Phosphate and Sulfate Residue Control in Display Edge Bead Removal

In display photolithography, edge bead removal is executed on flat-panel substrates ranging from Gen 5 (1100 mm × 1300 mm) to Gen 10.5 (2940 mm × 3370 mm) using radial solvent dispense at the substrate periphery while the coater chuck rotates at 500–2000 rpm. The EBR solvent dissolves the photoresist bead that forms at the edge after spin coating, preventing transfer of partially dried resist to robot end effectors, vacuum chucks, and downstream process tools. Phosphate and sulfate contamination enters the EBR system from multiple sources: phosphoric acid used in metal etch and panel cleaning, phosphate ester surfactants in resist and EBR formulations, sulfuric acid-peroxide mixtures used for organic removal before resist coating, and sulfate from sulfonated photoacid generators in chemically amplified resists. The resulting anionic residues are nonvolatile and remain at the edge exclusion boundary after solvent evaporation. In TFT-LCD and OLED backplane fabrication, these residues can cause localized pH shifts during wet etching, metal complexation in gate and data metal layers, and adhesion loss in subsequent organic planarization. Control requires an integrated approach that links EBR solvent selection, point-of-use filtration, rinse water quality, reclaim chemistry, and analytical verification. Standards that govern this integration include ISO 10304-1:2007 for anion quantification in water, ASTM D4327-17 for suppressed conductivity ion chromatography, SEMI F63 for ultrapure water distribution in semiconductor and display facilities, and ISO 14644-1:2015 for cleanroom classification. In practice, the EBR process window is often narrower than the resist coating process itself because residue tolerances at the substrate edge are set by downstream wet etch uniformity, not by visual edge quality alone. Production coaters for display substrates may use EBR solvent volumes of 50–150 mL per substrate, depending on substrate size, edge bead width, and head speed. The resulting waste stream is frequently recycled through central solvent recovery, which can reconcentrate phosphate and sulfate unless anion-specific purification is added.

What Drives the Retention of Phosphate and Sulfate at the Edge Bead Boundary During Spin-Off Drying?

Residual phosphate and sulfate at the edge bead boundary are not simply a function of influent concentration. The drying dynamics after EBR dispense create a concentrating zone. When the EBR solvent is a PGMEA/ethyl lactate blend with a vapor pressure near 3.8 mmHg at 25 °C for PGMEA, solvent evaporation at the edge quickly increases the concentration of nonvolatile ions. The drying front moves inward from the substrate periphery at velocities that typically fall between 0.1 mm/s and 3.0 mm/s depending on spin speed, exhaust velocity, and solvent composition. Sulfate ions, introduced as sulfuric acid or sulfonated residues, may remain dissolved in the thin aqueous layer at the edge because sulfuric acid is highly hygroscopic. Phosphate ions, however, show lower solubility in organic-rich films and may precipitate as calcium or magnesium salts if multivalent cations are available from glass, polishing slurry, or rinse water. Calcium sulfate dihydrate has a solubility near 2.4 g/L in water at 25 °C, whereas calcium phosphate phases exhibit solubilities below 0.02 g/L, so phosphate tends to persist as a film-bound precipitate rather than as a mobile ion. The pH at the drying edge can range from 3.5 to 6.5 due to carbon dioxide uptake and acidic residues. Under those pH conditions, sulfate exists mostly as HSO₄⁻ and SO₄²⁻, while phosphate is partitioned between H₂PO₄⁻ and HPO₄²⁻. The substrate surface charge also influences retention: at pH values above the isoelectric point of silicon oxide or ITO, the surface is negatively charged and anionic species are repelled, but at the acidic edge pH, surface silanol groups may be partially protonated, reducing repulsion. In addition, if the glass substrate contains aluminum oxide or silicon nitride barrier layers, positively charged surface sites below their point of zero charge can attract phosphate and sulfate. This electrostatic contribution explains why a simple increase in rinse water volume without pH adjustment often fails to reduce edge residue below the detection limit. Equipment variables such as nozzle angle, exhaust balance, and the delay between EBR dispense and aqueous rinse must therefore be treated as chemical variables, not merely mechanical settings. The absence of a unified standard for acceptable residual phosphate and sulfate at the display substrate edge means that process limits are typically derived from downstream defect data and supplier-qualified EBR formulations; published data for this specific configuration is limited but the mechanistic pathways are well established.

Monitoring Aqueous and Solvent-Based EBR Streams by Ion Chromatography

Anion quantification in EBR fluids and rinse water is performed using suppressed ion chromatography with conductivity detection, as specified in ISO 10304-1:2007 and ASTM D4327-17. Aqueous samples from the post-EBR rinse manifold, coater exhaust condensate, and solvent reclaim water are filtered through a 0.45 µm polyethersulfone membrane and injected onto a high-capacity anion exchange column such as an AS11-HC or AS19 column with a corresponding guard column. For sulfate and phosphate in dilute aqueous samples, a hydroxide eluent gradient from 1 mM to 60 mM with a flow rate of 1.0 mL/min and chemical suppression at 30 mA to 60 mA provides method detection limits below 0.5 µg/L for sulfate and below 1.0 µg/L for phosphate when a 100 µL injection loop is used. Solvent-based EBR samples require additional preparation because direct injection of PGMEA or ethyl lactate damages the analytical column. The usual procedure is a 1:10 to 1:100 dilution into ultrapure water followed by solid-phase extraction or liquid-liquid extraction to remove polymer and photoactive compounds that would foul the column. For trace work, a preconcentration column can be used to load 5–20 mL of diluted sample, bringing the lower reporting limit for sulfate into the sub-0.1 µg/L range. On-line ion chromatography is increasingly installed on EBR solvent distribution loops and waste manifolds because it provides time-resolved profiles of phosphate and sulfate breakthrough from ion exchange columns, filter media, or solvent reclaim systems. The sampling interval in on-line IC is typically 5–20 min, allowing a coater line to detect batch-to-batch contamination before substrate processing is affected. Calibration is performed using certified mixed-anion standards traceable to NIST or equivalent national metrology institutes, and the calibration range is normally 0.5 µg/L to 1000 µg/L for both analytes. Quality control samples at 5 µg/L and 50 µg/L are analyzed in duplicate at the beginning and end of each run to verify instrument drift. The table below summarizes the common analytical configurations for EBR process fluids and rinse waters.

On production coaters for Gen 8.5 glass substrates, the EBR solvent is distributed from a central chemical supply through fluoropolymer lines to the coating cup. Deadlegs in EBR distribution can accumulate phosphate and sulfate, especially when the solvent is hygroscopic and absorbs moisture from the coat exhaust. A bleed-and-feed protocol maintains EBR bath conductivity below 2 µS/cm, corresponding to a total anionic contamination of less than 5 mg/L in the solvent. Conductivity alone, however, cannot distinguish chloride from sulfate or phosphate; it must be paired with weekly ion chromatography or in-line capillary electrophoresis. Point-of-use filtration with a 0.05 µm PTFE or UPE membrane removes particulate phosphate salts but not dissolved ions. For dissolved ion control, some fabs insert a cation exchange column in the EBR reclaim loop, but anion removal from nonaqueous solvents is limited because standard anion exchange resins require aqueous solvation. Instead, the most effective control is replacement of EBR working fluid before accumulation exceeds the salt solubility limit. The absence of a published standard for EBR solvent ionic purity makes supplier-specific limits the only available benchmark; published data for this specific configuration is limited. The table below summarizes representative process control points used in display fab EBR operations.

When Sulfuric Acid-Peroxide Clean Residues Migrate Into the EBR Rinse Water

Sulfuric acid-peroxide mixture is used in display fabs for organic removal before photoresist coating. The cleaning bath typically operates at 70–120 °C with a H₂SO₄:H₂O₂ volume ratio between 3:1 and 6:1. After SPM, substrates pass through a quick dump rinse and a final ultrapure water rinse. If the final rinse resistivity after SPM is below 18 MΩ·cm, sulfate is likely present at the substrate surface. The sulfate is transferred to the coater by the substrate carrier and can be redissolved by the EBR solvent or the post-EBR rinse, then concentrated at the edge. Sulfate aerosol from SPM exhaust can also deposit on coater surfaces and be re-entrained onto the substrate. Segregation of acid cleaning exhaust and coater exhaust reduces this cross-contamination. In shared exhaust systems, sulfate can be detected in coater exhaust condensate at concentrations up to 20 µg/L during SPM bath maintenance. The EBR rinse water should therefore be sampled at the point of use rather than at the central water plant because contamination is often introduced in the tool. Ultrapure water at the point of use should maintain resistivity above 18.2 MΩ·cm at 25 °C, and sulfate concentration below 5 µg/L. Phosphate in SPM cleaning is less common but can appear when the SPM bath is used to strip resists containing phosphate-based adhesion promoters or phosphate ester surfactants. The combination of sulfuric and phosphoric acid residues at the substrate edge is particularly damaging for aluminum gate lines because sulfate can initiate pitting corrosion while phosphate can form insoluble aluminum phosphate films that interfere with subsequent wet etch. Wet etch uniformity on Gen 8.5 aluminum gate layers can be degraded by residual aluminum phosphate islands with heights above 2 nm. These islands are not always detected by optical inspection because their reflectance is similar to the underlying aluminum. Edge bead residue audits must therefore include spatially resolved ion analysis rather than visual inspection alone.

Residue Desorption Kinetics in Post-EBR Aqueous Rinse Programs

Removal of phosphate and sulfate from the EBR edge after drying is mass-transfer limited. In an aqueous rinse, the diffusion coefficient of sulfate at 25 °C is approximately 1.0 × 10⁻⁵ cm²/s, while phosphate diffuses more slowly at roughly 0.6 × 10⁻⁵ cm²/s due to its larger hydrated radius. A spin rinse with DI water dispensed at 1.0–2.5 L/min through a fan nozzle positioned 100–150 mm from the substrate generates a radially moving rinse front that reduces the boundary layer thickness and increases convective transport. Without chemical buffering, the rinse water pH drops from 7.0 to near 4.5 after contact with residual sulfate, reducing electrostatic repulsion between anionic species and the substrate. Adding a dilute volatile base such as 0.05–0.2% ammonium hydroxide to the aqueous rinse raises the pH to 8.0–9.0, deprotonating surface silanol groups and promoting desorption of phosphate and sulfate. However, ammonia can complex copper; on copper gate arrays, the rinse chemistry is replaced by carbonated DI water or a 0.1% tetramethylammonium hydroxide formulation. The optimal rinse time after EBR is typically 15–30 s at 800–1200 rpm; extending beyond 45 s can re-deposit extracted ions onto the substrate due to evaporation-induced upward wicking from the backside. Rinse temperature also matters. At 35–45 °C, the desorption rate of sulfate from oxidized silicon surfaces increases by a factor of 1.5–2.0 compared with 22 °C, but heated rinse may promote phosphate precipitation with calcium if water hardness is not controlled. The use of a final low-flow meniscus rinse after the main spin rinse can reduce residue at the edge by an additional 20–50% because it creates a moving solvent wedge that sweeps desorbed ions off the substrate. These rinse parameters must be validated with edge extraction tests rather than only rinse water conductivity. Published data for polyimide substrates is limited, but for glass and silicon oxide surfaces the trend is reproducible.

Drying after the post-EBR rinse is a critical residue transfer step. In solvent-based EBR, the rim of the substrate experiences evaporative cooling, which can reduce surface temperature by 8–15 °C below ambient and cause condensation of humid air. The resulting microdroplets solubilize acidic residues and spread them inward. Maintaining coater exhaust relative humidity at 45–55% and airflow velocity at 0.35–0.50 m/s across the substrate suppresses condensation. In high-humidity regions where exhaust RH exceeds 60%, substrate pre-drying before coat is required. On glass substrates with a thin silicon nitride or silicon oxide barrier, sulfate retention is lower than on bare glass; published data comparing phosphate retention on polyimide backplanes is limited. In addition, exhaust condensation on the EBR nozzle itself can drip back onto the substrate, and the nozzle should be purged with nitrogen between substrates when the fab is in a high-RH season. Nitrogen purge at 2–5 L/min through the EBR nozzle body reduces moisture accumulation and prevents phosphate and sulfate salts from building up at the nozzle lip.

Edge bead residue auditing on organic planarization layers and polyimide substrates requires extraction and ion quantification at the panel edge. A common procedure uses a masked extraction cell with an area of 1 cm² placed on the edge zone, filled with 10 mL of ultrapure water or 0.1% isopropanol, and sonicated for 30 min at 40 kHz. The extract is then analyzed by ion chromatography. For organic planarization layers, the extraction solvent may need to be a water-miscible organic solvent such as 50:50 methanol/water to wet the surface; the dilution factor must be accounted for in the final ion concentration. Residue levels are expressed as nanograms per square centimeter. On hardened color filter photoresist edges, phosphate and sulfate can be trapped in the pigment dispersion. In such cases, the extraction may require heating to 60 °C and adding a chelating agent to release phosphate bound to pigment surfaces. The edge bead removal process on color filter lines is particularly sensitive to sulfate because sulfonated pigment dispersants may be present in the resist; the EBR solvent can co-dissolve these dispersants and leave sulfur-containing films at the edge. If these residues are not removed before the ITO sputter step, the ITO film can show local sheet resistance shifts of 2–5% along the substrate edge. Such shifts are detectable only by mapping four-point probe measurements across edge exclusion zones. Process control at this level requires integration of EBR solvent chemistry, rinse pH, drying humidity, and analytical extraction methods. Without this integration, anion residues at the edge persist as a variable that can shift wet etch performance and metal film quality even when the visual edge bead removal appears acceptable.