c‑Axis‑Oriented Ba‑Doped BiCuSeO Thin Films Deliver Record Power Factors

Abstract

We report the epitaxial growth of c-axis‑oriented Bi_1−xBa_xCuSeO thin films (0 ≤ x ≤ 10 %) and examine how Ba substitution influences their structure, valence states, and thermoelectric behavior. X‑ray photoelectron spectroscopy shows partial reduction of Bi³⁺ to a lower valence state upon Ba doping, while Cu remains Cu⁺ and Se remains Se²⁻. Increasing Ba content raises hole concentration, lowering both resistivity and Seebeck coefficient. A peak power factor of 1.24 mW m⁻¹ K⁻² at 673 K was achieved in the 7.5 % Ba‑doped film—1.5× higher than bulk counterparts—suggesting a high ZT is attainable thanks to the films’ low thermal conductivity.

Background

Addressing the global energy crisis and environmental challenges requires clean, renewable solutions. Thermoelectric (TE) materials convert waste heat into electricity and enable solid‑state cooling. The figure of merit, ZT = (S²/ρκ)T, depends on the Seebeck coefficient (S), electrical resistivity (ρ), thermal conductivity (κ), and temperature (T). Thus, high ZT can be achieved by boosting the power factor S²/ρ and/or reducing κ.

BiCuSeO, a quaternary oxyselenide, has emerged as a promising TE material due to its intrinsically low κ. Its tetragonal ZrCuSiAs structure (P4/nmm) alternates insulating (Bi₂O₂)²⁺ layers with conductive (Cu₂Se₂)²⁻ layers stacked along the c axis. Recent bulk studies have improved BiCuSeO’s performance via c‑axis texturing, band‑gap tuning, vacancy engineering, grain‑boundary control, nano‑inclusions, and magnetic ion doping. For example, c‑axis‑textured Ba‑doped BiCuSeO bulk reached ZT ≈ 1.4 at 923 K, while dual‑vacancy and Pb/Ca dual‑doped systems achieved ZT ≈ 0.84–1.5 at elevated temperatures.

Thin‑film TE devices offer high power‑density cooling and fast response, and are compatible with micro‑electromechanical systems. However, BiCuSeO thin‑film growth is challenging due to volatile Bi and Se and stoichiometry control. Few reports exist on BiCuSeO thin‑film thermoelectric properties. Here we grow c-axis‑oriented Ba‑doped BiCuSeO films on SrTiO₃ (001) by pulsed laser deposition (PLD), investigate the impact of Ba on structure, valence, and TE performance, and demonstrate a record power factor.

Methods

Bi_1−xBa_xCuSeO films (x = 0, 2.5, 5, 7.5, 10 %) with ~50 nm thickness were deposited on commercial SrTiO₃ (001) substrates by PLD under high‑purity Ar. The lattice mismatch (<0.6 %) between BiCuSeO (a = b = 0.3926 nm) and SrTiO₃ (a = b = 0.3905 nm) facilitates epitaxy. Targets were sintered in vacuum‑sealed quartz tubes. Growth conditions: laser fluence ≈ 1.0 J cm⁻², 5 Hz repetition, 50 mm target–substrate distance, 0.1 Pa Ar, 330 °C substrate temperature.

Structural analysis used Cu Kα XRD, surface morphology was examined by FEI XL30 SEM (15 kV), and microstructure by TEM (Tecnai G2 F20). Valence states were probed by XPS (PHI Quantera SXM) under ~2×10⁻⁷ Pa, with a 5‑min Ar⁺ etch. Hall effect measurements employed a PPMS‑9 in van der Pauw geometry. Electrical resistivity and Seebeck coefficient were measured from 300 K to 700 K in He using a Linseis LSR‑800 with 5 K min⁻¹ heating rate.

Results and Discussion

Figure 1a shows XRD θ–2θ scans of Bi_1−xBa_xCuSeO films. All peaks correspond to (00l) reflections of the tetragonal BiCuSeO phase, confirming perfect c-axis alignment. Peak broadening with higher Ba content indicates reduced grain size, likely due to dopant pinning of grain boundaries. The shift to lower 2θ values reflects incorporation of larger Ba²⁺ (1.42 Å) ions into Bi³⁺ sites, expanding the c lattice—values closely match bulk references.

Figure 1b presents a pole figure at 2θ = 33.75°, capturing the BiCuSeO (111) peak and the SrTiO₃ (110) peak. Analysis with STEREOPOLE confirms a fully epitaxial relationship: [010] BiCuSeO//[010] SrTiO₃ and [001] BiCuSeO//[−100] SrTiO₃. Figure 1c shows a 4‑fold symmetric φ‑scan of the (103) peak, confirming tetragonal symmetry.

Cross‑sectional TEM (Fig. 2) reveals a flat surface and interface. A few‑nanometer bright layer at the film/substrate interface likely arises from crystallographic mismatch. High‑resolution images show the expected alternating Bi–O and Cu–Se layers along c. Selected‑area diffraction confirms epitaxy.

XPS spectra (Fig. 3) confirm valence states: Bi 4f peaks at 159.1 and 164.4 eV (Bi³⁺) with lower‑binding‑energy shoulders indicating reduced Bi (Bi^3−x), Ba 3d peaks at 780.4 and 795.8 eV (Ba²⁺), Cu 2p peaks at 933.2/953.0 eV (Cu⁺), and Se 3d peaks at 54.2/55.0 eV (Se²⁻). O 1s at 530.2 eV indicates clean oxide surfaces. These findings suggest increased hole carriers in heavily doped films.

Hall measurements (Fig. 4a) confirm p‑type conduction. The undoped film exhibits a carrier concentration of ~6.6×10¹⁹ cm⁻³, higher than bulk values, likely due to intrinsic vacancies. Ba doping increases n in proportion to Ba content: 3.62×10²⁰, 7.25×10²⁰, 1.08×10²¹, and 1.45×10²¹ cm⁻³ for 2.5, 5, 7.5, 10 % Ba, respectively. Slightly higher measured concentrations at ≥5 % Ba suggest additional vacancies. Mobility decreases from 8.3 to 1.3 cm² V⁻¹ s⁻¹ with increasing Ba due to enhanced scattering, yet remains comparatively high, comparable to Mg‑doped LaCuSeO thin films, likely due to strong covalency and low effective mass.

Electrical resistivity and Seebeck coefficient (Fig. 4b,c) both increase with temperature, indicative of metallic‑like transport. Ba doping reduces both parameters by raising hole concentration. Compared to polycrystalline ceramics, the ab‑plane resistivity is markedly lower, reflecting the anisotropic layered structure where in‑plane conduction is superior.

Consequently, the power factor (Fig. 4d) rises sharply with Ba content, reaching 1.24 mW m⁻¹ K⁻² at 673 K for the 7.5 % Ba film—2.8× the undoped film and 1.5× bulk Ba‑doped or c‑axis‑textured samples. This enhancement stems from low resistivity driven by high carrier concentration and c‑axis orientation. Estimating ZT using the Wiedemann–Franz law for electronic thermal conductivity and bulk phonon values (0.55 W m⁻¹ K⁻¹ at 300 K, 0.35 at 673 K) yields ZT ≈ 0.26 at 300 K and 0.93 at 673 K. Since thin films typically exhibit even lower phonon thermal conductivity due to surface and interface scattering, the true ZT may be higher.

Density functional theory calculations (Fig. 5) show that Ba substitution minimally perturbs the band dispersion but shifts the Fermi level into the valence band, confirming hole introduction to the Cu–Se layer. The rigid‑band behavior explains the observed power‑factor increase.

Conclusions

Bi_1−xBa_xCuSeO films (0–10 %) were successfully grown epitaxially on SrTiO₃ (001) by PLD. XRD and TEM confirm c‑axis orientation and in‑plane epitaxy. Ba substitution increases hole concentration, reducing resistivity and Seebeck coefficient, but the resulting power factor surpasses bulk materials, achieving 1.24 mW m⁻¹ K⁻² at 673 K for the 7.5 % Ba film. Combined with expected low thermal conductivity, these films hold promise for high‑performance thin‑film thermoelectrics.

Abbreviations

PLD:

Pulsed laser deposition

PPMS:

Physical property measurement system

SAED:

Selected area electron diffraction

SEM:

Scanning electron microscope

TE:

Thermoelectric

TEM:

Transmission electron microscope

XPS:

X‑ray photoelectron spectroscopy

XRD:

X‑ray diffraction