IPhO 1993 · Theory problem 1

Atmospheric Electricity

General Tabulated Information

Quantity Symbol Value
Earth’s mean radius \(R_E\) \(6.4\times10^6\ \mathrm{m}\).
acceleration due to gravity \(g\) \(9.8\ \mathrm{m\,s^{-2}}\).
Newtonian gravitational constant \(G\) \(6.67\times10^{-11}\ \mathrm{N\,m^2\,kg^{-2}}\).
permittivity of vacuum \(\varepsilon_0\) \(8.85\times10^{-12}\ \mathrm{C^2\,N^{-1}\,m^{-2}}\).
permeability of vacuum \(\mu_0\) \(8.85\times10^{-7}\ \mathrm{N\,A^{-2}}\) [sic; correct value \(4\pi\times10^{-7}\ \mathrm{N\,A^{-2}}\)].
speed of light in vacuum (or air) \(c\) \(3.00\times10^8\ \mathrm{m\,s^{-1}}\).
elementary charge \(e\) \(1.60\times10^{-19}\ \mathrm{C}\).
mass of electron \(m_e\) \(9.11\times10^{-31}\ \mathrm{kg}\).
mass of proton \(m_p\) \(1.67\times10^{-27}\ \mathrm{kg}\)
Planck constant \(h\) \(6.63\times10^{-34}\ \mathrm{J\,s}\).
Avogadro constant \(N_A\) \(6{,}02\times10^{23}\ \mathrm{mol^{-1}}\).
Boltzmann constant \(k\) \(1.38\times10^{-23}\ \mathrm{J\,K^{-1}}\).
molar gas constant \(R\) \(8.31\ \mathrm{J\,mol^{-1}\,K^{-1}}\).

Theoretical Problem 1

ATMOSPHERIC ELECTRICITY

From the standpoint of electrostatics, the surface of the Earth can be considered to be a good conductor. It carries a certain total charge \(Q_0\) and an average surface charge density \(\sigma_0\).

1) Under fair-weather conditions, there is a downward electric field, \(E_0\), at the Earth’s surface equal to about \(150\ \mathrm{V/m}\). Deduce the magnitude of the Earth’s surface charge density and the total charge carried on the Earth’s surface.

2) The magnitude of the downward electric field decreases with height, and is about \(100\ \mathrm{V/m}\) at a height of \(100\ \mathrm{m}\). Calculate the average amount of net charge per \(\mathrm{m^3}\) of the atmosphere between the Earth’s surface and \(100\ \mathrm{m}\) altitude.

3) The net charge density you have calculated in (2) is actually the result of having almost equal numbers of positive and negative singly-charged ions per unit volume (\(n_+\) and \(n_-\)). Near the Earth’s surface, under fair-weather conditions, \(n_+\approx n_-\approx6\times10^8\ \mathrm{m^{-3}}\). These ions move under the action of the vertical electric field. Their speed is proportional to the field strength:

\[ v\approx1.5\times10^{-4}\cdot E, \]

where \(v\) is in \(\mathrm{m/s}\) and \(E\) in \(\mathrm{V/m}\). How long would it take for the motion of the atmospheric ions to neutralize half of the Earth’s surface charge, if no other processes (e.g. lightning) occurred to maintain it?

4) One way of measuring the atmospheric electric field, and hence \(\sigma_0\), is with the system shown in the diagram. A pair of metal quadrants, insulated from ground but connected to each other, are mounted just underneath a grounded uniformly rotating disk with two quadrant-shaped holes cut in it. (In the diagram, the spacing has been exaggerated in order to show the arrangement.) Twice in each revolution the insulated quadrants are completely exposed to the field, and then (\(1/4\) of a period later) are completely shielded from it. Let \(T\) be the period of revolution, and let the inner and outer radii of the insulated quadrants be \(r_1\) and \(r_2\) as shown.

Figure 1

Take \(t=0\) to be an instant when the insulated quadrants are completely shielded.

Obtain expressions that give the total charge \(q(t)\) induced on the upper surface of the insulated quadrants as a function of time between \(t=0\) and \(t=T/2\), and sketch a graph of this variation.

[The effects of the atmospheric ion current can be ignored in this situation.]

(5) The system described in (4) is connected to an amplifier whose input circuit is equivalent to a capacitor \(C\) and a resistor \(R\) in parallel. (You can assume that the capacitance of the quadrant system is negligible compared to \(C\).) Sketch graphs of the form of the voltage difference \(V\) between the points \(M\) and \(N\) as a function of \(t\) during one revolution of the disk, just after it has been set into rotation with period of revolution \(T\), if:

a) \(T=T_a\ll CR\);

b) \(T=T_b\gg CR\).

Figure 2

[Assume that \(C\) and \(R\) have fixed values; only \(T\) changes between situations (a) and (b).]

Obtain an expression for the approximate ratio, \(V_a/V_b\), of the largest values of \(V(t)\) in cases (a) and (b).

6) Assume that \(E_0=150\ \mathrm{V/m}\), \(r_1=1\ \mathrm{cm}\), \(r_2=7\ \mathrm{cm}\), \(C=0.01\ \mathrm{\mu F}\), \(R=20\ \mathrm{M\Omega}\), and suppose the disk is set into rotation at \(50\) revolutions per second.

Approximately, what is the largest value of \(V\) during one revolution in this case?

Official English text of the problem.

Answers

Take \(V=V_M-V_N\). For parts 4–6, define the fully exposed area and the magnitude of its induced charge by

\[ A_m=\frac{\pi}{2}(r_2^2-r_1^2), \qquad Q_m=\varepsilon_0E_0A_m>0. \]

  1. The Earth’s surface charge is negative: \[ \boxed{\sigma_0\simeq-1.3\times10^{-9}\ \mathrm{C\,m^{-2}}}, \qquad \boxed{Q_0\simeq-6.8\times10^5\ \mathrm C}. \] The requested magnitudes are the absolute values.

  2. The average atmospheric charge density is \[ \boxed{\overline{\rho}\simeq+4.4\times10^{-12}\ \mathrm{C\,m^{-3}}}. \]

  3. Counting only positive ions arriving at the surface, and taking their concentration as approximately constant, \[ \boxed{t_{1/2}=\frac{\varepsilon_0\ln2}{e n_+(1.5\times10^{-4})} \simeq4.3\times10^2\ \mathrm s\simeq7.1\ \mathrm{min}}. \]

  4. The induced upper-surface charge is \[ \boxed{ q(t)= \begin{cases} -Q_m\,\dfrac{4t}{T},&0\le t\le T/4,\\[5pt] -Q_m\left(2-\dfrac{4t}{T}\right),&T/4\le t\le T/2. \end{cases}} \] Its graph joins \((0,0)\), \((T/4,-Q_m)\), and \((T/2,0)\) by straight lines and repeats every \(T/2\).

  5. For the first revolution, starting with \(V(0)=0\): - (a) \(V\simeq-q/C\): two positive triangular pulses, each of height \[ \boxed{V_a\simeq\frac{Q_m}{C}}. \] - (b) \(V\) alternates between approximately \(+V_b\) and \(-V_b\), with exponential transitions of time constant \(RC\), where \[ \boxed{V_b\simeq\frac{4RQ_m}{T_b}}. \] Thus \[ \boxed{\frac{V_a}{V_b}\simeq\frac{T_b}{4RC}}. \]

  6. The largest voltage during the first revolution is \[ \boxed{V_{\max}\simeq1.0\times10^{-3}\ \mathrm V=1.0\ \mathrm{mV}}. \] Including the finite resistance gives \(0.989\ \mathrm{mV}\), reached at \(t=T/4=5.0\ \mathrm{ms}\).

Solution

1) Earth’s surface charge

The electric field inside a conductor in electrostatic equilibrium is zero. Gauss’s law applied to a small pillbox across its surface gives

\[ E_{\perp,\mathrm{outside}}=\frac{\sigma_0}{\varepsilon_0}, \]

where the normal points outward from the conductor. Since the field points downward into the Earth,

\[ \sigma_0=-\varepsilon_0E_0. \]

Hence

\[ \sigma_0=-(8.85\times10^{-12})(150) =-1.3275\times10^{-9}\ \mathrm{C\,m^{-2}}. \]

Approximating the Earth as a sphere and using this as its average surface charge density,

\[ Q_0=4\pi R_E^2\sigma_0 =-4\pi(6.4\times10^6)^2(1.3275\times10^{-9}) \simeq-6.8\times10^5\ \mathrm C. \]

Thus the surface charge density has magnitude approximately \(1.3\ \mathrm{nC\,m^{-2}}\), and the total charge has magnitude approximately \(6.8\times10^5\ \mathrm C\).

2) Atmospheric space charge

Let \(z\) increase upward. The signed vertical fields at the two heights are

\[ E_z(0)=-150\ \mathrm{V/m}, \qquad E_z(100\ \mathrm m)=-100\ \mathrm{V/m}. \]

Over a height of \(100\ \mathrm m\), the Earth’s curvature is negligible because \(100/R_E\ll1\). Gauss’s law for a vertical column of horizontal cross-sectional area \(A\) therefore gives

\[ \overline{\rho}Ah =\varepsilon_0A\bigl[E_z(h)-E_z(0)\bigr]. \]

Consequently,

\[ \overline{\rho} =\varepsilon_0\frac{-100-(-150)}{100} =(8.85\times10^{-12})(0.50) =4.425\times10^{-12}\ \mathrm{C\,m^{-3}}. \]

The charge is positive, consistent with the downward field becoming weaker as height increases.

3) Time to neutralize half the surface charge

The Earth is negatively charged. Positive ions drift downward and can reach its surface; negative ions drift upward. Thus the surface-neutralization current contains the contribution of the positive ions only.

Write the ion mobility as

\[ \mu=1.5\times10^{-4}\ \mathrm{m^2\,V^{-1}\,s^{-1}}, \qquad v=\mu E. \]

Taking the supplied near-surface concentration \(n_+\) to remain approximately constant in this drift estimate, the positive charge arriving per unit area per unit time is

\[ j_+=e n_+v=e n_+\mu E. \]

Let \(S(t)=|\sigma(t)|\) denote the magnitude of the remaining negative surface charge density. The conductor boundary condition gives

\[ E(t)=\frac{S(t)}{\varepsilon_0}. \]

Since each arriving positive ion reduces \(S\),

\[ \frac{dS}{dt} =-j_+ =-\frac{e n_+\mu}{\varepsilon_0}S. \]

The discharge is therefore exponential:

\[ S(t)=S(0)e^{-t/\tau_+}, \qquad \tau_+=\frac{\varepsilon_0}{e n_+\mu}. \]

Numerically,

\[ e n_+\mu =(1.60\times10^{-19})(6\times10^8)(1.5\times10^{-4}) =1.44\times10^{-14}\ \mathrm{S\,m^{-1}}, \]

and

\[ \tau_+=\frac{8.85\times10^{-12}}{1.44\times10^{-14}} \simeq615\ \mathrm s. \]

Setting \(S(t_{1/2})=S(0)/2\),

\[ \boxed{t_{1/2}=\tau_+\ln2\simeq426\ \mathrm s\simeq7.1\ \mathrm{min}}. \]

Adding the conductivity of the upward-moving negative ions would incorrectly count a current that does not enter the Earth’s surface.

4) Induced charge on the quadrants

Neglect fringing at the edges and treat the shielding as complete wherever the grounded disk covers the quadrants. On an exposed upper surface, the downward field requires surface charge density

\[ \sigma_{\mathrm{upper}}=-\varepsilon_0E_0. \]

The total area of the two annular quadrants is

\[ A_m=2\left[\frac{\pi}{4}(r_2^2-r_1^2)\right] =\frac{\pi}{2}(r_2^2-r_1^2). \]

The disk’s angular speed is

\[ \omega=\frac{2\pi}{T}. \]

During the first quarter-revolution, each opening exposes an angular width \(\theta=\omega t\). The total exposed area is therefore

\[ A(t)=2\left[\frac12(r_2^2-r_1^2)\theta\right] =\frac{2\pi(r_2^2-r_1^2)}{T}t, \qquad 0\le t\le T/4. \]

It reaches \(A_m\) at \(T/4\), then decreases at the same rate to zero at \(T/2\). Thus

\[ A(t)= \begin{cases} A_m\,\dfrac{4t}{T},&0\le t\le T/4,\\[5pt] A_m\left(2-\dfrac{4t}{T}\right),&T/4\le t\le T/2. \end{cases} \]

Since \(q(t)=-\varepsilon_0E_0A(t)\),

\[ \boxed{ q(t)= \begin{cases} -\dfrac{2\pi\varepsilon_0E_0(r_2^2-r_1^2)}{T}\,t, &0\le t\le T/4,\\[7pt] -\dfrac{2\pi\varepsilon_0E_0(r_2^2-r_1^2)}{T} \left(\dfrac T2-t\right), &T/4\le t\le T/2. \end{cases}} \]

The sketch is a downward triangular pulse: it starts at zero, decreases linearly to \(-Q_m\) at \(T/4\), then increases linearly to zero at \(T/2\). There are two identical pulses per revolution.

Here \(q\) is the charge on the upper surface. An isolated quadrant system can acquire this surface charge by redistributing its existing charge.

5) Voltage across the amplifier input

Take \(V=V_M-V_N\), with \(N\) grounded. Before rotation begins, the quadrants are shielded and the resistor has discharged the capacitor, so

\[ q(0)=0,\qquad V(0)=0. \]

Neglecting the quadrant system’s capacitance compared with \(C\), the charge associated with the connected input conductor is \(q+CV\). Positive charge leaves this conductor through the resistor at rate \(V/R\), so charge conservation gives

\[ \frac{d}{dt}(q+CV)=-\frac VR, \]

or

\[ \boxed{C\frac{dV}{dt}+\frac VR=-\frac{dq}{dt}}. \]

As the quadrants become exposed, their upper surfaces acquire negative charge. This leaves positive charge on the capacitor’s upper plate, so the voltage initially rises positive.

From part 4,

\[ -\frac{dq}{dt}= \begin{cases} +I_0,&0<t<T/4,\\ -I_0,&T/4<t<T/2,\\ +I_0,&T/2<t<3T/4,\\ -I_0,&3T/4<t<T, \end{cases} \qquad I_0=\frac{4Q_m}{T}. \]

Thus the circuit is driven by a square-wave current.

For completeness, put \(\tau=RC\), \(B=RI_0\), and \(t_k=kT/4\). On quarter \(k=0,1,2,3\), the voltage is

\[ V(t)=(-1)^kB+ \left[V(t_k)-(-1)^kB\right]e^{-(t-t_k)/\tau}. \]

This expression, together with \(V(0)=0\), specifies the continuous startup waveform.

(a) Fast rotation: \(T_a\ll RC\)

During the first revolution, little charge leaks through \(R\). To leading order,

\[ C\frac{dV}{dt}\simeq-\frac{dq}{dt}. \]

Using the initial conditions,

\[ V(t)\simeq-\frac{q(t)}{C}. \]

The voltage therefore consists of two positive triangular pulses:

\(t/T_a\) \(0\) \(1/4\) \(1/2\) \(3/4\) \(1\)
\(V/(Q_m/C)\), to leading order \(0\) \(1\) \(0\) \(1\) \(0\)

Join successive entries by straight lines. The largest voltage is approximately

\[ \boxed{V_a\simeq\frac{Q_m}{C}}. \]

For finite \(T_a/RC\), the segments are slightly curved, the minima become slightly negative, and the second positive peak is slightly lower than the first. The positive triangles above are the leading approximation for the requested first revolution.

(b) Slow rotation: \(T_b\gg RC\)

Except during short transitions, the capacitor current is small and

\[ \frac VR\simeq-\frac{dq}{dt}. \]

Hence the voltage approaches alternating levels

\[ \boxed{V\simeq\pm V_b,\qquad V_b=\frac{4RQ_m}{T_b}}. \]

The sketch has the following features:

  • It starts at zero and rises exponentially toward \(+V_b\), with time constant \(RC\).
  • After \(t=T_b/4\), it falls exponentially toward \(-V_b\).
  • After \(t=T_b/2\), it rises toward \(+V_b\).
  • After \(t=3T_b/4\), it falls toward \(-V_b\).

Each quarter is long compared with \(RC\), so most of it is an almost horizontal plateau. The voltage remains continuous at the switching times. For example, immediately after the first switch,

\[ V(t)\simeq -V_b+2V_b e^{-(t-T_b/4)/(RC)}, \]

so it crosses zero approximately \(RC\ln2\) after that switch.

The requested ratio is therefore

\[ \boxed{ \frac{V_a}{V_b} \simeq \frac{Q_m/C}{4RQ_m/T_b} =\frac{T_b}{4RC}. } \]

6) Numerical voltage

The fully exposed area is

\[ A_m=\frac{\pi}{2}\left(0.07^2-0.01^2\right) =7.54\times10^{-3}\ \mathrm{m^2}. \]

Thus

\[ Q_m=\varepsilon_0E_0A_m =(8.85\times10^{-12})(150)(7.54\times10^{-3}) \simeq1.00\times10^{-11}\ \mathrm C. \]

The relevant times are

\[ T=\frac1{50}=0.020\ \mathrm s, \]

\[ RC=(20\times10^6)(0.01\times10^{-6}) =0.20\ \mathrm s. \]

In particular, the time to the first peak is only

\[ \frac T4=0.0050\ \mathrm s=0.025\,RC. \]

Consequently, the fast-rotation approximation gives

\[ \boxed{ V_{\max}\simeq\frac{Q_m}{C} =\frac{1.00\times10^{-11}}{1.00\times10^{-8}} =1.0\times10^{-3}\ \mathrm V. } \]

The finite-resistance correction follows from the first-quarter solution:

\[ V_{\max} =\frac{4RQ_m}{T}\left(1-e^{-T/(4RC)}\right) =\frac{Q_m}{C}\frac{1-e^{-0.025}}{0.025} \simeq0.989\ \mathrm{mV}. \]

It occurs at \(t=T/4=5.0\ \mathrm{ms}\); the second positive peak is smaller.

The solution was checked against the official solution.