Large Thermoelectric Power Factor in Pr-Doped SrTiO3−δ Ceramics

Article pubs.acs.org/cm

Large Thermoelectric Power Factor in Pr-Doped SrTiO3−δ Ceramics via Grain-Boundary-Induced Mobility Enhancement Arash Mehdizadeh Dehkordi,*,† Sriparna Bhattacharya,‡ Taghi Darroudi,⊥ Jennifer W. Graff,† Udo Schwingenschlögl,§ Husam N. Alshareef,§ and Terry M. Tritt*,†,‡ †

Department of Materials Science and Engineering, Clemson University, Clemson, South Carolina 29634, United States Department of Physics and Astronomy, Clemson University, Clemson, South Carolina 29634, United States ⊥ Electron Microscope Facility, Clemson Research Park, Clemson University, Clemson, South Carolina 29634, United States § King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia ‡

S Supporting Information *

ABSTRACT: We report a novel synthesis strategy to prepare high-performance bulk polycrystalline Pr-doped SrTiO3 ceramics. A large thermoelectric power factor of 1.3 W m−1 K−1 at 500 °C is achieved in these samples. In-depth investigations of the electronic transport and microstructure suggest that this significant improvement results from a substantial enhancement in carrier mobility originating from the formation of Pr-rich grain boundaries. This work provides new directions to higher performance oxide thermoelectrics as well as possibly other properties and applications of this broadly functional perovskite material.

■

INTRODUCTION The interest in the electronic transport properties of SrTiO3 (strontium titanate)-based perovskite systems originated with the report of superconductivity in oxygen-deficient semiconducting strontium titanate.1 In the following years, the functionality of SrTiO3-based materials has expanded over a wide range of intriguing phenomena and properties including quantum paraelectricity,2 giant dielectric permittivity,3,4 roomtemperature ferroelectricity,5 room-temperature photoluminescence,6,7 mixed ionic-electronic conductivity,8,9 bistable electrical resistance switching,10,11 and interfacial two-dimensional electron gas12,13 and more recently has exhibited some very interesting thermoelectric properties.14 Exploration of thermoelectric properties of strontium titanate and other oxides began after the report of Terasaki et al. in 1997 showed that p-type NaCoO2 single crystal possesses a higher room temperature power factor (defined herein as PF = α2σT, where α is the Seebeck coefficient, σ is the electrical conductivity, and T is the temperature in Kelvin) of ∼1.5 W m−1 K−1 as compared to ∼1.2 W m−1 K−1 for Bi2Te3, a state-of-the-art commercial thermoelectric material.15 Further investigations of oxide thermoelectrics highlighted layered alkali or alkaline-earth cobaltite compounds (Na x CoO 2 16 and Ca 3 Co 4 O 9 ), 17 KxRhO2,18 and recently BiCuSeO19 as the most promising ptype oxide materials, while SrTiO3, CaMnO3,20 ZnO,21 and very recently (SrxBa1−x)Nb2O622,23 were being investigated as potential n-type candidates. © 2014 American Chemical Society

The interest in the thermoelectric (TE) properties of SrTiO3 (“STO”) arose when Okuda et al. reported a high roomtemperature power factor of ∼1.08 W m−1 K−1 for heavily Ladoped STO single crystals.14 The electronic transport in SrTiO3 can be tuned over a wide range of physical properties. The electrical conductivity of n-type STO materials can be achieved over a very broad range: from insulating to metallic through a combination of either of the two following doping mechanisms: (i) substitutional doping of Sr2+ or Ti4+ sites with higher valence elements (e.g., La3+ for Sr2+ sites or Nb5+ for Ti4+ sites) and/or (ii) creating oxygen vacancies. Moreover, a large carrier effective mass (m* ∼ 2−16me)24−26 and conduction band degeneracy leads to a large Seebeck coefficient at high carrier concentrations. However, a low carrier mobility (μ ∼ 6 cm2 V−1 s−1 at 300 K for single crystals)27−29 and large total thermal conductivity (κ ∼ 12 W m−1 K−1 at 300 K for single crystals)29 have so far seriously limited their potential for thermoelectric applications. The large total κ is dominated by a large lattice contribution (κL) along with the low carrier mobility that detrimentally affects the thermoelectric performance of these STO materials. The potential of a material for thermoelectric application is primarily evaluated by a dimensionless number called the figure of merit, ZT, defined as ZT = (α2σT)/κ, where PF = α2σT is defined here as the Received: December 12, 2013 Revised: January 21, 2014 Published: January 22, 2014 2478

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

power factor and κ is the total thermal conductivity. The highest ZT values reported for bulk polycrystalline SrTiO3 fall between 0.3 and 0.35 at approximately 1000 K with maximum power factor values of PF ≤ 1.0 W m−1 K−1 (see Figure S1 of the Supporting Information). In order for this oxide thermoelectric material to compete with other high-temperature thermoelectrics, a more pronounced increase in the power factor and/or decrease in lattice thermal conductivity are required. The primary approach that has been adopted thus far, in order to achieve a significant enhancement in the thermoelectric performance, has mainly focused on the reduction of the thermal conductivity via lattice distortions and mass fluctuation scattering mechanisms through dopant selection and optimization of various synthesis methods. Single-30−37 or double-doping38−40 of the Sr2+ sites with rare-earth elements and double-doping of both the Sr2+ and Ti4+ sites with rare-earth elements and transition metals41−43 comprise the main efforts with respect to this direction. Synthesis of layered perovskite-type Ruddlesden− Popper structures is also another attempt to reduce thermal conductivity.44 In these approaches, the synthesis parameters were optimized in order to achieve a lower thermal conductivity, with much less attention being given to the optimization of electronic transport properties. It should be noted that, due to the very small phonon mean free path in SrTiO3 (lph ∼ 2 nm at 300 K for single crystal),45 reduction in thermal conductivity through grain boundary scattering using nanostructuring techniques is not a viable option for bulk ceramics. This fact highlights and further strengthens the importance of the need for mechanisms for the enhancement of the power factor in order for further improvements of the TE performance for SrTiO3-based oxides. In order to have a significant improvement in the power factor of these structures, mobility-enhancing mechanisms are the most desirable. Recently, dielectric-constant modulation was introduced as a method with which to improve the low carrier mobility of single crystal complex oxides below room temperature.46 Deltadoping47 and strain modulation of SrTiO3 thin and epitaxial films are among the methods reported to enhance the lowtemperature carrier mobility of SrTiO3.48 However, to the best of our knowledge, there has been no report of any significant (more than 10%) enhancement of the mobility of SrTiO3 oxides at or above room temperature. In this paper, we present a detailed report on the electronic transport properties of praseodymium (Pr)-doped SrTiO3 ceramics and show that Pr doping can bring about an additional enhancement in the thermoelectric performance of doped SrTiO3. Bulk polycystalline Sr1−xPrxTiO3 samples with x = 0.125 show, to the best of our knowledge, the highest ever reported values of the power factor of ∼1.3 W m−1 K−1 at 500 °C among n-type doped SrTiO3 ceramics and single crystals. This significant enhancement in power factor is a result of a much improved carrier mobility (by a factor of ∼2) in these ceramic materials. The synthesis strategy utilized in this work may open new horizons and opportunities to other properties and applications of this broadly functional perovskite material system.

■

nanopowder (99.5% ; Aldrich) were mixed, cold pressed into pellets, and then calcined in air at temperatures of 1300−1400 °C with intermediate grinding. The pellets were quenched in the furnace to room temperature after each calcination process. The calcined pellets were subsequently pulverized into powders using a mortar and pestle. The resulting powders were then wrapped in graphite foil and loaded into graphite dies. Powders were solidified into disks (≈12.7 diameters and 3 mm thick) using the spark plasma sintering (SPS) technique (Dr. Sinter Lab, SPS-515S) under a vacuum at 1400−1500 °C for 5 min. Samples were SPSed at ∼300 °C min−1 heating rates (except the sample discussed in Figure 2, which was SPSed at ∼100 °C min−1). All samples were sintered under the same current pulse pattern on/off ratio. All samples were polished down for 0.25−0.5 mm from each side to ensure the complete removal of the graphite foil. Densities of all the samples were determined using the Archimedes method, and they were all higher than 95% of their theoretical values. Transport Measurements. Rectangular bars (10 mm × 2 mm × 2 mm) were cut from the disks for high and low temperature measurements of electrical conductivity (σ) and Seebeck coefficient (α). Low temperature electrical conductivity and Seebeck coefficient measurements were performed using a custom-designed four-probe measurement system from 20 to 300 K.71 High temperature measurements of electrical conductivity (σ) and Seebeck coefficient (α) were conducted using an Ulvac-Riko ZEM-2 (300−800 K) system. Carrier concentrations and carrier mobilities of the samples were determined with a Quantum Design Hall system using a five-probe configuration under low (5000 Oe) and high (3 T) magnetic fields (10−300 K). All the transport measurements reported for a composition in this work were performed on the same sample. Uncertainty in the measurements of high-temperature electrical conductivity, Seebeck coefficient, and thermoelectric power factor are ±3, ±2, and ±7%, respectively. Structural/Chemical Analysis. The structure and morphology of the powders and the bulk samples were studied using a Rigaku Ultima IV high resolution X-ray diffractometer with Cu Kα doublet radiation and a Hitachi SU-6600 field emission scanning electron microscope. Energy dispersive scanning X-ray spectroscopy (EDS, Oxford Instruments) was also performed to investigate the chemical composition of the specimens and their grain boundaries. Rietveld refinement analysis was performed on collected X-ray diffraction (XRD) patterns using a PDXL software package (Rigaku, version 1.8.1.0). The oxidation state of Pr in ceramics was characterized by Xray photoelectron spectroscopy (XPS) with a Kratos Axis Ultra DLD spectrometer using a monochromatic Al Kα X-ray source.

■

RESULTS AND DISCUSSION The X-ray diffraction profiles of the Sr1−xPrxTiO3 powders for x = 0, 0.05, 0.075, 0.1, and 0.125 are shown in Figure 1a. All the diffraction peaks of the undoped powder (i.e., x = 0) can be indexed to the SrTiO3 cubic perovskite structure with the space group Pm3̅m, and no other phases were observed. However, with increasing the nominal Pr content of the powders above x > 0.05, small diffraction peaks corresponding to the Pr5O9 phase (monoclinic, space group P21/c) were identified.49 The intensity of these peaks increases with increasing Pr concentration, as shown in Figure 1b. To the best of our knowledge, there are no reports on the solid solubility of Pr2O3 in SrTiO3 (SrO−TiO2 system). A shift to higher angles is observed in all the STO peaks with an increase in the Pr content of the powders prepared in this work, which indicates the shrinkage of the lattice by substitution of smaller Pr3+ ions (0.112 nm for the coordination number 8)50 in Sr2+ (0.126 nm for the coordination number 8)50 sites. Refined lattice parameters calculated from Rietveld analysis exhibit this shrinkage of the lattice from a = 0.3905(7) nm for x = 0.05 to a = 0.3903(5) nm for x = 0.125.

EXPERIMENTAL SECTION

Synthesis. Pr-doped strontium titanate powders Sr1−xPrxTiO3 (x = 0, 0.025, 0.05, 0.075, 0.1, 0.125, and 0.15) were prepared using a solidstate reaction process. Stoichiometric amounts of SrCO3 powder (99.9%; Aldrich), Pr2O3 sintered lumps (99.9%; Alfa Aesar), and TiO2 2479

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

oxidation state of Pr in the ceramic materials made with the optimum SPS heating rate and the PrO2 (Pr4+) signature peak is absent in the Pr spectrum (Supporting Information, Figure S2). However, the effect of the multivalent nature of Pr, which differentiates Pr from other reported A-site dopants, on the kinetics of phase formation is not yet fully understood and needs to be further investigated. The effect of SPS heating rate on the modification of electronic transport, particularly the temperature dependence of electrical conductivity, is shown in Figure 2a for two ceramics prepared from the same Sr1−xPrxO3‑δ powder with x = 0.075. It is observed that the two samples possess similar temperature-independent carrier concentrations (n ∼ (1.10 ± 0.2) × 1021 cm−3) and Seebeck coefficient (thermopower) values (Figure 2a, inset). However, a significant improvement in the electrical conductivity (∼60% improvement at room temperature) was achieved through enhancing the carrier mobility. It should be noted that both samples possess the same density. Such an increase in the electron mobility and electrical conductivity thus leads to a significant enhancement (∼30%) in the thermoelectric power factor over the whole temperature range that has been achieved primarily through tuning the SPS heating rate. The proposed doping mechanism is depicted schematically in Figure 2a for both samples. As it is shown in the backscattered electron (BSE) micrographs in Figure 2b and c, the Pr5O9 particles present by the grain boundaries in the asprepared powder (Figure 2b, inset) form Pr-rich grain boundaries during SPS in both samples. However, the Pr5O9 species, in the form of PrOy (1.5 < y < 1.8) phases, are occasionally observed in the grain boundaries for the samples made by typical SPS heating rates of ∼100 °C min−1 (corresponding to 50 A min−1 current rate), as shown in Figure 2b and the schematic on the left. The EDS line scan performed across the bright praseodymium oxide region shows the typical sharp increase in the Pr counts (see Figure S3 of the Supporting Information for detailed line scan spectra). We found that by increasing the SPS heating rate (up to 300 °C min−1) enough energy can be provided for the Pr5O9 particles in the as-prepared powder to fully and locally dope the grain boundary region, as shown in Figure 2c. Such in situ Pr-doped SrTiO3 samples, in essence, can be thought of as twocomponent “core−shell-like” composites with both core and shell features being a Pr-doped SrTiO3, with the grain boundary phase having higher concentrations of Pr dopants. As we will see, these Pr-rich grain boundaries (shells) play a decisive role in improving the carrier mobility and hence the thermoelectric power factor. It is quite obvious in the inset of Figure 2c that such boundaries were not observed for the samples doped with other dopants (e.g., La) following the same recipe. Much thicker average Pr-rich grain boundaries (∼1 μm) in Figure 2c, versus that of the sample made with lower heating rate (∼200 nm), suggest a more localized distribution of dopants in the vicinity of the grain boundaries in this sample. Diffusion of the dopants from the grain boundaries to within the grains in the sample made with the lower SPS heating rate is believed to be the reason behind the reduction in the average grain boundary thickness. We believe this was facilitated by much lower heating rate and thus consequently longer SPS process time at elevated temperatures. Temperature dependence of electronic transport properties of Sr1−xPrxTiO3−δ ceramics as a function of x is plotted in

Figure 1. (a) X-ray diffraction (XRD) profiles of Sr1−xPrxTiO3−δ powders before SPS as a function of nominal Pr content. (b) Close-up view of the dashed rectangle in part a, showing an increase in the praseodymium oxides content of the powders with increasing doping concentration. (c) Comparison of XRD profiles of Sr1−xPrxTiO3−δ with x = 0.075 before SPS and after high-heating-rate SPS. Photographed images of cold-pressed powder after solid-state reaction and the corresponding SPSed ceramics are shown. The change in color is due to a change in the oxidation state of Ti4+ to Ti3+. (d) Close-up view of the dashed rectangle in part c, showing the disappearance of the praseodymium oxide peaks after high-heating-rate SPS.

Bulk ceramics in this work were solidified from the asprepared powders using the spark plasma sintering (SPS) technique. We found that the SPS heating rate can play a crucial role with respect to the modification of the electronic transport in these ceramics. To the best of our knowledge, no such analogous observations have been reported for oxide ceramic materials. Figure 1c shows the X-ray diffraction profiles of the as-prepared powder with x = 0.075 before SPS and the one of the ceramic solidified with optimized SPS conditions, corresponding to the sample shown with triangle markers in Figure 2a. A shift to higher angles is observed for all the STO peaks, implying a further shrinkage of the lattice in the ceramic samples. This might suggest a further incorporation of Pr ions into the STO lattice and/or creation of oxygen vacancies during SPS. Figure 1d exhibits the complete disappearance of the main Pr oxide peaks observed in the powder after SPS using an experimentally optimized heating rate. It is known that the solid state reaction kinetics in these oxides is sensitive to the reaction atmosphere and temperature as well as the choice and size of raw materials. We have taken advantage of the highly reducing atmosphere and high heating rate of the SPS technique to control the kinetics of the reaction of the residual Pr oxide in the as-prepared powder with the surrounding partially doped SrTiO3 grains (Figure 2b, inset). X-ray photoelectron spectroscopy (XPS) results confirmed that Pr3+ is the dominant 2480

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

Figure 2. (a) Temperature dependence of electrical conductivity and Seebeck coefficient (inset) for Sr1−xPrxTiO3−δ ceramic with x = 0.075 prepared using low (100 °C min−1) and high (300 °C min−1) SPS heating rates. The schematic on the right shows the presence of praseodymium oxide (PrOy) particles in the partially formed Pr-rich grain boundaries upon applying low heating rate (corresponding to 50 A min−1 current rate). The schematic on the left indicates the complete dissolution of the praseodymium oxide particles in the Pr-rich grain boundaries under high heating rate (corresponding to 250 A min−1). (b) Backscattered electron (BSE) micrograph of the ceramic made under a SPS low heating. A typical Pr spectrum of the EDS line scan across a PrOy particle is shown. The inset shows the BSE micrograph of corresponding powder showing Pr5O9 particles sitting by the grain boundaries of Pr-doped SrTiO3 grains. (c) Backscattered electron micrograph of the ceramic made under a high SPS heating rate. A typical Pr spectrum of EDS line scan across two grains, grain 1 and grain 2, is shown. The inset depicts the BSE micrograph of the Sr0.95La0.05TiO3 ceramic prepared following the same recipe showing no such grain boundaries.

In fact, our experimental results show a more relaxed temperature dependence for Hall mobility (T−1.5), close to room temperature, than reported in the literature (T−M, 2.0 < M < 3.2)27,29,53−56 (see Figure S6 of the Supporting Information). In view of the composite picture, theoretical efforts were made to explain the effective electrical conductivity of the composite from that of its constituents, namely, the grain and the Pr-rich grain boundary phase. However, effective medium theories such as Bergman−Fel57 fail to explain the effective properties (see sections S7 and S8 of the Supporting Information). This might suggest the presence of a charge transfer mechanism or a contribution from the phase interface which were not accounted for in the Bergman−Fel model. The temperature dependence of the Seebeck coefficient (α) is depicted in Figure 3b. Diffusive-like thermopower is apparent for all the samples over the entire temperature range of this investigation (25−500 °C). No sign of any minority carrier

Figure 3. These samples were all prepared with a high SPS heating rate (∼300 °C min−1). Figure 3a shows electrical conductivity (σ) as a function of temperature for these samples. All samples exhibit a degenerate semiconducting behavior, i.e., a decreasing electrical conductivity with increasing temperature. It is also observed that the electrical conductivity increases with increasing Pr content. However, this increase is more significant for temperatures close to the room temperature, T < 200 °C (473 K). The dominant charge carrier scattering mechanism in STO at this temperature region is a combination of polar optical phonon scattering, originating from the ionic nature of the STO lattice (Debye temperature, ΘD ∼ 390−413 K for undoped STO),51,52 and scattering from the deformation potential of acoustic phonons.29 It is possible that the presence of the Pr-rich boundaries has alleviated the polar optical phonon scattering of carriers at these temperatures. 2481

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

material system (see Figure S1 of the Supporting Information). It should be noted that the PF exhibits values above 1 W m−1 K−1 over a wide temperature range (T > 200 °C) for samples with x > 0.075. All the samples are reproducible and the transport measurements are repeatable with very good accuracy up to 500 °C, the upper limit of our investigation, under a helium atmosphere. A maximum power factor value as high as PF ∼ 2 W m−1 K−1 can be predicted at 1000 °C by fitting the experimental transport properties, if the measurement is to be performed under a highly reducing atmosphere (see Figure S9 of the Supporting Information). In order to investigate the origins of the excellent electronic transport properties in Pr-doped SrTiO3−δ ceramics, Hall measurements were performed on these samples. Figure 4a shows the carrier concentration as a function of nominal Pr content. All samples were synthesized under the same solidstate reaction and spark plasma sintering (SPS) conditions (as described in the Experimental Section). A linear increase in carrier concentration is observed with increasing Pr concentration in the samples. It is also apparent that no unusual behavior, such as self-compensation or a change in Pr incorporation as a function of doping concentration, is observed. Excellent agreement between Hall carrier concentration and as predicted by simple electron counting is observed, suggesting the incorporation of the majority of Pr dopants in Sr sites. The measured Hall carrier concentration for undoped STO (red marker) suggests the creation of a large density of oxygen vacancies during the SPS process. Figure 4b visualizes the dependence of effective mass (m*) on carrier concentration (n) and Pr doping. Seebeck coefficients of undoped (oxygen deficient) and Pr-doped samples were plotted versus their corresponding Hall carrier concentrations at 300 K along with a so-called “Pisarenko plot”58 solid line which is described by59,60 ⎛ 8π 2k 2 ⎞ ⎛ π ⎞2/3 B ⎜ ⎟ ⎟ m T (1 + r ) α=⎜ * 2 ⎝ 3n ⎠ ⎝ 3eh ⎠

(1)

where kB is the Boltzmann constant, e is the electronic charge, h is the Planck constant, m* is the effective mass, n is the carrier concentration, and r is the scattering parameter. The scattering parameter, r, determines the energy dependence of the relaxation time in the power law form, τ(E) = τ0Er−1/2. Deviations from this baseline curve described by eq 1 are easily observed if m* changes. It is found that the experimental transport data are well-described by a single parabolic band model (solid curve) for r = 0.529,58 (for ionic lattices) and m* = 3.83me at 300 K (see section S10 of the Supporting Information). It is also apparent that a slight increase in the effective mass occurs upon doping of the SrTiO3−δ lattice with Pr (for pure SrTiO3−δ ceramic, an effective mass of m* = 3me is estimated from eq 1 at 300 K with r = 0.5). By comparing the α (μV K−1) vs n (cm−3) values of Pr-doped ceramics to our other STO samples doped with different dopants (e.g., La61 or Tb62) as well as other reports in the literature (see Figure S5 of the Supporting Information), a similar effective mass behavior was observed for ceramics doped with different dopants. It can be concluded that the improvement in the mobility of these ceramics is not rooted in any band structure modification mechanisms through Pr doping. Figure 4c shows the room temperature mobility values calculated from Hall carrier concentration and electrical conductivity (σ = neμ, where σ is the electrical conductivity, n is the Hall carrier concentration, e

Figure 3. Temperature dependence of (a) electrical conductivity (σ), (b) Seebeck coefficient (α), and (c) power factor (defined as PF= α2σT) for Sr1−xPrxTiO3−δ ceramics as a function of Pr content. Power factor values of PF > 1.0 W m−1 K−1 were achieved over a wide temperature range and doping concentration (x > 0.05).

contributions and bipolar effects are observed, in agreement with what would be expected due to the large band gap of this material. Low temperature (10−300 K) measurements of the transport properties using a custom-designed four-probe measurement system71 show excellent agreement with our high temperature measurements, and the diffusive-like nature of the thermopower is extended to these low temperatures (see Figure S4 of the Supporting Information). The thermoelectric power factor (PF = α2σT) of these samples as a function of temperature and Pr content is plotted in Figure 3c. The power factor is increasing with an increase in Pr concentration. Sr1−xPrxTiO3−δ ceramics with x = 0.125 show, to the best of our knowledge, the highest ever reported power factor of ∼1.3 W m−1 K−1 at 500 °C for a doped SrTiO3 2482

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

concentrations. This behavior is observed over the whole temperature range from 20 K to room temperature (see Figure S6 of the Supporting Information). However, it is known that the carrier mobility of a material is typically always higher in a single crystal versus its polycrystalline counterpart due to carrier scattering by the defects and the grain boundaries in polycrystalline samples. This counterintuitive observation suggests the presence of a “process-dependent mobility-enhancing mechanism”. As a result of such a seemingly small increase in mobility, the thermoelectric power factor was enhanced from the previously reported maximum values of approximately 1 W m−1 K−1 at ∼800 °C to 1.3 W m−1 K−1 at 500 °C. Preliminary thermal conductivity measurements of these samples resulted in a ZT value of ∼0.4 at 500 °C which, to the best of our knowledge, is the highest reported for STO-based oxides. Further optimization of the thermal conductivity of these samples will surely improve the figure of merit. Maximum ZT values as high as 0.7 can be predicted at 1000 °C by fitting the experimental electronic and thermal transport data, if the measurements are to be performed under a highly reducing atmosphere. These projections need to be validated experimentally; however, they can provide us with a semiempirical roadmap to higher performance STO-based thermoelectrics. It is worth mentioning that such high ZT values are achieved with high electrical conductivity which makes these ceramics desirable candidates for device fabrication due to minimal electrical contact problems.

■

CONCLUSIONS In conclusion, results on polycrystalline Pr-doped SrTiO3 as prepared using a new synthesis strategy via the spark plasma sintering (SPS) technique were presented. It is found that the SPS heating rate can play a crucial role in the modification of the electronic transport properties in these ceramics through the formation of a Pr-rich grain boundary phase. The nature of these grain boundaries and their role in improving the mobility of the samples are not yet fully understood. However, it can be concluded from the transport data that there needs to exist a “carrier-mobility-enhancing mechanism” in order to be able to explain the improved carrier mobility of polycrystalline samples prepared in this work versus the reported values for their single crystal counterparts. To the best of our knowledge, there has been no report of such a signif icant (a factor of ∼2) enhancement of the mobility of SrTiO3 oxides at or above room temperature. Bulk polycystalline Sr1−xPrxTiO3 samples with x = 0.125 show the highest ever reported values of the power factor of ∼1.3 W m−1 K−1 at 500 °C among n-type doped SrTiO3 ceramics as well as single crystalline STO. Understanding the origins and underlying mechanisms of the enhancement reported in this work can quite possibly result in further improvements of the thermoelectric properties of SrTiO3-based ceramics. The synthesis strategy employed in this work may open new horizons to other properties and applications of this broadly functional perovskite material system.

Figure 4. Electronic transport data for ceramics prepared under high SPS heating rates: (a) Hall and calculated carrier concentration as a function of nominal Pr content. The red marker represents undoped oxygen deficient SrTiO3 sample, and the blue ones show Pr-doped ceramics. Solid and dashed lines are just guides to the eye. (b) The “Pisarenko plot” at room temperature. Solid curve is based on the single parabolic band model, described by eq 1, with m* = 3.83me (300 K). (c) Room temperature mobility of Pr-doped (blue squares) SrTiO3 ceramics as a function of carrier density. Reported mobility values in the literature were also shown for comparison (square markers represent polycrystalline samples and diamonds single crystalline).27,29,56,63−70 The dashed line indicates the average of the reported values. All samples exhibit improved mobility values versus comparative reported values for single crystals. To the best of our knowledge, there has been no report on the carrier mobility of Prdoped SrTiO3 single crystals.

■

ASSOCIATED CONTENT

S Supporting Information *

is the elementary charge, and μ is the electron mobility) for Prdoped SrTiO3−δ ceramics. Reported mobility values in the literature are also shown for comparison. An enhancement in the mobility of all polycrystalline samples prepared following our synthesis strategy is observed versus the reported values for single crystals with similar carrier

Overview of the maximum reported power factor values in the literature for polycrystalline SrTiO3, analysis of the oxidation state of Pr in the samples, EDS line scans across grain boundary regions, low-temperature electronic transport properties, overview of room-temperature Pisarenko plot of SrTiO3-based thermoelectrics, temperature dependence of Hall mobility, 2483

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

(22) Lee, S.; Wilke, R. H. T.; Trolier-McKinstry, S.; Zhang, S.; Randall, C. A. Appl. Phys. Lett. 2010, 96, 031910. (23) Lee, S.; Bocka, J. A.; Trolier-McKinstry, S.; Randall, C. A. J. Eur. Ceram. Soc. 2012, 32, 3971. (24) Mattheiss, L. F. Phys. Rev. B 1972, 6, 4718. (25) Ahrens, M.; Merkle, R.; Rahmati, B.; Maier, J. Physica B 2007, 393, 239. (26) Ravichandran, J.; Siemons, W.; Scullin, M. L.; Mukerjee, S.; Huijben, M.; Moore, J. E.; Majumdar, A.; Ramesh, R. Phys. Rev. B 2001, 83, 035101. (27) Tufte, O. N.; Chapman, P. W. Phys. Rev. 1967, 155, 796. (28) Ohta, H.; Kim, S.; Mune, Y.; Mizoguchi, T.; Nomura, K.; Ohta, S.; Nomura, T.; Nananishi, Y.; Ikuhara, Y.; Hirano, M.; Hosono, H.; Koumoto, K. Nat. Mater. 2007, 6, 129. (29) Ohta, S.; Nomura, T.; Ohta, H.; Koumoto, K. J. Appl. Phys. 2005, 97, 034106. (30) Muta, H.; Kurosaki, K.; Yamanaka, S. J. Alloys Compd. 2003, 350, 292. (31) Obara, H.; Yamamoto, A.; Lee, C.; Kobayashi, K.; Matsumoto, A.; Funahashi, R. Jpn. J. Appl. Phys. 2004, 43, L540. (32) Ito, M.; Matsuda, T. J. Alloys Compd. 2009, 477, 473. (33) Kikuchi, A.; Okinaka, O.; Akiyama, T. Scr. Mater. 2010, 63, 407. (34) Shang, P.; Zhang, B.; Li, J.; Ma, N. Solid State Sci. 2010, 12, 1341. (35) Okinaka, L.; Zhang, L.; Akiyama, T. ISIJ Int. 2010, 50, 1300. (36) Shang, P.; Zhang, B.; Liu, Y.; Li, J.; Zhu, H. J. Electron. Mater. 2011, 40, 926. (37) Wang, Y.; Fan, H. Scr. Mater. 2011, 65, 190. (38) Wang, H. C.; Wang, C. L.; Su, W. B.; Liu, J.; Zhao, Y.; Peng, H.; Zhang, J. L.; Zhao, M. L.; Li, J. C.; Yin, N.; Mei, L. M. Mater. Res. Bull. 2010, 45, 809. (39) Wang, H. C.; Wang, C. L.; Su, W. B.; Liu, J.; Sun, Y.; Peng, H.; Mei, L. M. J. Am. Ceram. Soc. 2001, 94, 838. (40) Sun, Y.; Wang, C. L.; Wang, H. C.; Peng, H.; Guo, F. Q.; Su, W. B.; Liu, J.; Li, J. C.; Mei, L. M. J. Mater. Sci. 2011, 46, 5278. (41) Tinh, N. T.; Tsuji, T. In Physics and Engineering of New Materials; Cat, D. T., Pucci, A., Wandelt, K., Ed.; Springer Proceedings in Physics; Springer-Verlag Heidelberg: Berlin, 2009; Vol. 127, p 209. (42) Wang, H. C.; Wang, C. L.; Su, W. B.; Liu, J.; Peng, H.; Zhang, J. L.; Zhao, M. L.; Li, J. C.; Yin, N.; Mei, L. M. J. Alloys Compd. 2009, 486, 693. (43) Cui, Y.; He, J.; Amow, G.; Kleinke, H. Dalton Trans. 2010, 39, 1031. (44) Wang, Y.; Hyoung Lee, K.; Ohta, H.; Koumoto, K. J. Appl. Phys. 2009, 105, 103701. (45) Koumoto, K.; Wang, Y.; Zhang, R.; Kosuga, A.; Funahashi, R. Annu. Rev. Mater. Res. 2010, 40, 363−94. (46) Siemons, W.; McGuire, M. A.; Cooper, V. R.; Biegalski, M. D.; Ivanov, I. N.; Jellison, G. E.; Boatner, L. A.; Sales, B. C.; Christen, H. M. Adv. Mater. 2012, 24, 3965. (47) Kozuka, Y.; Kim, M.; Ohta, H.; Hikita, Y.; Bell, C.; Hwang, H. Y. Appl. Phys. Lett. 2010, 97, 222115. (48) Jalan, B.; Allen, S. J.; Beltz, G. E.; Moetakef, P.; Stemmer, S. Appl. Phys. Lett. 2011, 98, 132102. (49) Zhang, J.; Von Dreele, R. B.; Eyring, L. J. Solid State Chem. 1995, 118, 141. (50) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. (51) Lawless, W. N. Ferroelectrics 1977, 15, 61. (52) Lawless, W. N. Phys. Rev. B 1978, 7, 1458. (53) Frederikse, H. P. R.; Thurber, W. R.; Hosler, W. R. Phys. Rev. 1964, 134, A442. (54) Lee, C.; Yahia, J.; Brebner, J. L. Phys. Rev. B 1971, 3, 2525. (55) Uematsu, K.; Sakurai, O.; Mizutani, N.; Kato, M. J. Mater. Sci. 1984, 19, 3671. (56) Moos, R.; Hardtl, K. H. J. Appl. Phys. 1996, 80, 393. (57) Bergman, D. J.; Fel, D. J. J. Appl. Phys. 1999, 85, 8205. (58) Ioffe, A. F. Semiconductor Thermoelements and Thermoelectric Cooling; Infosearch Ltd.: London, 1957; p 155.

discussions of single/multivalley transport and conduction band edge picture, discussions of single/multichannel transport and effective medium theory, and high-temperature electronic transport projection (PDF). This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Author Contributions

All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors wish to acknowledge the Faculty Initiated Collaboration (FIC) competitive grant from KAUST. The authors also wish to acknowledge M. N. Hedhili for XPS measurements and Dr. Jian He for many thoughtful conversations and discussions.

■

REFERENCES

(1) Schooley, J. F.; Hosler, W. R.; Cohen, M. L. Phys. Rev. Lett. 1964, 12, 474. (2) Muller, K. A.; Burkard, H. Phys. Rev. B 1979, 19, 3593. (3) Ang, C.; Yu, Z. Appl. Phys. Lett. 2007, 90, 202903. (4) Liu, C.; Liu, P.; Zhou, J.; He, Y.; Su, L.; Cao, L.; Zhang, H. J. Appl. Phys. 2010, 107, 094108. (5) Haeni, J. H.; Irvin, P.; Chang, W.; Uecker, R.; Reiche, P.; Li, Y. L.; Choudhury, S.; Tian, W.; Hawley, M. E.; Craigo, B.; Tagantsev, A. E.; Pan, X. Q.; Streiffer, S. K.; Chen, L. Q.; Kirchoefer, S. W.; Levy, J.; Scholm, D. G. Nature 2004, 430, 758. (6) Kan, D.; Terashima, T.; Kanda, R.; Masuno, A.; Tanaka, K.; Chu, S.; Kan, H.; Ishizumi, A.; Kanemitsu, Y.; Shimakawa, Y.; Takano, M. Nat. Mater. 2005, 4, 816. (7) Okamoto, S.; Kobayashi, H.; Yamamoto, H. J. Appl. Phys. 1999, 86, 5594. (8) Fagg, D. P.; Kharton, V. V.; Kovalevsky, A. V.; Viskup, A. P.; Naumovich, E. N.; Frade, J. R. J. Eur. Ceram. Soc. 2001, 21, 1831. (9) Marina, O. A.; Canfield, N. L.; Stevenson, J. W. Solid State Ionics 2002, 149, 21. (10) Szot, K.; Speier, W.; Bihlmayer, G.; Waser, R. Nat. Mater. 2006, 5, 312. (11) Janousch, M.; Meijer, G. I.; Staub, U.; Delly, B.; Karg, S. F.; Anderson, B. P. Adv. Mater. 2007, 19, 2232. (12) Ohtomo, A.; Hwang, H. Y. Nature 2004, 427, 423. (13) Ohta, H.; Mizuno, T.; Zheng, S.; Kato, T.; Ikuhara, Y.; Abe, K.; Kumomi, H.; Nomura, K.; Hosono, H. Adv. Mater. 2012, 24, 740. (14) Okuda, T.; Nakanishi, K.; Miyasaka, S.; Tokura, Y. Phys. Rev. B 2001, 63, 113104. (15) Terasaki, I.; Sasago, Y.; Uchinakora, K. Phys. Rev. B 1997, 56, 12685. (16) Lee, M.; Viciu, L.; Li, L.; Wang, Y.; Foo, M. L.; Watauchi, S.; Pascal, R. A., Jr; Cava, R. J.; Ong, N. P. Nat. Mater. 2006, 5, 537. (17) Van Nong, N.; Pryds, N.; Linderoth, S.; Ohtaki, M. Adv. Mater. 2001, 23, 2484. (18) Saeed, Y.; Singh, N.; Schwingenschlögl, U. Adv. Funct. Mater. 2012, 22, 2792. (19) Lan, J.-L.; Liu, Y.-C.; Zhan, B.; Lin, Y.-H.; Zhang, B.; Yuan, X.; Zhang, W.; Xu, W.; Nan, C.-W. Adv. Mater. 2013, 25, 5086. (20) Lan, J.; Lin, Y.; Fang, H.; Mei, A.; Nan, C.; Liu, Y.; Xu, S.; Peters, M. J. Am. Ceram. Soc. 2010, 93, 2121−2124. (21) Ohtaki, M.; Araki, K.; Yamamoto, K. J. Electron. Mater. 2009, 38, 1234. 2484

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485

Chemistry of Materials

Article

(59) Ravich, Y. I.; Efimova, B. A.; Smirnov, I. A. In Semiconducting Lead Chalcogenides; Stil’bans, L. S., Ed.; Plenum: New York, 1970; Vol. 5. (60) Cutler, M.; Leavy, J. F.; Fitzpatrick, R. L. Phys. Rev. 1964, 133, A1143. (61) Mehdizadeh Dehkordi, A.; Bhattacharya, S.; Alshareef, H. N.; Tritt, T. M. Unpublished. (62) Mehdizadeh Dehkordi, A.; Bhattacharya, S.; Alshareef, H. N.; Tritt, T. M. Unpublished. (63) Ohta, S.; Ohta, H.; Koumoto, K. J. Ceram. Soc. Jpn. 2006, 114, 102. (64) Perluzzo, G.; Destry, J. Can. J. Phys. 1978, 56, 453. (65) Moos, R.; Mcnesklou, W.; Hardtl, K. H. Appl. Phys. A: Mater. Sci. Process. 1995, 61, 389. (66) Yamada, H.; Miller, G. R. J. Solid State Chem. 1973, 6, 169. (67) Gerthsen, P.; Hardtl, K. H.; Csillag, A. Phys. Status Solidi A 1972, 13, 127. (68) Fukuyado, J.; Nariki, K.; Akaki, M.; Kuwaharayo, H.; Okuda, T. Phys. Rev. B 2012, 85, 075112. (69) Baniecki, J. D.; Ishii, M.; Aso, H.; Kurihara, K.; Ricinschi, D. J. Appl. Phys. 2013, 113, 013701. (70) Spinelli, A.; Torija, M. A.; Liu, C.; Jan, C.; Leighton, C. Phys. Rev. B 2010, 81, 155110. (71) Pope, A. L.; Littleton, R. T., IV; Tritt, T. M. Rev. Sci. Instrum. 2001, 72, 3129.

2485

dx.doi.org/10.1021/cm4040853 | Chem. Mater. 2014, 26, 2478−2485