Concepts•Jun 2026•3 min read

Direct Write Lithography vs Focused Ion Beam

Two ways to pattern at the nanoscale without a mask set. One writes with a photon or electron beam to define structures; the other machines matter atom-by-atom with a beam of gallium ions. They're not interchangeable, and treating them as rivals is half the confusion. Here's the decisive read on when each one earns its place.

The short answer

Direct Write Lithography over Focused Ion Beam for most cases. For defining patterns — making devices, prototyping circuits, R&D layouts — Direct Write Lithography wins because it patterns resist additively and cleanly,.

  • Pick Direct Write Lithography if defining device patterns, prototyping without a photomask, doing maskless R&D lithography, or need clean resist patterning across a wafer at reasonable throughput
  • Pick Focused Ion Beam if editing an existing chip, milling cross-sections for TEM, doing circuit edit/repair, or need true 3D site-specific material removal and deposition
  • Also consider: Resolution targets, whether you're building vs. editing, gallium contamination tolerance, sample size, and how much throughput you can sacrifice.

— Nice Pick, opinionated tool recommendations

What they actually are

Direct Write Lithography is maskless patterning: a focused beam — laser in DWL/laser writers, or electrons in EBL — scans across a resist-coated substrate, exposing it pixel by pixel so a develop step leaves your pattern behind. No photomask, no stepper. You change the design in CAD, you run it again. Focused Ion Beam is a different beast entirely: a beam of (usually gallium) ions physically sputters atoms off the surface, milling material away, or, with a precursor gas, depositing it. FIB doesn't pattern resist — it machines the actual sample. That distinction is the whole ballgame. One defines a template for fabrication; the other directly subtracts or adds matter. People pit them against each other because both are slow, serial, beam-driven, and live in the same cleanroom. The physics, and the purpose, could not be further apart.

Resolution and quality

Electron-beam direct write is the resolution king here — sub-10 nm features are routine, and that's why EBL still makes the photomasks everyone else copies. Laser direct write is coarser, typically a few hundred nanometers to a micron, but fast and forgiving. FIB resolution sits in between, roughly 5–50 nm depending on the column, but it carries a tax direct write doesn't: gallium implantation. Every FIB cut sprays Ga ions into your sidewalls and a surface amorphization layer, wrecking electrical and crystallographic properties near the edit. For failure analysis nobody cares; for building a working device you'll hate it. Direct write into resist leaves the substrate chemically untouched until you etch. So if pristine material and the smallest clean features matter, EBL wins outright. If you need to cut a trench in finished silicon and see what's inside, FIB's destructiveness is the feature, not the bug.

Throughput and cost

Both are serial — they write one spot at a time, which is why neither threatens optical lithography for volume production. But the economics differ. Laser direct write systems are comparatively cheap and quick, good for prototyping mask-free runs and microfluidics overnight. EBL is slower and pricier but earns it on resolution. FIB is the most expensive per useful operation: a dual-beam FIB-SEM is a six-figure-plus instrument, the gallium source needs babying, and each cross-section is painstaking manual work measured in hours per site. Nobody runs FIB to make a thousand of anything. It's a per-sample, hand-tuned procedure. Direct write at least dreams of small-batch device fabrication. FIB's value is concentrated entirely in single high-stakes edits — a $50,000 chip you need to rewire — where its cost is trivial against the alternative of a respin. Match the tool to the batch size or you'll bleed money.

The honest tradeoff

Stop framing these as competitors and the answer gets obvious. Direct Write Lithography is for creating patterns — additive intent, clean substrate, the front of the workflow. Focused Ion Beam is for modifying and inspecting what already exists — subtractive, destructive, site-specific surgery at the back end. A real fab uses both: EBL writes the mask or the prototype, FIB later cross-sections a defect or edits a prototype trace. If a vendor or a grad student tells you to pick FIB to 'fabricate' a device array, they're abusing a repair tool because it was the instrument on hand — and they'll wonder later why their transistors are leaky with gallium. Use direct write to build. Use FIB to fix, fail-analyze, and slice. The only people genuinely choosing between them are choosing because they only own one machine, and that's a budget problem, not a technical verdict.

Quick Comparison

FactorDirect Write LithographyFocused Ion Beam
Primary purposeDefine/create patterns (additive workflow)Edit, mill, cross-section existing samples
Best resolutionSub-10 nm with EBL~5–50 nm, plus gallium damage
Substrate contaminationResist patterning leaves substrate cleanImplants Ga ions, amorphizes surface
Throughput / batch scaleSerial but viable for small batches/prototypesHours per site, one-off only
3D site-specific machining2D pattern definition onlyTrue 3D mill and deposit at a point

The Verdict

Use Direct Write Lithography if: You're defining device patterns, prototyping without a photomask, doing maskless R&D lithography, or need clean resist patterning across a wafer at reasonable throughput.

Use Focused Ion Beam if: You're editing an existing chip, milling cross-sections for TEM, doing circuit edit/repair, or need true 3D site-specific material removal and deposition.

Consider: Resolution targets, whether you're building vs. editing, gallium contamination tolerance, sample size, and how much throughput you can sacrifice.

Direct Write Lithography vs Focused Ion Beam: FAQ

Is Direct Write Lithography or Focused Ion Beam better?

Direct Write Lithography is the Nice Pick. For defining patterns — making devices, prototyping circuits, R&D layouts — Direct Write Lithography wins because it patterns resist additively and cleanly, scales to wafers, and doesn't poison your silicon with embedded gallium. FIB is a surgeon's scalpel for editing and cross-sectioning, not a tool for building. If your goal is fabrication rather than failure analysis or chip repair, you start with direct write.

When should you use Direct Write Lithography?

You're defining device patterns, prototyping without a photomask, doing maskless R&D lithography, or need clean resist patterning across a wafer at reasonable throughput.

When should you use Focused Ion Beam?

You're editing an existing chip, milling cross-sections for TEM, doing circuit edit/repair, or need true 3D site-specific material removal and deposition.

What's the main difference between Direct Write Lithography and Focused Ion Beam?

Two ways to pattern at the nanoscale without a mask set. One writes with a photon or electron beam to define structures; the other machines matter atom-by-atom with a beam of gallium ions. They're not interchangeable, and treating them as rivals is half the confusion. Here's the decisive read on when each one earns its place.

How do Direct Write Lithography and Focused Ion Beam compare on primary purpose?

Direct Write Lithography: Define/create patterns (additive workflow). Focused Ion Beam: Edit, mill, cross-section existing samples.

Are there alternatives to consider beyond Direct Write Lithography and Focused Ion Beam?

Resolution targets, whether you're building vs. editing, gallium contamination tolerance, sample size, and how much throughput you can sacrifice.

🧊
The Bottom Line
Direct Write Lithography wins

For defining patterns — making devices, prototyping circuits, R&D layouts — Direct Write Lithography wins because it patterns resist additively and cleanly, scales to wafers, and doesn't poison your silicon with embedded gallium. FIB is a surgeon's scalpel for editing and cross-sectioning, not a tool for building. If your goal is fabrication rather than failure analysis or chip repair, you start with direct write.

Related Comparisons

Disagree? nice@nicepick.dev