Hidden-Sector Searches

Established A hidden sector is any set of particles that couples to ordinary matter through a small parameter rather than through the known forces. The canonical examples are the axion, invented to solve the strong-CP problem and later found to be a dark-matter candidate; the dark photon, which mixes kinetically with the ordinary photon; light scalars mediating a fifth force; and sterile neutrinos. Each has a two-dimensional plot — mass against coupling — and the experimental programme consists of deleting regions of those plots.

Established Nothing has been found, and the exclusions are the result. This is not a rhetorical softening. A limit curve is a measurement: it says that if a particle of this mass existed with a coupling above this line, an instrument of known sensitivity would have seen it and did not. The axion-photon plane, the dark-photon kinetic-mixing plane and the short-range-gravity plane are among the best-mapped parameter spaces in physics, and they were mapped by experiments that each published a null.

Frontier The honest structure of the field is that sensitivity is bought in narrow slices and theory prefers a window that instruments cannot yet reach. Resonant cavity searches achieve the sensitivity needed to reach the QCD axion band but scan slowly and only over the frequency range a cavity of buildable size supports. Broadband methods cover decades of mass at once and lose sensitivity doing it. Lattice simulations of axion string networks point to a mass window in the tens-to-hundreds of micro-electronvolt range, which is above where cavity haloscopes work well and below where helioscopes are optimal.

Established Two of the field’s most-cited motivating anomalies dissolved within four years. The lepton-universality discrepancies in B-meson decays returned to Standard Model values in a 2022 reanalysis with corrected particle identification, and the muon anomalous magnetic moment ceased to be a discrepancy in 2025 when the theory consensus adopted lattice determinations of the hadronic contribution. Both reversals came from inside the field and were published by the same collaborations that had reported the tensions.

Established A note on sourcing. This brief was commissioned in September 2026 from the Institute’s research base. Reading-list entries without links are cited from the bibliographic record rather than re-fetched, and claims are dated no later than early 2026 unless carried by a linked source.

Established The axion haloscope is the field’s best-developed instrument and it has reached the theory band over a narrow slice. The Axion Dark Matter eXperiment places a tunable microwave cavity in a magnet of roughly 7 to 8 tesla at about 100 millikelvin, reads it out with near-quantum-limited amplifiers, and searches for the conversion of galactic-halo axions into photons. Across its published runs it excluded axions with couplings down to the more pessimistic of the two benchmark models over roughly the 2.7 to 4.2 micro-electronvolt range. That is a decisive exclusion of a specific, well-motivated model over about half an octave of mass.

Established The scan-rate scaling is the governing constraint, and it is brutal. The rate at which a haloscope can cover frequency goes as the square of the magnetic field squared times the cavity volume, the quality factor, and the inverse square of the system noise temperature. Doubling the covered mass range at fixed sensitivity is not an incremental ask. This is why decades of effort have produced coverage measured in fractions of a decade of mass rather than in decades, and why quantum-limited and squeezed-state receivers — which cut the noise term — produced the largest single improvements of the last ten years.

Established Squeezed-vacuum readout has delivered a measured, replicated speed-up. The HAYSTAC experiment reported a roughly twofold increase in scan rate from injecting squeezed states into its receiver, the first use of quantum metrology to accelerate a dark-matter search rather than merely to improve a single measurement. The gain is real and it is a factor of two, not a factor of a thousand; the instrumentation frontier here is incremental by construction.

Frontier Above roughly ten micro-electronvolts the cavity approach fails and dielectric haloscopes take over, unproven. A higher axion mass means a higher frequency and a smaller resonant cavity, which shrinks the volume term to nothing. The MADMAX concept replaces the cavity with a stack of movable dielectric discs in a strong dipole field, converting a resonance problem into a boosted-emission problem. Prototype runs have been performed and have produced exclusion limits, including on dark-photon dark matter; the full experiment requires a dipole magnet of a class that does not exist.

Established The helioscope and light-shining-through-a-wall branches are model-independent and less sensitive. A helioscope looks for axions produced in the solar core and reconverted in a laboratory magnet; the CERN Axion Solar Telescope set a photon-coupling limit around 6.6 × 10⁻¹¹ per giga-electronvolt, comparable to the astrophysical bound from horizontal-branch stars. Light-shining-through-a-wall experiments produce and reconvert the particle entirely in the laboratory, which removes all astrophysical assumptions at the price of a coupling that enters to the fourth power. Neither has detected anything.

Established Direct dark-matter detection reached the neutrino background, which is the end of a road rather than a stage on it. Liquid-xenon time projection chambers now report spin-independent cross-section limits in the low 10⁻⁴⁸ square centimetre range near 40 giga-electronvolts, from exposures of a few tonne-years. At that sensitivity solar neutrinos scattering coherently off xenon nuclei produce an irreducible signal, and coherent elastic neutrino-nucleus scattering from solar boron-8 has been observed in these detectors. Further progress requires directional information or much larger exposures for logarithmic gains.

Established The single longest-running positive claim in the field has failed independent replication with the same target material. One experiment has reported an annually modulating signal in sodium iodide for over two decades. Two independent experiments using sodium iodide crystals specifically to test it have reported no modulation, with incompatibility significances growing past three standard deviations as exposure accumulated. This is the field’s clearest example of replication working as intended: same target, same technique, different result, published.

Established The xenon electron-recoil excess of 2020 was resolved as a background, by the same collaboration, within two years. A roughly three-and-a-half-sigma excess at low electron-recoil energies was widely interpreted as solar axions or a neutrino magnetic moment. The successor detector, with lower background, saw no excess; tritium contamination is the accepted explanation. The episode is worth carrying because it shows a three-sigma anomaly in a well-run experiment being retired by better instrumentation rather than by argument.

Established The muon anomalous magnetic moment is no longer a discrepancy with the Standard Model, and this is the most consequential change of the last five years. The Fermilab measurement reached the roughly 100-parts-per-billion level with its final dataset. The 2025 theory white paper shifted the hadronic vacuum polarisation input from electron-positron annihilation data to lattice quantum chromodynamics, and the resulting prediction agrees with the measurement. What remains is a disagreement between two ways of computing the same hadronic quantity — a serious problem, but a theory-internal one.

Established Laboratory tests of the gravitational inverse-square law now reach tens of micrometres. Rotating torsion-balance experiments have tested separations from about 52 micrometres to 3 millimetres, excluding gravitational-strength Yukawa interactions with ranges above a few tens of micrometres. Below that the Casimir force between the test masses dominates and the measurement becomes an electromagnetic problem. The wall is physical, and it is why proposals to probe shorter ranges are proposals to control or subtract the Casimir background rather than to build a better balance.

Frontier Precision clocks and atom interferometers are now competitive detectors for the lightest hidden sectors. Ultralight scalar dark matter would make fundamental constants oscillate, producing a periodic beat between clocks of different sensitivity; optical clocks now hold systematic uncertainties below one part in 10¹⁸, which converts directly into coupling limits. A 100-metre vertical atom interferometer under construction at Fermilab is designed to search this space and the mid-band gravitational-wave window at once. The corpus treats the instrument class in Precision Quantum Sensing.

Established Accelerator searches have closed the simplest dark-photon parameter space. Missing-energy searches with electron beams, electron-positron collider searches for dark photons produced with a photon or a dark Higgs, and fixed-target beam dumps together exclude the kinetic mixing values that the vanilla dark photon would have needed to explain the historical muon anomaly. The anomaly no longer needs explaining, and the parameter space is excluded anyway; this is the two-sided closure the field rarely gets.

Frontier Astrophysical and cosmological limits are stronger than laboratory limits over most of the plane, and they carry model dependence laboratory limits do not. Stellar cooling arguments, supernova energy-loss bounds and constraints on extra radiation at the epoch of nucleosynthesis exclude couplings orders of magnitude below any terrestrial reach. They depend on stellar models and on one supernova’s neutrino burst. Neutrino observatories extend the same logic to the highest energies; Neutrino Astronomy documents how tight those astrophysical constraints have become and at what statistical cost.

Frontier Where is the axion, if it exists? If the relevant symmetry broke after inflation, simulations of the resulting string network predict a mass in the tens-to-hundreds of micro-electronvolt range, with the central estimates clustering in the low hundreds. If it broke before inflation, the mass is essentially a free parameter and the theory predicts nothing useful. The first case gives the field a target; the second explains why the field has been searching for forty years without one.

Frontier Is the post-inflationary mass prediction robust? It comes from numerical simulations of string networks with adaptive mesh refinement, and different groups extrapolate the string tension differently, producing spreads of a factor of a few. A factor of a few in mass is a factor of a few in frequency, which is the difference between an experiment that exists and one that does not. Independent simulation with different methods is the cheapest high-value work available in this field.

Frontier Which hadronic vacuum polarisation is right? Lattice calculations and electron-positron annihilation data disagree at a level that matters for the muon anomaly, and within the annihilation data one modern measurement of the two-pion cross-section disagrees with the older set. Since the Standard Model prediction depends on which is used, the question of whether there is any muon anomaly at all is now a question about a hadronic cross-section measurement.

Speculative Are the remaining neutrino anomalies hidden-sector physics? Short-baseline excesses motivated a light sterile neutrino for two decades. A liquid-argon experiment found no corresponding electron-like excess, disfavouring the simplest reading, while gallium-source experiments report a deficit at high significance that conflicts with reactor limits. The pattern — several anomalies, mutually inconsistent parameter requirements — is what an accumulation of unrelated systematics looks like, but it is not proof of one.

Speculative Does a light boson appear in nuclear transitions? Anomalous angular correlations in internal pair creation from excited beryllium and helium nuclei have been reported repeatedly by one laboratory and interpreted as a roughly 17 mega-electronvolt boson. A dedicated search at an independent experiment with far better resolution found no signal in the relevant transition. The claim stays in this brief because omitting it hands it to worse sources; the evidence stands at one laboratory reporting, one independent search not confirming.

Established Magnets are the binding constraint for every magnetic conversion experiment. Sensitivity to the axion-photon coupling scales with the field and the volume, so the reachable coupling is set by what magnet money can buy. Haloscopes use solenoids of 7 to 12 tesla; dielectric haloscopes need a dipole with a bore large enough to hold a disc stack, in a geometry accelerator magnets are not built for. Recycled accelerator dipoles are used precisely because purpose-built ones are unaffordable.

Frontier High-temperature superconducting tape could relax that constraint and currently does not. Tape capable of higher fields exists in kilometre quantities and its production is now driven by fusion programmes rather than by physics experiments, as documented in High Temperature Superconductors. Whether a search-magnet programme can buy into that supply at a price a mid-scale experiment can pay is an open procurement question, not an open physics question.

Established Noise temperature is the other lever, and it is nearly exhausted. Once a receiver is at the standard quantum limit, further gains require evading that limit — squeezing, photon counting, or entanglement-assisted readout. Squeezing has bought a factor of two. Single-microwave-photon counting promises more and has not yet been demonstrated at the duty cycle a scan requires. There is no third lever of comparable size.

Established Cryogenic plant and dilution refrigerator capacity limits how many experiments can run at once. Every competitive haloscope needs millikelvin temperatures in a large bore inside a strong field, which is an unusual combination and a small supplier base. The constraint is rarely stated in papers because it appears as a schedule rather than as a sensitivity limit.

Frontier Backgrounds set the floor for everything non-resonant. Direct detection has reached the neutrino fog; short-range gravity has reached the Casimir wall; beam-dump searches are limited by neutrino-induced events and by muon halos. Each of these is a background that cannot be shielded away, only measured and subtracted, and subtraction at the required precision is where the remaining systematic risk in the field lives.

Established The whole programme depends on theory inputs it does not control. The local dark-matter density, near 0.4 giga-electronvolts per cubic centimetre with roughly a factor-of-two spread, multiplies every haloscope limit. The halo velocity distribution sets the expected line shape. A change in either rescales published limits without any experiment being rerun, and the astrophysical inputs are less certain than the instrumental ones.

Established Quantum measurement technology is now an upstream dependency rather than a spin-off. Josephson parametric amplifiers, squeezed-state generation and superconducting qubit readout came from quantum information research and were imported. The same instrument stack appears in Precision Quantum Sensing and its materials constraints in Quantum Materials; hidden-sector searches are a demanding customer of both and a driver of neither.

Frontier Lattice quantum chromodynamics has become load-bearing for what counts as an anomaly. First-principles lattice calculations of hadronic quantities are now precise enough to decide whether an experimental result is a discrepancy, which is a reversal of the historical direction of authority between calculation and measurement. Anyone who takes the muon result seriously is taking a supercomputer calculation seriously as an arbiter.

Established The field depends on a publishing culture that rewards nulls, and that culture is fragile. Limit papers are the field’s product; they are also harder to fund, less cited and less career-making than discovery papers. The structural incentives are documented in Scientific Funding Models. A hidden-sector programme that cannot publish nulls prominently degrades into a machine for generating unrefuted claims.

Established The decisive experiment is a scan of the tens-to-hundreds of micro-electronvolt axion mass window at a sensitivity that reaches the pessimistic QCD axion band. This is the window that post-inflationary string-network simulations point to, and it is the one region where a null result would be genuinely informative about the QCD axion rather than about one more slice of an unbounded parameter space. Covering it requires either dielectric haloscopes in a large-bore dipole above roughly nine tesla, or plasma haloscopes, or a helioscope of the next generation — and the magnet for the first of those does not exist.

Frontier The second decisive result is a measurement, not a search: an independent replication of the two-pion electron-positron cross-section. The disagreement between one modern measurement and the older body of data is now what determines whether the data-driven and lattice predictions of the muon anomalous magnetic moment can be reconciled. Several experiments can perform this measurement with existing beams, and radiative-return analyses at a B-factory are already underway. It is the cheapest high-consequence result available anywhere in this brief.

Established The third is already running and requires no new hardware: continued exposure by the sodium iodide replication experiments. Each additional year of data sharpens the incompatibility with the long-standing modulation claim. This is a natural experiment in replication as much as in dark matter, and its outcome is close to settled.

Frontier The fourth is the next-generation beam-dump programme. A dedicated proton beam-dump facility approved at CERN would extend sensitivity to feebly interacting particles in the mega-electronvolt-to-giga-electronvolt range by orders of magnitude in coupling, covering regions accessible by no other method. It is funded and sited, and it delivers physics in the 2030s rather than this decade.

Speculative The fifth would be the only unambiguous confirmation: two independent experiments, different techniques, same mass and coupling. The field’s history — a modulation seen by one group for twenty years, a nuclear anomaly reported by one laboratory, an electron-recoil excess that became tritium — makes single-experiment discovery claims uninterpretable. The prior on any first detection should be low until it is seen twice.

Established A haloscope is an engineering problem of holding a resonance stable while moving it. A tuning rod inside a cold cavity inside a bore must step in kilohertz increments, settle mechanically, and hold quality factor and mode purity while the thermal environment shifts. Mode crossings destroy sensitivity at specific frequencies and appear as gaps in published limits. The published exclusion curve is, read correctly, a record of mechanical engineering as much as of physics.

Established Dielectric haloscopes convert that into a precision positioning problem. Dozens of discs held parallel to micrometre tolerances in a cryogenic environment, each repositionable to retune the boost factor, with the emitted power focused onto a detector. Prototype work has demonstrated the principle at reduced scale and has published limits; the step to a full-scale booster in a full-scale magnet is the one that has not been taken.

Frontier Broadband low-mass searches are lumped-element radio engineering at millikelvin temperatures. Below a micro-electronvolt the axion field behaves as a slowly oscillating current source, read out with a superconducting pickup and an amplifier near the quantum limit. Prototypes have published nulls; the scaling path is larger volumes and lower noise, both of which are cryogenic-plant problems rather than new physics.

Established Fifth-force instruments are dominated by the suppression of everything except gravity. A torsion balance test requires electrostatic shielding with a stretched conducting membrane, magnetic screening, thermal gradient control at the millikelvin level across the apparatus, and a stiff rotation stage. Every published short-range limit is a systematics budget first and a measurement second.

Frontier Quantum sensors are moving from readout devices to detectors in their own right. Levitated masses, matter-wave interferometers and mechanical oscillators at their quantum ground state are being proposed as direct probes of feeble couplings, and the same hardware programme underwrites tabletop tests of gravity itself, as described in Quantum Gravity. The engineering is shared; the physics targets are different and should not be conflated in funding cases.

Established Neutrino observatories are hidden-sector instruments by another name. A cubic kilometre of instrumented ice built to find astrophysical neutrinos also sets some of the tightest existing limits on space-time-violating operators and on heavy dark-matter decay. The cross-use is real and documented; Neutrino Astronomy treats the detectors and their statistics, and this brief takes their limits as inputs.

Frontier Black-hole superradiance is an astrophysical haloscope with no hardware. An ultralight boson with a Compton wavelength comparable to a black hole’s size would extract angular momentum and form a cloud, and the observed spin distribution of black holes therefore excludes mass windows for such bosons. It is a constraint bought with telescope time rather than magnet time, and its assumptions are treated in Black Hole Physics Applications.

Established Atomic clocks and interferometers cover the ultralight end that magnets cannot. Below the nano-electronvolt scale the field oscillates too slowly for resonant conversion but fast enough to modulate atomic transition frequencies, which makes a clock comparison a dark-matter experiment. This is the cheapest sensitivity in the whole field per unit of new parameter space, and it exists because metrology was funded for other reasons.

Frontier The energy frontier is complementary and is not a substitute. Colliders produce heavy mediators directly and are blind to the feebly coupled light states these searches target, while beam dumps and fixed-target experiments sit in between. A field whose entire budget went to one of the three would leave large regions unprobed, which is the strongest argument for the current portfolio and is rarely made explicitly.

Established This is a mid-scale field, and mid-scale is the hardest size to fund. A haloscope costs millions to tens of millions and runs for a decade with a staff of tens — too large for a single grant, too small for a dedicated international organisation. The result is a landscape of university-hosted experiments dependent on national agency cycles, with magnets often inherited from cancelled or completed programmes.

Frontier The theory-experiment feedback loop has a publication-bias problem in one direction. Every new exclusion generates theory papers explaining why the particle could sit just beyond it. This is legitimate model-building and it is also unfalsifiable in aggregate: a framework with a free coupling and a free mass can always retreat. A discipline norm requiring proponents to state in advance which region would count as decisive refutation would cost nothing and does not exist.

Established Replication has been organised deliberately here, which is rare in physics. Two experiments were built specifically to test another experiment’s claim with the same target material, funded on the explicit basis that a twenty-year-old positive result required independent confirmation. That is a working institutional answer to a single-laboratory claim, and it should be cited whenever another field asks how replication could be organised.

Frontier Data and limit-setting conventions are a governance issue in a field whose product is curves. Published exclusions depend on halo density assumptions, statistical procedures and look-elsewhere corrections that vary between collaborations, which makes the composite plots reproduced in review articles less commensurable than they appear. There is no standards body for this, and the community’s compiled limit plots are maintained largely by volunteers.

Established The honest public case for this work is curiosity, and inflating it is the main integrity risk. There are no applications of axion detection. The instruments have spin-offs — quantum amplifiers, cryogenics, low-background materials — and those spin-offs do not require the search. Funding cases that promise technological returns invite an evaluation the field will lose.

Frontier Announcing anomalies responsibly is an ethical question with a track record. Two recent excesses were announced with appropriate statistical caution and were nonetheless reported publicly as discoveries, then retired. The collaborations behaved well and the communication system around them did not. The remedy is not less openness but pre-registered thresholds and a stated replication requirement before the word discovery is used.

Established Underground laboratories and rare-isotope supply carry real, small externalities. Ultra-low-background experiments need ancient lead, low-radioactivity copper, xenon inventories measured in tonnes and deep excavation. These are modest impacts by industrial standards and non-zero; the xenon inventory in particular ties a physics programme to an industrial gas market it does not control.

Established The atlas of nulls is the durable output and it is genuinely valuable. Excluded parameter space does not expire. A limit set in 1997 still constrains models written in 2026, which is a property very little scientific output has. Treating null results as a permanent civilizational asset rather than as failed searches is the correct accounting, and it is the accounting the field itself uses internally while describing itself outwardly in the language of discovery.

Frontier If the QCD axion exists and is found, the consequences are conceptual rather than practical. It would solve the strong-CP problem, identify the dark matter, and confirm that a symmetry broken in the early universe left a detectable relic. None of that changes any technology. A civilization that funds this is buying knowledge and should say so.

Speculative If nothing is found across the theoretically preferred windows, the field faces a genuine crisis of direction. Dark matter would remain observed gravitationally and unidentified in the laboratory, and the strong-CP problem would remain unsolved, with the most economical solution excluded. The honest response would be to reallocate toward gravitational and astrophysical probes, and there is no institutional mechanism that performs such reallocations quickly.

These horizons track parameter-space coverage and the instruments that would deliver it, not the probability of detection, which nobody can quantify.

• 10 yr: Frontier Cavity haloscopes extending coverage through the single-digit micro-electronvolt range at benchmark sensitivity; next-generation xenon detectors operating into the neutrino fog; the two-pion cross-section disagreement resolved or entrenched; the dedicated beam-dump programme under construction.
• 25 yr: Speculative Either the tens-to-hundreds of micro-electronvolt window is covered at QCD-band sensitivity by dielectric, plasma or helioscope instruments, or it is not and the field has no decisive target left. Which of these happens is decided by one magnet procurement rather than by any scientific result.
• 50 yr: Speculative Directional dark-matter detectors and gravitational probes take over from cross-section exposure, which has hit its background floor. Quantum-sensor networks cover the ultralight decades at sensitivities no current experiment approaches.
• 100 / 250+ yr: Handwave Proposals to use astrophysical environments or space-based baselines as detectors are coherent and have no engineering basis or cost estimate. Any claim about detection on this horizon is a claim about funding institutions surviving, not about instruments.

• Depends on results already on this map: the quantum-limited readout chain documented in Precision Quantum Sensing, the conductor supply described in High Temperature Superconductors, and the astrophysical constraint machinery in Neutrino Astronomy. No result on the map blocks the next decade of scanning; the limits are magnet-bore and money.
• Requires (not on this map) a purpose-built dipole magnet above roughly nine tesla with a bore large enough to hold a dielectric disc stack, without which the preferred axion window cannot be scanned; an independent replication of the two-pion electron-positron annihilation cross-section, which now decides whether the muon anomaly exists; dilution refrigerators and near-quantum-limited amplifiers available at the scale of many simultaneous experiments rather than a few; an independently simulated post-inflationary axion mass window using different numerical methods, to test the factor-of-a-few spread in the current prediction; and a funding line whose evaluation criteria treat a published null as a delivered result.
• Enables identification of the dark matter if a detection occurs, a solution to the strong-CP problem, and — whatever the outcome — permanent constraints on any future model with a feebly coupled light sector.
• Adjacent Quantum Gravity for the tabletop programme that shares this hardware, Black Hole Physics Applications for the superradiance constraint, and Quantum Materials for the device physics the readout chain depends on.

Established “Decades of searching found nothing, so the searches were wasted.” The searches produced exclusion limits that constrain every model written since, permanently. The correct criticism of the field is not that it found nothing but that it sometimes describes its exclusions as failures, which both misrepresents the product and invites the criticism.

Frontier “The muon g−2 result proved new physics.” It did not, and by 2025 the theory consensus removed the discrepancy by changing the hadronic input rather than by changing the measurement. The measurement is superb and stands; the interpretation moved. Anyone citing the anomaly as motivation is citing a 2021 state of knowledge.

Established “Dark matter has been detected; mainstream physics ignores it.” The long-standing annual-modulation claim was not ignored: two experiments were built with the same target material specifically to check it, and both report incompatibility. Attention was paid, at cost, and the check failed. That is the opposite of suppression.

Speculative “The 17 mega-electronvolt boson is a fifth force.” One laboratory has reported anomalous angular correlations in several nuclear transitions over more than a decade. An independent experiment with superior resolution searched the key transition and saw no signal. The claim is not dead and it is not confirmed; it is a single-laboratory result awaiting replication, which is exactly how it should be described.

Handwave “Axion detection would give us new energy technology.” Conversion of halo axions in a strong magnetic field yields powers of order 10⁻²² watts. There is no engineering route from that to anything useful, and no serious proponent claims one. The claim appears in popular coverage and nowhere in the literature.

Frontier “Direct detection is nearly done because we have reached the neutrino floor.” The neutrino background is a floor for one technique at one set of masses, not for the search. Directional detection, lower thresholds for light dark matter, and entirely different couplings remain open. What is true is that the era of order-of-magnitude improvements per detector generation in spin-independent nucleon scattering has ended.

Speculative “Sterile neutrinos are established by multiple anomalies.” Several anomalies exist and they do not agree with each other on parameters, and the cleanest test of the most-cited one returned a null. A set of mutually inconsistent anomalies is weaker evidence than a single consistent one, not stronger, and the arithmetic of combining them is often done incorrectly in popular accounts.

Handwave “Hidden sectors explain unidentified aerial phenomena, anomalous propulsion, or vacuum energy extraction.” The couplings that define a hidden sector are, by construction, feeble — that is why the particles are hidden. A sector weak enough to have escaped fifty years of dedicated searching cannot exert macroscopic forces on vehicles. The two claims are in direct contradiction and cannot both be held.