Classical treatment of electron

gas – Drude model

Questions from this chapter
• List the main assumptions of the Drude model for metals. Explain the
momentum relaxation time and electron mean free path.
• How fast do the electrons move in the Drude model, and how does
their speed depend on the temperature?
• Describe the electrical conduction in the Drude model. Where does
the electrical resistance come from?
• When an electric field is applied, how does the additional speed of
the electrons compare to their thermal speed (at room temperature)?
• What is the Hall effect, and what can be measured by it?
• Why don’t metals transmit light? Is this so for all frequencies?
So – quantum mechanical or classical
approach?
As we concluded based on simple estimations:
ü Classical approach holds well for ions at all temperatures (maybe above the
lowest for the lightest elements like H and He);
ü Quantum mechanical approach must be used for the electrons up to
temperatures of thousands K.

But maybe we could to apply classical theory to electrons as well, to describe at

least some phenomena.
This is done in Drude theory, which gives satisfactory qualitative results for
electrical conductivity, Hall effect, optical reflectivity of metals, and
Wiedemann-Franz law.
Assumptions of Drude model
ü The electrons in a solid behave like a classical ideal gas. They
interact with the ions, but do not interact with each other
(independent electron approximation)
ü With the ions, the electrons interact elastically and
instantaneously during collisions
ü No interaction with the ions between the collisions (free
electron approximation)
ü The electrons reach thermal equilibrium with the lattice
Paul Drude 1863-1906 through the collisions with the ions
Main parameters of Drude theory What is the velocity value at RT?

Electron energy at equilibrium in 3D.

vt – thermal velocity of electrons

l – mean free path – an average
distance between the ions
t – relaxation time – time between the
collision

!
𝜏=
"!
Wikipedia
DC Electrical conductivity
v is the drift velocity! Integration over time:

Inserting t:

Estimate drift velocity at 1 V/cm

DC electrical conductivity
E
Amount of particles (electrons) per time unit
Area A
Amount of charge per time unit
Electron density n
Speed v Current density. Remembering:

One obtains:

Electrical conductivity

Carrier mobility
Exercise: are electrons being scattered on
ions?

Compare thermal velocity, electron mobility in, say, silicon ( you measure it in
Laboratory work No 2) or germanium (in Laboratory work No 3), and interatomic
distance in solids.

If not, on what then?

DC electrical conductivity

What would be the temperature dependence of electrical conductivity according to Drude model? Think about
the exercise in the previous slide
176 A. c. PRIOR

length is normally short enough for injection germanium, does it seem likely that an avalanche
effects not to be serious. With the shorter speci- occurred within the body of the specimen. Where a
mens, particularly those under 1 mm in length, a sharp rise in current was observed, this increased

Dependence of mobility on electric field

rise in current with time was sometimes observed much more gradually both with voltage and with
and attributed to injection. With the specimens of time than the characteristic effect observed on the
lower resistivity, valid measurements could still be same apparatus with low-resistivity n-type ger-
made by extrapolating the current back to the manium; the values of the field at which much more

J. Phys. Chem. Solids Pergamon Press 1959. Vol. 12. pp. 175-180. Printed in Great Britain.

THE FIELD-DEPENDENCE OF CARRIER MOBILITY IN

SILICON AND GERMANIUM
A Smoothedtints plcttsdin fig.3
A. C. PRIOR
Royal Radar Establishment, Malvem, England

(Received 3 June 1959)

Abstract-The variation of mobility with electric field has been measured for n- and p-type silicon
and germanium with fields up to lo5 V/cm. For $-type silicon the variation is found to depend on
hole concentration. For the other materials any variation with concentration must be smaller, and
these experiments are inconclusive as to its existence.

1. INTRODUCTION voltage/current characteristic of bar-type speci-

THE variation of carrier drift velocity with electric mens is measured with short pulses, was similar
field has been measured by RYDER(~)for the four to that described by GuNNQ). He used german-
materials n- and p-type germanium and n- and ium of 2-Q-cm resistivity in which the displace-
P-type silicon with fields of the order of 104 V/cm, ment current was negligible in comparison with the
and by GUNN@) on n-type germanium with fields conduction current; most of the present measure-
Why mobility depends on electric field?
up to 7 x 104 V/cm. In the course of some high- ments were, however, made on material of con-
field measurements by the author on high- siderably higher resistivity, and this necessitated
resistivity p-type gold-doped silicon, it appeared the use of pulses of 6 x 10-s set duration-three
that the current/voltage characteristic was more times longer than those used by GUNN. Even then,
nearly ohmic than was indicated by RYDER’Sdata with the higher resistivities, the oscillations re-
3
IOZ IO’ 1s or by the present author’s earlier measurements on sulting from the displacement current and the lead
lower-resistivity material. A systematic series of inductances were not completely damped within
Electric field, V/h measurements using samples of p-type silicon the duration of the pulse, and this substantially re-
FSG. 1. Variation of current density with electric field in 80-Q-cm n-type having various resistivities was therefore under- duced the accuracy of the measurements. This
silicon. taken to investigate the effect of resistivity on high- difficulty placed the upper limit to the resistivity
field mobility. A significant variation was in fact that could be measured at N 500 Q-cm.
found, but corresponding measurements on n-type The range of field strengths was covered by
start of the pulse, but with those of higher re- gradual rises occurred varied for different speci-
silicon and on both n- and p-type germanium using three or four bars of each material with
sistivity, where substantial oscillation occurred, mens, and injection from the end contactsshowedseemsonly a minor variations of doubtful experi- lengths varying from 0.05 to 1.5 cm. The cross-
mental significance. In the course of this work the sections were adjusted to give an effective resist-
this procedure was not possible and the existence more likely explanation. Other limiting factors were
observations for n- and p-type silicon and for p- ance at the field being used of N 500 L2. On the
of injection was difficult to detect. Trouble caused sparking &rough the air between the ends of the
type germanium were extended to higher fields germanium specimens, contacts, ohmic at low
by injection was no worse with the silicon speci- specimen, or simply output limitationsthanof have the previously been reported, and the fields, were made to the whole end faces by solder-
 1)—
mination, however, made use of early values of drift
and for lattice scattering by optical modes:
".
mobility and assumed lattice scattering to have a tem-
perature dependence of T The expression to be =ET '(e ~
(2)
obtained in this paper is significantly diGerent from
the old one because recent mobility values are higher where A, 8, and 0 are constants. Lattice-scattering
at room temperature, and it is known that the tem- mobility has been determined' between 100' and

Dependence of mobility on temperature

300'K and the temperature dependence of mobility
perature dependence of lattice-scattering
departs from the T "law.
" mobility
'" found to be T ' for electrons and T for holes.
These results are shown as plotted points in Fig. I. For
105
comparison the temperature dependence of p, p is also ,
shown in Fig. 1. This has been computed using 0= 520'
„
8 l

s.
~ OPTICAL MODE
PHYSICAL REVIEW as determined by VOLUM Shockley. 4
E 94, NUMBER An absolute 6 value JUNE for p,15, 19S4
2
%
TEMPERATURE
DEPENDENCE
5 ELECTRON-HOLE
li SCATTERING
is not given
Conductivity Hall Effect the
and because in thevalueIntrinsic 8
forRangeisofunknown. Ger-i~anium It is
Z' probable, however, that J.optical
F. J. Moor@ AND P. MAxTA mode scattering is
Bell Telephone Laboratories, Affray Hill, New Jersey
810
srl
comparable with (Received acoustical mode scattering at tem-
February 4, 1954)
I- e peratures as high as 1000'. Therefore, since the curves
Conductivity and Hall effect have been measured in the intrinsic range of germanium from 250' to 1000'K.
0 Fromfor p, p andmobility
lattice-scattering
centration is determined:
p„do
np=3.
not diverge
and conductivity
10X10"T'exp( — with decreasing
below 500 K, a new empirical expression for carrier con-
0.785/kT). An rapidly
estimate is made of the contribution of
temperature, it is also probable that
optical modes to the lattice-scattering mobility. Conductivity from 500' to 1000' and Hall effect from 250'
to 1000' are computed and compared with experiment. Included in the computationoptical modes
are: the empirical
CV

V LATTICE
expression an appreciable
makefor carrier
trostatic interaction of charge carriers, extrapolated empirical lattice to
contribution scattering
concentration modified by the change in intrinsic
theenergy
ionization produced by elec-
scattering
mobility, scattering byover
Z 10' SCATTER IN G the temperature range where lattice-scattering mobility
electron-hole collisions, and an extrapolation of the ratio Hall mobility/conductivity mobility.

"
8
e has been measured. If it is assumed that the departure
4 1. INTRODUCTION
0 / from the T law foundwithfor experiment. Included in the computation are the
electrons is due to an
EXTRAPOLATED LATTICE FFORTS to understand the semiconducting prop- empirical expression for carrier concentration modi6ed
SCATTERING PLUS ~ erties of optical
germanium have mode been contribution,
con6ned for the by the
the changetwo inscattering
intrinsic ionization processes
energy produced
most part to the range Why mobility depends on temperature?
ELECTRON HOLE by electrostatic interaction of charge carriers; extra-
of impurity conductivity;
102
SCATTER IN G are separable.
literature on the intrinsic range is meager. Theory
Lattice-scattering
is polated mobility
lattice-scattering mobility; is and given
scattering by
electron-hole collisions. Hall effect from 250' to 1000'
8 incomplete, but approximately by
enough is known to say that the
number of processes involved in conduction are too is computed and compared with experiment; making use
of — —
6
information
1 obtained in computing conductivity plus
numerous to be separated empirically from IJI =@ac Map 1of the ratio: Hall mobility/conductivity (3)
I
4 measure-
I I
an extrapolation
100 150 200 300 400 600 800 1000 1500 ments of conductivity and Hall effect. This paper, mobility. This ratio has been determined between 100'
TEMPERATURE IN DEGREES KELVIN therefore, is not a complete analysis of the intrinsic
When Eqs. (1), (2), and (3) are combined and measured and 300' and found for holes to be increasing with

FIG, 1.Conductivity mobility es temperature.

range but tries to indicate the relative importance of
some of the conduction processes by comparing experi-
ment with available theory. values of i. and T inserted, the constants
p, and
temperature above the theoretically predicted constant
1.18. Along with the departure fromA the T law, this
effect is expected to disappear at some temperature
"8
Conductivity and Hall effect in the intrinsic range
above 300' and the ratio to decrease to some constant
s
J. Bardeen and W. Shockley, Phys. Rev. 80, '12— 80 (1950). W. Shockley,
of germanium are4 reported for the temperature
250' to 1000' Kelvin. The data are analyzed in two
range
Bell System Tech. J.
value. Since 30,
the actual (1951).
1025 behavior of the ratio has not
been determined, a linear extrapolation of the measured
steps, the results below 500'K being considered Grst. It
is believed that below 500' the conduction mechanism ratio from 300' to 1000'K is used in computing the
contains fewer unknowns and that a reliable expression Hall coe%cient. Computed conductivity and Hall
for carrier concentration can be obtained from con- effect agree well with experiment. This is encouraging
 Classical Hall experiment

On a New Action of the JIa19flet onz Electric Currents.

BY E. H. HALL, Fellow of the Johns Hopkins University.

SOMETIME during the last University year, wlhile I was reading Max-
well's Electricity and Magnetism in connection with Professor Rowland's
lectures, my attention was particularly attracted by the following passage in
Vol. II, p. 144:
"It must be carefully remembered, that the mechanical force which
urges a conductor carrying a current across the lines of magnetic force, acts,
not on the electric current, but on the conductor whiclh carries it. If the
conductor be a rotating disk or a fluid it will move in obedience to this force,
and this motion miiayor may not be accompanied wvitha change of position
of the electric current which it carries. But if the current itself be free to
choose any path through a fixed solid coniductor or a network of wires, theil,
when a constant magnetic force is made to act on the system, the path of the
current through the conductors is not permanently altered, but after certain
transienit phenomenia,called induction currents, have sulsided, the distribu-
tion of the current will be found to be the same as if no magnetic force were https://en.wikipedia.org/wiki/Hall_effect#cite_note-1
in action. The only force which acts on electric currents is electromotive
force, which must be disting,uished froml the mechanical force which is the
subject of this chapter."
This staternent seemed to mneto be contrary to the most natural supposi-
tion in the case considered, taking into account the fact tlhat a wire not bearing
a current is in general not affected by a mag,net and that a wire bearing a
current is affected exactly in proportion to the strengrthof the current, while
the size and, in general, the material of the wire are matters of indifference.
Moreover in explaining the phenomena of statical electricity it is customriary
to say that charged bodies are attracted towvardleach other or the contrary
solely by the attraction or repulsion of the clharges for each otlher.
 ARTICLE https://doi.org/10.1038/s41586-019-1632-2

Photo-Hall experiment Carrier-resolved photo-Hall effect

Oki Gunawan1*, Seong Ryul Pae2, Douglas M. Bishop1, Yudistira Virgus1, Jun Hong Noh4,5, Nam Joong Jeon4, Yun Seog Lee1,3,

W
Xiaoyan Shao1, Teodor Todorov1, David B. Mitzi6,7 & Byungha Shin2*

IE
Majority and minority carrier properties such as type, density and mobility represent fundamental yet difficult to access

EV
parameters governing semiconductor device performance. Obtaining this information simultaneously under light
illumination would unlock many critical parameters vital for optoelectronic devices and solar cells, but realizing this

Mobility of both types of carriers can be goal has remained elusive. Here we demonstrate a carrier-resolved photo-Hall technique that rests on a new equation
relating hole–electron Hall mobility difference (∆µH), Hall coefficient (H) and conductivity (σ): ∆µH = d(σ2H)/dσ. This
discovery, together with advances in a.c.-field Hall measurement using a rotating parallel dipole line system, allows us

PR
determined! to unlock a host of critical parameters for both majority and minority carriers such as recombination lifetime, diffusion
length and recombination coefficient. We successfully apply this technique to a variety of solar absorbers, including high-
performance lead-iodide-based perovskites. By mapping the Hall data against varying light intensities, unprecedented
simultaneous access to these parameters is demonstrated. This information, buried in the photo-Hall measurement1,2, has
been elusive for 140 years since the original discovery of the Hall effect3. Beyond historical significance, the applications

LE
of simultaneous majority/minority carrier measurement are broad, including photovoltaics and other optoelectronic
devices.
RESEARCH ARTICLE

C
The Hall effect is one of the most important characterization tech- a parallel dipole line (PDL) a.c. Hall system (Fig. 1a, b)18. In the classic
niques for electronic materials and has become the basis of funda- Hall measurement without illumination, one can obtain three param-

TI
mental advances in condensed matter physics, such as the integer and eters for majority carriers: (i) the type (P or N), from the sign of Hall
fractional quantum Hall effect4,5. The technique reveals fundamental coefficient H; (ii) the density (nC = r/He); (iii) the Hall mobility (µH =

AR
information about the majority charge carrier, that is, its type (P or σH), where e is the electron charge, nC is the carrier density and r is the
N), density and mobility. In a solar cell, the majority carrier parame- Hall scattering factor. The key challenge in the photo-Hall transport
ters determine the overall device architecture, width of the depletion problem, that is, to extract the majority and minority carrier informa-
region and bulk series resistance. The minority carrier properties, tion, is to solve for three unknowns at a given illumination level: hole
however, determine other key parameters that directly impact over- and electron (drift) mobility (µP, µN) and their photocarrier densities
all device performance, such as recombination lifetime (τ), (minority (∆n, ∆p), which are equal under steady state conditions. Unfortunately,

ED
carrier) diffusion length (LD) and recombination coefficients (kn). we only have two measured quantities: σ and H, as a function of illumi-

W
Unfortunately, the standard Hall measurement only yields majority nation. The key insight in solving this problem is illustrated in Fig. 1c.
carrier information. Studies to collect both majority and minority car- Consider two P-type systems with the same majority carrier density
rier properties for high-performance light absorbing materials have (p0) and mobility (µP) but different minority carrier mobility (µN).

AT
been attempted, but require a wide range of experimental techniques, When these systems are excited with the same photocarrier density,

IE
which typically use different sample configurations and illumination ∆nmax, they will produce different σ–H curves owing to the increasing
levels, thereby presenting additional complications in the analysis6–14 role of the minority carrier to the total conductivity, even though they
(see Supplementary Information sections F and G). start from the same point in the dark. Therefore, the characteristics of
ER

EV
Extraction of reliable carrier information is particularly sought the σ–H curves, specifically the slope (dH/dσ), contains information
Fig. 1 | The carrier-resolved photo Hall measurement. (a) The parallel (c)in
after Theoretical
the study ofcalculation of twohybrid
organic-inorganic systems with thea same
perovskites, familymajority
of about mobility
the detailed mobilities of the two systems. We show that the
dipole line (PDL) photo Hall setup for a complete photo-Hall experiment. (µP) butgarnering
materials different minority
intense mobility
attention due to(µ N)progress
fast under increasing
in achievingillumination,
(drift) mobility difference, ∆µ = µP − µN = ∆µH/r, with ∆µH (the
(b) The rotating PDL magnet system that generates a unidirectional and high performance
which solar cells—with
yield different the current coefficient
conductivity-Hall record power(σ–H)
conver- HallThe
curves. mobility difference) is given as (SI B):
15
single harmonic ac magnetic field at the center (see animation in Video S1). sion efficiency
slope of the (PCE) of 23.3%
σ–H curve —as well
contains the as promising applications
information of ∆µH. 16
EL
for other optoelectronics devices including light emitting diodes d(σ 2H )  d ln H 

PR
∆µ H = = 2 + σH (1)
and photodetectors17. Full understanding of charge transport prop- dσ  d ln σ 
erties of perovskites will help unveil the operating principles of the
high-performance optoelectronic devices, thereby guiding further Note that σ and H are experimentally obtained as a function of var-
C

improvement. ying light intensity or photocarrier density ∆n, but fortuitously the
In this work we present a carrier-resolved photo-Hall (CRPH) meas- ∆n term cancels out of Eq. 1. There are two equivalent expressions
AC

urement technique capable of simultaneously extracting both majority for Eq. 1 where the former and the latter are used for low and high
and minority carrier mobilities, densities and subsequent derivative injection level analysis respectively. The term dlnH/dlnσ has special
E
parameters (τ, LD) as a function of light intensity. This technique rests experimental meaning, as shown for the perovskite example discussed
on two key elements: a new equation that yields the majority/minority later. This equation applies to both carrier types and assumes that the
Hall effect
Hall effect

Must be true for electron to be able

to transit through the structure

Using the definition of current

density:
Hall effect

Hall measurements allow to

determine:
ü Density of carriers
ü Type (sign) of carriers
ü Mobility of carriers
Optical properties of metals
Light is a plane electromagnetic wave

k is a wave-vector

N is a complex refractive index, or it can be expresses as

a complex dielectric function e.

Remembering that:
Optical properties of metals

Motion equation (second Newton law)

We seek solution of this type

After finding second derivative x’’(t), one finds A

Now we need polarization P(t) and dielectric displacement field D(t)

Optical properties of metals

After inserting P(t) from previous slide, one obtains:

Plasma frequency

Let’s see, how the plane wave from previous slide:

Propagate through a media, when (i) w < wp and (ii) w > wp.
Optical properties

www.physics.stackexchange.com M. Fox. Optical properties of solids

Wiedemann-Franz law

Experimentally observed law has been explained very precisely by Drude model
(due to a lucky mistake!)

L is Lorentz number

It will be discussed more later

Drawbacks of Drude model

Since Drude model considers electrons as a classical and non-interacting particles,

it can not be totally precise. It fails to describe a number of effects:
ü Temperature dependence of electrical conductivity
ü Thermal capacity of metals
ü Electrical conductivity of alloys
etc.