3D thermal-dust maps
Gaia extinction cubes say how much dust sits at which distance. They are not millimetre emission maps. They still tell you which sightlines are one cloud and which are a stack.
Three models · one cosmology prompt
Dedusting CMB B-modes when the spectral index will not sit still.
Grok 4.6, GPT-5.6 Sol, and Claude Fable 5.1 independently, then in drive pairs, then as critic and optimist, answered how 3D thermal-dust maps, grain-size physics, and the sky-projected magnetic field can clean polarized Galactic dust out of primordial B-mode searches.
Explain how 3D maps of the thermal dust emission, together with knowledge on the dust grain size distribution and the sky-projected magnetic field can help dedusting of CMB whose current B-mode polarization studies' suffer from decorrelation of the spectral index.
Why a single spectral index fails
Two dust layers with different temperatures and a misaligned field. Drag frequency: the net polarization angle moves. A one-template cleanup cannot follow that rotation.
Why this question exists
Inflationary gravitational waves would leave a curl pattern in CMB polarization — B-modes — with amplitude set by the tensor-to-scalar ratio r. Galactic thermal dust is brighter than the r ~ 10⁻³ signal at the frequencies where ground and space experiments look. If dust at 150 GHz is not a scaled copy of dust at 353 GHz, a single modified-blackbody template leaks residual B-modes that look like a primordial detection.
Three ingredients, three factors of a complex number
Grain size and alignment
Only large, aligned grains polarize. Size sets temperature and who is aligned; millimetre β is mostly dielectric. That is why polarized and total-intensity SEDs need not match.
Sky-projected B
Starlight tomography and HI fibres give the plane-of-sky angle of each layer. Misaligned layers plus different SEDs rotate the net polarization with frequency. That rotation is what a scalar spectral index cannot represent.
How the matrix was run
Independent
Each model wrote alone. No other essay was visible.
Drive
Every model wrote briefs for the other two; the driven model wrote the essay. Six directed pairs.
Critic + optimist
Each independent draft was reviewed twice (roles swapped), then the primary synthesized. Six revised essays.