Hand-considered hypotheses, 4201–4300

Written one at a time, not composed. Each entry names the logical form it uses as well as the process it concerns: some claim a measured quantity is misattributed, some assert a bound, some invert a standard practice, some describe an artefact that mimics a real effect. None has been tested. These are hypotheses 4201–4300 of the series, confined to the process physics of electrode coating, drying, calendering, slitting, winding and stacking, electrolyte filling, formation, and end-of-line testing. Estimates are marked; no measurement is claimed.

Slurry and mixing

Misattribution — gas as viscosity
H4201. Batch-to-batch viscosity drift in a planetary mixer is usually charged to binder hydration, but entrained and dissolved air changes the measured torque without changing the solids network at all. Splitting one batch and measuring half under vacuum and half at ambient pressure at two gap settings separates the two, because air makes the apparent viscosity gap-dependent while hydration does not.
Hidden variable — mixing energy
H4202. Two slurries mixed to the same endpoint at different speeds are not the same slurry, since agglomerate breakup and binder adsorption are both driven by cumulative specific energy rather than by elapsed time. A matrix of speed and duration at constant specific energy should collapse viscosity and adhesion onto one curve, and points that fail to collapse mark a second mechanism, most likely binder chain scission.
Percolation threshold — drying-determined
H4203. Below a carbon content estimated near 2 wt% (estimate), electrode conductivity is decided less by how much carbon is present than by where drying leaves it, because a late redistribution can build or break the percolating cluster. Conductivity versus carbon content on freeze-dried and oven-dried films should share a high-content plateau and diverge below the threshold, and the divergence is the drying contribution made visible.
Inversion — over-mixing
H4204. Over-mixing is normally treated as a viscosity problem, yet its more damaging consequence may be the loss of elastic structure that keeps a wet film from dewetting the foil. A mixing series that holds viscosity constant by adjusting solids and varies only the elastic modulus would show defect density tracking modulus rather than viscosity, which inverts the usual rheology specification.
Identity — two agglomerate measures
H4205. Rheology and electron microscopy both report agglomerates, but rheology weights the largest few by hydrodynamic volume while microscopy counts by number, so the two distributions cannot be the same object even when they share a mean. A tracer experiment with fluorescent particles, sized by both methods on the same sample, should find the rheological upper end populated by clusters that microscopy calls isolated primary particles. The distinction matters because a formulation change that shrinks the rheological upper end can leave the microscopic population untouched.
Conserved quantity — solvent ledger
H4206. The solids content written on a slurry card is a setpoint rather than a measurement, because solvent evaporates from an open mixer throughout dispersion and the batch that reaches the coater is usually a little drier than the one specified. Weighing the condenser catch or running the same formulation in a closed mixer turns the nominal formulation into a range and accounts for part of the viscosity spread otherwise assigned to raw materials.
Selection — top-of-tank sample
H4207. Samples drawn from the top of a settled tank are selected against the dense, coarse fraction, so a viscosity rise observed over a long batch can be a sampling artefact rather than a reaction. Depth-resolved sampling at fixed elapsed time, or a continuous in-line measurement, is the minimal control, and without it the batch-time trend is not interpretable. Tank geometry changes the size of the bias without changing the chemistry.
Scaling — Kolmogorov mismatch
H4208. Geometric similarity between a lab mixer and a production mixer does not preserve the Kolmogorov scale, so matching tip speed reproduces the largest eddies and misrepresents the smallest ones that actually break agglomerates. Matching specific energy and peak shear rate instead is over-constrained in most geometries, which means a lab optimum transfers as a direction rather than as a number.
Artefact — rheometer slip
H4209. The strong shear-thinning reported for high-solids battery slurries at high shear rates is partly wall slip, because the suspension dewets a smooth fixture and slides as a plug. Roughened or serrated fixtures should shift the apparent viscosity at high shear while leaving the low-shear plateau unchanged, and the size of that shift is the error carried into any slot-die pressure calculation.
Limiting case — dilute convergence
H4210. At low solids all reasonable formulations converge toward the solvent's viscosity, so a comparison made in that regime is a comparison of solvents and dispersant chemistry. Differences among binders and carbon grades appear only above a solids threshold estimated near 55–60 wt% (estimate), which is why rankings from dilute rheology so often fail at coating solids.
Causal chain — pH drift
H4211. Water-based binder solutions are acid–base systems, and the pH of a CMC or polyacrylate slurry drifts as the polymer dissolves, changing the ionization of whichever dispersant is present. If viscosity tracks pH rather than mixing time, a buffered formulation should flatten the batch-time drift without changing any other ingredient, and that result would also identify which dispersant is actually active.
Bound — maximum solids
H4212. There is a maximum solids loading for a given particle-size distribution that no dispersant can exceed, because the adsorbed layer fixes a minimum separation and therefore a minimum free volume. A dispersant can shift the bound by roughly the layer thickness, but a claim that a new additive raises solids far beyond the packing limit implies either particle deformation or a measurement that is not what it claims. It is a packing calculation before it is a screening result.
Counterfactual — mixing order
H4213. If only the final adsorption equilibrium mattered, adding binder as a dry powder or as a pre-dissolved solution should give the same electrode at the same final adsorption. Any systematic difference between the two routes measures competitive adsorption, carbon blocking binder sites or binder coating carbon, and that difference rather than the order itself is the mechanism worth reporting.
Two-regime — kneading crossover
H4214. At low stress, agglomerates erode slowly from their surfaces; above a stress set by their internal coordination, they fracture into fragments, and the two regimes give different size distributions for the same energy input. A mixer that can be run in both regimes, with the crossover located by a dimensionless stress, would explain why some scale-ups change the slurry's texture while consuming the same energy per kilogram.
Symmetry — addition order
H4215. Swapping the addition order of carbon and binder should be exactly equivalent if the two adsorb independently, so any measured difference is a direct measure of competition rather than of mixing history. The test is cheap, and the sign of the difference identifies which species wins the surface.
Path dependence — binder thermal history
H4216. A binder solution dissolved hot and cooled quickly has a different coil size, and therefore a different thickening efficiency, from one dissolved at a lower temperature and equilibrated. Holding formulation and final solids fixed while varying only the dissolution thermal history should shift slurry elasticity and, if the effect survives drying, the electrode's adhesion. If it does not survive drying, the thermal history is a slurry-handling issue only.

Coating and drying

Misattribution — pump stripes
H4217. Cross-web stripes in slot-die coating are usually blamed on pump pulsation, but a gear pump and a syringe pump delivering the same mean flow can produce different stripe patterns if the die body or its mounting is resonating. Recording die-inlet pressure simultaneously with thickness separates pump-driven from die-driven stripes, because only the former should appear in the pressure trace at the stripe frequency.
Hidden term — die compliance
H4218. The die body is not rigid: its compliance stores liquid and low-pass filters the pressure variations delivered to the slot, so wet thickness reflects the die's elastic response as much as the pump's. A die of different stiffness, run at the same mean pressure and flow, should change the amplitude but not the frequency of residual thickness waves, which is how the compliance term can be isolated.
Instability onset — air entrainment
H4219. Below a certain vacuum level at the upstream lip, the coating bead pulls air in at the downstream lip, and the resulting pinholes are often read as particulate contamination. The onset should scale with the ratio of vacuum pressure to the coating flow's dynamic pressure rather than with vacuum alone, so a speed–gap map at constant vacuum collapses the onset onto one curve. A map of that kind is a better scale-up tool than a single vacuum recommendation.
Inversion — speed and defects
H4220. At low coating speeds, dewetting and edge defects dominate because the wetting line has time to recede, while increasing speed suppresses them until inertial and air-entrainment defects take over. The result is a defect-density minimum at an intermediate speed, which means a coater run slower than necessary can be dirtier than one run at its design point.
Bound — drying rate
H4221. Binder migration to the free surface begins when the evaporative Péclet number exceeds a value near one, so there is a drying rate above which surface enrichment is unavoidable at a given thickness and diffusivity. The bound belongs to the binder–coating pair rather than to the oven, and no nozzle arrangement removes it without changing the formulation.
Identity — migration versus skinning
H4222. Surface binder enrichment can come from migration during drying or from a skin that forms at the interface and traps binder beneath it, and the two give the same surface analysis but opposite depth profiles. Sectioning the electrode at two drying rates should show the enrichment front moving inward in one case and staying pinned at the surface in the other.
Conserved quantity — solvent flux
H4223. In an impingement oven the solvent flux is limited by the gas-side boundary layer, so raising nozzle velocity improves drying only until the boundary layer is thin enough that internal diffusion becomes limiting. An energy balance across the oven, heat in, solvent evaporated, air exhausted, would show whether a claimed drying improvement is a change in heat transfer or merely a change in the exhaust's solvent load.
Selection — visible defects
H4224. Optical inspection sees the surface, so a defect distribution built from what it sees is conditional on visibility: pinholes and scratches are over-represented while subsurface voids and binder-rich lenses are absent. The correct inference is not that internal defects are rare but that the inspection cannot count them, and a cross-section or X-ray sample from the same web is needed to estimate the hidden fraction.
Artefact — edge curl
H4225. Curl at a coated edge is attributed to surface tension, but a through-thickness drying stress gradient can produce the same sign and a similar magnitude. Coating the same formulation at two oven temperatures and measuring curl at fixed residual solvent separates them, because the stress gradient scales with drying rate while surface tension does not.
Scaling — Péclet mismatch
H4226. A lab coater that dries a film in a few seconds and a production oven that takes tens of seconds operate at different Péclet numbers, so the same formulation can have a uniform binder profile in one and a migrated profile in the other. Scaling by Péclet number, rather than by line speed or oven length separately, is the only way to claim that a lab coating represents production.
Limiting case — thick-coating collapse
H4227. Above a thickness where the drying front becomes skin-controlled, cracking and binder distribution stop depending on the formulation's chemistry and start depending only on shrinkage stress and the skin's permeability. In that limit, comparisons among formulations made on thick films are comparisons of one failure mode, and thin-film data are needed to recover the chemistry. Only thin films can then rank one formulation against another.
Causal chain — mud cracking
H4228. Cracking in a dried electrode is often attributed to poor adhesion to the foil, but a free-standing film of the same formulation can crack with the same spacing, which identifies constrained shrinkage rather than interfacial failure as the cause. The adhesion hypothesis predicts that a release layer should eliminate cracking, while the shrinkage hypothesis predicts it should not change the spacing much.
Symmetry — one-sided drying
H4229. Drying a coated foil from one side imposes a through-thickness solvent gradient that cannot be symmetric, so the two faces of a single-sided coating should differ in binder content and porosity. Flipping the web in the oven, or drying both sides simultaneously, should change the face-to-face difference only if the gradient is significant, so the size of the change measures how much the second side matters.
Competition — settling versus convection
H4230. Particles above a size set by the ratio of settling velocity to the drying-front velocity segregate toward the substrate, while smaller ones are carried with the receding meniscus. A bimodal formulation with the same mean size but different coarse fractions should therefore show different through-thickness compositions, a formulation variable that is almost never recorded.
Inversion — solids and uniformity
H4231. Raising solids is usually justified as a way to reduce drying energy, but it also suppresses convection and sedimentation in the wet film, so a higher-solids coating can be more uniform even though it loads the die more heavily. The trade is not monotonic: beyond the point where the slurry stops flowing uniformly in the die, uniformity collapses, so the optimum belongs to the die as much as to the formulation. The die sets the upper end of the window; the chemistry sets the lower end.
Confounding — foil roughness descriptor
H4232. Two copper foils with the same average roughness can wet differently if their roughness distributions have opposite skewness, because the wetting line pins on tall outliers rather than on the mean. A foil study that reports only Ra is therefore under-specified, and a profilometry metric weighted toward peak density should predict coating defects better than Ra does.
Count — two dimensionless groups
H4233. An oven has many setpoints, zone temperatures, air velocities, and dew points, yet the dried film's binder profile is governed by only two dimensionless groups, the Péclet number and an evaporation number. A model that fits a separate coefficient per zone is fitting the oven's wiring rather than the physics, and a two-group collapse should predict a new zone profile without refitting.
Path dependence — drying interruption
H4234. If drying is stopped and restarted, the first stage can form a skin that the second cannot redissolve, so the final electrode retains a memory of the interruption that no bulk measurement of residual solvent reveals. A controlled stop at different residual-solvent levels, followed by identical finish drying, should produce different adhesion and impedance even when the final solvent content is the same. Final inspection cannot see this memory.

Calendering and densification

Misattribution — capacity loss after calendering
H4235. The capacity loss observed after calendering is usually assigned to particle fracture, but part of it can be pore closure that leaves regions of the electrode unwetted, especially at high density. Replacing the electrolyte with one of lower surface tension, or calendering with a porogen that is later removed, should recover capacity without changing the particle-size distribution, and if it does, fracture was not the whole cause.
Inversion — calendering rate
H4236. Calendering should improve rate capability by improving particle contact, but it also raises ionic tortuosity, and the contact benefit saturates while the tortuosity penalty keeps growing. A density series measured at two rates should cross, with the high-rate optimum at a lower density than the low-rate optimum, an inversion that a single-C-rate study cannot see.
Threshold — fracture strain
H4237. Particle fracture should begin where the compressive strain in the coating exceeds the strain at which the particle's flaw population propagates, which makes the onset a function of particle size and internal porosity rather than of line load alone. A series that varies particle size at matched final density should shift the fracture onset in the direction predicted by the flaw distribution, and BET surface area before and after calendering is the readout.
Hidden term — elastic recovery
H4238. Gap-set and pressure-set calendering are not equivalent, because the electrode stores elastic strain and springs back after the nip, so the density measured offline is the unloaded density rather than the density the particles experienced. A gap-set and a pressure-set sample matched on final thickness should differ in particle contact and impedance if recovery is significant, and that difference is the hidden term.
Identity — fracture modes
H4239. Transgranular fracture and intergranular separation can produce similar laser-diffraction size distributions while creating very different new surface area, so a fracture study that reports only size may be reporting the wrong mode. BET surface area before and after calendering, paired with the size distribution, distinguishes them, since transgranular fracture raises area per unit size reduction more than intergranular separation does.
Conserved quantity — volume
H4240. Calendering conserves mass, so any thickness reduction must be accounted for as porosity reduction plus foil thinning, and a density increase larger than that balance implies either mass loss or a measurement that includes the foil. Running the balance explicitly on a coated sample before and after calendering is a cheap way to catch density numbers that are not physically possible. A five-minute balance prevents a class of impossible specifications.
Selection — edge-excluded density
H4241. Density is usually measured at the centre of the web because edges are difficult, but the edges experience a different roll gap and a different constraint, so the reported density is a selected statistic. A full-width density map on a production roll would show whether the centre is representative, and if it is not, the specification is controlling the mean of a distribution it never measures.
Scaling — roll diameter
H4242. At the same line load, a large production roll and a small lab roll have different contact lengths, so the electrode sees a different strain rate and a different residence time in the nip. Matching line load is therefore insufficient for scale-up, and contact length or strain rate should be the criterion, although the two cannot both be matched with a geometrically scaled roll.
Artefact — foil elongation
H4243. Part of the thickness reduction attributed to the coating is actually elongation of the foil, which is soft relative to the roll and can thin by a measurable fraction under calendering. Measuring the bare foil through the same nip, or measuring coating thickness by a method that references the foil surface, removes a systematic error that biases density high.
Saturation — high compaction
H4244. At very high compaction, electronic conductivity saturates because the particle network is already fully connected, while ionic tortuosity keeps rising as pores close. The last increment of density therefore buys almost no electronic benefit and costs rate capability, so the optimum density is one-sided and should be chosen at the knee of the conductivity curve rather than at its maximum.
Symmetry — two coating faces
H4245. Double-coated electrodes are usually calendered through one gap on the assumption that both sides see the same nip, but if the two coatings differ in initial thickness, the same gap imposes different compression on each side. Sectioning the electrode and measuring pore size on both faces should show a density difference, which would make the incoming thickness difference a hidden requirement of the calendering step rather than an independent coating specification.
Path dependence — calendering temperature
H4246. Calendering warm and calendering cold to the same final density can leave different binder distributions, because binder mobility during the nip determines whether it is pushed aside or deformed with the particles. Two temperature series matched on final density, one measured for adhesion and one for impedance, should separate the mechanical from the transport consequence of the binder's relocation.
Bound — adhesion
H4247. Peel strength cannot be increased indefinitely by calendering, because beyond a pressure where the coating is fully consolidated the failure path moves from the coating to the foil interface and the measured strength becomes a property of the foil's surface chemistry. A pressure series that plateaus has identified the bound, and further pressure only damages the foil. That plateau is a materials specification, and pressure beyond it is only damage.
Causal chain — springback
H4248. The density of a calendered electrode continues to change after the roll because the binder relaxes and the particles rearrange, so a density measured immediately and one measured a day later can differ. A specification written on the short-time unloaded density will therefore drift, and the drift should scale with binder content rather than with calendering pressure.
Counterfactual — re-porosification
H4249. If calendering's main benefit is particle contact rather than porosity, a calendered electrode that is subsequently perforated to restore some porosity should keep its low resistance while recovering rate capability. The experiment is a clean counterfactual, since perforation changes pores without changing contacts, and the direction of the result identifies which mechanism dominates.
Optimality — volumetric energy
H4250. The line load that maximizes volumetric energy density is not the line load that maximizes density, because rate capability falls as porosity closes and the cell's usable capacity depends on the rate it is asked to deliver. A cell-level calculation using measured rate-dependent capacity should place the optimum at a lower density than an electrode-level conductivity measurement suggests, and the gap between the two optima is the cost of ignoring the application.

Slitting, winding, and stacking

Misattribution — burr and blade wear
H4251. Burr height after slitting is usually tracked as a blade-wear variable, but it also depends on the coating's local hardness and on whether the cut passes through a particle or a binder-rich region. Cutting two coatings of different particle size with the same blade at the same wear level should produce different burr distributions, which would make blade-life specifications coating-specific rather than tool-specific.
Confounding — web tension
H4252. Slitting tension changes the stress state at the cut edge, so a web run too tight can produce microcracks that a slack web does not, even with an identical blade. The burr measurement is therefore partly a tension measurement, and a tension series at fixed blade and speed is needed before blade life can be interpreted.
Probabilistic threshold — burr penetration
H4253. A burr causes a short only if its height and sharpness are sufficient to penetrate the separator under stack pressure, so there is a critical burr size below which the defect is benign. That threshold should depend on separator thickness and modulus rather than on burr height alone, which means an absolute micrometre specification is measuring the wrong variable.
Inversion — dull blade
H4254. A slightly dull blade can produce a rounded edge that folds rather than fractures the coating, giving fewer free metal particles than a freshly sharpened blade that cuts cleanly but throws debris. If true, blade-life policy should aim for a controlled edge radius rather than maximum sharpness, and particle count per metre would be non-monotonic in blade wear. A radius target would serve better than a blade age.
Identity — burr versus particle
H4255. Edge burrs and embedded metal particles both produce self-discharge, and a cell-level test cannot tell them apart; only the location of the damage does. Dismantling failed cells and mapping damage relative to the electrode edge distinguishes the two mechanisms, and their ratio measures whether slitting or coating is the dominant contamination source.
Conserved quantity — particle balance
H4256. Every metal particle created by slitting must end up captured by a magnet, embedded in the coating, deposited on the web, or airborne, so a claim of zero free particles is a claim that one sink is perfect. Counting the sinks, magnet swabs, tape lifts, and filter residues, closes the balance and usually shows that the zero was a detection limit rather than an absence.
Selection — hi-pot alignment
H4257. A hi-pot test preferentially detects defects that happen to align with the applied field and are large enough to pass the current threshold, so the tested population is selected for a particular geometry of defect. Rotating the field or testing at two polarities should change which cells fail, and a cell that passes both is not necessarily defect-free but a cell whose defect is misaligned. The escape population is therefore geometric and can be estimated by rotating the field.
Artefact — winding tension
H4258. The deformation of a jelly roll after winding is often attributed to separator compression, but residual winding tension can produce the same ovality when the roll is released from the mandrel. Winding the same materials on mandrels of different stiffness, or with a controlled tension ramp, separates the two, because compression should not depend on the mandrel while residual tension should.
Scaling — winding speed
H4259. At high winding speeds, air is entrained between layers, and the amount depends on nip pressure and web speed in a way that winding tension alone does not capture. A dimensionless speed based on the entrained air film's thickness should collapse porosity-versus-speed data, and if it does, a lab winder run at low speed systematically overestimates the roll's final density.
Symmetry — winding direction
H4260. Reversing the winding direction reverses the sign of the bending stress in the electrode, so a symmetric cell should perform identically in both directions and any measured difference is evidence of separator or electrode anisotropy. The test is simple, and the size of the asymmetry sets a limit on how much cell-to-cell variation can be attributed to the winding process itself.
Path dependence — stacking sequence
H4261. In a z-fold stack a small misalignment is carried forward and can be partly corrected by the next fold, while in a pick-and-place stack each electrode is placed independently and errors do not accumulate the same way. The two processes should therefore produce different distributions of overlap area at the same placement accuracy, and that difference is a process signature rather than a quality difference.
Bound — stacking rate
H4262. The rate of stacking is bounded by how fast the separator can be handled without wrinkling rather than by how fast the electrode can be placed, because the separator is the compliant member and its wrinkles are irreversible. A stacking study that raises electrode placement speed while holding separator tension constant should show no gain beyond the separator's limit, which means throughput must come from web handling rather than from robotics. The separator's wrinkle threshold, not its tensile strength, is the number to specify.
Window — weld energy
H4263. Ultrasonic welding of tabs has an energy window: below it the interface does not form a metallurgical bond, and above it the foil work-hardens and cracks, so the window's width depends on foil thickness and anvil stiffness. A weld-strength map across energy and pressure should show a ridge, and the ridge's position is a more useful specification than a single energy setpoint.
Omitted variable — case stiffness
H4264. Two cells built from the same electrodes and separator but housed in cases of different stiffness experience different stack pressures, so their impedance and cycle life differ even though the electrode is nominally identical. Case gauge is therefore a hidden process variable, and a stack-pressure measurement on the finished cell should predict the performance difference better than any electrode measurement.

Electrolyte filling and wetting

Misattribution — wetting time
H4265. Differences in filling time between electrolytes are usually assigned to viscosity, but a low-viscosity electrolyte can fill more slowly if it traps gas in the separator's pores. Filling the same cells under vacuum and at ambient pressure separates the two, and a change in ranking identifies gas entrapment rather than viscosity as the controlling process.
Hidden term — smallest throat
H4266. The time to fill a porous electrode is set by the narrowest throat the wetting front must pass rather than by the mean pore size, because the front waits at the throat while the rest of the pore fills. A separator with a small fraction of tight pores can therefore dominate filling time out of proportion to its average porosity, which is why porosimetry that reports only a mean misses the manufacturing bottleneck. For a serial process, the mean is the wrong statistic.
Regime crossover — vacuum
H4267. Below a certain vacuum level, filling is limited by removal of gas from the pores; above it, the limiting process becomes capillary imbibition, and further vacuum buys nothing. The crossover should move with pore size and electrolyte surface tension, so a vacuum series at two pore sizes would locate it and identify which regime a given cell format actually operates in.
Non-monotonicity — hot fill
H4268. Heating the electrolyte lowers its viscosity and should speed filling, but it also lowers surface tension and increases gas solubility, both of which can slow the displacement of gas from fine pores. The filling rate should therefore peak at an intermediate temperature, and the peak should disappear in a gas-free system, which is a clean test of the mechanism. The peak is a process window and should be mapped, not avoided.
Conserved quantity — electrolyte ledger
H4269. The electrolyte added to a cell must be found in the electrode pores, the separator, the headspace, and the dead volume of the fill port, and any claimed starvation must close that ledger before it can be believed. Weighing the cell before and after filling, and measuring the displaced volume, is enough to detect the common error of counting the fill amount as the pore-fill amount.
Selection — easy-fill cells
H4270. Cells that fill easily are those with larger or better-connected pores, so a study that uses only well-wetted cells is selected on the very transport property it intends to measure. Filling a batch and comparing the slowest quartile with the fastest should show a rate-capability difference that is a pore-geometry effect rather than an electrolyte effect.
Artefact — pressure-driven uptake
H4271. In vacuum-assisted filling, the initial rapid uptake is often read as capillary wetting, but it is pressure-driven flow from the applied vacuum and stops when the pressures equalize even though the pores are not fully wet. A fill interrupted at the moment uptake stops and then held at ambient should continue to take up electrolyte if capillary wetting is incomplete, which distinguishes the two driving forces. The distinction decides whether the fill recipe needs more vacuum or more dwell.
Scaling — fill path
H4272. Filling time scales with the square of the longest path the electrolyte must travel, so a cell format doubled in linear dimensions takes roughly four times as long to fill at the same capillary pressure. That geometric penalty cannot be removed by an additive, and it explains why large-format cells need vacuum and long dwell times that small cells do not.
Identity — bubble versus dry spot
H4273. A trapped bubble and a dry spot both raise impedance and both look like a region of high resistance in an impedance map, but only the bubble moves when the cell is tilted or vibrated. A before-and-after orientation test on the same cell distinguishes them, and if bubbles dominate, the fill process rather than the electrode is the place to intervene. One cell and one minute are enough to run it.
Bound — diffusion length
H4274. Even with perfect wetting, the electrolyte's diffusion length limits how fast formation can proceed in a thick electrode, because lithium-ion flux must be carried through the pore network to the reaction front. The bound is set by porosity, tortuosity, and formation current, and it should appear as a maximum formation rate that is independent of bulk conductivity once the pores are saturated.
Path dependence — fill sequence
H4275. Filling under vacuum and then releasing the vacuum is not the same as releasing the vacuum and then filling, because the first pulls gas out of the pores before liquid arrives while the second traps gas behind the advancing front. The two sequences should leave different residual gas volumes at the same final electrolyte amount, which makes the fill sequence a process variable rarely recorded in the process route.
Symmetry — port location
H4276. Two fill ports placed symmetrically do not guarantee symmetric filling, because gravity and cell orientation break the symmetry: the lower port fills while the upper vents gas. Filling identical cells in two orientations and comparing their impedance and weight gain would show the asymmetry, and if it exists the second port is a vent rather than a fill path, so designing it as a fill path wastes electrolyte.

Formation, degassing, and aging

Misattribution — first-cycle loss
H4277. First-cycle irreversible capacity is assigned to SEI formation on the anode, but part of it can be oxygen loss and surface reconstruction at the cathode, which is not SEI and need not consume lithium. A cathode given a shallow conditioning cycle before being paired with a fresh anode should show a smaller first-cycle loss if the reconstruction term is real, and the difference is the cathode's contribution.
Inversion — formation rate
H4278. Slow formation is conventionally preferred for a dense SEI, but below a certain current the film can grow porous because solvent reduction at low overpotential produces a more organic and less compact deposit. The result is an optimum formation rate, and a rate series at fixed temperature and charge should show minimum first-cycle loss and best retention at an intermediate rate rather than at the slowest one.
Threshold — formation uniformity
H4279. Above a formation current set by the electrode's ionic overpotential, the SEI forms non-uniformly because the reaction front sees a distribution of local currents rather than a uniform potential. The threshold should scale with electrode thickness and tortuosity, so a thick electrode needs a proportionally lower formation current to stay below it, a constraint usually discovered only at scale. Because the constraint is geometric, it survives changes in electrolyte chemistry.
Confounding — rack position
H4280. Formation is exothermic, and heat is removed through tabs and surfaces, so cells at different positions in a rack experience different temperature histories under a single nominal setpoint. Instrumenting the rack at the same positions used for electrical testing should show whether the formation protocol is actually a temperature protocol, and if it is, rack design is a formation variable.
Conserved quantity — gas ledger
H4281. Gas generated during formation must end up as cell swelling, dissolved gas in the electrolyte, or gas lost through the seal, and a swelling measurement alone cannot distinguish them. A cell with known headspace, punctured under a calibrated collection volume, closes the ledger, and if the gas recovered is less than the swelling predicts, the difference is dissolved or leaked, which implicates the seal.
Selection — efficiency outliers
H4282. Cells with unusually high first-cycle efficiency are often kept as good, but a high efficiency can mean a thin or incomplete SEI that will grow later, so selecting on efficiency can enrich for unstable films. Following the high-efficiency outlier group through aging, rather than discarding it, would test whether the selection is inverted. The comparison is between outlier groups, not between means.
Artefact — voltage plateau
H4283. A voltage plateau during formation is often read as a phase transition, but a battery cycler with a slow current loop can produce a plateau that is the controller's response rather than the cell's. Running the same cell on two cyclers of different bandwidth, or with an added reference electrode, distinguishes a real thermodynamic plateau from a control artefact.
Scaling — formation time
H4284. Formation time scales with the square of electrode thickness if it is limited by electrolyte diffusion, but only linearly if it is limited by the interfacial reaction, so the scaling exponent identifies the limiting process. Measuring formation time at three thicknesses is enough to choose between the two, and the answer decides whether faster formation should come from thinner electrodes or from a better electrolyte. The exponent is a diagnostic rather than a fitting parameter.
Limiting case — low-current plateau
H4285. At very low formation currents the SEI approaches its equilibrium structure and first-cycle loss becomes independent of current, so further slowing buys no improvement in film quality. The current at which loss stops changing is the practical lower bound on formation time, and it should appear as a plateau in loss versus current rather than as a monotonic trend.
Path dependence — rest schedule
H4286. Two formation protocols with the same total charge and current but different rest schedules can produce different SEI compositions, because rest allows concentration gradients to relax and the next pulse to start from a different surface state. Holding charge and current fixed while varying only the rest pattern is the controlled test, and the resulting difference in first-cycle loss is the rest's contribution. Rest is an active ingredient of the protocol.
Symmetry — formation direction
H4287. Forming a cell by charging from a discharged state and forming it by discharging from a prelithiated state are not symmetric, because the SEI grows under different surface potentials and the two directions expose different facets. A prelithiated cell formed in discharge should show a different first-cycle loss and impedance from the same cell formed in charge, which makes formation direction a design choice rather than a convention.
Bound — heat removal
H4288. The throughput of a formation line is bounded by heat removal rather than by the cycler's current capability, because internal resistance turns formation current into heat that must be conducted out through a small tab cross-section. The bound scales with the tab's thermal conductance and the allowed temperature rise, so improving the cycler without improving the thermal path cannot shorten formation. The thermal path, not the cycler, is the design variable.

End-of-line testing and defect propagation

Misattribution — weld resistance
H4289. An elevated cell resistance is often charged to a poor weld, but current crowding in the foil near the tab can produce the same signature without any weld defect. Measuring resistance with the probe at two distances from the tab separates the weld's contribution from the foil's spreading resistance, because only the latter changes with probe position.
Artefact — fixture pressure
H4290. The contact resistance measured at end of line includes the fixture's probe-to-tab interface, which depends on probe pressure and on the tab's oxide layer. A four-wire measurement at two pressures, or a fixture with a Kelvin probe, should show a pressure-dependent offset that belongs to the fixture, and without that control cell-to-cell comparisons contain a term larger than the effect being studied.
Critical-size threshold — defect propagation
H4291. A defect propagates only if the local current density it concentrates exceeds a threshold that raises local temperature enough to accelerate adjacent degradation, so there is a critical defect size below which the cell lives normally. The critical size should scale with aspect ratio and cooling, which means a defect specification written as an area is less predictive than one written as a local current density. A current-density map would be a better specification than a size cutoff.
Inversion — self-discharge screening
H4292. A cell with a large but passivated defect can pass a self-discharge screen while a cell with a small active defect fails, so the screen selects for passivation rather than for defect size. The passed large defects can later reactivate under cycling or thermal stress, which means the screen's survivors are enriched in a failure mode the screen cannot see, an inversion of the intended selection.
Identity — two capacity-loss routes
H4293. Capacity lost during storage can be lithium consumed by SEI growth or charge lost to self-discharge, and both appear as a capacity offset at the next measurement. They differ in resistance, since SEI growth raises it while self-discharge leaves it roughly unchanged, so a storage series tracking capacity and resistance together separates the two without dismantling the cell.
Conserved quantity — capacity ledger
H4294. End-of-line capacity must equal formation capacity minus the capacity consumed by SEI, self-discharge, and any irreversible structural change, and a discrepancy larger than measurement uncertainty means one term is missing. Closing the ledger on a sample of cells is a better audit of formation than a capacity specification, because it localizes the loss instead of only flagging it.
Selection — grading noise
H4295. When capacity bins are narrower than the gauge's repeatability, a meaningful fraction of cells is assigned to a bin they do not belong to, and the bins' later performance regresses toward the mean. Widening the bins to the gauge's resolution, or grading on a second independent measurement, would test whether the apparent bin-to-bin life difference is real or an artefact of assignment noise. The bin boundary is a measurement, not a property.
Artefact — cable inductance
H4296. At the high-frequency end of an impedance spectrum, cable and fixture inductance can dominate measured reactance, so the apparent high-frequency resistance is partly the leads rather than the cell. Measuring a shorting block of the same geometry, or shortening the leads and checking that the intercept moves, is necessary before any high-frequency feature is assigned to the cell.
Scaling — inspection speed
H4297. The smallest defect an optical or X-ray inspection can reliably detect grows with line speed, because exposure time falls and motion blur rises. The detection limit is therefore a function of throughput, so a plant that raises line speed without changing inspection loses sensitivity to exactly the small defects that propagate, and the escape rate rises without any change in the process. Inspection speed and process speed are one specification.
Limiting case — relaxation convergence
H4298. After a long enough rest, cells formed under different protocols converge to nearly the same open-circuit voltage and capacity, because concentration gradients relax and the measurement loses the differences the protocols created. Comparisons made after a long rest therefore understate formation differences, and the informative measurement is the one made before relaxation rather than after it. A short rest preserves information that a long rest destroys.
Counterfactual — formation as wetting
H4299. If formation's main effect were wetting rather than SEI formation, a cell held at open circuit for the same time and temperature with no current should reach a similar impedance and first-cycle loss. The counterfactual is easy to run, and its outcome decides how much of formation is electrochemistry and how much is just time, a distinction that changes whether formation optimization should target current or dwell.
Optimality — test time
H4300. End-of-line test time trades escape rate against throughput, and the optimum is not the longest test the equipment can afford but the point where the marginal defect detected per hour equals the marginal cost of the test. Because defect prevalence varies by batch, the optimum test time should vary by batch too, which means a fixed protocol is either over-testing clean batches or under-testing dirty ones. The protocol should follow the batch's defect prevalence rather than a constant.

Form distribution

18 form families across 100 entries: Threshold 9 · Artefact 8 · Hidden variable 8 · Inversion 8 · Conserved quantity 7 · Misattribution 7 · Scaling 7 · Selection 7 · Bound 6 · Identity 6 · Path dependence 6 · Symmetry 6 · Limiting case 5 · Causal chain 3 · Counterfactual 3 · Optimality 2 · Competition 1 · Count 1.

Status and plan

100 of the series, 4201–4300. Each was written individually and given its own logical form; the labels differ structurally rather than only topically, and the distribution is reported with the file rather than hidden in the prose. Every entry is a proposed test, not a result.

The value of a hypothesis list is the ideas, and the value of a test is separate. A reader who wants one experiment from this batch should take the one below, because it is cheap and it discriminates between two mechanisms rather than between two materials.

If one of these is worth an experiment, it is H4249 — the re-porosification counterfactual. It needs a calender, a perforation tool, and two impedance measurements, and it decides in an afternoon whether calendering works through particle contact or through the pore network it closes.

Incoprea Research · hand-considered hypotheses 4201–4300 · 3 November 2026 · ~/Desktop/Nov3/