Download The Hall Effect: A Comprehensive Guide to Charge Transport in Metals and Semiconductors and more Lecture notes Solid State Physics in PDF only on Docsity!
The Hall Effect
1 Background
In this experiment, the Hall Effect will be used to study some of the physics of charge transport in metal and semiconductor samples.
In 1879 E. H. Hall observed that when an electrical current passes through a sample placed in a magnetic field, a potential proportional to the current and to the magnetic field is developed across the material in a direction perpendicular to both the current and to the magnetic field [1]. This effect is known as the Hall effect, and is the basis of many practical applications and devices such as magnetic field measurements, and position and motion detectors.
With the measurements he made, Hall was able to determine for the first time the sign of charge carriers in a conductor. Even today, Hall effect measurements continue to be a useful technique for characterizing the electrical transport properties of metals and semiconductors. Indeed, the failure of the simple model of metallic conductivity, which we discuss below, to account for many experimental measurements of the Hall effect has been one of the principal motivators leading to a better understanding of electronic properties of materials [5, pp. 58–62].
1.1 The simple theory of the Hall effect
Consider a conducting slab as shown in Fig. 1 with length L in the x direction, width w in the y direction and thickness t in the z direction.
Figure 1: Geometry of fields and sample in Hall effect experiment.
Assume the conductor to have charge carrier of charge q (can be either positive or negative or both, but we take it to be of just one sign here), charge carrier number density n (i.e., number of carriers per unit volume), and charge carrier drift velocity vx when a current Ix flows in the positive x direction. The drift velocity is an average velocity of the charge carriers over the volume of the conductor; each charge carrier may move in a seemingly random way within the conductor, but under the influence of applied fields there will be a net transport of carriers along the length of the conductor. The current Ix is the current density Jx times the cross-sectional area of the conductor
wt. The current density Jx is the charge density nq times the drift velocity vx. In other words
Ix = Jxwt = nqvxwt. (1)
The current Ix is caused by the application of an electric field along the length of the conductor Ex. In the case where the current is directly proportional to the field, we say that the material obeys Ohm’s law, which may be written
Jx = σEx , (2)
where σ is the conductivity of the material in the conductor.
Now assume that the conductor is placed in a magnetic field perpendicular to the plane of the slab. The charge carriers will experience a Lorentz force q~v × B~ that will deflect them toward one side of the slab. The result of this deflection is to cause an accumulation of charges along one side of the slab which creates a transverse electric field Ey that counteracts the force of the magnetic field. (Recall that the force of an electric field on a charge q is q E~.)
When steady state is reached, there will be no net flow of charge in the y direction, since the electrical and magnetic forces on the charge carriers in that direction must be balanced. Assuming these conditions, it is easy to show that
Ey = vxBz , (3)
where Ey is the electric field, called the Hall field, in the y direction and Bz the magnetic field in the z direction.
In an experiment, we measure the potential difference across the sample—the Hall voltage VH — which is related to the Hall field by
VH = −
∫ (^) w
0
Eydy = −Eyw. (4)
Thus, from equations (1), (3) and (4) we obtain
VH = −
( 1 nq
) IxBz t
The term in parenthesis is known as the Hall coefficient:
RH =
nq
It is positive if the charge carriers are positive, and negative if the charge carriers are negative. In practice, the polarity of VH determines the sign of the charge carriers. Note that the SI units of the Hall coefficient are [m^3 /C] or more commonly stated [m^3 /A-s].
Exercise 1 Work through the math to derive Eq. (5). Now consider that an electric current in the positive x direction can be created by positive charges moving positive along the x axis or negative charges moving negative along the x axis. Draw diagrams showing the electric and magnetic forces on the charges to convince yourself that, in spite of this fact, positive charge carriers will produce a positive Hall coefficient and negative charge carriers will produce a negative Hall coefficient.
The angular frequency ωc is known as as the “cyclotron frequency”. It is the frequency of rota- tion of a charge in a magnetic field, and can be taken as a measure of the strength of the field. The combination ωcτ is used to characterize an experimental situation: if the magnetic field is weak and/or the relaxation time short, ωcτ ≪ 1 and our experiment is in the “weak-field limit”; alternately if ωcτ ≫ 1 the experiment is in the “strong-field limit”. A number of materials show strikingly different behavior between the weak- and strong-field limits; aluminum is one.
In our classical model of the Hall effect with a single type of charge carrier, however, there is no such crossover between the weak and strong field. This can be proved by working out the following exercises.
Exercise 2 Show that a classical charged particle of charge q, mass m and speed v would execute a circular orbit of angular frequency ωc if it moves under the influence of a magnetic field B~. Assume that no other forces act on the particle.
Exercise 3 Solve equations 10 under the conditions that Jy = Jz = 0 to show two results: Jx = σEx and Jx = σEy/(ωcτ ). Define the hall coefficient as RH ≡ Ey/(BJx) and show that RH = 1 /(nq).
Note that this model predicts two things: the Hall coefficient is independent of the magnetic field strength and that there is no dependence of the sample resistance on the magnetic field either. This second effect is called “magnetoresistance”, and it was this effect that Hall originally tried (and failed) to find [1, 5].
The classical theory of the Hall effect presented above assumes that the electric current is the result of many charge carriers moving independently of each other and responding to applied fields as classical particles. But we know that electrons are quantum particles, specifically fermions, and they have wavelike properties. Curiously, the act of changing the model from classical independent particles moving freely to quantum independent particles moving freely changes little in the results so far presented. The “free-electron quantum gas” model still predicts a hall coefficient of 1/nq and zero magnetoresistance [5].
The benefit of using a quantum approach becomes apparent when it is coupled with a more realistic model of solid matter, specifically, crystalline. In a crystal, the atoms are arranged in a periodic lattice. Electrons in the lattice feel the effect of a periodic potential on their motion. The strongest effect occurs for those electrons in the outer atomic orbitals—the “valence” electrons, and espe- cially those valence electrons whose deBroglie wavelength is close to the spacing of the potential’s periodicity.
Within the periodic potential the allowed energies of the valence electrons are broken into into a series of energy bands with energy gaps between them. If the number of valence electrons per unit cell of the crystal is exactly enough to fill a band, the solid will be a poor conductor, since by symmetry at each energy there will be filled momentum states pointing in opposite directions. Conduction can occur only if an electron can jump a gap into an unoccupied state. If the gaps are large, the necessary energy may be too high, and the solid is an insulator. If the gap is small, then thermal energy may be enough to cause sufficient electrons to jump the gap, and the solid is called a semiconductor. On the other hand, if the number of valence electrons per unit cell is not enough to fill a band, then many unfilled momentum states lie within easy energy reach, and the solid is a good conductor—a metal.
However, the shape of the energy surfaces—the band structure—has a strong effect on the type of conduction that can occur. Electron states with energies near the “bottom” of a band—the lowest allowed energy in the band—behave like free electrons, except that their response to an applied field may be that of a heavier or lighter particle. It as if the electron mass m has been changed to an effective mass m∗. While this statement seems strange, stranger still are the responses of electrons with energies near the “top” of a band; these electrons act as if their mass is negative! In other words, they accelerate oppositely a free electron when acted upon by an applied field. Indeed, it is easier to imagine particle states of this type as acting like positive charges. Such states are called “hole” states.
Another way to think of these effects is that the electron states with wavenumbers lying near the band extremes are diffracted by the periodic lattice in the same way that light is diffracted by a grating: the motion of an electron wave is greatly altered, even reversed, analogous to the way that a grating can reflect light of a specific color at specific angles.
In a semiconductor the band gap is relatively small, and electrons may be excited by thermal energy to jump the gap. This process allows an electron-like state near the bottom of an upper band and a hole-like state near the top of a lower band to come into existence. Both states carry current, with the hole state acting as a positive charge. The number of these current-carrying states depends on the temperature in a roughly exponential way: the number is proportional to the Boltzmann factor e−Eg^ /kT^ where Eg is the energy of the gap.
The idea of simultaneous electron+hole states has yielded a useful model of the Hall effect called the “two-band model”. The mathematics works through just like in Exercise 3, except that the total current is the sum of contributions from the holes and the electrons. If we let each type of carrier have a Hall coefficient: Re (for electrons) and Rh (for holes), as given by the Eq. (6) and we assume the conductivity σ = σe + σh, where σe and σh are given by Eq. (11), then we can derive an expression for the total Hall coefficient as
RH =
σ h^2 Rh + σ^2 e Re (σh + σe)^2
In the above, the effective mass m∗^ is substituted for m and the charge q taken as positive for holes and negative for electrons. This equation follows from Eqs. (10) in the limit that ωcτ ≪ 1: the low-field limit.
In the case of semiconductors, it has become customary to separate out the carrier density n from the overall conductivity formula for σ, and define a new quantity called the mobility μ:
μ =
∣∣ ∣∣^ qτ m∗
∣∣ ∣∣. (14)
Thus, in our two-band model, the conductivity would look like
σ = μe|qe|ne + μh|qh|nh , (15)
and the low-field Hall coefficient would look like
RH =
|q|
nhμ^2 h − neμ^2 e (nhμh + neμe)^2
where we have assumed that the charge is the same magnitude for both types of carriers.
Material Thickness t (m) Width w (mm) Length ℓ (mm) Resistance (Ω) Au 1. 37 ± 0. 16 × 10 −^7 12. 7 ± 0. 05 30. 5 ± 0. 05 0. 770 ± 0. 006 Al 2. 34 ± 0. 17 × 10 −^7 12. 7 ± 0. 05 30. 5 ± 0. 05 0. 872 ± 0. 018 InAs 1. 26 ± 0. 02 × 10 −^4 0. 692 ± 0. 004 1. 542 ± 0. 005 1. 3 ± 0. 3
Table 1: Characteristics of the Hall effect probes. The thicknesses of the metal samples were determined optically by comparing fringes created by thin-film interference. (See Optics, 2nd edition, by E. Hecht, pp. 381,382 for an explanation of this method.) The length and width were measured with calipers. The dimensions of the of the InAs probe were determined by the use of a measuring microscope. The resistance of the probes was measured along the length ℓ of the sample by a the four-point-probe method using a low frequency AC excitation and lock-in amplifiers.
Mount the probe into the clamp attached to the stand so that it can slip between the magnet pole pieces. Make sure that the probe is well aligned between the poles: it should be perfectly vertical with the plane of the paddle perpendicular to the table and centered both vertically and horizontally between the poles.
Connect the probe wires according to color following the diagram. Note that the voltage measure- ment wires are attached to a dual banana-plug connector that can fit into the DMM sockets. Note that the meter which measures current should use the white and black connection points, while the meter that measures voltage should use the red and black connection points. Do NOT use the connection points marked SENSE Ω 4 WIRE.
Turn on the meters and the 0-12 volt power supply. Set the voltage on the 0-12 volt supp;y to 10 volts. Set the voltage measurement meter to measure DC volts by pressing the “DCV” button, and the current measurement meter to measure DC amps by pressing the “DCI” button.
Several precautions must be followed:
- Do not exceed 200 mA current through either of the two metal samples (Au and Al).
- Do not exceed 100 mA current through the InAs sample.
- Turn down the Current Adjust knob all the way down to ZERO and turn the Current Output switch OFF before connecting or disconnecting any of the Hall samples. Because the current source attempts to maintain the same current regardless of load resistance, a sudden increase in the resistance (say, by disconnecting the wires) can cause a spike in the output voltage.
To this point, Vy has been equated with VH. In practice this does not hold true because of asymmetries in the probe element and voltage measurement contacts. Thus, even for Bz = 0, when a current Ix flows a voltage Vy will be measured due to the voltage drop along the length of the sample or differences in the resistivity of various parts of the sensing element.
Check this out with the sample outside the magnet: Turn the current up and note that you see a change in Vy even though there is a near-zero magnetic field (relative to the strong one between the magnet poles).
One way to eliminate the effect of this offset voltage is to measure the Vy for one direction of Bz , remove the sample from the magnet, flip the sample over 180◦, put the sample back in the magnet and measure Vy again. The difference in the two readings will be twice VH.
KIETHLEY 2000
MULTIMETER
VOLTS COM AMPS
HALL PROBE
VOLTAGE MEASUREMENT
WHITE GREEN
BETWEEN MAGNET POLES
MULTIMETER
AMPS
COM
KIETHLEY 2010 VOLTS
CURRENT MEASUREMENT
BLACKRED GREENYELLOW BLUEWHITE
− + − + − +
BLACK
RED
Constant Current Supply
0−12V^ + PWR SUPPLY
− GND
Figure 2: Schematic of Hall Effect experiment. The blade of the probe should be slipped between the magnet pole pieces so that the sensing element is centered.
When using this technique it is important that the current through the probe remain exactly the same for the two orientations of the probe.
You should follow this procedure: Before making any measurements, adjust the probe, clamp and stand so that the feet of the stand are flush against a piece of wood pressed against the magnet’s base. Then make sure that the probe element is well centered between the poles of the magnet by sighting directly down the terminal pins that the element is connected to and comparing these positions to the magnet pole pieces. Finally set the set the collar rings on the ring stand and probe mounting bar to be flush against the ring-stand’s clamp. By using the collar rings and feet of the stand, you can quickly place the probe in good alignment with the magnet between measurements. Then do the following:
- Set the current to the desired level.
- Wait until the voltage reading settles to within the desired precision (3–4 significant figures is usually adequate).
- Record the voltage, making sure to note the orientation of the probe.
- Pull the probe stand back, loosen the clamp, rotate the probe by 180◦, make sure the collar ring is flush against the clamp and and tighten the clamp.
- Slide the probe stand back into position with its feet against the wood piece which is flush against the magnet’s base.
- Check that the current has not changed significantly (i.e., less than 1 mA).
From the stated values of the dimensions and resistances of the InAs sample, it is possible to determine the sample conductivity σ:
σ =
Rwt
where R is the resistance of the sample. This formula assumes that the resistivity of the material is uniform over the sample and that the dimensions of the sample accurately reflect its geometry.
From the conductivity σ and the Hall coefficient RH it is possible to compute the “Hall mobility” μH of the charge carriers, defined as
μH ≡ |vx| |Ex|
Use your values of RH and σ to compute μH (and the error) for the InAs sample. Then look up mobilities for InAs in texts or handbooks and compare your value with theirs.
Discussion questions to research, ponder and discuss in your report:
- Find some values for mobility in tables that are listed in texts, such as Kittel’s or other references, such as the Landolt-B¨ornstein series of books available in the library. Compare your results with the tabulated values. How close are your results to the ones cited for InAs? Can you explain any differences?
- In the case of the aluminum sample, you will find in such texts as Kittel’s “Introduction to Solid State Physics” that the assumed charge carriers are “1-hole per atom”. Why should we assume that the charge carrier in aluminum is a hole? This is not predicted by the simple “Drude” theory of electrons in metals.
- Your results for aluminum are likely to be different from the predicted or tabulated values, which are usually quoted as being measured at low temperature and high magnetic field. See what you can find out about the Hall coefficient for a range of temperatures and field strengths, and discuss how your measurement fits in with these. Reference [6] is especially useful.
References
[1] Hall, E. H., “On a new action of the magnet on electric currents”, American Journal of Math- ematics, 2 , No. 3, pages 287–292 (1879). The original paper by Hall which describes the effect. An interesting historical read.
[2] Preston, D. W. and E. R. Dietz, The Art of Experimental Physics, John Wiley & Sons (1991), pages 303–315. Excellent discussion of the Hall effect in semiconductors.
[3] Melissinos, A. C., Experiments in Modern Physics, Academic Press (1966), pages 85–87.
[4] Kittel, C., Introduction to Solid State Physics, 6th edition, John Wiley & Sons (1986), pages 147–151. Standard undergraduate text. Contains useful tables, but does not give a detailed account of the Hall effect for metals.
[5] Ashcroft, Neil W., and N. David Mermin, Solid State Physics, Saunders College, Philadelphia (1976). Chapter 1 gives the elementary theory of the Hall Effect. Chapter 3 discusses the fail- ures of the free-electron model. Chapters 12 and 15 treat the Hall effect in the high field low temperature limit for metals.
[6] Hurd, Colin M., The Hall Effect in Metals and Alloys, Plenum Press, New York (1972). A thorough overview of the Hall effect in metals. Predates the quantum Hall effect.
[7] Silsbee, Robert H., and J¨org Dr¨ager, Simulations for Solid State Physics, Cambridge University Press (1997). The book and computer simulations illustrate many important concepts in solid state physics. The simulations are available on a computer in the lab.
[8] Ziman, J. M., Principles of the Theory of Solids, Cambridge University Press (1969). An excel- lent, readable and clear text covering much of the same material as Ashcroft and Mermin.
Prepared by D. B. Pengra, J. Stoltenberg, R. Van Dyck and O. Vilches hall_effect_15.tex -- Updated 19 June 2015
