New Approach for Large-Area Thermoelectric Junctions with a Liquid

Letter Cite This: Nano Lett. XXXX, XXX, XXX−XXX

pubs.acs.org/NanoLett

New Approach for Large-Area Thermoelectric Junctions with a Liquid Eutectic Gallium−Indium Electrode Sohyun Park and Hyo Jae Yoon* Department of Chemistry, Korea University, Seoul 02841, Korea

Nano Lett. Downloaded from pubs.acs.org by UNIV PARIS-SUD on 11/13/18. For personal use only.

S Supporting Information *

ABSTRACT: A challenge in organic thermoelectrics is to relate thermoelectric performance of devices to the chemical and electronic structures of organic components inside them on a molecular scale. To this end, a reliable and reproducible platform relevant to molecularlevel thermoelectric measurements is essentially needed. This paper shows a new, efficient approach for thermoelectric characterization of a large area of molecular monolayers using liquid eutectic gallium− indium (EGaIn). A cone-shaped EGaIn microelectrode permits access to noninvasive, reversible top-contact formation onto organic surfaces in ambient conditions, high yields of working devices (up to 97%), and thus statistically sufficient thermoelectric data sets (∼6000 data per sample in a few hours). We here estimated thermopowers of EGaIn (3.4 ± 0.1 μV/K) and the Ga2O3 layer (3.4 ± 0.2 μV/K) on the EGaIn conical tip and successfully validated our platform with widely studied molecules, oligophenylenethiolates. Our approach will open the door to thermoelectric large-area molecular junctions. KEYWORDS: Molecular thermoelectrics, large-area junction thermopower, EGaIn, self-assembled monolayers, soft top-contact

O

indium alloy (EGaIn)19 covered with a conductive, selfpassivating Ga2O3 layer (nominal thickness of ∼1 nm).20,21 The EGaIn conical tip enabled reversible, noninvasive, and defined thermoelectric top contacts on delicate organic surfaces of organic thin films such as self-assembled monolayers (SAMs) in ambient conditions (Figure 1a). This advantageous feature of the EGaIn top electrode made it possible to reliably and reproducibly collect large amounts of thermoelectric data, in a reasonable amount of time (∼6000 data in a few hours), enough to draw statistically robust

rganic-based thermoelectric materials promise environmentally friendly energy harvest from situations where a temperature difference exists, taking the advantages of lowcost, bendable and stretchable characteristics, processability, and synthetic tailorability.1−3 To realize applications of thermoelectric devices based on organic materials, how chemical and electronic structures of active organic components are related to thermoelectric performance of devices should be elucidated on a molecular scale.4−7 Achieving this goal is, however, difficult because many organic thermoelectric devices contain molecules (and polymers) with complicated structures, their solid-state structures on the surface are uncertain, and interfacial characteristics between molecules and between the molecule and electrode are ill-defined.1,8−11 Thermoelectric characterization of single molecules or monolayers could allow access to an atomic-detailed structure−property relationship in organic thermoelectrics. Recently, efforts have gone into the fundamental understanding of how thermopower is associated with chemical and electronic structures of organic and bioorganic molecules and their interfaces with electrodes.1,6,7,10−18 Despite these elegant and stimulating studies, the field of molecular thermoelectrics is still relatively unexplored as compared to molecular electronics. For molecular-level thermoelectric study, an ideal platform should guarantee convenience and ease in operation and fabrication of device, high yields of working devices, and reliability and reproducibility in measurements. Here we show an efficient platform that satisfies these needs on the basis of a large-area junction architecture. Our approach relies on coneshaped non-Newtonian liquid metal, a eutectic gallium− © XXXX American Chemical Society

Figure 1. (a) Schematic describing our thermoelectric measurement system. (b) Molecules used in this study. (c) Schematic describing the structure of the large-area thermoelectric junction. Received: August 22, 2018 Revised: November 8, 2018 Published: November 12, 2018 A

DOI: 10.1021/acs.nanolett.8b03404 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Segalman and Majumdar research groups.7,12 The measured ΔV/ΔT for the junction was −2.4 ± 0.1 μV/K (Figure 2b). Given that our thermopower analysis of the circuit in Figure S2 indicates the measured ΔV = −(SEGaIn − SW tip)ΔT (where SEGaIn and SW tip are the thermopowers of EGaIn and tungsten tips, respectively; SW tip = 1.0 μV/K as shown in Table S1), the value of SEGaIn was 3.4 ± 0.1 μV/K. We further measured the thermopower of the Ga2O3 layer on the EGaIn conical tip by forming a junction of HOPG//Ga2O3/EGaIn (HOPG: highly oriented pyrolytic graphite), in which a van der Waals (vdW) contact was formed while retaining the Ga2O3 layer (Figure 2a).20 The measured ΔV/ΔT of the HOPG junction was −5.8 ± 0.2 μV/K (Figure 2b); as shown in Figure S3, ΔV = −(Stop electrode − SW tip)ΔT, where Stop electrode = SEGaIn + SGa2O3. Hence, the values of Stop electrode and SGa2O3 were 6.8 ± 0.2 and 3.4 ± 0.2 μV/K, respectively. Consequently, the values of SEGaIn and SGa2O3 were identical, and similar to those of conventional metals (e.g., gold, silver, and copper, ∼1.4−1.8 μV/K, as summarized in Table S1). From a molecular point of view, the surface of the EGaIn conical tip is rough.20 To determine if the roughness of EGaIn conical tip contributes to thermoelectric measurements, we further measured the value of SGa2O3 with the spherical drop of EGaIn on HOPG. The EGaIn drop would have a smoother surface than the conical tip.20 The thermopower of the EGaIn drop//HOPG junction was 5.5 ± 1.4 μV/K, indistinguishable from that of the analogous junction formed with the EGaIn conical tip (Figure 3a). This suggests that the thermopower

inferences on the structure−property relationship. We validated our platform through thermopower measurements of widely studied molecules in the research of thermoelectrics, oligophenylenethiolates (S(Ph)n; n = 1, 2, 3; Figure 1b), incorporated into large-area junctions with the structure AuTS/ S(Ph)n//Ga2O3/EGaIn (Figure 1c; AuTS is the ultraflat template-stripped gold substrate22). Figure 1a illustrates our measurement system, which incorporates five parts: (i) a micromanipulator to fabricate an Ga2O3/EGaIn conical tip and form junctions, (ii) a thermocouple to monitor the in situ temperature change of the bottom electrode, (iii) a nanovoltmeter to measure the thermoelectric voltage (ΔV) across the junction, (iv) a hot chuck to control the temperature of the bottom electrode and create a temperature differential (ΔT) in the junction, and (v) a tungsten (W) tip as a ground electrode. SAMs were formed on AuTS substrates to minimize the degree of structural defects resulting from the roughness of the substrate.22 The fabrication of the EGaIn conical tip has been reported elsewhere.19,23 The conductive, self-passivating Ga2O3 on the surface of EGaIn plays a role as a protective layer in improving the mechanical stability of the electrode and yield of the working junctions, and thus fabrication and operation of junctions can be done in ambient conditions.24 Upon creation of temperature differential, we measured the value of ΔV in microvolts and estimated the Seebeck coefficients (S, μV/K; S = −ΔV/ΔT). The Supporting Information contains detailed experimental procedures. In a system of thermoelectric measurement, thermopower of all the components inside it has to be probed.7 Hence, we measured the thermopower of EGaIn by forming a shorting junction on bare AuTS (Figure 2a). Upon shorting, direct

Figure 3. (a) Seebeck coefficients for HOPG//Ga2O3/EGaIn junctions of the EGaIn drop and the conical tip. (b) Plot of the thermoelectric voltage as a function of geometrical contact area in the AuTS/S(Ph)2//Ga2O3/EGaIn junction at ΔT = 5 K.

measurement seems not significantly sensitive to the shape change of EGaIn, or the variation of the Seebeck coefficient by the different shapes of EGaIn is likely dominated by the variation arising from tip-to-tip roughness (i.e., contact area) variation for each tip. Using the HOPG junction, we also examined if thermoelectric measurement depends on the geometrical contact area of the EGaIn conical tip. We formed separate junctions having various contact areas using EGaIn conical tips of various sizes. As shown in Figure 3b, the value of ΔV did not correlate with the contact area, which implies that the measurement of thermoelectric voltage through large areas using the EGaIn conical tip is not significantly affected by the contact area. These findings are reasonable given that thermopower is the intrinsic property of the molecule and hence insensitive to the number of molecules in the junction. In a typical experiment, we kept the geometrical contact areas

Figure 2. (a) Schematic describing the junction structures of the EGaIn conical tip with bare AuTS and HOPG (highly ordered pyrolytic graphite). (b) Plots of thermoelectric voltage (ΔV, μV) as a function of temperature differential (ΔV, K) for the AuTS and HOPG junctions.

contact of EGaIn with gold occurs. To estimate the thermopower of EGaIn (and others such as the Ga2O3 layer and SAM; see below), we performed thermoelectric analysis of the circuit corresponding to the junction as described in the Supporting Information (Figures S2−S4). Briefly, the thermopowers of all the parts in the circuit and the temperature profile across it were considered to confirm what the measured output voltage corresponded to, following the stimulating work by B

DOI: 10.1021/acs.nanolett.8b03404 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters Table 1. Summary of Data of Thermoelectric Junction Measurements n in S(Ph)n

ΔT (K)

no. of samples

no. of tips

no. of junctions

counts

n=1

5 9 13 5 9 13 5 9 13

5 5 3 3 4 4 4 4 4

17 14 8 9 11 12 12 11 13

43 55 51 55 68 77 73 71 59

3744 4869 3227 4960 6697 6415 6533 6644 5403

n=2

n=3

less than 3.0 × 103 μm2 to avoid short junctions caused by too large contact areas. The EGaIn-based thermoelectric platform was validated with oligophenylenethiolate SAMs. We obtained statistically significant values of ΔV for the AuTS/S(Ph)n//Ga2O3/EGaIn large-area junctions. For data acquisition, we introduced the following standard protocol. We collected ∼100 data points per junction (∼50 data points for the SAM of short molecule, SPh) and at least 10 junctions in different places per sample at a certain temperature differential (ΔT). An EGaIn conical tip was used for measuring 3−10 junctions, and thereafter a new tip was formed to minimize the complexity arising from the contribution of the surface contamination of the tip. While this protocol led to statistical data reflecting any type of possible variations such as trace-to-trace, junction-to-junction, tip-totip, and sample-to-sample, it avoided overweighting one or some of them. Table 1 summarizes the data of thermoelectric junction measurements. The yields of working junctions were 51−81% for the short molecule, SPh, and 90−97% for the long molecules, S(Ph)2 and S(Ph)3. It has been established that the benzenethiolate SAM is usually more defective than oligophenylenethiolate SAMs due to the weak lateral interaction;25,26 thus, the low yield in the short molecule could be attributed to the defective structure in the SAM. In Figure 4a, we found independence of our data from the possible variations. Indeed, the histograms of ΔV were fit to single Gaussian curves (Figure 4b), from which the mean value (ΔVmean) and standard deviation (σΔV) were extracted. As shown in Table 1, all the data exhibited indistinguishable mean and median values, indicative of the statistical significance of our data. The value of σΔV increased with the increasing value of ΔV and the length of molecule; similar trends have been observed previously and attributed to the increased degree of conformational change (e.g., variations in ring−ring torsion angle and internal vibrations and rotations) at high temperature and for long molecules and to the corresponding variation in the electronic structure.7,27 Figure 4c shows the plots of ΔVmean against the temperature differential (ΔT) where error bars correspond to ±σΔV. As indicated above, we performed the circuit analysis over the SAM-based junction in Figure S4 and found the value of ΔV was expressed as follows: ΔV = −(SSAM − S Wtip)ΔT

ΔVmean ± σΔV −25 −55 −78 −41 −77 −111 −91 −132 −192

± ± ± ± ± ± ± ± ±

7 14 14 10 19 25 30 30 54

ΔVmedian

yield (%)

−23 −55 −79 −39 −76 −111 −91 −133 −195

81 69 51 97 95 93 96 92 90

Figure 4. (a) Exemplary thermoelectric data of covering trace-totrace, junction-to-junction, electrode-to-electrode, and sample-tosample variations. (b) Exemplary histograms of ΔV for S(Ph)3 SAM. (c) Plot of ΔVmean as a function of ΔT. (d) Plot of SSAM as a function of length of molecule (n in S(Ph)n).

results measured in single-molecule7 and small-area (101−102 molecules) junctions.10 Figure 4d shows the plot of SSAM as a function of molecular length (n in S(Ph)n). Both the Seebeck coefficients linearly increased with the increasing number of phenylene moieties, consistent with previous experimental and theoretical results.7,10,29,30 Quek et al. have shown that the linear trend of the Seebeck coefficient for oligophenylenes could be analyzed by a simplified equation (eq 2),29 derived from the transmission function through a junction and the Landauer formulism.4 SSAM = SC + n·β S

(1)

(2)

Here, n is the length of molecule (the number of phenylene unit in our work; βS is the rate of change of thermopower with n; and SC is the thermopower of hypothetical junction where n = 0 (i.e., nonshorting junction without SAM), reflecting the thermopower of nonmolecular (here, nonphenylene) components in the junction. The slope (βS) and y-intercept (SC) in the plot of Figure 4d were 2.1 ± 0.3 and 5.6 ± 0.5, respectively,

With this equation, we estimated the Seebeck coefficients (SSAM) of S(Ph)n SAMs: 7.8 ± 0.4, 9.8 ± 0.2, and 12.9 ± 1.5 μV/K for n = 1, 2, 3, respectively. The positive polarity of the SSAM values suggests that the closest molecular orbital to the Fermi level of Ga2O3/EGaIn (−4.3 eV)28 was the highest occupied molecular orbital (HOMO). The magnitude and polarity of the SSAM values were consistent with the previous C

DOI: 10.1021/acs.nanolett.8b03404 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters consistent with the literature values.7,29,31 Provided that SSAM = S(Ph)n + SS−Au (thermopower of the gold−thiolate moiety), the value of SC corresponds to the value of SS−Au (5.6 ± 0.5 μV/ K). Different interfaces have exhibited different SC values: single-molecule junctions possessing diamine and dithiolate interfaces over gold electrodes have shown the SC values of 0.4 ± 0.2 and 6.4 ± 0.5 μV/K, respectively.31 The difference between these and our value confirms that SC depends on the interfacial characteristics of junctions. In summary, we have demonstrated an efficient thermoelectric platform harnessing large-area junctions of the EGaIn microelectrode. By measuring the thermopowers of EGaIn and the Ga2O3 layer in the EGaIn tip, we have revealed that the native oxide of Ga2O3 in the EGaIn conical tip does not influence the thermoelectric junction measurements, and our approach allows access to statistically sufficient data sets in high yields (up to 97%). To date, two types of techniques are limitedly available for thermoelectric research at the molecular level: (i) electrode-molecule-electrode junctions on singlemolecule7,12,15,17,32 and small-area10,16 scales, and (ii) noncontact optical measurements of heat transfer rates from electrode to molecules.5,6,11,14,18 Table S3 summarizes these. All the previous approaches rely on special instruments such as scanning probe microscopies and laser spectroscopy; no or little statistical information is available for some cases. Our work has shown, for the first time, large-area thermoelectric molecular junctions, which adds to the repertoire of molecular thermoelectrics.

■

(4) Paulsson, M.; Datta, S. Phys. Rev. B: Condens. Matter Mater. Phys. 2003, 67, 241403. (5) Ge, Z.; Cahill, D. G.; Braun, P. V. Phys. Rev. Lett. 2006, 96, 186101. (6) Wang, R. Y. Appl. Phys. Lett. 2006, 89, 173113. (7) Reddy, P.; Jang, S.-Y.; Segalman, R. A.; Majumdar, A. Science 2007, 315, 1568−1571. (8) Dresselhaus, M. S.; Chen, G.; Tang, M. Y.; Yang, R.; Lee, H.; Wang, D.; Ren, Z.; Fleurial, J.-P.; Gogna, P. Adv. Mater. 2007, 19, 1043−1053. (9) Bubnova, O.; Crispin, X. Energy Environ. Sci. 2012, 5, 9345− 9362. (10) Tan, A.; Balachandran, J.; Sadat, S.; Gavini, V.; Dunietz, B. D.; Jang, S.-Y.; Reddy, P. J. Am. Chem. Soc. 2011, 133, 8838−8841. (11) Losego, M. D.; Grady, M. E.; Sottos, N. R.; Cahill, D. G.; Braun, P. V. Nat. Mater. 2012, 11, 502−506. (12) Yee, S. K.; Malen, J. A.; Majumdar, A.; Segalman, R. A. Nano Lett. 2011, 11, 4089−4094. (13) Li, Y.; Xiang, L.; Palma, J. L.; Asai, Y.; Tao, N. Nat. Commun. 2016, 7, 11294. (14) Wang, Z.; Carter, J. A.; Lagutchev, A.; Koh, Y. K.; Seong, N.-H.; Cahill, D. G.; Dlott, D. D. Science 2007, 317, 787−790. (15) Baheti, K.; Malen, J. A.; Doak, P.; Reddy, P.; Jang, S.-Y.; Tilley, T. D.; Majumdar, A.; Segalman, R. A. Nano Lett. 2008, 8, 715−719. (16) Tan, A.; Balachandran, J.; Dunietz, B. D.; Jang, S.-Y.; Gavini, V.; Reddy, P. Appl. Phys. Lett. 2012, 101, 243107. (17) Kim, Y.; Jeong, W.; Kim, K.; Lee, W.; Reddy, P. Nat. Nanotechnol. 2014, 9, 881−885. (18) Majumdar, S.; Malen, J. A.; McGaughey, A. J. H. Nano Lett. 2017, 17, 220−227. (19) Chiechi, R. C.; Weiss, E. A.; Dickey, M. D.; Whitesides, G. M. Angew. Chem., Int. Ed. 2007, 47, 142−144. (20) Simeone, F. C.; Yoon, H. J.; Thuo, M. M.; Barber, J. R.; Smith, B.; Whitesides, G. M. J. Am. Chem. Soc. 2013, 135, 18131−18144. (21) Cademartiri, L.; Thuo, M. M.; Nijhuis, C. A.; Reus, W. F.; Tricard, S.; Barber, J. R.; Sodhi, R. N. S.; Brodersen, P.; Kim, C.; Chiechi, R. C.; Whitesides, G. M. J. Phys. Chem. C 2012, 116, 10848− 10860. (22) Weiss, E. A.; Kaufman, G. K.; Kriebel, J. K.; Li, Z.; Schalek, R.; Whitesides, G. M. Langmuir 2007, 23, 9686−9694. (23) Yoon, H. J.; Bowers, C. M.; Baghbanzadeh, M.; Whitesides, G. M. J. Am. Chem. Soc. 2014, 136, 16−19. (24) Barber, J. R.; Yoon, H. J.; Bowers, C. M.; Thuo, M. M.; Breiten, B.; Gooding, D. M.; Whitesides, G. M. Chem. Mater. 2014, 26, 3938− 3947. (25) Frey, S.; Stadler, V.; Heister, K.; Eck, W.; Zharnikov, M.; Grunze, M.; Zeysing, B.; Terfort, A. Langmuir 2001, 17, 2408−2415. (26) Ishida, T.; Mizutani, W.; Azehara, H.; Miyake, K.; Aya, Y.; Sasaki, S.; Tokumoto, H. Surf. Sci. 2002, 514, 187−193. (27) Malen, J. A.; Doak, P.; Baheti, K.; Tilley, T. D.; Majumdar, A.; Segalman, R. A. Nano Lett. 2009, 9, 3406−3412. (28) Yoon, H. J.; Liao, K.-C.; Lockett, M. R.; Kwok, S. W.; Baghbanzadeh, M.; Whitesides, G. M. J. Am. Chem. Soc. 2014, 136, 17155−17162. (29) Quek, S. Y.; Choi, H. J.; Louie, S. G.; Neaton, J. B. ACS Nano 2011, 5, 551−557. (30) Ke, S.-H.; Yang, W.; Curtarolo, S.; Baranger, H. U. Nano Lett. 2009, 9, 1011−1014. (31) Malen, J. A.; Doak, P.; Baheti, K.; Tilley, T. D.; Segalman, R. A.; Majumdar, A. Nano Lett. 2009, 9, 1164−1169. (32) Widawsky, J. R.; Darancet, P.; Neaton, J. B.; Venkataraman, L. Nano Lett. 2012, 12, 354−358.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.8b03404. Detailed experimental procedures, discussion of junction measurements, discussion of thermoelectric techniques and temperature gradients in junctions, figures (histograms of thermoelectric voltage, thermopower analysis, graphs of thermoelectric voltage vs number of junctions), and tables (Seebeck coefficients, thermal conductivities, molecular level thermoelectric measurement platforms from the literature) (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*H. J. Yoon. E-mail: [email protected]. ORCID

Hyo Jae Yoon: 0000-0002-2501-0251 Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This research was supported by the National Research Foundation of Korea (NRF-2017M3A7B8064518). REFERENCES

(1) Cui, L.; Miao, R.; Jiang, C.; Meyhofer, E.; Reddy, P. J. Chem. Phys. 2017, 146, 092201. (2) Zhang, Q.; Sun, Y.; Xu, W.; Zhu, D. Adv. Mater. 2014, 26, 6829− 6851. (3) Russ, B.; Glaudell, A.; Urban, J. J.; Chabinyc, M. L.; Segalman, R. A. Nat. Rev. Mater. 2016, 1, 16050. D

DOI: 10.1021/acs.nanolett.8b03404 Nano Lett. XXXX, XXX, XXX−XXX