Non-noble metal based broadband photothermal absorbers for cost effective interfacial solar thermal conversion

1 Introduction

Interfacial solar vapor generation (ISVG) has triggered the revival of solar thermal conversion and related applications, such as solar desalination [1], [2], [3], [4], solar water purification and sterilization [5], [6], [7]. Plasmonic absorbers [8], [9], [10], [11], [12], [13] have been widely investigated and regarded as one of the most promising photothermal materials for ISVG due to the unique capability of subwavelength light confinement as well as flexibility of light absorption manipulation [10]. However, compared with conventional carbon based photothermal materials [14], [15], [16], [17], [18], ideal plasmonic metals with relatively high figure of merit of ε^′/ε″ (the ratio of real and imaginary part of dielectric constant of materials) are much more challenging for broadband light absorption [19]. Careful designs of gradually steered continuous metasurfaces and/or multiple discontinuous resonant building blocks [8, 20] are the current main pathways for enabling broadband light harvesting. In the past few years, Zhou et al. reported a new three dimensional gold nanoparticles within a nanoporous substrate with physical vapor deposition process [11], in which a high density of plasmon resonances are realized in close-packed particle assembly for broadband absorption. Such kind of metallic structures possess not only high absorption efficiency but also a couple of unique advantages stemmed from the high porosity, such as reduced surface impedance, enhanced internal light scattering as well as sufficient water supply channels, making them good candidates for interfacial solar thermal conversion. However, most of them are still suffering from high cost of raw materials. In recent years, many metals such as Ti, V, W, Ni [19, 21], [22], [23], [24] are found possessing plasmonic behavior and the intrinsic material absorption losses of them in a wide frequency range can naturally provide a broadband absorption response. In this work, we demonstrate a new type of photothermal absorber via assembling the non-noble metal nickel (Ni) particles into the three-dimensional porous structure. Thanks to both the ultrabroadband intrinsic optical absorption from the single electron excitation of the material and the weak plasmonic absorption from the structure, combined with the unique high porosity structure reduced surface reflection, our Ni nanoparticle based photothermal absorbers can enable solar weighted solar absorption up to ∼97% with considerable tolerance of fabrication process. The proposed structure further achieves an interfacial solar evaporation with a considerable evaporation rate of ∼0.9 kg m⁻² h⁻¹. Besides, we demonstrate that the as-repaired absorber can serve a potential candidate for cost effective and personalized solar water purification.

2 Materials and methods

Ni nanoparticle absorbers based on porous substrate including a porous anodic aluminum oxide membrane (AAO) substrate at the bottom, randomly distributed Ni nanoparticles (Ni NPs) within the nanopores in the middle and a continuous but porous Ni film on the depositing surface is schematically shown in Figure 1A. Both the single electron excitation (due to the intrinsic interband transition, Figure 1B) and collective excitation (due to the particle plasmon, Figure 1C) contribute to its effective light harvesting. The fabrication process of this plasmonic absorber contains two steps [11]: At first, AAO is prepared with aluminum foil(purity of ∼99.997%, thickness of ∼0.25 mm) through a two-step anodic oxidation process. The average pore size of the AAO can be well controlled via the applied voltage. The periodicity of these nanopores and the thickness (∼1s μm) of the AAO can be controlled through a variety of processing methods. As a result, a transparent and porous AAO (Figure 1D) with nanoscale pores (diameter of ∼200–350 nm, period of ∼ 450 nm, thickness of ∼60 μm) is obtained. Then an ion sputtering deposition process is employed for Ni particle deposition. The nanoporous AAO is used as a scaffold of Ni NPs through the ion sputtering process (Leica EM ACE600), after which Ni NPs with diameters ∼24 nm can be observed form the pore entrances. Because of the unique high porosity of the AAO substrate, the particles around the entrance are larger while those deep inside the channels are smaller. By putting a flat fused silica substrate inside the deposition chamber, we estimate that the effective deposition thickness is ∼60 nm. Figure 1E depicts the as-prepared Ni NPs/AAO absorber, which turns to be completely black, indicating the pronounced light absorption in visible range. The detailed cross sectional and top view SEM images are shown in Figure 1F–G respectively, revealing the penetration depth of Ni NPs ∼<1 μm.

Figure 1:

(A) Schematic of Ni nanoparticle absorbers based on porous substrate induced physical self-assembly. (B) Single electron excitation. E_F is Fermi level. (C) Collective oscillation of free electrons. E is the electric field. (D-E) Optical images of baer AAO and Ni NPs/AAO before and after the ion sputtering process. (F-G) Cross sectional (F) and bottom (G) view SEM images of the prepared Ni NPs/AAO structure, respectively.

3 Results and discussion

To reveal the dual-mechanism enabled broadband light absorption of the proposed Ni NP/AAO structure, we will theoretically analyze the optical absorption processes from both microscopic and macroscopic viewpoints, which fully consider the unique material and structure properties. In the microscopic view, light harvesting of metals is mostly related to the light excitation of the high density of free electrons in the close proximity of metal surfaces, which can be divided into single electron excitation and collective excitation. On the one hand, when light illuminates on the metal surfaces, free electrons nearby the Fermi energy can obtain energy from the penetrated photons, resulting in the single electron excitation through interband (Figure 1B) and/or intraband transition with phonon involved. On the other hand, the incident light waves will couple with the collective oscillation of free electrons (Figure 1C), leading to the plasmon excitations such as bulk plasmon (BP), surface plasmon polariton (SPP) and localized surface plasmon (LSP), which can decay either radiatively or non-radiatively by creation of electron-hole pairs. Basically, light absorption by the former is intrinsically dominating by the material properties, while absorption by the latter can be steered by microstructures (especially for the LSPs). In the macroscopic view, an ideal absorber should be of not only high absorption capability (large imaginary part of dielectric function ε^″) but also of well-matched impedance with environment so that strong electric field E can penetrate sufficiently into the volume V of materials, which determines the absorption power P_abs inside metals as [19]:

We first demonstrate that the nickel metal is more superior for broadband light harvesting because of the single electron excitation via interband transition determined by the band structure. Figure 2A shows the schematic band diagrams of three representative plasmonic metals Ag, Al and Ni, respectively. It is clearly demonstrated that, Ni metals possesses narrow but partially filled d band that overlaps with sp band [25], [26], [27]. In addition, the bandwidth of the d band is much narrower than sp band, which means much higher density of free electrons around Fermi level and thus stronger interband transition. Therefore, the intrinsic optical absorption originating from single electron excitation can be much broader in wavelength range compared with conventional plasmonic metals Ag and Al. The band structures and density of states of free electrons around the Fermi levels reveals that interband transition can enable light absorption from visible up to mid-infrared region (0.4∼4.6 μm). More quantitative band structures and density of states of free electrons around the Fermi levels calculated by the first principle method (Supporting Information, Note S1, Figure S1)

Figure 2:

(A) Schematic diagram of band structures of Ag, Al and Ni, respectively. (B, C) Dielectric functions as a function of wavelength for the three metals. (D) Calculated figure of merit ε^″/ε^′ and (E) Zero order reflectance from metal/air interface as a function of light wavelength for the three metals.

We further compare the experimental dielectric function of three representative plasmonic metals based on the Palik book [28], as shown in Figure 2B–C. Note that the most promising plasmonic metal Ag has larger real part |ε^′| and smaller imaginary part ε^″, indicating higher figure of merits of SPP or LSP [29]. Quite differently, Ni is far from a plasmonic metal (shorter propagation length of plasmon, lower near field confinement) as indicated by the reverse figure of merit in Figure 2C. However, such intrinsic properties can make Ni metal possesses lower absorption coefficient and absorption depth (Supporting Information, Figure S3), which means larger Ni NPs with deeper light penetration and thus higher more hot carriers can contribute to the photothermal process. In addition, because of the weaker spatial confinement in the transverse direction versus the metallic interface, the optical loss from the surface reflection of Ni is much lower compared with conventional photothermal metals (Figure 2E), indicating the material reduced surface reflectance. It means that Ni based metamaterials is more favorable for ultrabroadband absorption from the view point of intrinsic material properties.

Except for the strong absorption capability stemmed from the intrinsic material nature, the highly porous three-dimensional nanoparticle structure also plays important roles to further reduce surface reflection and promote light broadband absorption. Figure 3A–B illustrate the two key features of the prepared Ni/AAO structure, i. e., the close packed particle distribution profile and high porosity as well as low filling ratio of metals. The calculated absorption cross section curves of eight close packed metal nanoparticles (normalized by the geometry cross section) are demonstrated in Figure 3C. One may find that, the plasmonic effect may play some roles in the visible wavelength range, which is still much weaker than conventional plasmonic metals. However, thanks to the intrinsic absorption from the interband transition that can be enhanced by the multiple scattering effect as well, the absorption cross section of Ni particles can be much higher than Al counterpart for λ > 700 nm. Apart from the plasmon coupling enabled absorption enhancement, we further demonstrate the role of the highly porous structure. Figure 3D demonstrates that the effective refractive index of Ni/AAO and Al/AAO can be greatly reduced to ∼1.1 according to the Maxwell-Garnett theory [30]. It is quite approaching to that of air ambient and thus enables sufficient penetration and rather low reflection loss (∼< 1%).

Figure 3:

(A) Representative configuration of the close packed Ni NPs assembly with N=8. The diameter of Ni particle is set as 24 nm. (B) Schematic profile of the highly porous and low filling ratio Ni NPs in AAO nanochannels. The porosity is ∼ 60% and filling of metal particle is ∼0.1%. (C) Calculated absorption cross section of (A) as normalized by the geometry cross section. (D) Calculated effective refractive index based on the Maxwell-Garnett theory. (E) FDTD simulated and (F) measured absorption the Ni NPs/AAO and Al NPs/AAO structure, respectively.

By fully considering all the above material and structure properties, the absorption spectroscopy of the three-dimensional Ni NPs/AAO and Al NPs/AAO structure are calculated by the full wave FDTD method, as depicted in Figure 3E. The absorption spectra were measured by an ultraviolet visible near-infrared spectrophotometer with an integrated sphere (UV 3600, Shimadzu), as shown in Figure 3F. Thanks to the intrinsic material absorption as well as the high porous structure, our Ni NPs/AAO structure demonstrates the broadband light absorption over the entire solar spectrum. The solar weighted absorption efficiency reaches up to ∼97% from 400 to 2500 nm wavelength range, which is comparable or even higher than the alumina and silver counterpart [12, 31]. In addition, one may find that Ni NPs/AAO exhibits much broader absorption in the infrared range than Al NPs/AAO structure, which further suggests the ultrabroadband interband absorption for Ni. Thanks to the much stronger intrinsic absorption capability of Ni material, the effective deposited thickness of Ni deposited on Ni NPs/AAO structure is merely 60 nm (slightly larger than two times of skin depth of plasmonic metals), which is even smaller than the noble metal counterparts reported previously (∼90 nm or more) [11, 12, 31]. Figure S5B and C further demonstrates that, an extra 20 nm Ni NPs deposition makes little changes to the broadband absorption, indicating that less material consumption and better structure tolerance on the particle distribution profiles.

Apart from the broadband solar absorption performance, we demonstrate that the prepared Ni NPs/AAO absorber can be an excellent cost-effective photothermal for interfacial solar thermal conversion ISVG. Firstly, the mass density of the solar absorbers is crucial as well for the efficient ISVG. Compared with noble metals such as gold and silver, Ni metal possesses lower mass density. Therefore, when combined with highly porous substrate, Ni NPs/AAO structure can be self-floatable on water surface, enabling the interfacial vapor generation configuration. Secondly, the massive nanopores of Ni NPs/AAO can provide sufficient channels for water supply and vapor escape. By taking all of the above advantages, the Ni NPs/AAO absorber shows promising potential in industrial water-treatment technologies, for example, solar desalination and/or purification [32, 33].

To verify the ISVG performance of our Ni NPs/AAO absorber, we performed the standard solar vapor evaporation experiment [34], with solar irradiation of different illumination intensity (from 1 kW m⁻² up to 5 kW m⁻²) with a xenon lamp (MEXESOLAR-500). As the Ni NPs/AAO evaporator self-float on the interface of water and air, the photons penetrate inside the photothermal material converted into phonons in ∼ ps time scale and highly localized in the thin film, then resonantly couple to nearby water molecules, which ensures efficiently thin layer-heating of water and vapor generation. In the experiment, the optical images of the Ni NPs/AAO absorber with the solar irradiations off/on are shown in Figure 4A–D, respectively. Under the irradiation of the solar simulator, the temperature starts to rise from 23.5 °C (Figure 4B) and finally reached the steady-state temperature of ∼36.0 °C (Figure 4D), which can be ascribed to the light induced thermal localization effect. Most of the input energy was utilized, resulting in relatively low temperature increment. Furthermore, the temperature curve of the real time solar vapor is quickly rising as a function of evaporation time, as shown in Figure 4E. One may find that, the vapor temperature can quickly reach the steady-state temperature within 8 min. The time-dependent mass changes of steam generation under different solar illuminations are shown in Figure 4F. The interfacial evaporation process is gradually accelerated with increased illumination intensity. The evaporation rate reached about ∼ 0.9 kg m⁻² h⁻¹ under one sun (1 kW m⁻²) (Supporting Information, Figure S7), which is almost comparable to the conventional metal counterparts.

Figure 4:

(A–D) Optical (A, C) and thermal images (B, D) of the Ni NP/AAO absorber floating on water before (A, B) and after (C, D) solar illumination, respectively. (E) Measured solar vapor temperature as a function of evaporation time. (F) Mass change curves as a function of evaporation time under different illumination intensity. The ambient temperature is ∼23 °C and the humidity is ∼50%.

Apart from the low-cost material and appealing photothermal performance, finally, we demonstrate that the Ni NPs/AAO exhibit excellent water purification performance for salty water. It can be found that the concentrations of all the primary ions in the simulated seawater (Na⁺, Mg²⁺, Ca²⁺, and B³⁺) are significantly reduced (Supporting Information, Figure S8A), which are well above the World Health Organization (WHO) standard for drinking water [35]. Furthermore, we tested the cycling performance of Ni NPs/AAO absorber and demonstrated the long-term stability of the Ni NPs/AAO for ISVG (Supporting Information, Figure S8B). One may find that, the vapor production is relatively stable in long time operation, which shows the promise for applicable solar water purification.

4 Conclusion

By the straightforward physical metal deposition via the ion sputtering deposition process, we demonstrated a three-dimensional Ni nanoparticle based broadband solar absorber due to both broadband intrinsic material absorption and highly porous metallic structure. The Ni NPs/AAO absorber can exhibit an effective absorption (∼97%) over 400–2500 nm wavelength range. It can be used in the field of solar vapor generation and realizes an evaporation rate of ∼0.9 kg m−² h⁻¹ under one solar illumination. The considerable solar vapor generation performance combined with cycle stability, low cost material and simple fabrication process suggests that it could be a promising photothermal material for personalized solar water purification applications in the future.