Design and Optimization Strategies for Flexible Qua-si‑Solid‑State Thermo-Electrochemical Cells

1. Introduction

Effective utilization of low-grade thermal energy is currently a great challenge. Thermoelectric (TE) effects are an extremely promising way of generating electrical power from a temperature difference. Being noiseless and emission-free is the core advantages of thermoelectric conversion technology. Moreover, a complete thermoelectric device consisting of simply a thermoelectric material and electrodes can be integrated by circuit design, which can be applied to small electronic and wearable devices, allowing for the collection and utilization of thermal energy over a wider range of temperature differences. Thermoelectric materials are categorized into electronic thermoelectric materials, ionic thermoelectric materials, and thermo-electrochemical cells (TECs), based on their mechanism of operation [1,2,3,4].

The majority of research has focused on electronic thermoelectric materials, which are categorized as p-type or n-type depending on whether the carriers are holes or electrons. However, the Seebeck coefficient (S) of the electronic thermoelectric materials remain on the scale of microvolts per Kelvin, and several hundred pairs of p-n legs need to be connected in series to make the devices have a high voltage output for real application (Figure 1) [5,6,7,8].

Ionic thermoelectric materials are based on the Soret effect, which involves an uneven distribution of ions in a conductor due to a temperature gradient. The difference in migration rates of anions and cations leads to a different stacking of ions at the hot and cold terminations, thus resulting in a potential difference between the two electrodes. However, ions cannot enter the external circuit and only depend on the accumulation of ions to generate an induced electromagnetic potential, therefore, the operation principle of ionic thermoelectric devices is similar to a capacitor, which cannot directly supply power to the outside world, and can store thermal energy for converting to electrical energy (Figure 2) [9,10,11,12].

TECs use a redox couple contained in an electrolyte like metal-ion batteries. When a temperature difference occurs between two electrodes, the anion and cation at the electrode gain and lose electrons, respectively. Additionally, the internal anion and cation migrate via the concentration difference diffusion and thermal diffusion constantly in a cyclic manner, thereby generating uninterrupted electricity supply to the exterior [13,14,15]. The Seebeck coefficients of TECs tend to be > 1000 μV K^-1, two orders of magnitude higher than those of electronic thermoelectric materials. Meanwhile, the reaction inside the TECs is usually carried out in the solution system, which has the advantages of low cost, easy fabrication, etc. With the unique advantages of TECs, the optimization methods are gradually diversified as the research on TECs progresses and the understanding of the principle of operation of thermoelectric batteries is deepened accordingly [16,17]. However, the present energy conversion efficiencies of TECs are not satisfactory, therefore, the optimization of the TECs for further increasing the thermoelectric conversion efficiency remains necessary.

Electrolyte, electrode, and mechanism layout are critical for optimizing TECs [18,19]. The optimization of electrode materials and the tuning of redox ions and electrolyte solvents have been focused on in the pursuit of high thermoelectric conversion efficiencies. However, traditional liquid electrolytes suffer from complicated encapsulation and integration problems for wearable applications. To circumvent these issues, a possible approach is to consolidate the liquid electrolyte into a quasi-solid hydrogel electrolyte [20,21]. However, there have been relatively few studies in this field, and this article reviews the advancement of gel-based TECs in recent years. Meanwhile, some problems and challenges are pointed out, and some new perspectives are offered for future research.

2. Wearable Thermo-Electrochemical Cells

2.1. Thermo-Electrochemical Cells

The TECs are constructed with electrolytes containing a redox couple with two electrodes and connected via an external circuit (Figure 3). A temperature gradient is applied to the electrode terminals with a changed electrode potential for redox ions due to the change in temperature. The high-potential electrode is the anode where oxidation reaction occurs, providing electrons to the external circuit; the low-potential electrode is the cathode where reduction reaction occurs, obtaining electrons from the external circuit. The valence state of ions changes at the electrode, which results in the formation of a concentration difference between the two terminals of the electrode, and the ions continuously migrate within the electrolyte through concentration difference diffusion and thermal diffusion, thus making the redox reaction continue and maintaining a consistent output of current and voltage. The operation of TECs involves two critical processes: 1) redox reactions at the electrodes and 2) ion transport processes in the electrolyte. In particular, the redox reaction at the electrode is associated with the TECs Seebeck coefficient, and the ion transport in the electrolyte is associated with the TECs conductivity and thermal conductivity [22,23].

2.1.1. The Seebeck Coefficient of TECs

With the given temperature difference, the magnitude of the potential difference capable of being generated is one of the dominant factors in the energy conversion efficiency of TECs, i.e., the Seebeck coefficient.

$$S=\frac{\Delta V_{\mathrm{OC}}}{\Delta T}\left(1.1\right)$$

In formula 1.1, ΔV_OC is the open-circuit voltage and ΔT is the temperature difference between the two electrode terminals. This is the definition equation for the Seebeck coefficient of a thermoelectric material which applies to any thermoelectric material. For any redox reaction in a TEC [24,25]:

$$\mathrm{A}+n{\mathrm{e}}^{-}\to B \left(1.2\right)$$

The Seebeck coefficient (S_e) of the TECs is defined as:

$$S_{\mathrm{e}}={\left(\frac{\partial E}{\partial T}\right)}_{t=\infty }\left(1.3\right)$$

In formula 1.3, E is the electrode potential, which can be calculated using the Nernst equation, and T is the temperature. If the electrolyte is homogeneous internally, there exists [16,19,25,26]:

In formula 1.4, n is the number of charges transferred in the redox reaction, F is Faraday’s constant, and S_A and S_B are the partial molal entropy of the ions, which result from taking one partial derivation of the total entropy of the ions with respect to the molar quantity of the ions. Ŝ_A and Ŝ_B are Eastman entropy, primarily derived from the interaction of the ions and their solvated shell structures with the surrounding solvent molecules as they move, which would be ignored in the majority of solutions. S_e is the transport entropy of electrons in the external circuit and usually only at the order of microvolts per Kelvin which is similarly negligible. Therefore, equation (1.4) can be simplified to:

$$S_{\mathrm{e}}={\left(\frac{\partial E}{\partial T}\right)}_{t=\infty }\approx \frac{{\mathrm{S}}_{\mathrm{B}}-S_{\mathrm{A}}}{n\mathrm{F}}$$

(1.5)

In formula 1.5, when the partial molar entropy of the reduced ions is greater than that of the oxidized ions the Seebeck value is positive for the p-type TECs ion couple. Whereas, the partial molar entropy of the reduced ions is less than that of the oxidized ions when the Seebeck coefficient is negative for n-type TEC’s ion couple [27].

2.1.2. Performance Index of TECs

During TECs operation, the Seebeck coefficient is not the only determinant of device performance; the electrical conductivity (δ) and thermal conductivity (κ) of a unit cell deserve to be considered in equal measure. Three primary sources of overpotentials exist in TECs: 1) Ohmic overpotentials, which are mainly caused by the internal resistance of the cell itself, the electrode resistance, and the circuit resistance. 2) Charge transfer overpotentials, which are related to the kinetics of redox charge transfer at the electrode surfaces. 3) Mass transfer overpotentials, which are related to the rate of movement of ions through the electrolyte, and which encompasses diffusion of the ions, migration, and convection of the overall solution. Combined, these three factors influence the conductivity of the cell device [28].

The thermal conductivity of an electrolyte is also an important factor in the performance of TECs. For the liquid electrolyte, the thermal conductivity of the solution, convection, heat transfer, and ion mobility are all factors that influence the overall thermal conductivity. If the thermal conductivity is too high, the temperature gradient cannot be maintained between the two electrode terminals, resulting in a decrease in the temperature difference, a decrease in the voltage and current output, and eventually, a temperature equilibrium TECs ceases to operate. Therefore, the performance of thermoelectric materials usually is evaluated by a dimensionless parameter, the thermoelectric merit value ZT [29].

$$ZT=\frac{S^2\sigma }{\kappa }T \left(1.6\right)$$

Similarly, thermoelectric device performance may be evaluated by the energy conversion efficiency η:

$$\eta =\frac{P_{\mathrm{out}}}{Q_{\mathrm{h}}}$$

(1.7)

In formula 1.7, P_out is the output power of the device and Q_h is the thermal energy supplied by the hot terminal. The output power is parabolic to the magnitude of the load resistance of the external circuit, and the output power reaches the maximum when the battery resistance is equal to the load resistance. The maximum energy conversion efficiency η_max can be calculated from the ZT value:

$${\eta }_{\max }=\frac{T_{\mathrm{hot}}-T_{\mathrm{cold}}}{T_{\mathrm{hot}}}\Delta \frac{\sqrt{1+ZT}-1}{\sqrt{1+ZT}+\frac{T_{\mathrm{cold}}}{T_{\mathrm{hot}}}}\left(1.8\right)$$

In equation 1.8, T_hot is the temperature of the hot terminal of the device, T_cold is the temperature of the cold terminal of the device, and the temperature adopted in the calculation of ZT value is the average temperature of the hot and cold terminals. Thermoelectric devices are still heat engines essentially, and their maximum energy conversion efficiency is limited by the Carnot efficiency [26].

2.2. Electrochemical Thermogalvanic Effect

The electrochemical thermogalvanic effect consists of two main processes: 1) the oxidation-reduction reaction occurring at the electrode surface and 2) the electrolyte migration. The conversion efficiency of a thermoelectrochemical cell is intimately related to the ZT value, hence the Seebeck coefficient and electrical conductivity can be increased or the thermal conductivity can be decreased according to Eq. 1.6. The magnitude of S is determined by the thermal power of the redox couple of materials in the electrolyte; the magnitude of σ is dependent on the resistance of the redox reaction occurring at the electrode surface and the transport resistance of the electrolyte, whereas the magnitude of κ is related to both the thermal conductivity in the presence of a temperature difference and the convection of the electrolyte [27,30].

S is dependent on the solvation-structure entropy difference (∆S) and concentration difference (∆C_r) between redox substances. The absolute value of the charge of the redox substance in the electrolyte and the type of solvent and solute surrounding it deeply affect the magnitude of ∆S. Among the studies in the liquid thermocells system [Fe(CN)₆]^3-/[Fe(CN)₆]^4-, Fe²⁺/Fe³⁺, and I¯/I₃¯ are the most interesting. In general, redox couple with large absolute charge values and complicated complex structures possesses large ∆S [28,31,32].

For Fe^2+/3+ redox ions, their anions with different coordination sites have a bigger effect on the Seebeck coefficient. Kyunggu et al. have specifically investigated the effect of anions on the Seebeck coefficient of three common iron salts: Fe₂(SO₄)₃/FeSO₄, FeCl₃/FeCl₂, and Fe₂(ClO₄)₃/FeClO₄. The excellent performance of Fe^2+/3+ perchlorate is attributed to the uncoordinated nature of its perchlorate anion, which inhibits the reduction of S and prevents the formation of ionic couples [33]. Fe(CN)₆]^3-/[Fe(CN)₆]^4- is the redox couple that has achieved the highest thermal power to date, and it remains possible to alter the solvent environment of the ions to increase the Seebeck coefficient. Kim et al. reported that with the addition of an organic solvent with appropriate solubility parameters to the aqueous electrolyte of [Fe(CN)₆]^3-/[Fe(CN)₆]^4-, the electrochemical thermopower can be more than doubled to 2.9 mV K^-1. The addition organic solvent results in a noticeable rearrangement of the solvation shells, which in turn leads to an increase in the entropy change of the whole redox system, thereby increasing the electrochemical thermopower [34]. Prediction of heat power (i.e., thermoelectric temperature coefficient) with molecular dynamics simulations could allow for simpler and more convenient optimization of redox couples. Chen et al. noticed the S of Fe^2+/3+ can reach 3.8 ± 0.5 mV K^-1 in a mixture of acetone-water solvent with molecular dynamics simulation, which matches the experimental value. The discovery provided insight into the design of solvation shell sequences to develop electrolytes with high S. Apart from changing the solvent environment, the addition of other additives which modify the redox ion hydration shell to optimize the Seebeck coefficients is commonly employed [35]. Duan et al. introduced guanidine salt with high ionic sequence and amide derivative urea with high polarity into Fe(CN)₆^4-/3- aqueous solution, and their synergistic effect resulted in the enhancement of the Seebeck coefficient of Fe(CN)₆^4-/3- from 1.4 mV K^-1 to 4.2 mV K^-1 and the growth of power density from 0.4 mW K^-2 m^-2 to 1.1 mW K^-2 m^-2. Guanidine salts are one of the highest cationic salts in the chaotropic sequence which can destabilize non-covalent bonding forces or destroy the structure of macromolecular proteins [36].

Besides increasing ∆S, an alternate way to increase S_e is to increase ∆C_r [37]. However, redox couples cannot permanently maintain a state of concentration difference between the hot and cold ends. Because the concentration difference state is unstable from a thermodynamic perspective, it spontaneously decays to a homogeneous state. The ∆C_r equals zero while the electrolyte is in a stable state. Zhou et al. exploited the temperature-sensitive properties of cyclodextrins and the host-guest interaction with I^3- to create an I^¯/I^3- concentration difference between the hot and cold ends, resulting in an increase in the Seebeck coefficient from 0.86 mV K^-1 to 1.97 mV K^-1. Figure 5 shows that at the cold terminal, the hydrophobic property of the inner ring of α-CD is exploited to form an α-CD-I^3- complex by combining with the also hydrophobic I^3-, which prevents I^3- ions from participating in the reaction and decreases the concentration at the cold terminal. However, the α-CD-I^3- composite has a temperature-sensitive property and releases I^3- ions upon dissolution at the hot terminal, consequently resulting in a different concentration level of I^3- at the hot and cold terminals, increasing the Seebeck coefficients [38]. Yu et al. employed guanidine salts and Fe(CN)₆^4- to form thermosensitive crystals that reduced the concentration of Fe(CN)₆^4- at the cold terminal and resolved at the hot terminal, with no effect on the rate of the redox reaction. This results in the formation of a continuous concentration gradient in the solution, which increases the Seebeck coefficient from 1.4 mV K^-1 to 3.73 mV K^-1. Meanwhile, the solid crystals formed also effectively suppress the thermal conductivity of the liquid, and ultimately increase the relative Carnot efficiency to 11% [39]. Furthermore, concentration theory may also be applied to change the sign of the Seebeck coefficient, i.e., to change the type of reaction that occurs at the hot and cold terminals. Duan et al. achieved an increase in the absolute value of the Seebeck coefficient of the I¯/I^3- ion pair and a change in the sign of the Seebeck coefficient from a p-type to an n-type thermocell via the incorporation of a temperature-sensitive nano-microgel (PNIPAM) into an aqueous solution of I¯/I^3-. PNIPAM has a hydrophilic to hydrophobic phase transition at around 32 ℃, which also changes the gel polymer chain backbone from stretching to condensation and controls the equilibrium of the I¯/I^3- redox reaction. The hydrophobic phase of PNIPAM dominates at the hot terminal, and the I^3- ion combines with PNIPAM due to the hydrophobic effect, thus the concentration of I^3- ion decreases at the hot terminal, and the oxidation reaction of conversion from I^3- to I^- occurs at the hot terminal. The hydrophilic phase of PNIPAM dominates the backbone stretching at the cold terminal, PNIPAM-I^3- releases I^3- ions, which leads to the reduction reaction of I¯ to I^3- conversion at the cold terminal, thus changing the original I¯/I^3- redox direction and altering the sign of the Seebeck coefficient. The concentration difference constructed in this manner resulted in a higher absolute Seebeck value, from 0.71 mV K^-1 to -1.91 mV K^-1 [40]. Concentration difference effects focus on regulating the Seebeck coefficient, which requires a specific ion in the redox couple to combine with the additive to form a temperature-sensitive substance in order to enable the formation of concentration difference effects of ions at the hot and cold terminals of the electrodes. This effect has a great ability to regulate the Seebeck coefficient and can also change the direction of the redox reaction, however, the resulting conjugates may influence the rate of the redox reaction resulting in irreversible side-reactions during the thermal cell cycling, which causes a decrease in the cycling performance and finally an attenuation of the output power [27,41,42,43]

2.3. Quasi-Solid-State Electrolyte

In practical application of TECs, the leakage issue of liquid batteries and non-portability are significant hindrances to the process of adoption. Therefore, quasi-solid electrolytes are gradually attracting attention due to their advantages of self-encapsulation, leakage prevention, and flexible characteristics. Meanwhile, thermal convection is ignored in gel electrolytes by the property of the quasi-solid state, hence thermal conduction becomes the dominant form of heat transfer in gel electrolytes, which reduces the thermal conductivity significantly [44,45,46,47,48].

PVA is a promising candidate for quasi-solid electrolyte substrates due to its biocompatibility, non-toxicity, non-corrosivity, and excellent water solubility. Zhou et al. synthesized a PVA-FeCl_2/3 gel electrolyte with high mechanical strength and low charge transfer resistance, which uses HCl as a supporting electrolyte, with a S_e of 0.8 ± 0.02 mV K⁻¹, current density of 16.1 A m^{− 2} and power density of 63.7 mW m^{− 2} at a ΔT of 20 K [49]. Liu et al. combined stretch-induced crystallization with the thermoelectric effect to present a high-strength quasi-solid stretchable PVA thermoelectric thermocouple (SPTC) with a tensile strength of 19 MPa and a thermopower of 6.5 mV K^-1. The SPTC has a high tensile strength of 1300%, an ultra-high toughness of 163.4 MJ m^-3, and an output power density of up to 1969 μW m ^-2 K^-2 [50]. Gao et al. designed PVA anisotropic polymer networks to produce aligned channels for ionic conduction and hierarchically assembled crystalline nanoprotofibres for crack passivation. The ionic conductivity of the anisotropic thermocouples was increased by more than 400% and the power density was comparable to that recorded for state-of-the-art quasi-solid thermocouples. Furthermore, compared to available quasi-solid thermocouples with the best mechanical properties, the material achieves bionic strain stiffness, with toughness and strength improved by more than 1100% and 300%, respectively [51].

In addition to PVA, gelatin, cellulose, agar, sodium polyacrylate, polyacrylamide, and other polymers are also promising materials for composing hydrogel matrices, which possess more abundant functional groups and may offer more possibilities for optimizing the performance of the TECs. Chen et al. proposed a flexible quasi-solid-state TEC via the rational design of a hydrogel electrolyte, which simultaneously modulates the thermoelectric effect and mechanical robustness by redox-coupled multivalent ions (Figure 6) [52].

Chen et al. also introduced high electrochemical potential of redox couple (Sn⁴⁺/Sn²⁺) into a flexible and stretchable (within a strain of 100%) composite hydrogel (polyacrylamide/acidified SWCNTs) and constructed a gel-state with a large and stable S_e of 1.59 ± 0.07 mV K^-1. The strain sensitivity originated from the well-dispersed acidified SWCNTs network and polyacrylamide hydrogel matrix endows the TEC with additional role as a self-powered strain sensor for monitoring various human motions relating to finger, wrist, and elbow [53]. The introduction of redox couple not only provides the hydrogel with excellent thermoelectric conversion capability but also acts as ionic cross-linking agents to generate double cross-linking structures, which result in the formation of reversible bonds for effective energy dissipation. With a high Seebeck coefficient of 1.43 mV K^-1 and a significantly improved fracture toughness of 3555 J m^-2, the optimized TECs are able to maintain stable thermoelectrochemical properties under various harsh mechanical stimuli [52]. Furthermore, Chen et al. designed an aqueous eutectic gel electrolyte based on a concentrated lithium bis(trifluoromethane) sulfonimide (LiTFSI) solution, which can be used to achieve freezing resistance and self-humidification capability by regulating the hydrophobicity in the hydrogel. It also exhibits long-term environmental stability without the requirement for encapsulation or packaging. Hydrogel electrolyte collision properties can inhibit ice crystallization, and molecular dynamics simulations suggest that the strong coordination effect between lithium ions and water molecules across a wide range of temperatures is an important potential mechanism [54].

2.3. Electrode

In general, Pt has been employed as an electrode to maintain the simplicity and reverse ability of the redox reaction in the HCF electrolyte (hexacyanoferrates), however, Pt remains rare and expensive, which prevents the commercialization of TECs. Some non-precious metals, including copper, nickel, tungsten, and stainless steel, have also been utilized as materials for electrodes [19,32,55,56]. Carbon is a widely available material and therefore carbon electrodes have attracted widespread attention recently, especially nanostructured carbon materials which typically have high electrical conductivity, rapid redox kinetics, large electrochemically active surface area (ESA), and energetic behavior in relation to HCF redox couples [57,58,59,60]. However, the manufacturing process for these carbon-based electrodes is complicated, not to mention the inherent hydrophobicity which hinders ion transport and thus prevents further performance improvements. Hence, a simplification of the fabrication process of carbon-based electrodes is necessary to make them easily scalable and to improve their hydrophilicity for better penetration into the electrolyte [61].

MXenes, a two-dimensional transition metal carbide and nitride material, have emerged to provide a promising approach to the construction of high-performance TEC electrodes. The high conductivity and hydrophilicity of Mxenes allow for high electron transport speeds and mass transfer. Meanwhile, the reduction activity of the transition metal atoms on the surface is more favorable for electrochemical processes, with a large surface area providing a high ESA. The layered architecture of MXenes, which facilitates the insertion of molecules and ions, is beneficial for the regulation of properties and the assembly of multilayers [62,63,64,65].

Wei et al. constructed flexible thin-film electrodes with ternary composites of Ti₃C₂T_x, polyaniline (PANI), and single-walled carbon nanotubes, which showed significantly enhanced thermoelectrochemical properties compared to the widely used precious metal platinum electrodes. A porous layered structure with a large electrochemically active surface area was formed in the ternary composite electrode. Results of experiments and simulations indicate that the synergetic effect of Ti₃C₂T_x and PANI promotes the mass and charge transport at the electrolyte-electrode interface, generating a TEC with an output power of 13.15 µW cm^-2 at a ΔT of 40 K. TEC can also respond rapidly to minute temperature differences between human bodies and the environment, indicating that it has great potential for harvesting low-grade heat to power small electronic devices [61].

3. Device Integration and Applications

Besides enhancing the efficiency of a single cell, a further research focus is on device integration and applications. The main integration methods are Z-shaped, a serrated connection of identical single cells in series, and Π-shaped, a combination of p-type and n-type cells in series (Figure 7). For thermocouples where oxidation reaction occurs at the hot electrode are typically defined as p-type, whereas thermocouples in which reduction reaction proceeds at the hot electrode are defined as n-type. The Z-shaped connected device has the advantage of achieving maximum power output with a single state-of-the-art cell [40,66,67].

Shi et al. designed a fatigue-resistant and highly conductive hydrogel thermocouple with photothermal conversion capability for non-contact self-powered applications. At a temperature difference of 20 K, the output voltage rises from ≈ 0.05 V to 0.85 V when individual thermocouples are assembled into an array of 20 cells [21]. However, the device integration process using Z-shaped connections is complex with a large contact resistance between the electrodes and the Z-shaped wires. The Π-shaped connection, on the other hand, simplifies the integration process and enables good contact between the electrodes and the collector. Xu et al. designed a p-n pair hydrogel electrolyte by choosing Fe(ClO₄)₃/Fe(ClO₄)₂ as the n-type ion couple. By integrating and fabricating a conformal portable thermal battery device, 14 pairs of p-n-connected cells achieved an output voltage of 0.16 V at ΔT = 4.1 K [68]. Unfortunately, integrated devices via Π-shaped connections remain inefficient for a lack of high-performance n-type batteries.

4. Summary and Outlook

TECs represent an emerging technology for thermoelectric conversion which has attracted extensive attention from researchers in both academia and industries. Despite the significant advances in TECs which have been achieved via the optimization of electrolytes, electrodes, insulation materials, and device modules, numerous challenges remain to be overcome.

Regarding the electrolyte, the modification of existing redox pairs or the development of new ionic pairs, the alteration of the solvent environment and the addition of additives to form a concentration difference or the destruction of the ionic solvent shells, in addition to the use of quasi-solid electrolytes, have all contributed to the great enhancement of the Seebeck coefficient of the thermal batteries or the output power. However, the following shortcomings still exist and require further development and optimization:

The interaction of redox ions with solvent molecules is one of the important factors determining the Seebeck coefficient. However, none of the relevant solvent parameters exhibit regularity in the effects on the ionic Seebeck coefficients, therefore the specific mechanism of the solvent molecules’ influence on the redox ions has not been clarified, with a lack of exploration into in-depth mechanism of exactly how the solvent affects the Seebeck coefficients.

Integrated device reliability is also one of the factors to be considered. In general, thermoelectric devices require the integration of multiple p-type and n-type TECs to obtain a stable voltage output. Although existing p-type TECs (e.g., the [Fe(CN)₆]^3-/[Fe(CN)₆]^4- system) have been relatively sophisticated, the development of corresponding n-types are still unsatisfactory. Further research on high-performance n-type redox couple is critical for the optimization of devices for TECs. For example, perchlorate redox pair (Fe²⁺/Fe³⁺) exhibited a high Seebeck coefficient of 1.76 mV K^-1 and high solubility (> 1 M) in aqueous electrolytes [33,34]. In addition, another focus is on the stability of TEC devices in extreme environment. The effects of cryogenic and high temperatures on material properties should be taken into account in the design of TECs, the widening of operating temperature range of TECs in particular should be one more focus.

The application of flexible thermoelectric devices for thermal batteries in wearable devices lacks extensive research, especially the adaptation of flexible electrodes and gel electrolytes, the improvement of the electrodes for the qualities of gel electrolytes, while the comprehensive performance of the thermoelectric devices after their integration remains to be explored.

It is suggested that the optimization of TECs, as well as large-scale integrated devices should consider the balance between performance and cost, thus ensuring a cost-effective design of TECs in which the low-grade heat harvested can be converted to electrical energy in a commercially accepted mode.