# Nine research/testing dossiers

No physical experiments have been performed. All models are conditional; no global novelty is certified.

# S1-0014 — Trapped-air capillary delay

A curved capillary branch should delay liquid invasion through air compression before breakthrough.

**Why selected.** Clear sealed-versus-vented intervention, inexpensive transparent apparatus, and a parameter with a model-predicted sign. Strong prior art makes this a replication/extension candidate rather than a novelty claim.

**Design.** Fabricate a transparent curved branch of measured area A with exchangeable terminal air volumes V0 and a separate vent. Record meniscus position at sufficient frame rate to resolve the slowest observed front. Keep liquid batch, branch surface preparation, inlet head, temperature, and starting meniscus fixed. Use at least five independently assembled branches per design, randomize test order, and repeat three starts per branch. Determine final replication after a variance pilot; these counts are planning choices, not a power calculation.

**Endpoints.** Primary: time to a preregistered meniscus station, seconds. Secondary: arrest distance, pressure at release, and residual gas volume. Define failure before testing: no positive relationship between terminal-gas confinement and delay above measurement uncertainty.

**Controls.** Same channel vented; sealed straight branch with equal hydraulic resistance; known external pressure ramp; leak test; independent contact-angle and temperature checks. Curvature itself may alter wetting, so do not attribute a sealed/vented difference uniquely to curvature.

**Conditional analysis.** Assume isothermal ideal gas, no dissolution/leakage, uniform area, and meniscus advance x with 0<Ax<V0. Then p_g(x)=p0*V0/(V0-Ax), and dp_g/dx=p0*V0*A/(V0-Ax)^2>0. Thus gas opposition increases with invasion within this assumed model. This derivative does not prove delayed breakthrough in a real curved branch: viscous losses, dynamic contact angle, compliance, and finite forcing determine whether breakthrough occurs at all. Adiabatic and dissolving-gas alternatives must be fitted separately.

**Analysis plan.** Fit branch-level delay versus V0 using uncertainty in position and pressure; report raw traces, confidence intervals, and all non-invasion trials rather than discarding them. Compare sealed/vented contrasts within matched branches. Reject the claimed useful effect if the interval includes the preregistered minimum useful delay or if a leak/wetting explanation accounts for it.

**Limitations.** Unproved physical proposal. Literature already supports the broad trapped-gas mechanism; no discovery or global novelty claim. No physical data were collected.

**Primary source/background.**

[Dynamics of a Capillary Invasion in a Closed-End Capillary](https://pubs.acs.org/doi/10.1021/la501927c). Publisher abstract read. The authors report theoretical and experimental closed-capillary invasion with compressed gas; their predicted early oscillations were not observed experimentally. This is direct related work and a warning against treating an ideal transient model as physical proof.


# S1-0039 — Opposed diffraction channels for bend/temperature separation

A flexible optical grating on a cylindrical support should distinguish bending from uniform temperature through opposite-side diffraction shifts.

**Why selected.** A paired-signal experiment admits a transparent identifiability calculation and direct nuisance-variable controls.

**Design.** Place nominally matched diffraction films on opposite sides of a cylindrical elastic support. Illuminate at fixed geometry and track diffraction angles with a calibrated camera. Independently vary curvature using a measured bending jig and temperature in a slow, uniform enclosure; include a crossed grid of curvature and temperature conditions. Use five independently mounted pairs and randomized condition order, followed by held-out combined perturbations.

**Endpoints.** Primary: held-out curvature error at changing temperature. Secondary: temperature estimation error and condition number of the measured 2x2 sensitivity matrix. An initial design goal is at least a twofold reduction of thermal cross-error against either channel alone; this is a proposed criterion, not an achieved result.

**Controls.** Rigid unbent reference, reversed illumination, exchanged top/bottom films, thermal-gradient measurement, camera scale reference, and independent curvature gauge. Track adhesive creep and support expansion separately.

**Conditional analysis.** For ideal matched channels y+=a*T+b*k and y-=a*T-b*k, T=(y++y-)/(2a) and k=(y+-y-)/(2b), provided a and b are nonzero. The determinant is -2ab, so the assumed linear map is invertible. Real coefficients must be independently calibrated; unequal thermal coefficients and gradients leave residual error. The algebra proves only identifiability of this assumed map, not the physical symmetry or measurement accuracy.

**Analysis plan.** Estimate the full matrix on training conditions, propagate angle uncertainty through its inverse, and publish held-out prediction errors against single-channel baselines. Reject useful separation if the matrix is effectively singular at the measured noise level or if calibration fails across thermal cycles.

**Limitations.** Unproved geometry-specific claim; simultaneous sensing is established prior art. Source access was limited to metadata for this paper. No data or novel theorem claimed.

**Primary source/background.**

[Simultaneous Vector Bend and Temperature Sensing Based on a Polymer and Silica Optical Fibre Grating Pair](https://doi.org/10.3390/s18103507). Primary publication metadata located; PMC access challenged. The title and metadata establish prior work on simultaneous bend/temperature sensing using paired gratings. They do not establish performance of the proposed opposite-side surface grating geometry.


# S1-0065 — Constant-torque cam and compression spring

A cam-guided compression spring should deliver approximately constant output torque across a designed angular range.

**Why selected.** The energy calculation is unusually strong under explicit assumptions, while a low-cost torque test can expose frictional departures. This is an engineering verification direction with obvious prior art.

**Design.** Use a measured linear compression spring and a CNC/printed cam driven quasistatically through a roller follower. Select finite angle interval [0,Theta] and positive preload x0 to avoid a singular cam slope. Measure spring compression, angle, and torque on loading/unloading. Build five independent cams, test three slow speeds, and compare with a circular cam at matched stored energy.

**Endpoints.** Primary: torque range divided by mean torque over the central designed interval. Secondary: hysteresis-loop work, peak contact force, and wear-induced drift. Proposed success criterion: at least 50% less normalized torque variation than the circular control without loss of stable contact.

**Controls.** Bare spring calibration; circular cam; unloaded bearing torque; reverse rotation; roller versus sliding follower; repeated mounting. Preload and follower geometry must be measured rather than treated as exact CAD values.

**Conditional analysis.** For a frictionless quasistatic linear spring U=k*x(theta)^2/2. Choose x(theta)=sqrt(x0^2+2*tau0*theta/k), with k>0, x0>0, tau0>0. Then dU/dtheta=k*x*dx/dtheta=tau0 exactly. This proves constant generalized torque only in the assumed lossless spring-and-cam model. It does not prove a manufactured cam has that profile, negligible friction, linear stiffness, or acceptable contact stress. A passive mechanism cannot deliver positive work indefinitely without recharging the spring.

**Analysis plan.** Compare measured torque against tau0 plus independently measured bearing/friction contributions; plot complete hysteresis loops and uncertainty bands. Reject the physical claim if torque flatness or contact stability fails the preregistered criterion. Keep model predictions distinct from observations.

**Limitations.** The conditional calculation is standard energy mechanics, not an original mathematical discovery. Physical performance remains untested; substantial related prior art exists.

**Primary source/background.**

[Design and experimental validation of a cam-based constant-force compression mechanism with friction considered](https://journals.sagepub.com/doi/10.1177/0954406218806015). Publisher abstract read. Li, Cheng, and Xie describe cam/spring constant-force prototypes and explicitly treat sliding and rolling friction. This supports feasibility of the broader design family and directly limits novelty of a cam-shaped output proposal.


# S1-0401 — Dust versus displacement in an optical cavity

A passive optical cavity should distinguish dust deposition from mirror motion through unequal changes in linewidth and resonance frequency.

**Why selected.** Two different observables may separate geometric path change from optical loss, but the proposal has a clear confounding test and could fail decisively.

**Design.** Use an enclosed low-power scanned optical cavity with a calibrated translation stage on one mirror. Introduce removable inert dust-coated witness surfaces at controlled positions without aerosolizing particles. Independently vary displacement and deposition, then test mixed conditions. Fit resonance center, linewidth, and contrast from the full scan rather than a single intensity reading. Use independently prepared witness samples and blind the mixed-condition labels during fitting.

**Endpoints.** Primary: misclassification rate between dust-only and displacement-only perturbations matched in peak-intensity effect. Secondary: displacement error under dust and sensitivity-matrix conditioning. Predefine a useful error target after measuring instrument noise, before testing mixed cases.

**Controls.** Clean witness exchange; laser-frequency reference; angular misalignment; source-power drift; temperature variation; deliberate scatter-free displacement. Dust can shift phase and alignment as well as loss, so linewidth is not assumed dust-specific.

**Conditional analysis.** Locally let the observable vector be y=(resonance shift,linewidth change) and perturbations z=(displacement,dust load). If the calibrated Jacobian J has nonzero determinant, z is locally identifiable in the linear approximation: z=J^-1*y. With noise covariance Sigma, propagated covariance is J^-1*Sigma*(J^-1)^T. This is a conditional inversion result, not a theorem that dust and displacement yield independent signatures. A nearly singular J invalidates useful separation.

**Analysis plan.** Estimate J from separate perturbations, evaluate singular values, then test held-out mixtures and report confusion matrices with uncertainty. Reject the proposal if angular or thermal controls recreate the dust signature or if held-out estimates are unstable.

**Limitations.** Unproved physical claim. Retrieved primary background treats cavity dynamics, not the proposed dust classifier. No physical data, classification success, or novelty established.

**Primary source/background.**

[Dynamic response of a Fabry–Perot interferometer](https://opg.optica.org/josab/abstract.cfm?uri=josab-16-4-523). Publisher abstract read. The authors predict and measure cavity-field response to cavity-length and laser-frequency changes. This supports length sensitivity as background, but supplies no direct dust/displacement discrimination evidence.


# S1-0413 — Phase-shifted heating for thermal-contact identifiability

A phase-shifted heating pair should identify thermal contact resistance independently of sample conductivity.

**Why selected.** The core uncertainty is parameter identifiability, which can be tested before expensive apparatus construction.

**Design.** Clamp two benign solid coupons at controlled force and use two small resistive heaters with independent sinusoidal phase control. Measure surface temperature amplitudes/phases at several positions using calibrated thermistors or an infrared camera with emissivity controls. Vary interface resistance with measured spacer films and separately vary bulk thickness. Pilot frequencies should span the observed thermal diffusion time; no single frequency is assumed sufficient.

**Endpoints.** Primary: held-out error of inferred contact resistance when bulk conductivity is independently changed. Secondary: parameter covariance and smallest sensitivity singular value. Proposed benefit: narrower uncertainty than in-phase heating at equal total heat input.

**Controls.** Single heater; in-phase heating; bonded reference contact; known thermal resistance standard; repeated remounting; convection shielding; heater-power measurement. Check that temperature modulation stays in the linear-property regime.

**Conditional analysis.** For a discretized linear heat model C*Tdot+K(k,R)*T=q(phi,t), local identifiability requires independent columns of the sensitivity matrix S=[dT/dk,dT/dR]. If S has full column rank, S^T*W*S is positive definite for positive-definite measurement weights W. This is proved by v^T*S^T*W*S*v=(S*v)^T*W*(S*v)>0 for nonzero v. Phase shifting may change S but does not guarantee full rank or good conditioning. No geometry-independent physical theorem is defensible here.

**Analysis plan.** Run sensitivity calculations with measured geometry, report singular values versus frequency and phase, then lock the chosen protocol before blind standards. Fit competing boundary-loss models and report residuals. Reject the claimed separation if confidence regions remain elongated or standards fail recovery.

**Limitations.** Unproved physical protocol. Primary related source was metadata-only; no claim about its results. Proposed simulations, if later run, must be labeled synthetic, and cannot establish experimental validity.

**Primary source/background.**

[Electrothermal static contact: inverse analysis and experimental approach](https://www.tandfonline.com/doi/full/10.1080/17415970903036777). Primary title/URL metadata; full page denied access. Related inverse contact-analysis work was located. Its inaccessible contents are not used as evidence that the proposed two-heater phase protocol succeeds.


# S1-0701 — Contact-triggered sign change in transverse lattice response

A folded lattice should develop negative incremental Poisson response only after its hinges contact.

**Why selected.** Direct imaging can locate contact onset and compare it with a measurable strain sign change. Strongly related prior art makes transparent replication essential.

**Design.** Build planar folded cells with a measurable hinge-contact gap and low-friction joints. Compress a multicell sheet while imaging fiducials. Compare three initial gaps and a no-contact control whose otherwise matched hinges cannot touch within the stroke. Use multiple independent sheets and track boundary versus interior cells.

**Endpoints.** Primary: incremental transverse/axial strain ratio before and after first interior contact. Secondary: contact strain, force slope, hysteresis, and repeatability. Define the Poisson sign convention explicitly: nu=-d(e_transverse)/d(e_axial). Success requires a reproducible sign crossing tied to contact, not merely a positive/negative total-strain ratio.

**Controls.** Gap-blocked control; lubricated versus dry contacts; alternate loading direction; camera distortion check; rigid-body motion removal; boundary-cell exclusion specified in advance. Contact friction and out-of-plane buckling are competing causes.

**Conditional analysis.** For a one-degree-of-freedom hinged model with axial length L(q) and width W(q), incremental nu=-(L/W)*(dW/dq)/(dL/dq) when dL/dq is nonzero. Contact changes the admissible kinematic branch; a sign reversal requires the derivative ratio to change sign. Contact alone does not imply that condition. This relation is a model identity, not a theorem that the constructed lattice will become auxetic.

**Analysis plan.** Estimate local derivatives using preregistered smoothing windows and report sensitivity to window choice. Align curves by measured contact onset, retain failed sheets, and compare sign-change confidence intervals against controls. Reject the hypothesis if sign change precedes contact or disappears after controlling out-of-plane deformation.

**Limitations.** Unproved geometry-specific extension; closely related published mechanism is known. No new theorem or measured auxeticity claimed.

**Primary source/background.**

[Tunable tension–compression asymmetry and auxeticity in lattice structures by harnessing unilateral contact](https://www.sciencedirect.com/science/article/pii/S0263822321011600). Publisher search-accessible abstract and highlights read. Peng and Bargmann report lattice contact changing connectivity and tunable Poisson behavior, including sign switching. The broad mechanism is already published; the proposed folded geometry is not established as novel.


# S1-0714 — Tendon-opened vent for low-work suction release

A soft suction foot should detach with lower energy when a tendon opens an edge vent.

**Why selected.** Release work is directly measurable and the intervention is reversible, making strong matched controls possible.

**Design.** Mount an elastomer suction foot on smooth nonbiological test plates. Establish equal initial internal pressure and contact area, then compare axial pulling with tendon-actuated edge venting. Measure actuator force-displacement, pressure, detachment timing, and total work including tendon work. Test multiple independent feet and randomized plate conditions.

**Endpoints.** Primary: total mechanical release work, joules. Secondary: peak plate load, release time, and reattachment reliability. Proposed success criterion: lower total work and lower peak load than axial release at equal initial holding force, without increased release failures.

**Controls.** A tendon routed without opening a vent; an externally opened vent; unchanged cup with axial pull; roughness and wetness blocks; leak-rate baseline. Include stored elastic energy and tendon friction in the accounting.

**Conditional analysis.** For a fixed-area ideal cup, the suction contribution to holding force is Delta_p*A. If opening a vent monotonically drives Delta_p toward zero before separation, that ideal contribution falls. With a simple isothermal fixed-volume conductance model d(Delta_p)/dt=-Delta_p/tau, Delta_p(t)=Delta_p0*exp(-t/tau). This proves decay only for that model; deformation, seal peeling, trapped liquid, and tendon losses can reverse the total-work comparison. No unconditional release-energy theorem is claimed.

**Analysis plan.** Integrate measured force over displacement for every actuator, subtract only separately measured instrument baselines, and compare paired total-work distributions. Report cases where the tendon fails to admit air. Reject the useful effect if venting lowers axial force but raises total work or damages repeatability.

**Limitations.** Unproved device-specific performance; suction release and self-sealing grippers have extensive prior art. No human-subject testing is proposed.

**Primary source/background.**

[Adaptive Self-Sealing Suction-Based Soft Robotic Gripper](https://advanced.onlinelibrary.wiley.com/doi/full/10.1002/advs.202100641). Publisher abstract and release-related sections read. Song and colleagues demonstrate a conformable membrane suction design and discuss controlled contact mechanics. This supports the broader suction-gripper context, not the energy benefit of the particular tendon vent.


# S1-0746 — Pressure-cycle mechanical counter

A compliant stepping mechanism should count pressure cycles without electronic power.

**Why selected.** A discrete state-transition claim permits exhaustive bench challenges and exact conditional reasoning, even though the broad pneumatic-step principle is established.

**Design.** Build a diaphragm-driven pawl with a compliant detent and a visible indexed wheel. Drive one inlet with recorded pressure cycles and read the wheel optically only for validation. Vary amplitude, frequency, rise time, dwell, partial unloading, and missed-pulse gaps. Use at least five independent counters; specify the operating envelope before endurance testing.

**Endpoints.** Primary: count error after a preregistered sequence of 1000 full cycles. Secondary: skipped/double steps, minimum reset pressure, wear drift, and energy per step. Proposed target is zero observed count errors within the declared envelope; this is not proof of a zero underlying error rate.

**Controls.** Subthreshold pulses; long sustained pressure; double peaks without full reset; reversed pressure; orientation changes; externally imposed vibration. Record actual inlet pressure rather than commanded values.

**Conditional analysis.** Assume each valid rising excursion moves a pawl exactly one index, a detent prevents backslip, and reset returns the pawl without moving the wheel. By induction, count after n valid cycles equals initial count+n modulo wheel capacity: the base case is immediate, and each transition adds one. This is a conditional state-machine proof, not a proof that the compliant hardware satisfies the transition assumptions. If zero errors occur in N independent trials, the one-sided 95% binomial upper bound is 1-0.05^(1/N); independence must be justified.

**Analysis plan.** Publish full pulse traces and wheel-state video, classify every error, and report error bounds by operating condition. Do not pool dependent repeated cycles as independent devices. Reject the design if partial pulses create false counts or reset requires external intervention.

**Limitations.** Unproved physical counter; no electronics-free performance demonstrated here. Prior pneumatic stepping makes global novelty implausible without much deeper searching.

**Primary source/background.**

[A New Type of Motor: Pneumatic Step Motor](https://pubmed.ncbi.nlm.nih.gov/21528106/). Primary article abstract and bibliographic metadata read. Stoianovici and colleagues report discrete pneumatic stepping with sequenced ports; the described system also includes electronic control and optical encoding. It supports pneumatic discrete motion but does not validate a single-port compliant passive counter.


# S1-0786 — Pressure-volume loops as delamination precursors

A soft actuator should reveal internal delamination through a changed pressure-volume loop before gross rupture.

**Why selected.** Damage can be introduced at known locations, permitting blind detection and explicit comparison with nuisance drift.

**Design.** Prepare matched soft actuator coupons with intact bonds, seeded small debonds, and sham fabrication changes. Measure pressure and displaced volume during slow inflation/deflation using a calibrated syringe system. Quantify debond area independently by transparent imaging or sectioning separate witness coupons. Blind the loop analyst to damage labels and reserve independent actuators for held-out testing.

**Endpoints.** Primary: held-out detection of seeded debonds below a preregistered area threshold. Secondary: loop area, tangent compliance, residual volume, and false alarms caused by temperature or cycling. No detection threshold should be tuned on held-out devices.

**Controls.** Temperature cycles; intact fatigued samples; controlled external load; small leaks without debonding; varying inflation rate; sham bond inserts. Leakage, viscoelasticity, and trapped gas can all change loop shape.

**Conditional analysis.** In a quasistatic elastic approximation, tangent compliance C=dV/dp is positive for a stable chamber. A debond may increase effective free wall area and hence C, but it can also constrain motion or cause leakage; no universal monotonicity theorem follows. Loop work W=integral p dV quantifies net pressure-volume work over a cycle but does not identify its dissipation mechanism. The physical diagnostic therefore remains unproved.

**Analysis plan.** Predefine loop features from a training set, evaluate held-out device-level sensitivity/specificity with uncertainty, and compare against leak-only and temperature controls. Track detection time relative to independently observed crack growth. Reject the early-warning claim if detection occurs only after visible leakage or if false alarms are comparable to true positives.

**Limitations.** No physical data or successful classifier generated. Existing endurance research is background, not proof of the proposed precursor. Independent-device validation is required.

**Primary source/background.**

[Endurance tests for a fabric_reinforced inflatable soft actuator](https://www.frontiersin.org/journals/materials/articles/10.3389/fmats.2023.1112540/full). Publisher endurance/burst-test methods read. The study evaluates cyclic use, burst behavior, and material properties of reinforced inflatable actuators. It motivates durability measurements but does not establish that pressure-volume loops uniquely detect early delamination.
