Accepted Manuscript

A review of solar photovoltaic-thermoelectric hybrid system for electricity

generation

Guiqiang Li, Samson Shittu, Thierno M.O. Diallo, Min Yu, Xudong Zhao, Jie Ji

PII: S0360-5442(18)31074-0

DOI: 10.1016/j.energy.2018.06.021

Reference: EGY 13063

To appear in: Energy

Received Date: 23 April 2018

Accepted Date: 04 June 2018

Please cite this article as: Guiqiang Li, Samson Shittu, Thierno M.O. Diallo, Min Yu, Xudong Zhao,
Jie Ji, A review of solar photovoltaic-thermoelectric hybrid system for electricity generation, Energy
(2018), doi: 10.1016/j.energy.2018.06.021

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 A review of solar photovoltaic-thermoelectric hybrid

2 system for electricity generation
3

4 Guiqiang Li1*, Samson Shittu1, Thierno M.O. Diallo1, Min Yu1, Xudong Zhao1*, Jie

5 Ji2

6 1 School of Engineering, University of Hull, Hull HU6 7RX, UK

7 2 Department of Thermal Science and Energy Engineering, University of Science and

8 Technology of China, 96 Jinzhai Road, Hefei City, 230026, China;

10 Abstract

11 Solar energy application in a wider spectrum has the potential for high efficiency

12 energy conversion. However, solar cells can only absorb photon energy of the solar

13 spectrum near the solar cell band-gap energy, and the remaining energy will be

14 converted into thermal energy. The thermoelectric generator is a good choice to utilize

15 this thermal energy. This paper analyses the feasibility of photovoltaic-thermoelectric

16 (PV-TE), and reviews the current types and performance of PV-TE. Furthermore, it

17 presents the optimization and development of PV-TE. In addition, this paper presents

18 the challenge and efficient improvement of PV-TE in actual application. Therefore, this

19 paper would provide a valuable reference for further research into the field of PV-TE

20 and its applications.

21

22 Keywords: PV-TE, Wide solar spectrum, Electricity generation, Hybrid solar system
1
 ACCEPTED MANUSCRIPT

23 *Corresponding authors. E-mail: Guiqiang.Li@hull.ac.uk (G. Li);

24 Xudong.zhao@hull.ac.uk (X. Zhao);

25 Contents

26 1. Introduction .............................................................................................................3

27 2. Groundwork and feasibility.....................................................................................4

28 2.1 A broader spectrum design ...................................................................................4

29 2.2 PV cell ..................................................................................................................6

30 2.3 TE material ...........................................................................................................8

31 3. Types and characteristics of PV-TE hybrid system for electricity generation........9

32 3.1 Spectrum splitting PV-TE system ........................................................................9

33 3.1.1 Spectrum splitting PV-TE system structure.................................................10

34 3.1.2 Spectrum splitting device.............................................................................11

35 3.1.3 Splitting PV-TE system performance ..........................................................12

36 3.2 Integrated PV-TE system....................................................................................14

37 3.2.1 Integrated PV-TE system structure ..............................................................14

38 3.1.2 Integrated PV-TE system performance ........................................................16

39 4. Optimization and development of PV-TE.............................................................19

40 4.1 Material Optimization.........................................................................................19

41 4.2 Structural Optimization ......................................................................................21

42 4.2.1 PV geometry optimization ...........................................................................22

43 4.2.2 TE geometry optimization............................................................................23

44 4.2.3 Concentration ratio.......................................................................................24

2
 ACCEPTED MANUSCRIPT

45 4.2.4 Cooling Medium ..........................................................................................25

46 4.3 Others..................................................................................................................27

47 4.3.1 Load resistance optimization........................................................................27

48 4.3.2 New system optimization.............................................................................28

49 5. Challenge and efficient improvement in actual application..................................29

50 5.1 Enhancing the system efficiency ........................................................................29

51 5.2 Improve the cost performance ............................................................................30

52 6. Conclusion.............................................................................................................31

53 References ....................................................................................................................33

54

55 1. Introduction

56 With the growing concerns about environmental pollution, climate change, and the

57 global fossil energy crisis, research and development of renewable clean energies has

58 received more attention. The sun as one of the renewable energy sources is the most

59 potent source for human kind. This is because the solar radiation that reaches the Earth

60 surface is about 1.2 × 105 TW, which is far greater than the energy consumed by humans

61 [1-2].

62 The most common way to utilize solar energy is to convert it into two easily

63 harnessed forms; electricity and thermal energy. Apart from photovoltaic (PV) which

64 can convert solar radiations to electricity directly, thermal energy also can be converted

65 to electricity, and one promising method is utilizing the thermoelectric generator (TEG).

66 Thermoelectric (TE) devices have many advantages such as gas-free emissions, solid-

3
 ACCEPTED MANUSCRIPT

67 state operation, maintenance-free operation without any moving parts and chemical

68 reactions, vast scalability, a long life span of reliable operation and no damage to the

69 environment. Therefore, the combination of PV and TE could be considered to produce

70 more electricity.

71 Combining a photovoltaic module and a solar thermoelectric generator would

72 enable photons outside the range of a particular solar cell’s narrow absorption

73 wavelength to be directed to the TE modules which generates electricity by the

74 thermoelectric effect. Doing this would allow energy conversion efficiency to be

75 increased while simultaneously reducing the heat dissipated by the PV module.

76 This paper presents a detailed review of the current state of art in solar photovoltaic-

77 thermoelectric hybrid system for electricity generation. It begins with the analysis of

78 the groundwork and feasibility of PV-TE system. An overview of the two main types

79 and characteristics of PV-TE hybrid system for electricity generation is presented in

80 detail. Furthermore, suggestions on how to optimize and develop PV-TE systems are

81 provided. In addition, the current challenge faced in the application of PV-TE system

82 is discussed and solutions are provided. Therefore, this paper will act as a valuable

83 reference material for research and development of solar PV-TE hybrid systems for

84 electricity generation.

85

86 2. Groundwork and feasibility

87 2.1 A broader spectrum design

88 The incident broadband solar spectrum is between about 280nm and 4000nm. Most
4
 ACCEPTED MANUSCRIPT

89 solar cells are more efficient when the photon energy of the solar spectrum is near the

90 solar cell band-gap energy. Though the PV panel can directly convert solar energy into

91 electricity, the solar cells cannot fully harness the photons of energy smaller or larger

92 than band-gap energy. For example, Silicon thin-film solar cell (STC) because of its

93 material abundance, promising high performance, potential ability to be low cost and

94 its environmental friendly performance has gained rapid development in recent years.

95 However, its conversion efficiency has the theoretical upper limit estimated to be no

96 greater than 30% [3].

97

98 As shown in Fig.1, photons below the band-gap energy are dissipated as waste

99 heat when they pass through the active area. The photons above the band-gap energy

100 can only be partially utilized while the rest of the energy is converted into waste heat.

101 With the temperature of the panel rising, its efficiency drops as shown in Fig. 2. For

102 example, STC can only absorb sunlight with wavelength up to 800nm, the remaining

103 sunlight greater than 800nm cannot be absorbed by PV and would be converted into

104 waste heat which increases the temperature of the panel [5]. This not only decreases the

105 efficiency but also reduces the lifespan of the panel.

106 Thermoelectric generator (TEG) can directly convert thermal energy into electric

107 power, thus the PV-TE hybrid system was proposed. Radiation energy near the band-

108 gap is directly converted to electricity by PV panel and simultaneously, infrared energy

109 is utilized by the TEG to convert heat to electricity. Consequently, more electricity can

110 be produced by the hybrid system than the electricity produced by a single PV or TE

5
 ACCEPTED MANUSCRIPT

111 system.

112 The two most common technologies for combining PV and TEG to achieve full

113 solar spectrum utilization are: spectrum splitting photovoltaic-thermoelectric system

114 and integrated photovoltaic-thermoelectric system. Elsarrag et al. developed a spectrum

115 splitting photovoltaic-thermoelectric (PV-TE) system and observed an improved

116 overall performance [6] while Fisac et al. developed an integrated photovoltaic-

117 thermoelectric system to obtain an improved overall efficiency under severe

118 temperature conditions [7].

119 The combination of PV panel with TEG module can exhaustively absorb a wide

120 range of solar radiation spectrum either in form of infrared energy or ultraviolet energy.

121 Therefore, the concept of photovoltaic–thermoelectric hybrid system is feasible since

122 the method of combining PV with TEG can fully utilize the solar spectrum in theory.

123 This is the greatest advantage of the hybrid system over single PV and TEG system

124 operation respectively.

125

126 2.2 PV cell

127 The PV materials widely used are Silicon, III–V Cells, Thin film chalcogenide,

128 Amorphous/microcrystalline Si, Dye sensitized, Organic, Multi-junction devices. For

129 each PV type, its efficiency is limited. Presently, the common highly efficient

130 photovoltaic cells are the Si (crystalline) of 25.6± 0.5% [8], GaAs (thin film) of

131 28.8 ± 0.9% [9] and Multi-junction devices, InGaP/GaAs/InGaAs of 37.9 ± 1.2% [10]

132 under the global AM1.5 spectrum (1000 W/m2) at 25 °C. Nevertheless, these PV
6
 ACCEPTED MANUSCRIPT

133 efficiencies were all affected by the operating temperature significantly.

134 The temperature increase of solar cells represents a significant issue because of

135 their efficiency decrease [11-12], mostly for solar concentrated systems. An electrical

136 efficiency decrease of about 0.2–0.5% for every 1oC rise in cell temperature was

137 reported due to the temperature dependence of the open-circuit voltage of the cell

138 depending on the PV material [13-15]. This PV property is known as the temperature

139 coefficient of the PV cell. Table 1 shows the temperature coefficients of various PV

140 technologies along with their typical efficiency [16].

141

142 Virtuani et al. measured and compared the relative temperature coefficient of

143 power (γrel) for different technologies (a-Si based single or multi-junctions, CdTe, CIS,

144 thin-film silicon). They found that, except for the thin film Si device (γrel = -0.48 %/°C),

145 compared to the c-Si wafer-based module (γrel = -0.45 %/°C) all the studied

146 technologies have lower values of relative temperature coefficient for power. The a-Si

147 single-junction (γrel = -0.13 %/°C) device showed the less pronounced decrease with

148 temperature, followed by CdTe (γrel = -0.21 %/°C). The microcrystalline and CIGS

149 device gave very close temperature coefficients (γrel = -0.36 %/°C) [20]. Fig. 3 shows

150 the impact of temperature on PV maximum power outputs for different PV technologies.

151

152 Thus, for common solar cells, such as Si based solar cells, their price is cheap but

153 their temperature coefficient is large. Multi-junction solar cells are a possible choice

154 for high temperature operation due to their high efficiency and low temperature

7
 ACCEPTED MANUSCRIPT

155 coefficient, but their price is high.

156 2.3 TE material

157 There are many ways to classify TE materials, one of which is based on temperature.

158 Therefore, TE materials can be divided into three sections. At high temperatures (800-

159 1300K), ceramics like BaUO3 (ZT=1.8) [21] and semiconductor SnSe single crystal

160 (ZT=2.6±0.3) [22] can be used. At middle temperatures (300-800K), semiconductors

161 like Bi2Te2.85Se0.15 (ZT=2.38) [23] can be used. At the low temperatures (200-400K),

162 semiconductors like Bi2Te3 (ZT=2.4) [24] and polymers like 3, 4-

163 ethylenedioxythiophene (ZT=0.4) [25] at room temperature can be used.

164 For TE, its theoretical efficiency can be estimated by the following equation:

TH
T c 1  ZT  1

T 1  ZT  T C
max
H

165 T H (1)

166 Where TC is the cold side temperature, TH is the hot side and ZT is the Figure of merit.

167 Applying this equation for low temperatures such as room temperature, for widely

168 used materials which have ZT=1, the theoretical efficiency for that ZT between the

169 temperature range of 293K and 373K is 4%. Considering the highest ZT value of 2.4,

170 the theoretical efficiency for the same condition is 6%. Thus it is of great importance

171 to improve the value of ZT. The dimensionless parameter, ZT, is usually used to

172 determine the performance of a thermoelectric material, as shown in Eq. (2).

ZT  
2

T
173  (2)

174 Where α is the Seebeck coefficient, σ is electrical conductivity,  is thermal and T is

8
 ACCEPTED MANUSCRIPT

175 temperature. As shown in Eq. (2), ZT relates three physical properties intrinsic to the

176 material: the electrical resistivity, σ, the thermopower or Seebeck coefficient, α, and the

177 thermal conductivity, κ. The traditional way to improve ZT is to modify the crystalline

178 structure of the material. It can decrease the thermal conductivity by alloying and/or

179 inserting foreign species to gain good electricity property. Another way is to design

180 novel materials with inherently low thermal conductivity values which would leave the


2

181 power factor  the only relevant parameter [26]. It is said that ZT must be greater than

182 3 to be competitive with traditional generator. The problem is for TE materials, α, σ, κ

183 are dependent. Although theoretically, there is no limitation for developing a TE

184 material with its ZT greater than 3, no such material has ever been found, especially for

185 low temperature. The largest ZT for low temperature ever found is 2.4 at 300K by

186 growing phonon blocking electron transmitting hetero-structures by low temperature

187 metal organic chemical vapor deposition (MOCVD) technique [27] but the process is

188 complex and the price for that material is high.

189 For wider application, the method for the development of TE materials must cause

190 the ZT to a attain a high value at a low price. Until now, no such material has ever been

191 found.

192

193 3. Types and characteristics of PV-TE hybrid system for electricity generation

194 3.1 Spectrum splitting PV-TE system

195 When photon energy is close to the solar cell band-gap energy, spectrum

196 response of several solar cells is usually more efficient. Therefore, photons whose
9
 ACCEPTED MANUSCRIPT

197 energy is smaller or larger than the solar cell band-gap energy can only be used partly.

198 To solve this problem, spectrum splitting PV-TE systems have been developed because

199 they can enable the utilization of the full solar energy in a very broad solar spectrum

200 wavelength range. With regards to the issue of reducing the PV operating temperature,

201 attention has been drawn to a solar spectrum splitting system in which the PV cells and

202 the thermal processes run in parallel [28]. A spectral beam splitter would only allow

203 the short wavelength solar spectrum to reach the PV receiver, thus substantially

204 reducing both the heat load and the operating temperature of the cell [3].

205

206 3.1.1 Spectrum splitting PV-TE system structure

207 Temperature has a significant opposing effect on the efficiency of the hybrid

208 system given that high temperature difference will lead to a high efficiency for the TEG,

209 however, it would also cause the temperature of the PV to rise and consequently

210 decrease its efficiency. Therefore, the spectrum splitting system was proposed.

211 Generally, the PV-TE splitting system consists of the solar concentrator, the spectrum

212 splitting device, PV module, TE module, and cooling systems as show in Fig. 4.

213

214 The spectrum splitting device can divide the sun wavelength into two parts: the

215 suitable wavelength for PV, and the remaining out of range energy for TE. Thus, the

216 temperature of PV can be increased inconspicuously in theory, and the solar

217 concentrator can also be utilized to obtain high heat flux for the TE via Seebeck effect.

218 This would enhance the electricity efficiency of the whole system, and the temperature

10
 ACCEPTED MANUSCRIPT

219 of the PV plate can be reduced from this procedure.

220 Li et al. proposed a power system concept for spectrum splitting PV-TE hybrid

221 system. As shown in Fig. 5, the solar spectrum splitter was used to split the incident

222 sunlight for the PV and TE. In order to overcome the discontinuous nature of the solar

223 radiation, a thermal energy storage unit was coupled to the system [33].

224

225 3.1.2 Spectrum splitting device

226 The spectrum splitting device is the core of the splitting system. Many types of

227 spectrum splitting techniques have been proposed over the years for solar systems,

228 some of which are: (1) Dichroic splitters (2) Prisms (3) Absorption filters (4)

229 Luminescent concentrators and (5) Holographic concentrators. All these types of

230 splitters mentioned can be used for the spectrum splitting hybrid system with a separate

231 PV receiver and TE generator if different system configurations are employed.

232 In Fig. 6, the sunlight concentrated by the Fresnel lens was focused on the dichroic

233 beam splitter which split the light into two parts at the cut-off wavelength (λs). The

234 wavelength of the photon below λs was directed to the concentrated photovoltaic (CPV)

235 cell while the remainder was directed to the TE. The cut-off wavelength λs was

236 determined by the different spectrum splitting devices, which were required to match

237 the PV to reach high efficiency. The converted power fraction of each subsystem when

238 the cut-off wavelength point increases can be seen from the inset [35].

239 Skjølstrup et al. designed a thin-film multilayer spectral beam splitter for hybrid

240 PV-TE. Thin-film layers of SiO2 and Si3N4 which were deposited on an N-BK7 glass
11
 ACCEPTED MANUSCRIPT

241 were used to construct the beam splitter (shown in Fig. 7). Subsequently, the authors

242 studied the total achievable efficiency versus the number of layers used in the beam

243 splitter while considering from app. 20 to app. 200 layers [32].

244 3.1.3 Splitting PV-TE system performance

245 For the spectrum splitting photovoltaic-thermoelectric (PV-TE) system, several

246 studies have been carried out. Zhang et al. designed a PV-TE power generating system

247 for a project named “Nano and graded thermoelectric materials/photovoltaic-

248 thermoelectric-wind power generation” [37]. Vorobiev et al. presented the possibility

249 of using spectrum splitting in a PV-TE system and consequently showed this kind of

250 system to be a practical and efficient one [38-39]. In addition, Kraemer et al. presented

251 a general optimization methodology which could be employed in spectrum splitting

252 PV-TE systems. The authors also presented the optimum methodology for cut-off

253 wavelength of such splitting systems and the power fractions of each segment were

254 analysed. Through this study, they found out that the short wavelength region of the

255 solar spectrum which is usually directed to the solar TEG is only a small portion of the

256 total solar spectrum thus, it could be neglected [28]. Furthermore, Fleurial JP stated that

257 highly concentrated PV-TE systems using the spectrum splitting technology can

258 maximize conversion efficiency and improve the overall thermal management [40].

259 Ju et al. studied the influence of concentration ratio and cooling condition on

260 hybrid PV-TE systems. The authors studied a GaAs-CoSb3 hybrid PV-TE system and

261 the relationships between cut-off wavelength, concentration ratio and heat sink

262 performance were discovered. Subsequently, guidelines to follow when designing and
12
 ACCEPTED MANUSCRIPT

263 optimizing a PV-TE hybrid system were presented. The results showed that the band

264 gaps of the solar cells mainly determined the optimized cut-off wavelength of the whole

265 system. It was discovered that the hybrid system showed a good electrical performance

266 and operated in low temperature range when the solar concentration ratio was ranged

267 between 550 to 770 and the heat transfer coefficient h was between 3000 and 4500

268 W/m2K. Furthermore, the TEG subsystem contributed about 10% of the total output

269 power of the GaAs-CoSb3 hybrid system therefore, it played a significant role in the

270 power generation. In comparison with the single PV system, the hybrid PV-TE system

271 performed better under high concentration conditions. Considering the cooling

272 condition hcool = ∞, maximum solar cell efficiency of the PV-TE hybrid system was

273 around 40% whereas it was only 26% for the single GaAs solar cell [35].

274

275 As shown in Fig. 8, the hot side temperature of the TEG sharply decreases with the

276 increase of cut-off wavelength, which decreases the electrical efficiency of the whole

277 system. The hybrid system’s efficiency increases with the increase of the concentration

278 ratio and the optimized cut-off wavelength are almost the same value (Fig. 9). In

279 addition, the maximum overall efficiency increases from 26% to 28%. Compared with

280 the PV-only system, the maximum efficiency of the PV-TE hybrid system increases

281 from 26.62% to 27.49% with the h rising from 3000 to 4500 (Fig. 10). Li et al. used a

282 thermal energy storage unit coupled to the system and their results showed that for high

283 overall efficiency, the system configuration and optimization are the most important

284 factors. Also, their results showed that just by adding a high-grade cold energy storage

13
 ACCEPTED MANUSCRIPT

285 system under reasonable working conditions, a 30% improvement in the output power

286 could be achieved. The simulation also revealed that a suitable single band-gap PV cell

287 could produce good performance at both peak and off-peak times because it is

288 independent of the operating temperature [33]. The performance of many spectrum

289 splitting PV-TE systems is shown in Table 2.

290

291 3.2 Integrated PV-TE system

292 Another type of PV-TE system is the integrated PV-TE system and as shown in Fig.

293 11, the system consists of a PV module and a TEG module. For some systems a

294 concentrator is usually used to increase the heat flux of the sun irradiation.

295

296 3.2.1 Integrated PV-TE system structure

297 Many research has been done for this type of PV-TE system. These researchers [53-

298 57] attached thermoelectric generators to the back side of PV modules for energy

299 harvesting and the performance of these hybrid systems was investigated theoretically

300 and experimentally in their studies. Van Sark investigated the PV-TE system which

301 attached the TEG to PV modules to convert waste heat into electricity, as shown in Fig.

302 12. The author also performed a simulation for the simple and idealized model [55].

303

304 Deng et al. studied an integrated design of a solar-driven hybrid generation system

305 (HGS) and the system consisted of a silicon thin-film solar cell (STC), thermoelectric

306 generators (TEGs) and a heat collector. As seen from Fig. 13, the heat collector collects
14
 ACCEPTED MANUSCRIPT

307 waste heat from the STC and part of the solar energy, subsequently, this waste heat is

308 transferred to the TEG for thermoelectric conversion. Therefore, it is imperative to

309 include a properly designed heat collector possessing an optimized absorbing layer,

310 conducting layer and insulating layer between the STC and the TEG. The function of

311 the absorbing layer is to collect both solar radiation energy and reflection energy. While

312 the insulating layer prevents heat loss and the conducting layer allows for effective heat

313 conduction [5].

314

315 Mohsenzadeh et al. designed a CPV/T+TE hybrid system that included a receiver,

316 parabolic trough reflecting surface and a sun tracking system. The structure of the

317 receiver was such that an equilateral triangle duct was placed in the middle while TE

318 modules and PV arrays were placed at the two lateral sides of the triangular duct and

319 the outer glass cover as shown in Fig. 14 [58].

320

321 Also, as regards system structure, Soltani et al. presented a tri-generation unit that

322 consisted of a parabolic trough solar collector which was integrated with a solar cell

323 and a thermoelectric generator. The four main parts of the evacuated tube were; glass

324 envelope, photovoltaic tube, thermoelectric tube and absorber pipe (Fig. 15).

325 Furthermore, a TEG module was placed at the back of the PV cell to utilize the waste

326 heat produced from the PV cell. In addition, a single-phase working fluid was flowing

327 inside the absorber tube and this created the temperature gradient across the two ends

328 of the TEG [41].

15
 ACCEPTED MANUSCRIPT

329

330 Marandi et al. designed and fabricated a solar cavity packed with hybrid PV-TEG

331 modules (Fig. 16). It was discovered that reduction of the re-radiation loss of the solar

332 radiative power in the cavity receiver led to a temperature increase in the module.

333 Furthermore, the system was tested under solar irradiance that was varied daily for

334 several hours. The authors discovered that the efficiency of the hybrid system reached

335 an all-time high efficiency of around 21.9% at the start of production during the

336 morning hours when the system was exposed to real sunlight [59].

337

338 3.1.2 Integrated PV-TE system performance

339 Deng et al. developed a solar-driven hybrid generation system (HGS) and the

340 results showed that the performance of the TEG and STC are both enhanced for the

341 integrated design. The whole system generated power of 393mW which was increased

342 by 107.9% compared with that of the single STC [5] and the output power and its

343 increase rate in HGS is shown in Table 3.

344 Furthermore, results obtained by Van Sark showed that the annual performance of

345 the system considered, for the given condition increased by 14.7% and 11%, for Malaga

346 and Utrecht, respectively. In addition, the PV efficiency at 1000W/m2 was 10.78% for

347 the given condition. Also, under simulation conditions, the TE efficiency was 3.20%.

348 The research also included the impact of high ZT value on efficiency and it was found

349 that for ZT value of about 3, efficiency could be increased up to 50% [55].

350 Liao et al. investigated a theoretical model of a hybrid power generation device

16
 ACCEPTED MANUSCRIPT

351 consisting of a low concentrated photovoltaic (CPV) module and a thermoelectric

352 generator (TEG). The conclusion from this research was that there is an optimum

353 operating condition for some key parameters, such as the thermal conductance between

354 the CPV and the TEG, the current of the CPV, the solar irradiation the concentrating

355 ratio and the figure of merit of the TEG. In addition, the authors found that the

356 efficiency of the system could reach 15% [61].

357 Xu et al. analysed the performance of a combined system consisting of

358 concentrating photovoltaic/ thermal collector and thermoelectric generators. It should

359 be noted that this system had a cooling channel to cool the cold side of the TEG. The

360 results showed that compared with the PV/T system, the PV-TE/T system led to an

361 increase of about 8% efficiency [62]. This result is quite opposite to that of [3]. It

362 suggests that to achieve a better electrical output, cooling of the TEG cold side should

363 be done. The results also showed that the figure of merit has a significant impact on the

364 whole system efficiency. For Z=0.00275 K-1, PV-TE electrical efficiency of 18.2% was

365 achieved, which was 28% less than that of Z=0.00534 K-1. This means that exploring

366 new TE materials with high ZT is a promising way to improve the overall performance

367 of the PV-TE hybrid system and decrease the cost per watt. In addition, the results

368 showed that thickness of the TE layer has a positive effect on the whole system. It is

369 true that the thicker the layer is, the less efficient the PV panel is because of the rising

370 temperature of the PV panel. However, the reverse is the case for TEG as the thicker

371 the layer is, the more efficient the TEG becomes. In addition, the result showed that the

372 increase of the TE’s efficiency is larger than the amount of the PV’s efficiency decrease

17
 ACCEPTED MANUSCRIPT

373 therefore, the overall efficiency is increased. However, there is a balance between the

374 system temperature and the thickness of the layer because high temperature causes high

375 thermal stress, which decreases the operational life and reliability of the system [62].

376 Naiafi et al. modelled and analysed a combined photovoltaic-thermoelectric

377 power generation system. The results showed that with the increase of the solar

378 radiation, the efficiency of the PV and the efficiency of the TEG show opposite trends

379 for rising temperature values. The performance of the system for a typical summer

380 revealed that the TEG modules generated 28.398 Wh electricity during that day, which

381 was about 1.84% of the total generated electricity by the PV panel during the same

382 time. This means the conversion efficiency of the TEG is rather small. High figure of

383 merit value for TE can significantly improve its conversion efficiency. The results also

384 revealed that the number of the TEG modules has an optimal value [63].

385 Wu et al. compared the performance of PV-TE with and without glass cover, and

386 the results showed that when Z = 0.0021 K-1, the efficiency of the unglazed PV-TE

387 system was higher than that of the glazed system however, when Z = 0.0085 K-1, the

388 difference in efficiency between both systems decreased gradually as concentration

389 ratio increased [64].

390 For recent research, the total efficiency of hybrid systems is not largely boosted due

391 to the low conversion efficiency of the TEG. Some of them even show a decrease in

392 the electricity output of the whole system as a result of the poor heat exchange condition

393 of the TEG. Therefore, it is very important to find TE materials with high figure of

394 merit and low cost. Although recent research show that nanomaterials such as super-

18
 ACCEPTED MANUSCRIPT

395 lattices and nanowires can obtain high ZT, the complex process involved and high price

396 make those material still not widely used. Another thing that can be done to improve

397 the electrical efficiency of the hybrid system is to cool the cold side of the TEG as this

398 can strongly affect the whole system’s efficiency. For the structure of the integrated

399 system, it seems that the thickness of the TEG has a positive effect on the conversion

400 efficiency. However, the thermal stress can affect the lifespan of the system. Therefore,

401 it is important to find a balance point taking into consideration all these factors affecting

402 the hybrid system. The performance of many integrated PV-TE systems is shown in

403 Table 4.

404

405

406 4. Optimization and development of PV-TE

407 The ways to improve the performance of a hybrid PV-TE system are; the use of

408 higher figure of merit (ZT) material for TEG, the use of PV cells with higher efficiency

409 and optimizing thermal management design of the hybrid system [5]. Therefore, PV-

410 TE performance optimization can be classified into two main categories; 1) Material

411 optimization 2) Structural optimization. The performance of hybrid PV-TE system is

412 mainly evaluated in terms of conversion efficiency and electrical power output. The

413 recent optimization and development efforts in the field of PV-TE is shown in Table 5.

414 4.1 Material Optimization

415 The choice of PV and TE materials used can greatly affect the performance of the

416 hybrid system. In the case of PV, the key parameter to be considered is the conversion

19
 ACCEPTED MANUSCRIPT

417 efficiency of the cell while for TE, figure of merit and temperature range are key

418 parameters to be considered.

419 The following TE materials can be selected based on the required operating

420 temperature in the hybrid system: a) Bismuth telluride (Bi2Te3): X < 500 b) Lead

421 telluride (PbTe): 500 ≤ X ≤ 900 c) Silicon germanium (Si-Ge): 900 < X ≤ 1300; where

422 X is Temperature in Kelvin [83]. Hybrid system efficiency can be improved by utilizing

423 higher ZT materials. However, the efficiency would remain poor if the thermal loss is

424 still high even if higher ZT materials are used [83]. Improvements in material ZT values

425 by a factor of 2 at least is required to enhance the wide use of TE technology on a large

426 commercial scale [40]. This suggestion is in agreement with [66] and research to

427 develop higher ZT materials is still on-going. Beeri et al. suggested an efficiency

428 enhancement of 50% compared to single PV cell when material with high ZT value of

429 2.4 is used [72]. Wu et al. argued that not all larger figure of merit values are favourable

430 to a hybrid PV-TE system, depending on some specific system parameters [64]. This

431 finding contradicts the preceding findings discussed therefore, it might be important to

432 carefully choose a material with optimal ZT in order to improve overall system

433 performance.

434 Different types of PV cells have been used by various researchers in the past with

435 each researcher arguing the comparative advantage of the cell used over others.

436 Perovskite solar cell has been reported by Zhang et al. to achieve a higher efficiency

437 and lower temperature coefficient compared to silicon solar cell and dye-sensitized

438 solar cell (DSSC). The authors estimated the efficiency of a hybrid PV-TE system using

20
 ACCEPTED MANUSCRIPT

439 perovskite solar cell to be 18.6% while that of a single perovskite solar cell was 17.8%

440 [74]. Kossyvakis et al. found through theoretical investigation that performance

441 enhancement was about 22.5% for the poly-Si and 30.2% for the dye-sensitized based

442 hybrid PV-TE systems respectively at 86oC [75]. For avoidance of repetition, the

443 different types of PV cells used by various researchers and the corresponding hybrid

444 system performance are shown in Table 5. In particular, silicon and dye-sensitized solar

445 cells have been mostly used in the literatures reviewed.

446 4.2 Structural Optimization

447 The performance of a hybrid PV-TE system can be optimized by changing its

448 structural parameters and the influence of these parameters are discussed in this section.

449 Different methods have been used to optimize the structure of a hybrid PV-TE system

450 and the common ones are; Theoretical model (including numerical simulation) and

451 Experimental model. However, Chen et al. used an uncommon method to optimize the

452 structure of a hybrid PV-TE system. The authors used an Equivalent circuit method and

453 argued that the structure for the best performance of PV and TEG many not be

454 consistent with that for the optimal efficiency of the whole hybrid system [68].

455 Therefore, it is imperative to determine the structure for the optimal performance of the

456 hybrid system.

457 Nanostructure has been proposed in recent years and Luo et al. fabricated a solar

458 cell based on heterojunction structure consisting of nanostructured CdTE and Bi2Te3.

459 The efficiency of this hybrid system increased by 23% after 1-minute illumination [76].

460 Rai et al. developed an embedded fishnet meta-structure in a hybrid PV-TE system and
21
 ACCEPTED MANUSCRIPT

461 an efficiency improvement of around 11 folds was observed. The authors also stated

462 that fabricating a 2-D fishnet structure improves efficiency without increasing the size

463 of the system [84].

464 The essential components of a hybrid PV-TE system are; PV cell and TE module.

465 However, more components (e.g. cooling medium, concentrator etc.) can be added to

466 improve the performance of the overall system. The geometric parameters of each of

467 the components in the hybrid system need to be optimized in the interest of obtaining

468 improved overall performance.

469

470 4.2.1 PV geometry optimization

471 Solar energy conversion efficiency can be improved by using nanostructured

472 surfaces because they absorb more irradiance. The cone nanostructures are periodically

473 distributed on the PV surface.

474 Da et al. used a bio-inspired moth-eye nanostructured surface to reduce the

475 reflection for full solar spectrum photons and an enhanced transmission film was used

476 to improve the transmission for photons with energy below the band-gap energy of the

477 PV cells thereby enabling a substantial utilization of the solar energy spectrum, shown

478 in Fig. 17 [50].

479 Zhou et al. investigated the effect of the full-spectrum characterisitcs of the

480 nanostructure on the PV-TE hybrid system performances (Fig. 18). Furthermore, moth-

481 eye nanostructures were used on the PV surface to improve the overall conversion

482 efficiency of the PV-TE system (Fig. 19) [85,86].

22
 ACCEPTED MANUSCRIPT

483

484 4.2.2 TE geometry optimization

485 Geometric parameters of the TE (e.g. thermoelement leg length and cross-sectional

486 area) can influence the performance of a hybrid system. The use of thermoelements

487 with shorter length can result in improved power outputs when the hybrid PV-TE

488 system is operated under sufficient illumination [75]. Reducing the side spacing

489 between the thermoelements would result in higher power output from the hybrid

490 system [68]. Surface area reduction in TE modules can increase thermal contact

491 resistance and consequently increase the temperature difference between the different

492 components in the hybrid system [77]. In addition, Hashim et al. discovered that TE

493 modules with smaller cross-sectional area can generate more electrical power than those

494 with a larger cross-sectional area [73]. The geometric parameters used by the authors

495 are shown in Table 6 and the corresponding performance graph obtained is shown in

496 Fig. 20.

497

498 Furthermore, Li et al. found that the overall efficiency of the hybrid system decreases

499 as TEG height increases however, the efficiency can increase when the cross sectional

500 area increases to the same TEG height [87]. Therefore, a relationship exists between

501 the performance of a hybrid system and the geometric parameters of the TEG and this

502 is shown in Fig. 21.

503 To elucidate the significance of finding an optimal geometry for the hybrid system,

504 Li et al. argued that the maximum power output and efficiency of a hybrid PV-TE

23
 ACCEPTED MANUSCRIPT

505 system occurs when the thermoelements (n and p type) footprint (area) is symmetric

506 however, this is not true for a stand-alone TEG [88]. In the case of a TEG only system,

507 optimal efficiency is achieved when the leg geometry is a trapezoidal shape i.e. Area

508 ratio is not symmetrical [89]. Therefore, it is important for researchers to remember that

509 the optimal design for a TEG only system might not be the same for a hybrid PV-TE

510 system. Finally, future developments in the area of optimizing the thermoelement

511 thickness and module substrate can lead to increase in hybrid system power generation

512 and conversion efficiency [36].

513

514 4.2.3 Concentration ratio

515 Thermal concentration can influence the performance and the cost of hybrid PV-TE

516 systems. Zhang et al. stated that altering thermal concentration can reduce the volume

517 of TEG materials used and consequently reduce the cost of the hybrid system. In

518 addition, the authors found that variation in thermal concentration does not lead to a

519 significant change in the perovskite solar cell thus; the influence of thermal

520 concentration on the efficiency of the hybrid system can be neglected [74]. Cheknane

521 et al. found that the efficiency of InGaP based solar cells used in concentration systems

522 reduces quickly with higher temperature, while working at low insolation levels [90].

523 Beeri et al. experimentally demonstrated that for concentration factor X ≤ 200,

524 maximum system efficiency of ~ 32% can be obtained for the hybrid PV-TE system

525 with the multi-junction PV cell contributing about ~ 40% of this efficiency at X ≈ 200

526 [72]. Fig. 22 provides a graphical illustration of the influence of concentration factor

24
 ACCEPTED MANUSCRIPT

527 on the performance of the system.

528 Incident solar radiation can be adjusted by varying the concentration ratio [28].

529 Consequently, concentration ratio affects the performance of a hybrid PV-TE system.

530 Zhang et al. found that the efficiency of hybrid system employing different types of

531 solar cells vary with concentration ratio. For example, systems using silicon PV and

532 CIGS PV had an increase in total efficiency when their concentration ratio were

533 increased respectively. However, systems using polymer PV had a reduced total

534 efficiency when concentration ratio was increased [70]. The authors used Fig. 23 to

535 illustrate their findings.

536 Ju et al. found that the hybrid PV-TE system is more suitable for working under

537 high concentrations in comparison with a PV only system [35]. This finding is in

538 agreement with [81] as it can be seen clearly from Fig. 24 that the hybrid PV-TE system

539 performs better than the single PV system under high concentrations.

540

541 Increasing optical concentration ratio can lead to the improvement of hybrid system

542 performance. This is because an increase in optical concentration ratio causes an

543 increase in maximum power output of the hybrid PV-TE system [61]. In addition,

544 optical concentration as an effect on the hot side temperature and efficiency of a TEG

545 as shown in Fig. 25.

546

547 4.2.4 Cooling Medium

548 Cooling of PV cells is a major design requirement as the cells may experience both

25
 ACCEPTED MANUSCRIPT

549 short term (efficiency loss) and long-term (irreversible damage) degradation due to high

550 temperature [92]. The integration of a cooling medium into hybrid PV-TE system is

551 very important as it can improve the performance of the whole system. Different

552 cooling mediums have been used by researchers in the past and the two most common

553 are; water cooling and air cooling. This can also be seen from the Table 5 which shows

554 the cooling medium used in different hybrid system design and the corresponding

555 performance achieved. Kil et al. found that active cooling below ambient temperature

556 is not desirable as it consumes extra energy [81]. Usually, the cooling medium (e.g.

557 water, air) and cooling device (heat sink) are implemented at the bottom of the TEG.

558 Therefore, the cooling methods employed at the cold side of the TEG, affects its heat

559 dissipation capability [5] thus; performance is affected due to the temperature

560 difference created.

561 Nanofluids have been proposed as an alternative cooling medium. Huminic et al.

562 found through experimental studies that thermal conductivity and viscosity both

563 increase with the addition of nanoparticles [93]. This finding is in agreement with

564 Hajmohammadi et al. who found that the use of Ag and Cu nanoparticles improves

565 cooling performance [94]. Nanofluid improves the system efficiency in comparison

566 with water therefore, it performs better than water [64]. The authors found that

567 nanofluid increased efficiency by 0.35% and 0.48% for the unglazed and glazed

568 systems respectively (Fig. 26).

569 Zhang et al. proposed a new cooling system that can adjust the number of TEG

570 cooled by water in order to decrease the temperature fluctuation caused by the change

26
 ACCEPTED MANUSCRIPT

571 in solar irradiance. The authors introduced an adjustable cooling block into a hybrid

572 PV-TE system and observed that the power output could be improved [48]. In terms of

573 cooling device, aluminium, copper and iron are some of the materials that have been

574 used in designing a heat sink.

575 4.3 Others

576 4.3.1 Load resistance optimization

577 The influence of load resistance on the performance of a TEG has been examined by

578 various researchers but there is not a substantial amount of research on its influence in

579 the case of hybrid PV-TE systems. A TEG can achieve its maximum power output

580 when its internal resistance is equal to its load resistance [5, 87]. Optimizing load

581 resistance of a TEG alone is not suitable for obtaining maximum PV-TE power output.

582 Therefore, the internal resistance of a TE module cannot be used as a reference to match

583 the load resistance for a hybrid PV-TE system [43]. Consequently, it is paramount for

584 researchers to know that the optimal design parameters for a single TEG might not be

585 the same for a hybrid PV-TE system. Lin et al. discovered that smaller TEG load

586 resistance is beneficial to the performance of the hybrid PV-TE system and for optimal

587 performance, TEG load resistance must be greater than TEG internal resistance [71].

588 This proves the argument that the TEG optimal load resistance is not the same with

589 hybrid PV-TE optimal load resistance.

590 Kil et al. observed that the efficiency of a CPV cell in a hybrid CPV-TE system

591 varies with load resistance connected to the TEG thus; the efficiency decreases as the

592 load resistance increases [81]. A solution to reduce the influence of load resistance on
27
 ACCEPTED MANUSCRIPT

593 a hybrid PV-TE system is to implement lossless coupling between the PV and TE

594 devices. Park et al. developed a hybrid PV-TE system with lossless coupling and an

595 overall efficiency improvement of ~30% at 15oC temperature gradient was achieved

596 [67]. Therefore, load resistance not only affects the power output of a hybrid system

597 but it also affects its efficiency.

598 Zhang et al. developed the power conditioning circuit using maximum power point

599 tracking so that the output power of the proposed TE–PV hybrid energy system can be

600 maximized, shown in Fig. 27 [95].

601

602 4.3.2 New system optimization

603 Li et al. designed a novel PV-TE system using a flat plate micro-channel heat pipe

604 (MCHP) which could reduce the quantity of TEG used. The PV modules in the system

605 were attached to the upper surface of the MCHP evaporator while the TE modules were

606 attached to the condenser lower surface. During operation, solar energy was imposed

607 on the PV modules located on the MCHP evaporator upper surface, then thermal energy

608 was conveyed to the MCHP condenser by evaporation of the working fluid inside the

609 MCHP. In the condenser, heat was disposed by condensation of the MCHP working

610 fluid and this heat was then transferred to the attached TE modules. This configuration

611 greatly reduced the amount of TEG modules used and also reduced a significant part of

612 the cost thereby showing its comparative advantage over the TE system without heat

613 pipe in series [79].

614 Cui et al. constructed a novel PV-PCM-TE hybrid system to reduce temperature

28
 ACCEPTED MANUSCRIPT

615 fluctuations in the PV cells and the TE modules consequently, keep the hybrid system

616 at a fixed operating condition (Fig. 28) [77].

617

618 5. Challenge and efficient improvement in actual application

619 PV and TE materials improvement would play a critical role in the future, however

620 the current challenge is that, efficiency increase by a large margin cannot be

621 accomplished in a short time. Therefore, for actual application, the current realizable

622 efficient improvement based on existing PV and TE modules is significant. There are

623 two methods to overcome this challenge: one is to increase the system efficiency, and

624 the other is to improve the cost performance.

625 5.1 Enhancing the system efficiency

626 The key to enhancing the system performance is to couple PV and TE in a way that

627 ensures optimum total conversion efficiency. There has been some different results that

628 indicated that the performance of the coupled system was worse than that of the

629 photovoltaic system alone [96, 39]. Bjørk et al. investigated the hybrid system with four

630 different types of PV cells and a universal bismuth telluride TE module theoretically.

631 For c-Si, CIGS and CdTe (Cadmium telluride) PV cells, the hybrid system had a worse

632 performance compared to the pure PV system. In addition, the output power of the

633 coupled system only increased slightly when the a-Si PV cell was used [96].

634 Therefore, thermal resistance needs to be optimized in actual working conditions

635 (either the thermal resistance between the PV and ambient environment, or the thermal

636 resistance between the PV and TEG, or the thermal resistance between the TEG and
29
 ACCEPTED MANUSCRIPT

637 ambient environment) so as to overcome the low PV-TE efficiency. Wu et al. analysed

638 a coupled system with and without a glass cover. The authors found that the glazed

639 system could have a better performance than the unglazed coupled system under certain

640 conditions [64].

641 Zhang et al. studied the impact of thermal contact resistance (TCR) and found that

642 when a small TCR was used, the total efficiency of the hybrid system was higher than

643 that of the PV system alone however, an opposite result was obtained when the TCR

644 was large [70].

645 Yin et al. studied two main TCRs which were; the TCR between PV cell and

646 TE generator as well as the TCR between TE generator and cooling system. The results

647 obtained showed for different PV cells used, the impacts of the TCR on the hybrid

648 system were similar. In addition, the results showed that a-Si PV cell and polymer PV

649 cell are better for a concentrated hybrid system. Furthermore, it was found that thermal

650 resistance increase can significantly improve the performance of a hybrid system using

651 a-Si PV cell or polymer PV cell [45].

652 In addition, many researchers have studied the influence of cooling system on

653 hybrid system performance and found that a good cooling system can increase the total

654 coupled system output power and efficiency [77, 46].

655 5.2 Improve the cost performance

656 It is clear that PV-TE has a higher price than PV alone and it has been mentioned

657 that the TE has a limited contribution to the PV-TE power output since TE has a low

658 electrical efficiency. Therefore, the additional cost of TE still makes the overall system
30
 ACCEPTED MANUSCRIPT

659 cost high even though the PV price is reducing significantly. Thus, for either the

660 splitting system or integrated system, the system looks economically unfeasible due to

661 the significant higher cost and lower output power of the TE module in comparison to

662 the PV module. However, Li et al. developed a conceptual design of a novel

663 photovoltaic thermoelectric system that saved a lot of TEG modules used by employing

664 the micro-channel heat pipe (Fig. 29). The study indicated that the recovery cost of the

665 extra investment of TE could be achieved in six years which means from the sixth year,

666 the new PV-TE system would be more cost effective compared to the conventional PV

667 system [79].

668

669

670 6. Conclusion

671 PV-TE is a solution for solar energy in a wide spectrum, because it can take full

672 advantage of the different power generation principles of PV and TE. The field of PV-

673 TE has received more attention by researchers recently therefore it is experiencing

674 continuous improvements. This review presented the feasibility of PV-TE, common

675 types, material, optimization and development of PV-TEG system, challenge and

676 efficient improvement in actual application. The decrease in cost and increase of

677 conversion efficiency of the PV-TE system positions it to potentially play a great part

678 in contributing to the global efforts for energy conservation and pollutant reduction. In

679 addition, some appropriate future research directions were proposed by the authors.

680 New system designs with high quality processing technology will also enhance the cost

31
 ACCEPTED MANUSCRIPT

681 effectiveness of PV-TE, which will enable it gain a wide scale application based on the

682 existing materials. In particular, a novel PV-TE system using a flat plate micro-channel

683 heat pipe has shown potential to help reduce the quantity of TEG used significantly and

684 thus, reduce the overall cost of the system. Also, nanofluids as an alternative cooling

685 medium as produced a better performance in comparison with water thus, it could be

686 considered in place of conventional cooling. This review as also shown that the optimal

687 design parameters for a TEG only system might not be the same for a hybrid PV-TE

688 system. Furthermore, nanostructure is gaining more attention and it could significantly

689 improve the efficiency of the PV-TE system. Comprehensive tables showing the most

690 important performance parameters of different PV-TE systems reported in literature

691 have also been presented in this review. In summary, this research has thoroughly

692 reviewed the current state of art in photovoltaic-thermoelectric hybrid system for

693 electricity generation.

694

695 Acknowledgment

696 This study was sponsored by the Project of EU Marie Curie International Incoming

697 Fellowships Program (745614), National Key Research and Development Project

698 (2016YFE0124800), Key Projects of International Cooperation of Chinese Academy

699 of Sciences (211134KYSB20160005).

700

701

32
 ACCEPTED MANUSCRIPT

702 References:

703 [1] Crabtree GW and Lewis NS. Solar energy conversion, Physics Today

704 2007;60:37-42

705 [2] Thirugnanasambandam M, Iniyan S, Goic R. A review of solar thermal

706 technologies 2010;14: 312–22.

707 [3] Henry CH. Limiting efficiencies of ideal single and multiple energy gap

708 terrestrial solar cells. Journal of Applied Physics 1980;51:4494-4500.

709 [4] Imenes AG, Mills DR. Spectral beam splitting technology for increased

710 conversion efficiency in solar concentrating systems: a review. Solar Energy

711 Materials & Solar Cells 2004;84:19-69.

712 [5] Deng Y, Zhu W, Wang Y, Shi Y. Enhanced performance of solar-driven

713 photovoltaic–thermoelectric hybrid system in an integrated design. Solar

714 Energy 2013;88:182–191.

715 [6] Elsarrag E, Pernau H, Heuer J, Roshan N, Alhorr Y, Bartholome K. Spectrum

716 splitting for efficient utilization of solar radiation: a novel photovoltaic–

717 thermoelectric power generation system. Renewables: Wind, Water, and Solar

718 2015;2:1-11.

719 [7] Fisac M, Villasevil FX , López AM. High-efficiency photovoltaic technology

720 including thermoelectric generation. Journal of Power Sources 2014;252:264-

721 269.

722 [8] Panasonic Press Release. Panasonic HIT® solar cell achieves world's highest

723 energy conversion efficiency of 25.6% at research level,

33
 ACCEPTED MANUSCRIPT

724 http://panasonic.co.jp/corp/news/official.data/data.dir/2014/04/en140410-

725 4/en140410-4.html; 2014.

726 [9] Kayes BM, Nie H, Twist R, Spruytte SG, Reinhardt F, Kizilyalli IC, Higashi

727 GS. 27.6% conversion efficiency, a new record for single-junction solar cells

728 under 1 sun illumination. Proceedings of the 37th IEEE Photovoltaic Specialists

729 Conference 2011.

730 [10] Sasaki K, Agui T, Nakaido K, Takahashi N, Onitsuka R, Takamoto T.

731 Proceedings, 9th international conference on concentrating photovoltaics

732 systems, Miyazaki, Japan 2013.

733 [11] Wysocki J and Rappaport P. Effect of Temperature on Photovoltaic Solar

734 Energy Conversion. Journal of Applied Physics 1960;31:571-578.

735 [12] Skoplaki E and Palyvos JA. On the temperature dependence of photovoltaic

736 module electrical performance: A review of efficiency/power correlations.

737 Solar Energy 2009;83:614-624.

738 [13] Radziemska E. The effect of temperature on the power drop in crystalline silicon

739 solar cells. Renewable Energy 2003;28:1–12.

740 [14] King DL., Eckert PE. Characterizing (rating) the performance of large

741 Photovoltaic arrays for all operating conditions. Proceeding of the 25th IEEE

742 PV specialists conference, Washington, USA 1996;1385–1388.

743 [15] Moshfegh B, Sandberg M. Flow and heat transfer in the air gap behind

744 Photovoltaic panels. Renewable and Sustainable Energy Reviews 1998;2:287–

745 301.

34
 ACCEPTED MANUSCRIPT

746 [16] Makki A, Omer S, Sabir H. Advancements in hybrid photovoltaic systems for

747 enhanced solar cells performance. Renewable and Sustainable Energy Reviews

748 2015;41:658–68.

749 [17] Kern Jr EC, Russell MC. Combined photovoltaic and thermal hybrid collector

750 systems. Proceedings of the 13th IEEEPV specialist conference 1978;1153-1.

751 [18] Zhang X, Zhao X, Smitha S, Xub J, Yuc X. Review of R&D progress and

752 practical application of the solar photovoltaic/thermal(PV/T) technologies.

753 Renewable and Sustainable Energy Reviews 2012;16:599–61.

754 [19] Del Cueto JA. Comparison of energy production and performance from flat

755 plate Photovoltaic module technologies deployed at fixed tilt. Proceeding of

756 the29th IEEE PV specialists conference, New Orleans, USA 2002.

757 [20] Virtuani A, Pavanello D, Friesen G, Overview of temperature coefficients of

758 different thin film photovoltaic technologies, 25th European Photovoltaic Solar

759 Energy Conference and Exhibition / 5th World Conference on Photovoltaic

760 Energy Conversion, Valencia, Spain 2010.

761 [21] Kurosaki K, Matsuda T, Uno S, Kobayashi S, Yamanaka S. Thermoelectric

762 properties of BaUO3. Journal of Alloys and Compounds 2001;319:271–275.

763 [22] Zhao LD, Lo S, Zhang Y, Sun H, Tan G, Uher C, et al. Ultralow thermal

764 conductivity and high thermoelectric figure of merit in SnSe crystals

765 2014;508:373-377.

766 [23] Kim TS, Kim IS, Kim TK, Hong SJ, Chun BS. Thermoelectric properties of p-

767 type 25%Bi2Te3 +75%Sb2Te3 alloys manufactured by rapid solidification and

35
 ACCEPTED MANUSCRIPT

768 hot pressing. Materials Science and Engineering 2002;90:42–46.

769 [24] Venkatasubramanian R, Siivola E, Colpitts T, O’Quinn B. Thin-film

770 thermoelectric devices with high room-temperature figures of merit. Nature

771 2001;413:597–602.

772 [25] Kim GH, Shao L, Zhang K, Pipe KP. Engineered doping of organic

773 semiconductors for enhanced thermoelectric efficiency. Nature Materials

774 2013;12.

775 [26] Wang S, Qing L, Li L. High-temperature thermoelectric properties of Cd1-

776 xPrxO. Scripta Materialia 2013;69:533-536.

777 [27] Seo JH, Lee DM, Park CLK, Kim JH, Nishida IA. Microstructural and

778 thermoelectric properties of hot-extruded p - t y p e Bi0.5Sb1.5Te3.

779 Functionally Graded Materials 1996;545–550.

780 [28] Kraemer D, Hu L., Muto A, Chen X, Chen G, Chiesa M. Photovoltaic-

781 thermoelectric hybrid systems: A general optimization methodology. Applied

782 Physics Letters 2008;92:243503.

783 [29] Bjørk R, Nielsen KK. The maximum theoretical performance of unconcentrated

784 solar photovoltaic and thermoelectric generator systems. Energy Conversion

785 and Management 2018;156:264-268.

786 [30] Contento G, Lorenzi B, Rizzo A, Narducci D. Efficiency enhancement of a-Si

787 and CZTS solar cells using different thermoelectric hybridization strategies.

788 Energy 2017;131:230-238.

789 [31] Sibin KP, Selvakumar N, Kumar A, Dey A, Sridhara N, Shashikala HD et al.

36
 ACCEPTED MANUSCRIPT

790 Design and development of ITO/Ag/ITO spectral beam splitter coating for

791 photovoltaic-thermoelectric hybrid systems. Solar Energy 2017;141:118-126.

792 [32] Skjølstrup EJH, Søndergaard T. Design and optimization of spectral

793 beamsplitter for hybrid thermoelectric-photovoltaic concentrated solar energy

794 devices. Solar Energy 2016;139:149-156.

795 [33] Li Y, Witharana S, Cao H, Lasfargues M, Huang Y, Ding Y. Wide spectrum

796 solar energy harvesting through an integrated photovoltaic and thermoelectric

797 system. Particuology 2014;15:39–44.

798 [34] Shou C, Luo Z, Wang T, Shen W, Rosengarten G, Wei W et al. Investigation of

799 a broadband TiO2/SiO2 optical thin-film filter for hybrid solar power systems.

800 Applied Energy 2012;92:298-306.

801 [35] Ju X, Wang Z, Flamant G, Li P, Zhao W. Numerical analysis and optimization

802 of a spectrum splitting concentration photovoltaic–thermoelectric hybrid

803 system. Solar Energy 2012;86:1941–1954.

804 [36] Mizoshiri M, Mikami M, Ozaki K. Thermal–Photovoltaic Hybrid Solar

805 Generator Using Thin-Film Thermoelectric Modules. Japanese Journal of

806 Applied Physics 2012;51:06FL07.

807 [37] Zhang, QJ, Tang, XF, Zhai PC, Niino M, Endo C. Recent development in nano

808 and graded thermoelectric-materials. Materials Science Forum 2005;492-

809 493:135-140.

810 [38] Vorobiev Y, González-Hernández J, Vorobiev P, Bulat L. Thermal-

811 photovoltaic solar hybrid system for efficient solar energy conversion. Solar

37
 ACCEPTED MANUSCRIPT

812 energy 2006;80:170–176.

813 [39] Vorobiev Y, González-Hernández J, Kribus A. Analysis of potential conversion

814 efficiency of a solar hybrid system with high temperature stage. Journal of Solar

815 Energy Engineering 2006;128:258–260.

816 [40] Fleurial JP. Thermoelectric power generation materials: technology and

817 application opportunities. Journal of the Minerals, Metals and Materials Society

818 2009;61:79–85.

819 [41] Soltani S, Kasaeian A, Sokhansefat T, Shafii MB. Performance investigation of

820 a hybrid photovoltaic/thermoelectric system integrated with parabolic trough

821 collector. Energy Conversation and Management 2018;159:371-380.

822 [42] Yin E, Li Q, Xuan Y. One-day performance evaluation of photovoltaic-

823 thermoelectric hybrid system. Energy 2018;143:337-346.

824 [43] Li G, Zhou K, Song Z, Zhao X, Ji J. Inconsistent phenomenon of thermoelectric

825 load resistance for photovoltaic–thermoelectric module. Energy Conversion and

826 Management 2018;161:155–61.

827 [44] Li D, Xuan Y, Li Q, Hong H. Exergy and energy analysis of photovoltaic-

828 thermoelectric hybrid systems. Energy 2017;126:343-351.

829 [45] Yin E, Li Q, Xuan Y. Thermal resistance analysis and optimization of

830 photovoltaic-thermoelectric hybrid system. Energy Conversion and

831 Management 2017;143:188-202.

832 [46] Soltani S, Kasaeian A, Sarrafha H, Wen D. An experimental investigation of a

833 hybrid photovoltaic/thermoelectric system with nanofluid application. Solar

38
 ACCEPTED MANUSCRIPT

834 Energy 2017;155:1033-1043.

835 [47] Teffah K, Zhang Y. Modeling and experimental research of hybrid PV-

836 thermoelectric system for high concentrated solar energy conversion. Solar

837 Energy 2017;157:10-19.

838 [48] Zhang J, Xuan Y. Performance improvement of a photovoltaic- Thermoelectric

839 hybrid system subjecting to fluctuant solar radiation. Renewable Energy

840 2017;113:1551–8.

841 [49] Lamba R, Kaushik SC. Modeling and performance analysis of a concentrated

842 photovoltaic–thermoelectric hybrid power generation system. Energy

843 Conversion and Management 2016;115:288-298.

844 [50] Da Y, Xuan Y, Li Q. From light trapping to solar energy utilization: A novel

845 photovoltaic–thermoelectric hybrid system to fully utilize solar spectrum.

846 Energy 2016;95:200-210.

847 [51] Rezania A, Sera D, Rosendahl LA. Coupled thermal model of photovoltaic-

848 thermoelectric hybrid panel for sample cities in Europe. Renewable Energy

849 2016;99:127-135.

850 [52] Verma V, Kane A, Singh B. Complementary performance enhancement of PV

851 energy system through thermoelectric generation. Renewable and Sustainable

852 Energy Reviews 2016;58:1017-1026.

853 [53] Yu HY, Li YQ, Shang YH, Su B. Design and Investigation of Photovoltaic and

854 Thermoelectric Hybrid Power Source for Wireless Sensor Networks.

855 Proceedings of the 3rd IEEE Inter. Conf. on Nano/Micro Engineered and

39
 ACCEPTED MANUSCRIPT

856 Molecular Systems, Sanya, China 2008:196-201.

857 [54] Leonov V, Torfs T, Vullers RJM, Hoof CV. Hybrid thermoelectric-photovoltaic

858 generators in wireless electroencephalography diadem and electrocardiography

859 shirt. Journal of Electronic Materials 2010;39:1674-1680.

860 [55] Van Sark WGJHM. Feasibility of photovoltaic–thermoelectric hybrid modules.

861 Applied Energy 2011; 88:2785-2790.

862 [56] Yang D, Yin H. Energy conversion efficiency of a novel hybrid solar system

863 for photovoltaic, thermoelectric, and heat utilization. IEEE Transactions on

864 Energy Conversion 2011;26:662-670.

865 [57] Wang N, Han L, He H, Park NH, Koumoto K. A novel high-performance

866 photovoltaic-thermoelectric hybrid device. Energy and Environmental Science

867 2011;4:3676-3679.

868 [58] Mohsenzadeh M, Shafi MB, Mosleh HF. A novel concentrating

869 photovoltaic/thermal solar system combined with thermoelectric module in an

870 integrated design. Renewable Energy 2017;113:822-834.

871 [59] Marandi OF, Ameri M, Adelshahian B. The experimental investigation of a

872 hybrid photovoltaic-thermoelectric power generator solar cavity-receiver. Solar

873 Energy 2018;161:38-46.

874 [60] Guo X, Zhang Y, Qin D, Luo Y, Li D, Pang Y, Meng Q. Hybrid tandem solar

875 cell for concurrently converting light and heat energy with utilization of full

876 solar spectrum. Journal of Power Sources 2010;195:7684–7690.

877 [61] Liao T, Lin B, Yang Z. Performance characteristics of a low concentrated

40
 ACCEPTED MANUSCRIPT

878 photovoltaic- thermoelectric hybrid power generation device. International

879 Journal of Thermal Sciences 2014;77:158–64.

880 [62] Xu X, Zhou S, Meyers MM, Sammakia BG, Murray BT. Performance Analysis

881 of a Combination System of Concentrating Photovoltaic/ Thermal Collector and

882 Thermoelectric Generators. Journal of Electronic Packaging 2014. 136(4),

883 041004.

884 [63] Najafi H, Woodbury KA. Modeling and Analysis of a Combined Photovoltaic-

885 Thermoelectric Power Generation System. J.Sol. Energy Eng 2013;031013:1-

886 8.

887 [64] Wu Y-Y, Wu S-Y, Xiao L. Performance analysis of photovoltaic-thermoelectric

888 hybrid system with and without glass cover. Energy Conversion and

889 Management 2015;93:151-159.

890 [65] Chen T, Guai GH, Gong C, Hu W, Zhu J, Yang H, Yan Q, Li CM.

891 Thermoelectric Bi2Te3-improved charge collection for high-performance dye-

892 sensitized solar cells. Energy & Environmental Science 2012;5:6294–6298.

893 [66] Zhang Y, Fang J, He C, Yan H, Wei Z, Li Y. Integrated energy-harvesting

894 system by combining the advantages of polymer solar cells and thermoelectric

895 devices. Journal of Physical Chemistry C 2013;117:24685–91.

896 [67] Park KT, Shin SM, Tazebay AS, Um HD, Jung JY, Jee SW, et al. Lossless

897 hybridization between photovoltaic and thermoelectric devices. Scientific

898 Reports 2013;3:1–6.

899 [68] Chen H, Wang N, He H. Equivalent circuit analysis of photovoltaic –

41
 ACCEPTED MANUSCRIPT

900 thermoelectric hybrid device with different TE module structure. Advances in

901 Condensed Matter Physics 2014:1–16.

902 [69] Hsueh T, Shieh J, Yeh Y. Hybrid Cd-free CIGS solar cell / TEG device with

903 ZnO nanowires. Progress in Photovoltaics: Research and Applications 2015;23:

904 507–512.

905 [70] Zhang J, Xuan Y, Yang L. Performance estimation of photovoltaic-

906 thermoelectric hybrid systems. Energy 2014;78:895–903.

907 [71] Lin J, Liao T, Lin B. Performance analysis and load matching of a photovoltaic-

908 thermoelectric hybrid system. Energy Conversion and Management

909 2015;105:891–9.

910 [72] Beeri O, Rotem O, Hazan E, Katz EA, Braun A. Hybrid photovoltaic-

911 thermoelectric system for concentrated solar energy conversion: Experimental

912 realization and modeling. Journal of Applied Physics 2015;118:115104.

913 [73] Hashim H, Bomphrey JJ, Min G. Model for geometry optimisation of

914 thermoelectric devices in a hybrid PV/TE system. Renewable Energy

915 2016;87:458–63.

916 [74] Zhang J, Xuan Y, Yang L. A novel choice for the photovoltaic-thermoelectric

917 hybrid system: the perovskite solar cell. International Journal of Energy

918 Research 2016;40:1400–9.

919 [75] Kossyvakis DN, Voutsinas GD, Hristoforou E V. Experimental analysis and

920 performance evaluation of a tandem photovoltaic-thermoelectric hybrid system.

921 Energy Conversion and Management 2016;117:490–500.

42
 ACCEPTED MANUSCRIPT

922 [76] Luo B, Deng Y, Wang Y, Gao M, Zhu W, Hashim HT, et al. Synergistic

923 photovoltaic-thermoelectric effect in a nanostructured CdTe/Bi2Te3

924 heterojunction for hybrid energy harvesting. RSC Advances 2016;6:114046–51.

925 [77] Cui T, Xuan Y, Li Q. Design of a novel concentrating photovoltaic-

926 thermoelectric system incorporated with phase change materials. Energy

927 Conversion and Management 2016;112: 49–60.

928 [78] Zhou Z, Yang J, Jiang Q, Li W, Luo Y, Hou Y et al. Large improvement of

929 device performance by a synergistic effect of photovoltaics and thermoelectrics.

930 Nano Energy 2016;22:120–128.

931 [79] Li G, Zhao X, Ji J. Conceptual development of a novel photovoltaic-

932 thermoelectric system and preliminary economic analysis. Energy Conversion

933 and Management 2016;126:935-943.

934 [80] Cotfas DT, Cotfas PA, Machidon OM, Ciobanu D. Investigation of the

935 photovoltaic cell / thermoelectric element hybrid system performance. IOP

936 Conference Series: Materials Science and Engineering 2016;133:1–10.

937 [81] Kil T-H, Kim S, Jeong D-H, Geum D-M, Lee S, Jung S-J, et al. A highly-

938 efficient, concentrating-photovoltaic/thermoelectric hybrid generator. Nano

939 Energy 2017;37:242–7.

940 [82] Rowe DM. Thermoelectrics, an environmentally-friendly source of electrical

941 power. Renewable Energy 1999;16:1251–6.

942 [83] Zhu W, Deng Y, Wang Y, Shen S, Gulfam R. High-performance photovoltaic-

943 thermoelectric hybrid power generation system with optimized thermal

43
 ACCEPTED MANUSCRIPT

944 management. Energy 2016;100:91–101.

945 [84] Rai P, Oh S, Ramasamy M, Varadan VK. Photonic nanometer scale

946 metamaterials and nanoporous thermoelectric materials for enhancement of

947 hybrid photovoltaic thermoelectric devices. Microelectronic Engineering

948 2015;148:104–9.

949 [85] Zhou YP, Li, MJ, Yang WW, He YL. The effect of the full-spectrum

950 characteristics of nanostructure on the PV-TE hybrid system performances

951 within multi-physics coupling process. Applied Energy 2018;213:169-178.

952 [86] Zhou YP, He YL, Qiu Y, Ren Q, Xie. Multi-scale investigation on the absorbed

953 irradiance distribution of the nanostructured front surface of the concentrated

954 PV-TE device by a MC-FDTD coupled method. Applied Energy 2017;207:18-

955 26.

956 [87] Li G, Chen X, Jin Y. Analysis of the primary constraint conditions of an

957 efficient photovoltaic-thermoelectric hybrid system. Energies 2017;10:1–12.

958 [88] Li G, Chen X, Jin Y, Ji J. Optimizing on thermoelectric elements footprint of

959 the photovoltaic-thermoelectric for maximum power generation. Energy

960 Procedia 2017;142:730–5.

961 [89] Al-Merbati AS, Yilbas BS, Sahin AZ. Thermodynamics and thermal stress

962 analysis of thermoelectric power generator: Influence of pin geometry on device

963 performance. Applied Thermal Engineering 2013;50:683–92.

964 [90] Cheknane A, Hilal HS, Charles JP, Benyoucef B, Campet G. Modelling and

965 simulation of InGaP solar cells under solar concentration: Series resistance

44
 ACCEPTED MANUSCRIPT

966 measurement and prediction. Solid State Sciences 2006;8:556–9.

967 [91] Hajji et al. 2017] [Hajji M, Labrim H, Benaissa M, Laazizi A, Ez-Zahraouy H,

968 Ntsoenzok E, et al. Photovoltaic and thermoelectric indirect coupling for

969 maximum solar energy exploitation. Energy Conversion and Management

970 2017;136:184–191.

971 [92] Royne A, Dey CJ, Mills DR. Cooling of photovoltaic cells under concentrated

972 illumination: A critical review. Solar Energy Materials and Solar Cells

973 2005;86:451–83.

974 [93] Huminic G, Huminic A. Application of nanofluids in heat exchangers: A review.

975 Renewable and Sustainable Energy Reviews 2012;16:5625–38.

976 [94] Hajmohammadi MR, Maleki H, Lorenzini G, Nourazar SS. Effects of Cu and

977 Ag nano-particles on flow and heat transfer from permeable surfaces. Advanced

978 Powder Technology 2015;26:193–9.

979 [95] Zhang X, Chau KT. An automotive thermoelectric–photovoltaic hybrid energy

980 system using maximum power point tracking. Energy Conversion and

981 Management 2011;52:641-647.

982 [96] Bjørk R, Nielsen KK. The performance of a combined solar photovoltaic (PV)

983 and thermoelectric generator (TEG) system. Solar Energy 2015;120:187–94.

984
985
986
987
988

45
 ACCEPTED MANUSCRIPT

989 Figure captions

990 Fig. 1. Splitting the solar spectrum into components for PV and thermal energy

991 conversion [4].

992 Fig. 2. Relationship between solar cell temperature and efficiency [18].

993 Fig. 3. Impact of temperature on PV maximum power output [20].

994 Fig. 4. The PV-TE hybrid system: (a) Schematic diagram of the hybrid system; (b)

995 Energy flow chart of the hybrid system [35].

996 Fig. 5. Schematic of the PV-TE hybrid power system [33].

997 Fig. 6. Segmentation of AM1.5D spectrum in two regions for the CPV subsystem and

998 TEG subsystem [35].

999 Fig. 7. Schematic of initial design. The first two layers from left represents MgF2

1000 (orange) and SiO2 (dark green), respectively [32].

1001 Fig. 8. Working temperature of the hybrid system vs. cut-off wavelength, Cg = 200, h

1002 = 10000 W/m2K-1 [35].

1003 Fig. 9. Hybrid system efficiency vs. cut-off wavelength for different concentration

1004 ratio, h = 10000 W/m2K-1 [35].

1005 Fig. 10. Comparison of the efficiency between the PV-only system and the PV-TE

1006 hybrid system [35].

1007 Fig. 11. The Schematic diagram of an integrated system.

1008 Fig. 12. Schematic overview of the PV-TE hybrid system with multiple thermoelectric

1009 generators consisting of n- and p-type doped semiconductor legs [55].

1010 Fig. 13. Schematic diagram of heat flux in HGS [5].

46
 ACCEPTED MANUSCRIPT

1011 Fig. 14. Schematic diagram of the cross-section of CPV/T+TE hybrid collector [58].

1012 Fig. 15. Cross-sectional view of the evacuated tube [41].

1013 Fig. 16. PV-TEG hybrid system (a) cavity (b) hybrid PV-TEG module [59].

1014 Fig. 17. (a) Schematic diagram of the PV-TE hybrid system with an optical concentrator;

1015 (b) the structure of the PV cell within one period [50].

1016 Fig. 18. Detail illustration of the all-back-contact N-type silicon solar cell [85].

1017 Fig. 19. A tandem PV-TE hybrid device with nanostructured surface [86].

1018 Fig. 20. Power output of different thermoelement length at ambient temperature [73].

1019 Fig. 21. Hybrid system overall efficiency at different geometric values [87].

1020 Fig. 22. Electrical efficiency (PV and Hybrid) variation with concentration factor (X)

1021 and TEG hot side temperature (Th) [72].

1022 Fig. 23. Efficiency of PV-TE variation with concentration ratio for (a) c-Si PV, (b) p-

1023 Si TFPV, (c) polymer PV, (d) CIGS [70].

1024 Fig. 24. Comparison of total efficiency with solar concentration [81].

1025 Fig. 25. Optical concentration effect on (a) TEG hot side temperature (b) TEG

1026 efficiency [91].

1027 Fig. 26. Variation of system efficiency with concentration adopting different cooling

1028 medium [64].

1029 Fig. 27. Proposed Ćuk–Ćuk MIC [95].

1030 Fig. 28. Schematic diagram of the novel photovoltaic-thermoelectric system [79].

1031 Fig. 29. The proposed PV–PCM–TE hybrid system diagram [77].

1032

47
 ACCEPTED MANUSCRIPT

1033 Table lists

1034 Table 1 Temperature coefficients of different PV cell technologies [16].

1035 Table 2 Selected spectrum splitting PV-TE systems

1036 Table 3 The output power and its increase rate in HGS [5].

1037 Table 4 Selected integrated PV-TE systems.

1038 Table 5 Selected performance data for some hybrid PV-TE systems

1039 Table 6 Geometric parameters of modules investigated [73].

1040

48
 ACCEPTED MANUSCRIPT

1041
1042
1043
1044

1045

1046 Fig. 1. Splitting the solar spectrum into components for PV and thermal energy

1047 conversion [4].

1048

49
 ACCEPTED MANUSCRIPT

1049

1050 Fig. 2. Relationship between solar cell temperature and efficiency [18].

1051

1052

1053

1054

1055 Fig. 3. Impact of temperature on PV maximum power output [20].

1056

1057

50
 ACCEPTED MANUSCRIPT

1058

1059

1060 Fig. 4. The PV-TE hybrid system: (a) Schematic diagram of the hybrid system; (b)

1061 Energy flow chart of the hybrid system [35].

1062

1063

1064 Fig. 5. Schematic of the PV-TE hybrid power system [33].

1065

51
 ACCEPTED MANUSCRIPT

1066
1067 Fig. 6. Segmentation of AM1.5D spectrum in two regions for the CPV subsystem and

1068 TEG subsystem [35].

1069

1070 Fig. 7. Schematic of initial design. The first two layers from left represents MgF2

1071 (orange) and SiO2 (dark green), respectively [32].

1072

52
 ACCEPTED MANUSCRIPT

1073
1074 Fig. 8. Working temperature of the hybrid system vs. cutoff wavelength, Cg = 200, h

1075 = 10000 W/m2K-1 [35].

1076

1077

1078 Fig. 9. Hybrid system efficiency vs. cutoff wavelength for different concentration

1079 ratio, h = 10000 W/m2K-1 [35].

1080

53
 ACCEPTED MANUSCRIPT

1081

1082 Fig. 10. Comparison of the efficiency between the PV-only system and the PV-TE

1083 hybrid system [35].

1084

54
 ACCEPTED MANUSCRIPT

Sun

Solar Radiation

Photovoltaic Panel

Heat Electricity

Thermoelectric Module

Heat

1085

1086 Fig. 11. The Schematic diagram of an integrated system.

1087

1088

1089

1090 Fig. 12. Schematic overview of the PV-TE hybrid system with multiple thermoelectric
55
 ACCEPTED MANUSCRIPT

1091 generators consisting of n- and p-type doped semiconductor legs [55].

1092

1093

1094

1095 Fig. 13. Schematic diagram of heat flux in HGS [5].

1096

1097

1098

1099 Fig. 14. Schematic diagram of the cross-section of CPV/T+TE hybrid collector [58].

1100
56
 ACCEPTED MANUSCRIPT

1101

1102

1103 Fig. 15. Cross-sectional view of the evacuated tube [41].

1104

1105

1106

1107 Fig. 16. PV-TEG hybrid system (a) cavity (b) hybrid PV-TEG module [59].

1108

1109

57
 ACCEPTED MANUSCRIPT

1110

1111 Fig. 17. (a) Schematic diagram of the PV-TE hybrid system with an optical concentrator;

1112 (b) the structure of the PV cell within one period [50].

1113

1114

1115

1116 Fig. 18. Detail illustration of the all-back-contact N-type silicon solar cell [85].

1117

1118

58
 ACCEPTED MANUSCRIPT

1119
1120 Fig. 19. A tandem PV-TE hybrid device with nanostructured surface [86].

1121

1122

1123 Fig. 20. Power output of different thermoelement length at ambient temperature [73].

59
 ACCEPTED MANUSCRIPT

1124

1125 Fig. 21. Hybrid system overall efficiency at different geometric values [87].

1126

1127

1128 Fig. 22. Electrical efficiency (PV and Hybrid) variation with concentration factor (X)

1129 and TEG hot side temperature (Th) [72].

60
 ACCEPTED MANUSCRIPT

1130

1131 Fig. 23. Efficiency of PV-TE variation with concentration ratio for (a) c-Si PV, (b) p-

1132 Si TFPV, (c) polymer PV, (d) CIGS [70].

1133

1134

1135 Fig. 24. Comparison of total efficiency with solar concentration [81].

1136

61
 ACCEPTED MANUSCRIPT

1137

1138

1139

1140

1141 Fig. 25. Optical concentration effect on (a) TEG hot side temperature (b) TEG

1142 efficiency [91].

62
 ACCEPTED MANUSCRIPT

1143

1144

1145 Fig. 26. Variation of system efficiency with concentration adopting different cooling

1146 medium [64].

1147

1148 Fig. 27. Proposed Ćuk–Ćuk MIC [95].

1149

1150

1151

1152

1153

63
 ACCEPTED MANUSCRIPT

1154

1155

1156

1157 Fig. 28. Schematic diagram of the novel photovoltaic-thermoelectric system [79].

1158

1159
1160 Fig. 29. The proposed PV–PCM–TE hybrid system diagram [77].
64
 ACCEPTED MANUSCRIPT

1161
1162 Table 1 Temperature coefficients of different PV cell technologies [16].
ηTref References
o
PV technology Tref ( C) (%) -1
βref (C )
Mono-cSi 25 16-24 0.0041 [17]
Poly-cSi 25 14-18 0.004 [18]
a-Si 25 4-10 0.011 [18]
CIS 25 7-12 0.0048 [19]
CdTe 25 10-11 0.00035 [19]
1163
1164
1165 Table 2 Selected spectrum splitting PV-TE systems
PV-TE
Publicati Affiliation Efficien Referenc
on year Country PV type TE type cy es
c-Si, a-Si,
CIGS and
2018 Denmark CdTe N/A ~34% [29]
a-Si
(Amorphous
2017 Italy silicon) N/A ~10% [30]
Cu2ZnSnS4
2017 Italy (CZTS) N/A ~11.25% [30]

2017 India N/A N/A N/A [31]

Microcrystalli
ne silicon
2016 Denmark (mc-Si) N/A 20.20% [32]
a-Si
(Amorphous
2016 Denmark silicon) N/A 19.40% [32]
Monocrystalli Bi2Te3 (HZ-
2015 Qatar ne PV cell 2) N/A [6]

2014 UK/China N/A N/A 34% [33]

China/Austra Si-N (Uxuan

2012 lia Inc.) (c-Si) N/A 15.30% [34]
China/Austra SSRC-C50 (c-
2012 lia Si) N/A 16.80% [34]

China/Austra Heterostructur
2012 lia e solar cell N/A 17.97% [34]

65
 ACCEPTED MANUSCRIPT

2012 China/France GaAs CoSb3 27.49% [35]

Thin film
(Bi0.5Sb1.5T
a-Si e3 and
(Amorphous Bi2Te2.7Se0.
2012 Japan silicon) 3) N/A [36]
1166
1167
1168 Table 3 The output power and its increase rate in HGS [5].
PHGS Increase rate
Module PPV (mW) (mW) (%) Reference
STC/TEG 189 393 107.9 [5]
PV/TEG - - 14.7 [55]
DSC/TEG 9.39 13.8 46.8 [57]
DSC/TEG 39.72 44.26 23 [60]
1169
1170
1171 Table 4 Selected integrated PV-TE systems.
PV-TE
Publicatio Affiliation TE Efficienc Reference
n year Country PV type type y s

TEC1-
2018 Iran c-Si 12706 ~8.25% [41]

c-Si, a-Si, CIGS

2018 Denmark and CdTe N/A ~34.2% [29]

2018 China N/A N/A 16.70% [42]

2018 China/UK c-Si N/A 11.04% [43]

2018 China/UK GaAs N/A 22.94% [43]

2017 China CIGS N/A 21.60% [44]

2017 China Thin film silicon N/A 13.10% [44]

a-Si
(Amorphous
2017 China silicon) N/A 9.10% [45]

2017 China Polymer N/A 8.80% [45]

66
 ACCEPTED MANUSCRIPT

Iran/China/U TEC-
2017 K c-Si 1206 ~15.5 [46]
GaInP2/GaAs/G
2017 China e Bi2Te3 N/A [47]

2017 Italy a-Si N/A ~11% [30]

Cu2ZnSnS4
2017 Italy (CZTS) N/A ~12.25% [30]
Bi2Te3
(TEC1-
2017 China GaAs 03103) 19.10% [48]

2016 India Siemens SP75 N/A 5.80% [49]

2016 China GaAs N/A 18.51% [50]

Silicon (Si) solar
2016 Denmark cell Bi2Te3 ~16.6% [51]
MAX60 solar Be2Te
2016 India cell 3 N/A [52]

67
Table 5

Selected performance data for some hybrid PV-TE systems

PV Additional System Methodol System Reference

Properties TEG Properties Components Operating Conditions ogy Performance s Remarks
Tem Figu Overal
perat re of Heat l
ure Meri Sink/Co Conc Overall Power
Mate Diffe t oling Irrad Temperat . Efficien Outpu
Type rial rence (1/K) Medium Others iance ure Ratio cy t
Concentra
tor, Solar The use of TEG
Multi- Tracking Theoretica increased efficiency
junction N/A N/A 4 N/A System N/A 300K N/A l 30% N/A [38] by about 5-10%.

Near- Maximum power

Air infrared output and efficiency
Dye- cooled (NIR) of the hybrid system
sensitized copper focusing 100 were 11.2% and 10%
solar cell based lens, Hot mW/ Experime 10% 40.02 higher than the single
(DSSC) N/A N/A N/A heat sink mirror cm2 N/A N/A ntal increase mW [60] DSSC respectively.

The use of SSA

Solar increased the overall
Dye- Selective 100 12.8 efficiency of the
sensitized Absorber mW/ Experime mW/c hybrid system to
(DSSC) N/A N/A N/A N/A (SSA) cm2 N/A N/A ntal 12.80% m2 [57] 13.8%.

68
 Annual energy
increase of about
14.7% and 11% was
Multi- Air observed for Malaga
crystalline Bi2Te cooled 1000 Idealized ~130 and Utrecht
silicon 3 58oC 0.004 heat sink N/A W/m2 300K N/A model 23% W/m2 [55] respectively.
The hybrid system
obtained an efficiency
increase of 28%
Dye- 100 compared to
sensitized Bi2Te mW/ Experime 28% conventional TiO2-
(DSSC) 3 N/A N/A N/A N/A cm2 N/A N/A ntal increase N/A [65] based DSSC.

Single Skutt Concentra

junction erudit tor, Performance values of
Gallium es Spectral the hybrid system
Arsenide CoSb 468.5 Active beam 25oC 770 Numerical 0.1905 shown were in the
(GaAs) 3 K 1.4 cooling splitter (ambient) suns model 27.49% W [35] optimized range.

The shape of the heat

Water Numerical collector (Bowl-
cooled simulation shape) ensured a large
Amorphous aluminiu 60 and temperature difference
silicon thin- m fin Heat mW/ 310K Experime 393 between the TEG
film N/A 27K N/A heat sink collector cm2 (ambient) N/A ntal N/A mW [5] sides.

Output power of
Polymer Indium hybrid system
solar cell Iron heat Tin Oxide 100 11.29 increased by 46.6%
(P3HT/IC60 Bi2Te sink and (ITO) mW/ Experime mW/c compared to single PV
B) 3 9.5oC N/A Ice water glass cm2 N/A N/A ntal N/A m2 [66] system.

Hybrid system
Theoretica efficiency and power
Crystalline Thermal 100 25oC l and output increased by
Silicon (c- Bi2Te Passive conductiv mW/ (room Experime 65.2m 30% compared to PV
Si) 3 15oC N/A heat sink e paste cm2 temp.) N/A ntal 16.30% W [67] cells (12.5%, 50mW).

69
 Smaller spacing
between
Dye- Equivalent thermoelements gives
sensitized circuit 1.389 higher output power of
(DSSC) N/A 6K N/A N/A N/A N/A N/A N/A method 24.60% mW [68] TEG.
Cold For hybrid CIGS
side PV/TEG, Voc
CuInGaSe2 Bi2T temp ZnO 1000 Experime 22% increased from 0.64V
(CIGS) e3 = 5oC N/A Water nanowires W/m2 N/A N/A ntal increase N/A [69] to 0.84V
Hybrid system
Crystalline Heat Optical efficiency higher than
Silicon (c- sink with concentrat Theoretica PV efficiency (
Si) N/A N/A N/A a fan or N/A N/A 16 l model 18.6% N/A [70] ̴18.4%).
Poly
crystalline
silicon thin Heat Optical Hybrid system
film (p-Si sink with concentrat Theoretica efficiency higher than
TFPV) N/A N/A N/A a fan or N/A N/A 12 l model 14% N/A [70] PV efficiency ( ̴11%).

Heat Optical Hybrid system

sink with concentrat Theoretica efficiency higher than
Polymer PV N/A N/A N/A a fan or N/A N/A 5 l model 12% N/A [70] PV efficiency ( ̴4%)
Copper
indium Hybrid system
gallium Heat Optical efficiency higher than
selenide sink with concentrat Theoretica PV efficiency(
(CIGS) N/A N/A N/A a fan or N/A N/A 30 l model 23.5% N/A [70] ̴21.5%).

For figure of merit

27% value above 0.004K-1,
Single- nano- higher the efficiency of the
crystalline Bi2T CUO thin 1000 300K Theoretica than hybrid system can be
silicon e3 N/A 0.003 Air film W/m2 (ambient) N/A l model 13% PV [71] improved to 30%.
Experime Increase in
Multijunctio 100m ntal and concentration ratio
n (InGaP, Bi2Te Conc. W/c Simulatio 0.190 leads to decrease in
InGaAs, Ge) 3 4.5K 1 N/A Light m2 N/A 20 n 32.09% W [72] overall efficiency.
70
 Cold
side Power output can be
temp Passive Thermal Numerical increased when PV-
= water concentrat 1000 298.15K simulation 163m TE is operated in
N/A N/A 298K N/A cooling or W/m2 (ambient) N/A (Matlab) 10.20% W [73] vacuum.

Glass 1000
Bi2Te Nanoflui cover W/m 298K Theoretica Nanofluid increased
N/A 3 N/A 0.002 d (Glazed) 2 (ambient) 2 l model ~0.095 ~150W [64] efficiency by 0.48%.

Bi2Te Nanoflui 1000 298K Theoretica Nanofluid increased

N/A 3 N/A 0.002 d Unglazed W/m2 (ambient) 2 l model ~0.012 ~152W [64] efficiency by 0.35%.

Only Cold Efficiency of hybrid

p- side system (18.6%) was
type temp Air SSA, 3-D higher than efficiency
Perovskite elem = (Convect Conc. 300K numerical of single perovskite
solar cell ent 300K N/A ion) light N/A (ambient) N/A model 18.60% N/A [74] solar cell (17.8%).
22.50
%
PV cell increas
temp. e at Shorter
35oC varied 86oC thermoelement leg
(Exp between (Theor length cools the hybrid
Bi2Te erime 1000 25oC and Experime etical system better than
poly-Silicon 3 nt) N/A Water N/A W/m2 86oC N/A ntal N/A value) [75] longer leg length.

86oC 30.20 Shorter

35oC (Maximu Experime % thermoelement leg
(Exp m ntal and increas length cools the hybrid
Dye- Bi2Te erime 1000 operating Theoretica e at system better than
sensitized 3 nt) N/A Water N/A W/m2 temp.) N/A l N/A 86oC [75] longer leg length.

71
Heterojuncti
on solar cell n-
(p type type
CdTe Bi2Te
nanorod, n- 3 The efficiency value
type Bi2Te3 nanos 100m obtained (23.30%)
nanostructur tructu W/c Experime was after 1 min
e) re N/A N/A N/A FTO/CdS m2 N/A N/A ntal 23.30% N/A [76] illumination.

Optical Total efficiency

concentrat 800 reduced with increase
Water or W/m2 in temperature no
cooled (Fresnel), (max 100 Theoretica matter the value of ZT
c-Si N/A N/A 0.77 heat sink PCM ) 300K sun l model 20.10% N/A [77] used (0.77 or 1.5).
800
Water W/m2 No optical
cooled (max Theoretica concentration was
CIGS N/A N/A 0.77 heat sink ) 300K N/A l model 20.5 N/A [77] used.

Optical
concentrat Performance of the
or, PCM 800 hybrid system was
Single- Water (Phase W/m2 superior to the single
junction cooled change (max 500 Theoretica GaAs PV cell and the
GaAs N/A N/A 1.5 heat sink material) ) 425K sun l model 28.09% N/A [77] GaAs PV-TE systems.

Addition of PCM
Optical helped to maintain
concentrat 800 system at optimal
GaInP/InGa Water or W/m2 operating temperature
As/Ge (III- cooled (Fresnel), (max 500 Theoretica by reducing irradiance
V) N/A N/A 0.77 heat sink PCM ) 300K sun l model 38.90% N/A [77] fluctuation influence.

72
 p-
type
Bi0.4S
b1.6T Efficiency of the
e3 hybrid system was
n- enhanced by 20.6%
type and 725.5% in
Dye- Bi2Se comparison with
sensitized 0.15Te 1000 Experime separate DSSC and
(DSSC) 3 N/A N/A N/A N/A W/m2 323K 1 sun ntal 9.08% N/A [78] TEG respectively
Efficiency increased
by ~30% compared to
conventional PV.
Micro- System efficiency
channel varied with wind
Poly- Bi2Te Heat heat pipe 1000 25oC Numerical velocity and ambient
crystalline 3 30 N/A sink (MCHP) W/m2 (ambient) N/A model ~10.75 N/A [79] temperature.
Simulatio
Water n
cooled (LabVIE m-Si provided the best
Monocrysta aluminiu W) and 3.40m system performance
lline silicon Bi2Te m heat cRIO 920 31.8oC Experime W compared to p-Si and
(m-Si) 3 5.9oC 2.4 sink module W/m2 (PV) N/A ntal 18.93% (TEG) [80] a-Si
Simulatio
Water n
cooled (LabVIE Power gain of 7%
Polycrystall aluminiu W) and 3.44m when irradiance was
ine silicon Bi2Te m heat cRIO 1020 32.9oC Experime W increased to 1000
(p-Si) 3 5.6oC 2.4 sink module W/m2 (PV) N/A ntal 16.71% (TEG) [80] W/m2
Simulatio
Water n
cooled (LabVIE Power gain of 5%
Amorphous aluminiu W) and 1.12m when irradiance was
silicon (a- Bi2Te m heat cRIO 720 30.8oC Experime W increased to 1000
Si) 3 3.4oC 2.4 sink module W/m2 (PV) N/A ntal 2.88% (TEG) [80] W/m2

73
 Hybrid system
Alumini efficiency increased
Single um heat by ~ 3% compared to
junction Bi2Te sink with Fresnel 1000 50 Experime single CPV cell (~
GaAs 3 41oC N/A a fan lens W/m2 N/A suns ntal ~ 23.2% N/A [81] 22.5%)

74
 ACCEPTED MANUSCRIPT

Table 6 Geometric parameters of modules investigated [73].

Module type N ATE [mm2]

I 62 0.64
II 62 1.44
III 62 1.96
IV 62 2.56
V 100 2.56
VI 150 2.56
VII 200 2.56
VIII 250 2.56

Nomenclature

Cg Geometrical concentration ratio CIGS Copper Indium Gallium Selenide

hcool Heat transfer coefficient of CPV Concentrated Photovoltaic

cooling system (W/m2 K)

T Temperature (T) DSSC Dye-Sensitized Solar Cell

TC Cold side temperature (K) GaAs Gallium Arsenide

TH Hot side temperature (K) HGS Hybrid Generation System

Greek Symbols InGaAs Indium Gallium Arsenide

 Thermal conductivity (W/m K) InGaP Indium Gallium Phosphide

α Seebeck coefficient (V/K) MCHP Micro-Channel Heat Pipe

σ Electrical conductivity (S/m) mc-Si Microcrystalline Silicon

 Efficiency (%) PCM Phase Change Material

γrel Temperature coefficient of PV Photovoltaic

power (%/oC)

λs Cut-off wavelength (nm) PV-TE Photovoltaic Thermoelectric

Abbreviations SSA Solar Selective Absorber

75
 ACCEPTED MANUSCRIPT

AM Air Mass STC Silicon Thin-film Solar Cell

a-Si Amorphous Silicon TCR Thermal Contact Resistance

c-Si Crystalline Silicon TE Thermoelectric

CdTe Cadmium Telluride TEG Thermoelectric Generator

76
 ACCEPTED MANUSCRIPT

 The performance and recent advances in PV-TE are presented.

 The challenge faced in the field of PV-TE and solutions are indicated.

 Current state of art in PV-TE for electricity generation is presented.

 Optimization and development of PV-TE is demonstrated.