Determination of the Magnetic Field of Earth and the Elementary Charge

Damian Boh
8
th
December 2012

Abstract
The horizontal and vertical components of the Earths magnetic field were measured. This was done using the
maximum deflection y
max
of a current carrying wire in the field. Experimental data suggest a positive linear
relationship between the horizontal y
max
and current I while an additional quadratic relationship was observed for the
vertical y
max
and I. These effects were analysed and the total magnitude of the Earths magnetic field was calculated to
be 3436T.

Millikans oil droplet experiment was also reproduced to determine the value of the elementary electric charge. The
terminal velocities of charged oil droplets in known electric fields between a parallel plate capacitor were measured.
These were analysed using fluid dynamic forces to give the radius a and charge q of the droplets. The deviation of the
data from Stokes law was also investigated and this was found to be little. The most accurate value of q was
determined to be (1.80.2)10
-19
C

and this was compared to the known value.

1. Introduction
A current carrying wire will experience a force and deflect in a magnetic field. The aim of the first experiment is to
determine the vertical magnetic field B
z
of the Earth and horizontal field B
x
using the maximum horizontal and vertical
deflections of a fine, current carrying wire
.
respectively. Experimental data suggest a positive linear relationship
between the horizontal y
max
and current I while an additional quadratic relationship was observed for the vertical y
max
and I. The linear relationship was expected and will be discussed in Section 2. The quadratic relationship was
interpreted to be due to the sagging of the wire as it expands under the power dissipated as heat in the wire. Further
complications arose when the tension caused by sagging of the wire built up and overcame the tension of a pulley
connected to one of the wires ends. These effects were analysed and factored out before the Earths magnetic field
was calculated to be 3436T.

The measurement of the elementary charge on an electron was achieved by Robert Millikan in 1909 in his famous oil
droplet experiment[1]. This was reproduced in the second experiment. The motions of charged oil droplets within an
electric field between a parallel plate capacitor were analysed. The terminal velocity v
g
at free fall under gravity and
the terminal velocity v
E
as it rises under both the influence of gravity and an oppositely directed electric field E were
measured. These were analysed using fluid dynamic forces to give the radius a and charge q of the droplets. Results
were plotted on scatter diagrams of charge against radius and charge against v
g
-1/2
. Experimental data suggest a general
positive correlation between the magnitude of charges and radii of the droplets. Clustering of charges were also
observed around different radii suggesting that the charge is a discrete quantity. Stokes law was used in this
experiment and the deviation of the observed q from its prediction was studied and found to be little. The most
accurate value of q was determined to be (1.80.2)10
-19
C

and this was compared to the known value.

2.1 Experimental Method (Earths Magnetic Field)
In the first experiment a tungsten wire was made to pass over fixed pulleys mounted onto an aluminium optical rail at
both ends (Figure 4, page 62). It was held under tension by a small weight T
0
at one end of the wire below the left
pulley while the right end of the wire was fixed to the insulating right pulley stand. An ammeter and power source was
connected to the wire as shown in Figure 5 of page 63 to produce a current flowing towards the right direction. The
positions of the left and right ends of the wire were labeled as x=-L/2 and L/2 respectively, with that of its centre as
x=0. A microscope with a graticule scale was mounted at the centre of the wire to observe for its deflection as the
deflection is very minute. The apparatus was arranged in an East-West direction to maximize the vertical deflections
of the wire. The wire was very sensitive to movement in the lab and the wind in that it vibrated very vigorously when
seen under the microscope when there was movement and wind. Hence such movement and air flow were minimized
and the deflection reading was only recorded when the wire vibration was minimal.

According to theory it should deflect in an arc with the maximum deflection y
max
at its centre when a current passes
through it[2]. Figures 2 and 3 on page 61 show how different forces are balanced at a position x along the wire.
Clearly, for the wire to be in equilibrium, the tension T(x) should balance the magnetic force F
x
in the negative x-
direction and T
0
in the negative y-direction. For small deflections the total magnetic force from x=0 to x is just F
x

=IB
z
x. Balancing the forces gives T(x)sin= IB
z
x [Equation 1] and T(x)cos=T
0
[Equation 2]
.
Diving equation 3 by
equation 4 gives
0
tan
T
x IB
dx
dy
z
= = u
[Equation 3] and integrating gives
I
T
x B
y
z
0
2
max
2
=
[Equation 4] where y
max
is the
maximum horizontal deflection at the middle of the wire. Equation 4 predicts a positive linear trend between y
max
and I.
Hence in this experiment, measurements of y
max
at different I were taken and a graph of y
max
against I was plotted. The
vertical magnetic field B
z
was determined from its slope. The horizontal magnetic field B
x
causes a vertical deflection
z
max
in the same way. Hence a graph of z
max
against I was plotted and analyzed as well.

2.2 Experimental Method (Elementary Electrical Charge)
In the second experiment Milikans oil droplet experiment was reproduced with an experimental set-up shown in
Figure 13 on page 88. A power source whose voltage can be varied was connected to a circular parallel plate capacitor
surrounded by a Perspex air shield. The electric field E between the parallel plates can be adjusted by varying the
power source voltage. This field was directed downwards so that it created a force opposite to the weight of the oil
droplets. A nebulizer containing oil was connected to the set-up via holes in the air shield.

Oil droplets negatively charged by friction were introduced in between the parallel plates when the nebulizer bulb is
squeezed. Their motions were observed using a microscope connected to the front of the Perspex air shield via a hole.
The terminal velocities v
g
and v
E
as mentioned in Section 1 was determined by measuring the time taken for the oil to
travel a distance of a few deflections on the microscope graticule scale under different E. The voltage V
0
required to
exactly cancel gravity and bring the droplet to rest can also be recorded to serve as a cross check for the charge on
each droplet q.

From theory, the terminal velocity depends on the viscous drag of the air given by Stokes Law[3] to be F=6av
[Equation 5] where is the viscosity of air and a is the radius of a spherical droplet. As v
E
is the terminal velocity
where viscous drag balances the combined influences of the electric force and weight of droplet, qE-mg=6av
E

[Equation 6] where m is the mass of a droplet and g is the gravitational constant. Similarly v
g
is measured when E=0,
hence the weight of droplet is mg=6av
g
[Equation 7]. The mass m can be determined from the droplets radius a and
density by m=(4/3)a
3
[Equation 8]. Further manipulations to these equations will give the droplet radius as
g
v
a
g

q
2
9
=
[Equation 9] and the observed charge on each droplet as
E
v v v
g
q
E g g
obs
) (
2
9
6
2 / 1 2 / 3
+
=
q

t
[Equation 10]. V
0
can be used to cross check this value of q using the equation mg=qE [Equation 11] when the electric force exactly
balances the weight to bring the droplet to rest. In this case E is V
0
/d where d is the distance between the parallel plates
of capacitor.

Stokes law is actually based on a continuous medium and it breaks down when the molecular mean free path l of air
is comparable with the droplet radius a. Hence it is important not to use the results of droplets that have too small a
values. This will be further discussed in Section 3. On the other hand, droplets that have high a will fall quickly
resulting in v
g
being more difficult to measure accurately. Thus it is important reach a compromise between both of
the effects where the a values of droplets measured are not too high or low.

3.1 Results and Discussion (Earths Magnetic Field)
A graph of the measured y
max
against I was plotted by Orgin in Graph 2 of page 71. As predicted by Equation 4,
experimental data suggest a positive linear trend between the 2 quantities. Equation 4 suggests that the slope of this
graph should be B
z
x
2
/2T
0
where x=L/2 as in Section 2. x=L/2 was measured to be 0.970.02 m while T
0
(weight of
mass)

was measured to be 0.1330.001N. The value of the vertical component of Earths magnetic field B
z
was hence
calculated as
T
x
slope T
B
z
1 32
2
2
0
~

=
. The errors in this result include errors from the measurement of L/2 with
a metre rule and also the measurement of T
0
with a mass balance. The main source of error comes from the vibration
of the wire as described in Section 2.

It is interesting to note that the slope of the graph is not entirely uniform. This is due to the power P=I
2
R being
dissipated as heat in the wire, causing the wire to expand and sag. Though the sagging is in the vertical direction and
should not be observed in y
max
, this affects the T
0
which affects the slope of the graph
.
As current I was increased in the
wire, the heat dissipated increased and the wire expanded more. This led to an increase in the tension in addition to the
original T
0
in the wire. From Equation 4, an increase in T
0
due to an increase in I will decrease the slope of the graph.

All these are due to the fact that the left pulley is not ideal and frictionless. If it were ideal, any increase in length of
the wire caused by the expansion would slip off to the left of the pulley and the tension would always remain at T
0.
Further complications arose when the increased in tension in the wire was large enough to overcome the static friction
at the pulley when I was large enough. In this case the pulley slipped (Figure 10, page 73) and the tension became T
0

again and built up until the additional tension was high enough to overcome the friction yet again. The change in slope
of the graph would thus be inconsistent at these values of I. Hence the slope in the graph is not uniform.

The maximum vertical deflection z
max
was also plotted against I in Graph 4 on page 80 using Origin. The thermal
sagging of the wire, being in the vertical direction, affected the results to a huge extent. As mentioned earlier, the heat
causing the expansion depends on the power P=I
2
R and hence a quadratic effect (dependence on I
2
) is expected. This,
in addition to the magnetic deflection of the wire, will lead to a linear and quadratic I dependence in the form z
max
=b
1
I
+ b
2
I
2
[Equation 12] (refer to Figure 12, page 77). The constant b
1
should correspond to B
x
x
2
/2T
0
like how the slope of
Graph 2 corresponds to B
z
x
2
/2T
0
.

The graph was observed to be divided into 3 regions of quadratic shapes. The quadratic shapes suggest that the
sagging effect outweighs the magnetic effect. Furthermore the fact the graph is divided into 3 regions arises from
exactly the same pulley slipping effect as described earlier. As the pulley slips at high enough I (for both positive and
negative values), the sagging and hence z
max
will suddenly decrease and there will be discontinuities for z
max
values in
the graph at the values of I where the slipping occurs. Equation 12 was thus fitted onto the middle region of Graph 4
using Origin before the slipping occurs (before any discontinuities) and b
1
was used to determine B
x.
B
x
was calculated
to be
T
x
b T
B
x
7 13
2
2
1 0
~

=
. The errors involved were similar to those involved in the determination of B
z
, i.e. the
measurements of L/2 and T
0.
The vibrations of the wire in the vertical plane were much higher than that in the
horizontal plane, leading to a much larger fractional error in the B
z
value.

The overall magnetic field of the earth was calculated to be B =B
x
2
+B
z
2
=3436T with the large error mainly from
B
x
. The direction of the field with respect to the horizontal plane was calculated to be = tan
-1
(B
z
/B
x
) =1.20.3 rad
=7020. The reference value of the magnetic field in London is 48T

[4] and this is within 1 from our results.
However we would expect the magnetic field in Blackett laboratory to differ from the Earths magnetic field due to
metal structures within the building. The accuracy and precision of the experiment could also be further improved
with proper shielding of the whole tungsten wire from any air movement in the lab. The convectional currents caused
by heating of the air due to the light source being placed closed to the wire could have affected the results of this
experiment as well.

3.2 Results and Discussion (Elementary Electrical Charge)
The scatter diagram of charge q
obs
(calculated from Equation 10) of each droplet against the droplet radius a was
plotted in Graph 5 on page 91 by Origin. Clearly, the data suggest a positive correlation between q
obs
and a. This is not
surprising as bigger droplets are more likely to be more highly charged by friction when squeezed through the
nebuliser since they have bigger surface areas. Clustering of q
obs
around certain q
obs
values and radii can also be seen.
This shows that electrical charges have discrete values which corresponds to theory.

It is important to take note of the limitations of Stokes law in this experiment. Equation 1 works on the assumption
that the air causing viscous drag on the oil droplet is a continuous medium. However, when the radius of the droplet a
is comparable to the molecular mean free path l of air, Stokes law breaks down. Hence the value of q
obs
will be the
true value of the charge only when l/a tends to 0. As shown in Graph 5, the value of a for the droplets used ranges
from 0.4 m to more than 1m. The reference value of l is 0.06635m which is significantly smaller than the a values.
Hence it is reasonable to assume a small value of l/a in this experiment.

This allows Equation 10 to be expressed as a power series in l/a in which the first order approximation to the
correction of Stokes law would be enough. This first approximation would hence be linear in the form of
q
obs
= A(l/a) + B [Equation 13] where A and B are constants and A(l/a) is the additional term caused by the deviations
from Stokes law. The value of B would therefore be the true value of q as l/a tends to 0 and Stokes law becomes
perfectly valid. We can conclude from Equation 9 that l/a is proportional to v
g
-1/2
. Hence a graph of q
obs
against v
g
-1/2

could be plotted with its y-intercept corresponding to B in Equation 13.

The graph of q
obs
against v
g
-1/2
is plotted in Graph 6 of page 9. The clustering of q
obs
values around the lowest value
corresponds to oil droplets being charged with 1e, where e is the elementary electric charge. Similarly the clustering
that occurs at the second lowest value corresponds to those charged with 2e. Droplets that have charges higher than 2e
were not used in this experiment for data analysis. A straight line was fitted onto the 1e cluster by Origin and the true
value of q was obtained from its intercept to be q
a
=(1.90.1)10
-19
C. This value is higher than the known value of
e = 1.60210
-19
C and furthermore the known value does not even lie within 1 of q
a
. As the random error would
already have been minimised by using a lot of data points in the 1e cluster, this suggest some form of systematic error
which may be present causing all the measured values of q to be higher than the actual values.

Hence a second method might be more useful. This method concerns the determination of the value of 2e from the
intercept of a line fitted to the 2e cluster data points. This value would then be subtracted from q
a
to obtain a value for
1e. Such a method of differences is useful in eliminating and checking for any systematic errors and might prove to be
more accurate. The value of the intercept for the line fitted to 2e cluster is shown in Graph 6 to be
B = (3.70.2)10
-19
C. The value of the elementary charge found from this method is thus obtained to be q
b
=B q
a
=
(1.80.2)10
-19
C. This value has a higher error as that of q
a
as the error from both intercepts are added up in
quadrature. Also, the known value of e lies within 1 of q
b
and this method is thus concluded to be more accurate.

The fact that this method is more accurate suggests that a systematic error is indeed present. This can be due to any of
the factors in Equation 10. For example the value of viscosity of air might have been overestimated resulting in too
high values of q
obs
calculated for every droplet. Similarly, the density of oil or the electric field E might have been
underestimated. Other sources of systematic errors include that due to an ambient E field being present in the lab as a
result of the electric power cables in the laboratory.

An even more accurate correction to Stokes law was discovered by Millikan[5] who suggested that Equation 5 should
be replaced by
a bl
v a
F
/ 1
6
+
=
q t
[Equation 14] where b is a constant. This results in the correction q
obs
2/3
= q
2/3
(1+bl/a)
[Equation 15] where q is the true value of the charge. Again noting that l/a is proportional to v
g
-1/2
, a graph of q
obs
2/3

was plotted against v
g
-1/2
in Graph 7 of page 96 and its intercept could be used to determine q
2/3
and hence q.

Similar methods as those described earlier were used to determine 2 further values for the elementary charge, q
c
and q
d
.
The intercept from the 1e cluster in this graph was used to determine q
c
to be q
c
=(1.90.1)10
-19
C. Similarly q
d
was
found by the difference from the calculated 2e and 1e to be q
d
=(1.80.2)10
-19
C. q
c
and q
d
has exactly the same
values and errors as q
a
and q
b
determined by their corresponding methods respectively. This suggests that the more
accurate correction to Stokes law is not necessary in the context of this experiment as the first approximation is good
enough. This further confirms the assumption that l/a is small is justified for the less accurate method to work.
Furthermore the slopes of the lines in graphs 6 and 7 correspond to the magnitude of the correction terms to Stokes
law. The fact that these slopes are of the orders of magnitudes 10
-23
and 10
-16
further suggests that the correction to
Stokes law is very little.

The most accurate measurement of e was taken to be that of q
d
, hence e =q
d
=(1.80.2)10
-19
C
.
The theoretical value
of e is within 1 of this measured value. q
d
was taken to be the most accurate measurement as it was determined by the
method of differences which eliminates systematic errors involved and it was also calculated using the more accurate
method that Millikan came up with. The errors in this measurement were mainly from the human reaction time in
measuring the time taken for the oil droplet to travel a particular distance and also the determination of the position of
the droplet with respect to the microscope graticule scale. The scale has divisions very close together and it was
difficult determining the positions of the droplets. Moreover the fact that q
d
is higher than the theoretical value
suggests that systematic errors were present. These could be due to ambient electric field in the lab as mentioned
earlier and also the Brownian motion of air particles and the wind affecting the motion of the air droplets.

Millikans approach to this experiment was somewhat similar in his paper[5]. He, too, selected droplets that were not
too small or big so that Stokes law did not deviate too much and the droplets did not fall too quickly for accurate
measurements. His result for the elementary charge was e=(4.7740.009)10
-10
electrostatic units which corresponds
to (1.5920.004) 10
-19
C. This value was far closer to the known value of 1.602 10
-19
C. However, Millikans
experiment was conducted over 60 days with 58 droplets. Each droplet was numbered and carefully analysed.
Furthermore he considered other factors like pressure of the air around the droplets. He also took great care to keep the
air dust free and to achieve complete stagnancy of the air by absorbing heat rays using a water cell and immersing his
whole experimental vessel into a constant temperature bath[5]. These precautions could have been adopted in the
experiment to achieve a far greater accuracy comparable to that of his original experiment.

4. Conclusion
The Earths magnetic field was measured using the maximum deflections of a current carrying wire. Also, the
elementary charge was measured by subjecting the motion of oil droplets within an electric field to analysis using
fluid dynamical forces. The Earths magnetic field was measured to be 3436T while the elementary charge was
measured to be (1.80.2)10
-19
C. The known values of the field 48T and the elementary charge 1.60210
-19
C fall
within 1 of both the measured values. In each experiment, complications from other factors such as the pulley
slipping effect and the deviation from Stokes law arose. These had to be analysed carefully before they can be
factored out to give an accurate result.

Though not in the main aim of these experiments, they have provided interesting insights into factors like the thermal
sagging effect of a current carrying wire and the accuracy of Stokes law. These suggest that a broad knowledge of
physics outside the field of work of an experimentalist is vital in handling experiments like these. Finally, the accuracy
of the experiments could have been improved by using better apparatus such as those which gives proper shielding of
the wire from any air movement and those used in the original Millikan experiment.

5. References
1. Robert Millikan The Oil-Drop Experiment, http://www.juliantrubin.com/bigten/millikanoildrop.html, last
updated July 2012
2. Imperial College London Second Year Laboratory, Charges and Fields Experiment Manual,
September 2012
3. Stokess Law, http://www.cord.edu/faculty/ulnessd/legacy/fall1998/sonja/stokes.htm, accessed December
2012
4. Magnetometers, http://www.barnardmicrosystems.com/L4E_magnetometer.htm, accessed December 2012
5. R.A. Millikan, On the Elementary Electrical Charge and the Avogadro Constant, Phys. Rev. 2, 109143 (1913)
6. S.G. Jennings, The Mean Free Path in Air, Journal of Aerosol Science, Vol, 19, No. 2, 159-166, 1988