and a signal is taken from the antenna placed in the plasma volume and the frequency of plasma oscillations is determined by which calculates the electron density value. In US patent N6744211 "Plasma density information measuring method, probe used for measuring plasma density information, and plasma density information measuring apparatus", publ. 06/01/2004, a probe is proposed that is an antenna for introducing a tunable frequency microwave signal into the plasma region. The antenna is isolated from the plasma by a dielectric housing. The frequency of the microwave signal is determined at which the intensity of the reflected wave drops sharply (total absorption of the incident wave), which occurs when the plasma frequency, determined by the electron density in the plasma, is reached. An improved version of this system is described in US patent N7061184 "Plasma electron density measuring and monitoring device", publ. 06/13/2006, according to which a microwave probe, consisting of transmitting and receiving antennas placed in parallel in one plasma-protecting dielectric tube, allows you to simultaneously radiate and receive a probe microwave signal. Subsequent analysis reveals the frequency at which the received microwave signal has a minimum amplitude and calculate the electron density in the plasma. The probe can be moved to measure inhomogeneous plasma parameters. A more complex version of the algorithm for calculating the electron density of plasma from probe data, which, thanks to the use of computer signal processing methods, allows measurements with increased accuracy, is described in US patent N7339656 "Method and apparatus for measuring electron density of plasma and plasma processing apparatus", publ. March 4, 2008. It describes a method for determining the electron density from the data of a microwave transceiver probe, when a frequency dependence of the complex reflection coefficient of a microwave wave from a plasma is formed from a receiving antenna signal and the resonant frequency is defined as the frequency at which the imaginary part of the complex coefficient reflection goes over zero. A variant of this method, in which the probe is moved to determine the spatial distribution of electron density, is described in US Pat. No. 7,462,293, Method and apparatus for measuring electron density of plasma and plasma processing apparatus, publ. 12/09/2008

Another typical example of the use of modern computer technology for the design of the probe is described in US patent N7878045 "Apparatus and use of the apparatus for the determination of the density of a plasma", publ. 02/09/2011, Langmuir probe for measuring electron density by the resonant frequency of the plasma, in which the measuring head is made in the form of a triaxial ellipsoid and contains a shell and core, on the surface of which electrodes are formed to supply voltage of opposite polarity to them. By varying the ratio of the shell and core dimeters and the geometrical parameters of the regions insulating the electrodes, it is possible to achieve zeroing of all response components except the dipole, thereby eliminating distortions in its measurement.

But a planar capacitively coupled ion flow sensor is also known, described in US Pat. No. 7,282,463, Method and apparatus to detect fault conditions of a plasma processing reactor, publ. November 9, 2010, to which a pulsed RF voltage is applied and which allows measuring the saturation of the ion current and then calculating other plasma parameters. Tracking sharp deviations in the value of the plasma parameter allows you to identify the occurrence of violations of the operating mode.

Another approach to measuring electron density is based on the fact that the presence of free electrons in a medium changes the effective value of the dielectric constant of the medium and, accordingly, the parameters of the resonant circuit of the probe. The electron density in microwave plasma according to the method described in US patent N6541982 "Plasma density measuring method and apparatus, and plasma processing system using the same", publ. 04/01/2003, is calculated from the measured and obtained by numerical simulation of the intensity distributions of surface waves excited in a dielectric (quartz) plate in contact with the plasma. The method is based on the effect of the dependence of the speed of propagation of a surface wave and, accordingly, its wavelength on the concentration of electrons above the dielectric plate. In another embodiment, the difference between the resonant frequencies of the microwave power reflection coefficient with and without plasma can be determined, and the local electron density in the discharge can be calculated from its change. This method is described, for example, in US patent N7582182 "Method and apparatus for measuring electron density of plasma and plasma processing apparatus", publ. 09/01/2009, U.S. Patent No. 8040138, "Planar type frequency shift probe for measuring plasma electron densities and method and apparatus for measuring plasma electron densities", publ. 10/18/2011, a probe in the form of a plate with a notch forming a petal of a dipole antenna with a length close to the expected value of λ / 4 resonant plasma oscillations is described. The change after the plasma is turned on of the resonant frequency of such a dipole antenna, previously measured in vacuum, allows us to calculate the plasma frequency and, accordingly, the electron density in the plasma. A similar approach is implemented in US patent N8190366 "LC resonance probe for determining local plasma density", publ. 05/29/2012, according to which the measurement of the natural resonant frequency in vacuum and in the presence of a plasma, but the resonant circuit consists of an inductance and a capacitor, the plasma can freely penetrate into the region between the plates. The capacitor can be either in the form of two plane-parallel plates at the end of the probe connected to the electrodes of a cable of different polarity, or it can be formed by a central core and mesh (so that the plasma can freely penetrate into the capacitor) with a coaxial braid electrode.

However, the use of Langmuir probes inevitably introduces distortions in the uniformity of the plasma, which negatively affects the uniformity of processes on the surface. Such probes cannot be used for a long period of time as a tool for monitoring the technological regime, since the surface of the probe will gradually change under the influence of chemically active plasma particles, and its readings will be distorted. Moreover, even slight spraying of the probe surface, which inevitably occurs under the influence of plasma ion bombardment and plasma-chemical reactions on its surface, even if it is made of refractory metal, such as tungsten, leads to technologically unacceptable levels of contamination of the treated surface and, as a result this, to the deterioration of the operating characteristics of the formed semiconductor structures. The use of protective coatings allows avoiding contact of the measuring surface. So, for example, in US patent N6894474 "Non-intrusive plasma probe", publ. 05/17/2005, also describes a wire probe coated with an insulating inorganic or polymer protective layer that excludes the effect of plasma on the probe and minimizes its effect on plasma. However, when such a probe measures in a plasma-chemical deposition medium, functional layers, especially dielectric layers, due to the formation of a poorly conducting or insulating film on the probe surface, a strong distortion of the measured currents is observed. In addition, modern technologies, as a rule, use high-frequency plasma of various frequencies or plasma in a magnetic field, which additionally complicates the interpretation of the values of currents and displacements measured by the probe. This circumstance requires further complication of the design of the probe, for example, as described in US patent N7015703 "Radio frequency Langmuir probe", publ. March 21, 2006, according to which inductors are located inside the legs of the Langmuir probe for measuring the characteristics of an RF discharge plasma, and external electrodes forming capacitors so as to form a voltage divider and a frequency filter are located outside, which allows measurements in discharges of different frequencies -Frequency RF discharges (for example, 2 MHz and 27 MHz).

The problems associated with the interaction between the measuring probe and the plasma can be eliminated by moving the measuring sensor away from the region of plasma existence. The best option is to embed such a sensor in the wall of the plasma reactor. This approach is described, in particular, in US patent N5705931 "Method for determining absolute plasma parameters", publ. 01/06/1998, and N5861752 "Method and apparatus for determining absolute plasma parameters", publ. 01/19/1999, according to which the measuring electrode is placed as an insulated element of small area in the grounded wall of the plasma of a highly asymmetric high-frequency reactor, which avoids the introduction of any foreign elements into the plasma volume, the inter-period profile of the current flowing through this measuring electrode is measured, and the spectrum of its harmonics is determined. The numerical solution of the differential equation describing the nonlinear motion of charges in the near-electrode layer is used to find the absolute values of the plasma parameters, including electron density. The same principle lies in the design and measurement method described in US patent N7994794 "Methods for measuring a set of electrical characteristics in a plasma", publ. 08/09/2011, according to which a disk probe for measuring the ion flux from plasma to the surface of the chamber of the plasma-chemical reactor in contact with it is placed directly in this wall flush with its surface, which eliminates the introduction of local distortions in its characteristics. The voltage, phase and current are measured, which can then be used to determine the characteristics of the plasma.

Of particular interest are methods for measuring plasma characteristics based on non-contact measurements, for example, various options for microwave sounding of plasma. There are three approaches to the use of microwave methods for determining electron density in a plasma, which are based, respectively, on obtaining and measuring the parameters of the obtained interference distributions of the probe microwave radiation, as described, for example, in US Pat. No. 6,818,844 to Electron density measurement and plasma process control system using changes in the resonant frequency of an open resonator containing the plasma ", publ. 03/01/2005, in which a tunable microwave generator is used, which, smoothly changing the frequency, finds the frequencies of all resonance modes in the cavity before and after turning on the plasma, after which the order of interference maxima, the frequency shifts of all resonance modes are calculated, and the electron plasma density in the cavity; measuring the reflection and absorption coefficients of probing microwave radiation in a plasma, as described, for example, in US Pat. No. 7,404,991 to “Device and control method for microwave plasma processing”, publ. 07/29/2008, according to which, according to the measured power of the reflected wave and the frequency deviation of the microwave generator, a system for automatically adjusting the power input in the microwave discharge excited in the plasma reactor is activated; or measuring changes in the resonant frequencies of a resonator when an electric discharge plasma is excited therein, as described in US Pat. No. 6,537,731 to "Electron density measurement and control system using plasma-induced changes in the frequency of a microwave oscillator", publ. 06/03/2003, according to which the plasma region is located on the path of a microwave beam excited in a generator with self-excitation. Accordingly, the excitation frequency of such a microwave generator varies depending on the plasma density, and comparing it with the frequency of the reference microwave generator and measuring the frequency of the difference signal allows us to calculate the electron density in the plasma along the direction of the excited microwave beam. The features of a stabilized microwave generator used to generate a difference frequency with microwave oscillations excited in the open resonator containing the plasma excitation region, and then to generate a signal for changing the power of the plasma excitation generator, are described in detail in US Patent N6646386 "Stabilized oscillator circuit for plasma density measurement ", publ. 11.11.2003. In this case, an implementation option of the excitation system of a microwave generator with an open resonator with tuning to a different mode after the initial mode leaves the resonance conditions when the plasma is turned on, as described in US patent N6741944 "Electron density measurement and plasma process control system using a microwave oscillator locked to an open resonator containing the plasma ", publ. 25.95.2004, Since the wavelength of microwave radiation, as a rule, is much shorter than the characteristic dimensions of the plasma chemical reactor, the possibility of making measurements with spatial resolution is of particular interest. So, in US patent N6713969 "Method and apparatus for determination and control of plasma state", publ. March 30, 2004, an open resonator is described that contains a plasma excitation region and is used to measure the detuning of its resonant frequency, which is movable, or many such resonators are installed with the cavity axis being parallel to the surface of the plate being processed in the plasma. And in US patent "Method and apparatus for electron density measurement" N7544269, publ. 06/09/2009, a variant is proposed when each microwave resonator is formed by a mirror and a processed substrate so that its axis is perpendicular to the processed surface. The frequency dependences of the transmission coefficient of the microwave signal in the formed multimode resonators in vacuum and in plasma are measured. By changing the resonant frequency, the electron density at different points along the plate being processed is calculated. The obtained measurement data make it possible to restore the spatial distribution of electron density in the reactor volume and, for example, to adjust the RF power put into maintaining the electric discharge or the gas supply to the reactor so that the specified process parameters remain unchanged.

An interesting version of the method for determining the characteristics of plasma from changes in the parameters of the probing signal is proposed in US patent N8214173 "Plasma system and measurement method", publ. 08/14/2012, according to which a disturbing signal of various shapes modulates the supply voltage of a plasma DC reactor and measures the impedance, phase currents and determines not only the occurrence of a failure in the operating mode, but also its type (breakdown of insulation, electrode wear, etc. .).

However, it is important to note that all the microwave methods described above, although they are non-contact, nevertheless presuppose a certain, though relatively low degree of effect on the plasma. Therefore, a measurement option that would completely exclude the effect of a probing signal on the plasma would be ideal. Closest to this could be methods based on the analysis of emission spectra of plasma. In practice, either simplified methods of optical diagnostics are used, or they are combined with traditional electrophysical methods of measurement. Possible options for this approach and a comparison of the applicability of different plasma diagnostic methods are described, for example, in the article "An overview of diagnostic methods of low-pressure capacitively coupled plasmas", Thin Solid Films, vol. 521 (2012), p.141-145.

Very often, methods of emission spectroscopy are used simply to monitor the normal course of the process, as, for example, described in the group of US patents under the general name "Method and apparatus for monitoring plasma processing operations" (N6134005, publ. 17.10.2000, N6221679, publ. April 24, 2001, N6419801, published July 16, 2002, N6805810, published October 19, 2004) a system for recording and automatic analysis of optical emission spectra of plasma during the processing of silicon wafers. The method is based on comparing the observed spectrum with the control one and allows one to determine the moments of the end of the process, the irregularity of the process, the appearance of contaminants in the reactor chamber, etc. This also takes into account the distortions introduced during the passage of radiation through the window in the chamber, including contaminated. For a more accurate determination of the state of plasma in a reactor in US patent N 7695987 "Method for automatic determination of semiconductor plasma chamber matching and source of fault by comprehensive plasma monitoring", publ. 04/13/2010, a solution is proposed when it is controlled by a combination of optical spectral and high-frequency characteristics and is compared with previously accumulated measurement results in a control reactor using the method of multivariate analysis to detect significant deviations in the operating mode of the reactor. In the same row is the radiation monitoring system for different lines or other plasma parameters so that, based on the measurements, it is possible to quickly adjust the matching device to maintain the level of power deposited in the discharge, described in US patent N 8144329 "Low power RF tuning using optical and non -reflected power methods ", publ. 03/27/2012

Typical examples of the approach of combining optical and electrophysical measurements are spectrometry methods in which probe structures are used to record local optical radiation. For example, according to US patent N6034781 "Electro-optical plasma probe", publ. 03/07/2000, an optical fiber is additionally located inside the leg of the Langmuir wire probe to collect plasma radiation. The tip of the wire probe at the exit from the carrier leg has a U-shape so that its end overlaps the optical axis of the fiber, which allows you to collect radiation only from the area directly adjacent to the Langmuir probe. And according to US patent "Particle density measuring probe and particle density measuring equipment" N7782463, publ. 08/24/2010, the optical probe at the end of which hangs a holder with a reflector at some distance from the end of the fiber, so that the light from the fiber passes a given distance through the plasma and is reflected back. The absorption spectrum determines the concentration of atoms or molecules in the plasma. A system based on fiber-optic sensors used to construct the spatial distribution of the plasma radiation intensity for different parts of the spectrum is also described in US patent N7532322 "Method and apparatus for measuring electron density of plasma and plasma processing apparatus", publ. May 12, 2009

But despite the fact that optical signals are not used in optical probes and therefore they do not directly distort the distribution and movement of charged plasma particles, the very fact of placing a foreign body in the plasma is undesirable. Therefore, the systems of optical diagnostics are of the greatest interest, in which information on the spatial distributions of the intensity of the plasma glow is recorded using sensors located outside the plasma region or even outside the reactor. The complexity of implementing this approach lies in the fact that in order to extract information about the electron density from optical spectra, it is necessary to use complex algorithms and computer programs that implement numerical simulation of processes in plasma.

Known, for example, is a method for measuring the electron temperature of a plasma, described in US patent N8214173 "Plasma electron temperature measuring method and device", publ. 07/03/2012. According to this method, the density of these metastable states is determined by measuring the absorption changes of the metastable states of helium in the plasma tuned to the absorption line, and the plasma electron temperature is calculated from the added intensity of the radiative transitions from the laser-excited upper states. However, a significant drawback of this method is that it is inherently operable only in a plasma in which helium is present and effectively excited. But such gas mixtures have very limited use.

, v=0→ $$X^2{\displaystyle \sum {}_g^{+}}$$

, v=0 первой отрицательной системы $$N_2^{+}$$

с максимумом 391,4 нм, описанный в статье "Intensity ratio of spectral bands of nitrogen as a measure of electric field strength in plasmas", Journal of Physics D: Applied Physics, vol. 38 (2005), p.3894-3899. Недостатком этого способа является использование для диагностики сравнительно слабой линии электронно-возбужденного молекулярного иона $$N_2^{+}$$

, концентрация которого значительно меньше концентраций электронно-возбужденных состояний нейтральных молекул N₂.There is also a method of determining the magnitude of the electric field in a nitrogen plasma based on measuring the ratio of the intensities of the bands of spectral transitions C ³ P _u , v = 0 → B ³ P _g , v = 0 and C ³ P _u , v = 2 → B ³ P _g , v = 5 of the second positive system N ₂ with maxima of 337.1 nm and 394.3 nm, respectively, and spectral transition bands $$B^{}$$$${}^2{\displaystyle \sum {}_g^{+}}$$

, v = 0 → $${\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="00000006.png"\; state="\{\{state\}\}">X</figure-callout>}}^2{\displaystyle \sum {}_g^{+}}$$

, v = 0 of the first negative system $$N_{}^{}$$$${}_2^{+}$$

with a maximum of 391.4 nm, described in the article "Intensity ratio of spectral bands of nitrogen as a measure of electric field strength in plasmas", Journal of Physics D: Applied Physics, vol. 38 (2005), p. 3894-3899. The disadvantage of this method is the use for the diagnosis of a relatively weak line of an electronically excited molecular ion $$N_{}^{}$$$${}_2^{+}$$

whose concentration is much lower than the concentrations of electronically excited states of neutral molecules N ₂ .

) и второй (С³П_u→В³П_g) положительной систем азота, как это описано, например, в статье "Determining the electron temperature in inductively coupled nitrogen plasmas by optical emission spectroscopy with molecular kinetic effects", Physics of Plasmas, vol.12, 103501 (2005), а колебательная и газовая температуры определялись методом сравнения экспериментально наблюдаемых и смоделированных спектров.A known method of measuring electron density in a plasma by optical spectroscopy, which measures the intensities of spectral lines and bands corresponding to optical transitions with different upper and lower states of plasma particles in different spectral ranges, processes these intensities using numerical simulation of the plasma using a system of kinetic equations describing the processes of excitation and relaxation of such states of plasma particles, determine the electronic, vibrational, and gas plasma temperature and the ratio of the intensities of the two lines calculate the electron density in the plasma. In most detail, this method was developed for the diagnosis of nitrogen plasma. The case of low-pressure nitrogen plasma is described in the article "A novel method to determine electron density by optical emission", Physics of Plasmas, vol. 13, 123501 (2006). The case of high pressure nitrogen plasma is described in the article "Electron density and temperature measurement method by using emission spectroscopy in atmospheric pressure nonequilibrium nitrogen plasmas", Physics of Plasmas, vol. 13, 093501 (2006). In this case, the ratio of the total intensities of the first ( $$B^3{P}_g\to A^3{\displaystyle \sum {}_u^{+}}$$

) and the second (C ³ P _u → B ³ P _g ) positive nitrogen systems, as described, for example, in the article "Determining the electron temperature in inductively coupled nitrogen plasmas by optical emission spectroscopy with molecular kinetic effects", Physics of Plasmas, vol.12, 103501 (2005), and the vibrational and gas temperatures were determined by comparing the experimentally observed and simulated spectra.

The application of the method for measuring electron density in plasma by optical spectroscopy using the ratio of the intensities of two lines of the argon emission spectrum is described for the case of low-pressure plasma in the article “Using OES to determine electron temperature and density in low-pressure nitrogen and argon plasmas”, Plasma Sources Science and Technology, vol.17, 024002 (2008), and for the case of high pressure plasma in the article "Measurement of the electron density in atmospheric-pressure low temperature argon discharged by line-ratio method of optical emission spectroscopy", Journal of Physics D: Applied Physics, vol. 42, 142003 (2009).

The key element of this method, determining the conditions of its applicability, is the numerical model of plasma used. Various plasma modes and the models used for the implementation of this method and variants of its experimental implementations are discussed in detail in the article "Optical emission spectroscopy in low-temperature plasmas containing argon and nitrogen: determination of the electron temperature and density by the line-ratio method", Journal of Physics D: Applied Physics, vol. 43, 403001 (2010).

A significant disadvantage of this method is the limited application of it only to the conditions of a stationary plasma, which exists outside the surface of the near-electrode layer formed near the plasma processed by the plasma. Thus, the measurements described in the article "Determination of the electron temperature and density in the negative glow of a nitrogen pulsed discharge using optical emission spectroscopy", Journal of Physics D: Applied Physics, vol. 43, 015202 (2010), showed that the application of this method for points 6 mm and 18 mm away from the electrode with an interelectrode gap of 67 mm leads to significant errors in determining the electrophysical parameters of the plasma.

However, the measurement of plasma parameters precisely near the electrode on which the substrate under treatment is located is the most important factor determining the improvement of the technologies of plasma-chemical technologies of modern microelectronics. Moreover, as described, for example, in the articles "Phase resolved optical emission spectroscopy: a non-intrusive diagnostic to study electron dynamics in capacitive radio frequency discharges", Journal of Physics D: Applied Physics, vol. 43, 124016 (2010) and "Measurement of electron temperatures and electron energy distribution functions in dual frequency capacitively coupled CF4 / 02 plasmas using trace rare gases optical emission spectroscopy", J. Vac. Sci. Technol. A, vol. 27, 2009, p.1159-1165, the use of multipart capacitive discharges and complex gas mixtures to excite plasmas greatly complicates the coupling between the optical spectra and the dynamics of plasma electrons and requires the use of additional experimentally measured parameters, namely, measurements of spatial distributions the intensities of the lines of the optical emission spectrum and taking into account these distributions during numerical simulation of the near-electrode plasma regions.

Various methods are known for measuring spatial intensity distributions of lines of the optical emission spectrum. An optical system for collecting radiation with spatial resolution based on a narrowly collimated lens system is described in US patent N5654796 "Apparatus and method for mapping plasma characteristics", publ. 08/05/1997, the plasma Radiation is output through a small hole of small diameter in the wall of the reactor and a rotary mirror is sent to the optical receiver system. The rotation of the mirror allows you to change the direction of observation and calculate the three-dimensional distribution of the intensity of individual lines. The distribution of the radiation intensity along the line of sight during angular scanning is calculated using a correspondingly modified Abel transform.

Another variant of the optical system, which includes multiple points for monitoring the emission spectrum of the plasma along the reactor, is described in US patent N6677604 "Optical system and method for plasma optical emission analysis", publ. 04/30/2004, Of interest and method according to US patent N8144328 "Methods and apparatus for normalizing optical emission spectra", publ. 03/27/2012, in which it is proposed to use a calibrated light source, the radiation of which is recorded in one measurement cycle with the studied optical emission spectra of the plasma and is used to relative normalize the intensity of plasma radiation. This allows you to numerically measure changes in intensity when changing the regime or conditions in the plasma. For the purposes of the present invention, any optical system can be used to measure the intensity profiles of lines or bands of optical radiation selected for diagnosing the plasma electrophysical parameters in the required spectral range.

A method for measuring the spatial distribution of the electron density of a plasma, including measuring the intensities of spectral atomic lines or molecular bands with different wavelengths emitted from plasma regions of different coordinates, calculating the distribution of the intensity ratios of one or more pairs of selected lines along this coordinate, conducting a numerical simulation of the plasma and calculating the distribution of electron density in a plasma with respect to the intensities of the mentioned pair of lines is described in the article "Spatially resolved op tical emission spectroscopy investigation of E and H modes in cylindrical inductively coupled plasmas ", Journal of Physics D: Applied Physics, vol. 40 (2007), p. 5112-5116. In this article, using this method, we measured the spatial, with a step of 2 mm, intensity distribution of various spectral lines and bands in the spectral range from 330 to 850 nm along the axis of an inductively coupled RF discharge in an Ar / N ₂ mixture with an excitation frequency of 13.56 MHz at 36 mm long and at various levels of enclosed power.

This method of measuring electron density distributions in plasma is adopted as a prototype, however, the approach implemented in it is that a full-scale simulation is performed at each spatial point, which requires either the use of simplified numerical models, which can lead to insufficient measurement accuracy, or the use of redundant computational and temporary resources.

The aim of the invention is to achieve the necessary accuracy and reliability of the measurement of electron density distributions in the near-electrode region of RF discharges that determine the composition and main parameters of particles interacting with the processed substrate and, at the same time, lacking the above-mentioned disadvantages. The specified technical result is achieved through the use of numerical simulation of the uncalculated distribution of the intensity ratio of a pair of selected lines, and the unique spatial intensity profiles of such lines.

Disclosure of invention

The specified technical result is achieved as follows: the plasma chemical reactor is equipped with the necessary tools (optical system) to measure the intensity profiles of the emission radiation of the plasma in the selected spectral ranges. For example, if it is supposed to use the radiation of lines in the ultraviolet range for diagnostics, then the used optical elements and detectors must ensure the passage and registration of radiation in this range. In particular, the material from which such elements are made must be correctly selected. Or, if during the operation of a plasma chemical reactor, the optical properties of such elements are degraded, then measures can be taken to clean their working surfaces or additional calibration, since the degradation rate of the properties of the elements of the optical system for different wavelengths can occur unevenly and at different speeds. An important condition for achieving the technical result is to ensure spatial selectivity of radiation collection by the optical system, which excludes the exposure of the detector to radiation other than the observed plasma regions or reflections from the electrodes or walls of the reactor chamber. Also, the diagnostic system may include means for measuring the electrophysical parameters of plasma, including electron density, and in other ways. For example, voltage meters and electric currents flowing through a plasma. In addition, probes of any aforementioned or other design, means for microwave diagnostics of electron density, or even means for determining electron density by the method described above, the ratio of the intensities of a pair of lines, can be located in areas remote from the treated substrate that can be used to realize a stationary plasma density. The results obtained by such means of measuring the electrophysical parameters of plasma can be used to calibrate or bind the electron density profiles obtained by the present method in the plasma to reference values and thereby improve the accuracy of measurements. Then, an electric discharge is initiated in the plasma chemical reactor. In this case, the substrate to be processed can be located on the electrode or in another place of the reactor. Changing the settings of the optical system so that its detector sequentially or parallelly records the spatial distributions of the intensities selected for the implementation of the inventive method for the diagnosis of plasma atomic lines or molecular bands emitted by electronically excited plasma particles, obtain profiles of their intensity. After this, numerical simulation of the plasma is carried out, while the complexity and detail of the model are selected on the basis of the need to ensure the necessary or sufficient measurement accuracy, and also taking into account the composition of the gas mixture used and the burning regimes of the electric discharge that forms the plasma in the plasma-chemical reactor. A distinctive feature of the numerical simulation used in the implementation of the proposed method is the calculation of the intensity profiles of the lines or bands of the electric field profiles selected for plasma diagnostics in the interelectrode gap, and the electron density distribution in the plasma is calculated on the basis of such electric field profiles in the interelectrode gap. In order to further optimize the computational and temporal resources used in the numerical simulation for diagnostics, one selects lines or bands of the plasma emission spectrum so that their intensities are mainly determined by excitation of the corresponding states by direct electron impact or by fast cascade processes compared to the period of the RF field. In another embodiment of the inventive method, to increase the accuracy of measuring electron density, numerical modeling is performed using line or strip profiles with close values of the energy of the emitting levels and line profiles or lines with significantly different values of the energy of the emitting levels.

Brief Description of the Drawings

The invention is illustrated by drawings.

Figure 1 shows a diagram of an experimental setup.

Figure 2 presents the emission spectra of RF plasma in the Ar / N ₂ gas mixture, recorded near the electrode, in the near-electrode layer and the center of the interelectron gap.

Figure 3 shows the measured intensity profiles of various lines and bands in the emission spectrum of the plasma.

Figure 4 presents the results of numerical simulation of the distribution of the amplitude of the electric field in the interelectrode gap.

Figure 5 shows the distribution of electron density obtained by numerical simulation.

The implementation of the invention

The present invention can be carried out using various types of electric discharges, compositions of working gas mixtures in which an electric discharge plasma is excited, various methods for detecting the spectral and spatial distributions of optical plasma radiation.

In an example implementation of the proposed method used an experimental setup, a diagram of which is shown in Fig.1. A voltage with a frequency of 5.28 MHz was supplied from the high-frequency (HF) electric generator 1 to the electrode system of the chamber of the plasma-chemical reactor 3. The constant bias voltage was shunted by a frequency filter 2, which is a piece of a short-circuited cable of length λ / 4. The amplitude of the voltage applied to the electrodes and the current flowing through the plasma was measured according to the standard scheme using a capacitive voltage divider and a measuring resistor, respectively. Flat cylindrical electrodes with an area of 50 cm ^{2 were used} ; the width of the interelectrode gap in which the plasma was excited was 22 mm. The plasma was ignited in argon Ar with the addition of 1-3% nitrogen N ₂ at pressures in the range P = 0.5–9 Torr.

The intrinsic emission spectra of the plasma were recorded by a multichannel optical analyzer 4, the entrance slit of the polychromator of which was set parallel to the plane of the electrodes. To suppress flare, an additional slit diaphragm was placed between the discharge chamber and the analyzer. By moving the analyzer and the diaphragm slit, a spatial scanning of the discharge gap and obtaining distributions of optical plasma radiation with a resolution of ~ 1 mm were provided.

The plasma intrinsic emission spectra recorded at the center of the discharge gap in the layer and near the electrode are shown in FIG. 2.

The profiles of spatial distributions of the radiation intensity of the Ar lines λ = 696.5 nm and 811.5 nm with the energy of the emitting level E _exc = 13.3 and 13.1 eV, respectively, of the band (5 → 2) of the first positive (1+) nitrogen system with E _exc = 8.5 eV in the region λ = 668 nm and the band (0 → 2) of the second positive (2+) nitrogen system with E _exc = 11.0 eV in the region λ = 380 nm are shown in Fig. 3 (points) . The parameters of the electric discharge in the example embodiment of the method amounted to: pressure 2.5 Torr, voltage 110 V and current density 0.4 mA / cm ² . The spatial heterogeneity of the plasma gas temperature was estimated by the degree of change in the shape of the envelope of the strip (0 → 2) of the second positive nitrogen system with an unresolved rotational structure and a maximum of λ = 380 nm. The difference in gas temperature at the center of the discharge and at the electrode surface did not exceed 50 K. Its absolute calibration was not carried out.

The existing unambiguous relationship between the discharge current density measured in the experiment and the concentration of charged plasma particles, on the one hand, and the relationship of their profiles and the electric field profile in the plasma with intensity profiles of its luminescence, on the other hand, make it possible to calculate the magnitude and spatial dependences of the concentration of charged particles.

To restore the profile of the spatial distribution of the plasma electron density, we used numerical modeling of the radiation intensity profiles of Ar and N ₂ , followed by their comparison with those measured in the experiment. Various numerical models can be used for modeling, but the described example used a numerical model developed to describe elementary processes in multicomponent gas mixtures in RF discharges and described in the article "Experimental and Theoretical Studies of Radical Production in RF CCP Discharge at 81 MHz Frequency in Ar / CF ₄ and Ar / CHF ₃ Mixtures ", IEEE Trans, on Plasma Science, Vol. 37, p. 1683-1696, 2009.

The result of modeling the radiation intensity profiles Ar and N ₂ for the described case is presented in Fig. 3 (solid lines). Their good agreement with the experimental data indicates the adequacy of the numerical model used, since the large difference in the excitation energy of the emitting levels of Ar and N ₂ impose rather stringent restrictions on the variations in the distribution of electron density that are acceptable in the simulation. The physical meaning of this statement becomes clear from the resulting numerical simulation of the distribution of the amplitude of the electric field in the interelectrode gap shown in Figure 4. Indeed, taking into account the strong and substantially nonlinear dependence of the rate of excitation of the emitting levels on the amplitude of the electric field, which is directly related to the concentration gradient of charged plasma particles, the use of intensity profiles as a modeling object allows one to achieve the necessary accuracy and reliability of measuring electron density distributions. Corresponding to the radiation intensity profiles Ar and N ₂ shown in FIG. _{3, the} reconstructed plasma electron profile is shown in FIG. 5.

Claims (2)

1. A method of measuring the spatial distribution of the electron density of a plasma, including measuring the intensity of plasma radiation from different coordinate regions of the interelectrode gap at a wavelength corresponding to a spectral atomic line or molecular band, which is chosen so that the radiation intensity of such a line or band is mainly determined by the excitation of the radiating states by direct electron impact or fast cascade processes compared to the period of the RF field, with after uyuschim determine the spatial distribution of the electron density of the plasma method of numerical simulation of plasma. 2. The method according to p. 1, characterized in that the numerical simulation is carried out using line profiles or bands with close values of the energy of the emitting levels and profiles of lines or bands with significantly different values of the energy of the emitting levels.

Images