New features of the Franck-Hertz experiment

Gerald Rapior, Klaus Sengstock, and Valery Baev

Citation: Am. J. Phys. 74, 423 (2006); doi: 10.1119/1.2174033

View online: http://dx.doi.org/10.1119/1.2174033
View Table of Contents: http://ajp.aapt.org/resource/1/AJPIAS/v74/i5
Published by the American Association of Physics Teachers

Additional information on Am. J. Phys.

Journal Homepage: http://ajp.aapt.org/
Journal Information: http://ajp.aapt.org/about/about_the_journal
Top downloads: http://ajp.aapt.org/most_downloaded
Information for Authors: http://ajp.dickinson.edu/Contributors/contGenInfo.html

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
New features of the Franck-Hertz experiment
Gerald Rapior, Klaus Sengstock, and Valery Baeva兲
Institut für Laser-Physik, Universität Hamburg, Luruper Chaussee 149, 22761 Hamburg, Germany
共Received 29 September 2005; accepted 20 January 2006兲
The fundamental properties of the signal structure in Franck-Hertz experiments are analyzed. The
central result is that the spacings between the minima in Franck-Hertz curves are not equidistant but
increase linearly with the number of minima. This increase is especially pronounced at low atomic
pressure. We suggest that the increase of the spacings is caused by the additional acceleration of
electrons over their mean free path after the excitation energy is reached. Our model accurately
estimates the lowest excitation energies of mercury 共4.67 eV兲 and neon 共16.6 eV兲 atoms and the
mean free path of electrons in standard Franck-Hertz experiments. These results contradict the usual
assumption that the spacings between successive minima or maxima are equal. We demonstrate that
a standard Franck-Hertz apparatus can be upgraded to do more advanced experiments. © 2006
American Association of Physics Teachers.
关DOI: 10.1119/1.2174033兴

I. INTRODUCTION collisions of each of the free electrons with atoms. This pro-
cess repeats with the increase of the voltage U2, and several
The Franck-Hertz experiment on electron-mercury colli- current dips can be observed at nearly integer multiples of
sions is one of the key demonstrations of the quantum be- the excitation energy of the atoms in the current-voltage dia-
havior of atoms and provides a direct nonoptical demonstra- gram 共the Franck-Hertz curve兲.
tion of the existence of discrete stationary energy levels in It is generally assumed that all the maxima or minima
atoms. In 1925 Franck and Hertz received the Nobel Prize spacings in Franck-Hertz curves are equal and correspond to
for this work.1,2 It is widely used in undergraduate physics the first excitation energy of atoms.4–8 It is even suggested
teaching laboratories. Usually the experiment is limited to that the magnitude of the lowest excitation energy can be
the determination of the energy required to excite the first calculated by using the mean value of the maxima8 or
energy levels of mercury or neon atoms. minima9 spacings. Depending on the pressure in the tube and
A typical arrangement of the Franck-Hertz experiment the number of measured spacings, these determinations for
with a Hg or Ne tube is shown in Fig. 1. The tube consists of Hg atoms range from 4.8 to 5.1 eV. This result contradicts
an indirectly heated cathode C, two grids G1 and G2 sepa- the expected value10 of 4.67 eV for the lowest excitation in
rated by a distance L, and an anode A. A small voltage U1 Hg atoms, 6 1S0 → 6 3 P0 共see Fig. 2兲. Higher values for the
can be applied between the cathode C and grid G1 to control lowest excitation energy of Hg atoms determined from the
the emission of electrons. The presence of this voltage is not experimental data have usually been identified with the tran-
critical and in some tubes G1 is absent. An accelerating volt- sitions to the second 6 1S0 → 6 3 P1 共4.89 eV兲 or to the third
age U2 is applied between the two grids, where electrons can 6 1S0 → 6 3 P2 共5.46 eV兲 excited levels, which are claimed to
gain enough energy to create inelastic collisions with atoms. be stronger.7
A small retarding voltage U3 is applied between grid G2 and It is not generally realized that the spacings between the
anode A so that an electron that has undergone an inelastic maxima and the minima in the experimental records are not
collision close to G2 has insufficient energy to reach the an- equidistant, although the continuous increase of these spac-
ode. The mercury tube needs to be heated to a temperature ings as a function of the order of the minima or maxima is
between 140 ° C and 200 ° C so that mercury pressure is suf- usually visible.4–9 In this article we present data that demon-
ficiently high. strates the increase of the spacings and introduce a model
When the electron energy is high enough to overcome the explaining this increase. Our model provides accurate deter-
retarding potential U3, they reach the anode and are included minations of the energies of the first excitation levels of mer-
in the measured anode current I. Electrons with an energy cury and neon atoms that compare well with published val-
less than eU3 are unable to reach the anode and are collected ues. In addition, we obtain information on the mean free path
by the grid G2 instead. For small accelerating voltages U2 the of the electrons and on the cross section of inelastic colli-
anode current characteristics of a Franck-Hertz tube are simi- sions of electrons with atoms.
lar to that of a triode. For greater voltages U2, electrons are
accelerated between the grids until they have enough energy II. EXPERIMENTS WITH A MERCURY TUBE
to excite an atom. At this voltage the anode current decreases
and passes through a minimum when almost every electron The Franck-Hertz experiment with Hg atoms has been per-
has suffered an inelastic collision. Subsequently the excited formed with a commercial experimental apparatus.9 It is
atoms return to their ground state by the spontaneous emis- similar to the apparatus shown schematically in Fig. 1 except
sion of a photon. At the voltage corresponding to the current the grid G1 is absent. The distance between the cathode and
dip a light emission in the tube near the second grid can be grid G2 is L = 7 mm. Figure 3 shows a Franck-Hertz curve
observed.3,4 A further increase of U2 leads to the additional with the Hg tube at the temperature T = 170 ° C. The anode
acceleration of electrons until they gain enough energy to current increases and oscillates as the voltage U2 increases
excite an atom again. As a result the anode current passes and shows 12 dips of the anode current. The separation be-
through its second minimum, corresponding to two inelastic tween the 4th and the 12th dip is 39.1 V. The first three dips

423 Am. J. Phys. 74 共5兲, May 2006 http://aapt.org/ajp © 2006 American Association of Physics Teachers 423

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 Fig. 3. Typical Franck-Hertz curve recorded with Hg tube at 170 ° C.

Fig. 1. Schematic diagram of the Franck-Hertz experiment.

servation supports our model, because the mean free path of
the electrons decreases with the atomic density and therefore
are not very well defined and are excluded from the data with the tube temperature.
analysis. The mean spacing is 共39.1 V兲 / 8 = 4.89 V and is
larger than the first excitation energy in mercury, 4.67 eV.
An accurate evaluation of the individual spacings between III. MODEL OF INELASTIC COLLISIONS
the minima as well as between the maxima reveals their sys-
tematic increase. The spacing between the 4th and the 5th Figure 5 shows the motion of an electron between two
minimum is 4.78 eV, whereas the spacing between the 11th grids in a Hg tube in the presence of the accelerating poten-
and the 12th minimum is 5.03 eV. We will show that this tial U2. While it accelerates the electron gains energy and
increase is due to the additional acceleration of electrons collides with mercury atoms. If the electron energy is smaller
over the mean free path after the excitation energy has been than the lowest excitation energy of the mercury atoms, the
reached, but before inelastic collisions with atoms occur. collisions are elastic and the energy loss by the electron is
The observed increase of the spacing between the maxima very small because of the large mass difference between the
and minima varies with the temperature of the Hg tube. colliding particles. If the electron energy reaches the excita-
Figure 4 shows that three spacings in the Franck-Hertz curve tion threshold of Hg atoms, inelastic collisions may occur.
at 145 ° C correspond to 3.25 spacings at 200 ° C. This ob- Before the inelastic collision takes place, an electron must
come close to a mercury atom. The average distance that an
electron moves before the inelastic collision takes place is
the mean free path ␭. The electrons continue to gain energy
over a distance equal to the mean free path and can excite not
only the lowest but also one of the higher energy states of the

Fig. 4. Franck-Hertz curves recorded with Hg atoms at two different tube

temperatures. The curves are shifted horizontally so that two of the maxima
Fig. 2. Lowest energy levels in Hg 共Ref. 10兲. coincide.

424 Am. J. Phys., Vol. 74, No. 5, May 2006 Rapior, Sengstock, and Baev 424

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 that there are many other energy states in the atom above Ea
that can be excited. This assumption is justified for both Hg
and Ne atoms.
Figure 6共b兲 illustrates how an electron gains energy at a
higher accelerating potential in comparison to Fig. 6共a兲. Be-
cause the electric field is higher, electrons gain more energy
共␦2 ⬎ ␦1兲 along the mean free path ␭. Electrons inelastically
collide twice with atoms and their total energy gained in the
electric field between two grids is E2 = 2Ea + 2␦2. This case
corresponds to the second minimum in a Franck-Hertz curve.
For n inelastic collisions the energy gained by the elec-
trons is
En = n共Ea + ␦n兲. 共1兲
At typical tube pressures, the mean free path of the electrons
is much less than the distance between two grids, ␭ Ⰶ L. With
Fig. 5. Schematic of the energy transfer from electrons to atoms. this assumption we have ␦n Ⰶ Ea and
␭
␦n = n Ea . 共2兲
L
atoms. This phenomenon significantly modifies the Franck- If we use Eqs. 共1兲 and 共2兲, the spacing between two minima
Hertz curves and has to be taken into account when analyz- in a Franck-Hertz curve is given by


ing the experimental data. ␭
Figure 6共a兲 shows the energy gain of a free electron mov- ⌬E共n兲 = En − En−1 = 1 + 共2n − 1兲 Ea . 共3兲
ing in the tube between the two grids. The potential U2 is set L
slightly above the first excited state of the atoms so that Equation 共3兲 shows that the spacing ⌬E共n兲 between the
electrons, after reaching the lowest excitation energy Ea, minima increases linearly with the minimum order n. The
move an additional distance ␭ before they reach the second lowest excitation energy Ea derived from Eq. 共3兲 is
grid G2. Over this distance the electrons gain the additional
energy ␦1 and with a high probability inelastically collide Ea = ⌬E共0.5兲. 共4兲
with atoms. We assume that an electron loses most of its This value corresponds to the minima spacing ⌬E共n兲 ex-
energy after an inelastic collision, corresponding to the idea
trapolated to n = 0.5. As a consequence the lowest excitation
energy of atoms cannot be directly measured from Franck-
Hertz curves as is usually suggested, because this energy is
smaller than the first measured spacing at n = 2, that is, be-
tween the first and the second minimum. 共The spacing for
n = 1 is usually not evaluated because it depends on tube
parameters.兲
The mean free path of the electrons can also be derived
from Eq. 共3兲,
L d⌬E共n兲
␭= . 共5兲
2Ea dn
We have assumed that the electric field between the grid G2
and the anode is much stronger than the field between the
two grids. In a typical Franck-Hertz experiment this condi-
tion is satisfied because the distance between the two grids is
usually much larger than the distance between the second
grid G2 and the anode, but in the following we will give an
example where this condition is not satisfied.

IV. THE LOWEST EXCITATION ENERGY

OF Hg ATOMS AND THE MEAN FREE PATH
OF THE ELECTRONS
Our measured results for the Franck-Hertz experiment
with a mercury tube at different temperatures are shown in
Fig. 7. The measured spacings ⌬E between the minima of
Fig. 6. Electron energy between grids G1 and G2 in a Franck-Hertz tube
the Franck-Hertz curves are shown as a function of the mini-
with an accelerating voltage sufficient for one 共a兲 and two 共b兲 inelastic mum order n for four temperatures 145 ° C ⬍ T ⬍ 190 ° C of
collisions. Ea is the lowest excitation energy of atoms, and ␦1 and ␦2 are the tube, together with the linear fits according to Eq. 共3兲.
additional energies gained by the electrons along the mean free path ␭. These results show a linear increase of the spacings with n

425 Am. J. Phys., Vol. 74, No. 5, May 2006 Rapior, Sengstock, and Baev 425

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 Table I. The values of the mean free path ␭ of the electrons in the Hg tube
for different temperatures.

T 共°C兲 145 160 175 190

d⌬E / dn 共V兲 0.091 0.049 0.039 0.02

␭ 共mm兲 0.097 0.052 0.041 0.022

phisticated versions of the Franck-Hertz experiment do not

resolve the lowest excited state 6 3 P0 in mercury.6,12
The determination of the lowest excitation energy of mer-
cury is often performed by measuring the spacings between
the maxima in Franck-Hertz curves. To verify this approach
we determined the values of Ea from the spacings between
Fig. 7. Spacings ⌬E between the minima in the Franck-Hertz curves mea-
sured with a Hg tube at four temperatures as a function of the minimum
the maxima. Figure 8 shows that these values are lower than
order n. The corresponding linear fits 共solid lines兲 according to Eq. 共3兲 are expected and vary with the tube temperature. The reason is
also shown. that the maxima correspond to inelastic collisions of rela-
tively high energy electrons. The distribution of the electron
energy depends on various tube parameters and varies with
for all temperatures according to Eq. 共3兲. We have not ob- the maximum order. Moreover, the accuracy of the measure-
served a decrease of the spacings as expected from the alter- ments of the positions of the maxima is affected by the over-
native model, which takes into account the nonuniform dis- lap of the oscillations and a rapid general increase of the
tribution of the electric field in the experiments.11 The slope anode current. Therefore, it is more appropriate to determine
of the linear fits of ⌬E共n兲 in Fig. 7 decreases with the tem- the spacings between the minima of the current. The minima
perature. This decrease is expected from our model because correspond to inelastic collisions of the electrons possessing
the mean free path of the electrons in the mercury tube de- the most probable energy and hence to the local maxima of
creases with atomic density and therefore with the tube tem- the light emission in the tube.4
perature. The mean free path ␭ of the electrons for inelastic colli-
Figure 7 shows that all the linear fits to the experimental sions in the tube is determined according to Eq. 共5兲 by the
data converge at approximately n = 0.5 共dashed line兲. Accord- slope of the linear fit of ⌬E共n兲. The values corresponding to
ing to Eq. 共4兲 the value of ⌬E共0.5兲 corresponds to the lowest the data in Fig. 7 are given in Table I. The mean free path
excitation energy Ea of the atoms. The values of ⌬E共0.5兲 decreases with the temperature and thus with the atomic den-
obtained from the linear fits of the data for nine temperatures sity N as10
in the range 140 ° C ⬍ T ⬍ 200 ° C are shown in Fig. 8. The 1 k BT
error bars show a lower limit because they only include sta- ␭= = , 共6兲
tistical errors obtained from the linear fits. Within the ex- N␴ p␴
pected accuracy the minima spacings show no dependence where ␴ is the cross section for inelastic collisions, kB is
on the temperature; their mean value 4.65± 0.03 V corre- Boltzmann’s constant, p is the pressure of the mercury vapor,
sponds to the lowest excitation energy, 4.67 eV, in Hg atoms and T is the tube temperature expressed in Kelvin. In the
共6 1S0 → 6 3 P0兲.10 To our knowledge, this experiment is the temperature range from 300 to 500 K the mercury pressure p
first determination of the lowest excitation energy of Hg at- 共in Pa兲 is approximated13
oms made with a standard Franck-Hertz experiment. Note
p = 8.7 ⫻ 10共9−共3110/T兲兲 . 共7兲
that even the energy resolved measurements with more so-
Figure 9 shows the values of the mean free path ␭ deter-

Fig. 8. Lowest excitation energy in Hg-atoms determined with Eq. 共4兲 by

evaluating minima 共filled square兲 or maxima 共쎲兲 spacings as a function of Fig. 9. Dependence of the mean free path ␭ on the tube temperature. The
the tube temperature T. curve represents the fit of the experimental data to Eq. 共6兲.

426 Am. J. Phys., Vol. 74, No. 5, May 2006 Rapior, Sengstock, and Baev 426

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 10. Franck-Hertz curves recorded with a Ne tube at several retarding Fig. 12. Spacings between minima in the Franck-Hertz curve measured with
voltages U3. a Ne tube as a function of the minimum order n. The corresponding linear fit
共solid line兲 according to Eq. 共3兲 is also shown.

mined in the experiment as a function of the temperature.

The curve is the fit of the experimental data using Eqs. 共6兲
and 共7兲. The cross section for inelastic collisions obtained ergy level Ea1 of Ne atoms, but also one of the higher excited
from this fit is ␴ = 共2.1± 0.1兲 ⫻ 10−19 m2. This value agrees states, for example, Ea2. Therefore, the minima in the
with the cross section for the electron excitation of the mer- Franck-Hertz curves are divided into local dips correspond-
cury state 6 3 P0 given in Ref. 7. ing to the energy separation between Ea1 and Ea2. The num-
ber of the dips increases with the minimum order, because
the electrons exciting Ea2 and Ea1 levels have different initial
V. EXPERIMENTS WITH A NEON TUBE conditions for the next inelastic collision. The evidence for
the excitation of Ea2 levels in neon is provided by the visible
We also performed Franck-Hertz experiments with neon.
light emitting zones inside the Franck-Hertz tube in the range
This experimental setup 共Leybold Didactic GmbH, model
from 540 to 744 nm. This emission corresponds to the spon-
555870兲 is similar to the setup shown in Fig. 1. The curves
taneous transition in neon atoms from Ea2 to Ea1 levels. We
measured with a neon tube at different values of the retarding
see that a standard commercial Franck-Hertz experiment al-
potential U3 are shown in Fig. 10. The minima of these
lows us to resolve different energy states in neon, which can
curves reveal a systematic substructure. We explain this sub-
be easily incorporated into the observations by undergradu-
structure by the excitation of additional energy levels of neon
ate students.
atoms above the lowest excited state. Figure 11 shows that
The presence of local dips in Franck-Hertz curves offers
the first 14 excited levels in neon are divided into two
an alternative way of evaluating the mean free path of elec-
groups, Ea1 and Ea2, with a spacing of about ⌬E = 1.7 eV.14
trons. Figure 10 shows that the second local dip in the third
According to our model, electrons gain additional energy
minimum of the Franck-Hertz curves becomes dominant,
over the mean free path and excite not only the lowest en-
which means that the mean free path is large enough for
electrons to gain the additional energy of ⌬E = 1.7 eV. There-
fore, the value of ␭ can be estimated as:

⌬E
␭= L. 共8兲
eU2

With ⌬E = 1.7 eV, U2 = 60 V 共third minimum兲, and L

= 6 mm, the mean free path of the electrons for inelastic
collisions in the neon tube is ␭ = 0.17 mm.
The mean free path of the electrons in the neon tube can
be estimated in the same way as for the mercury tube by
using Eq. 共5兲, assuming a homogeneous distribution of the
lower excited levels in atoms, and neglecting the local dip
substructure in Franck-Hertz curves. The spacing between
the main minima in the Franck-Hertz curves for the neon
atoms is shown in Fig. 12 as a function of n. Similar to the
mercury tube, the spacings increase linearly with n. Accord-
ing to Eq. 共4兲 the energy of the lowest excited level of neon
atoms can be determined from the linear fit of ⌬E共n兲 as
Ea1 = ⌬E共0.5兲 = 16.5± 0.2 eV. This value compares well with
the data in Fig. 11. The slope of the linear fit of ⌬E共n兲 in Fig.
Fig. 11. Selected energy levels of neon 共Ref. 14兲. 12 is d⌬E / dn = 1 V. From this slope and Eq. 共5兲, the mean

427 Am. J. Phys., Vol. 74, No. 5, May 2006 Rapior, Sengstock, and Baev 427

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 We also presented the first accurate determination of the en-
ergy of the first excited levels in mercury and neon using a
standard Franck-Hertz experiment. Franck-Hertz curves ob-
tained with a mercury tube at different temperatures show a
reciprocal dependence of the mean free path of the electrons
on the atomic density and permit us to determine the cross
section of inelastic collisions of electrons with atoms. Our
approach upgrades a typical Franck-Hertz experiment from a
demonstration to an experiment well suited for advanced un-
dergraduate laboratories.

ACKNOWLEDGMENTS
Fig. 13. Electron energy E between the grid G1 and anode A at 共a兲 low and The authors appreciate the support of Phywe Systeme
共b兲 high values of the retarding potential U3. Ea is the lowest excitation GmbH & Co. KG, which supplied a new version of the
energy of the atoms and ␦1 is the additional energy gained by the electron
Franck-Hertz apparatus.
along the mean free path ␭.

a兲
Electronic mail: [email protected]
free path of the electrons in the neon tube is ␭ = 0.18 mm. 1
J. Franck and G. Hertz, “Über Zusammenstöße zwischen Elektronen und
This value is in good agreement with our estimate from Molekülen des Quecksilberdampfes und die Ionisierungsspannung des-
Eq. 共8兲. selben,” Verh. Dtsch. Phys. Ges. 16, 457–467 共1914兲.
2
Nobel Lectures, Physics, 1922–1941 共Elsevier, Amsterdam, 1965兲, pp.
Figure 10 shows that an increase in the retarding potential
98–129.
U3 shifts the Franck-Hertz curves to the right, which means 3
J. S. Huebner, “Comment on the Franck-Hertz experiment,” Am. J. Phys.
that a higher accelerating voltage U2 is required to reach the 44, 302–303 共1976兲.
4
same minima in the Franck-Hertz curve. The reason is that W. Buhr and W. Klein, “Electron impact excitation and UV emission in
the excitation of neon atoms occurs not only in front of the the Franck-Hertz experiment,” Am. J. Phys. 51, 810–814 共1983兲.
5
grid G2, but also behind this grid, especially if the electric D. R. A. McMahon, “Elastic electron-atom collision effects in the
Franck-Hertz experiment,” Am. J. Phys. 51, 1086–1091 共1983兲.
field between G2 and the anode is less than the field between 6
F. H. Liu, “Franck-Hertz experiment with higher excitation level mea-
the two grids, as is shown in Fig. 13共a兲. At high values of U3 surements,” Am. J. Phys. 55, 366–369 共1987兲.
关see Fig. 13共b兲兴 the excitation occurs mostly before the grid 7
G. F. Hanne, “What really happens in the Franck-Hertz experiment with
G2, which corresponds to the approximation of our model. mercury,” Am. J. Phys. 56, 696–700 共1988兲.
8
For many minima, the influence of the retarding potential U3 W. Fedak, D. Bord, C. Smith, D. Gawrych, and K. Linderman, “Automa-
is important for the last excitation only and Eq. 共3兲 becomes tion of the Franck-Hertz experiment and the Tel-X-Ometer x-ray machine
accurate for both cases. using LABVIEW,” Am. J. Phys. 71, 501–506 共2003兲.
9
M. Brai, R. Butt, A. Grünemaier, K. Hermbecker, and O. Schenker,
“Laboratory experiments: Physics,” PHYWE series, LEP 5.1.03. See
VI. SUMMARY 具http://www.fizika.org/skripte/of-prakt/5_1_03.pdf典.
10
H. Haken and H. C. Wolf, The Physics of Atoms and Quanta, 6th ed.
We have discussed a new observation for the Franck-Hertz 共Springer, Heidelberg, 2000兲, p. 305.
experiment for mercury and neon tubes; that is, the spacings 11
F. Sigeneger, R. Winkler, and R. E. Robson, “What really happens with
between the minima in a Franck-Hertz curve increase lin- the electron gas in the famous Franck-Hertz experiment?,” Contrib.
early with the number of minimum. To explain this effect we Plasma Phys. 43, 178–197 共2003兲.
12
have taken into account the additional acceleration of the P. Nicoletopoulos, “Critical potentials of mercury with a Franck-Hertz
tube,” Eur. J. Phys. 23, 533–548 共2002兲.
electrons over their mean free path after the excitation en- 13
A. N. Nesmeyanov, Vapor Pressure of the Chemical Elements, edited by
ergy is reached. The model is consistent with experimental R. Gary 共Elsevier, Amsterdam, 1963兲.
data and allows an accurate estimate of the lowest excitation 14
C. E. Moore, Atomic Energy Levels 共National Bureau of Standards, Wash-
energy of the atoms and the mean free path of the electrons. ington, D.C., 1949兲, Vol. I, p. 76.

428 Am. J. Phys., Vol. 74, No. 5, May 2006 Rapior, Sengstock, and Baev 428

Downloaded 03 Jun 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission