Exploring Peltier effect in organic thermoelectric films

Organic materials are emerging thermoelectric candidates for flexible power generation and solid-cooling applications. Although the Peltier effect is a fundamental thermoelectric effect that enables site-specific and on-demand cooling applications, the Peltier effect in organic thermoelectric films have not been investigated. Here we experimentally observed and quasi-quantitatively evaluated the Peltier effect in a poly(Ni-ett) film through the fabrication of thermally suspended devices combined with an infrared imaging technique. The experimental and simulation results confirm effective extraction of the Peltier effect and verify the Thomson relations in organic materials. More importantly, the working device based on poly(Ni-ett) film yields maximum temperature differences as large as 41 K at the two contacts and a cooling of 0.2 K even under heat-insulated condition. This exploration of the Peltier effect in organic thermoelectric films predicts that organic materials hold the ultimate potential to enable flexible solid-cooling applications.

To obtain the temperature distribution of a poly(Ni-ett)-based film, we measured the relationship between T IR (measured by IR camera) and T PT100 (measured by a PT100 temperature sensor). All influence factors caused by the camera, IR window, and the properties of film were included in the calibration parameter. The poly(Ni-ett) film for temperature calibration was utilized without further treatment. The calibration was as follows: (1) Set the default parameters of the IR camera.
(2) Measure the temperature of poly(Ni-ett) in steady-state using the IR camera and PT100 temperature sensor.
(3) Fit T IR and T PT100 to obtain the working curve. (4) Calibrate the temperature of the tested device pixel by pixel.
The calibration was performed in the same apparatus and testing environment (6×10 -4 Pa). All data were fitted by a concatenate fitting mode because of the favorable repeatability of the five tested devices. After calibration, the change in temperature induced by the biased current was measured using the FLIR X6530sc directly. The result are comparable to those obtained using the PT100 temperature sensor.

Supplementary Notes 2 | Finite element simulation of lateral thin-film device
To testify the transient temperature distribution of the device, we simulated the transient work condition of the device using thermoelectric module of the COMSOL Multiphysics software. The simulation results based on the model and boundary conditions are described below.
The poly(Ni-ett) was set as a cuboid (2.3 mm×1.5 mm×2.15 μm). Two Au electrodes (1500 μm×150 μm× 95 nm) were placed above the conducting material on two sides of the sample. All transient simulation results below are based on this device geometry. The model is schematically illustrated in Figure S9a. In our simulation system, the Au/poly(Ni-ett) interface was assumed to have idea thermal contacts. Interfacial thermal conductance between the device and the atmosphere was omitted because the test was performed under 6×10 -4 Pa. The initial temperature condition was set as constant (T 0 = 298.15 K). All surfaces of the model are radiating active. In simulations, various currents were applied to the left gold electrode, whereas another gold electrode was grounded.
All physical properties of gold were based on built-in parameters. The temperature-dependent Seebeck coefficient and electrical conductivity of poly(Ni-ett) film were measured by SB100 and Keithley 4200SC, respectively. The heat specific heat capacity was measured using differential scanning calorimeter (TA Q2000 with compacted block sample). All relevant parameters of the materials are summarized in Supplementary Table 1

Supplementary Notes 3 | Lock-in IR measurement
Lock-in thermography can overcome the limitations of temperature resolution associated with the IR camera by long-term data acquisition and data processing. Here, we used this technique to distinguish the Peltier effect and Joule heating of poly(Ni-ett) film in small temperature modulation.
The temperature modulation of the OTE device was caused by the combined influence of the Peltier effect and Joule heating. When an a.c. current was applied to the device, the temperature modulations caused by these two effects are the first and second harmonic signals, respectively. Therefore, the extracted amplitudes of first and second harmonic represented the temperature changes caused by the Peltier effect and Joule heating, which can be used to calculate the Peltier coefficient.
In our experiments, two synchronous cosine currents were simultaneously applied to the device and IR camera. The frequency of the applied current was selected to limit the thermal diffusion length of the heat wave1. The detected temperature of the device was almost unchanged during the experiment duo to the very low driving a.c. current. After data processing, the first and second harmonic signal was both substantial. The extracted first harmonic signal represented the Peltier temperature modulation and was limited near the two electrode/poly(Ni-ett) interfaces, consistent with the fact that the Peltier effect is an interface effect. In addition, the phase difference of these two signals is 180°, indicating that one electrode/poly(Ni-ett) contact is cooled while another one is heated. The extracted second harmonic signal is uniform in the entire device, implying that the Joule heating is uniform in the entire device without any phase difference 1,2 .

Supplementary Notes 4 | Finite element modeling simulation of ultrathin vertical device
We simulated a single leg ultrathin device to predict the Peltier cooling ability of poly(Ni-ett). The cross-section of the device was set as a square (1 ×1 mm 2 ) and the thickness of the organic material was varied from 2 to 14 µm. All TE parameters of the poly(Ni-ett) (the maximum performance in literature) 3  In our model, all thermal contacts and electrical contacts were assumed to be ideal.

are listed in Supplementary
The upper ceramic functioned as cooling surface and the bottom one worked as heat sink with maintained temperature of 300 K. In addition, all interfaces were regarded as radiating active. In the simulation, the ΔT and heat transport capacity were evaluated by applying various current densities (10−100 Amm -2 ) to the upper gold electrode, whereas the bottom electrode was grounded. The simulated cold side temperature and temperature difference between two ceramic slices based on this thermoelectric material is shown in Figure S13b and S13c.
For the cooling device, the optimal currents ( opt = c ⁄ ) for the maximum ∆ (∆ max = 2 2 ⁄ ), varies depending on TE performance of the device 4 . The ∆ increases with increasing current and reaches the maximum ∆ at opt . However, a further increase in the current degrades the ∆ because of rapidly increased Joule heating.
Transported heat flux, an important parameter to characterize the performance of Peltier cooling device, is given by where l is the thickness of the thermoelectric material, S the Seebeck coefficient, σ is the electrical conductivity, and κ is the thermal conductivity. Theoretically, it is favorable for a thin-film OTE cooling device to have a high heat flux because of its intrinsically low thermal conductivity. Figure S14d shows

Supplementary Figures
Supplementary Figure 1  Thermal processes in a conventional TE device with lateral device geometry. Peltier effect, Joule-heating, internal heat transfer within TE film, interlayer heat conduction to the substrate, heat convection to the air, and thermal radiation occurs concurrently in the device. Notably, heat dissipation occurs "vertically" into the substrates (interlayer heat conduction) and the air (heat