Effect of Intertube Junctions on the Thermoelectric Power of

Article pubs.acs.org/JPCC

Effect of Intertube Junctions on the Thermoelectric Power of Monodispersed Single Walled Carbon Nanotube Networks Mingxing Piao,†,⊥ Min-Kyu Joo,†,‡,⊥ Junhong Na,† Yun-Jeong Kim,† Mireille Mouis,‡ Gérard Ghibaudo,‡ Siegmar Roth,§ Wung-Yeon Kim,† Ho-Kyun Jang,† Gary P. Kennedy,∥ Urszula Dettlaff-Weglikowska,*,† and Gyu-Tae Kim*,† †

School of Electrical Engineering, Korea University, 145, Anam-ro, Seongbuk-gu, 136-701 Seoul, South Korea IMEP-LAHC, Grenoble INP, Minatec, CS 50257, 38016 Grenoble, France § Sineurop Nanotech GmbH, Muenchner Freiheit 6, 80802 München, Germany ∥ School of Electrical Engineering, Electronics and Computer Science, University of Liverpool, Ashton Street, Liverpool L69 3BX, United Kingdom ‡

ABSTRACT: The thermoelectric power (TEP) of single walled carbon nanotube (SWCNT) thin films in pure metallic SWCNT (m-SWCNT) and pure semiconducting SWCNT (sSWCNT) networks as well as in m- and s-SWCNT mixtures is investigated. The TEP measured on the pure s-SWCNT film (≈88 μV/K) was found to be almost 7 times higher than that of the m-SWCNTs (≈13 μV/K). Moreover, a quasilinear increase of TEP of the mixed SWCNT networks was observed as the fraction of s-SWCNTs is increased. The experimentally determined relationship between TEP and the fraction of s-SWCNTs in the mixture allows fast and simple quantitative analysis of the s:m ratio in any as-prepared heterogeneous SWCNT network. Furthermore, a semiempirical model analyzing the effect of the intertube junctions is proposed and applied to describe the thermoelectric behavior of the prepared SWCNT networks. The results of calculations match well with the experimental data and clearly demonstrate that the measured TEP of thin SWCNT films is principally controlled by the intertube junctions. The fundamental role of junctions in generating thermoelectric power is not limited to only SWCNT networks as discovered here, but also could be applied to systems where nanoparticles/nanotube form percolating paths in thin films and composite materials.

1. INTRODUCTION

understanding of the transport mechanism in SWCNT networks. The electrical conductivity of SWCNT thin films with homogeneous electronic type and precisely tuned ratios of mand s-SWCNTs has been studied by several research groups.10,16−18 However, the effect of electronic type (m or s) on the thermoelectric power (TEP) of thin films was not yet studied systematically so far, although the importance of the intertube contacts was realized by analysis of bundles of aligned multiwalled carbon nanotubes, by comparison with those in randomly oriented networks.19 Moreover, the commonly fabricated SWCNTs were assumed to be a heterogeneous mixture of m-SWCNTs and s-SWCNTs in a ratio of 1:2. The measured TEP of such samples was a result of an ensemble effects stemming from both m- and s-SWCNTs. As a consequence, large differences of TEP arising from diverse m:s ratios in heterogeneous SWCNT networks were observed.20−22 This paper describes the thermoelectric properties of the SWCNT thin films of monodispersed m- and s-SWCNTs and

Single walled carbon nanotube (SWCNT) thin film networks were investigated for use in a variety of applications such as solar cells,1,2 field-effect transistors,3,4 photovoltaics,5 flat-panel displays,6 and thermoelectric power generators.7−9 In particular, SWCNT thin films are considered as attractive candidates to replace traditional transparent conducting oxides such as In2O3:Sn, SnO2:F, or ZnO:Al10 due to their low cost fabrication, high flexibility, and nontoxicity. In order to fully realize the potential of SWCNT networks for flexible devices, a fundamental understanding of the inherent electrical and thermal transport properties is essential. For this reason, many research groups have considered SWCNT network morphology,11 chemical dopants,12,13 surface defects,14 tube chiralities,15 and band structure15 in their studies. However, some questions are yet to be clarified concerning the random SWCNT network, especially the effect of intertube junctions on the electrical and thermal transport properties of the network. The electronic nature of the tubes (m- or s-type) and the ratio between m-SWCNTs and s-SWCNTs in a network can significantly modify its electrical and thermal properties. Here, thin films fabricated from monodispersed SWCNTs with their high electronic homogeneity contribute to a better © 2014 American Chemical Society

Received: June 8, 2014 Revised: October 1, 2014 Published: October 27, 2014 26454

dx.doi.org/10.1021/jp505682f | J. Phys. Chem. C 2014, 118, 26454−26461

The Journal of Physical Chemistry C

Article

the sample was adjusted between room temperature and 373 K by changing the current supply to the soldering iron, while the other end of the sample was exposed to air. Thus, a temperature difference (ΔT) was established and maintained across the SWCNT film. Two platinum (Pt 100Ω) resistors were clamped with the electrical contacts at both ends of the film strip to measure the sample temperature. A schematic picture showing the experimental arrangement of electrical contacts and the positions of the temperature sensors on a strip of the PC membrane, supporting the SWCNT films, is shown in Figure 1a. The Seebeck coefficient S = −ΔV/ΔT was easily

their mixtures with defined compositions. Our investigations contribute to the elucidation of transport mechanisms through thin SWCNT networks by considering the effect of intertube junctions (m−m and s−s) on the TEP in order to optimize the thermoelectric properties of the SWCNT networks. The TEP measurements were performed on a series of SWCNT flexible networks prepared by vacuum filtration of homogeneous electronic type (m or s) SWCNT dispersions and their mixtures with precisely controlled m:s ratios. The following section presents and discusses the experimental data in conjunction with the results obtained from TEP modeling. The model developed here can be applied to any heterogeneous network comprising of metallic or semiconducting nanomaterials.

2. EXPERIMENTAL SECTION 2.1. Monodispersed SWCNT Suspensions. The monodispersed m- and s-SWCNTs with greater than 99% homogeneity were purchased directly from the manufacturer (NanoIntegris Inc.) and used in the following study as-supplied without any further treatment. The mean length of the sSWCNTs was ≈1 μm, while that of the m-SWCNTs was ≈0.5 μm. The average diameter of the both types of tubes was ≈1.4 nm. The SWCNT concentration in both dispersions (m- and sSWCNTs) was 0.001 mg/mL. 2.2. Preparation of the SWCNT Thin Films. The flexible SWCNT thin films of m- and s-SWCNTs and their mixtures with varying m:s ratios were prepared by vacuum filtration of 10 mL of each electronic type SWCNT dispersions and 10 mL of defined mixture using a polycarbonate (PC) membrane (pore size 0.4 μm, Millipore). Then the surfactant was removed by rinsing several times with deionized water. Finally, the SWCNT thin films deposited on the membrane were dried at 80 °C for 30 min to evaporate water. The thickness of the SWCNT thin films was roughly obtained by using the density of a thick bucky paper (used for calibration), and the known amount of the filtrated dispersion was found to be 300 ± 20 nm. 2.3. Characterization and Measurement Methods. The surface morphology of the as deposited monodispersed SWCNT thin films was observed by using a scanning electron microscope (SEM, Hitachi S-4800). The SWCNT diameter and chirality was determined by Raman spectrometer (JobinYvon HR800UV) equipped with three different laser lines (514, 633, 785 nm). To confirm the ratio of the m- and s-SWCNTs in the prepared mixtures, the UV−vis−NIR absorption spectra (Cary 5000) were analyzed. Besides, to measure the electrical conductivity, the SWCNT thin films supported by the PC membrane were cut into narrow strips, typically 10 mm × 30 mm. Each sample was then pressure-contacted with four parallel metallic wires to perform the standard four probe measurement at room temperature. Current−voltage (I−V) sweeps were carried out by using a Keithley 238 as the current source and a 34401A multimeter from Hewlett-Packard (HP) as the voltage measurement. The electrical conductivity of the SWCNT samples was then determined from the sample geometries and the slope of the I−V curve. Afterwards, the measurement of the Seebeck coefficient was performed in ambient air according to the method described in detail elsewhere for flexible films.23 To this end, the SWCNT thin films supported by the PC membrane were contacted with a copper clip, in which a soldering iron tip as the heat source was mounted inside the copper clip. The temperature at one end of

Figure 1. (a) Schematic view demonstrating the TEP measurement of the SWCNT films supported by the PC membrane. (b) An example for the Seebeck coefficient determination using the relationship ΔV versus ΔT: S = −ΔV/ΔT.

determined from the slope of the measured thermoelectric voltage ΔV versus ΔT, while ΔT increased in 5 K intervals from room temperature to 373 K, as demonstrated in Figure 1b. The TEP measurements were carried out several times for each film sample. The error bars represent differences in measured values from different samples with the same kind of film.

3. RESULTS AND DISCUSSION 3.1. Characterization of Monodispersed SWCNT Thin Films. First, we characterized the properties of monodispersed m- and s-SWCNT networks in order to exclude the Schottky junction effect between metallic and semiconducting tubes. Figure 2a,b shows the SEM images of m- and s-SWCNT thin films deposited on a PC membrane, respectively. Their corresponding photographs in each inset display the typical green color of the m-SWCNT thin films and the brown color of the s-SWCNT thin films, respectively. The microscopic structure, especially the surface morphology, appears similar for both types of SWCNT films. Long, narrow, and interwoven SWCNT bundles forming uniform, pure, impurity free networks with a high density of intertube junctions are clearly visible. As-deposited SWCNT films were characterized by Raman spectroscopy to identify the diameter and possible (n,m) chiralities of tubes present in the networks. Figure 2c shows the typical Raman spectra of m- and s-SWCNT thin films. The wide shape of the G-mode in the spectrum of m-SWCNTs is a characteristic feature of m-SWCNTs and is related to the 26455



Article

Figure 2. Representative SEM images of high-purity (a) m-SWCNT and (b) s-SWCNT thin film, respectively. Insets show their photographs after filtration, (c) Raman spectra of the m- and s-SWCNT thin films, (d) the RBM spectral region for m-SWNCT thin film, and (e) for s-SWCNT thin film, identifying possible SWCNT chiralities present in the thin films under investigated.

Bright−Wigner−Fano resonance.24 The high G:D ratio visible for both m- and s-SWCNT thin films indicates the presence of high purity of SWCNT films with a low concentration of defects. Analysis of the radial breathing vibrations (RBM), performed with the use of the three different laser excitation energies, results in the determination of the diameter and SWCNT chiralities which are presented in Figure 2d,e. The range of diameters, 1.31−1.59 nm, for m-SWCNTs, is associated with nanotube chiralities (8,8), (9,9), (12,12), (14,8), and (16,1), while that of s-SWCNT with chiralities (13,6), (14,3), (14,9), (15,7), and (17,6) has been observed to have diameters between 1.39 and 1.59 nm. To prove the purity of the homogeneous electronic type we measured the SWCNT absorption of the purchased dispersions in the UV−vis−NIR range. The corresponding spectra are presented in Figure 3a. The two absorption peaks which are clearly separated with a deep plateau are attributed to the electronic transitions between the van Hove singularities for semiconducting tubes, S22 (900−1270 nm) and S33 (450−630 nm); such a spectrum is characteristic for monodispersed semiconducting samples.25,26 Also, the high intensity peak, M11 (600−850 nm), without any side bands in the spectrum of metallic tubes indicates high purity metallic samples. On the basis of these results we measured the UV−vis−NIR absorption spectra of SWCNT mixtures with varying m:s ratios. The systematic changes in the spectra exhibiting two isosbestic points clearly demonstrate the presence of two chemical species showing absorption proportional to their concentration (see Figure 3b). As a consequence, the ratio of integrated area underneath the M11 and the S22 absorption bands10,27 also reflects the m:s ratio in the mixed SWCNT samples. 3.2. Junction TEP of Monodispersed As-Deposited SWCNT Thin Films. The TEP values expressed by the Seebeck coefficient (S) and electrical conductivity of thin films formed by separate m- and s-SWCNT dispersions are presented in Figure 4. The electrical conductivities of our SWCNT films were measured about 270S/cm for s-SWCNTs and 300S/cm for m-SWCNTs, producing a conductivity ratio σs/σm = 0.9

Figure 3. (a) UV−vis−NIR absorbance of separate m- and s-SWCNT dispersions and (b) that of m- and s-SWCNT mixtures with varying m:s ratios.

(detailed discussions will be presented in section 3.3). A high TEP was measured for s-SWCNTs (88 μV/K) in contrast with a relatively low TEP for m-SWCNT thin films (13 μV/K). Hence, the TEP for m-SWCNT thin films is comparable with that of pure graphite (10 μV/K).28 Data in the literature reveal that the measured TEP of an individual semiconducting SWCNT is ≈42 μV/K.29 While the TEP of the m-SWCNT network is similar to that of graphite, there is a big discrepancy between the TEP of an individual s-SWCNT and that of a sSWCNT network measured here. Several effects can affect the TEP of the SWCNT network, like SWCNT synthesis procedure, enrichment process, film deposition technique, 26456



Article

St =

Rt =

1 (∑ R sSs + R t i=1

∑ R jiSji) i=1

∑ R s + ∑ R ji i=1

i=1

(1)

(2)

where Rs and Ss are the thermal resistance and the TEP of a sSWCNT, Rji and Sji are that of ith junction, respectively, and Rt is the total thermal resistance of the conduction path in series as displayed in Figure 5. Furthermore, similar to Ohm’s law, the temperature difference (ΔTd) is proportional to the thermal resistance (R), which can be given by

Figure 4. Results of the TEP and electrical conductivity measurements of the as deposited m- and s-SWCNT thin films.

film postprocessing leading to physisorption of insulating molecules,9 chemical doping,23,30 or chemisorption with covalent bonding.31 Thermal conductivity measurements on SWCNT networks provide the evidence that intertube junctions in the SWCNT networks are responsible for 97− 99% of the total thermal resistance.32 In this study, we demonstrate that the high value of TEP measured for the sSWCNT films prepared by vacuum filtration of the as-supplied pure s-SWCNT dispersions after successful removal of surfactant residues can be explained by a contribution from the physical intertube junctions. To understand thermoelectric transport through SWCNT thin films, we developed a simplified TEP model assuming that a typical monodispersed SWCNT thin film is composed of numerous individual or bundled nanotubes with junctions between them which are connected in series. Both the tubes and the junctions can contribute to the total TEP. In order to visualize the real situation of s-SWCNT random network, a numerical simulation was employed. The details of this methodology are explained elsewhere.33 A representative schematic of a network randomly generated by the numerical simulation is depicted in Figure 5. In the simulation the density

Q̇ =

ΔTd R

(3)

where Q̇ is the thermal current induced by the heat transport. In fact, both phonons and charge carriers contribute to the heat transport. In our study, a random SWCNT network was regarded as a heterogeneous structure consisting of ropes, bundles, and individual tubes connected by physical junctions. Such networks have been demonstrated to have the potential to be used as thermoelectrics as a result of the inverse relationship between the Seebeck coefficient and electrical conductivity due to charge carrier concentration and mobility.8 Besides, the heterogeneous structure of SWCNT networks is of particular interest, because it allows for the slight decoupling of the competing thermoelectric factors. One of the possible origins of TEP in carbon nanotube networks is the phonon scattering at the boundaries resulting in a lower thermal conductivity due to blocked heat flow across the junctions by comparison with that of crystalline nanotube ropes.35 For these reasons, we consider the charge carrier contribution to the thermal conductivity of SWCNT networks to be dominant, although some papers have indicated that phonons appear to dominate the thermal transport.32,35−37 With a combination of eqs 1−3, St can be summarized as St = Ss +

1 ΔT

n

∑ ΔTji(Sj − Ss) i=1

(4)

where ΔT and ΔTji are the temperature differences of the entire SWCNT network and ith junction, respectively. To quantify the junction contribution to the total TEP in detail, we established a specific cell in a SWCNT network which consists of half of two adjacent individual nanotubes with a junction as illustrated in the top of Figure 6. The temperature difference of the cell is ΔTsc, and η = ΔTj/ΔTsc describes the coefficient related to phonon scattering and electronic tunneling at each junction. By assuming that numerous identical specific cells (N) are connected in series, as in the bottom of Figure 6, the temperature difference of the network can be written as

Figure 5. Representative random network generated by a numerical simulation (top) and one of dominant paths in s-SWCNT network for TEP calculation (bottom).

of tubes is 20/μm2, the length of a tube is 1 μm (according to the measurement), and the film area is defined as 2 μm (width) × 10 μm (length). We assume for the sake of simplicity that the chirality (n,m), diameter, and tube length are identical. The TEP generated by an individual s-SWCNT is denoted as Ss, whereas the TEP arising from a junction is described by Sj. According to Kaiser,34 the TEP of a heterostructured material composed of elements connected in series is weighted by the thermal resistance of each component. Thus, by considering the tubes and junctions separately, the total TEP of the thin film (St) can be expressed as

ΔT = ηΔTscN + (1 − η)ΔTsc(N + 1)

(5)

The first term of eq 5 represents the sum of temperature drops at each junction, and the second term represents the sum of temperature drops along nanotubes in a single path in the network. An additional parameter α was introduced to describe the effect of junctions; it describes a contribution due to the temperature difference arising from all junctions to the total temperature gradient in the network. 26457



Article

is worth noticing that the TEP for one junction was found to be ≈1 mV/K; however, this value has no physical meaning in this study since the value of η is likely to be close to 1 (as shown further). The temperature drop across a single junction associated with phonon scattering and the electronic tunneling effect is determined by the interfacial thermal resistance between nanotubes.38−40 As reported by Zhong et al.,37 the heat transfer across the interface can be expressed as

ΔTj =

1 ΔTt

N

∑ ΔTji = i=1

ηΔTscN ηΔTscN + (1 − η)ΔTsc(N + 1) (6)

After insertion of eq 6 into eq 4, the total TEP of the thin films can be expressed as St = Ss + α(Sj − Ss)

(8)

where Ri is the thermal interfacial resistance, A is the area of the interface, and q is the heat flow rate across the interface. The contact area A has to be extremely small because of the tubular structure of nanotubes. Also, the distortion of the nanotube lattice at the contact point results in a dramatic increase in phonon scattering. Hence, the heat transport through lattice vibrations along the tube is severely blocked at the junction. In addition, the high intertube barriers localize the majority of carriers at the junction causing a significant decrease in electronic heat transport. Both of these effects acting simultaneously increase the thermal interfacial resistance. If these two main effects are acting together, we conclude that the temperature drop across the interface ΔTj is dominant compared with that across a specific cell ΔTsc; thus, the value of η is likely to be ≈1. Taking into account the numerous junctions in the dense SWCNT networks, the junction TEP in s-SWCNTs was estimated to be 90−120 μV/K, while the junction TEP of m-SWCNTs networks was about 13.3−15 μV/ K. In the literature, there is a strong interest in TEP with much research on a variety of nanomaterials due to favorable carrier scattering mechanisms. The research analyzing the effect of junctions on the total TEP demonstrated here has a general character and is valid for any heterostructured system not only for carbon nanotubes and their composites. 3.3. TEP of Defined Ratios between m- and s-SWCNTs. The SWCNT networks can be described as highly conductive quantum wires with tunneling barriers between individual wires.10 In the mixed type of SWCNTs three types of junctions composed of m-m, s-s, and m-s SWCNTs exist. Tunneling ohmic contacts form between tube junctions of the same electronic type (e.g., m-m and s-s), while tunneling Schottky contacts form between m-s junctions, as evidenced by a junction resistance which is expected to be orders of magnitude higher than that found at m-m and s-s junctions.41 In networks of mixed m- and s-SWCNTs numerous percolation paths are possible for thermal transport. However, mobile charge carriers always move via the least resistive percolation paths.33 For this reason, we can assume that constructed conduction paths in thin films with mixed m- and s-SWCNTs can be preferentially made through percolation paths in parallel composed of m-m or s-s SWCNTs, resulting in the fraction of these percolation paths being similar to the m:s ratio in the mixture. Thus, the TEP in parallel connection mode can be calculated from the following equations:

Figure 6. Definition of a specific cell (top), and a schematic of the specific cell in one ideal path (bottom), with the whole film composed of such numerous identical cells. Thus, the total TEP is obtained from the contribution of TEP from each cell.

α=

R iq A

(7)

In principle, the TEP as an intrinsic property of a bulk material is independent of material dimensions. Therefore, St and Ss must be constant in the ideal case when individual nanotubes and junctions have an identical TEP. On the basis of eq 7, the TEP of junctions is a function of η and the number of junctions N in the network. As already mentioned, the TEP of our pure s-SWCNT network was ≈88 μV/K, which is twice as high as the TEP of one individual s-SWCNT (≈42 μV/K) at 300 K29 suggesting that intertube junctions make a significant contribution to the measured TEP of s-SWCNT thin films. We analyzed the dependence of the junction TEP, Sj, on the parameter η, and the number of junctions N, which are plotted in Figure 7. In our networks the minimum number of junctions estimated through dividing the length of film by the length of individual s-SWCNT is about 104. It can be clearly seen that the junction TEP does not change with the number of junctions, except for N < 10, and only slightly changes for 0.5 < η < 0.8. It

St =

Figure 7. Calculated junction TEP as a function of the number of junctions (N) and factor η = ΔTj/ΔTsc. 26458

1 σt

N

∑ σjSj j=1

(9) dx.doi.org/10.1021/jp505682f | J. Phys. Chem. C 2014, 118, 26454−26461


Article

Figure 8. (a) Experimental and calculated TEP as a function of the s-SWCNT fraction in the mixed thin films, (b) the fitting curves of the TEP in the mixed SWCNT films for different γ = (σs/σm), (c) an example of the simulated TEP for a SWCNT film with a m:s ratio equal to 1:2 as a function of γ, (d) and the corresponding fitting curves of TEP versus γ for varying fractions of s-SWCNTs in the prepared films. N

σt =

presented in Figure 8b. Significant deviations from the experimental results are clearly visible. These calculations reveal that the difference of electrical conductivity between mand s-SWCNT networks plays an important role in the determination of the TEP of mixed SWCNT thin films. Generally, as-produced SWCNTs consist of a heterogeneous mixture of nanotubes with a ratio of m- and s-SWCNTs being about 1:2. Therefore, we also calculated the TEP of these commonly fabricated SWCNT thin films as a function of γ, and plotted the results in Figure 8c. The calculated values of TEP at this fixed ratio are in the range 35−62 μV/K for γ > 0.25, which is in a good agreement with our previously measured TEP for bucky paper (≈40 μV/K).30 Finally, the TEP values for a series of our artificially mixed SWCNT samples with various electrical conductivity ratio γ were simulated and plotted in Figure 8d. The numerical simulation curves of the TEP show no dependence on γ in the blue marked region 0.25 < γ < 1 for networks consisting mostly of s- or m-SWCNTs, while the TEP of intermediate mixtures is increased as a function of γ. The electrical conductivity ratio γ > 0.25 is in agreement with previously reported data for practically fabricated SWCNT films.33,41 It seems that γ = 0.25 is the lowest conductivity ratio limit of real as-prepared heterogeneous SWCNT networks. Using the experimental and simulated data shown in Figure 8a, the ratio of s- to mSWCNTs in a mixed network of any SWCNT sample can be estimated from the measured TEP. Moreover, such a model can be applied to any nanowire networks for theoretical analysis of junction TEP in thermoelectric materials. Our results highlighting the importance of the junctions in the transport properties of SWCNT films are in good agreement with the recently published data.31

∑ σj j=1

(10)

where σj and Sj are the electrical conductance and TEP of the jth percolated conduction path, and σt is the conductance of the entire thin film. It is known that the electrical conductivity of m- and s-SWCNT networks is dominated by junction barriers rather than the number of delocalized charge carriers.10 Furthermore, transmission probabilities through junctions in s-SWCNTs and m-SWCNTs are comparable, and the electrical conductivity ratio between the s- and m-SWCNT thin films γ = σs/σm is commonly reported to be between 0.25 and 1.41 Actually, the direct measurement of electrical conductivity of our m- and s-SWCNT films in this study yields a ratio σs/σm = 0.9. However, the best fit of the experimental data demonstrating the changes of the TEP with increasing fraction of s-SWCNTs shown in Figure 8a (solid line) has been achieved by using eqs 9 and 10 with the ratio σs/σm = 0.77. One of the possible origins for that deviation is the difficulty to determine precisely the dimensions of the conduction channel (length, width, and thickness) in thin SWCNT films. Especially, the precise determination of thickness of such networks is critical for the accurate estimation of electrical conductivity. Another plausible possibility is the spatial inhomogeneity in the thin film. On the other hand, the TEP determination does not require the sample dimensions, and thus can be very precisely derived from the experimentally measured thermoelectric voltage and temperature difference. Hence, the value of the conductivity ratio σs/σm = 0.77, as the best coincidence with the experimental data, was used in the analysis. For the purpose of comparison, we calculated the TEP of mixed SWCNT networks by taking into account several electrical conductivities ratios from the range 0.25−1 as was suggested in treating SWCNT networks like bucky papers.41 The obtained TEP curves as a function of varying s-SWCNT fractions are 26459



Article

Field Emission Flat Panel Displays. Nanotechnology 2008, 19, 135703/ 1−4. (7) Yu, C. H.; Murali, A.; Choia, K.; Ryu, Y. Air-Stable Fabric Thermoelectric Modules Made of N- and P-Type Carbon Nanotubes. Energy Environ. Sci. 2012, 5, 9481−9486. (8) Hewitt, C. A.; Kaiser, A. B.; Roth, S.; Craps, M.; Czerw, R.; Carroll, D. L. Multilayered Carbon Nanotube/Polymer Composite Based Thermoelectric Fabrics. Nano Lett. 2012, 12, 1307−1310. (9) Piao, M. X.; Na, J. H.; Choi, J. W.; Kim, J. S.; Kennedy, G. P.; Kim, G. T.; Roth, S.; Dettlaff-Weglikowska, U. Increasing the Thermoelectric Power Generated by Composite Films Using Chemically Functionalized Single-Walled Carbon Nanotubes. Carbon 2013, 62, 430−437. (10) Blackburn, J. L.; Barnes, T. M.; Beard, M. C.; Kim, Y. H.; Tenent, R. C.; McDonald, T. J.; To, B.; Coutts, T. J.; Heben, M. J. Transparent Conductive Single-Walled Carbon Nanotube Networks with Precisely Tunable Ratios of Semiconducting and Metallic Nanotubes. ACS Nano 2008, 2, 1266−1274. (11) Skákalová, V.; Kaiser, A. B.; Woo, Y. S.; Roth, S. Electronic Transport in Carbon Nanotubes: From Individual Nanotubes to Thin and Thick Networks. Phys. Rev. B 2006, 74, 085403/1−085403/10. (12) Skákalová, V.; Kaiser, A. B.; Dettlaff-Weglikowska, U.; Hrnčariková, K.; Roth, S. Effect of Chemical Treatment on Electrical Conductivity, Infrared Absorption, and Raman Spectra of SingleWalled Carbon Nanotubes. J. Phys. Chem. B 2005, 109, 7174−7181. (13) Dettlaff-Weglikowska, U.; Skákalová, V.; Jhang, S. H.; Kim, B. H.; Lee, H. J.; Ley, L.; Park, Y. W.; Berber, S.; Tománek, D.; Roth, S. Effect of SOCl2 Treatment on Electrical and Mechanical Properties of Single-Wall Carbon Nanotube Networks. J. Am. Chem. Soc. 2005, 127, 5125−5131. (14) Simoneau, L. P.; Villeneuve, J.; Aguirre, C. M.; Martel, R.; Desjardins, P.; Rochefort, A. Influence of Statistical Distributions on the Electrical Properties of Disordered and Aligned Carbon Nanotube Networks. J. Appl. Phys. 2013, 114, 114312/1−8. (15) Avouris, P. Carbon Nanotube Electronics. Chem. Phys. 2002, 281, 429−445. (16) Barnes, T. M.; Blackburn, J. L.; van de Lagemaat, J.; Coutts, T. J.; Heben, M. J. Reversibility, Dopant Desorption, and Tunneling in the Temperature-Dependent Conductivity of Type-Separated, Conductive Carbon Nanotube Networks. ACS Nano 2008, 2, 1968−1976. (17) Jackson, R. K.; Munro, A.; Nebesny, K.; Armstrong, N.; Graham, S. Evaluation of Transparent Carbon Nanotube Networks of Homogeneous Electronic Type. ACS Nano 2010, 4, 1377−1384. (18) Yanagi, K.; Udoguchi, H.; Sagitani, S.; Oshima, Y.; Takenobu, T.; Kataura, H.; Ishida, T.; Matsuda, K.; Maniwa, Y. Transport Mechanisms in Metallic and Semiconducting Single-Wall Carbon Nanotube Networks. ACS Nano 2010, 4, 4027−4032. (19) Baxendale, M.; Lim, K. G.; Amaratunga, G. A. J. Thermoelectric Power of Aligned and Randomly Oriented Carbon Nanotubes. Phys. Rev. B 2000, 61, 12706/1−12706/4. (20) Hone, J.; Ellwood, I.; Muno, M.; Mizel, A.; Cohen, M. L.; Zettl, A. Thermoelectric Power of Single-Walled Carbon Nanotubes. Phys. Rev. Lett. 1998, 80, 1042−1045. (21) McIntosh, G. C.; Kaiser, A. B. Calculating Thermopower Due to Fluctuation-Assisted Tunnelling with Application to Carbon Nanotube Ropes. Current Appl. Phys. 2001, 1, 145−148. (22) Romero, H. E.; Sumanasekera, G. U.; Mahan, G. D.; Eklund, P. C. Thermoelectric Power of Single-Walled Carbon Nanotube Films. Phys. Rev. B 2002, 65, 205410/1−205410/6. (23) Piao, M. X.; Kim, G. T.; Kennedy, G. P.; Roth, S.; DettlaffWeglikowska, U. Thermoelectric Properties of Single Walled Carbon Nanotube Networks in Polycarbonate Matrix. Phys. Status Solidi B 2013, 250, 1468−1473. (24) Brown, S. D. M.; Jorio, A.; Corio, P. Origin of the Breit-WignerFano Lineshape of the Tangential G-band Feature of Metallic Carbon Nanotubes. Phys. Rev. B 2001, 63, 155414/1−155414/8. (25) Eliseev, A. A.; Yashina, L. V.; Verbitskiy, N. I.; Brzhezinskaya, M. M.; Kharlamova, M. V.; Chernysheva, M. V.; Lukashin, A. V.; Kiselev, N. A.; Kumskov, A. S.; Freitag, B.; et al. Interaction Between Single

4. CONCLUSIONS The TEP of monodispersed m-SWCNT and s-SWCNT thin films was investigated theoretically and experimentally. SWCNT thin films consisting of precisely adjusted ratios of m- and s-SWCNTs showed a quasilinear enhancement of TEP with an increasing fraction of s-SWCNTs in the prepared mixtures. The experimental data were supported by theoretical calculations which involved the ratio of pure m- and sSWCNTs in a network and the ratio of the resulting electrical conductivity (σs/σm ≈ 0.77). A theoretical TEP model was developed to demonstrate the junction effect in SWCNT networks and its contribution to the total TEP of SWCNT thin films. The theoretical results show a good agreement with the experimental data, and indicate that the total TEP is dominated by the TEP of the intertube junctions. Our proposed model is beneficial for estimation of junction TEP in any type of SWCNT network. The relationship between TEP and the fraction of s-SWCNTs in SWCNT films, determined here, allows a simple estimation of the s:m ratio for any mixed SWCNT film by measuring its TEP. Finally, the TEP model developed here demonstrates that the junction effect in SWCNT networks can be applied to any system with percolating nanoparticle/nanotube paths.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: udettlaff[email protected]. *E-mail: [email protected]. Phone: +82 3290 3801. Fax: +82 953 3780. Author Contributions ⊥

These authors contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (Global Frontier Research Program, No.20110031638) and by the second stage of the Brain Korea 21 Plus Project in 2014. The authors also thank Alan Kaiser for his comments and valuable discussion.

■

REFERENCES

(1) van de Lagemaat, J.; Barnes, T. M.; Rumbles, G.; Shanheen, S. E.; Coutts, T. J. Organic Solar Cells with Carbon Nanotubes Replacing In2O3:Sn As the Transparent Electrode. Appl. Phys. Lett. 2006, 88, 233503/1−233503/3. (2) Contreras, M. A.; Barnes, T.; van de Lagemaat, J.; Rumbles, G. Replacement of Transparent Conductive Oxides by Single-Wall Carbon Nanotubes in Cu(In,Ga)Se2-Based Solar Cells. J. Phys. Chem. C 2007, 111, 14045−14048. (3) Javey, A.; Guo, J.; Wang, Q.; Lundstrom, M.; Dai, H. Ballistic Carbon Nanotube Field-Effect Transistors. Nature 2003, 424, 654− 657. (4) Biswas, C.; Lee, S. Y.; Ly, T. H.; Ghosh, A.; Dang, Q. N.; Lee, Y. H. Chemically Doped Random Network Carbon Nanotube p−n Junction Diode for Rectifier. ACS Nano 2011, 5, 9817−9823. (5) Pasquier, A. D.; Unalan, H. E.; Kanwal, A.; Miller, S.; Chhowalla, M. Conducting and Transparent Single-Wall Carbon Nanotube Electrodes for Polymer-Fullerene Solar Cells. Appl. Phys. Lett. 2005, 87, 203511/1−203511/3. (6) Chen, Z.; Zhu, F.; Wei, Y.; Jiang, K. L.; Liu, L.; Fan, S. S. Scanning Focused Laser Activation of Carbon Nanotube Cathodes for 26460



Article

Walled Carbon Nanotube and 1D Crystal in CuX@SWCNT (X = Cl, Br, I) Nanostructures. Carbon 2012, 50, 4021−4039. (26) Li, N.; Ma, Y. F.; Wang, B.; Huang, Y.; Wu, Y. P.; Yang, X.; Chen, Y. S. Synthesis of Semiconducting SWCNTs by Arc Discharge and Their Enhancement of Water Splitting Performance with TiO2 Photocatalyst. Carbon 2011, 49, 5132−5141. (27) Huang, L. P.; Zhang, H. L.; Wu, B.; Liu, Y. Q.; Wei, D. C.; Chen, J. Y.; Xue, Y. Z.; Yu, G.; Kajiura, H.; Li, Y. M. A Generalized Method for Evaluating the Metallic-to-Semiconducting Ratio of Separated Single-Walled Carbon Nanotubes by UV−vis−NIR Characterization. J. Phys. Chem. C 2010, 114, 12095−12098. (28) Uher, C. Thermopower of Exfoliated Graphites Between 1.7 and 300 K. Phys. Rev. B 1982, 25, 4167−4172. (29) Yu, C. H.; Shi, L.; Yao, Z.; Li, D.; Majumdar, A. Thermal Conductance and Thermopower of an Individual Single-Wall Carbon Nanotube. Nano Lett. 2005, 5, 1842−1846. (30) Piao, M. X.; Alam, M. R.; Kim, G. T.; Dettlaff-Weglikowska, U.; Roth, S. Effect of Chemical Treatment on the Thermoelectric Properties of Single Walled Carbon Nanotube Networks. Phys. Status Solidi B 2012, 249, 2353−2356. (31) Tian, X.; Moser, M. L.; Pekker, A.; Sarkar, S.; Ramirez, J.; Bekyarova, E.; Itkis, M. E.; Haddon, R. C. Effect of Atomic Interconnects on Percolation in Single Walled Carbon Nanotube Thin Film Networks. Nano Lett. 2014, 14, 3930−3937. (32) Itkis, M. E.; Borondics, F.; Yu, A.; Haddon, R. C. Thermal Conductivity Measurements of Semitransparent Single-Walled Carbon Nanotube Films by a Bolometric Technique. Nano Lett. 2007, 7, 900− 904. (33) Joo, M. K.; Mouis, M.; Jeon, D. Y.; Kim, G. T.; Kim, U. J.; Ghibaudo, G. Static and Low Frequency Noise Characterization of Ntype Random Network of Carbon Nanotubes Thin Film Transistors. J. Appl. Phys. 2013, 114, 154503/1−154503/10. (34) Kaiser, A. B. Thermoelectric Power and Conductivity of Heterogeneous Conducting Polymers. Phys. Rev. B 1989, 40, 2806− 2813. (35) Huang, H.; Liu, C.; Wu, Y.; Fan, S. Aligned Carbon Nanotube Composite Films for Thermal Management. Adv. Mater. 2005, 17, 1652−1656. (36) Hone, J.; Whitney, M.; Piskoti, C.; Zettl, A. Thermal Conductivity of Single-Walled Carbon Nanotubes. Phys. Rev. B 1999, 59, R2514−R2516. (37) Zhong, H. L.; Lukes, J. R. Interfacial Thermal Resistance Between Carbon Nanotubes: Molecular Dynamics Simulations and Analytical Thermal Modeling. Phys. Rev. B 2006, 74, 125403/1− 125403/10. (38) Dresselhaus, M. S.; Eklund, P. C. Phonons in Carbon Nanotubes. Adv. Phys. 2000, 49, 705−814. (39) Meng, F. Y.; Ogata, S.; Xu, D. S.; Shibutani, Y.; Shi, S. Q. Thermal Conductivity of an Ultrathin Carbon Nanotube with an XShaped Junction. Phys. Rev. B 2007, 75, 205403/1−205403/6. (40) Cummings, A.; Osman, M.; Srivastava, D.; Menon, M. Thermal Conductivity of Y-Junction Carbon Nanotubes. Phys. Rev. B 2004, 70, 115405/1−115405/6. (41) Fuhrer, M. S.; Nygard, J.; Shih, L.; Forero, M.; Yoon, Y. G.; Mazzoni, M. S. C.; Choi, H. J.; Ihm, J.; Louie, S. G.; Zettl, A. Crossed Nanotube Junctions. Science 2000, 288, 494−497.

26461


Effect of Intertube Junctions on the Thermoelectric Power of

Recommend Documents