Aero-Ga2O3 Nanomaterial Electromagnetically Transparent from Microwaves to Terahertz for Internet of Things Applications

In this paper, fabrication of a new material is reported, the so-called Aero-Ga2O3 or Aerogallox, which represents an ultra-porous and ultra-lightweight three-dimensional architecture made from interconnected microtubes of gallium oxide with nanometer thin walls. The material is fabricated using epitaxial growth of an ultrathin layer of gallium nitride on zinc oxide microtetrapods followed by decomposition of sacrificial ZnO and oxidation of GaN which according to the results of X-ray diffraction (XRD) and X-ray photoemission spectroscopy (XPS) characterizations, is transformed gradually in β-Ga2O3 with almost stoichiometric composition. The investigations show that the developed ultra-porous Aerogallox exhibits extremely low reflectivity and high transmissivity in an ultrabroadband electromagnetic spectrum ranging from X-band (8–12 GHz) to several terahertz which opens possibilities for quite new applications of gallium oxide, previously not anticipated.


Introduction
The materials transparent for a certain electromagnetic bandwidth are key components for many industries such as aeronautic, space, telecommunications, etc. [1,2]. They are called radomes and are configured in the form of various enclosures depending on the applications; their role is to protect antennas from various agents such as rain, snow, dust, heat, etc. The radomes can be seen, for example,

Materials and Methods
The technological route for the fabrication of aero-Ga 2 O 3 is as follows. Initially the aero-GaN was obtained by growing an ultra-thin layer of GaN on sacrificial ZnO templates [10]. The ZnO templates represented networks of interpenetrated ZnO microtetrapods obtained using the flame transport synthesis approach, as previously described in ref. [11]. GaN was grown in a hydride vapor phase epitaxy (HVPE) horizontal reactor containing distinct source and reaction zones. Metallic gallium as well as ammonia (NH 3 , 99.99%, EG No. 231-653.3), hydrogen chloride (HCl, 99.9995%) and hydrogen (H 2 , 99.999%) acquired from Geschaftsbereich Linde Gas, Germany, were used as source materials and carrier gases during the growth process. In the source zone, at high temperature (T = 850 • C) the GaCl is formed as a result of chemical reactions between gaseous HCl and liquid Ga. The gaseous GaCl and NH 3 reacted with each other in the react zone, where initially the temperature was kept at 600 • C for 10 min to initiate nucleation of GaN on the surface of ZnO microtetrapods, and then increased up to T g = 850 • C for 10 min to produce GaN layer. In the process of GaN growth, the HCl, NH 3 and H 2 flow rates were equal to 15, 600 and 3600 smL/min, respectively. In the process of HVPE growth of GaN, the ZnO sacrificial template is being decomposed due to corrosive atmosphere at high temperature leading to the formation of microtubular structures representing aero-GaN [10]. Previously, we demonstrated the high crystalline quality of the resulting GaN microtubes as well as the existence of ZnO traces (at the level of about 2%) on the inner surface of GaN microtube walls [10,12]. At the final step of the technological route, the aero-GaN is subjected to annealing in air at 950 • C for 60 min and, as a result, is transformed into aero-Ga 2 O 3 or Aerogallox.
The chemical composition at surfaces was studied by means of X-ray Photoemission Spectroscopy, XPS (AXIS ULTRA, DLD Kratos Analytical, Manchester, UK) equipped with a monochromatic Al Kα source (hν = 1486.6 eV) operating at 150 W (h-Planck's constant, ν-frequency). Spectra were recorded from a 100 × 100 µm 2 analysis area and at 160 eV (survey) pass energy, whereas core level spectra were recorded using pass energy of 20 eV. Operating pressure was 6 × 10 −7 Pa. Prior to the measurement, the surfaces were sputtered using an Ar + beam operated at 2 keV and 0.84 µA for 2 min. Surface charging was compensated using low energy (~5 eV) electrons and adjusted using the charge balance plate on the instrument. Three different spots were analyzed for each sample. All the spectra were processed with CasaXPS (ver. 2.3.20). Spectra were calibrated setting hydrocarbon C1s at 285.0 eV. The surface composition was evaluated from the survey spectra, after a Tougaard U3-type background subtraction, using relative sensitivity factors provided by the manufacturer. Peak fitting was performed with no preliminary smoothing. Symmetric Gaussian-Lorentzian (70% Gaussian and 30% Lorentzian) product functions were used to approximate the line shapes of the fitting components.
For the microwaves and terahertz characterizations of the material, bulk samples were prepared in the form of rectangular pellets (20 mm × 10 mm × 2 mm). The freestanding samples were exposed to electromagnetic radiation. The microwave characterization of the aero-Ga 2 O 3 pellets was carried out by means of a VNA (Vector Network Analyzer) connected to a WR90 waveguide-based set-up suitable for measurements in the X-band (i.e., 8.2-12.4 GHz) [13]. A schematics of the waveguide is presented in Figure 1, where a = 22.86 mm and b = 10.16 mm. Since the dimensions of the cavity are a bit larger than the aero-Ga 2 O 3 pellet, we made use of a supporting flange to fix the sample in the cavity, as to avoid electromagnetic radiation due to imperfect coupling between the two waveguides, hence affecting the reliability of the performed measurements. Nanomaterials 2020, 10, x FOR PEER REVIEW 3 of 11 The chemical composition at surfaces was studied by means of X-ray Photoemission Spectroscopy, XPS (AXIS ULTRA, DLD Kratos Analytical, Manchester, UK) equipped with a monochromatic Al Kα source (hν = 1486.6 eV) operating at 150 W (h-Planck's constant, νfrequency). Spectra were recorded from a 100 × 100 µ m 2 analysis area and at 160 eV (survey) pass energy, whereas core level spectra were recorded using pass energy of 20 eV. Operating pressure was 6 × 10 −7 Pa. Prior to the measurement, the surfaces were sputtered using an Ar + beam operated at 2 keV and 0.84 µ A for 2 min. Surface charging was compensated using low energy (~5 eV) electrons and adjusted using the charge balance plate on the instrument. Three different spots were analyzed for each sample. All the spectra were processed with CasaXPS (ver. 2.3.20). Spectra were calibrated setting hydrocarbon C1s at 285.0 eV. The surface composition was evaluated from the survey spectra, after a Tougaard U3-type background subtraction, using relative sensitivity factors provided by the manufacturer. Peak fitting was performed with no preliminary smoothing. Symmetric Gaussian-Lorentzian (70% Gaussian and 30% Lorentzian) product functions were used to approximate the line shapes of the fitting components.
For the microwaves and terahertz characterizations of the material, bulk samples were prepared in the form of rectangular pellets (20 mm × 10 mm × 2 mm). The freestanding samples were exposed to electromagnetic radiation. The microwave characterization of the aero-Ga2O3 pellets was carried out by means of a VNA (Vector Network Analyzer) connected to a WR90 waveguide-based set-up suitable for measurements in the X-band (i.e., 8.2-12.4 GHz) [13]. A schematics of the waveguide is presented in Figure 1, where a = 22.86 mm and b = 10.16 mm. Since the dimensions of the cavity are a bit larger than the aero-Ga2O3 pellet, we made use of a supporting flange to fix the sample in the cavity, as to avoid electromagnetic radiation due to imperfect coupling between the two waveguides, hence affecting the reliability of the performed measurements. For terahertz measurements the samples were fixed on metallic holders covering 6 mm aperture, the radiation passing through the sample in free space. Samples with different densities from 70 to 110 mg/cm 3 of aero-Ga2O3 were investigated. Room-temperature terahertz spectra of complex dielectric permittivity ε*(ν) = ε'(ν) + iε"(ν) were measured in the range ν = 10-100 cm -1 in a quasioptical arrangement (in an open space with no waveguides used) with the help of commercial timedomain spectrometer TeraView TPS 3000 (TeraView, Cambridge, UK). The spectra of the real and imaginary parts of dielectric permittivity are determined in the transmission geometry via measurements of the complex transmission coefficient (amplitude and phase) of the plane-parallel samples; standard expressions for electrodynamics of a plane-parallel layer are used [14]. In addition, the spectra of transmission coefficients of the same samples were measured at frequencies up to 7000 cm −1 using a standard Fourier-transform spectrometer Bruker Vertex 80v. For terahertz measurements the samples were fixed on metallic holders covering 6 mm aperture, the radiation passing through the sample in free space. Samples with different densities from 70 to 110 mg/cm 3 of aero-Ga 2 O 3 were investigated. Room-temperature terahertz spectra of complex dielectric permittivity ε*(ν) = ε'(ν) + iε"(ν) were measured in the range ν = 10-100 cm -1 in a quasi-optical arrangement (in an open space with no waveguides used) with the help of commercial time-domain spectrometer TeraView TPS 3000 (TeraView, Cambridge, UK). The spectra of the real and imaginary parts of dielectric permittivity are determined in the transmission geometry via measurements of the complex transmission coefficient (amplitude and phase) of the plane-parallel samples; standard expressions for electrodynamics of a plane-parallel layer are used [14]. In addition, the spectra of transmission coefficients of the same samples were measured at frequencies up to 7000 cm −1 using a standard Fourier-transform spectrometer Bruker Vertex 80v.      [16,17]. Data show that O 1s peak (Figure 4c) need to be resolved using two components. The main component BE at 530.98 eV can be attributed to Ga-O bonding in the oxide while the higher BE contribution at 533.17 eV can be assigned to O-vacancies sites and/or Ga suboxides [18][19][20]. This peak presence is related to the Ar ions sputtering process used to clean the sample surface in order to remove adventitious carbon contamination. Presence of reduced Ga can be seen also in the low energy satellite peak (at BE 19.16 eV) of Ga 3d (not shown). However, sputter cleaning of the sample was necessary for reliable estimation of atomic concentrations; before Ar + bombardment C atoms surface concentration was estimated to be around 11% while after cleaning dropped down to 2%. Measured Ga/O ratio is thus 0.62 with an estimated relative error of 6.21% (stoichiometric value 0.67). The presence of Zn 2p doublet (Figure 4d) suggests some residual (<1 at. %) of metallic impurities left from the sacrificial ZnO matrix used in the fabrication process. It is to be noted that the chemical composition study using energy dispersive X-ray analysis disclosed the stoichiometric composition of the Ga 2 O 3 nanoarchitecture, at the same time traces of Zinc at the level of about 1.5 at. % were shown.   Figure 4b), Ga 3p and Ga 3d (not shown) and several Auger lines in the 400-600 eV regions. All Ga doublets show energy shifts with respect to metallic gallium binding energy (BE), consistent with literature reported values for β-Ga2O3 [16,17]. Data show that O 1s peak (Figure 4c) need to be resolved using two components. The main component BE at 530.98 eV can be attributed to Ga-O bonding in the oxide while the higher BE contribution at 533.17 eV can be assigned to O-vacancies sites and/or Ga suboxides [18][19][20]. This peak presence is related to the Ar ions sputtering process used to clean the sample surface in order to remove adventitious carbon contamination. Presence of reduced Ga can be seen also in the low energy satellite peak (at BE 19.16 eV) of Ga 3d (not shown). However, sputter cleaning of the sample was necessary for reliable estimation of atomic concentrations; before Ar + bombardment C atoms surface concentration was estimated to be around 11% while after cleaning dropped down to 2%. Measured Ga/O ratio is thus 0.62 with an estimated relative error of 6.21% (stoichiometric value 0.67). The presence of Zn 2p doublet (Figure 4d) suggests some residual (<1 at. %) of metallic impurities left from the sacrificial ZnO matrix used in the fabrication process. It is to be noted that the chemical composition study using energy dispersive Xray analysis disclosed the stoichiometric composition of the Ga2O3 nanoarchitecture, at the same time traces of Zinc at the level of about 1.5 at. % were shown.

Characterization of Aero-Ga2O3 at Microwaves.
The S-parameters at the two ports of the VNA were measured to provide the reflection (S11/S22) and transmission (S12/S21) coefficients at/between the two ports, respectively. The measured Sparameters of the Aerogallox sample (Figure 2c) with the density of 110 mg/cm 3 are depicted in Figure 5.

Characterization of Aero-Ga 2 O 3 at Microwaves.
The S-parameters at the two ports of the VNA were measured to provide the reflection (S 11 /S 22 ) and transmission (S 12 /S 21 ) coefficients at/between the two ports, respectively. The measured S-parameters of the Aerogallox sample (Figure 2c) with the density of 110 mg/cm 3 are depicted in Figure 5.

Characterization of Aero-Ga2O3 at Microwaves.
The S-parameters at the two ports of the VNA were measured to provide the reflection (S11/S22) and transmission (S12/S21) coefficients at/between the two ports, respectively. The measured Sparameters of the Aerogallox sample (Figure 2c) with the density of 110 mg/cm 3 are depicted in Figure 5.  In Figure 5a,b we show only S 11 and S 21 , since S 22 is identical with S 11 , and S 12 with S 21 thanks to reciprocity and symmetry of the scattering matrix (typical for passive components). One can notice that the reflection coefficient is better than −10 dB all over the band of interest, whereas the insertion loss has a maximum value better than −0.12 dB in both cases without and with aero-Ga 2 O 3 . In other words, the presence of the aero-Ga 2 O 3 inside the cavity between the two X-band waveguides does not affect at all the transmission of the microwave signal, which means the aero-Ga 2 O 3 is completely (within our accuracy) electromagnetically transparent in the X-band.

Characterization of Aero-Ga 2 O 3 in the Terahertz Region
Two samples of Aerogallox with different densities were studied-the low density equals to 70 mg/cm 3 and the high density was 110 mg/cm 3 . Comparative analyses with aero-GaN samples with the density of 15 mg/cm 3 were performed. Broad-band terahertz-infrared transmissivity spectra for two samples with low and high densities are shown in Figure 6a. As expected, more dense samples have lower transmissivity. A decrease in the transmission coefficient for frequencies growing up to 200 cm −1 is caused by intensive absorption between 200 and 800 cm −1 (see inset in Figure 6a). Our attempts to extract information about absorption mechanisms in this range by measuring the spectra of reflection coefficient failed due to very low intensity reflected by the samples, i.e., extremely small value of reflectivity, as discussed below. Below 200 cm −1 , there are two narrow absorption resonances located at~155 and~176 cm −1 whose parameters (frequency position and intensity) do not depend noticeably on the sample density. A few more resonances are observed above 800 cm −1 , most intensive at~3500 cm −1 . Origin of observed narrow absorptions seems to be related to the presence of impurities such as hydrogen [21]. In Figure 6b,c we present the spectra of real and imaginary permittivity of the two samples, together with the reflection coefficient. It is seen that THz characteristics of our aero-Ga2O3 material drastically differ from those of parent bulk Ga2O3 crystal as well as from any other bulk semiconductor. For example, typical semiconductors Ge, GaAs and Si have refractive indexes at 300 GHz n = 3.99, 3.59 and 3.43, respectively [23], while the refractive index of aero-Ga2O3, n ≈ 1.07 (sample with density 70 mg/cm 3 , Figure 7), only slightly exceeds that of vacuum (n = 1). From Figure 6b,c one can see that there is pronounced difference between terahertz response of aero-Ga2O3 and previously studied aero-GaN [22]. While in aero-Ga2O3 real permittivity ε′ is dispersionless and imaginary permittivity ε" is small and approaches zero with frequency decrease, both indicating absence of any absorption process, aero-GaN demonstrates strong increase of ε′ and ε" toward low frequencies. The origin of corresponding pronounced absorption is associated with the polarizability of the 3D architecture of mutually interpenetrated GaN aerotetrapods, with the ZnO-GaN interfaces, and finally with dynamics of relatively big complexes of tetrapods. No such contribution exists in the present Aerogallox nanomaterial. Aero-Ga2O3 is characterized by very small imaginary permittivity In Figure 6b,c we present the spectra of real and imaginary permittivity of the two samples, together with the reflection coefficient. It is seen that THz characteristics of our aero-Ga 2 O 3 material drastically differ from those of parent bulk Ga 2 O 3 crystal as well as from any other bulk semiconductor. For example, typical semiconductors Ge, GaAs and Si have refractive indexes at 300 GHz n = 3.99, 3.59 and 3.43, respectively [23], while the refractive index of aero-Ga 2 O 3 , n ≈ 1.07 (sample with density 70 mg/cm 3 , Figure 7), only slightly exceeds that of vacuum (n = 1). From Figure 6b,c one can see that there is pronounced difference between terahertz response of aero-Ga 2 O 3 and previously studied aero-GaN [22]. While in aero-Ga 2 O 3 real permittivity ε is dispersionless and imaginary permittivity ε" is small and approaches zero with frequency decrease, both indicating absence of any absorption process, aero-GaN demonstrates strong increase of ε and ε" toward low frequencies.
The origin of corresponding pronounced absorption is associated with the polarizability of the 3D architecture of mutually interpenetrated GaN aerotetrapods, with the ZnO-GaN interfaces, and finally with dynamics of relatively big complexes of tetrapods. No such contribution exists in the present Aerogallox nanomaterial. Aero-Ga 2 O 3 is characterized by very small imaginary permittivity ε" and dielectric losses tanδ = ε"/ε (Figure 7a) that are responsible for radiation absorption, and by real part of permittivity ε and real part n of complex refractive index n* = n + ik (Figure 7b) close to 1. Both factors lead to a very low reflection coefficient (inset in Figure 6b) R = (n − 1) 2 + k 2 / (n + 1) 2 that is as small as R ≈ 0.1% and can be made close to R ≈ 0.01%.
We can compare these extremely low reflectivity values with those of typical bulk semiconductors, mentioned above, Ge (n = 3.99), GaAs (n = 3.59) or silicon (n = 3.43); corresponding reflection coefficients fall in the range 30-36%. Along with high value of transmissivity, this low reflectivity of the material may be beneficial for future applications of Ga 2 O 3 .
Nanomaterials 2020, 10, x FOR PEER REVIEW 9 of 11 1. Both factors lead to a very low reflection coefficient (inset in Figure 6b)

Conclusions
We developed a highly porous ultra-lightweight three-dimensional nanoarchitecture consisting of interconnected microtubes of Ga2O3 with nanometer thick walls, and demonstrated that the gallium oxide skeleton is of crystalline β-phase with almost stoichiometric composition. The new nanomaterial is shown to exhibit ultra-low reflectivity and high transparency in an extremely wide range of the electromagnetic spectrum, covering the X-band and THz region, up to 3 THz. The disclosed novel properties of aero-Ga2O3 open possibilities, in premiere, for the use of gallium oxide in IoT applications.

Conclusions
We developed a highly porous ultra-lightweight three-dimensional nanoarchitecture consisting of interconnected microtubes of Ga 2 O 3 with nanometer thick walls, and demonstrated that the gallium oxide skeleton is of crystalline β-phase with almost stoichiometric composition. The new nanomaterial is shown to exhibit ultra-low reflectivity and high transparency in an extremely wide range of the electromagnetic spectrum, covering the X-band and THz region, up to 3 THz. The disclosed novel properties of aero-Ga 2 O 3 open possibilities, in premiere, for the use of gallium oxide in IoT applications.

Conflicts of Interest:
The authors declare no conflict of interest.