Tailoring of structural and optical properties of electrosprayed β-Ga2O3 nanostructures via self-assembly

The present article demonstrates the fabrication of various β-Gallium Oxide (Ga2O3) nanostructures (NSs) by low-cost and scalable electrospraying (ES) methods via self-assembly (SA). The effect of the annealing sequences on the self-assembled β-Ga2O3 NSs has been detailed. The comparative studies of NSs and nanoflakes (NFs) growth with annealing effect at different stages of self-assembly time (SAT) have been explored further. The fabricated NSs and NFs have been investigated using X-ray diffraction, scanning electron microscopy, and transmission electron microscopy. The comparative study of categorized growth samples shows better crystalline quality when moving from the category 1 (C1) sample to the category 3 (C3) sample. The annealing sequences with SAT play a significant role in crystal formation and its quality, optical properties, and morphology of the Ga2O3 NSs. The optical properties of NSs have been derived from the normal incidence of absorbance measurements. The maximum observed energy band gap is ~ 5.40 eV. Traps and impurities play a critical role in the formation and deformation of energy bands in crystals. The photoluminescence spectra further reveal the variation in the intensity of luminescence of different emission bands due to the variation in the number of defect states and impurities in the NSs.


Introduction
Gallium Oxide (Ga 2 O 3 ) is a rapidly emerging wide bandgap semiconductor, that finds a wide range of applications in several regions such as luminescent phosphors, high-temperature sensors, anti-reflection coatings, solar cells, solar-blind photodetectors, high power electronics applications (Xu et al. 2019;Zhou et al. 2019). Ga 2 O 3 is also recognized as a deep ultraviolet transparent conducting oxide (UV-TCO), which makes the material a potential candidate for transparent electrode applications in UV optoelectronics (Stepanov et al. 2016). The same property has been used for photodiodes in several reports to enhance photoresponsivity by a light-trapping mechanism (Kumar and Bag 2019).
To date, several types of β-Ga 2 O 3 structures have been synthesized to realize different applications (Du et al. 2016;Weng et al. 2010). Various methods to deposit the β-Ga 2 O 3 thin film or nanostructures (NSs) have already been reported according to their practical applications in different areas (Yadav et al. 2020;Patil-Chaudhari et al. 2017;Huang et al. 2017;Nakagomi et al. 2013). Micro-flakes , nano-sheets (Feng et al. 2014), nano-webs (Graham et al. 2003), and NSs Kumar et al. , 2022Qiao et al. 2020;Xu et al. 2007) have been fabricated and implemented for different devices. Apart from the various conventional growth techniques, self-assembly has also been proven to be one of the inexpensive growth techniques for the synthesis of β-Ga 2 O 3 NSs (Kumar et al. 2022;Qiao et al. 2020;Xu et al. 2007).
The self-assembled nature of β-Ga 2 O 3 nanoflakes helps in the formation of a twodimensional (2D) structure of the semiconductor (Graham et al. 2003;. When a drop of dilute colloidal nano-sphere suspension spreads on a flat substrate, after evaporation of the solvent, a well-ordered 2-D hexagonal close-packed (hcp) colloidal crystal can be obtained. The intermolecular forces govern the particle interaction in selfassembled systems rather than ionic or covalent bonds. The ionic or covalent bond can lock the assembly into non-equilibrium structures which are not desirable for self-assembly (Xu et al. 2007). The Self-Assembly Phenomenon (SAP) for the growth of β-Ga 2 O 3 NSs has been applied along with the combination of electrospraying (ES) and low-pressure chemical vapor deposition (LPCVD) in many literatures (Kumar and Bag 2019;Zainizan Sahdan et al. 2010;Park et al. 2012). Graham et al. reported the self-assembly of β-Ga 2 O 3 for the very first time in 2003 (Graham et al. 2003). Since then, several other articles have also been reported on the self-assembly technique for the growth of β-Ga 2 O 3 NSs. Ying Guo et al. found a similar kind of growth of β-Ga 2 O 3 NSs and variation in their morphology along with optical properties with different growth conditions (Guo et al. 2008). Some other studies have also been done on the growth of β-Ga 2 O 3 which reveals new dimensions in search of an inexpensive and more practical way for the fabrication of β-Ga 2 O 3 (Winkler et al. 2019;Saha et al. 2019;Zatsepin et al. 2018).
It has been also found after several investigations that the material properties of the specific semiconductor such as energy bandgap and luminescence behavior could also vary with different deposition techniques. Lu Huang et al. investigated the bandgap tuning of Ga 2 O 3 thin film with the amount of O 2 exposure during deposition (Huang et al. 2017). In this context, the present research explains three different growth conditions with the combination of self-assembly time and different annealing sequences. The effect of growth conditions on morphologies, crystallinity, energy bandgap, and photoluminescence properties of β-Ga 2 O 3 NSs have also been studied in detail.

Electrospraying (ES)
Both electrospraying (ES) and electrospinning are solution-based techniques. They differ only in the concentration of the solution under consideration. If the solution is very dilute in comparison with the standard concentration of solution for electrospinning, there will be electrospraying rather than electrospinning (Zatsepin et al. 2018). Presently, the aqueous solution for the deposition of nanofibers by ES has been prepared by mixing PVP (Poly Vinyl Pyridine) and Gallium Chloride (GaCl 3 ) in a 5:1 weight ratio with 10 ml of de-ionized (~ 18 MΩ) water. Then the solution has been taken into a syringe and the substrate has been placed on an aluminum-wrapped drum (collector). A high electric field has been applied between a needle of the syringe and the rotating spindle, where both act as electrodes. The Si-Substrate has been taken for deposition of all samples.
The electrospraying parameters for the nanofibers deposition are as follows: syringe diameter = 13.8 mm, electrode spacing = 12 cm, applied voltage between electrodes = 25.6 kV, flow rate = 0.05 μL/s. The high electric field between the needle of the syringe and collector, the small distance between the syringe and collector, and the very low flow rate of the solution allow diluting and colloidal nano-sphere suspension to spread on a flat substrate of the collector. The time of electrospraying was ~ 4 h.

Samples and self-assembly
The samples are then kept in a dry vacuum (desiccator) for 15 days which provides an ample amount of time for self-assembly. Fifteen days of storing time in the vacuum (in a desiccator) is named self-assembly time (SAT). In 2012, Woon Ik Park et al. demonstrated forced self-assembly and studied the phenomenon by varying time and temperature (Park et al. 2012). In this paper, the self-assembly time of 15 days has been selected after several initial trials. The morphological study has been done before 15 days, after 15 days, and before 1 month. There was no significant improvement in the raw deposition before 15 days. The deformation from the actual structure has been seen after one month. It has been observed that the SAT of 15-20 days is more appropriate.
Nano-sphere-deposited samples with ES were strategically divided into three categories as in Table 1. The first category of samples (C1) was stored in the desiccator for 15 days just after the ES. There was no external intervention in these sets of samples apart from the initial ES.
The second category of the sample (C2) was observed for the annealing effect with selfassembly. The annealing was performed before and after the 15 days of SAT for these category samples. The third category of the sample (C3) was observed for only the post-annealing effect after the SAT. C3 samples were stored in a desiccator for 15 days just after the ES (or just after the deposition). Samples were taken out for annealing after 15 days of SAT.

Annealing sequence and NFs formation
Samples were placed into a horizontal tubular furnace for annealing under a mixed environment of Argon (Ar) and Oxygen (O 2 ) with a 50 sccm and 15 sccm flow rate respectively. The annealing was shown a significant effect on the growth of nanoflakes (NFs) which is discussed in the next section. The annealing temperature profile is shown in Fig. 1. The deposited (GaCl 3 + PVP) NSs get oxidized while annealing in the presence of O 2 . The following reaction is expected to take place while annealing the NSs which has been verified by XRD.
Samples were annealed at 300 °C, 500 °C, and 700 °C for 2 h each as shown in Fig. 1. Annealing is a continuous and sequential process that helps to adjust the thickness of the flakes also (Kumar et al. , 2022. The SAT and annealing details of the three categories of samples are summarized in Table 1.
The electrospraying process is explained in Fig. 2 which explains the physics of the growth process with ES in three steps explaining initial growth procedures and factors affecting the shape and morphologies of NSs.
(1) 4GaCl 3 + 3O 2 = 2Ga 2 O 3 + 6Cl 2 Fig. 1 Annealing profile for the samples of categories C2 and C3. The temperature has been scaled in three stages. The first stage (300 °C) is to remove the polymers from the samples and then 500 °C and 700 °C to modify the morphology of NSs (1) The high electric field between the syringe needle and the moving spindle accelerates spontaneously emitting a colloidal nano-sphere more towards Si-substrate.
(2) The electric field provides nano-spheres enough momentum to get flattened on the Si-substrate and form a circular 2-D-like possible structure as shown in Fig. 2. (3) The Si-substrate stuck on the moving spindle was already in motion while the deposition of nano-spheres. The struck nono-spheres experience two types of forces due to the moving of the spindle: (a) the static friction between the deposited nano-sphere and Si-substrate which holds the deposited structure with the Si-substrate, and (b) the centrifugal force which stretches the upper portion of the structure on the Si-substrate in the direction of the spindle movement. The process makes the punctured nano-sphere thinner and gives initial growth direction which becomes most important during the SAT.
Annealing time duration has been kept constant throughout the experiment, but the annealing has been done on three different sequences to see the morphological changes in the samples.
The structural properties of GONFs were studied using an X-ray diffractometer (Smart lab, Rigaku Japan; XRD) equipped with high-intensity Cu-Ka radiation (λ = 0.154 nm). The sample was characterized by a 2θ range from 10° to 90° at 45 kV and 100 mA. The omega for maximum intensity has been optimized at 1.5° with a step size of 0.04 and 1.5 s per step. Transmission electron microscopy (TEM) (Techn-aiG2, FEI) was used to determine the morphology and crystallinity of the NFs with the selected area electron diffraction (SAED) pattern of the lattice. The morphology and size of the flakes were identified using scanning electron microscopy (SEM, Raith). The optical properties have been observed using photoluminescence (PL) (LabRAM HR visible, Horiba Jobin Vyon) and UV-VIS (LAMBDA 750 UV/Vis/NIR, PerkinElmer, USA). The He-Cd laser used for these spectra was having a wavelength of 325 nm; the excitation intensity was in the range of 5-40 mW.

Material characterization
The SEM was done for all three categories of samples to understand morphological evolutions. Figure 3 shows two different regions of the C1 category of the sample. Figure 3a shows the flakes at a 10 µm scale whereas Fig. 3b shows flakes at a 3 µm scale. The shape of flakes is random in this category of samples. Therefore, it has been named a pre-matured flake. The size of flakes is asymmetrically ranging from micrometers to nanometres. The C1 category of the sample was not annealed. Therefore, it is expected to have a dominating self-assembly effect, but it has been observed from SEM that the growth of flakes was random.
The morphology of the category 2 (C2) sample is shown in Fig. 4a and b. In these samples, nanowires and beads are observed as a more dominant nanostructure. Since selfassembly requires no external interference and hence the pre-nucleation was not done  by self-assembly; these grown structures have more annealing effect than self-assembly (Kumar et al. 2022). Figure 5 shows the SEM image of the third category of samples. The morphology of samples was found more leaf-like structures as shown in Fig. 5a and b. These samples have been analyzed just after 15 days of SAT and before annealing. Figures 5c and d show the morphology of NFs after annealing. These flakes are categorized and named as matured flakes. The shape of the flakes is again confirmed with the TEM image shown in Fig. 9a. The nucleation has been done by self-assembly during SAT time from which annealing took its role to give a directional growth as a well-shaped nanoflake. The annealing parameters have been shown in Fig. 1 for C2 and C3 samples. Since all the samples are nanostructures, it was difficult to measure the exact thickness but the average thickness of samples has been estimated as ~ 100 nm by observation of several deposited samples and their duration of deposition time.
The Si peaks in all three categories of samples have not been shown in Fig. 6, because of their very high intensity in comparison to other visible peaks of Ga 2 O 3 . The correlation between the morphologies of all three categories of samples and their XRD plots infer the change in the crystalline structure of Ga 2 O 3 under different growth conditions. From C1 to C3, the diffraction peak of 2θ at 33.4° has emerged stronger in the absence of other peaks. Thus, the crystal structure has improved from a polycrystalline to a single-crystal structure. Therefore, it is evident that self-assembly (SA) and annealing seem to be helpful in the growth of a single orientation-dominated crystalline structure by providing the pre-nucleation for the growth of β-Ga 2 O 3 along with annealing under the mixed environment of Ar and O 2 for the formation NFs.
Further investigation of the crystal properties of all three categories of samples has been done by Transmission electron microscopy (TEM). It includes TEM imaging to observe the morphology of the samples, Fast Fourier transform (FFT) image which Fig. 6 a XRD plot of C1 samples, b XRD plot of C2 samples, and c XRD plot of C3 samples which confirms the differently oriented crystalline structure of Ga 2 O 3 NSs. Comparing the XRD of all three samples, the C3 category of samples is more toward the single-crystal structure than the other two. The Si peak was around 69º which has not been presented to avoid its large intensity Page 9 of 18 437 shows the crystalline nature of the samples, high-resolution TEM imaging to provide d-spacing of the crystals, and selected area electron diffraction (SAED) which attributes the crystal orientation of the samples along with crystallinity. Figure 7a depicts the TEM image of C1 samples which confirms the premature morphology of the NFs. Figure 7b represents the FFT image of the same sample which shows the polycrystalline nature of this category of samples with multiple spots in the figure. Figure 7c reveals the d-spacing of NFs as ~ 0.48 nm as reported in Zainizan Sahdan et al. (2010). The indexed SAED with a dominating circular pattern further confirms the polycrystalline nature of the C1 category samples as shown in Fig. 7d.
Similarly, Fig. 8a represents the morphology of C2 samples. The FFT image as in Fig. 8b, reveals the improved crystalline quality of the samples in comparison with C1 samples. Figure 8c reveals the d-spacing of NFs as ~ 0.26 nm. The indexed SAED pattern further confirms the better crystalline nature of the C2 category samples in the absence of circular patterns, as shown in Fig. 8d (Muhammad et al. 2011). Figure 9a delineates the morphology of C3 samples which agrees with the SEM image shown in Fig. 5 a and b. Figure 9a again confirms the leaf-like shape of matured NFs. The FFT image is shown in Fig. 9b, showing the improved crystalline quality in comparison with C1 samples. Figure 9c reveals the d-spacing of NFs as ~ 0.26 nm and ~ 0.23 nm which corresponds to (−111) and (−311) planes correspondingly. The two different d-spacings may be attributed to two different sets of planes for two different overlapped NFs. The spotty and sharp SAED pattern further confirms the crystalline  Fig. 9d. Therefore, TEM characterization vividly reveals the monotonic improvement of Ga 2 O 3 crystal quality from C1 to C3 and C2.
As the morphology of all categories of samples has been observed through SEM and TEM, Fig. 10 shows a conclusive schematic. It explains the experimental procedure and its outcomes in terms of morphological evolution. It has also been the key point for categorizing the samples into different groups at the beginning. The C1 category samples have not been annealed at any stage. Hence, in this category of samples, the self-assembly has only been increasing the size of deposited nanodots in a random shape. The crystallinity of the C1 category of samples has also been low without annealing. The impurities level is also high due to polymer content which has been observed through photoluminescence analysis. The C2 category of samples has been annealed before and after SAT. The initial annealing (before SAT) of the C2 category of samples locks the deposited nano-dots into random shapes after the electrospraying. Therefore, the self-assembly effect is negligible in the C2 category of samples, but the crystallinity of the samples has been enhanced due to annealing. The C3 category of samples has gone through annealing after the SAT. Therefore, the initial nucleation has been done during the SAT. Then, the annealing provided the final shape to the matured NFs for the C3 category of samples with better crystallinity. Figure 11a shows the absorbance spectra of the C1, C2, and C3 samples. All the samples represent almost similar absorbance spectra. The magnitude of the absorbance intensity is the only visible difference. The difference has been again confirmed by comparing  The amorphous and nanocrystalline Ga 2 O 3 films show high transparency in the spectral range except where the incident radiation is absorbed across the bandgap (E g ) (Stepanov et al. 2016). It indicates the transparent nature of all the Ga 2 O 3 films. Further analysis of the optical spectra is performed to understand the effect of morphology and crystallinity of microstructure on the optical properties which leads to deriving a quantitative structural property and co-relationship. For β-Ga 2 O 3 with a direct bandgap, the absorption follows a power law of the form (Yu et al. 2015), where hν is the energy of the incident photon, α absorption coefficient, B is the absorption edge width parameter, and E g is the bandgap. The optical absorption coefficient 'α' of the β-Ga 2 O 3 nanostructures is evaluated using the relation (Kokubun et al. 2007), where T is the transmittance, R is the reflectance, and t is the thickness of the grown nanostructures on a silicon substrate.
The absorption spectra and their plots for absorption intensity with variation in the wavelength of incident light have been plotted for all three categories of samples (from C1 to C3) of Ga 2 O 3 NSs which are shown in Fig. 11a. It is evident that (αhν) 2 vs hν results in linear plots in the high absorption region, suggesting directly allowed transitions of electrons across E g of Ga 2 O 3 nanostructures. The corresponding plots of (αhν) 2 vs hν are . 11 a Absorbance spectra of C1, C2, and C3 categories of samples. After comparing all three absorbance spectra, only a slight difference has been observed which has been confirmed later by comparing the tau plot to see the difference in energy bandgap. 9 b shows (αhν) 2 vs hν plots for C1, C2, and C3 samples respectively. Extrapolating the linear region of the plot to hν = 0 provides the bandgap (E g ) shown in Fig. 11b, where the E g value can be obtained by extrapolating the linear portion to the photon energy axis which has been indicated with an arrow. The E g variations can be observed for Ga 2 O 3 nanostructures (GONSs) grown in the different conditions as in Fig. 11b for the C1, C3, and C2 categories of samples respectively. The E g value of GONSs grown without annealing shows the highest E g (5.40 eV). The higher values of E g measured for GONS may be due to the polycrystalline nature, high impurities of polymers, and excess oxygen because of the high porosity of the nanostructures (NSs) (Mohamed et al. 2011;Shi et al. 2019). Apart from these two factors, the impurity in the NSs also plays a pivotal role to increase the disorders in the lattice which may lead to a slightly higher bandgap as measured for NSs grown at room temperature (Kokubun et al. 2007). As the NSs of the C1 sample did not go through annealing to remove the polymers from the sample, the impurities in the sample are possibly much higher than the other two categories of samples causing the highest bandgap.
It also reveals the decrement in E g from 5.40 to 5.29 eV as the quality of NSs improved with changes in the growth environment and annealing sequences. These outcomes are in good agreement with the previously reported results by Kokubun et al. (2007). Hence, it can be concluded from earlier and this experiment that polymers-based electrospraying techniques carry polymer residue as impurities after deposition of NSs which can be deduced with a planned annealing scheme. The annealing scheme also improves the morphologies and decreases the impurities level to get closure to the energy bandgap up to the pure β-Ga 2 O 3 .
Although Fig. 11b shows the lowest bandgap is 5.29 eV which corresponds to the C2 category of samples not the C3 category of samples because the C2 sample has been annealed twice. The C2 samples have been annealed just after the deposition and then after the SAT which may lead to less trapped and suspended impurities inside the C2 in comparison with C1 and C3.
It has already been observed that the quantum confinement effects in micro-crystallites occur when the particle size is very small (typically smaller than the excitons Bohr radius). Ruan et al. reported the excitons Bohr radius for Ga 2 O 3 is 3.29 nm (Ruan et al. 2017) and therefore, the present increase of the optical bandgap cannot be caused by quantum confinement effects because the nanostructure size of the samples is comparatively very large.
The deposition of NFs is mainly incorporated by the SA which leads to high concentrations of unsaturated defects which means the SA growth happens horizontally due to the high intermolecular force of the material which leads to a high surface-to-volume ratio. The XRD and TEM also show the crystallinity of NSs has been improved from C1 to C2 and C3 as the self-assembly is guided by annealing in O 2 ambient which results in a gradual decrease in defects and possible unsaturated O 2 at the surface of NSs. Therefore, the decrement of the optical bandgap can potentially be attributed to the decrease in the density of defect states.

Material physics
In the case of NSs, the surface-to-volume ratio is very high compared to the thin film. More surface area leads to more dangling bonds which again leads to more surface defects. UV luminescence is very sensitive to these surface defect states because the conduction band electrons (responsible for this luminescence band) are trapped in these defect states and recombine non-radiatively (Ruan et al. 2017;Kumar and Singh 2013). The UV luminescence peak is almost invisible in the room temperature PL spectra shown in Fig. 12a-c for the C1, C2, and C3 categories of samples respectively. Band-to-band transition in the deep UV region is not expected in the spectrum due to the excitation with the 325 nm wavelength of the light source. Figure 12a represents the room temperature PL spectra of the C1 sample in which a UV luminescence is almost invisible, blue, and green emission band is observed at 404 nm and 532 nm respectively which is lower than the emission bands of C2 and C3 samples shown in Fig. 12b and c respectively. Unlike the PL spectra of C2 and C3, the PL spectra of C1 also possess a red emission band due to a higher impurities level than the other two categories of samples. Comparing the PL spectra of all three categories of samples, the intensity of blue luminescence is almost unaffected in all three categories of samples because blue emission happens due to radiative recombination of trapped excitons in the donor (V O++ ) and acceptor level (V Ga , V Ga-and V O++ ) as shown in Fig. 13. The surface defects state does not affect the blue luminescence (Binet and Gourier 1998).
Green luminescence is potentially arising due to the presence of some impurities such as Be, Ge, Sn, Li, Zr, and Si, but this was a very early hypothesis given by L. Binet et al. in Fig. 12 a PL analysis of the C1 sample. The luminescence has been observed in three bands blue, green, and red except in the UV band due to very high surface defects. b PL analysis of the C2 sample. The luminescence has been observed in two bands blue and green with good intensity. The intensity of red band luminescence has been very low due to a very lesser impurity level than the C1 sample. c PL analysis of the C3 sample. The luminescence has been observed in two bands blue and green with good intensity. The intensity of red band luminescence has been very low due to a very lesser impurity level than the C1 and C2 categories of samples 1998 (Binet and Gourier 1998). It has been recently reported that interstitial impurities of the oxygen atoms in the lattice of β-Ga 2 O 3 are responsible for the green luminescence and it has also been seen in a few articles on the growth of β-Ga 2 O 3 that the intensity of green emission increases with an increase in the oxygen partial pressure of the growth environment (Kumar and Singh 2013;Binet and Gourier 1998). The details of various luminescence emission bands have been explained in Fig. 13. The present hypothesis supports the PL spectra shown for all three categories of samples. Since the first category (C1) of the sample has not been annealed, the oxygen interstitial impurities are relatively less so than the intensity of green emission. C2 has been annealed twice in the oxygen ambient, hence the intensity of the green emission is the highest of all. The third category of the sample (C3) has been annealed once; hence the intensity of green luminescence is greater than C1 but less than C2 (comparing Fig. 12a-c)). The dimensions of NSs studied in this work are far greater than the Bohr radius. Hence, quantum confinement is not applicable in this study of the β-Ga 2 O 3 NSs. The red luminescence is also visible in the C1 category sample but not in the other two categories. The red emission mainly happened due to the presence of foreign elements. The red luminescence due to doping of In 3+ and Cr 3+ has already been reported in 2016 by Li et al. (2016). Since only the first category (C1) of the sample was not annealed, it contains more impurities of polymers than the other two categories of samples. This corroborates the hypothesis of red luminescence only in the first category of the sample due to foreign elements. Jangir et al. had shown a similar type of comparative study with photoluminescence (Jangir et al. 2016). In 2001, Liang et al. as well as some other reports studied the photoluminescence behavior of β-Ga 2 O 3 NWs similarly (Liang et al. 2001).

Conclusion
This paper has shown the self-assembly nucleated growth of β-Ga 2 O 3 NSs and NFs which has been guided by annealing in the O 2 ambient. The maturity of grown NSs and NFs differs due to different annealing conditions which have been verified with Fig. 13 Proposed a Model of Luminescence Spectra for Gallium Oxide morphological analysis such as SEM and TEM. The quality of the crystalline structure has been enhanced through well-timed annealing with SAT. It has also been observed that the premature shape of NSs without annealing has a greater number of impurities and polycrystalline tending nature compared to the other categories of samples which may be responsible for the possible increment in energy band gaps (E g ). The PL spectra have also been studied further and it has been concluded that the intensity of green emission spectra may either decrease or increase with a decrease or increase in interstitial impurities. However, the blue emission is not sensitive to surface defects, and it remains unchanged throughout the process. All the characterization data shows the formation of Ga 2 O 3 with varying bandgap. As the fabricated material which has been annealed (i.e., C2 and C3) shows a very good crystalline structure. Hence, the grown Ga 2 O 3 NSs can be implemented in photodetectors, power devices, solar cells, fluorescence lamps, and some other optical devices.