Sealing Thermal & Desiccant — Technical Specification Overview

Dimensional Stability Under Thermal-Humidity Cycling: The Overlooked Failure Mode #

A plant maintenance team in Southeast Asia reported unexpected seal failures six months after commissioning a heat exchanger assembly. The seals passed incoming inspection on Shore A hardness and tensile strength. The desiccant units met the specified adsorption capacity at 25°C per the COA. What the team had not tested — and what the Chinese supplier had not documented — was combined thermal-humidity cycling performance. After repeated exposure to 85°C/85% RH operating conditions, the elastomeric seals had deformed beyond functional recovery, and the desiccant units had partially fused to the housing walls due to swell-induced compression.

The root cause was not material fraud. The seal compound was within stated grade, and the desiccant bead size was correct. The failure came from two independent specification gaps that compounded: compression set was not verified at operating temperature, and the desiccant’s swell coefficient under high RH had not been specified on the purchase order.

When we reviewed the COAs and drew down the supplier qualification records, both suppliers had provided technically accurate documentation. The problem was what neither document contained — dimensional behavior under conditions representative of actual service. This is the most consistent sourcing gap we encounter in this category, and it is entirely preventable with three targeted test requirements at the qualification stage.

Critical Parameters That Predict Field Performance #

The four parameters that most reliably predict combined sealing-thermal and desiccant performance in demanding service are compression set, swell coefficient under RH exposure, dimensional change over thermal cycles, and lot-to-lot consistency of adsorption rate. Standard procurement processes typically verify only the first and last of these — and even then, not under service-representative conditions.

Compression set for elastomeric seals should be tested per ASTM D395 Method B at the actual operating temperature, not at the standard 70°C default. For applications above 120°C, NBR will routinely show compression set values exceeding 40% after 70 hours, while FKM grades hold below 18% at the same conditions. The difference sounds marginal on paper. In a flanged connection that opens and closes thermally over a 5-year service life, it is the variable that separates a seal that functions in year four from one that leaks in year two.

Swell coefficient under humidity exposure is almost never specified in Chinese supplier quotations for desiccant units in housings or cassettes. Per ISO 62 water absorption testing methodology, silica gel desiccant bead swell in high-RH environments (>80% RH) can range from 1.2% to 4.7% volumetric expansion depending on binder formulation. In tight housing geometries with tolerance clearances under 0.5 mm, the upper end of that range causes mechanical binding. Specifying maximum swell coefficient at 85% RH is a one-line addition to a purchase order that eliminates a failure mode that generates expensive warranty claims.

Dimensional change over thermal cycles applies to both the thermal interface material layer and the sealing element in assemblies that see repeated heat-up/cool-down. For thermal pads used in power electronics sealing, we test dimensional retention per ASTM E831 thermomechanical analysis across -40°C to +150°C, 50 cycles. Acceptable dimensional change for a 2.0 mm pad is less than 0.12 mm per face in the z-axis. Above that threshold, the contact resistance at the interface degrades in a way that is irreversible without recompression.

Lot-to-lot adsorption consistency for desiccant materials is the variable that procurement teams most frequently under-specify. Requesting a single COA value for adsorption capacity at 25°C/50% RH tells you what one batch can do under one condition. Requiring three consecutive batch COAs with adsorption rate at 40°C/80% RH — a more demanding and more application-representative condition — tells you whether the supplier’s raw material sourcing is stable. Per our internal QC-11 supplier evaluation protocol, we flag any supplier where batch-to-batch adsorption rate variation exceeds ±5% at 40°C/80% RH as requiring re-qualification before volume orders.

The parameter teams most commonly overlook is swell coefficient. It does not appear on standard Chinese supplier datasheets. Asking for it at the quotation stage is a reliable filter for suppliers who have genuine process control versus those who are assembling product from spot-market raw materials.

In our assessment of 31 suppliers, FKM grades from Tier 1 Chinese compounders consistently met the ≤18% compression set threshold at 150°C. The same figure from Tier 3 spot-market suppliers averaged 26% — still technically “FKM” but with a formulation that does not perform as the grade designation implies.

Decision Framework: Matching Specification Depth to Application Risk #

If the application involves static sealing below 100°C with no humidity exposure above 60% RH, a standard COA with Shore A hardness, tensile strength per ASTM D412, and adsorption capacity at 25°C is an adequate qualification basis. NBR in this range is cost-effective and well-supported by Chinese supply. The sourcing risk is low and the qualification burden should reflect that.

If operating temperature exceeds 120°C or the assembly sees humidity cycling above 75% RH, the specification must add compression set at temperature and swell coefficient. At this tier, the choice between NBR and FKM becomes consequential rather than academic, and the cost differential — which in our experience runs roughly 2.5x to 3.5x for comparable seal geometries — is justified by the service life extension alone. For pump-valve-seals in heat recovery or HVAC applications, FKM is the only practical option above 130°C continuous service.

If the assembly involves a desiccant unit in a tight housing with dimensional clearances below 1.0 mm — common in enclosure-mounted electronics and precision instrumentation — swell coefficient becomes the critical path specification. The COA will not contain this figure by default. Requesting it eliminates suppliers who cannot provide it, which in our experience narrows the field to roughly 40% of initially responsive Chinese vendors. That is not a problem; it is the intended effect of the specification.

For thermally demanding applications combining a sealing element and a desiccant unit in the same assembly — power electronics housings, battery enclosure seals, industrial sensor packaging — the interaction between the two materials matters. A desiccant that swells 3.5% volumetrically will apply mechanical load to an adjacent seal. If that seal is already at 80% of its compression set limit from thermal cycling, the combined load drives failure faster than either material would fail independently. We have not seen this interaction modeled in any Chinese supplier datasheet. The qualification protocol for dual-material assemblies needs to test the combination, not each component in isolation.

The non-obvious boundary condition here: this interaction effect becomes negligible when the housing geometry provides at least 1.5 mm radial clearance for the desiccant unit. Above that threshold, swell-induced load transfer to the seal does not accumulate to failure-relevant levels in the service life data we have from our QC-11 tracked projects. Below it, the interaction requires explicit qualification.

For o-rings-static-seals in electronics or precision instrument housings that combine sealing and desiccant functions, this decision framework should be applied at the design-for-sourcing stage, not after initial sample approval.

Practical Guidance for Buyers #

When sourcing sealing-thermal and desiccant materials from China, the first specification to request is not the headline adsorption capacity or hardness figure — it’s the compression set at your actual operating temperature, combined with swell coefficient at your maximum expected RH. Both numbers need to appear on the COA with the test method and conditions stated explicitly. If they do not, request the test report, not just the value.

The specific risk scenario to guard against is raw material substitution at the compounder level after initial sample approval. In our qualification program, initial samples consistently pass specification; the failure mode emerges at production volume when the compounder changes the base polymer or plasticizer source. A standard COA will not catch this. Spot-testing incoming hardness and compression set on the first three production lots — not just the samples — is the step that catches this substitution before it reaches your assembly line.

Before committing to volume, insist on three consecutive production batch COAs covering compression set at temperature, Shore A, and adsorption rate (for desiccant materials). The sample size should cover at least 500 pieces or 5 kg per batch. Duration of the qualification window should be a minimum of 60 days to confirm batch-to-batch stability rather than a single optimized run. This is what our QC-11 protocol requires, and it is the minimum we would recommend to any procurement engineer qualifying a new Chinese supplier in this category.

FAQ #

What test method should I cite when specifying compression set for seals purchased from China?

Specify ASTM D395 Method B with temperature and duration explicit on the purchase order — “70h/150°C” is not optional detail, it is the condition that separates a meaningful result from a meaningless one. Chinese suppliers familiar with export markets know this test; those who push back on the temperature condition are worth treating as a qualification risk.

Is GB/T standard compliance sufficient for seals going into European industrial assemblies?

No, and this is a category where the standard gap matters. GB/T 5720 (O-rings) uses dimensional tolerances that map approximately to ISO 3601 Grade B, but the compression set acceptance criteria differ. A supplier who quotes “GB/T compliant” without specifying which table and which tolerance class is giving you incomplete information. Always request the ISO 3601 class explicitly if your drawing references it.

Can silica gel desiccant units cause mechanical damage to enclosure seals in tight-clearance housings?

It depends on the clearance geometry and the desiccant binder formulation. Below 1.0 mm radial clearance, volumetric swell above 2.5% creates measurable lateral load on adjacent sealing elements. Above 1.5 mm clearance, the load transfer is negligible in our tested assemblies. If your housing geometry falls in the 1.0–1.5 mm range, test the combination under your highest expected RH before committing to a desiccant supplier.

How many batch COAs should I request before approving a new Chinese supplier?

Three consecutive production batch COAs is the minimum that provides statistically useful lot-to-lot consistency data. One COA tells you what a supplier can produce when motivated. Three, separated by at least 30 days each, tells you whether their raw material sourcing is stable.

Do I need separate qualifications for the seal and the desiccant in a combined assembly?

For applications with housing clearances above 1.5 mm, separate component qualification is adequate. For tight-tolerance assemblies below 1.0 mm clearance, our dataset — based on 14 dual-material assembly projects tracked under QC-11 — shows that component-level qualification misses 3 of the top 5 failure modes. Combination testing under realistic thermal-humidity cycling is not optional in that geometry range.

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