Sealing Thermal & Desiccant — Supplier Qualification Guide

Failure Modes That Trigger a Qualification Review #

You open a batch of desiccant pouches and half are already at 40% saturation before installation. Or you receive thermal interface pads that measure 1.8 W/m·K on the sample COA but perform at roughly 1.1 W/m·K in your junction temperature testing. Or your sealing strips deform at 90°C when the COA says continuous service to 130°C. Each of these is a real scenario, and each traces back to a qualification gap, not a manufacturing accident.

Observable symptoms map to different root causes depending on where in the supply chain the problem originates.

Pre-saturation and Rth drift are the two symptoms that generate the most escalations in our QC-07 material risk procedure. They are also the ones most suppliers attribute to “shipping conditions” — which is a deflection, not a diagnosis.

The Root Cause Teams Consistently Misdiagnose #

The failure that generates the most misdiagnosis calls is thermal conductivity drop in silicone-based thermal interface pads. Teams blame the test method, the application torque, the surface finish of the heat spreader. In roughly two-thirds of confirmed cases we have reviewed, the actual cause is filler particle size distribution shift at the compounder level, not anything downstream.

Here is the mechanism. Silicone thermal pads achieve their rated conductivity through densely packed alumina, boron nitride, or aluminum oxide filler — typically at 60–75% by weight for pads rated above 3 W/m·K. That filler content is expensive. When a compounder reduces cost by dropping filler loading from, say, 68% to 58%, the change does not always show up as a Shore A hardness deviation because the silicone matrix itself absorbs the difference in mechanical feel. What does change is thermal conductivity, which drops nonlinearly with filler volume fraction. A 10-point drop in filler loading can translate to a 25–35% reduction in effective thermal conductivity at the bulk pad level.

The COA will still show a passing hardness value. It may still show a passing tensile strength. If the supplier is testing thermal conductivity via hot-disk method rather than ASTM D5470 stack resistance measurement, the reported value will reflect a single-point material measurement, not a system-level Rth — and those two numbers are not interchangeable. Hot-disk values are consistently 15–30% more optimistic than ASTM D5470 values for the same pad under application conditions, because they do not capture contact resistance at the interface.

To confirm this root cause, you need two things: a cross-section SEM image of the pad (which will show filler particle density and distribution) and an ASTM D5470 Rth measurement at the specified bondline thickness and clamping pressure. The pass threshold we use is: Rth must be within ±10% of the COA-stated value at 50 psi clamping pressure and 0.2 mm bondline thickness. A deviation greater than ±10% at these conditions flags a filler loading problem, not a test method problem.

For desiccant, the equivalent misdiagnosed failure is adsorption rate vs. adsorption capacity. Teams see fast initial moisture uptake and conclude the product is performing. What actually matters for sealed packaging applications is equilibrium capacity at the service humidity — typically measured at 25°C, 90% RH per ISO 9001 referencing GB/T 10294 conditions. A desiccant can reach 50% saturation quickly and then plateau well below its rated capacity if the silica gel pore structure is compromised by high-temperature storage or contaminated raw silicate feedstock. The symptom looks like pre-saturation. The root cause is a degraded pore structure, and the two have very different corrective paths.

Corrective Actions Ranked by Impact and Feasibility #

The following actions address the root causes above. They are ordered roughly by impact-to-effort ratio, not by ease of implementation.

1. Require ASTM D5470 (thermal) or gravimetric adsorption testing (desiccant) on every incoming lot, not just at qualification. This is the highest-impact action and the one most procurement teams delay because it adds 3–5 working days to goods receipt clearance. The alternative — discovering the failure in production — costs significantly more. For thermal pads, the AQL sampling level should be Level II, 2.5% maximum defective. For desiccant, sample size per DIN 55473 Annex A, with gravimetric retest on 10% of units per lot.

2. Add a filler content declaration to the thermal pad PO. Specify minimum filler loading by weight percentage, not just thermal conductivity W/m·K. A supplier who cannot or will not declare filler loading on the COA is a supplier whose formulation you cannot audit. This single PO clause eliminated the majority of pad performance deviations in the accounts we manage.

3. Require three consecutive production lot COAs before volume commitment. Not three sample submissions — three production lots, shipped against commercial orders, with full traceability to batch numbers. This is the filter that catches suppliers who pass qualification with a specially prepared sample and revert to standard production material at volume. Four of the eleven suppliers in our most recent evaluation cohort failed at this stage.

4. Audit raw material sourcing at the compounder level for desiccant. Silica gel adsorption capacity is directly tied to the quality of the sodium silicate feedstock. Chinese compounder-grade silica gel varies significantly by region of production. Request a raw material origin declaration and ask whether the compounder issues a separate COA for incoming silicate. Most do not — but the request signals to the supplier that you audit at this depth.

5. Switch evaluation metric from unit price to total-cost-of-ownership. For thermal pads, the cost of a failed lot is not just the pad cost — it is the rework, the junction temperature re-qualification, and potentially the warranty exposure downstream. For desiccant, a pre-saturated lot that passes incoming inspection visually but fails in use can void the seal warranty on the packaged product. The math on per-unit price optimization usually dissolves when you include a realistic rejection rate.

Prevention — What to Specify Upfront #

The specification errors we see most often are errors of omission, not errors of value. Buyers specify thermal conductivity but not test method. They specify adsorption capacity but not the temperature and humidity conditions at which it was measured. They specify Shore A hardness but not compression set after thermal aging.

For every sealing thermal or desiccant PO, the spec sheet should include: (a) the test method and conditions for each critical parameter, not just the target value; (b) lot-to-lot consistency tolerance — we use ±8% for adsorption capacity and ±10% for thermal Rth; (c) raw material declaration requirements; and (d) packaging and storage conditions with humidity exposure limits before use.

The document to request before any volume commitment is a Process FMEA covering raw material substitution risk. A supplier who has never prepared one will take time to generate it — that delay alone is informative.

Practical Guidance for Buyers #

When sourcing sealing thermal and desiccant materials from China, start with the COA field for compression set (sealing materials) or equilibrium adsorption capacity at 90% RH (desiccant) — not hardness, not tensile strength. Those fields are easier to optimize for and do not predict service performance as reliably. A sealing material with Shore A 60 and compression set of 38% after 70 hours at 100°C will fail in a static seal application regardless of how many other COA fields pass.

The specific risk scenario to plan for is raw material substitution at the compounder level between your qualification lot and first production shipment. This is not always intentional — compounder-level raw material supply chains in China are less consolidated than Western buyers typically assume, and a silicate feedstock switch can happen without any change to the product name or grade designation. The ±12% adsorption deviation threshold mentioned above is the earliest measurable signal of this substitution.

Before volume commitment, insist on three consecutive production lot COAs with full batch traceability — not three submissions from the same production run relabeled as separate lots. Pair this with incoming spot-testing on the first two commercial shipments using ASTM D5470 for thermal interface materials or gravimetric retest per DIN 55473 for desiccant. Sample size: minimum 5 units per lot for thermal pads, 10% of pouches per carton for desiccant. If you are also qualifying pump valve seals or elastomeric components from the same supplier family, the compression set threshold applies directly and cross-qualification data is worth requesting.

For related materials in adjacent categories, our guidance on adsorption and desiccant media covers activation temperature, pore size selection, and pharmaceutical-grade compliance separately.

Is ASTM D5470 the right test for all thermal interface materials?

For pads and phase-change materials used under clamping pressure, yes — it is the most application-relevant method because it captures contact resistance, not just bulk material conductivity. For thin films applied with adhesive, the bondline contribution changes the calculation and you need to test the full stack, not the pad alone.

What COA fields are most commonly falsified or omitted by Chinese suppliers?

Compression set for sealing materials and equilibrium adsorption capacity for desiccants. Both require time-consuming test cycles, which creates pressure to transcribe from a previous lot or from a target value rather than actual measurement. Request the raw test data file, not just the summary table.

Should we requalify a supplier after a single out-of-spec lot?

One out-of-spec lot triggers a corrective action request and incoming hold — not automatic disqualification. Disqualification applies when three consecutive lots fail the same parameter, or when the supplier cannot provide a credible root cause within 10 working days. The threshold matters because lot-to-lot variation at the ±8% level is manageable; systematic drift is not.

Does DIN 55473 apply to desiccant pouches sourced from China for non-European markets?

DIN 55473 is a German standard adopted broadly for desiccant unit weight and adsorption testing, but it is not a legal requirement outside the EU. Several Chinese manufacturers test against it anyway because it is the most commonly cited international benchmark. If your end market does not require it, you can specify adsorption capacity conditions directly — 25°C, 90% RH, equilibrium — and achieve equivalent incoming test coverage without formal standard compliance.

What is the minimum sample size for incoming thermal pad inspection?

For lots under 500 units, we use a minimum of 5 pads per lot tested to ASTM D5470 at the specified bondline thickness and clamping pressure. For lots above 500 units, AQL Level II at 2.5% maximum defective, which translates to 32 units inspected for a lot size of 500–1,200. A single thermal conductivity outlier more than 15% below COA value in the sample is sufficient to hold the lot pending full retest.

Published by sinoraw.com Technical Team | Request a sourcing consultation