Inverse Reactive Ion Etching Lag Research Articles

Assessing neutral transport mechanisms in aspect ratio dependent etching by means of experiments and multiscale plasma modeling

Since the onset of pattern transfer technologies for chip manufacturing, various strategies have been developed to circumvent or overcome aspect ratio dependent etching (ARDE). These methods have, however, their own limitations in terms of etch non-idealities, throughput or costs. Moreover, they have mainly been optimized for individual in-device features and die-scale patterns, while occasionally ending up with poor patterning of metrology marks, affecting the alignment and overlay in lithography. Obtaining a better understanding of the underlying mechanisms of ARDE and how to mitigate them therefore remains a relevant challenge to date, for both marks and advanced nodes. In this work, we accordingly assessed the neutral transport mechanisms in ARDE by means of experiments and multiscale modeling for SiO2 etching with CHF3/Ar and CF4/Ar plasmas. The experiments revealed a local maximum in the etch rate for an aspect ratio around unity, i.e. the simultaneous occurrence of regular and inverse reactive ion etching lag for a given etch condition. We were able to reproduce this ARDE trend in the simulations without taking into account charging effects and the polymer layer thickness, suggesting shadowing and diffuse reflection of neutrals as the primary underlying mechanisms. Subsequently, we explored four methods with the simulations to regulate ARDE, by varying the incident plasma species fluxes, the amount of polymer deposition, the ion energy and angular distribution and the initial hardmask sidewall angle, for which the latter was found to be promising in particular. Although our study focusses on feature dimensions characteristic to metrology marks and back-end-of-the-line integration, the obtained insights have a broader relevance, e.g. to the patterning of advanced nodes. Additionally, this work supports the insight that physisorption may be more important in plasma etching at room temperature than originally thought, in line with other recent studies, a topic on which we recommend further research.

Atomic-Scale Cellular Model and Profile Simulation of Si Etching: Analysis of Profile Anomalies and Microscopic Uniformity

Reactive ion etching (RIE) has been used in the manufacture of semiconductor integrated circuit devices. However, the formation mechanisms of profile anomalies and microscopic uniformity have been poorly understood until now. In this paper, we focus on the analysis of formation mechanisms of profile anomalies and microscopic uniformity during plasma etching of Si in Cl2 plasmas, using our own atomic-scale cellular model (ASCeM). The numerical results indicated that high neutral-to-ion flux ratios result in microtrench formation. Moreover, RIE lag tends to occur at low neutral-to-ion flux ratios (<50), whereas inverse RIE lag occurs at high neutral-to-ion flux ratios in typical low-pressure and high-density plasmas. In particular, the etch rates for narrow patterns (<70 nm) increase significantly with increasing neutral-to-ion flux ratio. The synergistic effects between ion-enhanced etching and neutral shadowing in microstructural features play a significant role in the formation of profile anomalies.

Simulation of SiO2 and Si feature etching for microelectronics and microelectromechanical systems fabrication: A combined simulator coupling modules of surface etching, local flux calculation, and profile evolution

A combined simulator linking gas flux data from a plasma reactor (experimental or simulated) to the feature profile evolution during etching/deposition processes is described. This combined simulator results from the coupling of surface etch, local flux calculation, and profile evolution modules. It is a modular tool, in the sense that different phenomena, surface models, and structures can be incorporated. In this work the combined simulator is applied in two processes: (a) in SiO2 feature etching to simulate reactive ion etching (RIE) lag and inverse RIE lag and (b) in etching of high aspect ratio Si trenches using the Bosch process. The profile evolution algorithm of the combined simulator, namely, the level set method, is applied to a process where etching and deposition occur, and tracking of two materials (Si and polymer formed during the Bosch deposition step) is needed.

Effect of sidewall roughness on the bottom etch properties of an SiO2 trench produced in a CF4 plasma

The effect of sidewall roughness on the bottom etch profile and the etch rate of an SiO2 trench produced in a CF4 plasma were examined using a specially designed apparatus that involved a Faraday cage, which permits the observation of microfeatures on an enlarged scale under practical processing conditions. A microtrench was produced on the bottom of a trench-shaped structure with a low aspect ratio when its sidewalls were smooth because ions reflected from the sidewall enhanced the bottom etch rate near the sidewall. However, the microtrench was not produced in a trench-shaped structure with a high aspect ratio and smooth sidewalls because the effects of ions reflected from two sidewalls were overlapped, nor in step-shaped and trench-shaped structures with rough sidewalls because the contribution of reflected ions to the bottom etch rate was negligible. The overlap in ion reflection effects in a trench having a high aspect ratio and smooth sidewalls was responsible for the inverse reactive ion etching lag phenomenon observed in this study.

Geometrical pattern effect on silicon deep etching by an inductively coupled plasma system

The etching rate in silicon deep reactive ion etching (RIE) is related to pattern geometry and a frequently seen defect, RIE lag, appears in feature sizes up to hundreds of micrometers. Different feature dimensions of rectangles, squares and circles/doughnuts are designed to realize how the geometrical pattern affects RIE lag in the inductively coupled plasma (ICP) etching process. Experimental results reveal that the primary dominating factor in RIE lag is feature width and secondary factors are feature area, shape and length-to-width ratio. Etching rates of rectangular trenches are sensitive to width while ring trenches are sensitive to both width and area. Process parameters are also adjusted to control RIE lag magnitude and realize its mechanism. The inverse RIE lag phenomenon appears at a much higher pressure of APC (auto pressure control) 75% at constant area features. The formation and removal of passivation film at the trench bottom will delay Si etching by F radical density, which will start earlier in a small width than a large one. It will be more obvious at higher pressure and lead to the reduction of RIE lag. This indicates that the cause of RIE lag in ICP etching is primarily attributed to the formation and removal of passivation film at the bottom of the trench, together with feature geometry. The RIE lag-eliminated trenches with constant area are obtained at a higher pressure of APC 70%. Deep and high aspect ratio silicon microstructures can be controlled by ICP etching with different pattern geometry.

Simulation of inverse reactive ion etching lag

Inverse reactive ion etching (RIE) lag was considered by a proposed model, which includes processes of adsorption, heterogeneous reactions, desorption, activation, sputtering, and stochastic mixing. Inverse RIE lag was explained by considering the neutral flux distribution at the groove bottom and assuming that arrival of C atoms from a plasma retards etching by causing surface passivation, whereas ion bombardment leads to etching. The thickness and composition of the altered layer at different points of groove during silicon etching in a fluorocarbon plasma were calculated. It was found that thickness of the altered layer increases with the increase of the microstructure width due to increased flux of C atoms at the groove bottom. Increased concentration of C atoms at the groove bottom reduces etching rate in the vertical direction.

Silicon dioxide etching yield measurements with inductively coupled fluorocarbon plasmas

Oxide etching yield has been measured directly with inductively coupled fluorocarbon plasmas. The yields measurement technique of this work can provide useful information for feature profile evolution modeling, which is essential to understand various issues in oxide etching such as reactive ion etching (RIE) lag, inverse RIE lag, etch stop, microtrenching, bowing, etc. Etching and deposition yields per ion were measured using quartz crystal microbalance (QCM) as a function of ion bombardment energy, ion-to-neutral flux ratio, and ion-impinging angle. C2HF5, C2F6, C2H4F2, and C4F8 were used for the oxide etching. Oxide etching mechanism with those gases is complex because etching and deposition are involved at the same time. In highly selective processes fluorocarbon deposition plays important role in determining etching characteristics. Two fluorocarbon deposition mechanisms are identified in this work: neutral deposition and ion-enhanced deposition. The low-energy ions are believed to enhance the deposition rates by creating active sites and fluorocarbon neutrals deposit on the active sites with higher sticking probability. A surface kinetic model is suggested to explain the ion-enhanced mechanism and shows good agreement with experimental data. Angular yield measurement shows that when fluorocarbon deposition is relatively severe, etching yield decreases significantly as the incident angle increases and deposit fluorocarbon at a high incident angle above 60°.

Etching of SiO2 features in fluorocarbon plasmas: Explanation and prediction of gas-phase-composition effects on aspect ratio dependent phenomena in trenches

A model to calculate etching rates in SiO2 features in fluorocarbon plasmas is presented. The model can predict several aspect ratio dependent phenomena such as reactive ion etching (RIE) lag, etch stop, inverse RIE lag, and aspect ratio independent etching (ARIE) at least for a limited range of aspect ratio values. The model includes three components: (a) a surface model for open area etching of SiO2 (and Si) [Gogolides et al., J. Appl. Phys. 88, 5570 (2000)]; (b) a flux calculator, which calculates local fluxes on each elementary surface of the feature being etched; and (c) a coupling of the two models (a) and (b), the focal point of coupling being the simultaneous calculation of the neutral species fluxes and the corresponding effective sticking coefficients. The model is applied for trench etching and the gas phase conditions considered correspond to a generic fluorocarbon gas. A different approach is presented by which the gas phase composition is divided (i.e., mapped) into regions leading to (a) deposition, (b) RIE lag with no etch stop, (c) intense RIE lag and etch stop, (d) inverse RIE lag, and (e) ARIE. Based on the proposed model an explanation of the aspect ratio dependent phenomena and ARIE is attempted, and a comparison with experimental data is done. Two parameters were found to be important in this explanation: the polymer surface coverage at the bottom of the etched feature and the effective sticking coefficients of the neutral species on the sidewalls of the etched feature.

Profile Evolution during Cold Plasma Beam Etching of Silicon

Processing characteristics of cold, directional plasma beams have been studied by numerical simulation of the profile evolution of trenches etched in silicon with chlorine, where the cold plasma beams were characterized by highly directional fluxes of neutrals as well as ions with their low temperatures or random thermal energies. The model included ion and neutral transport in microstructures and ion-assisted surface chemistry for the chlorine-silicon system. The numerical results demonstrate that the cold plasma beams achieve high etch anisotropy and microscopic uniformity: a tapered profile, inversely tapered profile, and reactive-ion-etching (RIE) lag, which often occur in usual plasma etching environments, are significantly suppressed owing to preferential incidence of neutral reactants as well as ions onto the bottom of microstructural features. In addition, the increased directionality of neutral fluxes not only offsets the RIE lag, but also further causes a weak inverse RIE lag without relying on surface inhibitors. Experimental techniques are also discussed for generation of such cold, directional beams of reactive plasmas.

Effects of O2 Addition on BCl3/Cl2 Plasma Chemistry for Al Etching

Effects of low level (0.5–20%) O2 addition on BCl3/Cl2 plasma chemistry have been investigated using several diagnostic tools: optical emission spectroscopy, microwave interferometry, and mass spectrometry. Experiments were performed using a magnetically enhanced planar 13.56-MHz rf plasma reactor, where aluminum etching was also performed using samples masked with a photoresist pattern of lines and spaces. As O2 was added into a BCl3/Cl2 plasma, the Al etch rate first increased and then dropped above ≈3% O2 added; a transition from reactive-ion-etching (RIE) lag to inverse RIE lag occurred at an O2 percentage of ∼8%. Optical and mass spectrometric measurements indicated that the Cl concentration increases as O2 is added into a BCl3/Cl2 plasma, and that above a critical O2 percentage (∼6% O2) Bx Oy species are formed in the plasma through a reaction between boron chlorides and oxygen and then deposit onto the wafer surface during etching. The Al etching characteristics obtained in BCl3/Cl2/O2 plasmas are interpreted in terms of competitive effects of increased concentrations of Cl and Bx Oy .

BSM 7: RIE lag in high aspect ratio trench etching of silicon

While etching high 3 aspect ratio trenches into silicon with reactive ion etching (RIE) using an SF6O2 chemistry it is observed that the etch rate is depending on the mask opening. This effect is known as RIE lag and is caused by the depletion of etching ions and radicals or inhibiting neutrals during their trench passage. In order to decide which source is the main cause, we constructed special “horizontal trenches” where only radicals are controlling the etching. The experiment showed that radicals are not responsible for RIE lag. Inhibitor depletion will result in inverse RIE lag. This effects is not found during our experimentation which leaves us with ion depletion to explain RIE lag. Depletion of ions is caused by ions captured by the sidewalls due to the angular distribution of incoming ions into the trench opening and the deflection of ions in the trench due to electrostatic fields. The analysis given in this paper indicates that the influencing field causes ion deflection, ion depletion, and therefore RIE lag in micron-sized Si trenches for low-energetic ions. In all cases, thus independent of the feature size, the angular distribution of incoming ions is thought to have a major contribution to RIE lag at higher pressures. These phenomena will be treated theoretically and simulated using a program, written in c++ under windows, in order to give a quantitative analysis of RIE lag.

Observation of inverse reactive ion etching lag for silicon dioxide etching in inductively coupled plasmas

The slowdown of the oxide etch rate with width of submicrometer structures is known as reactive ion etching (RIE) lag and has been explained by ion shadowing and differential charging of the sidewalls, among other effects [R. A. Gottscho and co-workers, J. Vac. Sci. Technol. B 10, 2133 (1992)]. Here we show for an inductively coupled high density plasma reactor working in the pressure regime from 6 to 20 mTorr that inverse RIE lag is primarily observed, i.e., the etch rates increase as the width of the microstructures decrease. Inverse RIE lag, which was first discussed by Vitkavage et al. [Tegal Plasma Proceedings Symposium, San Francisco, 1991 (unpublished)], may be explained by considering the neutral flux distribution at the structure bottom. The neutral flux has a stronger dependence on the aspect ratio than the ion flux due to its isotropic velocity distribution. The neutral flux distribution has been modeled and is consistent with etching profiles observed at high pressure.

Substrate bias effects in high-aspect-ratio SiO2 contact etching using an inductively coupled plasma reactor

High-aspect-ratio sub-half-micron contacts were etched in SiO2 using a high-density C2F6 plasma generated by an inductively coupled, low-pressure reactor. As expected, oxide and resist etch rates increased with increasing bias. The aspect ratio (depth/width) dependence of the etch process was also shown to vary with bias power; at high bias power conditions, reactive ion etching (RIE) lag was observed, where high-aspect-ratio features etch slower than low-aspect-ratio features, and at low bias powers, inverse RIE lag was observed. Increasingly vertical sidewall angles and increasing microtrenching depth were also observed for higher rf bias conditions on contacts, trenches, and steps. Topography modeling shows that the microtrench profiles are consistent with specular ion reflection from sidewalls and the increase in their depths with rf bias is consistent with the energy-dependent etching yields in high-density, fluorocarbon plasma etch chemistries reported by Oehrlein et al. [J. Vac Sci. Technol. A 12, 323, 333 (1994)].