Home
Search
Collections
Journals
About
Contact us
My IOPscience
Compositional disorder and its effect on the thermoelectric performance of Zn3P2 nanowire–copper nanoparticle composites
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 125402 (http://iopscience.iop.org/0957-4484/25/12/125402) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 129.2.19.100 This content was downloaded on 04/02/2015 at 12:53
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 125402 (8pp)
doi:10.1088/0957-4484/25/12/125402
Compositional disorder and its effect on the thermoelectric performance of Zn3P2 nanowire–copper nanoparticle composites Lance Brockway1 , Venkata Vasiraju2 and Sreeram Vaddiraju1,2 1
Artie McFerrin Department of Chemical Engineering, Texas A&M University, 3122 TAMU, College Station, TX 77843, USA 2 Materials Science and Engineering Department, Texas A&M University, College Station, TX 77843, USA E-mail: [email protected] Received 7 November 2013, revised 13 January 2014 Accepted for publication 20 January 2014 Published 27 February 2014
Abstract
Recent studies indicated that nanowire format of materials is ideal for enhancing the thermoelectric performance of materials. Most of these studies were performed using individual nanowires as the test elements. It is not currently clear whether bulk assemblies of nanowires replicate this enhanced thermoelectric performance of individual nanowires. Therefore, it is imperative to understand whether enhanced thermoelectric performance exhibited by individual nanowires can be extended to bulk assemblies of nanowires. It is also imperative to know whether the addition of metal nanoparticle to semiconductor nanowires can be employed for enhancing their thermoelectric performance further. Specifically, it is important to understand the effect of microstructure and composition on the thermoelectric performance on bulk compound semiconductor nanowire–metal nanoparticle composites. In this study, bulk composites composed of mixtures of copper nanoparticles with either unfunctionalized or 1,4-benzenedithiol (BDT) functionalized Zn3 P2 nanowires were fabricated and analyzed for their thermoelectric performance. The results indicated that use of BDT functionalized nanowires for the fabrication of composites leads to interface-engineered composites that have uniform composition all across their cross-section. The interface engineering allows for increasing their Seebeck coefficients and electrical conductivities, relative to the Zn3 P2 nanowire pellets. In contrast, the use of unfunctionalized Zn3 P2 nanowires for the fabrication of composite leads to the formation of composites that are non-uniform in composition across their cross-section. Ultimately, the composites were found to have Zn3 P2 nanowires interspersed with metal alloy nanoparticles. Such non-uniform composites exhibited very high electrical conductivities, but slightly lower Seebeck coefficients, relative to Zn3 P2 nanowire pellets. These composites were found to show a very high zT of 0.23 at 770 K, orders of magnitude higher than either interface-engineered composites or Zn3 P2 nanowire pellets. The results indicate that microstructural composition of semiconductor nanowire–metal nanoparticle composites plays a major role in determining their thermoelectric performance, and such composites exhibit enhanced thermoelectric performance. Keywords: nanowires, Zn3 P2 , nanoparticle, composites, thermoelectrics, interface engineering, carrier filtering (Some figures may appear in colour only in the online journal)
0957-4484/14/125402+08$33.00
1
c 2014 IOP Publishing Ltd
Printed in the UK
Nanotechnology 25 (2014) 125402
L Brockway et al
1. Introduction
by Neophytou et al indicated that high level doping of nanocrystalline silicon with boron leads to a simultaneous increase in the Seebeck coefficient and the electrical conductivity [24]. In view of the widely varying results in these reports, it is essential to know the effect of metal nanoparticle inclusion on the thermoelectric performance of metal–semiconductor composites. More specifically, it is essential to determine the effect of microstructure and compositional uniformity on the thermoelectric performance of metal–semiconductor composites made by consolidating mixtures of metal and semiconductor nanopowders. The material chosen for this study is zinc phosphide (Zn3 P2 ), a compound semiconductor composed of only earth abundant elements. The thermoelectric properties of Zn3 P2 reported previously were optimistic, but varied widely [25, 26]. In this work, the effect of microstructure and compositional uniformity on the thermoelectric performances of Zn3 P2 nanowire–copper nanoparticle composites is presented. By the end of the manuscript, it will be shown that the compositional uniformity at microscale plays a very critical role in determining their overall thermoelectric performance. It will also be demonstrated that compositional uniformity/non-uniformity in such composites can be employed to engineer and greatly enhance the thermoelectric performance (or zT values) of materials.
Thermoelectric modules are solid-state devices that convert waste heat into electricity [1]. The fact that they have no moving parts and require minimal maintenance makes them indispensable for continued power generation in locations that are off the grid, such as satellites [2–4]. For the routine use of thermoelectrics in terrestrial applications, it is essential to enhance the performance of thermoelectrics beyond that obtained by the current state of the art. The performance of solid-state thermoelectrics can be gauged from their figure 2σ T [1, 2, 4, 5]. Here, S of merit, zT , given by zT = κSe +κ l is the Seebeck coefficient, σ is the electrical conductivity, and κe and κl are the thermal conductivities of the material from electronic and lattice contributions, respectively. Previous studies aimed at the enhancing the zT values of materials relied on circumventing the Wiedemann–Franz law through a selective reduction of their κl [4, 6–8]. This intense effort to enhance the zT by reducing κl through the creation of compositional inhomogenities or through the intentional addition of impurities to materials led to the fabrication of p-type Bi2 Te3 /Sb2 Te3 superlattices with a high zT of 2.4 [9], bulk material of AgPbm SbTe2+m with a zT of 2.2 [10] and Bi-doped n-type PbSeTe/PbTe quantum dot superlattice with zT as high as 3.5 [11]. It was also theoretically predicted [12, 13] and experimentally observed [14, 15] that materials in nanowire form offer the ability to tune the magnitudes of electrical and thermal transport through them. In fact, nanostructuring was employed to increase the room temperature zT value of rough silicon nanowires to 0.6 [15]. Another set of studies relied on enhancing the zT of materials by increasing their power factors, in addition to lowering their κl [16–18]. These studies employed the use of either two different semiconductors or metal–semiconductor composites for achieving this task. For instance, variation of the ratio of copper to titania in Cu/TiO2−x composites led to varying electrical conductivities and the Seebeck coefficients in the composites [19]. Here, the increase in electrical conductivities of the composites led to a decrease in their Seebeck coefficients [19]. The behavior of the composites in this scenario was explained using rule of mixtures [19]. Additionally, Faleev and Leonard predicted that metallic nanoinclusions within PbTe matrix selectively scatters electrons at the grain boundaries and leads to an enhancement in the Seebeck coefficient of the composites, relative to bulk PbTe [20]. They suggested that this Seebeck coefficient enhancement occurs only at very high doping levels [20]. Martin et al experimentally demonstrated this energy-barrier scattering concept in Ag-doped PbTe composites [21]. However, Zebarjadi et al showed that inclusion of ErAs nanoparticles inside InGaAlAs matrix only enhances the electrical conductivity and not the Seebeck coefficient [22]. In another report, they have demonstrated that semimetallic inclusions in semimetal/semiconductor composites can cause energy-dependent scattering and not only reduce the lattice thermal conductivities, but also simultaneously increase their power factor and lead to an overall enhancement of their zT values [23]. Finally, a recent study
2. Experimental details
Zn3 P2 nanowires necessary for the fabrication of the semiconductor nanowire–metal nanoparticle composites have been synthesized using direct reaction of zinc foils with phosphorus supplied via the vapor phase in a hot walled chemical vapor deposition (HWCVD) chamber [27]. The experimental procedure employed for this purpose was discussed in detail in a previous publication [27]. Following the synthesis, the nanowires were brushed off the foils and collected as powders. Similar procedure was also employed for obtaining Zn3 P2 surface functionalized with 1,4-benzenedithiol (BDT) [27, 28]. These surface functionalized nanowires will henceforth be referred to as BDT functionalized Zn3 P2 nanowires in this manuscript, while nanowires without any surface functionalization will be referred to as unfunctionalized Zn3 P2 nanowires. For the preparation of the composites, 1 g of Zn3 P2 nanowire powder was mixed with 40 mg of commercially-available copper nanoparticle powder (average diameter = 40 nm, obtained from US Research Nanomaterials, Inc.) in a ball mill. The obtained mixture was consolidated using hot uniaxial pressing at a temperature of 650 ◦ C and at a pressure of 120 MPa for 1 h into 12 mm-diameter, 1 mm-thick cylindrical pellets. Pellets of two different types of composites have been fabricated by hot uniaxial pressing copper nanoparticle powder mixed with either unfunctionalized or BDT functionalized nanowire powders. Thermoelectric performance of the resulting BDT functionalized Zn3 P2 nanowire–copper nanoparticle composite (referred to as composite A in this manuscript) pellets and unfunctionalized Zn3 P2 nanowire–copper nanoparticle composite (referred to as composite B in this manuscript) pellets were calculated 2
Nanotechnology 25 (2014) 125402
L Brockway et al
from the following data: heat capacity of the pellets determined using differential scanning calorimetry (DSC) [29], the thermal diffusivity determined using a laser flash apparatus (LFA) [30], electrical conductivity determined using the Van der Pauw 4-point probe method [31], and the Seebeck coefficient determined by imposing a temperature gradient across the pellet cross-section and measuring the resulting voltage [32]. For deducing the effect of pellet composition on their thermoelectric performance, the nanowire powders obtained and the composite pellets fabricated were also characterized for morphology, phase and chemical composition using scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HR-TEM), x-ray diffraction (XRD), and energy dispersive spectroscopy (EDS). For the sake of brevity, the characterization of the nanowire powders is not discussed in detail in the manuscript. For a detailed discussion of the morphology, phase and chemical composition of the nanowire powders, the reader is referred to previously published work [27].
Figure 1. XRD patterns of composite-A and composite-B. In both the cases, the addition of copper to Zn3 P2 nanowires and consolidation of the mixtures into high density composite pellets led to the formation of only one additional phase, Cu0.2 Zn0.8 . Major presence of α-Zn3 P2 parent phase was also observed in the spectra.
3. Results and discussion
were observed to be the same [28]. Similarly, the trend of the variation of the Seebeck coefficients with temperature of both composite-B pellets and the unfunctionalized Zn3 P2 nanowire pellet were observed to be the same. In cases where BDT functionalized nanowires were employed (i.e., composite-A pellets and BDT functionalized nanowire pellets), the trend of the Seebeck coefficient was observed to be along that lines of that expected of a non-degenerate semiconductor. The deviation from non-degenerate behavior in the other two samples, composite-B pellets and unfunctionalized Zn3 P2 nanowire pellets, at higher temperatures is owed to the carrier transport contribution of the surface oxide layers surrounding the Zn3 P2 nanowires. In slight contrast, the electrical conductivities of both composite-A pellet and composite-B pellet were observed to be higher than that observed in unfunctionalized Zn3 P2 nanowire pellet (figure 2(b)). So, the addition of copper nanoparticles to unfunctionalized Zn3 P2 nanowires enhanced the electrical conductivity of the composite and resulted in a reduction of its Seebeck coefficient. This is in line with the classical behavior of thermoelectric materials where an increase in the Seebeck coefficients results in a decrease of their electrical conductivities [2]. In sharp contrast, the addition of copper nanoparticles to BDT functionalized nanowires led to an increase in the Seebeck coefficient and the electrical conductivity of the composite-A relative to unfunctionalized Zn3 P2 nanowires. Such increase in the Seebeck coefficient and electrical conductivity was observed in BDT functionalized nanowire pellets in a previous study by our group [28]. Therefore, use of BDT functionalized nanowires for the fabrication of pellets always led to an increase in both the Seebeck coefficient and the electrical conductivity of the pellets, irrespective of the addition of copper nanoparticles. The Seebeck coefficients, electrical conductivities, thermal conductivities and zT values observed in both composite-A and composite-B at a temperature of 770 K are summarized in table 1. A maximum zT value of 0.23 was observed in composite-B.
Hot uniaxial pressing of mixtures of either unfunctionalized or BDT functionalized Zn3 P2 nanowires and copper nanoparticles led to the formation of 12 mm-diameter, 1 mm-thick cylindrical pellets. The densities of all the pellets were measured geometrically and confirmed using Archimedes principle [32]. In each case, the density of the pellets obtained was found to be ≥98% of the respective theoretical composite density. In other words, the porosity of the composite pellets was ≤2%. XRD analysis of the both composite-A and composite-B are depicted in figure 1. The analysis indicated that both the composite pellets are mainly composed of the α-Zn3 P2 nanowire phase [33], with a minor contribution from the copper–zinc alloy phase (specifically, the hexagonal Cu0.2 Zn0.8 phase with lattice parameters of a = 0.273 nm, c = 0.429 nm) [34]. 2θ reflection corresponding to the Cu0.2 Zn0.8 alloy phase is represented by a ‘*’ in figure 1. No presence of pure unreacted copper was detected in the XRD spectra of the composite pellets. Additionally, no presence of ZnS and Cu2 S/CuS phases, owed to the reaction of sulfur from BDT with either copper of the nanoparticles or the Zn of Zn3 P2 , was observed in the composite-A pellet. Thermoelectric performance of the pellets was deduced by independently measuring the Seebeck coefficient, the electrical conductivity, and the thermal conductivity and are presented in figure 2. For comparison, the Seebeck coefficient, the electrical conductivity and the thermal conductivity data obtained from highly dense unfunctionalized Zn3 P2 nanowire pellet reported in a previous study is also presented in figure 2 [28]. When compared to the data from unfunctionalized Zn3 P2 nanowire pellet, the Seebeck coefficient of the composite-A pellet was observed to be higher (figure 2(a)). However, the Seebeck coefficient of composite-B pellet was observed to be lower than that observed from unfunctionalized nanowire pellet (figure 2(a)). The trend of the variation of the Seebeck coefficients with temperature of both the composite-A pellets and the BDT functionalized Zn3 P2 nanowire pellets 3
Nanotechnology 25 (2014) 125402
L Brockway et al
Table 1. Values of Seebeck coefficients, electrical conductivities, thermal conductivities, and zT values of composite-A and composite-B at
770 K. A maximum zT value of 0.23 was observed in composite-B. Sample
Seebeck coefficient (µV K−1 )
Electrical conductivity (S m−1 )
Thermal conductivity (W m−1 K−1 )
Maximum zT
Composite-A Composite-B
362.96 246.44
38.52 8457.82
0.92 1.64
0.0042 0.23
functionalized and unfunctionalized nanowires mixed with copper nanoparticles (in the same weight fractions employed for the fabrication of composite pellets) using a mortar and pestle was performed. A micrograph of the BDT functionalized nanowires mixed with copper nanoparticles is presented in figure 3(a). In this case, the process of mixing led to the uniform decoration of BDT functionalized nanowire surfaces with copper nanoparticles. This is owed to the formation of bonds between the free –SH groups of the BDT molecules present on top of the functionalized nanowires and the copper nanoparticles (figure 3(b)) [27]. This indicates that on hot uniaxial pressing, mixtures of BDT functionalized Zn3 P2 nanowires and copper nanoparticles lead to the formation of interface-engineered composites that are uniform in composition. It is believed that each pellet is composed of copper-doped Zn3 P2 nanowires interfaced with Cu0.2 Zn0.8 metal alloy particles (figure 3(b)). The BDT functional molecule barrier at the nanowire interfaces is expected to prevent, at least partially, the doping of the Zn3 P2 nanowires with copper in this composite. A completely different scenario was observed when unfunctionalized nanowires are mixed with copper nanoparticles. No decoration of the nanowires with copper nanoparticles was observed in the micrograph of the samples (figure 3(c)). The copper nanoparticles and Zn3 P2 nanowires exist as individual units within the mixture as observed in the micrograph (figure 3(c)). Owing to the presence of individual unit of copper nanoparticles and Zn3 P2 nanowires, the sample is non-uniform in composition across the sample at microscale (figures 3(c) and (d)). Hot uniaxial pressing of this mixture is believed to lead to the formation of Cu0.2 Zn0.8 metal alloy particles interspersed within the copper-doped Zn3 P2 nanowire matrix. It is further believed that the composite is non-uniform in composition with varying ratios of the amount of Cu0.2 Zn0.8 metal alloy and copper-doped Zn3 P2 nanowires (figure 3(d)) through its cross-section. Destructive analysis of the composite pellets confirmed the above analysis and indicated that the composite pellets are mainly composed of single-crystalline Zn3 P2 nanowires (figure 4). These experiments were performed by chipping pieces of the pellets directly onto nickel TEM grids, followed by analyzing the grids using a HR-TEM. The analysis of the chipped pieces indicated that single-crystalline Zn3 P2 nanowires are still present in the composite pellets (figures 4(a) and (c)). It is essential to recall here that flexibility of the nanowires allows for their high density packing into dense pellets, a concept that has been described in detail previously [28]. The selected area electron diffraction (SAED) analysis of a nanowire chipped from composite-A (figure 4(b)) confirmed that it is still single-crystalline in nature. Similar
Figure 2. Plots indicating the variation of (a) Seebeck coefficients, (b) electrical conductivities, (c) thermal conductivities, and (d) zT values of composite-A and composite-B. For comparison, data obtained from unfunctionalized Zn3 P2 nanowire pellets obtained by our group in a previous study [28] was also included in (a)–(c).
Thermal conductivity of the pellets indicated that the addition of copper to the both unfunctionalized and BDT functionalized Zn3 P2 nanowires led to an increase in the thermal conductivity of the pellets. However, the increase in the thermal conductivity of composite-A pellet was observed to be minor, compared to that observed in unfunctionalized Zn3 P2 nanowire pellet. The thermal conductivity of composite-A pellet composite was observed to vary between 1.49 and 0.92 W m−1 K−1 in the 300–775 K temperature regime. Contrastingly, composite-B pellet exhibited a relatively high thermal conductivity of 2.54–1.64 W m−1 K−1 in the 300– 775 K temperature regime. Finally, the variation of the zT values of both the composite pellets with temperature is presented in figure 2(d). A maximum zT value of 0.23 was achieved in compositeB pellet, orders of magnitude higher than the maximum zT value of 0.0042 achieved in composite-A pellet. Such high zT values of 0.23 have never been reported in Zn3 P2 materials system. To understand and explain the variations in the thermoelectric behavior of the composites and deduce composition-thermoelectric property interrelationships in them, additional microscopic composition analysis of the pellets was performed. Both destructive and non-destructive techniques were employed for deducing the variations in the chemical compositions of the pellets at microscale. First and foremost, morphological analysis of both BDT 4
Nanotechnology 25 (2014) 125402
L Brockway et al
Figure 3. (a) Scanning electron micrograph of copper nanoparticle decorated Zn3 P2 nanowires obtained by simply mixing BDT functionalized Zn3 P2 nanowires with copper nanoparticles in a mortar and pestle. (b) A schematic representing the steps underlying the formation of composite-A. The decoration of BDT functionalized Zn3 P2 nanowire surfaces with copper nanoparticles during the nanomaterial mixing phase leads to the formation of interface-engineered composites during the consolidation phase. (c) Scanning electron micrograph of mixtures of unfunctionalized Zn3 P2 nanowires and copper nanoparticles. No decoration of the nanowires with copper nanoparticles occurred. The nanomaterials remained as individual units after mixing. (d) A schematic representing the steps underlying the formation of composite-B. Here, consolidation leads to the formation of non-uniform composites that has metal alloy interspersed between copper-doped Zn3 P2 nanowires.
confirmation was also obtained from the SAED analysis of a chipped nanowire from composite-B pellet (figure 4(d)). In both the composites, the formation of zinc-rich Cu0.2 Zn0.8 metal alloy phase is believed to be the because of the partial decomposition of Zn3 P2 nanowires occurring during the consolidation process [35]. Evidence of the formation of such zinc-rich copper–zinc alloys during the reaction of Zn, Cu and P or Sb exists in the literature [36]. Attempts to image the cleaved composite pellets using back-scattered SEM to ensure their compositional uniformity/non-uniformity were not successful. The inability of the detector in resolving the presence of 20 nm nanoparticles prevented the accomplishment of this task. It is however believed that copper diffusion under the conditions employed for consolidating nanoparticle–nanowire mixtures is very low [37, 38]. As a result, it is believed that the compositional uniformity/non-uniformity in the original nanoparticle–nanowire mixtures is retained even after their consolidation into composite pellets. From the overall material composition analyses and the thermoelectric performance measurements of the composite pellets, it is clear that composite-A is uniform in composition and is composed of copper-doped Zn3 P2 nanowires interfaced with Cu0.2 Zn0.8 alloy. These pellets exhibited higher electrical conductivity and Seebeck coefficients, when compared to
unfunctionalized Zn3 P2 nanowire pellets. This behavior is unlike classical semiconductors, where an increase in Seebeck coefficient leads to a decrease in their electrical conductivity. The thermal conductivity of the composite remained close to that observed in unfunctionalized Zn3 P2 nanowire pellets. On the other hand, composite-B exhibited lower Seebeck coefficients and higher electrical conductivities, when compared to unfunctionalized nanowire pellets. This behavior is similar to that expected in classical semiconductors, where an increase in Seebeck coefficient leads to a decrease in their electrical conductivity. In essence, the type of procedure employed for the fabrication of composites dictates their ultimate thermoelectric performance. The thermoelectric performance of composite-B pellet can be explained by invoking the rule of mixtures [19] as the system is a mixture of Cu0.2 Zn0.8 nanoparticle and copper-doped Zn3 P2 nanowire components. The total electrical conductivity of this composite is then the sum of electrical conductivity contributions from the copper-doped nanowire component and the Cu0.2 Zn0.8 nanoparticle component [19]. The room temperature electrical conductivities of Cu–Zn alloys were previously reported to be in the 1.83 × 107 –3.33 × 107 S m−1 at 300 K [39]. At 700 K, the alloys exhibit electrical conductivities in the range of 6.77 × 106 –10.04 × 107 S m−1 [39]. 5
Nanotechnology 25 (2014) 125402
L Brockway et al
Figure 5. (a) A diagram representing the bandstructure of
composite-B. Although the barrier at the nanowire–nanowire interfaces is expected to lower electrical conductivities in the composite, the presence of metal alloy nanoparticle within it leads to enhanced electrical conductivity in the composite. (b) A diagram representing the bandstructure of composite-A. The formation of low energy the barrier at the nanowire–nanowire interfaces is expected to scatter low energy carrier at the interfaces and lead to an increase in its electrical conductivity and Seebeck coefficient.
thermoelectric performances by themselves were observed to be reproducible, as expected. A completely different scenario is in play in composite-A, where consolidation ultimately leads to the formation of copper-doped Zn3 P2 nanowires interfaced with Cu0.2 Zn0.8 alloy. The interface-engineered composite exhibited an increase in both the Seebeck coefficient and the electrical conductivity relative to unfunctionalized Zn3 P2 nanowires. The prevention of native oxide formation due to the functionalization of nanowires is believed to be responsible for the increase in electrical conductivity in this composite. Furthermore, it is believed that scattering of low energy charge carriers at the interfaces in the interface-engineered composites, is responsible for increase in the Seebeck coefficient in the composite [20]. This concept is pictorially represented in figures 5(a) and (b). Simple consolidation of Zn3 P2 nanowires leads to the formation of barrier for charge transport at the interfaces between the nanowires (figure 5(a)). The height of this barrier reduces because of the doping the Zn3 P2 nanowires with copper. The doping of the Zn3 P2 nanowires with copper (p-type doping [26, 36]), in conjunction with interface engineering is believed to reduce the overall barrier height to about a fraction of an eV [41]. This p-type doping of Zn3 P2 with copper increases the number of charge carriers [36]. The low energy barriers at the interfaces of the composite aid in enhancing the transport of charge carriers across them and hence lead to an increase in the electrical conductivity of the composite (figure 5(b)) relative to unfunctionalized nanowires (figure 5(a)). At the same time, the low energy barrier at the interfaces scatters low energy charge carriers and prevents any backward movement of the charge carriers. This helps in the preferential movement of charge carriers only in the forward direction and increases the overall Seebeck coefficient in the composite (figure 5(b)) [21]. The changes to the thermoelectric performance observed in composite-A, relative to that observed in unfunctionalized nanowire pellets by our group in a previous publication [28], is not expected to be due to carbon/sulfur doping of the nanowires. This can be clearly inferred from our previous study of the thermoelectric performance of BDT functionalized nanowire pellets [28].
Figure 4. (a) and (b) HR-TEM micrograph and the corresponding SAED diffraction pattern of a Zn3 P2 nanowire obtained by chipping a composite-A pellet. The analysis indicated that the nanowire still remains single-crystalline and exhibits the α-Zn3 P2 phase. (c) and (d) HR-TEM micrograph and the corresponding SAED diffraction pattern of a Zn3 P2 nanowire obtained by chipping a composite-B pellet. The analysis indicated that the nanowire still remains single-crystalline and exhibits the α-Zn3 P2 phase.
Our previous measurements indicate that the electrical conductivities of Zn3 P2 nanowire pellets are in the range of 6.65 × 10−5 –0.032 S m−1 in the 300–700 K temperature regime [28]. Even if the enhancement in the electrical conductivities of Zn3 P2 nanowires on doping with copper is taken into consideration, the composites are expected to exhibit electrical conductivities values that lie between those of Cu–Zn alloys and copper-doped Zn3 P2 nanowires, indicating that rule of mixtures clearly explain their thermoelectric performance. In such metal–semiconductor composites, the increase in electrical conductivities (figure 2(b)) afforded on addition of a metal component to the semiconductor component also leads to a decrease in their Seebeck coefficient (figure 2(a)). This is indeed the case in composite-B. The composition of composite-B also explains the increase in their thermal conductivity. The addition of Cu0.2 Zn0.8 nanoparticles to Zn3 P2 nanowires led not only to an increase in the electrical conductivity of the composite, but also its thermal conductivity (figure 2(c)). However, the increase in thermal conductivity is not as high as that expected from the known value of the thermal conductivity of Zn-0.7 wt% Cu alloy, 160–141.81 W m−1 K−1 in the 375–673 K temperature regime [40]. This is believed to be the result of enhanced phonon scattering at boundaries in the composite due to nanostructuring [4]. Although the composition of this composite is non-uniform at microscale, macroscale pellets were employed for the measurement of their thermoelectric performance. The macro sizes of the pellets allowed for the measurement of their average thermoelectric performance over the entire cross-section. The measured average 6
Nanotechnology 25 (2014) 125402
L Brockway et al
4. Conclusion
References
The interrelationship between the composition of metal– semiconductor composites and their ultimate thermoelectric performance were deduced. It was observed that use of functionalized nanowires, surface functionalized with bifunctional conjugated molecules, leads to the formation of interface-engineered composites that are uniform in composition all across their cross-sections, even at the microscale. The bifunctional molecules play a critical role and lead to the uniform decoration of each nanowire with metal nanoparticles during the mixing process. Hot uniaxial pressing of these metal nanoparticle decorated nanowires will then lead to the formation of interface-engineered composites, composed of doped nanowires interfaced with a metal (or metal alloy). Such composites reduce the barrier for charge transport across the nanowire interfaces and lead to an increase in their Seebeck coefficient and their electrical conductivities, especially when compared to that obtained in undoped nanowires. This behavior is unlike that observed in classical semiconductors. In sharp contrast, simple mixing of unfunctionalized nanowires and metal nanoparticles leads to the formation of composites that are non-uniform in composition at the microscale. The addition of metal leads to enhanced electrical conductivity in these composites. However, this enhancement comes at a loss of their Seebeck coefficient. This behavior is similar to that observed in classical semiconductors. Overall, the results indicate that addition of metal nanoparticles to semiconductor nanowires, either functionalized or unfunctionalized, leads to an increase in their overall thermoelectric performance. The magnitude of this enhancement is different and depends on the type of nanowires employed for the formation of composites, functionalized or unfunctionalized. In the case of composites fabricated using Zn3 P2 nanowires and copper nanoparticles, use of unfunctionalized nanowires offered the best thermoelectric performance. Specifically, a high zT value of 0.23 is achieved in the compositionally non-uniform composite-B pellets. These results indicate that high thermoelectric performance can be achieved even in non-traditional, inexpensive semiconductor materials by performing the following: (a) synthesizing and using the nanowire form of the semiconductor materials, and (b) coupling the semiconductor nanowires with metal nanoparticles and forming metal–semiconductor composites.
[1] Rowe D M 2006 Thermoelectrics Handbook Macro to Nano (Boca Raton, FL: CRC Press/Taylor and Francis) [2] Snyder G J and Toberer E S 2008 Complex thermoelectric materials Nature Mater. 7 105–14 [3] Caillat T, Fleurial J P, Snyder G J and Borshchevsky A 2001 Development of high efficiency segmented thermoelectric unicouples Proc. 20th Int. Conf. on Thermoelectrics pp 282–5 [4] Chen G and Shakouri A 2002 Heat transfer in nanostructures for solid-state energy conversion Trans. ASME J. Heat Transfer 124 242–52 [5] Mahan G D 1989 Figure of merit for thermoelectrics J. Appl. Phys. 65 1578–83 [6] Balandin A and Wang K L 1998 Effect of phonon confinement on the thermoelectric figure of merit of quantum wells J. Appl. Phys. 84 6149–53 [7] Balandin A and Wang K L 1998 Significant decrease of the lattice thermal conductivity due to phonon confinement in a free-standing semiconductor quantum well Phys. Rev. B 58 1544–9 [8] He J, Kanatzidis M G and Dravid V P 2013 High performance bulk thermoelectrics via a panoscopic approach Mater. Today 16 166–76 [9] Venkatasubramanian R, Siivola E, Colpitts T and O’Quinn B 2001 Thin-film thermoelectric devices with high room-temperature figures of merit Nature 413 597–602 [10] Hsu K F, Loo S, Guo F, Chen W, Dyck J S, Uher C, Hogan T, Polychroniadis E K and Kanatzidis M G 2004 Cubic AgPbm SbTe2+m : bulk thermoelectric materials with high figure of merit Science 303 818–21 [11] Heremans J P, Jovovic V, Toberer E S, Saramat A, Kurosaki K, Charoenphakdee A, Yamanaka S and Snyder G J 2008 Enhancement of thermoelectric efficiency in PbTe by distortion of the electronic density of states Science 321 554–7 [12] Hicks L D and Dresselhaus M S 1993 Thermoelectric figure of merit of a one-dimensional conductor Phys. Rev. B 47 16631–4 [13] Dresselhaus M S, Chen G, Tang M Y, Yang R G, Lee H, Wang D Z, Ren Z F, Fleurial J P and Gogna P 2007 New directions for low-dimensional thermoelectric materials Adv. Mater. 19 1043–53 [14] Boukai A I, Bunimovich Y, Tahir-Kheli J, Yu J K, Goddard W A and Heath J R 2008 Silicon nanowires as efficient thermoelectric materials Nature 451 168–71 [15] Hochbaum A I, Chen R K, Delgado R D, Liang W J, Garnett E C, Najarian M, Majumdar A and Yang P D 2008 Enhanced thermoelectric performance of rough silicon nanowires Nature 451 163-U5 [16] Bergman D J and Fel L G 1999 Enhancement of thermoelectric power factor in composite thermoelectrics J. Appl. Phys. 85 8205–16 [17] Broido D A and Mingo N 2006 Theory of the thermoelectric power factor in nanowire-composite matrix structures Phys. Rev. B 74 195325 [18] Zebarjadi M, Joshi G, Zhu G, Yu B, Minnich A, Lan Y, Wang X, Dresselhaus M, Ren Z and Chen G 2011 Power factor enhancement by modulation doping in bulk nanocomposites Nano Lett. 11 2225–30 [19] Lu Y, Sagara K, Matsuda Y, Hao L, Rong Jin Y and Yoshida H 2013 Effect of Cu powder addition on thermoelectric properties of Cu/TiO2−x composites Ceram. Int. 39 6689–94
Acknowledgments
The authors wish to thank Drs J P Fleurial and Sabah Bux of Jet Propulsion Laboratory for providing access to the tools necessary for the fabrication of nanowire–nanoparticle composite and the measurement of the thermoelectric performance of the composites. Aid of Dr Jacek Jasinski of the Conn Center for Renewable Energy at the University of Louisville with the TEM characterization of the nanowires and composites is gratefully acknowledged. Finally, financial support from the NSF/DOE thermoelectric partnership (NSF CBET #1048702) is gratefully acknowledged. 7
Nanotechnology 25 (2014) 125402
L Brockway et al
[20] Faleev S V and L´eonard F 2008 Theory of enhancement of thermoelectric properties of materials with nanoinclusions Phys. Rev. B 77 214304 [21] Martin J, Wang L, Chen L and Nolas G S 2009 Enhanced Seebeck coefficient through energy-barrier scattering in PbTe nanocomposites Phys. Rev. B 79 115311 [22] Zebarjadi M, Esfarjani K, Shakouri A, Bahk J-H, Bian Z, Zeng G, Bowers J, Lu H, Zide J and Gossard A 2009 Effect of nanoparticle scattering on thermoelectric power factor Appl. Phys. Lett. 94 202105 [23] Zide J M O et al 2010 High efficiency semimetal/ semiconductor nanocomposite thermoelectric materials J. Appl. Phys. 108 123702 123702-5 [24] Neophytou N, Zianni X, Kosina H, Frabboni S, Lorenzi B and Narducci D 2013 Simultaneous increase in electrical conductivity and Seebeck coefficient in highly boron-doped nanocrystalline Si Nanotechnology 24 205402 [25] Babu V S, Vaya P R and Sobhanadri J 1988 Electrical and thermoelectrical properties of Zn3 P2 films grown by the hot wall epitaxy technique J. Appl. Phys. 64 1922 [26] Nagamoto Y, Hino K, Yoshitake H, Koyanagi T and Ieee I 1998 Thermoelectric Properties of α-Zn3 P2 (New York: IEEE) pp 354–7 [27] Brockway L, Van Laer M, Kang Y and Vaddiraju S 2013 Large-scale synthesis and in situ functionalization of Zn3 P2 and Zn4 Sb3 nanowire powders Phys. Chem. Chem. Phys. 15 6260–7 [28] Brockway L, Vasiraju V, Asayesh-Ardakani H, Shahbazian-Yassar R and Vaddiraju S 2014 Thermoelectric properties of bulk Zn3 P2 nanowire assemblies Nanotechnology at press [29] O’neill M 1966 Measurement of specific heat functions by differential scanning calorimetry Anal. Chem. 38 1331–6 [30] Bux S K, Fleurial J-P and Kaner R B 2010 Nanostructured materials for thermoelectric applications Chem. Commun. 46 8311–24
[31] Schroder D K 2006 Semiconductor Material and Device Characterization (New York: Wiley) [32] Bux S K, Yeung M T, Toberer E S, Snyder G J, Kaner R B and Fleurial J P 2011 Mechanochemical synthesis and thermoelectric properties of high quality magnesium silicide J. Mater. Chem. 21 12259–66 [33] Massalski T B, Okamoto H, Subramanian P R and Kacprzak L 1990 Binary Alloy Phase Diagrams vol 3 (Materials Park, OH: ASM International) pp 2995–6 [34] Massalski T B and King H W 1962 The lattice spacing relationships in H.C.P. And η phases in the systems Cu–Zn, Ag–Zn; Au–Zn and Ag–Cd Acta Metall. 10 1171–81 [35] Schoonmaker R C, Venkitaraman A R and Lee P K 1967 The vaporization of zinc phosphide J. Phys. Chem. 71 2676–83 [36] Valset K, B¨ottger P H M, Taftø J and Finstad T G 2012 Thermoelectric properties of Cu doped ZnSb containing Zn3 P2 particles J. Appl. Phys. 111 023703 [37] Kullendorff N, Jansson L and Ledebo L Å 1983 Copper-related deep level defects in III–V semiconductors J. Appl. Phys. 54 3203–12 [38] Vaganov D V et al 2006 Determination of copper self-diffusion coefficients on the base of high-temperature creep data Defect Diffus. Forum 249 115–8 ¨ [39] Kaya H, C ¸ adırlı E and Ulgen A 2011 Investigation of the effect of composition on microhardness and determination of thermo-physical properties in the Zn–Cu alloys Mater. Des. 32 900–6 [40] Kaya H, B¨oy¨uk U, Engin S, C ¸ adirli E and Maras¸li N 2010 Measurements of microhardness and thermal and electrical properties of the binary Zn-0.7 wt% Cu Hypoperitectic Alloy J. Electron. Mater. 39 303–11 [41] Seto J Y W 1975 The electrical properties of polycrystalline silicon films J. Appl. Phys. 46 5247–54
8
Search
Collections
Journals
About
Contact us
My IOPscience
Compositional disorder and its effect on the thermoelectric performance of Zn3P2 nanowire–copper nanoparticle composites
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 125402 (http://iopscience.iop.org/0957-4484/25/12/125402) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 129.2.19.100 This content was downloaded on 04/02/2015 at 12:53
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 125402 (8pp)
doi:10.1088/0957-4484/25/12/125402
Compositional disorder and its effect on the thermoelectric performance of Zn3P2 nanowire–copper nanoparticle composites Lance Brockway1 , Venkata Vasiraju2 and Sreeram Vaddiraju1,2 1
Artie McFerrin Department of Chemical Engineering, Texas A&M University, 3122 TAMU, College Station, TX 77843, USA 2 Materials Science and Engineering Department, Texas A&M University, College Station, TX 77843, USA E-mail: [email protected] Received 7 November 2013, revised 13 January 2014 Accepted for publication 20 January 2014 Published 27 February 2014
Abstract
Recent studies indicated that nanowire format of materials is ideal for enhancing the thermoelectric performance of materials. Most of these studies were performed using individual nanowires as the test elements. It is not currently clear whether bulk assemblies of nanowires replicate this enhanced thermoelectric performance of individual nanowires. Therefore, it is imperative to understand whether enhanced thermoelectric performance exhibited by individual nanowires can be extended to bulk assemblies of nanowires. It is also imperative to know whether the addition of metal nanoparticle to semiconductor nanowires can be employed for enhancing their thermoelectric performance further. Specifically, it is important to understand the effect of microstructure and composition on the thermoelectric performance on bulk compound semiconductor nanowire–metal nanoparticle composites. In this study, bulk composites composed of mixtures of copper nanoparticles with either unfunctionalized or 1,4-benzenedithiol (BDT) functionalized Zn3 P2 nanowires were fabricated and analyzed for their thermoelectric performance. The results indicated that use of BDT functionalized nanowires for the fabrication of composites leads to interface-engineered composites that have uniform composition all across their cross-section. The interface engineering allows for increasing their Seebeck coefficients and electrical conductivities, relative to the Zn3 P2 nanowire pellets. In contrast, the use of unfunctionalized Zn3 P2 nanowires for the fabrication of composite leads to the formation of composites that are non-uniform in composition across their cross-section. Ultimately, the composites were found to have Zn3 P2 nanowires interspersed with metal alloy nanoparticles. Such non-uniform composites exhibited very high electrical conductivities, but slightly lower Seebeck coefficients, relative to Zn3 P2 nanowire pellets. These composites were found to show a very high zT of 0.23 at 770 K, orders of magnitude higher than either interface-engineered composites or Zn3 P2 nanowire pellets. The results indicate that microstructural composition of semiconductor nanowire–metal nanoparticle composites plays a major role in determining their thermoelectric performance, and such composites exhibit enhanced thermoelectric performance. Keywords: nanowires, Zn3 P2 , nanoparticle, composites, thermoelectrics, interface engineering, carrier filtering (Some figures may appear in colour only in the online journal)
0957-4484/14/125402+08$33.00
1
c 2014 IOP Publishing Ltd
Printed in the UK
Nanotechnology 25 (2014) 125402
L Brockway et al
1. Introduction
by Neophytou et al indicated that high level doping of nanocrystalline silicon with boron leads to a simultaneous increase in the Seebeck coefficient and the electrical conductivity [24]. In view of the widely varying results in these reports, it is essential to know the effect of metal nanoparticle inclusion on the thermoelectric performance of metal–semiconductor composites. More specifically, it is essential to determine the effect of microstructure and compositional uniformity on the thermoelectric performance of metal–semiconductor composites made by consolidating mixtures of metal and semiconductor nanopowders. The material chosen for this study is zinc phosphide (Zn3 P2 ), a compound semiconductor composed of only earth abundant elements. The thermoelectric properties of Zn3 P2 reported previously were optimistic, but varied widely [25, 26]. In this work, the effect of microstructure and compositional uniformity on the thermoelectric performances of Zn3 P2 nanowire–copper nanoparticle composites is presented. By the end of the manuscript, it will be shown that the compositional uniformity at microscale plays a very critical role in determining their overall thermoelectric performance. It will also be demonstrated that compositional uniformity/non-uniformity in such composites can be employed to engineer and greatly enhance the thermoelectric performance (or zT values) of materials.
Thermoelectric modules are solid-state devices that convert waste heat into electricity [1]. The fact that they have no moving parts and require minimal maintenance makes them indispensable for continued power generation in locations that are off the grid, such as satellites [2–4]. For the routine use of thermoelectrics in terrestrial applications, it is essential to enhance the performance of thermoelectrics beyond that obtained by the current state of the art. The performance of solid-state thermoelectrics can be gauged from their figure 2σ T [1, 2, 4, 5]. Here, S of merit, zT , given by zT = κSe +κ l is the Seebeck coefficient, σ is the electrical conductivity, and κe and κl are the thermal conductivities of the material from electronic and lattice contributions, respectively. Previous studies aimed at the enhancing the zT values of materials relied on circumventing the Wiedemann–Franz law through a selective reduction of their κl [4, 6–8]. This intense effort to enhance the zT by reducing κl through the creation of compositional inhomogenities or through the intentional addition of impurities to materials led to the fabrication of p-type Bi2 Te3 /Sb2 Te3 superlattices with a high zT of 2.4 [9], bulk material of AgPbm SbTe2+m with a zT of 2.2 [10] and Bi-doped n-type PbSeTe/PbTe quantum dot superlattice with zT as high as 3.5 [11]. It was also theoretically predicted [12, 13] and experimentally observed [14, 15] that materials in nanowire form offer the ability to tune the magnitudes of electrical and thermal transport through them. In fact, nanostructuring was employed to increase the room temperature zT value of rough silicon nanowires to 0.6 [15]. Another set of studies relied on enhancing the zT of materials by increasing their power factors, in addition to lowering their κl [16–18]. These studies employed the use of either two different semiconductors or metal–semiconductor composites for achieving this task. For instance, variation of the ratio of copper to titania in Cu/TiO2−x composites led to varying electrical conductivities and the Seebeck coefficients in the composites [19]. Here, the increase in electrical conductivities of the composites led to a decrease in their Seebeck coefficients [19]. The behavior of the composites in this scenario was explained using rule of mixtures [19]. Additionally, Faleev and Leonard predicted that metallic nanoinclusions within PbTe matrix selectively scatters electrons at the grain boundaries and leads to an enhancement in the Seebeck coefficient of the composites, relative to bulk PbTe [20]. They suggested that this Seebeck coefficient enhancement occurs only at very high doping levels [20]. Martin et al experimentally demonstrated this energy-barrier scattering concept in Ag-doped PbTe composites [21]. However, Zebarjadi et al showed that inclusion of ErAs nanoparticles inside InGaAlAs matrix only enhances the electrical conductivity and not the Seebeck coefficient [22]. In another report, they have demonstrated that semimetallic inclusions in semimetal/semiconductor composites can cause energy-dependent scattering and not only reduce the lattice thermal conductivities, but also simultaneously increase their power factor and lead to an overall enhancement of their zT values [23]. Finally, a recent study
2. Experimental details
Zn3 P2 nanowires necessary for the fabrication of the semiconductor nanowire–metal nanoparticle composites have been synthesized using direct reaction of zinc foils with phosphorus supplied via the vapor phase in a hot walled chemical vapor deposition (HWCVD) chamber [27]. The experimental procedure employed for this purpose was discussed in detail in a previous publication [27]. Following the synthesis, the nanowires were brushed off the foils and collected as powders. Similar procedure was also employed for obtaining Zn3 P2 surface functionalized with 1,4-benzenedithiol (BDT) [27, 28]. These surface functionalized nanowires will henceforth be referred to as BDT functionalized Zn3 P2 nanowires in this manuscript, while nanowires without any surface functionalization will be referred to as unfunctionalized Zn3 P2 nanowires. For the preparation of the composites, 1 g of Zn3 P2 nanowire powder was mixed with 40 mg of commercially-available copper nanoparticle powder (average diameter = 40 nm, obtained from US Research Nanomaterials, Inc.) in a ball mill. The obtained mixture was consolidated using hot uniaxial pressing at a temperature of 650 ◦ C and at a pressure of 120 MPa for 1 h into 12 mm-diameter, 1 mm-thick cylindrical pellets. Pellets of two different types of composites have been fabricated by hot uniaxial pressing copper nanoparticle powder mixed with either unfunctionalized or BDT functionalized nanowire powders. Thermoelectric performance of the resulting BDT functionalized Zn3 P2 nanowire–copper nanoparticle composite (referred to as composite A in this manuscript) pellets and unfunctionalized Zn3 P2 nanowire–copper nanoparticle composite (referred to as composite B in this manuscript) pellets were calculated 2
Nanotechnology 25 (2014) 125402
L Brockway et al
from the following data: heat capacity of the pellets determined using differential scanning calorimetry (DSC) [29], the thermal diffusivity determined using a laser flash apparatus (LFA) [30], electrical conductivity determined using the Van der Pauw 4-point probe method [31], and the Seebeck coefficient determined by imposing a temperature gradient across the pellet cross-section and measuring the resulting voltage [32]. For deducing the effect of pellet composition on their thermoelectric performance, the nanowire powders obtained and the composite pellets fabricated were also characterized for morphology, phase and chemical composition using scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HR-TEM), x-ray diffraction (XRD), and energy dispersive spectroscopy (EDS). For the sake of brevity, the characterization of the nanowire powders is not discussed in detail in the manuscript. For a detailed discussion of the morphology, phase and chemical composition of the nanowire powders, the reader is referred to previously published work [27].
Figure 1. XRD patterns of composite-A and composite-B. In both the cases, the addition of copper to Zn3 P2 nanowires and consolidation of the mixtures into high density composite pellets led to the formation of only one additional phase, Cu0.2 Zn0.8 . Major presence of α-Zn3 P2 parent phase was also observed in the spectra.
3. Results and discussion
were observed to be the same [28]. Similarly, the trend of the variation of the Seebeck coefficients with temperature of both composite-B pellets and the unfunctionalized Zn3 P2 nanowire pellet were observed to be the same. In cases where BDT functionalized nanowires were employed (i.e., composite-A pellets and BDT functionalized nanowire pellets), the trend of the Seebeck coefficient was observed to be along that lines of that expected of a non-degenerate semiconductor. The deviation from non-degenerate behavior in the other two samples, composite-B pellets and unfunctionalized Zn3 P2 nanowire pellets, at higher temperatures is owed to the carrier transport contribution of the surface oxide layers surrounding the Zn3 P2 nanowires. In slight contrast, the electrical conductivities of both composite-A pellet and composite-B pellet were observed to be higher than that observed in unfunctionalized Zn3 P2 nanowire pellet (figure 2(b)). So, the addition of copper nanoparticles to unfunctionalized Zn3 P2 nanowires enhanced the electrical conductivity of the composite and resulted in a reduction of its Seebeck coefficient. This is in line with the classical behavior of thermoelectric materials where an increase in the Seebeck coefficients results in a decrease of their electrical conductivities [2]. In sharp contrast, the addition of copper nanoparticles to BDT functionalized nanowires led to an increase in the Seebeck coefficient and the electrical conductivity of the composite-A relative to unfunctionalized Zn3 P2 nanowires. Such increase in the Seebeck coefficient and electrical conductivity was observed in BDT functionalized nanowire pellets in a previous study by our group [28]. Therefore, use of BDT functionalized nanowires for the fabrication of pellets always led to an increase in both the Seebeck coefficient and the electrical conductivity of the pellets, irrespective of the addition of copper nanoparticles. The Seebeck coefficients, electrical conductivities, thermal conductivities and zT values observed in both composite-A and composite-B at a temperature of 770 K are summarized in table 1. A maximum zT value of 0.23 was observed in composite-B.
Hot uniaxial pressing of mixtures of either unfunctionalized or BDT functionalized Zn3 P2 nanowires and copper nanoparticles led to the formation of 12 mm-diameter, 1 mm-thick cylindrical pellets. The densities of all the pellets were measured geometrically and confirmed using Archimedes principle [32]. In each case, the density of the pellets obtained was found to be ≥98% of the respective theoretical composite density. In other words, the porosity of the composite pellets was ≤2%. XRD analysis of the both composite-A and composite-B are depicted in figure 1. The analysis indicated that both the composite pellets are mainly composed of the α-Zn3 P2 nanowire phase [33], with a minor contribution from the copper–zinc alloy phase (specifically, the hexagonal Cu0.2 Zn0.8 phase with lattice parameters of a = 0.273 nm, c = 0.429 nm) [34]. 2θ reflection corresponding to the Cu0.2 Zn0.8 alloy phase is represented by a ‘*’ in figure 1. No presence of pure unreacted copper was detected in the XRD spectra of the composite pellets. Additionally, no presence of ZnS and Cu2 S/CuS phases, owed to the reaction of sulfur from BDT with either copper of the nanoparticles or the Zn of Zn3 P2 , was observed in the composite-A pellet. Thermoelectric performance of the pellets was deduced by independently measuring the Seebeck coefficient, the electrical conductivity, and the thermal conductivity and are presented in figure 2. For comparison, the Seebeck coefficient, the electrical conductivity and the thermal conductivity data obtained from highly dense unfunctionalized Zn3 P2 nanowire pellet reported in a previous study is also presented in figure 2 [28]. When compared to the data from unfunctionalized Zn3 P2 nanowire pellet, the Seebeck coefficient of the composite-A pellet was observed to be higher (figure 2(a)). However, the Seebeck coefficient of composite-B pellet was observed to be lower than that observed from unfunctionalized nanowire pellet (figure 2(a)). The trend of the variation of the Seebeck coefficients with temperature of both the composite-A pellets and the BDT functionalized Zn3 P2 nanowire pellets 3
Nanotechnology 25 (2014) 125402
L Brockway et al
Table 1. Values of Seebeck coefficients, electrical conductivities, thermal conductivities, and zT values of composite-A and composite-B at
770 K. A maximum zT value of 0.23 was observed in composite-B. Sample
Seebeck coefficient (µV K−1 )
Electrical conductivity (S m−1 )
Thermal conductivity (W m−1 K−1 )
Maximum zT
Composite-A Composite-B
362.96 246.44
38.52 8457.82
0.92 1.64
0.0042 0.23
functionalized and unfunctionalized nanowires mixed with copper nanoparticles (in the same weight fractions employed for the fabrication of composite pellets) using a mortar and pestle was performed. A micrograph of the BDT functionalized nanowires mixed with copper nanoparticles is presented in figure 3(a). In this case, the process of mixing led to the uniform decoration of BDT functionalized nanowire surfaces with copper nanoparticles. This is owed to the formation of bonds between the free –SH groups of the BDT molecules present on top of the functionalized nanowires and the copper nanoparticles (figure 3(b)) [27]. This indicates that on hot uniaxial pressing, mixtures of BDT functionalized Zn3 P2 nanowires and copper nanoparticles lead to the formation of interface-engineered composites that are uniform in composition. It is believed that each pellet is composed of copper-doped Zn3 P2 nanowires interfaced with Cu0.2 Zn0.8 metal alloy particles (figure 3(b)). The BDT functional molecule barrier at the nanowire interfaces is expected to prevent, at least partially, the doping of the Zn3 P2 nanowires with copper in this composite. A completely different scenario was observed when unfunctionalized nanowires are mixed with copper nanoparticles. No decoration of the nanowires with copper nanoparticles was observed in the micrograph of the samples (figure 3(c)). The copper nanoparticles and Zn3 P2 nanowires exist as individual units within the mixture as observed in the micrograph (figure 3(c)). Owing to the presence of individual unit of copper nanoparticles and Zn3 P2 nanowires, the sample is non-uniform in composition across the sample at microscale (figures 3(c) and (d)). Hot uniaxial pressing of this mixture is believed to lead to the formation of Cu0.2 Zn0.8 metal alloy particles interspersed within the copper-doped Zn3 P2 nanowire matrix. It is further believed that the composite is non-uniform in composition with varying ratios of the amount of Cu0.2 Zn0.8 metal alloy and copper-doped Zn3 P2 nanowires (figure 3(d)) through its cross-section. Destructive analysis of the composite pellets confirmed the above analysis and indicated that the composite pellets are mainly composed of single-crystalline Zn3 P2 nanowires (figure 4). These experiments were performed by chipping pieces of the pellets directly onto nickel TEM grids, followed by analyzing the grids using a HR-TEM. The analysis of the chipped pieces indicated that single-crystalline Zn3 P2 nanowires are still present in the composite pellets (figures 4(a) and (c)). It is essential to recall here that flexibility of the nanowires allows for their high density packing into dense pellets, a concept that has been described in detail previously [28]. The selected area electron diffraction (SAED) analysis of a nanowire chipped from composite-A (figure 4(b)) confirmed that it is still single-crystalline in nature. Similar
Figure 2. Plots indicating the variation of (a) Seebeck coefficients, (b) electrical conductivities, (c) thermal conductivities, and (d) zT values of composite-A and composite-B. For comparison, data obtained from unfunctionalized Zn3 P2 nanowire pellets obtained by our group in a previous study [28] was also included in (a)–(c).
Thermal conductivity of the pellets indicated that the addition of copper to the both unfunctionalized and BDT functionalized Zn3 P2 nanowires led to an increase in the thermal conductivity of the pellets. However, the increase in the thermal conductivity of composite-A pellet was observed to be minor, compared to that observed in unfunctionalized Zn3 P2 nanowire pellet. The thermal conductivity of composite-A pellet composite was observed to vary between 1.49 and 0.92 W m−1 K−1 in the 300–775 K temperature regime. Contrastingly, composite-B pellet exhibited a relatively high thermal conductivity of 2.54–1.64 W m−1 K−1 in the 300– 775 K temperature regime. Finally, the variation of the zT values of both the composite pellets with temperature is presented in figure 2(d). A maximum zT value of 0.23 was achieved in compositeB pellet, orders of magnitude higher than the maximum zT value of 0.0042 achieved in composite-A pellet. Such high zT values of 0.23 have never been reported in Zn3 P2 materials system. To understand and explain the variations in the thermoelectric behavior of the composites and deduce composition-thermoelectric property interrelationships in them, additional microscopic composition analysis of the pellets was performed. Both destructive and non-destructive techniques were employed for deducing the variations in the chemical compositions of the pellets at microscale. First and foremost, morphological analysis of both BDT 4
Nanotechnology 25 (2014) 125402
L Brockway et al
Figure 3. (a) Scanning electron micrograph of copper nanoparticle decorated Zn3 P2 nanowires obtained by simply mixing BDT functionalized Zn3 P2 nanowires with copper nanoparticles in a mortar and pestle. (b) A schematic representing the steps underlying the formation of composite-A. The decoration of BDT functionalized Zn3 P2 nanowire surfaces with copper nanoparticles during the nanomaterial mixing phase leads to the formation of interface-engineered composites during the consolidation phase. (c) Scanning electron micrograph of mixtures of unfunctionalized Zn3 P2 nanowires and copper nanoparticles. No decoration of the nanowires with copper nanoparticles occurred. The nanomaterials remained as individual units after mixing. (d) A schematic representing the steps underlying the formation of composite-B. Here, consolidation leads to the formation of non-uniform composites that has metal alloy interspersed between copper-doped Zn3 P2 nanowires.
confirmation was also obtained from the SAED analysis of a chipped nanowire from composite-B pellet (figure 4(d)). In both the composites, the formation of zinc-rich Cu0.2 Zn0.8 metal alloy phase is believed to be the because of the partial decomposition of Zn3 P2 nanowires occurring during the consolidation process [35]. Evidence of the formation of such zinc-rich copper–zinc alloys during the reaction of Zn, Cu and P or Sb exists in the literature [36]. Attempts to image the cleaved composite pellets using back-scattered SEM to ensure their compositional uniformity/non-uniformity were not successful. The inability of the detector in resolving the presence of 20 nm nanoparticles prevented the accomplishment of this task. It is however believed that copper diffusion under the conditions employed for consolidating nanoparticle–nanowire mixtures is very low [37, 38]. As a result, it is believed that the compositional uniformity/non-uniformity in the original nanoparticle–nanowire mixtures is retained even after their consolidation into composite pellets. From the overall material composition analyses and the thermoelectric performance measurements of the composite pellets, it is clear that composite-A is uniform in composition and is composed of copper-doped Zn3 P2 nanowires interfaced with Cu0.2 Zn0.8 alloy. These pellets exhibited higher electrical conductivity and Seebeck coefficients, when compared to
unfunctionalized Zn3 P2 nanowire pellets. This behavior is unlike classical semiconductors, where an increase in Seebeck coefficient leads to a decrease in their electrical conductivity. The thermal conductivity of the composite remained close to that observed in unfunctionalized Zn3 P2 nanowire pellets. On the other hand, composite-B exhibited lower Seebeck coefficients and higher electrical conductivities, when compared to unfunctionalized nanowire pellets. This behavior is similar to that expected in classical semiconductors, where an increase in Seebeck coefficient leads to a decrease in their electrical conductivity. In essence, the type of procedure employed for the fabrication of composites dictates their ultimate thermoelectric performance. The thermoelectric performance of composite-B pellet can be explained by invoking the rule of mixtures [19] as the system is a mixture of Cu0.2 Zn0.8 nanoparticle and copper-doped Zn3 P2 nanowire components. The total electrical conductivity of this composite is then the sum of electrical conductivity contributions from the copper-doped nanowire component and the Cu0.2 Zn0.8 nanoparticle component [19]. The room temperature electrical conductivities of Cu–Zn alloys were previously reported to be in the 1.83 × 107 –3.33 × 107 S m−1 at 300 K [39]. At 700 K, the alloys exhibit electrical conductivities in the range of 6.77 × 106 –10.04 × 107 S m−1 [39]. 5
Nanotechnology 25 (2014) 125402
L Brockway et al
Figure 5. (a) A diagram representing the bandstructure of
composite-B. Although the barrier at the nanowire–nanowire interfaces is expected to lower electrical conductivities in the composite, the presence of metal alloy nanoparticle within it leads to enhanced electrical conductivity in the composite. (b) A diagram representing the bandstructure of composite-A. The formation of low energy the barrier at the nanowire–nanowire interfaces is expected to scatter low energy carrier at the interfaces and lead to an increase in its electrical conductivity and Seebeck coefficient.
thermoelectric performances by themselves were observed to be reproducible, as expected. A completely different scenario is in play in composite-A, where consolidation ultimately leads to the formation of copper-doped Zn3 P2 nanowires interfaced with Cu0.2 Zn0.8 alloy. The interface-engineered composite exhibited an increase in both the Seebeck coefficient and the electrical conductivity relative to unfunctionalized Zn3 P2 nanowires. The prevention of native oxide formation due to the functionalization of nanowires is believed to be responsible for the increase in electrical conductivity in this composite. Furthermore, it is believed that scattering of low energy charge carriers at the interfaces in the interface-engineered composites, is responsible for increase in the Seebeck coefficient in the composite [20]. This concept is pictorially represented in figures 5(a) and (b). Simple consolidation of Zn3 P2 nanowires leads to the formation of barrier for charge transport at the interfaces between the nanowires (figure 5(a)). The height of this barrier reduces because of the doping the Zn3 P2 nanowires with copper. The doping of the Zn3 P2 nanowires with copper (p-type doping [26, 36]), in conjunction with interface engineering is believed to reduce the overall barrier height to about a fraction of an eV [41]. This p-type doping of Zn3 P2 with copper increases the number of charge carriers [36]. The low energy barriers at the interfaces of the composite aid in enhancing the transport of charge carriers across them and hence lead to an increase in the electrical conductivity of the composite (figure 5(b)) relative to unfunctionalized nanowires (figure 5(a)). At the same time, the low energy barrier at the interfaces scatters low energy charge carriers and prevents any backward movement of the charge carriers. This helps in the preferential movement of charge carriers only in the forward direction and increases the overall Seebeck coefficient in the composite (figure 5(b)) [21]. The changes to the thermoelectric performance observed in composite-A, relative to that observed in unfunctionalized nanowire pellets by our group in a previous publication [28], is not expected to be due to carbon/sulfur doping of the nanowires. This can be clearly inferred from our previous study of the thermoelectric performance of BDT functionalized nanowire pellets [28].
Figure 4. (a) and (b) HR-TEM micrograph and the corresponding SAED diffraction pattern of a Zn3 P2 nanowire obtained by chipping a composite-A pellet. The analysis indicated that the nanowire still remains single-crystalline and exhibits the α-Zn3 P2 phase. (c) and (d) HR-TEM micrograph and the corresponding SAED diffraction pattern of a Zn3 P2 nanowire obtained by chipping a composite-B pellet. The analysis indicated that the nanowire still remains single-crystalline and exhibits the α-Zn3 P2 phase.
Our previous measurements indicate that the electrical conductivities of Zn3 P2 nanowire pellets are in the range of 6.65 × 10−5 –0.032 S m−1 in the 300–700 K temperature regime [28]. Even if the enhancement in the electrical conductivities of Zn3 P2 nanowires on doping with copper is taken into consideration, the composites are expected to exhibit electrical conductivities values that lie between those of Cu–Zn alloys and copper-doped Zn3 P2 nanowires, indicating that rule of mixtures clearly explain their thermoelectric performance. In such metal–semiconductor composites, the increase in electrical conductivities (figure 2(b)) afforded on addition of a metal component to the semiconductor component also leads to a decrease in their Seebeck coefficient (figure 2(a)). This is indeed the case in composite-B. The composition of composite-B also explains the increase in their thermal conductivity. The addition of Cu0.2 Zn0.8 nanoparticles to Zn3 P2 nanowires led not only to an increase in the electrical conductivity of the composite, but also its thermal conductivity (figure 2(c)). However, the increase in thermal conductivity is not as high as that expected from the known value of the thermal conductivity of Zn-0.7 wt% Cu alloy, 160–141.81 W m−1 K−1 in the 375–673 K temperature regime [40]. This is believed to be the result of enhanced phonon scattering at boundaries in the composite due to nanostructuring [4]. Although the composition of this composite is non-uniform at microscale, macroscale pellets were employed for the measurement of their thermoelectric performance. The macro sizes of the pellets allowed for the measurement of their average thermoelectric performance over the entire cross-section. The measured average 6
Nanotechnology 25 (2014) 125402
L Brockway et al
4. Conclusion
References
The interrelationship between the composition of metal– semiconductor composites and their ultimate thermoelectric performance were deduced. It was observed that use of functionalized nanowires, surface functionalized with bifunctional conjugated molecules, leads to the formation of interface-engineered composites that are uniform in composition all across their cross-sections, even at the microscale. The bifunctional molecules play a critical role and lead to the uniform decoration of each nanowire with metal nanoparticles during the mixing process. Hot uniaxial pressing of these metal nanoparticle decorated nanowires will then lead to the formation of interface-engineered composites, composed of doped nanowires interfaced with a metal (or metal alloy). Such composites reduce the barrier for charge transport across the nanowire interfaces and lead to an increase in their Seebeck coefficient and their electrical conductivities, especially when compared to that obtained in undoped nanowires. This behavior is unlike that observed in classical semiconductors. In sharp contrast, simple mixing of unfunctionalized nanowires and metal nanoparticles leads to the formation of composites that are non-uniform in composition at the microscale. The addition of metal leads to enhanced electrical conductivity in these composites. However, this enhancement comes at a loss of their Seebeck coefficient. This behavior is similar to that observed in classical semiconductors. Overall, the results indicate that addition of metal nanoparticles to semiconductor nanowires, either functionalized or unfunctionalized, leads to an increase in their overall thermoelectric performance. The magnitude of this enhancement is different and depends on the type of nanowires employed for the formation of composites, functionalized or unfunctionalized. In the case of composites fabricated using Zn3 P2 nanowires and copper nanoparticles, use of unfunctionalized nanowires offered the best thermoelectric performance. Specifically, a high zT value of 0.23 is achieved in the compositionally non-uniform composite-B pellets. These results indicate that high thermoelectric performance can be achieved even in non-traditional, inexpensive semiconductor materials by performing the following: (a) synthesizing and using the nanowire form of the semiconductor materials, and (b) coupling the semiconductor nanowires with metal nanoparticles and forming metal–semiconductor composites.
[1] Rowe D M 2006 Thermoelectrics Handbook Macro to Nano (Boca Raton, FL: CRC Press/Taylor and Francis) [2] Snyder G J and Toberer E S 2008 Complex thermoelectric materials Nature Mater. 7 105–14 [3] Caillat T, Fleurial J P, Snyder G J and Borshchevsky A 2001 Development of high efficiency segmented thermoelectric unicouples Proc. 20th Int. Conf. on Thermoelectrics pp 282–5 [4] Chen G and Shakouri A 2002 Heat transfer in nanostructures for solid-state energy conversion Trans. ASME J. Heat Transfer 124 242–52 [5] Mahan G D 1989 Figure of merit for thermoelectrics J. Appl. Phys. 65 1578–83 [6] Balandin A and Wang K L 1998 Effect of phonon confinement on the thermoelectric figure of merit of quantum wells J. Appl. Phys. 84 6149–53 [7] Balandin A and Wang K L 1998 Significant decrease of the lattice thermal conductivity due to phonon confinement in a free-standing semiconductor quantum well Phys. Rev. B 58 1544–9 [8] He J, Kanatzidis M G and Dravid V P 2013 High performance bulk thermoelectrics via a panoscopic approach Mater. Today 16 166–76 [9] Venkatasubramanian R, Siivola E, Colpitts T and O’Quinn B 2001 Thin-film thermoelectric devices with high room-temperature figures of merit Nature 413 597–602 [10] Hsu K F, Loo S, Guo F, Chen W, Dyck J S, Uher C, Hogan T, Polychroniadis E K and Kanatzidis M G 2004 Cubic AgPbm SbTe2+m : bulk thermoelectric materials with high figure of merit Science 303 818–21 [11] Heremans J P, Jovovic V, Toberer E S, Saramat A, Kurosaki K, Charoenphakdee A, Yamanaka S and Snyder G J 2008 Enhancement of thermoelectric efficiency in PbTe by distortion of the electronic density of states Science 321 554–7 [12] Hicks L D and Dresselhaus M S 1993 Thermoelectric figure of merit of a one-dimensional conductor Phys. Rev. B 47 16631–4 [13] Dresselhaus M S, Chen G, Tang M Y, Yang R G, Lee H, Wang D Z, Ren Z F, Fleurial J P and Gogna P 2007 New directions for low-dimensional thermoelectric materials Adv. Mater. 19 1043–53 [14] Boukai A I, Bunimovich Y, Tahir-Kheli J, Yu J K, Goddard W A and Heath J R 2008 Silicon nanowires as efficient thermoelectric materials Nature 451 168–71 [15] Hochbaum A I, Chen R K, Delgado R D, Liang W J, Garnett E C, Najarian M, Majumdar A and Yang P D 2008 Enhanced thermoelectric performance of rough silicon nanowires Nature 451 163-U5 [16] Bergman D J and Fel L G 1999 Enhancement of thermoelectric power factor in composite thermoelectrics J. Appl. Phys. 85 8205–16 [17] Broido D A and Mingo N 2006 Theory of the thermoelectric power factor in nanowire-composite matrix structures Phys. Rev. B 74 195325 [18] Zebarjadi M, Joshi G, Zhu G, Yu B, Minnich A, Lan Y, Wang X, Dresselhaus M, Ren Z and Chen G 2011 Power factor enhancement by modulation doping in bulk nanocomposites Nano Lett. 11 2225–30 [19] Lu Y, Sagara K, Matsuda Y, Hao L, Rong Jin Y and Yoshida H 2013 Effect of Cu powder addition on thermoelectric properties of Cu/TiO2−x composites Ceram. Int. 39 6689–94
Acknowledgments
The authors wish to thank Drs J P Fleurial and Sabah Bux of Jet Propulsion Laboratory for providing access to the tools necessary for the fabrication of nanowire–nanoparticle composite and the measurement of the thermoelectric performance of the composites. Aid of Dr Jacek Jasinski of the Conn Center for Renewable Energy at the University of Louisville with the TEM characterization of the nanowires and composites is gratefully acknowledged. Finally, financial support from the NSF/DOE thermoelectric partnership (NSF CBET #1048702) is gratefully acknowledged. 7
Nanotechnology 25 (2014) 125402
L Brockway et al
[20] Faleev S V and L´eonard F 2008 Theory of enhancement of thermoelectric properties of materials with nanoinclusions Phys. Rev. B 77 214304 [21] Martin J, Wang L, Chen L and Nolas G S 2009 Enhanced Seebeck coefficient through energy-barrier scattering in PbTe nanocomposites Phys. Rev. B 79 115311 [22] Zebarjadi M, Esfarjani K, Shakouri A, Bahk J-H, Bian Z, Zeng G, Bowers J, Lu H, Zide J and Gossard A 2009 Effect of nanoparticle scattering on thermoelectric power factor Appl. Phys. Lett. 94 202105 [23] Zide J M O et al 2010 High efficiency semimetal/ semiconductor nanocomposite thermoelectric materials J. Appl. Phys. 108 123702 123702-5 [24] Neophytou N, Zianni X, Kosina H, Frabboni S, Lorenzi B and Narducci D 2013 Simultaneous increase in electrical conductivity and Seebeck coefficient in highly boron-doped nanocrystalline Si Nanotechnology 24 205402 [25] Babu V S, Vaya P R and Sobhanadri J 1988 Electrical and thermoelectrical properties of Zn3 P2 films grown by the hot wall epitaxy technique J. Appl. Phys. 64 1922 [26] Nagamoto Y, Hino K, Yoshitake H, Koyanagi T and Ieee I 1998 Thermoelectric Properties of α-Zn3 P2 (New York: IEEE) pp 354–7 [27] Brockway L, Van Laer M, Kang Y and Vaddiraju S 2013 Large-scale synthesis and in situ functionalization of Zn3 P2 and Zn4 Sb3 nanowire powders Phys. Chem. Chem. Phys. 15 6260–7 [28] Brockway L, Vasiraju V, Asayesh-Ardakani H, Shahbazian-Yassar R and Vaddiraju S 2014 Thermoelectric properties of bulk Zn3 P2 nanowire assemblies Nanotechnology at press [29] O’neill M 1966 Measurement of specific heat functions by differential scanning calorimetry Anal. Chem. 38 1331–6 [30] Bux S K, Fleurial J-P and Kaner R B 2010 Nanostructured materials for thermoelectric applications Chem. Commun. 46 8311–24
[31] Schroder D K 2006 Semiconductor Material and Device Characterization (New York: Wiley) [32] Bux S K, Yeung M T, Toberer E S, Snyder G J, Kaner R B and Fleurial J P 2011 Mechanochemical synthesis and thermoelectric properties of high quality magnesium silicide J. Mater. Chem. 21 12259–66 [33] Massalski T B, Okamoto H, Subramanian P R and Kacprzak L 1990 Binary Alloy Phase Diagrams vol 3 (Materials Park, OH: ASM International) pp 2995–6 [34] Massalski T B and King H W 1962 The lattice spacing relationships in H.C.P. And η phases in the systems Cu–Zn, Ag–Zn; Au–Zn and Ag–Cd Acta Metall. 10 1171–81 [35] Schoonmaker R C, Venkitaraman A R and Lee P K 1967 The vaporization of zinc phosphide J. Phys. Chem. 71 2676–83 [36] Valset K, B¨ottger P H M, Taftø J and Finstad T G 2012 Thermoelectric properties of Cu doped ZnSb containing Zn3 P2 particles J. Appl. Phys. 111 023703 [37] Kullendorff N, Jansson L and Ledebo L Å 1983 Copper-related deep level defects in III–V semiconductors J. Appl. Phys. 54 3203–12 [38] Vaganov D V et al 2006 Determination of copper self-diffusion coefficients on the base of high-temperature creep data Defect Diffus. Forum 249 115–8 ¨ [39] Kaya H, C ¸ adırlı E and Ulgen A 2011 Investigation of the effect of composition on microhardness and determination of thermo-physical properties in the Zn–Cu alloys Mater. Des. 32 900–6 [40] Kaya H, B¨oy¨uk U, Engin S, C ¸ adirli E and Maras¸li N 2010 Measurements of microhardness and thermal and electrical properties of the binary Zn-0.7 wt% Cu Hypoperitectic Alloy J. Electron. Mater. 39 303–11 [41] Seto J Y W 1975 The electrical properties of polycrystalline silicon films J. Appl. Phys. 46 5247–54
8
