Thermal Sunyaev-Zel'dovich Cluster Counts as a Cosmology Probe

What it is: hot gas casting a microwave shadow

A galaxy cluster contains a diffuse atmosphere of ionized gas — the intracluster medium (ICM) — with electron temperatures of 10⁷–10⁸ K (thermal energies kT_e ≈ 1–15 keV) and electron densities around 10⁻³ cm⁻³. As cosmic microwave background (CMB) photons pass through, roughly 1% or less scatter off these fast electrons via inverse Compton scattering. On average the cold 2.725 K photons gain a little energy, so the CMB spectrum seen through the cluster is distorted: photons are shifted from the Rayleigh-Jeans (low-frequency) side to the Wien (high-frequency) side.

• Below 217 GHz the cluster appears as a decrement — a cold spot against the CMB.
• Above 217 GHz it becomes an increment — a hot spot.
• At the crossover null (217 GHz) the effect nearly vanishes for non-relativistic gas.

Crucially, this is a fractional distortion of a fixed-temperature background, so its surface brightness does not dim with distance. That makes tSZ a nearly redshift-independent way to find clusters out to z > 1.5.

The mechanism: the Compton-y parameter

The strength of the distortion at a point on the sky is the dimensionless Compton-y parameter:

y = (σ_T k_B / m_e c²) ∫ n_e T_e dl

where σ_T is the Thomson cross-section, n_e the electron density, T_e the temperature, and the integral runs along the line of sight. Physically, y measures the total thermal energy of electrons integrated through the cluster — essentially the pressure of the gas, P_e = n_e k_B T_e, projected on the sky. Typical cluster central values are y ≈ 10⁻⁵ to 10⁻⁴.

The temperature change in the non-relativistic limit is ΔT/T_CMB = y · f(x), with x = hν/k_B T_CMB and f(x) = x·coth(x/2) − 4, which equals −2 in the Rayleigh-Jeans limit and passes through zero at x ≈ 3.83 (217 GHz). For very hot clusters, relativistic corrections shift the null by a few GHz and modify f(x) at the few-percent level.

Integrating y over the cluster's solid angle gives Y_SZ = ∫ y dΩ, a measure of the cluster's total thermal energy — and hence its mass.

Key quantities: the Y–M scaling relation and a worked example

Cosmology from cluster counts needs to convert the observable Y_SZ into a halo mass. Self-similar collapse predicts a tight power law:

E(z)^(−2/3) · D_A² · Y_500 ∝ M_500^(5/3)

where M_500 is the mass within a radius enclosing 500× the critical density, D_A is the angular-diameter distance, and E(z) = H(z)/H_0 encodes expansion. The 5/3 slope follows because Y ∝ M·T and the virial temperature scales as T ∝ M^(2/3).

• A M_500 ≈ 5×10¹⁴ M_☉ cluster at z ≈ 0.3 has kT_e ≈ 5 keV and integrated Y_500 ≈ 10⁻⁴ Mpc² — comfortably above Planck's detection floor of a few ×10¹⁴ M_☉.
• The most massive clusters reach ~2×10¹⁵ M_☉ with kT_e ≈ 12–15 keV.

The abundance of such clusters is set by the halo mass function (Press-Schechter / Tinker), which is exponentially sensitive to σ₈ — a 5% change in σ₈ can change the number of massive clusters by tens of percent.

How it's observed: mm-wave surveys and cluster catalogs

Detecting tSZ requires arcminute-resolution millimeter telescopes with multiple frequency bands to separate the SZ spectrum from the primary CMB, dusty galaxies, and radio sources. Three surveys dominate:

• Planck (all-sky, 2013–2015): its multi-frequency coverage (100–857 GHz) exploited the 217 GHz null to build a catalog of ~1,650 SZ sources, ~1,200 confirmed clusters, mostly at z < 1.
• ACT (Atacama Cosmology Telescope, Chile): ~1.4 arcmin resolution, thousands of clusters over ~13,000 deg², reaching z > 1.5.
• SPT (South Pole Telescope): deep, high-resolution surveys (SPT-SZ, SPTpol, SPT-3G) that pioneered SZ-selected cluster cosmology with samples like the SPT-SZ 2500 deg² catalog.

Clusters are then followed up with optical/IR photometry for redshifts and with X-ray (Chandra, XMM, eROSITA) and weak-lensing data to calibrate mass. The observable-mass calibration — especially from weak gravitational lensing — is now the pacing item for these analyses.

Comparison with related probes and regimes

tSZ counts sit alongside several cousins, each with distinct strengths:

• X-ray counts depend on n_e² and suffer (1+z)⁴ surface-brightness dimming, so they excel at low z but fade at high z where tSZ still shines.
• Kinetic SZ (kSZ) comes from the bulk peculiar motion of the gas (Doppler, ∝ v/c · τ_e), has a blackbody spectrum with no 217 GHz null, and probes velocities rather than temperature — it is ~10–100× weaker than tSZ.
• Weak-lensing counts measure mass directly but with low signal-to-noise per object; they are the gold standard for calibrating SZ masses rather than a competing selection.
• tSZ power spectrum and bispectrum use the diffuse, unresolved SZ background (C_ℓ^yy ∝ σ₈^~8) as a complementary, calibration-light probe.

Because tSZ abundance tracks late-time growth while the primary CMB fixes conditions at z ≈ 1100, comparing them tests whether structure grew as ΛCDM predicts — a powerful lever on dark energy and modified gravity.

Significance, tensions, and open questions

tSZ cluster counts deliver competitive constraints on the combination σ₈(Ω_m/0.27)^0.3 ≈ 0.78. ACT+WMAP found σ₈ ≈ 0.83 and Ω_m ≈ 0.29; SPT and Planck reach similar precision. But a famous puzzle emerged: Planck's SZ cluster counts preferred a lower σ₈ (~0.75) than Planck's own primary-CMB fit (~0.83).

• The discrepancy hinges on the hydrostatic mass bias (1−b): X-ray masses may underestimate true masses by 20–40% because of non-thermal pressure (turbulence, bulk motions) the assumption of hydrostatic equilibrium ignores.
• Reconciling SZ counts with the primary CMB requires (1−b) ≈ 0.6, i.e. a ~40% bias — larger than most weak-lensing calibrations find.

This mass-calibration uncertainty is the central open problem. Ongoing weak-lensing programs (DES, HSC, KiDS, Euclid, Rubin/LSST) and next-generation SZ surveys (SPT-3G, Simons Observatory, CMB-S4) aim to pin down (1−b), sharpening whether the σ₈ tension is a systematic or a genuine crack in ΛCDM.

Thermal SZ compared with related cluster-detection and CMB probes

Probe	Signal / observable	Redshift dependence	Primary cosmology sensitivity
Thermal SZ counts	CMB spectral distortion, Compton-y ∝ ∫ n_e T_e dl	Surface brightness independent of z	σ₈, Ω_m via cluster abundance growth
X-ray cluster counts	Bremsstrahlung, L_X ∝ n_e² Λ(T)	Dims as (1+z)⁴ (cosmological dimming)	σ₈, Ω_m; strong at low z
Kinetic SZ	Doppler shift ∝ (v/c) τ_e	Independent of z	Peculiar velocities, reionization
Weak-lensing counts	Shear of background galaxies	Depends on source-lens geometry	σ₈, Ω_m; direct mass, low S/N per cluster
Primary CMB (Planck)	Acoustic peaks at z ≈ 1100	Single early-universe snapshot	Ω_m, σ₈ extrapolated from z ≈ 1100