Femtosecond-Laser Nanostructuring of Black Diamond Films under Different Gas Environments

Irradiation of diamond with femtosecond (fs) laser pulses in ultra-high vacuum (UHV) conditions results in the formation of surface periodic nanostructures able to strongly interact with visible and infrared light. As a result, native transparent diamond turns into a completely different material, namely “black” diamond, with outstanding absorptance properties in the solar radiation wavelength range, which can be efficiently exploited in innovative solar energy converters. Of course, even if extremely effective, the use of UHV strongly complicates the fabrication process. In this work, in order to pave the way to an easier and more cost-effective manufacturing workflow of black diamond, we demonstrate that it is possible to ensure the same optical properties as those of UHV-fabricated films by performing an fs-laser nanostructuring at ambient conditions (i.e., room temperature and atmospheric pressure) under a constant He flow, as inferred from the combined use of scanning electron microscopy, Raman spectroscopy, and spectrophotometry analysis. Conversely, if the laser treatment is performed under a compressed air flow, or a N2 flow, the optical properties of black diamond films are not comparable to those of their UHV-fabricated counterparts.


Introduction
In the field of concentrated solar energy conversion, black diamond is gaining increasing attention, thanks to its outstanding solar absorptance values, even exceeding 90% [1][2][3], as well as for its suitability for operation at very high temperatures in thermionic emission-based devices [4][5][6]. "Black" diamond is obtained by a surface treatment of native semi-transparent diamond with linearly-polarized ultrafast large area can be obtained on single-crystal diamond samples fs-laser treated in air. Conversely, in this work we demonstrate that treatments in air, as well as under a N 2 flow, are not the optimal solution for polycrystalline diamond samples, whereas the use of a He flow ensures the fabrication of black diamond films with optical properties comparable to those obtainable with UHV treatments.

Materials and Methods
Three freestanding 10 × 10 × 0.3 mm 3 polycrystalline CVD (Chemical Vapour Deposition) "thermal management (TM) grade" diamond samples, provided by Element Six Ltd. (Didcot, UK), were selected for the present work. For convenience, the three samples were labelled as TM1, TM2, and TM3.
Before starting the laser treatments, all the samples were subjected to a standard cleaning procedure, aimed at removing possible non-diamond contents, as well as residual contaminants derived from the polishing procedure. Diamond plates were dipped in a strongly oxidizing mixture (Carlo Erba Reagents S.r.l., Cornaredo, Milan, Italy), composed by HNO 3 (purity 70%), H 2 SO 4 (purity 96%), and HClO 4 (purity 71%) in the 1:1:1 volume ratio, for 15 min at boiling point. Subsequently, the samples were subjected to ultrasound cleaning in acetone for 5 min, then in 2-propanol for a further 5 min. Samples were finally rinsed in deionized water.
Laser processing sessions were performed at room temperature and atmospheric pressure with a linearly polarized fs-pulsed laser beam (wavelength λ fs = 800 nm), generated by a mode-locked regeneratively amplified "chirped-pulse" Ti:sapphire laser system. Repetition rate, f, was set to 1 kHz. Pulse duration was set to~100 fs. Diamond samples were placed on an x-y translational stage with micrometric resolution and automatic control capabilities (Laser µFAB microfabrication Workstation, from Newport, Irvine, CA, USA). For every diamond plate, the laser beam, with a spot-size of diameter (1/e 2 width) equal to 2w = 150 µm, scanned a delimited surface area (namely, a 8 × 8 mm 2 square) according to a raster pattern. Figure 1 reports a sketch of the scanning process; as can be visualized, the laser beam sweeps horizontally left-to-right at a constant speed, v x , then switches off and quickly moves back to the starting point; after that, it shifts vertically by a distance ∆y, switches on again, and starts a new scan along x-axis. The sequence is repeated until a 8 × 8 mm 2 square is processed. For a single scan along the x-axis, the accumulated horizontal laser fluence, Φ x , is defined as Φ where N x is the average number of pulses irradiating the surface unit (exemplified by the point P in Figure 1), and Φ p = E p /(πw 2 ) is the single pulse laser fluence, with E p and w being the pulse energy and the radius of the circular laser spot on the focal plane, respectively. However, to obtain the total accumulated laser fluence delivered to the surface unit, Φ A , we also have to take into account the vertical overlap ∆y between two consecutive horizontal scans, i.e., Φ A = (2w/ ∆y) Φ x . In our case, we obtain Φ A = 7.5 Φ x , being ∆y = 20 µm and 2w = 150 µm.
The same value of total accumulated laser fluence was delivered to each of the treated samples. In particular, we chose Φ A = 5.0 kJ/cm 2 , corresponding to the best periodic 1D structures (in terms of structural integrity and geometric regularity) obtained in UHV-fabricated black diamond samples of the same crystalline grade, as previously reported [2]. In all the different laser treatments, the single pulse energy was kept constant. More precisely, it was set to E p = 785 µJ, which corresponds to a single pulse laser fluence Φ p = 4.44 J/cm 2 , which is slightly above the ablation threshold for polycrystalline CVD diamond, estimated to be about 3 J/cm 2 with a single 100 fs pulse irradiation at 800 nm [26]. A constant value of single pulse laser fluence allowed us to adjust the accumulated horizontal laser fluence Φ x by simply varying the laser scanning speed v x , which, together with the repetition rate f, determines the average number of pulses N x irradiating the surface unit through the relationship N x = 2wf/v x . By choosing v x = 1 mm/s, which represents a good trade-off between effectiveness and speed of the laser treatment, we obtained N x = 150 and Φ x = 0.67 kJ/cm 2 , resulting in the desired value of total accumulated laser fluence Φ A = 5.0 kJ/cm 2 .

Figure 1.
Laser scanning patterns used for the treatment of the diamond samples. The laser beam scans the sample ("laser on") along x, than blanks ("laser off"), moves back to the starting point, shifts along y by a distance Δy, switches on again, and starts a new scan along x. For the sake of clarity, only the first laser spot of horizontal scans #2 to #7 is shown. Note that the point P is subjected to five laser scans along x (rows 1 to 5), but it is not irradiated when the laser scans rows 6 and 7.
The same value of total accumulated laser fluence was delivered to each of the treated samples. In particular, we chose ΦA = 5.0 kJ/cm 2 , corresponding to the best periodic 1D structures (in terms of structural integrity and geometric regularity) obtained in UHV-fabricated black diamond samples of the same crystalline grade, as previously reported [2]. In all the different laser treatments, the single pulse energy was kept constant. More precisely, it was set to Ep = 785 μJ, which corresponds to a single pulse laser fluence Φp = 4.44 J/cm 2 , which is slightly above the ablation threshold for polycrystalline CVD diamond, estimated to be about 3 J/cm 2 with a single 100 fs pulse irradiation at 800 nm [26]. A constant value of single pulse laser fluence allowed us to adjust the accumulated horizontal laser fluence Φx by simply varying the laser scanning speed vx, which, together with the repetition rate f, determines the average number of pulses Nx irradiating the surface unit through the relationship Nx = 2wf/vx. By choosing vx = 1 mm/s, which represents a good trade-off between effectiveness and speed of the laser treatment, we obtained Nx = 150 and Φx = 0.67 kJ/cm 2 , resulting in the desired value of total accumulated laser fluence ΦA = 5.0 kJ/cm 2 .
Throughout the duration of the laser treatment, a constant gas flow was directed towards the sample surface. Aimed at ensuring the same exposure conditions on the whole surface, the gas flow direction formed an angle with the surface normal very close to 90°, so to avoid the formation of a boundary layer with a spatially dependent thickness. Therefore, the only experimental condition which was varied from one treatment session to another was the type of gas used: compressed air (standard ISO 8573-1:2010 [27], purity class 1) for TM1, N2 (purity > 99.998%) for TM2, and He (purity Laser scanning patterns used for the treatment of the diamond samples. The laser beam scans the sample ("laser on") along x, than blanks ("laser off"), moves back to the starting point, shifts along y by a distance ∆y, switches on again, and starts a new scan along x. For the sake of clarity, only the first laser spot of horizontal scans #2 to #7 is shown. Note that the point P is subjected to five laser scans along x (rows 1 to 5), but it is not irradiated when the laser scans rows 6 and 7.
Throughout the duration of the laser treatment, a constant gas flow was directed towards the sample surface. Aimed at ensuring the same exposure conditions on the whole surface, the gas flow direction formed an angle with the surface normal very close to 90 • , so to avoid the formation of a boundary layer with a spatially dependent thickness. Therefore, the only experimental condition which was varied from one treatment session to another was the type of gas used: compressed air (standard ISO 8573-1:2010 [27], purity class 1) for TM1, N 2 (purity > 99.998%) for TM2, and He (purity > 99.5%) for TM3. After the treatments, all the sample were subjected to the same cleaning procedure previously described (HNO 3 :H 2 SO 4 :HClO 4 = 1:1:1, 15 min at boiling point), aimed at removing debris caused by the ablation process.
The morphology of the fabricated black diamond samples was evaluated with a "Leo-Supra35" field emission scanning electron microscope (FE-SEM) (Zeiss, Oberkochen, Germany). Prior to starting SEM characterization, an ultra-thin (~2 nm) Au coating was deposited on the surface of each sample by DC sputtering to avoid sample charging. After the end of SEM analysis, the Au coating was simply removed by dipping the samples in aqua regia (HNO 3 :HCl in the 1:3 volume ratio, 5 min at boiling point), then samples were rinsed in deionized water.
Aimed at investigating the laser-induced structural modifications, a comprehensive Raman characterization was successively performed on the black diamond samples. Raman spectra in the 1000-1800 cm −1 range were acquired at room temperature and in back-scattering geometry with an Ar + laser (514.5 nm wavelength). A Dilor XY triple spectrometer (Dilor Instruments SA, Edison, NJ, USA) was used, equipped with an Olympus confocal microscope (Olympus Corporation, Tokyo, Japan) and a liquid nitrogen-cooled charge coupled device (CCD) detector. The laser spot-size was set to 2 µm. Data were then processed with Thermo-Grams Suite v9.2 software (Thermo Fisher Scientific Inc., Waltham, MA, USA), aimed at calculating widths and intensities of the Raman peaks. In particular, peak profiles were fitted by using Voigt functions.
The optical properties of nanotextured black diamond samples were finally investigated by spectrophotometric analysis, aimed at evaluating their absorptance in the solar wavelength range, whereas a detailed defect spectroscopy study is beyond the scope of this work. Room temperature spectrophotometry in the wavelength range 0.25-2.5 µm was performed by measuring optical transmittance (τ) and hemispherical reflectance (ρ) with a "Lambda900" double-beam spectrophotometer (PerkinElmer, Waltham, MA, USA) equipped with a Spectralon ® -coated integrating sphere (150 mm diameter). Spectra were acquired at quasi-normal incidence conditions. Absorptance (α) was obtained from transmittance and reflectance values through the conservation of energy relationship [28]:

Results
In the following subsections, the results of the morphological, structural, and optical characterization of the three black diamond samples (TM1, TM2, TM3) fabricated under different gas environments (air, N 2 , He, respectively) will be introduced and compared to those previously reported on their UHV-fabricated counterparts [2,3].

SEM Characterization
The left side of Figure 2 shows three high-magnification SEM images taken from the surface of the three nanostructured black diamond films. As can be seen, no significant difference in LIPSS morphology can be highlighted among the samples treated under a compressed air (Figure 2a), a N 2 (Figure 2c), or a He (Figure 2e) flow: in all cases, well-formed and rather regular LIPSS can be noticed, with a periodicity, Λ, of about 170 ± 10 nm, which perfectly matches the theoretical value of 166 nm given by the relationship Λ = λ fs /2n, where n = 2.41 is the refractive index of diamond at the laser wavelength λ fs = 800 nm used for the treatments [29]. LIPSS depth, as inferred from SEM images recorded in tilted configuration, was found to be about 480 ± 20 nm. Periodic structures well resemble those obtained on UHV-fabricated black diamond films of the same crystalline quality, treated at the same laser wavelength, and with the same value of total accumulated laser fluence [2]. So we can conclude that the gas environment of the laser treatment has little influence on the morphological and geometrical features of LIPSS.
Conversely, looking at the right side of Figure 2, which shows representative lower magnification SEM images of larger areas of the laser-treated samples, it can immediately be seen that the surfaces of the samples treated under a compressed air (Figure 2b) or a N 2 flow (Figure 2d) differ significantly from the surface of the sample treated under a He flow (Figure 2f). In the case of TM1-air and TM2-N 2 , surfaces have indeed a terraced morphology, with irregularly shaped treads and risers of different heights, whereas TM3-He has a substantially flat surface, with shallow scratches most probably caused by the mechanical polishing procedure performed on the untreated sample. Most significantly, while LIPSS are rather uniformly distributed throughout the whole surface of TM3-He, as in the case of a UHV-fabricated black diamond film, the same cannot be said for TM1-air and TM2-N 2 samples. As can be more clearly seen from Figure 3, showing a detail of the surface of TM1-air (but the same considerations can be made for TM2-N 2 ), most of the treads show regular LIPSS (green dot), but in some cases LIPSS degenerate into irregular square-like nanostructures (blue dot), or they do not even form at all (red dot). Conversely, looking at the right side of Figure 2, which shows representative lower magnification SEM images of larger areas of the laser-treated samples, it can immediately be seen that the surfaces of the samples treated under a compressed air (Figure 2b) or a N2 flow (Figure 2d) differ significantly from the surface of the sample treated under a He flow (Figure 2f). In the case of TM1-air and TM2-N2, surfaces have indeed a terraced morphology, with irregularly shaped treads and risers of different heights, whereas TM3-He has a substantially flat surface, with shallow scratches most probably caused by the mechanical polishing procedure performed on the untreated sample. Most significantly, while LIPSS are rather uniformly distributed throughout the whole surface of TM3-He, as in the case of a UHV-fabricated black diamond film, the same cannot be said A possible explanation for the irregular morphology of TM1-air and TM2-N 2 surfaces, as well as for the poor uniformity of LIPSS distribution, can be given if we observe that deep cracks (see Figure 2b,d) and grooves (see Figure 3) do not follow the laser scanning direction, but rather randomly-oriented curved paths, which can extend for several tens of micrometers, i.e., the typical grain size of a CVD "thermal management grade" polycrystalline diamond sample with a thickness of a few hundreds of micrometers [30]. Therefore, such curved paths run most likely along the grain boundaries, which are the preferential incorporation sites of nitrogen [31]. Our hypothesis is that when the Gaussian laser beam, during the raster scan process, irradiates with its tails (i.e., at a fluence between the modification threshold and the ablation one), a grain boundary exposed to a compressed air or a N 2 flow, incorporation of nitrogen is strongly enhanced by local heating. This leads to bond distortions [32], induces a significant mechanical stress in the surrounding lattice, and increases the number of vacancy defects [33]. As a consequence, the ablation threshold decreases in correspondence of a grain boundary, as well as in its neighborhood, so that the subsequent full-spot irradiation (i.e., at the peak fluence) results in a much higher ablation rate. A high fluence ablation regime [34] may indeed be triggered, with the formation of deep cracks and the ejection of differently sized shivers [35], which are then re-deposited as debris or carried away with the gas flow. Such re-deposited shivers, by partially masking the sample surface from laser irradiation, may be responsible for the presence of untreated areas in TM1-air and TM-N 2 samples (see Figure 2b,d and Figure 3). for TM1-air and TM2-N2 samples. As can be more clearly seen from Figure 3, showing a detail of the surface of TM1-air (but the same considerations can be made for TM2-N2), most of the treads show regular LIPSS (green dot), but in some cases LIPSS degenerate into irregular square-like nanostructures (blue dot), or they do not even form at all (red dot). A possible explanation for the irregular morphology of TM1-air and TM2-N2 surfaces, as well as for the poor uniformity of LIPSS distribution, can be given if we observe that deep cracks (see Figure  2b,d) and grooves (see Figure 3) do not follow the laser scanning direction, but rather randomlyoriented curved paths, which can extend for several tens of micrometers, i.e., the typical grain size of a CVD "thermal management grade" polycrystalline diamond sample with a thickness of a few hundreds of micrometers [30]. Therefore, such curved paths run most likely along the grain boundaries, which are the preferential incorporation sites of nitrogen [31]. Our hypothesis is that when the Gaussian laser beam, during the raster scan process, irradiates with its tails (i.e., at a fluence between the modification threshold and the ablation one), a grain boundary exposed to a compressed air or a N2 flow, incorporation of nitrogen is strongly enhanced by local heating. This leads to bond distortions [32], induces a significant mechanical stress in the surrounding lattice, and increases the number of vacancy defects [33]. As a consequence, the ablation threshold decreases in correspondence of a grain boundary, as well as in its neighborhood, so that the subsequent full-spot irradiation (i.e., at the peak fluence) results in a much higher ablation rate. A high fluence ablation regime [34] may indeed be triggered, with the formation of deep cracks and the ejection of differently sized shivers [35], which are then re-deposited as debris or carried away with the gas flow. Such redeposited shivers, by partially masking the sample surface from laser irradiation, may be responsible for the presence of untreated areas in TM1-air and TM-N2 samples (see Figures 2b,d and 3).
It is worth stressing here that femtosecond laser treatments performed in air on high-quality single crystal diamond plates result in regular LIPSS uniformly distributed over the whole sample surface, which substantially retains its original flatness without the formation of deep cracks and/or It is worth stressing here that femtosecond laser treatments performed in air on high-quality single crystal diamond plates result in regular LIPSS uniformly distributed over the whole sample surface, which substantially retains its original flatness without the formation of deep cracks and/or terraces [25]. This may support our hypothesis on the role that grain boundaries play in hampering a reproducible and reliable fabrication process of black diamond films in air or in a N 2 -rich environment. A He environment, conversely, ensures a uniform treatment over large areas even on polycrystalline diamond samples. Figure 4a shows the Raman spectra of the three fabricated black diamond films in the full analyzed range 1000-1800 cm −1 . As can be seen, the only feature appearing in all the spectra is a sharp strong line centered at about 1332 cm −1 , which is the well-known first-order peak related to crystalline diamond. As in the case of an untreated diamond sample, as well in those of all the black diamond films laser-treated in UHV at different values of total accumulated fluence, no other carbon-related feature (e.g., graphite, diamond-like carbon) is present. Linewidths (Full Width Half Maximum, FWHM) are all in the 7-10 cm −1 range regardless of the gas environment, thus partially overlapping the Materials 2020, 13, 5761 8 of 12 4-8 cm −1 range measured on their UHV-fabricated counterparts [2]. In black diamond films, the large FWHM value of the 1332 cm −1 peak, always exceeding 4-5 cm −1 in the case of polycrystalline samples, is due to the laser-induced creation of intra-grain defects, which degrade the structural quality of crystalline grains. Figure 4a shows the Raman spectra of the three fabricated black diamond films in the full analyzed range 1000-1800 cm −1 . As can be seen, the only feature appearing in all the spectra is a sharp strong line centered at about 1332 cm −1 , which is the well-known first-order peak related to crystalline diamond. As in the case of an untreated diamond sample, as well in those of all the black diamond films laser-treated in UHV at different values of total accumulated fluence, no other carbon-related feature (e.g., graphite, diamond-like carbon) is present. Linewidths (Full Width Half Maximum, FWHM) are all in the 7-10 cm −1 range regardless of the gas environment, thus partially overlapping the 4-8 cm −1 range measured on their UHV-fabricated counterparts [2]. In black diamond films, the large FWHM value of the 1332 cm −1 peak, always exceeding 4-5 cm −1 in the case of polycrystalline samples, is due to the laser-induced creation of intra-grain defects, which degrade the structural quality of crystalline grains.  Figure 4a in the Raman shift range around the reference diamond peak position at 1332.0 cm −1 ; square dots represents raw data, solid lines indicate the Voigt functions used to fit the peak profiles, whereas dashed lines indicate the diamond peak position of the three different black diamond films.

Raman Characterization
Most interestingly, the gas environment of the laser treatment was found to strongly influence the Raman shift position of the diamond peak, which always differs from the reference value of 1332.0 cm −1 . In particular, diamond peak positions were measured to be 1326.3 cm −1 , 1326.9 cm −1 , and 1335.1 cm −1 for TM1-air, TM2-N2, and TM3-He, respectively. This implies that samples treated under a compressed air flow or a N2 flow are subjected to approximately the same amount of tensile stress, whereas the sample treated under a He flow is subjected to a compressive stress. It is worth mentioning here that an upshifted diamond Raman peak, denoting a compressive stress as in the case of TM3-He, has been observed in all the UHV-fabricated black diamonds, showing the best LIPSS in terms of geometrical regularity and structural integrity [2,3,6,24]. Conversely, a downshifted diamond Raman peak, denoting a tensile stress as in the case of TM1-air and TM2-N2, has only been  Most interestingly, the gas environment of the laser treatment was found to strongly influence the Raman shift position of the diamond peak, which always differs from the reference value of 1332.0 cm −1 . In particular, diamond peak positions were measured to be 1326.3 cm −1 , 1326.9 cm −1 , and 1335.1 cm −1 for TM1-air, TM2-N 2 , and TM3-He, respectively. This implies that samples treated under a compressed air flow or a N 2 flow are subjected to approximately the same amount of tensile stress, whereas the sample treated under a He flow is subjected to a compressive stress. It is worth mentioning here that an upshifted diamond Raman peak, denoting a compressive stress as in the case of TM3-He, has been observed in all the UHV-fabricated black diamonds, showing the best LIPSS in terms of geometrical regularity and structural integrity [2,3,6,24]. Conversely, a downshifted diamond Raman peak, denoting a tensile stress as in the case of TM1-air and TM2-N 2 , has only been observed in "thermal management grade" black diamond plates UHV-treated at high values of total accumulated laser fluence (>10 kJ/cm 2 ), which resulted in a very damaged and irregular surface [2,24]. Figure 5 shows the absorptance spectra of the three fabricated black diamond samples, as well as of a reference untreated sample of the same crystalline quality. Noticeable is the increase of absorptance in the whole wavelength range for all the treated samples with respect to the reference diamond, thus validating the effectiveness of all the fabrication processes at room temperature and atmospheric pressure under different gas environments. As expected from the SEM and Raman characterizations, showing very similar results for TM1-air and TM2-N 2 , absorptance curves measured for the samples treated under a compressed air or a N 2 flow are almost superimposable, with slightly higher values for TM1-air in the near-infrared (NIR) region. Figure 5 shows the absorptance spectra of the three fabricated black diamond samples, as well as of a reference untreated sample of the same crystalline quality. Noticeable is the increase of absorptance in the whole wavelength range for all the treated samples with respect to the reference diamond, thus validating the effectiveness of all the fabrication processes at room temperature and atmospheric pressure under different gas environments. As expected from the SEM and Raman characterizations, showing very similar results for TM1-air and TM2-N2, absorptance curves measured for the samples treated under a compressed air or a N2 flow are almost superimposable, with slightly higher values for TM1-air in the near-infrared (NIR) region.

Figure 5.
Absorptance spectra of the three black diamond films laser-treated under different gas environments, obtained by spectrophotometric analysis according to Equation (1). For comparison, the absorptance spectrum of an untreated sample (gray curve) with the same crystalline quality is shown.
Most significantly, TM3-He, laser-treated under a He flow, shows absorptance values around 85% in the ultraviolet (UV) range, between 85 and 90% in the visible (VIS) range, and between 80% and 90% in the NIR range: this curve retraces almost perfectly the absorptance spectrum measured for a polycrystalline black diamond sample of the same crystalline quality UHV-treated at the same ΦA value (5.0 kJ/cm 2 ) [2,3]. The solar absorptance, αS, of the different black diamond films can be obtained from the following equation: where α(λ) is the absorptance as a function of the exciting wavelength λ, and W(λ) is the global-tilt (GT) 1.5 air mass (AM) solar irradiance [36,37]. Calculated αS values, according to Equation (2), are 79.1%, 78.4%, and 86.2% for TM1-air, TM2-N2 and TM3-He, respectively. In particular, the solar Figure 5. Absorptance spectra of the three black diamond films laser-treated under different gas environments, obtained by spectrophotometric analysis according to Equation (1). For comparison, the absorptance spectrum of an untreated sample (gray curve) with the same crystalline quality is shown.
Most significantly, TM3-He, laser-treated under a He flow, shows absorptance values around 85% in the ultraviolet (UV) range, between 85 and 90% in the visible (VIS) range, and between 80% and 90% in the NIR range: this curve retraces almost perfectly the absorptance spectrum measured for a polycrystalline black diamond sample of the same crystalline quality UHV-treated at the same Φ A value (5.0 kJ/cm 2 ) [2,3]. The solar absorptance, α S , of the different black diamond films can be obtained from the following equation: where α(λ) is the absorptance as a function of the exciting wavelength λ, and W(λ) is the global-tilt (GT) 1.5 air mass (AM) solar irradiance [36,37]. Calculated α S values, according to Equation (2), are 79.1%, 78.4%, and 86.2% for TM1-air, TM2-N 2 and TM3-He, respectively. In particular, the solar absorptance value obtained for TM3-He (86.2%) is very close to the one (88.4%) measured on its UHV-treated counterpart [2], therefore confirming that the use of a He flow at ambient conditions is a valid alternative to a UHV chamber for the fabrication of polycrystalline black diamond films with enhanced optical absorption properties.

Discussion
As mentioned in the introductory section, absorptance enhancement in black diamond samples is both due to a more effective light trapping and the introduction of defect-related energy levels within the bandgap, which promote sub-bandgap photon absorption. The geometrical quality and structural integrity of the laser-induced LIPSS has of course a strong influence on light trapping, because a well-defined diffraction grating, uniformly distributed over the sample surface, ensures an efficient coupling of impinging photons. Conversely, as it is easy to deduce, more defected LIPSS reflect into a larger density of defect-related sub-bandgap energy levels. Now, by observing the higher magnification SEM images of the LIPSS fabricated under the three different gas environments (Figure 2a,c,e), no substantial morphological differences can be noticed; therefore we can deduce that microscopically (i.e., within a LIPSS-containing domain), both the light trapping capability and the distribution of defect-related energy levels within the bandgap are approximately the same. What really marks the difference between the three cases can be easily inferred from the lower magnification SEM images, showing strongly irregular surfaces (alternating LIPSS-containing domains, flat zones, and deep grooves) for TM1-air ( Figure 2b) and TM2-N 2 (Figure 2d), and a uniform regular distribution of LIPSS for TM3-He (Figure 2f). As a result, an effective trapping of photons impinging all over the surface (as in the case of a UHV-treated sample) is only ensured for laser treatments performed under a He flow; this may be the reason of the increased absorptance of TM3-He ( Figure 5) with respect to TM1-air or TM2-N 2 . The lack of significant discontinuities in the LIPSS spatial distribution may also explain why a compressive stress, typical of polycrystalline black diamond films UHV-treated at optimal Φ A values, is only present in TM3-He (Figure 2b), whereas TM1-air and TM2-N 2 show a tensile stress typical of strongly damaged surfaces. Of course, the extended defects (e.g., deep grooves) of TM1-air and TM2-N 2 may surely imply local enhancements of photon absorption by introducing a high density of sub-bandgap energy levels, but apparently this cannot compensate for the lack of a widely distributed uniform optical grating (as in the TM3-He case) able to increase light trapping efficiency.
In conclusion, while the total accumulated laser fluence is the key parameter to obtain LIPSS with high geometrical quality and structural integrity, the gas environment the laser treatment is performed within turns out to be crucial for the homogeneous distribution of LIPSS on the surface of the resulting black diamond film. From this point of view, laser treatments under a He flow are very effective, and are demonstrated to be the starting point for a cost-effective and reproducible fabrication of polycrystalline black diamond films with superior properties in terms of optical absorption. Directions for future research are based on a further refinement of the optimal value of the total accumulated laser fluence used for the treatments, aimed at a simultaneous enhancement of absorptance and quantum efficiency of the treated material. Moreover, the possibility of exploiting a He flow for implementing more complex nanostructures (e.g., two-dimensional LIPSS with enhanced light-trapping capabilities [38]) which have already been demonstrated with UHV treatments, will be explored.