Thermoelectric devices are solid-state energy converters that directly transform temperature gradients into electrical voltage and vice versa. They have attracted intense research interest because they enable clean power generation from waste heat, provide localized cooling without refrigerants, and offer maintenance-free operation in remote or harsh environments. Despite these advantages, the widespread adoption of thermoelectric technology has historically been limited by relatively low conversion efficiency. That efficiency hinges on a delicate interplay among three material properties: the Seebeck coefficient (S), electrical conductivity (σ), and thermal conductivity (κ). Among these, electrical conductivity is particularly susceptible to modification through a process known as material doping. Doping, the intentional introduction of impurities into a semiconductor or semimetal, is the single most powerful tool for tuning the charge carrier concentration and, consequently, the electrical conductivity of thermoelectric materials. Understanding the relationship between doping, carrier density, and electrical transport is therefore essential for designing next-generation thermoelectric devices with high performance.

Understanding Material Doping in Thermoelectric Semiconductors

At its core, doping is the controlled addition of foreign atoms into a host crystal lattice. In thermoelectric materials—which are typically heavily doped semiconductors or semimetals—doping alters the electronic band structure by introducing states near the conduction or valence bands. When dopant atoms donate electrons to the conduction band, the material becomes n-type (negative charge carriers). Conversely, dopants that accept electrons create holes in the valence band, yielding p-type (positive charge carriers) behavior.

The key parameter affected by doping is the charge carrier concentration (denoted n for electrons or p for holes). In an undoped intrinsic semiconductor, carrier concentrations are low and dominated by thermally excited electron–hole pairs. Doping can increase n or p by several orders of magnitude, often from 1016 cm–3 in intrinsic material to 1019–1021 cm–3 in heavily doped thermoelectrics. This massive shift directly controls the electrical conductivity through the fundamental Drude relation:

σ = n e μ

where e is the elementary charge and μ is the carrier mobility. Doping thus has a dual effect: it increases n (beneficial for conductivity) but often reduces μ due to increased scattering from ionized impurities. The net outcome depends on the doping level, the host material, and the specific dopant.

The Direct Relationship Between Doping Level and Electrical Conductivity

In the low-to-moderate doping regime, electrical conductivity rises monotonically as dopant concentration increases. More charge carriers become available to transport current, and the mobility typically falls only slowly because scattering is weak. For example, in bismuth telluride (Bi2Te3)—the most widely used thermoelectric material near room temperature—adding antimony (Sb) as a p-type dopant progressively raises the hole concentration, boosting conductivity from about 103 S/m in lightly doped samples to over 105 S/m at optimal doping levels.

However, the relationship is not linear indefinitely. At very high doping levels, several mechanisms conspire to limit further increases—or even reverse the trend. Ionized impurity scattering becomes dominant, severely reducing carrier mobility. Additionally, the solubility limit of the dopant may be reached; beyond that point, secondary phases or precipitates form, which act as additional scattering centers and can degrade electrical connectivity. In extreme cases, excessive doping can also shift the Fermi level into a region of the band structure where the density of states is low, limiting further gains in conductivity.

The result is a characteristic “doping curve” for each material: initial increases in doping yield strong gains in conductivity, but eventually a plateau or maximum is reached. Identifying this optimum—where σ is high while μ is still reasonable—is a central challenge in thermoelectric design. This optimum also depends on operating temperature because the relative contributions of phonon scattering, ionized impurity scattering, and other mechanisms shift with temperature.

Optimizing Thermoelectric Performance: The Role of the Figure of Merit

Electrical conductivity does not exist in isolation. Thermoelectric efficiency is captured by the dimensionless figure of merit ZT = (S2σ / κ) T. The power factor (S2σ) rewards high σ, but S itself tends to decrease with increasing carrier concentration. This trade-off means that blindly maximizing doping to raise σ can reduce the Seebeck coefficient, sometimes dramatically. The product S2σ therefore exhibits a maximum at an intermediate carrier concentration—typically ranging from 1019 to 1020 cm–3 for most thermoelectric materials.

Moreover, the denominator of ZT—thermal conductivity—is also affected by doping. Increased electrical conductivity comes with a higher electronic contribution to thermal conductivity (via the Wiedemann–Franz law, κe = L σ T, where L is the Lorenz number). At the same time, doping can alter the lattice thermal conductivity by introducing point defects that scatter phonons. In some materials, heavy doping actually reduces lattice thermal conductivity, partially compensating for the rise in electronic thermal conductivity. The net effect on κ must be accounted for when selecting doping strategies.

The Pisarenko relation—which links S to carrier concentration through the density of states effective mass—provides a theoretical framework for predicting the optimal doping level. For a given material, one can calculate the carrier concentration that maximizes ZT by balancing these competing trends. Experimentally, this is achieved by systematically varying dopant type and concentration, then measuring S, σ, and κ at the target temperature.

Practical Doping Strategies in Common Thermoelectric Materials

A wide variety of thermoelectric compounds have been developed, each with tailored doping approaches. Below are representative examples illustrating how electrical conductivity is controlled through material doping.

Bismuth Telluride (Bi2Te3)

Bi2Te3 is the workhorse of room-temperature thermoelectrics. For p-type doping, antimony (Sb) is substituted on bismuth sites; for n-type doping, selenium (Se) replaces tellurium, or iodine (I) is introduced as a donor. Optimal carrier concentrations for both polarities lie near 1019 cm–3, yielding ZT values around 1.0 at 300 K. Recent work has explored doping with copper (Cu) or silver (Ag) to further increase conductivity by intercalating between van der Waals layers, though long-term stability remains a concern.

Lead Telluride (PbTe)

PbTe is a classic mid-temperature thermoelectric (500–900 K). P-type doping is achieved with sodium (Na) or thallium (Tl), while n-type doping uses iodine (I) or bismuth (Bi). A particularly successful strategy is “resonant level” doping with Tl, which distorts the density of states to maintain a high Seebeck coefficient even at elevated carrier concentrations. This approach has yielded ZT > 1.5 at around 750 K.

Skutterudites

Skutterudites (e.g., CoSb3) feature large cages that can be filled with guest atoms such as rare earth elements (Ce, Yb, Ba). Filling these voids introduces “rattling” modes that scatter phonons, reducing lattice thermal conductivity. Simultaneously, the filler atoms donate electrons, tuning the carrier concentration. P-type skutterudites are obtained by doping with Fe on Co sites. Optimized compositions achieve ZT values exceeding 1.4 at 800 K.

Half-Heusler Compounds

Half-Heusler alloys (e.g., TiNiSn, ZrNiSn) are mechanically robust and thermally stable, making them attractive for high-temperature power generation. They naturally have low carrier concentrations; doping with elements like Sb, Ta, or V (for n-type) and Sc, Y, or Ti (for p-type) raises n into the 1021 cm–3 range. Despite the high conductivity, lattice thermal conductivity is relatively high, keeping ZT around 0.5–1.0. Nanostructuring and intense doping are being combined to improve performance.

Tin Selenide (SnSe)

SnSe has recently attracted attention for its ultra-low lattice thermal conductivity and record ZT > 2.6 at 923 K in single crystals. Doping SnSe with sodium (Na) or silver (Ag) introduces holes and increases electrical conductivity along the b-axis. However, anisotropic properties mean that doping and conductivity are direction-dependent, complicating optimization.

Advanced Doping Techniques to Overcome Limitations

As conventional substitutional doping reaches its limits, researchers have developed sophisticated strategies to decouple carrier concentration from mobility.

Modulation Doping

In modulation doping, charge carriers are spatially separated from the dopant atoms that supply them. A common approach is to embed quantum dots or nanoscale precipitates of a heavily doped semiconductor inside a matrix of the same material. Carriers spill from the precipitates into the matrix, where they experience less ionized impurity scattering because the charged dopants remain confined to the precipitates. This technique has been demonstrated in SiGe and BiSbTe systems, achieving up to a 30% increase in power factor without raising thermal conductivity.

Delta Doping

Delta doping confines dopants to a single atomic plane within the material. The carriers enter the adjacent layers and move in a high-mobility channel. Delta doping has been used in thin-film thermoelectric superlattices, such as PbTe/PbSe structures, to boost conductivity while preserving the Seebeck coefficient.

Alloying and Solid Solution Formation

Creating solid solutions (e.g., Bi2Te3-Sb2Te3) amounts to a form of “pseudo-doping” that effectively adjusts the band gap and the carrier concentration. Alloying also introduces point defects that scatter short-wavelength phonons, reducing lattice thermal conductivity—a beneficial side effect that can enhance ZT beyond what pure doping achieves.

Nanostructuring Combined with Doping

Nanostructuring—creating grains or inclusions on the nanometer scale—introduces numerous grain boundaries that scatter mid- to long-wavelength phonons. When combined with heavy doping, one can simultaneously achieve high carrier concentration and low thermal conductivity. The key is to ensure that the grain boundaries do not severely scatter charge carriers. Proper passivation or modulation doping can mitigate that risk.

Challenges and Future Directions in Doping Thermoelectrics

Despite decades of progress, several fundamental challenges remain. Dopant solubility in the host lattice is often limited; exceeding that limit leads to secondary phases that degrade performance. Moreover, the solubility itself is temperature-dependent, meaning the optimum doping at synthesis temperature may not be the equilibrium optimum at the operating temperature. Thermal stability is another issue: dopants can diffuse or precipitate out over time under thermal cycling, causing performance drift.

Another challenge is the convergence of electronic and thermal design. Ideally, doping should maximize electrical conductivity while simultaneously reducing lattice thermal conductivity. Some dopants, such as rare earth elements in skutterudites, play a dual role by both donating carriers and rattling to scatter phonons. Identifying or designing such “multifunctional” dopants is an active area of research.

Finally, manufacturing constraints limit the range of practical doping techniques. For large-scale applications (e.g., automotive waste heat recovery), doping must be reproducible, cost-effective, and compatible with existing processing routes such as powder metallurgy or melt spinning. Scalability remains a gating factor between laboratory records and commercial products.

Looking ahead, machine learning and high-throughput computational screening are being used to accelerate the discovery of optimal dopants. Databases of doping properties combined with density functional theory calculations can predict the most promising combinations of host and dopant for a given temperature range. Experimental validation of these predictions is rapidly advancing, and several new dopant–host systems with improved ZT have already been reported.

Conclusion

Material doping is the most direct and powerful lever for controlling electrical conductivity in thermoelectric devices. By adjusting carrier concentration, doping can raise conductivity by orders of magnitude, but it simultaneously affects the Seebeck coefficient and thermal conductivity. The art of designing a high-performance thermoelectric material lies in finding the doping level that optimally balances these competing parameters to maximize the figure of merit ZT. From classic materials like Bi2Te3 and PbTe to emerging systems such as SnSe and doped half-Heuslers, the principles remain the same: introduce the right dopant at the right concentration to achieve a high power factor without penalizing thermal transport. Advanced techniques such as modulation doping and delta doping promise to push beyond the trade-offs of conventional doping. Continued research into new dopants, combined with computational screening and better understanding of scattering mechanisms, will further improve the efficiency and reliability of thermoelectric generators and coolers. As global efforts to harvest waste heat intensify, mastering the relationship between doping and electrical conductivity will remain at the core of thermoelectric innovation.

For further reading: Nature: Advances in thermoelectric materials | ScienceDirect: Doping strategies for thermoelectrics | Chemical Reviews: Thermoelectric doping mechanisms | Journal of Materials Science: Modulation doping review.