VHP Villa Laboratory Build and Operating Concept

1 Villa layout and room allocation

The laboratory occupies the ground floor and one upper room of the villa property. Rooms were chosen so that gas, high voltage, and radiation each own a dedicated space with a dedicated wall to the outside.

Room-to-room rules:

• No hydrogen-bearing line enters the Control Room or Cell Room. Gas delivery to cells is through a single Swagelok stub line with excess-flow check valve, entering at the Gas Room/Cell Room wall.
• Detector signals travel Cell Room to Control Room on RG-223 coax through a bulkhead plate; all triggers traveling the other way are fiber optic.
• The Cell Room floor carries painted standoff rings at 1.0 m and 2.0 m radius from the primary cell position. Fixed detector mounts anchor at the 1.0 m ring: the moderated helium-3 counter north, the two liquid-scintillator tubes east and west at 1.0 m separation (time-of-flight baseline), the NaI(Tl) gamma detector south behind a shadow bar.
• The Drive Room and Cell Room share a single star ground point on the Cell Room wall: one 6 mm2 copper strap from each room's equipment racks to one grounded bus bar. Nothing else grounds to building steel through a second path.

2 Test cells

Two cell families, matching the two carrier systems of the concept: suspension cells (deuterated oil or glycerol, stage-two aqueous heavy-water electrolyte) and dry-bed cells (ex-situ deuterided powder plus piezoelectric powder in a sealed volume). Both families share ports, sensors, and mounting so the calorimetry and diagnostics never need re-rigging between variants.

2.1 Stainless suspension cells

• 316L stainless body, 2 L nominal, conflat (CF) flanged, DN100 CF main flange and DN40 CF side ports.
• A central dip tube carries the ultrasound sonotrode tip (Hielscher UP400St sonotrode, 24 kHz drive, see section 6) to 10 mm above the cell bottom; the dip tube seats through a metal-sealed feedthrough.
• Two opposed sapphire viewports (DN40 CF, 25 mm clear aperture) for optical observation: sonoluminescence imaging and scattered-light monitoring of suspension state.
• One DN40 CF port carries the headspace gas line (to mass spectrometry, section 4.5) with integrated filter to retain powder aerosol.
• One DN16 CF instrument well hosts a cartridge heater sleeve, type K thermocouple, and Pt100 probe in a single removable assembly.
• Fill fraction: 60 to 70 percent liquid, headspace at least 300 mL so gas sampling never sucks liquid.
• Working pressure 0 to 8 bar absolute (deuterium headspace over the suspension); test pressure 12 bar; burst margin beyond 24 bar by vessel design.

2.2 Quartz suspension cells

• Fused quartz, 1 L, wall 5 mm, rated 0 to 4 bar gauge at room temperature (manufacturer certificate required; quartz pressure ratings fall with temperature, so the quartz cells run liquid-only experiments near ambient).
• Optical access around 300 degrees of circumference; used for camera-based sonoluminescence and for watching shear-thickening state, clouding, and particle settling during excitation.
• Quartz cells sit inside the same outer containment vessel as the stainless cells (section 2.4) so that a quartz failure is a containment event, not a room event.

2.3 Dry-bed cells

• 316L, 1 L, quick-open clamp closure (KF/ISO class clamp with metal seal), designed for powder exchange inside the glovebox: the powder basket lifts out as a closed cartridge.
• Gas loading port, relief valve (10 bar), and the same instrument well as the stainless suspension cells.
• Bed composition per the concept documents: 50 volume percent metal deuteride powder (1 to 10 micrometer, TiD or ZrD primary, PdD reference), 10 volume percent piezoelectric powder (PZT, barium titanate, zinc oxide, lithium niobate across the four-composition matrix), balance deuterated carrier or dry argon headspace.
• Dry-bed cells are driven by the pulsed-discharge bank and by an external clamp-mounted piezo stack (section 6), not by the immersion sonotrode.

2.4 Outer containment and pressure ratings

• Every quartz and custom cell mounts inside an outer containment vessel: a Parr Instrument series 4520-class bench-top reactor vessel (1 L, manufacturer-rated to 1900 psi, about 131 bar) or the 4540-class 600 mL high-pressure variant where we need headroom. The Parr vessel provides the certified pressure boundary; the inner cell is an insert.
• Custom stainless cells (2.1) are their own primary boundary, hydrostatically tested at 1.5 times working pressure with certificate on file, plus a 10 bar relief valve and a rupture disc vented into the flare line.
• Rule: no cell runs without a certified primary boundary (own hydro test certificate or Parr vessel) plus a relief path into the flare.

2.5 Cell instrumentation (all variants)

• Cartridge heaters: 500 W, 316L sheath, 230 V, in the instrument well; driven by the heater PID loop (section 6.3) for temperature-programmed loading, degassing, and calorimeter calibration.
• Temperature: one type K thermocouple at bulk, one Pt100 Class A 4-wire at wall inside the well, plus Pt100 pairs on the calorimetry envelope (section 3).
• Pressure: Omega PX309-class stainless transmitters on every headspace, 0 to 10 bar gauge at cells, 0 to 100 bar gauge on the loading manifold; 4 to 20 mA output into the acquisition system.
• Optical: one USB-class machine-vision camera per cell aimed through the sapphire viewport, frame-triggered by the acquisition system, used to catch sonoluminescence flashes and suspension state.

2.6 Operating concept for cells

Every experimental run is a matrix entry, not an improvised fill. A run card fixes: composition (hydride, piezo phase, carrier), loading ratio history, fill volume, excitation program (ultrasound power and duty cycle, discharge program, heater program), controls run in interleaved order (hydrogen analog, no-piezo, no-hydride, excitation-off), and the pre-registered analysis threshold defined in section 9. Cells are filled in the glovebox or hood, sealed, leak-checked with a sniffer against 1x10-7 mbar L/s class sensitivity, then mounted into the calorimeter envelope.

3 Calorimetry

Two independent calorimeters share the Cell Room bench: an isoperibolic Seebeck-envelope calorimeter for the highest sensitivity, and a flow calorimeter for steady-state absolute power. Both run continuously; we never rely on a single thermal number.

3.1 Isoperibolic calorimeter (Seebeck envelope class)

Layout: the test cell sits inside a thick-walled aluminum can; the can wall carries a thermopile: a series chain of type E thermocouple junctions (400 junction pairs, glass-insulated wire) bonded alternately to the can outer wall and to a reference can. The reference can sits inside an isothermal enclosure: a second aluminum can with its own cartridge heaters and its own Pt100, held by a proportional-integral-derivative (PID) loop at 25.00 C setpoint, all inside a lagged box.

Wiring: thermopile output (microvolts) travels on twisted shielded pair to a nanovolt-class input channel on the acquisition chassis; thermopile leads are electrically floating relative to the cell so excitation currents cannot inject into the thermal signal. Enclosure heaters run from the same clean DC bus as cell heaters, switched by solid-state relays in the Drive Room.

Heat balance, logged continuously at 1 Hz:

P_thermal(t) = U * V_tp(t) + C_th * dT_in/dt

where U is the envelope conductance (watts per volt of thermopile signal) and C_th the envelope heat capacity. U and C_th are determined by Joule calibration: a precision resistor in the cell dissipates a known electrical power P_J = I^2 R at six setpoints from 0.1 to 20 W, and we fit U and C_th by least squares. Calorimeter noise floor target after 1000 s averaging: 50 mW (design target, to be demonstrated in commissioning). Excess heat is reported as P_thermal - P_input, with P_input measured at the cell feedthrough by a dedicated voltage and current channel pair, not read off a supply display.

3.2 Flow calorimetry

Layout: a second test cell jacket carries a constant-flow water circuit: Julabo F25-ME refrigerated/heating circulator feeding the jacket, flow measured by a Bronkhorst mini CORI-FLOW M15 Coriolis mass-flow meter (direct mass flow, independent of water properties), and inlet/outlet temperatures read by a matched Pt100 pair (two probes matched to 0.03 K at 25 C) in stirred thimbles. The circuit runs at 20 to 60 g/min, giving 0.2 to 3 K rise at our expected power range.

Heat balance, logged continuously at 1 Hz:

P_flow(t) = m_dot(t) * c_p * (T_out(t) - T_in(t)) + C_eff * dT_mean/dt

with m_dot from the Coriolis meter, c_p taken from a water property table by temperature, and C_eff the effective heat capacity of jacket plus cell (calibrated by the same Joule method: power steps into the cell with flow running, fit the transient). A condenser on the headspace returns vapor to the cell so no latent heat leaves through the vent; the condensate return temperature is logged and its correction term is bounded below 10 mW in normal operation.

3.3 What the calorimetry logs

Every second, per cell: thermopile voltage, enclosure setpoint and current, inlet and outlet temperature, Coriolis flow, cell bulk and wall temperature, headspace pressure, cell voltage and current (input power), excitation state flags (ultrasound on, discharge armed/fired), room temperature, barometric pressure. Calibration runs (Joule only, no active excitation) happen weekly and after any cell rebuild; a run without a same-week calibration on the same envelope is not a reportable run.

4 Radiation and particle diagnostics

Diagnostics are ordered so that the artifact-resistant pair (neutron time-of-flight plus CR-39) is always live, and every other channel adds specificity.

4.1 Neutron counting: moderated helium-3 proportional counter

• Instrument: Ludlum Measurements Model 15 neutron counter with moderated helium-3 proportional detector (Model 42-14H or 44-7 class, polyethylene-moderated tube, the configuration shipped with the Model 15).
• Position: fixed mount on the 1.0 m ring north of the primary cell, moderator side toward the cell. The polyethylene moderator also lives under the detector permanently; background campaigns (section 9) are run with exactly this geometry so background and run data share geometry.
• Role: independent, gamma-insensitive total neutron monitor. It integrates; it does not resolve. It is the tie-breaker when the spectroscopic pair (4.2) is saturated or offline.
• Add-on monitors: Bubble Technology Industries BD-PND bubble detector sets (12 units) deployed at the 1.0 m and 2.0 m rings and at the operator station; read by bubble count after each run. Bubble detectors are immune to electromagnetic interference and give a room-averaged check on the electronics-based counters.

4.2 Neutron and gamma spectroscopy pair with time of flight

• Detectors: two Eljen Technology EJ-309 liquid scintillator cells (2 inch by 2 inch, pulse-shape-discrimination grade), each coupled to an ET Enterprises 9821B 78 mm photomultiplier tube (PMT) in a light-tight housing with gain-stabilized high-voltage base.
• Electronics: both PMTs feed a CAEN DT5730S desktop digitizer (8 channel, 14 bit, 500 MS/s) running its on-board digital pulse processing firmware for charge-integration pulse-shape discrimination (PSD). Neutron and gamma events separate by pulse shape; event timestamps enable time-of-flight (ToF) between the two detectors.
• Geometry: detectors east and west of the cell at the 1.0 m ring, 2.0 m apart (ToF baseline), on rigid stands, heights matched to cell center. A shadow bar (borated polyethylene on the south mount) blocks the direct line from cell to the NaI detector so the gamma spectrum channel sees scattered background representative of the room.
• ToF relation used in analysis: t = d * sqrt(E_meV^2 + 0.876)^-1/2 * 7.83 ns/m class formula; in practice we histogram time differences and require coincidence windows consistent with 2.45 MeV neutrons crossing the 2.0 m baseline (about 105 ns flight time).
• Acquisition windows: during pulsed-discharge firing the digitizer is either trigger-inhibited or run in a double-blind schedule (fire recorded, live data discarded) for a randomized fraction of windows, so electromagnetic pickup cannot masquerade as events (section 6.4).

4.3 Gamma spectroscopy

• Detector: Scionix Holland 2 inch by 2 inch thallium-activated sodium iodide (NaI(Tl)) scintillation detector assembly with integrated bialkali PMT, south of the cell behind the shadow bar.
• Electronics: Ortec (AMETEK) digiBASE, a PMT base integrating bias supply, preamplifier, and digital multichannel analyzer, on the Control Room rack; spectra accumulated per run segment with live-clock normalization.
• Role: gamma lines over background during and after excitation, and a radionuclide screen on every loaded batch before it enters a cell.

4.4 Charged particles

• In-cell tracks: CR-39 (polyallyl diglycol carbonate track-etch plastic, PADC) foils from Track Analysis Systems (TASTRAK grade), mounted inside each cell in open and masked pairs: one foil fully exposed, its twin behind a metal mask that admits only selected solid angles and stops charged particles below a threshold energy. Foils are etched in the Wet Lab (6 molar sodium hydroxide, 70 C, timed bath) and counted under the microscope. CR-39 is immune to electromagnetic noise and integrates over the whole run: it is the quiet partner of the ToF pair.
• Active charged-particle spectroscopy: a Mirion (Canberra) passivated implanted planar silicon (PIPS) detector, A-300 class, behind a thin polyimide or beryllium window on a DN40 CF port, watching the 3.02 MeV proton branch of the D-D reaction; read out through a charge-sensitive preamplifier (Mirion 2018EB class) into the CAEN digitizer's fourth channel. The detector sits outside the primary boundary, hence the thin window and a stated detection-efficiency calibration per cell geometry.
• Personnel and area dosimetry: Landauer Neutrak CR-39 neutron dosimetry service for room monitors and badges, exchanged quarterly.

4.5 Headspace helium mass spectrometry

• Primary instrument: Hiden Analytical HPR-20 QIC quadrupole gas-analysis mass spectrometer (bench-top, heated capillary inlet, closed ion source) sampling cell headspace through the filtered gas port.
• Secondary instrument: Pfeiffer PrismaPlus QMG 220 compact residual gas analyzer (mass range 1 to 200 u) on the vacuum manifold for bake-out checks, outgassing trends, and leak diagnostics.
• The hard measurement: separating helium-4 (mass 4) from D2 (mass 4, doubly scattered) at unit resolution. Protocol: (1) seal the cell with getters (section 5) active so deuterium partial pressure is stripped over the accumulation period; (2) sample headspace through the getter into the Hiden inlet; (3) quantify helium against a calibration made by expanding a known helium dose into an identical empty cell; (4) require the mass-4 signal to track the mass-20 (neon spike, added as an internal reference in some runs) dilution and to persist after additional gettering, which consumes residual D2 but not helium. This protocol is written into every dry-bed run card; its false-positive modes (D2 tail, background helium permeation) are quantified in commissioning with hydrogen-analog cells.

5 Gas handling

Layout: all gas lives in the Gas Room. One deuterium line, one hydrogen line, one vacuum line, one flare line. The Cell Room sees only the single stub line to each cell.

• Storage: D2 and H2 cylinders in the outdoor grounded cage; stainless two-stage regulators; lines enter the Gas Room through a wall sleeve. Cylinder change is a two-person procedure with leak check before opening the room valve.
• Manifold: 1/4 inch 316L tubing with Swagelok VCR metal gasket face-seal fittings for make-and-break joints and Swagelok tube fittings elsewhere; bellows-sealed diaphragm valves at every branch; the deuterium and hydrogen branches share only the flare line, never a volume.
• Vacuum chain: Pfeiffer HiCube 80 Eco pumping station (HiPace 80 turbomolecular pump with integrated diaphragm backing pump and DCU 002 control unit) for cell evacuation and dry-bed gas loading; an Agilent IDP-3 oil-free dry scroll pump as the dedicated backing pump for the PrismaPlus residual gas analyzer and as an oil-free roughing alternative; an Edwards nXDS10i dry scroll pump on the loading station. Oil-sealed pumps are banned in the Gas Room: hydrocarbon backstreaming would poison getters and contaminate the mass spectra.
• Pressure measurement: Pfeiffer TPG 362 dual-channel controller with PKR 361 full-range (Pirani plus cold cathode) gauges on manifold and cells, and a CMR 362 ceramic capacitance diaphragm gauge (100 hPa full scale) where loading pressure accuracy matters; Omega PX309-class transmitters above atmosphere on every headspace.
• Flow control: Alicat MWB-series low-flow mass-flow controllers (hydrogen-calibrated) meter D2 and H2 into cells at 1 to 500 standard cubic centimeters per minute; Omega FMA-1600A mass-flow meters on the flare and purge legs for mass closure.
• Getters: a SAES Getters SORB-AC MK5-series cartridge non-evaporable getter (NEG) pump with activation heater sits on each cell's recirculation loop. Getters strip O2, H2O, N2 and (slowly) D2 from the headspace; helium passes through untouched, which is what makes the section 4.5 protocol work. Activation happens under vacuum at the manufacturer's temperature program, logged in the run file.
• Vent and flare: relief valves and rupture discs from every cell, the manifold overpressure path, and the purge legs all terminate in one flare line that rises above the villa roof. The flare line has a flame arrestor at the roof termination and a normally-closed solenoid valve energized only when ventilation is proven.
• Detection and interlocks: three Draeger Polytron 5200 CAT transmitters (catalytic sensor, 0 to 100 percent lower explosive limit, LEL): one at Gas Room ceiling, one at bench height in the Gas Room, one in the Cell Room stub-line corner. Logic in section 8.1.
• Operating concept: loading a dry bed means evacuate, backfill D2 to the loading pressure, hold with getters in the loop, sample to the PrismaPlus until D2 purity is stable, then seal. Hydrogen analogs run on the mirror branch so D and H plumbing never share wetted volume. Every gas action (valve state, setpoint, pressure, flow) is logged by the acquisition system, not by hand.

6 Power drive

The Drive Room drives the cells through two mechanical-excitation paths and one thermal path, with isolation engineered so excitation power cannot talk to the diagnostics.

6.1 Ultrasound drive

• Suspension cells: Hielscher UP400St ultrasonic processor (400 W, 24 kHz, automatic frequency tuning) with sonotrodes sized per cell; the generator's own frequency tracking follows the load as the suspension shear-thickens. The UP400St amplitude is remote-programmed so the acquisition system owns the duty cycle.
• Direct piezo drive: for cells with a mounted piezo stack or embedded piezo particles probed externally, a Siglent SDG2122X function/arbitrary waveform generator (dual channel, 120 MHz) synthesizes burst programs; an E&I 240L radio-frequency (RF) power amplifier (40 W linear, 10 kHz to 12 MHz, class A, specified by the maker for ultrasonics) amplifies them into a tuned network.
• Impedance matching: a manual L/C matching network between amplifier and transducer, tuned with an OMICRON Lab Bode 100 vector network analyzer (1 Hz to 40 MHz) at installation and after any transducer change; reflected-power is monitored at the amplifier and logged. Target: voltage standing-wave ratio below 1.5 at the working frequency so amplifier power actually enters the acoustic load.
• Wiring: all ultrasound power runs in its own conduit, physically separated from signal cable trays; sonotrode and stack returns return to the Drive Room bus, never through chassis grounds.

6.2 Pulsed-discharge bank

• Energy store: 10 mF film capacitor bank, maximum 1.5 kV charge (up to about 11 kJ), assembled from self-healing film cans with per-can fusing, discharge resistors, and a shorting bar that is inserted with a hook stick whenever the bay door opens.
• Charging: AMETEK Sorensen XHR 60-18-class programmable DC supply feeds the bank through a charging resistor; bank voltage readback is a dedicated isolated channel on the acquisition system.
• Switching: a Behlke HTS-series fast high-voltage MOSFET switch (3.1 kV class) fires the bank into the cell electrode pair through 2 m of coax; switch trigger comes from the acquisition system over a fiber-optic link, so no copper couples the trigger into the signal side.
• Interlocks: bank charger contactor is in the emergency-stop chain (section 8.4); the cabinet door opener shorts the bank through the dump resistor mechanically before the door swings past 10 degrees; a bank "armed" stack light is visible from the Cell Room.
• Operating concept: discharge programs (voltage, count, interval) are run-card items; the acquisition system logs every fire with bank voltage before and after, so delivered energy per shot is known, not assumed.

6.3 DC supplies and heater PID loop

• Cell and envelope cartridge heaters: driven from the Sorensen XHR-class supply through solid-state relays (SSRs); each heater zone has its own Omega CN7500-series 1/32 DIN PID controller wired to its own SSR and thermocouple, with the controller's retransmitted output and setpoint also logged by the acquisition system. PID runs locally (survives PC faults); the acquisition system supervises.
• Bias and low-noise analog: an Aim-TTi PL303 dual-output precision DC supply powers preamplifiers and photomultiplier bias dividers that need clean DC; the digiBASE and PIPS preamp high voltages are generated internally by their own modules.
• Grounding: one star point (section 1); supplies float with respect to signal ground except at the star; no daisy-chained outlet strips.

6.4 Isolation and EMI (electromagnetic interference) control for diagnostics

• Triggers and arm/disarm commands travel on fiber-optic links only.
• Detector signal cable is RG-223 double-shielded coax, single-braid grounded at the Control Room bulkhead; PMT housings float; all detector racks sit on the uninterruptible power supply (UPS) with an isolation transformer upstream.
• The Drive Room has its own line filters at the panel; the capacitor bank charging pulses ride their own circuit with a dedicated neutral.
• During discharge programs the acquisition system runs interleaved blind windows (section 4.2): a randomized subset of post-fire windows are recorded but excluded from unblinding unless they reproduce across blind and open windows with physically consistent timing. This is our structural answer to the triboelectric and pickup artifact classes reported across this field.
• Weekly: cable-bundle integrity check, ground-strap resistance measurement (target below 0.1 ohm to bus), and a pickup census with all excitations firing into a dummy load while detectors count; pickup counts above twice the quiet-room baseline trigger a wiring investigation before any run resumes.

7 Data acquisition

Platform: National Instruments CompactDAQ (cDAQ-9178, 8-slot USB chassis) per rack, two chassis total, with module allocation:

Timing and storage: both chassis share one host PC; samples are timestamped by the host with network time synchronization, and detector systems (CAEN, digiBASE, Ludlum) each keep their internal clock disciplined to the same host at session start, with a recorded offset. Raw data land on a redundant-disk workstation in the Control Room at 1 Hz for thermal and gas channels, event-mode for detectors, and mirror nightly to a second disk set. Dashboards: one operator screen shows calorimeter heat balance, detector rates, gas states, and interlock health; another shows spectroscopic live views; both are read-only live views of the same logging stream. Every run writes a self-describing file (run card hash, channel list, calibration constants in force, operator, timestamps).

8 Safety

8.1 Hydrogen detection and interlocks

• Three Draeger Polytron 5200 CAT transmitters (section 5) with two alarm steps each.
• Step one at 25 percent LEL: local alarm, ventilation to maximum, cell gas flows to minimum, operator acknowledges.
• Step two at 50 percent LEL: automatic closure of all cell gas solenoids (normally-closed valves), flare line solenoid opens, all heaters and the discharge bank charger drop out through the safety relay chain, room lights switch to the hazardous-area circuit, and the room is left to purge for 30 minutes before re-entry by two people with a portable hydrogen detector.
• Interlock chain health (sensor fault, valve fault, chain broken) is a latching state on the acquisition dashboard; a run cannot start with the chain unhealthy.

8.2 Pressure

• Every vessel carries the certification chain of section 2.4; relief paths terminate in the flare; rupture discs are dated and replaced on schedule; the loading manifold has its own relief at the regulator setpoint plus 10 percent.
• Annual external inspection of vessels, relief valves, and the flare line; records live in the Workshop.

8.3 Radiation and particles

• The Cell Room is a registered radiation area with the local authority from day one (neutron-emitting runs are only prospective, but registration precedes first excitation).
• Fixed monitors: Ludlum helium-3 counter and the EJ-309 pair run during every excitation; bubble detectors redeploy per run; the Ludlum Model 3 survey meter with Model 44-2 NaI(Tl) detector performs the documented weekly gamma survey and any-response checks.
• Personnel: Landauer Neutrak CR-39 badges quarterly for everyone listed on the project, plus electronic dose alarms for cell-room work.
• Access: Cell Room door contact and badge reader log entries; nobody works in the Cell Room during discharge programs without a second person in the building.

8.4 Emergency-stop architecture

One Pilz PNOZ s4-class safety relay chain monitors: four emergency-stop pushbuttons (Cell Room door, Drive Room door, Gas Room door, Control Room bench), the hydrogen step-two alarm, the bank cabinet door switch, and the fume-hood airflow switch for the etch bench. The chain contactors drop, in order: cell gas solenoids, heater SSRs, RF amplifier and ultrasound generator mains, bank charger contactor. Vacuum pumps and ventilation stay powered (they are the safe state for hydrogen). The chain resets only from the Control Room with a key switch, and every trip is logged with timestamp and source.

8.5 Personal protective equipment and isotope rules

• PPE (personal protective equipment): safety glasses everywhere; face shield plus alkali-resistant apron and gloves for etching; hearing protection when the ultrasound generator runs above 50 percent amplitude; antistatic wrist strap for board-level work; closed shoes and lab coats in all labs.
• Isotope handling: D2 is treated as a flammable compressed gas first and an isotope second: outdoor storage, two-person cylinder changes, no D2 release indoors except through the flare. Deuterated liquids (heavy water, deuterated oils) are bought in inventory-controlled bottles, decanted in the hood, and logged; deuterided powders are handled only in the glovebox under argon and stored in sealed containers in a ventilated cabinet (deuterided powders can release gas on air exposure and some are pyrophoric when finely divided).
• Waste: etch solutions and powder residues to labeled sealed drums in the Workshop store; disposal through a licensed contractor; nothing deuterated goes down a drain.

8.6 Training requirements

No person performs a procedure alone that appears in the two-person list (cylinder change, bank arming, hydrogen recovery from a cell, any run with discharge program active).

9 Operating concept: run modes and accountability

Three modes:

1. Calibration mode: Joule-only heat steps, detector efficiency checks (calibrated sources for NaI channels, cosmic background for the scintillator pair), mass-spectrometer helium dose calibration, pickup census. All weekly or after rebuilds.
2. Control mode: full rig, full instrumentation, hydrogen analog compositions or excitation-off interleaves, logged exactly like an experiment.
3. Experiment mode: the deuterated matrix entries from the run card.

Interleave discipline: within any 48 hour block, an experiment run is bracketed by control runs in randomized order, and the analysis threshold is fixed before the block starts: five sigma above the simultaneous control distribution, pre-registered in the run file, applied to the calorimetric excess, the neutron coincidence excess, the CR-39 track density difference, and the helium accumulation independently. A result is a result when it survives the same threshold in the same file that generated it; no post-hoc threshold moves.

Commissioning order (first 90 days of occupancy): gas room leak-back and purity; calorimeter Joule calibration and noise floor; detector background campaigns (two weeks, full geometry, excitation off) to fix room baselines; pickup census; then first hydrogen-analog suspension runs to exercise every channel end to end. Only after a clean hydrogen-analog cycle does the first deuterated dry-bed run enter the calendar.

10 Interfaces

• OUTLINE.md, SCIENCE-AND-FOUNDATIONS.md: concept, targets, controls, thresholds that this lab exists to serve.
• prior-art/*: the reported experimental record that motivates each diagnostic pair and every control we run.
• simulation/TOOLCHAIN.md and simulation/MARBELL-SIMULATION-QUEUE.md: modeling precedes hardware; cell geometries here are the geometries the simulation models.
• materials/SPECS.md, materials/SUPPLIERS-AND-PRICES.md: powders, carriers, and loading supplies that flow through section 2 cells.
• LAB-BOM.md: the itemized bill of materials for this build, with vendor, model class, quantities, and estimated prices.