Gyeo-Re Lee

Angular dependence of SiO2 etching in a fluorocarbon plasma

SiO2 etch rates in a CF4 plasma were obtained at various surface angles using a Faraday cage with pinholes on the upper plane through which ions are incident on the substrate fixed at various angles inside the cage. The reactive ion etching experiments were performed at 5 mTorr in a wide bias-voltage range from −100 to −800 V. The normalized etch-yield curves showed virtually the same angular dependence regardless of the ion incident energy. The curve shape was similar to that of physical sputtering except that the ratio of the maximum yield to that at 0° was as low as about 1.3. This is the deviation from the cosine dependence characteristic of chemical sputtering, which is the main mechanism of SiO2 etching in a fluorocarbon plasma. This deviation is partly attributed to the fluorocarbon polymer film, which existed as a few monolayers-thick film on the substrate surface at low angles near 0° but as a submonolayer at high angles between 45° and 75°. The effect of the film-coverage difference on the devia...

Redeposition of etch products on sidewalls during SiO2 etching in a fluorocarbon plasma. I. Effect of particle emission from the bottom surface in a CF4 plasma

The effect of etch-product redeposition on sidewall properties during the etching of step-shaped SiO2 patterns in a CF4 plasma was examined using a Faraday cage located in a transformer coupled plasma etcher. Sidewall properties were observed for two cases: with and without particles emitted from the bottom surface in normal contact with the sidewall. Particles sputtered from the bottom surface were redeposited on the sidewall, which contributes to the formation of a passivation layer on the surface of the latter. The passivation layer consisted of silicon oxide, SixOy, and fluorocarbon, CxFy, the latter comprising the major species. Ar plasma experiments confirmed that CxFy or a fluorocarbon polymer must be present on the sidewall in order for the SixOy species to be deposited on the surface. The redeposited particles, which were largely F-deficient fluorocarbon species, as evidenced by x-ray photoelectron spectroscopy analyses, functioned as precursors for fluorocarbon polymerization, resulting in a rou...

Angular dependence of the redeposition rates during SiO2 etching in a CF4 plasma

The angular dependence of the redeposition rates during SiO2 etching in a CF4 plasma was studied using three types of Faraday cages located in a transformer coupled plasma etcher. The SiO2 substrates were fixed on sample holder slopes that have different angles to the cathode. The substrate was subjected to one of three processes depending on the design of the Faraday cage, i.e., redeposition of sputtered particles from the SiO2 bottom surface (case I), substrate etching by incident ions (case II), or simultaneous etching and redeposition (case III). Both the redeposition and the etch rates were measured by changing the substrate–surface angle and the self-bias voltage in the range of −100 to −800 V. The redeposition-only rates (case I) at −450 and −800 V closely followed the quadratic curve of the angle whereas the rates at −100 V followed the cubic curve, indicating different mechanisms of the bottom SiO2 etching depending on the energy regimes. The steep increase of the redeposition rate with the angle...

Angular dependence of etch rates in the etching of poly-Si and fluorocarbon polymer using SF6, C4F8, and O2 plasmas

The dependences of etch rates on the angle of ions incident on the substrate surface in four plasma/substrate systems that constitute the advanced Bosch process were investigated using a Faraday cage designed for the accurate control of the ion-incident angle. The four systems, established by combining discharge gases and substrates, were a SF6/poly-Si, a SF6/fluorocarbon polymer, an O2/fluorocarbon polymer, and a C4F8/Si. In the case of SF6/poly-Si, the normalized etch rates (NERs), defined as the etch rates normalized by the rate on the horizontal surface, were higher at all angles than values predicted from the cosine of the ion-incident angle. This characteristic curve shape was independent of changes in process variables including the source power and bias voltage. Contrary to the earlier case, the NERs for the O2/polymer decreased and eventually reached much lower values than the cosine values at angles between 30° and 70° when the source power was increased and the bias voltage was decreased. On the other hand, the NERs for the SF6/polymer showed a weak dependence on the process variables. In the case of C4F8/Si, which is used in the Bosch process for depositing a fluorocarbon layer on the substrate surface, the deposition rate varied with the ion incident angle, showing an S-shaped curve. These characteristic deposition rate curves, which were highly dependent on the process conditions, could be divided into four distinct regions: a Si sputtering region, an ion-suppressed polymer deposition region, an ion-enhanced polymer deposition region, and an ion-free polymer deposition region. Based on the earlier characteristic angular dependences of the etch (or deposition) rates in the individual systems, ideal process conditions for obtaining an anisotropic etch profile in the advanced Bosch process are proposed.The dependences of etch rates on the angle of ions incident on the substrate surface in four plasma/substrate systems that constitute the advanced Bosch process were investigated using a Faraday cage designed for the accurate control of the ion-incident angle. The four systems, established by combining discharge gases and substrates, were a SF6/poly-Si, a SF6/fluorocarbon polymer, an O2/fluorocarbon polymer, and a C4F8/Si. In the case of SF6/poly-Si, the normalized etch rates (NERs), defined as the etch rates normalized by the rate on the horizontal surface, were higher at all angles than values predicted from the cosine of the ion-incident angle. This characteristic curve shape was independent of changes in process variables including the source power and bias voltage. Contrary to the earlier case, the NERs for the O2/polymer decreased and eventually reached much lower values than the cosine values at angles between 30° and 70° when the source power was increased and the bias voltage was decreased. On th...

Sidewall-angle effect on the bottom etch profile in SiO2 etching using a CF4 plasma

The dependence of the bottom etch profile on the sidewall angle in the CF4 plasma etching of an SiO2 film was investigated using a Faraday cage, which allowed ions to impinge on the sidewall at specified angles. The bottom etch profile obtained at −500 V was not affected by the sidewall when the angle between the sidewall and the bottom surface was 45° but showed microtrenching when the angle was greater than 60°. The microtrench depth increased until the angle reached 80°, beyond which the local etch rate was drastically reduced to allow the corner rounding of the bottom profile. As the sidewall angle increases, the etch rate of the bottom surface near the corner is controlled by two opposing factors. The decreasing number of incident ions on the sidewall surface and the increasing shadowing of the bottom surface from ions and neutrals by the sidewall contribute to the reduced etch rate, whereas the increasing kinetic energy and the narrower ion-angular distribution of the ions reflected from the sidewal...

Angular dependence of Si3N4 etch rates and the etch selectivity of SiO2 to Si3N4 at different bias voltages in a high-density C4F8 plasma

The dependence of Si3N4 etch rates and the etch selectivity of SiO2 to Si3N4 on ion-incident angles was studied for different bias voltages in a high-density C4F8 plasma. A Faraday cage and specially designed substrate holders were used to accurately control the angles of incident ions on the substrate surface. The normalized etch yield (NEY), defined as the etch yield obtained at a given ion-incident angle normalized to that obtained on a horizontal surface, was unaffected by the bias voltage in Si3N4 etching, but it increased with the bias voltage in SiO2 etching in the range of −100to−300V. The NEY changed showing a maximum with an increase in the ion-incident angle in the etching of both substrates. In the Si3N4 etching, a maximum NEY of 1.7 was obtained at 70° in the above bias voltage range. However, an increase in the NEY at high ion-incident angles was smaller for SiO2 than for Si3N4 and, consequently, the etch selectivity of SiO2 to Si3N4 decreased with an increase in the ion-incident angle. The ...

Dependences of bottom and sidewall etch rates on bias voltage and source power during the etching of poly-Si and fluorocarbon polymer using SF6, C4F8, and O2 plasmas

The dependences of bottom and sidewall etch rates on the bias voltage and source power in four plasma/substrate systems constituting the advanced Bosch process were investigated using a Faraday cage and a step-shaped substrate specially designed for the accurate observation of lateral and vertical etch rates. The four systems, established by combining discharge gases and substrates, were SF6/poly-Si, SF6/fluorocarbon polymer, O2/fluorocarbon polymer, and C4F8/Si. For etch systems using SF6/poly-Si, SF6/polymer, and O2/polymer, the degree of anisotropy showed a higher dependence on the bias voltage than on the source power. As the bias voltage was increased, the degree of anisotropy obtained in SF6/poly-Si decreased while that for the SF6/polymer and O2/polymer increased. The contribution of spontaneous etching by reactive radicals to the etch rates increased in the order of SF6/polymer<O2/polymer<SF6/poly-Si, while that of ion-enhanced chemical etching by ions to the degree of anisotropy increased in the ...

Angular dependence of SiO2 etch rate at various bias voltages in a high density CHF3 plasma

The dependence of the SiO2 etch rate on the angle of ions incident on the substrate surface was studied over a bias voltage range from −20 to −600 V in a high-density CHF3 plasma using a Faraday cage to control the ion incident angle. The effect of the bottom plane on the sidewall etching was also examined. Differences in the characteristics of the etch rate as a function of the ion angle were observed for different bias voltage regions. When the absolute value of the bias voltage was smaller than 200 V, the normalized etch rate (NER) defined as the etch rate normalized by the rate on the horizontal surface, changed following a cosine curve with respect to the ion incident angle, defined as the angle between the ion direction and the normal of the substrate surface. When the magnitude of the bias voltage was larger than 200 V, the NER was deviated to higher values from those given by a cosine curve at ion angles between 30° and 70°, and then drastically decreased at angles higher than 70° until a net depo...

Contribution of bottom-emitted radicals to the deposition of a film on the SiO2 sidewall during CHF3 plasma etching

The contribution of radicals emitted from the bottom of an etched pattern to the deposition of a film on the sidewall in a CHF3 plasma was investigated using a specially designed experimental system. In the experimental system, a target (1), simulating the bottom of an etched pattern, emits various types of radicals after collision with incident ions. Among the radicals and ions that are incident on a target (2), which is located in close proximity to the target (1), to simulate the sidewall of an etched pattern, ions were prevented from arriving at the surface by placing a double grid structure in front of the target (2). The deposition rate and the composition of the film formed on the target (2) were measured for the cases where the target (1) was made of different materials and subject to impingement by ions at different bias voltages. In another set of experiments, a rectangular tube of a predetermined length was located in front of the target (2) for the purpose of estimating the probability of radi...

Etch Characteristics of Silsesquioxane-based Low Dielectric Constant Material in Fluorocarbon Plasma

The etch characteristics of hydrido-organo-siloxane-polymer (HOSP), a typical silsesquioxane-based low-dielectric material, were compared with those of silicon dioxide in CF4 and CHF3 plasmas. The etch-rate ratios of the two materials are more significantly affected by the types and pressure of plasma gases than by plasma bias voltage, indicating that the relative etch rates are determined largely by the density of radicals than by the energy of ions incident on the substrate surface. The etching of silsesquioxane is accompanied by an increase in the CF2 radical density and a significant decrease in F radical density, indicating that silsesquioxane is etched via the sequential dissociation of Si–CH3 and cage-like Si–O bonds by reaction with F radicals. Based on the findings herein, we propose that the relative amounts of cage-like and network Si–O bonds remaining in silsesquioxane after etching can be controlled by adjusting the parameter, (F radical density)2/(CF2 radical density).