Tailoring Electro/Optical Properties of Transparent Boron-Doped Carbon Nanowalls Grown on Quartz

Carbon nanowalls (CNWs) have attracted much attention for numerous applications in electrical devices because of their peculiar structural characteristics. However, it is possible to set synthesis parameters to vary the electrical and optical properties of such CNWs. In this paper, we demonstrate the direct growth of highly transparent boron-doped nanowalls (B-CNWs) on optical grade fused quartz. The effect of growth temperature and boron doping on the behavior of boron-doped carbon nanowalls grown on quartz was studied in particular. Temperature and boron inclusion doping level allow for direct tuning of CNW morphology. It is possible to operate with both parameters to obtain a transparent and conductive film; however, boron doping is a preferred factor to maintain the transparency in the visible region, while a higher growth temperature is more effective to improve conductance. Light transmittance and electrical conductivity are mainly influenced by growth temperature and then by boron doping. Tailoring B-CNWs has important implications for potential applications of such electrically conductive transparent electrodes designed for energy conversion and storage devices.


Introduction
Most commercial transparent electrodes are usually made of indium tin oxide (ITO) due to its good electrical performance (~10 Ω/square), while the limited amount of indium resources raises concerns in the European Union as well as globally. For this reason, the scientific community has moved to seek new solutions [1,2]; among those, carbon-based transparent conducting materials are extensively applied nowadays in many optoelectronic [3,4], biosensing [5,6], energy storage, and conversion devices [7][8][9]. Carbon nanostructures are particularly attractive due to their unique three-dimensional structure with a high surface area and specific surface interactions [10]. Alternative nanomaterials, such as carbon nanowalls (CNWs), have held the attention of the scientific community due to their structural characteristics in mono-, bi-and tri-dimensional arrangements [11][12][13], including excellent electrical and thermal conductivity. Their unique geometrical shape and morphology cause CNWs to possibly be an interesting material for field emitters [14]. Field emitters based

Surface Morphology
Scanning electron microscopy via a FEI Quanta FEG 250 scanning electron microscope (SEM, FEI, Hillsboro, OR, USA), using a 15 kV beam accelerating voltage with a secondary electron-Everhart-Thornley (SE-ETD) detector working in high vacuum mode (pressure 10 −4 Pa) was used to observe the structure of the B-CNW surfaces. The program used for data visualization and analysis was Gwyddion (v.2.40, Czech Metrology Institute, Brno, Czech Republic).

Raman Spectroscopy
The molecular composition of the deposited films was studied by means of Raman spectroscopy using a Raman microscope (InVia, Renishaw, UK). Spectra were recorded over the range 200-3500 cm −1 with an integration time of 5 s (10 averages) using an argon ion laser emitting at 514 nm and operating at 5% of its total power (50 mW). Data were smoothed (using a Savitzky-Golay method: 15 points, second polynomial order) and the baseline (approximated with a cubic polynomial) subtracted and normalized.

Spectroscopy Ellipsometry
Spectroscopic ellipsometry analysis was conducted with a Jobin-Yvon UVISEL phase-modulated ellipsometer (HORIBA Jobin-Yvon Inc., Edison, Rosemead, CA, USA). The investigated wavelength region was between 250 and 800 nm. The experiments were carried out at room temperature using an angle of incidence fixed at 60 • and a compensator set at 45 • . DeltaPsi software (v. 2.4.3, HORIBA Jobin-Yvon Inc., Edison, Rosemead, CA, USA) was employed to determine the spectral distributions of refractive index n (λ) and extinction coefficient k (λ) of the B-CNW films.

Electrical Conductivity
Electrical conductivities were measured at room temperature by a four-point probe placed in a straight line with equal spacing (s = 1.5 mm). Needle-like probes with a radius of 100 µm were utilized. The size of the samples was 1 × 1 cm 2 . A correction factor of 0.86 was considered during conductivity estimation, due to the finite sample size, and the layer thickness was estimated by spectroscopic ellipsometry measurements. The current was gradually increased from 0 up to 300 µA, with a step of 10 µA, with a source meter, and the voltage on the internal probes was measured with a Keithley 2400.

UV Visible Spectroscopy
UV-visible spectroscopy was measured using a double beam spectrophotometer (UV-9000 Metash, Shanghai, China) for the 200-1000 nm range with a scan step of 1 nm and a scan filter of 10. In this spectrophotometer, a deuterium lamp and a tungsten halogen lamp are used as light sources, a double beam optical system with 1200 lines/mm grating is implemented, and silicon photodiodes play the role of detectors. The wavelength repeatability is about 0.2 nm and the wavelength accuracy is equal to ±0.3 nm. Photometric accuracy is equal to about ±0.3 %T, and photometric repeatability is equal to 0.2 %T, where %T is percent transmittance.
In order to evaluate the band gap energy (E g ) of the samples, (αhν) n were plotted against the photon energy (hν), according to the well-known relation [26] where A is a constant, hν is the photon energy, and n depends on the type of allowed transition in the material. It has a value equal to 1/2, 2, and 3/2 for directly allowed transitions, indirectly allowed transitions, and directly forbidden transitions, respectively. While both allowed direct and indirect transitions were computed, those of a direct type are proposed in this study.

Results
The achieved transparency of B-CNW growth on quartz is shown in Figure 1, which shows a picture of the sample growth at 700 • C with a 2000 ppm [B]/[C] ratio.
Materials 2019, 12 FOR PEER REVIEW 4 Electrical conductivities were measured at room temperature by a four-point probe placed in a straight line with equal spacing (s = 1.5 mm). Needle-like probes with a radius of 100 μm were utilized. The size of the samples was 1 × 1 cm 2 . A correction factor of 0.86 was considered during conductivity estimation, due to the finite sample size, and the layer thickness was estimated by spectroscopic ellipsometry measurements. The current was gradually increased from 0 up to 300 μA, with a step of 10 μA, with a source meter, and the voltage on the internal probes was measured with a Keithley 2400.

UV Visible Spectroscopy
UV-visible spectroscopy was measured using a double beam spectrophotometer (UV-9000 Metash, Shanghai, China) for the 200-1000 nm range with a scan step of 1 nm and a scan filter of 10. In this spectrophotometer, a deuterium lamp and a tungsten halogen lamp are used as light sources, a double beam optical system with 1200 lines/mm grating is implemented, and silicon photodiodes play the role of detectors. The wavelength repeatability is about 0.2 nm and the wavelength accuracy is equal to ±0.3 nm. Photometric accuracy is equal to about ±0.3 %T, and photometric repeatability is equal to 0.2 %T, where %T is percent transmittance.
In order to evaluate the band gap energy (Eg) of the samples, (αhν) n were plotted against the photon energy (hν), according to the well-known relation [26] αhν where A is a constant, hν is the photon energy, and n depends on the type of allowed transition in the material. It has a value equal to 1/2, 2, and 3/2 for directly allowed transitions, indirectly allowed transitions, and directly forbidden transitions, respectively. While both allowed direct and indirect transitions were computed, those of a direct type are proposed in this study.

Results
The achieved transparency of B-CNW growth on quartz is shown in Figure 1, which shows a picture of the sample growth at 700 °C with a 2000 ppm [B]/[C] ratio.

Investigation of the Growth Process
Morphology of boron-doped carbon nanowalls plays an important role in the light transmission process and also affects the conductivity of the deposited material. It was found that a randomly vertically aligned wall structure belonged to the substrate only for samples with a temperature of process higher than 700 °C. B-CNWs are homogeneously distributed

Investigation of the Growth Process
Morphology of boron-doped carbon nanowalls plays an important role in the light transmission process and also affects the conductivity of the deposited material.  It was found that a randomly vertically aligned wall structure belonged to the substrate only for samples with a temperature of process higher than 700 • C. B-CNWs are homogeneously distributed all over the substrate and they have sharp top edges.
For B-CNWs with a temperature of process of 400 or 550 • C the structure resembles nanoislands-one of the growth stages suggested by Hiramatsu et al. [27].
In our studies, a significant effect of temperature on B-CNW length was observed ( Figure 3a). Samples with a temperature of process of 850 • C contain nanowalls approximately two times longer than for 700 • C samples. The distance between longer B-CNWs is greater than between the shorter ones. Temperature causes a decrease in the thickness of B-CNWs, which is correlated with the length of the nanowalls.

Absorbance Spectrum
Absorbance spectra for sample growth at different temperatures are shown in Figure 4a.  Boron incorporation, in different ratios at the same growth temperature, into the B-CNWs lattice causes a slight decrease in the thickness of B-CNWs (Figure 3b), which is in good agreement with research suggested by Sobaszek et al. [6].
It is worth pointing out that wall length is slightly positively influenced by the boron level in the gas phase. The sample 5000 ppm [B]/[C] at 700 • C has approximately 20 nm longer walls than the sample 0 ppm [B]/[C] at the same temperature of process.

Absorbance Spectrum
Absorbance spectra for sample growth at different temperatures are shown in Figure 4a. The absorbance for wavelengths of between 400 nm and 700 nm ranges between 0.2 and 0.8 with increasing growth temperature. In particular, a prominent change in average absorbance is observed in the 200 nm to 400 nm wavelength region (Figure 4b). This is due to a change in the morphology and composition of the B-CNW layer, as confirmed by SEM ( Figure 2) and Raman ( Figure 5). As the temperature rises, we see a more distinct shape of the walls. Also as the temperature rises, we get an absorption peak at the wavelength of 200-350 nm. The absorption for 850 • C for 270 nm is 2.4, which corresponds to a transmission below 0.002%. This spectroscopic feature consistently observed in the UV range is known to originate from the peak in the joint density of states for a simple graphene layer [28] which occurs at the M point of the Brillouin zone at E = 2γ 0 and is usually known as the plasmon resonance. For carbon nanotubes, this feature becomes shifted and split due to the more complex band structure, as has been predicted by several authors [29,30].

Absorbance Spectrum
Absorbance spectra for sample growth at different temperatures are shown in Figure 4a. The absorbance for wavelengths of between 400 nm and 700 nm ranges between 0.2 and 0.8 with increasing growth temperature. In particular, a prominent change in average absorbance is observed in the 200 nm to 400 nm wavelength region (Figure 4b). This is due to a change in the morphology and composition of the B-CNW layer, as confirmed by SEM ( Figure 2) and Raman ( Figure 5). As the temperature rises, we see a more distinct shape of the walls. Also as the temperature rises, we get an absorption peak at the wavelength of 200-350 nm. The absorption for 850 °C for 270 nm is 2.4, which corresponds to a transmission below 0.002%. This spectroscopic feature consistently observed in the UV range is known to originate from the peak in the joint density of states for a simple graphene layer [28] which occurs at the M point of the Brillouin zone at E = 2γ0 and is usually known as the plasmon resonance. For carbon nanotubes, this feature becomes shifted and split due to the more complex band structure, as has been predicted by several authors [29,30]

Raman Spectroscopy
Raman spectra of B-CNWs were analyzed and are reported in Figure 5. In all samples, two main bands at ∼1580 cm −1 and ∼1350 cm −1 are found, which are generally attributed to the G and D bands, respectively. In particular, Ferrari and Robertson have suggested that the G and D peaks are due to sp 2 only [31]. The band centered near 1350 cm −1 corresponds to the breathing mode of sp 2 atoms in a ring (D band), and Figure 5a and its intensity correlates with the amount of disorder of sp 2 -hybridized atoms in graphite-like materials. The different samples morphology, previously described and reported in Figure 2, is also highlighted by the presence of the D peak at 1350 cm −1 in the Raman spectra for the B-CNW growth at a higher temperature, which corresponds to distortion of the sp 2 crystal structure;

Raman Spectroscopy
Raman spectra of B-CNWs were analyzed and are reported in Figure 5. In all samples, two main bands at ∼1580 cm −1 and ∼1350 cm −1 are found, which are generally attributed to the G and D bands, respectively. In particular, Ferrari and Robertson have suggested that the G and D peaks are due to sp 2 only [31].
The band centered near 1350 cm −1 corresponds to the breathing mode of sp 2 atoms in a ring (D band), and Figure 5a and its intensity correlates with the amount of disorder of sp 2 -hybridized atoms in graphite-like materials. The different samples morphology, previously described and reported in Figure 2, is also highlighted by the presence of the D peak at 1350 cm −1 in the Raman spectra for the B-CNW growth at a higher temperature, which corresponds to distortion of the sp 2 crystal structure; by contrast, for the first two samples (at 400 • C and 550 • C), a peak appears earlier, at 1334 cm −1 , which can be attributed, instead, to the diamond structure [32].
The wide band in the range 1500-1600 cm −1 is dominated by overlapping G and D' bands. The G band centered at 1580 cm −1 is associated with bond stretching of sp 2 atoms in rings and chains (G band, Figure 5b). The G peak position is shifted to a higher frequency with higher temperatures and the full-width at half-maximum (G-FWHW) is also found to decrease with temperature. Marchon et al. have reported a similar trend and have linked it to a decrease in sp 3 content [33].
The D' band observed near 1500 cm −1 is usually caused by defects generated via double-resonance processes. The origin of the D' band in the plasma processed graphite nanomaterials can be associated with nitrogen defects incorporated into the graphene lattice [34,35]. The intensity of the D' band changes with plasma process temperature. The maximal intensity is observed for B-CNW film synthesized at 400 • C and the band disappears for films deposited at temperatures above 700 • C. Increasing the process temperature results in a reduction of defect concentration, which is also reflected in the narrowing of the D band.
The band at 1220 cm −1 is reported to be caused by the effect of boron on the diamond lattice [36]. However, this peak also appears in boron-free CNWs, which makes another hypothesis reasonable. Chen et al. [37] merged their results with theoretical analysis and experimental results from others to identify two additional Raman peaks at around 1168 cm −1 and 1271 cm −1 , in tetrahedral amorphous carbon, linked to the sp 3 bonding. In our results, reported in Figure 5c, it is possible to observe a distinct peak, ahead of the D band, which redshifts with increasing temperature. This peak can be assigned to the sp 3 carbons or disordered diamonds and is related to the phonon density of states (PDOS) band [38,39]. Moreover, from the Raman spectra it may be observed that when the boron concentration was increased further from a 2000 ppm [B]/[C] the G band peak upshifted by 6 cm −1 . A similar result was noted by McGuire et al. [40] for boron-doped single-walled carbon nanotubes.

Optical and Electrical Properties
In Figure 6a it is possible to observe how higher electrical conductivity belongs to sample growth at a higher temperature, in a similar fashion to nanowall length, as this parameter particularly affects charge mobility. Longer wall structure delivers higher conductivity, mostly due to the minimized effect of carrier scattering at the inter-wall regions and surface defects. Next, the optical bandgap shows a maximum for the sample growth at 700 • C, which is much larger than the values reported by Kawai et al. [41]. This effect is unrevealed, though it could be attributed to the specific molecular structure caused by boron doping. Boron incorporation into the carbon lattice slightly positively affects nanowall length, which is not directly translated into an enhancement of the electrical conductivity; however, the latter, jointly with the optical bandgap, depends on the [B]/[C] ratio (Figure 6b). Overall, the incorporation of boron into the CNW lattice structure induces the unique effect of enhanced electrochemical performance and improved charge transfer [42,43]. (PDOS) band [38,39]. Moreover, from the Raman spectra it may be observed that when the boron concentration was increased further from a 2000 ppm [B]/[C] the G band peak upshifted by 6 cm −1 . A similar result was noted by McGuire et al. [40] for boron-doped single-walled carbon nanotubes. In Figure 6a it is possible to observe how higher electrical conductivity belongs to sample growth at a higher temperature, in a similar fashion to nanowall length, as this parameter particularly affects charge mobility. Longer wall structure delivers higher conductivity, mostly due to the minimized effect of carrier scattering at the inter-wall regions and surface defects. Next, the optical bandgap shows a maximum for the sample growth at 700 °C, which is much larger than the values reported by Kawai et al. [41]. This effect is unrevealed, though it could be attributed to the specific molecular structure caused by boron doping. Boron incorporation into the carbon lattice slightly positively affects nanowall length, which is not directly translated into an enhancement of the electrical conductivity; however, the latter, jointly with the optical bandgap, depends on the [B]/[C] ratio ( Figure 6b). Overall, the incorporation of boron into the CNW lattice structure induces the unique effect of enhanced electrochemical performance and improved charge transfer [42,43].

Optical and Electrical Properties
The refractive index of B-CNWs, evaluated at 550 nm, ranges between 2.10 and 2.86 for sample growth at 700 °C, respectively, at 0 ppm and 5000 ppm boron doping levels; these results are comparable to those found by others [44,45]. Temperature also affects the refractive index; however, the variation is smaller and it falls within the previously reported range. The extinction coefficient shows a similar trend, reaching its minimum, equal to 0.83, with the sample growth at 700 °C and 0 ppm and 7500 ppm boron doping levels, and its maximum, equal to 1.24, with the 2000 ppm sample.
For sample growth at a temperature of 300 °C and 450 °C, which does not exhibit a nanowall structure, the extinction coefficient is much lower, indicating that the structure is similar to that of diamond-like carbon (DLC) film [46].

Discussion
Process duration affects mainly the thickness of the nanowall layer. Longer times lead to thicker layers, together with longer and more electrically conductive nanowalls, while light transmittance is The refractive index of B-CNWs, evaluated at 550 nm, ranges between 2.10 and 2.86 for sample growth at 700 • C, respectively, at 0 ppm and 5000 ppm boron doping levels; these results are comparable to those found by others [44,45]. Temperature also affects the refractive index; however, the variation is smaller and it falls within the previously reported range. The extinction coefficient shows a similar trend, reaching its minimum, equal to 0.83, with the sample growth at 700 • C and 0 ppm and 7500 ppm boron doping levels, and its maximum, equal to 1.24, with the 2000 ppm sample. For sample growth at a temperature of 300 • C and 450 • C, which does not exhibit a nanowall structure, the extinction coefficient is much lower, indicating that the structure is similar to that of diamond-like carbon (DLC) film [46].

Discussion
Process duration affects mainly the thickness of the nanowall layer. Longer times lead to thicker layers, together with longer and more electrically conductive nanowalls, while light transmittance is negatively affected. For these reasons, the right thickness is the starting point for obtaining a conductive and transparent composite.
Increasing nanowall length and layer thickness in accordance with boron doping are in contradiction to that reported by Takashi et al. [47] but in agreement with others [31,35,48].
Raman spectra and SEM pictures show that growing B-CNWs in low temperatures brings more defects than with elevated temperatures, with difficulties in developing a well-developed nanowall network on the substrate for temperatures lower than 700 • C, as well as a thinner B-CNW layer. Nanowall length and electrical conductivity are favorable with elevated growth temperatures.
The drop in growth rate versus stage temperature is caused by different surface pyrolysis reaction pathways involving H, CHx, HCN, and CN radicals [6]. The higher temperature favors bonding of CHx radicals towards the formation of a nanocrystalline diamond sp 3 phase observed in the Raman spectra. The growth rate of the diamond phase is much smaller than the nanowall formation kinetics [6].
The dependency of B-CNW average absorbance in the 400-700 nm range, as a function of the conductance, is reported in Figure 7. It is interesting to observe how the conductance and absorbance are slightly negatively correlated for different boron concentrations (blue •), while they are positively correlated for different growth temperatures (orange ).
the Raman spectra. The growth rate of the diamond phase is much smaller than the nanowall formation kinetics [6].
The dependency of B-CNW average absorbance in the 400-700 nm range, as a function of the conductance, is reported in Figure 7. It is interesting to observe how the conductance and absorbance are slightly negatively correlated for different boron concentrations (blue •), while they are positively correlated for different growth temperatures (orange △ .

Conclusions
In this study, a successful deposition of B-CNWs on a quartz glass substrate was conducted and characterization was carried out in order to study two parameters which strongly affect B-CNW growth and morphology: growth temperature and boron doping.
Process duration affects not only the thickness of the nanowall layer but also the development of the nanowall structure.
Raman spectrum and SEM pictures show that growing B-CNW in low temperatures brings more defects than with elevated temperatures, with difficulties in developing a well-developed nanowall network on the substrate observed for temperatures lower than 700 °C. Nanowall length and electrical conductivity are favorable with elevated growth temperatures.
Boron incorporation into the carbon lattice slightly positively affects nanowall length, which is not explicitly translated into an enhancement of the electrical conductivity.
These results suggest that it is possible to operate with both parameters to obtain a transparent and conductive film; however, boron doping is a preferred solution to maintain the transparency in the visible region, while a higher growth temperature is more effective to improve conductance.

Conclusions
In this study, a successful deposition of B-CNWs on a quartz glass substrate was conducted and characterization was carried out in order to study two parameters which strongly affect B-CNW growth and morphology: growth temperature and boron doping.
Process duration affects not only the thickness of the nanowall layer but also the development of the nanowall structure.
Raman spectrum and SEM pictures show that growing B-CNW in low temperatures brings more defects than with elevated temperatures, with difficulties in developing a well-developed nanowall network on the substrate observed for temperatures lower than 700 • C. Nanowall length and electrical conductivity are favorable with elevated growth temperatures.
Boron incorporation into the carbon lattice slightly positively affects nanowall length, which is not explicitly translated into an enhancement of the electrical conductivity.
These results suggest that it is possible to operate with both parameters to obtain a transparent and conductive film; however, boron doping is a preferred solution to maintain the transparency in the visible region, while a higher growth temperature is more effective to improve conductance.