Cosmology · Complementary probes of cosmic history

Supernovae as standardisable candles

Apply a supplied width–color correction and trace the common calibration that remains in absolute luminosity distance.

Preparation: ratios, powers of ten and ordinary arithmetic. Physical quantities and notation are introduced as they are needed.

Back to the chapter

After reading, you can

Explain why standardization differs from assuming identical raw events.
Calculate a corrected magnitude under explicitly supplied coefficients.
Distinguish relative-distance information from an absolute zero-point calibration.

A standardized candle needs a response model

A supernova observation supplies changing light intensity and spectral or color information. Turning those measurements into a distance estimate requires a relation between the event’s corrected brightness and a luminosity calibration.

Standardisable means that observable properties can help account for brightness differences. It does not mean that every event has the same uncorrected apparent or intrinsic brightness.

Comparisons must refer to compatible emitted bands or include the relevant spectral and redshift corrections. Raw detector-band flux is not automatically bolometric flux or a rest-frame magnitude in the desired band.

Use a supplied compatible rest-frame-band magnitude convention below, with those band corrections already made. No actual supernova sample, fitted correction coefficient or selection model is supplied.

Keep every chosen term visible

Choose an absolute reference magnitude M₀ = −19 and two true distance moduli μ = 35 and 40. These are toy calibrations and positive-distance groups, rather than observed luminosities or inferred cosmological parameters.

For a dimensionless width coordinate w_shape and a color residual C in magnitudes, stipulate m = M₀+μ−αw_shape+βC+b. Choose α = 1.5 magnitudes per width unit, β = 2 and common zero-point shift b = 0.05 magnitudes.

The width symbol is a light-curve coordinate, not the cosmological pressure-to-density ratio. The chosen color coefficient is not a universal dust law or an estimated empirical correlation.

Apply the supplied correction m_corr = m+αw_shape−βC. It removes exactly those two stipulated terms, leaving m_corr = M₀+μ+b.

Synthetic compatible-band events; identical distance groups retain the same supplied common zero point
True μ	w_shape	Color C (mag)	Reported m	Corrected m_corr	Inferred μ if b is ignored
35	−0.1	0.02	16.24	16.05	35.05
35	0.1	−0.01	15.88	16.05	35.05
40	−0.1	0.02	21.24	21.05	40.05
40	0.1	−0.01	20.88	21.05	40.05

Standardization does not determine every calibration

The corrected values within each distance group agree in the noise-free toy. Agreement does not remove the common b: subtracting the supplied M₀ yields μ+b for every event.

For the same compatible corrected magnitude convention, μ = 5log₁₀(D_L/10 pc). Ignoring b multiplies the inferred luminosity distance by 10^(b/5) = 10^0.01 ≈ 1.023293.

Averaging more identical events does not reduce this deterministic common shift. The model contains no random scatter from which a decreasing standard error could be inferred.

If the absolute magnitude calibration were unknown instead, it could also affect the overall distance normalization. Standardizing the relative light curves does not automatically identify that absolute scale.

This original four-event construction assumes compatible corrected rest-frame-band magnitudes, M₀ = −19, α = 1.5 magnitudes per unit dimensionless width w_shape, β = 2 and common b = 0.05 magnitudes. It supplies m = M₀+μ−αw_shape+βC+b; C is a color residual in magnitudes. Panel A uses two categorical positions, Reported and Corrected, rather than physical time. Its linear residual ordinate m−(M₀+μ), or m_corr−(M₀+μ), runs from −0.15 to 0.3 magnitudes. Solid blue circles for w_shape = −0.1, C = 0.02 go from 0.24 to 0.05. Dashed orange squares for w_shape = 0.1, C = −0.01 go from −0.12 to 0.05. Connecting segments indicate only the stipulated algebraic correction; they do not imply an evolution or an intermediate measurement. The corrected circle and larger hollow square intentionally share the exact 0.05 endpoint. Each pair occurs in both supplied true-modulus groups μ = 35 and 40, so the graph shows the two distinct residual pairs rather than concealing duplicate points. The native table gives all four actual rows (μ, w_shape, C, m, m_corr, μ_inf if b is ignored): (35, −0.1, 0.02, 16.24, 16.05, 35.05), (35, 0.1, −0.01, 15.88, 16.05, 35.05), (40, −0.1, 0.02, 21.24, 21.05, 40.05), (40, 0.1, −0.01, 20.88, 21.05, 40.05). The declared correction m_corr = m+αw_shape−βC removes exactly the two stipulated property terms and leaves M₀+μ+b. Ignoring b gives μ_inf = μ+b and multiplies every positive luminosity distance by 10^(b/5) = 10^0.01 ≈ 1.023293. The group difference remains 5 magnitudes, giving true distance ratio 10; shared M₀ and b cancel only in that difference. The supplied true moduli correspond to 10⁸ and 10⁹ parsecs, not measured cosmological distances. No random scatter is supplied, so more identical events do not reduce this deterministic common zero point. These chosen coefficients are not a fitted standardizer or universal dust law. No raw detector-band to bolometric substitution, observed sample, H₀ result or acceleration inference is made. Shape, dash, direct labels and the full table convey all distinctions without color. Original Ark teaching model.

A common shift can cancel in a difference

The two corrected groups differ by 21.05−16.05 = 5 magnitudes. The common b cancels from this difference, as does the shared M₀.

Their supplied true distance ratio is therefore 10^(5/5) = 10. The chosen moduli correspond to 10⁸ pc and 10⁹ pc before the common calibration shift is applied.

This cancellation does not make the two absolute distances separately calibrated. It identifies a relative-distance combination under the same standardization and zero-point assumptions.

A redshift-dependent calibration or population change would not necessarily cancel in the same way. It would need its own response model and evidence rather than being assumed equivalent to a constant b.

Use the candle together with its uncertainty model

A real analysis must account for measurement noise, selection, population variation, dust and calibration, as well as the compatible-band corrections. The four-event table fits none of these effects.

Shared calibration produces dependencies between events and analyses. Different papers or event labels do not establish independent absolute-distance information if they reuse the same calibration.

The supplied groups have no associated redshifts or expansion-history fit. No H₀ estimate, acceleration significance or claim that a calibration residual explains actual data follows from the toy.

The transferable reasoning is to standardize the specified observable, identify the residual calibration and ask which absolute or relative combinations the observations can constrain under the adopted model.

Worked reasoning

Correct the first event

What remains after the supplied width and color correction?

The first reported magnitude is −19+35−1.5(−0.1)+2(0.02)+0.05 = 16.24.
Add 1.5(−0.1) and subtract 2(0.02), giving m_corr = 16.05.
Subtract M₀ = −19 to obtain μ_inferred = 35.05, still 0.05 above the true modulus.

The stipulated width and color terms cancel, while the common zero-point shift remains.

Try the reasoning

By what factor does ignoring b = 0.05 magnitudes change the inferred distance?
Show the answer and reasoning
10^0.01 ≈ 1.023293.
A modulus offset b changes distance by 10^(b/5).
What is the ratio of the second group’s true distance to the first group’s?
Show the answer and reasoning
10.
The shared M₀ and b cancel, leaving a five-magnitude difference.
Will adding identical events remove the deterministic b in this toy?
Show the answer and reasoning
No.
The same shift affects every event; no independent zero-point information was added.

What to carry forward

Standardization requires a declared relation between light-curve properties and brightness.
A common zero point can remain after width and color corrections.
Relative distances and an absolute distance scale carry different calibration requirements.

References

Sources supporting the explanations, physical distinctions and examples in this lesson.