Tuning the Thermoelectric Properties of Ca9Zn4+xSb9 by Controlled

Sep 19, 2016 - ... 823 K. Hall effect measurements on Ca9Zn4.5–xAl2x/3Sb9 (x = 0, 0.5, 1), which nominally feature .... Materials Today Physics 2017...
1 downloads 0 Views 1MB Size
Subscriber access provided by Northern Illinois University

Article

Tuning the thermoelectric properties of Ca9Zn4+xSb9 by controlled doping on the interstitial structure Zhen Wu, Jun Li, Xin Li, Min Zhu, Ke-Chen Wu, Xu-Tang Tao, Bai-Biao Huang, and Sheng-qing Xia Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.6b02498 • Publication Date (Web): 19 Sep 2016 Downloaded from http://pubs.acs.org on September 19, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Tuning the thermoelectric properties of Ca9Zn4+xSb9 by controlled doping on the interstitial structure Zhen Wu, a Jun Li, b Xin Li, a Min Zhu, a Ke-chen Wu, *b Xu-tang Tao, a Bai-Biao Huang, a and Sheng-qing Xia *a

State Key Laboratory of Crystal Materials, Institute of Crystal Materials, Shandong University, Jinan, Shandong 250100, People’s Republic of China AUTHOR EMAIL ADDRESS: [email protected]; [email protected] CORRESPONDING AUTHORS FOOTNOTE: a

State Key Laboratory of Crystal Materials, Institute of Crystal Materials, Shandong University, Jinan,

ShanDong 250100; Phone: (531) 883-62519, Fax: (531) 883-62519. b

Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian,

350002, People’s Republic of China

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

Abstract

The complex Zintl phase Ca9Zn4.5Sb9 is a promising thermoelectric material due to its low thermal conductivity. Initial studies revealed that undoped Ca9Zn4.5Sb9 has a maximum zT value of 0.37 at 823 K. Hall effect measurements on Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1), which nominally feature identical carrier concentration, suggested that control of the interstitial chemistry is critical in affecting the mobility of this system. Removal of the interstitial atoms results in rapid decrease in the carrier mobility, which has negative effects on the overall thermoelectric performance. Further material optimization was successfully carried out via Cu-doping in Ca9Zn4.5Sb9, with the aid of theoretical predictions. This resulted in high concentration of interstitial atoms as in Ca9Zn4.5-xCuxSb9 (x = 0.05, 0.1, 0.15, 0.2). For the optimized composition Ca9Zn4.35Cu0.15Sb9, a maximum of the figure of merit (zT = 0.72) was obtained at 873 K, which is increased by more than twice compared to the undoped material Ca9Zn4.5Sb9.

2

ACS Paragon Plus Environment

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Introduction

Since thermoelectric materials can realize the energy conversion between heat and electricity directly, they are especially useful in applications such as waste energy recovery, portable device cooling, and remote power supplies. However, the large-scale use of such materials is seriously hampered by their low efficiency, generally lower than that of competing existing technologies. The evaluation of a thermoelectric material is done via the figure of merit, zT = (α2T)/(ρκ), in which α, ρ and κ represent Seebeck coefficient, electrical resistivity and thermal conductivity, respectively. Since these parameters are correlated to each other through carrier concentration, an optimized thermoelectric material is usually a highly doped semiconductor. The overall performance, clearly, is a compromised result of above properties. For complex Zintl phases, low thermal conductivities are expected due to their complex crystal structures. 1-7 Such characteristic makes Zintl phases especially convenient for designing high zT materials by simply tuning their electronic properties. This strategy has been already demonstrated for Yb14MnSb11, an emerging thermoelectric material in the high temperature range.8 By tuning the carrier concentration with Al (formally electron doping with Al3+ on the Mn2+ site), a maximum zT of 1.1 was achieved at 1223 K.9 A different class of Zintl phases for thermoelectrics development is represented by several antimonide compounds with the CaAl2Si2 structure type. The thermoelectric performance in this case is generally strong because of their high carrier mobility.10,11 Another Zintl system that has attracted much attention in the last few years is Ca9Zn4+xSb9 related intermetallics. Although the first report on this structure appeared more than 30 years ago,12 the accurate determination and understanding on the structures of such phases were not fulfilled until recently.13,14 Based on previous crystallographic and theoretical studies, the homogeneity range and electronic structure of these phases are delicately governed by the interstitial transition metal site.14 Thus, by utilization of various doping atoms as well as the modification 3

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

of doping levels, these materials could be turned into very promising thermoelectric candidates. For example, not long ago, Yb9Mn4.2Sb9 was found with a zT = 0.7 at 950K.15 Further efforts such as the substitution of Mn2+ by Zn2+ would lead to the decreased efficiency (zT = 0.18 at 975 K).16 Since the original thermal conductivity of these materials is very low, the carrier concentration optimization was also tried for the Eu9Cd4+xSb9 series by inclusion of doping atoms such as Ag+ and Cu+, however, no significant improvement on thermoelectric performance was presented through these attempts.17,18 In this paper, we carried out a systematic study on the relationship between the interstitial structure and the corresponding thermoelectric properties. Ca9Zn4+xSb9 was chosen for both experimental and theoretical investigation, with the materials constructed delicately, the relationships between the crystal structure and the corresponding physical properties were studied carefully. The results revealed that the interstitial structure in this compound was critical in affecting the carrier mobility of these materials. With the combination of theoretical and experimental studies, Cu-doped materials Ca9Zn4.5-xCuxSb9 were designed. Tuning the carrier concentration afforded a material with a high figure of merit 0.72 at 873 K, which is increased by more than twice compared to Ca9Zn4.5Sb9. Results and Discussion Synthesis. All manipulations were performed in an argon-filled glovebox. Starting materials were used as received: Ca (Alfa, 99%), Zn (Alfa, 99.99%), Al (Alfa, 99.999%), Cu (Alfa, 99.999%) Sb (Alfa, 99.999%). The reactants were loaded in Nb tubes in a stoichiometric ratio. The Nb tubes were enclosed by arc-welding in the glovebox. Then they were sealed in fused silica tube under vacuum. The reactants were first heated to 1173 K in 6 h and dwelled at this temperature for 24 h, and then slowly cooled down to 873 K at a rate of 6 K/h. Finally the reactions were kept at this temperature for 6 h and subsequently cooled down to 573 K at a rate of 10 K/h. At this temperature, the furnaces were shut down and the tubes were opened in the glovebox. The resultant products were finely ground for SPS experiments. SPS experiments. The powdered samples were placed into the graphite mould with the diameter of 4

ACS Paragon Plus Environment

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

12.6 mm and then sintered at 973 K through spark plasma sintering (SPS 1050: Sumitomo Coal Mining Co, Ltd.) for 7 min with pressure of 40 MPa. The density of all the pellets was about 95% of the theoretical value. After SPS, the pellets were checked by powder X-ray diffraction and they were consistent with the calculated. The samples were saved in the glovebox for subsequent property measurements. Based on powder X-ray diffraction measurements, no significant crystal orientation was observed for all prepared samples. Powder X-ray diffraction. Powder X-ray diffraction patterns were collected at room temperature using a Bruker AXS X-ray powder diffractometer employing Cu-Kα radiation. The data were recorded in a 2θ mode from 20 to 70° with a step size of 0.02° and counting time of 10 seconds. Differential thermal analysis and thermogravimetry measurements (DTA/TG). The measurements were taken on the polycrystalline sample with a Mettler-Toledo TGA/DSC/1600HT instrument under the protection of high-purity argon gas with a heating rate of 10 K/min applied. Thermal Conductivity. Thermal conductivity was measured by using a NETZSCH LFA457 instrument in the argon atmosphere. A pyroceram sample 9606 (Ø12.7 × 1.98 mm) was used as a standard and the measured temperature range is from 323 to 873 K. Seebeck Coefficient and Resistivity. After thermal conductivity measurements, the sample pellets were cut into bars with suitable sizes to measure Seebeck coefficient and resistivity. These two properties were measured simultaneously with a Linseis LSR-3/1100 instrument system at helium atmosphere over a temperature range of 323 to 873 K. Carrier concentration and mobility. Measurements of carrier concentration as well as mobility were performed by using a MMR K2500 Hall effect test system with a temperature range from 300 to 573K. The samples were cut into squares of 4 × 4 mm with gold-plating at each corner. All samples were polished until the thickness is ca. 0.5 mm. Computational details. The BoltzTrap program19 is employed to investigate the thermoelectric (TE) properties based on the analytical expressions of the electronic bands, with rigid band approach and

5

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 25

constant scatting time approximation adopted. The electronic structure calculations are done within density functional theory (DFT) and plane wave pseudopotential technique, as implemented in the Vienna Ab-initio Simulation Package (VASP).20 The generalized gradient approximation of PerdewBurke-Ernzerhof (PBE)21 for the exchange-correlation potential and the projector augmented wave (PAW) method22 are employed in this code. The kinetic energy cutoff of wave functions is 500 eV, with the energy convergence sets as 10-4 eV/atom. A Monkhorst-Pack k-point mesh of 3 × 8 × 3 is used to sample the Brillouin zones and the K-point mesh convergence test is performed to guarantee the accurate prediction of the TE properties. Results and Discussion Structure Description

Figure 1. Crystal structure of Ca9Zn4+xSb9, viewed along c axis. The Ca, Zn, Sb atoms are shown in orange, blue and green colors. The interstitial Zn atoms are emphasized in red.

Since the detailed description of the crystal structure has already been well discussed in previous references,13,14,23 here only a concise introduction will be provided with the most important features mentioned. A schematic view of the structure is shown in Figure 1. If the interstitial Zn sites are removed, emphasized in red color, the structure of this compound can be visualized as tetramerized Zn4Sb9 chains. However, such an “empty” structure corresponds to a charge-imbalanced system, namely

6

ACS Paragon Plus Environment

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

[Ca2+]9[Zn2+]4[Sb3-]9. This electron count does not match the electronic requirements for a Zintl phase. Thus, by introduction of an interstitial Zn atom, even at low site occupancy, this compound becomes electron precise, i.e., [Ca2+]9[Zn2+]4.5[Sb3-]9. Moreover, with such an interstitial site introduced, the Zn4Sb9 ribbons will be joined in a quasi-2D-layered structure. As frequently reported, Zintl phases with chain structures usually exhibit poor electrical conductivity. Several relevant examples that illustrate this are Ca5Al2Sb624 and Eu5Sn2As625, among others. At the same time, layered structures based on ZnSb4 tetrahedra, such as YbZn2Sb2,26 offer high thermoelectric performance that comes about from the excellent electrical transport. Thus, the Ca9Zn4+xSb9 structure seems to offer a combination of these two desirable characteristics. Material design of Ca9Zn4.5-xAl2x/3Sb9

Figure 2. Powder X-ray diffraction patterns of materials Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1.0). The simulated pattern of Ca9Zn4.5Sb9 is provided for comparison.

With above analysis, the interstitial atom in Ca9Zn4+xSb9 may play an important role on related electronic structures as well as the corresponding physical properties. However, a study on Ca9Zn4+xSb9 with different defect levels will lead to a complicated situation since with the concentration of interstitial Zn atom varied, the carrier concentration levels will change as well. The properties measured will be a synergetic result of these two factors. In order to separate above two parameters, we designed a series of Al-doped materials Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1.0). With such a strategy, the nominal valence electrons are maintained identical as in 7

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

Ca9Zn4.5Sb9, in which case the properties are assumed to be predominantly affected by the variation of “x” in Ca9Zn4.5-xAl2x/3Sb9. Four compositions with decreasing Al concentrations from Ca9Zn3AlSb9 to Ca9Zn4.5Sb9 were tried and the products were pure based on powder diffraction data. However, Ca9Zn3AlSb9 decomposes after SPS, which hampers the property studies. Higher Al concentrations were also tried and in those cases, Ca11Sb10 appears as a side product. Thus, only three materials, as shown in Figure 2, were selected for further measurements Thermal stability of Ca9Zn4.5Sb9

Figure 3. TG-DSC measurements taken on the polycrystalline sample of Ca9Zn4.5Sb9. The sample quickly decomposes above ca. 1050K.

According the TG-DSC data (Figure 3), Ca9Zn4.5Sb9 decomposes above 1050 K, which corresponds to a significant mass loss of about 14%. The product after decomposition was further checked by powder X-ray diffraction, which suggested a mixture of several binaries such as Ca2Sb, Ca11Sb10, Ca5Sb3. The Al-doped samples show similar thermal stability, but the decomposition products have not been established. Thermoelectric properties of Ca9Zn4.5-xAl2x/3Sb9 Hall effect measurements were firstly carried out to verify the carrier concentrations of Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1), as shown in Figure 4a. The measurements indicated that these materials are p-type semiconductors with holes as the major carrier. At room temperature, the 8

ACS Paragon Plus Environment

Page 9 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

hole concentrations are about 1.9×1020, 1.7×1020 and 1.4×1020 cm-3 for Ca9Zn4.5Sb9, Ca9Zn4Al0.33Sb9 and Ca9Zn3.5Al0.67Sb9, respectively. These results support above idea on the design of Ca9Zn4.5-xAl2x/3Sb9 with similar carrier concentrations. Under this circumstance, the measured properties will be mainly related to the effects contributed by the defect levels. For all three samples, there is no evidence indicating that the carrier concentrations are temperaturedependent in the measured range (RT to 573 K).

Figure 4. a) Carrier concentration and b) mobility of materials Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1.0). The measured temperature range is from RT to 573 K.

Worthy of a specific mention here is the mobility changes in these materials (Figure 4b). As discussed above, the concentration of interstitial atoms are critical in constructing a quasi-2D structure of these compounds, which may be critical in affecting the corresponding electrical transport properties. As indicated in the graph, the mobility increase from 0.74 cm2/Vs to 6.73 cm2/Vs, improved by about 10 times with the interstitial atom concentrations varied from 0.17 9

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

in Ca9Zn3.5Al0.67Sb9 to 0.5 in Ca9Zn4.5Sb9. Normally, if the carrier concentration is fixed, an increase on the mobility will result in an improvement on the electrical conductivity, which may also enhance the thermoelectric performance if the electronic band structure follows a rigid band model after doping.

Figure 5. a) Electrical resistivity and b) Seebeck coefficient of materials Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1.0). These two parameters were measured simultaneously over a temperature range between RT to 873 K.

More detailed results on the thermoelectric properties of materials Ca9Zn4.5-xAl2/3xSb9 were provided in Figure 5 and 6. For the electrical resistivity measurements, it is clear that with the concentration of the interstitial atoms increased, i.e., from Ca9Zn3.5Al0.67Sb9 to Ca9Zn4.5Sb9, the resistivity decreases from 47.9 mΩ·cm to 4.0 mΩ·cm at RT, which is in good agreement with the Hall effect measurements above (Figure 4a). However, a slight decrease in the Seebeck coefficient was also observed, which is probably caused by the change of the effective mass near the Fermi level when Zn is substituted by Al. Based on the electronic band structure 10

ACS Paragon Plus Environment

Page 11 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

calculations on Ca9Zn4.5Sb9 (Figure S1 in Supporting Information), the hole and electron effective mass are 0.11 mo and 0.07 mo, respectively, where m0 is the free electron mass. These theoretical results suggest that higher Seebeck coefficient should be obtained through hole-doping. In addition, the small energy separation ∆ε between two valence bands near Fermi level also indicates high Seebeck coefficient. However, for Al-doped material Ca9Zn4Al0.33Sb9, the decreased degeneracy of bands and increased band dispersion actually lead to smaller effective mass, and the calculated m* becomes much smaller (0.016 m0 ), as shown in Figure S2 (Supporting Information). Another direct way to visualize the doping effects on the change of effective mass is Pisarenko plot, which shows the Seebeck coefficient as a function of carrier concentration based on various predicted effective mass m*. As indicated in Figure S3, Al doping will lead to a decreasing on effective mass from 1.06 m0 to 0.62 m0. Note that these values were evaluated through the SPB (Single Parabolic Band) model, thus they are different from those calculated from actual electronic band structures. However, both methods suggest diminished effective mass in Al-doped Ca9Zn4.5Sb9. For Ca9Zn4.5Sb9, a maximum Seebeck coefficient value of ~150 µV·K-1 is achieved at 823 K, which

is

comparable

to

those

of

other

well-know

antimonide-based

Zintl

phase

thermoelectrics.15, 24, 27, 28 The thermal conductivity of these materials is generally very low, all below 1.0 W·m-1·K-1. This phenomenon is very common for Zintl phases if taking into account their very complex crystal structures as well as the unique electronic structures.29-31 Based on the electrical transport properties, p-type semiconducting behavior can be assumed for Ca9Zn4.5Sb9; by employing Eg = 2αmax·Tmax, the band gap can be roughly evaluated (About 0.25 eV), which is a little higher compared to that of theoretical prediction (an underestimation of band gap is very common based on GGA method). Interestingly, with more Al doped into the Zn site, the thermal conductivity also drops, due to the increased disorder between Al and Zn. For Ca9Zn3.5Al0.67Sb9, the average total thermal conductivity is around 0.4 W·m-1·K-1. This value compares favorably with the previously

11

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

reported values in Yb14MnSb11,8 Yb9Mn4+xSb9,15 and Eu9Cd4+xSb9.18 The lower thermal conductivity in more Al-doped samples may be due to lower electronic thermal conductivity contribution as they provide higher electrical resistivity, which can be clearly seen by decomposing the electronic and lattice components (Figure S4). In addition, Al doping also promotes the structure disorder, with which a lower lattice thermal conductivity was also observed.

Figure 6. a) Thermal conductivity and b) figure of merit evaluated for materials Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1.0) over a temperature range from RT to 873 K.

All above results imply high potential of Ca9Zn4+xSb9 as new thermoelectric candidates. The calculated zT values were evaluated in Figure 6b for Ca9Zn4.5-xAl2x/3Sb9 (x = 0, 0.5, 1.0). The maximum figure of merit reaches 0.37 at 823 K in Ca9Zn4.5Sb9. However, other two materials with Al doped into the Zn sites result in negative effects on the overall thermoelectric

12

ACS Paragon Plus Environment

Page 13 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

performance. Although the thermal conductivity seems optimized, the reduction in the electrical conductivity, which mainly originates from the carrier mobility, will impede the electrical transport properties. Theoretical studies on Ca9Zn4+xSb9 Since detailed theoretical studies on the isostructural compound Eu9Cd4Sb9 have been reported in previous reference,

18

here only a brief discussion on the electronic band structure of Ca9Zn4.5Sb9 was

provided and current theoretical results are aiming to help the material design of Ca9Zn4.5-xSb9. The calculated total and partial density of states (TDOS and PDOS) on Ca9Zn4.5Sb9 are presented in Figure 7. According to the calculations, material Ca9Zn4.5Sb9 should be semiconducting, which exhibit a narrow band gap in the corresponding electronic band structure. However, owing to the obvious problem of GGA for the exchange-correlation potential, this theoretically predicted gap value may be underestimated. Generally, large Seebeck coefficients are usually associated with the slope of DOS curve. By a comparison of valence bands as well as the conduction bands, it is indicative that higher Seebeck coefficient may be achieved by hole-doping. According to the analysis of various PDOS, the 5p-obitals of Sb contribute dominate the states right below the Fermi level, whereas the 4p-states of Zn have very minor contribution. For the conduction bands, significant mixing takes place between the orbitals from various constituent atoms, which makes the bands much more dispersive. In such a case, the p-type materials of Ca9Zn4.5Sb9 tend to exhibit larger Seebeck coefficient than the n-type ones. Above theoretical results also coincide with those calculated from Eu9Cd4Sb9.18 The calculated electronic band structure of Ca9Zn4.5Sb9 was input into BoltzTrap for property prediction. The results are summarized in Figure 8. Note that such calculations are not aiming at prediction of the exact figure of merit of these materials, but the correlation and understanding between the structures and physical properties. Since the maximum figure of merit was achieved at 823 K in Ca9Zn4.5Sb9, the data were extracted at temperature 800 K for better comparison. For the calculated Seebeck coefficient, with the hole concentration increased

13

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

from 1.4×1020 cm-3, the values increase rapidly at first and then reach the maximum value, ~130 µV.K-1 at a concentration level of 2.5×1020 cm-3. However, with carrier concentration continuously increased, the Seebeck coefficient diminished quickly. The calculated optimized power factor also shows a peak value of 300 µW·K-2·m-1, which corresponds to a highest zT of 0.31. This result indicates that the electronic structure of Ca9Zn4+xSb9 is probably very different from regular semiconductors and the carrier concentration has to be tuned in a narrow range.

Figure 7. Calculated total and partial density of states (TDOS and PDOS) for Ca9Zn4.5Sb9. The Fermi level was chosen as energy reference.

Figure 8. Theoretical prediction on the thermoelectric properties of Ca9Zn4+xSb9. With optimized carrier concentration of 2.2×1020 cm-3, a maximum zT of 0.31 is achieved at 800 K.

Material optimization on Ca9Zn4.5-xCuxSb9 14

ACS Paragon Plus Environment

Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

With above experimental and theoretical results combined, the material optimization of Ca9Zn4+xSb9 can be conducted. Based on the Hall effect measurements, the undoped Ca9Zn4.5Sb9 already possesses a suitable hole concentration, which corresponds to a relatively high Seebeck coefficient, as well as good electrical conductivity. An optimization by heavy hole-doping will probably reduce the Seebeck coefficient of the materials. However, for Ca9Zn4.5Sb9, the measured carrier concentration is still a little lower compared to the theoretical predictions. Previous doping experiments also indicate that the inclusion of Al into the Zn sites will lead to the decrease of the Seebeck coefficient due to the change of the effective mass. Thus, Cu was chosen as the doping atom in further optimization procedure. In addition, since the interstitial structure is very critical to the carrier mobility of this system, high interstitial concentration levels were configured on purpose in all materials. Since nominally Cu+ replaces Zn2+, the hole concentration will be further increased, for which the electrical conductivity thus can be enhanced significantly.

Figure 9. Powder X-ray diffraction patterns of materials Ca9Zn4.5-xCuxSb9 (x = 0.05, 0.1, 0.15, 0.2). The simulated pattern of Ca9Zn4.5Sb9 is provided for comparison.

Based on this reasoning, materials Ca9Zn4.5-xCuxSb9 (x = 0.05, 0.1, 0.15, 0.2) were designed and prepared. With the Cu concentration increased, the diffraction peaks shift towards lower angles, as shown in Figure 9. This observation supports Cu/Zn mixing, as the expansion of the unit cell correlates with the larger size of Cu+. All samples are phase pure with the prepared compositions and no significant secondary phases were detected in the corresponding powder diffraction patterns.

15

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25

Thermoelectric properties of Ca9Zn4.5-xCuxSb9 Hall effect measurements were also performed to verify the corresponding carrier concentrations of Cu-doped materials Ca9Zn4.5-xCuxSb9 (x = 0, 0.05, 0.1, 0.15, 0.2) and the results are presented in Figure 10a. With Zn replaced by Cu, the carrier concentrations increase with the increasing Cu content. Typically, at room temperature, the hole concentrations are about 2.4×1020, 3.2×1020, 4.3×1020 and 5.3×1020 cm-3 for Ca9Zn4.45Cu0.05Sb9, Ca9Zn4.4Cu0.1Sb9, Ca9Zn4.35Cu0.15Sb9 and Ca9Zn4.3Cu0.2Sb9, respectively. All materials have positive hall coefficients, indicating that holes are the majority carriers over the whole measured temperature range. Similar to materials Ca9Zn4.5-xAl2x/3Sb9, the temperature dependence of the carrier concentrations of Ca9Zn4.5-xCuxSb9 is also not evidential. Constant carrier concentration versus temperature has also been observed for the Cu, Ag or Au-substituted materials based on Eu9Cd4Sb9.18 Unlike materials Ca9Zn4.5-xAl2x/3Sb9, Cu-doping has less effect on the mobility changes of materials Ca9Zn4.5-xCuxSb9 (Figure 10b). As proved by above studies, the defect levels are critical in affecting the mobility of Ca9Zn4.5Sb9-based materials. Thus, with defect levels theoretically identical in materials Ca9Zn4.5-xCuxSb9, high mobility can be maintained after doping. Actually, within the doping level x ranged from 0.05 to 0.2, the mobility of Ca9Zn4.5-xCuxSb9 is slightly larger than that of undoped Ca9Zn4.5Sb9. However, with the increasing Cu content, the mobility tends to decrease slightly. The measured mobility of Ca9Zn4.5-xCuxSb9 (x = 0, 0.05, 0.1, 0.15, 0.2) varies from 6.73 to 12 cm2/Vs at room temperature. The reason for such a mobility behavior with various Cudoping contents is very difficult to explain, and an in-depth theoretical study would have been helpful if the disorder structure were not so complicated in so low doping levels. Apparently, to accurately localize the doping atoms in the structure will be a very difficult issue, which requires substantial experimental syntheses and crystallographic analyses. However, a Seebeck Pisarenko plot of Cu-doped materials may still provide some useful information (discussed below) and the close relationship between the interstitial structure and the electronic transport properties can be

16

ACS Paragon Plus Environment

Page 17 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

ascertained. The electrical transport properties of materials Ca9Zn4.5-xCuxSb9 were presented in Figure 11. For the electrical resistivity, it is very obvious that doping Cu into the Zn site will significantly improve the

Figure 10. a) Carrier concentration and b) mobility of materials Ca9Zn4.5-xCuxSb9 (x = 0, 0.05, 0.1, 0.15, 0.2).

electrical conductivity due to the increased carrier concentrations. For example, compared with undoped Ca9Zn4.5Sb9, the resistivity of Ca9Zn4.5-xCuxSb9 is reduced from 8 mΩ·cm to 3.5 mΩ·cm at 800 K. However, as suggested by the Hall effect measurements, an increasing on the carrier concentration also correspond to the decreasing of the mobility. In such a case, no clear trend was observed for the electrical resistivity changes with the varied Cu concentrations. This is probably due to the unusual electronic structure of these materials, which can also correspond to a very complicated conducting mechanism. Very interestingly, Cu substitution in Ca9Zn4.5Sb9 does not diminish the Seebeck coefficient in current doping range. All Ca9Zn4.5-xCuxSb9 materials achieve considerable values of ~150 µV·K-1 around 800 K. Since the doping of Cu into Zn results in the increasing of carrier concentration n, thus 17

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25

the unchanged Seebeck coefficient should be possibly related to the increasing effective mass m* after Cu-doping. This assumption was confirmed by the corresponding Seebeck Pisarenko plot (Figure S3), which also indicates possible multiple band or nonparabolic band effects in Ca9Zn4.5-xCuxSb9. More accurate theoretical analyses are difficult to provide since the ordered models of such Cu-doping materials are too difficult to construct. According to some references on filled skutterudites, nonparabolic band effect was also observed and the effective mass m* increases with increasing carrier concentration.7 Besides, there are also reports indicating that with the valence bands near Fermi level composed of localized transition metal orbitals, the bands become flat and heavy, which also coincide with our results.32

Figure 11. a) Electrical resistivity and b) Seebeck coefficient of materials Ca9Zn4.5-xCuxSb9 (x = 0.05, 0.1, 0.15, 0.2) over a temperature range from RT to 873 K.

As illustrated in Figure 12a, for the Cu-doped materials the thermal conductivity increases with the increasing temperature above 700 K. This can be explained by the bipolar effect of the materials, for 18

ACS Paragon Plus Environment

Page 19 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

which the carrier concentration increases after Cu-doping. This argument can be further supported by aforementioned thermal conductivity data of Al-doped samples, in which the bipolar effect becomes less significant with the carrier concentration decreased. The measured thermal conductivities of the

Figure 12. a) Thermal conductivity and b) figure of merit on materials Ca9Zn4.5-xCuxSb9 (x = 0.05, 0.1, 0.15, 0.2) over a temperature range from RT to 873 K.

Ca9Zn4.5-xCuxSb9 materials are higher than those of Ca9Zn4.5Sb9. This is normal if taking into account the fact that the substitution of Zn2+ by Cu+ should lead to the increase of carrier concentration, which can favor the electronic contribution (the electronic thermal conductivity was generally doubled if comparing the Cu-doped materials to undoped Ca9Zn4.5Sb9 in Figure S5). Although the doping of Cu will also lead to a slight increase in the thermal conductivity, a significant improvement on the thermopower still result in considerable enhancement of the final figure of merit. As indicated in Figure 12b, material with an optimized composition Ca9Zn4.35Cu0.15Sb9 achieves the highest zT of 0.72 at 19

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 25

873K, which is doubled compared to the undoped Ca9Zn4.5Sb9 at the same temperature. As the figure of merit still does not reach the maximum with the increasing temperature, limited by the thermal stability, it can be expected that the thermoelectric performance of these materials will be even better if they can be modified to sustain higher temperature. Conclusions

In conclusion, we carried out a systematic study on the thermoelectric properties of Ca9Zn4+xSb9 Zintl phases. With the combined theoretical and experimental studies, the close relationship between the interstitial structure and the physical properties of these materials was established, which suggests that the interstitial structure should play a critical role in affecting the electrical transport properties of these materials. Based on this knowledge, efficient optimization strategy was conducted by introduction of Cu-doping in Ca9Zn4.5Sb9. Material Ca9Zn4.35Cu0.15Sb9 achieves the highest zT of 0.72 at 873 K, which is increased by more than twice compared to the undoped material Ca9Zn4.5Sb9.

Acknowledgements S.Q. Xia and X.T. Tao acknowledge the financial support from the National Natural Science Foundation of China (Grant Nos. 51271098, 51272129). K.C. Wu acknowledges the Supercomputing Center of CNIC and Special Program for Applied Research on Super Computation of the NSFC-Guangdong Joint Fund (the second phase) for providing the computer resources. B.B. Huang acknowledges the support from Taishan Scholars Program of Shandong Province.

Supporting Information Electronic band structures calculated on the hypothetical ordered models of Ca9Zn4.5Sb9 and Ca9Zn4Al0.33Sb9; Pisarenko plot showing the experimental Seebeck coefficient and predicted single parabolic band effective mass for doped and undoped Ca9Zn4.5Sb9 materials; Electronic and lattice

20

ACS Paragon Plus Environment

Page 21 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

contribution of the thermal conductivity for Al- and Cu-doped Ca9Zn4.5Sb9. This information is available free of charge via the internet at http://pubs.acs.org.

References (1) Snyder G. J. and Toberer E. S., Complex Thermoelectric Materials. Nat. Mater., 2008, 7, 105−114. (2) Ohno S., Zevalkink A., Takagiwa Y., Bux S. K. and Snyder G. J., Thermoelectric Properties of the Yb9Mn4.2−xZnxSb9 Solid Solutions. J. Mater. Chem. A, 2014, 2, 7478−7483. (3) Aydemir U., Zevalkink A., Ormeci A., Bux S., Snyder G. J., Enhanced Thermoelectric Properties of the Zintl Phase BaGa2Sb2 via Doping with Na or K. J. Mater. Chem. A, 2016, 4, 1867−1875. (4) Toberer E. S., Zevalkink A., Snyder G. J., Phonon Engineering through Crystal Chemistry. J. Mater. Chem., 2011, 21, 15843−15852. (5) Shuai J., Liu Z., Kim H. S., Wang Y., Mao J., He R., Sui J., Ren Z., Thermoelectric Properties of Bi-based Zintl Compounds Ca1−xYbxMg2Bi2. J. Mater. Chem. A, 2016, 4, 4312−4320. (6) Kauzlarich S. M., Brown S. R. and Snyder G. J., Zintl Phases for Thermoelectric Devices. Dalton Trans., 2007, 21, 2099−2107. (7) Toberer E. S., May A. F. and Snyder G. J., Zintl Chemistry for Designing High Efficiency Thermoelectric Materials. Chem. Mater., 2010, 22, 624−634. (8) Brown S. R., Kauzlarich S. M., Gascoin F., and Snyder G. J., Yb14MnSb11: New High Efficiency Thermoelectric Material for Power Generation. Chem. Mater., 2006, 18, 1873−1877. (9) Cox C. A., Toberer E. S., Levchenko A. A., Brown S. R., Snyder G. J., Navrotsky A., Kauzlarich S. M., Structure, Heat Capacity, and High-Temperature Thermal Properties of Yb14Mn1−xAlxSb11. Chem.

21

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 25

Mater., 2009, 21, 1354−1360. (10) Gascoin F., Ottensmann S., Stark D., Haïle S. M., Snyder G. J., Zintl Phases as Thermoelectric Materials: Tuned Transport Properties of the Compounds CaxYb1–xZn2Sb2. Adv. Funct. Mater., 2005, 15, 1860−1864. (11) Toberer E. S., May A. F., Melot B. C., Flage-Larsen E., Snyder G. J., Electronic Structure and Transport in Thermoelectric Compounds AZn2Sb2 (A= Sr, Ca, Yb, Eu). Dalton Trans., 2010, 39, 1046−1054. (12) Brechtel E., Cordier G., Schäfer H., New Compounds with the Ca9Mn4Bi9 Structure: Ca9Cd4Bi9, Sr9Cd4Bi9, and Ca9Zn4Sb9. Z. Naturforsch. 1981, 36, 1099−1104. (13) Bobev S., Thompson J. D., Sarrao J. L., Olmstead M. M., Hope H. k. and Kauzlarich S. M., Probing the Limits of the Zintl Concept: Structure and Bonding in Rare-Earth and Alkaline-Earth ZincAntimonides Yb9Zn4+xSb9 and Ca9Zn4. 5Sb9. Inorg. Chem., 2004, 43, 5044−5052. (14) Xia S. Q. and Bobev S., Interplay between Size and Electronic Effects in Determining the Homogeneity Range of the A9Zn4+ xPn9 and A9Cd4+xPn9 Phases (0≤x≤0.5), A = Ca, Sr, Yb, Eu; Pn = Sb, Bi. J. Am. Chem. Soc., 2007, 129, 10011−10018. (15) Bux S. K., Zevalkink A., Janka O., Uhl D., Kauzlarich S., Snyder G. J., Fleurial J. P., Glass-like Lattice Thermal Conductivity and High Thermoelectric Efficiency in Yb9Mn4.2Sb9. J. Mater. Chem. A, 2014, 2, 215−220. (16) Ohno S., Zevalkink A., Takagiwa Y., Bux S. K., Snyder G. J., Thermoelectric Properties of the Yb9Mn4.2− xZnxSb9 Solid Solutions. J. Mater. Chem. A, 2014, 2, 7478−7483. (17) Kazem, N., Hurtado, A., Klobes, B., Hermann, R. P., Kauzlarich, S. M., Eu9Cd4–xCM2+ x–yySb9: Ca9Mn4Bi9-type Structure Stuffed with Coinage Metals (Cu, Ag, and Au) and the Challenges with 22

ACS Paragon Plus Environment

Page 23 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Classical Valence Theory in Describing these Possible Zintl Phases. Inorg. Chem., 2014, 54, 850−859. (18) Kazem N., Zaikina J. V., Ohno S., Snyder G. J., Kauzlarich S. M., Coinage-Metal-Stuffed Eu9Cd4Sb9: Metallic Compounds with Anomalous Low Thermal Conductivities. Chem. Mater. 2015, 27, 7508−7519. (19) Madsen G. K. H., Singh D. J., BoltzTraP. A Code for Calculating Band-structure Dependent Quantities. Comput. Phys. Commun., 2006, 175, 67−71. (20) Kresse G., Furthmüller J., Efficient Iterative Schemes for Ab-initio Total-energy Calculations Using a Plane-wave Basis Set. Phys. Rev. B, 1996, 54, 11169−11186. (21) Perdew J. P., Burke K., Ernzerhof M., Generalized Gradient Approximation Made Simple. Phys. Rev. Lett., 1996, 77, 3865−3868. (22) Blöchl P. E., Projector Augmented-wave Method. Phys. Rev. B, 1994, 50, 17953−17979. (23) Xia S. Q. and Bobev S., New Manganese-bearing Antimonides and Bismuthides with Complex Structures. Synthesis, Structural Characterization, and Electronic Properties of Yb9Mn4+xPn9 (Pn = Sb or Bi). Chem. Mater., 2009, 22, 840−850. (24) Zevalkink A., Toberer E. S., Bleith T., Flage-Larsen E., Snyder G. J., Improved Carrier Concentration Control in Zn-doped Ca5Al2Sb6. J. Appl. Phys., 2011, 110, 013721. (25) Wang J., Xia S. Q. and Tao X. T., A5Sn2As6 (A = Sr, Eu). Synthesis, Crystal and Electronic Structure, and Thermoelectric Properties. Inorg. Chem., 2012, 51, 5771−5778. (26) Wang X. J., Tang M. B., Chen H. H., Yang X. X., Zhao J. T., Burkhardt U. and Grin Y., Synthesis and High Thermoelectric Efficiency of Zintl Phase YbCd2-xZnxSb2. Appl. Phys. Lett., 2009, 94, 092106.

23

ACS Paragon Plus Environment

Chemistry of Materials

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 25

(27) Wang J., Liu X. C., Xia S. Q., Tao X. T., Ca1–xRExAg1–ySb (RE = La, Ce, Pr, Nd, Sm; 0≤x≤1; 0≤y≤1): Interesting Structural Transformation and Enhanced High-temperature Thermoelectric Performance. J. Am. Chem. Soc., 2013, 135, 11840−11848. (28) Zhang H., Zhao J. T., Grin Y., Wang X. J., Tang M. B., Man Z. Y., Chen H. H., Yang X. X., A New Type of Thermoelectric Material, EuZn2Sb2. J. Chem. Phys., 2008, 129, 164713−164713. (29) Zevalkink A., Toberer E. S., Zeier W. G., Flage-Larsen E., Snyder G. J., Ca3AlSb3: An Inexpensive, Non-toxic Thermoelectric Material for Waste Heat Recovery. Energy Environ. Sci., 2011, 4, 510−518. (30) Suekuni K., Avila M. A., Umeo K. and Takabatake T., Cage-size Control of Guest Vibration and Thermal Conductivity in Sr8Ga16Si30−xGex. Phys. Rev. B, 2007, 75, 195210. (31) Toberer E. S., Zevalkink A., Crisosto N., Snyder G. J., The Zintl Compound Ca5Al2Sb6 for Lowcost Thermoelectric Power Generation. Adv. Funct. Mater., 2010, 20, 4375−4380. (32) Yang J., Liu R., Chen Z., Xi L., Yang J.-H., Zhang W., Chen L.-D., Power Factor Enhancement in light Valence Band p-type Skutterudites. App. phys. Lett., 2012, 101, 022101.

24

ACS Paragon Plus Environment

Page 25 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Table of Contents Synopsis

25

ACS Paragon Plus Environment