Human body heat-driven thermoelectric generators as a sustainable power supply for wearable electronic devices: Recent advances, challenges, and future perspectives

Abstract

Thermoelectric generators are devices that directly convert heat flux or the temperature difference between two hot and cold surfaces into electricity. With the advancement of the Internet of Things (IoT) technology and the development of wearable and portable devices, the issue of providing a sustainable power source is one of the main challenges in the development path of these tools. Creating electric power by harvesting the waste heat from the human body is one of the effective solutions in this way. For this reason, the development and improvement of the technology of wearable thermoelectric generators have received much attention recently. Due to the low-temperature difference between the two sides of wearable thermoelectric generators and the high thermal resistance between the skin and the heated surface of these modules, the performance of these systems is highly dependent on their structural parameters and environmental factors. In this paper, it has tried to review all the previous studies regarding the impact of structural factors (such as the matching of internal and external thermal resistances, geometrical parameters of the module, design of heat source and sink, and flexibility of thermoelectric module) and environmental parameters (including the effect of ambient air temperature and humidity, skin temperature, and the interaction of power consumers with thermoelectric modules). Based on the studies, it seems that in optimizing the performance of wearable thermoelectric generators (WTEGs), it is necessary to consider the effect of the human body's thermoregulatory responses, such as skin temperature and sweating rate. The change in skin temperature directly affects the performance of WTEGs, and the change in sweating rate can also affect the thermal resistance between the skin and the hot plate and overshadow the matching of thermal resistances during operation.

Nomenclature

$\mathrm{C}\mathrm{O}\mathrm{M}\mathrm{S}\mathrm{O}\mathrm{L}$

A software for Multiphysics simulation

$\mathrm{B}\mathrm{i}\mathrm{T}\mathrm{e}$

Bismuth Tellurium Composition

$\mathrm{F}\mathrm{P}\mathrm{C}\mathrm{B}$

Flexible printed circuit board

$\mathrm{F}\mathrm{T}\mathrm{E}\mathrm{G}$

Flexible thermoelectric generator

$\mathrm{I}\mathrm{o}\mathrm{T}$

Internet of Things

$\mathrm{L}\mathrm{C}\mathrm{D}$

Liquid crystal display

$\mathrm{P}\mathrm{A}\mathrm{N}\mathrm{I}$

Polyaniline

$\mathrm{P}\mathrm{C}\mathrm{M}$

Phase change material

$\mathrm{P}\mathrm{D}\mathrm{M}\mathrm{S}$

Poly dimethyl siloxane

$\mathrm{P}\mathrm{I}$

Polyimide

$\mathrm{P}\mathrm{V}\mathrm{D}\mathrm{F}$

Polyvinylidene fluoride

$\mathrm{T}\mathrm{E}$

Thermoelectric

$\mathrm{T}\mathrm{E}\mathrm{G}$

Thermoelectric generator

$\mathrm{W}\mathrm{S}\mathrm{N}$

Wireless sensor network

$\mathrm{W}\mathrm{T}\mathrm{E}\mathrm{G}$

Wearable thermoelectric generator

1. Introduction

With the rapid development of technology, numerous electronic devices are used for different applications, from various industrial systems and distant communications to health care monitoring systems and entertainment [1,2]. All mentioned applications have raised the demand for usage of the notion “Internet of Things” (IoT) in data exchange between users and electronic devices via the internet. According to remarkable advances in the concept of IoTs, and the advent of Wireless Sensor Networks (WSNs), various types of smart gadgets and autonomous wearable devices have gained substantial attention [3]. Wearable devices are widely used in different areas, such as personal health monitoring systems, electronic sports sensors, smartwatches, and clothing [4]. The main concern in evolving IoT technology embedded in smart wearable devices is to find a sustainable power supply. Three different solutions have been proposed so far to overcome this obstacle. The first is to achieve an enhanced design with a higher battery capacity. The second is replacing batteries with a permanent power supply, and the third is optimizing wearable devices so that they need less input power [2]. Batteries cannot be a suitable power source for new generations of wearable gadgets due to their many practical limitations [2,4,5]. In addition to being bulky, batteries have a limited capacity and often either must be charged periodically or replaced after a short period of operation [6,7]. Limitations on the use of batteries are especially more evident in health care monitoring systems and implanted medical devices [4,8]. In these cases, the batteries must be able to provide the conditions of continuous and uninterrupted operation of the mentioned systems. Because the power consumption of wearable devices is often in the range of microwatt to milliwatt [3], the use of heat dissipated from the human body as a source of electrical power required by wearable devices has attracted the researchers’ interest. It is anticipated that replacing bulky, low-capacity conventional batteries with body-heat-driven micro-generators could pave the way for further development of smart wearable technologies.

In recent years, various approaches have been introduced to extract energy from the human body. Some of these methods exploit physical body motions, and some of them benefit from body heat dissipation [4,5]. Available body energy from daily activities can be converted to electricity utilizing piezoelectric devices [5], electrostatic-based harvesters [5,9], electromagnetic generators [10,11], and triboelectric generators [12].

The skin temperature is generally higher than the ambient temperature. Depending on the amount of physical activity, about 60 to 180 W of heat generated by metabolism in the body is exchanged with the surrounding environment through heat convection and radiation. The waste heat recovery from the human body can be a reliable way to produce electric power for supplying wearable devices. This can be done through various methods such as the conversion of infrared emissions from the human body to electricity [13] and methods based on four scientific phenomena, i.e., the electrokinetic effect [14], pyroelectricity [15], thermoelectricity (Seebeck effect) [16] and ionic thermoelectricity (Soret effect) [17].

This article aims to review the latest research on thermoelectric generators (TEGs) and, specifically, wearable thermoelectric generators (WTEGs) and the parameters affecting their structure and performance. So far, several great review articles have been written about this topic, but all of them have focused on the issues of materials, geometry, and configuration of thermoelectric legs, and finally, the mutual effects of these factors. In this study, in addition to investigating the effects of the factors mentioned above, it has been tried to investigate the effects of factors such as environmental conditions, the matching of internal and external thermal resistance of thermoelectrics. Morover, the effect of changes in skin temperature and the body's thermophysiological responses to environmental temperature stimuli has been investigated on the performance of wearable thermoelectric generators. In the end, the challenges of the existing studies are reviewed and the future perspectives in the development of wearable thermoelectrics are drawn. Investigating the importance of the mutual effects of environmental, individual, and physiological parameters on the correct choice of material and the precise design of the geometrical structure of thermoelectrics is one of the most significant achievements of the present study, which distinguishes it from other existing review articles.

2. Research strategy

By employing search and identification strategies, a total of 3500 articles were selected by searching related keywords and their Boolean combinations in Scopus and Web of Science databases. Then, in the screening step, the articles that were less related to the article's objectives were discarded. In the next stage, by studying the summary and conclusion sections of the articles and applying the eligibility criteria, 354 articles that were more in harmony with the responsibility of the present research remained. Out of this number of articles, 248 of them were removed for not answering the main questions of the research. Finally, 106 articles were included in the background of the research. The PRISMA diagram of this research review strategy was shown in Fig. 1.

Fig. 1.

The PRISMA diagram of the procedures for bibliographic research.

It is necessary to highlight that this manuscript was designed as a review paper, and no experimental tests were performed on human samples and participants by the authors.

3. Methods of human body energy harvesting

Body energy harvesters are often divided into active and passive methods depending on their work principles [4]. Active strategies usually generate electricity from harnessed body motions during daily activities such as running, walking, and even chest movement due to breathing. In contrast, passive methods use the human body's heat loss to generate the power required by wearable devices. A summary of the available methods for harvesting energy from the human body is presented in Table 1.

Table 1.

Different methods for energy harvesting from human body.

	Type	Category	Strategy
Human body energy harvesting	Active (Kinetic Energy)	contact	Piezoelectric
		contact	Electrostatic
		contact	Electromagnetic
		contact	Triboelectric
	Passive (Thermal Energy)	non-contact	IR-Harvesting
		contact	Electrokinetic
		contact	Pyroelectric
		contact	Thermoelectric

3.1. Active human body energy harvesters

Active energy harvesters are defined as systems that require external mechanical stimuli to generate electricity [4,18]. Walking, running, bending, and stretching are examples of intentional external mechanical stimuli in the body that can be used to generate electricity. Even involuntary movements of the body, such as movements caused by breathing [19] or blood flow [20], can be used to stimulate active systems and generate electricity [18].

The active human body energy harvesters include generators based on piezoelectric [5,21,22], electrostatic [22,23], electromagnetic [5,22,23], and triboelectric [18,24] effects. In the following, each method's structural advantages and disadvantages are examined.

In addition to generating potential difference and high electric power, piezoelectric generators do not need an initial electric potential difference to start up; However, despite the structure's scalability and simplicity, combining these generators with other types is difficult. Also, the production of low electric current and spontaneous electric discharge in low-frequency external stimulations are among the disadvantages of these types of generators.

Generators based on the electrostatic effect can produce significant potential difference and electric power density using low-frequency external excitations, but they need initial electric potential difference to start working. The electric current produced by these tools is also relatively low. Also, the dielectric breakdown phenomenon is one of the other disadvantages of these generators.

Electromagnetic generators do not need an initial electric potential difference to start working and can produce a high electric current. However, on the other hand, when they receive low-frequency external stimuli, they produce a slight electric potential difference. These generators have a complex structure and combining them with other generators is difficult.

By using low-frequency external excitations, triboelectric generators produce significant electrical power density, and at the same time, they have high energy conversion efficiency and flexibility; But, in their nature, there are problems such as low electric current generation, low durability, and the difficulty of integrating them with other generators.

In addition to the structural disadvantages mentioned for the above generators, the irregular and random movement of body parts [2] as the external driving force of these wearable generators harms their output stability. In addition, the low frequency of body movements [2] causes a decrease in the performance of piezoelectric and electromagnetic generators. On the other hand, if the user is unable to move for any reason, including old age, disability, or illness, these generators will not be able to produce electricity. Also, the output power of generators that use involuntary body movements such as breathing is insignificant. Therefore, in general, the use of generators based on the operating mechanism to provide the electrical power required by smart wearable devices, especially in medical care, has received less attention.

The advantages and disadvantages of the above wearable generators are summarized in Table 2.

Table 2.

Advantages and disadvantages of generators based on the active mechanism.

Generator type	Pros	Cons
Piezoelectric	•Production of high electric potential difference [5] •High electric power production [21] •No need to initial potential difference to start working [5] •The simplicity of structure and scalability [21]	•Spontaneous electrical discharge at low frequency of external stimuli [22] •Low-output electric current [22] •High electrical impedance [22] •The difficulty of integration with other generators [5]
Electrostatic	•Production of high potential difference and power density while the dimensions are small [23] •No need to enter high stimulation frequency to start working [23] •Structural simplicity [22]	•The need for the initial potential difference to start working [22] •Low-output electric current [22] •High electrical impedance [22] •The possibility of dielectric breakdown [23]
Electromagnetic	•No need for the initial potential difference to start working [5] •High-output electric current [22] •Low electrical impedance [22] •Low mechanical damping [5]	•The low output voltage, especially when the frequency of external stimulation is low [22,23] •The complexity of the manufacturing process and miniaturization [22] •The difficulty of integration with other generators [5]
Triboelectric	•Production of high-power density [24] •High energy conversion efficiency [24] •No need to enter high stimulation frequency to start working [18] •Flexibility [18]	•Low-output electric current [18] •The difficulty of integration with other generators [24] •Low durability [24]

3.2. Passive human body energy harvesters

The body core temperature is usually kept constant in a healthy person by thermoregulatory mechanisms. For this purpose, the excess heat produced in the body is transferred through the skin in both forms of sensible and latent heat (sweat evaporation) to the surrounding ambient. Table 3 represents the contribution of each mechanism to the total heat loss from the body during different activities. Passive energy harvesting methods are commonly used to harness the sensible thermal energy lost from the skin. Passive systems can be divided into two categories: contact and non-contact systems.

Table 3.

The contribution of different mechanisms of heat dissipation from the human body to the environment during various activities [25].

Activity	Sensible $\left(\mathrm{W}\right)$	Latent $\left(\mathrm{W}\right)$	Total $\left(\mathrm{W}\right)$
Seated at rest	60	40	100
Sedentary work	65	55	120
Seated eating	75	95	170
Walking at 3 $\left(\mathrm{m}\mathrm{p}\mathrm{h}\right)$	100	205	305
Heavy work	165	300	465
Athletics	185	340	525

3.2.1. The non-contact system

In this category, physical contact between the scavenger and the human body is not required. It has been previously studied that heat loss from the human body through radiation, as the main heat dissipation mechanism from the body at room temperature [26], can be harvested by photovoltaic structures. The human body is an almost ideal infrared emitter source for powering wearable and implanted devices at a skin emissivity of 0.98 for infrared wavelength ranging from 1 to 14 $\mathrm{\mu }\mathrm{m}$ [27]. Ghomian et al. [28] fabricated a device capable of harvesting infrared emission from the human body, generating 2.2 μW from each square centimetre of skin area.

3.2.2. Contact systems

Harvesters in this category require direct contact with human skin since they exploit the dissipated heat from the skin as a heat source by means of a conduction mechanism to generate electricity. Methods employing this approach will be explained bellow.

3.2.2.1. Electrokinetic effect

It has been proved that due to the electrokinetic effect, when water flows in narrow channels in porous carbon structures, a double layer of opposite electrical charges is formed on the interface of the solid structure and the adjacent flowing water, generating electrical voltage [14]. This phenomenon can be a sustainable source of electricity for small wearable gadgets only if water is consistently running within the micro channels of the carbon structure [29]. Creating natural evaporation is an ideal suggestion to drive capillary water flow inside the porous carbon [13]. Liu and co-researchers [30] have presented a wearable nanogenerator bracelet that uses the body's skin temperature to evaporate the capillary deionized water absorbed by the porous carbon. The cold side of the bracelet, which is nearly at the environment temperature, liquefies the water vapor, constructing a closed cycle of natural evaporation. Fig. 2-a and 2-b demonstrate the bracelet design and generated voltage over time while worn on the wrist, respectively.

Fig. 2.

Wearable nanogenerator bracelet: a) design and working principle; b) generated voltage to time while the skin temperature is nearly at 31 $°C$, room temperature is set at 18 $°C$ with a relative humidity of 70% [30], with permission.

3.2.2.2. Pyroelectricity

When exposed to time-dependent temperature fluctuations, the spontaneous polarization of anisotropic crystalline materials is known as the pyroelectric effect [31]. When a nearly uniform temperature distribution condition exists, pyroelectricity is the best choice to exploit time-variant temperature to generate electricity [15,32]. The generated electric current by pyroelectrics is directly related to the rate of temperature fluctuations [33]. Also, the pyroelectric coefficient and Curie temperature1 are influential factors in selecting pyroelectric materials [34]. For example, Potnuru and Tadesse [35], studied two types of PZT2 to choose the best material, owing to higher pyroelectric coefficient and lower Curie temperature, for powering a computer mouse, taking hand palm as the heat source. Kim et al. [36] studied the performance of a honeycomb-shaped pyroelectric harvester on the human body at a temperature of 31 $°C$ and achieved a maximum power of 0.06 and 0.16 $\mathrm{\mu }\mathrm{W}$, at the environmental temperature of 24 $°C$, under wind velocities of 2 and 4 $\mathrm{m}/\mathrm{s}$, respectively.

3.2.2.3. Thermoelectricity

Thermoelectricity is the direct conversion of spatial temperature difference to electricity and vice versa [5]. Thermoelectric generators are solid-state devices based on the Seebeck effect [37]. Advantages such as having a simple structure with no moving parts, and being maintenance-free, durable, and reliable [38], have broadened the applications of TEGs in different areas, especially in wearable devices harvesting wasted heat from the body [39]. Flexibility is a chief factor in wearable TEGs to maintain the wearer's comfort and to create the possibility of maximum heat harvesting from the body [40]. Using flexible TE materials and substrates in WTEGs [41,42] is a common trend. Ionic thermoelectric materials [13], are another choice to fabricate flexible thermoelectric devices in which temperature difference between the hot and cold sides of TE devices roots for thermo-diffusion of ionic charges [43]. Since ion charges cannot pass through electrodes, ionic thermoelectric type is often used as supercapacitors; as a result, their output electric current is transient and reaches zero due to the discharge of supercapacitors [44]. However, a solid-state thermoelectric, which works based on the Seebeck effect, can generate a continuous electric current. Therefore, most thermoelectric devices work on this basis in different applications.

Generally, thermoelectric generators can be classified to macro and micro-TEGs with regards to the size of their geometry and required manufacturing technology. Macro TEGs usually have less than 100 pair of thermoelectric legs with rather large cross section area in their structure. Also, relatively long length of TE legs benefits macro-TEGs in maintaining an adequate temperature difference between their hot and cold terminals. Fabrication technologies which are used in this type TEGs are typical methods such as Quick ohm and Thermal force [45]. On the other hand, the structure of micro thermoelectric generators consists of a high density of thermoelectric legs with relatively small cross section area and short length. With respect to their miniature structure, different micro manufacturing technology (e.g., green TEG [45], thick and thin technology [46], IC technology and MEMS [46], etc.) can be applied in their fabrication process. Micro-TEGs take advantage of light weight, low cost of fabrication, and adjustability in shape and size but due to the small thickness, they have problems with maintaining temperature difference between hot and cold plates. A quantitative comparison among the mentioned characteristics of macro and micro-TEGs are presented in Table 4.

Table 4.

Comparison between the characteristics of macro and micro TEGs [45,46].

Characteristic	Macro TEGs	Micro TEGs
TC number/density	Low (< $100\mathrm{c}{\mathrm{m}}^{-2}$)	High ($\ge 100\mathrm{c}{\mathrm{m}}^{-2}$)
Thermoleg cross section	Large ($\ge 0.2\mathrm{m}{\mathrm{m}}^2$)	Small ($<0.2\mathrm{m}{\mathrm{m}}^2$)
Thermoleg length	Long ($\ge 900\mathrm{\mu }\mathrm{m}$)	Short ($<900\mathrm{\mu }\mathrm{m}$)

3.2.3. Pros and cons of passive energy harvesters

The generators that work based on the passive mechanism do not need body movement as a driving force to produce electric power and only use the waste heat of the body. In this way, if the user remains still, it will not affect the performance of these generators. On the other hand, these generators do not have any moving parts, so they are more durable and usually do not require maintenance. In the following, the advantages and disadvantages of generators that absorb infrared rays from the body and generators that work based on electrokinetic, pyroelectric, and thermoelectric phenomena and in the category of generators based on passive work are given.

Generators that absorb infrared rays from the body can produce a suitable electric potential difference in the room temperature range. However, their efficiency decreases significantly due to the significant decrease in the amount of radiant heat transfer from the body in hot environments. Also, having a complex and expensive structure is another problem for these generators.

The electrokinetic generators have a flexible structure and can produce a high electric potential difference. However, the requirement for producing an electric voltage in these generators is to create a permanent pressure difference along the microchannels in their carbon structure. Also, the amount of electric potential difference produced is proportional to the speed of fluid movement in micro-channels, which limits the performance of these generators.

Pyroelectric generators have a higher energy conversion efficiency than other generators based on passive mechanisms. However, the dependence of the amount of potential difference and electric current produced by these generators on the rate of temperature changes with high oscillation frequency is one of the disadvantages of these generators.

Ionic thermoelectric generators can produce a significant electric potential difference, and because their internal thermal resistance is relatively high, the temperature difference between the electrodes is maintained for a long time. However, the high value of electric resistance and the inability to produce continuous electric current are the most important disadvantages of these generators. Also, some characteristics of common ion conductors (such as Silver Nafion & Silver Polystyrene-sulfonate), which are used in the structure of these generators, limit their use. The low thermoelectric property in electrolyte solutions (e.g., 80 mmol O₂Te and 20–60 mmol Bi₂O₃ in 2 M Hno3 (65%)) and ionic liquids (e.g., 1-Ethyl-3-methyl-imidazolium-dicyanamide) and the complex and expensive structure of ionic gels are some examples in this regard [49,50]. In contrast to these generators, there are thermoelectric generators based on the Seebeck effect, which have low electrical resistance and can produce continuous electrical current. Also, they have a simple and durable structure that can be easily integrated with other generators. However, low thermal resistance and fragility of semiconductors that are conventionally used in the structure of these generators are considered as their disadvantages. Finally, by comparing the advantages and disadvantages of these two types of thermoelectric generators, it can be said that thermoelectric generators based on the Seebeck effect, due to their considerable advantages, are a more suitable choice for providing electrical power for smart wearable devices. Therefore, leaving aside ion thermoelectric generators, from this part onwards, thermoelectric generators based on the Seebeck effect are called thermoelectric generators for brevity.

The advantages and disadvantages of the passive generators are summarized in Table 5.

Table 5.

Advantages and disadvantages of generators based on the passive mechanism.

Generator type	Pros	Cons
IR generators	•Production of high electric potential difference around room temperature [26,28] •Continuous and continuous power generation [28]	•Low energy conversion efficiency [28] •Lack of high efficiency in hot environments (temperature above 33 $°C$) [26,28] •Complex and expensive structure [28]
Electrokinetic	•Ability to produce high electric potential difference [30] •Flexibility [47]	•The need for a pressure difference along narrow carbon channels to produce an electric potential difference [48] •Dependence of the produced electric potential difference on the fluid movement speed in the cavities of the porous carbon structure [48]
Pyroelectric	•High relative energy conversion efficiency [36] •Long life and high durability [36] •Suitable for working in adverse weather [36]	•Dependence of potential difference and produced electric current on temperature fluctuation rate [15,33] •The need for a high-frequency temperature fluctuation rate [49]
Ionic thermoelectric	•Production of high electric potential difference [50] •High thermal resistance in the structure [50] •High flexibility [50]	•Transient output electric current [51] •High structural and electrical resistance [50] •Low thermoelectric properties in some ionic conductors used in the structure of these generators, such as electrolyte solutions and ionic liquids [50] •It is complicated and expensive to make these generators with ion gels [50]
Thermoelectric	•Continuous electric current generation [50] •Low electrical resistance [50] •Ability to integrate with other generators [5,13,52] •Scalability [4] •The simplicity of the structure [38] •High stability and durability [38] •Suitable for working in adverse weather [3]	•Decrease in temperature difference between hot and cold plates due to low thermal resistance [50] •The fragility of some conventional thermoelectric materials [36]

In general, by examining the advantages and disadvantages of the above wearable generators, which are more fully classified in Table 5, It can be concluded that the generators absorbing infrared rays and generators based on the electrokinetic effect are not capable of producing stable electric power. Also, since the body temperature changes during the day and night are around 1 $°C$ [5], pyroelectric generators can produce electric power for a limited number of wearable smart devices. However, due to the constant temperature difference between the body and the environment, thermoelectric generators can produce considerable electrical power throughout the day and night. This advantage and other advantages have caused them to be chosen as an ideal option, among the other mentioned generators, to provide electrical power for smart wearable devices and equipment.

Body motions are generally random, irregular, and low in frequency which hinder the requirement of active harvesters for magnification of external oscillation [2]. Also, they hold conversion efficiency problems in generating the requisite output voltage from low-exerted external vibrations. On the other hand, if the user of an active wearable harvester is in a stationary state for any reason, i.e., being old, handicapped, or suffering from illnesses, these generators cannot generate electricity. Active harvesters that take advantage of involuntary body motions, e.g., breathing, cannot generate enough power for wearable devices.

Passive harvesters, evaporative-driven generators, and body IR harvesters can only produce power in the range of nanowatts to a few microwatts, respectively, which is too low for many wearable gadgets. Pyroelectrics utilize time-dependent temperature changes to generate electricity. The fact that human body temperature varies by about 1 $°C$ during a day makes pyroelectrics power generation limited [5]. Thermoelectric generators have gained plain attention due to their superiorities against other methods. They are independent of body movement and capable of generating electricity when the user is motionless. While being capable of generating the required power for many wearable devices and miniaturized smart gadgets [8] at room temperature, thermoelectric generators are known to be noiseless, environment friendly, and sustainable. Alongside long operating lifetime, they consist of no moving parts [5,8] and can work under harsh weather [3]. Another important advantage of TEGs is the possibility of being integrated with other harvesters, creating hybrid generators [13]. For example, a thermoelectric-photovoltaic energy harvester was presented in Ref. [53].

Due the above-mentioned advantages for TEGs as energy supplies for wearable devices, it is predicted that by further progress in the technology of smart wearables and textiles, the investment in the wearables global market will reach $9.4 B by 2024 [4]; which is drawing even more extensive attention toward WTEGs.

4. Wearable thermoelectric generators (WTEGs) for body heat energy harvesting

Since the late 1990s, much research has been carried out on heat harvesting from the body employing wearable thermoelectric generators to produce the required electric power for low-consumption wearable devices, equipment, and sensors. In the first part of this section, the history behind the advent of TEGs in wearable applications has been studied. Then, the conducted research on influential factors and parameters in the performance of WTEGs has been investigated. After that, studies on implementing these generators in various wearable applications have been reviewed, and finally, several review papers on different thermoelectric materials, configurations of WTEGs, and their applications have been mentioned.

4.1. The advent of thermoelectric generators in wearable applications

Power generation from dissipated body heat via WTEGs started with wrist watches (Fig. 3-a). Seiko thermic wristwatch (Fig. 3-b), providing its power from dissipated body heat, was the first TEG-integrated product to make the path to market in 1998. The operating TEG in the wristwatch could generate 25 $\mathrm{\mu }\mathrm{W}$ over a 1.5 $°C$ temperature difference [54].

Fig. 3.

Integrating a TEG with a wristwatch, a) the schematic idea, b) the picture of the first commercialized wristwatch by Seiko from Ref. [4], with permission.

One year later, in 1999, Citizen Watch Co., Ltd. commercialized its first wristwatch (Fig. 4-a [55]), which was an integrated TEG-powered. The wristwatch was comprised of 1242 pairs of P and N-type Bi₂Te₃ (Bismuth-Telluride) thermoelectric legs in a 7.5 × 7.5 $\mathrm{m}{\mathrm{m}}^2$ module (Fig. 4-b [51] generating 14 $\mathrm{\mu }\mathrm{W}$ power and 640 $\mathrm{m}\mathrm{V}$ open circuit voltage under 1 $°C$ temperature difference [51]. In another Seiko thermic wristwatch, an integrated TEG with 1040 thermoelectric legs was used which could produce an electric voltage of 200 $\mathrm{m}\mathrm{V}$ under 1 $°C$ temperature difference [51].

Fig. 4.

a) A picture of CITIZEN first commercialized wristwatch with an integrated TEG [55], and b) a picture from its Bismuth-Telluride TE legs [51], all with permission.

The successful implementation of thermoelectric generators in wristwatches made researchers seek other possible wearable applications. Therefore, research groups tried to identify the influential factors on the performance of wearable thermoelectric generators and optimized them for a specific application.

4.2. Influential parameters on the performance of WTEGs

The quantity of produced power in a thermoelectric generator is directly dependent on the temperature difference between its hot and cold plates, and the temperature difference between the human body and the environment is limited. Alongside introducing thermoelectric materials with a higher figure of merit, many researchers started to identify the structural parameters in wearable thermoelectric generators and external factors that play a substantial role in the performance of these generators. The structural parameters of WTEGs include finding matched values for thermal resistance [53], geometric parameters [56], hot plate and heat sinks design and parameters affecting flexibility (e.g., flexible materials for legs and electrodes and designs integrated with fabrics and cloth) of these generators. The external factors comprise the boundary conditions on the hot and cold plates of WTEGs and the mutual influence of the power converters connected to these generators. In the following subsections, the conducted many pieces of research on optimizing the mentioned parameters have been reviewed.

4.2.1. Matching of thermal resistance

The temperature difference between the terminals of wearable thermoelectric generators is small owing to the low skin thermal conductivity and high thermal resistance at the interface of skin and the hot plate of the WTEGs and at the interface of the WTEGs heat sink and the ambient. Hence, to maintain the small temperature difference between the hot and cold plates of wearable thermoelectric generators, it is essential to optimize the internal thermal resistance of these generators; in a way that matches the thermal resistance at the contact surface of the skin and the hot plate of the generators, as well as the thermal resistance between the cold plate of the generators and the environment [57]. For this purpose, it is necessary to optimize the geometric parameters of wearable thermoelectric generators so that the internal thermal resistance increases as much as possible to find suitable designs and materials for hot plates and heat sinks of these generators in order to reduce the thermal resistance at the interface of hot and cold plates and to increase the flexibility of thermoelectric generators in order to extract the maximum waste heat from the body.

4.2.2. Geometric parameters

The geometric parameters include the height, shape, and cross-sectional area of thermoelectric generator legs, the number of TE legs, the filling factor (i.e., the occupied planar area by thermoelectric legs divided by the total surface area of the generator), and filling materials between the bases. Here, the conducted studies on these parameters have been investigated.

In 2013, Sahin and Yilbas [58], after analyzing the various cross-sections at fixed heights for thermoelectric generating legs in theoretical research, showed that in order to produce maximum electric power, a rectangular cross-section was the best choice for TE legs in small temperature differences.

Suarez et al. in 2016 [59], presented a quasi-3D model of the structure of a thermoelectric generator in which the effect of heat loss from the substrates and filler materials between the TE legs of the generator was included. Using the presented model, they found the optimal values for the number, cross-sectional area, height of the thermoelectric legs, and the filling factor and chose the appropriate material to fill the space between them. Then, based on the obtained results, they presented a wearable thermoelectric generator with an area of 0.97 ${\mathrm{c}\mathrm{m}}^2$ with legs of 2.2 $\mathrm{m}\mathrm{m}$ height, which produced 20 $\mathrm{\mu }\mathrm{W}$ of electric power when placed on the wrist at room temperature and in an environment without air flow. At last, by comparing the performance of the presented generator with a commercialized thermoelectric generator at four different speeds for air flow (zero, 0.3, 0.6, and 0.9 $\mathrm{m}/\mathrm{s}$), They showed that the output power of the optimized TEG was nearly three times of the obtained output power from the commercialized generator.

In 2017, Lee et al. [60], conducted a research study combining theoretical and experimental work. First, they modelled four thermoelectric generators with filling factor coefficients of 6%, 11%, 21%, and 33%, regarding the fact that filling factor directly influences the internal thermal resistance of TEGs. Then, to investigate the effect of heat sinks, they assumed three finned plates with a different number of fins as heat sinks and studied them in free convection conditions and two other convection conditions with either parallel or impinging air flow. Since the generator with the filling factor of 11% and a 6-fin heat sink had the best performance among the other investigated models, by placing it on the wrist in a temperature difference of 6 $°C$ between the generator plates, they obtained power densities of 6.5 $\mathrm{\mu }\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ in a free convection condition and 20 $\mathrm{\mu }\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ when the air flow was either in parallel or impinging state on the heat sink.

In 2019, Nozariasbmarz et al. [61], investigated the effect of the filling factor on the output of WTEGs. Therefore, they made two thermoelectric generators with 12% and 36% filling factor coefficients with heat sinks (Fig. 5-a). Then, by testing these two generators on the arm, forearm, wrist, and foot (Fig. 5-b), they observed that the highest electric power density was obtained when the generators were placed on the forearm. Fig. 5-c illustrates the comparison of the generated electric power density from these two generators. Other obtained results from this research showed that at an ambient temperature of 17 $°C$, thermoelectric generators with filling factor coefficients of 12% and 36% could produce temperature differences of 2.5 $°C$ and 1.2 $°C$ between the hot and cold plates of the generators, respectively.

Fig. 5.

a) Comparison of the measured power density, b) Location of the measurement points, c) Power output vs. time at forearm position, all from Ref. [61] with permission.

In 2020, Liu et al. [62], implemented a comprehensive 3D modelling of the structure of a thermoelectric generator using the finite element method and validated this model with experimental results. Then, by assuming the temperature of the hot plate and ambient to be constant, they investigated the effects of heat loss from the thermoelectric generator, the thermal conductivity coefficient of the encapsulation materials that held the TE legs and electronic parts of the generator, as well as the height of the thermoelectric legs and the size of its square cross-section on TEG outputs. Finally, they showed that if the optimal values for the mentioned parameters were applied to the construction of the WTEG, under a temperature difference of 15 $°C$, its maximum produced power would reach 169.97 $\mathrm{\mu }\mathrm{W}$, which demonstrated a growth of 598.1% compared to its original design.

Yuan et al. [63], presented a systematic method to achieve an optimal value for power density in flexible thermoelectric generators (FTEGs) considering the number of materials consumed in the structure of an FTEG and its electric matching with smart devices. They used this method to obtain an optimal number of TE legs and the filling factor coefficient; then, relying on the results, they built an FTEG with an optimized structure and Bi₂Te₃-based thermoelectric legs which were placed on a Polyimide (PI) substrate. This generator was able to produce a power density of 3.5 $\mathrm{\mu }\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ in a windless environment. Finally, they utilized this generator in a smart bracelet equipped with multi-purpose sensors (consisting of temperature, humidity, and acceleration sensors), which could show collected and processed data on a liquid crystal display (LCD).

In 2021, Tanwar et al. [64], investigated the effect of structural parameters on the output of WTEGs, which work in limited temperature differences and are used in medical devices. This research simulated a pair of thermoelectric legs using the finite element method. Then, they evaluated the effect of the shape of the TE legs on the generator's output power by comparing three different cross sections with square, hexagonal, and circular shapes of equal area and found that square legs performed better than the other two shapes. Also, according to the construction cost, they found optimal values for the height of the TE legs and the thickness of their connecting materials. In addition, they investigated the effect of the type of material filling the space between the legs. Finally, by inserting the obtained optimal values in a thermoelectric generator model, they achieved electric power of 0.796 $\mathrm{m}\mathrm{W}$ and 3.18 $\mathrm{m}\mathrm{W}$, under temperature differences of 5 $°C$ and 10 $°C$, respectively.

Sathiyamoorthy et al. [65], utilized the finite element method to investigate how the geometry of Bismuth-Telluride (Bi₂Te₃) thermoelectric legs influenced WTEG outputs. Therefore, they considered cylindrical, conical, and rectangular shapes for their thermoelectric generator legs, compared the generator outputs for different leg heights and areas and realized that TE legs of rectangular shape showed the best output in the generator. After that, using the optimized data obtained from the numerical calculations, they built a thermoelectric generator with four pairs of TE legs with a rectangular cross-section, which could produce a voltage of 0.9 $\mathrm{m}\mathrm{V}$ when placed on the wrist at room temperature and under a temperature difference of 1 $°C$.

Considering the effect of filler materials between thermoelectric legs on the internal thermal resistance and flexibility of the thermoelectric generators, in 2021, Jung et al. [66] presented a flexible wearable thermoelectric generator and the space between its rigid legs was filled with polydimethylsiloxane (PDMS) material. In order to reduce the heat transfer coefficient of PDMS material, which is a common material for filling the empty space between thermoelectric legs, they made a porous structure (sponge-like) for this substrate. As a result, its thermal transfer coefficient was reduced to 0.08 $\mathrm{W}/\mathrm{m}\mathrm{K}$. Then, this spongy substrate was injected between the 47 pairs of thermoelectric legs. The manufactured generator (Fig. 6) was able to produce a power density of 130 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ under a temperature difference of 8 $°C$.

Fig. 6.

A picture of the manufactured thermoelectric generator by Jung et al. [66], with permission.

Among other materials used as fillers in wearable thermoelectric generators, silica aerogels [67] and air [68] are more suitable options.

In 2022, Jeong et al. [69], presented a wearable thermoelectric generator consisting of a dual system that produced electric power by harvesting body heat and ambient light in order to overcome the low-temperature difference between the skin and the environment. To optimize the number and cross-sectional area of the thermoelectric legs before manufacturing the generator, they utilized a simulation in COMSOL software and assumed four widths of 0.6, 0.8, 1.2, and 1.4 $\mathrm{m}\mathrm{m}$ for the sides of TE legs; then, they compared the generated power and voltage of the simulated generator. At last, the generator with 18 pairs of thermoelectric legs and a width of 0.8 $\mathrm{m}\mathrm{m}$ showed the best performance. In the next step, they started the process of fabricating their thermoelectric generators with rigid thermoelectric legs and to maintain the flexibility of their structures, they placed the TE legs between two PI sheets and then inserted PDMS material between them. After that, on one side, under the thermoelectric bases, a silicone-based solar light-absorbing material was drawn on the polyamide sheet, which was able to absorb 95% of the ambient light in the ultraviolet to the infrared frequency range. Then, a heat sink with a copper sponge structure was used on top of the thermoelectric legs. In the end, by placing this generator on the arm (Fig. 7), the power densities of 2.19, 8.28, and 15.33 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$, with regards to three different working modes: (1) extracting body heat alone, (2) absorption of ambient light alone, and (3) a state where both systems worked simultaneously, at room temperature, were recorded.

Fig. 7.

A view of the introduced thermoelectric generator in the work of Jeong et al. [69], open access article.

In another research, Fan et al. [70], studied the influence of the height of thermoelectric legs as well as the relative velocity of air on the performance of WTEGs. They presented four thermoelectric generators with different leg heights and removed their ceramic substrate to create a flexible structure and provide space to increase the height of the thermoelectric legs. These generators consisted of 48 pairs of thermoelectric legs with an area of 19 × 12 ${\mathrm{m}\mathrm{m}}^2$ and were placed on the wrist to conduct experiments. They compared the produced open circuit voltages from the neck, temple, wrist, and back of the hand using a generator with a leg height of 3.14 $\mathrm{m}\mathrm{m}$. The results showed that the hand's neck, wrist, temples, and back produced the highest to the lowest open circuit voltage, respectively.

4.2.3. Heat sink design

With regards to the slight temperature difference between the skin and the environment, taking advantage of a proper design and materials [6] in the construction of the hot plate and heat sinks of wearable thermoelectric generators results in generating a more considerable temperature difference between the hot and cold plates of the generator. Therefore, in the fabrication of WTEGs hot plate, flexible structures and materials with high thermal conductivity and low weight must be used to ensure full contact of the wearable generators with the skin surface to harvest the maximum heat from the body and to provide users convenience. At the same time, materials used to fabricate heat sinks and their design must also be thermally conductive and flexible to conquer the high relative thermal resistance between the cold plate of the generators and the environment and to maintain the user's comfort. Heat sinks mounted on wearable thermoelectric generators are divided into heat spreaders, finned heat sinks, and containers of phase change materials (PCM) in terms of design. Heat spreaders and containers of PCM benefit from higher flexibility and less weight; in this sense, they are suitable for wearable applications [4]. The use of the proposed designs can increase the output power up to 50% at low air speeds and up to 100% at a speed of about 2 m per second, depending on the area of the spreader [4]. On the other hand, since the user of the WTEGs is usually in a stationary position, finned heat sinks take advantage of higher heat transfer coefficients compared to the other two categories, and this has increased the application of this type of heat sink. Therefore, many pieces of research have been carried out to optimize the design of blades in terms of shape, height, number, and distance between them when arranged on the cold plate of wearable generators [71,72]. Here, conducted studies focusing on the effect of heat sinks on the performance of WTEGs have been investigated.

Settaluri et al. [73], utilized simple modelling of a WTEG structure with a heat sink, assuming constant temperatures for skin and ambient, to optimize the height of the thermoelectric legs, and the heat sinks with three surface designs of flat, grooved, and checkerboard. They made three thermoelectric generators with Tellurium–Bismuth legs, and copper heat sinks of the three mentioned designs in the form of wristbands. As illustrated in Fig. 8-a and 8-b, they recorded voltages of 108, 94, and 85 $\mathrm{m}\mathrm{V}$ and power densities of 28.5, 21.6, and17.6 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$, in a steady state, for generators with heat sinks with surface designs of flat, grooved, and checkerboard, respectively [71,73].

Fig. 8.

a) Voltages and b) power densities obtained from thermoelectric generators with a heat sink with surface designs of flat, grooved, and checkerboard in the study of Settaluri et al. from Ref. [71], with permission.

In 2018, Shi et al. [74], investigated the effect of a heat sink and its shape on the output of WTEGs. They considered three models of thermoelectric generators, one of which did not have a heat sink, and the other two had copper heat sinks in spongy and finned plate designs. First, they predicted the performance of the three mentioned generator models in various temperature differences by analytical modelling of the internal thermal resistance of the generator. Then, they fabricated these three generators, whose pictures are also given in Fig. 9-a to 9-c, and experimentally obtained their voltage and electric power. They concluded that the thermoelectric generator with a sponge heat sink, concerning the generator's weight, produces more electric power. So, they utilized it to power an accelerometer (2.4 $\mathrm{\mu }\mathrm{W}$) on a wristband (Fig. 9-d).

Fig. 9.

Wearable thermoelectric generator, a) without heat sink, b) with a sponge-shaped heat sink, c) with a finned plate heat sink, d) with a sponge-shaped heat sink connected to an accelerometer, all from Ref. [74] open access article.

In 2018, Kim et al. [75], in an innovative design, introduced a thermoelectric generator with a flexible heat sink made of a superabsorbent polymer material. The cold plate of the generator was cooled through water evaporation. They showed that, by placing this generator on an artificial arm that simulated the temperature of a human arm, the power density of 38 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ at the first 10 min of the experiment and more than 13 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ even after 22 h of continuous testing was produced. They proved that this WTEG could provide the required power of an Electrocardiogram (ECG).

In 2019, Lee et al. [76], designed a flexible heat sink for a WTEG using n-octadecane (C₁₈H₃₈) phase change material (Fig. 10-a). The material used in the heat sink of this generator (Fig. 10-b) was solid at room temperature and by taking heat from the cold plate of the generator, its phase changed from solid to liquid and cooled down the cold plate of the generator. In order to test the performance of presented WTEG, they used an artificial hand, which had a structure similar to human arm, and the temperature of its core was kept at 37 with a heater. So, by placing the generator on this artificial arm, they recorded power density of 20 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ in a period of 33 min at ambient temperature of 25 $°C$.

Fig. 10.

a) A schematic of the presented thermoelectric generator by Lee et al. and b) the structure of its heat sink, reproduced from Ref. [76], with permission.

In 2021, Attar and Albatati [77], presented a simple analytical model on the working principle of thermoelectric generators by assuming the temperature of the skin and the environment to be constant at 32 and 22 $°C$, respectively. Then, by means of this model and assuming two heat sinks with different cross sections and thermal resistance, they obtained the power and voltage for a TEG. After that, they experimentally validated their proposed model with a commercialized TEG. In order to increase the production of electric power, they utilized the provided analytical model to optimize the amount of external electric load, the number of thermoelectric legs, and their geometric coefficient (which was defined as the division of the cross-sectional area of the thermoelectric legs by their length), in two generators, with a different cross-sectional area for their heat sinks. The generator's output power with smaller and larger heat sinks increased by 19.76% and 360%, respectively, compared to the original state.

4.2.4. Flexibility

Flexibility in wearable thermoelectric generators is critical to harvest the maximum amount of wasted heat from the body, considering that most body surfaces are curved [78]. Achieving flexible structures in WTEGs is usually possible in four ways: (1) by providing flexible designs for TEGs in wearable applications, (2) using flexible thermoelectric materials as TE legs in WTEGs, (3) the use of flexible substrates and electrodes and (4) integration in fabric or clothing. Here, the conducted research studies in each case have been reviewed.

4.2.4.1. Flexible design

To investigate the performance of wearable thermoelectric generators on the curved surface of body skin, in 2017, Wang et al. [78], presented a three-dimensional numerical model of a generator using the finite element method in which a flexible thermal interface layer was assumed at the connection surface of the TEG with skin. Utilizing this model, they studied the effects of the curvature radius of a hot surface, the thickness, and the thermal interface layer's heat transfer coefficient on the thermoelectric generator's performance. After that, they validated this model with experimental tests by manufacturing a TEG consisting of 24 pairs of TE legs. At last, they recorded open circuit voltages of 10 and 3 mV in a steady state by placing the thermoelectric generator on the forearm, in standing and walking states under an ambient temperature of 19 $°C$, respectively.

In another research, Shi et al. [79], presented a wearable thermoelectric generator whose thermoelectric legs were made of organic materials such as poly (3,4- ethylenedioxythiophene): poly(styrenesulfonate) or (PEDOT: PSS). In order to increase the flexibility of this generator, they placed 12 pairs of TE legs in the embedded holes on a flexible printed circuit board (FPCB). Then, placing it on the wrist under an ambient temperature of 10 $°C$, they recorded the open circuit voltage of 10.5 $\mathrm{m}\mathrm{V}$, enough to light up a light-emitting diode (LED).

In the other study, Eom et al. [80], presented a wearable thermoelectric generator in which rigid and brittle legs of Bismuth and Tellurium were used. In order to increase the flexibility of their generator, they placed each pair of thermoelectric legs in a polymer link unit and the TE legs were connected to each other with copper wires and a heat sink made of copper was placed on the upper part of each polymer link. Then, by connecting 10 units of polymer links to each other, using thin shafts that passed through the holes embedded in the thermoelectric legs and polymer links, they made a thermoelectric generator in the shape of a bracelet. In the best case, this generator was able to produce 80 $\mathrm{\mu }\mathrm{W}$ of electric power from the wrist while the research participant was running at a slow speed, under an ambient temperature of 20 $°C$ and relative humidity of 40%. Eom et al. drew the relationship of the maximum power and voltage output of a flexible thermoelectric versus the temperature difference (Fig. 11).

Fig. 11.

The relationship between maximum power and voltage output with temperature difference for a FTEG [80], with permission.

Park et al. [72], provided a flexible structure for a wearable thermoelectric generator by placing conventional rigid and fragile legs of Bismuth-Telluride in holders made of Bakelite (a type of plastic made of formaldehyde and phenol) as presented in Fig. 12-a. Then, they connected the legs in series with soft wires, and as a result, they achieved a highly flexible structure for their generator. This generator could produce an electric voltage of 8.8 $\mathrm{m}\mathrm{V}$, 138.67 $\mathrm{\mu }\mathrm{W}$ power density of 5.6 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ under an ambient temperature of 24 $°C$ and relative humidity of 40%. A view of the thermoelectric generator used in this experiment is shown in Fig. 12-b.

Fig. 12.

a) A schematic view of the flexible thermoelectric generating components presented by Park et al. and b) A real view of the thermoelectric generator used by Park et al. from Ref. [81], with permission.

4.2.4.2. Flexible materials for legs

In 2014, We et al. [82], presented a wearable thermoelectric generator using a combination of inorganic thermoelectric materials and conductive polymers, which, as shown in Fig. 13-a, consisted of 15 pairs of thermoelectric legs on a Polyimide (PI) substrate. It could produce a voltage of 12.1 $\mathrm{m}\mathrm{V}$ under a temperature difference of 5 $°C$ between the body skin and the environment. A general view of the structure and components of the used system along the cross-section C'-C is drawn in Fig. 13-b.

Fig. 13.

a) A view of the flexible thermoelectric generator presented by We et al. [82]; b) A schematic view of the components of the system along the cross-section C'-C from Ref. [82], with permission.

In 2019, Xu et al. [42], by means of material engineering, were able to produce flexible PEDOT: PSS films with a high-power factor. In their research, they reported that the power density of 1 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ was produced by placing the polymer films on the arm.

Conductive polymer thermoelectric materials are known as the most widely used and cheapest organic thermoelectric materials; however, in order to compete with conventional solid thermoelectric semiconductors in wearable applications, it is necessary to improve their thermoelectric properties. In 2021, Cao et al. [83], reviewed the various studies on the progress in the structure, mechanism, and compositions, besides mechanical and thermoelectric characteristics of conductive thermoelectric polymers.

4.2.4.3. Flexible substrates and electrodes

In 2006, Weber et al. [39], placed 900 thermoelectric legs on a flexible polyimide (PI) strip as a substrate by means of the sputtering method and then achieved a design like Fig. 14 for the generator by screwing it into a ring. This thermoelectric generator was able to produce an electric voltage of 0.8 $\mathrm{V}$ and power of 0.8 $\mathrm{\mu }\mathrm{W}$ at a temperature difference of 5 $°C$, which was adequate to provide the required electric power of a wristwatch.

Fig. 14.

A schematic of the TEG presented by Weber et al. [39] (the dimensions of this generator were the size of a coin), with permission.

In another research to produce a flexible structure for a wearable thermoelectric generator in 2018, Kim et al. [84], utilized polymer substrate and solid thermoelectric legs based on Telluride to present WTEGs with two different leg heights of 0.8 and 2.5 $\mathrm{m}\mathrm{m}$, in three different filling factor coefficients of 15.1%,19.8%, and 27.2% (Fig. 15-a). In order to test the performance of their TEGs, they used an artificial arm that simulated the temperature of an accurate human arm at room temperature. By setting a constant temperature of 33.9 $°C$ on the artificial arm, they recorded the values for open circuit voltages and power densities of the generators. Then, they presented a theoretical model in which the heat transfer in a TEG structure was simulated one-dimensionally, validated it with experimental results, and utilized it to find the optimal value of the filling factor coefficient (1.5 to 3%) and height of TE legs in different environmental conditions. A view of the prototype of the flexible TE generator used in this research is shown in Fig. 15-b.

Fig. 15.

The flexible thermoelectric generator structure made by Kim et al. [84]; a) A schematic view, b) A picture of the prototype of this generator, with permission.

In 2022, Na et al. [85] made a flexible thermoelectric generator based on a composite of Bismuth-Telluride particles and polymer. They attached the new thermoelectric generator on a common face mask and investigated its performance to generate electricity during breathing. They achieved a power of 12.5 $\mathrm{n}\mathrm{W}$ at a temperature difference of 25 $°C$.

Zhang et al. [86] developed a high-performance ultralight-weight flexible micro-TEG through pulse electroplating, microfabrication, and peeling off processes. They achieved an output power density of 14.3 $\mathrm{m}\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ at a temperature difference of 29.9 $°C$. They also argued that the voltage factor and normalized power to weight ratio of their flexible micro-TEG are over three times higher than the common flexible TEGs in the literature.

Toan et al. [87] proposed a flexible TEG based on apolyimide substrate and an electrodeposited thermoelectric material for self-powered wireless Bluetooth sensing systems. They reported the maximum output power density of 610.8 $\mathrm{\mu }\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ under a temperature difference of 52 $°C$.

Kuang et al. [88] showed that the use of heteromorphic electrodes can increase the output power of flexible thermoelectric generators (FTEGs) up to 44 times compared to FTEG with conventional electrodes. They achieve to power density of 21.3 $\mathrm{\mu }\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ under natural convection and 116.1 $\mathrm{\mu }\mathrm{W}{\mathrm{c}\mathrm{m}}^{-2}$ under forced convection with the air speed of 2.1 $\mathrm{m}/\mathrm{s}$. They also claimed that the use of flexible thermoelectric with heteromorphic electrodes can be a capable solution for obtaining energy from the human body.

4.2.4.4. TEGs integrated with fabrics or clothing

The integration of thermoelectric generators into fabric or clothing will produce more electrical power due to the extended contact surface to extract waste heat from the body. It was concluded that integrating TEGs in fabrics or clothing has a negligible effect on their outputs [89].

After Leonov et al. [90], implemented the idea of integrating TEGs into clothing for the first time in 2009, Kim et al. [91], in 2013, presented a TEG integrated into a fabric with high flexibility (Fig. 16-a). This generator consisted of 20 pairs of thermoelectric legs that were connected through thin conductive threads in the fabric (Fig. 16-b) and could produce 178 $\mathrm{n}\mathrm{W}$ of electric power from the temperature difference between the skin (32 $°C$) and the environment (5 $°C$).

Fig. 16.

a) A view of the thermoelectric generator integrated into fabric in presented by Kim et al. b) Magnified view of thermoelectrics implanted in fabric texture, from Ref. [91], with permission.

Min-Ki Kim et al. [92], presented a texture-integrated thermoelectric generator in which 12 pairs of TE legs were connected by thin conductive threads (Fig. 17) and by placing it on the chest, under an ambient temperature of 5 $°C$, 46 $\mathrm{n}\mathrm{W}$ of electric power was produced [71,92].

Fig. 17.

A schematic of the presented integrated thermoelectric generator in fabric texture by Min K. Kim et al. from Ref. [71], with permission.

Another group of researchers led by Kim et al. [93], presented a flexible generator on a piece of glass fabric, similar to a medical bandage in appearance, which consisted of 11 pairs of thermoelectric legs and when placed on the wrist (Fig. 18), it was capable of producing an open circuit voltage of 2.9 $\mathrm{m}\mathrm{V}$ and electric power of 3 $\mathrm{\mu }\mathrm{W}$ at an ambient temperature of 15 $°C$.

Fig. 18.

A view of the thermoelectric generator presented by Sun Jin Kim et al. [93], this figure is published in a RSC (Royal Society of Chemistry) journal, no permission required.

In 2017, Wu and Hu [94], presented a novel structure for wearable thermoelectric generators with high flexibility, in which impregnated yarns with thermoelectric materials, as thermoelectric legs, were sewn into a 3D textile fabric, as the substrate of the TEG. This generator, as shown in Fig. 19, with a sandwich-like structure, converted the temperature difference along the thickness of the fabric into electricity. Finally, by placing the generator on the wrist, voltages of 0.025 and 0.1242 $\mathrm{m}\mathrm{V}$ under temperature differences of 6.4 and 10.7 $°C$ were recorded, respectively [94,95].

Fig. 19.

A schematic of the thermoelectric generator structure sewn into the 3D fabric texture in the research of Wu and Hu from Ref. [95], with permission.

Cao et al. [96] overview the latest progress on the up-to-the-date FTEDs with their unique designs. They summarized the structure-related principles and factors that determine the performance of FTEDs and the advanced strategies for improving their utilities. They also pointed out the current challenges, controversies, and prospects of flexible thermoelectric modules.

4.2.5. Boundary conditions on the hot and cold plates of WTEGs

As mentioned earlier, the temperature difference between the skin and the surrounding is limited. As a result, it is crucial to identify parameters that determine the skin temperature and investigate the influence of environmental conditions on the cold surface of these generators.

Skin temperature, which usually acts as a heat source for WTEGs, is influenced by various parameters, such as the type of physical activity and metabolic rate, the thermal resistance of the skin in different parts of the body, clothing factor, and the discrepancy of the worn garments, as well as environmental conditions, including the temperature and relative humidity of the air. Environmental conditions affect the thermal resistance of the skin and influence the heat transfer rate from the cold plate of TEGs [97]. With the provided explanations, to achieve the optimal design for WTEGs, utilizing appropriate models to simulate the conditions on the terminals of these generators is a smart idea to save time and money and reduce uncertainties in the optimization process. Here, the conducted research studies on the factors affecting the temperature of the hot and cold plates of thermoelectric generators have been studied.

In order to provide an innovative thermal structure for a wearable thermoelectric generator, Leonov et al. [57], conducted experimental research on the local temperature, the amount of heat flux, and the thermal resistance of the wrist. Then, based on the results of their research, they fabricated a WTEG in the form of a bracelet that was placed on a radial artery. This generator consisted of 6144 pairs of thermoelectric legs and could produce an electric voltage of 1.5 $\mathrm{m}\mathrm{V}$ in a windless environment and a temperature of 22 $°C$. Following this research, in another work led by Leonov [53] in 2007, two WTEGs were introduced. One of them, named “Mini-Matrix 2R″ with a waterproof design for outdoor use and consisted of 24 pairs of thermoelectric legs, and the other, named “five to seven”, consisting of 32 pairs of thermoelectric legs and pin heat sink, was designed for indoor environments and could produce electric power of 250 $\mathrm{\mu }\mathrm{W}$ in the daytime. Then, by utilizing the designed generator “five to seven”, this group compared the obtained electric power by analytical and experimental methods in four areas of the wrist. Furthermore, they reported the achieved voltages in different physical activities. Finally, they used this generator to provide the required power of a wireless sensor node (in the range of 50 to 100 $\mathrm{\mu }\mathrm{W}$). The extra amount of produced power was stored in a 1.2-V rechargeable battery.

In 2009, in another research led by Leonov [90], the effect of the temperature difference between the human body and ambient on the output of wearable thermoelectric generators was experimentally investigated using “five to seven” TEGs, while the research participant was sitting or walking. It was observed that by reducing the temperature difference between the skin and the ambient (between 7 and 9 $°C$), the output power decreased, but its average never reached zero. Even if the temperature of the skin was lower than the ambient temperature, the change in the direction of heat flow in the generator would lead to the alteration of the direction of produced electric current in the generator.

Wang et al. [98], obtained the thermal resistance of thermoelectric legs under working conditions by utilizing a finite element method and applying the boundary conditions of heat flux and constant temperature on the hot and cold plates of the generator, respectively. Then, they achieved the optimum design for the generator by utilizing the previously achieved results in an analytical method. In this method, they assumed body core and the ambient temperatures to be of 37 $°C$ and 22 $°C$, respectively, and set a constant thermal resistance between the body core and the hot plate of the generator. Finally, by taking advantage of the optimal design and with the help of micromachinery technology, they presented a generator with 2350 pairs of thermoelectric legs, which was able to produce an electric voltage of 0.15 $\mathrm{V}$ and power of 0.3 $\mathrm{n}\mathrm{W}$ from the wrist while the research participant was sitting in an office.

With regards to the effect of clothing on the amount of dissipated heat from the body, in 2015, Myers and Jur [99], experimentally and numerically, investigated the structure of two types of fabric with Rib (Fig. 20-a) and Jersey textures (Fig. 20-b) in terms of thermal resistance and found that Rib textured fabric, had a lower heat transfer coefficient due to its porous structure. As a result, they concluded that by utilizing a TEG on the surface of skin covered with this type of fabric, the hot plate of the generator would receive more heat from the body, and consequently, the outputs of the TEG would increase.

Fig. 20.

Pictures of the structure of a) Rib and b) Jersey textures fabric presented by Myers and Jur [99], open access article.

Hyland et al. [6], carried out a series of experiments to find the optimal size and material for the heat sink of the generator in order to optimize the performance of their thermoelectric generator. Then, they tried to find the best body location to place the optimized TEG; therefore, they measured the output power of the generator by placing it on the wrist, upper arm, and chest and also integrating it into a T-shirt, in different airspeeds while the research participant was walking and as it is illustrated in Fig. 21, they came to the conclusion that the highest to the lowest electric power density was achieved by placing the generator on the upper part of the arm, wrist, chest, and integrating it into the T-shirt, respectively.

Fig. 21.

The diagram related to recording the power density produced by placing the generator on the upper arm, wrist, chest, and the generator integrated in the T-shirt in the experiments conducted by Hyland et al. [6], with permission.

In 2017, Myers et al. [100], studied the effects of dry bulb temperature, wet bulb temperature, and relative humidity of air on WTEGs performance in different physical activities. Using a substrate with a flexible structure, they made a TEG which was integrated into the fabric. After conducting many experiments, they concluded that with increased physical activity, not only the body's metabolic rate would increase, but the amount of the convection heat transfer coefficient on the cold plate of the generator would also escalate due to the increase of relative velocity of the body, which had a positive effect on the performance of the thermoelectric generator. In addition, they observed that dry bubble temperature and relative air humidity did not significantly affect the power density of the generator; meanwhile, the wet bulb temperature was an influential factor in the performance of the TEG.

In another research in the same year, Proto et al. [101], tested a TEG on the arm and Gastrocnemius muscle and measured the amount of electrical power produced by the generator in different physical activities, such as sitting, walking, running, and cycling. Finally, by recording electric power of about 5 $\mathrm{\mu }\mathrm{W}$ from the sitting position and 50 $\mathrm{\mu }\mathrm{W}$ while cycling, they concluded that the maximum amount of power production in the thermoelectric generator was from the Gastrocnemius muscle.

Wijethunge et al. [102], investigated the importance of assuming a constant or variable value for the thermal resistance of the skin in the temperature distribution on the skin surface and, as a result, the effect on the performance of a WTEG under different environmental conditions. Since blood perfusion rate plays a vital role in heat transfer inside the body, they used Pennes's bio-heat equation to find the temperature distribution inside the body tissue by assuming the body core temperature to be constant. In this modelling, they also considered evaporation from the skin surface and convective heat transfer as mechanisms of heat transfer from the body to the surrounding environment. Then, this modelling was validated with experimental results. In the next step, they used the obtained results of skin temperature from their modelling as input data to model the structure of a TEG in the form of a network of thermal resistances, ultimately assuming the thermal resistance of the skin to be constant or variable, resulting in a deviation between 10% and 60% in the output power from the generator.

In another research, Park et al. [72], presented a wristwatch design for a WTEG. The introduced generator was able to produce a power density of 6.97 $\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2$ and an open circuit voltage of 12.1 $\mathrm{m}\mathrm{V}$ while walking. Then, in order to more accurately simulate the behaviour of the skin, using COMSOL software, they solved the Pennes's bio-heat equation by assuming the body core temperature to be constant and utilized its results as inputs in a modelling of their TEG structure which was based on the energy conservation equation. They investigated the effect of thermal resistances on the open circuit voltage of the generator and then showed that if the optimal values are used for the height of the thermoelectric legs and filling factor and a suitable structure for the heat sink, the output of the generator would be significantly improved.

Zhang et al. [103], investigated the importance of skin tissue structure modelling in obtaining a better design for WTEGs. They divided the skin tissue geometrically into a multi-part structure, including fat, dermis, and epidermis. Also, they modelled a TEG structure on the skin surface, in which the substrates, thermoelectric legs, filling materials, and an encapsulating layer were considered. Then, they utilized a mathematical model based on one-dimensional heat transfer in the mentioned structure for skin tissue and thermoelectric generator by assuming the body core temperature to be constant, to investigate the effect of metabolic rate and blood perfusion rate, thermal resistance between the skin and the generator, and the convection heat transfer coefficient of ambient on the power density of the generator. Also, they obtained optimal values for the filling factor coefficient and thermal resistance of the filling material in several different TE leg heights.

4.2.6. Electrical matching

The temperature difference between the hot and cold terminals of wearable thermoelectric generators may change during the electricity production process, which causes the electric voltage to be low and non-uniform. On the other hand, devices powered by these generators require a certain amount of input electric voltage to start working. Therefore, the electric voltage produced by WTEGs must be boosted to an expected value to provide the required voltage of wearable devices [104], which is done by a DC-DC converter embedded in the electrical subsystem of TEGs [105]. After that, a voltage regulator adjusts the boosted voltage to power an electronic device or to be stored. Some of the DC-DC converters cannot adapt to the electric input load from the thermoelectric generator to which they are connected, which causes the energy conversion factor to drop, and as a result, the generator will be out of optimal working conditions. Therefore, considering the effect of this type of DC-DC converter on the performance of WTEGs, in order to establish the electrical resistance matching is essential.

In this regard, Vostrikov et al. [106], in 2021, successfully presented a TEG's mathematical model connected to a DC-DC converter. They utilized an iterative-based method in their model and predicted the produced electric power of a thermoelectric generator with a deviation of 9% compared to the experimental results.

4.3. Comparison of performance parameters of flexible/wearable thermoelectric generators

So far, a lot of research has been done on the development of thermoelectric wearable generators. Most of these research articles have focused on material selection and geometric structure design of wearable thermoelectric generators. In Table 6, the performance of several wearable thermoelectric generators that have been developed in recent years has been compared. In this comparison, the effect of material, substrate, and temperature difference on important performance parameters of wearable thermoelectric generators such as output voltage, output power and output power density have been investigated.

Table 6.

Performance comparison of several flexible/wearable thermoelectric generators made in recent years [83].

Materials		Substrate	Couple Number	$\Delta T$ $\left(\mathrm{K}\right)$	Output Voltage $\left(\mathrm{m}\mathrm{V}\right)$	Output power $\left(\mathrm{\mu }\mathrm{W}\right)$	Power density $\left(\mathrm{\mu }\mathrm{W}/{\mathrm{c}\mathrm{m}}^2\right)$
p-type	n-type
Bi/Te + CNTs	Bi/Te + CNTs	PDMS	–	333.15	920	0.57	4.5
Bi2Te3+FPCB	Bi2Te3+FPCB	FPCB	–	12	48	0.1306	0.67
Bi0.5Sb1.5Te3	Bi0.5Sb1.5Te3	PI	1	35	10.5	23	4.75
Bi2Te3	Bi2Te3	FPCB	52	50	37.2	0.18	16.8
Bi0.5Sb1.5Te3	Bi2Se0.5Te2.5	FPCB/PDMS	52	18	1600	2.4	13
Sb2Te3	Bi2Te2.7Se0.3	PI	10	–	151	2.9	3.44
Bi0.5Sb1.5Te3	Bi2Te1.8Se0.2	PI	–	287.15	70	0.23	3.5
PANI	Ag2Se	PVDF	–	30	7.9	0.835	0.0233
Bi0.4Sb1.6Te3	Ag2Se	PI	13	40	11.5	0.25	0.0008

As can be seen from Table 6, in most cases, thermoelectric wearable generators developed with the same number of couples and under equal temperature differences have not been tested. This makes it impossible to provide any detailed comments and conclusions about the optimal choice of materials and geometric structure.

5. A look at the commercial area of available products in micro-TEGs and wearable technologies

The review of research conducted in wearable thermoelectric generators showed that handmade modules were used in all these studies. In other words, it can be said that the technology of wearable thermoelectric generators is still in the stages of research studies and has not been commercialized. However, in a few companies, the initial steps of commercializing micro thermoelectricity technology have been taken, both in micro cooler and microgenerator applications. For example, “TEC Microsystems” [107] is one of the active and leading companies in manufacturing and commercializing miniature thermoelectric generators for small-scale energy harvesting. In this company, standard micro-TEGs, advanced micro-TEGs with high density (up to 1200 pellets per square centimetre) and even micro-TEGs with customized and special applications are designed and manufactured.

The Hi-Z Technology Company [108] is another commercial company that designs and manufactures small-scale thermoelectric generators on a customized basis.

The list of companies active in the field of thermoelectric production is given in Ref. [109].

6. Critical review of challenges and future perspectives

Reviewing the background of the research conducted in the field of wearable thermoelectric generators shows that most of these studies have focused on the issues of choosing suitable materials and optimizing the structural and geometrical design of the thermoelectric legs of these modules. In addition, thermal design and improving the geometry of heat sinks, investigating the effect of the placement of thermoelectric generators on the body, investigating the effect of free and forced heat convection on the thermal effectiveness of heat sinks, and studying climatic conditions such as temperature and humidity on the performance of thermoelectric generator modules are also among the topics of interest to researchers. Although extensive and valuable research has been done so far in the field of development of wearable thermoelectric generators to provide power for stand-alone wireless sensors, wireless sensor networks, or wearable devices, there are still weaknesses and shortcomings that their correct identification can cause more acceleration in the development of this technology. Some of the challenges are briefly stated in the following so that it may open a new perspective for future research.

• -The wearable thermoelectric generators introduced in various studies have been tested under different temperature differences, making it impossible to compare their performance and provide a definitive opinion regarding the optimal choice of material and structure. Also, in some existing research reports, wearable thermoelectric generators have been tested under very high-temperature differences, which are practically far from the actual operating conditions.
• -Although a lot of research has been done on increasing flexibility to reducing the contact resistance of wearable thermoelectric generators with the skin, not many studies have been done on the effect of the connection method and the contact pressure of the modules with the skin surface on their performance. Modules not sticking to the skin surface, which is possible in daily use, can affect their performance. This is even though, in most of the studies, the contact of the thermoelectric module with the hot source is assumed to be ideal.
• -Sweating from the skin surface can affect the thermal resistance between the skin and the hot thermoelectric surface. In the studies, this issue has not been paid much attention, so it is necessary to investigate the effects of sweating on the thermoelectric performance of wearable generators with more precision and detail.
• -In most of the existing studies, the performance of wearable thermoelectric generators has been investigated assuming a constant skin temperature. Skin temperature is not constant and varies with physical activity and metabolic rate, the structure of skin tissue in different locations on the body, which relates to the thermal resistance of the skin, clothing factor and the discrepancy of the worn garments, ambient temperature, relative humidity, and air velocity. In several previous studies, the skin temperature has been calculated by solving the Pennes bioheat equation to solve this shortcoming. However, in this method, the effect of the body's thermoregulatory responses to environmental thermal signals such as skin vasodilation, vasoconstriction, and sweating has not been considered. Therefore, it seems that using thermoregulatory models to calculate skin temperature can increase the accuracy of predictions.
• -Power converters that are incapable of matching electrically to TEGs also affect the performance of WTEGs and must be considered in simulations.

7. Conclusion

In this research, a complete review of the studies conducted in the field of thermoelectric wearable generators was done. At first, different methods of extracting kinetic and thermal energy from the body were briefly introduced and their strengths and weaknesses were compared. Then, a complete overview of the discussions regarding the thermoelectricity of wearable generators was presented. The effects of geometry, structure, and material were investigated. Also, the effect of filling material and filling factor, the effect of the geometric shape of heat sinks, the effect of airflow speed on heat sinks was investigated. Also, the factors affecting the performance of thermoelectric generators combined with fabrics and clothes were investigated. Finally, based on the results of the reviewed studies, the weaknesses and challenges in this field were examined, and suggestions for future research were presented.

It seems that most of the research so far has been done with the assumption that the skin temperature is constant and regardless of the body's thermoregulatory responses, such as the expansion and contraction of skin vessels and sweating. Meanwhile, skin temperature changes according to environmental conditions, which can affect the performance of wearable thermoelectrics in daily use conditions. Also, sweating on the skin's surface can affect the contact thermal resistance between the skin and the thermoelectric and overshadow the matching of the thermal resistances.

The results of this research can clarify the way to conduct more targeted and realistic research in optimizing wearable thermoelectric generators.

Ethics statement

This article is designed and organized as a review paper. Therefore, all the experimental data cited in this study have been reported by other researchers, and no experimental tests were performed by the authors of this article on human samples and participants or any other living organisms.

Author contribution statement

Zahrasadat Tabaie: Conceived and designed the experiments; Performed the experiments; Analyzed and interpretated the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Amir Omidvar: Analyzed and interpretated the data; Conceived and designed the experiments; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement

No data was used for the research described in the article.

Declaration of interest’s statement

The authors declare no conflict of interest.