Virial Theorem for Molecular Clouds

Every star that has ever formed began as an overdensity inside a molecular cloud — the coldest, densest, most fragile phase of the interstellar medium. Whether such a cloud collapses into stars, hovers in quasi-equilibrium, or is shredded back into diffuse gas is decided by a single competition: self-gravity, pulling relentlessly inward, versus a coalition of supports — thermal pressure, supersonic turbulence, magnetic tension, and the confining weight of the surrounding medium. The scalar virial theorem is the accountant that tallies this competition. It is not a new force law; it is the ordinary MHD momentum equation, multiplied by position and integrated over the whole cloud, until the local statement of Newton's second law becomes a global energy budget. The theorem's central object is the moment of inertia $I=\int\rho r^2\,\mathrm{d}V$ , a proxy for the cloud's size. Its second time derivative $\ddot I$ is driven by the imbalance of energies: gravity ( $W<0$ ) drives $\ddot I<0$ and collapse, while the positive kinetic, thermal, and magnetic terms drive expansion, and equilibrium is the knife-edge $\ddot I=0$ where they cancel. Because each term scales differently with mass, radius, temperature, field strength, and external pressure, the virial equation cleanly separates the regimes of star formation — thermally supported cores, turbulence-dominated giant clouds, and magnetically supported subcritical clumps — and hands observers a way to weigh a cloud from nothing but its linewidth and size. This article builds the theorem from the momentum equation, identifies every term, derives the uniform-sphere and magnetic results in full, and connects them to what CO maps and Zeeman measurements see.

Key takeaways

The scalar virial theorem is the moment of the MHD momentum equation: $\tfrac{1}{2}\ddot I = 2(T-T_s)+2U+W+\mathcal{M}$ , balancing kinetic $T$ , surface pressure $T_s$ , internal $U$ , gravity $W$ , and magnetic $\mathcal{M}$ energies.
In equilibrium ( $\ddot I=0$ ) the negative gravitational term must be balanced by the positive supports; too much gravity means $\ddot I<0$ and runaway collapse.
For a uniform sphere $W=-\tfrac{3}{5}GM^2/R$ , and setting $2T=|W|$ gives the virial mass $M_{\rm vir}=5\sigma^2 R/G$ ; the dimensionless virial parameter $\alpha_{\rm vir}=5\sigma^2 R/GM=M_{\rm vir}/M$ then measures boundedness ( $\alpha_{\rm vir}\lesssim1$ bound and collapse-prone, $\gg1$ unbound).
Magnetic support is scale-free and set by the mass-to-flux ratio: clouds above the critical value $(M/\Phi_B)_{\rm crit}=c_\Phi/\sqrt{G}$ (with $c_\Phi\approx0.16$ ) are supercritical and cannot be held up by any field.
Key idea: multiply the fluid momentum equation by position and integrate over the cloud — the result is a global energy balance $\tfrac{1}{2}\ddot I = 2(T-T_s)+2U+W+\mathcal{M}$ , so equilibrium requires the negative gravitational term to be balanced by the positive kinetic, thermal, surface, and magnetic terms.

The core insight

Multiply the fluid momentum equation by position and integrate over the cloud: the result is a global energy balance $\tfrac{1}{2}\ddot I = 2(T-T_s)+2U+W+\mathcal{M}$ , so equilibrium ( $\ddot I=0$ ) requires the negative gravitational term to be balanced by the positive kinetic, thermal, surface, and magnetic terms.

Where this lives today

The virial theorem is the daily working tool of star-formation research. Wide-field CO surveys — from the Milky Way plane with the CfA and Nobeyama telescopes to whole galaxies with ALMA — catalogue thousands of clouds, with a size, linewidth, and virial parameter for each. The $\alpha_{\rm vir}$ distributions are a central diagnostic: most massive clouds cluster near $\alpha_{\rm vir}\sim1$ – $2$ , consistent with marginal binding, while the smallest clumps scatter to high values, marking them as transient and pressure-confined. On the magnetic side, Zeeman splitting of OH and CN lines, combined with dust-polarization maps from Planck, ALMA, and SOFIA/HAWC+, places most dense cores at $\mu_\Phi\sim2$ – $3$ — supercritical, as Example 2 found — supporting a picture in which fields regulate but do not prevent collapse. The open problems are live: does supersonic turbulence truly support clouds, or merely delay and fragment their collapse? Is $\alpha_{\rm vir}\approx1$ equilibrium or a fleeting stage of global contraction? The theorem does not settle these debates, but every side is argued in its language.

The physical picture

Picture the cloud as a single object pulled in four directions at once. Gravity is the only attractive player: unopposed, it would drive the moment of inertia to zero in a free-fall time. Against it stand three expansive agents. Thermal pressure pushes isotropically outward with a strength set by temperature. Turbulence acts like an effective pressure with a "sound speed" — the dispersion $\sigma$ — many times the true sound speed, because the gas motions are supersonic. Magnetic tension resists compression across field lines but not along them, so it holds up only a limited mass per unit flux. A fourth agent, the surface pressure of the ambient ISM, pushes inward and helps gravity confine the smallest cores. The virial theorem adds these up as energies rather than forces.

Figure

The virial tug-of-war: gravity W = −3/5 GM²/R pulls inward while thermal 2U, turbulent 2T, and magnetic ℳ push outward and ambient pressure squeezes the surface. Equilibrium (d²I/dt² = 0) demands 2(T − T_s) + 2U + W + ℳ = 0; excess gravity makes the cloud collapse

The convenience of the theorem is that near equilibrium the gravitational and kinetic terms are comparable, $GM^2/R\sim M\sigma^2$ , so $M\sim\sigma^2 R/G$ up to a factor of a few. An astronomer who measures a cloud's radius and velocity dispersion can therefore weigh it without assuming how gas turns into light, and comparing that dynamical mass to the CO-luminosity mass is a pillar of the claim that giant molecular clouds are roughly bound.

Figure

The virial parameter α_vir = 5σ²R/GM sorts clouds into regimes: below 1 strongly bound and collapsing, near 1–2 virialized, above 2 unbound and dispersing. A cloud with σ = 3 km/s and R = 20 pc has virial mass M_vir = 5σ²R/G = 2.1×10⁵ M⊙ and, with a comparable CO mass, sits near α_vir ≈ 1

The core equations

The whole theory is one master equation plus the energy terms that fill it: the second derivative of the moment of inertia as a sum of global energies — exact for an ideal magnetized fluid, with no assumption of equilibrium yet.

\frac{1}{2}\frac{\mathrm{d}^2 I}{\mathrm{d}t^2} = 2(T-T_s) + 2U + W + \mathcal{M} \qquad (1)

where: $I=\int\rho r^2\,\mathrm{d}V$ — moment of inertia (a size proxy); $T=\tfrac{1}{2}\int\rho v^2\,\mathrm{d}V$ — total kinetic energy, bulk plus turbulent; $T_s=\tfrac{1}{2}\oint P_{\rm ext}\,\mathbf{r}\cdot\mathrm{d}\mathbf{S}$ — surface (external-pressure) term; $U$ — internal thermal energy, with $2U=3\int P\,\mathrm{d}V$ ; $W$ — gravitational potential energy (negative); $\mathcal{M}=\int (B^2/8\pi)\,\mathrm{d}V$ — magnetic energy. Reading it: equilibrium is $\ddot I=0$ ; if the negative $W$ overwhelms the positive supports, the cloud accelerates inward. The gravitational term is the engine of collapse. For a uniform-density sphere it has an exact, famous coefficient, obtained by assembling the sphere shell by shell.

W = -\frac{3}{5}\frac{GM^2}{R} \qquad (2)

where: $G=6.674\times10^{-8}\,\mathrm{cm^3\,g^{-1}\,s^{-2}}$ ; $M$ — total cloud mass; $R$ — cloud radius. Reading it: more centrally concentrated clouds have a coefficient larger than $3/5$ , so this is a conservative (least-bound) estimate; the $M^2/R$ scaling is universal. Setting the turbulent kinetic energy against gravity turns equation (2) into a measurable mass — the most-used result in molecular-cloud astronomy.

M_{\rm vir} = \frac{5\,\sigma^2 R}{G} \qquad (3)

where: $\sigma$ — one-dimensional velocity dispersion (from the line width, $\sigma=\Delta v/2.355$ for a Gaussian FWHM $\Delta v$ ); $R$ — cloud radius. Reading it: two observables deliver a dynamical mass. A common refinement replaces the leading $5$ by $5a$ with $a\sim1$ for centrally concentrated clouds; the uniform-sphere value is the anchor. Dividing the kinetic energy by the gravitational binding gives a dimensionless verdict on boundedness.

\alpha_{\rm vir} \equiv \frac{2T}{|W|} = \frac{5\sigma^2 R}{GM} = \frac{M_{\rm vir}}{M} \qquad (4)

where: $\alpha_{\rm vir}$ — virial parameter; $M$ — the actual (independently measured) cloud mass. Reading it: $\alpha_{\rm vir}\approx1$ is equilibrium, $\lesssim1$ bound and collapse-prone, $\gg1$ unbound; since $\alpha_{\rm vir}=M_{\rm vir}/M$ it is exactly the ratio of dynamical to true mass. Magnetic support obeys a scale-free rule: because both the magnetic and gravitational energies scale as $1/R$ , whether a field can hold a cloud up depends only on how much mass threads a given flux.

\left(\frac{M}{\Phi_B}\right)_{\rm crit} = \frac{c_\Phi}{\sqrt{G}}, \qquad c_\Phi=\sqrt{\frac{5}{18\pi^2}}\approx0.17 \;\;(\text{uniform sphere}) \qquad (5)

where: $\Phi_B=\pi R^2 B$ — magnetic flux threading the cloud; $B$ — mean field strength (Gaussian units); $c_\Phi$ — dimensionless, $\approx0.17$ (uniform sphere), $\approx0.16=1/2\pi$ (flux-freezing), $\approx0.13$ (Mouschovias–Spitzer). Reading it: a cloud above critical is supercritical — no field can stop it — while a subcritical cloud is magnetically supported and contracts only by ambipolar diffusion. Finally the two "soft" supports have simple closed forms for comparing their magnitudes against gravity.

2U = 3\int_V P\,\mathrm{d}V = 3\,\frac{M}{\mu m_H}\,kT, \qquad \mathcal{M} = \int_V \frac{B^2}{8\pi}\,\mathrm{d}V \approx \frac{B^2 R^3}{6} \qquad (6)

where: $P=nkT$ — thermal pressure; $\mu\approx2.3$ — mean molecular weight; $m_H$ — hydrogen mass; $k=1.381\times10^{-16}\,\mathrm{erg\,K^{-1}}$ ; $T$ — gas temperature (10–20 K in dense clouds). Reading it: the thermal term grows only linearly with $T$ , which is why cold clouds are so easy for gravity to beat and why turbulence and magnetism do the real support work.

Derivation ADeriving the scalar virial theorem from the momentum equation

The goal is to convert the local statement of Newton's second law for a magnetized fluid element into a single global energy balance for the whole cloud. We start from the ideal-MHD momentum equation, take its first moment by dotting with the position vector $\mathbf{r}$ and integrating over the cloud volume, and identify each resulting integral as a recognizable energy — kinetic, thermal, gravitational, magnetic, plus a surface term. The payoff is equation (1), which holds whether or not the cloud is in equilibrium.

Step 1. Write the ideal-MHD momentum equation, $\rho\,\mathrm{d}\mathbf{v}/\mathrm{d}t = -\nabla P - \rho\nabla\Phi + \tfrac{1}{4\pi}(\nabla\times\mathbf{B})\times\mathbf{B}$ , where the three forces per unit volume are the thermal-pressure gradient, self-gravity through the potential $\Phi$ , and the Lorentz (magnetic tension plus pressure) force. Why: this is the complete force law for a molecular cloud treated as a perfectly conducting fluid — every virial term descends from one of these three forces.

Step 2. Take the moment: dot the equation with $\mathbf{r}$ and integrate over the volume, forming $\int\rho\,\mathbf{r}\cdot(\mathrm{d}\mathbf{v}/\mathrm{d}t)\,\mathrm{d}V$ on the left. Using $I=\int\rho r^2\,\mathrm{d}V$ and mass conservation, this is exactly $\tfrac{1}{2}\ddot I - 2T$ , with $2T=\int\rho v^2\,\mathrm{d}V$ . Why: differentiating $I$ twice yields one kinetic-energy term and one acceleration moment, so the inertia side of Newton's law becomes a rate of change of size minus a kinetic energy.

Step 3. Reduce the pressure force by parts: $-\int\mathbf{r}\cdot\nabla P\,\mathrm{d}V = 3\int P\,\mathrm{d}V - \oint P\,\mathbf{r}\cdot\mathrm{d}\mathbf{S}$ . The volume piece is $2U=3\int P\,\mathrm{d}V$ , the surface piece the confining term $-2T_s$ . Why: integration by parts splits pressure into an internal support that pushes out and a boundary term that lets the ambient medium push in — the two must never be conflated.

Step 4. Reduce the gravity force: $-\int\rho\,\mathbf{r}\cdot\nabla\Phi\,\mathrm{d}V = W$ , the gravitational potential energy $W=-\tfrac{1}{2}\int\rho\Phi\,\mathrm{d}V$ , negative for any self-bound configuration. Why: the moment of self-gravity is precisely the potential energy, so gravity enters as a single negative number — the term that drives collapse.

Step 5. Reduce the Lorentz force via the Maxwell stress tensor: $\tfrac{1}{4\pi}\int\mathbf{r}\cdot[(\nabla\times\mathbf{B})\times\mathbf{B}]\,\mathrm{d}V = \mathcal{M} + (\text{surface field terms})$ , the interior piece being $\mathcal{M}=\int (B^2/8\pi)\,\mathrm{d}V$ . Why: magnetic tension resists compression like an anisotropic pressure, and its moment is the positive field energy — but only up to the flux it carries.

Step 6. Assemble: $\tfrac{1}{2}\ddot I = 2T + 2U + W + \mathcal{M} - \oint P\,\mathbf{r}\cdot\mathrm{d}\mathbf{S}$ , and fold the surface integral into $T_s$ to reach the standard form $\tfrac{1}{2}\ddot I = 2(T-T_s) + 2U + W + \mathcal{M}$ . Why: collecting the moments of all three forces yields the global ledger; $\ddot I=0$ is the statement that the cloud neither expands nor contracts. *As a consistency check: with no pressure, field, or external medium the balance reduces to $2T+W=0$ , the classical gravitational virial theorem; every term has dimensions of energy.*

Go deeper · Derivation B · 1 minGravitational energy of a uniform sphere and the virial mass

Derivation BGravitational energy of a uniform sphere and the virial mass

The goal is to pin down two of the most-used numbers in the subject: the exact coefficient $3/5$ in the self-gravity of a uniform sphere, and the resulting virial-mass estimator $M_{\rm vir}=5\sigma^2 R/G$ . We build the sphere shell by shell to get $W$ , then balance it against the turbulent kinetic energy in the regime where turbulence, not thermal pressure or magnetism, provides the support — the regime of giant molecular clouds. The result is a mass measured from a linewidth and a size alone.

Step 1. Bring a shell of mass $\mathrm{d}m$ from infinity to radius $r$ against the gravity of the mass $m(r)$ already inside: $\mathrm{d}W = -G\,m(r)\,\mathrm{d}m/r$ . Why: assembling the sphere piece by piece counts every pair of mass elements exactly once, which is what potential energy means.

Step 2. Insert uniform density $\rho=3M/(4\pi R^3)$ , so $m(r)=\tfrac{4}{3}\pi r^3\rho$ and $\mathrm{d}m=4\pi r^2\rho\,\mathrm{d}r$ , giving $W=-\tfrac{16\pi^2 G\rho^2}{3}\int_0^R r^4\,\mathrm{d}r$ . Why: every quantity becomes an explicit power of $r$ , so the assembly integral is elementary.

Step 3. Integrate, $\int_0^R r^4\,\mathrm{d}r = R^5/5$ , giving $W=-\tfrac{16\pi^2 G\rho^2 R^5}{15}$ , and re-express $\rho$ in terms of $M$ to reach $W=-\tfrac{3}{5}GM^2/R$ . Why: the $R^5$ combines with $\rho^2\propto M^2/R^6$ to leave the universal $M^2/R$ scaling and the clean $3/5$ .

Step 4. Balance turbulence against gravity: with $\langle v^2\rangle=3\sigma^2$ , the kinetic energy is $2T=M\langle v^2\rangle=3M\sigma^2$ ; setting $2T=|W|$ gives $3M\sigma^2=\tfrac{3}{5}GM^2/R$ . Why: virial equilibrium for a turbulence-supported cloud is exactly $2T=|W|$ , the relevant limit for a GMC where $\sigma\gg c_s$ .

Step 5. Cancel one power of $M$ and solve: $M_{\rm vir}=5\sigma^2 R/G$ — only the observed dispersion and radius appear. Why: the mass drops out once, leaving a formula that weighs the cloud with no assumed light-to-mass conversion — the whole point of a virial mass. *As a check: the $3/5$ coefficient is the standard uniform-sphere result, and the numbers $\sigma=3\,\mathrm{km\,s^{-1}}$ , $R=20\,\mathrm{pc}$ give $M_{\rm vir}=2.1\times10^5\,M_\odot$ .*

Go deeper · Derivation C · 1 minThe critical mass-to-flux ratio for magnetic support

Derivation CThe critical mass-to-flux ratio for magnetic support

The goal is to show that magnetic support is fundamentally different from thermal support: it does not depend on the cloud's size at all, only on the ratio of its mass to the flux threading it. We compute the magnetic energy of a uniformly magnetized sphere, express it through the conserved flux, balance it against self-gravity, and watch the radius cancel — leaving a critical mass-to-flux ratio that decides, once and for all, whether a field can prevent collapse.

Step 1. Take a uniform field $B$ through a sphere of radius $R$ , so $\mathcal{M}\approx (B^2/8\pi)\cdot\tfrac{4}{3}\pi R^3 = B^2 R^3/6$ . Why: the field energy is the energy density $B^2/8\pi$ times the volume; the exact prefactor shifts with geometry but not the scaling.

Step 2. Under flux freezing the flux $\Phi_B=\pi R^2 B$ is conserved as the cloud contracts, so $B=\Phi_B/\pi R^2$ and $\mathcal{M}=\Phi_B^2/(6\pi^2 R)$ . Why: it is the flux, not $B$ , that is fixed for a given cloud, so rewriting $\mathcal{M}$ in terms of $\Phi_B$ exposes how support behaves under compression.

Step 3. Note the crucial coincidence: $\mathcal{M}\propto 1/R$ and $W=-\tfrac{3}{5}GM^2/R\propto 1/R$ share the same radial scaling, so their ratio is independent of $R$ . Why: compressing the cloud strengthens the field and deepens the potential in lockstep, so magnetic support neither gains nor loses as the cloud shrinks.

Step 4. Set $\mathcal{M}+W=0$ : $\Phi_B^2/(6\pi^2 R)=\tfrac{3}{5}GM^2/R$ ; the radius cancels and $M/\Phi_B=\sqrt{5/(18\pi^2 G)}\approx0.17/\sqrt{G}$ . Why: with $R$ gone, magnetic support reduces to one dimensionless number, $M/\Phi_B$ measured against $c_\Phi/\sqrt{G}$ .

Step 5. Interpret: above this critical value the cloud is supercritical — gravity wins at every radius; below it the cloud is subcritical and magnetically supported, contracting only as ambipolar diffusion lets neutrals drift past ions and slowly raises $M/\Phi_B$ . Why: the sign of $\mathcal{M}+W$ is fixed once and for all by $M/\Phi_B$ , so a field either always holds a given cloud up or never does. *As a check: the uniform-sphere coefficient $0.17$ brackets the flux-freezing value $1/2\pi=0.16$ and the Mouschovias–Spitzer value $0.13$ , confirming the scale-independence of magnetic support.*

Limiting cases

The full balance $2(T-T_s)+2U+W+\mathcal{M}=0$ contains every classical star-formation criterion as a special case: drop the terms that are small in a given regime and a familiar result falls out.

Case	Result	Meaning
Thermal only (Jeans limit)	$2U+W=0 \Rightarrow M_J\sim c_s^3/(G^{3/2}\rho^{1/2})$	Neglect turbulence, field, and surface pressure: recovers the Jeans mass; collapse when $M>M_J$ .
Turbulence-dominated (GMCs)	$2T+W\approx0,\; T=\tfrac12 M\sigma^2,\; \sigma\gg c_s$	Supersonic linewidths supply the support; yields $M_{\rm vir}=5\sigma^2 R/G$ and $\alpha_{\rm vir}\approx1$ .
Magnetically supported (subcritical)	$\|\mathcal{M}\|>\|W\| \Leftrightarrow M/\Phi_B<(M/\Phi_B)_{\rm crit}$	Field prevents collapse at any density; only ambipolar diffusion allows slow contraction.
Pressure-confined clump	$2U+W=2T_s$ , $T_s=\tfrac12\oint P_{\rm ext}\mathbf{r}\cdot\mathrm{d}\mathbf{S}$	Bonnor–Ebert regime: ambient pressure confines an isothermal sphere, stable below $M_{\rm BE}\approx1.18\,c_s^4/(G^{3/2}P_{\rm ext}^{1/2})$ .

Common trapThe virial theorem proves a cloud is in equilibrium

It proves no such thing. The scalar virial theorem is a necessary global constraint, not a sufficient local one, and $\alpha_{\rm vir}\approx1$ is at best consistent with equilibrium, not a demonstration of it. A cloud can satisfy the global balance while sub-regions collapse into stars; an unbound cloud ( $\alpha_{\rm vir}\gg1$ ) can still form stars in its densest cores. Two further traps live in the same equation. First, do not confuse the volume term $2U=3\int P\,\mathrm{d}V$ (internal pressure, which supports) with the surface term $T_s=\tfrac12\oint P_{\rm ext}\,\mathbf{r}\cdot\mathrm{d}\mathbf{S}$ (ambient pressure, which confines) — opposite senses. Second, the kinetic term contains all motions: in a GMC the supersonic turbulent part dwarfs the thermal one, so identifying $\sigma$ with the sound speed underestimates the support by an order of magnitude.

Worked examples

1 · Weighing a giant molecular cloud (fully worked)

A CO survey maps a cloud of radius $R=20\,\mathrm{pc}$ and velocity dispersion $\sigma=3\,\mathrm{km\,s^{-1}}$ . Find its virial mass. If the CO-luminosity mass is $M_{\rm CO}\approx2\times10^5\,M_\odot$ , is the cloud bound?

Step 1. Convert to CGS: $\sigma^2=9\times10^{10}\,\mathrm{cm^2\,s^{-2}}$ and $R=20\times3.086\times10^{18}=6.17\times10^{19}\,\mathrm{cm}$ .

Step 2. Apply equation (3): $M_{\rm vir}=5\sigma^2 R/G = 5(9\times10^{10})(6.17\times10^{19})/(6.674\times10^{-8}) = 4.16\times10^{38}\,\mathrm{g}$ .

Step 3. In solar masses, $M_{\rm vir}=4.16\times10^{38}/1.989\times10^{33}=\mathbf{2.1\times10^{5}\,M_\odot}$ — a typical GMC.

Step 4. Form the virial parameter with the CO mass: $\alpha_{\rm vir}=M_{\rm vir}/M_{\rm CO}=2.1\times10^5/2\times10^5\approx\mathbf{1.05}$ . The dynamical and luminosity masses agree, so the cloud is close to virial equilibrium and gravitationally bound — a value this near unity is why GMCs are said to be roughly virialized.

2 · Is the core magnetically supported? (fully worked)

Zeeman observations of a dense core give a line-of-sight field $B=30\,\mu\mathrm{G}$ and a column density $N(\mathrm{H_2})=1\times10^{22}\,\mathrm{cm^{-2}}$ . Is the core sub- or supercritical?

Step 1. The mass-to-flux ratio in units of the critical value, with $c_\Phi=1/2\pi$ , is the widely used $\mu_\Phi=7.6\times10^{-21}\,N(\mathrm{H_2})/B$ , with $N$ in $\mathrm{cm^{-2}}$ and $B$ in $\mu$ G.

Step 2. Substitute: $\mu_\Phi = 7.6\times10^{-21}(10^{22})/30 = 76/30 = \mathbf{2.5}$ .

Step 3. Since $\mu_\Phi>1$ the core is supercritical by a factor of about $2.5$ : its self-gravity exceeds what the field can resist. This is typical of observed cores, which cluster around $\mu_\Phi\sim2$ – $3$ — evidence they have already crossed the critical threshold and are set to collapse.

3 · Thermal support cannot hold up a GMC (fully worked)

Compare the sound speed of $T=10\,\mathrm{K}$ molecular gas ( $\mu=2.3$ ) to the turbulent dispersion of Example 1.

Step 1. The isothermal sound speed is $c_s=\sqrt{kT/\mu m_H}=\sqrt{(1.381\times10^{-16})(10)/(2.3\times1.673\times10^{-24})}=\mathbf{0.19\,km\,s^{-1}}$ .

Step 2. The turbulent dispersion is $\sigma=3\,\mathrm{km\,s^{-1}}$ , so $\sigma/c_s\approx16$ and the kinetic-energy ratio is $(\sigma/c_s)^2\approx\mathbf{250}$ . Turbulence supplies roughly 250 times the kinetic energy of thermal motions, so in a GMC the thermal term is negligible and the virial mass uses $\sigma$ , not $c_s$ . In a cold, quiescent prestellar core, by contrast, turbulence decays to near-thermal levels and the classical Jeans/Bonnor–Ebert analysis takes over.

Physics of the portrait

The virial theorem reaches molecular clouds along a two-century relay: a thermodynamicist coined the "virial" for gases, and astrophysicists later added turbulence and magnetism to explain star birth.

Origins · 1 minRudolf Clausius and how this was discovered

Rudolf Clausius

1822–1888, German physicist · thermodynamics and the mechanical virial of a system of particles

Rudolf Clausius is remembered above all for the second law of thermodynamics and for coining the word entropy, but the virial theorem is equally his. In 1870, seeking to connect the average kinetic energy of a gas to the forces within it, he introduced what he called the virial — from the Latin vis, viris, "force" — defined as $-\tfrac12\sum_i \mathbf{F}_i\cdot\mathbf{r}_i$ , and proved that for a steady bounded system the time-averaged kinetic energy equals the virial. For particles bound by an inverse-square force this is precisely $2\langle T\rangle + \langle W\rangle = 0$ . He meant it as a tool of the kinetic theory of gases and could not have foreseen that the same accounting would one day weigh clouds of interstellar gas. Poincaré and Eddington carried the theorem into astrophysics, applying it to the stability of stars and clusters. The decisive step came in 1953, when Subrahmanyan Chandrasekhar and Enrico Fermi added the magnetic energy term and used the theorem to estimate interstellar field strengths from cloud sizes and linewidths — the ancestor of every mass-to-flux measurement made today. The modern form, with its supersonic turbulent term and critical mass-to-flux ratio, completes a line of reasoning Clausius began with nothing grander in mind than the pressure of a gas.

History

1870

— Clausius introduces the virial theorem and coins the term virial

1953

— Chandrasekhar and Fermi add the magnetic term and apply the virial theorem to interstellar clouds

1956

— Bonnor and Ebert find the maximum mass of a pressure-bounded isothermal sphere

1976

— Mouschovias and Spitzer derive the critical mass-to-flux ratio for magnetic support

2007

— McKee and Ostriker synthesize the turbulent, magnetized theory of star formation

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 12 (interstellar medium and star formation) · Binney & Tremaine, Galactic Dynamics, 2nd ed., §4.8 (the virial theorem) · Stahler & Palla, The Formation of Stars, Ch. 3 and Ch. 9 · McKee & Ostriker, Annu. Rev. Astron. Astrophys. 45, 565 (2007), "Theory of Star Formation" · Mouschovias & Spitzer, ApJ 210, 326 (1976), critical mass-to-flux ratio. Key numerical anchors used throughout this article: the uniform-sphere gravitational energy $W=-\tfrac{3}{5}GM^2/R$ ; the virial-mass estimator $M_{\rm vir}=5\sigma^2 R/G$ , which gives $2.1\times10^5\,M_\odot$ for $\sigma=3\,\mathrm{km\,s^{-1}}$ and $R=20\,\mathrm{pc}$ ; the critical mass-to-flux coefficient $c_\Phi\approx0.13$ – $0.17$ ; the normalized mass-to-flux ratio $\mu_\Phi=7.6\times10^{-21}N(\mathrm{H_2})/B$ , giving $\mu_\Phi=2.5$ for $N=10^{22}\,\mathrm{cm^{-2}}$ and $B=30\,\mu\mathrm{G}$ ; the isothermal sound speed $c_s=0.19\,\mathrm{km\,s^{-1}}$ at $T=10\,\mathrm{K}$ with $\mu=2.3$ ; and the Bonnor–Ebert coefficient $1.18$ .

Prerequisites. Molecular Clouds and Giant Molecular Clouds · The Jeans Instability and Jeans Mass · Turbulence in the Interstellar Medium · Interstellar Polarization and Magnetic Fields. Leads to. CO as a Molecular Gas Tracer and the X-Factor · Phases of the Interstellar Medium.

Ask the tutor: Give me a cloud's linewidth, size, mass, and field strength, and I will walk you through its virial parameter and mass-to-flux ratio to decide whether it collapses, holds, or disperses.

Is the virial theorem the same as saying kinetic equals potential energy?

Only in the simplest limit. The bare gravitational virial theorem is $2T+W=0$ , so $T=\tfrac12|W|$ — kinetic energy is half the magnitude of the (negative) potential energy, not equal to it. The full molecular-cloud version adds the thermal $2U$ , magnetic $\mathcal{M}$ , and surface-pressure $T_s$ terms on top of $2T$ , and the clean $2T=|W|$ balance survives only when those three are negligible — the turbulence-dominated GMC limit. In a magnetized or pressure-confined cloud, $2T=|W|$ alone gives the wrong mass and the wrong verdict on stability.

Why is the virial mass written with a 5, and how reliable is that number?

The 5 combines two exact uniform-sphere pieces: the gravitational coefficient $3/5$ and the isotropy factor $\langle v^2\rangle=3\sigma^2$ . Setting $3M\sigma^2=\tfrac{3}{5}GM^2/R$ leaves $M_{\rm vir}=5\sigma^2R/G$ . Real clouds are centrally concentrated and often filamentary, so the true coefficient depends on the density profile — for $\rho\propto r^{-2}$ the factor drops from 5 to about 3. Practitioners absorb this into a geometry parameter, $M_{\rm vir}=5a\,\sigma^2R/G$ with $a\sim1$ . The virial mass is therefore good to a factor of order unity, usually enough to decide whether a cloud is bound.

What is the difference between the virial parameter and the mass-to-flux ratio?

They answer two different questions about the same cloud. The virial parameter $\alpha_{\rm vir}=5\sigma^2R/GM$ asks whether kinetic energy — mostly turbulence — can resist gravity; $\alpha_{\rm vir}\lesssim1$ means it cannot. The mass-to-flux ratio $M/\Phi_B$ asks whether the magnetic field can; above $c_\Phi/\sqrt{G}$ it cannot. A cloud can be bound by one measure and supported by the other, which is why a full assessment needs both. Crucially, $\alpha_{\rm vir}$ as usually defined omits the magnetic term, so a cloud with $\alpha_{\rm vir}<1$ might still be held up by a strong subcritical field.

How can supersonic turbulence support a cloud if it dissipates so quickly?

This is a genuine puzzle the virial theorem exposes rather than solves. Supersonic turbulence generates shocks, which radiate energy away in roughly a crossing time — comparable to the free-fall time — so the turbulent energy in $2T$ should decay and leave the cloud to collapse. One resolution is that the turbulence is continuously driven: by galactic-scale flows, by expanding H II regions and supernova blast waves, and by outflows from the very stars the cloud is forming. An alternative view holds that clouds are not in equilibrium at all but caught in global, hierarchical collapse, with $\alpha_{\rm vir}\approx1$ a transient coincidence rather than a stable balance.

Does external pressure make a cloud more or less stable?

It cuts both ways, and the surface term $T_s$ is where this lives. External pressure pushes inward on the boundary, so a small core in a warm medium is held together partly by the weight of gas pressing on it — bound at a smaller mass than an isolated Jeans analysis requires. But that confinement has a limit: pile mass onto a pressure-bounded isothermal sphere and it becomes unstable above the Bonnor–Ebert mass $M_{\rm BE}\approx1.18\,c_s^4/(G^{3/2}P_{\rm ext}^{1/2})$ , beyond which no hydrostatic solution exists. So external pressure stabilizes light cores and destabilizes heavy ones, with $M_{\rm BE}$ the turnover.

Why do observers trust the virial mass if it assumes equilibrium?

They trust it as a dynamical mass, not a proof of equilibrium. What $M_{\rm vir}=5\sigma^2R/G$ estimates is the mass required for the observed motions to be gravitationally bound — a statement about energetics, not a claim that $\ddot I=0$ . Its value is that it is independent of the CO-to-mass conversion (the X-factor): comparing $M_{\rm vir}$ with the CO-luminosity mass tests that conversion and flags unbound clouds. When the two agree across a large sample, as they roughly do for Galactic GMCs, that agreement is itself the evidence that the clouds are near virial balance.