Thermodynamic Limits of Thermoelectric Efficiency: Carnot Bound, Figure of Merit, and Practical Generator Performance
1. Why thermodynamic efficiency matters
A thermoelectric generator converts part of a heat flow directly into electrical work. Its performance is governed by two distinct questions:
- What is the maximum efficiency permitted by thermodynamics?
- How closely can a real thermoelectric leg or module approach that limit?
The first question establishes an absolute ceiling. The second depends on the material transport coefficients, their temperature dependence, device geometry, electrical and thermal contacts, and operating conditions.
The central material parameter is the dimensionless figure of merit,
[ ZT=, ]
where (S) is the Seebeck coefficient, () is electrical conductivity, () is total thermal conductivity, and (T) is absolute temperature. A high value of (ZT) is necessary for high thermoelectric efficiency, but it is not by itself a complete description of a finished generator.
This article derives the principal limits and shows how to use them without confusing an ideal material-level estimate with measured device efficiency.
2. The Carnot limit: the universal upper bound
Consider a heat engine operating between a hot reservoir at (T_h) and a cold reservoir at (T_c), with (T_h>T_c). If the engine absorbs heat (Q_h), rejects (Q_c), and delivers work (W), the first law gives
[ W=Q_h-Q_c. ]
The thermal efficiency is therefore
[ ==1-. ]
For a reversible engine, the total entropy change of the two reservoirs is zero:
[ -+=0. ]
Thus,
[ =, ]
and the reversible, or Carnot, efficiency becomes
[ ]
No heat engine operating between the same two reservoir temperatures can exceed this value. Real thermoelectric generators are irreversible, so their efficiency must remain below the Carnot limit.
Temperatures in these expressions must be in kelvin. A temperature difference in kelvin has the same numerical value as a difference in degrees Celsius, but absolute temperatures do not.
3. How a thermoelectric generator converts heat into work
A thermoelectric leg experiences a temperature difference along its length. The Seebeck effect generates an electromotive force. For a material with a temperature-dependent Seebeck coefficient, the open-circuit voltage is
[ V_{}=_{T_c}^{T_h} S(T),dT, ]
with the sign determined by the adopted voltage and carrier conventions.
When a load is connected, current flows. The generator then transports heat through several coupled mechanisms:
- Fourier heat conduction, which transfers heat from hot to cold even without useful electrical output.
- Peltier heat, associated with charge carriers entering or leaving a junction or region.
- Joule heating, generated by electrical resistance.
- Thomson heat, which can arise when current flows through a material with temperature-dependent Seebeck coefficient.
The balance among these contributions determines the heat absorbed at the hot side, the heat rejected at the cold side, and the electrical power delivered to the load. The Kelvin relations connect the reversible thermoelectric coefficients; see Kelvin Relations in Thermoelectric Materials.
4. The thermoelectric figure of merit
The dimensionless figure of merit is
[ ]
The numerator contains the power-factor contribution, (S^2), multiplied by temperature. The denominator is the total thermal conductivity.
The total thermal conductivity is commonly separated into electronic and lattice contributions:
[ =_e+_L, ]
where (_e) is electronic thermal conductivity and (_L) is lattice thermal conductivity. In some materials, bipolar or other contributions may also be relevant and should be included in the measured total.
A useful material-development strategy is therefore to increase the power factor while reducing heat conduction that does not contribute to electrical power. These objectives are coupled: changing carrier concentration can alter (S), (), and (_e) simultaneously.
For a deeper discussion of the transport terms, see Thermoelectric Transport and Thermodynamics of the Seebeck Effect.
5. Maximum efficiency for constant properties
For a single thermoelectric leg with constant transport properties, idealized one-dimensional operation, and appropriately optimized electrical load, the maximum efficiency can be written as
[ . ]
Here (ZT_{}) is an effective average figure of merit over the operating temperature interval. In the simplest constant-property model, it is evaluated using representative constant values. The equation is often used as a first estimate of the maximum conversion efficiency of a thermoelectric material operating between two specified temperatures.
The first factor is the Carnot efficiency. The second factor is a thermoelectric performance factor smaller than unity for finite positive (ZT). Consequently,
[ _{}<_C ]
for finite (ZT), while the idealized expression approaches the Carnot limit as (ZT) becomes very large.
Important limitation: this compact expression is not an exact solution for a real material whose (S), (), and () vary substantially with temperature. In that case, the coupled transport equations must be solved with temperature-dependent properties and suitable boundary conditions. A single arithmetic average of (ZT(T)) may not reproduce the exact device efficiency.
6. Interpreting the equation
The expression separates the temperature-reservoir limit from the material contribution.
6.1 Temperature span
The Carnot factor increases as the temperature difference increases for a fixed hot-side temperature. However, increasing the operating temperature can also change the material properties, accelerate chemical degradation, and increase thermal stresses.
6.2 Figure of merit
The square-root term shows that increasing (ZT) improves the ideal maximum efficiency, but the improvement is not linear. Doubling (ZT) does not double efficiency.
6.3 Hot- and cold-side temperatures
The ratio (T_c/T_h) appears in the denominator. The absolute reservoir temperatures therefore matter, not only their difference.
6.4 Why power factor alone is insufficient
A large Seebeck coefficient or electrical conductivity does not guarantee high efficiency if thermal conductivity is also large. The relevant material balance is represented by (ZT), while the actual output power additionally depends on geometry, load, and operating conditions.
7. Worked example: Carnot efficiency and thermoelectric efficiency
Consider an illustrative generator operating between
[ T_h=800;,T_c=400;, ]
with a representative constant-property value
[ ZT_{}=1.0. ]
The Carnot efficiency is
[ _C==0.50=50%. ]
For the idealized thermoelectric estimate,
[ ==1.414. ]
Therefore,
[ _{}=0.50 . ]
The estimated maximum efficiency is approximately 10.8%. This is well below the 50% Carnot ceiling.
The result is an idealized material-level estimate, not a prediction for a particular module. Contact resistance, parasitic heat leakage, nonuniform temperature distribution, and temperature-dependent properties can reduce the measured efficiency.
8. How efficiency changes with ZT
At fixed hot- and cold-side temperatures, a larger effective (ZT) increases the idealized maximum efficiency. The following values illustrate the trend for the same (800;) and (400;) reservoirs.
| (ZT_{}) | Idealized maximum efficiency |
|---|---|
| 0.5 | 7.4% |
| 1.0 | 10.8% |
| 1.5 | 13.3% |
| 2.0 | 15.2% |
| 3.0 | 18.0% |
These values are calculated from the constant-property expression above. They should not be interpreted as universal module efficiencies: the assumed reservoir temperatures and model assumptions are part of the result.
9. Load matching and maximum power are not the same as maximum efficiency
The external electrical load affects the current and voltage delivered by a generator. In a simplified constant-property circuit, the internal electrical resistance is (R_i), the load resistance is (R_L), and the current is
[ I=. ]
The electrical power delivered to the load is
[ P_L=I^2R_L. ]
For the elementary Thevenin model, maximum electrical power occurs when (R_L=R_i). However, the load that maximizes power does not generally coincide with the load that maximizes conversion efficiency. The optimum depends on the heat balance and the thermoelectric properties.
A generator designed for maximum energy conversion efficiency may therefore operate at a different electrical operating point from one designed for maximum power output. The design objective must be stated explicitly.
10. Why temperature-dependent properties matter
In practical materials, the transport coefficients vary with temperature:
[ S=S(T),=(T), =(T). ]
As a result,
[ ZT(T)=. ]
A material may have a high peak (ZT) over a narrow temperature range but a lower effective performance across the full device temperature gradient. Conversely, a material with a moderate but broad (ZT(T)) profile may be attractive for a particular operating window.
For non-constant properties, local transport equations are solved along the leg. In one common one-dimensional description, the electrical current density is coupled to the electrochemical potential gradient and temperature gradient, while the heat flux contains both conductive and thermoelectric terms. The energy equation must account for Joule and Thomson contributions where applicable.
A practical computational workflow is:
- Measure or establish (S(T)), ((T)), and ((T)) over the complete operating interval.
- Fit smooth, physically reasonable functions to the data.
- Solve the coupled one-dimensional transport and energy equations.
- Apply hot- and cold-side thermal boundary conditions and electrical load conditions.
- Calculate hot-side heat input and load power, then evaluate (=P_{}/Q_h).
- Repeat over load and geometry to identify the desired operating optimum.
The thermal conductivity measurement methodology is discussed in Thermal Conductivity: Measurement and Calculation.
11. Segmented thermoelectric legs
A single material rarely maintains optimal transport properties across a very wide temperature interval. A segmented leg uses different materials in different temperature regions, selecting each material for the temperature range in which it performs best.
Segmentation can improve the overall temperature-dependent performance, but it introduces additional design constraints:
- Electrical and thermal compatibility at the segment interface.
- Contact resistance and interfacial thermal resistance.
- Chemical compatibility during long-term operation.
- Mechanical compatibility under thermal cycling.
- The need to optimize the coupled current and heat flow through all segments.
The performance of a segmented device cannot be estimated by simply adding the individual efficiencies of its segments. The segments share current and are thermally coupled, so the complete structure must be modeled as one device.
12. Practical losses that separate a material estimate from a module
The idealized efficiency expression assumes a simplified thermoelectric element. A real generator may incur additional losses.
Electrical contact and interconnect resistance dissipate part of the generated electrical power as Joule heat.
Parasitic thermal conduction through supports, electrical interconnects, insulation gaps, and packaging transfers heat without producing useful electrical work.
Non-ideal heat exchangers create temperature drops between the external reservoirs and the thermoelectric material. The material therefore experiences a smaller temperature difference than the nominal system temperature span.
Radiative heat transfer can become important at high temperatures, depending on surface emissivity, view factors, and enclosure design.
Mechanical and chemical degradation can change contact quality, geometry, and transport properties during operation.
For meaningful reporting, distinguish among material-level maximum-efficiency estimates, single-leg measurements, module efficiency, and complete system efficiency. State the temperatures at the material boundaries, the heat-input measurement method, and whether auxiliary heat-exchange losses are included.
13. Implications for oxide thermoelectrics
Oxide thermoelectrics are often investigated for elevated-temperature applications. Their potential operating environment makes thermodynamic efficiency only one part of material selection.
A useful oxide thermoelectric candidate should be evaluated for:
- Transport performance: temperature-dependent Seebeck coefficient, electrical conductivity, and total thermal conductivity.
- Atmosphere stability: changes in oxygen chemical potential can alter oxygen nonstoichiometry, carrier concentration, and transport behavior.
- Phase stability: the material should remain in the required phase under the operating temperature and atmosphere.
- Microstructure: grain boundaries and secondary phases can influence both electrical and thermal transport.
- Thermal cycling: expansion mismatch and interfacial stresses can affect mechanical integrity.
- Contact compatibility: electrodes and interconnects must remain chemically and electrically stable.
In oxides, oxygen exchange can couple defect chemistry to electronic transport. A change in oxygen partial pressure may change carrier concentration and therefore both (S) and (). Thus, a room-temperature or single-atmosphere value of (ZT) is not sufficient to characterize a high-temperature generator.
14. A practical evaluation strategy
For a research-scale material assessment, the following sequence keeps the thermodynamic and experimental questions separate.
- Establish the operating window. Define realistic hot- and cold-side material temperatures, not just furnace and ambient temperatures.
- Measure transport properties. Obtain (S(T)), ((T)), and ((T)) using consistent specimen geometry and temperature calibration.
- Calculate (ZT(T)). Propagate measurement uncertainties, particularly those associated with thermal conductivity and specimen dimensions.
- Check stability. Repeat measurements after thermal exposure and under the relevant oxygen partial pressure.
- Estimate the idealized limit. Use the constant-property expression only when its assumptions are appropriate, and label the result clearly.
- Model the actual leg. Use temperature-dependent properties, realistic contacts, and boundary conditions.
- Validate experimentally. Measure electrical output and heat input independently and report the uncertainty in efficiency.
This workflow prevents a common mistake: treating the highest measured (ZT) value as if it were the efficiency of a working thermoelectric generator.
15. Common interpretation errors
- Calling Carnot efficiency a thermoelectric efficiency. Carnot efficiency is the reversible upper bound for every heat engine between the same reservoirs.
- Using Celsius in efficiency equations. Absolute temperatures must be expressed in kelvin.
- Using peak (ZT) for a broad temperature interval. Device performance depends on transport properties throughout the leg.
- Equating maximum power with maximum efficiency. These are different optimization objectives.
- Ignoring contact and parasitic losses. A material-level estimate is not a module or system measurement.
- Using only lattice thermal conductivity in the denominator. The figure of merit uses total thermal conductivity; electronic and other relevant contributions must be included.
- Comparing efficiencies without matching temperature boundaries. The material temperature span and the external reservoir span may differ.
16. Key equations
| Quantity | Expression |
|---|---|
| Carnot efficiency | (_C=1-T_c/T_h) |
| Figure of merit | (ZT=S^2T/) |
| Total thermal conductivity | (=_e+_L) (when these are the relevant contributions) |
| Open-circuit thermovoltage | (V_{}=_{T_c}^{T_h}S(T),dT) |
| Constant-property maximum efficiency | (_{}=) |
| Generator efficiency | (=P_{}/Q_h) |
17. Further reading
This article connects to the following topics in the thermodynamics and thermoelectric series:
18. Conclusions
The Carnot efficiency defines the universal thermodynamic ceiling for a generator operating between two fixed reservoir temperatures. The thermoelectric figure of merit determines how closely an idealized thermoelectric element can approach that ceiling. For constant properties, the maximum-efficiency expression provides a useful first estimate; for real materials, temperature-dependent transport coefficients and coupled heat and charge flow must be treated explicitly.
For oxide thermoelectrics, the evaluation must extend beyond a peak (ZT): oxygen chemical potential, phase stability, microstructure, contacts, and thermal cycling can all affect generator performance. The most meaningful path from material characterization to device efficiency is therefore a combination of complete temperature-dependent transport data, thermodynamic modeling, and direct measurements of heat input and electrical output.