Dark Matter: Early Considerations

Scientific Paper
Title	Dark Matter: Early Considerations
Read in full	Link to paper
Author(s)	Jaan Einasto
Keywords	Dark matter, galaxies, clusters of galaxies
Published	2004
No. of pages	18

Read the full paper here

Abstract

A review of the study of dark matter is given, starting with earliest studies and finishing with the establishment of the standard Cold Dark Matter paradigm in mid 1980-s. Particular attention is given to the collision of the classical and new paradigms concerning the matter content of the Universe. Also the amount of baryonic matter, dark matter and dark energy is discussed using modern estimates.

Overview

This is a historical review, written as a School lecture and circulated as astro-ph/0401341, in which Jaan Einasto traces the emergence of the Dark Matter concept from Ernst Öpik's 1915 estimate of the density of matter near the galactic plane to the acceptance of the Cold Dark Matter (CDM) model in the mid-1980s. Einasto is not a neutral chronicler: he was one of the principals. The 1974 Tartu paper of Einasto, Kaasik and Saar — which argued in Nature that all giant galaxies are surrounded by massive non-stellar "coronas" — is one of the two papers he identifies as having settled the question, the other being Ostriker, Peebles and Yahil (1974) from Princeton.

The review's organising claim is that dark matter was never "discovered" in a single observation. Einasto presents it explicitly as a Kuhnian paradigm shift: "as often in a paradigm shift, there is no single discovery, the new concept was developed step-by-step." He insists on separating two problems that are often conflated — local dark matter in the disk of our own Galaxy, and global dark matter around galaxies and clusters — and concludes that the first largely evaporated while the second did not. For a wiki whose contributors are frequently sceptical of dark matter, the chief value of the paper is that it is a first-hand account of how the case was actually assembled, including the objections raised against it at the time and the reasons they were dropped.

The argument

Local dark matter: a problem that dissolved

Öpik (1915) found no evidence for large amounts of invisible matter near the galactic plane. Oort (1932) disagreed, obtaining a total density up to twice the density of visible stellar populations — the so-called Oort limit. Grigori Kuzmin at Tartu reworked the problem using the ratio of vertical velocity and coordinate dispersions, C = σ_z/ζ_z, obtaining C = 68 km s⁻¹ kpc⁻¹ and a density ρ = 0.08 M_sun pc⁻³, in agreement with the known populations alone. Kuzmin's students Eelsalu and Jõeveer confirmed him; Hill (1960) and Oort (1960) did not. Einasto reports that Gilmore, Wyse & Kuijken (1989) eventually vindicated Kuzmin, and states flatly that "there is no evidence for the presence of large amounts of dark matter in the disk of the Galaxy." Any local invisible matter, he argues, must be baryonic — faint stars or jupiters — because non-baryonic matter is dissipationless and cannot settle into a flat disk.

Galaxy modelling and the M/L anomaly

Einasto's route into the global problem came through mass modelling rather than cosmology. He describes the generalised exponential density law he introduced for stellar populations,

ρ(a) = ρ(0) exp[−(a/ka₀)^1/N],

where a is the semi-major axis of the isodensity ellipsoid and N a shape parameter (N = 4 recovering de Vaucouleurs' law, N = 1 the exponential disk, N = ½ a Gaussian). Fitting M31 with populations whose mass-to-light ratios were constrained independently — by colours, spectra, and his own model of the physical evolution of stellar populations (Einasto 1971, developed in parallel with Beatrice Tinsley's) — produced a contradiction. Rotation data taken at face value forced M/L > 1000 at the periphery of M31, while every known old metal-poor halo population has M/L ≈ 1.

Einasto is candid that he first resolved this the wrong way, by postulating non-circular motions at large radii (Einasto 1969b): "As I soon realised, this was a wrong decision." Flat rotation curves made that escape untenable.

Coronas

The alternative — a population of unknown nature — was pressed on him by Enn Saar. Einasto coined "corona" rather than "halo" to mark that the component need not be stellar. He gives three reasons why a stellar corona fails: it has no place in the otherwise continuous kinematic-physical sequence of known populations; star formation goes as the square of local density, so such stars would have formed centrally and could not have been expanded outward, since contracting all other populations to zero radius would not supply the required energy; and star formation converts only about 1% of a gas cloud into stars, so producing so massive a population from primordial gas is implausible. Using companion galaxies as distant test bodies, the Tartu group found corona radii and masses exceeding those of the parent galaxies by an order of magnitude, and Einasto reported at the 1974 Elbrus Winter School that dark matter must be at least 90% of all matter.

The objections, and how they were answered

Einasto reproduces the contrary case fairly. Ambartsumian (1958) proposed that clusters are simply unstable and expanding rather than massive. Burbidge (1975) and Materne & Tammann (1976) argued that companions might not be physical members, that groups are bound with conventional masses, that cluster masses could be attributed to dominant cD galaxies, and that Big Bang nucleosynthesis plus the smoothness of the Hubble flow favour Ω ≈ 0.05. The replies he records are: morphological segregation of companions (ellipticals close in, gas-rich spirals further out, with the segregation radius scaling with primary luminosity) shows companions are real members; Faber et al. (1977) measured M/L = 3 for the Sombrero bulge, confirming that luminous populations are less massive than previously assumed; and Rubin, Ford & Thonnard (1978, 1980) established flat rotation curves for essentially all spirals.

From hot to cold dark matter

Neutrino dark matter (Cowsik & McClelland 1973, Szalay & Marx 1976, Rees 1977) was attractive because non-baryonic perturbations can grow during the radiation era while baryonic ones are damped, permitting amplitudes of order 10⁻³ at recombination and thus small Cosmic Microwave Background fluctuations. Einasto's own wedge-diagram work with Jõeveer (1978) — chains, superclusters and voids up to ≈ 70 h⁻¹ Mpc across, "cells of a honeycomb" — supported Zeldovich's pancake scenario built on hot dark matter, but showed two failures: simulated voids contained particles that real voids did not (an early indication of biased galaxy formation), and real superclusters are filamentary where the models were diffuse. The cold alternative (Blumenthal, Pagels & Primack 1982; Bond, Szalay & Turner 1982; Peebles 1982; Melott et al. 1983; Blumenthal et al. 1984) matched connectivity, multiplicity and correlation statistics, and became standard.

How much

Einasto assembles independent estimates: Ω_m = 0.28 ± 0.05 from distant supernovae, 0.31 ± 0.05 (h/0.65)^−1/3 from cluster gas fractions, 0.4 ± 0.1 from cluster abundance evolution, weighted mean Ω_m = 0.32 ± 0.03. From SDSS combined with WMAP (Tegmark et al. 2003) he quotes Ω_m = 0.30 ± 0.04, h = 0.70 ± 0.04 and h²Ω_b = 0.0232 ± 0.0012, giving Ω_b = 0.047 and hence Ω_DM = 0.25, with the remainder Dark Energy.

Assessment

The paper's real strength is documentary. Einasto records the internal logic of a research programme rather than its tidied-up textbook version, and he is unusually willing to describe his own errors — the non-circular-motion hypothesis, the missed contact with Tinsley's work. He also records the sceptics' arguments in their own terms instead of caricaturing them, which is rarer than it should be in reviews of this subject. The most robust part of the argument is negative and often overlooked: the case for dark matter strengthened not because the dynamical masses rose but because the luminous masses fell, once population synthesis and better velocity dispersions showed that M/L of real stellar populations is a few, not a few tens. That is a genuine measurement result, and any alternative account of galaxy dynamics must accommodate it.

The weaknesses are those of an advocate's history. First, the review treats the failure of stellar and gaseous coronas as establishing a non-baryonic substance, when what the arguments actually establish is that the gravitational discrepancy is not accounted for by known baryonic populations. The step from "the attraction is larger than the visible matter implies" to "there exists a new species of particle" is asserted rather than derived, and Einasto's closing sentence concedes the cost: "we still do not know of what non-baryonic particles the dark matter is made of." Two decades of direct-detection experiments since 2004 have not changed that.

Second, the epistemic standard invoked is candidly circular. Einasto quotes with approval Eddington's maxim, "No experimental result should be believed until confirmed by theory," and Zeldovich's banquet declaration of a "holy feeling of understanding the secrets of Nature". Read against the same paper's account of how Zwicky's 1933 Coma result was ignored for a quarter century, this reads less as method than as an account of how a community decides what to notice.

Third, the review sets aside without discussion the possibility that the dynamical laws themselves are at issue. Modified-inertia and modified-gravity accounts are dismissed in a single clause referring the reader to Sanders (1990); the systematic regularity that motivates them — the tight coupling between rotation-curve shape and baryonic surface density — is not mentioned. Nor are the Plasma Cosmology proposals that attribute large-scale filamentary structure to electromagnetic forces, though Einasto's own honeycomb result is precisely the sort of structure at stake. Readers of this wiki will also note that the concordance figures quoted at the end (Ω_m = 0.30, Ω_DM = 0.25) are derived within the CDM framework from CMB power-spectrum fitting, so they cannot serve as independent confirmation of it; Einasto presents them as convergence, but the convergence is partly definitional.

Taken on its own terms — as an account of how one paradigm displaced another, written by a participant — the paper is honest and valuable. Taken as an argument that dark matter exists as a substance, it shows clearly where the inference is empirical (falling M/L of stars, flat rotation curves, cluster dynamics) and where it is theoretical necessity imposed by structure formation.

See also

• Jaan Einasto
• Dark Matter
• Dark Energy
• Fritz Zwicky
• Big Bang
• Cosmic Microwave Background
• Hubble Constant
• Cosmological Constant
• Expanding Universe
• Gravitational Lensing
• Plasma Cosmology
• Neutrino