Thermodynamics of the Seebeck Effect: Chemical Potential, Carrier Statistics, and Thermoelectric Transport
Thermodynamics of the Seebeck Effect: Chemical Potential, Carrier Statistics, and Thermoelectric Transport
The Seebeck effect is the generation of an electrical voltage when a material experiences a temperature gradient. It is one of the central thermoelectric phenomena and provides a direct link between thermal energy, electronic structure, carrier statistics, and charge transport.
For materials scientists, the Seebeck coefficient is more than a voltage-per-temperature number. Its magnitude and sign contain information about the dominant carrier type, the position of the chemical potential relative to electronic band edges, and the energy dependence of carrier transport. In semiconducting oxides, these relationships are particularly useful because defect chemistry and oxygen nonstoichiometry can strongly change carrier concentration and electronic conductivity.
This article develops the physical and thermodynamic foundations of the Seebeck effect, derives the transport expressions commonly used to interpret measurements, and explains how carrier concentration and electronic structure influence thermopower. The discussion distinguishes general transport equations from simplified semiconductor models, whose assumptions must be checked before applying them to real materials.
1. The Seebeck effect and sign convention
Consider a homogeneous sample with a temperature difference (T) between its ends. Under open-circuit conditions, no net electrical current flows. A voltage develops as charge carriers redistribute in response to the temperature gradient.
The Seebeck coefficient is conventionally defined as
[ ]
In the local linear-response description,
[ =ST ]
for this stated convention, where () is the electric field. Equivalent expressions can appear with different signs if voltage polarity, electric-field direction, or the definition of the temperature gradient is reversed. A measurement report should therefore state the sign convention and electrode orientation.
The SI unit is V K(^{-1}), but thermopower is commonly reported in ().
A negative Seebeck coefficient is commonly associated with electron-dominated transport, while a positive coefficient is commonly associated with hole-dominated transport. This interpretation is reliable for a single dominant carrier contribution, but can be misleading in multiband or mixed-conduction materials.
2. Why a temperature gradient produces a voltage
A temperature gradient changes the distribution of carriers over energy. Carriers at the hot side have a broader energy distribution than carriers at the cold side. Because the population and transport probability depend on energy, the two sides do not exchange carriers symmetrically.
The resulting redistribution produces an electric field. Under open-circuit conditions, this field builds until the electrical contribution balances the net thermally driven charge flow.
The process involves three connected ingredients:
- Carrier statistics: the occupation probability of electronic states depends on temperature and chemical potential.
- Electronic structure: the density of available states and their energies determine which carriers can participate.
- Transport probability: carrier velocity, scattering time, and band structure determine how effectively carriers move.
Consequently, thermopower is not determined by carrier concentration alone. It depends on how the carriers that contribute to conduction are distributed in energy.
3. Chemical potential and the Fermi–Dirac distribution
The electronic chemical potential () determines the equilibrium occupation of electronic states. For non-interacting fermionic quasiparticles, the Fermi–Dirac distribution is
[ f(E,,T)= {+1} ]
where (E) is the state energy, (k_B) is Boltzmann’s constant, and (T) is absolute temperature.
The chemical potential is closely related to the Fermi level in electronic materials. In a semiconductor, its position relative to the conduction-band minimum (E_C) and valence-band maximum (E_V) strongly influences electron and hole populations.
For a nondegenerate semiconductor, the approximate carrier concentrations are
[ n=N_C]
and
[ p=N_V]
where (N_C) and (N_V) are the effective densities of states in the conduction and valence bands.
Moving the chemical potential toward the conduction band generally increases electron concentration; moving it toward the valence band generally increases hole concentration. The Seebeck coefficient responds to this shift because the energy distribution of the conducting carriers changes.
4. Thermopower as an energy-weighted transport response
A useful general description of electronic thermoelectric transport uses transport integrals. In an isotropic or appropriately averaged representation, define
[ L_n= _{-}^{} (E)(E-)^n (-)dE ]
where ((E)) is the transport distribution function. It contains the contributions from the electronic states, their velocities, and their scattering times.
For a single carrier channel with the conventional electron-charge magnitude (e>0), the Seebeck coefficient can be written
[ ]
under the usual electron transport convention and with the corresponding definition of the transport integrals.
The factor (-f/E) is strongly concentrated within an energy range of a few (k_BT) around the chemical potential. Thus, thermopower is particularly sensitive to the energy dependence of transport near ().
The ratio (L_1/L_0) is the average transport energy measured relative to the chemical potential. If the transport distribution is symmetric around (), positive and negative energy deviations can cancel and the Seebeck coefficient can be small. A substantial thermopower requires an energy asymmetry in the conducting states and their transport probabilities.
5. The transport distribution function
In the semiclassical Boltzmann description within the relaxation-time approximation, the transport distribution can be expressed schematically as
[ (E)= D(E)v^2(E)(E) ]
for an isotropic three-dimensional system. Here (D(E)) is the density of states, (v(E)) is the carrier group velocity, and ((E)) is the relaxation time.
For anisotropic crystals, the transport distribution is tensorial and contains directional velocity components. The scalar expression is therefore a useful model, not a universal representation for every crystal symmetry.
This expression clarifies why two materials with similar carrier concentrations can have different Seebeck coefficients. Their density of states, band velocities, and scattering mechanisms may have different energy dependences.
6. Physical interpretation: entropy and energy carried by charge
Charge carriers transport both electrical charge and energy. A temperature gradient drives a redistribution of carriers, and the open-circuit electric field reflects the energy carried per unit charge by the mobile population.
This motivates an intuitive description of thermopower as an entropy-related quantity per carrier. However, the Seebeck coefficient is not generally equal to a simple equilibrium entropy divided by charge. The measured value is a transport coefficient and is weighted by the energies and velocities of carriers that actually conduct.
The distinction matters when interpreting strongly correlated materials, narrow-band systems, oxides with multiple conduction channels, and materials where scattering varies rapidly with energy.
7. Nondegenerate semiconductor approximation
For a nondegenerate semiconductor with a parabolic band and a specified energy dependence of the relaxation time, simplified analytical expressions can be obtained.
For electrons in a single parabolic conduction band,
[ S_n= - ]
For holes in a single parabolic valence band,
[ S_p= + ]
The dimensionless terms (A_n) and (A_p) depend on the band model and scattering mechanism. For a common three-dimensional parabolic-band model with ((E-E_C)^r) for electrons, the corresponding transport factor is (A_n=r+5/2). The analogous hole expression uses the appropriate valence-band energy dependence.
These expressions show that the magnitude of thermopower often increases as the chemical potential moves farther into the band gap, while the carrier concentration decreases. This is one reason high Seebeck coefficients are frequently observed in lightly doped semiconductors.
The approximation is not appropriate when the semiconductor is degenerate, when multiple bands contribute substantially, or when the density of states and scattering cannot be represented by the assumed parabolic-band model.
8. Carrier concentration and the Seebeck coefficient
For a nondegenerate n-type semiconductor, the carrier concentration is approximately
[ n=N_C]
Taking the logarithm,
[ = () ]
Substitution into the single-band expression gives
[ S_n= - ]
Similarly, for a nondegenerate p-type semiconductor,
[ S_p= + ]
These equations demonstrate the qualitative relationship between carrier concentration and thermopower: increasing the majority-carrier concentration generally reduces the magnitude of (S) in the nondegenerate regime, if the other model parameters remain fixed.
The relationship is not a universal one-variable law. Effective density of states, band degeneracy, scattering, temperature, and multiband effects can all change the result.
9. Degenerate semiconductors and the Mott relation
When the chemical potential lies within or close to a band, the nondegenerate approximation may fail. For a degenerate conductor at sufficiently low temperature relative to the electronic energy scale, the Mott relation gives
[ ]
This expression shows that thermopower is governed by the energy derivative of the transport distribution at the chemical potential.
A large density of states alone does not guarantee a large Seebeck coefficient. What matters in this approximation is the energy asymmetry of the full transport distribution, including velocity and scattering contributions.
The Mott relation assumes a sufficiently smooth transport distribution near the chemical potential and a degenerate regime where the Sommerfeld expansion is valid. It should not be applied indiscriminately to nondegenerate semiconductors or systems with sharp energy features on the scale of (k_BT).
10. Why band structure matters
The electronic band structure determines the available states and their group velocities. Several band features can influence thermopower:
- Band curvature: affects the effective mass and carrier velocity.
- Band degeneracy: multiple equivalent valleys can increase the density of states.
- Band convergence: nearby band extrema may contribute simultaneously.
- Strong energy dependence: sharp changes in the density of states or transport distribution can enhance thermopower.
- Anisotropy: different crystallographic directions can have different Seebeck coefficients.
A high density-of-states effective mass can increase thermopower in suitable band models, but it can also reduce mobility if it is accompanied by low band velocities or enhanced scattering. The transport distribution, not one band parameter in isolation, controls the result.
11. Electron and hole contributions in a mixed conductor
If electrons and holes both contribute to electrical conduction, the measured Seebeck coefficient is a conductivity-weighted average:
[ ]
where (_n) and (_p) are the partial conductivities.
Because electron and hole Seebeck coefficients usually have opposite signs, their contributions can partially cancel. A material may therefore exhibit a small measured Seebeck coefficient even when both carrier channels individually have substantial thermopower.
This is important for oxides near a change from n-type to p-type behaviour, for materials with thermally activated minority carriers, and for mixed ionic-electronic conductors where the measured electrical response may contain additional contributions.
12. Temperature dependence of thermopower
The Seebeck coefficient can vary strongly with temperature because several quantities change simultaneously:
- The chemical potential may shift with temperature.
- Carrier concentrations may increase through thermal excitation.
- The dominant scattering mechanism may change.
- Additional bands may become thermally accessible.
- Bipolar conduction may become important.
- Defect equilibria may change the electronic carrier concentration.
For an extrinsic semiconductor, thermopower may vary gradually over a temperature range where dopants dominate the carrier population. At higher temperatures, intrinsic excitation or defect-controlled changes can alter both the magnitude and sign of (S).
A temperature-dependent Seebeck curve should therefore be interpreted together with electrical conductivity and, where possible, carrier concentration and defect-chemistry information.
13. Connection to oxide defect chemistry
In functional oxides, electronic carriers are often controlled by point defects and gas–solid equilibria. A representative oxygen-loss reaction is
[ ]
Under conditions where this reaction produces mobile electrons, lowering oxygen partial pressure can increase the electron concentration. This can change the Seebeck coefficient as well as the electrical conductivity.
In acceptor-doped oxides, charge compensation and oxygen incorporation or loss can produce different relationships between oxygen activity and carrier concentration. The sign and magnitude of the thermopower therefore depend on the dominant defect equilibria, not simply on the nominal dopant valence.
A rigorous interpretation combines the electronic transport model with charge neutrality, mass-action relations, oxygen nonstoichiometry, and the relevant temperature and oxygen-pressure range.
14. A practical example: interpreting an n-type oxide
Suppose an oxide exhibits a negative Seebeck coefficient over a temperature interval. This is consistent with electron-dominated thermoelectric transport.
If a controlled reduction treatment increases oxygen deficiency and the measured conductivity rises, one possible interpretation is that oxygen loss generates additional electrons. If the magnitude of the negative Seebeck coefficient decreases at the same time, that trend may be consistent with an increase in electron concentration within a nondegenerate single-band regime.
However, this interpretation is not conclusive by itself. Reduction can also change mobility, phase composition, grain-boundary barriers, and the energy dependence of scattering. A useful experimental analysis therefore compares:
- Seebeck coefficient as a function of temperature.
- Electrical conductivity under the same atmosphere.
- Oxygen nonstoichiometry or mass change, where measurable.
- Phase composition before and after treatment.
- Reversibility on reoxidation.
The combination is more informative than assigning a carrier concentration from the Seebeck coefficient alone.
15. Measurement considerations
The measured Seebeck coefficient is obtained from the voltage generated by a known temperature difference. In practice, the result can be affected by:
- Thermocouple calibration and placement.
- Temperature-gradient stability.
- Electrical contact quality.
- Lead-wire thermopower and reference junctions.
- Sample geometry and temperature uniformity.
- Atmosphere-dependent changes during measurement.
- Non-ohmic contacts or parasitic voltage offsets.
For high-temperature oxide measurements, the atmosphere and equilibration time should be reported because oxygen exchange can change the sample’s defect concentration during the measurement.
A reliable result should include the sign convention, temperature range, atmosphere, heating or cooling protocol, and whether the reported value is steady-state or transient.
16. Seebeck coefficient and electrical conductivity must be considered together
A large Seebeck coefficient does not by itself indicate a high-performance thermoelectric material. Electrical conductivity is also essential.
The power factor is
[ ]
where () is electrical conductivity.
Increasing carrier concentration often increases () while reducing the magnitude of (S). Thermoelectric optimization therefore involves balancing these competing trends rather than maximizing either quantity independently.
The power factor is one component of thermoelectric performance. Thermal conductivity and the dimensionless figure of merit must also be considered when evaluating a complete thermoelectric material.
17. Common interpretation errors
- Assuming the Seebeck sign proves a single carrier type: multiband and mixed conduction can complicate the sign.
- Treating thermopower as a direct carrier-density measurement: extracting carrier concentration requires a validated band and scattering model.
- Using a nondegenerate expression for a degenerate material: the appropriate transport regime must be established first.
- Assuming a large density of states guarantees large thermopower: the energy dependence of the full transport distribution is decisive.
- Ignoring atmosphere effects in oxides: oxygen nonstoichiometry can change carrier populations during measurement.
- Comparing data with different sign conventions: voltage polarity and temperature-gradient direction must be consistent.
- Optimizing (S) alone: electrical conductivity and thermal conductivity are also needed to assess thermoelectric performance.
18. Further reading
The following articles develop related concepts and experimental methods:
- Thermoelectric Transport: Seebeck Coefficient, Electrical Conductivity, and Power Factor
- Thermodynamic Potentials in Materials Science
- Chemical Potential and Activity in Solid Materials
- Thermodynamic Factors in Ionic Transport
- Oxygen Chemical Potential and pO2-Dependent Defect Chemistry
- Defect Formation Energy and Equilibrium Concentrations in Oxides
- Oxygen Nonstoichiometry Measurement in Functional Oxides
19. Final perspective
The Seebeck coefficient is a transport response that connects temperature gradients to the energy distribution of mobile charge carriers. Its sign often identifies the dominant carrier type, while its magnitude reflects the chemical potential, electronic structure, and energy dependence of transport.
For functional oxides, the interpretation becomes more powerful when electronic transport is connected to defect thermodynamics and oxygen activity. Combining Seebeck measurements with conductivity, atmosphere control, and structural or nonstoichiometry measurements provides a more defensible picture of the mechanisms governing thermoelectric behaviour.