SN 1987A

Summary

• One supernova settles the distant-starlight question better than any argument about the universe at large.  SN 1987A exploded in the Large Magellanic Cloud.  Its distance is fixed twice over by geometry, with no cosmology and no standard candle in either measurement.  The two answers were obtained 28 years apart, by methods that share no physics, and they agree.

• The neutrinos arrived first, and that is exactly what core collapse predicts.  Three detectors on three continents recorded a burst on 23 February 1987 within seconds of each other:  Kamiokande-II in Japan, IMB in Ohio, and Baksan in the Caucasus.¹²³  The first optical evidence came hours later.  Albert Jones looked at that patch of sky on February 23.39 UT and saw nothing;  Robert McNaught's photograph at February 23.443 UT shows the star at about magnitude 6.5.⁴⁴  Neutrinos leave a collapsing core immediately.  Light cannot leave until the shock reaches the star's surface.

• The star that exploded had been photographed for ninety years.  Sanduleak −69 202 sat on Harvard patrol plates from 1896 to 1954 with no variation at the 0.3-magnitude level, and on CTIO plates from 1974 to 1983 as a normal B3 I supergiant.⁵  After February 1987 it was gone.

• The circumstellar ring measures its own physical size with a stopwatch.  The explosion flash swept outward and lit a ring of gas the star had shed about 20,000 years earlier.  The near edge of the ring reported first, at 86 ± 11 days;  the far edge last, peaking at 408 ± 54 days.⁶  Those two numbers give the ring's true diameter, (1.27 ± 0.07) × 10¹⁸ cm.  Divide by the angular diameter Hubble measured and the distance falls out:  51.2 ± 3.1 kpc, about 167,000 light years.⁶⁶  This is a timing measurement, not a model.

• Detached eclipsing binaries in the same galaxy give 49.59 ± 0.09 (statistical) ± 0.54 (systematic) kpc.⁷⁷  That is one per cent, from twenty binary systems, using surface-brightness relations and orbital mechanics.  It agrees with the ring echo well inside the ring echo's errors.  Neither result uses a Cepheid, a supernova standard candle, or a value of the Hubble constant.

• The late glow is radioactive decay, and the isotopes emit at their laboratory energies.  Cobalt-56 powered the light curve for the first few years.  Titanium-44 took over after that.  In 2012 INTEGRAL detected the two direct-escape lines of ⁴⁴Ti from the remnant at 67.9 and 78.4 keV, the same energies that isotope emits on a laboratory bench.⁸  A Chandra analysis published in 2025 found the ⁴⁴Sc daughter line at 4.09 keV in the same remnant.⁹⁹  Nuclear and atomic energy levels 167,000 light years away, and 167,000 years ago, are the ones measured on a laboratory bench.

• The follow-up record is now four decades long and still growing.  The ring's hotspots lit up one by one as the blast wave reached it.¹⁰¹⁰  Hubble's refurbished spectrograph mapped the reverse shock in 2010.¹¹¹¹  ALMA weighed the dust the explosion made.¹²¹²  In 2024 JWST found narrow argon and sulphur lines at the centre, the best evidence yet for the neutron star the neutrino burst implied.¹³¹³

• Created light in transit fails here, and young-earth astronomers say so themselves.  On that model the star never exploded, the neutrinos were manufactured in flight, and the ring echo times a bounce that never happened.  Danny Faulkner's verdict on created-light eclipses applies with more force here:  "This is deceptive."¹⁴  Jason Lisle printed the same conclusion under a photograph of this supernova.¹⁵

• The anisotropic synchrony convention survives the light echo but not the appearance of age.  Lisle works the ring geometry correctly:  the light that lit the ring travelled away from Earth first, so it is slow under ASC, and the observed delay comes out right.¹⁶  What ASC cannot remove is the object's own history.  The ring is 0.67 light years across and expanding at 10 kilometres per second, so it was launched about 20,000 years before the explosion.⁵  Under ASC that lands in about 18,000 BC, before Day Four.  The ring has to have been created already in flight.

• A decaying speed of light survives everything here, by construction, and its author says so.  Barry Setterfield's answer is that "the slowing of light exactly counteracts the faster decay rate," so the observed decay looks normal.¹⁷  Follow the arithmetic and his compensation leaves the distance to SN 1987A unchanged;  it shortens only the transit time.  He also grants that in his model the fine-structure constant never changes at all, so the line energies come out right automatically.  A model tuned so that no measurable quantity ever differs has made itself unfalsifiable, which is a defeat rather than a defence.

• What this one object does not prove, stated plainly.  It is a single supernova 167,000 light years away.  It shows that light took far longer than 6,000 years to get here and that nuclear physics runs at the same rates there.  It says nothing directly about billions of years, about the age of the Earth, about the Big Bang, or about anything beyond the Local Group.  Overclaiming it would be the same error this project criticises elsewhere.

Why this one object carries so much weight

Most arguments about distant starlight are arguments about categories.  How far away are galaxies in general?  How reliable is the distance ladder as a whole?  Does redshift mean what astronomers say it means?  Those questions are worth answering, and the cosmic distance ladder answers them.  But they are answered with statistics over populations, and a determined skeptic can always propose that the population is misunderstood.

SN 1987A is different.  It is one object.  Everything about it was measured, not inferred from a class.  Its distance was measured by geometry, twice.  Its explosion was timed by clocks on three continents.  Its progenitor star is on photographic plates going back to 1896.  Its debris has been imaged, spectroscopically dissected, weighed and mapped for forty years, and the observations are still coming in.

That makes it a test case rather than an argument.  A young-earth model of the sky does not have to explain "galaxies in general."  It has to explain this.

The article walks through the observations in the order they were made, then asks what each young-earth answer has to say about them.  The conclusion is that one answer fails outright, two survive but at a price their own advocates have named, and none of them gets to treat SN 1987A as a minor difficulty.

23 February 1987: the neutrinos came first

Three detectors, three continents, one burst

At 07:35:35 UT on 23 February 1987, the Kamiokande-II detector under a mountain in Japan recorded eleven electron events in thirteen seconds, at energies from 7.5 to 36 MeV.  The first two pointed back toward the Large Magellanic Cloud, at 18° ± 18° and 15° ± 27°.¹

The same burst hit a salt mine in Ohio.  The Irvine-Michigan-Brookhaven detector logged eight events in six seconds, at 20 to 40 MeV.²  A third detector, the Baksan Underground Scintillation Telescope in the Caucasus mountains, recorded a signal on the same date.³

The three detectors were built by different groups, used different technologies, sat under different mountains, and had different energy thresholds.  None of them was watching for a supernova.  Kamiokande was looking for proton decay.

Francesco Vissani's 2014 re-analysis compiles all the events with their relative times and energies in one table, which is the cleanest place to see the structure of the burst:¹⁸

Vissani's own summary of what these 29 events are worth is worth quoting, because it comes from a particle physicist with no stake in any of this:

"The events observed in underground neutrino detectors a few hours before SN1987A explosion are still the only direct validation of our understanding of what happens in the first instants after a gravitational collapse."¹⁸

Why the neutrinos beat the light, and by how much

Here is the part that matters, and it is not a technicality.

When a massive star runs out of fuel, its iron core collapses in about a second.  The collapse releases an enormous amount of gravitational binding energy, and roughly all of it leaves as neutrinos, because neutrinos barely interact with matter.  They pass straight out through the star.

Light cannot do that.  The photons produced in the collapse are trapped.  What gets out is the shock wave, which has to plough through the entire body of the star before it reaches the surface.  Only when the shock breaks out does the star brighten.  For a compact blue supergiant, that transit takes a few hours.  For a bloated red supergiant it takes a day or more.

So the prediction is specific:  a neutrino burst, then a delay set by the size of the star, then the light.  That is exactly what happened.

The neutrinos arrived at February 23.32 UT.  Albert Jones, an amateur astronomer in New Zealand, was monitoring variable stars in the Large Magellanic Cloud and looked at that field on February 23.39 UT.  He saw nothing.  Robert McNaught photographed the same field at February 23.443 UT and the supernova is on the plate at about magnitude 6.5.⁴

Arnett, Bahcall, Kirshner and Woosley put the timing in one sentence in their 1989 review:

"The neutrinos arrived at February 23.32 UT (55, 179), and the first optical brightening was recorded at February 23.443 UT (234, 235), less than 3 hr later."⁴

And the physical reading of that gap, in the same review:

"The neutrinos were seen first, and the brightening followed, as expected if the neutrinos signal the deposition of energy in the star from a collapsing core.  The very rapid brightening of SN 1987A [by a factor of 100 in 3 hr, rather than days as seen in other Type II supernovae (248)] indicated that the atmosphere of this supernova was more compact than the red supergiants that give rise to normal Type II supernovae."⁴

Read that twice.  The three-hour delay is not an awkward fact that had to be accommodated.  It is a measurement of the size of the star, taken by comparing two independent messengers, and it agreed with what the photographic plates already said about the star's spectral type.  Two lines of evidence, taken by unrelated instruments, converged on "the star was small and blue."

What the burst rules out about the explosion

Three things follow from the neutrino detection that nothing else could establish.

First, the core actually collapsed.  Before 1987 the neutrino signature of core collapse was theory.  After 1987 it was data, and the data matched the theory in total energy, in duration, and in spectrum.

Second, a compact object formed.  The duration of the burst indicates a neutron star rather than prompt collapse to a black hole.  Fransson and colleagues state the inference plainly in their 2024 JWST paper:  "A neutrino burst was detected coinciding with SN 1987A; its duration indicates that the compact object is a NS."¹³  That prediction sat unconfirmed for 37 years.  Section Forty years of follow-up covers what happened to it.

Third, neutrinos and photons travelled the same 167,000 light years at the same speed and along the same path, to a precision nothing on Earth can match.  Leo Stodolsky pointed out that the near-coincidence of the two arrival times bounds any difference between the speed of light and the limiting velocity for neutrinos.  Arnett's review states the result:

"With a generous estimate of 10 hr for the uncertainty of the coincidence in arrival time between the first photons and the intense burst of neutrinos, Stodolsky noted that the speed of photons and of neutrinos cannot differ by more than 1 part in 10⁸, which is close to the accuracy with which the speed of light is known."⁴

The same coincidence tests the weak equivalence principle over intergalactic distances:

"The observation of a neutrino burst within a few hours of the associated optical burst from SN 1987A provides a new test of the weak equivalence principle, demonstrating that neutrinos and photons follow the same trajectories in the gravitational field of the Galaxy (204, 212). … Their results show that photons and neutrinos move along the same trajectories to an accuracy of about 0.5% or better."⁴

Notice what has happened to the logic.  The bound on the neutrino-photon speed difference is derived from the assumption that both left the same event 167,000 years ago.  A young-earth reader is free to reject the premise.  But then the coincidence of two messengers, of completely different physical character, arriving within three hours of each other after a journey that on the standard account took 167,000 years, is a coincidence with no explanation at all.

The star that used to be there: Sanduleak −69 202

Ninety years of photographs

Nobody had ever known, before 1987, which star a supernova came from.  This time they did, and the reason is boring in the best way:  somebody had catalogued it.

Nicholas Sanduleak surveyed the Large Magellanic Cloud with an objective prism in 1970 and listed a twelfth-magnitude star as spectral class OB.  It became Sanduleak −69 202.  The observational record afterwards is thin, because nothing about the star drew attention:⁵

"Unfortunately, the observational record of Sk -69 202 is limited, partly because no spectral peculiarities or variability called attention to Sk -69 202 during the decades prior to its demise."⁵

Thin is not the same as absent.  Parthasarathy, Branch, Baron and Jeffery list what exists:

"Isserstedt (1975) obtained UBV photoelectric photometry between 1971 and 1973:  V = 12.24, B − V = +0.04, and U − B = −0.65, characteristic of spectral type B3 I.  Walborn et al. (1987) analyzed eight CTIO 4–m prime–focus plates obtained between 1974 and 1983 that covered the wavelength range from blue to near–IR.  Sk -69 202 appeared to be normal, and no near–IR excess (no hot dust) was detected.  Also, there was little or no light variability between 1974 and 1983 (Walborn et al. 1987) and between 1970 and 1981 (Blanco et al. 1987).  Plotkin & Clayton (2004) examined a large number of Harvard blue patrol plates obtained from 1896 to 1954 and found no variability at the 0.3 magnitude level."⁵

So there is a photographic record of this star, at this position, holding steady for ninety-one years.  Then it exploded and is not there any more.

Their conclusion about its state just before the explosion:

"Sk -69 202 was a normal B3 I supergiant shortly before exploding."⁵

The blue supergiant problem, which is a real problem

Here is where honesty is more useful than triumph.  SN 1987A did not behave the way theory said a Type II supernova should behave, and the reason is that the star that exploded was the wrong colour.

Standard stellar evolution says the progenitor of a Type II supernova should be a red supergiant, a huge cool star.  Sanduleak −69 202 was a blue supergiant, hot and comparatively small.  Parthasarathy and colleagues state the consequence:

"SN 1987A was classified as a Type II supernova (SN II) in view of the strong hydrogen lines in its optical spectrum, but because it was the explosion of a blue supergiant (BSG) rather than a red one (RSG), it was an atypical SN II:  its light curve did not reach maximum until three months after core collapse and at maximum it was only about 10 percent as luminous as most SNe II."⁵

Nobody has settled why.  The three candidate explanations each have a defect, and the paper lists them without pretending otherwise.  The single-star story is that the star became a red supergiant, blew off the ring material as a slow wind, then returned to the blue.  The trouble is that the return to blue requires assumptions chosen to produce it:

"Within the current uncertainties of stellar evolution it is possible to make assumptions that cause models of single stars to return to the blue, but it is not clear that these assumptions are natural (Woosley et al. 2002)."⁵

The binary-merger story is that Sanduleak −69 202 was two stars that merged.  That produces a blue supergiant and a complicated circumstellar system, and it is the currently favoured account, but it needs the merger to have happened at a suspiciously convenient moment:

"Another possible difficulty with the merger model is that it requires fine tuning to get the merger to happen just 20,000 years before the supernova, which would make SN 1987A an uncommon event (Woosley et al. 2002).  Also, if Sk -69 202 was a merger product, rapid rotation of a few hundred km s−1 is expected.  As mentioned above, the spectra of Sk -69 202 showed no evidence for rapid rotation."⁵

The third option is that the star was never red at all, which then leaves the low-velocity circumstellar rings with no source.  Their closing line on the whole question:

"The history of Sk -69 202 continues to be a mystery."⁵

That sentence should be read carefully by both sides.  Stellar evolution theory took a real hit here.  A model that could not predict the colour of the most-studied supernova progenitor in history is not a model anyone should treat as finished.  A young-earth reader is entitled to say so.

But notice what the failure does not touch.  The distance is measured by geometry and does not depend on which evolutionary track the star followed.  The neutrino burst does not depend on it.  The radioactive decay rates do not depend on it.  The blue supergiant puzzle is a problem for stellar evolution theory.  It is not a problem for the light travel time, and using it as one would be a bait and switch.

The distance, measured twice by geometry

The ring is a stopwatch

The explosion lit up a ring of gas the star had thrown off about 20,000 years before it died.  That ring is the single most useful object in this whole story, because it turns the supernova into a self-measuring instrument.

Work through what happens when the flash goes off.

Light leaves the star in every direction at once.  Some of it comes straight to Earth, and that is the light seen on 23 February 1987.  Some of it goes sideways, hits the ring, and makes the ring's gas fluoresce in ultraviolet emission lines.  That glow then travels to Earth as well, but by a longer route:  star to ring, then ring to Earth.

The ring is tilted.  It is not seen face-on.  So the part of the ring nearest Earth has a head start, and the part farthest away is behind.  The near side lights up first.  The far side lights up last.  Everything in between reports in order, which is why the observed emission does not switch on and off but rises to a peak and falls.

Write the ring's radius as R and its tilt as i.  The extra distance the light has to cover, compared with light that came straight here, is R minus the head start, and the head start is R sin i at the near edge and minus R sin i at the far edge.  So the near edge reports at

t₀ = (R/c)(1 − sin i)

and the far edge at

t_max = (R/c)(1 + sin i)

Now add the two together.  The sin i terms cancel:

t₀ + t_max = 2R/c

The ring's true radius is the speed of light multiplied by half the sum of the two delays.  The tilt drops out.  No stellar model, no assumed luminosity, no cosmology.  Two dates and a stopwatch.

Doing the arithmetic

Panagia and colleagues measured the caustics of the ultraviolet light curves.  Andrew Gould, who later disagreed with their result, states their numbers cleanly:  t₋ = 83 ± 6 days and t₊ = 413 ± 24 days.¹⁹¹⁹

Add them:  496 days.  Half of that is 248 days.  So the ring's radius is 248 light days, and its diameter is 496 light days, which is 1.28 × 10¹⁸ cm.  Panagia's published value from the full fit is (1.27 ± 0.07) × 10¹⁸ cm.⁶

Subtract them instead, and divide by the sum:  330/496 = 0.665, so i = 41.7°.  That is the tilt of the ring, derived purely from timing.

Now compare it with the tilt derived from the shape of the ring in Hubble's image.  Peter Jakobsen's team measured the ellipse as 1″.66 ± 0″.03 by 1″.21 ± 0″.03.  The ratio of the axes gives i = 43.2° ± 1.8°.¹⁹

Two completely different measurements of the same tilt, one from a light curve and one from a photograph, agreeing to within about a degree and a half.  That agreement is what tells you the ring really is circular and the method really is working.  Panagia's paper says so:

"Our analysis confirms that the observed elliptical structure is indeed a circular ring seen at an inclination of i = 42.8° ± 2.6° and provides a determination of its absolute diameter of (1.27 ± 0.07) × 10¹⁸ cm.  Its ratio to the angular diameter of 1″.66 ± 0″.03 (Jakobsen et al.) gives an accurate determination of the distance to SN 1987A, i.e., d(1987A) = 51.2 ± 3.1 kpc."⁶

The last step is grade-school trigonometry.  A true diameter of 1.27 × 10¹⁸ cm subtending 1″.66 means the object is 1.27 × 10¹⁸ / (1.66/206265) centimetres away, which is 1.58 × 10²³ cm, or 51 kiloparsecs.  About 167,000 light years.

Nothing in that chain uses a Cepheid, a Type Ia supernova, a Tully-Fisher relation, or the Hubble constant.  It is the same trick a surveyor uses:  know the true size of a thing, measure the angle it covers, get the distance.  The only unusual step is that the true size was measured with a clock instead of a tape measure.

Andrew Gould's objection, which makes the distance smaller

The honest thing to report is that Panagia's result did not go unchallenged, and the challenger was right about the physics.

Andrew Gould argued that the ring gas does not light up the instant the flash hits it.  Recombination takes time.  A line can be optically thick.  Any such delay pushes the measured t₀ and t_max later than the true light-crossing times, which inflates R, which inflates the distance.  His conclusion:

"Finally, I consider the effect of relaxing the assumption that the fluorescence response is prompt, whether due to recombination or some other cause.  In any event, the general effect of a delayed response will be to cause the light curve caustics to be delayed relative to what they would be with prompt fluorescence.  Hence, if the light curves are analyzed using a prompt-response expression such as equation (2.1)–(2.3), then t₊ and t₋ will be overestimated.  This in turn will lead to an overestimate of D_SN."¹⁹

He also refitted the light curves with a proper ring reflection function instead of Panagia's top-hat, and got earlier caustics:  t₋ = 75.0 ± 2.6 days and t₊ = 390.0 ± 1.8 days.¹⁹  His result is therefore an upper bound, not an equality:

"I derive an upper limit to the distance to SN 1987A of D_SN < 46.77 ± 0.76 kpc, or a distance modulus μ_SN < 18.350 ± 0.035.  If SN 1987A lies in the plane of the Large Magellanic Cloud (LMC) and hence 500 pc in front of the LMC center, this implies μ_LMC < 18.37 ± 0.04. … If the response function of the supernova-ring gas is prompt, then these upper limits become equalities."¹⁹

Gould and Osamu Uza revisited it in 1997 with better spectral reductions and with evidence that the ring is slightly elliptical after all, axis ratio about 0.95, which raises the bound to μ_LMC < 18.44 ± 0.05.²⁰²⁰  That is 48.8 kiloparsecs.

So the ring echo, done as carefully as anyone has done it, says the Large Magellanic Cloud is closer than 48.8 kiloparsecs, not 51.2.  This is a genuine reduction, and it should be stated in exactly those terms rather than buried.  It is also, from a young-earth point of view, completely useless.  48.8 kiloparsecs is 159,000 light years.  The correction is four per cent.  The gap that has to be closed is a factor of 26,000.

The eclipsing binaries: a second geometry, 28 years later

The other geometric handle on the Large Magellanic Cloud has nothing to do with SN 1987A at all.

Grzegorz Pietrzyński's group spent two decades on late-type detached eclipsing binaries in the same galaxy.  A detached eclipsing binary is two stars orbiting each other, edge-on to Earth, with no mass transfer between them.  The eclipses give the ratio of the stellar radii to the orbital separation.  The Doppler shifts of the two spectra give the orbital velocities, and therefore the separation in kilometres.  Multiply and you have each star's radius in kilometres.  A calibrated surface-brightness relation gives the star's angular diameter from its colour.  Divide, and you have the distance.

Their 2019 result:

"We use this calibration to determine the geometrical distance to the Large Magellanic Cloud that is precise to 1 per cent based on 20 eclipsing binary systems.  The final distane is 49.59 +/- 0.09 (statistical) +/- 0.54 (systematic) kiloparsecs."⁷

Put the two numbers side by side.

Two of these were obtained 28 years apart.  They share no instruments, no calibration, no target objects, and no physics beyond the speed of light and Euclidean geometry.  The eclipsing binary result falls between the light echo's central value and Gould's upper bound.  That is what agreement looks like when the errors are honest.

The full treatment of how these fit into the wider distance scale, and why shrinking the universe does not work, is in the cosmic distance ladder.  It is not repeated here.

The radioactive clock

A supernova is lit by its own ashes

A supernova is bright for two reasons at two different times.  The first flash is the shock breaking out of the star.  That fades in days.  After that, the thing glowing is radioactive debris.

The explosion fuses a large amount of nickel-56.  Nickel-56 decays to cobalt-56 with a half-life of 6.08 days.  Cobalt-56 decays to iron-56 with a half-life of 77.2 days.²¹²¹  The gamma rays from those decays are absorbed by the surrounding debris, heat it, and come out as visible and infrared light.  So the brightness of a supernova months after the explosion is a direct readout of a nuclear decay rate.

That is a testable prediction with no free parameters.  If the light curve is powered by cobalt-56, it has to fall off exponentially at cobalt-56's rate.

Nicholas Suntzeff and Patrice Bouchet assembled the bolometric light curve from photometry spanning 0.32 to 20 micrometres, out to day 903:

"SN 1987A has declined exponentially with an e-folding time of 100 ± 5 days from day 126 to day 903.  The sum of the uvoir flux and the high-energy flux of the gamma-ray lines and x-ray continuum from model fits is consistent with the energy released by 0.055–0.090 [M☉] of ⁵⁶Co."²²²²

Cobalt-56's e-folding time is 111.3 days.²²  The observed 100 ± 5 days is slightly faster, and the reason is not mysterious:  as the debris thins out, more of the gamma rays escape without being absorbed, so the visible output drops a little faster than the decay itself.  Add the escaping high-energy flux back in, which Suntzeff and Bouchet did, and the books balance against about 0.07 solar masses of cobalt-56.

That is one full nuclear decay chain, running at its laboratory rate, watched for two and a half years, 167,000 light years away.

Then the gamma rays themselves

The decay of cobalt-56 emits gamma-ray lines directly.  SN 1987A was the first supernova ever caught doing it.  Grebenev's group states the fact in passing while setting up their own result:

"Note that this remnant was the first for which the X-ray and γ-ray emission from the decay of ⁵⁶Co and ⁵⁷Co was detected."⁸

So the argument does not rest on an inference from a light curve alone.  The isotope was identified by its own emission lines.

Titanium-44 takes over, and INTEGRAL catches it

Cobalt-56 and cobalt-57 are gone after three or four years.  Something else has to power the remnant, and the candidate has been titanium-44 since long before it was detected:

"It is assumed that the radioactive decay of ⁴⁴Ti powers the infrared, optical and UV emission of supernova remnants after the complete decay of ⁵⁶Co and ⁵⁷Co (the isotopes that dominated the energy balance during the first three to four years after the explosion) until the beginning of active interaction of the ejecta with the surrounding matter."⁸

Titanium-44's decay scheme is worth spelling out, because the detail is the argument.

Titanium-44 captures an electron and becomes scandium-44.  Its characteristic time for that step is 85.0 ± 0.4 years, measured on a laboratory bench in a fourteen-year study by Ahmad and colleagues.⁸  Scandium-44 then decays to calcium-44 in about 5.7 hours.  Along the way the chain emits photons at five specific energies, with five specific efficiencies:

"The only immediate way to determine the initial mass of ⁴⁴Ti (denoted M₄₄) in a supernova remnant is through its emission in the direct-escape lines at energies Eᵢ = 4.1, 67.9, 78.4, 511 and 1,157 keV.  This isotope decays according to the chain ⁴⁴Ti→⁴⁴Sc→⁴⁴Ca, with a characteristic time of t₄₄ ≃ 85.0 ± 0.4 years (ref. 16).  Photons in the three X-ray lines, namely 4.1, 67.9 and 78.4 keV, are emitted with efficiencies (average numbers of photons per decay) Wᵢ = 17.4%, 87.7% and 94.7% during the first stage of the decay, and photons in the γ-ray lines, namely 511 and 1,157 keV, are produced with Wᵢ = 188.8% and 99.9% during the second very quick (∼ 5.7 h) stage."⁸

Every number in that paragraph comes from nuclear physics done on Earth.

In 2010 and 2011 the INTEGRAL satellite stared at SN 1987A for about 4.5 million seconds.  It found an excess at the remnant's position, and only in the energy band containing the titanium-44 lines:

"Here we report observations of hard X-rays from the remnant of supernova 1987A in the narrow band containing two direct-escape lines of ⁴⁴Ti at 67.9 and 78.4 keV.  The measured line fluxes imply that this decay provided sufficient energy to power the remnant at late times.  We estimate that the initial mass of ⁴⁴Ti was (3.1 ± 0.8) × 10⁻⁴ M☉."⁸

That is the sentence this article exists to underline.  A specific isotope, identified by two emission lines at energies fixed by its own nuclear structure, decaying at a rate that matches a fourteen-year laboratory measurement, in an object 167,000 light years away whose light left 167,000 years ago.

How strong is the detection, honestly

Not overwhelming, and the paper says so.  The combined signal-to-noise on both lines together is 4.1.  Split into two narrow bands that each contain one line, the significances are 3.0 and 3.1, giving a combined chance probability of 1.3 × 10⁻⁶, which the authors describe as about 4.7σ.⁸

Two independent confirmations have arrived since.

NuSTAR observed the same lines and derived M₄₄ = (1.5 ± 0.3) × 10⁻⁴ M☉, about half the INTEGRAL value.⁹

In 2025 Roberta Giuffrida's group took a different route entirely.  Instead of the hard X-ray lines from the titanium, they looked for the atomic X-ray line from the daughter.  Electron capture leaves the newly made scandium atom with a hole in its innermost electron shell, and when that hole fills the atom emits a scandium K-shell X-ray at 4.09 keV:

"The electron capture results in a ⁴⁴Sc with a vacancy in its K-shell, which produces a K-shell transition with an X-ray emission line at 4.09 keV."⁹

They found it in Chandra spectra from 2016 to 2021 at about 3σ, giving M₄₄ = (1.6 ± 0.5) × 10⁻⁴ M☉, consistent with NuSTAR.⁹

This matters more than the mass number does.  4.09 keV is not a nuclear energy.  It is the energy difference between two electron orbitals in a scandium atom, which depends on the electron mass and the fine-structure constant.  So SN 1987A now reports laboratory-standard values for nuclear decay rates, nuclear line energies, and atomic line energies, all from the same place.

The disagreement, which should not be hidden

The four measurements of how much titanium-44 was made do not agree:

The spread is a factor of six between the extremes, and Seitenzahl's paper says so in its own abstract:  his best fit "is in disagreement with the much higher (3.1 ± 0.8) × 10⁻⁴ M⊙ recently derived from INTEGRAL observations."²¹

That is a live scientific dispute about how much of an isotope one supernova produced.  It should be reported, and it is reported here, and it changes nothing about the point at issue.  The disputed quantity is the amount.  The line energies are not in dispute by anyone.  Four groups using three instruments all found photons at the energies titanium-44 and scandium-44 emit in a laboratory.  Whether the star made half a ten-thousandth or three ten-thousandths of a solar mass of it does not touch the question of whether nuclear physics is the same there as here.

Forty years of follow-up

The strongest thing about SN 1987A is not any single measurement.  It is that the object has kept doing new things, on schedule, for four decades, and that people keep pointing better instruments at it.

The blast wave reaches the ring

The ring sits about 0.67 light years from the star.  The debris from the explosion was moving at a few thousand kilometres per second, so it took roughly a decade to get there.  When it arrived, the collision lit up.

Stephen Lawrence's group reported the sequence:

"We present the discovery of several new regions of interaction between the ejecta and equatorial circumstellar ring of SN1987A, an interaction leading to a much expanded development of the supernova remnant.  We also trace the development of the first such 'hot spot,' discovered in 1997, back to 1995.  Later hot spots seem to have emerged by early 1999."¹⁰

One hot spot in 1995, more by 1999, eventually a full necklace.  The ring lit up section by section as the shock swept around it, and the order and timing were watched happening.

The ring is now being destroyed, on a published schedule

Twenty years of Hubble and Very Large Telescope imaging showed the hot spots brighten almost exponentially to about day 8,000, then start to fade.  Claes Fransson's group tracked it:

"After an almost exponential increase of the shocked emission from the hotspots up to day ∼ 8,000 (∼ 2009), both this and the unshocked emission are now fading. … All of these observations indicate that the interaction with the supernova ejecta is now gradually dissolving the hotspots.  We predict, based on the observed decay, that the inner ring will be destroyed by ∼ 2025."²³²³

That is a dated prediction, made in 2015, about an object 167,000 light years away, and its due date has arrived.  Whatever else is true of SN 1987A, it is not a static picture.

The reverse shock

When fast ejecta slam into slower surrounding gas, two shocks form.  One runs outward into the gas.  The other runs backward into the ejecta, in the frame of the expanding debris.  The reverse shock shows up as broad hydrogen emission with a distinctive profile.

Hubble's refurbished Space Telescope Imaging Spectrograph mapped it in 2010, and the comparison against 2004 data gave the trend:

"Comparing observations taken in 2010 using the refurbished instruments on HST with data taken in 2004, just before the Space Telescope Imaging Spectrograph failed, we find that the Lyα and Hα lines from shock emission continue to brighten, while their maximum velocities continue to decrease."¹¹

Brightening emission, falling velocity.  That is what a shock decelerating into denser material does.

Dust, and a lot of it

Supernovae were long suspected of manufacturing interstellar dust.  SN 1987A is where that was checked.

ALMA imaged the remnant at four submillimetre and millimetre wavelengths in 2012:

"Longer wavelength emission is dominated by synchrotron radiation from shock-accelerated particles, shorter wavelengths by emission from the largest mass of dust measured in a supernova remnant (>0.2 M☉).  For the first time we show unambiguously that this dust has formed in the inner ejecta (the cold remnants of the exploded star's core)."¹²

Higher-resolution ALMA images in 2019 refined the number and the temperature:

"Fits to the far-infrared–millimeter spectral energy distribution give ejecta dust temperatures of 18–23K.  We revise the ejecta dust mass to M_dust = 0.2 − 0.4 M☉ for carbon or silicate grains, or a maximum of < 0.7 M☉ for a mixture of grain species."²⁴²⁴

Those same images turned up something else:  a spot where the dust is hotter than its surroundings, in the region where a neutron star was expected.  Cigan's group was careful about it:

"One of the possibilities is that a compact source provides additional heat at that location."²⁴

The neutron star, found in 2024

The neutrino burst of 1987 implied a neutron star.  For 37 years nobody could find it.  It was reasonable to worry that it had collapsed further into a black hole, or that the whole picture was wrong.

In July 2022 JWST pointed both its mid-infrared and near-infrared spectrographs at the remnant.  The result, published in 2024:

"In this work, we observe the supernova remnant with JWST spectroscopy finding narrow infrared emission lines of argon and sulphur.  The line emission is spatially unresolved and blueshifted in velocity relative to the supernova rest frame.  We interpret the lines as gas illuminated by a source of ionizing photons located close to the center of the expanding ejecta.  Photoionization models show that the line ratios are consistent with ionization by a cooling neutron star or pulsar wind nebula.  The velocity shift could be evidence for a neutron star natal kick."¹³

The lines are [Ar VI] at 4.529 micrometres and [Ar II] at 6.985 micrometres, plus [S IV] at 10.51 and [S III] at 18.71.  The narrow [Ar II] component is blueshifted by −259.6 ± 0.4 km s⁻¹ relative to the supernova's rest frame, and it is spatially unresolved, which is what a point source of ionizing radiation looks like.¹³

The authors do not overstate the result, and neither should this article:

"Given the uncertain model assumptions about the ionizing spectrum, density, elemental abundances and dust properties, we conclude that the observed lines can be explained by either the CNS or PWN photoionization models.  Both require the presence of a NS."¹³

A cooling neutron star or a pulsar wind nebula.  They cannot tell which.  Either way, something at the centre is emitting ionizing radiation, which is the prediction the neutrino burst made in 1987.

Still being measured in 2024

The XRISM satellite's Resolve calorimeter observed SN 1987A in 2024 and reported in 2025.  Its results include a temperature, an ionization state, Doppler line widths of 1,500 to 1,700 km/s from the bulk motion of the ejecta, and an independent upper bound on the titanium-44 mass consistent with earlier work.²⁵²⁵  It also found that the reverse shock has not yet reached the metal-rich core of the ejecta, which is a statement about where in a 38-year-old explosion the shock front currently sits.

There is no sign of this record closing.  The remnant is being watched with better instruments every decade, and every decade it does what the previous decade's physics said it would do.

What SN 1987A does to each young-earth answer

The published young-earth answers to distant starlight are catalogued and assessed in Young-Universe Cosmologies.  This section does not re-argue them.  It asks a narrower question of each:  what does this one object cost you?

Shrinking the distances: nothing left to shrink

The oldest move is to deny the distance.  SN 1987A closes it, and young-earth astronomers say so.

The ring echo needs one assumption, that the ring is intrinsically circular, and that assumption was cross-checked against the ring's photographed shape and passed.  Gould's reanalysis, which is the most hostile treatment the method has received, lowers the answer to under 48.8 kiloparsecs rather than raising doubt about the order of magnitude.  The eclipsing binaries agree to one per cent by a route with no overlap.

Danny Faulkner, an astronomer at Answers in Genesis, has ruled this family of answers out in print twice.  In 2013:  "for this solution to work, even the well determined trigonometric parallax method must be abandoned, but this is not physically supported."²⁶  In 2020:  "While all of these methods suffer from error, those errors would not reduce the distances down to just thousands of light-years."²⁷²⁷

He is right, and the shortfall is not close.  The required factor is about 26,000.

Light created in transit: this is where it dies

The mature-creation answer holds that God made the light beams already in flight, so no light ever left the star.  On that account SN 1987A did not explode.

What the beam then has to contain is not an image.  It is a synchronised, multi-messenger, four-decade record:

D. Russell Humphreys, a young-earth physicist who once held this view, set out the consequence in Starlight and Time:

"But according to the 'in-transit' theory, in spite of the images and particles astronomers observed, no actual supernova explosion would have ever happened!"²⁸²⁸

And the theological price he attached to it:

"It would make the study of distant stars into a kind of theological literary criticism — a study of the fiction God would have chosen to write for us in the sky."²⁸

Faulkner's test is sharper, and he built it out of his own observing career.  He has spent more than four decades watching eclipsing binaries.  His objection to created light is not that it is unlikely but that it makes God a forger of records:

"However, if the light of these stars were created in transit so that the light never left the stars, then we would see evidence of events (eclipses, for example) that would have never happened.  This is deceptive."¹⁴¹⁴

An eclipsing binary produces a repeating brightness dip.  SN 1987A produces a dated, cross-checked, four-decade history in four independent channels.  If an eclipse that never happened is deceptive, this is not a marginal case.

Jason Lisle printed the conclusion under a photograph of this supernova:  "If the light-in-transit model is correct, then this star never actually existed, and this explosion never really happened."¹⁵¹⁵  Donald DeYoung, defending mature creation, printed the Large Magellanic Cloud in his own paper with the caption:  "An exploding star observed here, named Supernova 1987A, is a challenge to the mature creation view."²⁹²⁹

The defence and the prosecution agree about the facts.  The full treatment, including the strongest defence available and where it still holds up, is in Light Created in Transit.

The anisotropic synchrony convention: survives the echo, pays elsewhere

Lisle's proposal is that Scripture uses a convention in which light arriving at an observer is treated as having travelled instantaneously, while light travelling away moves at c/2.  This is a legitimate coordinate choice;  the one-way speed of light genuinely is conventional in relativity.

The obvious objection is that the SN 1987A ring echo shows a real delay, so light cannot be instantaneous.  That objection fails, and it should be retired.  The light that lit the ring did not come toward Earth.  It went sideways and outward first, which under ASC is the slow direction.  Lisle works the geometry and gets the observed answer:

"The equation shows that the closer the light is to being directly aimed at the observer, the faster it will be. … If you use the actual angles (which are quite small), you will find that ASC predicts that the light along AE will reach point E one year after the light moving along AB reaches point B. … Hence, the observations of the light from the supernova seen in 2014 and 2015 is exactly what ASC predicts."¹⁶¹⁶

He is right, and he has to be, because the delay is a physical invariant and a change of coordinates cannot alter one.  Any objection to ASC that bottoms out in something a detector reads is dead before it starts.

ASC is also better than created light in the way that matters most:  under ASC the supernova really exploded.  Faulkner marks the distinction when defending his own model:  "all of the processes that we see playing out, such as SN 1987A, actually happened."¹⁴

What ASC cannot do is remove the object's own history.  Under ASC the explosion happened in 1987 by our clocks, because that is when we saw it.  Run the star's biography backwards from there and it does not fit inside a young universe.

The sharpest instance is the ring itself.  The ring is 0.67 light years in radius and expanding at 10 kilometres per second, which gives it a kinematic age of about 20,000 years before the explosion.  Two independent groups state it:  the ring system "appears to have been ejected by the progenitor some 20,000 years before Sk -69 202 exploded,"⁵ and "The ER is thought to have originated ∼ 20,000 years before the explosion."¹³

Both of those are measurements, not models.  The radius comes from the light echo.  The expansion velocity comes from Doppler shifts.  Divide one by the other.

So under ASC the ring was launched in about 18,000 BC.  That is before Day Four on any young-earth chronology, by a factor of three.  The ring therefore has to have been created already in flight, already 14,000 years into its expansion, already carrying nitrogen enriched by CNO-cycle processing to about ten times the solar ratio.⁵

That is created maturity, moved from the light beam to the gas.  The events we watch did happen;  the object we watch them in was made carrying a history it did not live.  Faulkner's own criterion was designed to separate those two cases, and ASC lands on the acceptable side of it by a hair.  But the appearance-of-age problem is not solved.  It is relocated.  The argument is worked out in full in The Anisotropic Synchrony Convention.

A decaying speed of light: compensated into silence

Barry Setterfield's proposal is that light was much faster in the past and has been slowing since.  SN 1987A is the obvious test, and Setterfield has answered it.

His answer is that everything else changed too, in step, so that no observation can ever detect the difference.  On the radioactive decay that powers the light curve:

"So what is happening when we see SN1987A and other supermova is that we are indeed seeing the process in slow motion, since the half-life of Cobalt isotopes was proportionally shorter when the explosion occurred.  In a word, the slowing of light exactly counteracts the faster decay rate, so we see it from earth in the same elapsed time as would normally occur."¹⁷¹⁷

On the light echo off the ring:

"However, what has been ignored in that criticism is that the light is being received at only 1/10th of the rate the original events occurred.  So we see events which nonetheless are in accord with a value for c equal to c now.  It is therefore impossible to deal with the speed of light issue in this fashion, as the slow-motion effect will always match the faster speed."¹⁷

Take him at his word and follow the arithmetic, because the result is not what either side usually claims.

If light was ten times faster when the flash left the star, the echo delays imposed at the ring were ten times shorter, and the reception rate stretches them back out by ten.  The ring radius derived from the observed delays is therefore unchanged.  So is the distance.  Setterfield's model does not move SN 1987A closer.  It leaves the supernova 167,000 light years away and shortens the transit time.

The same compensation covers the line energies.  Setterfield holds that hc is constant and that e²/ε₀ is constant, so the fine-structure constant "will never register any change, no matter how the speed of light or the other three quantities vary."¹⁷  A photon's energy is hc/λ, so if hc never changes, neither do the observed line energies.  The titanium-44 lines at 67.9 and 78.4 keV come out right by construction.

That is not a rescue.  It is a confession.  Setterfield has said in his own words that no observation of a distant object can bear on the question, and he is correct about his own model.  Jean-Philippe Uzan's standard treatment of varying constants names the general case:  a variation that leaves every dimensionless number unaffected "is actually just a redefinition of units."³⁰³⁰

There is a further cost that SN 1987A makes concrete.  The neutrinos and the photons arrived within three hours of each other.  If the transit time was compressed by a factor of ten million, the two messengers still had to track the same changing speed to that precision over the whole journey, and nothing in the model requires them to.  It is not an impossible demand.  It is one more thing the model has to be tuned to produce and cannot predict.

The history of the proposal, including George Aardsma's weighted refit that killed the statistical case and Setterfield's own 2001 withdrawal of the redshift mechanism, is in A Decaying Speed of Light.

Faulkner's Day Four miracle: survives, and admits it predicts nothing

Faulkner's own 2013 proposal is that God brought the starlight to Earth by a miracle during Day Four, with no physical mechanism and no change to the speed of light since.  SN 1987A costs him nothing, because on his account the events happened:

"While the dasha solution superficially resembles the mature creation solution, it differs in a fundamental way:  all of the processes that we see playing out, such as SN 1987A, actually happened.  With the mature creation solution, none of these distant events occurred."¹⁴

That is a real advantage and it should be granted.  It is also bought at a price Faulkner states himself:  "Since my modest proposal appeals to a miracle, there may be no physical predictions and hence nothing that we can test."²⁶²⁶

A model that survives every observation because it forbids itself from predicting any is safe.  It is not thereby supported.

Relativistic cosmologies: SN 1987A is not their problem

Humphreys' white-hole model and his later timeless-zone proposal, and Hartnett's Carmelian cosmology, all work by letting large amounts of time pass in distant space while little passes on Earth.  On those models SN 1987A exploded, its light really crossed 167,000 light years, and everything in this article happened as described.  SN 1987A raises no objection to them at all.

Their difficulties lie elsewhere, and they are substantial:  Humphreys abandoned the white-hole model, and Hartnett withdrew his Carmelian cosmology in an abandonment note dated 23 June 2021.  Both are catalogued in the parent article.

What SN 1987A does not prove

Overclaiming a good argument is the fastest way to lose it.  Here is what this object does not settle.

It is one supernova, in one nearby galaxy.  167,000 light years is a rounding error on cosmological scales.  The Large Magellanic Cloud is a satellite of the Milky Way.  Nothing about SN 1987A speaks directly to distances of megaparsecs, let alone gigaparsecs.  Those rest on the ladder and its ladder-free cross-checks, which is a separate case made in the cosmic distance ladder.

It says nothing about billions of years.  The measurement is that light took roughly 160,000 to 170,000 years to reach Earth.  That is fatal to a 6,000-year light travel time.  It is not evidence for 13.8 billion years, and citing it as such would be dishonest.  The age of the universe is argued from stellar populations, nucleosynthesis, the microwave background and expansion history, none of which appear in this article.

It says nothing about the age of the Earth.  Geology and radiometric dating are separate arguments with separate evidence.  A young-earth reader can accept everything here and still hold that the Earth is young, at the cost of accepting that the universe is at least a few hundred thousand years old as measured by light travel time.  Several young-earth models are built to allow exactly that.

It does not refute the Big Bang, and it does not confirm it.  SN 1987A is a stellar event.  It is neutral on cosmology.

It does not settle whether stellar evolution theory works.  It arguably shows the opposite.  The progenitor was the wrong colour, the explanations for that are all unsatisfying, and the amount of titanium-44 produced is still disputed by a factor of several.  A skeptic who wants to argue that astrophysicists overstate their confidence in stellar models has a real case here, and SN 1987A is exhibit A for it.

It does not tell you whether God could have made the sky differently.  The argument in this article is empirical, not theological.  If a reader holds that God created a universe with a detailed false history, no measurement can contradict that.  What the measurements establish is that on any account where the sky reports real events, the light from this one took far longer than 6,000 years to arrive.

What remains after all of those subtractions is still substantial.  One object, measured to death for forty years, pins its own distance by geometry twice over, runs nuclear physics at laboratory rates, and delivers a dated multi-messenger record that no young-earth model can absorb without either denying the events or invoking a miracle that predicts nothing.

Objections and responses

"The ring echo distance depends on assuming the ring is circular.  Assume otherwise and the distance changes."  It does, and the effect has been calculated.  Gould showed that for a ring with eccentricity e, the distance estimate changes as D → (1 − 0.4e⁴)D, which is tiny for any plausible ellipticity.¹⁹  When Gould and Uza later argued from new evidence that the ring really is slightly elliptical, axis ratio about 0.95, the effect was to move the upper bound from μ < 18.37 to μ < 18.44, about three per cent in distance.²⁰  More to the point, the circularity assumption was tested rather than asserted:  the tilt derived from the light-curve timing, 41.7° ± 2.0°, agrees with the tilt derived from the photographed shape, 43.2° ± 1.8°.¹⁹  A non-circular ring would not generally make those two agree.

"Two geometric methods agreeing proves nothing if both are wrong in the same way."  They cannot be, because they have no shared machinery.  The ring echo uses ultraviolet spectroscopy, a light-crossing time, and a Hubble image of one object.  The eclipsing binary method uses radial-velocity curves, photometric eclipse depths, and an infrared surface-brightness-colour calibration applied to twenty binary systems.  There is no calibration in common, no target in common, and no instrument in common.  For both to be wrong by the same factor would require a coincidence with no mechanism behind it.

"Nobody has directly measured the decay rate of titanium-44 in 1987A.  You are assuming the laboratory rate."  Partly true and worth separating carefully.  What was measured is the line energies and the fluxes.  The energies are unambiguous:  photons arrived at 67.9 and 78.4 keV, and independently at 4.09 keV, which are the energies that decay chain produces on Earth.  The rate enters when converting flux into an initial mass, and that step does assume the laboratory lifetime of 85.0 ± 0.4 years.⁸  So the strong claim is about energies, not rates.  That is still a direct measurement of nuclear and atomic structure at 167,000 light years, which is what a varying-constants model has to explain.

"The titanium mass estimates disagree by a factor of six.  Why should any of this be trusted?"  Because the disagreement is about a quantity that depends on models of how transparent the debris is, how the ejecta are distributed, and how much continuum sits under the lines, while the agreement is about photon energies, which do not.  The dispute is real and this article reports it in full.  It is also orthogonal to the point.

"The blue supergiant problem shows stellar evolution theory failed.  Why trust anything else it says?"  The premise is fair;  the inference is not.  Stellar evolution theory did fail to predict the progenitor's colour, and the leading explanations still require either unnatural assumptions or a finely timed binary merger.⁵⁵  But no step in the distance measurement, the neutrino detection, or the radioactive decay chain uses stellar evolution theory.  They use timing, trigonometry, and nuclear physics.  Discrediting one does not touch the others.

"The neutrino burst was only 29 events.  That is a small number."  It is, and the statistical analyses of it are correspondingly careful.  But the argument does not lean on the count.  It leans on three independent detectors on three continents recording events in the same seconds, pointing back at the same galaxy, hours before an optical brightening in that galaxy.  Vissani's assessment of the value of those events, after a quarter century of reanalysis, is that they remain "the only direct validation of our understanding of what happens in the first instants after a gravitational collapse."¹⁸

"Under ASC the light arrived instantly, so there is no light travel time problem here."  Correct, and this article grants it.  The ASC answer to the echo geometry is right.  What ASC does not do is give the ring a history that fits inside 6,000 years.  A ring 0.67 light years in radius expanding at 10 km/s was launched about 20,000 years before the explosion,⁵¹³ which under ASC means about 18,000 BC.  Either the ring was created already expanding, or the chronology is wrong.  Lisle's own answer to this general class of objection is that a created initial condition is always available:  "Not every sequence is a time-sequence."¹⁶  That is consistent, and it is the point at which the model stops making predictions.

"You are a young-earth publication and this article argues against a young earth."  It argues against several young-earth answers to one question, using the same standard applied everywhere else on this site:  what do the measurements say.  Two young-earth astronomers, Faulkner and Lisle, reject created light for the same reason given here, and Humphreys rejected it in 1994.  Reporting that a popular argument fails is not an attack on the position;  concealing it would be.

"Demanding a natural explanation for a supernatural creation week is inconsistent."  This is Faulkner's own objection against himself and it has force.¹⁴  The response is that the choice here is not between a miracle and no miracle.  Faulkner's Day Four proposal is a miracle, and it leaves the SN 1987A record truthful.  Created light is also a miracle, and it does not.  Both are supernatural.  Only one makes the sky a report rather than a composition.

Where this leaves things

SN 1987A is the best-observed event in the history of astronomy outside the solar system, and it is not finished.  Its distance was measured with a stopwatch and a photograph.  Its explosion was clocked by three neutrino detectors before anyone saw a thing.  Its progenitor is on plates going back to 1896.  Its ashes have been glowing at the decay rates of cobalt-56 and titanium-44 for forty years.  In 2024 the neutron star that the 1987 neutrinos implied finally showed itself.

Any account of the sky has to fit that.

Created light in transit does not fit it, and the young-earth astronomers who work in this field say so:  Humphreys in 1994, DeYoung in his own figure caption in 2010, Lisle under a photograph of the supernova in 2010, Faulkner repeatedly.  The position requires a manufactured record of an event that never occurred, in four independent messengers, agreeing with each other and with laboratory physics.

The anisotropic synchrony convention fits the light travel time and does not fit the ring's own age.  A 0.67-light-year ring expanding at 10 kilometres per second was launched about 20,000 years before the explosion, which lands before Day Four.  ASC keeps the events real and moves the appearance of age into the object.

A decaying speed of light fits everything, by construction, and its author has said that it will always fit everything.  That is the problem with it.

Faulkner's Day Four miracle fits everything and predicts nothing, which he states himself.

There is no dishonest way to hold a young universe against this object.  There are honest ways, and they cost something.  The point of laying the case out at this length is that the cost should be paid with open eyes rather than argued away.

Sources