This article presents an analysis of the emergence of etch profiles of nanopatterned SiO2 utilizing a CHF3/Ar plasma process.
First, the effect of electrode forward power, chamber pressure, and substrate electrode temperature were investigated. At lower pressure, the etch rate was lower, while the pattern fidelity was higher. The anisotropy and etch rate increase with higher forward power. Consequently, the process conditions could be effectively improved to simultaneously realize aspect ratio independent etching and high pattern quality.
Recently, atomic-scale surface engineering using sophisticated manufacturing technologies has attracted a great deal of interest. Innovative processes are being developed to facilitate the development of complex device structures, in which crucial dimensions and pitch shrinkage are required for higher precision and selectivity of etching.
Atomic layer etching (ALE) is a technique that offers unparalleled levels of control for etching performance to fulfill the latest technological nodes. It, therefore, has a considerable amount of potential to address and resolve the problems associated with contemporary fabrication methods.
An inert plasma is maintained all through the plasma-enhanced ALE process. Through alternating pulse injections of precursors and ion energy to the wafer, alternating cycles of deposition and etching are produced, respectively. To deposit angstrom-thick layers on SiO2, fluorocarbon chemistry is used to offer the reactant adsorption.
Initially, a fluorinated SiO2 mixed surface layer is created and then removed using low-energy Ar-ion bombardment. These ions are below the energy threshold for SiO2 sputtering; therefore, etching ceases once the mixed layer is removed. This provides the “self-limiting” criterion required for ALE processing.
Through a pulsed injection of CHF3, the fluorinated layers are reproduced with unbiased substrate conditions. After the layer is grown, a low-level RF bias is applied to speed up the Ar ions toward the substrate with the highest ion energy, while making sure that the RF bias is less than the energy threshold for the physical sputtering of SiO2. The sequence is repeatedly performed until the desired etched thickness is obtained.
The aim of this analysis was to optimize the SiO2 etching profiles for nanoscale-sized features (30–200 nm), during the atomic layer etch patterning of SiO2 utilizing CHF3/Ar plasma.
A sequence of experiments was then carried out where only a single plasma process parameter is adjusted at a time, paying attention to pattern fidelity transfer and the quality of vertical sidewall fabrication.
The significant end results are given below:
- At -10 °C, slow deposition rates are realized together with a directional transfer, flat-bottom surfaces, and vertical sidewalls.
- The profile was enhanced at a very low pressure of 5 mTorr.
- The directionality of the ion bombardment was considerably increased at higher forward bias power (FBP) of 4 W.
A lift-off process was employed for patterning of a chromium (Cr) hard mask, initially producing nano-sized lines on poly(methyl methacrylate) (PMMA) and then utilizing electron-beam lithography on Si wafers with a 250-nm SiO2 layer. The thickness of the PMMA was 60 nm, and the line width differed from 30 to 200 nm.
A Cr layer was subsequently added. The Cr lines were defined using a lift-off process and then by acetone cleaning in an ultrasonic bath. The ALE process was performed using a Plasmalab System 100 ICP etcher. To speed up the ions toward the substrate, 13.56 MHz power was then applied to the substrate electrode.
A schematic of the cyclic ALE process used is shown in Figure 1; this process involves repeating the deposition and etch steps. The process parameters that were varied were the process pressure, FBP, and substrate temperature.
Figure 1. Schematic of the cyclic ALE process. Image Credit: Impedans.
Ar gas was allowed to flow constantly at 100 sccm, where the ICP power was maintained at 300 W. For the deposition half-cycles, periodical injections of 10 sccm of CHF3 were introduced. The FBP delivered to the substrate electrode regulates the ion energy at the substrate.
FBP was then applied to create a DC bias ranging between −9 and −19 V. The Semion System, a retarding field energy analyzer from Impedans, was used to measure the ion energy distribution (IED). This IED was quantified as a function of the pressure, utilizing a FBP of 2 W and pure Ar plasma with 300 W inductively coupled source power.
Results and Discussions
To assess the effect of the plasma process parameters on the etch profiles, etching rate per cycle (EPC), sidewall angle (θ), and undercut were compared as shown in Figure 2.
Figure 2. Schematic cross-section of the etch profile including sidewall angle, degree of undercut, and etching per cycle. Image Credit: Impedans.
The etching depth of each cycle (EPC, Å/cycle) refers to the SiO2 thickness divided by the total number of cycles, that is, 60 cycles. The sidewall angle denotes the deviation from a perpendicular sidewall. The undercut percentage quantifies the amount of SiO2 eliminated from the region under the Cr mask. These parameters were quantified for different line widths (that is, 30, 40, 50, 100, 150, and 200 nm).
Effect of the Substrate Temperature on ALE Mechanism
In a previous study, the impact of temperature on SiO2 etch profiles was studied. The etching rate increased when the temperature was above T = −10 °C. This was due to the residual fluorine radicals emerging from the chamber walls. At temperatures ranging between T = −10 °C and +20 °C, there was an undesirable increase in the chemical etching of SiO2 by fluorine because of a reaction between the SiO2 surface and the fluorocarbon polymer.
This etch rate increase came at the cost of etch selectivity and film quality. Here, the authors carried out the ALE process at three different temperatures—T = −40 °C, -10 °C, and 20 °C. The FBP and pressure were 2 W and 10 mTorr, respectively. The corresponding DC bias was −9 V. Figure 3 shows the undercut and sidewall angle parameters.
Figure 3. Etched silicon oxide thickness per cycle and fluorocarbon film thickness per cycle deposited on SiO2 per ALE cycle (a). Sidewall angle and degree of undercut as a function of the substrate temperature (b). Image Credit: Impedans.
The EPC stayed constant across the temperature range, whereas the deposition per cycle decreased from 7 Å/cycle at T = −40 °C to 3 Å/cycle at T = 20 °C. At -40°C, the higher deposition rate results in unwanted rounding of the bottom of the feature when compared to 20 °C. But the flat surface achieved at T = 20 °C was accompanied by a more inclined sidewall and higher undercut value.
The authors observed that the −10 °C value delivered the optimum results for the deposition rate and allowed an optimal directional transfer.
Effect of the Pressure and Ion Energy Distribution
Operating pressure is one of the significant parameters in the ALE process. This parameter has a direct impact on the energy and flux of the Ar ions, and on the concentration of F, C, and Ar radicals in the plasma. The pressure was set to 5, 10, 25, and 40 mTorr for various ALE process experiments, while maintaining a fixed substrate temperature of T = −10 °C and FBP of 2 W. For each pressure level, the authors carefully defined the IED with the help of the Semion system. A single-peaked energy distribution was displayed at all the pressures analyzed. The peak energy and ion flux (area under the curve) were constant at pressures of 5 and 10 mTorr.
An increase in pressure causes the mean free path of the ions to become shorter than the plasma sheath width, and leads to lower peak energy as illustrated in Figure 4(a). The higher energy at lower pressure results in more anisotropic etching but reduced etch rate, as demonstrated in Figures 4(b) and 4(c).
Figure 4. IED for each pressure (a). Etched silicon oxide thickness per cycle and fluorocarbon film thickness per cycle deposited onto SiO2 per cycle (b). Sidewall angle and degree of undercut as a function of the pressure values (c). Image Credit: Impedans.
At higher pressure, the EPC increases as more radicals become available from the plasma, which, in turn, increases the etch rate by over 60% at 40 mTorr when compared to 5 mTorr. But it was observed that the sidewall angle was enhanced at lower pressure, demonstrating that anisotropic etching was facilitated at lower pressure.
Effect of the Bias Power on the Etching Directionality
The impact of FBP on the etched profiles was tracked. Other parameters were maintained constant, where the temperature was -10 °C and the pressure was 10 mTorr.
Figure 5 illustrates the etching rate as a function of the increasing negative bias corresponding to an increase in FBP bias power from 2 W (DC bias = -9 V) to 3 W (DC bias = −12 V) and 4 W (DC bias = -19 V).
Figure 5. Etched silicon oxide thickness per cycle as a function of the same negative DC bias values (a). Sidewall angle and degree of undercut as a function of the negative DC bias (b). Image Credit: Impedans.
The higher energy ions interacted with the mixed layer at the time of etching, resulting in a deeper SiO2 etching and thus higher EPC. At lower FBP, partial removal of the mixed layer at the bottom was observed.
Due to lower energies of the Ar ions, a longer time was required to fully eliminate the fluorocarbon layer. When the FBP is increased, more ions become incident on the electrode with a smaller deflection angle because of the smaller contribution of ion-neutral scattering at higher energies. This impact is illustrated in Figure 5.
The profiles turned out to be more vertical, with the angle increasing from 78° at -9 V DC bias to 85° at -19 V DC bias, and the undercut decreasing from 10% down to 6% for the same DC bias values, respectively. Therefore, -19 V DC bias (4 W) was identified as the optimum value to be used in future patterning design of SiO2.
Aspect Ratio Independent Etching
A range of Cr-masked SiO2 samples with different trench sizes was etched for 60 ALE cycles under the optimal ALE self-restricting conditions identified, that is, -10 °C, 5 mTorr, 19 V DC bias and tetch = 60 seconds.
Each feature had the same vertical profile, about 87° ± 1.5° on average, and a low undercut value of 3.7% ± 0.5%. This showed that the process helped to realize aspect ratio independent etching in a traditional ICP tool.
The authors also mentioned that features that have larger aspect ratios etch faster compared to those that have lower aspect ratios, irrespective of the feature width. One can rule out the contribution of ions scattered from the feature edges, because the sidewall profile does not vary considerably with the feature width.
To realize features without any undercut, a careful balance should be found between the fluorocarbon chemical reactant and the Ar ion parameters.
Using a traditional ICP tool, a cyclic CHF3/Ar ALE process was analyzed in-depth and characterized. The temperature range (T = -40 °C to 20 °C) does not have a major impact on the ALE behavior; instead, it is governed by simultaneous kinetic and chemical processes between the mixed layer and the accelerated Ar ions. Pressure had a major effect on the process, where low values of pressure (5 mTorr) led to high ion energy, as well as low density of C and F radicals.
The decreased EPC and the steep sidewalls (up to 86°) were induced by the anisotropic etching from the Ar ions. Higher forward power resulted in smaller deflection angle and increased ion energy, which, in turn, led to deeper etching and enhanced verticality of the etched features.
References and Further Reading
S. Dallarto et al., “Balancing ion parameters and fluorocarbon chemical reactants for SiO2 pattern transfer control using fluorocarbon-based atomic layer etching.” American Vacuum Society. doi: 10.1116/1.5120414
This information has been sourced, reviewed and adapted from materials provided by Impedans Ltd.
For more information on this source, please visit Impedans Ltd.