| HS Code | 188177 |
| Product | JSR GR-900 |
| Resist Type | Positive G-line photoresist |
| Exposure Wavelength | g-line (436 nm) |
| Resin Composition | Novolac resin |
| Photoactive Compound | Diazonaphthoquinone (DNQ) |
| Developer Compatibility | 2.38% aqueous TMAH developer |
| Typical Film Thickness | About 1.0 µm depending on spin-coating conditions |
| Resolution Class | Sub-micrometer, generally around 0.5 µm |
| Spectral Sensitivity | Sensitive in the 436 nm G-line region |
| Contrast Profile | High contrast for steep resist sidewalls |
| Storage Condition | Store sealed at 10–25°C away from light |
| Shelf Life | Typically 12 months from date of manufacture |
As an accredited G-line Photoresist JSR GR-900 factory, we enforce strict quality protocols—every batch undergoes rigorous testing to ensure consistent efficacy and safety standards.
| Packing | G-line photoresist JSR GR-900 is packaged in a sealed, light-resistant 4-liter bottle, with label and safety information. |
| Container Loading (20′ FCL) | G-line photoresist JSR GR-900 in sealed containers, loaded into 20-foot FCL, secured with dunnage, labeled properly, and documented for safe transport. |
| Shipping | Ship G-line Photoresist JSR GR-900 in tightly sealed, UN-approved containers, protected from light and moisture. Label as flammable/irritant, include SDS, and use grounded, temperature-controlled transport. Avoid extreme heat or freezing, secure against shifting, and ensure compatible segregation from oxidizers and foodstuffs. |
| Storage | Store G-line Photoresist JSR GR-900 in its original, tightly sealed container in a cool, dry, well-ventilated area away from direct sunlight, UV light, heat, and ignition sources. Maintain recommended temperature (typically 5–25°C). Keep away from oxidizers and incompatible materials. Avoid freezing. Ensure container remains upright and check for leaks regularly. |
| Shelf Life | Shelf life is typically 6 months from manufacture if stored sealed in the original container at 5–30°C, away from light. |
A 0.45 NA g-line reduction stepper at 436 nm can hold isolated line CD of 0.8 µm only when the resist swing curve is mapped against the substrate optical stack. For JSR GR-900, published swing-curve data on polysilicon and field-oxide substrates is limited; the process window below reflects the g-line DNQ-novolac class and must be verified on the production track. Dehydration bake at 120°C for 120 s is followed by HMDS vapour prime in a closed chamber at 120°C for 60 s. The film is dynamically dispensed and spun to 1.1 µm thickness at 3,100 rpm on a coater track with 18 MΩ·cm DI rinse. Softbake on a proximity hotplate at 100°C for 90 s is preferred over convection ovens because edge-to-centre solvent depletion shifts by ±0.08 µm when bake uniformity exceeds ±2°C. Exposure dose in the 65–95 mJ/cm² range is established by swing curve; isolated line, dense line, and contact hole features are measured on a Hitachi CD-SEM using 0.8 µm test structures. Puddle development with 2.38% TMAH at 23°C for 60 s, followed by DI rinse and spin dry, defines the pattern. Hardbake at 120°C for 120 s improves dry etch resistance for polysilicon and silicon dioxide etches. Production tracks with recirculating developer baths require chillers holding 23°C ± 0.5°C; deviation below 21°C is a known cause of 0.2–0.3 µm CD drift at the wafer edge. The terminal application is a 0.8–1.2 µm non-critical CMOS layer on 150 mm wafers. Compliance for this integration is anchored to ISO 14644-1:2015 Class 4 cleanroom conditions, SEMI S2 for equipment safety, and SEMI S6 for exhaust and chemical handling.
On MEMS lines where sacrificial oxide release precedes metal patterning, the resist must survive buffered oxide etch chemistry long enough to protect aluminium bond pads and silicon nitride passivation. A 2.5 µm film is spin-coated on the MEMS wafer at 1,800 rpm and softbaked at 115°C for 180 s on a vacuum hotplate with ±1°C uniformity. Exposure on a contact aligner with 436 nm filter at 100–130 mJ/cm² clears the sacrificial oxide open areas; puddle development in 2.38% TMAH for 90 s produces 20 µm anchor pad openings. After post-develop inspection under yellow light, the wafer is descummed in an oxygen plasma at 50 W for 60 s because residual DNQ photoacid clusters on oxide inhibit wet etch initiation. Buffered oxide etch in 6:1 NH₄F:HF at 23°C removes sacrificial SiO₂ at 80 nm/min; the hardened resist film must withstand 10–15 min immersion without lifting. Cantilever structures have exhibited edge lifting when the hardbake ramp exceeds 10°C/min and when relative humidity during coating exceeds 60%. A hardbake at 130°C for 150 s with ramp rate of 5°C/min eliminates this failure mode. In plasma-released MEMS structures, the same film is used as a dry-etch mask in an Oxford Instruments PlasmaPro 100 ICP etcher with SF₆/O₂ chemistry at 3.5 mTorr, coil power 800 W, platen power 50 W; resist erosion is 60–100 nm/min, giving selectivity of approximately 15:1 over silicon. Thermal budget conflicts arise when sacrificial oxide is released before metal deposition, because resist must be stripped before aluminium sputtering but residue on oxide complicates the subsequent lift-off profile. Published GR-900-specific adhesion data on MEMS substrates is limited; the above values are class-typical starting conditions and require confirmation by cross-section SEM on the actual device stack. The terminal structure is a released MEMS cantilever or membrane with protected metal anchors. Cleanroom classification is ISO 14644-1:2015 Class 5, and chemical handling follows SEMI S2 within a ventilated acid hood.
Wafer-level redistribution flows use a single-coat template that must hold sidewall angles above 70° through acid copper electroplating at 25°C. On a sputtered Ti/Cu seed layer, a 4.5 µm film is obtained at 1,600 rpm and softbaked at 110°C for 120 s. Exposure on a broadband contact aligner with 436 nm filter at 120 mJ/cm² clears the redistribution trace and bump pad openings. Development in 2.38% TMAH for 75 s is followed by descum at 100 W O₂ plasma for 45 s to remove seed-layer oxide and DNQ residue. Acid copper plating in a vertical bath containing 50 g/L CuSO₄·5H₂O and 10 vol% H₂SO₄ at 10–20 mA/cm² deposits 3 µm traces; resist cracking is observed when current density exceeds 25 mA/cm² or bath temperature drifts above 28°C. For gold bump templates, a cyanide-free sulfite gold bath at 60°C and pH 9.5 demands resist resistance to alkaline immersion for 30–60 min; edge lifting on a 100 mm wafer has been traced to incomplete HMDS priming of the copper seed ring. Hardbake at 125°C for 150 s is used only after copper plating because a hardbake before plating increases stripper residue in narrow 10 µm spaces. Resist stripping in NMP at 80°C for 10 min followed by O₂ ash at 150 W for 120 s avoids metal attack. Adhesion verification follows ASTM D3359-17 tape peel method; the target classification is 4B or better on the plated copper surface. The terminal product is a wafer-level chip-scale package with 3–5 µm Cu RDL traces or shallow gold bump pads. GR-900-specific adhesion data on sputtered Cu seed is limited; a seed-layer surface energy check with contact angle below 20° after HMDS prime is recommended before volume lot release. Compliance for bump template processing is governed by SEMI S8, SEMI S10 for plating safety, and the final device RoHS Directive 2011/65/EU.
In 100 mm gallium phosphide rectifier production, contact aligner exposure dose drifts upward as mercury arc lamps age past 800 h. A 1.5 µm film is dispensed at 2,800 rpm and softbaked at 105°C for 90 s. The aligner’s 436 nm bandpass filter is replaced at 1,000 h; dose at the wafer plane is checked with a radiometer calibrated to a NIST-traceable 436 nm standard. Puddle development in 2.38% TMAH for 45 s clears the mesa pattern; hardbake at 135°C for 180 s precedes wet etching of the GaP mesa. In LED metal contact lithography on arsenide/phosphide substrates, residual resist in 5 µm anode vias has been traced to excessive softbake temperature above 110°C, which accelerates novolac crosslinking and lowers develop solubility. The process window is narrowed by the ±3°C hotplate tolerance of the contact aligner bake unit. Terminal devices are passivated planar p-n junction GPP rectifiers and LED mesa chips. Published GR-900-specific data on GaP and GaAs substrates is limited; contrast and adhesion must be re-established with each substrate lot. Cleanroom requirement is ISO 14644-1:2015 Class 5, and process chemical handling follows SEMI S2 with local exhaust face velocity of 0.5 m/s.
Resist edge topology over 0.4 µm Ra grain boundaries controls whether wet etching leaves shunt current paths in thin-film nichrome resistor networks. A 1.8 µm film is spin-coated at 2,200 rpm on 96% alumina substrates; the high edge-bead produced by surface roughness is removed with a perfluorinated solvent edge rinse. Softbake at 105°C for 90 s on a ceramic-specific hotplate with vacuum channels is used because contact heating on porous alumina varies by ±5°C. Exposure on a 1:1 mask aligner at 100 mJ/cm² prints resistor track widths of 25 µm; development in 2.38% TMAH for 60 s clears the pattern over the grain boundaries. Ion-beam milling of 50 Ω/sq nichrome at 400 eV beam energy and 0.3 mA/cm² current density is hardmasked by the resist; resist erosion of 80 nm/min requires a 1.8 µm starting thickness to guarantee protection of the underlying metal. After etching, resist stripping with NMP at 80°C followed by oxygen plasma at 150 W for 120 s removes residue. Laser trim stability is tested according to MIL-PRF-38534 Class K; excess carbon residue from incomplete ash shifts the temperature coefficient of resistance by ±2 ppm/°C. The terminal product is a thin-film hybrid microcircuit resistor network with 25 µm track pitch on alumina. GR-900-specific data on ceramic substrates is not publicly established; the above thickness and etch selectivity values are class-typical and require a substrate-specific qualification. Compliance for hybrid microcircuit assembly also references ISO 14644-1:2015 Class 5, ASTM D3359-17 for tape adhesion, and NASA-STD-8739.6 for workmanship when the resist is used in space-grade hybrid assembly.
| Standard / test | Scope in this document | Typical requirement |
|---|---|---|
| ISO 14644-1:2015 | Cleanroom classification for coat/develop areas | Class 4, Class 5, or Class 6 depending on substrate |
| SEMI S2 | Equipment safety and environmental performance | Exhaust face velocity 0.5 m/s for dispense hood |
| SEMI S6 | Photolithography equipment exhaust and drainage | Solvent vapour below 20% LEL |
| SEMI S8 | Ergonomic and manual handling of wafer tracks | Manual cassette loading below 25 kg |
| SEMI S10 | Electroplating safety and chemical handling | Interlocked plating bath exhaust |
| SEMI S26 | Flat panel display substrate handling and safety | Glass cassette robotic interlock |
| ASTM D3359-17 | Cross-cut tape adhesion on metal seed and ceramic | Classification 4B or better |
| MIL-PRF-38534 | Hybrid microcircuit assembly qualification | Class K or Class H depending final use |
| 2011/65/EU | RoHS recast in final packaged device | Device-level lead restriction compliance |
| REACH 1907/2006 | Chemical import and use documentation | Article 33 communication for SVHC if present |
When interdigital transducer fingers drop below 1.2 µm on 128° Y-cut lithium niobate, line-end shortening in a 1:1 contact aligner introduces electrode frequency scatter. A 1.0 µm film is spin-coated at 3,500 rpm; the lithium niobate substrate is pre-baked at 120°C for 60 s to remove surface-bound water that degrades resist adhesion. Softbake at 100°C for 90 s is followed by exposure at 90 mJ/cm² through a chromium photomask with 0.8 µm finger openings. Development in 2.38% TMAH for 50 s produces finger electrodes for aluminium deposition; the resist is used as a liftoff mask. A post-exposure bake is intentionally omitted because elevated temperature would reflow the resist edge and increase finger width by 0.1 µm. After electron-beam evaporation of 120 nm aluminium, liftoff in NMP at 80°C under ultrasonic agitation at 40 kHz removes the resist. Electrical test of the transducer frequency response is performed with a vector network analyzer to ±5 MHz; CD variation above 0.05 µm shifts the resonance beyond the passband. This liftoff flow is sensitive to solvent content in the resist; incomplete softbake leaves residual solvent that redissolves the photoactive compound and creates rounded finger ends. The terminal device is a band-pass SAW filter. Published GR-900-specific data on pyroelectric substrates is limited; a contrast enhancement layer may be needed to suppress standing waves. Cleanroom classification is ISO 14644-1:2015 Class 5, and handling follows SEMI S2 with static discharge control on lithium niobate cassettes.
The softbake hotplate on a Gen 4 TFT line exhibits 8°C edge-to-centre temperature spread unless vacuum contact and exhaust baffle settings are rebalanced after every shut-down. A 1.3 µm film is slit-coated on 550 mm × 650 mm glass at a line speed of 50 mm/s; spin coating is not used for this area. Proximity exposure on a g-line aligner at 436 nm uses a 50 µm proximity gap and 80 mJ/cm² dose. Puddle development in 2.38% TMAH for 60 s clears contact holes for TFT source/drain vias. Hardbake at 125°C for 120 s precedes wet etching of Mo/Al/Mo interconnects with a PAN etchant at 40°C; resist delamination is observed when the hardbake temperature drops below 120°C because the film absorbs water from the etch bath. Film thickness uniformity is measured with a spectroscopic reflectometer at 13 points per substrate; the specification for TFT array patterning is ±0.05 µm. A process conflict emerges when cleanroom exhaust changes during production shift, shifting the proximity gap by 5 µm and broadening the exposure gap. The terminal product is a TFT array layer on display glass. GR-900-specific data on display-grade glass is limited; display-specific lot qualification is required before volume use. Equipment compliance follows SEMI S2 and SEMI S26 for flat panel display substrate handling; cleanroom class is ISO 14644-1:2015 Class 6.
Competitive G-line Photoresist JSR GR-900 prices that fit your budget—flexible terms and customized quotes for every order.
For samples, pricing, or more information, please contact us at +8615365186327 or mail to admin@ascent-chem.com.
We will respond to you as soon as possible.
Tel: +8615365186327
Email: admin@ascent-chem.com
Flexible payment, competitive price, premium service - Inquire now!
JSR GR-900 is a positive-tone g-line photoresist formulated around a diazonaphthoquinone (DNQ) photosensitizer dispersed in a novolak resin matrix. The product is designated for primary exposure at 436 nm and is used on reduction projection steppers, contact aligners, and proximity exposure tools where feature dimensions do not require deep-ultraviolet resolution. In exposed regions, DNQ undergoes photochemical rearrangement to an indene carboxylic acid, increasing the dissolution rate of the exposed resist in aqueous alkaline developer; unexposed novolak-rich regions remain significantly less soluble. Lot-specific physical constants, including kinematic viscosity, non-volatile content, water content, trace metal concentrations, and particle counts, are controlled by JSR and reported in the certificate of analysis. Because JSR does not publish a complete numerical datasheet for this specific product, the class-typical ranges cited in this document are engineering references for g-line DNQ/novolak photoresists and must be verified against the GR-900 certificate of analysis.
Material handling is performed in a cleanroom complying with ISO 14644-1:2015, typically at Class 5 or better. The resist is supplied as a light-sensitive viscous liquid in solvent-resistant containers. Storage at 5–15 °C is common for DNQ/novolak resists of this class; containers should be brought to dispense temperature before opening to avoid condensation. Solvent loss through repeated freeze–thaw cycling or loose container seals changes solids content and spin-speed film thickness, and photosensitizer precipitation may occur if the material is stored below the solvent-cloud point. Filtration at 0.2 µm absolute is typical in the supply chain, but the user should confirm the lot-specific particle specification from the certificate of analysis.
The main process applications for GR-900 are aluminium interconnect patterning, polysilicon gate and passivation layer definition, MEMS sacrificial and structural layers, and wafer-level bump or pillar electroplating templates. The material is not intended for sub-0.25 µm front-end lithography, where chemically amplified deep-UV resists are required because of shorter exposure wavelength and higher imaging contrast.
Incoming inspection should confirm lot-specific viscosity, water content, solids content, trace metal contamination, and particle counts before production release. Kinematic viscosity can be measured at 25.0 °C in accordance with ASTM D445-24; density can be checked with ASTM D4052-22; water content can be measured by Karl Fischer titration following ASTM D6869-17. Viscosity is used to predict spin-speed film thickness on a given wafer track. A change in viscosity outside the controlled lot range shifts film thickness and may alter subsequent development time, dimensional bias, and residue levels. Water content is monitored because excess water can affect photosensitizer solubility and development rate in DNQ/novolak systems. Trace metal analysis is critical for applications involving aluminium or copper metallization, where mobile-ion contamination must remain below the specified limits.
| Attribute | Description |
|---|---|
| Model | JSR GR-900 |
| Resist tone | Positive |
| Chemistry | DNQ/novolak |
| Primary exposure wavelength | 436 nm |
| Solvent family | PGMEA-based |
| Developer type | Aqueous alkaline, typically 2.38 wt% TMAH |
| Lot release data | Certificate of analysis |
Published data for this specific configuration is limited; therefore the following ranges are class-typical values for g-line DNQ/novolak photoresists and not substitute for JSR’s lot release limits. Typical film thickness for spin coating at 1,000–5,000 rpm falls between 0.5 µm and 3.0 µm, depending on solvent balance and dispense volume. A prebake of 90–110 °C for 60–120 s on a contact hotplate is used to remove residual solvent and stabilize the film. Exposure energy for 436 nm exposure is commonly in the 80–200 mJ/cm² range for g-line DNQ/novolak resists, but the actual lot-specific sensitivity must be established by a contrast curve or clearing-dose array on the production exposure tool.
Prebake latitude is a critical threshold. If prebake temperature is more than ±2 °C from the optimum for a fixed time, residual solvent content changes enough to alter development rate and linewidth. In production tracks, hotplate uniformity is therefore specified at ±1 °C across the wafer. Hotplate lid closure and exhaust balance prevent solvent vapor condensation on the backside. A prebake hotplate with poor contact can produce radial thickness and sensitivity gradients; this is observed as a center-to-edge linewidth difference after development.
Adhesion promotion is necessary on metallic and polysilicon surfaces. A vapor prime with hexamethyldisilazane (HMDS) at 120–150 °C for 30–60 s is standard on automated wafer tracks. On aluminium layers, the high reflectivity at g-line wavelengths creates a standing-wave pattern in the resist film. The resulting constructive and destructive interference produces sinusoidal variation in the absorbed dose through the resist depth, which can appear as notches in the sidewall profile and periodic linewidth changes as a function of resist thickness. Swing curves on aluminium substrates are used to select a film thickness at a minimum of the reflectance swing, or a bottom anti-reflective coating is added to reduce substrate reflectivity below 10 %. In production, reflective notching is minimized by controlling film thickness within ±3 % of the target and by maintaining developer normality through a point-of-use dispense system.
On polysilicon gates, the dominant process risk is substrate topography. Step coverage must be sufficient to avoid edge thinning over field-oxide steps; class-typical g-line DNQ/novolak resists require a thickness at least 1.5 times the nominal step height for acceptable coverage. If the film is too thin at a step edge, over-development can produce breaks in the resist line and allow etch attack. Batch-to-batch viscosity differences are controlled by solvent addition and are verified by spin curve. Production-scale experience on 100 mm and 150 mm wafer tracks shows that film thickness variation from edge bead removal and exhaust flow can be larger than lot-to-lot viscosity variation when dispense volume is not optimized.
Exposure on g-line steppers and mask aligners is typically performed without post-exposure bake for DNQ/novolak systems. Development is carried out in aqueous tetramethylammonium hydroxide, with 2.38 wt% TMAH as the standard concentration for high-purity semiconductor processing. Single-puddle and double-puddle development recipes are used depending on feature density and dark erosion requirements. The development endpoint is influenced by developer temperature, which should be controlled to ±0.5 °C. Higher developer temperature increases dark erosion and may reduce adhesion; lower temperature increases development time and may leave residue in low-dose regions. The contrast of g-line DNQ/novolak resists is sufficient for linear features in the 0.8–1.5 µm range on typical exposure tools, but optical proximity effects and mask biasing become significant below 1.0 µm.
| Process step | Class-typical range | Measurement/control |
|---|---|---|
| Dehydration bake | 120–150 °C, 30–60 s | Hotplate thermocouple |
| HMDS vapor prime | 120–150 °C, 30–60 s | Vapor prime chamber |
| Spin coating | 1,000–5,000 rpm, ramp 10,000 rpm/s | Enclosed spin bowl, exhaust flow |
| Prebake | 90–110 °C, 60–120 s | Contact hotplate, ±1 °C uniformity |
| Exposure | 80–200 mJ/cm² at 436 nm | Exposure tool radiometer |
| Development | 2.38 wt% TMAH, 30–60 s single puddle | Temperature ±0.5 °C |
| Postbake/hardbake | 100–140 °C, 60–120 s | Hotplate uniformity |
JSR GR-900 is distinguished from i-line resists primarily by spectral sensitivity and intended exposure wavelength. I-line photoresists are formulated for 365 nm exposure and are used on higher-numerical-aperture steppers, allowing smaller feature sizes. G-line DNQ/novolak resists operate at 436 nm and are used on older reduction steppers, mask aligners, and proximity systems where resolution requirements are less demanding. Because g-line exposure uses a longer wavelength, the optical resolution limit for a given numerical aperture is larger than that of i-line imaging. In practical terms, i-line resists may resolve features at or below 0.35 µm on advanced i-line steppers, whereas g-line products of the GR-900 class are generally used for feature sizes near or above 0.8 µm.
Compared with chemically amplified resists, DNQ/novolak materials do not require a post-exposure bake to drive acid-catalyzed deprotection. This simplifies process flow and reduces sensitivity to post-exposure delay. DNQ/novolak resists are generally less sensitive than chemically amplified resists, and they are not suitable for deep-UV wavelengths because the novolak matrix absorbs strongly at 248 nm and 193 nm. The operational stability of DNQ/novolak materials is relevant in contact and proximity aligners where environmental base contamination is less tightly controlled than in deep-UV tracks. However, the resolution ceiling and dose latitude of DNQ/novolak g-line materials are lower than those of chemically amplified 193 nm systems.
Compared with broadband g-line/h-line resists, a g-line-specific product may show reduced sensitivity at 365 nm; this is advantageous when i-line emission produces ghost images in mask aligners. Some broadband resists are formulated to clear at multiple wavelengths, which can increase linewidth variance when the exposure source spectrum drifts. A monochromatic g-line formulation reduces the contribution of shorter wavelengths to the latent image, improving linewidth stability on mixed-wavelength tools. Published data for JSR GR-900 specific spectral absorbance is limited, so a spectral sensitivity test on the specific aligner is required.
Regulatory documentation for JSR GR-900 includes a Safety Data Sheet prepared under EC No. 1272/2008 and REACH Regulation EC No. 1907/2006. Users in electrical and electronic equipment applications should verify compliance with RoHS Directive 2011/65/EU and any customer-specific halogen or metal contamination limits. The solvent system is combustible; dispensing and coating operations must follow local fire codes and extraction requirements. Waste developer containing dissolved novolak and DNQ should be segregated from solvent waste and disposed through an approved chemical waste contractor.
Post-development hardbake is used to improve adhesion and chemical resistance for wet etching and electroplating. In g-line DNQ/novolak resists, thermal flow begins near the glass transition of the novolak matrix. For this class, hardbake temperatures are usually between 100 °C and 140 °C. Baking above 140 °C can round features and reduce linewidth control, particularly in isolated lines and small contact holes. In electroplating applications, a crosslinked hardbaked film may be required to resist acidic copper or nickel plating baths; the same hardbake can increase difficulty of final stripping. Plasma ashing or solvent strippers should be selected after evaluating the degree of crosslinking produced by the hardbake. If the pattern is used for lift-off, a negative slope or controlled re-entrant profile is often generated by using a g-line resist with a slower development rate at the surface, but published data for this specific configuration is limited and process development on the target substrate is required.
Process monitoring on production aligners with mercury arc lamps should include periodic measurement of 436 nm output power, since lamp aging changes the dose delivered to the wafer. If dose compensation is not used, linewidth shifts because DNQ/novolak resists have finite contrast. Exposure dose is therefore checked with a radiometer traceable to a national metrology institute, and the exposure time is adjusted when lamp output falls below the established control limit. On production tracks, developer concentration and temperature are monitored with conductivity and temperature sensors; a loss of developer normality can produce incomplete clearing in dense features while a higher normality can increase dark erosion. These process controls are standard for g-line DNQ/novolak materials and should be applied to GR-900 on the specific exposure tool.
