I. The Drawing That Makes Most Shops Say “No”
You’ve sent out your RFQ to five different machine shops. Four of them came back with the same answer:
“We’re sorry, but Ra 0.01μm is beyond our current capabilities.”
The fifth one said yes—but three weeks later, delivered parts with a surface finish of Ra 0.08μm and called it “close enough.”
Sound familiar?
If you work in semiconductor equipment manufacturing, this scenario is frustratingly common. Ra 0.01μm surface roughness sits at the absolute frontier of precision machining. It’s not a spec that most shops can hit. And more importantly, it’s not a spec you can compromise on—because in semiconductor applications, surface quality isn’t just an aesthetic requirement. It directly determines whether your component performs, or fails.
So what separates the shops that can consistently deliver Ra 0.01μm from the ones that can’t? That’s exactly what we’re going to break down in this article.
II. First, Let’s Put Ra 0.01μm Into Perspective
Before we dive into the “how,” let’s make sure we’re aligned on the “what.”
Ra stands for Roughness Average—the arithmetic mean of the absolute values of surface profile deviations measured from a mean line. In simpler terms: it’s the average height of the “hills and valleys” on a machined surface.
Here’s a quick reference to understand where Ra 0.01μm sits in the machining world:
To put Ra 0.01μm in physical terms:
“If a standard human hair is approximately 75 micrometers in diameter, Ra 0.01μm represents a surface variation of just 1/7,500th of that hair. We’re talking about surface features that are invisible not just to the naked eye—but to most optical microscopes.”
This is the level of smoothness required for semiconductor wafer chucks, vacuum stages, photolithography components, and precision optical mounts—parts where even the slightest surface irregularity can scatter light, trap particles, or compromise vacuum sealing performance.
III. Why Standard CNC Machining Cannot Get You There
This is the part most shops won’t tell you. Let’s be brutally honest about why conventional CNC processes fundamentally cannot achieve Ra 0.01μm—no matter how good the programmer is, or how new the machine is.
Problem #1: Spindle Runout Is Already Too Large
A standard high-quality CNC machining center has a spindle runout of approximately 3 to 10 micrometers. That means the cutting tool itself is already wobbling by 3-10μm during every single rotation.
When your target surface roughness is Ra 0.01μm—which equals 0.01 micrometers—the spindle runout alone is already 300 to 1,000 times larger than your target finish.
You cannot cut more precisely than your machine can hold position. It’s a physical impossibility.
Problem #2: Every Cutting Tool Leaves a Signature
Every machining process—milling, turning, grinding—leaves behind a characteristic surface texture. These are called feed marks, and they are the geometric imprint of the tool path, feed rate, and cutting depth.
Even with fine grinding using a 320-grit wheel, the resulting surface finish typically sits between Ra 0.2μm and Ra 0.4μm. The individual abrasive grains in the wheel are simply too large and too irregular to produce a surface below Ra 0.1μm without subsequent processing steps.
Think of it this way: you can sand a piece of wood with 600-grit sandpaper all day, but you’ll never achieve the surface of polished glass. You need to change the process entirely—not just refine it.
Problem #3: Thermal Drift Destroys Dimensional Consistency
A typical machine shop operates at an ambient temperature of somewhere between 18°C and 28°C, with daily fluctuations of ±3 to ±5°C.
For steel with a thermal expansion coefficient of approximately 12×10−6 /°C, a temperature shift of just 1°C causes a 100mm part to expand or contract by:
When you are chasing Ra 0.01μm surface roughness on a part with dimensional tolerances of ±0.1μm, a 1.2μm thermal shift doesn’t just cause problems—it invalidates your entire measurement baseline. You could be polishing a surface that has already moved beyond your tolerance window, and not even know it.
Problem #4: Vibration—The Invisible Enemy
Standard machine shops are rarely isolated from external vibration. Forklifts moving nearby, HVAC systems cycling on and off, neighboring machines running simultaneously—all of these generate micro-vibrations that transmit through the floor and into the workpiece during ultra-fine polishing.
At the Ra 0.01μm level, a vibration of just 0.5μm amplitude will leave visible surface artifacts. This is why ultra-precision machining facilities invest heavily in vibration-damped foundations and isolated machine plinths.

IV. The Modo Rapid Approach: A Complete Precision Ecosystem
Achieving Ra 0.01μm consistently isn’t about one machine or one technique. It requires building what we call a complete precision ecosystem—where every single variable, from the spindle to the shop temperature to the cleaning chemistry, is under strict control.
Here is exactly how we do it.
Step 1: Material Selection and Pre-Treatment
Before a single cut is made, the choice of workpiece material and its preparation state are critical.
For semiconductor components requiring Ra 0.01μm, we predominantly work with:
- Invar (FeNi36): With a thermal expansion coefficient of just 1.2×10−6 /°C—ten times lower than standard steel—Invar is the material of choice for components where dimensional stability across temperature ranges is non-negotiable.
- Silicon Carbide (SiC): Exceptional hardness (Mohs 9.5), chemical inertness, and near-zero porosity make SiC ideal for wafer processing components. However, its brittleness demands specialized grinding strategies.
- Reaction-Bonded Silicon Carbide (RB-SiC): Offers improved machinability compared to sintered SiC while retaining most of its thermal and mechanical properties.
- Precision Aluminum Alloys (7075-T6): For components where weight is critical, properly heat-treated 7075-T6 provides sufficient hardness and stability for Ra 0.01μm finishing.
Pre-treatment is not optional. Before any precision grinding begins, all workpieces undergo:
- Stress relief annealing — to release internal residual stresses from prior machining operations that would otherwise cause the part to “spring back” after fine material removal.
- Dimensional pre-inspection — full CMM scan to establish a baseline before entering the precision workflow.
- Surface cleaning — ultrasonic degreasing to remove all cutting fluids, oils, and contaminants from prior operations.
Skipping pre-treatment is one of the most common reasons why parts fail inspection after polishing. The stress was always there—you just couldn’t see it until you removed enough material.
Step 2: Ultra-Precision Grinding
This is where the Ra journey truly begins.
Modo Rapid’s precision grinding operations are performed on dedicated surface grinding machines with hydrostatic spindle bearings, achieving spindle runout of < 0.5μm—compared to the 3-10μm of standard CNC centers.
Our grinding sequence for Ra 0.01μm targets:
Stage A: Coarse Grinding (Ra target: 0.4-0.8μm)
- Wheel specification: Vitrified CBN, 80-120 grit
- Coolant: Water-soluble synthetic, temperature-controlled to ±1°C
- Purpose: Remove bulk material, establish geometric form
Stage B: Fine Grinding (Ra target: 0.1-0.2μm)
- Wheel specification: Resin-bond diamond, 400-600 grit
- Feed rate: Reduced to 30% of Stage A
- Purpose: Remove grinding marks from Stage A, approach near-net form
Stage C: Precision Grinding (Ra target: 0.05-0.08μm)
- Wheel specification: Resin-bond diamond, 1200-2000 grit
- Spark-out passes: Minimum 5 passes with zero infeed
- Purpose: Achieve consistent surface texture ready for lapping
Key principle: Each stage must fully remove the surface damage layer created by the previous stage. Rushing between stages—a common shortcut in lower-tier shops—means carrying forward deeper sub-surface damage that will re-emerge during polishing as pitting or scratching.

Step 3: Precision Lapping
Lapping is the bridge between grinding and polishing. It’s also the step that most shops either skip entirely or execute incorrectly.
What is lapping? Unlike grinding (where the abrasive is fixed to a wheel), lapping uses free abrasive particles suspended in a fluid, applied between the workpiece and a flat lapping plate. The random motion of abrasive particles produces an isotropic surface texture—meaning the surface has no directional lay, which is critical for vacuum sealing and optical applications.
Modo Rapid’s lapping sequence:
- Lapping Compound 1: Silicon Carbide, 9μm particle size — Ra target: 0.03-0.05μm
- Lapping Compound 2: Alumina (Al₂O₃), 3μm particle size — Ra target: 0.015-0.025μm
- Lapping Compound 3: Diamond suspension, 1μm particle size — Ra target: 0.010-0.015μm
Critical process parameters:
- Lapping pressure: 20-50 g/cm² (excessive pressure causes sub-surface cracking in brittle materials)
- Lapping plate material: Cast iron for metals; ceramic-coated for SiC and ceramics
- Plate flatness: Must be maintained to within 0.5μm across the working surface
- Compound change protocol: Full cleaning between compound grades to prevent cross-contamination of abrasive sizes
Step 4: Ultra-Fine Polishing
If lapping is the bridge, polishing is the destination.
At this stage, we are no longer “cutting” the surface in any conventional sense. We are performing chemo-mechanical planarization (CMP)—a process where both chemical and mechanical action combine to achieve atomic-level material removal.
Modo Rapid’s polishing process:
- Polishing pad: Polyurethane foam pad, Shore A hardness 60-70
- Polishing slurry: Colloidal silica (SiO₂), 20-50nm particle size, pH 9.5-10.5
- Polishing pressure: 5-15 g/cm² (significantly lower than lapping)
- Relative velocity: 20-40 m/min
- Duration: 30-90 minutes depending on material and starting roughness
Why colloidal silica? The slightly alkaline pH of the slurry creates a thin, soft hydrated layer on the workpiece surface. The nano-scale SiO₂ particles then remove this softened layer preferentially from the “peaks” of the surface profile, gradually leveling the surface to Ra 0.01μm and below.
Final polishing verification: After polishing, surface roughness is measured using a white light interferometer (WLI) capable of resolving features as small as 0.1nm. This is not optional—it is part of our standard process, not an extra service.
Step 5: Controlled Environment — The Factor Nobody Talks About Enough
Everything described above happens inside Modo Rapid’s Class 1000 cleanroom, maintained under the following conditions:
- Temperature: 20°C±0.1°C, 24 hours per day, 7 days per week
- Humidity: 45%±5% RH
- Particle count: < 1,000 particles ≥ 0.5μm per cubic foot
- Vibration isolation: Machine plinths isolated from building structure via pneumatic anti-vibration mounts
Why does this matter so much? Let’s go back to our thermal expansion calculation. For an Invar component with α=1.2×10−6 /°C:
Even with Invar—one of the most dimensionally stable materials available—a temperature variation of just 0.1°C causes a 100mm part to move by 0.012μm. That’s already 12% of our Ra 0.01μm target. For standard steel, the same 0.1°C shift causes a movement of 0.12μm—twelve times our target roughness.
This is why temperature control isn’t a luxury in ultra-precision machining. It is a prerequisite.
Step 6: Precision Cleaning and Final Inspection
The final step before delivery is often underestimated—but it can make or break a part that has been flawlessly machined.
Semiconductor components require ultra-clean surfaces, free from:
- Organic contamination (oils, fingerprints, polishing residues)
- Ionic contamination (sodium, potassium, chloride ions)
- Particulate contamination (any particle > 0.5μm)
Modo Rapid’s cleaning protocol:
- Ultrasonic degreasing in alkaline detergent solution (40kHz, 60°C, 10 minutes)
- DI water rinse (resistivity > 18 MΩ·cm)
- Ultrasonic cleaning in isopropyl alcohol (IPA) bath
- Final DI water rinse with spin-dry
- Cleanroom packaging in double-sealed, nitrogen-purged anti-static bags
After cleaning, 100% of parts undergo final inspection:
- Surface roughness: White light interferometry (WLI)
- Flatness and form: CMM with 0.1μm resolution probe
- Visual inspection: Under 100x magnification in oblique lighting
- Cleanliness verification: Particle count per unit area
Every part ships with a full inspection report—not a summary, not a certificate of conformance, but actual measurement data that you can verify against your drawing requirements.
V. Case Study: From “Impossible” to Delivered in 28 Days
The Challenge:
A semiconductor equipment manufacturer based in South Korea approached Modo Rapid in 2023 with a vacuum chuck component for a next-generation wafer inspection system. The drawing specified:
- Material: Reaction-Bonded Silicon Carbide (RB-SiC)
- Surface roughness: Ra ≤ 0.01μm (sealing surface)
- Flatness: ≤ 0.08μm over 150mm diameter
- Parallelism: ≤ 0.2μm
- Quantity: 10 pieces (NPI batch)
The component had already been rejected by two suppliers in Korea and one in Germany. The feedback was consistent: “The flatness and roughness combination on SiC is beyond our capability.”
The Modo Rapid Approach:
Our engineering team conducted a 3-day technical feasibility review before accepting the order. Key decisions made during this review:
- Workholding strategy: Designed custom vacuum fixture to hold the SiC blank without introducing clamping stress.
- Grinding wheel selection: Selected a fine-grade resin-bond diamond wheel specifically formulated for RB-SiC to minimize sub-surface fracture.
- Lapping plate conditioning: Reconditioned our cast iron lapping plate to within 0.3μm flatness before beginning the job.
- CMP slurry optimization: Adjusted colloidal silica concentration and pH specifically for SiC’s surface chemistry.
The Results:
The customer has since placed three follow-on orders, with volumes increasing to 50 pieces per batch.
VI. Common Misconceptions—Addressed Directly
“Ra 0.01μm means the surface is perfectly flat, right?”
Not exactly. Ra measures surface roughness—the micro-texture of the surface. Flatness is a separate geometric parameter that measures how much a surface deviates from a perfect plane. A surface can have Ra 0.01μm roughness but still have poor flatness if the part has bowed or warped. At Modo Rapid, we control both simultaneously, because semiconductor applications demand both.
“Can’t you just polish any surface to Ra 0.01μm?”
No. Polishing cannot fix geometric errors—it can only refine surface texture. If your part has tool marks 5μm deep from a prior grinding operation, polishing will reduce the roughness number, but those underlying features will remain as sub-surface stress concentrations. The correct approach is to remove each damage layer completely at each process stage before moving to the next.
“This level of precision must cost ten times more than standard machining.”
The cost premium for Ra 0.01μm work is real, typically 3 to 6 times the cost of standard precision grinding. However, the relevant comparison is not the machining cost in isolation—it’s the total cost of quality. A semiconductor tool that fails in the field due to surface contamination or seal leakage carries costs measured in production downtime, yield loss, and warranty claims that dwarf any machining cost difference.
“Only European or Japanese shops can do this.”
This was perhaps true fifteen years ago. Today, with the right investment in equipment, process development, and engineering talent, Chinese precision machining facilities are fully capable of delivering world-class surface finish quality. Modo Rapid has been delivering Ra 0.01μm components to customers in Taiwan, South Korea, Europe, and North America since 2018.
VII. Is Your Current Supplier Truly Capable? Ask These 5 Questions
Before your next RFQ, ask any potential supplier these five questions. Their answers will tell you everything:
- “What is the spindle runout specification on your precision grinding machine?” (Acceptable answer: < 1μm. Red flag: “I’m not sure” or “it’s very good.”)
- “What temperature control do you maintain in your precision machining area?” (Acceptable answer: ±0.5°C or better. Red flag: “We have air conditioning.”)
- “Can you provide a white light interferometry report for the surface roughness measurement?” (Acceptable answer: Yes, it’s standard. Red flag: “We use a contact profilometer” or “we can arrange it.”)
- “What is your typical yield rate for Ra 0.01μm components?” (Acceptable answer: > 90%. Red flag: Any hesitation, or “it depends.”)
- “Have you previously machined this material to Ra 0.01μm?” (Acceptable answer: Yes, with data. Red flag: “We’re confident we can figure it out.”)