ALScN Etchant Electronic/EL Grade is supplied as a ready-to-use aqueous formulation for removal of aluminum scandium nitride (Al₁₋ₓScₓN) thin films in piezoelectric microelectromechanical systems and bulk acoustic wave filter fabrication. Product designation ALScN-EL-200 is the standard shipping configuration; the liquid is filled in 1 L and 4 L PFA bottles inside an ISO 14644-1:2015 Class 4 cleanroom and sealed under nitrogen. Lot release includes ICP-MS trace metal analysis, particle counting at 0.2 μm and 0.5 μm, assay, density, and viscosity. The product is not a conventional aluminum etch mixture; its composition is designed for nitride removal with controlled attack on molybdenum and platinum electrode layers. The operational temperature range is 25 °C to 60 °C, with a recommended starting point of 40 °C for immersion tools and 45 °C for single-wafer spray processors. The formulation is free of added sodium and potassium salts to the extent measurable by ICP-MS and is classified as a strongly alkaline solution with a pH above 12.0 at 25 °C. Density at 20 °C is specified in the range of 1.06–1.10 g/cm³, and dynamic viscosity at 25 °C is below 5 mPa·s. The alkaline nature requires exhaust ventilation and fluoropolymer-wetted equipment. The product is compatible with sapphire and silicon substrates but not with exposed aluminum or AlCu electrodes unless protected by a dielectric barrier.

What distinguishes Electronic/EL Grade trace metal and particle specifications from standard wet etchants?

Electronic/EL Grade control focuses on impurities that shift acoustic properties or cause electrical leakage. In bulk acoustic wave filter production, trace cations such as sodium, potassium, iron, and copper can segregate at AlScN-electrode interfaces and produce resonance frequency outliers across a 200 mm wafer. Standard wet etchants often contain sodium above 500 ng/g; this grade holds sodium below 10 ng/g, potassium below 10 ng/g, iron below 5 ng/g, and copper below 5 ng/g. Total trace metals remain below 50 ng/g. Particle concentration at ≥0.2 μm is ≤ 100 counts/mL; at ≥0.5 μm it is ≤ 10 counts/mL. Chloride is specified below 20 ng/g, and sulfate below 30 ng/g. These lot-specific values are supported by laser light-scattering particle counters calibrated to ISO 21501-1:2024 and by ICP-MS analysis conducted under an ISO 17025-aligned quality system. The specification is intended for front-end semiconductor wet processing, not for general metal finishing.

Unmodified tetramethylammonium hydroxide and potassium hydroxide solutions used for silicon micromachining are not direct substitutes for this product. Those solutions produce higher attack rates on aluminum-containing electrodes and may generate excessive hydrogen evolution at the AlScN-electrode interface. The Electronic/EL Grade formulation includes a buffering agent and a corrosion inhibitor to moderate aluminum attack, and it is filtered to lower particle levels. This does not make it compatible with all electrode materials; exposed aluminum is still attacked at 15–40 nm/min at 40 °C.

Packaging is matched to the user's dispense system. Bulk chemical distribution uses 20 L fluoropolymer drums with 0.05 μm final filters; smaller facilities use 1 L bottles to reduce headspace exposure. The product shelf life is 12 months from the date of manufacture when stored at 10–30 °C and protected from direct sunlight. If the package is exposed to repeated opening outside a clean bench, point-of-use filtration at 0.1 μm is recommended before the liquid enters the etch tool.

Scandium fraction, film stress, and etch rate nonlinearity

Etch response is not constant across Al₁₋ₓScₓN compositions. Films with x = 0.10–0.20 typically etch slower than films with x = 0.28–0.35 because increased scandium content modifies bond polarity and grain boundary chemistry. Process data from 200 mm wafer monitor trials indicate that, under identical bath conditions at 40 °C, the removal rate can increase by a factor of 1.5–2.0 when x increases from 0.15 to 0.30. High residual compressive stress above 1.5 GPa also raises the etch rate, and center-to-edge variation on a 200 mm wafer may exceed 10 % if film stress is not uniform. For this reason, etch rate qualification is performed on the same sputter target composition and stress state as the production wafer, not on a generic ALScN coupon. Published data for scandium fractions above 0.42 are limited; those compositions may require a different dilution or lower bath temperature to avoid excessive sidewall attack.

In bulk acoustic wave filter fabrication, the ALScN layer is typically 300–900 nm thick and is patterned after bottom electrode deposition. A dry hard mask of silicon dioxide or silicon nitride is commonly used. The wet etch step must clear the open areas without significant undercut of the mask. At a bath temperature of 40 °C, a 500 nm AlScN film with x = 0.30 requires approximately 3–6 min of total immersion, depending on film stress and grain orientation. Overetch is limited to 20–30 % to preserve critical dimension. The etchant attacks silicon dioxide masks at less than 2 nm/min at 40 °C, so a 50 nm oxide hard mask can survive with sufficient margin. Silicon nitride masks show similar compatibility; silicon is not significantly attacked under the same conditions.

Etch selectivity to molybdenum, platinum, and silicon dioxide in device integration

Selectivity is evaluated by forming blanket etch rate monitors on 200 mm silicon wafers and measuring thickness loss by spectroscopic ellipsometry after immersion in a temperature-controlled quartz bath. At 40 °C, the removal rate for AlScN with scandium fraction x = 0.28 falls in the range of 80–160 nm/min. Under the same conditions, the attack rate on molybdenum is below 5 nm/min, on platinum below 1 nm/min, and on thermal silicon dioxide below 2 nm/min. The resulting selectivity ratios exceed 16:1 for molybdenum and 80:1 for thermal silicon dioxide. Selectivity degrades when bath temperature exceeds 60 °C or when the solution is contaminated with residual aluminum from prior processing. Galvanic coupling between the nitride and noble metal electrode can accelerate edge attack; the Electronic/EL Grade formulation incorporates a corrosion inhibitor to reduce this undercut mechanism without increasing total trace metals.

The etch proceeds by surface reaction at the nitride grain boundaries followed by dissolution of the reaction products. Because AlScN is chemically resistant in certain orientations, the process typically requires a finite overetch to clear rotated grains at the film-substrate interface. Endpoint detection by optical emission can underestimate residual nitride if the endpoint algorithm is trained only on blanket film calibration curves. Production lots should validate endpoint delay with cross-sectional scanning electron microscopy on a test wafer with the same thickness and stress.

In single-wafer spray processors, the etchant is dispensed through a fluoropolymer nozzle at a flow rate of 1.0–2.5 L/min onto a wafer rotating at 300–800 rpm. Endpoint detection is typically performed by optical emission or scatterometry; overetch margins are held to 20–30 % to minimize sidewall attack. The low-particulate formulation allows direct dispense without point-of-use re-filtration if the line is already equipped with 0.1 μm fluoropolymer filters. When used in single-wafer tools, the process window is usually ± 5 °C at the wafer surface; this narrow temperature control is required because the etch rate increases approximately 1.5× per 10 °C increment between 30 °C and 60 °C. Process engineers have observed center-to-edge etch nonuniformity below 8 % on 200 mm wafers when exhaust balance and chuck temperature are calibrated; without such calibration, nonuniformity can increase to 15 %. The exact activation behavior depends on scandium content and film stress.

When the etchant is used in recirculating immersion baths instead of single-wafer spray processors

Immersion bath operation introduces different constraints. The bath must be constructed of quartz, PTFE, or PFA, and heater elements must be protected from the alkaline liquid by fluoropolymer sheathing. Filtration through a 0.05 μm or 0.1 μm PFA filter loop at a recirculation rate of 5–10 turnovers per hour is recommended to prevent particle accumulation from wafer handling. Bath temperature should be maintained at 40 °C ± 2 °C for critical devices. Dissolved aluminum accumulates in the bath as wafers are processed; lot-specific data from production baths show that etch rate decreases by approximately 10–15 % after 50 wafer-sized 200 mm loads per liter, so bath life is normally defined by cumulative etch load rather than elapsed time. Ventilation should capture ammonia and hydrogen evolution at a capture velocity of 0.5–1.0 m/s across the bath opening.

Operational boundaries must be observed before process release. The etchant should not be mixed with acids or with hydrogen peroxide unless specifically directed by the manufacturer because exothermic reactions and release of toxic gases may occur. Wetted parts of dispensing pumps must be fluoropolymer; stainless steel and borosilicate glass are not acceptable for long-term exposure. The product is not recommended for direct etching of AlScN over aluminum or AlCu electrodes unless the electrode is protected by a dielectric barrier. Waste neutralization must follow site permits for alkaline process chemicals containing organic chelators. The product is supplied under REACH and RoHS documentation, but it is a process chemical, not a finished article subject to RoHS restrictions. For devices with scandium fractions above 0.35 or film stress above 1.5 GPa, published data for this specific formulation is limited, and on-site feasibility studies are required before committing to volume processing.