The datasheet says "HEPA H13, 99.95% efficiency" — but how was that number measured? At what particle size? Does "99.95%" apply to the entire filter or just the average? Understanding EN 1822 answers all of these.
Where EN 1822 Comes From
EN 1822 is the European standard for grading and testing high-efficiency air filters, published by CEN (European Committee for Standardization). Its full title is "High efficiency air filters — EPA, HEPA and ULPA." It covers filters with efficiency ≥ 95% — well above what typical office HVAC requires.
Before EN 1822, every region had its own system: the US used MIL-STD-282 (DOP method), Japan used JIS B 9927, Germany used DIN 24184. EN 1822 unified these fragments, and was later adopted internationally as ISO 29463.
For manufacturers and buyers, the EN 1822 grade is a common language — whether a Taiwanese plant ships to Germany or a Japanese maker delivers to a US fab, H13 means H13, tested the same way.
EPA / HEPA / ULPA — Three Families
EN 1822 splits high-efficiency filters into three families:
| Family | Grade Range | Plain English |
|---|---|---|
| EPA (Efficient Particulate Air) | E10 – E12 | "Efficient" but below the HEPA threshold |
| HEPA (High Efficiency Particulate Air) | H13 – H14 | The real "high-efficiency particulate air filter" |
| ULPA (Ultra Low Penetration Air) | U15 – U17 | "Ultra-low penetration," semiconductor grade |
Each grade is defined by two metrics — overall efficiency and local efficiency:
EN 1822 Complete Classification
EPA → HEPA → ULPA — 10 grades from E10 to U17; higher number = higher efficiency but also higher ΔP and cost
| Class | Type | Overall eff. ≥ | Local eff. ≥ | Typical application |
|---|---|---|---|---|
| E10 | EPA | 85% | — | General ventilation, AHU pre-filter |
| E11 | EPA | 95% | — | Medium stage, electronics secondary |
| E12 | EPA | 99.5% | — | Hospital general areas, pharma pre-stage |
| H13 | HEPA | 99.95% | — | Cleanroom ISO 7–8, surgical suites |
| H14 | HEPA | 99.995% | ≥ 99.975% | Semiconductor front-end, ISO 5, BSL-3 labs |
| U15 | ULPA | 99.9995% | ≥ 99.9975% | Advanced EUV process, ISO 3–4 |
| U16 | ULPA | 99.99995% | ≥ 99.99975% | Extreme clean environments, quantum labs |
| U17 | ULPA | 99.999995% | ≥ 99.9999% | Theoretical limit grade, special research |
Testing is based on MPPS (Most Penetrating Particle Size, ~0.12–0.25 μm) — the hardest particle size for any filter media to capture. "Overall" efficiency is the whole-filter average; "Local" efficiency is the worst single point found during scan testing. H13 and below require overall efficiency only; H14 and above also require local efficiency, ensuring no single point leaks.
Key takeaways:
- ▸E10 to E12 are tested on overall efficiency only — no scan test required
- ▸H13 and above must also pass a local efficiency test — the weakest spot on the filter must meet a separate, stricter threshold
- ▸Each step up is roughly 10× tighter on leakage: H13 allows 0.05% penetration, H14 allows only 0.005%
Analogy: H13 is like "class average above 95, and the lowest scorer cannot drop below 75." H14 is "class average above 99.5, lowest scorer above 99.75." ULPA is "nobody in class is allowed to fail."
MPPS — Why Not 0.3 μm?
You may have heard "HEPA filters 0.3 μm particles at 99.97% efficiency" — that is actually the US IEST definition. EN 1822 does not use a fixed 0.3 μm. Instead it uses MPPS (Most Penetrating Particle Size) — the specific particle size that is hardest for this particular filter to capture.
Why? Because different media have different weak spots:
- ▸Glass-fiber HEPA: MPPS typically 0.12 – 0.25 μm
- ▸PTFE membrane: MPPS may be 0.05 – 0.1 μm
- ▸Electret media: MPPS drifts over time as charge decays
MPPS: Where the Filter's "Achilles' Heel" Is
Mechanical interception catches large particles, diffusion catches small ones — the size in between that neither handles well is MPPS
The legacy US standard (IEST) uses a fixed 0.3 μm test particle, but 0.3 μm is already past the MPPS — so the measured efficiency looks better than the true worst case. EN 1822 tests at MPPS, which catches the real weakness. That is why the same filter tested under both standards will show a slightly lower number under EN 1822.
Actual MPPS varies by media type, fiber diameter, and face velocity; typical range is 0.12–0.25 μm. EN 1822 tests at the "hardest-to-catch" particle size, which is stricter and more realistic than testing at a fixed 0.3 μm.
MPPS is the filter's Achilles' heel — testing at that size reveals the worst-case efficiency. Testing at a fixed 0.3 μm lets some filters dodge their weak spot and report flattering numbers.
Key point: EN 1822 efficiencies are measured at MPPS, so the same filter will show a "lower-looking" efficiency number compared to a fixed-0.3 μm test — but the EN 1822 number is more honest.
Three-Stage Test Flow
The full EN 1822 test has three stages. Higher grades must pass more stages:
EN 1822 Test Flow: Three Gates from Raw Media to Shipment
Media passes a material-level efficiency test first, then assembled filters get an overall scan, and H14+ adds a per-unit local leak scan — all three gates must pass before shipping
Test raw media sample (not assembled filter) with DOP/PAO aerosol at MPPS
Assembled filter at rated airflow — DOP/PAO upstream vs downstream concentration ratio
Probe scans the downstream face in an S-pattern, flagging any point above the leak threshold
DOP/PAO aerosol is concentrated at MPPS (~0.12–0.25 μm), the hardest particle size for the media. Scan speed is typically 3–5 cm/s covering the full downstream face. H13 requires only the first two gates (media + overall); H14 and above require all three.
Stage 1 — Media efficiency test A small sample is cut from the finished filter and challenged with MPPS-sized aerosol in the lab. This establishes the baseline capability of the raw media.
Stage 2 — Overall efficiency test The complete filter is mounted in a test duct and challenged with the same aerosol. Upstream and downstream concentrations are compared. This measures "media + frame + sealant + assembly quality" as a whole.
Stage 3 — Scan leak test A probe traverses the downstream face of the filter point by point (like sweeping for mines), locating the single worst spot. Any point exceeding the local-efficiency threshold means the filter fails.
E10 – E12 need stages 1 and 2 only. H13 and above must pass all three. H14 and above face even stricter local-efficiency thresholds.
For scan test operational details, see PAO Scan Test Guide.
Overall Efficiency vs Local Efficiency
This is one of EN 1822's most important concepts:
- ▸Overall efficiency: average penetration across the entire filter face — measures "bulk performance"
- ▸Local efficiency: worst single-point penetration found during scanning — measures "weakest link"
Why two metrics? Because a filter can have an excellent average yet still have pinhole leaks. Common local weak points:
| Weak Point | Cause |
|---|---|
| Frame-to-media bond | Uneven sealant application |
| Pleat crests | Thinner media at fold lines |
| Separator edges | Gaps between aluminum separators and media |
| Media splices | Large filters require media joints; seams are weak spots |
In a cleanroom, a single pinhole leak can ruin the entire room's particle count — that is why H14 and above require scanning, with local-efficiency thresholds stricter than overall efficiency.
Practical Selection: EN 1822 Grade vs Cleanroom Class
The most common procurement question: "My cleanroom is ISO Class 5 — should I use H13 or H14?"
Quick reference (actual selection also depends on air change rate, return-air ratio, etc.):
| ISO 14644 Class | Suggested EN 1822 Grade | Typical Application |
|---|---|---|
| ISO Class 8 – 9 | E11 – E12 | General electronics assembly, food packaging |
| ISO Class 6 – 7 | H13 | Pharma filling, LCD panel |
| ISO Class 5 | H14 | Semiconductor front-end, TFT-LCD |
| ISO Class 4 and above | U15 – U17 | Advanced node (EUV lithography, wafer inspection) |
For more on cleanroom classification, see ISO 14644 Cleanroom Classification. For the basics of HEPA vs ULPA, see HEPA vs ULPA Difference Guide.
Practical tip: The grade is a minimum threshold. Semiconductor fabs typically specify additional requirements like "overall efficiency ≥ 99.999%" or "local penetration < 0.001%" — tighter than the H14 standard. When negotiating with suppliers, always state "per EN 1822 test method" + "efficiency value" + "overall or local" — all three pieces are essential.
FAQ
Q: How much difference is there between H13 and H14?
A: In penetration terms, H13 allows up to 0.05% penetration (99.95% efficiency); H14 allows up to 0.005% (99.995%) — a 10× difference in leakage. In a cleanroom, that factor of 10 often separates ISO Class 6 from ISO Class 5.
Q: The US says HEPA is 99.97%, Europe says 99.95% — which is correct?
A: Both are correct but use different test methods. The US IEST definition tests at a fixed 0.3 μm DOP, threshold 99.97%. EN 1822 H13 tests at MPPS, threshold 99.95%. Because MPPS is the hardest particle size to capture, 99.95%@MPPS and 99.97%@0.3 μm are roughly equivalent in real-world difficulty.
Q: Is MPPS always 0.3 μm?
A: No. MPPS varies by media, typically between 0.05 and 0.3 μm. The 0.3 μm figure comes from early research showing that most glass-fiber media have an MPPS near 0.3 μm. Modern media — especially PTFE membranes — often have a smaller MPPS. EN 1822 requires each filter's actual MPPS to be determined first, then used for the efficiency challenge.
Q: Does EPA (E10–E12) count as HEPA?
A: Strictly, no. EN 1822 explicitly separates EPA from HEPA. EPA has a lower efficiency threshold (E10 requires only ≥ 85%) and does not require scan testing. Some vendors market E11 or E12 as "sub-HEPA," but in formal specification documents, H13 is where HEPA begins.
Q: Can I downgrade from H14 to H13 when replacing filters?
A: It depends on your cleanroom validation protocol. If the protocol specifies "terminal filtration H14," downgrading will likely push the ISO Class out of spec, and you will need to re-validate airflow balance and particle counts after replacement. If cost savings justify the attempt, perform post-installation particle counting to verify that actual concentrations remain within ISO Class limits, and obtain written approval from QA or the end customer.
