Experimental Determination of The Speed of Light
Experimental Determination of The Speed of Light
Experimental Determination of The Speed of Light
Light
Alex Roberts, February 2016
Abstract
The speed of light in air is calculated using two methods; the first involving
a rotating mirror and building upon the ideas set forth by Lon Foucault in the
mid-19th century; the second involving the use of some more modern techniques
and technology, such as Optical Fibres and an oscilloscope. The value for the
speed of light obtained from the first method is found to be
= 299,792,458 m. s
299,792,458 m. s
[2].
M
R
Figure 1: A diagram demonstrating how a beam of light entering the apparatus
through slit S reflects off R, and after being focused by L onto M, the beam
of light returns through the apparatus to slit S. It is important to note that no
matter what angle the mirror R makes with the lens or spherical mirror, the
beam of light will always return to the exact same point (providing the lens L
focuses the light onto a spot on the spherical mirror M).
If mirror R is now rotated at a very high angular velocity, the situation
changes. In the short time it takes for a single photon to travel from R to M
and back, the rotating mirror R will have been tilted by a small angle, say . A
change in angle of the rotating mirror R by would result in the light bouncing
back towards slit S to be reflected at an angle of 2 from the rotating mirror
R. The photon may be detected at an angle 2 relative to the position of slit
S. This is shown below. Lens L is not pictured here for simplification.
Light entering system
S
M
R1
Light leaving system
S
2
M
R2
Figure 2: A diagram showing the path of a single photon when R rotates counter
clockwise at a speed .
If the distance between the mirrors R and M is h, the time between the
light hitting R for the first and second times is given as:
t=
2h
c
(1.
1)
2 h
c
(1.
2)
c=
2 h
(1.
3)
s ' (B+h)
2 hA
(1.
4)
Where: A is the distance between the lenses L 1 and L2, minus the focal length
of L1; B is the distance between L 2 and R; s is the displacement of the
focused spot of light on the microscope viewer (see Figure 3 below for reference).
may then be eliminated from the equation, giving an equation which
may be used to calculate a value for the speed of light using the Foucault
apparatus:
s' =
4 A h2
c (B+h)
(1.
5)
2. Fibre-Optic Method
Fibre optic cables are a valid modern method for measuring the speed of
light, as they allow a long length of cable to be carried in a small bundle, thus
giving the light a long distance to travel, hence allowing the time of propagation
to be measured, and the speed of light calculated from that.
The theory behind the Fibre-Optic method is much simpler than that of the
Foucault method. The time of propagation for a photon is simply given by:
t=t 0+
nx
c
(2.
1)
Quarterwave
plate
R
BeamL splitter L
s
PM
PL
Lase
r
Microsco
pe
Figure 3: The full version of the modern Foucault apparatus, as used in this
experiment. M is a stationary mirror; R is a mirror rotating at angular velocity
; L1 and L2 are lenses, and PL and PM are polarisers. s is the point where
the beam of light is focused when returning from the apparatus, and s is the
focal length of L1.
The apparatus was first aligned to ensure the laser light travelled correctly
through all pieces of equipment. This involved rotating R by hand to focus the
spot of laser light onto the fixed mirror M. M also required small adjustment to
make sure the photons returned down the path they came, hitting the stationary
R in the centre. It was ascertained for certain that the spot of light at s had
travelled from the laser to M and back by blocking the path of laser light at
several points along the way and observing the disappearance of the spot of
light. The Polarisers PL and PM were crossed so as to attenuate the laser beam,
and the Quarter-Wave Plate moved into position, in the path of the light. The
Microscope was then focused upon the spot of laser light, and the crosshairs
centred upon it. The Micrometre was zeroed, so that any movement of the spot
would be automatically recorded as s as opposed to s 2 (therefore eliminating
the need for later calculations). When the mirror R was rotated, it was observed
that the spot on the microscope was displaced horizontally.
Following the setup of the equipment, errors were calculated for the
1
, bringing
the microscope in and out of focus, and centring the microscope upon the spot
once more. The error in was also measured by keeping the motor running at a
constant speed and seeing how much the value for angular velocity changed
over the course of 40 seconds.
Measurements of how s varied with a change in were then recorded
1
a function of the angular velocity was then produced, and the gradient used to
calculate c [as according to equation (1.5)].
2. Fibre-Optic Method
The apparatus for this method was set up as shown below.
Signal
Generator
Oscillosco
pe
LED
Fibre-Optic
Cable
Photodiode
Figure 4: The equipment setup for the Fibre-Optic Method. Each thin line is a
fibre-optic cable. The Power supply for the LED Transmitter was omitted from this
diagram due to its relative unimportance (it simply supplies power to the LED
light in the transmitter.
After setting up the equipment as detailed in Figure 4, the signal generator
was set to square wave. This was done to reduce the reading error on the
Figure 5: How the signals appeared on the oscilloscope. The lighter vertical lines
shown are the oscilloscope cursors.
The oscilloscope trace was enlarged and two separate cursors aligned with
the start of each signal. The time difference between the two cursors was given
on the screen, and if aligned correctly, this corresponded to the time difference
between the signal which travelled straight from the Signal Generator to the
Oscilloscope, and the signal which travelled through the optical fibre. The fibre
optic cables were of various lengths. Each time an optical fibre was connected,
the new time difference was recorded. A graph of Time Difference against
Cable Length was plotted. The Propagation Delay corresponded to the
intercept, and the slope was used to calculate c [as according to equation (2.1)].
Results
1. Foucault Method (with quarter-wave plate)
The speed of light was calculated using the gradient of a graph which
plotted deflection, s, against rotational speed, . Equation (1.5) was applied,
and a value for c calculated using it. The graph was plotted in Origin, and
hence the gradient of the graph was determined by the built in least-squares
fitting method. The error bars for the error in are too small to be seen here.
Figure 6: The deflection of the small spot of laser light plotted against the
rotational speed of the mirror R.
7
. The error
here was calculated using the built in chi 2 algorithm. Using Equation (1.5) and
this number, the first value for the speed of light was found to be
8
c 1=2.9 E 1 E m. s
The error associated with this value was calculated using partial
differentiation. Letting x be the gradient of the graph, from equation (1.5):
2
c=
8 A h
x (B+ h)
(1.
6)
A correction factor of 2 was used here due to the units of the gradient being
1
m. s . v
as opposed to m. s . ra d
gives:
(1.
7)
(Error in c)2 =
Where: e1 = error in A,
error in h,
0.261 5 E4 m ; e2 = error in B,
0.542 5 E4 m ; e3 =
7
1.89 E m . s . ra d
2. Fibre-Optic Method
In this case the propagation delay, along with the speed of light, was
calculated using a graph of Time Difference versus Cable Length.
Figure 7: A graph showing how Time Difference (t) varies as a function of the
Length of Cable (x). The error bars for all t and x values are too small to be
seen here. The gradient of this graph is
6.1 E 1 E
10
was calculated using the least-squares algorithm, and the standard error was
given by chi2.
The y-intercept on Figure 7 corresponds to the value for t 0, as given by
equation (2.1). The propagation delay for this apparatus, therefore, is:
t 0=7.5 E8 6 E9 s
A value for the speed of light from this graph and equation (2.1) may also
be obtained by dividing the refractive index of the core by the gradient. The
standard error here is simply the sum of the percentage errors of each value. The
speed of light from this method was found to be
c 2=2.5 E 3 E m . s
Discussion
The value for the speed of light obtained from the Foucault method was
c 1=2.9 E8 1 E7 m. s1 , and the value obtained from the Fibre-Optic method was
8
found to be c 2=2.5 E 3 E m . s
boundaries of the quantity procured from the Foucault method, however, the
accepted value is not contained within the uncertainty in the Fibre-Optic result.
1
m. s
, roughly 2E7
too slow.
The error in c1 is roughly 3 times smaller than the error in c 2. This could be
for a multitude of reasons. For example, the error in c 1 was greatly reduced with
the inclusion of the Quarter-Wave plate, and the two polarisers. This is because it
helped focus the microscope more clearly on the spot, resulting in a reduction in
human error.
On the other hand, the error in c 2 could also be bigger for a wide variety of
reasons. For example, incorrectly calibrated equipment, human error in judging
exactly when the square-wave signal started on the oscilloscope, and lining up
the cursors exactly with that point. However, the method used of measuring light
speed through an optical fibre has a large systematic error. The small value for
c in this case was mostly due to the fact that the light was travelling a larger
distance than given by the length x. Using the fact that the numerical aperture
of the optical fibres is 0.47, and the refractive index of the core is 1.492, and the
following equation:
(2.
Numerical Aperture=nsin max
2)
It is realised that max = 18.4 degrees, meaning for each metre of straight optical
fibre, a photon of light may travel 1.05 metres.
Given this information, a larger value of c2 may be calculated, and along
with it a new error. However, the fact that each photon may travel anywhere
between x metres and 1.05*x metres not only leads to increased chance of
human error (the signal will become fuzzier, as photons leaving the LED
Transmitter at the same time stay in the optical fibre for different lengths of time,
possibly emerging at considerably different times; this will result in it being
harder to read the oscilloscope trace accurately), but also results in a larger error
as it is impossible to be sure how far the photon has actually travelled. This
inherent fault with the experiment is the reason the optical-fibre method has
such a large margin for error. If possible, using a Fibre-Optic cable with a much
smaller numerical aperture would reduce the number of possible bounces
inside the cable, resulting in a normal distribution of distance travelled by
photons with a smaller variance than before, meaning less error.
Despite the accurate value obtained in method 1, there are still some
improvements to the experiment which could be made to reduce the error of
3.4%. For example, by far the largest error which could easily be reduced was in
the measurement of h. By increasing the distance h, the standard error would
easily be reduced. Another simple way to reduce the error in the measurement of
c1 by reducing the error in h would be to use a straight ruler rather than plastic
tape (to prohibit bending in the middle) or even a CMM (co-ordinate measuring
machine), which would offer a very high degree of accuracy.
Conclusion
The values obtained were both acceptably close to the true value,
however it is clear that the measurement acquired through the Foucault method
is more accurate, reliable and precise than the value obtained from the FibreOptic method, as shown in the discussion.
Accuracy could be improved in the Foucault method (Foucaults error was
just 0.6% in 1862 [3]) by focusing on reducing the (large) percentage error in h.
This could be done by increasing the distance or using more accurate measuring
devices, such as a CMM.
The Fibre-Optic method was much less accurate, and improvements could
be made in many areas. Ways in which the error could be decreased include
taking the average distance travelled by photons and using that value (as
opposed to the minimum used here), or using an optical fibre with a smaller
numerical aperture, which would both decrease the variance for the distance
travelled by the photons, and reduce the signal noise caused by this
phenomenon.
Considering the improvements listed above and discussed earlier in the
8
close to the accepted value for the speed of light in air as stated in [3], c =
1
299,792,458 m. s
References
[1]Edmond Halley, "A demonstration concerning the motion of light,
communicated from Paris, in the Journal des Scavans, and here made English",
Philosophical Transactions of the Royal Society of London, 12, Pages 89394,
(1677)
[2]Gemma Winter, Level 4 Handbook, Page 56
[3]Stefan Hughes, Catchers of the Light: The Forgotten Lives of the Men and
Women Who First Photographed the Heavens, (2012), III.1 - The c Men.