Series Photothermoelectric Coupling Between Two Composite Materials for a Freely Attachable Broadband Imaging Sheet

Abstract

As flexible wearable sensors and imagers are receiving attention from diverse social sectors, the freely attachable photothermoelectric (PTE) conversion technique should be evaluated to develop a highly usable safety sensor network. Although carbon nanotube (CNT)‐related materials should be effective, the key parameters/structures that maximize PTE conversion have not been clarified, thus hindering optimum device design and practical use. Herein, the flexible, sensitive broadband photodetection operation based on a coupling configuration between the CNT film photo/heat/electron channel and metal electrode is evaluated. Experimental PTE measurements and steady‐state thermal distribution simulations reveal that a series coupling of a p‐type CNT film channel and a highly negative Seebeck coefficient counter metal electrode facilitate superior photodetection performances than those of a parallel coupling configuration. Furthermore, subsequent device designs provide sensitive broadband photodetection from the millimeter‐wave to visible light wavelength regions with a minimum noise equivalent power of 5 pWHz^−1/2 in an uncooled nonvacuum condition. Simultaneously, the mechanical flexibility of the proposed photodetector allows for its use in freely attachable sheet imager applications on curvilinear objects, and the nondestructive 3D photomonitoring of a defective intricately bent sample is demonstrated.

Introduction

With the emergence of the Internet of Things (IoT) society, flexible and freely attachable photothermoelectric (PTE) sensing and visualization applications are increasingly becoming a major topic in the material and device investigation fields. To make flexible PTE sensing/imaging devices practically available, namely, comprehensively assuring the safety/quality of the social components via multiview broadband uncooled photomonitoring, diverse materials have been studied, including graphene,^[] carbon nanotubes (CNTs),^[] MoS2,^[] SnSe,^[] NbS3,^[] and PEDOT:PSS.^[] Among these candidates, randomly stacked CNT films demonstrate remarkable properties. A large Seebeck coefficient over 200 μV K⁻¹ has been reported in enriched semiconducting CNT films,^[] and a larger value of 2500 μV K⁻¹ has been well theorized in individual semiconducting CNTs with small diameters.^[] Simultaneously, CNT film-based materials exhibit excellent broadband photoabsorption characteristics,^[] flexibility,^[] mechanical strength,^[] and corrosion resistance,^[] potentially leading to promising PTE photomonitoring applications. Indeed, we previously performed nondestructive, omnidirectional imaging inspections utilizing wearable, wide-frequency-band photodetectors with single-walled CNT (SWCNT) film photo/heat/electron channels.^[]

Contrary to significant potential and interest, the CNT film PTE photodetectors require further sensitivity improvement, which is attributed to their lower noise equivalent power (NEP) values^[] than those of the state-of-the-art solid-state detectors.^[] Thus, the underlying PTE conversion mechanism in CNT film photodetectors should be clarified to provide effective approaches for such aims. In particular, a deeper understanding of PTE conversion dynamics with a CNT film channel–metal electrode coupling structure is important for further photodetection sensitivity enhancement. While metal electrodes are essential to maintain photo-induced PTE responses (i.e., the PTE effect at the CNT film–metal electrode interface),^[] their use is thought to affect the PTE properties of CNT film–metal coupling^[] and hinder efficient PTE conversions. Although the earlier issue has not been resolved, namely, devising functional PTE coupling, PTE property degradation should be reduced at the CNT film–metal coupling, and subsequent photodetector designs should contain substantial sensitivity improvements.

This study develops a device that collectively satisfies sensitive broadband uncooled photodetection and free-form 3D photomonitoring, derived from a unique CNT film channel–metal electrode PTE coupling configuration, where each Seebeck coefficient can be fully utilized. A numerical analysis of the photodetection interface reveals that an interrelationship between the photo-induced heat diffusion direction and CNT film–metal coupling configuration governs the PTE effect in the CNT film photodetectors. Experimental photodetection measurement further specifies the significance of the CNT film–metal series coupling compared with a conventional parallel configuration, and the series configuration of a p-type CNT film channel and the highly negative Seebeck coefficient Bi counter electrode maximize the PTE properties. Subsequent device design following series PTE coupling, including thermal management, conformal coating, and miniaturization processing, provides sensitive broadband photodetection ranging from the millimeter-wave (MMW) to the visible light wavelength regions with a minimum uncooled nonvacuum NEP of 5 pWHz^−1/2. The obtained index represents a tenfold sensitivity enhancement over the CNT film PTE sensor with optimal parallel coupling configuration and is comparable with those of cutting-edge solid-state uncooled nonvacuum broadband PTE sensors.^[] Concurrently, the mechanical flexibility of the proposed PTE sensor sheet allows for the development of freely attachable device patches on curvilinear target objects, thereby demonstrating nondestructive photomonitoring of a defective intricately bent multilayered target. The fundamental PTE device design and conformable 3D photomonitoring application could be used to provide a promising target structure-independent social sensing platform, where PTE materials, flexible electronics, thermoelectric (TE) conversion, and visualization techniques are utilized.

PTE coupling between the CNT films and metal electrodes was evaluated to explore the key physical parameter governing the PTE effect and potentially maximizes the device functionality of the CNT film photodetector. Figure S1a,b, Supporting Information, shows the PTE effect at the CNT film–metal electrode interface, in which the photoinduced PTE response can be obtained^[] based on the following equation.^[]1$V=|S_{\text{CNT}}-S_{\text{Com}}|\times \mathrm{\Delta }T$where V, S_CNT, S_Com, and ΔT are the photoinduced PTE DC voltage response, Seebeck coefficient of the CNT film, that of the composite material between the CNT film and metal electrode, and photoinduced thermal gradient, respectively. Here, the following equation further describes S_Com.^[]2$S_{\text{Com}}=\frac{{\sigma }_{\text{CNT}}{t}_{\text{CNT}}{S}_{\text{CNT}}+{\sigma }_{\text{Metal}}{t}_{\text{Metal}}{S}_{\text{Metal}}}{{\sigma }_{\text{CNT}}{t}_{\text{CNT}}+{\sigma }_{\text{Metal}}{t}_{\text{Metal}}}$where σ_CNT, σ_Metal, t_CNT, t_Metal, and S_Metal are the electrical conductivities, thicknesses, and Seebeck coefficient of the CNT film and metal electrode, respectively. Equation () indicates that S_Com is significantly influenced by the ratio between t_CNT and t_Metal when the metal electrode is parallel to the CNT film plane along the photoinduced heat diffusion direction in the channel. Simultaneously, Equation () infers that S_Com approximates to S_CNT (Figure S1c, Supporting Information) when the CNT film is thicker than the metal electrode (typically, t_CNT and t_Metal range from 5 to 100 μm and 20 to 200 nm, respectively). With Equation (), the above situation, namely the parallel CNT film–metal electrode configuration, is thought to cause attenuation of the photoinduced PTE responses.

To avoid such photoresponse degradation, we proposed a different type of PTE configuration. The CNT film and metal electrode are serially coupled along the photoinduced heat diffusion direction (Figure ). The proposed structure advantageously allows full utilization of the fundamental values of S_CNT and S_Metal, thereby enhancing the photodetection PTE response. Here, each series/parallel metal electrode coupling structure was designed in the same CNT film channel (Figure S2, Supporting Information). Figure  shows the corresponding photoinduced PTE responses under light irradiation on each interface. The result indicates a photoresponse enhancement with the series PTE coupling by a factor of 4.7 compared with that with the parallel configuration, in good agreement with the theoretically estimated value (5 times, Table S1, Supporting Information, S_CNT: 62 μV K⁻¹, S_Com: 49.9 μV K⁻¹, S_Au: 1.6 μV K⁻¹). A reversed polarity of the photoresponses was confirmed in each coupling interface, in accordance with typical PTE effect behavior (Figure S3, Supporting Information, V_Series ≈ (S_CNT – S_Au) × ΔT, V_Parallel ≈ (S_Com – S_CNT) × ΔT). This verification further enabled us to properly select the types of series metal electrodes for original p-type CNT film channels in terms of the Seebeck coefficient. Figure  and Table S1, Supporting Information show the effective Seebeck coefficients S_Eff (S_Eff = S_CNT –  S_Metal) and photoresponses under light irradiation for different types of series electrodes (Au, Al, Ni, and Bi). The use of negative S_Metal series electrodes, such as Ni and Bi, brought further photoresponse enhancement (up to two times with Bi) compared with that of a positive S_Metal series electrode (Au). Thus, the presented effective use of S_CNT and S_Metal revealed substantial enhancement of S_Eff and incidental photoresponses in the series PTE coupling between the p-type CNT film channel and highly negative Seebeck coefficient Bi electrode.

[IMAGE OMITTED. SEE PDF]

The proposed CNT film–Bi series PTE coupling brings further opportunity for the subsequent photodetector device design. In particular, several multifaceted approaches, including fundamental thermal management, conformal coating for stable operations, and miniaturization processing, could potentially provide the emergence of promising photodetectors. Here, we first optimized the series Bi electrode length for thermal management of the present CNT film photodetector. While the NEP represents the photodetection sensitivity (NEP $\propto \sqrt{\mathrm{R}}$/ΔT, V_Sens $\propto$ ΔT, R: device electrical resistance, V_Sens: normalized photoresponse, see the Experimental Section “Noise Equivalent Power”), controlling the series electrode length is thought to govern the NEP of the CNT film PTE sensor based on its ΔT and R. Figure  shows the photodetection performance as a function of the series Bi electrode length. The results, similar to the simulated photoinduced thermal distribution analysis (see Figure S4, Supporting Information, and the Experimental Section “Steady-State Thermal Distribution Simulation”), demonstrated that the optimum series Bi electrode length is ≈3.5 mm, where the NEP change indicates the peak value. The optimal series Bi electrode length of 3.5 mm is of the same order of magnitude as the heat diffusion length in Bi of 1.3 mm; therefore, the observed experimental photodetection behavior (Figure ) can be determined by ΔT in the series electrode (see the Experimental Section “Theoretical Calculation of the Heat Diffusion Length in Bi”). As shown in Figure , when the series Bi electrode length is shorter than the photoinduced heat diffusion length, the PTE response is reduced due to the insufficient ΔT. The shorter electrode length condition is thought to bring about a reduction in thermal resistance, resulting in suppression of the thermal gradient. This concurrently leads to NEP degradation due to the inverse proportionality between the NEP and V_Sens, as described in the Experimental Section “Noise Equivalent Power.” In addition, when the series Bi electrode length is longer than the photoinduced heat diffusion, NEP degradation occurs due to an excessive increment in R. The incidental thermal noise voltage deterioration and its mutual proportionality to the NEP could be fatal for this longer-series electrode length situation (see Figure S5, Supporting Information, and the Experimental Section “Noise Equivalent Power”). Thus, the thermal management of the CNT film–Bi series PTE coupling, including the channel length optimization shown in Figure S6, Supporting Information, is important in the comprehensive device design for highly functional photoimager utilization.

[IMAGE OMITTED. SEE PDF]

Meanwhile, a conformal coating of the CNT film–Bi series PTE coupling should be developed to overcome the latent problems of the series Bi electrode: easy oxidation property^[] and weak adhesion to the channel due to the chemical modification instability of the SWCNT surface.^[] To resolve these issues and ensure photodetector durability, we devised a lamination series PTE coupling structure, as shown in Figure . The combined use of UV laser processing and selective-suction CNT film filtration allowed for bottom-up channel forming of the series electrode (see  “Selective Suction CNT Film Filtration”), enabling a Bi buffer layer (Au or Al)–CNT film sandwich structure design. Figure  shows an aging characteristic of R with and without the 20 nm-thick metal coating on the series Bi electrode of the CNT film PTE sensors. The proposed laminated-series PTE coupling structure demonstrated R suppression and aging stabilization, whereas the photodetector with a naked-series Bi electrode exhibited a higher R and aging degradation characteristics. Thus, the thin Au/Al coating on the series Bi electrode provided oxidation protection and adhesion to the CNT film channels. Furthermore, the aforementioned suction bottom-up channel formation simultaneously triggered physisorption at the series PTE coupling (930 and 2200 Ω for R with and without [Bi deposition on the CNT film interface] the suction bottom-up series PTE coupling). These efforts are essential to minimize the thermal noise, thereby making the long-term, low-noise, flexible, broadband photodetector operation feasible. Regarding the above conformal coating, the proposed suction bottom-up channel formation, where the size/position of the channels can be freely controlled, could provide high-density multipixel integration and further NEP improvement via channel width shortening (Figure S7, Supporting Information). The device design using the CNT film channel–Bi electrode series PTE coupling successfully provided an opportunity for flexible, sensitive, long-term, stably operable, broadband visualization tools.

[IMAGE OMITTED. SEE PDF]

The CNT film PTE sensor schematically shown in Figure  represents incorporation of all findings and technologies obtained. Here, we assess device performances, such as the photodetection sensitivity, spatial resolution, available frequency regions, and imaging applications, of the proposed CNT film flexible broadband PTE sensor. Figure  shows the I–V characteristic acquired by the proposed photodetector with and without light irradiation, and the minimum NEP value under an uncooled nonvacuum condition reached 5.06 pWHz^−1/2. The thermal device design of the CNT film–metal parallel coupling^[] and PN junction formation via chemical carrier doping^[] were previously evaluated for sensitive operations. The obtained index here equals tenfold and sixfold sensitivity enhancement over these two types of photodetectors, respectively. Simultaneously, as shown in Figure S8, Supporting Information, the proposed CNT film photodetector exhibits 1.7-times and 8.7-times spatial resolution improvement compared with that of the aforementioned CNT film PTE sensors. Therefore, the CNT film–metal series PTE coupling and its subsequent multifaceted device design provided the best functionality among the CNT film-based flexible broadband photodetectors. In addition, imaging applications (Figure : far IR [FIR], S9: MMW/sub-terahertz [sub-THz]) indicate that the proposed device sufficiently serves as a broadband photoimager and could potentially be used for arbitrary hierarchical image extraction of multilayered 3D curvilinear composite structures via multifrequency band sensing.

[IMAGE OMITTED. SEE PDF]

Quantifying the contributions of photodetection sensitivity enhancement could facilitate further CNT film PTE sensor design. We evaluated six approaches (CNT film–metal series PTE coupling, Bi as the series electrode material, length optimization of the series Bi electrode, suction bottom-up channel formation of the series electrode, conformal coating of the series Bi electrode, and channel width miniaturization). The photodetection sensitivity was enhanced totally by a factor of 110 via the above efforts, as shown in Figure S10, Supporting Information. Here, the corresponding NEP values before/after each step (560, 110, 53, 48, 27.2 , 17, and 5.06 pWHz^−1/2) provided the following effectiveness: 34%, 14%, 7%, 12%, 11%, and 22%, respectively. Thus, the series PTE coupling between the CNT film channel and highly negative Seebeck coefficient counter metal electrode, which provides nearly 50% of the total effectiveness, was dominant in the proposed substantial photodetection sensitivity enhancement. Indeed, the sensitivity limit is calculated as 45 fWHz^−1/2 by designing superior-series PTE coupling: semiconducting CNT films and Bi₂Te₃ and referring their ZT values (ZT_{Semi-CNT film}  = 0.33, ZT_{Mixed-CNT film} = 0.03, ZT_Bi2Te3  = 1, and ZT_Bi  = 0.1).^[] In addition, as represented by the experimentally obtained minimum NEP, our CNT film PTE sensor exhibits the photodetection sensitivity comparable with those of cutting-edge uncooled nonvacuum photodetectors^[] in the broadband frequency region ranging from MMW to visible light.

While the presented efforts potentially provide a diverse range of photomonitoring usages (such as the temperature-independent operation, as shown in Figure S12, Supporting Information), mechanical flexibility (Figure S13, Supporting Information), an advantageous functionality of the proposed CNT film PTE sensor, facilitated unique free-form 3D photomonitoring applications. Here, the use of a series PTE coupling-based 26 pixel 5 μm-thick imager sheet, firmly patched on the rear side of a complicatedly bent resin object, facilitated nondestructive 3D photomonitoring (Figure S14, Supporting Information, and -e). The measurement was carried out under full-face FIR radiation on the target sample, and the concealed defect (tine breakages and a metallic tape) on the high curvature surface was visualized by detecting local changes in the transmission signals. Therefore, the proposed freely attachable, thin-film, multipixel photoimager sheet, which is smoothly used for high-speed 2D flat photoscanning (Figure S15, Supporting Information), can be used to monitor curvilinear 3D objects. Here, the presented free-form photomonitoring on arbitrary-structured 3D targets is thought to bridge the gap between electromagnetic-wave sensing/imaging techniques and flexible electronics, thereby enabling incidental, compact, omnidirectional, high-speed sensitive broadband visualization applications.

The presented findings and technologies can potentially be used for highly functional active sensing/imaging applications and further developments in blackbody radiation (BBR)-based passive photomonitoring techniques, simultaneously. Although the current range of BBR-based applications^[] is limited, due to a weaker intensity than that of external photo sources,^[] passive photomonitoring techniques have received significant attention for their unique advantageous characteristics: photo-sourceless compact setup and wideband-available frequency regions (MMW–IR). In particular, the use of human/animal bodies as BBR sources could contribute to noninvasive passive healthcare photomonitoring. Here, the practical healthcare verification could potentially be promoted by the demonstrated functionality (i.e., sensitive, broadband, and flexible operation) of the proposed photodetector, which sufficiently governs the BBR low-loss freely attachable operation on living sources. The series PTE coupling-based CNT film photodetector could be used as a wearable, passive, noninvasive, telemedical photomonitoring module, satisfying the rising demand for redressing regional healthcare disparities.

Concurrently, the further enrichment of CNT film-based PTE coupling could be a breakthrough in improving photodetection performances as for sensitivity and spatial resolution. The recent development of the TE generator module design has crossed the boundary of flexible electronics,^[] and the printing of the high-density pixel integration of TE materials superior to Bi (e.g., Bi₂Te₃) on flexible substrates with a scale of 84 PN pairs/10 × 10 mm^2[] represents unique flexible TE conversion techniques. Thus, the built-in implementation of the CNT thin-film photoabsorbent channel pixels on steric π-shaped Bi₂Te₃ PN structures of flexible frameworks could lead to the emergence of freely attachable, ultrahighly sensitive, high-resolution, broadband 2D matrix cameras.

In conclusion, this paper proposed a multifaceted PTE device design based on a coupling configuration between CNT film channels and metal electrodes, providing freely attachable, sensitive, broadband photomonitoring. The series coupling configuration between the p-type CNT film and highly negative Seebeck coefficient Bi counter electrode maximized the effective Seebeck coefficient and overcame PTE property degradation posed by parallel PTE coupling. Subsequent efforts regarding operation noise suppression, higher spatial resolution, aging/temperature/strain stabilization, and microfabrication allowed the comfortable photodetector to collectively satisfy the mechanical flexibility and uncooled broadband photodetection with comparable sensitivity to that of existing solid-state devices. This concept synergizes PTE materials, flexible electronics, TE conversions, and visualization measurements. The demonstrated multispectral sensing-based simple target material analysis and target structure-independent, free-form, multipixel, 3D surface imaging technique potentially provide a roadmap for ubiquitous social sensing platforms.

Experimental Section

Series PTE Coupling

Inch-scale, 5- μm-thick semiconducting/metallic mixed SWCNT films (ZEON Co.) were used as the PTE channels of photodetectors shown in Figure , , S3b–S6c,d, S10a,b, Supporting Information (“Original parallel”). As shown in Figure S1a, the CNT films were cut into μm–mm-scale rectangular pieces using a cutter or scissors. To form the CNT film–metal series PTE coupling configuration, the oblique metal deposition was conducted utilizing a resistance heating vacuum evaporator (Al, AU, Bi/SVC700TMSG, Sanyu Electron Co.) and an electron beam vacuum evaporator (Ni/EBH-6s, ULVAC Inc.). The tilt angle of the sample holder inside the chamber was kept at 45°. Metal depositions were conducted under the vacuum condition of 5 × 10⁻⁵ Pa for Al, Bi, and Ni and 5 × 10⁻⁴ Pa for Au. For each metal deposition, the deposition rate was set at 2.0 $\stackrel{\cdot }{\mathrm{A}}$s⁻¹. The thicknesses of series electrodes and thin Au/Al coating layers (Figure ) were, respectively, 200 and 20 nm (Figure S17, Supporting Information). Figure S1a and S2, Supporting Information, shows the detailed conditions of the entire device fabrication process.

Selective-Suction CNT Film Filtration

Selective-suction CNT film filtration was used to form the bottom-up series PTE coupling configuration for the photodetectors shown in Figure , , S7b,c, S8, S9, S10a,b, Supporting Information (“Optimized series”), S12b-d, S13b,c, S14, S15b-d, Supporting Information. A semiconducting/metallic mixed-type SWCNT solution (Zeon Co.) was dripped on a membrane filter (70 μm-thick, 200 nm pore, C020A025A, ADVANTEC Ltd.), which was covered with a polyimide mask (5 μm-thick, Kapton, DU PONT-TORAY Co.) and then vacuumed (Vacuum pump: MVP015, Pfeiffer Vacuum Technology AG). The polyimide mask was processed by a UV laser (wavelength λ = 355 nm, LWL-3030-T06, SIGMAKOKI Co.), and the minimum processing resolution was 10 μm. The typical CNT film channel thickness was 1 μm.

Current–Voltage Measurement

In measuring the DC current–voltage characteristic of the photodetector, we utilized a current amplifier (1211, TOYO Co.) and a programmable voltage source (3245 A, Hewlett Packard Co.). The readout signal was recorded on a digital multimeter (34410 A, KEYSIGHT TECHNOLOGIES Inc.) and controlled by a LabVIEW program on a PC via a GPIB cable. The measurement resolution of the digital multimeter was 100 nV, and the CNT film PTE sensor and the digital multimeter were directly connected without passing the readout signals through an amplifier for measuring the photoinduced PTE DC voltage responses.

Measurement of the Seebeck Coefficient

To measure the Seebeck coefficient of CNT films, CNT film–metal parallel composites, and metal electrodes, we used a microceramic heater (MS-M1000, SAKAGUCHI E.H. VOC Co.), K-type thermocouples (T-35 K, SAKAGUCHI E.H. VOC Co), and the aforementioned digital multimeter (Figure S18, Supporting Information). One end of the target was locally heated (+5 °C) with the microceramic heater, and the voltage response was recorded on the digital multimeter via the K-type thermocouple probed at both ends of the sample. Seebeck coefficients of the target samples were calculated using the Seebeck coefficient of each alumel terminal and chromel terminal of the K-type thermocouple (S_Alumel  = −18 μV K⁻¹, S_Chromel  = 22 μV K⁻¹). We carried out the Seebeck coefficient measurement in a dark room under atmospheric pressure and room temperature.

Photo sources

In this study, we used eight types of photo sources: frequency multipliers in the MMW/sub-THz/THz bands (100 GHz, 4 mmφ collimated irradiation, Signal Generator Extension WR1.0/260 GHz, 10 mmφ collimated irradiation, Amplifier Multiplier Chain 573/300 GHz▪1 THz switchable, 10 mm/4 mmφ collimated irradiation, Transmitter Tx261, Virginia Diodes, Inc.), a gas laser in the FIR band (29 THz, 20 mmφ collimated irradiation/1 mmφ fiber irradiation, CO₂ Laser L4, Access Laser Co.), a radiator (Broadband FIR, 600 W, Carbon Heater DCT-J066, YAMAZEN Co.), a quantum cascade laser in the MIR band (39 THz, 1 mmφ focus lens irradiation, QCL L12007-1294 H-C, Hamamatsu Photonics K.K.), LEDs in the NIR band (λ = 870 nm, 5 mmφ collimated irradiation, L12170, Hamamatsu Photonics K.K.), and the visible light band (λ = 568 nm, 2 mmφ collimated irradiation, L-934SGD, Kingbright Electronic Co.). The light irradiation from these photosources was guided to and concentrated in the photodetection interface by lenses or fibers.

Noise Equivalent Power

The photodetection sensitivity is defined by the NEP. Here, the NEP can be expressed by the following equation^[]$\text{NEP}=\frac{V_{\text{Noise}}}{V_{\text{Sens}}}=\frac{\sqrt{4k_{\mathrm{B}}TR}}{S_{\text{Eff}}\times \mathrm{\Delta }T}\times P_{\text{Eff}}$3$\propto \frac{\sqrt{R}}{\mathrm{\Delta }T}$where V_Noise, V_Sens, k_B, T, R, S_Eff, ΔT, and P_Eff, are noise voltage spectral density, normalized photoinduced PTE response, Boltzmann constant, absolute temperature, electrical resistance, effective Seebeck coefficient, photoinduced thermal gradient, and effective output power of light irradiation on the photodetection interface. As shown in Figure S5, Supporting Information, the zero-bias voltage operation of the CNT film photodetector suppressed the shot noise and 1/f noise. Thereby, the noise voltage of our CNT film PTE sensor was approximated to the thermal noise ($\sqrt{4k_{\mathrm{B}}TR}$).

Steady-State Thermal Distribution Simulation

In designing the CNT film photodetector (series Bi electrode length optimization (Figure ), channel width shortening (Figure S7b,c, Supporting Information), we utilized ANSYS software as steady-state thermal analysis to investigate the heat distribution and temperature gradient in CNT film photodetectors during light irradiation (Figure S4, Supporting Information). By setting the thermophysical property values of CNT film channels and the metal electrodes, the following heat conduction equation can be solved using the finite element method.4$\rho C\frac{\partial \mathrm{T}}{\partial t}=k(\frac{{\partial }^2}{\partial x^2}+\frac{{\partial }^2}{\partial y^2}+\frac{{\partial }^2}{\partial z^2})T+Q$

Here, ρ, C, k, T, and Q are the density, heat capacity, thermal conductivity, absolute temperature, and total heat flow, respectively.^[] The thermophysical property values of the CNT film were k: 10 W mK^−1[] and emissivity ε: 0.9 (calculated from Kirchhoff's law of thermal radiation: ε = absorptivity α). Here, we set the outside temperature to a constant, 295 K.

Photoinduced PTE Response Mapping

A digital stepping motor (MOTORIZED STAGE, SIGMAKOKI Co.) was used to conduct photoresponse mapping, as shown in Figure S6c-d, S8, S14, Supporting Information. The CNT film photodetector was connected to the stepping motor and the aforementioned digital multimeter, and the photoresponse at each point was recorded on the digital multimeter while controlling the position of the CNT film PTE sensor at every 100 μm.

Single-pixel XY Scan Imaging

The single-pixel CNT film photodetector was connected to the aforementioned digital multimeter and the target object was attached on a two-axis digital stepping motor stage to carry out XY-axis 2D planer imaging measurements, as shown in Figure , S9, S10b, Supporting Information. The photoresponse at each point was recorded on the digital multimeter while controlling the XY position of imaging objects at every 100 μm.

Multipixel Array Scan Imaging

A multiplexer data logger (34980 A-34923 A T⁻¹, KEYSIGHT TECHNOLOGIES Inc.) was utilized to read out multipixel photoresponses and carry out one-axis array scan imaging measurements, as shown in Figure , S15b-d, Supporting Information. Here, the multipixel CNT film PTE image sensor array sheet was connected to the multiplexer and the target object was attached on an aforementioned one-axis digital stepping motor stage. Up to eight terminal blocks were mounted on one multiplexer data logger, and there can be up to 80 elements per terminal block. The fastest readout speed was 500 channels s⁻¹. However, there was a trade-off relationship with the readout resolution. In this study, the readout speed, stepping motor scan speed, and resolution were, respectively, set at 50 channels s⁻¹, 1 mm s⁻¹, and 100 nV. For the multipixel array scan imaging, photoresponses were recorded on the multiplexer while controlling the one-axis position of target objects at every 100 μm. By conducting the multipixel array scan imaging, the acquisition time of 2D images was over ten times shorter than that conducted utilizing a single pixel of the CNT film PTE sensor with the same spatial resolution.

Heat Resistance Evaluation of the CNT Film–Bi Electrode Series Coupling Configuration

To investigate the heat resistance of the proposed photodetector, a digital programmable hotplate (1-7565-01 Digital Hot Plate To 430 °C, HP-1SA, AS ONE Co.) was used. PTE signals of the photodetector induced by the external FIR light irradiation were continuously recorded on the multiplexer data logger during annealing. An FIR laser fiber was mounted on the digital stepping motor and suspended above the photodetection interface of the device. To avoid heat damage, the fiber was kept away from the hotplate except when taking the photoresponse measurements.

Bending Resistance Evaluation of the CNT Film–Bi Electrode Series Coupling Configuration

The proposed photodetector was connected with the aforementioned multiplexer data logger and suspended between two digital stepping pushing stages, which faced each other. PTE signals of the photodetector induced by the external FIR light irradiation were continuously recorded on the multiplexer. Here, the rear side of the photodetection interface was mounted on a thin supporting column, so that the optical path length could be maintained at a constant value throughout the device folding.

3D Image Reconstruction Via a Freely Attachable Multipixel PTE Image Sensor Array Sheet

In Figure , the target was designed via 3D CAD software (AUTODESK TINKER CAD) and fabricated using a 3D printer (Value3D Magix MF-2200D, MUTOH INDUSTRIES LTD). The proposed freely attachable thin-film 26 pixel CNT film PTE image sensor array sheet was patched on a supporting resin substrate, which was designed to firmly follow the target structure, and was fabricated using the same 3D printer (Figure S16a, Supporting Information). Both the target and monitoring module were mounted on the aforementioned digital stepping motor stage via screw holes (Figure S16b,c, Supporting Information), which were simultaneously designed on resin bodies. As shown in Figure S16a,Supporting Information, the used components were aligned with each other, and then the target was scanned every 100 μm under external full-face FIR radiation. Here, the 3D structure in the XYZ space (X: scanning direction, Y: pixel array direction, Z: light irradiation direction) was reconstructed based on the CAD data. Then, a 4D diagram was coordinated by referring to the obtained transmission PTE responses at each pixel. The differences in the transmission signals of each pixel due to their position shift along the Z-direction were calibrated using the photoresponse intensity ratios (Figure S16e, Supporting Information), which were acquired in advance. The minimum processing accuracy was 50 μm in the XY axis and 100 μm in the Z-axis. Polylactic acid and acrylonitrile–butadiene–styrene were used for the printing resins.

Theoretical Calculation of the Heat Diffusion Length in Bi

By substituting the reported thermal diffusivity of a 230 nm-thick Bi film at 300 K^[] into Equation (), the theoretical value of the heat diffusion length in Bi at a measurement frequency of 1 Hz is calculated.5$L=\sqrt{\frac{D}{\pi f}}$

Here, L, D, and f are the heat diffusion length, thermal diffusivity, and measurement frequency, respectively.^[]

Acknowledgements

The authors thank ZEON Corporation for providing CNT films. This work was supported in part by the Mirai Program and Center of Innovation Program from the Japan Science and Technology Agency, the Toray Science Foundation, and JSPS KAKENHI grant numbers JP17H02730, JP18H03766, JP19K22099, JP19H02199, and JP19H04539 from the Japan Society for the Promotion of Science.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

K.L. contributed to the device design, simulations, device fabrication, PTE measurements, and PTE nondestructive inspection demonstrations. D.S. contributed to the analysis of the measurement results and the application of the proposed CNT film devices. K.L. wrote the paper. Y.K. conceived the study, participated in its coordination, and assisted in writing the manuscript. All authors made significant contributions to the experiments, analysis of the data, and writing of the manuscript.

References

D. Zhang, Y. Song, L. Ping, S. Xu, D. Yang, Y. Wang, Y. Yang, Nano Res. 2019, 12, 2982.

X. He, N. Fujimura, J. M. Lloyd, K. J. Erickson, A. A. Talin, Q. Zhang, W. Gao, Q. Jiang, Y. Kawano, R. H. Hauge, F. Léonard, J. Kono, Nano Lett. 2014, 14, 3953.

K. J. Erickson, X. He, A. A. Talin, B. Mills, R. H. Hauge, T. Iguchi, N. Fujimura, Y. Kawano, J. Kono, F. Léonard, ACS Nano 2015, 9, 11618.

M. He, Nano Energy 2018, 49, 588.

Y. Zhong, L. Zhang, V. Linseis, B. Qin, W. Chen, L. D. Zhao, H. Zhu, Nano Energy 2020, 72, 104742.

W. Wu, Y. Wang, Y. Niu, P. Wang, M. Chen, J. Sun, N. Wang, D. Wu, Z. Zhao, ACS Appl. Mater. Interfaces 2020, 12, 14165.

M. Zhang, J. T. Yeow, Carbon 2020, 156, 339.

B. A. MacLeod, N. J. Stanton, I. E. Gould, D. Wesenberg, R. Ihly, Z. R. Owczarczyk, K. H. Hurst, C. S. Fewox, C. N. Folmar, K. H. Hughes, B. L. Zink, J. L. Blackburn, A. J. Ferguson, Energy Environ. Sci. 2017, 10, 2168.

Y. Nonoguchi, A. Tanaka, C. Goto, T. Kitano, T. Kawai, Sci. Technol. Adv. Mater. 2018, 19, 581.

Y. Nonoguchi, A. Tani, T. Kitano, T. Kawai, J. Appl. Phys. 2019, 126, 135108.

A. D. Avery, B. H. Zhou, J. Lee, E. S. Lee, E. M. Miller, R. Ihly, D. Wesenberg, K. S. Mistry, S. L. Guillot, B. L. Zink, Y. H. Kim, J. L. Blackburn, A. J. Ferguson, Nat. Energy 2016, 1, 16033.

S. Kivistö, T. Hakulinen, A. Kaskela, B. Aitchison, D. P. Brown, A. G. Nasibulin, E. I. Kauppinen, A. Härkönen, O. G. Okhotnikov, Opt. Exp. 2009, 17, 2358.

S. Park, M. Vosguerichian, Z. A. Bao, Nanoscale 2013, 5, 1727.

A. Cao, P. L. Dickrell, W. G. Sawyer, M. N. Ghasemi-Nejhad, P. M. Ajayan, Science 2005, 310, 1307.

P. Hammer, F. C. D. Santos, B. M. Cerrutti, S. H. Pulcinelli, C. V. Santilli, Prog. Org. Caot. 2013, 76, 601.

D. Suzuki, Y. Kawano, Carbon 2020, 162, 13.

D. Suzuki, S. Oda, Y. Kawano, Nat. Photonics 2016, 10, 809.

D. Suzuki, Y. Ochiai, Y. Kawano, ACS Omega 2018, 3, 3540.

D. Suzuki, Y. Ochiai, Y. Nakagawa, Y. Kuwahara, T. Saito, Y. Kawano, ACS Appl. Nano Mater. 2018, 1, 2469.

H. H. Yang, G. M. Rebeiz, IEEE Trans.Microw. Theory Tech. 2016, 64, 2129.

V. M. Nascimento, L. Méchin, V. Pierron, F. Starecki, C. Adamo, D. G. Schlom, B. Guillet, 44th Int. Conf. on Infrared, Millimeter, and Terahertz Waves, IEEE, Paris 2019, p. 1.

Y. Zhang, Y. Watanabe, S. Hosono, N. Nagai, K. Hirakawa, Appl. Phys. Lett. 2016, 108, 163503.

Z. Liu, Z. Liang, X. Zheng, Y. Jiang, 44th Int. Conf. on Infrared, Millimeter, and Terahertz Waves, IEEE, Paris 2019, p. 1.

B. C. St-Antoine, D. Ménard, R. Martel, Nano Res. 2012, 5, 73.

L. Leontie, M. Caraman, M. Alexe, C. Harnagea, Surf. Sci. 2002, 507–510, 480.

N. Karousis, N. Tagmatarchis, D. Tasis, Chem. Rev. 2010, 110, 5366.

Y. Nakai, K. Honda, K. Yanagi, H. Kataura, T. Kato, T. Yamamoto, Y. Maniwa, Appl. Phys. Lett. 2014, 7, 025103.

H. J. Goldsmid, Materials 2014, 7, 2577.

J. S. Son, K. Park, M. Han, C. Kang, S. G. Park, J. H. Kim, W. Kim, S. Kim, T. Hyeon, Angew. Chem. Int. Ed. 2011, 50, 1363.

J. Zhou, M. A. R. Miah, Y. Yu, A. C. Zhang, Z. Zheng, S. Damle, I. A. Niaz, Y. Zhang, Y. H. Lo, Opt. Exp. 2019, 27, 37056.

Q. Gao, A. Pospischil, M. Bhuiyan, H. Jiang, H. Talin, D. Farmer, B. Deng, C. Li, S. Han, H. Wang, Q. Xia, T. Ma, T. Mueller, F. Xia, Nano Lett. 2016, 16, 4648.

J. B. Wang, W. L. Li, B. Chu, C. S. Lee, Z. S. Su, G. Zhang, S. H. Wu, F. Yan, Org. Electron. 2011, 12, 34.

D. Y. Kim, K. O. Kenneth Electron. Lett. 2017, 53, 732.

R. Huang, X. Ji, Y. Liao, J. Peng, K. Wang, Y. Xu, F. Yan, Opt. Exp. 2019, 27, 23250.

M. Bauer, A. Rämer, S. A. Chevtchenko, K. Y. Osipov, D. Čibiraitė, S. Pralgauskaitė, K. Ikamas, A. Lisauskas, W. Heinrich, V. Krozer, H. G. Roskos, IEEE Trans. Terahertz Sci. Technol. 2019, 9, 430.

M. W. Ryu, J. S. Lee, K. S. Kim, K. Park, J. Yang, S. Han, K. R. Kim, IEEE Tarns. Electron Devices 2016, 63, 1742.

G. Auton, D. B. But, J. Zhang, E. Hill, D. Coquillat, C. Consejo, P. Nouvel, W. Knap, L. Varani, F. Teppe, J. Torres, A. Song, Nano Lett. 2017, 17, 7015.

X. Yang, A. Vorobiev, A. Generalov, M. A. Andersson, J. A. Stake, Appl. Phys. Lett. 2017, 111, 021102.

X. Zhang, W. Jin, L. Li, X. Wang, J. Chen, Y. Jia, Appl. Opt. 2018, 57, 3991.

Y. Meng, A. Qing, C. Lin, J. Zang, Y. Zhao, C. Zhang, Sci. Rep. 2018, 8, 7852.

Y. Cheng, Y. Wang, Y. Niu, Z. Zhao, Opt. Exp. 2020, 28, 6350.

R. Bogue, Sens. Rev. 2009, 29, 6.

S. J. Kim, J. H. We, B. J. Chou, Energy Environ. Sci. 2014, 7, 1959.

Y. Wang, L. Yang, X. Shi, L. Chen, M. S. Dargusch, J. Zou, Z. Chen, Adv Mater. 2019, 31, 1807916.

Y. Ekubaru, T. Sugahara, K. Ibano, A. Suetake, M. Tsurumoto, N. Kagami, K. Suganuma, Adv. Mater. Technol. 5, 1901128.

Y. Nonoguchi, M. Nakano, T. Murayama, H. Hagino, S. Hama, K. Miyazaki, R. Matsubara, M. Nakamura, T. Kawai, Adv. Funct. Mater. 2016, 26, 3021.

F. Völklein, E. Kessler, Phys. Status Solidi A 1984, 81, 585.

I. Hatta, Rev. Sci. Instrum. 1985, 56, 1643.

V. D. Das, N. Soundararajan, Phys. Rev. B 1987, 35, 5990.