Astronomy · Active galaxies and central black holes

Orbital and dynamical evidence for central masses

Compare inclination-conditioned circular dynamics, a lag–line-dispersion virial product and local uncertainty propagation, keeping compact-mass evidence separate from a direct horizon observation.

Continue in the approved chapter order. Necessary vocabulary, quantities and assumptions are explained in the lesson.

Back to the chapter

After reading, you can

Identify the tracer and geometry assumptions in a central enclosed-mass estimate.
Calculate conditional masses and a reverberation virial product using declared constants and width conventions.
Distinguish propagated input uncertainty from model ambiguity and causal-boundary evidence.

Different tracers require different dynamical models

Gravity can be investigated through the motions of stars or gas near a galactic centre. The directly measured quantities include image positions, spectral shifts and profiles. A mass is then inferred through a model connecting those observations to spatial positions, velocities and the gravitational field.

A thin gas disk on nearly circular orbits is different from a stellar population with a distribution of orbital directions. Stellar velocity dispersion is a spread of velocities, not one circular speed. The balance of radial and tangential motions, called velocity anisotropy, can affect stellar-dynamical inference.

Bright molecular maser emission can trace small gas structures with precise spectral and positional measurements. Herrnstein and collaborators’ nuclear gas-disk study illustrates an explicit warped-disk model with distance and inclination dependencies. It is evidence for modelled compact dynamics, not a license to treat every gas cloud as a circular test particle.

A central compact mass strongly constrained by dynamics is one stage of evidence. Inferring a black hole also considers how compact that mass is and the alternatives. A mass measurement by itself is not an image of an event horizon or a general proof that an ordinary extended distribution has a horizon.

Write the projection before solving for mass

For a separate original exercise, stipulate a thin circular disk dominated by a compact central mass in the Newtonian weak-field regime. Let i be its inclination, with i = 0° face-on and i = 90° edge-on. Observe along the projected disk’s major axis and subtract the systemic velocity of the galaxy.

If v_LOS is the magnitude of that line-of-sight velocity offset, the circular speed in this model is v_c = v_LOS/sin i for 0° < i ≤ 90°. At true orbital radius r, circular balance gives M_enc = rv_c²/G = r(v_LOS/sin i)²/G.

Supply physical radius r = 0.2 pc and v_LOS = 600 km s⁻¹. For the following arithmetic adopt 1 pc = 3.085677581491367 × 10¹⁶ m and G = 6.67430 × 10⁻¹¹ m³ kg⁻¹ s⁻² as teaching inputs, without treating G as an exact defined constant.

The radius is already supplied physically; a real angular radius would require an adopted distance. Noncircular flows, pressure, warps, extended stellar matter and instrumental blurring would need explicit modelling. The face-on limit cannot yield this velocity deprojection from a nonzero v_LOS.

Geometry changes the inferred enclosed mass

At i = 90°, v_c = 600 km s⁻¹ and M_enc ≈ 3.328720403 × 10³⁷ kg. At i = 60°, v_c ≈ 692.820323 km s⁻¹ and M_enc ≈ 4.438293871 × 10³⁷ kg. At i = 30°, v_c = 1200 km s⁻¹ and M_enc ≈ 1.331488161 × 10³⁸ kg.

All three rows use the same supplied physical radius and observed velocity offset. Their masses differ because the inclination assumption changes the circular speed, which enters squared. Relative to the edge-on row, the mass factors are 1, 4/3 and 4.

The table does not select the true inclination. Morphology and independent kinematic information can constrain it, but a projected image may leave ambiguity. Smaller velocity measurement errors alone do not remove that geometric dependency.

Kormendy and Ho discuss how stellar and gas dynamical models encounter projection, orbital distributions and spatial-resolution limits. The simple circular formula is useful within its domain; substituting one stellar line width for v_c without the corresponding model would change the inference.

Original thin circular major-axis gas-disk toy. Supplied r = 0.2 pc and systemic-subtracted |v_LOS| = 600 km s⁻¹; i is 0° face-on. Adopted pc and G inputs are stated in the text. Each row is conditional, not a separate observation or a selected true mass.
Assumed i	v_c (km s⁻¹)	M_enc (kg)	Relative to edge-on mass
90°	600	3.328720403 × 10³⁷	1
60°	692.820323	4.438293871 × 10³⁷	4/3
30°	1200	1.331488161 × 10³⁸	4

Keep the response scale and mass factor explicit

Reverberation combines continuum and variable-line time series under a response model. For an independent synthetic example, supply observed lag τ_obs = 24 days and host redshift z = 0.2. Use the supplied rest-frame conversion τ_rest = τ_obs/(1 + z) = 20 days; no cosmological distance parameter is being inferred.

Let ΔV = 4000 km s⁻¹ be the specifically defined rest-frame line dispersion, a velocity standard-deviation measure of the line. It is not silently replaced by full width at half maximum. Use c = 299792458 m s⁻¹, one day = 86400 s and the adopted G above.

The response-weighted length is cτ_rest = 5.18041367424 × 10¹⁴ m. The virial product VP = cτ_rest ΔV²/G is approximately 1.241877332 × 10³⁸ kg. A mass model uses M = fVP, where f accounts for geometry and kinematics under the chosen lag and width conventions.

For stipulated alternatives f = 1 and f = 4, the conditional masses are approximately 1.241877332 × 10³⁸ and 4.967509329 × 10³⁸ kg. Those f values are not measured or universal calibrations. The lag is response weighted, and cτ_rest is not a uniquely imaged gas boundary.

Original response-model arithmetic using an observed 24-day lag, supplied z = 0.2 and rest-frame line dispersion 4000 km s⁻¹. The 20-day rest lag gives the stated virial product. f = 1 and 4 are stipulated alternatives with the same width convention.
Quantity	Conditional value
Rest-frame lag	20 days
cτ_rest	5.18041367424 × 10¹⁴ m
Virial product VP	1.241877332 × 10³⁸ kg
Mass if f = 1	1.241877332 × 10³⁸ kg
Mass if f = 4	4.967509329 × 10³⁸ kg

Precision within a model does not remove model limits

Return to M = rv_LOS²/(G sin²i). Holding the adopted G fixed in this example, local fractional sensitivities are 1 for radius, 2 for velocity and −2cot i for inclination measured in radians. Here cot i = cos i/sin i. The sign matters when covariance terms are present.

Stipulate uncorrelated input standard uncertainties u(r)/r = 0.05, u(v_LOS)/v_LOS = 0.025 and u(i) = 2° = 0.03490658504 rad at i = 60°. Assuming a locally adequate first-order approximation gives [u(M)/M]² ≈ 0.05² + 4(0.025)² + 4cot²60°(0.03490658504)².

The relative variance is approximately 0.00662462624, giving relative standard uncertainty approximately 0.081391807, or 8.1391807 percent. This is the local propagated result under the supplied uncertainty model, not a probability interval with guaranteed coverage.

If the inputs share calibration or modelling dependencies, retain the signed covariance contributions rather than assuming independence. Model discrepancy, unknown geometry and unmodelled gas physics are not erased by a small propagated random uncertainty. NIST’s first-order propagation framework requires that adequacy and correlation be assessed.

A central-mass claim should retain its inference chain

A useful report identifies the measured positions and spectra, the distance and inclination assumptions, the tracer model and the mass quantity being inferred. It states how extended matter, instrumental response and alternative dynamics were handled.

Independent methods can strengthen a conclusion when they probe different assumptions, but they can also share distances or calibration dependencies. Agreement is more informative when those connections are made explicit.

A classical event horizon is a causal boundary in spacetime. Compact dynamical mass evidence, a line-response mass estimate and a reconstructed emission image contribute different constraints. None should be relabelled as a direct observation of the same radius without the appropriate relativistic and emission model.

The next lesson examines obscuration, viewing angle and interferometric images. It will distinguish an emission-ring diameter, a black-hole shadow and an event horizon, retaining the calibration steps needed to connect an image to physical scales.

Worked examples

Change geometry without changing the measured velocity

Why is the 30° conditional mass four times the edge-on mass?

The formula gives M proportional to 1/sin²i when r and v_LOS are fixed.
sin 90° = 1 and sin 30° = 1/2.
The ratio is 1/(1/2)² = 4, while the circular speed doubles.

The deprojected circular speed doubles and enters the mass estimate squared under the supplied circular model.

Separate a measured-model product from a geometry factor

Do the supplied lag and line dispersion alone select between f = 1 and f = 4?

They give the same rest lag 20 days and VP ≈ 1.241877332 × 10³⁸ kg.
The two assumed f values multiply that same product by 1 or 4.
Choosing f requires a geometry and kinematics interpretation consistent with the width and response definitions.

No. The product is fixed by the supplied inputs, while the final conditional mass still depends on f.

Try the reasoning

Can a stellar velocity dispersion always replace circular speed directly in M = rv²/G?
Show the answer and reasoning
No.
A stellar orbital distribution requires its own projection and dynamical model.
Should the 2° angle uncertainty be inserted as the number 2 in a radians-based derivative?
Show the answer and reasoning
No.
It must first be converted to approximately 0.03490658504 rad.
Does the 8.14-percent local standard uncertainty eliminate inclination or gas-model ambiguity?
Show the answer and reasoning
No.
It describes the supplied local input uncertainties, not every possible model discrepancy.
Is a compact mass inferred through dynamics itself a direct measurement of a classical event horizon?
Show the answer and reasoning
No.
The horizon is a causal spacetime boundary; connecting different evidence requires further physical modelling.

What to carry forward

Central masses are inferred through tracer-specific geometry and dynamics.
Circular deprojection and reverberation factors can dominate the interpretation of precise data.
Local uncertainty propagation and causal-boundary evidence are distinct parts of the argument.

References

Teaching resources and research supporting this lesson’s explanations and examples.