A client at a semiconductor plant (FOUP handler, EUV lithography prep area) calls on Friday evening: "Wafer yield dropped from 87% to 31% in 48 hours, the contamination scan shows an unidentified organic residue on last week's lot." After 9 days of investigation the cause is found: a technician replaced an M6 bolt on the vacuum chamber flange and used an ordinary stainless A4-70 instead of a silver-plated vacuum-baked variant. Outgassing from that one bolt contaminated 14 lots worth €2.3M.
This isn't a sci-fi story, it's a recurring scenario in semiconductor, optics and vacuum-technology production. This article is a primer on why "a bolt isn't just a bolt" in UHV and cleanroom environments, and where contamination budget is most often underestimated.
Three worlds of vacuum and three worlds of cleanroom
Vacuum levels
- **Low vacuum** — 10⁵ to 10² Pa (~100 mbar to 1 mbar). Vacuum packs, packaging. No special materials.
- **Medium vacuum** — 10² to 10⁻¹ Pa (~1 mbar to 10⁻³ mbar). Coating, vacuum brazing. Standard SS304, standard gaskets (Viton, EPDM).
- **High vacuum (HV)** — 10⁻¹ to 10⁻⁵ Pa. PVD coating, scanning electron microscopy. SS316L, Viton O-rings, no oil residue.
- **Ultra-high vacuum (UHV)** — 10⁻⁵ to 10⁻⁹ Pa. MBE, ALD, ion implantation, surface science. **CF (ConFlat) flanges, copper gaskets, vacuum-baked materials.**
- **Extreme high vacuum (XHV)** — < 10⁻⁹ Pa. Particle accelerators, fundamental research. Special bake-out processes 200–400 °C, NEG (Non-Evaporable Getter) pumps.
The boundary where everything changes: **HV → UHV**. Once you go below 10⁻⁵ Pa, **outgassing** (release of gas from materials) becomes the limiting factor, not pumping speed. Ordinary materials in UHV emit water, hydrocarbons, oxygen, nitrogen at rates the pump can't draw away.
Cleanroom classes under ISO 14644-1
- **ISO 9** — outdoor environment, > 35M particles ≥ 0.5 μm per m³. Ordinary hall.
- **ISO 8** — ordinary clean warehouse, ~3.5M particles ≥ 0.5 μm per m³. Cheap optics, mechanical assembly.
- **ISO 7** — 352,000 particles ≥ 0.5 μm per m³. Medical device manufacturing, OLED display assembly.
- **ISO 6** — 35,200 particles ≥ 0.5 μm per m³. Pharma sterile, hard disk drive assembly.
- **ISO 5** — 3,520 particles ≥ 0.5 μm per m³. Semiconductors (front-end-of-line wet etch), pharma aseptic fill.
- **ISO 4** — 352 particles ≥ 0.5 μm per m³. Semiconductors (photolithography area), space optics.
- **ISO 3** — 35 particles ≥ 0.5 μm per m³. EUV lithography, advanced node (7 nm and below).
- **ISO 1** — 10 particles ≥ 0.1 μm per m³. Border-line — research only, almost no production.
Semiconductor production in 2026 is standard ISO 3 or better for photolithography. A pharma fill suite is typically ISO 5 (Grade A under EU GMP Annex 1). Optical components for lithography: ISO 4.
Outgassing budget — a concept nobody states explicitly
For a UHV system `Q_chamber` (total outgassing rate, Pa·m³/s) is the sum of contributions of all surfaces and materials in the chamber. A pump with effective speed `S` (m³/s) holds the equilibrium pressure:
``` P_eq = Q_chamber / S ```
For P_eq = 10⁻⁸ Pa with a pump S = 1,000 l/s (turbo pump Twistorr 1000 or Pfeiffer HiPace 1500) the acceptable Q_chamber = 10⁻⁸ × 1 = 10⁻⁸ Pa·m³/s.
Typical outgassing rates (room temperature, unbaked):
| Material | Outgassing rate (Pa·m³/s/m²) | |----------|------------------------------| | 304L bake-out for 24 h | 10⁻¹² | | 316L electropolished, vacuum-baked | 10⁻¹¹ | | 304L electropolished, unbaked | 10⁻⁹ | | 304L plain machined | 10⁻⁸ | | Viton O-ring | 10⁻⁷ | | Viton after 24 h pump-down | 10⁻⁹ | | Standard silicone rubber seal | 10⁻⁴ | | Standard machine oil (Loctite, Molykote) | 10⁻³ | | Standard steel with organic residue | 10⁻⁵ |
**One M6 bolt with a surface of ~0.001 m²** and outgassing 10⁻⁵ Pa·m³/s/m² (an ordinary machine bolt with oil contamination) contributes 10⁻⁸ Pa·m³/s — **exhausts the chamber's entire outgassing budget.**
This is why in UHV you use:
- **Silver-plated bolts (M6, M8, A2-70 base + 5–10 μm Ag coating)** — silver has extremely low vapour pressure, doesn't oxidise during bake-out at 250 °C
- **Vacuum-baked screws** — 250–350 °C for 12–48 h in 10⁻⁵ Pa, removes absorbed water + organic residues
- **Bolts without oil residue** — vacuum-grade lubricant Apiezon-L or dry-film MoS2 instead of standard oil
A real-world story: a single bolt that destroyed a lot
Client: ASM Pacific Holding subcontractor, production of SiC power modules for EV traction inverters. Vacuum chamber for the die-attach process (silver sintering at 250 °C, 50 MPa pressure, 10⁻³ Pa). Yield baseline 87%.
**Incident:** a technician replaces an M8 bolt on the chamber port flange (damaged during maintenance). Uses an ordinary stainless A4-70 from the workshop stock — looked identical, "it's just a fixing bolt."
**Consequence:** - The bolt had absorbed surface water + machining oil residue (oil layer ~10 nm) - The bake-out cycle was not performed on the bolt itself (it was assumed the 350 °C 8 h chamber bake-out would resolve it) - On the first production cycle (250 °C, vacuum draw) the bolt released ~10⁻⁵ Pa·m³/s of outgassing over 4 hours - Outgassing products (mainly hydrocarbons from oil + Cr/Ni oxides from the surface) condensed on cooler surfaces in the chamber — including the workpieces - Silver sinter bond strength dropped from 60 MPa to 18 MPa — failed at the outgoing inspection
**Cost:** - 14 lots × 240 modules = 3,360 SiC power modules lost, average €680 / module = €2.28M - 9 days of downtime during investigation + decontamination (chamber bake-out 400 °C 72 h, replacement of all Viton gaskets, full re-baseline) - No harmful sequential effects (the following lot reached 89% yield)
**True cost of the incident:** ~€2.6M. **Cost of the right bolt:** silver-plated M8 vacuum-grade ~€18 vs. standard A4-70 ~€0.80 = €17 difference per bolt × 24 bolts on the chamber port = €410 for the whole port set. **ROI of the right part: 6,300×.**
Glove discipline — when latex contaminates, when nitrile is the requirement
Cleanroom gloves aren't "gloves" — they are certified tools with defined particle shedding rate, outgassing profile and extractables/leachables.
**Latex (natural rubber)** — medium particle generation (cornstarch in cheap variants), latex protein induces sensitisation in 10–15% of personnel after 6 months of daily wear. Use: ISO 7+ cleanrooms, pharma non-sterile.
**Nitrile (acrylonitrile-butadiene)** — low particle generation in cleanroom-grade variants, no latex proteins, excellent chemical resistance to solvents (IPA, acetone). Use: ISO 5–7 semiconductors, pharma sterile (post gamma irradiation), composite manufacturing.
**Cleanroom-grade Nitrile (FDA, USP Class VI)** — triple-washed, individually packed, sterile, < 100 particles ≥ 0.5 μm per glove. Use: ISO 3–5, pharma aseptic, optical assembly.
Latex is disqualified for UHV (outgassing from zinc oxide accelerators), for photolithography (particle shedding contaminates the wafer) and for EUV reticle handling (there only PVA finger cots or ESD-safe nitrile go). Nitrile is mandatory for solvent handling, semiconductor manufacturing and everything ISO 5 and better.
Bolts for UHV — a specification you must not shorten
**Standard specification for an M6 bolt on a UHV CF flange:**
- **Material:** A2-70 base (304L stainless) or A4-70 base (316L stainless)
- **Surface treatment:** vacuum bake-out 250 °C 24 h in 10⁻⁵ Pa **or** silver plating 5–10 μm
- **Cleaning protocol:** triple ultrasonic clean in deionised water + IPA, dry under filtered N2, packed in a cleanroom-grade bag
- **No lubrication except:** Apiezon-L (UHV-grade vacuum grease) or dry MoS2 film (5 μm DLC coating)
- **Supplier:** Lesker (Kurt J. Lesker Company), VAT Group, MDC Vacuum Products, Pfeiffer Vacuum — **never** a general industrial supplier (Würth, Inserco, Fabory)
**Price:** silver-plated M6×20 mm = €5–12 / piece (vs. €0.20 standard). For a chamber with 80 bolts = €400–960 vs. €16. The difference is trivial against a potential lost lot.
ESD and particle counter — two disciplines that overlap
**ANSI/ESD S20.20-2021** (EU equivalent IEC 61340-5-1) is the baseline for semiconductors, MEMS and sensitive electronics. Personnel grounding via wrist strap (1 MΩ to ground), ESD-safe footwear, worktops 10⁶–10⁹ Ω/sq, ionisers (Simco-Ion AeroBar) where contact grounding isn't guaranteed. Audit: monthly continuity test, quarterly ioniser balancing. ESD damage is "latent" — the component passes the post-build test, fails 6–18 months in the field. For an automotive ECU or medical implantable that means recall + liability + certification reset.
Particle monitoring via discrete counter (Lighthouse Solair, TSI AeroTrak): baseline should be < 30% of the class limit. For ISO 5 (3,520 particles ≥ 0.5 μm/m³) the robust baseline is < 1,000 particles/m³ — reserve for transient events (door opening, personnel movement). A cleanroom with a baseline of 2,800/m³ is "technically in class," but one bad event shifts it to OOC (Out of Class), which means stopping production and an investigation.
Five top mistakes that cost the client yield
1. **Substitute bolt** — an ordinary industrial bolt instead of vacuum-grade. Outgassing budget shot. 2. **Recycled cleanroom suit** — repeated laundering reduces particle holding capacity. Lifecycle max 30–50 cycles, then it sheds. 3. **Latex instead of nitrile for solvent handling** — glove degradation by IPA → particles + dissolved chemicals → contamination. 4. **No vacuum bake-out after a component swap** — a new part is added but the bake-out cycle is skipped. Outgassing rate overrides the baseline. 5. **HVAC filter overdue** — HEPA H14 / ULPA U15 filters have lifetime 12–36 months at ISO 5. After expiry leakage rate rises, particle count drifts up.
---
*We do contamination control audits for vacuum and cleanroom operations in SR/CZ/AT/DE. The first consultation (90 min) walks through your production pipeline and flags the 2–3 most likely places where outgassing budget or particle baseline is drifting — usually these are operational disciplines, not new hardware investments.*