Spin Coating & Development: A Complete Guide to Photoresist Processing

By NineScrolls Engineering · 2026-03-27 · 16 min read · Nanotechnology

Target Readers: Photolithography process engineers, cleanroom technicians, research scientists, lab managers, and technical decision-makers evaluating spin coating and development equipment. Newcomers will find the fundamentals and process flow sections valuable; experienced engineers can skip to the parameter optimization tables, troubleshooting guide, and photoresist selection chart.

TL;DR

1) Spin Coating Fundamentals

Spin coating is the dominant method for depositing uniform thin films of photoresist, dielectrics, and other solution-processable materials onto flat substrates. The technique is elegantly simple: dispense a liquid onto a substrate, spin the substrate at high speed, and let centrifugal force and solvent evaporation produce a uniform film. Despite this simplicity, the underlying physics involves a rich interplay of fluid dynamics, rheology, and mass transport.

1.1 The Physics of Spin Coating

When a viscous liquid is placed on a spinning substrate, three forces compete:

The spin coating process proceeds through four distinct phases:

  1. Deposition: Liquid is dispensed onto the substrate (static or dynamic dispense).
  2. Spin-up: The substrate accelerates to the target speed. During this phase, centrifugal force rapidly ejects excess liquid. The acceleration ramp rate determines how the initial fluid redistribution occurs.
  3. Spin-off: At constant speed, the film thins primarily through radial outflow. Excess liquid is flung off the substrate edge.
  4. Evaporation-dominated thinning: As the film becomes thin enough that viscous forces dominate centrifugal ejection, further thinning occurs primarily through solvent evaporation. This phase determines the final film thickness.
Four stages of spin coating: dispense, spin-up, spin-off, and evaporation with cross-section diagrams

1.2 The Film Thickness Equation

The Meyerhofer model provides the foundational relationship for spin-coated film thickness:

t = k · Cβ · ηγ · ω−α

where t = film thickness, k = calibration constant, C = solution concentration (% solids), η = viscosity (cP), ω = angular velocity (rpm), and α ≈ 0.5, β ≈ 2, γ ≈ 0.33 for most resist systems.

The critical exponent is α ≈ 0.5, meaning film thickness scales as the inverse square root of spin speed. Practical implications:

Most photoresist manufacturers provide spin speed vs thickness curves for their products. These empirically-derived curves are more accurate than the theoretical model because they account for specific solvent systems, polymer molecular weight distributions, and evaporation kinetics unique to each formulation.

Film thickness vs spin speed curves for different photoresists showing inverse square root relationship

1.3 Viscosity and Spin Speed Relationship

Understanding how viscosity interacts with spin speed is essential for process design:

Viscosity Range (cP) Typical Resist Type Thickness Range Spin Speed Range Notes
1–10 Dilute PMMA, thin imaging layers 50 nm – 500 nm 2000–6000 rpm Sensitive to humidity; dynamic dispense recommended
10–100 Standard positive resists (AZ, SPR) 0.5–5 µm 1000–5000 rpm Most common regime; well-characterized spin curves available
100–1000 Thick resists (AZ 4620, SPR 220) 5–30 µm 500–3000 rpm Multi-step spin profiles often needed; edge bead significant
1000–50000+ SU-8, thick KMPR, polyimide 20–500+ µm 500–3000 rpm Slow ramp rates critical; extended soft bake required

1.4 Coating Uniformity

Film thickness uniformity across the wafer is characterized by the 3σ variation (three standard deviations of thickness measured at multiple points). Key factors affecting uniformity include:

State-of-the-art spin coaters achieve <0.5% thickness uniformity (3σ) on 200 mm wafers with proper environmental control and optimized dispense/spin recipes.

2) The Lithography Process Flow

Spin coating and development are two steps in the larger photolithography process. Understanding the complete flow is essential because parameters at each step interact — an apparently good coat can produce poor results if bake or development conditions are wrong.

Photolithography process flow: HMDS prime, spin coat, soft bake, expose, PEB, develop, hard bake, inspect

2.1 Substrate Preparation

Proper substrate preparation is the most frequently overlooked step in lithography, yet it has an outsized impact on pattern quality:

2.2 Spin Coating

The resist is dispensed and spun to the target thickness. A typical recipe consists of:

  1. Dispense step: 1–5 mL of resist dispensed at substrate center (static dispense at 0 rpm, or dynamic dispense at 500–800 rpm).
  2. Spread step: Ramp to 500–800 rpm at moderate acceleration (200–500 rpm/s) for 2–5 seconds to spread the resist across the substrate.
  3. Spin step: Ramp to final speed (1000–6000 rpm) at controlled acceleration (500–8000 rpm/s) and hold for 30–60 seconds.
  4. Deceleration: Controlled ramp-down to 0 rpm. Abrupt stops can cause resist reflow at the substrate edge.

For thick resists (>10 µm), multi-step profiles with intermediate hold speeds help avoid striations and improve uniformity. For very thin films (<200 nm), high spin speeds (4000–8000 rpm) with dynamic dispense minimize material waste.

2.3 Soft Bake (Pre-Bake)

The soft bake removes ~80–95% of the resist solvent, converting the liquid film to a solid. This step is essential for:

Hotplate baking is preferred over convection oven baking because heat transfers from the substrate up through the resist, allowing solvent to evaporate from the top surface without trapping bubbles. Typical conditions:

Resist Type Soft Bake Temp Soft Bake Time Notes
Standard positive (AZ 1500, SPR 700) 90–100°C 60–90 s Hotplate; overbaking reduces sensitivity
Thick positive (AZ 4620, SPR 220) 100–115°C 2–5 min Ramp slowly to avoid nitrogen bubbling; may require multi-step bake
SU-8 (negative epoxy) 65°C → 95°C ramp 3–30 min (thickness-dependent) Two-step ramp critical; fast cooling causes stress cracking
PMMA (e-beam resist) 170–180°C 60–90 s Higher bake temp than optical resists; improves resolution

2.4 Exposure

UV light (or electron beam) transfers the mask pattern into the resist by triggering a photochemical reaction. The exposure dose (mJ/cm²) must be calibrated for the specific resist thickness, bake conditions, and development process. Under-exposure causes incomplete pattern transfer; over-exposure causes feature widening (positive resist) or feature narrowing (negative resist).

2.5 Post-Exposure Bake (PEB)

PEB is critical for chemically amplified resists (CARs) and SU-8, where it drives the acid-catalyzed crosslinking or deprotection reaction initiated by exposure. For conventional DNQ/novolac resists, PEB is optional but can reduce standing wave effects by allowing diffusion of the photoactive compound. Typical PEB temperatures are 5–15°C above the soft bake temperature, with tight uniformity (±0.5°C across the wafer) required for CD control.

2.6 Development

The developer selectively dissolves either the exposed (positive resist) or unexposed (negative resist) regions. Development is covered in detail in Section 6.

2.7 Hard Bake (Post-Bake)

A final bake at 110–150°C crosslinks the remaining resist, improving its resistance to subsequent etch or implant steps. Hard bake also drives off residual solvent and developer, improving adhesion. For resists that will be used as permanent structures (e.g., SU-8 MEMS), hard bake at 150–200°C provides maximum mechanical and chemical stability.

3) Key Process Parameters

Achieving reproducible lithography results requires precise control of multiple interacting parameters. This section quantifies the impact of each variable and provides practical control targets.

3.1 Spin Speed

Spin speed is the primary thickness control parameter. For production-grade results, speed stability of ±1 rpm is essential — speed fluctuations directly translate to thickness non-uniformity.

Rule of thumb: A 1% change in spin speed produces approximately a 0.5% change in film thickness (from the t ∝ ω−0.5 relationship). At 3000 rpm, a ±30 rpm variation causes ~±0.5% thickness variation — acceptable for most processes. At ±1 rpm control, thickness stability from spin speed alone is <0.02%.

3.2 Acceleration

The spin-up acceleration profile has a surprisingly large effect on coating quality:

3.3 Dispense Volume and Method

Dispense volume affects material consumption, coating uniformity, and defect density:

Substrate Size Static Dispense Volume Dynamic Dispense Volume Savings with Dynamic
2” (50 mm) 1–2 mL 0.5–1 mL ~50%
4” (100 mm) 2–4 mL 1–2 mL ~50%
6” (150 mm) 3–6 mL 1.5–3 mL ~50%
8” (200 mm) 4–8 mL 2–4 mL ~50%
12” (300 mm) 5–10 mL 2–5 mL ~50%

Dynamic dispense (dispensing onto a slowly spinning substrate at 500–800 rpm) reduces waste by approximately 50% compared to static dispense (dispensing onto a stationary substrate). Dynamic dispense also reduces the probability of trapped air bubbles under the resist puddle.

3.4 Bake Temperatures

Bake temperature uniformity is as important as the absolute temperature. A hotplate with ±0.5°C uniformity across the substrate is the minimum requirement for reproducible lithography.

3.5 Humidity and Environment

Ambient conditions during coating and baking have a measurable impact on lithographic performance:

4) Photoresist Selection Guide

Selecting the right photoresist for your application requires balancing resolution, thickness range, process compatibility, and cost. The following table compares the most commonly used photoresists across research and production environments:

Resist Tone Chemistry Resolution Thickness Range Developer Key Applications
AZ 1500 series Positive DNQ/novolac 0.5–1 µm 0.5–2.5 µm AZ 300 MIF (2.38% TMAH) General photolithography, contact/proximity
AZ 4620 Positive DNQ/novolac 3–5 µm 6–25 µm AZ 400K (diluted KOH) MEMS, deep etch masks, electroplating molds
SPR 220 Positive DNQ/novolac 2–4 µm 3–30 µm MF-26A (2.38% TMAH) Thick resist lithography, MEMS, TSV
SU-8 Negative Epoxy (CAR) 1–2 µm 1–500+ µm SU-8 Developer (PGMEA) MEMS structures, microfluidics, permanent dielectric
PMMA Positive Chain scission <10 nm (e-beam) 50 nm – 2 µm MIBK:IPA (1:3) E-beam lithography, nanoimprint, lift-off
AZ nLOF 2000 series Negative DNQ-based image reversal 0.5–1 µm 1–15 µm AZ 300 MIF (2.38% TMAH) Metal lift-off (undercut profile)
ma-N 1400 series Negative Crosslinking 0.5–1 µm 0.5–3 µm ma-D 533/S (TMAH-based) High-resolution negative patterning, wet etch masks
AZ 5214E Pos/Neg (image reversal) DNQ/novolac + crosslinker 0.8–1.5 µm 0.5–2 µm AZ 300 MIF (2.38% TMAH) Metal lift-off, image reversal for undercut

Positive vs Negative Resists: When to Use Each

5) Development Process Optimization

Development is the pattern-defining step of lithography. The developer selectively dissolves resist to reveal the desired pattern. Poor development control is one of the most common causes of lithographic failure, yet it receives less attention than exposure optimization in many labs.

5.1 Development Methods

Method Description Advantages Disadvantages Best For
Puddle (static) Dispense developer to cover wafer surface; hold static for timed interval; spin rinse with DI water Low developer consumption; gentle (minimal pattern erosion); excellent uniformity with proper puddle formation Slower; dissolved resist products can redeposit if puddle is held too long Thin resists, high-resolution patterns, CD-critical layers
Spray Spray fresh developer onto spinning wafer through a nozzle Faster; continuous fresh developer supply; better at clearing high-aspect-ratio features Higher developer consumption; more aggressive (can erode thin features) Thick resists, high-aspect-ratio features, production throughput
Multi-puddle Sequential puddle–rinse–puddle cycles (typically 2–3 cycles) Combines gentleness of puddle with fresh developer supply; minimizes redeposition Longer cycle time; more complex recipe programming Thick resists where single puddle leaves residue; high-AR features
Immersion Submerge wafer in developer bath with agitation Simple; good for batch processing small substrates Developer becomes loaded with dissolved resist; poor uniformity; not recommended for production R&D with small pieces; SU-8 development (long soak times)

5.2 Development Time Optimization

The optimal development time is a balance between complete pattern clearing and minimizing resist erosion:

Practical tip: Use the “10% over-development” rule as a starting point. If your nominal development time is 60 seconds, develop for 66 seconds. This small margin ensures complete clearing without significant feature erosion. Verify the exact margin needed for your specific resist/developer system through a development time matrix (expose a test wafer, cleave into strips, develop each strip for a different time).

5.3 Developer Temperature Control

Development rate is strongly temperature-dependent. For TMAH-based developers, a 1°C increase in developer temperature increases the development rate by approximately 5–10%. This means:

5.4 Rinse and Dry

After development, a thorough DI water rinse (10–30 seconds at 300–500 rpm) quenches the development reaction and removes dissolved resist products. Follow with a high-speed spin dry (2000–5000 rpm for 15–30 seconds). For high-aspect-ratio features, a brief IPA rinse before spin dry reduces pattern collapse caused by surface tension (Marangoni effect).

6) Troubleshooting Common Issues

This section provides a systematic guide to diagnosing and resolving the most common spin coating and development defects. Each defect is traced to its root cause with specific corrective actions.

Defect Appearance Root Causes Corrective Actions
Comets Radial streaks originating from particles, usually visible as tails pointing outward from a defect source Particles on substrate or in resist; clogged dispense filter; contaminated resist bottle Improve substrate cleaning; install/replace 0.2 µm inline dispense filter; filter or replace resist; clean dispense lines
Striations Radial or spiral thickness variations visible as color fringes under sodium lamp or thin-film interference Excessive acceleration for resist viscosity; premature resist drying during spread; low humidity; resist dispensed too far from center Reduce acceleration; use dynamic dispense; increase ambient humidity (target 40–50% RH); ensure centered dispense; increase dispense volume
Edge bead Thick resist ridge (5–20× nominal thickness) at the substrate edge, extending 1–5 mm inward Surface tension causes resist to accumulate at the edge during spin-off; worse with thick, viscous resists Use edge bead removal (EBR): dispense solvent at wafer edge during a dedicated EBR spin step; increase final spin speed; program a brief high-speed burst at the end of the spin recipe
Poor adhesion / delamination Resist lifts or peels from substrate during development; visible as pattern distortion, missing features, or “flag” defects Inadequate substrate cleaning; missing HMDS; insufficient dehydration bake; over-development; developer temperature too high Add or improve HMDS vapor prime; increase dehydration bake time/temperature; verify substrate cleanliness; reduce development time or developer concentration
Incomplete development (scum) Thin resist residue remaining in areas that should be fully cleared; may be invisible to the eye but detectable by profilometer or etch test Under-exposure; insufficient development time; developer depletion (immersion method); soft bake temperature too low (excess solvent inhibits dissolution) Increase exposure dose; extend development time; use multi-puddle development; increase soft bake temperature; verify developer concentration/freshness
Bubbles / pinholes Circular voids in the resist film, visible as bright spots under microscope; may be isolated or clustered Nitrogen outgassing during soft bake (thick resists); air trapped during dispense; contaminated substrate surface; resist degassed improperly Use slower soft bake ramp (1–2°C/min for thick resists); ensure dynamic dispense; degas resist before use; improve substrate cleaning
Standing waves Periodic undulations in resist sidewall (visible as “scalloped” profile in SEM cross-section); period equals λ/(2n) where n is resist refractive index Constructive/destructive interference of incident and reflected exposure light within the resist film Apply bottom anti-reflective coating (BARC); use PEB to diffuse photoproducts and smooth the standing wave pattern; adjust resist thickness to a half-wave condition
Pattern collapse High-aspect-ratio resist features bend or collapse during rinse/dry, often sticking to adjacent features Capillary forces during drying exceed mechanical strength of resist features; aspect ratio (height:width) >3:1 at risk Use IPA rinse before spin dry; reduce feature aspect ratio; use supercritical CO₂ drying for extreme AR; increase hard bake to strengthen resist
Thickness non-uniformity Center-to-edge thickness gradient; visible as color variation across wafer Asymmetric exhaust; off-center dispense; non-uniform chuck vacuum; temperature/humidity gradient; substrate not flat on chuck Verify centered dispense; check chuck vacuum uniformity; ensure symmetric exhaust; control environment; verify substrate flatness

7) Equipment Selection Guide

The right coater/developer system can mean the difference between reproducible, high-quality lithography and constant process troubleshooting. When evaluating equipment, consider the following critical specifications:

7.1 Spin Performance

7.2 Hotplate Performance

7.3 Dispense System

7.4 Substrate Compatibility

Verify the system accommodates your substrate formats:

7.5 Environmental Control

For critical lithography, an environmental enclosure maintaining 23±0.5°C and 45±5% RH is essential. This option significantly improves run-to-run reproducibility and eliminates a major source of process variability — especially in facilities without centralized cleanroom HVAC control.

7.6 Automation and Programmability

7.7 Modular Configuration

A modular coater/developer platform allows you to configure the system for your specific workflow — coating-only, development-only, or integrated coat-bake-develop in a single footprint. This flexibility is valuable for research environments where process requirements evolve over time.

8) Frequently Asked Questions

What spin speed should I use for a 1 µm thick photoresist film?

The required spin speed depends on the specific resist and its viscosity. As a rough guide, standard positive resists (AZ 1512, SPR 700) typically achieve 1 µm at 3000–4000 rpm. However, always consult the manufacturer’s spin speed vs thickness curve for your specific resist product — these curves are empirically derived and account for solvent system, molecular weight, and formulation-specific behavior. Start with the manufacturer’s recommended speed and fine-tune with profilometer measurements on test wafers.

How do I choose between puddle and spray development?

Use puddle development for thin resists (<5 µm), critical CD layers, and processes where minimal feature erosion is important. Use spray development for thick resists (>10 µm), high-aspect-ratio features that need aggressive developer penetration, and production environments where throughput matters. For thick resists in the 5–15 µm range, a multi-puddle approach (2–3 puddle cycles with intermediate DI rinse) often gives the best balance of clearing and CD control.

Why is my resist film thickness not matching the manufacturer’s spin curve?

Several factors can cause deviation from published spin curves: (1) Temperature: the spin curve was likely generated at 23°C — if your lab is warmer, resist viscosity is lower and the film will be thinner. (2) Humidity: high humidity causes water absorption that changes solvent evaporation dynamics. (3) Resist age: solvent evaporation from partially filled bottles increases viscosity over time, yielding thicker films. (4) Acceleration: different ramp rates change the amount of resist ejected during spin-up. (5) Exhaust conditions: strong exhaust accelerates solvent evaporation, increasing effective viscosity. Always characterize the spin curve on your specific equipment under your cleanroom conditions.

How important is HMDS for photoresist adhesion?

HMDS is critical for resist adhesion on silicon, SiO₂, and silicon nitride surfaces. Without HMDS, water molecules on the hydrophilic substrate surface create a weak boundary layer that causes resist delamination during development — especially in aqueous (TMAH) developers. Vapor-phase HMDS priming at 150°C is more effective and reproducible than liquid HMDS applied by spin coating. In our experience, the majority of “poor adhesion” complaints from lithography engineers are resolved by adding or improving the HMDS priming step.

Can I coat non-round substrates (square pieces, irregular samples)?

Yes, but with caveats. Non-round substrates create asymmetric airflow during spinning, which can cause thickness non-uniformity — particularly at the corners of square substrates. To mitigate this: (1) use a carrier wafer with the sample mounted at center using vacuum or thermal release tape; (2) increase the dispense volume to ensure full coverage; (3) reduce spin speed (the corners of a square experience higher effective velocity than the edges of a round wafer at the same RPM); (4) use an environmental enclosure to minimize airflow-induced evaporation asymmetry. Modern spin coaters with programmable chuck configurations can accommodate substrates from small pieces to 12” wafers and square substrates with appropriate fixturing.

What causes “comet” defects and how do I eliminate them?

Comets are caused by particles on the substrate surface or in the resist that act as nucleation points for thickness discontinuities during spinning. The centrifugal force drags resist away from the particle, creating a radial streak. Elimination strategy: (1) improve substrate cleaning (add O₂ plasma descum or an additional solvent rinse); (2) install or replace the inline point-of-use resist filter (0.2 µm); (3) purge the dispense line with several mL of resist before coating (the first dispense from a line often contains particles); (4) check resist bottle age and storage conditions — expired or improperly stored resist develops particle precipitates; (5) ensure the spin bowl is clean and free of dried resist that could flake onto the substrate.

How do I handle edge bead for thick resist processes?

Edge bead is inherent to spin coating and becomes progressively worse with increasing resist viscosity and thickness. For thick resists (>10 µm), the edge bead can be 50–100+ µm tall and extend 3–5 mm from the wafer edge. Solutions: (1) Edge bead removal (EBR): program a dedicated EBR step where solvent (typically PGMEA or EBR solvent) is dispensed at the wafer edge while spinning at moderate speed (500–1000 rpm). The solvent dissolves the excess resist at the edge. (2) Backside rinse: dispense solvent on the wafer backside during spinning to remove resist that wraps around the edge. (3) Optimize spin profile: a brief high-speed burst (e.g., 5000–8000 rpm for 2–3 seconds) at the end of the spin recipe ejects additional edge material. EBR is the most effective and reproducible method for edge bead control.

Do I need environmental control for my spin coater?

It depends on your process requirements. For R&D with relaxed CD tolerances (±0.5 µm), a cleanroom with standard HVAC (±2°C, ±10% RH) may be sufficient. For production lithography or any process where run-to-run thickness repeatability of <1% (3σ) is required, environmental control (23±0.5°C, 45±5% RH) is essential. The cost of an environmental enclosure is small compared to the cost of diagnosing and reworking yield losses caused by uncontrolled ambient conditions. If you are seeing unexplained day-to-day or season-to-season variation in your lithographic results, environmental control should be one of the first process improvements to investigate.

NineScrolls Coater/Developer Systems

Our coater/developer platform offers up to 8000 rpm spin speed with ±1 rpm accuracy, integrated hotplate with precise temperature distribution, modular coat/develop/bake configuration, optional environmental control (23±0.5°C, 45±5% RH), and programmable edge bead removal — everything you need for reproducible photolithography from R&D to production.