How Surface Roughness and Figure Error Affect Optical Mirror Reflectance Quality

When an optical mirror underperforms, engineers usually blame the coating. But the substrate underneath often tells a different story. Surface roughness and figure error—two distinct manufacturing parameters—determine how much light a mirror reflects, how cleanly it does so, and whether it can handle demanding laser or imaging applications. Understanding the difference between them is the first step toward specifying mirrors that actually meet your system's requirements.

This article explains how surface roughness and figure error affect optical mirror reflectance quality, what tolerances you should specify, and how to measure and verify both parameters before committing to a production run. We'll cover the physics behind scattering and wavefront distortion, practical inspection methods, and common specification mistakes that cost manufacturers time and money.

Key Takeaways

• Surface roughness drives scattered light and reduces total reflectance, while figure error distorts the reflected wavefront without necessarily lowering total reflectivity.
• RMS roughness below 1 nm is typically required for high-power laser mirrors; 2–5 nm may suffice for imaging systems.
• Figure error is expressed in waves or fractions of a wavelength (λ/10, λ/4) and directly impacts focal spot quality and imaging resolution.
• Interferometry measures figure error; profilometry and atomic force microscopy (AFM) measure roughness—each requires different sample preparation and environmental control.
• Specifying both parameters on your drawing prevents costly rework and ensures the mirror performs as intended in its final optical system.

What You Need Before Starting

Before you specify or evaluate an optical mirror, gather the following information:

• Operating wavelength and power density — determines how tight your roughness and figure tolerances must be. A 1064 nm laser mirror demands different specs than a visible-light imaging mirror.
• Substrate material — fused silica, BK7, calcium fluoride (CaF₂), or silicon each polish differently and have distinct hardness and thermal properties.
• Measurement equipment access — a Zygo or similar interferometer for figure error, and a profilometer or AFM for roughness. If you lack these, plan to send samples to a qualified lab.
• Coating specifications — the coating adds its own roughness and stress, so substrate specs must be coordinated with the coating design.

For the components themselves, manufacturers like Shandong Yanggu Constant Crystal Optics produce a wide range of precision optics. Our main products include optical window,prism, lens, beamsp littler, filter, wedge, blanks and etc., all available with custom tolerances to match your application.

Step 1 — Define Surface Roughness and Its Effect on Reflectance

What to Do

Surface roughness refers to the microscopic irregularities left on a mirror's surface after polishing. These are high-spatial-frequency errors—typically measured in nanometers—that scatter incident light rather than reflecting it specularly.

To specify roughness correctly:

• Choose a measurement parameter. RMS (root mean square) roughness is the industry standard for optical surfaces. Ra (arithmetic average) is sometimes used in mechanical contexts but is less meaningful for optics.
• Define the spatial frequency band. Roughness measured over a 10 µm scan differs from that measured over 1 mm. State your measurement area and instrument.
• Set a realistic tolerance. For most laser mirrors, RMS roughness of 0.5–1 nm is achievable with modern polishing. For imaging optics, 2–5 nm may be perfectly acceptable.
• Verify with the right tool. AFM resolves sub-nanometer features; optical profilometers work well for 1–10 nm ranges; stylus profilometers are faster but less sensitive.

Why This Matters

Scattering follows the Rayleigh–Rice relation: total integrated scatter (TIS) is approximately proportional to (4πσ/λ)², where σ is RMS roughness and λ is wavelength. At 632.8 nm, a 1 nm RMS surface scatters roughly 0.04% of incident light. At 355 nm, the same surface scatters about 0.12%. That may sound small, but in a multi-mirror laser cavity, losses compound quickly.

For high-reflectance mirrors (R > 99.9%), scattering is often the dominant loss mechanism after coating absorption. A mirror with 2 nm RMS roughness can lose 0.15–0.5% to scatter at visible wavelengths—enough to degrade laser efficiency or reduce signal-to-noise in sensitive detection systems.

Common Mistakes to Avoid

• Confusing Ra and RMS: Ra is the arithmetic mean deviation; RMS weights larger deviations more heavily. For the same surface, RMS is typically 1.1–1.3× Ra. Mixing them up can lead to under-specifying.
• Ignoring spatial frequency: A surface can have low RMS roughness but high mid-spatial-frequency error that still scatters light into small angles. Specify the measurement bandwidth.
• Measuring before coating: Coatings add their own roughness. Measure the finished mirror, not just the polished substrate.

Step 2 — Understand Figure Error and Wavefront Distortion

What to Do

Figure error describes low-spatial-frequency deviations from the ideal surface shape—the "macro" flatness or curvature error. It's measured in waves (λ) or fractions of a wave, such as λ/10 at 632.8 nm.

To specify figure error:

• Define the aperture. Figure error is meaningless without specifying the clear aperture over which it applies (e.g., λ/10 over the central 90% of the diameter).
• Choose the metric. Peak-to-valley (PV) is the most common, but RMS wavefront error better predicts system performance. A surface can have high PV but low RMS if the error is localized.
• Consider the test wavelength. A λ/10 specification at 632.8 nm corresponds to 63.3 nm of surface deviation. At 1064 nm, the same physical surface is only λ/17.
• Account for mounting. Clamping a mirror can distort its figure by several waves. Specify the mounting condition or allow for compensation.

Why This Matters

Figure error does not reduce total reflectance—nearly all the light still reflects—but it distorts the wavefront. In a laser cavity, this increases beam divergence and reduces focusability. In an imaging system, it blurs the image and reduces resolution.

A mirror with λ/4 figure error at 632.8 nm introduces up to 158 nm of wavefront distortion. For a diffraction-limited system requiring λ/10 total wavefront error, that single mirror consumes more than half the budget. This is why precision systems specify λ/10 or better for critical mirrors.

Common Mistakes to Avoid

• Specifying figure without aperture: "λ/10 flatness" is meaningless without defining the clear aperture. Always state both.
• Ignoring gravity and mounting: A mirror that measures λ/20 in a test mount can deform to λ/2 when bolted into your system. Test under representative conditions.
• Forgetting the coating's stress: Dielectric coatings can bend a thin substrate by several waves. For thin mirrors, specify the figure after coating.

Step 3 — Measure Both Parameters Correctly

What to Do

Measurement technique matters as much as the specification itself.

• Figure error: Use a Fizeau or Twyman-Green interferometer. The mirror is compared against a reference flat or sphere, and the resulting fringe pattern is analyzed to produce a surface map. Modern phase-shifting interferometers achieve repeatability of λ/100 or better.
• Roughness: Use a white-light interferometer, AFM, or stylus profilometer. AFM provides the highest resolution (down to 0.01 nm) but scans small areas (typically 1–100 µm). White-light interferometry covers larger areas (up to several mm) with resolution around 0.1 nm.
• Cross-check with scatterometry: For production QC, measuring total integrated scatter (TIS) with an integrating sphere can quickly flag roughness problems without full surface mapping.

Why This Matters

Both parameters interact with coating performance. A substrate with 0.5 nm RMS roughness can yield a coated mirror with 1–2 nm RMS roughness, depending on coating material and deposition method. Ion-assisted deposition (IAD) typically produces smoother, denser coatings than traditional thermal evaporation.

For demanding applications like Laser Optical System s, the combination of substrate quality and coating technology determines whether a mirror survives high-power operation or fails from localized absorption and thermal distortion.

Common Mistakes to Avoid

• Using the wrong reference: A poor reference flat can make a good mirror look bad—or vice versa. Calibrate your reference against a certified standard.
• Environmental noise: Air currents, vibration, and temperature drift corrupt interferometric measurements. Use vibration isolation and thermal stabilization.
• Sampling too small an area: A 10 µm AFM scan may miss mid-spatial-frequency errors. Combine AFM with a larger-area optical measurement.

Step 4 — Set Realistic Tolerances for Your Application

What to Do

Match your specifications to your system's actual requirements.

Application	RMS Roughness	Figure Error	Notes
High-power laser mirrors (1064 nm, >100 W)	0.5–1 nm	λ/10	Scatter and absorption are critical
UV laser optics (355 nm, 266 nm)	0.3–0.5 nm	λ/10 or better	Scatter scales with 1/λ²
Visible imaging systems	1–3 nm	λ/4–λ/8	Balance cost and performance
Infrared optics (3–5 µm, 8–12 µm)	5–20 nm	λ/4–λ/2	Longer wavelength relaxes roughness
Medical diagnostic instruments	1–2 nm	λ/8–λ/10	Repeatability and consistency matter

Why This Matters

Over-specifying drives cost up unnecessarily. A λ/20 mirror with 0.3 nm roughness might cost three times more than a λ/4 mirror with 2 nm roughness—yet deliver no measurable benefit in a system with other limiting factors. Under-specifying, meanwhile, leads to failed performance validation and expensive rework.

In Medical Instruments , for example, mirrors in diagnostic imaging systems need consistent, repeatable performance across production batches. Tight but achievable tolerances—typically λ/8 figure and 1–2 nm roughness—ensure every unit meets the same standard.

Common Mistakes to Avoid

• Copying specs from a datasheet: Your application is unique. Derive tolerances from your system's error budget.
• Ignoring environmental conditions: A mirror that works at 20°C may fail at 60°C if substrate and coating have mismatched thermal expansion.
• Not budgeting for coating effects: The coating can change both roughness and figure. Specify the final coated performance, not just the substrate.

Step 5 — Verify Performance in Your System

What to Do

Final validation should happen under conditions that match real use.

• Measure reflectance: Use a spectrophotometer for broadband reflectance or a laser-based setup for single-wavelength measurements. For high reflectors (>99.9%), cavity ring-down or similar techniques provide the needed sensitivity.
• Check wavefront in situ: If possible, measure the mirror's effect on the system wavefront, not just the isolated surface.
• Test at operating power: Low-power testing doesn't reveal thermal distortion or coating damage that appears only at full power.
• Document everything: Record measurement conditions, equipment, and results. This data becomes your quality baseline for future batches.

Why This Matters

A mirror can pass every bench test and still fail in the system. Thermal loading, mounting stress, and contamination all degrade performance. Verifying under realistic conditions catches these issues before they become field failures.

Common Mistakes to Avoid

• Testing at the wrong wavelength: A mirror optimized for 1064 nm may perform poorly at 532 nm if the coating design is narrowband.
• Ignoring polarization effects: Some coatings have different reflectance for S and P polarization. Specify and test the polarization state you'll actually use.
• Skipping environmental testing: Humidity, temperature cycling, and UV exposure can degrade coatings over time. Accelerated life testing is worth the investment for critical applications.

Pro Tips for Success

• Build a relationship with your supplier early: Share your system requirements, not just your drawing. A good manufacturer will flag unrealistic tolerances and suggest cost-effective alternatives.
• Request a sample before committing to volume production: Most reputable manufacturers, including SYCCO, can send samples for testing in a short time. Verify performance before scaling up.
• Keep a measurement log: Track roughness and figure data across batches. Trends can reveal polishing or coating process drift before it causes failures.
• Consider environmental factors in your specification: If your mirror operates at elevated temperature or in vacuum, specify tolerances at those conditions, not just at 20°C.
• Budget for both substrate and coating quality: The best substrate is useless with a poor coating, and vice versa. Specify the complete mirror, not just the glass.

Frequently Asked Questions

What is the difference between surface roughness and figure error?

Surface roughness describes microscopic, high-spatial-frequency irregularities (nanometer scale) that scatter light and reduce total reflectance. Figure error describes macroscopic, low-spatial-frequency deviations from the ideal shape (wavelength scale) that distort the wavefront without necessarily reducing total reflectance. Both affect optical mirror reflectance quality, but through different physical mechanisms.

What roughness is acceptable for a laser mirror?

For high-power laser mirrors, RMS roughness of 0.5–1 nm is typical. At 1064 nm, 1 nm RMS roughness scatters roughly 0.04% of incident light; at 355 nm, the same surface scatters about 0.12%. For lower-power systems, 2–5 nm may be acceptable. Always specify the measurement bandwidth and instrument.

How is figure error measured?

Figure error is typically measured with a Fizeau or Twyman-Green interferometer. The mirror is compared against a reference surface, and the interference pattern is analyzed to produce a surface map. Phase-shifting interferometers achieve repeatability of λ/100 or better. Results are expressed as PV or RMS wavefront error over a specified clear aperture.

Can a coating improve surface roughness?

A well-designed coating can slightly smooth a rough substrate, but it more often adds roughness. Ion-assisted deposition (IAD) produces denser, smoother coatings than traditional thermal evaporation. Always measure the final coated surface, not just the polished substrate, to know the true performance.

What figure error do I need for a diffraction-limited system?

For a diffraction-limited system, total wavefront error should be λ/10 or better. If a mirror contributes more than a fraction of that budget, it will limit system performance. For most applications, λ/10 figure error per mirror is a reasonable starting point, with tighter tolerances for critical elements.

Conclusion

Surface roughness and figure error affect optical mirror reflectance quality in fundamentally different ways—roughness scatters light and reduces total reflectance, while figure error distorts the wavefront and degrades image quality or beam focusability. Both parameters must be specified, measured, and verified to ensure your mirror performs as intended.

The practical approach is straightforward: define your system's error budget, derive realistic tolerances for both parameters, measure them with appropriate instruments, and validate performance under real operating conditions. Work with a manufacturer who understands these trade-offs and can deliver components that meet your specifications consistently.

Start by reviewing your current mirror specifications. Are both roughness and figure error defined with clear apertures and measurement conditions? If not, that's your first improvement. Then request samples from your supplier and verify performance before committing to volume production. The time invested in proper specification pays off in fewer field failures and better system performance.