KEHUA KPG-200 is specified as a positive-tone photoresist for the 436 nm g-line emission of mercury-discharge lithography lamps. The material belongs to the diazonaphthoquinone/novolak class and develops in aqueous alkaline solution, with exposed areas undergoing selective dissolution in tetramethylammonium hydroxide-based developers. The model designation KPG-200 identifies a g-line formulation for contact aligners, proximity lithography systems, and g-line steppers used in semiconductor, compound semiconductor, MEMS, and thin-film patterning. Published comparative data for this specific grade against competing resists remains limited; therefore, the numerical values given here are either representative ranges for conventional g-line DNQ/novolak positive resists or values derived from standard optical and processing principles. Lot-specific specifications should be obtained from the manufacturer’s certificate of analysis before a production process is fixed.

What Differentiates KPG-200 Within the 436 nm DNQ/Novolak Class?

Three process-relevant variables separate g-line DNQ/novolak photoresists: novolak molecular-weight distribution, diazonaphthoquinone ester loading, and solvent balance. KPG-200 is described as a general-purpose positive resist, but public technical data for its specific viscosity, solids content, and contrast curve are limited. In the absence of published contrast data, the material should be characterized on a production g-line stepper or contact aligner using a focus-exposure matrix. For a conventional 436 nm DNQ/novolak resist, practical resolution on low-numerical-aperture contact or proximity equipment is generally bounded near 0.8–1.5 µm; the exact value depends on mask-to-wafer gap, exposure mode, substrate reflectivity, and development conditions.

KPG-200 is not interchangeable with i-line photoresists tuned for 365 nm. Although both use DNQ/novolak chemistry, the sensitizer absorption and resin dissolution selectivity are optimized differently. Products in this class are normally supplied as filtered solutions with solids content between 20 % and 35 % by weight; viscosity at 25 °C generally spans 5–50 mPa·s depending on the target film thickness. If KPG-200 is offered as a single viscosity grade, its spin curve must be validated on the intended wafer track rather than extrapolated from formulations of similar appearance. Viscosity should be measured by cone-and-plate viscometry under ASTM D4287-19 or ISO 2884-2; the exact filtration retention rating and trace-metal limits must appear on the certificate of analysis.

A conventional process sequence for this chemistry begins with a dehydration bake on silicon or silicon dioxide at 150–200 °C for 30 min in an oven or 60 s on a vacuum hotplate, followed by hexamethyldisilazane vapor priming at 90–110 °C for 60 s. The resist is dispensed through a point-of-use filter and spin-coated at 1,000–4,000 min⁻¹ for 30 s, then softbaked at 90–110 °C for 60–90 s on a contact hotplate with surface uniformity better than ±1 °C. Hotplate processing is preferred over convection ovens because solvent removal is more repeatable across the wafer. On closed-bowl spin coaters with controlled exhaust, edge-bead removal may require a solvent nozzle; if the edge bead is not removed, hard-contact printing on aligners can generate defects and reduce mask life. The representative exposure dose at 436 nm for this class is 50–150 mJ/cm²; development in 2.38 % tetramethylammonium hydroxide for 45–60 s is the normal starting point, but the KPG-200 lot-specific optimum may require adjustment.

Film Thickness, Viscosity, and Spin Curve Considerations

For weakly shear-thinning photoresist solutions, film thickness after spin coating is approximately proportional to the inverse square root of spin speed. A KPG-200 film on a 150 mm or 200 mm wafer must be mapped with the actual dispensed volume, exhaust flow, and substrate diameter; edge-bead morphology differs from small coupons. Viscosity data obtained under ASTM D4287-19 or ISO 2884-2 should be recorded at 25 °C because a temperature drift of 2 °C can shift viscosity sufficiently to alter thickness by several tenths of a micrometer. Dispense lines and bottles should therefore be equilibrated to 20–25 °C before coating. Solvent evaporation during idle handling raises viscosity and promotes particles or striations; the bottle must be closed immediately and the dispense line should be kept free of stagnant resist. On production equipment with hotplate surface uniformity of ±1 °C, the softbake temperature profile should be verified with a substrate-embedded thermocouple or equivalent contact thermometry rather than inferred from the hotplate setpoint alone.

When the Resist Is Exposed at 436 nm

The DNQ/novolak response at 436 nm is not a simple linear bleaching process. The diazonaphthoquinone sulfonate ester absorbs near-ultraviolet light, undergoes photolysis to a ketene intermediate, and hydrolyzes to an indene carboxylic acid that increases the dissolution rate of exposed resist in aqueous base. Contrast is influenced by novolak molecular weight, degree of DNQ substitution, and residual solvent after softbake. Exposure doses for conventional g-line positive resists commonly occupy 50–150 mJ/cm² at 436 nm, but the exact KPG-200 dose must be derived from a production focus-exposure matrix. Dill A, B, and C parameters can be measured by spectrophotometric transmission and are required for lithography simulation; published Dill parameters for KPG-200 specifically are limited, so simulation should use measured film transmission rather than borrowed literature values.

Feature profiles in 2.38 % TMAH puddle development are sensitive to developer temperature, puddle time, and wafer rotation during rinse. A developer temperature variation of ±0.5 °C can shift linewidth by a measurable amount; therefore the developer bath or puddle module should include temperature control. Dilution of developer is not recommended without a full revalidation because it changes dissolution selectivity and may increase scum formation on oxide or nitride surfaces. Post-exposure delay longer than 30 min may allow environmental water or amines to affect the latent image; for DNQ/novolak systems the effect is generally smaller than for chemically amplified deep-ultraviolet resists, but in production lines with poor airflow it can still degrade linewidth control.

Compared with i-line photoresists, g-line photoresists generally exhibit lower resolution because the longer wavelength increases the minimum resolvable feature for a given numerical aperture. On a 0.28 NA g-line stepper with a coherence factor of 0.5 and a k1 factor of 0.65, the Rayleigh resolution approximates 1.0 µm; under the same k1 on a 0.45 NA i-line stepper, resolution improves to approximately 0.53 µm. KPG-200 should therefore be evaluated for legacy g-line equipment or for processes where thicker films and wet-etch resistance are more important than sub-half-micrometer resolution. Compared with chemically amplified deep-ultraviolet resists at 248 nm, KPG-200 and other DNQ/novolak systems are less sensitive to airborne amine contamination and can be processed with simpler environmental controls, but they require higher exposure doses and cannot print features below roughly 0.8 µm without resolution-enhancement techniques.

Thermal Stability Is Not a Substitute for Environmental Control

Novolak-based g-line resists are thermally robust relative to some chemically amplified DUV resists, but thermal stability cannot be assumed without wafer-level data. Hardbake temperatures above 120–140 °C can harden the film for wet etching; excessive temperature or duration may reduce stripability and cause feature flow or rounding. Because a published hardbake window for KPG-200 is not available in widely distributed industrial literature, a hotplate or convection-oven study should map linewidth change and film retention before the process is released. Adhesion on silicon, silicon dioxide, silicon nitride, and metal films should be verified with a pull-off or tape test under ASTM D3359-17, because photoresist adhesion is not equivalent to mechanical coating adhesion after developer and rinse exposure.

Humidity control is a separate boundary. Substrates should be dehydrated immediately before priming; rehydration delays as short as 30 min can reduce hexamethyldisilazane effectiveness. Processing in relative humidity above 60 % can alter development rate and adhesion, particularly on hydrophilic oxide surfaces. The resist itself is solvent-based; unprotected storage in alkaline or high-humidity environments may reduce sensitivity and increase surface skin formation. Failure modes observed on g-line lines include scumming in high-humidity conditions, foot formation from inadequate exposure, edge-bead delamination during etch, and linewidth drift after post-exposure delay.

Batch-to-batch variation in viscosity, solids content, and photospeed should be monitored by incoming quality control using a cone-and-plate viscometer, a UV-visible spectrophotometer, and a standardized wafer-level exposure test. A change of more than ±2 % in film thickness or ±10 % in exposure dose from the qualified reference lot indicates a process shift; the material should be requalified before production use. This is particularly important on g-line contact aligners where lamp aging and mask contamination can mask a resist lot change.

For Metal Lift-Off, Profile Control Depends on Development Stop

In metal lift-off applications, sidewall angle is more important than ultimate resolution. A conventional positive DNQ/novolak resist can be made to undercut by a controlled chlorobenzene soak before development, but the process is sensitive to soak time, softbake condition, and ambient temperature. Published data for KPG-200 in chlorobenzene-assisted lift-off is limited; evaluation should therefore include cross-sectional scanning electron microscopy of the developed profile and lift-off yield after metal deposition. Standard 2.38 % TMAH puddle development is the normal starting condition, but the puddle time must be adjusted only after critical dimension and residual scum are confirmed. Amine-based image-reversal additives should not be mixed into KPG-200 without manufacturer compatibility data because amine contamination can alter g-line sensitivity and storage stability.

Solvent-based photoresist products are flammable liquids; handling must follow local fire-code requirements and the safety data sheet. The lithography environment should meet ISO 14644-1 Class 5 or better for particle control, especially during coating and development. Compliance with RoHS Directive 2011/65/EU and REACH Regulation (EC) No 1907/2006 must be verified against the manufacturer’s certificate of compliance; the presence of novolak and diazonaphthoquinone derivatives does not by itself establish regulatory fitness for every downstream application. Semiconductor-grade g-line resists often specify trace metal levels below 1 ppm for sodium, potassium, and iron and below 100 ppb for mobile ion species; if KPG-200 is used in gate-level or metal-insulator-semiconductor processes, the certificate of analysis must confirm the required metal limits for the specific device.