SiGe Etchant Electronic/EL Grade is a pre-blended ammoniacal peroxide wet-etch formulation intended for selective removal of sacrificial Si_1-xGe_x layers in front-end-of-line semiconductor device fabrication and microelectromechanical systems release. The commercial designation SiGe Etchant Electronic/EL Grade functions as the product model identifier in this documentation; no separate model code is assigned. The Electronic/EL classification distinguishes the material from technical-grade etchants by imposing tighter trace metal, particle, and assay controls. The product is used principally in recirculating immersion wet benches and single-wafer spray processors where selective SiGe removal is required relative to monocrystalline Si, thermal SiO₂, and SiN stop layers. In production practice, etch-rate verification is performed on blanket SiGe and Si monitor wafers using spectroscopic ellipsometry or X-ray reflectometry; published data for this specific pre-blended electronic-grade configuration is limited, and process owners should not transfer etch recipes across suppliers without requalification.

Material certification boundaries for the Electronic/EL Grade

The grade is released against a lot-specific certificate of analysis that includes trace cation analysis by inductively coupled plasma mass spectrometry, anion analysis by ion chromatography, particle counts by laser light-scattering particle counter, hydrogen peroxide assay by iodometric titration, and pH by glass-electrode method per ASTM E70-19. Trace metal thresholds are aligned with SEMI C7 electronic-grade process chemical limits; advanced nodes below 7 nm commonly impose additional limits of ≤1 ppb for Fe, Ni, Cu, and Zn. Particle counts are controlled by SEMI F40-1118 laser particle inspection, with a target of fewer than 50 particles/mL at 0.2 µm or larger in the as-supplied liquid. Filtration is performed through 0.1 µm or 0.05 µm fluoropolymer membranes during packaging. Hydrogen peroxide concentration is stated on the lot certificate and is maintained within ±2% of nominal; pH is controlled within ±0.2 pH units of the target value. These values are representative of high-purity wet etchants in this class; user acceptance criteria may be tighter for gate-all-around release steps.

Certificate of analysis reports alkali and alkaline earth elements because residual Na and K are known to shift threshold voltage and degrade gate oxide integrity. For the Electronic/EL Grade, sodium and potassium are typically controlled below 1 ppb, while calcium and aluminum are controlled below 2 ppb. These limits are achieved by using high-purity subcomponents and by maintaining closed-loop packaging that avoids glass-lined reactors and metal fittings. The difference between the Electronic/EL Grade and ordinary technical-grade SiGe etchants is documented in the certificate of analysis. Technical-grade mixtures may carry metal cation concentrations of 10–100 ppb and particle burdens above 100 particles/mL, which are acceptable for cleaning fixtures but not for device fabrication. The Electronic/EL Grade is filtered and filled in a cleanroom environment, and each lot is assigned an expiration date based on accelerated aging data. Lot-specific certificates of analysis support change control in ISO 9001 and IATF 16949 quality systems; semiconductor fabs retain these records as part of process chemical qualification.

Storage is specified at 10–25°C in vented cabinets away from direct ultraviolet light and transition metal contamination. The etchant is packaged under nitrogen after final filtration to minimize dissolved oxygen and carbonate absorption. In a production-scale recirculating wet bench, the bath is typically held at 35–55°C depending on sacrificial SiGe thickness and the selectivity window of the device stack. Hydrogen peroxide concentration and pH are monitored at intervals not exceeding 4 h; oxidant decomposition accelerates if dissolved Fe exceeds 1 ppb, and rising dissolved Ge indicates progression of the SiGe removal step. Dilution or spiking is performed only with ultrapure water meeting ASTM D5127-13 Type E-1.2; the introduction of plant water raises particle counts and cation burden above the certified envelope. Utility water with resistivity below 18.2 MΩ·cm at 25°C or total organic carbon above 5 ppb is not suitable for dilution.

Shelf life is typically 12 months from the date of manufacture when stored in unopened containers at 10–25°C. The lot certificate includes the manufacturing date, recommended retest date, and bulk assay. Once the container is opened, the product must be transferred through fluoropolymer tubing under nitrogen; unused material must not be returned to the original container because of particle and cation contamination from the dispense line.

How does the Electronic/EL Grade differ from point-of-use SC1 and acidic HF–HNO₃ mixtures?

In wet-bench qualification data, the principal differences are seen in trace metal burden, particle count, and Si:SiGe selectivity. Point-of-use SC1 mixed from VLSI-grade ammonium hydroxide and hydrogen peroxide can meet bulk etch-rate targets but introduces batch-to-batch variation through manual pouring, container-rinsing, and trace metal pickup from pH adjusters. The Electronic/EL Grade is preblended and filtered at the supplier, so the oxidizer ratio and cation burden are defined before the chemical reaches the wafer environment. Acidic HF–HNO₃–CH₃COOH mixtures remove Si and SiGe at high isotropic rates but exhibit lower Si:SiGe selectivity and attack SiO₂ and SiN hard masks aggressively. In contrast, the ammoniacal peroxide chemistry of the Electronic/EL Grade oxidizes SiGe preferentially and removes the oxidized layer through dissolution, leaving monocrystalline Si and most dielectric stop layers with substantially less material loss. The trade-off is lower bulk etch rate than acidic isotropic etchants, which limits its use to sacrificial SiGe release rather than bulk silicon removal.

Etch selectivity arises from the difference in oxidation and dissolution rates between SiGe and Si in ammoniacal peroxide. In the oxidation step, hydrogen peroxide adsorbs on the SiGe surface and generates mixed germanium and silicon oxyhydroxides; ammonium hydroxide then dissolves the germanium-rich oxide component, while the silicon-rich surface remains passivated. The result is a blanket Si_0.7Ge_0.3 removal rate of 5–25 nm/min over 25–65°C, while undoped Si(100) removal remains below 0.5 nm/min. Selectivity is a function of Ge fraction, bath composition, temperature, and dissolved oxygen; it must be validated on fully processed short-loop wafers because blanket selectivity values do not capture loading effects in patterned nanosheet structures.

During gate-all-around nanosheet fabrication, the etch bath is introduced after the SiGe/Si superlattice has been exposed through anisotropic dielectric recess. The process operating point is selected from an etch-rate matrix that varies temperature from 25°C to 65°C and H₂O₂:NH₄OH ratio from 1:1 to 5:1. Selectivity to Si(100) typically exceeds 50:1 for Ge fractions above 0.25; selectivity decreases sharply when the Ge fraction falls below 0.15. A processing window of ±5°C is required for high-Ge SiGe release because the etch rate approximately doubles per 10°C temperature increase according to an Arrhenius-type relationship. Wafers are processed in PFA or PTFE cassettes, rinsed with ultrapure water, and dried by isopropyl alcohol vapor or low-spin drying. Released nanosheet structures are susceptible to capillary-force collapse if the final rinse is not maintained at low surface tension; IPA vapor drying is the accepted route for high-aspect-ratio release.

Operational boundaries include incompatibility with exposed Cu and Al interconnects because alkaline peroxide chemistries oxidize and dissolve these metals. For stacks with exposed NiPt or TiN hard masks, the etchant may cause unacceptable metal loss unless the mask is fully encapsulated. The chemistry is not intended for selective silicon removal; HF-based chemistries remain the industry baseline for bulk Si isotropic release. Thermal SiO₂ loss is typically below 1 nm for a 10 min immersion at 45°C, but the alkaline peroxide environment will thicken chemical oxide on hydrogen-terminated Si if the rinse delay exceeds 30 s. Spiking with organic surfactants or amines is not permitted because such additives raise total organic carbon and shift the pH outside the certified envelope.

The as-supplied formulation is controlled to a narrower oxidizer-to-base ratio than manual SC1 mixing. In front-end-of-line wet benches equipped with chemical concentration monitors, the oxidizer setpoint is maintained within ±2% of the target by mass-balance dosing; without a concentration monitor, the bath is recharged after a fixed number of wafers, typically 25–50 300 mm wafers per 50 L bath. The etch rate of Si_0.7Ge_0.3 at 45°C is commonly found in supplier application notes to fall between 8 and 15 nm/min; laboratory measurement on a blanket monitor wafer within the same lot should be used to set immersion time. Process owners should not rely on published values alone because wafer doping, Ge grading, and prior surface treatments alter the etch rate by up to 30%.

In single-wafer spray processors, the etchant is dispensed at a flow rate of 1–2 L/min through a point-of-use filter and a heated nozzle. Dispense temperature is maintained at 35–65°C, and spin speed is set between 300 and 800 rpm to maintain a uniform puddle. The high-purity fluid path uses PFA tubing with 1/2 in or 3/4 in diameter; PTFE seals and O-rings must be inspected after 500 chemical hours. These tool parameters are typical for 300 mm single-wafer wet etch processors and are supplied here to define the mechanical interface.

Particle, Trace Cation, and Fluoropolymer Compatibility Controls

Wetted components in the supply and recirculation loop are constructed from PFA, PTFE, or PVDF; stainless steel and titanium components are excluded because peroxide decomposition occurs on metal surfaces and releases transition metal ions. The fluid path is passivated with dilute nitric acid after installation and before the first production bath. Filtration at the dispense gun is maintained at 0.05 µm retention in many single-wafer tools. Particle counts are rechecked after each chemical drum change; batch-to-batch variation in particle burden is lower for preblended Electronic/EL Grade than for point-of-use mixed SC1 because manual pouring and container-rinsing operations are eliminated. In a 300 mm immersion bench with 60 L recirculating volume, a single batch change typically requires 2–4 h of recirculation through the filter before the bath reaches particle counts below 20 particles/mL. Supplier pre-filtration is not a substitute for tool-side dispense filtration, but it substantially reduces the initial burden on tool-side filters.

Failure modes observed on manufacturing lines include localized unetched defects caused by microbubble formation when peroxide decomposition is accelerated by iron contamination, and recurrent particle adders when drum pumps are not flushed after idle periods longer than 8 h. The supplier-recommended flush volume for a 200 L drum is 2–5 L through the dispense line before the first wafer lot after idle. These operational limits define the acceptable operating envelope.

When sacrificial SiGe release demands low silicon loss and post-etch surface integrity

Process engineers select the Electronic/EL Grade when the allowable Si loss in the channel is below 1 nm per release step and when surface roughening must remain below 0.5 nm RMS as measured by atomic force microscopy over a 5 µm scan. For suspended nanosheet stacks, the etchant must remove SiGe without collapsing the released silicide or source/drain epitaxy. The lower oxidizer concentration compared with aggressive acidic etchants limits the lateral etch rate and provides a process window of approximately ±10% on over-etch time. The trade-off is a lower blanket etch rate than HF–HNO₃ systems; when etch time exceeds bath life, a fresh bath is prepared and the spent etch solution is sent to fluoride-free waste treatment. The alkaline peroxide solution does not require calcium fluoride precipitation as do HF-based etchants, but it must be segregated from acidic waste to avoid exothermic neutralization and release of ammonia.