Top 100 Unsolved Mysteries of the Universe, Episode 35: The CMB Spectral Distortion Problem. Picture an almost perfect thermal receipt. When a giant furnace is running at its most stable setting, every color band lines up with uncanny regularity, like a factory reference spectrum with no burrs and no missing pieces. In mainstream cosmology, the cosmic microwave background looks almost exactly like that kind of receipt from the early universe: a blackbody fingerprint so smooth that it seems as if cosmic history ironed away nearly every wrinkle. But the part that keeps cosmologists restless is not how smooth the receipt is. It is whether, in places almost too faint to see, a few hairline creases are still hiding in it. Maybe one band sits slightly above the ideal curve, another sinks slightly below, the way a pane of glass can look perfectly clean until slanted light reveals a thin oily film. That is the real spectral-distortion question: beyond the overall near-perfect blackbody, does the CMB still carry tiny departures - mu, y, or more complicated residual patterns - that record when the early universe received extra heat, when it failed to thermalize completely, and when later hot electrons brushed across it? You can also picture a pot of soup taken off the fire long ago. From a distance the surface looks uniform, like cream smoothed flat. But under slanted light you may still catch a thin sugar film from an earlier hard boil, a later oily sheen from gentle reheating, and a faint in-between watermark that belongs cleanly to neither stage. The CMB spectrum may be just like that: not a broken blackbody, but a nearly perfect one with faint time-stamped surface memories. This is where mainstream cosmology gets awkward. The problem is not the lack of explanations. The problem is the crowd of them. Silk damping, particle decay, annihilation, hot-electron scattering during reionization, late heating, foreground leftovers, absolute-calibration drift, and bandpass errors can all carve plausible-looking roughness into the spectrum. So the logic gets uncomfortable fast. If you rush to promote a tiny excess into a message from new physics, foregrounds and systematics immediately ask why that bump is not just your own ruler shaking. But if you dump every leftover into the noise bin in advance, you may wash away the real thermal handwriting of the early universe along with the junk. And even if future instruments do find a nonzero distortion, that still will not automatically point to one unique origin story, because very different heat-injection histories can project onto very similar-looking templates over a limited frequency range. EFT rewrites the problem at a deeper level. It does not begin by asking, 'Is the blackbody perfectly exact or not?' It begins by asking, 'Which stage of cosmic history wrote itself into which template?' In EFT, CMB spectral distortion is not first an ID card for one sacred origin script. It is a time-evolution ledger of the early worksite. If the high-redshift thermalization window leaves a mu_CMB signal, that is like the earliest sugar film from the hottest stage. If the lower-redshift scattering window leaves a y_CMB signal, that is like a later streak of hot oil dragged across the surface. And the transition window in between should not be lazily blended away just because a simpler fit looks cleaner. EFT wants it kept honestly as a frozen residual template, R(nu). Then mu, y, and R stop being three symbols fed into a fitting engine and become three layered traces from different eras, different injection channels, and different rearrangement rhythms. That changes what counts as evidence. EFT is not mainly staring at whether one frequency point rises a little above the curve. It is asking whether the three template coefficients remain stable, whether the residual template shows the right single-node structure, whether the zero-crossing frequency lands where it should, and whether the whole pattern converges toward a stable nonzero platform under harsher masks, deeper foreground cleaning, and tighter absolute calibration. In other words, EFT upgrades the question from 'does the spectrum have a tiny edge defect?' to 'which stage of cosmic construction signed its name here, and in what layer?' This also requires hard guardrails. First, EFT is not downplaying the success of the blackbody fit. Quite the opposite: it is precisely because the spectrum is so close to perfect that any surviving distortion would be so valuable. Second, EFT is not saying every small residual is automatically the universe speaking in first person. Until foregrounds, bandpass systematics, temperature drift, and calibration errors have been audited to exhaustion, any little hump can still be an instrument fingerprint. Third, EFT is not trying to turn mu, y, and R into three holy seals. Their real value is whether they can close the books together with temperature anisotropies, polarization, 21-centimeter tomography, and the later structure chain, all on the same early-universe baseplate. Only then does this beautiful thermal receipt rise from a pretty frequency curve into a real historical ledger. So the sentence to pin down for this episode is simple: CMB spectral distortion is not a few disposable burrs on the edge of a blackbody curve. It is the time-layered ledger of early-universe heat injection, scattering, and rearrangement written into the spectrum itself. And the way to decide whether it is real, and what it testifies for, is not to worship one small bump, but to see whether the mu, y, and R families - together with their residual structure - keep telling the same story under a harder audit. Tap the playlist for more. Next episode: The CMB Lensing Anomaly and A Posteriori Tension Problem. Follow and share—our new-physics explainer series will help you see the whole universe more clearly.
EFT
Language