Astronomy · Our view of the Milky Way

Disk, bulge and halo as distinct components

Describe Galactic components as overlapping spatial, chemical and dynamical distributions, and calculate how a broad tracer distribution can become relatively prominent away from a plane.

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 useful distinctions among the disk, central bar/bulge and stellar halo.
Distinguish a component distribution from a rigid boundary or a unique individual-star assignment.
Compare absolute density with a component’s fraction in an explicitly defined mixture.

Different structures occupy the same Galaxy

The Milky Way’s disk is a flattened distribution of stars, gas and dust. Its appearance depends on the tracer and wavelength: light from short-lived bright stars emphasizes different regions from a sample selected by another stellar property. A distribution can be flattened without every member lying exactly in one plane.

The central bar is an elongated stellar structure. Its inner three-dimensional box or peanut-shaped part is associated with the central bulge. These descriptions should not be interpreted as unrelated solid objects, one inserted into the other. The central stellar distribution is reconstructed through obscuration and projection, with the adopted component model made explicit.

The stellar halo is a diffuse, more extended stellar distribution containing orbital and spatial substructure. It need not be a perfectly smooth sphere, and its stars do not all share one orbit or chemical composition. Streams and other structure carry information that a single smooth density law can omit.

A dark halo has a different meaning: it concerns an inferred gravitating component in a dynamical model. The stellar halo is a population of stars that can be selected and observed. The two phrases must not be exchanged simply because both use the word halo.

Geometry, motion and chemistry give related selections

Disk samples can be separated by vertical structure, velocity distributions or chemical patterns. These criteria are related, but the sets they select need not coincide exactly. Calling a sample geometrically thick does not prove that every chemically selected high-abundance-ratio star belongs to that same sample.

The structural review by Bland-Hawthorn and Gerhard discusses this dependence on tracer and definition. A model can use several overlapping components to describe the data without establishing one unique formation route for each component. A present structure and its historical explanation are different levels of inference.

An individual star with uncertain distance, motion or chemistry may have uncertain component membership. A population trend is not an exact age, composition or orbit for every member. The earlier cluster-membership lesson’s distinction between a group definition and an individual probability is useful here, even though a Galactic component is not a star cluster.

This also limits simple colour labels. A measured colour depends on intrinsic spectrum and attenuation, and a bright evolutionary stage is not the whole population. Claims about a component’s age distribution require stellar modelling and a selection description beyond the visible colour of its brightest members.

A narrow and a broad distribution can overlap everywhere

For an original calculation, choose one fixed cylindrical radius and let z be height relative to a specified plane. Define q = |z|/h with an arbitrary positive height unit h, and let n0 be a positive reference tracer number density. Two illustrative distributions are n_A/n0 = exp(−q) and n_B/n0 = 0.1 exp(−q/3).

These are abstract A and B populations. They are not fitted Galactic scale heights, mass densities or surface brightnesses. No radial profile, force law or self-consistent equilibrium is supplied. The use of |z| makes the example symmetric about its plane, including the exponential cusp at z = 0.

At the plane, A has ten times B’s local tracer density. Farther from the plane, both decline, but B declines more slowly. At q = 3, the densities are approximately 0.049787068 n0 and 0.036787944 n0. Neither has vanished, even though their local importance has changed.

The sum n_A + n_B describes this particular complete tracer mixture. Its normalization is a number-density convention, so it cannot be relabelled as total gravitating mass without a relation between tracer counts and mass.

Original vertical tracer profiles at one fixed cylindrical radius. q = |z|/h; n0 and h are arbitrary positive units. Rounded densities describe this illustrative mixture, not measured Galactic components.
q	n_A/n0	n_B/n0	(n_A + n_B)/n0	Fraction B
0	1.000000000	0.100000000	1.100000000	0.090909091
1	0.367879441	0.071653131	0.439532572	0.163021208
2	0.135335283	0.051341712	0.186676995	0.275029668
3	0.049787068	0.036787944	0.086575012	0.424925658
6	0.002478752	0.013533528	0.016012281	0.845196805

Relative prominence and absolute abundance differ

B’s fraction in the mixture is f_B = n_B/(n_A + n_B). It rises from about 0.0909 at q = 0 to about 0.8452 at q = 6, while B’s own absolute density falls from 0.1 n0 to about 0.01353 n0. Becoming dominant in a thinning mixture does not mean that the dominant component has become denser.

To locate equal densities, set exp(−q) = 0.1 exp(−q/3). Rearranging gives exp(−2q/3) = 0.1, so q = (3/2) ln(10) ≈ 3.453877639. At that height each distribution contributes half the local mixture.

The crossover is not a physical wall. Both functions remain positive at every finite q, and their relative contributions change continuously. A threshold chosen for a component label may be useful for a task, but the label does not create a discontinuity in the underlying distributions.

Nor does this local fraction determine a global fraction. To count a full spatial population one would need the radial distribution, spatial limits and the appropriate volume integration, along with observational selection. This example supplies none of those additional inputs.

Overlapping profiles, changing fractions

Scroll sideways to see the full diagram. Its explanation is below.

Original synthetic vertical tracer profiles at one fixed radius, not measured Milky Way data or a gravitational-equilibrium fit. Define q = absolute z divided by h, where h is an arbitrary positive height unit, and n0 is a positive reference tracer number density. Population A has nA/n0 = exp(minus q); population B has nB/n0 = 0.1 exp(minus q/3). The upper panel plots A as a solid blue curve with circle markers, B as a dashed orange curve with square markers, and their sum as a dotted violet curve with diamond markers. Its linear density axis runs from 0 to 1.1; its zero is explicit. The lower panel plots the B number fraction nB/(nA+nB) as a green curve with square markers, on a linear axis from 0 to 1. The panels share the horizontal q axis from 0 to 6. Markers at q = 0, 1, 2, 3 and 6 match the complete fallback table. A vertical dashed reference marks q = 1.5 ln(10) = 3.453877639491, where the two densities are equal and the B fraction is 0.5. At q = 0 the densities are 1 and 0.1, total 1.1, and the B fraction is 1/11. At q = 6 both densities are still positive, about 0.002478752177 and 0.013533528324, total 0.016012280500, with B fraction about 0.845196805251. Neither density has a hard edge, and the crossing is not a physical boundary. Fractions describe only this complete declared tracer mixture; they do not assign an individual star to a component or measure Galactic mass fractions, ages, compositions or formation histories. Colours and line styles identify mathematical series, not real star colours. No fit, time track, morphology map or observational precision is claimed.

A structural description starts a historical question

Comparing positions, motions and abundance patterns can reveal groups that a purely geometric decomposition misses. Conversely, a chemical grouping can extend across several spatial regions. Neither result is a contradiction if the selections concern different measured properties.

A component’s formation history can involve more than one process and more than one epoch. Present location is not automatically birthplace: orbital changes and transport can move stars. The history must be tested against several kinds of evidence rather than read directly from a disk or halo label.

The practical discipline is to state the tracer, coordinate convention and selection rule whenever quoting a component property. Keep measured distributions separate from the model used to divide them, and keep that division separate from its proposed historical cause.

The following lessons add dust, motion and chemistry to this structural picture. Each provides new constraints, while retaining overlap and uncertainty instead of forcing every object into an exact, universal category.

Worked examples

Find equal contributions without inventing a boundary

Where do the two supplied vertical tracer densities become equal?

Set exp(−q) = 0.1 exp(−q/3).
Divide by exp(−q/3) to obtain exp(−2q/3) = 0.1.
Take natural logarithms: q = (3/2) ln(10) ≈ 3.453877639.

The densities are equal at |z| ≈ 3.453877639 h. Both remain nonzero on either side, so this is a crossover rather than an edge.

Read a fractional increase correctly

Does B’s larger fraction at q = 6 mean that its absolute density increased away from the plane?

At q = 0, n_B = 0.1 n0 and f_B ≈ 0.0909.
At q = 6, n_B ≈ 0.013533528 n0 and f_B ≈ 0.8452.
A has declined faster, so B’s share rises even as its own density falls.

No. Relative prominence increased in a mixture whose total density declined.

Try the reasoning

Are the stellar halo and an inferred dark halo the same component by definition?
Show the answer and reasoning
No.
One denotes a stellar population; the other concerns an inferred gravitating component.
Does a disk or halo label uniquely determine an individual star’s age?
Show the answer and reasoning
No.
Population properties overlap and an individual age needs further evidence and modelling.
Must geometrically and chemically defined disk samples contain exactly the same stars?
Show the answer and reasoning
No.
They select by different properties, with related but overlapping distributions.
Does f_B at one height determine B’s fraction throughout the Galaxy?
Show the answer and reasoning
No.
A global inference also needs radial structure, spatial integration, tracer-to-mass information where relevant and selection.

What to carry forward

Disk, central bar/bulge and stellar halo describe useful overlapping distributions.
A rising local fraction can accompany a falling absolute density.
A component definition, an individual assignment and a formation history require distinct evidence.

References

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