Interfacial Distortion of Sb2Te3–Sb2Se3 Multilayers via Atomic Layer Deposition for Enhanced Thermoelectric Properties

Atomic layer deposition (ALD) is an effective technique for depositing thin films with precise control of layer thickness and functional properties. In this work, Sb2Te3–Sb2Se3 nanostructures were synthesized using thermal ALD. A decrease in the Sb2Te3 layer thickness led to the emergence of distinct peaks from the Laue rings, indicative of a highly textured film structure with optimized crystallinity. Density functional theory simulations revealed that carrier redistribution occurs at the interface to establish charge equilibrium. By carefully optimizing the layer thicknesses, we achieved an obvious enhancement in the Seebeck coefficient, reaching a peak figure of merit (zT) value of 0.38 at room temperature. These investigations not only provide strong evidence for the potential of ALD manipulation to improve the electrical performance of metal chalcogenides but also offer valuable insights into achieving high performance in two-dimensional materials.


INTRODUCTION
Thermoelectric (TE) materials have the capacity to directly convert heat energy into electrical power, rendering them highly promising for microenergy harvesting applications in Internet of Things devices. 1 The heat-to-electricity conversion efficiency of thermoelectric materials can be evaluated by the dimensionless figure of merit (zT), which is defined as zT = σS 2 T/κ, where σ, S, T, and κ are the electrical conductivity, Seebeck coefficient, absolute temperature, and total thermal conductivity, respectively.κ is a combination of electronic (κ E ) and lattice (κ L ) thermal conductivities.Enhancing the TE power factor (σS 2 ) in any material necessitates precise control of the critical transport parameters.A high power factor in thermoelectric materials indicates their efficient conversion of a temperature difference into electrical power. 2 This efficiency enhancement results in increased power output, enhanced energy conversion efficiency, reduced energy losses, and optimal power transfer, making high power factor thermoelectric materials highly sought after for various applications.However, these thermoelectric parameters are intrinsically intertwined, posing a significant hurdle in their independent optimization.
One of the effective strategies to enhance thermoelectric performance is to reduce the dimensions, thereby introducing the quantum confinement effect in nanostructured thermoelectric materials. 3The electronic states are confined by closed geometries at the nanometer scale, leading to quantized energy levels and enhancing the density of electron states per unit volume occurring for a small well within. 4In this case, the carrier effective mass and Seebeck coefficient can be elevated rapidly.Hicks et al. demonstrated a 3-fold increase in the Seebeck coefficient value of PbTe/Pb 0.927 Eu 0.073 Te compared to bulk PbTe, attributed to the quantum confinement effect in the nanostructured system. 5As an alternative, Gao et al. fabricated PbSe/SnSe multiple quantum well structures, achieving a high power factor of 25.7 μW cm −1 K −2 at 300 K, which is 4 times larger than that of a PbSe single layer. 6une et al. implemented a SrTi 0.8 Nb 0.2 O 3 (conductor) layer between SrTiO 3 (insulator) films, resulting in an improved Seebeck coefficient from 61 to 320 μV K −1 . 7However, as the insulator barrier does not contribute to electronic transportation, there was no improvement in the carrier concentration and electrical conductivity.
In addition to high S, κ is also a crucial parameter in determining the thermoelectric performance of a material.Ren and Dow demonstrated that the thermal conductivities of multilayer thin films are lower than those of their corresponding bulk materials, employing the classical Boltzmann transport equation approach with a quantum mechanical treatment of the scattering rate. 8The presence of high-density superlattice interfaces leads to strong phonon scattering, resulting in a significant reduction in lattice thermal conductivity.The Bi 2 Te 3 (1 nm)/Sb 2 Te 3 (5 nm) superlattice  structure constructed by Winkler yielded an ultralow lattice thermal conductivity of 0.23 W m −1 K −1 . 9Tang and colleagues recently fabricated the MoTe 2 /Bi 2 Te 3 system.The interfacial scattering of phonons has resulted in a notable reduction of κ L , with values in the produced superlattice films ranging from 0.31 to 0.39 W m −1 K −1 . 10−13 While artificially engineered thermoelectric nanostructures hold great promise for synergistically optimizing transport parameters, the rational design, reliable fabrication, and the mechanism responsible for optimizing electrical properties remain elusive and pose significant challenges.Here, nanostructured thin films of Sb 2 Te 3 −Sb 2 Se 3 were successfully synthesized using thermal ALD, 14 with the composition of this nanostructure precisely controlled by adjusting the ALD cycle numbers for Sb 2 Te 3 and Sb 2 Se 3 .A systematic study of the crystal structures and transport properties revealed an obvious improvement in the thermoelectric performance upon careful optimization of the Sb 2 Se 3 sublayer thickness.The sample Sb 2 Te 3 :Sb 2 Se 3 = 5:5 (nm) exhibited an impressive Seebeck coefficient of 174 μV K −1 at room temperature.Further investigation demonstrated a high power factor of 852 μW m −1 K −2 accompanied by a peak zT value of 0.38 at room temperature.

RESULTS AND DISCUSSION
Before the structures were synthesized, the growth behaviors of single-phase Sb 2 Te 3 and Sb 2 Se 3 films were investigated.For Sb 2 Te 3 films, a combination of SbCl 3 and (Et 3 Si) 2 Te precursors was employed, whereas for Sb 2 Se 3 films, SbCl 3 and Se(SnMe 3 ) 2 precursors were utilized. 15,16The Sb 2 Te 3 and Sb 2 Se 3 thin films were grown at 80 and 110 °C without any vacuum break, respectively (see more details in the Experimental Section).The structural characterizations of Sb 2 Te 3 −Sb 2 Se 3 are shown in Figure 1 and Figure S1 in the Supporting Information.Grazing incidence X-ray diffraction (GID) measurements were performed using a custom-built laboratory setup equipped with a Mo Kα source to analyze the structural characteristics of the samples.The corresponding data reveals that Sb 2 Te 3 exhibits a rhombohedral phase with a space group of R3̅ m, while Sb 2 Se 3 adopts an orthorhombic phase with a Pbnm space group.As the thickness of the Sb 2 Te 3 layer decreases, distinctive peaks emerge from the Laue rings, signifying a highly textured film structure achieving optimal crystallinity in the sample Sb 2 Te 3 :Sb 2 Se 3 = 4:2 (nm).On the other hand, the sample Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) has a lower diffraction intensity than the other samples, suggesting a smaller grain size.
To investigate the microstructure of the Sb 2 Te 3 −Sb 2 Se 3 nanocomposites, cross-sectional imaging at atomic resolution was performed using high-resolution transmission electron microscopy in both bright-field mode (HRTEM) and highangle annular dark-field scanning mode (HAADF-STEM).A representative HAADF-STEM image of a single grain taken along [010] zone-axis orientation according to the rhombohedral Sb 2 Te 3 crystal structure is shown in Figure 2a.The latter reveals an atomic zigzag arrangement between the mono-, bi-, and multilayers highlighted by a green dashed line indicating the existence of twin boundaries. 17Furthermore, a faint brightdark atomic (Z-) contrast modulation is visible with a periodicity in the horizontal direction of about 2 nm in which the Te atoms appear brighter (Z = 52) than the Se atoms (Z = 34).In fact, this coincides with the intended alternation between Sb 2 Te 3 and Sb 2 Se 3 deposition, where in this case the Sb 2 Se 3 adopts the R3̅ m crystal structure of the Sb 2 Te 3 .This observation is also confirmed by STEM energydispersive X-ray mapping (STEM-EDX), as illustrated in Figure S2 of the Supporting Information.It should be noted that ALD growth at relatively low temperatures leads to polycrystalline layers.For low-cycle ALD of Sb 2 Te 3 , the presence of dangling bonds at the layer edges preferentially adsorbs precursor molecules. 18This phenomenon may diverge from the conventional monolayer-by-monolayer arrangement, instead leading to the formation of an expanding alloy cluster (Volmer−Weber growth behavior). 18,19Furthermore, the exchange reaction at the gas−solid interface during the ALD can contribute to alloying.The exposure of the Se precursor onto the Sb 2 Te 3 surface leads to the exchange of Se by Te atoms at the interface, resulting in the formation of alloy clusters and atomic-scale distortions. 20For example, the HRTEM image in Figure 2b depicts different grains, two of which are marked with dashed squares in yellow and orange.The corresponding Fourier transforms (FTs) of these marked regions displayed in Figure 2c,d are consistent with two different structures.The FT of the upper grain (Figure 2c) shows reflections corresponding to the [324] orientation of the orthorhombic phase with a Pbnm space group, whereas the FT of the lower grain (Figure 2d) reveals, e.g., the 006 reflection, indicating a direction parallel to the c axis of the R3̅ m space group.The thicker Sb 2 Te 3 layer in the structure facilitates a larger potential drop at the Sb 2 Te 3 /Sb 2 Se 3 interface, leading to a higher carrier concentration compared to a thinner Sb 2 Te 3 layer. 6However, a slight decrease in the carrier concentration is observed when the Sb 2 Te 3 to Sb 2 Se 3 ratio is further reduced.Specifically, the carrier concentration of the Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) sample is lower than that of Sb 2 Te 3 :Sb 2 Se 3 = 4:2 (nm).In the Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) sample, the Sb 2 Te 3 layer is too thin to effectively shield the built-in potential, resulting in carrier pinning at the interface. 6Under these conditions, charge transfer occurs more readily in Sb 2 Se 3 with a thicker Sb 2 Te 3 layer, leading to a higher carrier concentration.The highest electrical conductivity of 327 S cm −1 was achieved at Sb 2 Te 3 :Sb 2 Se 3 = 6:2 (nm) (Figure 3b).The Hall mobility of the systems falls within the range of 90−120 cm 2 V −1 s −1 , which is lower than a value of 140 cm 2 V −1 s −1 observed in the Sb 2 Te 3 layer (Figure 3c).Interface scattering arises when carriers traverse between different layers, leading to mobility restrictions and impacting overall mobility. 25Additionally, potential barriers within the superlattice structure can obstruct the interlayer transport of charge carriers, further reducing mobility. 26Furthermore, interface effects can induce the formation of interface states, which act as scattering centers for carriers, contributing to additional charge scattering and a subsequent decrease in mobility. 27Conversely, an inverse trend is evident in the Seebeck coefficient results.For instance, as depicted in Figure 3d, the lowest Seebeck coefficient of 125 μV K −1 is obtained at Sb 2 Te 3 :Sb 2 Se 3 = 6:2 (nm), while the Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) sample exhibits the highest value of 146 μV K −1 .The room temperature transport properties correspond well with the theoretical calculation results based on replacement modes, which can be supported by density functional theory results in Table S2 and Figures S3 and S4 in the Supporting Information.The bandgap is minimally affected by the strain induced by the lattice mismatch between Sb 2 Te 3 and Sb 2 Se 3 .Instead, it is the states far from the Fermi energy that are primarily influenced.The strain can alter the thermoelectric coefficient slightly due to the changes in the electronic states.Further discussion on lattice mismatch can be found in the Supporting Information (Table S3 and Figure S5).
The single parabolic band (SPB) model is the most widely employed approach for evaluating the electrical transport properties of thermoelectric materials. 28,29The relationship between the Seebeck coefficient and carrier concentration at room temperature was determined using the Pisarenko plot, as shown in Figure S6 in the Supporting Information.The effective mass (m*) increases from 0.38m e to 0.65m e with the rising doping level of Sb 2 Se 3 into the Sb 2 Te 3 system.Notably, the m* of the Sb 2 Te 3 −Sb 2 Se 3 layer is significantly higher than that of the Sb 2 Te 3 layer, indicating the presence of band bending and the energy filtering effect in the heterostructure system.Additionally, the incorporation of the quantum confinement effect in the multilayer system further enhances m*, given that the thickness is smaller than the de Broglie wavelength of electrons. 30The power factor was calculated using the electrical conductivity and Seebeck coefficient data (Figure 3e).A maximum PF value of 605 μW m −1 K −2 was determined for the sample Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) at room temperature, primarily attributed to the high Seebeck coefficient.The sample Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) appears more semimetallic, as indicated by the DOS not approaching zero in the bandgap region (Figure 3f).Moreover, a slight decrease in the bandgap is observed, accompanied by a slight shift and enhancement of the Fermi level, contributing to the improved Seebeck coefficient.
To investigate the electrical properties of Sb 2 Te 3 −Sb 2 Se 3 nanostructures while minimizing the influence on their alloyed structure, samples with higher thickness ratios were fabricated.−37 According to Wiedemann−Franz law, the electronic (κ E ) and lattice (κ L ) thermal conductivities can be carried out by the following formulas: where L is the Lorenz number, T is the temperature, and S is the Seebeck coefficient.The calculated κ L is also shown in Figure 5a.In the Sb 2 Te 3 −Sb 2 Se 3 system, lattice thermal conductivity primarily contributes to the total thermal conductivity.Generally, increasing the interface density of thin films decreases κ L , which adequately explains the overall change trend of κ L . 10 Interestingly, the lattice thermal conductivity of the sample with a composition ratio of Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) (0.46 W m −1 K −1 ) is lower than that of the sample with a composition ratio of Sb 2 Te 3 :Sb 2 Se 3 = 5:5 (nm) (0.51 W m −1 K −1 ), suggesting that additional scattering occurs in the former system.This should be attributed to the different weightings of phonon scattering mechanisms in alloyed structures and quantum confinement effects in superlattice structures, resulting in distinct impacts on thermal conductivity.In systems where the average mean free path of phonons exceeds the limiting dimension of the sample, heat transport does not follow Fourier's law, exhibiting semiballistic phonon transport characteristics. 38,39This regime highlights the wave nature inherent in phonons, with the potential to sustain coherence and consequently reduce the κ. 40Low-dimensional nanostructures exhibit favorable TE performance due to quantum confinement effects. 41,42For the Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) sample, the Volmer−Weber growth mode results in the formation of structural features such as alloy compositions, point defects, or phase boundaries as well as superlattice structures.These structural characteristics collectively contribute to enhanced phonon scattering, consequently reducing the thermal conductivity (Figure 5b,c).In comparison, the Sb 2 Te 3 :Sb 2 Se 3 = 5:5 (nm) sample exhibits a more ordered structure, and the quantum confinement effect is the primary factor responsible for decreasing the thermal conductivity.The rational design of the work function difference will lead to facilitate directional hole injection into the coherent region through modulation doping.To establish charge equilibrium at the interface, carrier redistribution takes place, leading to band bending.Our theoretical simulations provided strong confirmation of the results (see Tables S5−S7 in the Supporting Information).Phonons with mid-to longwavelengths scattering are enhanced at the heterojunction interfaces, which is considerably more efficient than scattering at normal grain boundaries. 22,26,43Consequently, the m* increases and the thermal conductivity decreases simultaneously, leading to a high zT value of 0.38 for the sample Sb 2 Te 3 :Sb 2 Se 3 = 5:5 (nm).

CONCLUSIONS
In conclusion, we designed Sb 2 Te 3 −Sb 2 Se 3 nanostructured thin films by using thermal ALD.The transport properties, including the carrier concentration, electrical conductivity, Seebeck coefficient, and thermal conductivity, were evaluated.
In the Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) system, Volmer−Weber growth introduced alloy and defect structures that reduce the carrier concentration and Hall mobility.Through meticulous optimization of the sublayer thickness of Sb 2 Se 3 , a notable improvement was achieved with a high Seebeck coefficient of 174 μV K −1 , demonstrating a 35% enhancement compared to the pure Sb 2 Te 3 layer.As a result of the increase in S and σ for the Sb 2 Te 3 −Sb 2 Se 3 systems, a high power factor reaches up to 852 μW m −1 K −2 for the sample Sb 2 Te 3 :Sb 2 Se 3 = 5:5 (nm).
The results indicated that the relative contributions of phonon scattering mechanisms and quantum confinement effects could significantly influence the overall thermoelectric performance of nanostructured thin films.This study provides valuable insights into the design and fabrication of high-performance chalcogenides via the ALD technique.

EXPERIMENTAL SECTION
Fabrication of Sb 2 Te 3 and Sb 2 Se 3 Thin Films.The Sb 2 Te 3 and Sb 2 Se 3 thin films were grown using a thermal ALD reactor (Veeco Savannah S100) at 80 and 110 °C, respectively.SbCl 3 and (Et 3 Si) 2 Te were used for Sb 2 Te 3 films, while Se(SnMe 3 ) 2 and SbCl 3 were used for Sb 2 Se 3 synthesis.During the ALD process, the SbCl 3 , (Et 3 Si) 2 Te, and Se(SnMe 3 ) 2 were kept at 60, 77, and 60 °C, respectively.Highpurity N 2 was used as the carrier gas, and the chamber was kept at a flow rate of 10 sccm during the reaction process.The optimized pulse and purge times for one ALD deposition (precursor 1/N 2 /precursor 2/N 2 ) were 1/10/1/10 s.The detailed ALD process is shown in Table S1 (Supporting Information).The sample identifier Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) indicates that the individual thicknesses of Sb 2 Te 3 and Sb 2 Se 3 are both 2 nm.The total thickness for all samples is approximately 100 nm.
Characterization of Morphology, Electrical, and Thermal Properties.The thickness of thin films was measured using X-ray reflectometry (X'Pert MRD PRO).Gracing incident X-ray diffraction (GID) was performed using a custom-made laboratory setup with a Mo Kα source (λ = 0.713 Å).We aligned the surface of the films at an angle of 0.5°with respect to the incident X-ray beam.The samples were then rotated about their surface normal, and diffraction images for scattering angles up to 40°were recorded using a 2D pixel detector.The detector images were transformed into reciprocal space, and the diffracted intensity was regrouped into the scattering angle (2θ) and the polar angle along the Laue circles (Χ) using the pyFAI package. 44The morphology and atomic structure of thin films were characterized by transmission electron microscopy using a doublecorrected (aberration correction in scanning and bright-field mode) ThermoFisher Titan 80-300 operated at a 300 kV acceleration voltage.Thin films were deposited on SiO 2 and Si 3 N 4 substrates (TFA chip from Linseis company) for structural characterization and transport property measurements, respectively. 45,46The detailed platform measurement can be found in the Supporting Information.
Theoretical Calculations.The initial structures were created using the fully relaxed Sb 2 Te 3 bulk structure.Sb 2 Te 3 consists of three rhombohedral stacked layers each consisting of five atomic layers with a space group of R3̅ m.The heterostructure systems were made by replacing the Te with Se within one layer to form different mixing ratios of Sb 2 Te 3 /Sb 2 Se 3 layers.Later, the effect of precursor mixing was also studied by Se−Te substitutions within one layer in the 2:2 nm system (Table S2, Supporting Information).Additionally, we also investigated Se Te defects by replacing a single Te atom with Se.For creating various concentrations, 1 × 1 × 1, 2 × 2 × 1, and 3 × 3 × 1 supercells of the hexagonal unit cell were considered.The structures were relaxed until the forces were below 1 meV/Å.The electronic structures were obtained using density functional theory (DFT) with the Perdew−Burke−Ernzerhof (PBE) 47 exchange-correlation functional as implemented in FHI-aims. 48It is well known that hybrid functionals such as HSE06 can lead to an opening of the bandgap, 49 correcting (sometimes "over"-correcting) the underestimation by (semi)local functionals.On the other hand, the purpose of this study is to understand the effects of structures and interface variations on the thermoelectric coefficients.It has been shown that (semi)local functionals such as PBE can produce satisfactory results. 50,51Scalar relativistic corrections (ZORA), spin−orbit coupling (SOC) effects, and nonlocal many-body dispersions were included in the computations as described in the references. 52,53Subsequently, transport coefficients were calculated based on the Boltzmann transport equation within the constant−relaxation−time approximation employing BoltzTrap2. 54
Additional characterizations such as out-of-plane XRD patterns; EDS mapping for elemental distribution; detailed DFT simulations of defect models, lattice mismatch, and interfacial charge transfer; and transport property measurement platform (PDF)

Figure 2 .
Figure 2. Structural characterizations of Sb 2 Te 3 −Sb 2 Se 3 thin films.(a) HAADF-STEM image of a Sb 2 Te 3 −Sb 2 Se 3 multilayer grain taken along its [010] zone-axis orientation and c axis in the vertical direction (scale bar: 5 nm).The inset shows a zoom-in of the region marked by the red dashed box, in which the atomic Sb 2 Te 3 model is overlaid (scale bar: 1 nm).The orange arrows highlight the Sb 2 Se 3 regions, which appear slightly darker in the image due to their lower Z-contrast.(b) HRTEM images show different grains of the poly crystalline structure (scale bar: 10 nm).(c, d) Fourier transforms (FTs) of regions marked by orange (c) and yellow (d) dashed boxes in panel (b) (scale bars: 5 nm −1 ).(c) FT shows reflections corresponding to the [324] orientation of the orthorhombic phase with a Pbnm space group.(d) The FT shows, e.g., the 006 reflection, indicating a direction parallel to the c axis of the R3̅ m space group.

Figure 3
depicts the room temperature electrical transport properties of the Sb 2 Te 3 −Sb 2 Se 3 structures.The carrier concentration of the Sb 2 Te 3 thin film is 1.18 × 10 19 cm −3 and reaches a peak value of 2.61 × 10 19 cm −3 in the Sb 2 Te 3 :Sb 2 Se 3 = 4:2 (nm) structure.The carrier concentration of Sb 2 Se 3 is 3.6 × 10 18 cm −3 at room temperature.Considering the bandgaps for Sb 2 Te 3 and Sb 2 Se 3 as 0.3 and 1.14 eV, respectively, a band offset occurs for the Sb 2 Te 3 −Sb 2 Se 3 heterostructure. 21−23 The significantly higher work function difference between Sb 2 Se 3 and Sb 2 Te 3 results in a pronounced hole junction from Sb 2 Se 3 to Sb 2 Te 3 .

Figure 4 .
The Sb 2 Te 3 −Sb 2 Se 3 samples exhibit distinct transport behavior compared to the Sb 2 Te 3 layer.At room temperature, the highest n of 1.78 × 10 19 cm −3 was achieved for the sample Sb 2 Te 3 :Sb 2 Se 3 = 6:4 (nm).The electrical conductivity of Sb 2 Te 3 −Sb 2 Se 3 structures showed minimal variation compared to the Sb 2 Te 3 single layer.The highest σ of 295.45 S cm −1 was achieved for the Sb 2 Te 3 :Sb 2 Se 3 = 8:2 (nm) sample, closely followed by the value of 291.23 S cm −1 for Sb 2 Te 3 .All thin films exhibited positive S values, confirming their p-type semiconducting nature.A Seebeck coefficient of 174 μV K −1 was obtained at room temperature, representing a 35% enhancement compared to the Sb 2 Te 3 layer.This significant increase in the Seebeck coefficient for the Sb 2 Te 3 −Sb 2 Se 3 superlattice layers resulted in an exceptional power factor reaching up to 852 μW m −1 K −2 at room temperature.The thermal conductivity of Sb 2 Te 3 −Sb 2 Se 3 films was evaluated at room temperature using the 3-ω technique (Figure 5a).At room temperature, the Sb 2 Te 3 −Sb 2 Se 3 layers exhibited a lower total thermal conductivity (κ Total ) compared to pure Sb 2 Te 3 layers.The sample with a composition ratio of Sb 2 Te 3 :Sb 2 Se 3 = 2:2 (nm) demonstrated an impressive minimum thermal conductivity of 0.62 W m −1 K −1 , which is almost 2 times lower than the total thermal conductivity of Sb 2 Te 3 (1.23 W m −1 K −1 ).The obtained thermal conductivity is comparable to that of multilayered structures such as Sb 2 Te 3 −MoS 2 but lower than alloys like Te 0.85 (Sb 2 Se 3 ) 0.06 or bulk Sb 2 Te 3 or Sb 2 Se 3 (Table

Figure 5 .
Figure 5. (a) Total thermal conductivity and lattice thermal conductivity of Sb 2 Te 3 −Sb 2 Se 3 thin films and the comparison with reported data at room temperature. 10,31−37 The models of (b) alloy and (c) superlattice structures in the thin films.(d) zT values for various samples at room temperature.