CF2 Density Research Articles

Development of a probe-type optical absorption spectroscopic system for spatially resolved CF2 density measurement in inductively coupled C4F8/Ar plasmas.

A spatially resolvable optical probe is developed for the measurement of absolute CF2 radical density in inductively coupled plasmas using the broadband ultraviolet absorption spectroscopic method. The probe-type system is implemented inside a low-pressure inductively coupled plasma chamber and is built using a series of optical fibers, feedthroughs, collimators, and an ultraviolet light emitting diode. Using the CF2 A(0, v2', 0) ← X(0, 0, 0) absorption spectra, the spatial profile of CF2 density in the remote region of an inductively coupled C4F8/Ar plasma is investigated. The CF2 density is found to decrease from 4.06 × 1013 cm-3, around the chamber center, to 2.66 × 1013 cm-3, near the chamber wall. The spatial profile of the measured CF2 density, layer thickness, and composition of deposited films are compared. Plasma properties such as electron temperature and ion density are also discussed.

Experimental and theoretical study of RF capacitively coupled plasma in Ar–CF4–CF3I mixtures

Radio frequency capacitively coupled plasma (RF CCP) sustained at RF frequencies of 27 MHz in an Ar–CF4–CF3I gas mixture is studied experimentally and theoretically. The RF CCP in Ar–CF4–CF3I is simulated by using a 1D hybrid particle-in-cell–fluid numerical model. We pay special attention to the changes in plasma structure and fluxes to the electrode, negative ion and neutral radical production with admixture of CF3I in Ar–CF4 plasma. With CF3I admixture the plasma becomes strongly electronegative as a result of the high electron attachment rate to the CF3I molecule. The atomic fluorine density becomes extremely low with addition of CF3I molecules due to the large volume loss in the reaction CF3I + F → CF3 + IF. Optical emission spectrometry data on CF2 emission at the wavelength of 250 nm indicate the essential sources of CF2 in CF3I-containing plasma in the studied conditions, although direct dissociation channels for the CF3I molecule with CF2 production have not been studied. Evaluation of CF2 density in Ar–CF4–CF3I plasma was first carried out on the basis of the actinometry technique and numerical simulation. The possible mechanism of ultra-low-k film damage in the studied conditions is also discussed.

Comparison of CF4 and C4F8 gas etching profiles by multiscale simulation

The CF4 and C4F8 gas etching profiles of oxide films were compared by multiscale simulation that comprises gas reaction, sheath, and surface reaction models. The densities of CF3, CF2, and CF radicals in CF4/Ar or C4F8/Ar gas were measured and compared with those obtained by simulation using the gas reaction model. From this comparison, the electron temperatures were determined to be 2.8 and 4.5 eV for CF4 and C4F8 gases, respectively. In the sheath model, the behavior of ions in a sheath was simulated for sheath lengths calculated from these electron temperatures. In the surface reaction model, we simulated the formation of a polymer and active layers by CF2 radicals, determined the depth of etching resulting from ion bombardment, and obtained the etching profiles of the oxide films. The profile of a contact hole with a depth of 820 nm and an aperture diameter of 160 nm was simulated. The results showed that the photoresist height was approximately 70 nm greater and the bowing diameter was approximately 10 nm smaller in the case of using C4F8 gas than in the case of using CF4 gas. This is because the CF2 density in the C4F8 gas is approximately 30 times higher than that in the CF4 gas and the polymer layer more strongly protects the underlying film. When the etching profiles were simulated with a fixed density of positive ions but with various CF2 density, the bottom diameter was constant but the bowing diameter changed for CF2 densities between 1013 and 1014 cm−3.

Absolute CF2 density and gas temperature measurements by absorption spectroscopy in dual-frequency capacitively coupled CF4/Ar plasmas

Broadband ultraviolet absorption spectroscopy has been used to determine the CF2 radical density in dual-frequency capacitively coupled CF4/Ar plasmas, using the CF2 Ã1B1←X̃1A1 system of absorption spectrum. The rotational temperature of ground state CF2 and excited state CF was also estimated by using Ã1B1←X̃1A1 system and B2Δ−X2Π system, respectively. The translational gas temperature was deduced from the Doppler width of the Ar*(3P2) and Ar*(3P0) metastable atoms absorption line by using the tunable diode laser absorption spectroscopy. The rotational temperatures of the excited state CF are about 100 K higher than those of ground state CF2, and about 200 K higher than the translational gas temperatures. The dependences of the radical CF2 density, electron density, electron temperature, rotational temperature, and gas temperature on the high frequency power and pressure have been analyzed. Furthermore, the production and loss mechanisms of CF2 radical and the gas heating mechanisms have also been discussed.

Spectroscopy diagnostic of dual-frequency capacitively coupled CHF3/Ar plasma

A combined spectroscopic method of absorption, actinometry, and relative optical emission intensity is employed to determine the absolute CF2 density, the relative F and H densities, H atom excitation temperature and the electron density in dual-frequency (60/2 MHz) capacitively coupled CHF3/Ar plasmas. The effects of different control parameters, such as high-frequency (HF) power, low-frequency (LF) power, gas pressure, gap length and content of CHF3, on the concentration of radical CF2, F, and H and excitation temperature are discussed, respectively. It is found that the concentration of CF2 is strongly dependent on the HF power, operating pressure and the proportion of CHF3 in feed gas, while it is almost independent of the LF power and the gap length. A higher concentration ratio of F to CF2 could be obtained in dual-frequency discharge case. Finally, the generation and decay mechanisms of CF2 and F were also discussed.

Spatial Distributions of Electron, CF, and CF2 Radical Densities and Gas Temperature in DC-Superposed Dual-Frequency Capacitively Coupled Plasma Etch Reactor Employing Cyclic-C4F8/N2/Ar Gas

On a plasma etch reactor for a wafer of 300 mm in diameter, the spatial distributions of the absolute densities of CF and CF2 radicals, electron density (ne), and the gas temperature (Tg) of N2 were measured employing the dual frequency of negative dc voltage superposed to a very high frequency (VHF) of 60 MHz capacitively coupled plasma (DS-2f-CCP) with the cyclic- (c-)C4F8/Ar/N2 gas mixture. The dc bias was superposed on the upper electrode with a frequency of 60 MHz. The distributions of electron and radical densities were uniform within a diameter of about 260 mm, and took a monotonic decay in regions outside a diameter of 260 mm on the reactor for 300 mm wafers in the reactor. It was found that only CF2 density at the radial position between 150 and 180 mm, corresponding to the position of the Si focus ring, dropped, while CF density took a uniform distribution over a diameter of 260 mm. Additionally, at this position, the rotational temperature of N2 gas increased to be 100 K larger than that at the center position. CF2 radical density was markedly affected by the modified surface loss probability of the material owing to coupling with surface temperature.

Role of CF2 in the etching of SiO2, Si3N4 and Si in fluorocarbon plasma

The CF2 density and etch rate of SiO2, Si3N4 and Si are investigated as a function of gas pressure and O2 flow rate in fluorocarbon plasma. As the pressure increases, the self-bias voltage decreases whereas the SiO2 etch rate increases. Previous study has shown that SiO2 etch rate is proportional to the self-bias voltage. This result indicates that other etching parameters contribute to the SiO2 etching. Generally, the CF2 radical is considered as a precursor for fluorocarbon layer formation. At a given power, defluorination of fluorocarbon under high-energy ion bombardment is a main source of fluorine for SiO2 etching. When more CF2 radical in plasma, SiO2 etch rate is increased because more fluorine can be provided. In this case, CF2 is considered as a reactant for SiO2 etching. The etch rate of Si3N4 and Si is mainly determined by the polymer thickness formed on its surface which is dominated by the CF2 density in plasma. Etching results obtained by varying O2 flow rate also support the proposition.

Role of chamber dimension in fluorocarbon based deposition and etching of SiO2 and its effects on gas and surface-phase chemistry

It is well understood that chamber geometry is an influential factor governing plasma processing of materials. Simple models suggest that a large fraction of this influence is due to changes in basic plasma properties, namely, density, temperature, and potential. However, while such factors do play an important role, they only partly describe the observed differences in process results. Therefore, to better elucidate the role of chamber geometry in this work, the authors explore the influence of plasma chemistry and its symbiotic effect on plasma processing by decoupling the plasma density, temperature, and potential from the plasma-surface (wall) interactions. Specifically, a plasma system is used with which the authors can vary the chamber dimension so as to vary the plasma-surface interaction directly. By varying chamber wall diameter, 20–66cm, and source-platen distance, 4–6cm, the etch behavior of SiO2 (or the deposition behavior of fluorocarbon polymer) and the resulting gas-phase chemistry change significantly. Results from in situ spectroscopic ellipsometry show significant differences in etch characteristics, with etch rates as high as 350nm∕min and as low as 75nm∕min for the same self-bias voltage. Fluorocarbon deposition rates are also highly dependent on chamber dimension and vary from no net deposition to deposition rates as high as 225nm∕min. Etch yields, however, remain unaffected by the chamber size variations. From Langmuir probe measurements, it is clear that chamber geometry results in significant shifts in plasma properties such as electron and ion densities. Indeed, such measurements show that on-wafer processes are limited at least in part by ion flux for high energy reactive ion etch. However, in situ multipass Fourier transform infrared spectroscopy reveals that the line-averaged COF2, SiF4, CF2, and CF3 gas-phase densities are also dependent on chamber dimension at high self-bias voltage and also correlate well to the CFx overlayer stoichiometry under deposition conditions.

Transient and stable species kinetics in pulsed cc-rf CF4/H2 plasmas and their relation to surface processes

Fluorocarbon plasmas are widely used in applications and as model systems for fundamental investigations of complex plasmas. In recent years pulsing of the rf discharge has been used as an additional parameter for process control, because many plasma parameters, e.g. densities and temperatures, become time dependent when the rf power is modulated. In this work tunable diode laser absorption spectroscopy in the mid-IR (IR-TDLAS) was applied to measure time-resolved densities of the transient species CF and CF2 and that of the stable product C2F4 in pulsed CF4/H2 asymmetrical capacitively coupled radio-frequency plasmas at 13.56 MHz. Simultaneously, the thickness of amorphous thin fluorocarbon films (a-C:F) on the powered electrode was determined by means of in situ ellipsometry. Therefore, it was possible to study the correlation between gas phase species and thin film formation. The decay curves of the CF and CF2 densities in the off-phase of the pulsed rf plasma were fitted with a combination of first and second order processes involving the loss processes of these radicals in the gas phase and at the surfaces. Particularly, in the plasma off-phase, the loss of CF2 radicals forming C2F4 was found to be dominant in the CF2 kinetics, but of minor importance for C2F4 production. Plasma process parameters such as total pressure, gas composition, power and power modulation were varied to investigate the interaction between gas phase species and surfaces.

Dependences of Plasma Parameters on Dilution Gas Content in Inductively Coupled C2F6/Ar and C2F6/O2 Discharges

Electron energy probability functions (EEPFs) are investigated in inductive C2F6/Ar and C2F6/O2 discharges. The structure of the EEPFs changes from bi-Maxwellian to Druyvesteyn-like distributions through a Maxwellian one with increasing Ar content, whereas the EEPFs form the bi-Maxwellian structure at any oxygen content except for pure oxygen. Dependences of both the fluorine atom density measured by actinometry and the relative variation of CF2 density determined from an optical emission at 251.9 nm on the dilution gas content are also investigated. The densities of F and CF2 are independent of the Ar content, while they depend strongly on the oxygen content.

CF and CF2 radical kinetics and transport in a pulsed CF4 ICP

The kinetics of CF and CF2 radicals (in their electronic ground states) was investigated using space and time resolved laser induced fluorescence (LIF) in an inductively coupled discharge in CF4 gas at 33 mTorr, both in the steady state and in pulsed plasmas. The gas temperature was determined from the rotationally resolved LIF excitation spectra of the CF radical, reaching about 800 K in the centre of the chamber. The densities were put on an absolute scale with the aid of broad-band UV absorption spectroscopy. The density profiles were used to deduce the radical diffusive transport, and therefore the localized (gas-phase and surface) net production or destruction rates. However, when gas heating is significant, the radical transport becomes more complex due to gas density gradients and thermo-diffusion. In the afterglow, gas convection induced by gas cooling and contraction can play a significant role. The CF and CF2 axial fluxes estimated from measured density and temperature profiles indicate that these species are emitted from the top and bottom surfaces, and are destroyed in the gas phase at high rates (≈900 s−1 and 1500 s−1, respectively): for CF2 this process appears to be electron-impact induced. In the afterglow, the CF2 density at the reactor centre increases abruptly, reaching four times the steady-state value after 3 ms, followed by a slow decay (100 s−1). This phenomenon is the combined result of the return of CF2-rich gas as the central region cools and contracts, along with gas-phase reactions producing CF2. The central CF density shows only a slight initial increase in the afterglow, followed by a rapid decay (600 s−1), due to gas-phase chemical reactions, e.g. with F2 or CF3, converting it to CF2, processes which are also likely to be important in the steady state.

Monte Carlo sensitivity analysis of CF2 and CF radical densities in a c-C4F8 plasma

A Monte Carlo sensitivity analysis is used to build a plasma chemistry model for octacyclofluorobutane (c-C4F8) which is commonly used in dielectric etch. Experimental data are used both quantitatively and qualitatively to analyze the gas phase and gas surface reactions for neutral radical chemistry. The sensitivity data of the resulting model identifies a few critical gas phase and surface aided reactions that account for most of the uncertainty in the CF2 and CF radical densities. Electron impact dissociation of small radicals (CF2 and CF) and their surface recombination reactions are found to be the rate-limiting steps in the neutral radical chemistry. The relative rates for these electron impact dissociation and surface recombination reactions are also suggested. The resulting mechanism is able to explain the measurements of CF2 and CF densities available in the literature and also their hollow spatial density profiles.

Use of the ultraviolet absorption spectrum of CF2 to determine the spatially resolved absolute CF2 density, rotational temperature, and vibrational distribution in a plasma etching reactor.

Broadband ultraviolet absorption spectroscopy has been used to determine CF(2) densities in a plasma etch reactor used for industrial wafer processing, using the CF(2) A (1)B(1)<--X (1)A(1) absorption spectrum. Attempts to fit the experimental spectra using previously published Franck-Condon factors gave poor results, and values for the higher vibrational levels of the A state [(0,v(2),0), with v(2) (')>6] from the ground state were missing; hence new values were calculated. These were computed for transitions between low-lying vibrational levels of CF(2) X (1)A(1) to vibrational levels of CF(2) A (1)B(1) (v(1) ('),v(2) ('),0) up to high values of the vibrational quantum numbers using high level ab initio calculations combined with an anharmonic Franck Condon factor method. The Franck Condon factors were used to determine the absorption cross sections of CF(2) at selected wavelengths, which in turn were used to calculate number densities from the experimental spectra. Number densities of CF(2) have been determined in different regions of the plasma, including the center of the plasma and outside the plasma volume, and CF(2) rotational temperatures and vibrational energy distributions were estimated. For absorption spectra obtained outside the confined plasma volume, the CF(2) density was determined as (0.39+/-0.08)x10(13) molecule cm(-3) and the vibrational and rotational temperatures were determined as 303 and 350 K, respectively. In the center of the plasma reactor, the CF(2) density is estimated as (3.0+/-0.6)x10(13) molecules cm(-3) with T(rot) approximately 500 K. The fitted vibrational distribution in the CF(2) ground state corresponds to two Boltzmann distributions with T(vib) approximately 300 and T(vib) approximately 1000 K, indicating that CF(2) molecules are initially produced highly vibrationally excited, but are partially relaxed in the plasma by collision.

Surface and gas-phase observations of Ar-diluted c-C4F8 plasma by using real-time infrared spectroscopy and planar laser-induced fluorescence

Using infrared spectroscopy with attenuated-total-reflection methods and planar laser-induced fluorescence, we examined the relationship between the thickness of amorphous fluorinated carbon (a-C:F) film on the surface of a substrate and the distribution of CF2 radical density in an inductively coupled plasma (ICP) generated from an argon-diluted octafluoro-cyclo-butane (c-C4F8) gas. The axial profiles of the CF2 radical density were consistently concave, maximally close to the surface, and minimally present around the bulk plasma. The CF2 density initially increased with increasing the ICP power and then decreased as the plasma density exceeded ∼0.5×1011 cm−3, while the thickness of the a-C:F films deposited on a nonbiased surface consistently increased. When bias power was applied, the thickness of the a-C:F film began to decrease constantly due to the sputtering effect while the CF2 density exhibited a stepwise increase, followed by a gradual decrease in response to the reduction in the thickness of the a-C:F film and its vanishing. The amount of the stepwise increase in the CF2 density depended only on the rate of sputtering of the a-C:F film, and it was negligible at high c-C4F8/Ar ratios. The positive and negative correlations revealed here between the thickness of the a-C:F film and the density of CF2 radicals are discussed in comparison with mechanisms proposed in other studies. Electron-impact-induced destruction of CF2 radicals is demonstrated as the dominant factor affecting the distribution of CF2 radicals.

Monitoring of CF and CF2 Number Density by Optical Emission Intensity Corrected using Probe-Measured Plasma Parameters

The use of the CF and CF2 optical emission (OE) intensity corrected by the probe-measured electron density and temperature is proposed for the estimation of CF and CF2 relative number density in low-pressure, high-density fluorocarbon plasmas. The estimation of the CF and CF2 density by the corrected OE is mainly based on an assumption that the electron energy distribution function near the threshold energy of these OE is approximated by a Maxwellian using the probe-measured electron temperature. The comparison of the corrected OE intensity with CF and CF2 laser-induced fluorescence (LIF) intensity, which represents the CF and CF2 number density, confirms good proportionality between the LIF and the corrected OE intensity. Thus, the proposed procedure is expected to be useful for the monitoring of CF and CF2 radicals.

Effect of changing the electrode gap on the spatial and electrical properties of O2/CF4 plasmas

Planar laser-induced fluorescence (PLIF) measurements were made to determine two-dimensional spatial maps of CF2 density as a chemical marker of plasma uniformity in 9% O2/91% CF4 chamber-cleaning plasmas. Broadband optical emission and discharge current and voltage measurements were also made and compared to the PLIF results. Measurements were made in a capacitively coupled Gaseous Electronics Conference Reference Cell as pressure was varied from 13 (100 mTorr) to 133 Pa (1000 mTorr) and electrode gap was varied from 2.25 to 0.5 cm. Smaller gaps resulted in increased radial uniformity and extended the pressure range over which CF2 density and electrical properties remain insensitive to pressure. These effects are explained by a decrease in the bulk plasma resistance at narrower electrode gaps, which changes the path that rf current takes through the discharge, thus affecting where electron heating and CF2 production occur. These results provide insight into the optimization of chamber-cleaning processes and reactors as well as provide necessary data for validation of plasma simulations.

Fluorocarbon-based plasma etching of SiO2: Comparison of C4F6/Ar and C4F8/Ar discharges

A gas phase and surface chemistry study of inductively coupled plasmas fed with C4F6/Ar and C4F8/Ar intended for SiO2 etching processes was performed. Adding Ar to those fluorocarbon gases results in a strong increase of the ion current, by up to a factor of 5 at 90% Ar relative to the pure fluorocarbon gases. The fluorocarbon deposition rate is higher for C4F6/Ar than for C4F8/Ar, whereas the fluorocarbon etching rate is lower, and both quantities decrease as the amount of Ar is increased. For both C4F6/Ar and C4F8/Ar, the CF2 density is more than an order of magnitude greater than the CF density. The CF2 partial pressure decreases as more Ar is added to the C4F6/Ar plasmas. A comparison of these data with corresponding results obtained with C4F8/Ar shows that the CF2 partial pressure in C4F6 is higher for Ar-lean gas mixture than for C4F8/Ar. This remains true up to 40% Ar. Above 40% Ar the CF2 partial pressure in C4F8 is higher than for C4F6. The CF and COF2 partial pressures in C4F8 are higher than for C4F6. The SiO2 etch rate is higher for C4F8/Ar than for C4F6/Ar. This may be attributed in part to the higher F/C ratio of the steady-state fluorocarbon film formed on SiO2 surfaces for C4F8/Ar which was determined by x-ray photoemission spectroscopy (XPS). The etching selectivity of SiO2 over resist and silicon is increased by the addition of Ar to the fluorocarbon gases. Overall, the SiO2/resist and SiO2/Si etching selectivity are higher for C4F6/Ar (i.e., 4 and 9, respectively) at 90% Ar than for C4F8/Ar (i.e., 2 and 5, respectively) at 90% Ar and otherwise identical conditions. Both ellipsometry and XPS measurements show that the steady-state fluorocarbon layer thickness is greater for C4F6/Ar (∼4 nm) than for C4F8/Ar (∼2.8 nm). Argon addition leads to a strong decrease of the fluorine content of the steady-state fluorocarbon layers on both Si and SiO2 surfaces relative to films produced in pure fluorocarbon discharges, and this effect is related to the increase of the SiO2/Si and SiO2/resist etching selectivity.

Planar Laser-Induced Fluorescence of Fluorocarbon Radicals in Oxide Etch Process Plasma

Planar laser-induced fluorescence (PLIF) has been employed to clarify the two-dimensional distribution of CF2 radicals in an inductively coupled plasma (ICP). Axial profiles and contour plots for CF2 radical density were monitored for CF4 (100%), CF4/Ar (=10%/90%) and C4F8/Ar (=10%/90%) plasmas in order to examine the roles of electron impact, fluorocarbon film deposition and ion bombardment in the production and loss mechanisms of CF2 radicals. Concave distributions which were minimal around the bulk plasma were observed for the CF4/Ar and C4F8/Ar plasmas with high electron density, demonstrating that the dominant mechanism is destruction in the bulk plasma by electron impact, rather than production on the surfaces of the substrate and chamber wall exposed to ion bombardment. Substrate biasing resulted not in the enhancement of the surface production but rather in the decrease of the CF2 density, associated with the fluorocarbon film thickness reduction due to the sputtering effect of ion bombardment.

Gas-phase studies in inductively coupled fluorocarbon plasmas

Quantitative results from infrared laser absorption spectroscopy (IRLAS) of CF and CF2 radicals and COF2 products in inductively coupled plasmas fed with C2F6, CHF3 and C4F8 are presented and compared with results simultaneously obtained by mass spectrometry and optical emission spectroscopy. These plasma gas-phase analysis results are discussed and compared to fluorocarbon deposition and etching rates resulting from plasma–surface interactions at the substrate. It is found that COF2 species are being formed (1) during O2 plasma cleaning of a fluorocarbon contaminated reactor and (2) during SiO2 etching in fluorocarbon plasmas, which in this work occurred at the quartz coupling window as a result of capacitive coupling between the induction coil and the plasma. IRLAS results on CF and CF2 densities are compared to fluorocarbon deposition and etching rates and it is found that low CF and/ or CF2 density does not necessarily translate into a low fluorocarbon deposition rate. A relatively high deposition rate can be achieved at conditions with a high ion current density and low CF and CF2 densities.

Modeling the plasma chemistry of C2F6 and CHF3 etching of silicon dioxide, with comparisons to etch rate and diagnostic data

A detailed chemical reaction mechanism is reported that describes the C2F6 and CHF3 plasma etching of silicon dioxide, which is widely used in the fabrication of microelectronic devices. The gas-phase part of the C2F6 mechanism involves 28 species and 132 reactions, while the surface part involves 2 materials, 6 species, and 85 reactions. Rate parameters are generally taken from independent studies in the literature, or estimated from rates measured for related species. Zero-dimensional simulations using these mechanisms compare well with a large body of etch rate and diagnostic measurements in three different high-density plasma reactors. The diagnostic measurements include electron and negative ion absolute densities, CF, CF2, and SiF densities, gas temperatures, and ion current densities. An analysis of the dominant reaction paths shows the importance of gas-phase electron impact reactions and the need to include reactions of the etch-product species. On the surface, the etching reactions are dominated by ion-assisted processes; spontaneous etching by F atoms is relatively unimportant.