Reference taxonomy: move astronomy/ephemeris sources out of "Materials" into a numbered block

[Wiljea]

@c3d:
BLUF: Following up on my email re: the galaxy constants and the reference
numbering. Plan: merge #1700/#1701/#1702 as-is, then I'll submit one cleanup PR
that moves astronomy/ephemeris sources out of "Materials" into numbered
References (decision below). Flagging here for visibility — no action needed
from you beyond merging the three PRs when ready.

Why

The Materials reference list (constants.md) is accumulating sources that have
nothing to do with materials science:

• Materials 20 — Espenak, AstroPixels almanacs (perihelion/perigee dates)
• Materials 21 — Park et al. 2021, JPL DE440/441 ephemerides
• Materials 22 — Meeus, Astronomical Algorithms 2nd ed., ch. 50
• Materials 23 — NASA Planetary Fact Sheet (added in Pr a orbital #1700)

Two issues this causes:

1. Materials 21 is a byte-for-byte duplicate of Reference 4 (same
   DE440/441 paper); the giant-planet T₀ constants are tagged [4] [M21],
   citing one source twice.
2. Every new astronomy PR extends Materials with non-materials sources, so
   the namespace drifts from its meaning. (Pr c completeness #1702 did it the clean way, with a
   named WGS-84 reference alongside CODATA, Planck 2020, the PDG entries.)

Decision

Astronomy/ephemeris sources get their own numbered block:

• Reference 27 = Espenak (was M20)
• Reference 28 = Meeus, Astronomical Algorithms (was M22)
• Reference 29 = NASA Planetary Fact Sheet (was M23)
• drop Materials 21; retag its constants to the existing [4] (DE440/441)

Materials then reverts to genuine materials (1–19, contiguous); no existing
Reference 4–26 or Materials 1–19 is renumbered; WGS-84 stays as-is.

Sequencing

The three PRs branch off an older baseline and only touch citation tags, not
physics. Best merged first; I'll then do one cleanup PR on the merged tree,
allocating all new 27+ numbers in a single pass — both this reclassification and
the galaxy-constants fix (issue #1707):

• Astronomy: Reference 27 Espenak · 28 Meeus · 29 NASA Fact Sheet;
  drop Materials 21, retag to [4]
• Galaxies (issue Galaxy constants: Milky Way and Andromeda have identical masses and mis-attributed references #1707 (issue B galaxies)):
  Reference 30 Li & Riess 2021 · 31 Ibata 2005 · 32 Sick 2015 ·
  33 Licquia & Newman 2015 · 34 Bland-Hawthorn & Gerhard 2016 ·
  35 Nieten 2006; reassign the wrong [11]/[12] (LMC + M31-satellite-plane
  papers) to Watkins 2010 / Zhang 2024

All of 27–35 allocated once, collision-free. Editing the three PRs individually
would create merge conflicts on the reference list, which is why it's a single
follow-up pass.

(Minor nits folded into the same cleanup PR: G0F/θw tagged PDG but citing
CODATA, orphan Materials 6, over-cited Materials 5.)