Realizing a high-performance n-type thermogalvanic cell by tailoring the thermodynamic equilibrium

Abstract

Ionic thermogalvanic cells (TGCs) have attracted interest for their superior thermopower (α) compared to electronic systems. To maximize the thermopower and overall device performance, it is necessary to integrate both p- and n-type TGCs. However, while high-performance p-type TGCs have been well reported, there are few reports on n-type TGCs. Here, an innovative high-performance n-type TGC is proposed based on an anionic polymer (AP) and hydroquinone (HQ). The AP facilitates self-regulation of the pH in the polymer matrix, which controls the equilibrium between the HQ and its redox partner, benzoquinone (BQ). Moreover, the AP enables the selective transport of the target redox material, leading to the accumulation of HQ near the cold electrode and the spontaneous reaction of HQ to form BQ. The resulting n-type TGC exhibits a superior α of 4.29 mV K⁻¹ compared to previously-reported n-type quasi-solid systems. Moreover, a high Carnot-relative efficiency (1.05%) was achieved in n-type TGCs.

---

Broader context

Thermogalvanic cells (TGCs) have immense potential for wearable energy generators compared to thermocells based on electronic systems due to their superior thermopower (α). For the development of high-performance TGCs, it is crucial to integrate both p- and n-type TGCs. However, most research has focused on p-type TGCs, while very few results have been reported on n-type TGCs. Thus, the development of high-performance n-type TGCs is significant. Hence, anionic polymers (APs) are designed, which contribute to self-regulating the pH for the dominant specific redox material and selective transport of the target redox material. Consequently, an AP-based TGC exhibits a high n-type performance via thermodynamic equilibrium tailoring. Moreover, the wearable TGC worn on the wrist lights up an LED regardless of motion. Therefore, this work demonstrates the significance of the polymer matrix and thermodynamic equilibrium tailoring for high-performance n-type TGC as a wearable energy generator.

Introduction

Thermocells (TCs) have been extensively studied for the effective utilization of waste heat from diverse fields (e.g., waste industrial, solar, geothermal, and body heat).^1–4 In particular, as the human body releases heat almost continuously, wearable TCs that use body heat have attracted immense interest as a power source for wearable electronic devices.^4–8 However, the modest temperature difference of approximately 7 °C between the human body and its surrounding environment is an obstacle that must be overcome.^9,10 Therefore, it is essential to develop deformable quasi-solid state TCs with high thermopower (α) values.

Unfortunately, traditional TCs based on electronic systems, such as inorganic materials and semiconducting polymers, have exhibited extremely low α values (typically around several hundred μV K⁻¹ per unit pair).^11–14 Accordingly, ionic TCs have been considered as promising alternatives due to their high α values.^14–16 These ionic TCs are classified as either thermodiffusion cells (TDCs)^17–21 or thermogalvanic cells (TGCs).^22–25 In TDCs, the Soret effect leads to the transport of ions and their accumulation at the cold-side electrode, thereby generating a potential difference. However, as this does not directly promote electron transfer, continuous power generation is unavailable.^26,27 By contrast, TGCs are able to transfer electrons directly via redox reactions at the electrode interface.^28–30

As the potential difference in the TGC is determined by the temperature-dependent redox reaction, the choice of redox couple is critical to realizing a high-α TGC.^16,31 For example, Fe(CN)₆³⁻/Fe(CN)₆⁴⁻ has been extensively used as a p-type redox couple in TGCs due to its outstanding thermoelectric performance.^32–35 Moreover, recent dramatic improvements in the performance of the Fe(CN)₆³⁻/Fe(CN)₆⁴⁻-based TGCs have been achieved via the crystallization strategy (giving α = −6.5 mV K⁻¹)³³ and by the synergetic combination of a TDC and TGC (giving α = −17 mV K⁻¹).³⁶

However, the integration of both p- and n-type TGCs is eventually required to enhance the overall α value.^37,38 Consequently, the development of both high-performance p-, and n-type TGCs is crucial.³⁸ Nonetheless, only a few n-type TGCs have been reported. For example, Chen et al. reported a high α of 2.02 mV K⁻¹ in n-type quasi-solid TGCs via the selective crystallization strategy.³⁹ Nevertheless, the α values of n-type TGCs are generally much lower than those of p-type TGCs. This difference arises because p-type redox couples typically possess higher absolute charges, and, hence, higher coulombic strengths. This, in turn, leads to a higher entropy difference and, hence, a higher α value.¹⁵

Herein, a strategy is proposed for dramatically enhancing the n-type α value by tailoring the thermodynamic equilibrium of the redox reaction. Specifically, hydroquinone (HQ) was employed as the redox material. The reaction between HQ and benzoquinone (BQ) is reversible depending on the proton concentration (H⁺) via the proton-coupled electron transfer (PCET) reaction (1):^40–42

HQ ⇌ BQ + 2H⁺ + 2e⁻ (1)

This PCET reaction generates a potential difference according to the ratio of HQ to BQ (Note 1, ESI†).^43–45 Therefore, a maximized concentration of HQ is crucial for realizing a high-performance n-type TGC.

In addition, as another key parameter for controlling the PCET, the H⁺ concentration is adjusted by using an anionic co-polymer (AP) polymerized from the two monomers lithium 3-sulfopropyl methacrylate (LSP) and 2-acrylamido-2-methyl-1-propane sulfonic acid (AMPS). The LSP moieties absorbed water from the ambient environment, thereby aiding the dissociation of H⁺ from the AMPS. This, in turn, led to a decrease in the pH of the polymer matrix and an increase in the HQ concentration due to Le Chatelier's principle. The dominant HQ acted as the primary mobile species, and was transported to the cold electrode under the temperature gradient. Meanwhile, the transportation of H⁺ was impeded by electrostatic interaction with the AP. Consequently, the temperature gradient caused an increased concentration gradient of HQ on the cold side by the Soret effect, which led to a potential difference through effective redox reaction from HQ into BQ. Thereby, it generated two electrons per molecule at the cold-side electrode, corresponding to n-type thermoelectric behavior. As a result, the AP-based TGC exhibited an impressive n-type α value of 4.29 mV K⁻¹, along with a Carnot relative efficiency of 1.05% compared to the previously-reported n-type TGCs. In addition, the device exhibited excellent self-healing properties in only 20 min, along with a remarkable stretchability of 1703%, thereby increasing the practicality of the TGC as a wearable power generator. Moreover, when attached to the wrist, the as-fabricated TGC lit up a light-emitting diode (LED) by utilizing the body heat regardless of wrist motion. These results demonstrated the potential of the thermodynamic equilibrium tailoring strategy for high-performance TGCs. In addition, the AP was proven to be a key factor in determining the n-type thermoelectric properties.

Results and discussion

The schematic describing the phenomenon in this study is illustrated in Fig. 1a. The AP absorbed water from the ambient environment, which aided the dissociation of H⁺ from the AMPS. Consequently, the pH of the polymer matrix was decreased, thereby regulating the equilibrium between HQ and BQ in the redox couple to affect the thermoelectric properties. Thus, the polymer matrix requires a large number of acid groups. Therefore, AMPS was used in this study. In addition, a mobile lithium counterion (Li⁺) was used to regulate the water absorption properties of the polymer, as this ion is known to have the largest hydration shell.⁴⁶ To compare the AP properties, we synthesized 3 polymers, namely: poly(3-sulfopropyl methacrylate potassium) (PKSP), pol(3-sulfopropyl methacrylate lithium) (PLSP), and poly(2-acrylamide-2-methyl-1-propane sulfonic acid_0.5-r-3-sulfopropyl methacrylate lithium_0.5) (PLAM) (Fig. 1b and Note 3). The detailed synthesis procedures and characterizations by nuclear magnetic resonance (NMR) spectroscopy are described in Note 3 and Fig. S1 (ESI†).

Fig. 1 The design and properties of the anionic polymers (APs). (a) A schematic diagram of the pH control and H⁺ capture in the AP. (b) The chemical structures of the various as-fabricated APs and photographic images of the corresponding free-standing films. (c) The observed changes in pH according to the water retention in the polymer matrix at 50% RH and temperature 25 and 70 °C. The complex capacitance of the (d) PLSP and (e) PLAM at frequencies of 10⁰ and 10⁶ Hz at DC voltages of 0, 0.25, and 0.50 V.

The ion side chains in the polymers contained different alkali ions as counter ions, and these ions form diverse hydration shells through ion–dipole interaction.^46,47 The potassium, sodium, and lithium ions theoretically enable interaction with the 2, 3, and 4 water molecules in their respective hydration shells.⁴⁸ Therefore, PLSP exhibited a higher water absorption capacity compared to PKSP. Moreover, the saturation time was also faster in PLSP under 50% relative humidity (RH) and 25 °C conditions (Fig. S2, ESI†). Notably, the PLAM also exhibited significant water absorption, along with a comparable absorption rate to that of the PLSP (Fig. S2, ESI†).

The changes in the pH of the polymer matrix were investigated depending on the water absorption and evaporation. Here, both the PLSP and PKSP, which lack any acid groups, exhibited constant pH values as the water retention increased or decreased. By contrast, the presence of the acid group in the PLAM enabled a pH response according to the water retention.⁴⁹ This was because the water within the polymer matrix increased the dissociation of the H⁺, which caused the pH of the PLAM to decrease from 5.5 to 3.0 during the water absorption. Conversely, an increase in pH was observed as the evaporation of water within the polymer matrix by heat (Fig. 1c).

Due to the abovementioned H⁺ dissociation, two types of cations (H⁺ and Li⁺) are present in the PLAM. To reveal the behavior of these ions, the S–O stretching peak which interacts with the cations (1038 cm⁻¹) in the Fourier-transform infrared (FT-IR) spectrum was observed. As the temperature increases, the dissociation between cations and sulfonate groups becomes more facile, leading to a blue shift at the S–O stretching peak.¹⁸ In PLSP with only Li⁺, a significant blue shift was observed when the temperature rose. However, PLAM with H⁺ and Li⁺ exhibited a less blue shift (Fig. S3, ESI†). Consequently, Li⁺ had a weak interaction with sulfonate groups and readily separated to function as mobile ions, whereas H⁺ demonstrated a stronger interaction and had limitations for mobile ions. In addition, to support this, electrical double layer (EDL) capacitance was investigated (Note 2, ESI†).⁵⁰ The PLSP (which contains only Li⁺) exhibited a constant EDL capacitance regardless of the direct current (DC) voltage (Fig. 1d and Fig. S4, ESI†). Thus, most of the Li⁺ were already dissociated at 0 V DC, and were transported to the electrode interface. By contrast, the EDL capacitance of the PLAM increased from 0 V to 0.25 V, and remained constant thereafter as the voltage increased (Fig. 1e and Fig. S4, ESI†). This indicated that a DC voltage is required for H⁺ transportation. Consequently, only PLAM was appropriate for use in an n-type TGC due to its H⁺ transportation control capability.

The amount of HQ was added at ratios of 0.5, 1.0, 1.5, and 2.0 molar equivalent with regard to the amount of acid group in PLAM. Four samples were abbreviated as follows: PLAMHQ0.5, PLAMHQ1.0, PLAMHQ1.5, and PLAMHQ2.0. As shown schematically in Fig. 2a, the –OH group in HQ is expected to interact with the sulfonic acid and sulfonate groups via hydrogen bonding. This was confirmed by the FT-IR spectra.⁵¹ The O–H bending peak in HQ and S[double bond, length half m-dash]O stretching peak in PLAM were indicated at 1514 and 1366 cm⁻¹, respectively. The series of PLAM exhibited a red shift, which means the vibration suppression due to the interaction. However, the excessive amount of HQ (over PLAMHQ1.5) surpassed the threshold for interaction, thereby resulting in phase separation (Fig. 2a–c, and Fig. S5, ESI†). Notably, these interactions serve to increase the free volume between the polymers, as demonstrated by the temperature-dependent heat flow measurements (Fig. S6, ESI†). The glass transition temperature (T_g) values of the various PLAMHQ samples were seen to decrease as the amount of HQ was increased, thereby indicating increases in the free volume and softness of the polymer at ambient temperature. Consequently, the break of elongation of PLAMHQ0.5 and PLAMHQ1.0 dramatically increased (≈1512 and 1703%) (Fig. 2d).

Fig. 2 The interaction between PLAM and HQ, and the effect on the mechanical properties. (a) A schematic diagram showing the interaction between the PLAM and HQ according to the amount of HQ. (b) and (c) The FT-IR spectra of the HQ, PLAM, PLAMHQ0.5, PLAMHQ1.0, PLAMHQ1.5, and PLAMHQ2.0. (d) The tensile stress–strain (S–S) curves of the PLAM, PLAMHQ0.5, and PLAMHQ1.0. (e) The storage (G′) and loss (G′′) moduli of the PLAM, PLAMHQ0.5, and PLAMHQ1.0 under angular frequencies. (f) The segmental relaxation time (τ_s) values of the PLAM, PLAMHQ0.5, and PLAMHQ1.0 at 0.05 rad s⁻¹. (g) A quantitative analysis of the tensile S–S curves, along with an inset showing a qualitative analysis of optical microscope (OM) images of the PLAMHQ1.0 obtained before cutting and after 10 and 20 min of self-healing.

Meanwhile, the gel-like properties and segmental motions of the various PLAMHQ samples were investigated by rheological investigations. Here, each sample exhibited a larger storage modulus relative to its loss modulus at angular frequencies of 0.05 to 100 rad s⁻¹, thereby indicating stable elastic properties despite the increase in free volume (Fig. 2e and S7, ESI†). Furthermore, the segmental relaxation time (τ_s) values of the various samples were analyzed (Fig. S8, ESI†). These were calculated from relaxation time (λ) at 0.05 rad s⁻¹ by using eqn (2):^52,53

J′ = G′/([η*]ω)² = λ/[η*] (2)

where J′, G′, [η*], and ω are the storage compliance, storage modulus, complex viscosity, and angular frequency, respectively. The PLAMHQ0.5 and PLAMHQ1.0 exhibited shorter τ_s values of ∼15.9 and 15.7 s, respectively, compared to that of the PLAM (∼17.7 s) (Fig. 2f). This indicated that the ionic side chains in the PLAMHQ1.0 had the most vigorous motion, which enabled rapid self-healing (re-bonding between the ionic side chains) due to the reversible electrostatic interaction.^53,54 This effect was evaluated via a qualitative analysis of the optical microscope (OM) images of the various PLAM films on bare glass in the inset of Fig. 2g and in Fig. S9 (ESI†). Here, the cutting traces that were generated by a blade were observed over time. Although the cutting trace on the PLAM faded gradually, it remained visible even after 20 min. By contrast, the cutting trace on the PLAMHQ1.0 disappeared completely within 20 min, thereby demonstrating excellent self-healing. Moreover, a quantitative analysis of this effect was provided by the S–S curves of the cut free-standing films (Fig. 2g and Fig. S9, ESI†). After 20 min of self-healing time, the PLAM, PLAMHQ0.5, and PLAMHQ1.0 films exhibited 68.4, 79.1, and 99.8% break of elongation compared to the original (uncut) films, respectively.

As noted above, the incorporation of HQ into the PLAM led to an expansion of the free volume. This resulted in a reduction in the amount of Li⁺ within a unit volume of the PLAM. This, in turn, was expected to influence the electrical and thermoelectric properties of the various films. Thus, PLAM had the highest 1.25 mS cm⁻¹ ionic conductivity (σ_i). PLAMHQ0.5 and PLAMHQ1.0 exhibited over 1 mS cm⁻¹, implying rapid ion migration in the polymer matrix (Fig. S10, ESI†).

Because the PCET reaction between HQ and BQ is reversible, the HQ coexists with BQ in the polymer matrix.⁴² The changes in the relative ratios of HQ and BQ over time were compared by FT-IR. Here, the overall intensity of the peaks decreased over time due to the gradual adsorption of water. Therefore, the peaks of HQ were compared based on the C–N stretching peak in 1040 cm⁻¹, which is only in PLAM. After 1.5 h, the intensity of the O–H bending peak in 1512 cm⁻¹- and the C–O stretching peak in 1180 cm⁻¹ increased in the PLAMHQ (Fig. 3a).⁵⁰ In addition, the UV-vis absorption peaks of HQ and BQ in the various PLAMHQ samples were distinguished. The peak below 250 nm absorbed by BQ gradually decreased, while the 292 nm peak by HQ increased after 1.5 h (Fig. S11, ESI†).⁴² This indicated that the amount of HQ increased from 0 h to 1.5 h.

Fig. 3 The thermoelectric properties and mechanism. (a) The time-dependent FT-IR spectra of the various PLAMHQ films. (b) The thermopowers of the PLSP, PLAM, PLAMHQ0.5, and PLAMHQ1.0. (c) The thermopowers of previously-reported quasi-solid n-type TGCs.^{28–30,39,55–59} The complex capacitance values of the (e) PLAMHQ0.5 and (d) PLAMHQ1.0 at frequencies of 10⁰ to 10⁶ Hz and DC voltages of 0, 0.25, and 0.50 V. (f) A schematic diagram of an n-type TGC based on the PLAMHQ mechanism, voltage distribution in the polymer matrix, and PCET redox reactions at the hot and cold sides.

The HQ and BQ act as a redox couple, and the ratio between the two affects the type of thermoelectric behavior. In thermoelectrics, the driving force is the applied temperature gradient. This causes diffusion and convection of mobile ions and the redox couple (HQ/BQ) in the polymer matrix via the Soret effect.^60,61 The Li⁺, which is relatively light, reached the cold electrode rapidly and generated a voltage by establishing a potential difference between the hot and cold sides. To compensate for the positive charge of Li⁺, electrons are attracted to the cold side, thus leading to the formation of an EDL similar to the thermodiffusion cell mechanism. Therefore, PLSP and PLAM showed low −0.73 and −1.30 mV K⁻¹ p-type α, respectively (Fig. 3b). Meanwhile, the PLAMHQ0.5 and PLAMHQ1.0 exhibited n-type thermoelectric properties, with α values of 2.09 and 4.29 mV K⁻¹, respectively (Fig. 3b). Moreover, a comparison with Fig. 3c indicated that these α values were superior to those of previously-reported n-type polymer-based TGCs, which was attributed to the significantly high ratio of HQ to BQ.

The transport of H⁺ through the PLAMHQ0.5 and PLAMHQ1.0 matrix via the FT-IR spectra and EDL capacitance was investigated. As the temperature increased, the S–O stretching peak of PLAMHQ0.5 and PLAMHQ1.0 exhibited a subtle blue shift (Fig. S12, ESI†). Moreover, the EDL capacitance of the PLAMHQ0.5 and PLAMHQ1.0 increased from 0 V to 0.5 V DC voltage conditions (Fig. 3d, e, and Fig. S13, ESI†). These results indicated that H⁺ is captured by the anionic side chains. Consequently, only Li⁺ was transported towards the cold side. Moreover, after the rapid transport of Li⁺, a significant amount of HQ was transported to the cold side, where the PCET reaction of HQ to form BQ occurred in accordance with Le Chatelier's principle. The electrons generated from this reaction were transferred to the cold-side electrode, while the BQ and extra H⁺ were transported to the hot side due to the concentration gradient. Thus, the reaction of BQ to HQ occurred at the hot side and, hence, n-type thermoelectric behavior was observed in the PLAMHQ0.5 and PLAMHQ1.0 (Fig. 3f and Fig. S14, ESI†).

To confirm the performance of the TGC based on PLAMHQ1.0 according to water retention, PLAMHQ1.0 was investigated in various RH conditions (70% and 90% RH). As RH rose, PLAMHQ1.0 absorbed more water from the surroundings (Fig. S15, ESI†), thereby promoting the dissociation of more cations from the sulfonate groups. The cation dissociation was confirmed through the blue shift of the S–O stretching peak from a dry state to 90% RH in the FT-IR spectra, indicating the weaker interaction between the cations and sulfonate groups (Fig. S16, ESI†). Moreover, the increment of capacitance was very limited at high RH (e.g., 90%) compared to the 50% RH condition (Fig. S17, ESI†). This implies a higher concentration of H⁺ free from the sulfonate groups (i.e., low pH) (Fig. S18, ESI†). In addition, improved σ_i was observed at high RH (Fig. S19, ESI†). This enhancement allowed H⁺ to participate in the device working. Thus, the n-type TGC performance was degraded at high RH (Fig. S20, ESI†). On the other hand, when the polymer matrix was dried (namely, under dried conditions), the mass transport of all species was severely reduced. As a result, only a minute cell potential was established by minimal thermo-diffusion, instead of faradaic electrochemical reactions of HQ/BQ (Fig. S21, ESI†). Therefore, regulating H⁺ transport is considered a crucial parameter for TGC based on PLAMHQ1.0, and 50% RH was concluded to be the optimal condition.

Moreover, to determine the effect of the ratio between the HQ and BQ, PLSP with 1.0 molar equivalent of HQ (i.e., PLSP with 1.0 molar equivalent of HQ) was examined. PLSPHQ1.0 solution exhibited a distinct brown color in contrast to the PLAMHQ0.5 and PLAMHQ1.0 solutions, thereby indicating a substantially higher concentration of BQ inside it (Fig. S22, ESI†). Moreover, although the PLSPHQ1.0 exhibited n-type thermoelectric properties (Fig. S23, ESI†), its α value was extremely low (0.71 mV K⁻¹). Therefore, both anionic and pH-regulating moieties are required for an n-type TGC with a high α value.

For further investigation of the thermoelectric properties, the output voltage-current density curves of the PLAMHQ0.5 and PLAMHQ1.0 were obtained under various temperature conditions, respectively. The open-circuit voltage (V_oc) and short-circuit current density (J_sc) rose with temperature from 10.28 mV and 0.28 A m⁻² to 41.95 mV and 0.88 A m⁻² in PLAMHQ0.5 and from 21.41 mV and 0.43 A m⁻² to 80.44 mV and 1.64 A m⁻² in PLAMHQ1.0 (Fig. 4a and d). Furthermore, to evaluate the output performances of the PLAMHQ0.5 and PLAMHQ1.0, the specific output power density (P_max/ΔT²) of each sample was calculated from the measured J_sc and V_oc values, respectively (Fig. 4b and e). It was insensitive to temperature, as predicted by the theoretical eqn (3):^32,33

Fig. 4 The output performance of the PLAMHQ0.5 and PLAMHQ1.0. (a) and (d) Output voltage–current density–output power curve depending on the temperature difference, (b) and (e) Output power and maximum specific output power density, and (c) and (f) thermal energy conversion and Carnot-relative thermal efficiency depending on the temperature difference. (g) Thermopower and Carnot-relative efficiency of PLAMHQ1.0 and reported TGCs.^62–68 Voltage and thermopower at ΔT = 5 K (h) from 0 to 300% uni-axial stretching and (i) under repeated 100% uni-axial stretching. (h) Inset visual image of PLAMHQ1.0 TGC under 0 and 300% stretching.

In terms of TGC performance, the heat-to-electricity conversion efficiency is the most critical parameter. To compare the energy conversion efficiency, the thermal energy conversion efficiency (η) and Carnot relative efficiency (η_r) were calculated from eqn (4) and (5), respectively:^32,33

where k is the thermal conductivity. The maximum of η_r was 0.37% for the PLAMHQ0.5 and 1.05% for the PLAMHQ1.0 at a ΔT of 15 K (Fig. 4c and f). Notably, due to its remarkable α value (4.29 mV K⁻¹) and significantly low k value (0.014 W m⁻¹K⁻¹), the PLAMHQ1.0 exhibited a higher η_r value than the previously reported n-type TGCs (Fig. 4g and Table S1, ESI†).

Furthermore, rapid self-healing, high stretchability, and repeatability are also crucial properties because wearable TGCs are utilized in a dynamic environment. The PLAMHQ1.0, which had rapid self-healing and high stretchability, was reflected precisely as it is in the TGCs. The thermoelectric properties of the as-fabricated TGC measured at hourly intervals after cutting with a blade were measured (Fig. S24, ESI†). The PLAMHQ1.0-based device maintained its initial voltage at ΔT = 5 K throughout the self-healing. Moreover, the initial voltage was maintained after stretching once by ∼300% (Fig. 4h) and after repeated stretching from 0 to 100% (Fig. 4i). The deformation of the polymer alters the geometric structure (e.g. the cross-sectional area and the distance between each electrode), which mainly affects the kinetic properties of the mobile materials.¹⁷ Consequently, there was an increase in the duration required for the voltage to reach saturation under uniaxial stretching at ΔT = 5 K (Fig. S25, ESI†). Nevertheless, the total amount of redox couples was consistent. Hence, deformation would not change the voltage generated by the redox couple. In addition, PLAMHQ1.0 exhibited consistent performance during the 40th measurement (Fig. S26, ESI†).

The application of the PLAMHQ1.0-based TGC was further demonstrated in Fig. 5, where the device was used to operate an LED. The LED lit up at 5 K temperature difference by the Peltier devices and showed a consistent active state while the TGC was uni-axially stretched over 200% and recovered (Fig. 5a). In addition, to confirm the self-healing capability of the device, the TGC was divided into two parts by a blade and then reattached. The LED became inactive during this procedure. However, the LED lit up when the temperature difference of 5 K was applied after self-healing (Fig. 5b).

Fig. 5 The application of the wearable TGC. (a) Photographic images showing the continuous operation of an LED during 200% uniaxial stretching and recovery. (b) Photographic images showing the inactive state of the LED before self-healing, and the active state of the LED at ΔT = 5 K after self-healing. (c) A schematic diagram of the wearable TGC on the wrist, and a thermal image of the wrist and its surroundings. (d) Photographic images of the wearable TGC attached to the wrist and lighting up the LED by utilizing the body heat.

To generate continuous electrical energy, the TGC requires a continuous temperature gradient. The human body constantly dissipates heat to the surroundings, thereby generating a temperature gradient between the body and its surroundings.¹⁰ To demonstrate this, the wearable TGC was applied to the wrist, and the temperatures of the wrist and surroundings were measured by using a thermal imaging camera. There was an approximately 6 °C temperature difference (Fig. 5c), and this temperature difference was adequate for driving the TGC. When there was a disconnection between the wearable TGC and the LED, the LED remained in an inactive state. However, upon connecting the wearable TGC and the LED, the LED exhibited an active state and remained in this state. In addition, the LED worked successfully regardless of wrist motion (Fig. 5d and Video S1, ESI†). This working of the LED by utilizing waste body heat clearly demonstrated the tremendous potential of PLAM with HQ for a wearable thermoelectric energy generator.

Conclusions

In this study, we designed an AP with protonation, regulating the pH depending on the water absorption. The regulation of pH caused a large population of HQ compared to BQ. Moreover, the anionic moieties in the AP allowed for the selective transport of HQ (rather than H⁺), thereby leading to the PCET reaction of HQ to form BQ at the cold side in accordance with Le Chatelier's principle. Moreover, this strategy was successfully used to realize a TGC with an n-type α of 4.29 mV K⁻¹ for the PLAMHQ1.0. Notably, this was the highest reported value for quasi-solid state n-type TGCs. In addition, the calculated η_r of the PLAMHQ1.0 (1.05%) was higher than that of the previously-reported n-type TGCs, thereby indicating a higher conversion of heat to electric energy. Furthermore, the PLAMHQ1.0 exhibited a consistent voltage after cutting with a blade, upon stretching of up to 300%, and after repeated stretching from 0–100% and releasing. Finally, a wearable TGC was fabricated using the PLAMHQ1.0 and shown to operate an LED by utilizing body heat when attached to the wrist. Moreover, the LED remained in the active state despite the motion of the wrist. Overall, these results demonstrated the significance of tailoring the thermodynamic equilibrium in the ionic polymer matrix for the development of a high-performance n-type TGC. Moreover, the feasibility of the as-designed PLAMHQ was demonstrated for use as a wearable energy generator.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the Nano & Material Technology Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (RS-2023-00283244). This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MIST) (No. 2021R1A2C3004420).

Notes and references

1. Z. Miao, X. Meng and L. Liu, Appl. Therm. Eng., 2023, 228, 120480.
2. H. Yang, S. A. Khan, N. Li, R. Fang, Z. Huang and H. Zhang, J. Chem. Eng., 2023, 473, 145247.
3. L. Catalan, P. Alegria, M. Araiz and D. Astrain, Appl. Therm. Eng., 2023, 222, 119843.
4. J. Li, S. Chen, Z. Wu, Z. Han, X. Qu, M. Jin, Y. Jia, Z. Zhou and H. Wang, Nano Energy, 2023, 112, 108482.
5. D. Zhang, Y. Mao, F. Ye, Q. Li, P. Bai, W. He and R. Ma, Energy Environ. Sci., 2022, 15, 2974.
6. X. Fu, Z. Zhuang, Y. Zhao, B. Liu, Y. Liao, Z. Yu, P. Yang and K. Liu, ACS Appl. Mater. Interfaces, 2022, 14, 44792.
7. D. Zhang, Y. Zhou, Y. Mao, Q. Li, L. Liu, P. Bai and R. Ma, Nano Lett., 2023, 23, 11272.
8. T. H. Park, B. Kim, S. Yu, Y. Park, J. W. Oh, T. Kim, N. Kim, Y. Kim, D. Zhao, Z. U. Khan, S. Lienemann, X. Crispin, K. Tybrandt, C. Park and S. C. Jun, Nano Energy, 2023, 114, 108643.
9. J. Duan, B. Yu, L. Huang, B. Hu, M. Xu, G. Feng and J. Zhou, Joule, 2023, 5, 768.
10. C. Tian, C. Bai, T. Wang, Z. Yan, Z. Zhang, K. Zhuo and H. Zhang, Nano Energy, 2023, 106, 108077.
11. Y.-H. Pai, J. Tang, Y. Zhao and Z. Liang, Adv. Energy Mater., 2023, 13, 2202507.
12. Y. Chen, Y. Zhao and Z. Liang, Energy Environ. Sci., 2015, 8, 401.
13. R. Kita, P. Polyakov and S. Wiegand, Macromolecules, 2007, 40, 1638.
14. S. Sun, M. Li, X.-L. Shi and Z.-G. Chen, Adv. Energy Mater., 2023, 13, 2203692.
15. J. Zhang, C. Bai, Z. Wang, X. Liu, X. Li and X. Cui, Micromachines, 2023, 14, 155.
16. M. F. Dupont, D. R. MacFarlane and J. M. Pringle, Chem. Commun., 2017, 53, 6288.
17. S. Kim, M. Ham, J. Lee, J. Kim, H. Lee and T. Park, Adv. Funct. Mater., 2023, 2305499.
18. D. H. Ho, Y. M. Kim, U. J. Kim, K. S. Yu, J. H. Kwon, H. C. Moon and J. H. Cho, Adv. Energy Mater., 2023, 13, 2301133.
19. D. Zhao, A. Martinelli, A. Willfahrt, T. Fischer, D. Bernin, Z. U. Khan, M. Shahi, J. Brill, M. P. Jonsson, S. Fabiano and X. Crispin, Nat. Commun., 2019, 10, 1093.
20. H. Cheng, X. He, Z. Fan and J. Ouyang, Adv. Energy Mater., 2019, 9, 1901085.
21. Z. A. Akbar, J.-W. Jeon and S.-Y. Jang, Energy Environ. Sci., 2020, 13, 2915.
22. Z. Lei, W. Gao and P. Wu, Joule, 2021, 5, 2211.
23. P. Peng, J. Zhou, L. Liang, X. Huang, H. Lv, Z. Liu and G. Chen, Nano-Micro Lett., 2022, 14, 81.
24. X. Shi, L. Ma, Y. Li, Z. Shi, Q. Wei, G. Ma, W. Zhang, Y. Guo, P. Wu and Z. Hu, Adv. Funct. Mater., 2023, 33, 2211720.
25. J. Li, Z. Wang, S. A. Khan, N. Li, Z. Huang and H. Zhang, Nano Energy, 2023, 113, 108612.
26. H. Wang, D. Zhao, Z. U. Khan, S. Puzinas, M. P. Jonsson, M. Berggren and X. Crispin, Adv. Energy Mater., 2017, 3, 1700013.
27. M. Fu, Z. Sun, X. Liu, Z. Huang, G. Luan, Y. Chen, J. Peng and K. Yue, Adv. Funct. Mater., 2023, 33, 2306086.
28. Y. Liu, S. Zhang, Y. Zhou, M. A. Bukingham, L. Aldous, P. C. Sherrell, G. G. Wallace, G. Ryder, S. Faisal, D. L. Officer, S. Beirne and J. Chen, Adv. Energy Mater., 2020, 10, 2002539.
29. P. Yang, K. Liu, Q. Chen, X. Mo, Y. Zhou, S. Li, G. Feng and J. Zhou, Angew. Chem., Int. Ed., 2016, 55, 12050.
30. J. Duan, B. Yu, K. Liu, J. Li, P. Yang, W. Xie, G. Xue, R. Liu, H. Wang and J. Zhou, Nano Energy, 2019, 57, 473.
31. T. J. Abraham, D. R. MacFarlane and J. M. Pringle, Energy Environ. Sci., 2013, 6, 2639.
32. B. Yu, J. Duan, H. Cong, W. Xie, R. Liu, X. Zhuang, H. Wang, B. Qi, M. Xu, Z. L. Wang and J. Zhou, Science, 2020, 370, 342.
33. L. Liu, D. Zhang, P. Bai, Y. Mao, Q. Li, J. Guo, Y. Fang and R. Ma, Adv. Mater., 2023, 35, 2300696.
34. Y. Wang, Y. Zhang, X. Xin, J. Yang, M. Wang, R. Wang, P. Guo, W. Huang, A. J. Sobrido, B. Wei and X. Li, Science, 2023, 381, 291.
35. P. Yin, Y. Geng, L. Zhao, Q. Meng, Z. Xin, L. Luo, B. Wang, Z. Mao, X. Sui, W. Wu and X. Feng, J. Chem. Eng., 2023, 457, 141274.
36. C.-G. Han, X. Qian, Q. Li, B. Deng, Y. Zhu, Z. Han, W. Zhang, W. Wang, S.-P. Feng, G. Chen and W. Liu, Science, 2020, 368, 1091.
37. M. A. Maimani, J. J. Black and L. Aldous, Electrochem. Commun., 2016, 72, 181.
38. W. Li, J. Ma, J. Qiu and S. Wang, Mater. Today Energy, 2022, 27, 101032.
39. W. Gao, Z. Lei, C. Zhang, X. Liu and Y. Chen, Adv. Funct. Mater., 2021, 31, 2104071.
40. B. Guo, Y. Hoshino, F. Gao, K. Hayashi, Y. Miura, N. Kimizuka and T. Yamada, J. Am. Chem. Soc., 2020, 142, 17318.
41. M. Ishii, Y. Yamashita, S. Watanabe, K. Ariga and J. Takeya, Nature, 2023, 622, 285.
42. Y. Wang, Y. Dai, L. Li, L. Yu and W. Zeng, Angew. Chem. Int. Ed., 2023, 62, e202307947.
43. H. S. Harned and D. D. Wright, J. Am. Chem. Soc., 1933, 55, 4849.
44. Y. Song and G. R. Buettner, Free Radic. Biol. Med., 2010, 49, 919.
45. Y. Song, J. Xie, Y. Song, H. Shu, G. Zhao, X. Lv and W. Xie, Spectrochim. Acta A Mol. Biomol., 2006, 65, 333.
46. R. Buchner, W. Wachter and G. Hefter, J. Phys. Chem. B, 2019, 123, 10868.
47. J. Luo, S. Ye, T. Li, E. Sarnello, H. Li and T. Liu, J. Phys. Chem. C, 2019, 123, 14825.
48. H. Lu, W. Shi, J. H. Zhang, A. C. Chen, W. Guan, C. Lei, J. R. Greer, S. V. Boriskina and G. Yu, Adv. Mater., 2022, 34, 2205344.
49. H. Wang, T. He, X. Hao, Y. Huang, H. Yao, F. Liu, H. Cheng and L. Qu, Nat. Commun., 2022, 13, 2524.
50. I. You, D. G. Mackanic, N. Matsuhisa, J. Kang, J. Kwon, L. Beker, J. Mun, W. Suh, T. Y. Kim, J. B.-H. Tok, Z. Bao and U. Jeong, Science, 2020, 370, 961.
51. L. E. Mphuthi, M. R. Maseme and E. H. G. Langner, J. Inorg. Organomet. Polym. Mater., 2022, 32, 2664.
52. Y. Eom, S.-M. Kim, M. Lee, H. Jeon, J. Park, E. S. Lee, S. Y. Hwang, J. Park and D. X. Oh, Nat. Commun., 2021, 12, 621.
53. Y. M. Kim, J. H. Kwon, S. Kim, U. H. Choi and H. C. Moon, Nat. Commun., 2022, 13, 3769.
54. J. Lee, M. W. M. Tan, K. Parida, G. Thangavel, S. A. Park, T. Park and P. S. Lee, Adv. Mater., 2020, 32, 1906679.
55. W. Gao, H. Meng, Y. Chen and X. Liu, Appl. Phys. Lett., 2022, 121, 203902.
56. M. A. Buckingham, F. Marken and L. Aldous, Sustainable Energy Fuels, 2018, 2, 2717.
57. A. Taheri, D. R. MacFarlane, C. Pozo-Gonzalo and J. M. Pringle, ChemSusChem, 2018, 11, 2788.
58. A. Taheri, D. R. MacFarlane, C. Pozo-Gonzalo and J. M. Pringle, Electrochim. Acta, 2019, 297, 669.
59. A. Taheri, D. R. MacFarlane, C. Pozo-Gonzalo and J. M. Pringle, Sustainable Energy Fuels, 2018, 2, 1806.
60. D. Zhao, H. Wang, Z. U. Khan, J. C. Chen, R. Gabrielsson, M. P. Jonsson, M. Berggren and X. Crispin, Energy Environ. Sci., 2016, 9, 1450.
61. Y. Fang, H. Cheng, H. He, S. Wang, J. Li, S. Yue, L. Zhang, Z. Du and J. Ouyang, Adv. Funct. Mater., 2020, 30, 2004699.
62. H. Zhou, T. Yamada and N. Kimizuka, J. Am. Chem. Soc., 2016, 138, 10502.
63. P. F. Salazar, S. T. Stephens, A. H. Kazim, J. M. Pringle and B. A. Cola, J. Mater. Chem. A, 2014, 2, 20676.
64. K. Kim, S. Hwang and H. Lee, Electrochim. Acta, 2020, 335, 135651.
65. K. Kim and H. Lee, Phys. Chem. Chem. Phys., 2018, 20, 23433.
66. K. Kim and H. Lee, Energy Technol., 2019, 7, 1900857.
67. A. Gunawan, C.-H. Lin, D. A. Buttry, V. Mujica, R. A. Taylor, R. S. Prasher and P. E. Phelan, Nanoscale Microscale Thermophys. Eng., 2013, 17, 304.
68. U. B. Holeschovsky, Analysis of flooded flow fuel cells and thermogalvanic generators. Massachusetts Institute of Technology, 1994.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ee00768a

---