Placement decisions made during schematic design quietly determine whether AMC events in a modular semiconductor cleanroom are caught before yield loss or only after it. Teams that default to treating all chemical contamination in a single air path—typically the make-up air unit—discover the flaw months later when process-zone emissions from photoresist stripping or etch chambers pass through the recirculation loop unchecked, media life falls short of projections, and the first clear signal is a lithography excursion rather than a filter alarm. The harder retrofit is not adding media capacity; it is retrofitting monitoring infrastructure and access provisions into a module layout that was never designed for distributed chemical filter changeout. Understanding which source drives which air path, and how that maps to fan selection, pressure control, and breakthrough detection, is what allows an engineering or QA team to make a placement decision that holds through commissioning and into production operation.

Identify Whether The Chemical Source Is Outdoor, Facility, Or Process Generated

Source identification is not a documentation step—it is the constraint that determines which air path carries the filtration burden and whether one path alone is sufficient.

Outdoor air entering a make-up air handling unit (MAHU) carries ambient amines, sulfur dioxide, and organic compounds whose concentrations vary by site geography, industrial neighbors, and seasonal conditions. These contaminants enter the cleanroom through the fresh air fraction and, if untreated, distribute across the entire recirculation volume. The make-up air path is the correct interception point for this category because it is the only location where outdoor AMC can be removed before diluting into the clean space.

Process-generated AMC behaves differently. Amines released from photoresist chemistry, acid gases from etch processes, and organic solvent vapors from coating and strip steps originate inside the clean zone. These compounds enter the recirculation airstream and accumulate in proportion to process intensity and local exhaust capture efficiency. No amount of make-up air filtration addresses this load, because the contaminants never pass through the MAHU. For these sources, chemical filtration in the recirculation air handling unit (RAHU) or at localized return points near the tool interface is the relevant design response.

A third source category is often underweighted at the design stage: facility-generated AMC from construction materials, filter media binders, adhesive sealants, painted surfaces, and cable jacketing. This category tends to be highest during initial cleanroom qualification and declines over months of operation, but it can temporarily overwhelm chemical media capacity if not accounted for in media sizing. Unlike outdoor or process sources, facility outgassing may not be addressable through filtration placement alone—it may require material selection control during construction and an accelerated purge protocol before occupancy.

The classification framework most useful for structuring this analysis is SEMI F21, which defines allowable airborne molecular contaminant concentrations by compound class. Using those thresholds as design targets, teams can compare measured or estimated source loads against allowable limits to assess which sources drive exceedance risk. ISO 14644-8:2022 provides the assessment methodology for measuring chemical air cleanliness and is useful as a testing reference when validating that post-filtration concentrations meet those targets. Neither standard mandates a specific filter placement topology—they define what must be achieved, not where the media must sit.

The practical mistake is treating source identification as a binary choice between outdoor and process emissions. When both are significant—as they are in sites near industrial corridors running EUV or immersion lithography—both air paths require chemical filtration, and the media selection for each path may differ. Treating one as negligible without site-specific characterization data is an assumption that tends to surface as a media life shortfall or a persistent AMC exceedance that does not respond to make-up air media changes alone.

Compare Make-Up Air Treatment With Recirculation Filtration

The core trade-off between make-up air treatment and recirculation filtration is not about which approach is better—it is about which source dominates and whether layering both is justified by the contamination profile.

Make-up air typically represents roughly 5–15% of total cleanroom airflow in a recirculating design, but conditioning that fraction to cleanroom specification requires more energy per unit volume than the recirculation air. Chemical scrubbing in the MAHU is cost-effective when outdoor AMC loads are the primary risk, because intercepting contaminants before they dilute into the clean volume means the media in the RAHU is not doing double duty. When outdoor sources are light and the facility sits in a relatively clean ambient environment, heavy investment in MAHU chemical media may not return proportional protection.

Recirculation filtration handles the bulk of airflow and is better positioned to respond to contaminants released inside the clean space. However, if the make-up air path carries untreated outdoor contaminants, the RAHU chemical media absorbs both the indoor and the outdoor load—shortening service life and potentially creating localized exceedances near fresh air introduction points before the media reaches equilibrium. This is a common failure pattern in projects where HVAC contractors specify the MAHU and cleanroom module suppliers specify the RAHU without a shared source characterization.

Local AMC filtration at individual tool enclosures—most commonly applied at EUV scanner environments—represents a third tier that supplements central treatment rather than replacing it. This approach is application-specific; it is appropriate when a single tool generates or is uniquely sensitive to a particular compound class at concentrations that central recirculation filtration cannot reliably reduce to the required level. It is not a universal design requirement and should not be used as a substitute for addressing upstream source loads.

The most common layering decision arises when a site has measurable outdoor amine or SO₂ sources and also runs wet etch or photoresist processes. In that case, treating the make-up air path is necessary to protect the RAHU media from outdoor load, and RAHU chemical filtration is necessary to address internal emissions. Attempting to handle both with make-up air treatment alone typically results in media exhausted earlier than predicted and residual internal contamination that make-up air filtration structurally cannot reach.

Coordinate Chemical Media With FFU And Pressure Strategy

Chemical filtration media introduce pressure drop that must be carried through FFU selection and fan commissioning—not estimated at specification and then verified only at startup.

Fan filter units sized for HEPA or ULPA pressure drop alone will not maintain target air changes per hour once chemical media modules are integrated into the same airflow path or upstream of the FFU array. The fan curve must be confirmed against worst-case loaded filter differential pressure for both the particulate and chemical media stages combined. For FFUs using EC motors or variable frequency drives, the control range must cover the additional resistance across the full media service life, not just the initial clean state. If this verification is deferred to commissioning, the outcome is either insufficient ACH in the critical zone or fans running at speeds that compromise noise, energy, and bearing life targets.

A related but distinct problem is the conflation of particulate filter change triggers with chemical media management. Differential pressure monitoring is the standard indicator for particulate filter loading—replacement is typically triggered when pressure drop reaches 150–200% of the initial clean reading. Chemical media breakthrough does not follow the same pattern. A carbon or chemisorbent bed can be functionally exhausted while showing negligible change in differential pressure, because molecular-weight contaminants pass through the media without meaningfully increasing resistance. Treating differential pressure as a universal filter status indicator for chemical media is an incomplete maintenance strategy and one that routinely delays breakthrough detection until yield data provides the signal instead.

The pressure cascade interaction deserves explicit review at the design stage. Make-up air volume controls cleanroom pressurization relative to adjacent spaces. If chemical filtration added to the MAHU path increases resistance beyond the original fan sizing, the practical effect may be a reduction in make-up air delivery that compresses the pressure differential at zone boundaries. Automated damper control can compensate within limits, but if the chemical filter DP budget was not included in the original MAHU fan selection, the available compensation range may be insufficient. This should be a documented design review checkpoint—confirm that the damper control logic and MAHU fan capacity can maintain the required cascade gradient under worst-case loaded chemical filter conditions. For FFU-based recirculation systems, reviewing how chemical media interacts with fan speed control is part of the same coordination task; the Unitatea de filtrare a ventilatorului - FFU selection must account for the combined filter system’s resistance envelope across its service life.

Maintenance And Breakthrough Risks In Distributed Filter Layouts

Distributing chemical filtration across multiple modules—closer to process tools, away from central air handlers—solves a contamination response problem and creates a maintenance accountability problem simultaneously.

The operational advantage of distributed media placement is proximity to the contamination source. A chemical filter module integrated into a local return plenum near an etch cluster intercepts process emissions before they re-enter the full recirculation volume. Central RAHU filtration handles what escapes local capture. This tiered response improves localized control, but each filter location is a maintenance point that requires access provisions, a change-out schedule, and someone accountable for monitoring status. In modular cleanroom layouts assembled from components specified by different parties, distributed chemical filter locations can outnumber the access provisions designed into the module—a configuration mismatch that is inexpensive to correct at layout and expensive to retrofit.

The more serious operational risk is the latency between chemical media breakthrough and any observable consequence. Contamination symptoms—yield excursions, resist sensitivity shifts, optical surface fogging—may appear hours to days after the breakthrough event, making root cause diagnosis slow and correlation to a specific filter change event unreliable. This lag makes reactive management of chemical media a structurally weak strategy. The relevant operational response is proactive AMC monitoring using atmospheric pressure ionization mass spectrometry (APIMS) or equivalent instrumentation capable of detecting molecular-scale contaminants at concentrations relevant to SEMI F21 compound class thresholds. Differential pressure monitoring remains necessary for particulate filter management but should not be extended as a proxy for chemical media status. ISO 14644-8:2022 provides a testing framework for sampling location, analysis method selection, and cleanliness assessment that can support the design of a monitoring plan—though the specific alarm setpoints and replacement schedules must be defined as part of the facility’s own operational specification, not derived from the standard alone.

The failure pattern most difficult to diagnose in distributed layouts is partial breakthrough in one filter bank while other banks remain functional. Because AMC sensors are rarely positioned at every filter outlet, a localized breakthrough may contaminate a specific process zone while aggregate room-level measurements appear within tolerance. The monitoring plan must include sensor placement logic that reflects where a breakthrough would first reach a sensitive tool or process zone, not just where sensors are easiest to install. For lithography clusters, this typically means sensor placement inside or immediately upstream of tool enclosures, not at the module return. Teams working through FFU-based recirculation architecture will find that the distributed nature of fan filter arrays complicates this sensor placement further; coordinating chemical media locations with FFU array geometry and return air paths is addressed in more detail in the context of FFU vs. conventional HVAC system comparisons.

Decision Checklist For Chemical Filter Placement

The checklist below is a planning aid, not a guaranteed design recipe. Its value lies in forcing the source characterization and accountability questions to be answered before filter locations are committed to a module layout—not after equipment is on order.

The most common ambiguous case is when both outdoor and internal sources contribute meaningfully. A site adjacent to an industrial corridor that also runs high-volume photoresist processing cannot resolve its AMC risk through a single-path decision. Both the make-up air and recirculation paths require chemical filtration, and the media specification for each may differ in compound class priority—acid gas capacity in the MAHU to handle ambient SO₂, and amine-specific chemisorption in the RAHU to handle photoresist amines. Forcing a single-path decision in this scenario typically results in one source being undertreated, with the consequence deferred to commissioning or early production.

The SEMI F21 concentration thresholds referenced in the checklist—amines below 0.1 µg/m³, organics below 1 µg/m³, acid gases below 0.1 µg/m³—function as industry-accepted design targets for setting filter media selection and breakthrough criteria in the absence of facility-specific contractual limits. They are not universal regulatory mandates; their applicability depends on whether they have been incorporated into the facility specification or customer requirements. Using them as planning benchmarks is appropriate; presenting them as compliance thresholds without that contractual grounding is not.

The verification responsibility step in the checklist is where most project coordination gaps appear. When chemical filter banks are split across a MAHU (specified by the HVAC contractor), RAHU modules (specified by the cleanroom module supplier), and local tool enclosures (specified by the equipment vendor or process owner), no single party has natural ownership of the end-to-end AMC performance outcome. The checklist question is not rhetorical—it requires a named party, a defined monitoring method, a replacement trigger, and a documented handoff process before the project reaches commissioning. Without that structure, breakthrough events in production are likely to generate finger-pointing rather than rapid response, and the latency between contamination and yield impact will compound the diagnostic difficulty.

The most consequential placement decisions are made at the layout stage, when source characterization data may still be incomplete and the temptation is to defer media placement to a later design phase. By the time module fabrication is underway, access provisions, fan curves, and damper control logic are already fixed—and retrofitting chemical filter capacity or monitoring infrastructure into a completed module is disproportionately expensive relative to addressing those decisions at schematic design.

Before committing to a chemical filter placement strategy, confirm three things: that source characterization distinguishes outdoor, facility, and process AMC contributors rather than treating them as a single load; that chemical media pressure drop has been verified against FFU fan curves under worst-case loaded conditions; and that breakthrough monitoring responsibilities are assigned to a named party with a defined instrument, sensor location, and alarm threshold referenced to project-specific AMC targets. If any of these confirmations are outstanding when module procurement begins, the risk is not abstract—it is a pressure cascade problem or a yield-impacting contamination event that will be significantly harder to resolve in a commissioned cleanroom than in a design review.

Întrebări frecvente

Q: What if our cleanroom uses a once-through airflow design instead of recirculation?
A: In a once-through system, all chemical filtration must be placed in the make‑up air path. The recirculation‑versus‑make‑up trade‑off disappears, so the design effort shifts to correctly sizing the make‑up air chemical media for both outdoor contaminants and any fugitive internal emissions that re‑enter through airlocks or pressurisation leaks.

Q: After we settle on a placement strategy, what is the first concrete verification step?
A: Commission a site‑specific AMC baseline survey using the same measurement method (e.g., APIMS) that will be used for ongoing verification. This confirms that the outdoor, facility, and process source loads assumed during placement match real concentrations, letting you correctly size media beds and set meaningful breakthrough alarm thresholds before procurement.

Q: At what concentration does central recirculation filtration stop being enough, making local tool‑level filters essential?
A: When the modelled or measured contaminant concentration at the most sensitive tool inlet reaches approximately 50 % of the lower of the SEMI F21 recommended limit or the tool manufacturer’s exposure specification. Below that margin central treatment is usually sufficient; above it, adding local enclosure filtration is the safer way to prevent yield‑impacting excursions.

Q: Is it more cost‑effective to centralise chemical filters in the RAHU or to distribute them next to process tools?
A: Centralisation reduces the number of maintenance points and access costs but may require a larger media bank and can let local emission spikes reach multiple tools. Distributed modules lower total media volume and localise containment, but multiply replacement labour and monitoring points. The best balance depends on tool density, dispersion patterns, and service‑corridor access—tightly clustered tools often favour central RAHU treatment, while isolated sensitive tools justify local modules.

Q: If our outdoor air is consistently clean and chemical use is modest, how do we justify installing chemical filtration in both air paths?
A: If budget forces a phased approach, install recirculation filtration first to protect process tools from internal emissions. This is only safe when outdoor AMC levels are confirmed below 10 % of the SEMI F21 target for the most sensitive compound class and you commit to continuous outdoor monitoring—so you can add make‑up air filtration before any seasonal or industrial change prematurely exhausts the recirculation media.