Full-spectrum volumetric solar thermal conversion via photonic nanofluids

Xianglei Liu and Yimin Xuan *
School of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China. E-mail: ymxuan@nuaa.edu.cn; Tel: +86 25 84891512

Received 2nd June 2017 , Accepted 27th June 2017

First published on 4th July 2017

Volumetric solar thermal conversion is an emerging technique for a plethora of applications such as solar thermal power generation, desalination, and solar water splitting. However, achieving broadband solar thermal absorption via dilute nanofluids is still a daunting challenge. In this work, full-spectrum volumetric solar thermal conversion is demonstrated over a thin layer of the proposed ‘photonic nanofluids’. The underlying mechanism is found to be the photonic superposition of core resonances, shell plasmons, and core–shell resonances at different wavelengths, whose coexistence is enabled by the broken symmetry of specially designed composite nanoparticles, i.e., Janus nanoparticles. The solar thermal conversion efficiency can be improved by 10.8% compared with core–shell nanofluids. The extinction coefficient of Janus dimers with various configurations is also investigated to unveil the effects of particle couplings. This work provides the possibility to achieve full-spectrum volumetric solar thermal conversion, and may have potential applications in efficient solar energy harvesting and utilization.

1 Introduction

The global challenges of the decreasing storage of fossil fuels, increasing energy demand, and climate changes are the driving force for people to search for renewable energy sources. It is of great significance to better utilize the clean solar energy given its abundance with the amount impinging on the earth per hour even larger than the current annual global energy consumption. Besides converting solar energy into electricity via photovoltaic cells, the conversion of sunlight into heat is also crucial to deal with the current climate change and energy security problems considering that the global energy demand for heat represents 47% of the final energy use according to a report by the International Energy Agency.1 In addition to traditional applications of solar heating and cooling,2–4 solar thermal conversion has been shown to have great potential in solar thermal power generation,5–13 desalination,14–17 and solar water splitting.18,19 Solar thermal conversion methods can be categorized into the direct way (i.e., volumetric conversion) and the indirect way relying on the absorption of black surfaces and heat transfer to a working fluid. The indirect method generally suffers from additional thermal resistance between the absorbing surface and the working fluid, which may cause a large temperature difference between them, leading to a higher radiative loss.7 Therefore, the volumetric solar thermal conversion approach is more promising to achieve higher conversion efficiencies.13,20–27 Nevertheless, the commonly employed working fluids, such as water, methanol, and ethylene glycol, are almost transparent to sunlight, inhibiting the wide applications of volumetric solar thermal conversion.

Recently, nanofluids, base fluids with nanoparticles being dispersed inside, have been demonstrated to be promising alternatives.5,20,28 When the incident light's wavelength matches with the collective oscillation frequency of free electrons inside the nanoparticle, the well-known localized surface plasmon resonances (LSPRs) occur. The local electric fields in the vicinity of the nanoparticle are greatly enhanced, leading to a tremendous increase of the absorption and scattering cross-section of the nanoparticle. Subsequently, the nanofluid's absorption coefficient is improved, resulting in a higher solar thermal conversion efficiency. Tyagi et al.23 concluded that the solar absorption of pure water can be improved by more than nine times by adding aluminum nanoparticles, and the volumetric solar thermal conversion efficiency via aluminum nanofluids is about ten percent higher than that of conventional flat-plate solar collectors under similar operating conditions. Although the absorption of silver or aluminum nanoparticles is enhanced prominently at LSPRs, the absorption band is narrow especially when the material loss is small. Plasmonic materials with larger material losses or damping factors can help to broaden the absorption peak,29 but the absolute absorption cross-section drops dramatically. As a result, the particle concentration may have to increase to make up for the weak absorption, while a higher particle concentration will cause other problems, like larger pipe work being required due to the increase in flow viscosity, pipe wear, clogging, and so on. Another way to overcome the narrow-band nature of plasmon-enhanced absorption is adjusting the configuration and constitution of nanoparticles. It has been experimentally confirmed that silver nanoshells have a relatively broader plasmon-enhanced absorptance due to the hybridization of particle plasmons and cavity plasmons.30 However, achieving broadband especially full-spectrum volumetric solar thermal absorption is still a daunting challenge.

In this paper, the full-spectrum volumetric solar thermal absorption is theoretically demonstrated based on a novel design. The working principle of the proposed ‘photonic nanofluids’ lies in the integration of various resonances for different nanoparticles into one platform, which is enabled by the asymmetry of Janus nanoparticles. The narrowband nature of single resonances is thus overcome, leading to an efficient solar thermal conversion. The extinction efficiency of GZO (Gallium doped Zinc Oxide) nanospheres, GZO/Au core–shells, and GZO/Au Janus particles is investigated first, and the underlying mechanism for every absorption peak is unveiled. The absorbed solar power spectrum and solar thermal conversion efficiency of different types of particles are then discussed and compared. The extinction coefficient of various Janus dimers is analyzed at last to investigate the effects of particle couplings on the extinction of single Janus nanoparticles.

2 Theory and methods

The schematic of volumetric solar thermal conversion, including a glass container and photonic nanofluids consisting of nanoparticles and base fluids, is given in Fig. 1. As can be clearly seen, the employed nanoparticle is a composite consisting of a half of the GZO sphere on one side and a half of the GZO/Au core–shell on the other side. It is rotationally symmetric around the axis denoted by the dot-dashed line in Fig. 1. This special kind of nanoparticle is called Janus nanoparticles, named after the roman god Janus with two faces representing beginning and ending. Janus nanoparticles are usually composed of chemically or biologically different regions at two sides, and have attracted much attention in nanomotors, biological sensors, drug delivery, and electronic displays.31–33 The unique configuration and the associated broken symmetry of Janus nanoparticles will be employed to achieve full-spectrum solar thermal absorption.
image file: c7nr03912c-f1.tif
Fig. 1 Schematics of volumetric solar thermal conversion of photonic nanofluids. r1 and r2 are the radius of the GZO core and Au shell, respectively. The dot-dashed line represents the rotational symmetry axis of the GZO/Au Janus nanoparticle.

Before calculating the solar thermal conversion efficiency of nanofluids, we need to predict the extinction coefficient of a single Janus nanoparticle first. As shown in Fig. 1, r1 and r2 are the inner and outer radius of the shell, respectively, and are set to be 20 nm and 24 nm, respectively, as default. The governing Maxwell's equations are solved based on the finite-difference time-domain (FDTD) method, which is an explicit time marching algorithm running on a grid discretized on both time and space. It has been widely used due to its capability to deal with the radiative properties of nanoparticles with arbitrary geometry.34 A mesh size of 0.5 nm, for which converging results have been obtained, is used in calculating the extinction coefficients. The dielectric functions of both GZO and Au are described by a Drude-Lorentz oscillator model image file: c7nr03912c-t1.tif, where ω is the angular frequency. The Drude model, the second term at the right hand side, describes the response of conduction electrons with the electromagnetic waves. The Lorentz oscillator models consider the contribution of valence electrons. All the corresponding parameters (N is equal to 1 and 5 for GZO and Au, respectively) are obtained by fitting experiment data.35,36 The refractive index of the environment surrounding the nanoparticles is taken as 1.33, which is close to the common working fluids like water and methanol.37 The particle extinction coefficient can be defined as

image file: c7nr03912c-t2.tif(1)
where Cabs and Csca are the absorption and scattering cross-sections, respectively, obtained from FDTD approaches. Vp, the volume of particles, is equal to 4r31/3, 4r32/3, and 2(r31 + r32/3) for the GZO sphere, GZO/Au core–shell, and Janus nanoparticles, respectively.

3 Results and discussion

3.1 Broadband average extinction coefficient of Janus nanoparticles

Nanoparticles play a vital role in enhancing the solar absorption of the base fluid, and nanoparticles made of noble metals are popular candidates for solar thermal conversion. The wavelength of LSPRs usually lies in the visible and ultraviolet region while 45.3% of the solar energy (AM 1.5) is from the near-infrared part. Therefore, to achieve full-spectrum solar thermal conversion, it is desired to redshift the LSPRs. Doped semiconductors are known to have a relatively lower density of free electrons, and the resonant wavelength is tunable from near-infrared to mid-infrared by controlling the doping level.35,38,39 GZO is considered here due to its metallic behavior and low losses in the near-infrared wavelength range. As shown in Fig. 2, the particle extinction coefficient of the GZO sphere with a radius of 20 nm exhibits two peaks at 243 nm and 1710 nm, respectively. Both of them are caused by the excitation of LSPRs, and both peak wavelengths agree well with the excitation condition of LSPRs under the Rayleigh approximation, i.e., ε(ω) = −2εfluid. This is expected given that the incident wavelength is more than one order of magnitude larger than the radius. The first (244 nm) and second (1710 nm) absorption peaks are due to the resonant oscillation of valence electrons and free electrons, respectively. The location of the first peak is close to the Lorentz resonance wavelength of GZO, where the permittivity has an anomalous dispersion. That's why there are two wavelength positions satisfying ε(ω) = −2εfluid, leading to substantial extinction coefficients in the ultraviolet and near-infrared wavelength range. Nevertheless, the whole visible region exhibits a weak extinction. Given the strong light–matter interaction of noble metals in the visible wavelength range, depositing a thin layer of Au to make GZO/Au core–shells may affect the extinction coefficient of visible light. Indeed, the extinction coefficient of the GZO/Au core–shell peaks at 248 nm and 738 nm due to the excitation of the antisymmetric and symmetric mode of the hybridization of cavity plasmons and particle plasmons, respectively.40 Due to these two resonant peaks, the extinction coefficient of the whole visible light can be improved by more than one order of magnitude as can be clearly seen in Fig. 2. Nevertheless, it is surprising to see that the resonant absorption of the GZO core in the near-infrared region disappears after the deposition of the Au clad, precluding the realization of full-spectrum solar thermal conversion.
image file: c7nr03912c-f2.tif
Fig. 2 Extinction coefficient for different types of nanoparticles with r1 = 20 nm and r2 = 24 nm as default. The double sided arrow denotes the direction of electric fields. For Janus nanoparticles, the presented extinction coefficient is averaged considering different directions of electric fields.

Considering that the GZO sphere and GZO/Au core–shell particles have high extinction coefficients in the near-infrared and visible range, respectively, one may wonder whether it is possible to combine the two complementary extinction properties together. In the following part, we will consider the possibility by employing GZO/Au Janus particles with only half of the core covered by the clad, which can be readily obtained by the current nanofabrication techniques.32 Unlike isotropic spherical nanoparticles, the extinction efficiency of Janus nanoparticles depends on the relative orientation of the incident electric field (or polarization) with the rotation axis of the Janus nanoparticle. The average extinction coefficient can be obtained as

image file: c7nr03912c-t3.tif(2)
where σp,para and σp,perp represent the extinction coefficient when the electric field is parallel and perpendicular to the rotation axis, respectively. As can be clearly seen from Fig. 2, the average extinction coefficient of the GZO/Au Janus particle possesses several peaks, resulting in a broadband extinction coefficient across almost the whole visible and near-infrared wavelength range, making it promising in full-spectrum solar thermal conversion. Similar results can be obtained by replacing Au with cheap metals, such as Al, though the location of resonant peaks may change to some extent. Note that the extinction coefficient depends on the direction of the electric field, so that some resonant peaks may show up only for certain orientations of Janus particles. For example, the peaks at 650 nm and 878 m are excited when the electric field is parallel and perpendicular to the rotation axis, respectively. More discussions on individual σp,para and σp,perp will be given in the next section.

3.2 Resonant extinction mechanism of Janus nanoparticles

To unveil the resonant extinction mechanism of Janus nanoparticles, the separate components σp,perp and σp,para need detailed discussions. As given in Fig. 3a, σp,perp has three peaks at 244 nm, 878 nm, and 1625 nm, respectively. Since the induced charge distributions can provide a powerful tool for the microscopic interpretation of the resonant extinction, Fig. 3a, b, and c show the charge distributions for the three peaks at their corresponding resonant wavelengths. The induced charge is obtained as ρ = ε0∇·E, where ε0 is the vacuum permittivity. A mesh size of 0.4 nm is used for the charge density distribution as default. As shown in Fig. 3b, the induced charges at the inner and outer surface of the Au shell have opposite signs, suggesting the excitation of the antisymmetric mode of core–shell plasmons (hybridization of cavity plasmons and particle plasmons41). The second peak at 878 nm is dominated by the symmetric hybridized plasmons considering the symmetric distribution of induced charges in Fig. 3c. The charge distribution of the right half GZO core for the third resonant peak at 1625 nm resembles a dipole, so that it can be attributed to the excitation of LSPRs of the GZO core. Also note that the left GZO/Au shell seems to have opposite dipole moments, which may inhibit the GZO core plasmons to some extent. This can be confirmed by realizing that the extinction coefficient at 1625 nm is much smaller than that of a single GZO sphere at a resonance wavelength of 1710 nm as shown in Fig. 2.
image file: c7nr03912c-f3.tif
Fig. 3 Extinction coefficients (a) and induced charge distributions at the peak resonance wavelength of (b) 244 nm, (c) 878 nm, and (d) 1625 nm for the electric field perpendicular to the rotation axis. “+” and “−” represent positive and negative charges, respectively.

For the scenario with the electric field parallel to the rotation axis, the particle extinction coefficient is plotted in Fig. 4a. Similarly, there are also three resonant peaks but at different wavelengths. For the resonant peak at 245 nm, the charge is positive in the whole core while being negative on the outer shell surface as can be seen from Fig. 4c. Subsequently, both the core and the half shell play an important role in exciting this resonance. Interestingly, this resonant wavelength is almost coincident with the first peak of σp,perp at 244 nm, and the LSPRs of a single GZO at 243 nm. This is due to the anomalous dispersion of GZO around 244 nm, where its permittivity has a giant change from negative values to large positive values in a very narrow range. Positive and negative charges accumulate at the left and right end of the half Au shell in Fig. 4c, subsequently the resonant peak at 650 nm can be attributed to the LSPRs of the Au shell itself. As demonstrated in Fig. 4d, the right and left half GZO core are featured with negative and positive charges, respectively. Due to its large electric susceptibility in the absolute form, the right end of the Au shell has an even higher density of positive charges compared with the left half GZO. Nevertheless, the Au shell behaves just as a good conductor, and the resonance is still due to the LSPRs of the core although the charge distribution is modified by the half shell to some extent. The resonance wavelength lies at 1756 nm, which is very close to the LSPRs of the single GZO core, further confirming the dominant role of GZO in exciting this resonance.

image file: c7nr03912c-f4.tif
Fig. 4 Extinction coefficients (a) and induced charge distributions at the peak resonance wavelength of (b) 245 nm, (c) 650 nm, and (d) 1756 nm for the electric field parallel to the rotation axis.

3.3 Full-spectrum volumetric solar thermal conversion

After obtaining broadband extinction coefficients via the superposition of various resonances enabled by Janus nanoparticles and analyzing their resonant extinction mechanism, now it is time to investigate the solar thermal conversion properties of the proposed photonic nanofluids made of Janus nanoparticles. To minimize the possible abrasion or clogging problems caused by particle agglomerations,42 the volume concentration of nanoparticles fv should be as small as possible if the solar thermal conversion performance meets the needs. Therefore, fv is set to be a small value of 5 × 10−4 as default in this paper. According to Beer's law, the spectral absorptance of light over nanofluids with a thickness of L can be written as (1 − exp(σtL)), where the total extinction coefficient image file: c7nr03912c-t4.tif with κ as the imaginary refraction index of the base fluid. The assumption of independent scattering is adopted here given its very low volume concentration according to ref. 42 and 43. The absorbed solar power spectrum by a thin layer of photonic nanofluids without considering interface reflection losses can be expressed as
IA(λ) = (1 − exp(σtL))IAM1.5(3)
where IAM1.5 is the AM 1.5 solar power spectrum, and the solar thermal conversion efficiency can thus be obtained as image file: c7nr03912c-t5.tif. Different base fluids like water, methanol, and ethylene glycol may be used to achieve full-spectrum solar thermal conversion in systems with different temperatures. Subsequently, to reduce the distractions caused by different extinction coefficients of the base fluid although actually small, the fluid extinction coefficient 4πκ/λ is set to be zero temporally.

Fig. 5 shows the absorbed solar power spectrum of 1 mm thick nanofluids where different types of nanoparticles may be employed. For GZO nanospheres, the absorbed solar power approaches the incident amount when the wavelength is larger than 1.1 μm, but is small in the short visible range. As a result, the solar thermal conversion efficiency is only 38.4%. The GZO/Au core–shell on the other hand has a high and low absorptance when the incident wavelength is smaller and larger than 1.1 μm, respectively. However, the spectral solar power is much higher in the short wavelength range below 1.1 μm, so that the conversion efficiency can increase to 88.2% after covering a thin layer of the Au clad on the GZO sphere. For GZO/Au Janus nanoparticles, as can be clearly seen from Fig. 5, the absorbed power is close to the incident one for the whole solar spectrum, proving its capability of full-spectrum volumetric solar thermal conversion. This agrees with the broadband extinction coefficient in Fig. 2. The conversion efficiency becomes as high as 97.7%, which is 10.8% higher than that of core–shell nanofluids. Similar phenomena occur if replacing Au by other metals, for example, the solar thermal conversion efficiency of GZO/Al core–shell nanofluids can be enhanced by 11.1% by employing the corresponding GZO/Al Janus nanoparticles. The presence of the base fluid's intrinsic extinctions may affect the conversion efficiency. To check whether this effect is large or not, one base fluid with high fluid extinction coefficients in the infrared range, i.e., water, is chosen, whose extinction coefficients are obtained from ref. 37. The conversion efficiencies for nanofluids made of the GZO sphere, the GZO/Au core–shell, and the GZO/Au Janus particles become 39.0%, 94.8%, and 98.9%, respectively. Still, photonic nanofluids made of Janus nanoparticles have the highest solar thermal conversion efficiency, not to mention working fluids with smaller extinction coefficients. Also note that specific geometry parameters and material types can be further optimized for a given base fluid to achieve the best solar thermal conversion performances.

image file: c7nr03912c-f5.tif
Fig. 5 Absorbed solar power spectrum for different nanofluids made of GZO spheres, GZO/Au core–shells, or GZO/Au Janus nanoparticles. The thickness of nanofluids and the volume concentration of nanoparticles are 1 mm and 5 × 10−4, respectively.

3.4 Extinction coefficient of Janus dimers

Before concluding this work, we would like to show the effects of coupling between Janus nanoparticles separated by a nanoscale distance on the extinction coefficient since some particles may aggregate in practical applications of photonic nanofluids. The simplest case with only two Janus nanoparticles involved, i.e., the Janus dimer, is considered here to make the analysis simple. Fig. 6a investigates the influence of the coupling between σp,perp on the extinction coefficient of a Janus dimer with the gap distance d = 5 nm (set as a default value). When the electric field is parallel and perpendicular to the inter-particle axis, the symmetric mode of core–shell plasmons redshifts and blueshifts, respectively. Besides, the extinction coefficient is further enhanced or reduced when the redshift and blueshift occur, respectively. Similar observations have been reported for plasmonic dimers consisting of noble spheres.44–47 The underlying mechanism can be attributed to the weakened and enhanced repulsive forces for the surface charges, so that the energy (or frequency) required to drive the coupled plasmons becomes lower and higher for the electric field parallel and perpendicular to the inter-particle axis, respectively.45 Therefore, the corresponding resonance wavelength becomes longer and shorter accordingly. The near-infrared extinction peak has a similar but less obvious shift trend. The reason is that this peak is dominated by the GZO core resonance while the actual distance between the cores is 13 nm, which is much larger than d. Subsequently, the coupling between the two cores should be weak, making the peak location change only slightly.
image file: c7nr03912c-f6.tif
Fig. 6 Extinction coefficients of Janus dimers (two Janus nanoparticles in close proximity) by considering the resonance couplings for (a) σp,perp and (b) σp,para. The black solid line represents the extinction coefficient of a single Janus nanoparticle for comparison. The minimum gap spacing between nanoparticles d is set to be 5 nm. The double sided arrow denotes the direction of electric fields.

Fig. 6b shows the extinction coefficient of Janus dimers considering the coupling between σp,para. The principle that the redshift (blueshift) occurs when the electric field is along (perpendicular to) the inter-particle axis is applicable again for both the shell resonance at 650 nm and the core resonance at 1756 nm. The presence of one Janus nanoparticle in the close proximity of another one modifies the shell resonance only slightly except when the shell–shell distance is equal to the minimum distance d and the electric field is along the inter-particle axis, as can be seen from the green dashed line. A similar phenomenon occurs for the core resonance as being denoted by the red dotted line. The coupling between σp,para and σp,perp is also investigated but is found to be so small that the average extinction coefficient of such a dimer is almost the same as that of a single Janus nanoparticle, thus it is not shown here. For other configurations of a Janus dimer, the effect of particle couplings may also be small due to the mismatch of extinctions of each single Janus nanoparticle just as the case of very weak coupling between σp,para and σp,perp. The discussion in this section helps to further examine the broadband extinction mechanism of Janus nanoparticles and guide the design of photonic nanofluids by demonstrating how the extinction coefficient changes if nanoparticles aggregate to form dimers.

4 Conclusions

In summary, the full-spectrum solar thermal conversion is theoretically demonstrated based on a new design of photonic nanofluids made of composite Janus nanoparticles. When the electric field is perpendicular to the rotation axis, the antisymmetric and symmetric modes of hybridized core–shell plasmons and the core resonance are excited. For the electric field parallel to the rotation axis, both the shell resonance and the core plasmon are supported. The above resonant extinctions at different wavelengths compensate for the weak absorption of GZO spheres in the visible range and that of GZO/Au core–shells in the near-infrared region, leading to the full-spectrum solar thermal conversion. The solar thermal conversion efficiency of Janus photonic nanofluids can be 10.8% and 154% higher than that of the GZO/Au core–shell and GZO sphere nanofluids, respectively, at the same nanofluids’ thickness of 1 mm and the particle volume concentration of 5 × 10−4. Similar phenomena can occur by using different metal materials or working fluids, making the proposed design have general applicability. The effect of particle couplings on the extinction coefficient of Janus dimers with various configurations is investigated in the last section. Resonant extinction peaks redshift and blueshift when the electric field is parallel and perpendicular to the inter-particle axis, respectively, for two Janus nanoparticles with matching resonances, otherwise, the coupling effect is weak. This work will find applications in designing highly efficient volumetric solar collectors for solar power plants, solar desalination, and solar water splitting.


This work is mainly supported by the National Natural Science Foundation of China (Grant No. 51590901). XL also wants to acknowledge the support from Fundamental Research Funds for the Central Universities (Grant No. 56XIA17001).


  1. IEA, Technology Roadmap: Solar Heating and Cooling, 2012 Search PubMed .
  2. T. S. Ge, Y. J. Dai and R. Z. Wang, Renewable Sustainable Energy Rev., 2014, 39, 476–497 CrossRef .
  3. A. Allouhi, T. Kousksou, A. Jamil, P. Bruel, Y. Mourad and Y. Zeraouli, Renewable Sustainable Energy Rev., 2015, 44, 159–181 CrossRef .
  4. Y. Dai, X. Li and R. Wang, Energy Proc., 2015, 70, 418–426 CrossRef .
  5. R. Taylor, S. Coulombe, T. Otanicar, P. Phelan, A. Gunawan, W. Lv, G. Rosengarten, R. Prasher and H. Tyagi, J. Appl. Phys., 2013, 113, 011301 CrossRef .
  6. Q. Zhu and Z. M. Zhang, Radiative Properties of Micro/Nanoscale Particles in Dispersions for Photothermal Energy Conversion, CRC Press/Taylor&Francis Group, 2012 Search PubMed .
  7. A. Lenert and E. N. Wang, Sol. Energy, 2012, 86, 253–265 CrossRef CAS .
  8. Z. Luo, C. Wang, W. Wei, G. Xiao and M. Ni, Int. J. Heat Mass Transfer, 2014, 75, 262–271 CrossRef .
  9. P. Phelan, T. Otanicar, R. Taylor and H. Tyagi, J. Therm. Sci. Eng. Appl., 2013, 5, 021003 CrossRef .
  10. Y. Tian and C. Y. Zhao, Appl. Energy, 2013, 104, 538–553 CrossRef CAS .
  11. R. Z. Wang and X. Q. Zhai, Energy, 2010, 35, 4407–4416 CrossRef .
  12. L. Zhang, R. Li, B. Tang and P. Wang, Nanoscale, 2016, 8, 14600–14607 RSC .
  13. C.-C. Lai, W.-C. Chang, W.-L. Hu, Z. M. Wang, M.-C. Lu and Y.-L. Chueh, Nanoscale, 2014, 6, 4555–4559 RSC .
  14. G. Xiao, X. Wang, M. Ni, F. Wang, W. Zhu, Z. Luo and K. Cen, Appl. Energy, 2013, 103, 642–652 CrossRef CAS .
  15. O. Neumann, A. S. Urban, J. Day, S. Lal, P. Nordlander and N. J. Halas, ACS Nano, 2012, 7, 42–49 CrossRef PubMed .
  16. L. Zhou, Y. Tan, J. Wang, W. Xu, Y. Yuan, W. Cai, S. Zhu and J. Zhu, Nat. Photonics, 2016, 10, 393–398 CrossRef CAS .
  17. X. Li, G. Yuan, Z. Wang, H. Li and Z. Xu, Desalination, 2014, 351, 1–8 CrossRef CAS .
  18. T. Kodama and N. Gokon, Chem. Rev., 2007, 107, 4048–4077 CrossRef CAS PubMed .
  19. H. Duan and Y. Xuan, Appl. Energy, 2014, 114, 22–29 CrossRef CAS .
  20. V. Khullar, H. Tyagi, P. E. Phelan, T. P. Otanicar, H. Singh and R. A. Taylor, J. Nanotechnol. Eng. Med., 2012, 3, 031003 CrossRef .
  21. V. Khullar, H. Tyagi, N. Hordy, T. P. Otanicar, Y. Hewakuruppu, P. Modi and R. A. Taylor, Int. J. Heat Mass Transfer, 2014, 77, 377–384 CrossRef CAS .
  22. T. P. Otanicar, P. E. Phelan, R. S. Prasher, G. Rosengarten and R. A. Taylor, J. Renewable Sustainable Energy, 2010, 2, 033102 CrossRef .
  23. H. Tyagi, P. Phelan and R. Prasher, J. Sol. Energy Eng., 2009, 131, 143–155 CrossRef .
  24. Y.-T. Yen, C.-W. Chen, M. Fang, Y.-Z. Chen, C.-C. Lai, C.-H. Hsu, Y.-C. Wang, H. Lin, C.-H. Shen, J.-M. Shieh, J. C. Ho and Y.-L. Chueh, Nano Energy, 2015, 15, 470–478 CrossRef CAS .
  25. Q. Li, C. Zheng, S. Mesgari, Y. L. Hewkuruppu, N. Hjerrild, F. Crisostomo, G. Rosengarten, J. A. Scott and R. A. Taylor, Sol. Energy, 2016, 136, 349–364 CrossRef CAS .
  26. G. Ni, N. Miljkovic, H. Ghasemi, X. Huang, S. V. Boriskina, C.-T. Lin, J. Wang, Y. Xu, M. M. Rahman, T. Zhang and G. Chen, Nano Energy, 2015, 17, 290–301 CrossRef CAS .
  27. B. J. Lee, K. Park, T. Walsh and L. Xu, J. Sol. Energy Eng., 2012, 134, 100–110 CrossRef .
  28. J. Zeng, Y. Xuan and H. Duan, Sol. Energy Mater. Sol. Cells, 2016, 157, 930–936 CrossRef CAS .
  29. S. Ishii, R. P. Sugavaneshwar and T. Nagao, J. Phys. Chem. C, 2016, 120, 2343–2348 CAS .
  30. Y. Xuan, H. Duan and Q. Li, RSC Adv., 2014, 4, 16206–16213 RSC .
  31. M. Lattuada and T. A. Hatton, Nano Today, 2011, 6, 286–308 CrossRef CAS .
  32. A. Walther and A. H. Müller, Chem. Rev., 2013, 113, 5194–5261 CrossRef CAS PubMed .
  33. Q. Yang, M. H. de Vries, F. Picchioni and K. Loos, Nanoscale, 2013, 5, 10420–10427 RSC .
  34. C. Oubre and P. Nordlander, J. Phys. Chem. B, 2004, 108, 17740–17747 CrossRef CAS .
  35. J. Kim, G. V. Naik, N. K. Emani, U. Guler and A. Boltasseva, IEEE J. Sel. Top. Quantum Electron., 2013, 19, 4601907 CrossRef .
  36. A. D. Rakic, A. B. Djurisic, J. M. Elazar and M. L. Majewski, Appl. Opt., 1998, 37, 5271–5283 CrossRef CAS PubMed .
  37. E. D. Palik, Handbook of Optical Constants of Solids, Academic Press, 1998 Search PubMed .
  38. Z. M. Zhang, Nano/Microscale Heat Transfer, McGraw-Hill, New York, 2007 Search PubMed .
  39. L.-W. Chou, N. Shin, S. V. Sivaram and M. A. Filler, J. Am. Chem. Soc., 2012, 134, 16155–16158 CrossRef CAS PubMed .
  40. X. L. Liu and Y. M. Xuan, Sol. Energy, 2017, 146, 503–510 CrossRef CAS .
  41. E. Prodan and P. Nordlander, J. Chem. Phys., 2004, 120, 5444–5454 CrossRef CAS PubMed .
  42. R. A. Taylor, P. E. Phelan, T. P. Otanicar, R. Adrian and R. Prasher, Nanoscale Res. Lett., 2011, 6, 1–11 CrossRef PubMed .
  43. C. Tien, ASME J. Heat Transfer, 1988, 110, 1230–1242 CrossRef CAS .
  44. P. K. Jain and M. A. El-Sayed, Chem. Phys. Lett., 2010, 487, 153–164 CrossRef CAS .
  45. W. Rechberger, A. Hohenau, A. Leitner, J. Krenn, B. Lamprecht and F. Aussenegg, Opt. Commun., 2003, 220, 137–141 CrossRef CAS .
  46. L. Gunnarsson, T. Rindzevicius, J. Prikulis, B. Kasemo, M. Käll, S. Zou and G. C. Schatz, J. Phys. Chem. B, 2005, 109, 1079–1087 CrossRef CAS PubMed .
  47. I. Romero, J. Aizpurua, G. W. Bryant and F. J. G. De Abajo, Opt. Express, 2006, 14, 9988–9999 CrossRef PubMed .

This journal is © The Royal Society of Chemistry 2017