Maskless inverted pyramid texturization of silicon

We discovered a technical solution of such outstanding importance that it can trigger new approaches in silicon wet etching processing and, in particular, photovoltaic cell manufacturing. The so called inverted pyramid arrays, outperforming conventional pyramid textures and black silicon because of their superior light-trapping and structure characteristics, can currently only be achieved using more complex techniques involving lithography, laser processing, etc. Importantly, our data demonstrate a feasibility of inverted pyramidal texturization of silicon by maskless Cu-nanoparticles assisted etching in Cu(NO3)2 / HF / H2O2 / H2O solutions and as such may have significant impacts on communities of fellow researchers and industrialists.

In this case, the Si/Cu(NO 3 ) 2 /H 2 O 2 /HF system is composed of a corrosion-type redox couple: the cathodic reduction of Cu 2+ ions with H 2 O 2 as well as the anodic oxidation and dissolution of Si beneath the deposited Cu-NPs. Cu 2+ ions capture electrons from the vicinity of the Si substrate, aggregate and form NPs. In this manner, Cu-NPs nucleated originally on the Si surface attract electrons from Si and become negatively charged because Cu is more electronegative than Si. These negatively charged Cu-NPs grow further by attracting the Cu 2+ ions from the solution 19,20,26 . During this process, Si atoms underneath the Cu-NPs are continuously oxidized and etched by the HF. Simultaneously, the reduction of H 2 O 2 on the Cu-NPs is indispensable to induce anisotropic Cu-NPs' deposition onto the Si surface and guarantee the formation of inverted pyramid arrays, which will be interpreted in the following sections. The inverted pyramid arrays were cleaned using concentrated nitric acid in a sonication bath for at least 20 min to remove all of the residual Cu-NPs. After the nitric acid bath, no Cu-NPs were observed by scanning electron microscopy (SEM) analysis and no Cu peaks appeared in the energy-dispersive X-ray spectra of the inverted pyramid arrays.
The inverted pyramid arrays fabricated using this approach were regular and consistent throughout the batches and across large areas, up to wafer-scale. Fig. 1(a) Fig. 1(d) reveals that the angle between the facet of the inverted pyramid and the (100) surface is 54.7°, indicating that the facets are terminated with Si (111) planes. As shown in Fig. 1(b), the inverted pyramid arrays are random because of the original irregular morphology and saw damage on Si; however, the surface is fully covered by random inverted pyramids and is regular when considered as a whole. The length of the inverted pyramids' bottom side varies within 2-6 μ m, and the The key parameters of the reaction were identified using p-type (100)-oriented, nominally 1-3 Ω cm c-Si as the etching wafer, but the process was successfully tested for n-type material too, naturally adjusting the solvent. In general, the morphology of the inverted pyramid arrays was well controlled by adjusting the etching time, the concentration of Cu(NO 3 ) 2 /HF/H 2 O 2 and the etching temperature.
Up to date, the anisotropic wet chemical etching of c-Si has been done in alkaline solutions, resulting in random pyramid arrays, whereas isotropic wet chemical etching is commonly observed in HF based solutions 4 . In our technique, the anisotropic deposition of Cu-NPs from the Cu(NO 3 ) 2 /HF/ H 2 O 2 /H 2 O mixture is the key step in the synthesis of the inverted pyramid arrays without any mask. In Fig. 2(a), the early form of one inverted pyramid is demonstrated, where more and larger Cu-NPs are observed on Si (100) surface and fewer and smaller Cu-NPs are observed on Si (111) surfaces. This orientation-dependent population of Cu-NPs occurs due to a low reduction potential of Cu 2+ , exhibiting a limited ability to withdraw electrons directly from Si at room temperature 27 . Notably, rising temperature (to 50 °C) enhances the probability of the electron capturing by Cu 2+ ions, leading to the nucleation of Cu-NPs on c-Si. Importantly, the electronic properties of (100) and (111) planes in c-Si are different in terms of the surface bond densities 4,28,29 . Accounting that, the number of electrons "available" on Si (100) planes is greater than that on the Si (111) planes, Cu-NPs preferentially nucleate on Si (100) plane because of the higher probability of electron capturing by Cu 2+ ions. Further growth may be facilitated by the interaction between Cu 2+ ions and these negatively charged Cu-NPs, resulting in faster growth of Cu-NPs on Si (100) plane than that on (111) planes. Fig. 2(b) is a simulated diagram that shows the pattern of anisotropic Cu-NPs' deposition. As a result, Si atoms underneath the Cu-NPs are preferentially oxidized and then etched by HF, leading to a faster etching along Si < 100> directions.
Notably, H 2 O 2 plays an important and special role in the selective etching too, as previously mentioned. Energetically, the electrochemical potential of H 2 O 2 is much more positive than either the valence band of Si or the reduction potential of Cu 2+ /Cu-NPs. Thus, the presence of H 2 O 2 may promote holes injection into the valence band of Si as well as into the Cu-NPs. Although the Si etching by HF does occur when a Si substrate is subjected to an HF/H 2 O 2 solution, the etching rate is less than 10 nm/h in an etchant with a high concentration of H 2 O 2 . In practice, the presence of a noble metal is necessary for accelerating Si etching in HF/H 2 O 2 solutions 22 . In other words, the observed preferential reduction of H 2 O 2 is attributed to the surface of the metal rather than to the surface of the bare Si 22 . Therefore, the reduction of H 2 O 2 (labeled as "1" in Fig. 2(c)) in the Si/Cu(NO 3 ) 2 /H 2 O 2 /HF system is likely to occur on the surface of the Cu-NPs too. When moderate amounts of H 2 O 2 are introduced, Cu-NPs are deposited anisotropically, which assists in the etching of inverted pyramids by HF; thus, the injected holes are consumed via two pathways. The holes are consumed by the Cu-NPs, ensuring kinetic equilibrium between the deposition and dissolution of Cu NPs (as labeled by "2" in Fig. 2c). Some holes also diffuse through the Cu-NPs and are injected into the Si substrate (labeled by "3" in Fig. 2c) selectively accelerating the etching of c-Si beneath the Cu-NPs.
In the absence of H 2 O 2 , the reduction of Cu 2+ will be faster, allowing more Cu-NPs to be formed at the Si surface because of unlimited supply of Cu 2+ ions. The Cu-NPs quickly grow larger and cover the majority of the Si wafer surface 30 . This situation induces overlapping regions of oxidation and, accordingly, faster isotropic Si etching. A thin Cu film will be formed very quickly at 50 °C; this film will inhibit the etching of Si by acting as a barrier between the Si and the etchants. At high concentrations of H 2 O 2 , the oxidation of Cu-NPs is faster than the reduction of Cu 2+ , resulting in the deposition of a few small Cu-NPs randomly on the Si surface; therefore, the original morphology of the Si does not substantially change, except for the formation of a few shallow nano-pits. On the basis of the interaction between H 2 O 2 and Cu-NPs and their roles in the process, inverted pyramid arrays were obtained for the first time using a simple and low-cost maskless chemical etching.
After the etching process, the hemispheric total reflectance for normal incidence was measured using a Varian Cary 5000 spectrophotometer with an integrating sphere. The reflectance spectra of the inverted pyramid arrays prepared using different etching times are shown in Fig. 3, illustrating the tunability of our technique. Notably, our texturization reduced the mean reflectivity of c-Si (over the wavelength range from 300 to 1000 nm) down to ~4.4% without any antireflective coating as compared to much higher reflectivity from the normal pyramid arrays (see Fig. 3). The short wavelength spectral response analysis indicates that approximately 37% of the incoming light undergoes a triple bounce due to its interaction with the adjoining planes of the inverted pyramid geometry before being reflected away 11 . As shown in the inset in Fig. 3, the incident light, which strikes one face of the inverted pyramid, is reflected (the first bounce) onto an orthogonal plane, which is then reflected (the second bounce) onto the face complementary to the original point of impact and finally reflected away (the third bounce), while a double bounce of most incident light was observed in the pyramid geomtry 11 . The triple bounce increases the path length of the light, which induces an increased absorption of the incident light. The reduced front surface reflectance, together with the increased light absorption, makes inverted pyramid arrays an attractive and efficient light-trapping geometry. Moreover, the recessed structure is more suitable for covering and filling, such as for the coverage of amorphous Si in heterojunction solar cells and the filling of metal electrodes in photovoltaic devices.
In summary, we discovered a method of maskless texturization of c-Si with so-called inverted pyramid arrays that can be immediately uptaken for industrialization. In comparison with the pyramidal structure, which is widely used for Si solar cells, the inverted pyramidal structure has the same micrometer scale but lower reflectivity and higher light absorption, along with a shorter etching time and a lower etching temperature during the texturing process; compared with the B-Si obtained by the noble metal assisted chemical etching, inverted pyramidal structures are achieved by the cheaper metal, Cu-NPs assisted chemical etching. Moreover, inverted pyramid not only has superior structural characteristic but also ensures low recombination rates. Furthermore, the generality and scalability of our technique are very promising for the fabrication of other advanced Si-based devices. Last but not least, the underlying principles governing the etching process are understood. In particular, the preferential etching along < 100> directions is realized by a charge transfer assisted with Cu-NPs.

Methods
Inverted pyramid arrays were originally etched from commercial 156 mm × 156 mm, boron-doped (1-3 Ω cm), (100)-oriented c-Si wafers via Cu NPs assisted anisotropic etching. Before being etched in the solution, the wafers were thoroughly rinsed in acetone to remove any organic contaminants and then rinsed with deionized water. The inverted pyramid arrays were obtained when the wafers were etched for 15 min at 50 °C in a polytetrafluoroethylene container filled with 5 mM Cu(NO 3 ) 2 , 4.6 M HF and 0.55 M H 2 O 2 . Residual Cu-NPs were removed using concentrated nitric acid in a sonication bath for at least 20 min. The Si wafers were thoroughly rinsed with deionized water and dried under flowing nitrogen. The morphologies and structures of the wafers were characterized using a Hitachi S-4800 scanning electron microscope. The hemispheric total reflectance for normal incidence was measured using a Varian Cary 5000 spectrophotometer with an integrating sphere.