AbstractThe versatile range of applications for two-dimensional (2D) materials has encouraged scientists to engineer their properties. This is often accomplished by stacking atomically thin layered materials into complex van der Waals heterostructures. A less popular but technologically promising approach is alloying 2D materials. In this work, we demonstrate a first step towards tuning the intrinsic electronic properties of hexagonal boron nitride (hBN). We present a series of aluminum alloyed hexagonal boron nitride (hBAlN) samples grown by metal organic vapor phase epitaxy on 2-inch sapphire substrates with varying aluminum concentration. Importantly, the obtained samples revealed a sp2-bonded crystal structure and modifications in interband optical transitions. Optical absorption experiments disclosed two prominent peaks in the excitonic spectral range with absorption coefficients ~ 106 cm− 1. Their peak energies align closely with the energies of indirect and direct bandgap transitions in hBN. The presence of two absorption peaks can be attributed to mixing of electronic states in the K and M conduction band valleys, resulting in a substantial increase in the absorption coefficient for indirect transitions. The observed effects offer insights into hBN-based two-dimensional alloys, highlighting the potential for developing 2D material-based quantum well structures capable of operating in the challenging deep UV spectral range.
IntroductionThe initial research on graphene has motivated researchers to explore the properties of other atomically thin layered materials (2D materials) and their stacks, sometimes referred to as NanoLego structures1,2,3,4. This is typically accomplished by stacking different layers together into van der Waals heterostructures5,6. Combining these materials using epitaxial growth provides the additional opportunity to scale-up or create new structures by growing materials as alloys with appropriate composition7,8. In the case of hexagonal boron nitride (hBN), the natural choice is alloying hBN with aluminum and gallium, since they are in the same group of the periodic table as boron. Research on the novel hB1 − x−yAlxGayN material (where x and y stand for Al and Ga concentration in the alloy respectively) is interesting in terms of exploration of new physical phenomena as well as from a technological point of view9. hBN with its, unusual for III-N, strong sp2 covalent bonds in plane and weak van der Waals bonds out of plane exhibits extraordinary properties10. One of the intriguing facts is that in spite of the indirect character of the bandgap, hBN shows a photoluminescence (PL) emission intensity of the phonon-assisted transitions as strong as for direct semiconductors11,12,13. Thus, the possibility of tuning the hBN bandgap using Al opens up further prospects for applications in hBN-based quantum wells and optoelectronic devices. The same scheme could be used for hBN alloying with Ga. Tunable hBAlN, hBGaN or hBAlGaN alloys could become promising candidates for building blocks of efficient light sources in the deep UV spectral range (DUV, 200–280 nm, 4.4–6.2 eV). Such candidates are of high importance, since nowadays the efficiency of semiconductor light sources operating in DUV is very limited (~ 10% for 270–280 nm and ~ 3% for 250 nm)14. This spectral range is crucial, for example for the disinfection and sterilization of air, water and surfaces15,16.However, before hBN-based quantum structures and DUV light sources can be realized, more fundamental questions have to be addressed. Firstly, it is still unclear how to tune the hBN bandgap energy. This knowledge is of great importance for designing hBN-based quantum wells that could trap carriers and enhance emission from the structure. Moreover, it would be very beneficial for further improvement if one could change the nature of the hBN bandgap from indirect to direct17,18,19.In this work, we present results for boron nitride layers grown by metal organic vapor phase epitaxy (MOVPE) that were alloyed with aluminum and preserve a sp2-bonded crystal structure. The produced hB1 − xAlxN layers with aluminum concentration ranging from x = 0.01% to x = 1.1% exhibit strong absorption for two energies that coincide with energies of indirect and direct bandgap in hBN. Our approach that uses MOVPE to grow this novel material provides a perspective for further development of the hBAlN alloy composition control in terms of doping for conductivity, manipulation of the bandgap, and commercial up-scaling.MethodsSamplesThe samples used in these studies were grown using an Aixtron CCS 3 × 2’’ metal organic vapor phase epitaxy system. The growth was carried out on 2-inch sapphire c-plane substrates with 0.3° off-cut. For the growth of layered, sp2-bonded hB1 − xAlxN samples ammonia, triethylboron (TEB) and trimethylaluminum (TMAl) were used as nitrogen, boron, and aluminum precursors and hydrogen was used as carrier gas. Note that we used the hBAlN notation (with letter h) to highlight the importance of the hexagonal hBN-like crystal structure of the obtained material. The precursors were injected alternatively following the flow modulation epitaxy growth mode20. The scheme of pulses in a single cycle was as follows: TEB-NH3-TMAl-NH3. The growth temperature was kept at 1300 °C. The temperature value was controlled with an in situ multipoint optical pyrometer (ARGUS). NH3 and TEB flows were fixed for all samples (V/III ratio for TEB-NH3 pulses was ~ 15 000). The TMAl/III ratio (III = TEB + TMAl) in the process varied for the samples: Al0.02 (ratio 0.02), Al0.04 (ratio 0.04), Al0.07 (ratio 0.07), Al0.13 (ratio 0.13). To obtain the lowest ratio, the duration of consecutive pulses of TMAl and NH3 was reduced by half. This step was enforced by the minimal precursor flow limitation in our system. The number of pulsing cycles was chosen in such a way that all samples were grown 60 min.Experimental detailsThe crystal structure of the samples was examined with a Panalytical X’Pert diffractometer with standard CuKα X-ray source. The X-ray beam was formed by an X-ray mirror for Bragg-Brentano optics. Infrared reflectance spectra were collected with a Thermo Scientific Nicolet Continuum Infrared Microscope equipped with a 32x Schwarzschild infinity corrected objective (NA 0.65). All samples of 7 × 10 mm were measured with a perpendicular incident beam on five 70 × 70 μm areas placed in the center and four corners of the sample. Spectra were collected in the range of 650–4000 cm− 1 with a resolution of 0.482 cm− 1. The surface morphology was examined by scanning electron microscopy (SEM) using a FEI Helios NanoLab 600 system. This device equipped with gallium focus ion beam was also utilized to prepare electron-transparent cross sections of the samples for transmission electron microscopy (TEM). The sample surfaces were protected with electron-deposited Pt-C composite from a metalorganic source and underwent plasma cleaning before measurements. High-resolution TEM images were acquired using the FEI Titan 80–300 transmission electron microscope operating at 300 kV. Composition mapping was obtained using the FEI Talos F200X microscope operating at 200 kV equipped with energy-dispersive X-ray spectroscopy detector (Bruker). Scanning transmission electron microscopy (STEM) compositional maps were captured within a scattering angle range of 80 to 200 mrad and a converged semi-angle of 9.5 mrad for the incident beam. The secondary ion mass spectrometry (SIMS) depth profiling experiment was conducted using the TOF-SIMS.5 system (ION-TOF GmbH, Germany), equipped with a time-of-flight mass analyzer (TOF-SIMS). Samples were analyzed without special pretreatment upon receipt, and they were transferred directly to the analytical chamber, where the pressure was maintained at 9 × 10− 10 mbar. Elemental distribution was acquired using the dual beam mode, with the samples being sputtered by oxygen ions (operating conditions: 250 eV, 23 nA) over a 250 μm × 250 μm area. Subsequently, the internal layers of the sample were analyzed using a Bi3+ ion beam (operating conditions: 64 μm × 64 μm raster size, 30 keV, 0.5 pA). Internal mass calibration was performed using the mass of ions that are consistently present: BxO+ and AlxO+ clusters. Absorption spectra were measured using a Varian Cary 5000 UV-vis-NIR spectrophotometer in dual-beam mode with nitrogen purging. The spectral bandwidth was set to 1 nm.Although our samples are very homogeneous on the wafer scale21,22, we decided to measure each of the samples taken from the exact same area on the wafer.ResultsMaterial compositionFig. 1Cross-sectional high-resolution transmission electron microscopy pictures for samples (a) Al0.02, (b) Al0.04, (c) Al0.07, (d) Al0.13. Carbon (C) and platinum (Pt) caps were deposited in lamella preparation process. Atomic fraction profiles derived as integrated signal from energy-dispersive X-ray spectroscopy (EDX) maps for samples (e) Al0.02, (f) Al0.04, (g) Al0.07 (h) Al0.13. The distance on horizontal axis refers to map coordinate across the sample. Straight lines visualize trend for Al2O3 background signal. Secondary ion mass spectrometry (SIMS) results for aluminum ion for samples (i) Al0.02, (j) Al0.04, (k) Al0.07, l) Al0.13. The distance on horizontal axis refers to the distance from sample interface. Orange rectangle in (e)-(l) visualizes signal corresponding to BAlN layer. Al values denote to aluminum composition extracted from the measurement and δ is its uncertainty.Full size imageHigh-resolution transmission electron microscopy (HRTEM) images for each sample are presented in Fig. 1a-d. The BAlN layers are observed to consist of atomic planes aligned parallel to the sapphire substrate characteristic for epitaxial growth of layered, two-dimensional materials22. At the Al2O3-BAlN interface, there is an unintentional, transitional layer of approximately 2 nm thickness composed of amorphous AlON. This layer forms during the initial stages of growth, prior to the hBAlN layer covering and protecting the substrate from ammonia exposure as described in the literature21,23. Additionally, above the BAlN layer, there are carbon and platinum cap layers deposited during the process of sample preparation for TEM measurements. Material composition was assessed using energy-dispersive X-ray spectroscopy (EDX) mapping in cross-sectional scanning transmission electron microscopy (STEM) measurements (see Supplementary Materials). As depicted in Fig. 1e–h, an excess of aluminum signal is observed in the EDX map area corresponding to BAlN, superimposed on a monotonous background originating from the Al2O3 substrate signal. Upon background signal subtraction, aluminum composition values were obtained, correlating with the amount of TMAl in the gas phase during the growth process. Based on the EDX results, the estimated compositions for the grown samples are as follows: below 0.03% for Al0.02, 0.12% for Al0.04, 0.18% for Al0.07, and 0.55% for Al0.13. Assuming aluminum substitution for boron in the crystal lattice, we infer a nitrogen concentration of 50% at. Consequently, based on EDX results, the composition of hB1 − xAlxN ranges from x < 0.06% to x = 1.1%. Comparing the ratio of metalorganic precursors to the resulting material composition, we deduce an aluminum incorporation efficiency of only 3–8%.Alternatively, we conducted secondary ion mass spectrometry (SIMS) measurements to independently determine the aluminum composition of the as-grown samples. To do this, we assumed that the value derived from the EDX result for sample Al0.07 applies similarly to the SIMS measurement. This assumption enabled us to calculate the efficiency of Al+ ion spreading during sample milling. We selected sample Al0.07 because its EDX signal is robust and clearly distinguishable from the background, while the SIMS profile is smoother compared to sample Al0.13, which shows a stronger EDX signal. Subsequently, the obtained signal for Al+ ions for the remaining samples was recalculated to atomic fraction. The results are presented in Fig. 1i–l. The plateau observed at 0.35% atomic composition arises from the saturation of the detector, resulting in its nonlinear response. Consequently, we are unable to accurately monitor aluminum composition above 0.35% atomic. It is evident that for samples Al0.02 and Al0.13, the profiles exhibit raggedness. This is attributed to sample charging during measurement. The aluminum composition values obtained for the samples differ from those obtained through EDX analysis because we operate in very small concentrations regime. Nonetheless, the trends remain consistent. According to the SIMS results, we estimate aluminum composition as follows: 0.005% for Al0.02, 0.026% for Al0.04, 0.18% for Al0.07, and 0.21% for Al0.13. Applying the same assumption as in the case of EDX results to the SIMS data, we find that the material composition ranges from x = 0.01% to x = 0.42%, with an estimated efficiency of aluminum incorporation of approximately 0.5-5%.Both experimental methods exhibit varying detection sensitivity within this range of atomic composition. Additionally, the obtained results are interrelated, as the composition derived for one sample was utilized to calibrate the results for others. Consequently, pinpointing the exact composition for our samples is challenging. Nevertheless, it is evident that we are operating within a compositional range with aluminum concentrations not exceeding 1%.X-ray diffractionIn Fig. 2a we present X-ray diffraction 2θ/ω scans collected for all the studied samples. The peaks at 20.5°, 40° (Cu Kβ line) and 41.75° (Cu Kα line) correspond to the 0003 and 0006 planes of Al2O3. In Fig. 2b, c we present a zoom on the peaks at 26° and 36° that come from the reflection of the 0002 planes in sp2-BAlN and sp3-AlN, respectively. The parameters of the Gaussian curve fit performed on the data concerning hBAlN are included in Table 1. Note that the 0002 BAlN peak related to sample Al0.02 is asymmetric, so we fitted two Gaussian curves yielding two different components of the peak. The components originate from turbostratic (tBAlN, lower angles) and hexagonal (hBAlN, higher angles) phases of sp2-bonded BAlN21,23. In the further analysis we will focus only on the hexagonal phase. In Table 1 we do not observe significant changes in the peak position 2θB and the full width at half maximum (FWHM) for the hBAlN peak. This implies that the c lattice constant is comparable for all the samples (~ 3.41 Å) as well as the thickness of the crystal. The only monotonous trend can be found in the amplitude of the hBAlN peak. This means that the misorientation of the 0002 crystal planes is enlarged when increasing the TMAl flow.Fig. 2X-ray diffraction 2θ\ω scans of the studied samples: (a) scan in a wide 2θ range (semi-log scale), (b) zoom on the peak related to hBAlN and (c) zoom on the peak related to AlN.Full size imageTable 1 Parameters of the Gaussian curve fit to 0002 BAlN XRD peak. For sample Al0.02 two Gaussian curves were fitted, indicating the presence of both tBAlN and hBAlN.Full size tableThe signal associated to AlN consists of a very broad peak, which is alike for all the samples and a narrower component that changes its intensity. In fact, these two components have different origins. The broad component comes from an unintentional thin (~ 2 nm) AlON layer that is created on the hBAlN-sapphire interface during the initial growth stage. This layer is also observed in HRTEM results in Fig. 1a-d. This kind of layer is characteristic for MOVPE-grown hBN23,24. However, in accordance with the enhancement of the sharp component with an increase in TMAl, we can deduce the formation of AlN crystals in our samples. These crystals do not seem to influence the crystal structure of hBAlN as we do not observe any correlation between the intensity of the sharp peak at 36° and the values of 2θB and FWHM in Table 1 which are related to BAlN.Fourier-transform infrared spectroscopyIn Fig. 3 we present exemplary infrared reflectance spectra obtained for the studied samples. The peak observed at approximately 1368 cm− 1 corresponds to the E1u vibrational mode characteristic of hBN25. The observation of this peak provides direct evidence of the sp2 bonding in our samples. FTIR results for the E1u mode complement the analysis of the E2g Raman mode and offer similar insights into variations of stress within the hBAlN layers. However, the advantage of the FTIR technique lies in the absence of background photoluminescence signals, which can be disruptive in Raman spectroscopy. The high reflectance below 1000 cm− 1 comes from the sapphire substrate. Barely visible small and sharp peaks about 1550 cm− 1, 2300 cm− 1 and 3500 cm− 1 correspond to transitions in the atmospheric gases present during the experiment.Fig. 3Fourier-transform infrared reflectance spectrum for the Al0.02 sample with the lowest TMAl flow (black dots). The red line is the fitted curve. The inset shows a zoom on the peak of the E1u vibrational mode, which is strong evidence for a sp2 crystal structure for all hBAlN samples. The black dots represent the experimental data and the lines are the fitted curves: red for Al0.02, blue for Al0.04, green for Al0.07, purple for Al0.13.Full size imageTo make the FTIR results quantitative, we implemented a spectra analysis method, as described in the Ref26. In this method, it is assumed that the material is composed of harmonic oscillators with self-energy ω and damping parameter γ. It makes use of the Dynamic Dielectric Function (DDF) of the materials in the structure (boron nitride and sapphire in this case). The substantial advantage of this method is the possibility of characterizing the grown layer itself. The self-energy of the oscillator ωBAlN (phonon energy, peak position), its damping parameter γBAlN (peak broadening) and layer thickness provide information about strain, homogeneity and growth rate. In Table 2 we present the best fit parameters obtained as an average of 5 points measured for each sample.Table 2 Averages of the best fit parameters of the FTIR spectra obtained for the studied samples.Full size tableThe fitted parameters are monotonous for the samples Al0.04, Al0.07 and Al0.13. The increase in TMAl flow leads to an increase in phonon energy and peak broadening, which suggest the introduction of an inhomogeneous compressive strain within the layer. The changes of ωBAlN (~ 0.63 cm− 1 for the following samples) and γBAlN (~ 4 cm− 1 for the following samples) are significant in terms of FTIR measurements. Remarkably different is the trend for the layer thickness. The 4-time increase in amount of TMAl (samples Al0.04 and Al0.13) resulted in a layer that is only 1 nm thicker, which is equivalent to the increase of the growth rate by less than 0.02 nm/min.The sample Al0.02 clearly stands out and does not follow the same trends. The reason for this behavior is most likely due to the change of TMAl-NH3 pulse duration necessary to obtain a TMAl/III ratio equal to 0.02. This led to a two-fold increase in growth rate compared to the sample Al0.04 as can be seen in doubled value of the layer thickness. This result is reasonable, since the TMAl/III ratio for Al0.02 was twice lower than for Al0.04 yielding to increase of material synthesis rate. At the same time the number of TEB pulses increased by 33% aiming for uniform growth duration across all samples. So we find the competition between TEB-NH3 and TMAl-NH3 pulses to be more important for the growth rate rather than the change in number of TEB pulses in the growth process. This finding also opens up avenues for studying the effectiveness of aluminum incorporation depending on the timing of precursor injection.Morphology analysisThe morphology of the studied samples is presented in Fig. 4. For each sample, a characteristic wrinkle pattern is observed. The wrinkles are created during cooling down the material after the growth process and are due to the relaxation of strain caused by the difference in thermal expansion coefficients of the layer and the substrate26,27. Because of the mechanism of the wrinkle formation their presence is an evidence for the continuity of the hBAlN epitaxial layer. The wrinkles are most pronounced for sample Al0.02, which is due to a much larger thickness. The darker circular areas in the images are bubbles with hydrogen inside. They are created during SEM imaging and are the result of the radiolysis of interfacial water via electron irradiation28. The size of crystalline objects on the surface of the layer is correlated with the amount of TMAl present in the process. This observation is in good agreement with the XRD results for which the AlN narrow diffraction peak was observed for Al0.04, Al0.07 and Al0.13.Fig. 4Scanning electron microscopy images of the samples (a) Al0.02, (b) Al0.04, (c) Al0.07, (d) Al0.13. The scale bar corresponds to 1 μm for each image.Full size imageAdditionally, supported by atomic force microscopy (AFM) and nanomechanical mapping, we can rule out the possibility that the crystallites are other forms of boron nitride. Furthermore, Raman spectroscopy revealed a signal originating from the E2high characteristic mode in wurtzite aluminum nitride, with its intensity correlating with the number of crystalline objects observed in SEM. Additionally, wurtzite precipitations were observed in HRTEM images. All of these findings, detailed in the Supplementary Materials, indicate that the crystalline objects observed in SEM are sp3-bonded aluminum nitride. This poses challenges in aluminum incorporation into hBN, as BN and AlN, being stable in nature, exhibit significantly different allotropic phases. However, in EDX analysis, we selected areas without sp3-bonded AlN inclusions, suggesting that aluminum may also incorporate into the sp2-bonded crystal structure.UV-vis spectroscopyTo determine the absorption coefficient from the measured absorbance, we subtracted the absorbance spectrum obtained for bare sapphire. The thickness of the layer, dBAlN, extracted from Table 2, was used to calculate the absorption coefficient. The fitted values of dBAlN align well with the layer thickness observed in HRTEM (Fig. 1). Additionally, in FTIR measurements, we averaged the thickness over a much wider area in comparison to HRTEM results. Therefore, we consider these values as the best approximation of the average thickness necessary for interpreting the absorption spectra. The absorption coefficient spectra of the samples studied are presented in Fig. 5. The spectrum for the reference epitaxial boron nitride without aluminum sample was obtained in the same way. The reference sample is a thin BN layer grown at 1300 °C in continuous flow growth (CFG) regime analogously to the sample SI presented in the work in Ref21. The absorption coefficient for all the samples is of order of α = 2 × 106 cm− 1. This value is twice higher as compared to those previously reported for boron nitride10,29,30 which most likely originates from high crystalline quality of the material. In contrast to the reference sample, the samples with aluminum exhibit two well-resolved peaks. In previous studies only peak shifting was observed, which was accompanied by a broadening due to a decrease of sample quality31. In the case of our samples the two peak energies are close to the dash-dotted and dashed lines that are positioned at energies corresponding to the indirect (5.955 eV32) and direct (6.125 eV33,34) band-edge transitions in boron nitride, respectively. Depending on the TMAl flow, one can change the intensity of those peaks and the intensity ratio between higher and lower energetic peaks. However, this relationship does not correlate with the amount of TMAl in the growth process or aluminum composition in the BAlN layer.Fig. 5Absorption coefficient spectra of the samples studied. The solid gray line is the spectrum collected for the reference boron nitride sample without aluminum. It was multiplied by factor 0.45 for clarity. The reference sample was grown in analogy to the SI from Ref21. The uncertainty of the results is of order of 0.5 × 105 cm− 1. Black dashed and dash-dotted lines illustrate the energies of direct and indirect bandgaps in boron nitride.Full size imageThe lower energy peak, which is not observed for the reference sample, could be thought to be the result of the crystalline objects presented in the SEM images (Fig. 4). However, the peak intensity does not scale with the number of objects on the layer’s surface. Secondly, as has been mentioned before, the crystallites are identified as sp3-bonded AlN which is known to have a bandgap energy of 6.2 eV35. Possible alloying of sp3-bonded wurtzite-AlN (wAlN) with boron would, on the other hand, further increase the bandgap since wurtzite-BN (wBN) is a wider bandgap semiconductor18. As a result, the signal from AlN should be observed at higher, rather than lower energies when compared to the peak of pure boron nitride. To ensure that the observed effect is not induced by crystalline objects, we removed them by etching with NaOH. The NaOH solution dissolves AlN without damaging the BAlN layer. No changes were detected in the spectrum within the wAlN bandgap energy region (6.2 eV), which can be attributed to the significantly lower absorption coefficient of this material. After the removal of the crystallites, the lower-energy peak remains present, corroborating that the observed effect originates from the layer itself (see Figure S9).DiscussionAs presented in Fig. 2b; Table 1, the X-ray diffraction peak is observed at angles lower than expected for the 0002 hBN plane (26.764°36). This indicates a larger lattice constant in the c direction. However, the peak position does not change significantly between the samples, which is in contrast to the results reported in Ref31. in which authors observed peak position shifting towards higher angles with an increasing TMAl flow. According to theoretical DFT calculations presented in Ref37. , a decrease in the c lattice constant should be observed when hBN is alloyed with Al, since hexagonal aluminum nitride has calculated a smaller c lattice constant. We conclude that since we do not have a perfect hexagonal phase the main reason for the XRD peak position change in our samples is that it is attributed to random twists of subsequent atomic layers, which make the material more turbostratic. In this case we postulate that the peak position variations related to the alloying with small amounts of Al are just a higher order correction. This hypothesis seems to be confirmed by the fact that all samples were grown under the same growth conditions. Another prominent feature presented in Fig. 2c is the peak at ~ 36°, whose intensity increases with the increase in TMAl flow. The value of the peak position is in agreement with the XRD peak of 0002 AlN in the crystal structure of wurtzite38. This indicates the creation of sp3-bonded wAlN crystals, which scale with the amount of TMAl. This additional notable peak is a proof of the phase separation of sp2-bonded hBAlN and sp3-bonded wAlN. This conclusion is further supported by SEM results (Fig. 4), observation of E2high peak related to wAlN in Raman spectroscopy (Fig. S6) and wurtzite structure precipitations observed in HRTEM (Fig. S8). The number and size of crystalline objects correlate with the amount of TMAl, which provides strong evidence for the very limited solubility of aluminum in the hexagonal boron nitride layer at the temperature and pressure used in the growth process (1300 °C, 400 mbar). Nevertheless, using cross-sectional STEM with EDX and SIMS we were able to determine aluminum concentration within hB1 − xAlxN layers (x < 1%). By comparison of TMAl in a gas phase during growth process and final material composition we estimate aluminum incorporation rate to be approximately 3%.Although AlN crystallites can be observed, the sp2-bonded, layered structure of hBAlN is maintained, which is proved by HRTEM and FTIR measurement presented in Figs. 1a–d and 3. However, by introducing TMAl we modify the optical properties of the layer. This can be seen as a shift of the phonon energy ωBAlN towards higher energies that suggests compressive stress in the structure as demonstrated in Ref39. Furthermore, the increase in TMAl is followed by a broadening of the peak described by the parameter γBAlN, which indicates a defect-related inhomogeneity of the strain within the layer.The most notable finding of this study is the identification of an additional low-energy peak in the absorption spectra. While previous studies have reported features related to indirect bandgap transitions, primarily evidenced by phonon replicas in reflectivity and photoluminescence measurements, our work provides results of a peak in absorption that corresponds well with the expected energy of the indirect excitonic transition. The hBN conduction band is known to consist of minima at the K and M points of the Brillouin zone that are energetically close to each other11. They are responsible for direct and indirect band transitions, respectively. As an indirect transition is a three-particle (photon, electron, phonon) event, it has a much lower probability to occur than a direct transition, which requires only two particles (photon, electron). Consequently, the absorption coefficient related to indirect transitions is usually 2–3 orders of magnitude lower as compared to direct ones. However, in the case of our samples, we observe two peaks at energies which coincide with the energies of direct and indirect transitions in hBN (dash-dotted and dashed lines in Fig. 5). Furthermore, both have a very high value of the absorption coefficient (α ~ 106 cm− 1) typical for direct bandgap transitions. Notably that the lower-energy peak has even lower energy than expected for the indirect transition in hBN. This observation is in agreement with predictions about a decrease in bandgap energy when hBN is alloyed with Al17. To understand the mechanism that stands behind the observation of the two peaks in the absorption spectra, we need to be aware of the role of aluminum in the crystal structure. Al incorporation introduces short range disorder, which in turn would lead to the mixing of states with large k-vectors. As presented in Ref11. , conduction and valence bands along KH and ML points in hBN are very flat and almost parallel to each other. This feature leads to high oscillator strength and consequently very efficient absorption. A disorder induced by Al incorporation further modifies the band structures and the probability of electronic transitions, i.e. oscillator strength. Consequently, other absorption channels are enabled through the defect-related states. The modification of states caused by Al-related defects allows us to observe highly effective absorption for both energies. However, further increase of TMAl flow and limited solubility of Al in hBN lead to a deterioration of optical quality of the material which is observed in broadening of spectroscopic peaks in the spectra (same for FTIR and UV-Vis). In this case, two opposing effects influence the results: (i) an increase in oscillator strength, which enhances the efficiency of the indirect transition, and (ii) a degradation in optical quality, causing broader and less distinct absorption peaks. As a consequence, the relationship between absorption spectra and aluminum concentration is not straightforward. The optimal balance between these effects was achieved for the sample Al0.04.ConclusionsIn this work, we have shown the results for MOVPE grown hB1 − xAlxN layers with TMAl/III ratio ranging from 0.02 to 0.13. High-resolution transmission electron microscopy, X-ray diffraction and Fourier-transform infrared spectroscopy measurements confirmed that the obtained material maintained a sp2-bonded layered crystal structure, characteristic for hBN. Composition analysis through energy-dispersive X-ray spectroscopy (EDX) and secondary ion mass spectrometry (SIMS) revealed that our hB1 − xAlxN samples have aluminum compositions ranging from x = 0.01% to x = 1.1%, indicating an effectiveness of aluminum incorporation of approximately 3%. The variations in aluminum composition significantly impact the properties of the layers. Most importantly, we have shown two peaks of strong absorption (α ~ 106 cm− 1) typical for direct excitonic transitions in hBN. The peak energies coincide with the energy of the indirect and direct bandgap transition in hBN, indicating the activation of absorption channels through defects introduced by Al incorporation. These findings shed light on the manipulation of the boron nitride bandgap, affecting its indirect/direct nature and width. Such understanding is crucial for fabricating hBN-based structures capable of efficient deep UV emission.
Data availability
Data for this article, including the experimental data presented in the manuscript are available at the Repository for Open Data RepOD at https://doi.org/10.18150/J8UYQK.
ReferencesNovoselov, K. S. et al. Electric field in atomically thin carbon films. Science 306, 666–669 (2004). https://doi.org/10.1126/SCIENCE.1102896/SUPPL_FILE/NOVOSELOV.SOM.PDFNovoselov, K. S., Mishchenko, A., Carvalho, A. & Castro Neto, A. H. 2D materials and Van der Waals heterostructures. Science 353 (2016). https://doi.org/10.1126/SCIENCE.AAC9439.Gupta, A., Sakthivel, T. & Seal, S. Recent development in 2D materials beyond graphene. Prog Mater. Sci. 73, 44–126. https://doi.org/10.1016/J.PMATSCI.2015.02.002 (2015).Article
CAS
MATH
Google Scholar
Gibertini, M., Koperski, M., Morpurgo, A. F. & Novoselov, K. S. Magnetic 2D materials and heterostructures. Nat. Nanotechnol. 14, 408–419. https://doi.org/10.1038/s41565-019-0438-6 (2019).Article
ADS
CAS
PubMed
Google Scholar
Guo, H. W. et al. Stacking of 2D materials. Adv. Funct. Mater. 31, 2007810. https://doi.org/10.1002/ADFM.202007810 (2021).Article
ADS
CAS
Google Scholar
Liang, S. J. et al. Van der Waals heterostructures for High-Performance device applications: challenges and opportunities. Adv. Mater. 32, 1903800. https://doi.org/10.1002/ADMA.201903800 (2020).Article
CAS
Google Scholar
Yao, J. & Yang, G. Material alloys: synthesis and application in electronic and optoelectronic devices. Adv. Sci. 9, 2103036. https://doi.org/10.1002/ADVS.202103036 (2022).Article
MATH
Google Scholar
Singh, A. K. et al. Review of strategies toward the development of alloy two-dimensional (2D) transition metal dichalcogenides. IScience 24, 103532. https://doi.org/10.1016/j.isci.2021.103532 (2021).Article
ADS
CAS
PubMed
PubMed Central
MATH
Google Scholar
Vuong, P. et al. Control of the mechanical adhesion of III-V materials grown on layered h-BN. ACS Appl. Mater. Interfaces. 12, 55460–55466. https://doi.org/10.1021/acsami.0c16850 (2020).Article
CAS
PubMed
MATH
Google Scholar
Zunger, A., Katzir, A. & Halperin, A. Optical properties of hexagonal Boron nitride. Phys. Rev. B. 13, 5560. https://doi.org/10.1103/PhysRevB.13.5560 (1976).Article
ADS
CAS
MATH
Google Scholar
Elias, C. et al. Flat bands and giant Light-Matter interaction in hexagonal Boron nitride. Phys. Rev. Lett. 127, 1–11. https://doi.org/10.1103/PhysRevLett.127.137401 (2021).Article
MATH
Google Scholar
Watanabe, K., Taniguchi, T. & Kanda, H. Direct-bandgap properties and evidence for ultraviolet lasing of hexagonal boron nitride single crystal. Nat. Mater. 6, 404–409 (2004). https://doi.org/10.1038/nmat1134Schué, L. et al. Bright luminescence from indirect and strongly bound excitons in h-BN. Phys. Rev. Lett. 122, 067401. https://doi.org/10.1103/PHYSREVLETT.122.067401/FIGURES/4/MEDIUM (2019).Article
ADS
PubMed
MATH
Google Scholar
Amano, H. et al. The 2020 UV emitter roadmap. J. Phys. D Appl. Phys. 53, 503001. https://doi.org/10.1088/1361-6463/ABA64C (2020).Article
CAS
MATH
Google Scholar
Ramos, C. C. R. et al. Use of ultraviolet-C in environmental sterilization in hospitals: A systematic review on efficacy and safety. Int. J. Health Sci. (Qassim). 14, 52. https://doi.org/10.5281/ZENODO.3933425 (2020).Article
PubMed
MATH
Google Scholar
Chen, J., Loeb, S. & Kim, J. H. LED revolution: fundamentals and prospects for UV disinfection applications. Environ. Sci. (Camb). 3, 188–202. https://doi.org/10.1039/C6EW00241B (2017).Article
CAS
MATH
Google Scholar
Zhang, Q. et al. Phase transition and bandgap engineering in B1-xAlxN alloys: DFT calculations and experiments. Appl. Surf. Sci. 575, 151641. https://doi.org/10.1016/j.apsusc.2021.151641 (2022).Article
CAS
Google Scholar
Zhang, M. & Li, X. Structural and electronic properties of wurtzite BxAl1–xN from first-principles calculations. Phys. Status Solidi B Basic. Res. 254 https://doi.org/10.1002/pssb.201600749 (2017).Turiansky, M. E., Shen, J. X., Wickramaratne, D. & Van De Walle, C. G. First-principles study of bandgap bowing in BGaN alloys. J. Appl. Phys. 126 https://doi.org/10.1063/1.5111414 (2019).Kobayashi, Y., Akasaka, T. & Makimoto, T. Hexagonal Boron nitride grown by MOVPE. J. Cryst. Growth. 310, 5048–5052. https://doi.org/10.1016/J.JCRYSGRO.2008.07.057 (2008).Article
ADS
CAS
Google Scholar
Dąbrowska, A. K. et al. Two stage epitaxial growth of wafer-size multilayer h-BN by metal-organic vapor phase epitaxy - A homoepitaxial approach. 2d Mater. 8. https://doi.org/10.1088/2053-1583/abbd1f (2020).Ludwiczak, K. et al. Heteroepitaxial growth of high optical quality, Wafer-Scale Van der Waals heterostrucutres. ACS Appl. Mater. Interfaces. 13, 47904–47911. https://doi.org/10.1021/acsami.1c11867 (2021).Article
CAS
PubMed
PubMed Central
MATH
Google Scholar
Tokarczyk, M. et al. Effective substrate for the growth of multilayer h-BN on sapphire—substrate off-cut, pre-growth, and post-growth conditions in metal-organic vapor phase epitaxy. 2d Mater. 10. https://doi.org/10.1088/2053-1583/acb44a (2023).Iwański, J. et al. Delamination of large area layers of hexagonal Boron nitride grown by MOVPE. Acta Phys. Pol. A. 139, 457–461. https://doi.org/10.12693/APhysPolA.139.457 (2021).Article
ADS
MATH
Google Scholar
Geick, R., Perry, C. H. & Rupprecht, G. Hexagonal Boron. Phys. Rev. 146 543–547. (1966).Iwański, J. et al. Temperature induced giant shift of phonon energy in epitaxial Boron nitride layers. Nanotechnology 34. https://doi.org/10.1088/1361-6528/ac9629 (2023).Tatarczak, P. et al. Strain modulation of epitaxial h-BN on Sapphire: the role of wrinkle formation for large-area two-dimensional materials. Nanotechnology https://doi.org/10.1088/1361-6528/ad18e6 (2023).Article
MATH
Google Scholar
Binder, J. et al. Epitaxial hexagonal Boron nitride for hydrogen generation by radiolysis of interfacial water. Nano Lett. 23, 1267–1272. https://doi.org/10.1021/ACS.NANOLETT.2C04434/SUPPL_FILE/NL2C04434_SI_004.AVI (2023).Article
ADS
CAS
PubMed
PubMed Central
MATH
Google Scholar
Li, X. et al. Large-area two-dimensional layered hexagonal Boron nitride grown on Sapphire by metalorganic vapor phase epitaxy. Cryst. Growth Des. 16, 3409–3415. https://doi.org/10.1021/acs.cgd.6b00398 (2016).Article
CAS
MATH
Google Scholar
Li, J. et al. Dielectric strength, optical absorption, and deep ultraviolet detectors of hexagonal Boron nitride epilayers. Appl. Phys. Lett. 101, 45. https://doi.org/10.1063/1.4764533/24853 (2012).Article
MATH
Google Scholar
Vuong, P. et al. Single crystalline Boron rich B(Al)N alloys grown by MOVPE. Appl. Phys. Lett. 116, 1–6. https://doi.org/10.1063/1.5135505 (2020).Article
CAS
MATH
Google Scholar
Cassabois, G., Valvin, P. & Gil, B. Hexagonal Boron nitride is an indirect bandgap semiconductor. Nat. Photonics. 10, 262–266. https://doi.org/10.1038/nphoton.2015.277 (2016).Article
ADS
CAS
Google Scholar
Watanabe, K. & Taniguchi, T. Jahn-Teller effect on exciton States in hexagonal Boron nitride single crystal. Phys. Rev. B Condens. Matter Mater. Phys. 79, 193104. https://doi.org/10.1103/PHYSREVB.79.193104/FIGURES/6/MEDIUM (2009).Article
ADS
Google Scholar
Rousseau, A. et al. Determination of the optical bandgap of the Bernal and rhombohedral Boron nitride polymorphs. Phys. Rev. Mater. 5, 064602. https://doi.org/10.1103/PHYSREVMATERIALS.5.064602/FIGURES/5/MEDIUM (2021).Article
CAS
MATH
Google Scholar
Perry, P. B. & Rutz, R. F. The optical absorption edge of single-crystal AlN prepared by a close‐spaced vapor process. Appl. Phys. Lett. 33, 319–321. https://doi.org/10.1063/1.90354 (1978).Article
ADS
CAS
MATH
Google Scholar
JCPDS. Joint Committee on Powder Diffraction Standards; h-BN pattern 34-0421Ghorbanzadeh Ahangari, M., Fereidoon, A., Hamed, A. & Mashhadzadeh Interlayer interaction and mechanical properties in multi-layer graphene, Boron-Nitride, Aluminum-Nitride and Gallium-Nitride graphene-like structure: A quantum-mechanical DFT study. Superlattices Microstruct. 112, 30–45. https://doi.org/10.1016/J.SPMI.2017.09.005 (2017).Article
ADS
CAS
Google Scholar
JCPDS. Joint Committee on Powder Diffraction Standards; AlN pattern 25-1133.Androulidakis, C. et al. Strained hexagonal Boron nitride: phonon shift and Grüneisen parameter. Phys. Rev. B. 97, 241414. https://doi.org/10.1103/PHYSREVB.97.241414/FIGURES/3/MEDIUM (2018).Article
ADS
CAS
MATH
Google Scholar
Download referencesAcknowledgementsThis work was supported by the National Science Centre, Poland, under the decisions 2019/33/B/ST5/02766, 2020/39/D/ST7/02811 and 2022/45/N/ST5/03396.Author informationAuthors and AffiliationsFaculty of Physics, University of Warsaw, Pasteura 5, 02-093, Warsaw, PolandJakub Iwański, Mateusz Tokarczyk, Aleksandra K. Dąbrowska, Jan Pawłowski, Piotr Tatarczak, Marta Bilska, Johannes Binder & Andrzej WysmołekLaboratory of Electrochemistry, Faculty of Chemistry, University of Warsaw, Pasteura 1, 02-093, Warsaw, PolandMarcin StrawskiBiological and Chemical Research Centre, Faculty of Chemistry, University of Warsaw, Żwirki i Wigury 101, 02-089, Warsaw, PolandKamil SobczakInstitute of Physics, Polish Academy of Sciences, Aleja Lotników 32/46, 02-668, Warsaw, PolandMarta Bilska, Maciej Wójcik & Sławomir KretAuthorsJakub IwańskiView author publicationsYou can also search for this author inPubMed Google ScholarMateusz TokarczykView author publicationsYou can also search for this author inPubMed Google ScholarAleksandra K. DąbrowskaView author publicationsYou can also search for this author inPubMed Google ScholarJan PawłowskiView author publicationsYou can also search for this author inPubMed Google ScholarPiotr TatarczakView author publicationsYou can also search for this author inPubMed Google ScholarMarcin StrawskiView author publicationsYou can also search for this author inPubMed Google ScholarKamil SobczakView author publicationsYou can also search for this author inPubMed Google ScholarMarta BilskaView author publicationsYou can also search for this author inPubMed Google ScholarMaciej WójcikView author publicationsYou can also search for this author inPubMed Google ScholarSławomir KretView author publicationsYou can also search for this author inPubMed Google ScholarJohannes BinderView author publicationsYou can also search for this author inPubMed Google ScholarAndrzej WysmołekView author publicationsYou can also search for this author inPubMed Google ScholarContributionsStudy conceptualization and design: J.I., M.T., A.K.D., J.B., A.W.; data collection: J.I., M.T., A.K.D., J.P., P.T., M.S., K.S., M.B., M.W., S.K.; analysis and interpretation of results: J.I., M.T., J.P., M.S., K.S., S.K., J.B., A.W.; visualization: J.I, J.P.; funding acquisition: J.I, J.B., A.W.; writing – original draft: J.I; writing – review and editing: J.I., M.T., A.K.D., M.S., K.S., S.K., J.B., A.W.Corresponding authorCorrespondence to
Jakub Iwański.Ethics declarations
Competing interests
The authors declare no competing interests.
Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Electronic supplementary materialBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articleIwański, J., Tokarczyk, M., Dąbrowska, A.K. et al. Alloying hBN with aluminum influences absorption and electronic properties.
Sci Rep 15, 9904 (2025). https://doi.org/10.1038/s41598-025-92671-9Download citationReceived: 29 November 2024Accepted: 03 March 2025Published: 22 March 2025DOI: https://doi.org/10.1038/s41598-025-92671-9Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsBAlN alloyhBNEpitaxyMOVPEBandgap manipulationExcitonic absorption