{"schema_version":1,"assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","source_experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"invariant_mode_decomposition_design__nanotechnology","archetype_slug":"invariant_mode_decomposition_design","domain_slug":"nanotechnology","title":"Advisory modal-instability detection for nanoparticle synthesis","opportunity_summary":"Estimate coupled temporal modes from instrumented nanoparticle-synthesis runs, flag weakly damped or growing quality-relevant modes, and provide operator-approved damping guidance subject to residual, drift, spectral-gap, safety-limit, and quarantine controls. The existence of reproducible coupled instability and any improvement over raw-coordinate alarms or multivariate statistical monitoring remain hypotheses.","adopter_authorizer":"The adopter is a nanoparticle synthesis process and quality organization; the accountable process owner authorizes evidence work and any pilot, trained operators retain actuator control, and safety and quality personnel set hard limits and material disposition.","scores":{"meaningful_impact":{"score":3,"rationale":"Earlier detection could reduce irrecoverable off-spec batches, distribution widening, aggregation, and harmful cross-effects from coordinate-wise adjustments. However, the frequency, severity, and downstream exposure consequences of the proposed hidden instability are unsupported."},"stakeholder_pull":{"score":2,"rationale":"Operators, process engineers, quality personnel, and downstream users have identifiable interests in synthesis stability, but the packet contains no stakeholder testimony, adoption inquiry, demonstrated pain frequency, or willingness to allocate resources."},"incremental_advantage":{"score":3,"rationale":"Temporal stability classification, actuator-to-mode mapping, and explicit residual and drift vetoes could add decision value beyond univariate alarms and PCA-based correlated-deviation flags. No replay or pilot result establishes better lead time, fewer false alarms, or improved quality."},"distinctiveness_plausibility":{"score":2,"rationale":"The composition is specifically distinguished from the named PCA or multivariate-control rival by its temporal mode gains and actuator mapping, but prior-art status is UNSearched and no external distinctiveness evidence is available."},"technical_implementability":{"score":3,"rationale":"The proposal names measurable state vectors, local-transition estimation, comparators, bounded perturbations, and halt criteria. Implementability remains limited by unproven local stationarity, observability, sensor precision, mode conditioning, safe excitation, actuator identifiability, nonlinear boundaries, and non-normal transient growth."},"adoption_authority_feasibility":{"score":4,"rationale":"The packet identifies an accountable process owner, trained operators, and safety and quality authorities, preserves raw alarms, excludes autonomous control and material release, and supplies rollback and quarantine rules. Process-specific approvals are not yet documented."},"evidence_readiness":{"score":3,"rationale":"A safe sequence of historical replay followed conditionally by an instrumented small-scale pilot is specified, with baseline and nearest-rival comparisons and falsifiers. Data adequacy, repeated-run availability, thresholds, sensor specifications, and preregistered transient-growth diagnostics are still undefined."},"safety_net_benefit":{"score":4,"rationale":"The advisory could provide earlier warning while retaining operator decisions and existing alarms, and it includes residual, spectral-gap, mode-rotation, hard-limit, rollback, and quarantine safeguards. The warning's reliability and ability to avoid unsafe corrections have not been demonstrated."},"scalability":{"score":2,"rationale":"The framework could in principle be repeated across reactors, but modes, sensors, valid linearization windows, actuator mappings, thresholds, and safety validation are likely process-specific; no transferability evidence is present."}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"Curate and replay repeated historical runs, define state variables and outcomes, fit and condition-check modal models, preregister comparators and falsifiers, and, only if replay passes, conduct one instrumented small-scale pilot with preapproved sub-limit perturbations.","confidence":"LOW","assumptions":["Usable time-aligned historical reactor and particle-quality data already exist.","A small-scale instrumented reactor and authorized operators are available through a partner.","No major new analytical instrument or extensive toxicology program is required.","The pilot material is quarantined and not released."]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Validate and integrate an operator-advisory implementation for one synthesis process, including sensor and data integration, model qualification, alarm-interface work, operating procedures, training, safety and quality review, and prospective evaluation.","confidence":"LOW","assumptions":["Deployment is limited to one existing, instrumented process.","The system remains advisory and does not autonomously change set points.","Only moderate sensor or data-infrastructure upgrades are needed.","Process qualification can use existing safety and quality governance."]},"operational_launch":{"band_2026_usd":"1M_TO_5M","scope":"Launch the qualified advisory in routine operations for a limited facility scope, including multi-run validation, production integration, change control, monitoring, operator coverage, documentation, and material-disposition procedures.","confidence":"LOW","assumptions":["Several reactor or recipe contexts require separate validation.","Production integration and quality-system documentation are required.","No production-scale perturbations outside validated limits are performed.","The organization already possesses core synthesis and characterization equipment."]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Maintain one limited-facility deployment through data-quality monitoring, recalibration, drift and mode-rotation review, model requalification, software support, incident review, operator training, and quality oversight.","confidence":"LOW","assumptions":["The number of supported processes remains small.","Existing staff perform part of the monitoring and review.","Major sensor replacement, new characterization equipment, and broad process expansion are excluded.","Material remains governed by existing quality and exposure programs."]}},"research_burden":"HIGH","earliest_credible_horizon":"3_TO_12_MONTHS","pipeline_gates":{"recognizable_externally_supportable_problem":{"status":"YES","reason":"The candidate identifies a concrete synthesis failure pattern—aggregation or widening particle-size distributions emerging from coupled states while individual averages remain ordinary—and clearly states that reproducibility in any given process must be tested."},"identifiable_adopter_or_authorizer":{"status":"YES","reason":"The accountable process owner can approve replay and a pilot; trained operators retain actuator decisions, while safety and quality personnel control limits and material disposition."},"distinct_testable_incremental_claim":{"status":"YES","reason":"The testable claim is earlier, more reliable warning or better final quality from temporal modal guidance compared with both raw-coordinate monitoring and PCA or multivariate statistical process control, with out-of-sample failure criteria."},"bounded_next_evidence_step":{"status":"YES","reason":"Historical replay is bounded and non-interventional, and it can compare warning lead time, false alarms, residual validity, and held-out quality prediction against both named comparators before any pilot is authorized."},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The next retrospective step creates no live process intervention. Any later pilot requires accountable approval, operator control, preapproved sub-limit perturbations, preserved raw alarms, explicit veto conditions, recipe rollback, and material quarantine."},"implementation_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The packet bounds replay, small-scale piloting, and an advisory role, but does not specify historical-data condition, sensor upgrades, reactor count, validation obligations, integration effort, or partner terms; therefore only assumption-dependent broad bands are supportable."}},"blocking_evidence":["Repeated runs must show coupled temporal structure beyond measurement error rather than excursions fully explained by one observable variable.","The local transformation must remain sufficiently stationary, well-conditioned, and reproducible within a useful operating window.","A preregistered non-normal transient-amplification diagnostic must qualify or veto eigenvalue-based warnings.","Sensor precision, scaling stability, missingness, state observability, excitation coverage, and actuator-to-state rank must meet predefined thresholds.","Safe permitted perturbations must identify actionable modes without approaching process or exposure limits.","Held-out replay and any later bounded pilot must improve preregistered warning lead time or quality versus both raw-coordinate alarms and the PCA or multivariate rival without unacceptable false alarms or cross-effects.","Process-specific authorities must confirm data access, pilot approval, hard limits, rollback procedures, and quarantine capacity."],"next_evidence_step":"Run a preregistered retrospective replay on repeated, time-aligned historical synthesis runs, separating training and held-out runs. Compare the modal method with existing raw-coordinate alarms and the named PCA or multivariate-control rival on warning lead time, false alarms, residual and drift validity, transient amplification, and prediction of final aggregation or size-distribution failure. Stop before live guidance and reject the problem hypothesis if excursions lack reproducible coupled structure beyond measurement error or are fully predicted by one observable variable; reject the incremental claim if held-out performance does not beat both comparators.","research_questions":["Do repeated runs exhibit reproducible coupled temporal structure before aggregation or distribution widening?","What local operating windows satisfy residual, stationarity, conditioning, spectral-gap, and mode-rotation criteria?","Does non-normal transient amplification agree with, qualify, or invalidate the eigenmode warning rule?","Are sensor precision, sampling rate, scaling stability, missingness, and state observability adequate for the retained modes?","Do permitted perturbations provide sufficient excitation and actuator-to-state rank to identify safe damping directions?","Does the method improve held-out warning lead time, false-alarm performance, or final quality versus both existing alarms and PCA-based monitoring?","How sensitive are the inferred modes and guidance to nonlinear nucleation boundaries, contamination, sensor artifacts, recipe changes, and reactor transfer?","What process-specific approvals, validation documentation, infrastructure, and recurring review would adoption require?","Does external research reveal materially equivalent prior art or narrower competing methods?"],"recommendation":"VALIDATE_PROBLEM_FIRST","uncertainty_constraints":["Closed-book assessment provides no evidence of problem prevalence, stakeholder demand, market size, realized impact, or external prior art.","All causal benefits, mode reproducibility, actuator effects, and transferability are hypotheses.","Cost bands are resource-equivalent planning ranges, not quotes, and depend strongly on data condition, instrumentation, validation, and facility scope.","The 3-to-12-month horizon applies only to credible initial retrospective and possibly small-scale evidence when data and partner access are ready, not production deployment.","Abrupt nucleation or aggregation boundaries, non-normal dynamics, sensor drift, contamination, and unmeasured states may invalidate the modal approximation."],"closed_book_prior_art_boundary":"No external search was performed or permitted. Prior-art status remains UNSearched; novelty, prevalence, market position, and equivalence to existing temporal process-control or system-identification methods are unmeasured. The assessment credits only the proposal's internally specified distinction from its named rival."}