TiN etching cleaning agent Electronic/EL Grade is supplied as a filtered aqueous alkaline peroxide formulation designed for post-etch residue removal in copper/low-k damascene flows, aluminum interconnect flows, and through-silicon via liner cleaning where titanium nitride hardmasks, barrier liners, or anti-reflective coatings must be oxidized and complexed without re-depositing titanium dioxide on exposed dielectric sidewalls. Here, EL denotes electronic-grade, not electroluminescent; the grade is defined by the contamination limits in the specification table. Model EL-TiN-100 is qualified for batch immersion wet stations, and model EL-TiN-200 is qualified for single-wafer spray processors. The electronic/EL grade designation indicates that each lot is released against trace-metal, particle, and total organic carbon limits that are compatible with front-end-of-line gate-level cleaning and back-end-of-line copper/low-k integration. The as-supplied formulation contains 29.0–32.0 wt% hydrogen peroxide, 14.0–15.5 wt% ammonium hydroxide equivalent, a titanium-specific aminopolycarboxylate chelant at 0.2–0.6 wt%, and ultrapure water. The pH is maintained at 9.2–9.8 at 25 °C, and the liquid is filtered through 0.05 µm PTFE at point-of-use in critical single-wafer applications. The product is not intended as a bulk TiN blanket etchant; it is co-optimized for residue dissolution, particle removal, and surface-metal redeposition control.

What Distinguishes Electronic/EL Grade TiN Cleaning Agent from Conventional SC1 and Dilute HF Chemistries?

RCA SC1 mixtures typically use ammonia–hydrogen peroxide–water ratios of 1:1:5 to 1:4:20. SC1 oxidizes TiN surfaces through an alkaline peroxide mechanism, but the absence of a titanium complexant allows hydrolyzed titanium species to precipitate as TiO₂·xH₂O on via sidewalls, particularly when the bath is aged and titanium loading exceeds 1–5 ppm. The EL-grade formulation retains the same oxidation route but binds titanium in solution; the resulting wet residue is removed during the overflow rinse rather than forming a re-deposited layer. Dilute HF at 100:1 removes TiN and titanium fluoride residues by dissolution, but it also attacks aluminum bond pads at rates of 5–15 nm/min and can generate surface roughness on low-k SiCOH dielectrics. Compared with dilute HF, the EL-grade product is specified to reduce aluminum pad attack to 0.1–0.3 nm/min at 55 °C. Sulfuric acid–hydrogen peroxide mixtures operate at 120–150 °C and are effective for post-ash organic residue removal, but their TiN removal rate is typically below 1 nm/min, making them insufficient as a single-step TiN cleaning agent on hardmask residues. The EL-grade product therefore occupies a process niche: front-end wet cleaning of TiN-bearing residues where low aluminum loss, low particle addition, and low mobile-ion contamination are as important as etch rate.

Model EL-TiN-100 is released under the certificate-of-analysis limits shown below. The limits are verified by inductively coupled plasma mass spectrometry, ion chromatography, and laser particle counting after packaging in electronic-grade HDPE containers filled in an ISO 14644-1:2015 Class 2 cleanroom. On-site re-qualification is required after bulk distribution because pump shedding and transfer-line particle release can increase particle counts by 10–50 counts/mL above vendor limits. For point-of-use quality, inline particle counters should be installed downstream of the final filter and upstream of the dispense nozzle.

The immersion grade is suitable for batch tools with cassette-to-cassette handling, while the spray grade is filtered to 0.05 µm and formulated with a low-foam surfactant package. Foam height in recirculated tanks must remain below 5 mm to prevent pump cavitation and pressure oscillation in point-of-use filtration. The lower trace-metal limits for EL-TiN-200 are required because single-wafer tools process wafers with exposed gate dielectrics; sodium and potassium above 3 ppb can shift threshold voltage in high-k metal-gate stacks.

Particle Reduction and Metallic Impurity Profiles in 300 mm Single-Wafer Tools

In single-wafer spray cleaning tools, tighter particle control is required than in immersion benches because the wafer surface is exposed to aerosolized droplets and nozzle plate condensation. The EL-TiN-200 spray-grade lot-release particle limit of ≤25 counts/mL at ≥0.2 µm is below typical immersion-grade acceptance criteria. The product is dispensed through PFA lines at 0.15–0.25 MPa and 800–1500 rpm wafer rotation. In this flow regime, the dominant particle source is not the formulated liquid but the delivery system: sudden suck-back valve closure can create cavitation bubbles that collapse and shed 10–30 counts/mL of particles downstream. Production-scale 300 mm tools therefore use point-of-use filtration with 0.05 µm PTFE membranes and a slow suck-back speed of 5–10 mm/s to avoid pressure transients. Metallic impurity limits of ≤3 ppb for the spray grade reduce mobile ion contamination on high-k gate dielectrics. Reagent-grade ammonia–peroxide mixtures often contain Fe at 50–200 ppb and particle counts above 1000 counts/mL, which is unacceptable for gate-level cleaning. The electronic/EL grade is therefore not a particle specification alone; it includes a packaging and delivery verification protocol that includes cleanroom filling, leached-container testing, and point-of-use filter compatibility testing.

For immersion processing, EL-TiN-100 is diluted with ultrapure water meeting ASTM D5127-13 Type E-1. The dilution range is 1:5 to 1:20 by volume depending on residue thickness and thermal history. Process temperature is maintained at 55 ± 2 °C. Immersion time is 3–10 min for TiN hardmask residues and 5–12 min for aged post-etch polymers. Wafers are transferred to an overflow rinse of ultrapure water at 22–25 °C for 5–10 min until rinse water resistivity reaches 18.0 MΩ·cm. Drying is performed with isopropyl alcohol vapor or a Marangoni-assisted dryer. The bath is recirculated through 0.1 µm PTFE filters for immersion tools and 0.05 µm filters for spray tools. Wetted materials should be PVDF, PTFE, PFA, or quartz. Stainless steel and titanium components are incompatible because dissolved Fe and Ti ions catalyze hydrogen peroxide decomposition and degrade the chelant. Bath heaters should be quartz-sheathed or PVDF-sheathed; exposed titanium heater elements have been observed to reduce bath life by 50% in immersion tools.

Monitor wafer verification usually includes ellipsometric measurement of TiN thickness before and after cleaning, X-ray photoelectron spectroscopy for titanium and nitrogen signals, and scanning electron microscopy for via sidewall residue. A titanium signal remaining at the 2p peak at 458.5 eV after cleaning indicates incomplete oxidation or redeposited TiO₂. Surface roughness on low-k dielectrics is measured by atomic force microscopy; a root-mean-square roughness increase above 0.2 nm after 10 min immersion at 55 °C typically indicates a pH excursion or bath aging. The bath titanium concentration is monitored by ICP-MS; titanium loading above 5 ppm correlates with particle growth and redeposition in immersion tools.

When Bath Temperature Exceeds 65 °C, Titanium Peroxide Decomposition Accelerates

The temperature window is critical because TiN oxidation and hydrogen peroxide decomposition are competing reactions. At 50 °C, TiN removal rate is approximately 2.1 nm/min and TEOS oxide loss is 0.3 nm/min. At 60 °C, TiN removal rises to 4.8 nm/min while TEOS loss rises to 0.9 nm/min. Above 65 °C, hydrogen peroxide decomposition dominates, oxygen evolution increases, and the titanium complexant loses binding capacity; titanium oxide redeposition on via sidewalls becomes visible by XPS after 10 min of exposure. The processing window is therefore ≤±5 °C around the 55 °C setpoint. Production-scale immersion tools with 25-wafer cassettes exhibit bath life of approximately 8 hours or 200 wafers, whichever occurs first. The limiting factor is titanium loading in the bath, not peroxide depletion. At pH 9.0, soluble titanium species hydrolyze; at pH 10.0, aluminum bond pad attack increases. Automatic dosing with pH and H₂O₂ analyzers is required for lot-to-lot stability; manual replenishment can produce pH excursions above 0.3 pH units and increase pad attack. For single-wafer spray processing, chemical temperature at the nozzle should not exceed 40 °C to prevent peroxide decomposition before wafer contact.

Compared with conventional RCA SC1, dilute HF, and SPM, the EL-grade chemistry occupies a narrower process window but provides lower residue redeposition and lower metal contamination. The table below summarizes representative process data for TiN hardmask residue removal on silicon wafer test structures.

The comparison table is based on blanket film measurements on thermal oxide and TEOS test wafers. Device-level structures with porous low-k dielectrics or airgap integration require additional screening because the chelant and surfactant can penetrate pore structures and alter dielectric constant after subsequent anneal. A post-clean anneal at 350 °C for 30 min under nitrogen is used to test for carbon residue and moisture uptake.

The Spray-Delivery Variant Requires a Different Additive Package

EL-TiN-200 is not simply a filtered version of EL-TiN-100. The spray-grade formulation contains a nonionic low-foam surfactant at 0.01–0.05 wt% to wet high-aspect-ratio vias without generating foam in recirculating lines. Foam height under 1 L/min recirculation is specified at ≤5 mm. The product is qualified for dispense on 300 mm single-wafer tools with 0.15–0.25 MPa nozzle pressure and wafer rotation of 800–1500 rpm. Spray cleaning time is 30–90 s per wafer, followed by rinse and spin dry. The shorter contact time requires the chelant package to achieve complexation within the liquid film residence time of 1–3 s; therefore EL-TiN-200 uses a higher chelant-to-peroxide ratio than the immersion variant. Published data for long-term performance in EUV-processed wafers is limited, and qualification on device wafers is required before high-volume manufacturing.

Waste streams containing the EL-grade chemistry must be segregated from concentrated sulfuric acid and piranha waste because peroxide-acid mixing can generate violent oxygen evolution. Spent immersion baths are reduced with sodium metabisulfite before pH neutralization, and fluoride-containing rinse water is routed to a separate fluoride abatement system. The product is stable for 12 months in unopened HDPE containers stored at 15–25 °C, with vented caps to release oxygen generated by slow peroxide decomposition. Once opened, the product should be blanketed with filtered nitrogen and consumed within 30 days to preserve metal and particle specifications.