Distance to Supernova SN1987A and the Speed of Light
by Dave Matson
When supernova SN1987A exploded, its light soon struck a ring of gas some
distance from the star and illuminated it. As viewed from Earth, the ring
appeared around the supernova about a year after it exploded. Its angular
size combined with the time it took for the ring to be illuminated after
SN1987A was first observed allows a direct, trigonometric calculation of
the distance to that supernova with an error of less than 5%.
Oddly enough, if we use the older Newtonian physics (which most creationists
love because it allows them to play around with the speed of light) we find
that a change in the speed of light does not affect our calculations of the
distance to SN1987A! Gordon Davisson pointed out that interesting tidbit.
The distance is based on triangulation. The line from Earth to the supernova
is one side of the triangle and the line from Earth to the edge of the ring
is another leg. The third leg of this right triangle is the relatively short
distance from the supernova to the edge of its ring. Since the ring lit up
about a year after the supernova exploded, that means that a beam of light
coming directly from the supernova reached us a year before the beam of light
which was detoured via the ring. Let us assume that the distance of the ring
from the supernova is really 1 unit and that light presently travels 1 unit
If there had been no change in the speed of light since the supernova exploded,
then the third leg of the triangle would be 1 unit in length, thus allowing
the calculation of the distance by elementary trigonometry (three angles
and one side are known). On the other hand, if the two light beams were originally
traveling, say three units per year, the second beam would initially lag
1/3 of a year behind the first as that's how long it would take to do the
ring detour. However, the distancethat the second beam lags behind
the first beam is the same as before. As both beams were traveling the same
speed, the second beam fell behind the first by the length of the detour.
Thus, by measuring the distancethat the second beam lags behind the
first, a distance which will notchange when both light beams slow
down together, we get the true distance from the supernova to its ring. The
lag distance between the two beams, of course, is just their present velocity
multiplied by the difference in their arrival times. With the true distance
of the third leg of our triangle in hand, trigonometry gives us the correct
distance from Earth to the supernova.
Consequently, supernova SN1987A is about 170,000 light-years from us (i.e.
997,800,000,000,000,000 miles) whether or not the speed of light has
Still, the creationist has one ace of a sort remaining. Had the speed of light
slowed down, as often imagined by creationists who have not advanced beyond
Newtonian physics, the distance of SN1987A would still be 170,000 light-years
as indicated above. However, the /time/ that it would take for the light
to reach us need not be anywhere near 170,000 years. We might counter by
arguing that if the speed of light had changed then so would the decay rates
of cobalt-56 and cobalt-57, and since their decay rates have been observed
in SN1987A (and appear normal) that should settle it. After all, in observing
SN1987A we are seeing it as it was in the past. The decay rates of cobalt-56
and cobalt-57 haven't changed, so light hasn't slowed down. (The speed of
light is related to energy by E = mc2. Thus, if the speed of light
could somehow change, then energy would be affected. The end result would
also be a change in the radiometric decay rates.)
Unfortunately, this argument is based on the assumption that we are
observing the correct decay rates of the cobalt on SN1987A. In fact, if the
speed of light had slowed down according to some reasonable curve, we would
be seeing a slow motion replay of reality. The farther away objects are the
greater the slow motion effect. The actual decay rates of the cobalt
in SN1987A would have been much faster than what we observe today by looking
at SN1987A even though we are, in effect, seeing into the past. That is,
we would be seeing a slow motion replay of the decay rates of the two cobalt
isotopes, and those observed rates might just happen to match the
actual decay rates we observe today on earth. It would merely appear
to us that no change had occurred. Does this sound confusing? If it doesn't
then you are ahead of me! I'm still trying to put the pieces together!
To this one might say, "Get an education!" Relativity is central to modern
science and the speed of light is a fundamental constant. Light can't go
faster than about 186,282 miles a second and that's that. One could then
recite volumes of laboratory studies, experiments, and observations to impress
the reader with the power and reliability of special relativity. However,
that approach might seem rather dogmatic to someone lacking a good education
in the sciences. Thus, I will pretend that light once traveled much faster
than today (as might be imagined in Newtonian physics) and show that it still
won't help the young-earth creationist.
Our first argument is based on a straightforward observation of pulsars. Pulsars
put out flashes at such precise intervals and clarity that only the rotation
of a small body can account for it (Chaisson and McMillan, 1993, p.498).
Indeed, the more precise pulsars keep much better time than even the atomic
clocks on Earth! In the mid-1980s a new class of pulsars, called millisecond
pulsars, were discovered which were rotating hundreds of times each second!
When a pulsar, which is a neutron star smaller than Manhattan Island with
a weight problem (about as heavy as our sun), spins that fast it is pretty
close to flying apart. Thus, in observing these millisecond pulsars, we are
not seeing a slow motion replay as that would imply an actual spin rate which
would have destroyed those pulsars. We couldn't observe them spinning that
fast if light was slowing down. Consequently, we can dispense with the claim
that the light coming from SN1987A might have slowed down. Therefore, the
decay rates observed for cobalt-56 and cobalt-57 were the actual decay rates.
A more quantitative argument can also be advanced for those who need the details.
Suppose that light is slowing down according to some exponential decay curve.
An exponential decay curve is one of Mother Nature's favorites. It describes
radioactive decay and a host of other observations. If the speed of light
were really slowing down, then an exponential decay curve would be a reasonable
curve to start our investigation with. Later, we will be able to draw some
general conclusions which apply to almost any curve, including those favored
by creationist Barry Setterfield.
We want the light in our model to start fast enough so that the most distant
objects in the universe, say 10 billion light-years away, will be visible
today. That is, the light must travel 10 billion light-years in the 6000
years which creationists allow for the Earth's age. (A light-year is the
distance a beam of light, traveling at 186,282 miles per second, covers in
one year.) Furthermore, the speed of light must decay at a rate which will
reduce it to its present value after 6000 years. Upon applying these constraints
to all possible exponential decay curves, and after doing a little calculus,
we wind up with two nonlinear equations in two variables. After solving those
equations by computer, we get the following functions for velocity and distance.
The first function gives the velocity of light (light-years per year) t
years after creation (t=0). The second function gives the distance
(light-years) that the first beams of light have traveled since creation (since
1. V(t) = V0 e-Kt , and
2. S(t) = 1010*(1 - e-Kt) , with
V0 = 28,615,783 (the initial velocity for light in units of present
light years per year, equivalent to the factor by which the present speed
of light was larger at creation, t = 0)
K = 0.0028615783 = V0/1010 (the decay rate parameter,
units of inverse years)
With these equations in hand, it can be shown that if light is slowing down
then equal intervals of time in distant space will be seen on Earth as unequal
intervals of time. That's our test for determining if light has slowed down.
But, where can we find a natural, reliable clock in distant space with which
to do the test?
As it turns out, Mother Nature has supplied some of the best clocks around.
They are the pulsars. Pulsars keep time like the Earth does, by rotating
smoothly, only they do it much better because they are much smaller and vastly
heavier. The heavier a spinning top is the less any outside forces can affect
it. Many pulsars rotate hundreds of times per second! And they keep incredibly
precise time. Thus, we can observe how long it takes a pulsar to make 100
rotations and compare that figure to another observation five years later.
Thus, we can put the above creationist model to the test. Of course, in order
to interpret the results properly, we need to have some idea of how much
change to expect according to the above creationist model. That calculation
is our next step.
Let's start by considering a pulsar which is 170,000 light-years away, which
would be as far away as SN1987A. Certainly, we can see pulsars at that distance
easily enough. In our creationist model, due to the initial high velocity
of light, the light now arriving from our pulsar (light beam A) took
about 2149.7 years to reach Earth. At the time light beam A left the
pulsar it was going 487.4686 times the speed of light. The next day (24 hours
after light beam A left the pulsar) light beam B leaves; it
leaves at 487.4648 times the speed of light. As you can see, the velocity
of light has already decayed a small amount. (I shall reserve the expression
"speed of light" for the true speed of light which is about 186,282 miles
per second.) Allowing for the continuing decay in velocity, we can calculate
that light beam Ais 1.336957 light-years ahead of light beam B.
That lead distance is not going to change since both light beams will slow
down together as the velocity of light decays.
When light beam A reaches the Earth, and light is now going its normal
speed, that lead distance translates into 1.336957 years. Thus, the one-day
interval on our pulsar, the actual time between the departures of light beams
A and B, wrongly appears to us as more than a year! Upon looking
at our pulsar, which is 170,000 light-years away, we are not only seeing
2149.7 years into the past but are seeing things occur 488.3 times more slowly
than they really are!
Exactly 5 years after light beam A left the pulsar, light beam Y
departs. It is traveling at 480.5436 times the speed of light. Twenty-four
hours after its departure light beam Z leaves the pulsar. It is traveling
at 480.5398 times the speed of light. Making due allowances for the continual
slowing down of the light, we can calculate that light beam Y has
a lead in distance over light beam Z of 1.318767 light-years. Once
again, when light beam Y reached Earth, when the velocity of light
had become frozen at its present value, that distance translates into years.
Thus, a day on the pulsar, the one defined by light beams Y and Z,
appears in slow motion to us. We see things happening 481.7 times slower
than the rate at which they actually occurred.
Therefore, if the above creationist model is correct, we should see a difference
in time for the above two identical intervals, a difference which amounts
to about 1.3%. Of course, the above calculations could be redone with much
shorter intervals without affecting the 1.3% figure, being that the perceived
slowdown is essentially the same for the smaller intervals within one day.
As a result, an astronomer need only measure the spin of a number of pulsars
over a few years to get definitive results. Pulsars keep such accurate time
that a 1.3% difference--even after hundreds of years--would stand out like
a giant redwood in a Kansas wheat field!
So, what are the results of this definitive test? Many pulsars have been observed
which show nothing remotely close to a 1% change in their rotation rates
over a five year period. Although we have technically disproved only the
above model, we have, nevertheless, thrown a monkey wrench into the machinery
for decaying lightspeed. Every such scenario must have the slow motion effect
described above. Furthermore, the slow motion effect is directly related
to how fast the light is moving. If a model requires light in the past to
move one hundred times faster than observed today, then, at least for some
interval of time measured in that part of space, we would observe things
moving one hundred times as slow.
That's the fatal point which no choice of light-velocity decay curve can wholly
remedy. The creationist model, in order to be useful, must start with a high
velocity for light so that objects ten billion light-years away can
be seen in a universe a mere 6000 years old. Consequently, such a universe
must appear, in general, to be slowing down more and more the further we
look into the depths of space. And the further we look, in general, the more
dramatic the perceived slowdown should be.
It might seem that if we started out with a fantastically high velocity for
light, which then decayed precipitously, we could reduce the problems. Certainly,
that would produce a light-velocity decay curve with near normal velocities
for most of the years between t=0 and t=6000. However, the effect would be
to move the departure time of light beam A (in the above model) closer
to the creation time and to jack up its speed. Thus, the slow motion factor
would be even worse than the model we just examined! On the other extreme,
by abandoning an exponential decay curve, one can get the initial velocity
down to about 1.6 million light-years per year. But alas! The velocity of
light beam A is now 1.6 million light-years per year! We've gone from
the frying pan into the fire.
The problem, from a graphical point of view, is that we have a certain amount
of obligatory area under the velocitytime curve which must be distributed
in some way. That area represents the 10 billion light-years of space which
our initial light beams must cross in 6000 years. No matter where you put
that area, no matter how you poke or shape it, you have a problem.
The big question, then, is whether our general observations of the universe
fit such models. Do we, for example, observe pulsars spinning slower and
slower the further away they are? Do the rotation of galaxies, as determined
from the Doppler effect, grind to a near halt in the more remote regions
of space? Do dust clouds seem to collapse more slowly the farther away they
are? Do the closer novas and supernovas explode, on the whole, more quickly
than the more remote ones? Do galaxies appear to be traveling any slower
the farther away they are? The answer is no.
The alternative, if these light-velocity decay models are to be salvaged,
is that the more distant the object the faster it is moving. Thus, we would
have the illusion of seeing normal rates prevail everywhere, the slow motion
factor being cancelled by objects which are moving, in truth, faster and
faster the further we look into the depths of space. However, there is a
limit to how fast some things can go. Millisecond pulsars are already close
to flying apart. Their spin rates are no illusion! The distant galaxies,
if they were really rotating millions of times faster a few thousand years
ago, would have flown apart. We are led into absurdity. There is no reason,
for example, for believing that the distance of a gas cloud from us dictates
how fast it will collapse! We have no reason to believe that distant galaxies
once traveled millions of times faster than their observed rates. Had they
done so, they would surely have broken out of the great clusters of galaxies
which are bound by gravity. Their distribution today would have been more
or less random.
Light, itself, would have behaved differently at different speeds. The higher
the speed the more blueshifted, the more energetic it would be. Certain wavelengths
of light, for example, have the power to penetrate the galactic dust, thus
allowing us to see events going on in the core of our galaxy. If the wavelength
of such light was merely an illusion produced by the slow motion effect,
if those light waves actually existed at shorter frequencies back then, they
would have been absorbed or scattered differently by the galactic dust. That
is to say, astronomers would not see a logical correspondence between the
wavelengths they observe and their known properties. In the above example,
we might not see the galactic core at all by using the preferred wavelength
for dust penetration! Needless to say, astronomers don't have that problem.
Our conclusion, then, is that any model which would drive us to such views
is bankrupt. We can forget about those claims that light traveled much faster
in the past.
Once it's clear that the light-velocity decay models are bankrupt, not only
with respect to modern science but even within Newtonian physics, then there
is only one reasonable conclusion. The light coming from distant stars and
galaxies have not only traveled immense distances but have spanned ages as
well. In particular, the fact that supernova SN1987A is around 170,000 light-years
distant means that we are seeing an event which is around 170,000 years old.
A few creationists have argued that the universe really isn't that big. In
particular, Slusher, working for the Institute for Creation Research, argued
in 1980 that the universe is based on a Riemannian space which allowed no
point to be more than 15.71 light-years away. The great distances observed
would be an illusion based on mistaking the Riemannian space for Euclidean
This model, however, requires that the distance to supernova SN1987A be measured
at less than 15.71 light-years in contradiction to the 170,000 light-years
actually measured. Unexploded versions of SN1987A would be seen at the same
time, one of them being at a perceived distance of 170,000 light-years! A
few decades later, the light from the explosion would circle around again,
thus causing us to see SN1987A explode all over again! This is madness, not
science! See Strahler (1987, pp.114-116) for a thorough debunking of this
Riemannian space nonsense. (George Friedrich Bernhard Riemann, 1826-1866,
was a German mathematician whose work on curved space proved helpful to Einstein,
but not with the absurd radius of curvature assigned by Slusher!)
Yet another idea, advanced by Henry Morris and others, is that star light
was created /in situ/ during the Genesis creation week. However, we have
now left the realm of science for theology. There is no scientific way to
separate star light from its origin in a star. Not only is it theology, but
it's bad theology. God creates a universe which forces him to be a deceiver!
It goes beyond the need for any reasonable appearance of age as a result
of functionality. There is no need, for example, to see supernovae explode
before their time. An observer would ultimately see the supernova leap back
together and explode all over again when the light from the real explosion
finally arrived! It makes God out to be an idiot.
When the smoke blown about finally drifts away and the debate hall falls silent,
the young-earth creationist finds himself back on square one. He is looking
at stars many millions of light-years away, stars putting out light which
takes many millions of years to reach us! Attempts to speed up the
velocity of light or to shrink down the universe have come to naught. What
does remain is the old age of our universe.
Last updated: Friday, 13-Sep-2002 10:47:58
Copyright © 1995 Dave E. Matson.
Originally appeared here.