From a research thesis to a testable workbench.

Research audit and decisions

MAXWELLX — RESEARCH AUDIT AND DECISION MEMO
21 October 2026 | Research engineering handoff | Review scope, not independent replication

DECISION
Build a measurement-first workbench around Helios attribution, conservation accounting and controllable electrical boundaries. Use Casimir calculations as a separate, explicitly scoped modeling track. Maintain a mechanism-development notebook for propulsion and communication; neither is a demonstrated device program on the evidence reviewed here.

WHAT THIS AUDIT DID
Rechecked the earlier prompt's principal citations against publisher pages, author preprints and primary-paper records. The registry contains 19 literature records: 10 with 2026 journal/preprint dates and 9 earlier foundational records. These are assessed records, not 19 full-text validations. Access ranges from inspected methods to abstract-only and metadata-only. R08 is cross-checked against its author preprint but remains abstract-level; R19 is historical bibliography only. R11's full text was blocked; R03's full text requires subscription. Access limitations are explicit in evidence.json.

This audit does not repeat or certify the earlier claim of searching 195 countries. Country counts are not a measure of scientific validation. Only countries verified from affiliations are populated; empty lists mean unresolved affiliation extraction, not no country. The verified subset includes South Korea, Switzerland, Japan, the United States, Israel, the United Kingdom, Belgium and Canada. This is an international targeted review, not an exhaustive country-by-country census. No paper was experimentally replicated, no author contacted, and no MaxwellX raw observation was supplied.

VALIDATION HAS FOUR SEPARATE LEVELS
1. Bibliographic: title, authors, DOI and publication status match a primary record.
2. Claim: the inspected text actually supports the stated result under its conditions.
3. Numerical: a specified implementation passes analytic limits, conservation checks and convergence checks.
4. Physical: calibrated measurements with appropriate controls replicate the prediction.
Only levels actually completed may be shown. A journal publication is not independent replication. A code test is not a physical result. A publisher abstract is insufficient for component-level build instructions.

CORRECTIONS TO THE EARLIER HANDOFF
• R05 is now identified as Phys. Rev. D 114, 025016 (2026), DOI 10.1103/vn2g-ryw1. The old preprint-only classification is stale.
• R01's half-cell current is bias-assisted. Keep it separate from its full-cell polarization peak. Do not combine open-circuit voltage and short-circuit current into a claimed operating power.
• R06 demonstrates time refraction; its discussion of future optical time reflection is not a completed observation of that phenomenon.
• R09's reported signal gain is not a whole-system energy surplus. Optical modulation and bias belong in the resource accounting. It remains a preprint in the inspected record.
• R08 is a theoretical materials-control direction. Do not describe it as experimental confirmation or a spacecraft mechanism.
• The MaxwellX ideal-plate page currently labels its model educational and excludes real material corrections. This package adds a restricted analytic-response planar solver, not a calibrated model of the actual MaxwellX apparatus.

THE SEVEN QUESTIONS — ENGINEERING DISPOSITION

1. Quantum inequalities and propulsion [R04, R12, R13, R16]
The requested 'bypass' is not a specified mechanism. Before computing any inequality, state the field, quantum state, observer, sampling function, boundary conditions and reference energy. Then identify the applicable bound. For propulsion, define a control volume enclosing supports, power supplies and fields. Report external momentum flux, including radiation, alongside force. An asymmetric internal force distribution is not a calculation of sustained centre-of-mass acceleration. Nonequilibrium radiation is a distinct, resource-consuming possibility and deserves its own model. Deliverable: a momentum ledger and a research note, not a thrust slider derived from plate pressure.

2. Gap scaling and array geometry [R11, R18]
Ideal flat-plate pressure varies as d^-4, whereas energy per area varies as d^-3. Subdividing an area can help mechanical gap control but does not cancel the opposing reaction force. Ideal sphere-plane PFA is included only when d/R <= 0.01; that is a conservative software policy, not a universal 1% error guarantee. Pillar, hollow-cylinder and finite-array geometry need an appropriate scattering solver, converged basis/mesh and measured geometry. Their renderers must not produce guessed force multipliers. The supplied external-solver importer enforces provenance fields but cannot establish numerical truth by itself.

3. Conductivity and patches [R03, R15, R18]
The planar solver supports ideal, Drude, plasma and constant-dielectric response scenarios. It does not ship measured optical data for gold or any other actual coating. A device prediction requires epsilon(i xi), low-frequency extrapolation, thickness/substrate treatment, surface state, measured gap and an uncertainty budget. Patch estimates are a separate electrostatic calculation. Mean bias cancellation does not establish the absence of spatial patch fields. Inverse material inference should remain a later phase until forward modeling and experimental identifiability are demonstrated; fitting a curve alone can conceal parameter degeneracy.

4. Dynamic Casimir photon production [R05, R06, R09, R14]
Frequency matching is a kinematic condition; it is not a model of output flux or wall-plug efficiency. The package calculates pump-frequency matching, photon energies and thermal occupation. It makes no photon-count prediction. The additional classical LC model demonstrates modulation work explicitly and is not a quantum vacuum simulation. To progress toward photon statistics, obtain a sourced Hamiltonian/coupled-mode model, calibrated gain/noise chain, loss budget and pump-leakage controls. Cryogenic DCE hardware is not an appropriate first owned-bench purchase for this program.

5. Thermalization and array power [R07]
First name the usable resource: light, electrical modulation, a temperature difference, chemical free energy or initial stored energy. Then specify the replenishment process. Arrays need transport, thermal coupling and parasitic-load models. No efficiency parameter can substitute for this resource definition. The tank kernel accounts for finite storage, conversion losses, inverter losses, rate limits, curtailment and unmet demand. Unserved auxiliary power is exposed: a future process model must decide whether the biological process can continue under that condition.

6. Nonlocal communication [R17, R19]
For a trace-preserving local operation on Alice, Bob's reduced state is unchanged. The package computes this for valid two-qubit density matrices and Kraus channels. Postselected operations are rejected unless the conditioning information is treated separately. A demonstration of correlations is not a message channel. Wheeler–Feynman is a historically relevant theoretical reference, but the inspected record does not provide an engineerable instantaneous protocol. A research submission needs a transmitter variable, channel law, receiver observable, latency distribution and controllable bit test.

7. Helios controls and attribution [R01 and project records]
The currently accessible MaxwellX Casimir page separates Helios from Casimir and describes the initial response as unattributed to Casimir forces. Direct retrieval of the first-result and controls pages failed during this audit. Their earlier supplied descriptions are historical context, not a fresh verification of running tests. We cannot say which controls are operating now. Obtain raw files and apparatus metadata before making that claim. Biological electrochemistry is a possible mechanism, not something to remove by definition. The task is to distinguish biology-dependent conversion from abiotic chemistry, corrosion, stored charge, thermal effects and instrumentation.

PRODUCT OPPORTUNITIES — RECOMMENDATIONS, NOT MARKET VALIDATION
First: reproducible measurement analysis with visible provenance, control comparisons and honest exports. Useful internally immediately; external demand and willingness to pay remain untested.
Second: low-frequency programmable impedance experiments and validated design tools. A circuit that changes a predictable transfer response gives a tractable geometry/boundary result.
Third: geometry optimization of practical electrodes, after attribution and baseline stability. Optimize sustained net energy per illuminated projected area and lifetime, not maximum current alone.
Fourth: precision Casimir forward-modeling and eventually surface characterization. Attractive intellectually, but limited by materials data and instrumentation.
Hold: household energy claims, macro vacuum harvesting, reactionless thrust and instantaneous communication as product promises. They need mechanisms and measurements beyond the current package.

NEXT RESEARCH GATES
Retrieve full text/supplements for R02, R03, R08, R11 and R15 before any replication claim. Complete missing exact journal dates and affiliations in the registry; retain the audit history. Review independent replications and contrary results, not just the originating paper. Acquire or measure optical response data before material-specific Casimir predictions. Specify actual Helios electrodes, medium, culture, loading, measurement mode and instrument exports. None of these gaps is filled by synthetic demonstration data.

Primary-source register

Each record names what was inspected and what remains unverified. Links point to original papers or author preprint records.

Equations and implementation limits

MODEL CONTRACTS AND NUMERICAL VALIDATION
All inputs in SI except explicitly marked Wh/h dispatch quantities. Source: src/maxwellx.py, version 1.0.0. Read limitations before porting.

A. Ideal parallel plates
C0=pi^2 hbar c/720; U=-C0 A/d^3; P=-3 C0/d^4; F=P A; dF/dd=12 C0 A/d^5.
Zero temperature, perfect conductors, vacuum, negligible edge effects. Negative pressure means attraction. Differentiating U with respect to d recovers F. No power is calculated from a static gap. Independent small Gaussian input uncertainty gives sigma_F/|F|=sqrt((4 sigma_d/d)^2+(sigma_A/A)^2). Covariance and model error are excluded; inputs above 10% relative uncertainty are rejected. Mechanical stability output is only a local spring-gradient condition and assumes an equilibrium exists.
At 100 nm pressure is approximately -13.0013 Pa. At twice the gap its magnitude is divided by 16. Tests compare derivatives and analytic scaling, not only duplicated code expressions.

B. Ideal sphere-plane PFA
F=2 pi R (U/A)=-pi^3 hbar c R/(360 d^3). Restricted d/R <= 0.01. Real material corrections and quantified truncation error are absent. Arbitrary 3D arrays are not solved. External results require solver version, geometry, material provenance, convergence description, units and ordered numeric points. Import success never changes validation status to 'verified'.

C. Long-patch electrostatic limit
Pressure magnitude=epsilon0 Vrms_difference^2/(2 d^2). Vrms is the RMS difference between facing surface potentials in a long-correlation-length limit, not automatically the RMS on one measured surface. This is not a spectral patch model and should not be used to automatically subtract backgrounds. The 10 mV curve in the figure is an assumption.

D. Finite-temperature planar Lifshitz pressure
Identical local isotropic nonmagnetic semi-infinite half-spaces in vacuum. Let xi_n=2 pi n kB T/hbar, zeta_n=2 d xi_n/c, y=2 q d, w=sqrt(y^2+(epsilon(i xi_n)-1) zeta_n^2).
r_TE=(y-w)/(y+w); r_TM=(epsilon*y-w)/(epsilon*y+w).
P= -kB T/(8 pi d^3) sum_n' integral_{zeta_n}^infinity y^2 sum_p [r_p^2 exp(-y)/(1-r_p^2 exp(-y))] dy.
Prime halves n=0. Ideal r_TE^2=r_TM^2=1. Drude epsilon=1+omega_p^2/[xi(xi+gamma)], with n=0 TE=0, TM=1. Plasma uses gamma=0 and its finite TE zero-mode expression; it is not implemented as Drude with an arbitrarily tiny gamma. Constant dielectric uses n=0 TE=0, TM=(epsilon-1)/(epsilon+1). That dielectric scenario is a mathematical approximation, not a realistic all-frequency response dataset.
Quadrature and Matsubara truncation share y cutoff 50 by default. Re-run 40/60 and stricter quadrature tolerance for convergence. Returned quadrature error excludes truncation, material error and physical-model error. The included report checks a Drude scenario at 100 nm, 300 K, omega_p=1e16 rad/s, gamma=1e14 rad/s. These are hypothetical parameters. No gold label is permitted.
Analytic gates: ideal low-T/small-gap limit; ideal high-T pressure=-kB T zeta(3)/(4 pi d^3); Drude high-T half of ideal; no dielectric contrast gives zero; increasing plasma frequency approaches ideal. Positive T only; >10000 required terms fails explicitly. Nonlocal response, films, substrates, roughness, magnetic materials, curvature and arbitrary patterns are excluded. Call this numerically checked analytic-response solver, not a measured-material validated solver.

E. Energy and momentum
Residual=inputs-outputs-losses-DeltaStored. Named terms must cover one boundary and interval; null produces incomplete accounting. A negative DeltaStored is discharge. Zero balance does not prove all terms were measured. Three-sigma flag is only meaningful when the supplied combined standard uncertainty is justified, including covariance. Pure collimated emitted-radiation recoil is P/c; incident reflection is a different system. At 1000 W emitted power, recoil is 3.33564 microN.

F. Photon explorer
Pump f=f1+f2; photon E=h f; mean thermal occupation=1/expm1(h f/kB T), zero at T=0. No assumed conversion efficiency or flux prediction. Thermal occupation does not separate measured DCE photons from leakage or amplifier noise.

G. Helios CSV kernel
Canonical headers: time_s,voltage_V,current_A,run_id,condition. Additional columns retained only outside this analysis kernel in the original file. Strictly increasing elapsed seconds; one run per file. Hash original canonical text; retain the separately hashed instrument original when mapping units. Voltage and current must share a passive load and sampling epoch. For biased half-cells, integrated V*I is a terminal diagnostic and never delivered_load_energy_J. The UI must not provide passive-load confirmation by default.
Trapezoid integration excludes intervals touching missing samples, crossing condition labels or exceeding max_gap_s. It does not bridge outages, interpolate between runs or smooth. Coverage is covered duration/total span, not a count of samples. Negative powers are preserved; wiring sign conventions need review. This kernel does not compute biological attribution, confidence intervals or metrological uncertainty. High coverage does not eliminate aliasing; choose sampling based on instrument response and signal bandwidth.

H. Tank-to-home dispatch
Incident energy=irradiance * illuminated projected area * interval. Conversion efficiency is assumed, never inferred from tank volume. Direct DC generation serves auxiliary demand and then AC household demand through inverter. Excess charges battery subject to energy/rate limits; deficits discharge it. Charging and discharging are not simultaneous in an interval. Curtailment is explicit. Round-trip efficiency is eta_charge * eta_discharge, not either factor alone. Rate limits are at DC bus. No standby inverter drain beyond the specified auxiliary demand. No battery aging, temperature dependence, self-discharge or biological feedback. Piecewise-constant profiles require finer resolution for transients. Residual closes the electrical balance; incident optical energy minus generation is outside that electrical balance and remains a conversion loss.
Eight-foot diameter is 2.4384 m; top aperture is about 4.67 m2. Total inner electrode surface is not additional external solar aperture. A peak paper power density cannot be multiplied by arbitrary internal area to make a household-energy claim.

I. Communications
rho is a normalized Hermitian positive 4x4 density matrix; Alice indices come first. Local Kraus operators obey sum K†K=I. Bob is obtained by tracing over Alice. The before/after marginals agree for complete local channels. A postselection is not trace preserving and needs separately transmitted outcome information. Included tests use random mixed states and amplitude damping, not only a Bell state.

J. Classical temporal LC model
q'=i; L i'=-Ri-q/C(t); C(t)=C0[1+m sin(2 pi fm t)]. Energy=q^2/(2C)+Li^2/2. Modulation power=-q^2 C'/(2 C^2), resistor heat rate=R i^2. Therefore DeltaEnergy=modulation_work-heat. The numerical model integrates all terms and verifies the residual. It has one classical mode, no vacuum noise, no spatial propagation, no input source and no photon statistics. It is a preparatory energy-accounting exercise, not a replication of R02 or R09. A static impedance benchmark is also provided; it must never be labeled temporal reflection.

DEPENDENCIES AND REPRODUCIBILITY
Pinned runtime versions in requirements.txt are the versions actually used here, not claims about the latest releases. Run python run_validation.py from the package directory; Node must be available for browser parity tests. Tests use Python unittest. Figures and fixtures are rebuilt deterministically. See results/validation-summary.json and test logs. Browser tests execute ES modules under Node; live-site integration, accessibility and visual browser tests remain for the MaxwellX coding AI.

Experiments and equipment gates

EXPERIMENTS, EQUIPMENT REQUIREMENTS AND DECISION GATES
Planning estimates in USD, 21 October 2026. Estimates are engineering allowances, NOT vendor quotations. No purchases or researcher messages have been made. Build around owned equipment; do not depend on borrowing instruments or access to somebody else's laboratory. Exact models must be selected after signal and bandwidth requirements are confirmed.

TRACK H — HELIOS ATTRIBUTION BEFORE SCALING
Question: does a living preparation deliver sustained additional load energy relative to a matched organism-free preparation, and is the effect reproducible across separately prepared runs?
Primary outcome: signed delivered energy in a fixed predeclared passive-load interval, with duration, projected illuminated area and uncertainty. Secondary outcomes: drift, light/dark difference, electrode temperatures, pH, conductivity and oxygen where measurable. Open-circuit voltage is a separate diagnostic. A biased half-cell belongs in a separate series.

Apparatus record required before acquisition: existing organism/culture provenance; medium composition and age; cell volume; electrodes/material/area/spacing/preparation history; load; instrument make/model/range/integration time; wiring and shielding; illumination spectrum/irradiance; electrode temperatures; raw data export. Unknown fields remain unknown. No organism or electrode identity is inferred from the announcement.

Proposed sequence:
H0: Instrument qualification. Open input, shorted input, known reference signal and precision dummy load in the same wiring environment. Collect enough baseline data to quantify drift over the intended experiment duration. Compare ranges/integration times without selecting the most favorable trace. Save saturation/range-change flags.
H1: Matched passive-load cells. Prepare a live and no-organism blank with identical electrode history, medium volume, geometry, temperature and illumination. Randomize physical positions and channel assignment. Predeclare polarity convention, load resistance and analysis interval. Use the same timestamp system. Do not use one reference trace repeatedly as independent control replicates.
H2: Repeat preparations. A pilot of six independently assembled matched pairs across at least three preparation days is a planning starting point, not a power calculation. If the same culture batch supplies multiple pairs, preserve that clustering. Use pilot variance to set sample size for a confirmatory study before collecting it.
H3: Light/dark and thermal discrimination. Use predeclared cycles and equilibration windows. Log both electrode temperatures; include an abiotic heated/temperature-matched comparison where feasible. Light response alone cannot distinguish biology from photoelectrochemistry or heating.
H4: Stored-charge and corrosion checks. Extend passive loading; document rest periods and prior biasing; inspect electrode surfaces; assess whether a finite reservoir explains total charge. Weighing alone may lack sensitivity at the expected currents. If necessary obtain a separately quoted material analysis; do not silently mark corrosion excluded.
H5: Geometry experiment after baseline attribution. Change one geometry feature with matched material/preparation, fixed illuminated projected area and measured mass/loading. Include physical replicate identity and fabrication batch. Define an independent confirmatory comparison; avoid selecting the best geometry and reporting that same data as confirmation.

Analysis: integrate raw signed V*I with exclusions disclosed; keep zero/below-detection results. For paired runs, compute run-level energy differences over matched intervals. Confidence intervals must treat independent preparations as units and account for shared batch and drift; thousands of timestamps do not create thousands of biological replicates. Report effect size and uncertainty, not a single p value. The included kernel stops at run-level integration; confirmatory statistical modeling must be added after the sampling design is fixed.

Proposed gate (engineering choice, not a literature standard): do not advance to scale-up unless instrument checks pass, matched-run effect exceeds the expanded measurement uncertainty, the prespecified run-level interval excludes zero, and the direction repeats on a separate preparation day. These conditions support a narrower attribution claim; they do not prove the complete electron-transfer mechanism. Failed conditions trigger diagnosis, not deletion of inconvenient runs.

Owned-bench equipment specification and budget allowances:
• Low-current measurement: select range so expected signal is well above measured system leakage/drift, with headroom against overload. Guarded cabling, isolation and timestamped exports matter as much as display resolution. A Keithley 2450 is a candidate for characterization, not automatically sufficient for every sub-nA experiment. Its lowest-range accuracy includes offset terms; never treat 10 fA resolution as 10 fA accuracy. Allow $8,000–15,000 including relevant accessories/calibration, pending quote.
  Manufacturer datasheet: https://www.tek.com/en/datasheet/smu-2400-graphical-sourcemeter/model-2450-touchscreen-source-measure-unit-smu-instrument
• Biased electrochemistry, optional second acquisition: a true potentiostat with stable reference-electrode operation, documented low-current behavior, impedance capability if needed and exportable raw samples. Gamry Reference 620 is a candidate; manufacturer lists 11 current ranges from 60 pA to 600 mA. The older 600/600+ appear in legacy documentation, so do not assume current availability. Allow $15,000–30,000 pending quote. Do not buy both this and an SMU before resolving the initial measurement mode.
  Manufacturer: https://www.gamry.com/potentiostats/reference/reference-620/
• Matched vessels/electrodes, spares and fixtures: $600–2,000. Select exact materials from the existing apparatus record and corrosion analysis, not from the example paper alone.
• Calibrated illumination measurement and controllable source: $1,000–4,000. Sensor spectrum must match the source; a lux meter is not a general irradiance standard.
• Two electrode-adjacent temperature channels plus ambient logging: $300–1,200. Target resolution better than 0.1 C as a planning requirement; calibrate differential offset. Small resolution does not imply that thermal artifacts are excluded.
• Guarding/shielding/cables/load standards/reference source: $800–2,500. Characterize leakage with the actual wet setup.
• Basic chemistry logging and consumables: $500–2,000. Probe maintenance and calibration solutions included in allowance; analytical corrosion measurements may be separate.
• Calibration/service and shipping contingency: request quotes; reserve roughly 20% of the selected subtotal.
Starting allowance using one SMU path: about $13,400–32,000 including that contingency. This is not a purchase recommendation or a binding total. Existing qualified equipment can reduce the scope after its performance is documented.

TRACK W — PROGRAMMABLE ELECTRICAL BOUNDARIES
W0: Establish a static 50-ohm termination benchmark. Measure matched and intentionally mismatched loads, cable delay and reflections. The code gives Gamma=(ZL-Z0)/(ZL+Z0) only for real positive resistances; extend to complex impedances only with tested phase conventions.
W1: Measure the selected varactor's capacitance versus voltage, parasitic resistance, leakage and hysteresis over the chosen range. Include modulation-feedthrough measurements with the signal source off.
W2: Implement a low-frequency single-mode LC experiment, with current/voltage acquisition and modulation input measured. Compare the model's stored-energy change with integrated pump work and dissipation. This is a preparatory classical circuit study.
W3: Only after reviewing R02's complete circuit/supplement, define a distributed transmission line, cell pitch, effective-medium bandwidth, dispersion, termination, modulation synchronization and readout. Measure S-parameters in static bias states before time-varying trials.
W4: Separate the incident signal, reflected signal and modulation leakage using calibration and time gating. Compare static control, modulation-only, signal-only and synchronized combined runs. Report repeat runs and source/bias consumption alongside signal gain.

Owned-bench candidate for W0/W2: Keysight EDUX1052G, official page specifies 50 MHz and two analog channels with generator. This is a basic low-frequency candidate and is NOT adequate by assumption for the 200 MHz modulation preprint. Four synchronized channels or a separately synchronized digitizer may be required for full accounting.
https://www.keysight.com/sg/en/product/EDUX1052G/oscilloscope-50-mhz-2-analog-channels-waveform-generator.html
Plan $2,000–6,000 for scope/generator capabilities, $500–2,000 for probes, loads, fixtures and shielding, and $500–2,000 for prototype boards and spares. A distributed fast-modulation replication may require $10,000–30,000 or more in additional bandwidth/synchronization equipment; obtain a circuit-specific quote. Do not order from these allowances.

TRACK C — CASIMIR MODEL QUALIFICATION
C0: Run packaged analytic and convergence checks. C1: obtain material response data and its provenance; quantify low-frequency extrapolation uncertainty. C2: implement thickness/substrate transfer matrices and benchmark against an independent solver. C3: obtain a geometry-specific convergence study; compare PFA only within its demonstrated error. C4: design force metrology, electrostatic calibration, gap calibration, roughness measurement and patch characterization. The equipment specification depends on force range and geometry; no defensible complete hardware price exists yet. Keep hardware spend gated. An ordinary electrical bench does not establish nanometre-gap force metrology.

90-DAY SEQUENCE (RELATIVE TO ACTUAL START DATE)
Days 1–10: integrate provenance and unit validation; obtain Helios apparatus/data inventory; H0; finish missing primary-source methods. Exit: usable raw-data pipeline and identified instrument limitations.
Days 11–30: H1–H3 pilot and W0/W1 owned-bench characterization. Exit: measured repeatability, uncertainty and feasible bandwidth; select confirmatory endpoints and sample size.
Days 31–60: repeat independent Helios preparations; W2 energy-accounting test; C1 material-response ingestion if data available. Exit: model-to-data residuals and explanations of major confounders.
Days 61–90: confirmatory experiment defined before collection; one geometry change if earlier gates pass; review whether W3 is justified. Publish a methods/results note only for actually completed work. If controls or source access remain blocked, record a revised decision and continue the lower-dependency track.