Hertzsprung–Russell Diagram

The Hertzsprung–Russell diagram (abbreviated H–R diagram or HRD, and termed a Color–Magnitude Diagram [CMD] when plotted purely in observational photometric filters) is the foundational two-dimensional diagnostic chart of stellar astrophysics. It plots stars according to their intrinsic radiative power—expressed as bolometric luminosity (L, in units of solar luminosity L_☉ = 3.828 × 10²⁶ W) or absolute visual magnitude (M_V)—on the vertical axis against their surface thermal state—expressed as effective temperature (T_eff, plotted by universal convention decreasing from hot [~40,000–50,000 K] on the left to cool [~2,400 K] on the right), Harvard spectral class (O B A F G K M), or photometric color index (B − V or G_BP − G_RP)—on the horizontal axis.

Rather than scattering randomly across the luminosity–temperature plane, stars cluster into discrete, tightly defined loci corresponding to distinct hydrostatic nuclear-burning stages and stellar radii: the diagonal Main Sequence (comprising ~90% of extant stars, including the Sun), the upper-right Red Giant and Asymptotic Giant Branches, the top Supergiant domain, and the lower-left degenerate White Dwarf cooling sequence.

The diagram was constructed independently between 1905 and 1913 by the Danish chemical engineer and astronomer Ejnar Hertzsprung (who published tabular and graphical color–magnitude relationships for the Pleiades and Hyades open clusters in photography journals in 1905–1911 and discovered that red K and M stars split into luminous "giant" and dim "dwarf" populations) and the American astrophysicist Henry Norris Russell at Princeton University (who presented his landmark scatter plot of absolute magnitude versus Harvard spectral type at the Royal Astronomical Society in June 1913 and the American Astronomical Society in December 1913). Their work synthesized trigonometric parallax distances with the Harvard spectral classification sequence (O B A F G K M) developed by Annie Jump Cannon, Williamina Fleming, and Antonia Maury (whose narrow "c-characteristic" spectral line widths had first hinted at luminous supergiants), which Cecilia Payne-Gaposchkin (1925) and Meghnad Saha (1920) subsequently proved to be an ionization sequence governed purely by photospheric temperature T_eff.

---

Constant-Radius Diagonal Loci

The physical power of the H–R diagram stems from the Stefan–Boltzmann law of blackbody radiation. A spherical star of photospheric radius R and surface effective temperature T_eff radiates a total bolometric luminosity L given by its surface area 4*πR*² multiplied by the radiative flux F = σ_SBT_eff⁴:

L = 4 π R² σ_SB T_eff⁴

where σ_SB = 5.670374 × 10⁻⁸ W m⁻² K⁻⁴ is the Stefan–Boltzmann constant. Dividing by the solar values (L_☉ = 3.828 × 10²⁶ W, R_☉ = 6.957 × 10⁸ m, T_eff,☉ = 5,772 K) and taking base-10 logarithms yields a linear equation in the (log₁₀ T_eff, log₁₀ L) plane:

log₁₀( L / L_☉ ) = 2 log₁₀( R / R_☉ ) + 4 log₁₀( T_eff / T_eff,☉ )

Consequently, on a logarithmic H–R diagram where log₁₀(L/L_☉) is plotted vertically and log₁₀ T_eff is plotted horizontally (decreasing to the right), loci of constant stellar radius (R = constant) form straight parallel diagonal lines sloping downward from upper left to lower right with slope −4:

• Stars near the upper-right corner combine high luminosity (L ~ 10²–10⁵ L_☉) with low surface temperature (T_eff ~ 3,000–4,500 K), requiring enormous radii (R ~ 10¹–10³ R_☉: red giants and red supergiants).
• Stars near the lower-left corner combine low luminosity (L ~ 10⁻⁴–10⁻² L_☉) with high surface temperature (T_eff ~ 10,000–100,000 K), requiring compact planetary radii (R ≈ 0.01 R_☉ ≈ R_⊕: white dwarfs).

Spectral Classes and Yerkes (Morgan–Keenan) Luminosity Classes

The horizontal axis maps directly onto the Harvard spectral sequence (subdivided into decimal subtypes 0–9 from hottest to coolest, e.g., B0 to B9, G0 to G9, extended to cooler ultracool dwarfs L, T, and Y):

• O (30,000–50,000 K, blue): Ionized helium (He II λ4541, λ4686 Å) and multiply ionized C, N, O lines.
• B (10,000–30,000 K, blue-white): Neutral helium (He I λ4471, λ5876 Å) peaking at B2; e.g., Rigel (B8 Ia), Spica (B1 III–IV).
• A (7,500–10,000 K, white): Maximum strength of hydrogen Balmer absorption lines (Hα, Hβ, Hγ, peaking at A0 where T_eff ≈ 9,600 K); e.g., Vega (A0 V), Sirius A (A1 V).
• F (6,000–7,500 K, yellow-white): Weakening Balmer lines; strengthening singly ionized calcium (Ca II H & K at λ3934, λ3968 Å) and Fe II lines; e.g., Procyon A (F5 IV–V).
• G (5,200–6,000 K, yellow): Maximum strength of Ca II H & K lines alongside prominent neutral metal lines (Fe I, Na I D) and CH G-band; e.g., the Sun (G2 V, T_eff = 5,772 K, M_V = +4.83, B − V = +0.65), Alpha Centauri A (G2 V).
• K (3,700–5,200 K, orange): Dominated by neutral metallic lines (Fe I, Ca I λ4227 Å) and molecular CN/CH bands; e.g., Arcturus (K1.5 III), Epsilon Eridani (K2 V).
• M (2,400–3,700 K, red): Deep molecular absorption bands of titanium oxide (TiO) and vanadium oxide (VO); e.g., Proxima Centauri (M5.5 V), Betelgeuse (M1–M2 Ia–Iab).

In 1943, William Wilson Morgan, Philip C. Keenan, and Edith Kellman introduced the two-dimensional Yerkes (MK) spectral classification, appending a Roman numeral Luminosity Class based on pressure-sensitive spectral line widths (in giant atmospheres with low surface gravity g = GM/R², lower gas density reduces Stark pressure broadening, yielding narrower absorption lines than in high-gravity dwarfs of the same T_eff):

• Class 0 (or Ia⁺): Hypergiants (e.g., Cygnus OB2-12, VY Canis Majoris).
• Class Ia / Iab / Ib: Luminous, intermediate, and less-luminous Supergiants.
• Class II: Bright Giants.
• Class III: Normal Giants (Red Giant Branch, Horizontal Branch / Red Clump).
• Class IV: Subgiants (transitioning between the Main Sequence and Giant Branch).
• Class V: Main-Sequence Dwarfs (core hydrogen-burning stars).
• Class VI (sd): Metal-poor Subdwarfs lying 1–2 magnitudes below the solar-metallicity Main Sequence.
• Class VII (or D): White Dwarfs (DA, DB, DC, DO, DZ, DQ).

---

1. The Main Sequence (Luminosity Class V)

Stretching diagonally from hot, luminous O-type stars in the upper left (M ≈ 60 M_☉, L ≈ 10⁶ L_☉, T_eff ≈ 45,000 K) to cool, dim M-type red dwarfs in the lower right (M ≈ 0.08 M_☉ [the hydrogen-burning limit], L ≈ 10⁻⁴ L_☉, T_eff ≈ 2,500 K), the Main Sequence is the locus of stars in hydrostatic equilibrium fusing hydrogen into helium in their cores (stellar nucleosynthesis):

• Lower Main Sequence (M ≲ 1.3 M_☉): Core hydrogen fusion proceeds primarily via the proton–proton (p–p) chain (ε_pp ∝ ρT⁴). These stars possess radiative cores and convective outer envelopes.
• Upper Main Sequence (M ≳ 1.3 M_☉): Core temperatures exceed T_c ≈ 1.7 × 10⁷ K, where the CNO (carbon–nitrogen–oxygen) bi-cycle (ε_CNO ∝ ρT¹⁶) dominates energy generation. The extreme temperature sensitivity drives a convective core surrounded by a radiative envelope.

Along the Main Sequence, stellar position is determined primarily by initial birth mass M (Vogt–Russell theorem) via Arthur Eddington's (1924) mass–luminosity relation and the empirical mass–radius relation (R ∝ M^0.7–0.8):

L / L_☉ ≈ ( M / M_☉ )^α (where α ≈ 3.5 for 0.5 ≤ M/M_☉ ≤ 20; α ≈ 2.3 for M < 0.43 M_☉; α → 1 near the Eddington limit for M > 50 M_☉)

Because a star's available nuclear fuel is proportional to its mass (E_fuel ∝ M) while its fuel consumption rate equals its luminosity (L ∝ M^3.5), a star's main-sequence lifetime (τ_MS) scales inversely with a steep power of mass:

τ_MS ∝ M / L ∝ M^−2.5 ⇒ τ_MS ≈ 10¹⁰ years × ( M / M_☉ )^−2.5

Thus, a 25 M_☉ O-star exhausts its core hydrogen in just ~3–6 million years, a 1 M_☉ G2V star (the Sun) lives ~10 billion years, and a 0.1 M_☉ M-dwarf burns hydrogen convectively for >1–10 trillion years (exceeding the current 13.8 Gyr age of the universe).

2. Post-Main-Sequence Evolution: Subgiants, Red Giants, and the Horizontal Branch

When a solar- or intermediate-mass star (M ≲ 8 M_☉) exhausts hydrogen at its center, it leaves an inert helium core surrounded by a hydrogen-burning shell. Once the isothermal helium core exceeds the Schönberg–Chandrasekhar limit (~8–10% of the star's total mass), the core contracts rapidly, releasing gravitational energy that expands the outer envelope ("mirror principle"):

• Subgiant Branch (Class IV): The star moves horizontally to the right across the H–R diagram at nearly constant luminosity as its radius expands and T_eff drops from ~6,000 K to ~4,800 K.
• Red Giant Branch (RGB, Class III): As the outer envelope becomes deeply convective, opacity rises (dominated by H⁻ ions) and the star ascends nearly vertically along its Hayashi track, expanding to R ~ 50–200 R_☉ and L ~ 10²–2.5 × 10³ L_☉ at T_eff ≈ 3,500–4,500 K.
• Helium Flash and Horizontal Branch / Red Clump: For stars with M ≲ 2.0 M_☉, the contracting helium core becomes electron-degenerate until T_c reaches ~10⁸ K, igniting helium fusion into carbon via the triple-alpha process (3 ⁴He → ¹²C) in a thermal runaway (helium flash). Once degeneracy is lifted, the star settles onto the Horizontal Branch (in metal-poor globular clusters such as Messier 13) or the Red Clump (in metal-rich populations) at L ≈ 50 L_☉, stably burning core helium and shell hydrogen.
• Asymptotic Giant Branch (AGB) and Planetary Nebula: After exhausting core helium, the star forms a degenerate C–O core surrounded by concentric He- and H-burning shells, ascending the AGB a second time (R ~ 200–500 R_☉, L ~ 10⁴ L_☉). Thermal pulses and radiation-driven dust winds eject the envelope as an ionized planetary nebula (such as the Ring Nebula, M57), while the exposed hot core (T_eff > 100,000 K) crosses the top of the H–R diagram to the left.

3. White Dwarf Cooling Sequence (Luminosity Class VII / D)

So long as the remnant core mass lies below the Chandrasekhar limit (M < 1.44 M_☉, typically 0.55–1.0 M_☉), electron degeneracy pressure fixes its radius at a constant value (R ≈ 0.008–0.015 R_☉ ≈ 5,500–10,500 km). Deprived of nuclear fusion, the white dwarf slides diagonally downward and to the right along a constant-radius line (L ∝ T_eff⁴) in the lower-left quadrant of the H–R diagram (Mestel cooling law), fading from T_eff ≈ 100,000 K (L ~ 10 L_☉) to T_eff ≈ 4,000 K (L ~ 10^−4.5 L_☉) over 10 billion years.

4. The Instability Strip and Cluster Turnoff Isochrone Dating

Crossing the H–R diagram between the Main Sequence (at spectral classes A5–F2) and the Supergiant region (F5–G8 Ia/Ib) is the nearly vertical Cepheid Instability Strip. Stars traversing this temperature band (T_eff ≈ 5,500–8,000 K) experience radial acoustic pulsations driven by the Eddington κ-valve mechanism in the subsurface doubly ionized helium (He II → He III) partial ionization zone (including Classical Cepheids [P = 1–100 d], Type II Cepheids, RR Lyrae horizontal-branch stars [P = 0.2–1.0 d, M_V ≈ +0.6], and δ Scuti variables), obeying Henrietta Swan Leavitt's (1912) Period–Luminosity relation.

Because all stars in a gravitationally bound star cluster formed simultaneously from the same molecular cloud at the same distance, plotting a cluster's H–R diagram reveals a sharp Main-Sequence Turnoff (MSTO) point—the exact luminosity and temperature at which stars whose main-sequence lifetime τ_MS(M) equals the cluster's current age (t_cluster) are peeling off the Main Sequence onto the Subgiant Branch:

• Young open clusters like the Pleiades (M45) retain luminous blue B-type stars up to a high turnoff at B6 (M_TO ≈ 4.5 M_☉), dating the cluster to ~100–125 Myr.
• Ancient globular clusters like Messier 13 (NGC 6205) have lost all O, B, A, and early F stars; their turnoff sits near F8–G0 (M_TO ≈ 0.80 M_☉) with a prominent Red Giant Branch and Horizontal Branch, dating them to ~11.6–13.0 Gyr.

---

Why is temperature plotted backward (decreasing from left to right) on the Hertzsprung–Russell diagram?

When Ejnar Hertzsprung and Henry Norris Russell constructed the first diagrams in 1905–1913, surface temperatures were not yet directly measured; instead, they plotted stars horizontally from left to right according to the alphabetical Harvard spectral sequence (O, B, A, F, G, K, M) or increasing photometric color index (B − V, from blue = −0.3 to red = +1.8). Because O and B stars are the hottest (~30,000–50,000 K) and M stars are the coolest (~2,400–3,700 K), temperature naturally decreases from left to right.

How can two stars with the exact same surface temperature and spectral type have luminosities that differ by a factor of a million?

By the Stefan–Boltzmann law (L = 4*πR*²σ_SBT_eff⁴), stars of identical effective temperature T_eff emit the exact same radiative power per square meter of photospheric surface. Therefore, if a red supergiant like Betelgeuse (M1–M2 Ia, L ≈ 10⁵ L_☉) is 10⁸ times more luminous than a red dwarf of the same ~3,500 K temperature (L ≈ 10⁻³ L_☉), its surface area 4*πR*² must be 10⁸ times larger, meaning its radius R is 10⁴ times larger (~700 R_☉ vs. 0.07 R_☉).

How do astronomers use the H–R diagram to measure the age of a star cluster?

Because all stars in a cluster form simultaneously with a wide range of masses, and massive stars burn through their core hydrogen much faster than low-mass stars (τ_MS ≈ 10¹⁰ yr × [M/M_☉]^−2.5), the upper Main Sequence steadily erodes from top left to bottom right over time. The exact luminosity and temperature of the main-sequence turnoff (MSTO) point—where stars are currently exhausting core hydrogen—directly measures the cluster's age.

---

1. Hertzsprung, E. (1911). "Über die Verwendung photographischer effektiver Wellenlängen zur Bestimmung von Farbenäquivalenten." Publikationen des Astrophysikalischen Observatoriums zu Potsdam, 22(63), 1–40. https://doi.org/10.1002/asna.19111891402
2. Russell, H. N. (1914). "Relations between the spectra and other characteristics of the stars." Popular Astronomy, 22, 275–294. https://doi.org/10.1038/093227a0
3. Morgan, W. W., Keenan, P. C., & Kellman, E. (1943). An Atlas of Stellar Spectra, with an Outline of Spectral Classification. Chicago: University of Chicago Press. https://doi.org/10.1086/144522
4. Gaia Collaboration, Babusiaux, C., et al. (2018). "Gaia Data Release 2: Observational Hertzsprung-Russell diagrams." Astronomy & Astrophysics, 616, A10. https://doi.org/10.1051/0004-6361/201832843