Exploration · Evidence · A deeper universe
Investigating invisible mass through galaxy rotation curves, gravitational lensing, clusters and cosmology.
Start here
You do not need a cosmology textbook to enter the project. These two ideas are related only because both are still physically unexplained.
Galaxies and clusters behave as if extra gravitating mass is present. We infer it from motion, lensing and cosmic structure—but we still do not know what it is made of.
The universe is not only expanding; its expansion is accelerating. Dark energy is the label for whatever drives that late-time behaviour in our models.
The universe today
In a Planck-like base flat ΛCDM reference model, ordinary matter is only a small fraction of the total cosmic energy budget.
See how the percentages are calculated →H²(a) = H₀² [ Ωra−4 + Ωma−3 + Ωka−2 + ΩΛ ]
Radiation dilutes fastest; matter follows volume; a cosmological constant remains constant. The quantitative cosmology laboratory later in the page lets you move through this history.
Key areas
Four visual entry points into the project. Each opens into a quantitative section below.
Cluster-scale inference
Map total mass with gravitational lensing.
Observation → reconstruction → mass distribution
Galaxy dynamics
Compare visible matter with measured rotation.
175 galaxies · published SPARC data
Early-universe constraints
Ask why all gravitating matter cannot simply be baryonic.
CMB · BAO · expansion history
Scientific leap
Turn a future microscopic discovery into measurable engineering questions.
sensing · astronomy · energy · propulsion gates
The research frontier
We can map where unseen gravity acts. We can constrain how the dark component behaves. We still do not know what it is—and that is precisely where the open research begins.
Explore the open questions03
How can we see something dark?
Each arrow changes the epistemic status of the information. Assumptions are part of the result, not footnotes to hide.
Photons locate light. Motion probes gravity. Lensing maps projected mass. Cosmology tests content and history. Detectors search for microscopic interactions.
Observatory context · directly credited
Dark matter is inferred through different measurements. These credited images are context—not substitutes for the quantitative data below.
Blue maps the lensing-derived mass distribution; pink maps X-ray-emitting gas. Dashed circles locate the two colliding galaxy clusters.
Credit: X-ray NASA/CXC/SAO · NIR NASA/ESA/CSA/STScI · processing NASA/STScI/J. DePasquale
A real spiral galaxy used as visual context. The selectable rotation curves below come from the SPARC catalogue, not this photograph.
Credit: NASA, ESA, S. Beckwith (STScI) and Hubble Heritage Team (STScI/AURA)
Temperature fluctuations constrain the universe’s matter content and the growth conditions for later cosmic structure.
Credit: ESA/Planck Collaboration
Wide, repeated imaging lets researchers map enormous galaxy populations and test small gravitational distortions statistically.
Credit: Olivier Bonin/SLAC National Accelerator LaboratoryEvery image is a real mission or observatory product. Full credits and primary-source links appear directly on each image.
Interactive galaxy workbench
The same transparent model is applied across spirals and gas-rich dwarfs. A fit describes gravitational evidence; it does not identify a particle.
Lab mode shows every available measured, modelled and inferred layer.
Observation → model
The rotation-curve plot is loading.
Diagnostic
Residual diagnostics are loading.
Parameter response
Likelihood surface is loading.
Fit telemetry
Audit trail
| Radius (kpc) | Observed (km/s) | σ (km/s) | Gas (km/s) | Disc, scaled (km/s) | Halo (km/s) | Total (km/s) | Residual / σ |
|---|
—Phase 3–4 · galaxies as a population
Filter one shared population and compare a global baryonic scaling relation with all resolved acceleration measurements. Correlation is evidence for a regularity—not, by itself, a microscopic explanation.
The population plot is loading.
The acceleration-relation plot is loading.
BTFR: Mbar = 0.5 L3.6 + 1.33 MHI. Galaxies without a published positive Vflat are excluded from that panel. The stellar mass-to-light ratio is an explicit fixed assumption, not a measurement.
RAR: gobs = Vobs²/R and gbar = (Vgas|Vgas| + 0.5Vdisk² + 0.7Vbulge²)/R. The reference curve uses g† = 1.2 × 10−10 m s−2. Non-positive derived accelerations are excluded; distance and inclination systematics are not propagated here.
Subsets: low/high surface brightness split at the catalogue median effective surface brightness; gas dominated means the stored fixed-M/L gas fraction exceeds 0.5. Quality flags are SPARC’s own: 1 high, 2 medium, 3 low.
| Galaxy | Type | Q | Distance (Mpc) | Inclination (deg) | SBeff (L☉/pc²) | Gas fraction | Vflat (km/s) | N |
|---|---|---|---|---|---|---|---|---|
| Population data are loading. | ||||||||
These views reproduce relationships within the selected assumptions. They do not establish whether the regularities arise from dark matter, baryonic physics, modified dynamics, selection effects, or a combination.
SPARC data and primary referencesPhase E · challenge the model
Compare the selected halo configuration with baryons alone and two acceleration-based phenomenological relationships. This tests predictions; it does not declare a winning theory.
Alternative-model velocity comparison is loading.
Alternative-model residual comparison is loading.
Radius, circular velocity, random velocity uncertainty and photometric/gas contributions from SPARC.
Stellar mass-to-light ratios, axisymmetric circular motion and the selected functional relationship between baryonic and observed acceleration.
Independent distance and inclination constraints, vertical dynamics, lensing, satellites, clusters and cosmology test whether one galaxy-scale relation generalises.
Empirical RAR: gpred = gbar / [1 − exp(−√(gbar/g†))]. It compactly describes a galaxy-scale regularity but is not itself a microscopic theory.
Simple ν: gpred = [½ + √(¼ + a₀/gbar)]gbar. This is a MOND-like phenomenological interpolation; relativistic completion and external-field assumptions are outside this calculation.
Comparison: χ² values use only quoted independent random velocity errors. Lower residuals under these settings do not establish physical truth or account for model flexibility and systematics.
| Radius (kpc) | Observed | σ | Baryonic | Phenomenological | Current halo | Phen. residual/σ | Halo residual/σ |
|---|---|---|---|---|---|---|---|
| Comparison data are loading. | |||||||
Phase F · clusters as gravitational laboratories
Change the lens geometry, reveal each measured tracer and test why a separation between hot gas, galaxies and reconstructed total mass is informative.
Cluster reconstruction is loading.
Schematic, not a dark-matter photograph. Lensing contours represent a model-dependent inversion of many background-galaxy shapes; the gas layer represents an X-ray tracer.
Optical images locate galaxies, X-ray photons trace hot plasma and coherent distortions appear in background-galaxy shapes.
Shape distortions constrain projected gravitational potential. Recovering a convergence or mass map requires source redshifts, calibration and inversion assumptions.
Gas experiences ram pressure; stars are nearly collisionless. When most reconstructed mass follows galaxies instead of gas, the dominant gravitating component is not the plasma.
Critical density: Σcrit = c²Ds/(4πG DlDls). A lower value means the chosen lens–source geometry is more efficient.
SIS scale: θE = 4π(σv/c)²Dls/Ds. This compact proxy connects velocity dispersion to angular lens strength; real merging clusters require multi-component mass models.
Inference limit: Lensing responds to projected total gravity. Calling the residual component “dark matter” combines that map with baryonic tracers and a gravity model.
| System | Lens z | Source z | kpc/arcsec | Σcrit (M☉/pc²) | θE (arcsec) | Evidence status |
|---|---|---|---|---|---|---|
| Cluster data are loading. | ||||||
Phase G · early universe to experiments
Connect the cosmic matter budget to the very different instruments that test candidate particles, fields and compact objects. Null results constrain models; they are not evidence that dark matter is absent.
Cosmological component evolution is loading.
BAO distance response is loading.
Radiation ∝ a⁻⁴, matter ∝ a⁻³, curvature ∝ a⁻², and the selected dark-energy response supplies the remaining background term.
Ω values alter the component budget; w₀–wₐ changes the dark-energy density history; H₀ rescales the distance ladder; BAO redshift selects the comparison point.
The browser evaluates background equations only. It does not recompute CMB spectra, DESI likelihoods or Boltzmann transfer functions.
Open the percentage mathematics →Primordial nucleosynthesis and CMB acoustic structure constrain the baryon density. Replacing Ωcdm with baryons changes photon–baryon oscillations instead of reproducing the same sky.
Before recombination, baryons are coupled to photons. A non-baryonic component can begin building gravitational wells without participating in that acoustic fluid.
DESI DR2 full-shape Lyα measurements moved toward the Planck ΛCDM prediction. The site therefore treats w₀–wₐ exploration as a competing-model exercise, not a discovery of evolving dark energy.
DESI Collaboration contextA fitted Ωcdm quantifies a gravitating component within a model. It does not specify mass, spin, couplings or a laboratory detection channel.
Candidate atlas
| Programme | Channel | Observable | Class | Current result/status | Updated | Source |
|---|---|---|---|---|---|---|
| Experiment records are loading. | ||||||
Phase C · Bayesian inference
Set explicit uniform priors, run four deterministic Metropolis chains and inspect credible intervals and convergence before interpreting a fitted halo.
Run posterior chains to create marginal distributions and a halo-parameter covariance plot.
Run posterior chains to inspect trace mixing.
Run posterior chains to inspect autocorrelation.
The rotation plot gains a 68% posterior predictive interval after sampling. It combines fitted-parameter uncertainty with the quoted SPARC random velocity errors. The current distance and inclination sensitivity settings are held fixed rather than marginalised, and correlated systematics remain outside this browser model. A posterior-predictive p-value near 0 or 1 can flag mismatch; it is not the probability that a halo model is true.
| Parameter | 16th percentile | Median | 84th percentile | Split R-hat | ESS |
|---|---|---|---|---|---|
| No posterior samples yet. | |||||
| Radius (kpc) | Observed | Model q16 | Model median | Model q84 | Predictive q16 | Predictive q84 |
|---|---|---|---|---|---|---|
| No posterior predictive intervals yet. | ||||||
Phase H · scales, frontiers and constrained futures
Explore what changes with physical scale, calculate a deliberately simplified encounter rate and distinguish active engineering from conditional or speculative future capability.
Recoil, lensing shear, conversion photon, timing shift or another calibrated observable.
Instrument calibration and a stated model transform the raw signal into a mass, field or event-rate constraint.
A reproducible interaction must establish mass, coupling and consistency with the astrophysical abundance.
Detection is not enough: useful engineering requires controllable production, capture, confinement or state manipulation.
Only then can sensing, communication, energy or propulsion be evaluated with ordinary conservation laws.
Dynamics, streams, microlensing and lensing
If the microscopic physics became controllable
These are conditional engineering studies, not predictions. The images are generated concept art used to visualise the engineering questions—not observational evidence and not a claim that these devices exist. Each idea starts from an explicit physics gate that current evidence has not yet crossed.
If a reproducible non-gravitational interaction exists, directional detectors could map the local dark-matter velocity field rather than only count events.
Author perspective →
Lensing, streams and precision dynamics already let us map unseen gravitating structure without identifying a particle.
Author perspective →
A weakly interacting messenger could penetrate ordinary matter, but the same weak coupling makes generation and detection extraordinarily difficult.
Author perspective →
Future field or stream sensors could become an independent reference for local gravitational structure if a stable directional signal were ever demonstrated.
Author perspective →
Ambient density provides a mass-flux ceiling, but usable power collapses unless capture and conversion physics exist at practical efficiencies.
Author perspective →
Any engine must transfer directed momentum and close a complete mass, power and heat ledger. Reactionless thrust is not implied by dark matter.
Author perspective →A real exclusion or sensitivity curve must specify an interaction operator, target response, detector exposure and statistical treatment.
Hypothesis laboratory
No extra spatial dimension is required for matter to be electromagnetically dark. A particle in ordinary 3+1-dimensional spacetime can simply lack a useful electromagnetic coupling. Extra-dimensional models are separate, testable beyond-Standard-Model hypotheses.
Weak or absent electromagnetic coupling is enough to prevent ordinary imaging. Gravity can still reveal the mass distribution.
Kaluza–Klein, braneworld or bulk-state scenarios can introduce additional states or mediators, but must earn their place through observable deviations—not visual intuition.
Collider spectra, modified force laws, missing-energy signatures or other model-specific observables must differ from an ordinary weakly coupled dark sector. “We cannot see it” alone is not evidence for extra dimensions.
Particle Data Group reviewsDiscovery-dependent engineering
Only after specifying interaction, capture, energy transfer and momentum transfer. This calculator makes the missing physics explicit and treats a failed engine as a useful result.
Engine feasibility calculation is loading.
Research Frontier
| Horizon | Possible capability | Validation gate | Confidence |
|---|---|---|---|
| Research-frontier data are loading. | |||
Phase I · evidence architecture
Select a node to inspect what it claims, where it comes from and what can break the inference. Arrows are relationships, not proof.
Start with the observation nodes, then follow their equations to an inference. The final “microscopic identity” node remains a hypothesis because gravity alone cannot tell us which candidate exists.
| From | Relationship | To |
|---|---|---|
| Evidence graph is loading. | ||
Author perspective · long-horizon synthesis
These pages separate my interpretation from the measured evidence. They are first-person scientific reflections, not claims that unresolved physics has been solved.
Capability roadmap · 2026–2076
This is a staged research pathway, not a promise that discovery will arrive on a schedule. Each stage has a measurable validation gate.
Cross-calibrate Rubin, Euclid and Roman lensing maps; expand resolved galaxy dynamics; keep tonne-scale rare-event detectors accumulating exposure.
Gate: consistent halo structure across instruments, wavelengths and analysis pipelines.Scale ultra-low-background detectors, directional recoil sensing, cryogenic thresholds and axion-band coverage while combining CMB lensing with galaxy surveys.
Gate: a repeatable signal with target, direction or annual-modulation tests—or stronger exclusions that retire model families.Condition future observatories on what earlier null results leave open: quantum sensors, interferometric concepts, distributed timing and next-generation underground targets.
Gate: independent replication plus a shared astrophysical and laboratory explanation. Mission names and architectures remain conditional.