Animation Cel Preservation Guide: Storage Tiers & Lifetime Risk


Animation cels don’t simply “age.” They follow the same reliability curve that governs engineered materials: an early drift phase, a long stable window, and an accelerated wear-out period driven by chemical kinetics and environmental stress. Vinegar syndrome is not random decay — it is a chemical failure mode governed by heat, water activity, and feedback from trapped acetic acid.


Collectors are often told that storage is “good,” “archival,” or “museum-grade,” but those labels rarely correspond to measurable environmental control. A frame, a binder, or a box is only as safe as the physics inside it. Warm rooms accelerate acid generation. Sealed volumes trap it. Cold storage can be either highly protective or mechanically destructive depending on moisture and cycling. Marketing language cannot resolve these contradictions — but a risk-based engineering framework can.


This document applies tools normally used for mission-critical systems — reliability curves, functional requirements, failure modes, Arrhenius kinetics, and diffusion-limited mass transfer — to animation cel preservation. Instead of treating storage choices as opinions, we evaluate them through three independent likelihood vectors:


  • Chemical (Lchem) — hydrolysis rate from temperature, absolute humidity, and acid feedback.
  • Structural (Lstruct) — mechanical fatigue from compression, RH cycling, thermal cycling, and geometry.
  • Photochemical (Luv) — UV and visible-light dose.

Each vector is scored independently for how fast it moves a cel along the vinegar-syndrome curve. Preservation tiers are then defined by measured improvements in these vectors, not by product labels. The result is a transparent, physics-based way to classify frames, binders, boxes, and engineered microclimates — and to design future systems that are objectively safer.

What is the Preservation Tier Framework?
The Cel Nexus Preservation Tiers are an engineering standard for animation art storage. They range from Tier 5 (High Risk), such as unbuffered attics, to Tier 1 (Archival Cold Storage), which uses temperature and humidity control to extend cel lifespan by 400% or more.

Archival Lifecycle Context: Preservation, Conservation, and Restoration

Before introducing the preservation framework, it is useful to establish the broader archival lifecycle. Animation cels pass through three distinct phases of technical care, each with its own intent and method. Understanding these phases clarifies the scope of this document.

1. Preservation — maintaining stability and preventing degradation

Preservation focuses on maintaining a stable microenvironment so the acetate base never enters the degradation regime. For cellulose acetate, this primarily involves controlling the factors that dictate the rate of hydrolysis and the accumulation of acetic acid:

  • Temperature
  • Absolute humidity
  • Boundary-layer behavior
  • Off-gas removal pathways (venting or scavenging)

Effective preservation suppresses R (acid generation) and prevents D (acid accumulation). The framework that follows is centered entirely on this phase.

2. Conservation — intervening once early deterioration is detectable

Conservation addresses early chemical or physical changes while the cel remains structurally sound. It includes controlled conditioning environments, acid-reduction strategies, and stabilization measures designed to re-establish favorable conditions before significant damage develops.

This phase is referenced for context but is not the focus of this document.

3. Restoration — correcting visible or structural damage

Restoration is reserved for advanced deterioration that compromises visual or physical integrity. It includes repair of paint loss, correction of severe warping or brittleness, reinforcement of weakened material, and other invasive treatments.

This phase lies outside the scope of engineered preservation.

Purpose of defining this lifecycle

Defining the archival lifecycle establishes a clear boundary for the analysis that follows. The framework in this document does not attempt to address conservation or restoration workflows. It provides a structured preservation model focused on understanding acid generation, acid accumulation, and the environmental controls required to avoid entering the degradation regime.

CTA Preservation Risk & Mitigation Framework

This section turns cellulose triacetate (CTA) preservation into a formal risk problem instead of a set of vague rules (“cool and dry”, “use archival frames”). We treat the storage environment like an engineered system, using Reliability-Centered Maintenance (RCM) and FMEA-style logic to ask a simple question: “How hard is this environment pushing my cel toward failure?”


TL;DR – How this framework works

  • We separate environment aggressiveness (likelihood, L) from the cel’s current condition (consequence, C). A Tier-1 box is Tier-1 whether it holds a mint cel or a damaged one.
  • The environment is scored along three vectors: chemical (Lchem – hydrolysis & VS), structural (Lstruct – RH cycling & mechanics), and UV/light (Luv – dose, spectrum, duration).
  • Chemistry gets a physics model. Lchem is derived from Arrhenius temperature factors, absolute humidity / water activity, and microenvironment behavior (dilution vs. trapping), not from gut feel.
  • Structural and UV vectors are scored qualitatively for now, using engineering judgement (how much cycling? how much handling? how much light?), all mapped onto the same 1–5 likelihood ladder.
  • Each vector carries its own risk score. We deliberately don't blend them into one weighted number — the failure modes are physically different, so a cel can be low-risk chemically but high-risk structurally. Every vector gets its own L × C score, and the environment's overall tier is governed by the worst vector.
  • Tier assignment is still conservative. The overall environment tier is limited by the worst vector (a “veto rule”), so a setup with excellent chemistry but very bad mechanics cannot be scored as Tier-1 just because Lchem is low.
  • In dark storage, Luv effectively sits at the bottom of the scale and stops driving risk; for display, it rises and can become the dominant vector.

Acid Generation vs Acid Emission – The Two Axes of Preservation

CTA degradation has two coupled halves: generation (hydrolysis producing acetic acid inside the film, driven by temperature and absolute humidity) and emission (that acid diffusing back out, governed by enclosure geometry, scavengers, and the surrounding vapor concentration). Storage that only helps one axis — an open room that vents well but runs hot and humid, or a sealed box that buffers temperature but traps vapor — can still drive risk up. The chemistry additionally runs in two regimes: a slow Induction baseline and a self-accelerating Autocatalytic regime once internal acidity crosses a threshold — which is why a cel can carry real risk before it smells like anything.

The full kinetic model — generation rate R0(T)·H(D), the Pint mass balance, and the Induction/Autocatalytic regime transition — is derived in the Engineering Framework. This framework applies those results directly as the Lchem,env scoring table and correction factor in Section 3.1, without re-deriving the physics here.

Problem Definition

Vinegar Syndrome (VS) in triacetate cels is usually explained with over-simplified guidance, such as “keep things cool and dry” or “use archival materials.” These statements are true in a general sense but are not precise enough to support quantitative decision-making about real storage setups.

Collectors therefore struggle with questions like:


  • Is a sealed frame with “archival” backing safer than a portfolio in a good room?
  • Does putting a box in the fridge help, or does condensation offset the temperature benefit?
  • Is a dehumidified room better than a sealed container with a desiccant pack?
  • Does adding a MicroChamber® board automatically make a frame “safe” for VS?

In practice, the same piece of art can be stored in a way that feels safer (boxed, framed, slabbed, sealed), yet in reality may carry a higher chemical or structural risk than leaving it on a shelf in a well-behaved room. Many of the most common “protective” actions actually trade one failure mode for another:


  • Sealing a frame may reduce dust and handling but can trap acetic acid and moisture.
  • Moving a box into a fridge slows hydrolysis but may create saturated, cycling microclimates.
  • Adding scavenger materials can help, but only if they can see the boundary layer and have remaining capacity.


To sort out these trade-offs, we need a framework that:


  • Separates likelihood (how aggressive the environment is) from consequence (how damaged the cel is now).
  • Distinguishes between chemical drivers (FM1/FM2 – hydrolysis and VS), structural drivers (FM5 – fatigue & warping), and UV/light drivers (FM4 – fading, yellowing).
  • Accepts that we can model chemistry quantitatively but must treat mechanics and UV as structured judgement calls until more data exist.
  • Maps common storage setups (open room, box, frame, fridge, engineered cold) onto a shared 1–5 scale so collectors can compare options.


The remainder of this section builds that framework, starting from simple “room vs. box vs. fridge” questions and ending with a multi-vector likelihood model and environment tiers that can be plugged directly into tools, checklists, and case studies.



Acid Transport: Generation → Boundary Layer → Removal

Mechanically, acetic acid crosses two bottlenecks on its way out of a cel: first from the reacting substrate into the local boundary layer, then from that boundary layer out into the enclosure or room. Early in a cel's life, generation outpaces diffusion almost by definition — internal concentration has to build before any gradient exists to push acid out, which is why some initial acid buildup is physically unavoidable. Past that point, whether risk keeps climbing depends on the second bottleneck: if removal (scavenging, ventilation) keeps pace with what reaches the surface, the boundary layer stays low and the gradient keeps pulling acid out; if it doesn’t, vapor accumulates and re-acidifies the film.

The full transport model — flux equations, the Rtot enclosure resistance term, and the Perfect Sink approximation — lives in the Engineering Framework. This framework uses its output directly as the Menvironment correction factor in Section 3.1 (enclosure geometry and scavenger state are evaluated jointly as a single lookup, since both are proxies for the same Pint buildup; paint-layer emission resistance is noted qualitatively there but not scored).

The Animation-Cel Bathtub Curve

Reliability engineers use the "bathtub curve" to describe how failure rate changes over a system's life: high early (infant mortality), low and flat in the middle (useful life), then rising again (wear-out). Animation cels follow the same shape — but unlike a generic piece of hardware, we know roughly where the wear-out transition sits (the IPI's 50-year archival baseline), and we know that storage choices can shift it. The chart below plots that directly instead of using a generic textbook illustration.

Animation cel bathtub curve, showing how preservation delays wear-out onset EARLY-LIFE MID-LIFE PLATEAU WEAR-OUT 50 yrs — IPI archival baseline 0 25 75 100 yrs High Low Relative VS Risk Rate Preservation's goal: push this transition right Unmanaged / archival baseline (68°F/50%RH) Actively preserved (high tier)

Here's how each region of the bathtub curve maps directly to animation cels:

Early-Life Region

Early-Life Region – Period of Decreasing Failure Rate

Failure rate starts high and then drastically decreases. For animation cels, this corresponds to failures driven by factory and handling defects:

  • paint adhesion or formulation issues from the studio
  • variation in polymer quality, plasticizer content, or curing
  • trapped moisture in the acetate during manufacture
  • stuck layers from humid studio storage
  • unstable dyes or paints that were marginal from day one


Storage today does not change this region — these are baked-in, “infant mortality” issues.

Mid-Life Plateau

Mid-Life Plateau – Period of Stabilized Failure Rate

This is where most 1980s–2000s cels live right now. Hydrolysis is active but not yet catastrophic. Internal acidity is slowly climbing, paint is intact, and diffusion through the film is still reasonably fast. The key is: the slope here is almost entirely controlled by environment.


Your storage tier decides whether this plateau is gently sloped or quietly steep:

  • Low Tier: Hot, humid, stagnant → steep slope → early onset of VS.
  • Medium Tier: Reasonably controlled → “natural” slope → coasting.
  • High Tier: Cold, dry, low-ppm → very shallow slope → life extension.
Why cels can look stable for decades, then fail rapidly: stabilizer capacity depletion

A useful way to interpret the mid-life plateau is stabilizer capacity. Many acetate formulations include stabilizers or scavenging capacity from manufacture. Early in life, this capacity buffers acidity rise; over time it is gradually consumed. Once stabilizer capacity approaches zero, internal acidity can rise faster and the cel transitions into a wear-out regime.

Preservation slows stabilizer depletion (by reducing Q and limiting moisture/heat), but cannot regenerate stabilizer capacity once it is exhausted — which is why proactive tiers matter.

Wear-Out Region

Wear-Out Region – Period of Increasing Failure Rate

Once internal acidity crosses a threshold, the cel enters a true wear-out regime:

  • CTA modulus increases; the base film stiffens and shrinks.
  • Paint experiences more mechanical stress and begins to crack or lift.
  • Diffusion slows, so new acid is generated faster than it can escape.
  • Boundary layers worsen; acetic acid is generated, trapped, and reabsorbed.
  • Vinegar syndrome becomes self-accelerating and obvious to the collector.


This is analogous to aging and wear-out failures in hardware. Environment still matters — you can slow the slope — but you’re now managing an already unstable system. Reset protocols are required if you want to move a cel back toward a lower-acidity state.

The red curve is the unmanaged/archival baseline, with wear-out onset around the 50-year IPI reference point. The green curve shows the same cel under active preservation: the early-life region is unchanged (those failures are baked in at manufacture), but the mid-life plateau extends and wear-out is pushed well past 50 years. That rightward shift is the goal of high-tier preservation.

Functional Requirements, Functional Failures, and Failure Modes

To apply reliability-centered thinking to animation cels, we have to be explicit about what it means for a cel to be “working” and what it means for it to have “failed.” In equipment reliability, we start with functional requirements, define functional failures, and then enumerate the failure modes that cause those failures. The same structure works cleanly for cels.

Framework Logic – From Functions to Preservation Spectrum

CTA Preservation Framework Logic Flow

CTA Preservation Framework Logic Flow Diagram

This logic flow shows how we move from system functions (what a storage or display environment must achieve), to functional failures (what it means when those functions are not met), to failure modes (the chemical and mechanical mechanisms), and finally into a structured FMEA and the risk framework. The resulting insights drive the Preservation Tiers, which are engineered environments that explicitly control the dominant failure modes to target specific likelihood and consequence bands.

The preservation framework follows a classic reliability-centered logic:

  • FR – System Functions: define what a storage or display system must do for a cel (e.g., limit hydrolysis, avoid local acid build-up, prevent mechanical fatigue, avoid light damage).
  • FF – Functional Failures: describe what it looks like when those functions are not met (e.g., “environment allows VS feedback,” “support geometry allows curl and blocking,” “display causes fading”).
  • FM – Failure Modes: map each functional failure to a specific physical / chemical cause (FM1–FM7).
  • FMEA: scores each FM for likelihood, consequence, and detection, and identifies the controls needed to keep that FM benign.
  • Preservation spectrum & tiers: stack those controls into coherent environments (Tier 1–5) and ask whether the full set of FMs is controlled to the desired likelihood band over a ~50-year horizon.

In other words, the functions define what the system must achieve, the FMEA defines how it can fail, and the preservation tiers define how much engineering control is applied to keep those failures from occurring.

Functional Requirements (FRs)

Functional Requirements (FRs)

For the purposes of preservation, an animation cel is expected to satisfy the following core functions:

  • FR1 – Structural Integrity: the CTA base remains flexible, continuous, and not embrittled.
  • FR2 – Chemical Stability: internal acidity remains low enough to avoid rapid, self-accelerating deterioration.
  • FR3 – Visual Fidelity: paint remains adhered, legible, and reasonably close to original color and gloss.
  • FR4 – Dimensional Stability: the cel stays flat enough to be safely handled and displayed (no severe cupping, rippling, or shrinkage).
  • FR5 – Handling/Display Survivability: the cel tolerates normal archival handling, inspection, and display without damage.

    • These functional requirements are defined at the level of the cel as a whole. The paint layer is not given its own independent FR/FF/FM set in this first version of the framework; instead, paint behavior appears in FR3 (visual fidelity) and in the associated functional failures and failure modes. In other words, the CTA substrate drives the chemistry, and the paint layer expresses many of the consequences. A later, more detailed model may decompose paint-specific requirements separately once more data exist

In addition to the CTA polymer and paint layers, animation cels contain a mobile plasticizer system (e.g., triphenyl phosphate and phthalates) that is integral to both mechanical compliance and long-term visual behavior. Plasticizers do not directly generate acetic acid; however, their presence and gradual depletion influence polymer mobility, stress relaxation, diffusion behavior, and the expression of deformation. Plasticizer loss can indirectly impact visual fidelity (FR3) through increased curl, surface distortion, paint stress, and loss of gloss uniformity, even in the absence of severe chemical degradation.

In this first version of the framework, plasticizer behavior is treated as an internal modifier of CTA substrate performance rather than a standalone functional requirement. Its effects are therefore captured implicitly within FR1 (structural integrity), FR3 (visual fidelity), and FR4 (dimensional stability). Future revisions of the framework may elevate plasticizer state to an explicit modeled parameter once sufficient empirical data exist to support independent thresholds and verification criteria.

Functional Failures (FFs)

Functional Failures (FFs)

A functional failure occurs when one or more of those requirements is no longer met, even if the object is still physically present. In practice, collectors experience functional failures as:

  • FF1 – Structural Failure: cracking, brittleness, tearing, or severe base instability.
  • FF2 – Chemical Failure: active vinegar syndrome (strong odor, high acidity, rapidly advancing damage).
  • FF3 – Visual Failure: obvious fading, chalking, haze, paint loss, or irreversible blocking.
  • FF4 – Geometric Failure: strong cupping, shrinkage, rippling, or stuck/background-bonded layers.
  • FF5 – Handling Failure: damage from normal handling, including creases, scratches, or paint lift.
Failure Modes (FMs)

Failure Modes (FMs)

With functions and failures defined, we can talk about the failure modes — the characteristic ways a cel drifts into those failure states. These are driven by the chemistry and physics we covered in the previous sections:

  • FM1 – Intrinsic Hydrolytic Chain Scission (Normal Aging): the ever-present background hydrolysis of CTA that slowly lowers Degree of Polymerization and produces acetic acid, governed by temperature and absolute humidity. This is the “natural” aging mode and sets the pace for eventual FF2.
  • FM2 – Autocatalytic Hydrolysis (Runaway VS): a secondary, accelerated mode that appears once internal acidity and local pH drop far enough that acid-catalyzed hydrolysis feeds itself. This mode rapidly drives FF1, FF2, FF3, and FF4.
  • FM3 – Vapor-Phase Feedback / Emission Bottleneck: small, poorly vented volumes or blocked geometries that allow acetic acid to build up around the cel, reducing diffusion out of the film and recycling acid back into it. This mode amplifies FM2 and accelerates entry into the wear-out region.
  • FM4 – UV-Induced Photooxidation: exposure to UV and high-energy visible light that degrades pigments and binders, driving FF3 independently of vinegar syndrome. UV acts as an accelerator on top of the chemical baseline.
  • FM5 – RH Cycling and Dimensional Fatigue: repeated moisture-uptake and loss cycles that cause expansion and contraction of CTA and paint, leading to accumulated stress, curl, and distortion (FF4), especially later in life or when plasticizer levels are reduced.
  • FM6 – Sustained Compression and Creep: long-duration static loading from stacking, tight framing, overfilled binders, gravity loading, or prolonged contact pressure that drives viscoelastic creep, thickness loss, set-in deformation, and optical distortion. This mode primarily affects dimensional stability (FF4) and visual fidelity (FF3), and is strongly time- and temperature-dependent.
  • FM7 – Enclosure Contamination: acids, plasticizers, or solvents migrating from non-archival boards, foams, and plastics into the cel stack, accelerating FF2 and compounding FF3.
  • FM8 – Mechanical Damage: one-off events such as creasing, compression shocks, frame over-tightening, drops, or rough handling that trigger FF5 (and sometimes FF3 or FF4 if layers stick, shear, or delaminate).


What we call a “preservation tier” is, in reliability language, really just a shorthand for which failure modes are being controlled well, which are being controlled poorly, and which are being actively accelerated.

FMEA Bridge

From Failure Modes to FMEA

Once functions, failures, and modes are clear, we can step into a structured Failure Modes and Effects Analysis (FMEA) for cels. That’s where we assign severity, occurrence, and detectability rankings to each mode and deliberately decide which preservation controls are worth the effort.

The preservation tiers later in this article are essentially the “control strategies” flowing out of that FMEA: they show how different storage choices either mitigate or amplify specific failure modes.

Failure Modes (FMEA): How & Why Cels degrade

With the functional failures and failure modes on the table, we can lay out a high-level Failure Modes and Effects Analysis (FMEA). This isn’t meant to be a line-by-line lab worksheet; it’s a map of which physics dominate cel lifetime and how our storage choices interact with them.

The table links each failure mode (FM1–FM7) to:


  • Primary Effect – what actually happens to the cel when this mode dominates.
  • Failure Type – whether the mode is hidden, evident, predictable, or dormant in the RCM sense (can you see it early, or only after damage?).
  • Detection – practical ways a collector or conservator can observe or infer that mode (AD strips, odor, RH logs, visual inspection, mass change, etc.).
  • Key Controls – the environmental and design levers that keep that mode benign or push it into very low likelihood.


Later sections turn these controls into likelihood vectors for chemical and structural aging, and then into preservation tiers. In other words, if a tier claims to be “Tier 1,” it is really claiming that the key controls for all of these FMs are in place and, where required, verified.

# Failure Mode Primary Effect Failure Type Detection Key Controls
FM1 Intrinsic hydrolytic chain scission (normal aging) Slow increase in internal acidity; long-term drift toward FF2. Dormant / Predictable AD strip drift (long horizon), pH estimation model, elevated emission baseline. Temperature reduction, absolute humidity control, stable environments without extreme peaks.
FM2 Autocatalytic hydrolysis (runaway VS) Rapid onset of VS; warping, odor, severe structural/visual loss (FF1–FF4). Hidden → Evident (late) Sudden odor spike, strong AD strip response, rapid AD shift in sealed tests. Limit internal acidity buildup, enable emission, add local scavengers, keep temperature and moisture low.
FM3 Vapor-phase feedback / emission bottleneck AA accumulates near the cel, decreasing net diffusion and increasing reabsorption. Hidden AD strip stays elevated despite “good” storage; scavenger mass uptake indicates boundary-layer load; gradients by strip location. Engineered air paths, local sinks (scavengers), avoid tiny sealed volumes without sinks, avoid blocked geometries.
FM4 UV-induced photooxidation Color fading, binder oxidation, and pigment/binder damage driving long-term visual loss (FF3). Evident (with lag) Visual inspection, baseline photos under consistent lighting, color shift monitoring, spectro/colorimetry (optional). UV-filtered glazing, remove direct light, limit high-intensity display duration, rotate display time.
FM5 RH cycling / dimensional fatigue Curling, cockling, shrinkage, and accumulated stress cracking (FF4), amplified by paint constraint and late-life embrittlement. Hidden → Evident Environmental trend logs (RH swings), geometric distortion trend, edge lift, seasonal drift signatures. Buffer RH, avoid humidity swings, use microclimates, maintain stable temperature/RH, avoid repeated open-air exposure cycles.
FM6 Sustained compression and creep (static load) Long-duration loading from stacking, tight framing, overfilled binders, gravity loading, or prolonged contact pressure causing viscoelastic creep, set-in deformation, thickness loss, and optical distortion (FF4; often manifests as FF3). Hidden → Evident (time-dependent) Imprint/set patterns after storage, subtle plane deviation, texture/gloss change over months/years; deformation persists after “rest.” Limit local pressure (P = F/A), avoid tight stacks, add intentional spacers/air gaps, store valuable cels near top, avoid clamping/ridge contact at paint islands, control dwell time.
FM7 Enclosure contamination / material incompatibility Added acid/solvent/plasticizer load accelerates FF2/FF3; can also increase sticking/blocking risk and local haze/film deposits. Hidden Localized AD strip spike, odor within enclosure, residue/haze, material pH tests, supplier specs, “new foam/plastic smell.” Use archival-tested materials; avoid PVC/unknown foams; isolate questionable materials with barrier films; pre-condition/air-out components; keep scavengers away from direct contact surfaces.
FM8 Event-based mechanical damage (handling / assembly / shock) Creases, scratches, edge nicks, punctures, paint scuffs, clamp marks, or shear from mishandling (FF5; sometimes FF3/FF4 if layers stick or shear). Evident (usually immediate) Immediate visual inspection; compare before/after handling, reframing, or transport. Rigid supports, safe sleeves, controlled handling procedure, avoid over-tight framing points, avoid sliding contact, use spacers and controlled clearances.

FM5 covers slow geometric fatigue from RH cycling; FM6 covers time-dependent deformation from sustained load (creep/compression); FM8 covers discrete event-based handling/assembly damage. FM7 captures non-UV, enclosure-contamination risk from incompatible materials that can accelerate chemical or visual failure even without severe vinegar syndrome.

From FMEA to a Risk Framework

The FMEA tells us which failure modes exist and what controls keep each one benign. What it does not tell us is how these modes combine into an overall preservation risk. To move from a list of failure modes to a usable decision tool, we need a framework that separates two different questions:


  • Likelihood: How fast will the cel move toward the effects of a given FM under the current environment?
  • Consequence: If degradation progresses, how severe is the resulting loss?


This separation is essential. Some environments accelerate aging but do not immediately change the cel’s condition (high likelihood, low current consequence). Others do little harm chemically but allow mechanical damage to occur suddenly (low likelihood, high consequence). Treating likelihood and consequence independently lets us apply the same analytic structure used in aerospace and conservation engineering:

risk = f(likelihood, consequence).


The next section defines the likelihood and consequence scales and shows how each failure mode contributes to them. These two axes form the backbone of the risk framework that the preservation tiers are ultimately based on.

Degradation Progression – Chemical vs. Structural Tracks

With the FMEA complete and the risk framework defined, we now need to connect those failure modes to how degradation actually unfolds over time. Vinegar syndrome progresses along two interdependent tracks—a chemical lifetime track and a structural fatigue track—and each stage maps directly to the shared C₁–C₅ consequence scale. This progression is what anchors the later methodology for assigning likelihood (L) and consequence (C) in real-world storage environments.

Stage Chemical Lifetime Track
(Lchem)		 Structural / Mechanical Track
(Lstruct)		 Typical FMs / Consequences
1 — Stable Lchem: 1 (Very Low)
CTA near original condition; internal acidity at baseline. Hydrolysis generation rate (Q) is very low and diffusion easily keeps pace. No detectable odor; AD strips remain blue. 		Lstruct: 1 (Very Low)
Minimal RH cycling; film flat and compliant with no observable curl, cockle, or strain patterns. 		Dominant FM: FM1 (normal aging).
Consequence: C1–C2 (cosmetic or negligible in a 50-year horizon).
2 — Early Drift Lchem: 2 (Low)
Hydrolysis begins to outpace baseline; Q increases with temperature and absolute humidity. Internal pH slowly declines. Odor may be faint or intermittent. 		Lstruct: 2 (Low)
Early fatigue from repetitive RH swings produces micro-strain, but geometry remains visually normal. 		FMs: FM1 with emerging FM5.
Typical consequence: C2 (early chemical risk; mild structural impact).
3 — Early VS / Dimensional Instability Lchem: 3 (Neutral)
Detectable acetic acid odor; AD strips shift from blue toward green. Boundary layers can accumulate acid when emission (E) is weaker than generation (Q). VS onset is now measurable. 		Lstruct: 3 (Neutral)
Noticeable curl or cockling, especially under cycling or uneven moisture exposure. 		FMs: FM1, FM3 (vapor feedback), FM5.
Consequence: C2–C3 (visible degradation; early-stage reversible).
4 — Runaway & Structural Stress Lchem: 4 (High)
Autocatalytic VS: Q » internal diffusion. Acid accumulates in the film faster than it can migrate out. Poor ventilation or a saturated cavity accelerates internal breakdown. 		Lstruct: 4 (High)
Strong curl, shrinkage, blocking, paint instability, and cracking as mechanical strain exceeds the film’s tolerance. 		FMs: FM2 (runaway VS), FM3, FM4/FF3 (UV-linked), FM5, FM7.
Consequence: C3–C4 (major loss of function and appearance).
5 — End-of-Life Lchem: 5 (Very High)
Polymer is heavily deacetylated; internal acidity remains high even with strong vapor control. Chemical stability cannot be restored. 		Lstruct: 5 (Very High)
Film embrittlement, cracking, paint detachment, and possible image loss. Handling causes further damage. 		FMs: FM2, FM4, FM7 dominate.
Consequence: C4–C5 (severe / near-total loss).

In design terms, preservation work aims to keep the cel in Stage 1–2 for as long as possible by reducing chemical likelihood (lowering Q, maintaining a strong emission gradient E) and avoiding structural cycling that increases fatigue. When we assign a tier, we can now say both “this environment keeps the chemical track in Lchem ≈ 1–2” and “it limits structural loading to Lstruct ≈ 1–2,” and then read the resulting consequence vector directly from the shared C1–C5 scale.


Risk Framework

In the sections above we defined the functional requirements of a preservation system, the functional failures, the chemical and mechanical failure modes, and the shared consequence scale (C1–C5). We also mapped how degradation progresses across two tracks: a chemical lifetime track and a structural fatigue track. With these foundations in place, we can now define what risk means for animation cels.


Risk = Likelihood (L) × Consequence (C)


The likelihood component reflects how strongly an environment accelerates the relevant failure modes—primarily hydrolysis (VS) and structural fatigue—based on temperature, absolute humidity, handling, diffusion geometry, and microenvironment behavior. The consequence component reflects how far along the degradation progression a cel could move if exposed to that stress. This gives a traceable and physically grounded way to compare different storage and display environments: not by opinion, but by the physics driving chemical and mechanical deterioration.

5. Risk Scoring Logic – From Failure Modes to Tiered Risk

Risk Framework Logic Flow

CTA Preservation Framework Logic Flow Diagram

The framework evaluates a unified set of failure modes (FM1–FM8) across three independent likelihood vectors: chemical, structural, and UV/light. Each vector uses the same L1–L5 ladder, and all three map into the shared C1–C5 consequence scale.


  • Lchem — chemically driven progression (temperature, absolute humidity, acid feedback, diffusion geometry);
  • Lstruct — mechanically driven progression (RH cycling amplitude, dimensional fatigue, handling stress, enclosure mechanics);
  • Luv — UV/photooxidative progression (UV dose, spectrum, display duration);
  • C — resulting consequence for that FM (shared C1–C5 ladder).


Although all vectors use the same scale, the chemical vector is treated differently: Lchem is not subjectively assigned. It is calculated from Arrhenius temperature dependence, absolute humidity thresholds, and feedback accelerants such as boundary-layer stagnation. The derivation is detailed in the Appendices and forms the backbone of the Risk calculators.


The structural and UV vectors remain qualitative because their dominant drivers (RH cycling and UV dose) currently lack validated quantitative lifetime models. Maintaining a unified consequence ladder allows all three vectors to combine coherently.

Rather than collapsing these independent drivers into a single universal likelihood, the framework evaluates likelihood on a per–failure-mode basis. Each failure mode is governed by one or more dominant physical vectors (chemical, structural, or UV), and only those applicable vectors are used when assessing likelihood for that mode.


For each failure mode FMi, likelihood is therefore defined as:

LFMi = max( Lchem,i, Lstruct,i, Luv,i )

where the maximum is taken only over the vectors that physically drive that failure mode. For example, UV exposure is irrelevant to compression creep, and structural loading is irrelevant to intrinsic hydrolysis. This avoids artificial averaging between unrelated mechanisms and ensures the dominant driver governs the assessed risk.


Risk for each failure mode is then expressed as:

RiskFMi = LFMi × CFMi

The framework output is therefore a ranked risk register (FM1–FM8), rather than a single blended score. This reflects the physical reality that different environments tend to fail cels in different ways, and that preservation decisions should prioritize the failure mode most likely to dominate near-term or long-term survivability.

Important Rules and Clarifications

1. Dominant Vector Rule
For each failure mode, likelihood is governed by its dominant physical driver. Increasing control of non-dominant vectors cannot compensate for poor control of the dominant one. For example, excellent UV protection does not mitigate sustained compression creep, and stable geometry does not mitigate runaway autocatalytic hydrolysis.

2. Chemistry Priority Clarification
Chemical degradation in acetate is unavoidable unless actively suppressed through temperature and moisture control. As a result, chemical likelihood dominates FM1–FM3 and strongly influences long-term survivability even when structural and UV stresses are low. Structural and UV-driven failure modes, by contrast, are contingent on handling, display choices, and enclosure design.

3. Dark-Storage Interpretation
In environments where UV exposure is negligible, Luv is not considered for any failure mode. This does not reduce conservatism; it simply reflects that UV-driven mechanisms are inactive in dark storage and should not be allowed to dilute assessment of chemical or structural risk.


The preservation tiers then combine these vector-level results into a multi-vector risk profile. Tier 1 environments keep all vectors in low-likelihood, low-consequence bands. Higher tiers permit one or more vectors to rise into moderate or high risk. Using the shared consequence scale ensures that chemical, structural, and UV pathways remain directly comparable across all scenarios.

1. Functional Failures & Risk Vector Groups

We retain the same functional failures (FFs) defined in the main RCM framework. These represent the observable end states that matter from a preservation and usability standpoint:

  • FF1 – Structural failure: base film embrittlement, cracking, tearing.
  • FF2 – Chemical failure: active vinegar syndrome; high internal acidity and strong odor.
  • FF3 – Visual failure: paint loss, blocking, haze, gloss nonuniformity, or surface damage.
  • FF4 – Geometric failure: cupping, rippling, shrinkage, set-in deformation, stuck stacks.


The detailed failure modes from the FMEA do not contribute equally to every type of risk. Instead, they cluster into a small number of dominant physical risk vectors, each governed by different inputs and control strategies. These vectors are used to evaluate likelihood per failure mode, not to produce a single blended score.

Risk Vector Group Included Failure Modes (FM) Primarily Drives Dominant Inputs
Chemical / VS Progression FM1 – intrinsic hydrolysis (normal aging);
FM2 – autocatalytic hydrolysis (runaway VS);
FM3 – vapor-phase feedback / emission bottleneck 		FF2 (chemical), FF4 (geometric);
indirectly FF1 / FF3 at late stages 		Temperature, absolute humidity, internal acidity, diffusion geometry, boundary-layer behavior
Photochemical / Light Exposure FM4 – UV-induced photooxidation FF3 (visual) UV dose, spectrum, intensity, cumulative display duration
Structural / Mechanical Evolution FM5 – RH cycling and dimensional fatigue;
FM6 – sustained compression and creep;
FM7 – enclosure contamination / material incompatibility;
FM8 – event-based mechanical damage 		FF4 (geometric), FF3 (visual), FF1 (structural) RH cycling amplitude, static load and dwell time, plasticizer state, enclosure mechanics, handling and assembly practices

In later tools and calculators, environments are therefore described using a vectorized risk profile (e.g., chemical risk, structural risk, UV risk), rather than a single “good/bad” score. Preservation decisions prioritize the dominant failure mode for a given storage or display configuration.

2. Consequence Scale (C) – How Bad Is the Outcome?

Consequence answers: “If this vector drives a failure, how bad is it for the cel?” This is tied to the functional failures above and is used consistently across environments and vectors.

C Score Impact Description Examples in Cels
1 – Cosmetic Slight change, no real loss of function or life expectancy. Light board scuffs, tiny sleeve marks, mat burn on a non-image border.
2 – Local Localized damage; noticeable to you but cel still broadly stable. Small area of stuck paint, minor curl that relaxes in storage, slight localized haze.
3 – Significant Noticeable loss of function/aesthetics; intervention recommended. Moderate vinegar odor, visible cockling, early blocking, paint lift in important areas.
4 – Major Large impact on one or more functions; remaining life substantially shortened. Strong VS with warping, widespread blocking, major shrinkage or repeated stuck-layer separation.
5 – Critical System-level failure; the cel is chemically unstable in normal storage and functionally at end-of-life. Deacetylation cannot be reversed, but in a professional context the image may still be stabilised by moving to deep-frozen storage (e.g. < -20 °C). Severe VS, strong vinegar odor, brittle and warped base, blocking/sticking between layers, widespread paint loss or cracking through the image. In collector settings this is effectively “total loss”; in an institutional freezer it may be held in a “no further loss, but not restored” state.

This consequence scale applies across all degradation vectors. For any environment we can score chemical and mechanical outcomes separately using the same C1–C5 ladder. When we later show a “consequence vector” such as (C₂, C₃), it simply means: C₂ for the chemical lifetime track and C₃ for the structural/mechanical track, both using this shared scale.

3. Environment-Relative Likelihood (L) by Vector

Likelihood is always scored relative to the 68 °F / 50% RH baseline over a 50-year horizon. Each vector gets its own L-scale because they respond differently to the same environment.


3.1 VS Onset Likelihood (FM1 + FM3 Drivers) – Lchem,env

This captures the inherent aggressiveness of the storage environment for a stable CTA cel. It is highly sensitive to temperature and absolute humidity, and it is adjusted for microenvironment behavior (vapor paths, stagnant sealing, scavengers).

Scope: likelihood here means likelihood of VS onset. Runaway (FM2, autocatalytic hydrolysis) is deliberately not scored as a separate likelihood vector — a cel in runaway has already crossed onset, so its "likelihood" is no longer a meaningful question; the questions that matter at that point are consequence, rate, and intervention, which are handled quantitatively by the Engineering Framework's kinetic model (the D / Pint two-track equations). For this framework's purposes, Lchem,env is the single likelihood score for the chemical track.

LVS VS Progression vs Baseline (50 yrs) Heuristic
1 – Very Low Much slower than baseline; VS onset effectively beyond 50 yrs for stable cels. Deep-cold, buffered microclimate; AD strips stay blue; reset cels remain stable when managed correctly.
2 – Low Slower than baseline; VS still possible but clearly delayed on a 50-year horizon. Cool, dry rooms or mild cold storage with controlled RH.
3 – Neutral Comparable to baseline institutional storage. 68 °F / 50% RH, modest RH swings, no microclimates.
4 – High Faster than baseline; clear VS progression expected within 50 yrs. Warm, humid, or RH-cycling environments; poor vapor paths; no scavengers.
5 – Very High Much faster than baseline; obvious VS on 5–15 yr timescales. Hot + humid storage (attic / garage / unit) or sealed, unbuffered boxes in those spaces.


Note on independence of L and C. Although a severely degraded cel (high C) can influence its own microenvironment (for example by releasing more acid or trapping vapor in warped geometry), the L-scale in this framework is defined as the inherent aggressiveness of the environment itself. Keeping L independent of C makes tiers comparable between collectors and allows us to specify environments in a clear, testable way.

Any state-driven feedback (diffusion bottlenecks, reduced scavenger effectiveness, etc.) is handled via the correction factor covered directly below, not by changing the L-table. In other words, the environment sets the base L; the cel’s condition and feedback effects are accounted for when we adjust that base L and then combine it with C in Risk = L × C.

Microenvironment Correction Factor

FM3 effects reduce to one question: does acetic acid vapor (Pint) accumulate near the cel? Pint is the accelerator in the chemical track — everything about enclosure geometry and scavenging matters only insofar as it lets vapor build up or escape. A single factor, Menvironment, captures the combined enclosure + scavenger state.

Why one combined factor, not two multiplied

Enclosure geometry and scavenger presence aren't independent risks — both are proxies for the same thing: whether acid vapor (Pint) builds up near the cel. The nominal baseline already assumes Pint ≈ 0 (open, non-trapping). So:

  • Enclosure can only push likelihood worse, never better than nominal — open storage is what nominal already assumes; only a sealed/trapping enclosure deviates, and only upward.
  • A scavenger can only push back toward nominal, never below it — in open storage it has nothing to remove; in an enclosure an effective one offsets the trapping penalty, capped at nominal (we don't credit benefit beyond "cancels the enclosure").
  • A saturated or absent scavenger is neutral, not punished twice — it just fails to offset the enclosure penalty.

Menvironment — Combined Enclosure + Scavenger State

Enclosure Scavenger Menvironment Rationale
Open / vented (well-ventilated space, no stagnant dead zones) Any (irrelevant) 1.0 No assumed Pint buildup — this is the nominal case, whether or not a scavenger is present
Enclosed, not airtight (binder, box, sleeve without active venting) Effective (fresh/appropriately sized, not saturated) 1.0 Scavenger offsets the assumed buildup — capped at nominal, not credited below it
Enclosed, not airtight (binder, box, sleeve without active venting) None / saturated / undersized 1.2 Moderate assumed Pint buildup, nothing offsetting it
Sealed / airtight (frame, vacuum bag, no path to room air) Effective (fresh/appropriately sized, not saturated) 1.0 Scavenger offsets even a fully sealed enclosure — same conservative cap at nominal
Sealed / airtight (frame, vacuum bag, no path to room air) None / saturated / undersized 1.5 Full assumed Pint buildup; acid effectively recirculates with no removal path

This assumes scavenger efficiency doesn't degrade as it saturates — a real scavenger's actual removal rate trails off before it's visibly "used up," which this qualitative table doesn't capture. Treat the "effective" row as optimistic if you don't have a recent replacement or capacity check.

Implementation note — VS Time-to-Onset Estimator. The estimator applies this same cap-at-nominal logic quantitatively. Its storage-change projection adjusts the cel's measured Pint by a single combined storage multiplier, clamped at nominal from below (never less than 1.0) — so a scavenger can offset an enclosure penalty back to the as-is Pint but never below it, and earns no credit in open storage. This mirrors the Menvironment lookup above: one combined enclosure-plus-scavenger state, not two separately multiplied factors. The one deliberate exception is the estimator's "Strong" option, which invokes the Engineering Framework's Perfect Sink approximation (Pint → 0) as a theoretical lower bound on the timeline, not as a claim about achievable storage. The estimator's pristine-cel benchmark is likewise calibrated to this framework's baseline: R0(68 °F) is set so that at 68 °F / 50% RH — the IPI reference / Tier 3 / L = 3 neutral condition — a stable cel reaches DVS = 0.5% in exactly 50 years, matching the IPI reference.

A Note on Heavy Paint (Not Scored as a Factor)

Heavy, impermeable paint layers (thick acrylic, vinyl fills, multi-layer builds) act as vapor barriers on the painted side of the cel. This adds diffusion resistance to the cel's own emission pathway — acid generated inside the substrate has fewer routes to the surface-to-air interface, which raises internal accumulation independent of enclosure conditions. For simplicity, this framework does not score paint coverage as a correction factor; it is noted here as a qualitative risk amplifier, and its transport physics are handled in the Engineering Framework.

Practical mitigation — use the uncoated side. Production cels are painted on one side only, so even a heavily painted cel always retains a clear emission face. Placing the scavenger (or interleaving sheet with scavenging capacity) against the unpainted side preserves an effective escape path regardless of paint load. One caution carries over from the transport physics: a heavily painted cel can under-report on A-D strips ("warped but odorless" — internal accumulation with a muted external vapor signature), so don't let a clean strip reading on a heavily painted, visibly distorted cel override the physical indicators.

3.2 UV / Light Vector (FM4) – LUV

UV risk is driven by dose × time, not humidity. So we anchor its scale to display behavior.

LUV Relative UV Risk (50 yrs) Typical Pattern
1 – Very Low Essentially dark storage; UV events are rare and filtered. Stored in the dark; brief viewing under low-intensity, UV-filtered light.
2 – Low Mostly dark with short, controlled display windows. Framed but displayed a few weeks per year; UV glass; away from windows.
3 – Neutral Routine indoor lighting without direct sun. Wall display in a typical room; no UV glass but also no strong sunlight beam.
4 – High Frequent bright display and/or partial sun exposure. Near windows, spotlighted gallery lighting for long periods.
5 – Very High Direct or near-direct sun repeatedly hits the art. In front of unfiltered windows; strong daylight across the piece for hours per day.

LUV quantifies the likelihood of light-driven photochemical degradation to both paint and cellulose acetate. Unlike other failure modes, UV damage is governed by dose × time, not temperature, humidity, or mechanical loading.

UV exposure causes:

  • Paint fading and pigment breakdown
  • Yellowing and embrittlement of CTA
  • Loss of visual contrast and surface integrity
Key distinction: UV damage is cumulative and irreversible. Unlike RH or mechanical stress, there is no recovery once photochemical bonds are broken.
Why UV is scored qualitatively, and what it does/doesn't capture

In theory, UV risk could be calculated from wavelength, intensity, exposure time, and material sensitivity. In practice, collectors rarely know:

  • Actual UV content of room lighting
  • Lux levels at the cel surface
  • Total cumulative exposure over years or decades

For this reason, LUV is intentionally behavior-based, anchored to how and where the cel is displayed rather than attempting false precision.

Framework philosophy: UV is easier to control than chemistry or mechanics — so the scoring emphasizes preventability, not modeling complexity.

What LUV captures:

  • Captured: cumulative risk from routine display behavior
  • Captured: presence or absence of UV filtering
  • Captured: proximity to windows and daylight paths
  • Not captured: RH effects (handled in FM5)
  • Not captured: chemical degradation (FM1 / FM2)
  • Not captured: mechanical stress or compression (FM5 / FM8)

A cel may score low on UV risk while still being at high mechanical or chemical risk. UV is a separate vector and must be evaluated independently.

Scoring Guidance

Assign LUV based on the most representative long-term display condition, not occasional short-term handling.

  • LUV = 1–2: Storage-dominant, rare viewing, UV-filtered light
  • LUV = 3: Routine indoor display without direct sun
  • LUV = 4–5: Near windows, frequent bright display, or sun exposure
Conservative rule: If you are unsure whether direct or indirect sunlight reaches the cel, score one level higher.

For a deeper discussion of photochemical mechanisms, UV filtering limits, and display tradeoffs, see:
→ Blog: UV Exposure, Display Risk, and Why “UV Glass” Is Not a Cure-All

3.3 RH Cycling and Dimensional Fatigue (FM5)

FM5 covers accumulated deformation from repeated moisture uptake and loss cycles (expansion/contraction of acetate and paint). It is structurally dominated and maps heavily into FF4 (geometric) and sometimes FF3 (visual).

Lstruct,geom quantifies the likelihood of time-dependent geometric damage (bowing, curl, edge stress, corner creep, handling-induced bending) driven by two coupled inputs: (1) support geometry / constraint and (2) repeated water-activity (RH) cycling. This is a fatigue-style failure mode: damage accumulates from many cycles, even if no single cycle looks extreme.

This score does not include paint flattening or pressure imprinting (that is handled separately in Lstruct,comp).

Why RH cycling matters mechanically, and the physics behind it

What “RH cycling” actually means (and why “average RH” is not enough)

Collectors often think of RH as a single value (“I store at 50%”). In reality, the cel experiences a band of RH over time — and the cycling amplitude (ΔRH) is what drives mechanical fatigue. The cel is a composite: CTA base + paint islands. When RH changes, CTA absorbs/desorbs moisture and wants to swell/shrink. Paint responds differently (different uptake, stiffness, and expansion). That mismatch creates interfacial shear and bending moments.

Differential strain is the driver, not generic “fatigue.”

Paint and cellulose acetate do not expand/contract the same way. They have different coefficients of thermal expansion (CTE) and hygroscopic expansion (CHE). During temperature or humidity cycling, the substrate and paint attempt to move by different amounts, generating interfacial shear stress that can lead to paint lift, cracking, and flaking.

A simplified strain-mismatch expression is: Δε ≈ ΔT·(αfilm − αpaint) + Δaw·(βfilm − βpaint)

This is why cycling environments (even “cold cycling”) can be mechanically harsher than a stable room: cycling increases ΔT and Δaw, increasing interfacial stress. In tier terms: stability can reduce Lstruct even when average temperature is not minimal.

Why RH shifts during temperature drift (even if you didn’t “add moisture”)

A temperature drift can create RH cycling even when the amount of water in the air (absolute humidity) is constant, because saturation vapor pressure changes strongly with temperature:

One-line physics:
RH = e / esat(T)
where e is actual water vapor partial pressure and esat(T) rises with temperature.

Practical translation: the cel can see RH cycling from (a) HVAC drift, (b) fridge compressor cycling / door openings, and (c) absolute humidity “events” (cooking/shower) — even if your thermostat or setpoint looks “stable.”

Why this matters mechanically (failure mechanisms)

  • Interfacial fatigue: repeated mismatch strain drives micro-slip, microcracking, and eventual edge lift/delamination.
  • Bending / curl: transient moisture gradients through thickness create bending moments (especially with non-uniform paint).
  • Creep-set: moisture and acidity can soften layers, making deformation more time-dependent and permanent.
  • Geometry multiplier: tight mats/frames, constrained sleeves, or poor support converts “free strain” into stress.
Read the full physics + worked examples: RH Cycling, Vapor Pressure, and “Microclimate Fatigue” (examples, vapor-pressure tables, and interpretation)

Likelihood scale (ΔRH bands + examples)

Lstruct Relative Structural / Mechanical Risk (50 yrs) Examples (handling + RH cycling)
1 – Very Low RH cycling ΔRH < 5% per typical cycle (much tighter than a good room). Microclimate buffered so the effective aw band is extremely narrow (engineered small volume, buffered cold box, stable enclosure).
2 – Low RH cycling 5% ≤ ΔRH ≤ ~10%, slightly better than or equal to a best-case room. Room or microclimate held near a stable setpoint with modest drift; generally as good as a well-behaved 68 °F / 50% room.
3 – Neutral RH cycling ≈10–12% per typical cycle (reference case). Typical interior drift around 70 °F with constant AH (baseline “survivable” historical storage).
4 – High RH cycling ~12–20%, but still without sustained condensation. Large day–night swings, seasonal spikes, repeated excursions above ~60% or below ~30% RH, yet not routinely hitting saturation.
5 – Very High RH cycling > 20%, including condensation or very low RH episodes. Environments that push aw toward 1.0 (condensation, damp basements) or very low for long periods, combined with repeated T/RH shocks (unbuffered fridge cycling, attic/garage extremes).

Scoring note: Use Lstruct,geom when the failure mode is bowing/warping/curling, corner stress, edge creep, or “frame as a clamp.” Use Lstruct,comp when the failure mode is paint flattening, texture transfer, blocking patterns, or imprinting from localized contact pressure.

3.4 Sustained Compression and Creep, Static Load (FM6)

FM6 covers time-dependent deformation under sustained load: stacking weight, tight frames, overfilled binders, clamp points, gravity loading, and prolonged contact pressure. This is structurally dominated and strongly dependent on dwell time and temperature.

Lstruct,comp Pressure / Dwell Interpretation
1 — Very Low Green zone: P ≤ 2 psi at contact features, or effectively near-zero contact (large flat area, strong spacers, or no stacking). Multi-year dwell is unlikely to imprint or deform paint.
2 — Low Low Yellow: 2 < P ≤ 3 psi. Generally safe, but cumulative risk exists for thick or soft paint, or chemically weakened (VS-active) cels under long dwell.
3 — Neutral Upper Yellow: 3 < P ≤ 5 psi. Time-dependent risk band. Multi-month to multi-year dwell can contribute to subtle flattening, blocking patterns, or texture transfer.
4 — High Red onset: 5 < P ≤ 7 psi. Not instant failure, but undesirable as a long-term resting load, especially for raised linework, heavy paint islands, or any cel with active VS or surface stickiness.
5 — Very High Deep Red: P > 7 psi and/or repeated long dwell under load. High likelihood of permanent deformation or imprinting within years (or sooner if paint is softened).

Lstruct,comp quantifies the likelihood of time-dependent mechanical damage to paint and surface features caused by sustained localized contact pressure. This includes raised-paint flattening, texture transfer, blocking patterns, and pressure-driven creep — failure modes that are not captured by temperature, humidity, or chemical aging models.

Unlike bending or RH-cycling damage (handled under Lstruct,geom), compression damage is driven by the physics of pressure × time, where relatively modest forces become damaging when concentrated over very small contact areas (paint ridges, islands, stuck regions) and held for months or years.

Key distinction: Lstruct,comp is governed by local pressure at the contact patch (P = F / A), not by the total number of pages, binder thickness, or overall stack height.
Why this matters, and how the pressure bands are derived

Compression damage is often hidden and time-dependent. A cel may appear stable for years before subtle texture loss or blocking becomes visible — at which point the deformation is permanent. Because this failure mode is strongly amplified by chemical softening (e.g., vinegar syndrome), it frequently acts as a multiplier on other risks rather than an isolated issue.

Lstruct,comp is evaluated using a conservative, physics-based approach:

  • Estimate load (F): mass of pages and binder components acting on a given cel position.
  • Assume a worst-case contact area (A): small paint ridges or islands (typically a few cm²), not full-surface contact.
  • Compute local pressure: P = F / A.
  • Account for dwell time: pressures that are harmless for minutes of handling can cause damage under multi-year storage.
  • Map pressure bands to likelihood: Green / Yellow / Red zones correspond to increasing probability of long-term deformation.

The table below expresses these pressure bands directly as Lstruct,comp levels (1–5), enabling consistent scoring across storage formats (binders, stacks, portfolios, boxes).

Pressure bands use left-open, right-closed intervals to eliminate overlap; borderline cases are conservatively assigned to the higher band. They target multi-decade storage, not short-term handling — flat, low-relief cels see lower true risk than shown; raised or chemically softened paint sees higher.

Detailed Derivation & Examples

Full derivations, binder measurements, and worked examples — including vertical vs. horizontal storage tradeoffs — are documented separately: Vertical vs. Horizontal Cel Storage: Pressure, Diffusion, and Design. This section defines what is scored and how; the linked article documents why the bands were chosen.

3.5 Enclosure Contamination / Material Incompatibility (FM7)

FM7 captures chemical loading introduced by the enclosure itself: acids, solvents, plasticizers, or reactive volatiles emitted from boards, foams, adhesives, plastics, or coatings. This failure mode is typically chemically dominant, acting as a source term that elevates local acidity, interferes with emission or scavenging, and accelerates other chemical or mechanical failure modes.

Likelihood Enclosure Condition Interpretation
Unlikely All enclosure materials are known-compatible or tested archival (e.g., PET/Mylar, inert plastics, buffered boards, metals, glass). The enclosure does not contribute meaningful chemical loading. FM7 can be neglected as a primary risk driver.
Likely One or more enclosure components are chemically active, untested, or known-incompatible (e.g., PVC, rubber foams, unknown plastics, pressure-sensitive adhesives, acidic boards). The enclosure is a probable source of volatile acidity or solvent load. FM7 should be treated as an active chemical risk contributor.

FM7 is scored categorically rather than on a continuous scale. If incompatible or unverified enclosure materials are present, treat FM7 as Likely. If all materials are known-compatible or tested archival, treat FM7 as Unlikely.

FM7 does not receive its own numeric L-score. Instead, it acts as a modifier to chemical and mechanical likelihoods (e.g., elevating Lchem or accelerating FM8 when present). In the risk vector table (Section 4), it appears as Lmatl — a categorical flag (Likely / Unlikely) rather than a 1–5 numeric score. When FM7 is flagged Likely, apply a +1 modifier to Lchem,env and note elevated FM8 probability in the overall risk assessment.

Class Examples FM7 Logic Practical Note
Known-Compatible Uncoated PET (Mylar), Polypropylene (PP), Buffered Rag Board. Chemically Inert. No off-gassing. Gold Standard. Best for chemical isolation. Manage sticking risk via low-pressure stacking.
Conditional "Acid-free" wood pulp, Glassine, unknown archival foams. Potential for Acid Reversion. Glassine is "non-stick" but can become acidic and hygroscopic over time. Use with caution.
Incompatible PVC (Vinyl), Rubber, Cardboard, Pressure-sensitive tapes. Active Hazard. HCl & Plasticizers. Causes "Vinyl burn" and melts paint regardless of pressure. Remove immediately.
The "Pressure vs. Chemistry" Rule

Sticking (adhesion) is primarily a function of pressure, heat, and humidity. Using a "safe" material like Mylar does not cause sticking in a well-managed environment. Conversely, a "non-stick" material that is chemically active (like PVC or cheap paper) will destroy a cel even if it never touches the paint.

Fast Triage Rules
  • The Sniff Test: "New car" or "Pool toy" smell = PVC (Immediate Removal).
  • Isolation: If using "Conditional" materials, use a Mylar barrier between the cel and the material.
  • Stack Height: Keep cel stacks thin to prevent the mechanical adhesion that critics of Mylar fear.
Mitigation Philosophy: FM7 is about Exclusion. If a material's chemistry is unknown, it is a risk contributor. The goal is to ensure the enclosure acts as a protector, not a reactant.

3.6 Event-based Mechanical Damage (FM8)

FM8 covers discrete, handling-driven damage — drops, flexing, pinching, over-tightening, mishandling during reframing, or improper storage transitions — rather than slow accumulation over time. This is structurally dominated and driven by peak stress during an event, not by dwell time or average conditions.

Lstruct Relative Structural / Mechanical Risk (50 yrs) Examples (handling behavior)
1 – Very Low Rigid, gentle support; handled rarely. Rigid backing, oversized sleeves, no stack compression; archival handling only a few times per year.
2 – Low Occasional handling with good supports; low frequency of reframing or repositioning. Binder or box with careful page turning; frame not over-tightened.
3 – Neutral Regular handling or mediocre support. Loose binders, soft backing; occasional frame flex; guests handling frames.
4 – High Frequent handling or poor support. Thin sleeves, tight portfolios, frame points digging into stacks, frequent reframing.
5 – Very High Frequent rough handling with no protective margin. Cels stored loose; handled roughly or transported often; frames or sleeves used as physical clamps.
Why event-based damage matters, and how it's scored

Why Event-Based Damage Matters

Event damage is often binary: nothing happens… until something does. A single mishandling event can introduce:

  • Permanent creases or kinks
  • Local delamination or paint cracking
  • Edge stress that later evolves into curl or lift
  • Stress risers that amplify RH cycling or compression damage

Because FM8 damage is often invisible at first, it frequently acts as a risk multiplier for other failure modes rather than a standalone issue.

Methodology Summary (Qualitative)

Lstruct,event is evaluated qualitatively based on handling frequency, support adequacy, and event severity over long time horizons.

  • Frequency: How often the cel is handled, reframed, transported, or accessed
  • Support: Whether handling is rigidly supported or allows flexing
  • Constraint: Whether frames, binders, or sleeves act as clamps or pinch points
  • Margin: Whether clearance and slack exist to absorb mistakes
Engineering intuition: FM8 is governed by peak stress, not averages. Even rare events matter if the geometry concentrates load.

Relationship to Other Structural Likelihoods

  • FM5 (RH cycling): governs fatigue and slow curvature over time
  • FM6 (compression): governs pressure × time damage at contact patches
  • FM8 (events): governs discrete failures and damage initiation

In scoring, FM8 should be elevated whenever handling, reframing, or support transitions are frequent — even if RH and compression risks are otherwise well controlled.

Reference

For a deeper discussion of structural failure modes, support geometry, and how event damage couples with fatigue and compression, see:

Related article:
“Handling, Support Geometry, and Mechanical Failure in Animation Cels”
(Framework reference blog)

4. How Vectors Map Back to Functional Failures

To keep things traceable, we explicitly link each vector to the FFs it mostly affects:

Risk Vector (Likelihood) Included FMs Primarily Drives Dominant Inputs
VS Chemistry
Lchem		 FM1 — intrinsic hydrolysis (normal aging)
FM2 — autocatalytic hydrolysis (runaway VS)
FM3 — vapor-phase feedback / emission bottleneck 		FF2 (chemical), FF4 (geometric) → eventually FF1/FF3 Temperature, absolute humidity, vapor-phase acetic acid, enclosure diffusion/venting geometry, scavenger presence & placement
UV / Light
Luv		 FM4 — UV / photooxidation FF3 (visual) UV dose, high-energy visible, glazing/filtering, display duration, distance to light
RH Cycling / Bending
Lstruct,geom		 FM5 — RH cycling / dimensional fatigue FF4 (geometric), FF1 (structural) RH swing magnitude & frequency, support rigidity/flatness, sleeve fit, frame/binder geometry
Compression / Contact Pressure
Lstruct,comp		 FM6 — compression / creep (time-dependent) FF3 (surface/paint), sometimes FF4 (blocking/adhesion patterns) Local pressure (P = F/A) at paint ridges/islands, dwell time, stack configuration, interleaving/spacers
Material Compatibility
Lmatl		 FM7 — enclosure contamination FF2/FF3 (chemical acceleration + visual/paint impacts) Material chemistry (PVC/plasticizers, rubber sulfur, wood acids), adhesives/inks, VOC load, sealing, proximity/contact
Handling / Accidents
Lstruct,event		 FM8 — mechanical damage (event-based) FF1 / FF3 (scratches, creases, tears, paint loss) Handling frequency, procedure discipline, workspace, pets/kids risk, transport, reframing frequency

The intent is: for any storage setup, you can score each vector independently (Lchem, Luv, Lstruct,geom, Lstruct,event, Lstruct,comp, Lmatl) and pair each with the shared consequence scale (C1–C5).

Example – Same Cel, Different Environments, Different Risk

To see how the framework works in practice, consider a cel that currently reads blue–green on an AD strip (roughly Level 0.5–1). In the progression model, this corresponds to early chemical drift – some acid is present, but the cel is not yet in an autocatalytic runaway regime. On the shared consequence ladder, we can treat this as C = 2 (early VS exposure, no structural distortion).


We keep this starting condition fixed and compare two storage environments. The cel, its AD reading, and its present-day consequence level C = 2 do not change. What changes is the likelihood vectors (Lchem, Lstruct, Luv) that push the cel forward along each lifetime track.


Each vector is scored and carried separately. We do not blend them into a single weighted number, because the three failure modes are physically independent — a cel can sit at low chemical risk while carrying high structural risk. Each vector produces its own L × C score, and the environment's governing risk is the worst (dominant) vector, per the dominant-vector rule.


Environment A — Stable Room (Approx. Tier 3)

  • ~70 °F with daily drift ±2–3 °F
  • RH averages ~50 % with excursions toward 60 %
  • No scavenger; free emission into a large room volume
  • Indoor ambient light, but not direct sun

Chemical (Lchem): Arrhenius inputs are relatively high. Hydrolysis remains active at room temperature. This yields:
Lchem(A) = 3


Structural (Lstruct): Although temperature cycling is similar to a typical 70 °F interior room (±3 °F), this environment also sees RH excursions toward 60 %. That pushes the moisture cycling well beyond the neutral “T-only” fatigue band (≈10–12 % RH swing) and into the range where dimensional stress accumulates measurably over time. Handling remains ordinary but the RH behavior dominates the score.
Lstruct(A) = 4
(RH delta exceeds baseline; moisture variability beyond the neutral band)


UV (Luv): Ordinary indoor lighting with no focused display. No direct UV load.
Luv(A) = 2


Governing likelihood (dominant vector):
Lchem = 3, Lstruct = 4, Luv = 2 → the structural vector dominates, so L(A) = 4.
Note how a blended average (which would land near 3) would have masked the structural risk — exactly what the dominant-vector rule exists to prevent.


C = 2 (fixed starting condition)



Environment B — Controlled Cold Box (Tier 1 Conditions)

  • ~42–48 °F, drift ≤ 1 °F
  • Buffered absolute humidity (Art Sorb); no condensation
  • Pre-conditioned scavenger acting as a one-way sink
  • No cycling, static microclimate; minimal handling
  • Fully dark environment except brief access


Chemical (Lchem): Low temperature plus buffered aw and a sink push chemical progression to the lowest band:
Lchem(B) = 1


Structural (Lstruct): Very tight RH cycling (<5 % equivalent), static mechanics, no handling. This meets the “very low” band:
Lstruct(B) = 1


UV (Luv): Box is dark by design; UV dose is effectively zero:
Luv(B) = 1


Governing likelihood (dominant vector):
Lchem = 1, Lstruct = 1, Luv = 1 → all vectors low, so L(B) = 1.


C = 2 (same starting condition)


Risk Matrix – L × C Score Bands

C = 1 C = 2 C = 3 C = 4 C = 5
L = 1 1 2 3 4 5
L = 2 2 4 6 8 10
L = 3 3 6 9 12 15
L = 4 4 8 12 16 20
L = 5 5 10 15 20 25

Reading the matrix: scores near 1 (bottom-left) represent very low risk over a collector-relevant horizon; scores near 25 (top-right) represent environments where both the likelihood of VS progression and the consequence of that progression are severe.


Risk Comparison – Per-Vector Scoring

Environment Lchem Lstruct Luv Dominant L C (today) Governing Risk = Ldom × C
A – Stable Room (Tier 3) 3 4 2 4 (struct) 2 8
B – Cold Box (Tier 1) 1 1 1 1 2 2

Same cel, same starting condition (C = 2). Each vector is scored on its own; the governing risk is set by the worst vector, not an average. Environment A's structural vector (L = 4) drives a governing risk of 8, while the cold box keeps every vector at 1 for a governing risk of 2. A blended score would have reported A as a 6 and quietly hidden that its real problem is dimensional fatigue, not chemistry.


In this example, the AD level defines the starting consequence only. Likelihood — especially Lchem — is defined by the quantitative VS method in the Engineering Framework and applied via the Lchem,env table and correction factor in Section 3.1.


From Risk Framework to Preservation Tiers

Once likelihood and consequence are defined, we can evaluate any storage environment in terms of how aggressively it drives degradation and how vulnerable the cel is to that degradation. But collectors still need a practical way to answer a simpler, actionable question:

“Given this environment, what level of engineering control am I actually achieving?”

This is where the preservation tiers come in. The tiers translate the risk framework into clear engineering categories by asking:


  • Which failure modes are controlled? (derived directly from the FMEA)
  • Which controls are actually implemented? (temperature, absolute humidity, diffusion access, scavengers, RH stability, geometry)
  • Which controls are verified? (AD strips, sensors, test data, diffusion checks, 0 ppm validation)
  • What likelihood band does the environment fall into? (slow vs. fast aging)

In other words, tiers are not opinions about storage quality. Each tier represents a specific, increasingly stringent set of controls that suppress the relevant failure modes:


  • Tier 5–4: Little or no control; multiple FMs remain active.
  • Tier 3: Basic environmental control, but no strong mitigation of chemical drivers.
  • Tier 2: Engineered cool storage with controls in place for both chemical and structural vectors.
  • Tier 1–0: Same core controls as Tier 2, but with tighter margins and more stringent verification (logging, capacity checks, documented performance).

Tiers 0–2 represent engineering mandates, not just likelihood categories. These tiers are defined by measurable environmental controls rather than subjective scoring:


  • Vault Storage (Tier 0) – Deep Cold Institutional Storage: Film-vault style environments with deep cold, tight RH control, and verification suitable for long-term institutional retention.
  • Tier 1 – Engineered Archival Cold: Cold storage with measured RH stability, sized scavenger capacity, and documented microenvironment performance.
  • Tier 2 – Cool, Controlled Storage: Cooler than room with bounded RH and cycling, suppressing major chemical and mechanical FMs. The configuration is still intentionally engineered; verification and logging may be lighter-weight than the deep archival expectations in Tiers 1/0.

This distinction separates a simple “cold room” or cabinet from a verified, high-performance archival microenvironment.

The preservation tier table that follows is effectively a control specification built from the FMEA and risk framework. If an environment claims to be Vault Storage, 1, or 2, it is really claiming that the controls for all major failure modes are satisfied—and that the measured likelihood of degradation matches the expected band.


The next section defines each tier in detail and shows how typical storage setups— binders, frames, cold boxes, reset containers, and open shelving—map onto them.

Tier T – Temperature
(with typical Lbase)		 aw – Moisture / RH M – Mechanical / Cycling U – Light / UV
Vault Storage
(Static Institutional)		 Deep cold, tightly bounded.
Target ~40–52°F.
Typical krel < 0.25 → Hydrolysis slowed >4× vs Room. 		Buffered aw (45–50% RH).
Mandatory moisture buffer required. High risk of paint embrittlement if RH drops below 30% in this range. 		Static mechanics.
Zero handling permitted. Paint is in a "glassy" state; any flex, vibration, or rapid thermal shift is a fracture risk. aw swings per cycle are held to ≲ 5 % RH. 		Dark Storage.
Stored dark by default. Not intended for active viewing or hobbyist access.
Tier 1
Engineered archival cold		 Cold, tightly bounded.
Target ~53–60 °F, daily drift ≤ ±1 °F with logging, no warm excursions.
Typical krel ≲ 0.5 → Lbase in the ~2–4× band versus 68 °F / 50 %. 		Buffered aw (45–50% RH).
Moisture acts as a critical plasticizer here to keep the paint above its glass transition temperature.
Static mechanics.
No compressor/defrost cycling; minimal handling; no repeated thermal or RH shocks. aw swings per cycle are held to ≲ 5 % RH. 		Dark or well protected.
Stored dark by default. Any viewing uses UV-filtered glazing/sleeves.
Tier 2
Cool, controlled room / cabinet		 Cool, modest-drift.
Approx. 61–67 °F with modest daily swings (≈ ±2–3 °F); not tied to fridge-style compressor cycling, not near heaters or hot exterior walls.
Typical krel ≈ 0.5–0.9 → Lbase in the ~1.1–2× band versus 68 °F / 50 %. 		Buffered aw.
RH held in a safe mid-band at operating T (approx. 45–55 %) with explicit buffering (e.g., Art Sorb). Saturation and very low RH are both avoided.
Gentle cycling (6–10 % RH swing).
HVAC-style drift and slow seasonal changes; aw swings per cycle are bounded to roughly 6–10 %, with no hard hot–cold or wet–dry cycling. Handling is careful and infrequent. 		Light managed.
Not displayed in direct sun; occasional display behind UV-filter glazing or in low light.
Tier 3
Stable room preservation		 Good interior room.
Approx. 68–72 °F with modest daily swings (≈ ±2–3 °F); away from radiators, attics, or exterior hot spots.
Typical krel ≈ 0.9–1.1 → Lbase ≈ 1× (neutral vs 68 °F / 50 %). 		Moderate RH, limited excursions.
Average roughly 40–55 %, with only occasional, short excursions outside this band. No sustained damp (>60 %) or very dry (<30 %) periods. 		Normal household cycling (~≤ 10–12 % RH swing).
Ordinary HVAC cycling and seasonal drift; aw swings are comparable to a good 68 °F / 50 % room, with no deliberate rapid cycling or repeated condensation events. 		Light managed, not engineered.
Not parked in window light or under strong spots; normal indoor lighting for storage, occasional display.
Tier 4
Unbuffered / uncontrolled		 Warm or variable.
Frequently > 72 °F or subject to large daily swings (garage, attic, near exterior walls, poorly controlled rooms).
Typical krel ≳ 1.1 → Lbase < 1× (accelerated decay). 		Humid or swinging RH.
Regular episodes above ~60 % or below ~30 % with no buffering or logging. Seasonal dampness or pronounced dry periods are common. 		Unmanaged cycling (> ~10–12 % RH swing).
Strong day–night temperature changes, intermittent dampness, space heaters, open windows, etc. aw swings often exceed the ~10–12 % RH band of a controlled 68 / 50 room. 		UV risk likely.
Some exposure to daylight or bright artificial sources without UV mitigation (e.g., near windows, under bright LEDs).
Tier 5
High-risk / damaging		 Hot or extreme.
Prolonged high temperatures (hot attics, cars, near heaters) or repeated extremes (e.g., 60 °F → 90 °F cycles).
Typical krel ≫ 1 → Lbase ≪ 1× (VS strongly accelerated). 		Saturated or condensing.
Conditions that drive aw → 1.0 or cause repeated condensation and drying (e.g., sealed box dropped into a fridge, damp basements, leaks). 		Severe mechanical stress (large RH and T shocks).
Fridge/freezer cycling, direct HVAC blasts, frequent hot–cold swings or wet–dry pulses that drive aw swings well beyond ~15–20 % and cause warping, blocking, and delamination. 		Unsafe light exposure.
Direct sun, backlit display, strong spotlights, or long-term display with no UV or intensity limits.

How to read this table: These tiers describe the physical environment only (T, aw, cycling, and light). For a given tier, the temperature band maps to a typical chemistry-only likelihood band Lbase via krel. The final VS likelihood Lchem may be higher or lower than Lbase if microenvironment factors (sealed boxes, boundary-layer stagnation, scavengers, etc.) accelerate or mitigate acid generation and feedback, as defined by the correction factor in Section 3.1.

How the Preservation Tiers Are Calculated

Tier 3 is not arbitrary—it reflects the long-established archival baseline of 68 °F and 50% RH, which is the condition used across film-preservation literature, including IPI research, when reporting earliest vinegar syndrome onset. This becomes the “1× life factor” reference point for all other tiers.

Temperature tiers (Lchem / Lbase) are derived from Arrhenius life-factor calculations using CTA hydrolysis activation energy (~105 kJ/mol). Life factors are grouped as:

  • Tier 0: L > 4× (deep cold institutional storage).
  • Tier 1: L ≈ 2–4× (engineered archival cold).
  • Tier 2: L >1× but <2× (cool, controlled home or cabinet storage).
  • Tier 3: L ≈ 1× (archival baseline at 68 °F / 50% RH).
  • Tier 4–5: L < 1× (accelerated decay relative to baseline).

Tiers 0–2 are, by definition, engineered systems. Achieving meaningful life extension requires controlled temperature, controlled RH, sealed microenvironments, acid-absorption design, and sensing or verification appropriate to the tier. Even Tier 3—the archival baseline—requires a managed environment with basic stability and detection. Tier 4 is where most collectors naturally live, since open-air rooms, consumer frames, Itoyas, and sleeves do not provide engineered environmental control.

Tier 1 and Tier 0 are not practical targets for typical display products. They represent deep-storage preservation—tight RH control and cold temperatures suitable for long-term archival housing, not everyday home display. Preservation products developed under this framework therefore aim for Tier 2 engineered environments with life factors in roughly the 1.3–2× band while operating in a safe, moderate temperature range appropriate for display.

  • Structural Risk Is Quantified (UV Remains Qualitative): Thermal/RH cycling is anchored to a numeric psychrometric model (±3 °F → ~±11 % RH), which defines the objective tier boundaries for mechanical stress. RH swings ≤5 % map to Tier 0–1, 6–10 % to Tier 2, and the typical indoor ~11 % baseline defines Tier 3. These thresholds are fully quantitative within the framework.

    • Water-activity (aw) criteria reflect the minimum moisture availability required for CTA hydrolysis. Because absolute humidity decreases with temperature, colder environments naturally move into lower-reactivity tiers even at the same RH. The lower bound is limited by CTA and paint stability, and ongoing material testing will refine this boundary.

    UV tiers remain qualitative, as its risk is best managed through simple avoidance rather than detailed modeling. The scoring simply reflects relative exposure risk.

    This tier system is not stylistic—it is a structured, engineering-based classification derived from kinetics, diffusion considerations, and archival baselines.

    How to Use the Preservation Tier Table

    The preservation tier table translates the failure modes (FMs) from the FMEA into engineering controls that prevent or slow those failures. Each column in the table corresponds to one or more FMs and specifies the environmental or design requirement needed to keep that FM inactive or low-likelihood. In other words, the table is not describing “nice-to-have” storage features—it is a specification of what must be controlled to keep FM1–FM7 from progressing into their associated failure effects (FF1–FF5).

    Tiers are built by stacking these controls.

    Tier 5 has almost no controls, so multiple FMs can progress freely.
    Tier 4–3 apply partial control to some FMs, but not consistently or not to engineering thresholds.
    Tier 2 controls all major degradation pathways chemically and mechanically, but without full verification.
    Tier 1–0 control all major degradation pathways and require verification (scavenger performance, RH stability, diffusion access, and no local acid accumulation).

    This table should be used as a design and assessment tool: designers can use it to ensure their enclosure or frame actively suppresses each relevant FM, and collectors can use it to evaluate how their setup handles chemistry (Q), diffusion (D), moisture, geometry, and cycling stresses. When a tier is chosen, the associated controls become the minimum engineering requirements needed to maintain that tier’s likelihood band.

    Assumptions & Limitations of This Framework

    This preservation tier framework is based on current scientific and engineering understanding of cellulose acetate hydrolysis, acetic acid diffusion, and environmental stressors. The following assumptions are used when defining tier boundaries:


    • CTA hydrolysis follows Arrhenius behavior within the stable (non-autocatalytic) regime.
    • Tier 3 reflects the archival baseline of 68 °F / 50% RH used in historical vinegar syndrome research.
    • Temperature and RH cycling are evaluated using worst-case indoor drift of ±3 °F.
    • Water-activity thresholds assume typical CTA and paint stability ranges; empirical testing will refine these limits.
    • UV risk is qualitative due to the absence of a universal UV-dose threshold for paint or CTA failure.
    Mechanical Assumption

    The Paint Glass Transition (Tg) Floor

    A critical mechanical constraint in this framework is the Glass Transition Temperature (Tg) of common animation paints (Acrylics, Vinyl-Copolymers like Cel-Vinyl, and Gum Arabic/Gouache).

    • The Brittle Threshold: Most animation paints undergo a phase change near 52°F (11°C) at 50% RH. Above this temperature, the paint is flexible. Below it, the paint enters a "glassy" state, becoming a rigid ceramic-like layer prone to shattering if the acetate base flexes.
    • RH as a Plasticizer: Moisture is essential for flexibility. If RH drops below 30%, the Tg effectively rises, causing paint to become brittle even at warmer temperatures.
    • Tier Selection: Tier 1 (53°F+) is engineered to stay safely above this mechanical "cliff." Vault Storage (<52°F) is chemically superior for the plastic but physically hazardous for the paint, intended only for static items that will not be handled.

    As new data emerges—particularly from diffusion testing, pH modeling, and controlled reset experiments—the boundaries of each tier may be refined. The structure of the tier system, however, remains governed by physical principles rather than subjective preference.

    Clarification: Scavenger Sizing & System Responsibility

    “Sized scavenger capacity” may be met through either (1) a formal cel-to-scavenger mass ratio calculation, or (2) manufacturer-provided absorption data under relevant temperature and RH conditions. Either method is acceptable as long as performance is documented and verifiable for the intended microenvironment.

    System vs. Material: Scavenger verification is a requirement for the system, not the material. Products such as MicroChamber are engineered and effective when used correctly, but frames and display cases often employ these materials in undersized or untested configurations. The responsibility to demonstrate proper sizing, diffusion access, and overall system performance rests with the frame or product manufacturer, not the archival-material supplier. This framework therefore evaluates the preservation system as a whole.

    Minimum safe RH: Operating RH limits will continue to be refined with material-specific testing. For now, Tier 0–1 exclude saturation conditions and exclude ultra-dry RH levels that may induce mechanical stress or excessive water-vapor gradients.

    Preservation Management Lifecycle: Using Tiers Over Time

    The preservation tiers describe how safe a storage environment is right now. In practice, cel care is not a single decision — it is a management cycle. Cels move from stable, lower-risk conditions into higher risk as chemistry drifts, and effective collectors use a repeating loop of assessment, preventive care, monitoring, intervention, and review to keep R (acid generation) low and D (acid accumulation) controlled.

    Not every collector is aiming for a Tier 1 or Tier 2 environment. The “right” tier depends on collector goals: desired display vs. storage, acceptable risk, and the target lifetime for the cel. The Cel Nexus risk / lifetime tool helps quantify how each tier affects estimated years of stability, making it easier to choose a tier aligned with your preservation goals.

    Preservation management lifecycle aligned with collector goals Collector goals & target lifetime 1. Intake Baseline AD / photos Initial tier choice 2. Preventive care Tier set from goals Environment stabilized 3. Monitoring AD / sensors / ppm Check vs tier model 4. Intervention Condition / reset Rehouse if needed 5. Review Update risk & tier Adjust goals if needed Reassessment based on new data & goals
    Preservation management lifecycle

    1. Assessment / Intake

    By the time vinegar odor is obvious, chemical breakdown is already underway. Intake is where you establish a baseline so tier decisions are grounded in data instead of guesses:

    • AD strips or vapor tests in a known volume to estimate vapor-phase acidity trend.
    • Quick odor/ppm check to confirm whether D is already elevated.
    • Photographs and a simple condition log (warp, blocking, paint status, annotations).

    This baseline tells you whether a cel can reasonably stay in lower tiers for your goals or whether it is already in a Tier 3 / vinegar-positive state that needs more aggressive management or reset planning.

    2. Preventive Care

    Passive storage — an inert binder in a fluctuating room, or a basic frame with no RH control — is still slow chemical drift. Preventive care means moving the cel into an environment where Lchem is intentionally improved relative to your target life:

    • Cold storage that meaningfully increases lifetime relative to a 68 °F, 50 % RH baseline.
    • RH control that keeps the material within safe absolute humidity bounds.
    • Boundary-layer and enclosure designs that control how acetic acid escapes or is captured.
    • Inert sleeves and support materials that avoid adding new chemical risk.

    In tier language, preventive care is how you place a cel in the tier that best matches your goals: a display-first collector may accept Tier 0–2 tradeoffs, while a long-term preservation collector may target Tier 1 or better. The life-expectancy / risk tool is what connects these choices to estimated years of stability.

    3. Monitoring

    Without measurement, every storage choice looks fine until it fails. Monitoring is how you confirm that a chosen tier is actually being delivered in practice:

    • Periodic AD strip checks in representative containers or binders.
    • Basic sensor logging (temperature, RH) at the storage location instead of relying on room averages.
    • Occasional ppm tube tests for higher-risk or vinegar-positive cels, when available.

    Monitoring links the abstract tier labels back to real data. If a setup that was modeled as Tier 1 is consistently running warmer or wetter than intended, or if vapor acidity is rising, the framework tells you to re-classify the cel and adjust your plan or goals.

    4. Intervention / Reset

    Waiting for visible warping or paint loss leaves little room to recover. Intervention is where you deliberately reduce D (stored acid) and, where possible, improve R by rehousing. Depending on your goals and the cel’s condition, that may include:

    • Conditioning in a reset environment with controlled temperature and ~30% absolute humidity.
    • Using molecular sieves or engineered scavengers sized and placed for effective diffusion and absorption.
    • Rehousing into microenvironments or containers that better support the target tier and lifetime.

    The goal is not simply to “feel better” about storage, but to move the cel back toward a tier that gives a measurable lifetime benefit relative to its current chemical state and your target horizon (e.g., 10, 20, or 30 years).

    5. Review & Planning

    The final stage closes the loop so preservation stays proactive instead of reactive:

    • Schedule reassessments based on risk and goals — vinegar-positive or high-value cels are checked more often.
    • Update lifetime expectations as new AD data, pH estimates, or environment logs come in.
    • Decide whether to adjust goals (accept a lower tier) or invest in higher tiers and/or reset work.

    Over time, this turns preservation from “set and forget” into a managed program. The tier framework becomes not just a label for storage hardware, but a way to track where each cel sits on the degradation curve, what level of risk you are accepting, and how your actions are shifting its trajectory toward your chosen outcome.

    Collector Practice vs. Preservation Model

    The table below summarizes how the same lifecycle looks under casual collector habits versus a structured, goal-aligned tier-based approach.

    Stage Common Collector Practice Preservation Model
    Assessment / Intake Wait for vinegar smell; no systematic intake or logging. Baseline AD/ppm trend, photos, and condition log; assign an initial tier based on environment and goals.
    Preventive Care Inert binders/frames in uncontrolled rooms; “good enough” assumptions. Deliberate choice of Tier 0–2+ based on temperature, RH, diffusion, and target lifetime from the risk tool.
    Monitoring No periodic checks; issues found only when visually obvious. Regular AD strip checks, sensor logs, and spot ppm tests to confirm the tier is actually being delivered.
    Intervention / Reset Ad-hoc fixes after warping or paint issues appear. Planned conditioning, scavenger use, and rehousing when data shows rising acidity or risk relative to goals.
    Review & Planning None; storage rarely revisited once “done.” Scheduled reassessment, updated lifetime models, and targeted upgrades or goal adjustments as cels drift toward higher tiers.

    Closing Thoughts – Choosing Your Tier

    At the end of the day, every collector is choosing a tier, whether consciously or not. Leaving cels in warm, humid, unbuffered environments pushes them toward the right-hand side of the bathtub curve sooner. Basic archival steps put you in the middle lane. Only deliberate temperature, humidity, and vapor-phase control — plus smart geometry — actually push your cels into the high tier where the clock slows down.

    In future posts, we’ll walk through example setups and show how to move a specific cel from low → medium → high tier using real products, measured conditions, and test data.

    For the full risk-scoring methodology behind these tiers — likelihood vectors, correction factors, and how likelihood combines with consequence — see Section 3 above. For the underlying kinetic model that drives Lchem,env, see the Engineering Framework.