Fickian and Non Fickian Diffusion Basics in Sensor Encapsulation Polymers

Gradient

Water molecules penetrate polymers. Ingress into microelectronic encapsulation materials begins the moment a sensor leaves a nitrogen-purged bakeout chamber. Ambient moisture adsorbs onto the polymer exterior, dissolves into the surface matrix according to Henry’s law, and moves inward driven by concentration divergence.

For sensor packaging engineers, this movement sets the clock on zero-point wander, insulation resistance drop, and dielectric baseline drift across every calibrated device.

Mass transport follows classical Fickian behavior when the mechanical relaxation rate of the polymer chains occurs orders of magnitude faster than the penetrant diffusion rate. Under these conditions, the characteristic diffusion time far exceeds the macromolecular rearrangement time. The material maintains thermodynamic equilibrium locally throughout the sorption profile.

Fickian Transport across Protective Overcoats

Sorption continues indefinitely. The mathematical framework rests on two relations formulated by Adolf Fick. The first expression defines the instantaneous penetrant flux proportional to the spatial concentration slope along a single dimension:

J = -D (dC / dx)

Flux J measures the mass transferred per unit area per second. D represents the mutual diffusion coefficient expressed in square millimeters per second, while dC divided by dx defines the concentration slope across thickness coordinate x. The negative sign designates net mass motion from zones of higher chemical potential toward zones of lower concentration.

Transient accumulation within an infinitesimal volume element yields the second relation:

dC / dt = D (d^2 C / dx^2)

Assuming a constant diffusion coefficient across uniform concentration bands, integration across a plane sheet of thickness 2L with constant boundary concentration C_s and zero initial core concentration produces the classic infinite series for total fractional mass uptake M_t relative to equilibrium saturation mass M_inf:

M_t / M_inf = 1 – Sum((8 / ((2n + 1)^2 pi^2)) exp(-D (2n + 1)^2 pi^2 t / (4 L^2)))

During early exposure phases where M_t divided by M_inf remains below 0.5, the complex summation simplifies into a linear relationship against the square root of time:

M_t / M_inf = (4 / L) (D t / pi)^0.5

Plotting relative mass gain versus the square root of time normalized by thickness yields a straight line passing through the coordinate origin. This linear slope provides the primary baseline metric during incoming material qualification.

A two-millimeter silicone encapsulation layer exposed to eighty-five percent relative humidity at eighty-five degrees Celsius reaches ninety percent moisture saturation within forty-eight hours.

Governing Equations for Planar Diffusion Flux

Mass uptake in rubbery packaging materials scales with the square root of exposure time. Silicones, soft polyurethanes, and flexible aliphatic gels operating well above their glass transition temperature maintain rapid segmental chain motion. Water molecules slip through free volume cavities without encountering energetic resistance from rigid polymer backbones.

In these systems, Fick’s equations predict moisture arrival times at the active transducer die face with high fidelity.

Temperature shifts accelerate this transport process according to an Arrhenius relationship:

D(T) = D_0 exp(-E_d / (R T))

D_0 represents the reference diffusion pre-exponential constant, E_d denotes the activation energy for diffusion in joules per mole, R is the universal gas constant, and T defines absolute temperature in kelvin. For standard epoxy Novolac molding compounds, activation energies typically span 35 to 50 kilojoules per mole. An increase in operating ambient from 25 degrees Celsius to 85 degrees Celsius accelerates penetrant diffusion velocity by a factor of 12 to 18.

Laboratory determination of diffusion coefficients relies on standardized gravimetric sorption tracking under controlled ambient conditions. Test specimens molded into flat discs undergo initial desiccation until mass readings stabilize on an analytical microbalance resolving to 10 micrograms.

1. Bakeout desorption removes residual volatile fractions and manufacturing moisture at 125 degrees Celsius for 48 hours to secure a dry baseline tare mass.
2. Environmental conditioning places the prepared disc into an environmental chamber holding 85 degrees Celsius and 85 percent relative humidity per standard IEC 60068-2-67.
3. Periodic weighing tracks cumulative mass ingress at intervals scaled to the square root of elapsed hours, returning the coupon to the chamber within 30 seconds to prevent thermal exchange errors.
4. Linear slope regression extracts parameter D from the initial linear portion of the fractional absorption curve using the known specimen thickness.

When the measured data matches the theoretical Fickian model, the extracted diffusion constant remains invariant across test coupon thicknesses. A change in specimen thickness from one millimeter to three millimeters merely shifts the time axis by a factor of nine, leaving the calculated D value identical within measurement uncertainty limits.

Relaxation

Glassy matrices resist sudden expansion. When polymers operate near or below their glass transition temperature, penetrant ingress departs from classical square-root kinetics. Rigid macromolecular segments lack the immediate conformational mobility required to accommodate invading water molecules.

The polymer matrix must physically relax, uncoil, and expand its micro-cavity network before penetrant accommodation occurs.

This coupling between mechanical deformation and chemical transport introduces anomalous, non-Fickian kinetics into sensor potting compounds. The relative speed of molecular diffusion versus chain relaxation determines which kinetic regime governs moisture uptake.

Deborah Numbers and Viscoelastic Polymer Shifts

Chain rearrangement rates dictate how penetrant molecules distribute throughout the bulk. Transport classification uses the dimensionless diffusion Deborah number, defined as the characteristic material relaxation time lambda divided by the characteristic observation diffusion time theta:

De = lambda / theta = lambda D / L^2

Here L represents characteristic diffusion path length and D denotes the mutual diffusion constant. Three operational regimes define material behavior:

• Fickian domain occurs where the Deborah number is negligible compared to unity, meaning polymer chain relaxation happens far faster than penetrant movement. Water enters open cavities without waiting for conformational matrix rearrangement.
• Case II transport arises when the Deborah number significantly exceeds unity, reflecting rigid glassy structures where molecular relaxation proceeds far slower than penetrant penetration. Ingress displays a sharp swelling shock front propagating inward at constant velocity rather than a fading concentration tail.
• Anomalous transport develops when the Deborah number hovers around unity, forcing diffusion rates and viscoelastic relaxation rates to operate on identical timescales. Mass uptake scales non-linearly with fractional time exponents ranging between 0.5 and 1.0.

Polymers operating near their glass transition temperature exhibit anomalous swelling fronts that outpace standard square-root time predictions.

Can Viscoelastic Creep Invert Mass Absorption Rates?

Dual-stage uptake profiles display an initial plateau followed by a sudden secondary surge. During incoming qualification of amine-cured potting epoxies, sorption plots frequently exhibit an apparent early Fickian plateau where mass gain flattens around 40 hours. Quality technicians often interpret this inflection as full saturation and stop the evaluation.

Extended exposure reveals this plateau as a kinetic illusion. As penetrant molecules disrupt interchain hydrogen bonds, the local glass transition temperature depresses toward the operating ambient. Once Tg drops into the environmental exposure range, plasticization triggers a sudden acceleration in viscoelastic relaxation.

The polymer begins a secondary sorption phase that absorbs an additional 30 to 60 percent more water over the subsequent 500 hours.

Linear solvent fronts propagate through glassy encapsulants under Case II kinetics. The mass uptake expression abandons fractional exponents:

M_t / M_inf = (v t) / L

The variable v denotes front velocity in micrometers per hour. Behind this advancing boundary, the polymer transitions completely from a rigid glass to a swollen rubbery state, generating steep internal shear stresses at the interface with the unswollen core.

Anomalous Exponents in Langmuirian Sorption

Permeation models classify penetrant kinetics through the empirical power-law exponent:

M_t / M_inf = k t^n

A value of n equal to 0.50 marks ideal Fickian diffusion. An exponent equal to 1.00 confirms pure Case II transport. Exponents situated between 0.50 and 1.00 define anomalous viscoelastic diffusion.

Exponents exceeding unity represent Super Case II behavior, where severe internal crazing and micro-cavitation create accelerating ingress pathways.

Langmuir-type diffusion equations account for chemical affinity between migrating water and polar moieties within the polymer structure. Water molecules exist simultaneously in two distinct states: mobile penetrants diffusing through free volume voids, and bound penetrants immobilized at polar hydroxyl, amine, or ester sites through hydrogen bonding. Two exchange parameters govern the interchange:

gamma = probability rate of a mobile molecule becoming bound

beta_r = probability rate of a bound molecule breaking free through thermal excitation

When the immobilization rate gamma exceeds the liberation rate beta_r, standard Fickian equations fail to predict die-face moisture arrival. The encapsulant acts as a molecular sponge, delaying penetrant breakthrough at active sensor surfaces while building up internal localized mass that alters mechanical stress distributions long before saturation equilibrium occurs.

Whether continuous cyclic sorption and desorption under field conditions permanently shifts the ratio of bound to free sites remains an unsettled debate among polymer physicists studying sensor drift.

Seal

Thick coatings delay moisture arrival. Transducer packaging relies on polymer barriers to protect wirebonds, silicon dies, and passivated metallization from industrial atmospheric contaminants. Standard hermetic packaging utilizing welded kovar cans or gold-tin soldered ceramic headers provides helium leak rates below 1e-9 standard cubic centimeters per second, preventing moisture ingress over decades.

Automotive and consumer sensors substitute polymer encapsulation to compress manufacturing bills of materials.

Polymer barriers are semi-permeable membranes. No organic potting resin, silicone glob-top, or epoxy transfer molding compound forms a permanent moisture seal. Given sufficient exposure duration, water vapor penetrates every polymer barrier, arriving at the underlying silicon circuitry.

Encapsulant Morphology and Filler Interfaces

Fused silica spheres added to control thermal expansion alter internal path tortuosity. Unfilled epoxy resins exhibit a high coefficient of thermal expansion, spanning 55 to 70 parts per million per degree Celsius. Silicon dies exhibit thermal expansion around 2.6 parts per million per degree Celsius.

Unfilled resins tear wirebonds and crack silicon dies during thermal shock testing.

Formulators blend 70 to 90 weight percent micro-fused silica fillers into epoxy molding compounds to suppress thermal expansion down to 8 to 14 parts per million per degree Celsius. This mineral loading forces penetrant molecules into winding paths around impermeable silica particles. The effective diffusion coefficient drops according to a tortuosity factor tau:

D_eff = D_resin (phi / tau)

The variable phi represents the polymer volume fraction, while tau accounts for geometric path lengthening. Poor coupling between the mineral filler surface and the organic resin matrix completely reverses this benefit. If silane coupling agents degrade or fail to wet filler surfaces during compounding, continuous microscopic annular gaps form around every silica sphere.

These gaps act as moisture superhighways. Water vapor condenses into capillary channels, enabling rapid penetrant channeling along filler interfaces that bypasses the bulk resin diffusion resistance entirely.

Moisture Front Velocity toward Active Microstructures

Bond wires surrounded by molding resin experience localized chemical attack when penetrant molecules contact internal metallization. Water presence alone drops surface insulation resistance between adjacent leadframe fingers from 100 gigaohms down to several megaohms. Combined with ionic contaminants such as residual chloride ions leaching from unpurified epoxy feedstocks, moisture forms an active electrolyte that initiates galvanic corrosion on aluminum bond pads.

Delamination at the interface between the encapsulant and the silicon die creates an immediate moisture reservoir. A similar phenomenon occurs in commercial aerospace composite skins, where ultrasonic nondestructive testing tracks hidden water ingress along carbon ply interfaces following sub-zero thermal cycling. In sensor packaging, acoustic micro-imaging confirms that delamination voids fill with condensed liquid within hours of environmental exposure, eliminating the geometric diffusion barrier provided by package thickness.

Selecting an encapsulation material involves strict evaluation of chemical and mechanical boundary constraints:

• Hydrolytic stability prevents chain cleavage and acid generation when water molecules attack ester or anhydride linkage groups within the cured resin backbone.
• Adhesion promoter integrity preserves covalent bonding across silicon nitride die passivations and copper leadframe finishes under cyclic thermal shear loads.
• Ionic purity limits restrict extractable chloride, bromide, and sodium ion concentrations below 10 parts per million to suppress electrolytic corrosion cells.
• Glass transition headroom maintains the material operating envelope at least 30 degrees Celsius above peak environmental exposure to avoid triggering Case II relaxation fronts.

Material suppliers frequently argue that surface hydrophobic additives eliminate the need for hermeticity by preventing bulk wetting, yet moisture vapor transport through the polymer bulk continues unabated regardless of contact angle measurements on the exterior package surface.

Hysteresis

Moisture induces volumetric expansion. As water molecules occupy free volume spaces and form hydrogen bonds with macromolecular backbones, the polymer matrix swells. In unconstrained specimens, this deformation is characterized by the coefficient of hygroscopic swelling, defined as the strain induced per unit weight fraction of absorbed water:

epsilon_h = beta Delta C

The variable beta represents the swelling coefficient, typically ranging from 0.15 to 0.45 for epoxy molding compounds, and Delta C defines moisture concentration change. When an encapsulant adheres to a rigid silicon substrate, mechanical boundary conditions prevent free expansion along the interface plane.

This physical constraint converts volumetric swelling into severe compressive and shear stresses across the underlying die. Swelling produces asymmetric stress.

Hygroscopic Stress on Piezoresistive Bridges

Silicon piezoresistive pressure sensors and strain gauge bridges detect mechanical stress through resistivity shifts within doped semiconductor paths. The relative resistance change follows the fundamental piezoresistive tensor relation:

Delta R / R = pi_l sigma_l + pi_t sigma_t

The coefficients pi_l and pi_t denote longitudinal and transverse piezoresistive coefficients, while sigma_l and sigma_t represent stress components parallel and perpendicular to current flow. In an ideal transducer, these stresses originate purely from diaphragm deflection under applied external pressure.

Moisture ingress breaks this measurement isolation. In a typical piezoresistive pressure sensor die coated with a two-millimeter protective glob-top epoxy, absorbing 1.5 weight percent water generates package-induced stresses reaching 15 to 35 megapascals across the silicon surface. This parasitic stress produces a direct output shift.

The zero-point offset wanders, mimicking a genuine physical pressure change.

JEDEC standard J-STD-020 classifies moisture sensitivity levels to prevent internal package delamination during high-temperature solder reflow.

Will Dual Stage Sorption Distort Longterm Zeroes?

Capacitive drift reflects water absorption. Liquid water exhibits an extraordinarily high relative permittivity, around 78 to 80 at room temperature, whereas dry sensor packaging polymers exhibit dielectric constants between 2.8 and 4.2. As water molecules diffuse into the dielectric field of capacitive MEMS sensors, the effective dielectric permittivity of the packaging environment increases systematically.

A penetrant concentration increase of 1.0 weight percent inside an encapsulation layer increases its dielectric constant by 6 to 12 percent. In unshielded capacitive relative humidity sensors, capacitive barometric sensors, and micro-accelerometer readouts, this shift produces baseline drift that invalidates factory calibration settings.

Cyclic humidity exposure generates pronounced hysteretic loops between absorption and desorption phases. During adsorption in rising humidity, water enters dry micro-cavities. During desorption in falling ambient humidity, capillary condensation within nanometer-scale pores and hydrogen bonding with polymer polar sites retards water release.

The desorption diffusion coefficient frequently measures 20 to 50 percent lower than the absorption diffusion coefficient. Desorption rates lag absorption speeds. When a sensor cycles between 20 percent and 80 percent relative humidity, the transducer output traces a closed hysteresis loop.

The calibration offset measured at 50 percent relative humidity during an upward humidity swing differs significantly from the offset measured at 50 percent during a downward swing.

The desk cannot fully defend published literature claims stating that post-cure thermal annealing permanently stabilizes moisture absorption kinetics in transfer-molded epoxy sensors. Annealing removes residual curing stresses and drives off traces of unreacted monomer, but experimental data tracking long-term drift reveals that subsequent hydrothermal aging erodes this benefit within several hundred operational hours. Sourcing teams relying on annealing to bypass moisture screening find that sensor zero baselines drift unpredictably once field humidity cycles resume.

Capacitive Baseline Shifts from Permeated Moisture

Dielectric constant discrepancies between liquid water and dry polymers create signal offset errors. In high-precision capacitive transducers, fringe electric fields inevitably pass through portions of the protective encapsulation layer. When moisture enters these fringing zones, the capacitance baseline shifts independently of the primary physical variable being tracked.

This parasitic signal drift corrupts calibration traceability. The measurement error cannot be canceled through simple software temperature compensation because water diffusion kinetics lag thermal changes by hundreds of hours. A temperature excursion alters the saturation limit immediately, but the concentration profile shifts slowly according to Fickian or non-Fickian transport mechanics.

Engineering teams that treat moisture-induced baseline wander as random electrical noise miss the physical degradation mechanism, allowing shifting calibration baselines to pass into field service where the resulting measurement drift generates widespread field returns and expensive warranty claims.

Toll

Calibration certificates expire under drift. Sourcing managers frequently treat sensor calibration as an isolated factory milestone, buying parts based on initial catalog accuracy specifications. A pressure sensor certified to an accuracy of 0.05 percent of full-scale span on the factory floor will drift beyond 0.50 percent within six months if moisture-induced swelling alters the piezoresistive balance.

The commercial cost of non-Fickian diffusion appears in recurring field recalibration expenses, degraded factory test yields, and expanded warranty reserves. Mitigating these expenses requires adjusting procurement documents and incoming inspection procedures to match the physics of polymer mass transport.

Calibration Costs under Non Equilibrium Conditions

Environmental chambers consume factory floor time while waiting for plastic packages to saturate. Standard production schedules allocate between 10 minutes and two hours for thermal and humidity stabilization during multi-point sensor calibration trimming. For a packaged sensor encased in a two-millimeter epoxy shell, reaching true chemical sorption equilibrium requires hundreds of hours.

Factory trims executed while moisture profiles remain in transient flux freeze non-equilibrium concentration gradients into the digital compensation tables. As water continues to diffuse toward the die face in storage, the concentration profile flattens, shifting sensor zero-points before the device reaches an end customer. Dry nitrogen preserves zero baselines.

Consider a standard procurement batch of 10,000 barometric pressure sensors incorporating transfer-molded packaging. Calibrating these units under unconditioned factory ambient conditions introduces an initial calibration error band of 0.15 percent due to unmanaged humidity states. A specialized 96-hour pre-conditioning soak at 40 degrees Celsius and 75 percent relative humidity stabilizes the concentration profile, compressing the post-trim drift band down to 0.03 percent.

Chamber run-time and electrical power add a direct toll of 2.15 dollars per unit. Skipping this step avoids immediate production overhead, but historical warranty statistics across harsh industrial environments demonstrate a subsequent field service call rate of 4.2 percent, costing an average of 145 dollars per returned field transmitter.

Unbaked sensor assemblies calibrated immediately after humid transit produce baseline errors that disappear once the package dries out.

Incoming Batch Qualification and Desorption Windows

Receiving docks face hidden financial penalties when unconditioned shipments fail acceptance screening. Sensors shipped across maritime freight routes absorb ambient humidity inside standard packaging bags. If incoming quality verification tests occur immediately after opening crates, parts will fail zero-offset specifications due to transient hygroscopic stress.

Incoming quality agreements must specify controlled stabilization windows. Components require a 24-hour bakeout at 60 degrees Celsius followed by a 48-hour equilibration period inside a desiccant cabinet holding less than 10 percent relative humidity prior to formal bench testing. Temperature accelerates molecular mobility.

Without this step, incoming inspection lots will be falsely rejected, triggering supply chain disputes and delayed production builds.

Contractual procurement specifications must abandon vague references to general reliability standards and incorporate exact environmental test regimes tied to metrological traceability. Sourcing contracts covering plastic-encapsulated sensor assemblies should enforce the following operational clause:

Per IEC 60068-2-78 test criteria, sensor offset and span calibration drift shall remain within plus or minus 0.10 percent of full-scale range following 1000 hours of continuous exposure to 40 degrees Celsius and 93 percent relative humidity, with measurement verification conducted within 60 minutes of extraction from the chamber following an initial two-hour surface de-moisturizing stabilization in dry air. Incorporating this specific line into the purchase contract shifts the financial burden of incomplete polymer curing, improper filler silanization, and unverified packaging formulations back onto the component supplier before hardware ships.