# Fileset

[Bisignano-ACSAMI25_nano-FTIR.pdf](https://mdr.nims.go.jp/filesets/1d241a8d-2eba-4201-8b2b-583428194b39/download)

## Creator

[Ilario Bisignano](https://orcid.org/0009-0004-9964-2106), [Masataka Imura](https://orcid.org/0000-0002-4236-9549), [Nicholaus Kevin Tanjaya](https://orcid.org/0000-0003-4126-8540), Ming-Jyun Ye, Noriyuki Okada, [Satoshi Ishii](https://orcid.org/0000-0003-0731-8428)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Infrared Near-Field Spectroscopy of AlGaN/GaN Heterostructures for Probing Two-Dimensional Electron Gas](https://mdr.nims.go.jp/datasets/6991b1ee-5669-4cb5-96fb-f50f8333cd99)

## Fulltext

Infrared Near-Field Spectroscopy of AlGaN/GaN Heterostructures for Probing Two-Dimensional Electron GasInfrared Near-Field Spectroscopy of AlGaN/GaN Heterostructures forProbing Two-Dimensional Electron GasIlario Bisignano, Masataka Imura, Nicholaus Kevin Tanjaya, Ming-Jyun Ye, Noriyuki Okada,and Satoshi Ishii*Cite This: ACS Appl. Mater. Interfaces 2025, 17, 50077−50084 Read OnlineACCESS Metrics & More Article Recommendations *sı Supporting InformationABSTRACT: Infrared near-field spectroscopy, or nano-FTIR, offers nanoscale resolutionin three dimensions to probe the chemical and physical properties of samples, making it aunique characterization tool. This nanoscopic resolution in three dimensions is particularlysuitable to probe a two-dimensional electron gas (2DEG) where a 2DEG has an effectivethickness of a few nanometers and exists a few tens of nanometers below the capping layer.This work employs nano-FTIR spectroscopy to noninvasively probe the 2DEG of AlGaN/GaN heterostructures, which are crucial for high-power electronic devices and sensingapplications. Higher harmonic amplitude and phase of the nano-FTIR spectra are sensitiveenough to the carrier concentration of the 2DEGs, which is supported by analyticalcalculations based on the finite dipole model. A comparative analysis confirms thatincorporating the 2DEG layer into the model is essential to matching spectral features withexperimental observations. Furthermore, hyperspectral imaging of a cross-sectional sampleprovides a visual representation of the 2DEG. The findings demonstrate that nano-FTIRenables the characterization of 2DEG in AlGaN/GaN heterostructures with nanometric resolution under ambient conditions, henceexpanding its applicability in the study of such systems.KEYWORDS: near-field microscopy, infrared spectroscopy, AlGaN/GaN heterostructure, two-dimensional electron gas, nano-FTIR■ INTRODUCTIONAlGaN/GaN heterostructures have garnered significantattention for their exceptional electromagnetic propertiesover a wide frequency range,1,2 making them highly relevantfor various applications including sensing3−6 and THzdetection.7,8 In electronics, power devices find a wide use ofGaN heterostructures9 but are affected by temperaturedegradation, such as in AlGaN/GaN FinFETs,10 whichmakes exploring their heat transfer and emission propertiesan important point to be considered for future applications andcharacterizations. The plethora of applications of theseheterostructures requires an ever-increasing need for accurateand detailed characterization of the structure’s properties,especially at the nanoscale, with the increasing trend ofdownscale devices.11 This crucial need underlines theimportance and motivation of the present study.Of the commonly used characterization processes, theelectrical transport in 2DEG materials is measured via Halleffect measurements,12 magnetoresistance measurements,13split C−V measurements,14 Kelvin probe,15 or by exploitingthe oscillations in the conductivity at low temperatures andhigh magnetic fields in the Shubnikov−de Haas effect.16,17Optical characterization of the 2DEG is usually carried out byphotoluminescence experiments;18,19 however, identifying the2DEG’s most prominent feature requires the disappearance ofa peak after etching the AlGaN layer, making it a destructivetechnique. Moreover, the characterization of the electronenergy and momentum, which unveils information aboutdispersion relations, is carried on via angle-resolved photo-emission spectroscopy.20 Regarding the visualization of 2DEG,the only available techniques are differential phase contrastscanning transmission electron microscopy21 or scanningtunneling microscopy.22Nano-FTIR offers a different approach to characterizing2DEG samples; it combines scattering-type scanning near-fieldoptical microscopy (s-SNOM) with Fourier transform infrared(FTIR) detection. In nano-FTIR, a broad-band infrared (IR)laser is focused on the probe apex, generating a highlyenhanced near-field between the tip apex and the sample. Thebackscattered radiation enables the derivation of the sampleproperties by analyzing its spectroscopic amplitude andphase.23 Nano-FTIR has the merit of noninvasive character-ization of samples at the nanoscale in ambient conditions.24−26The nanoscale resolution of nano-FTIR is not limited to thelateral direction. Because of the near-field interaction betweenReceived: July 8, 2025Revised: August 11, 2025Accepted: August 15, 2025Published: August 22, 2025Research Articlewww.acsami.org© 2025 The Authors. Published byAmerican Chemical Society50077https://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−50084This article is licensed under CC-BY 4.0Downloaded via NATL INST FOR MATLS SCIENCE (NIMS) on October 9, 2025 at 03:36:30 (UTC).See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.https://pubs.acs.org/action/doSearch?field1=Contrib&text1="Ilario+Bisignano"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Masataka+Imura"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Nicholaus+Kevin+Tanjaya"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Ming-Jyun+Ye"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Noriyuki+Okada"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Satoshi+Ishii"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Satoshi+Ishii"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/showCitFormats?doi=10.1021/acsami.5c12417&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?goto=articleMetrics&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?goto=recommendations&?ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?goto=supporting-info&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=tgr1&ref=pdfhttps://pubs.acs.org/toc/aamick/17/35?ref=pdfhttps://pubs.acs.org/toc/aamick/17/35?ref=pdfhttps://pubs.acs.org/toc/aamick/17/35?ref=pdfhttps://pubs.acs.org/toc/aamick/17/35?ref=pdfwww.acsami.org?ref=pdfhttps://pubs.acs.org?ref=pdfhttps://pubs.acs.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://www.acsami.org?ref=pdfhttps://www.acsami.org?ref=pdfhttps://acsopenscience.org/researchers/open-access/https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/the probe and sample, the measured spectra containinformation from a definite depth within the sample. Thisprobing depth depends, among other parameters, on thedemodulation harmonics: as the demodulation order increases,the probing depth decreases, limiting the probing depth to afew tens of nanometers from the uppermost surface. Thispivotal point makes nano-FTIR a valid technique to character-ize multilayer samples and recover material properties.27−29Because of that, nano-FTIR has been used to investigate the2DEG in the past; reports mostly focus on oxide/interface2DEGs (LaAlO3/SrTiO3 and related)30−32 where particularfocus was given on identifying the carrier concentration andmobility, while THz s-SNOM has been used for plasmoniccharacterization of graphene FET channels at nanometerresolution.33 To the best of our knowledge, this work presentsthe first demonstration of nano-FTIR spectroscopy applied todetect 2DEG in AlGaN/GaN heterostructures. Furthermore, itis the first to prove a spatially resolved visualization of 2DEGthrough hyperspectral imaging on a cross-sectional sample.This work focuses on probing the 2DEG of AlGaN/GaNheterostructures through nano-FTIR measurements andcomparing the measured spectra with analytical calculations.The analytical method using finite dipole methods (FDMs)demonstrated that the 2DEG inclusion was necessary forreproducing the nano-FTIR spectra. While the limitations andapproximations inherent in the FDM model were acknowl-edged, this study achieves notable agreement between thesimulation and experimental data. The FDM model contrastswith the finite element method, which is heavily limited bycomputational constraints.34■ METHODSThe 2DEG samples were fabricated by initially growing 2.8 μm ofGaN buffers on sapphire substrates and subsequentially AlGaN layersvia a metal−organic chemical vapor deposition system (SR2328HT-RR, TAIYO NIPPON SANSO CORPORATION). Three sampleswith AlGaN thicknesses of 15, 30, and 45 nm were prepared andlabeled with their AlGaN thickness. In addition, to obtain a crosssection of the 45 nm thick AlGaN sample (Figure 1b), a carbon layerwas deposited on the AlGaN layer inside the focused ion beam (FIB)instrument (Ethos NX5000, Hitachi High-Tech Corporation) toextend the surface area of the cross section before the cross-sectionalfabrication. The cross section was obtained via FIB using argon (Ar)milling at a tilt angle of 10°. The Ar milling was performed by rotatingthe sample stage four times with 30 s of irradiation at each positionand repeated twice; thus, the total Ar milling time was 240 s.The X-ray diffraction (XRD) patterns were obtained by a Smartlab(Rigaku Holdings Corporation). The fitting was performed todetermine the thickness and Al content of the AlGaN layer, whichwere incorporated into the FDM model alongside the average sheetresistance. At the fitting, the in-plane lattice constant of the AlGaNlayer was assumed to match that of GaN due to the coherent AlGaNgrowth. Electron density and band edges were calculated usingNextnano, a technology computer-aided design software used tomodel and analyze nanoscale semiconductor structures and devices.To characterize the 2DEG of the samples, Nextnano, a commercialsoftware, was used. Nextnano basically solves self-consistent solutionsfor the Schrödinger−Poisson equation. The used electronic bandparameters are taken from Vurgaftman et al.,35 while Nickel Schottkybarriers are used as the boundary conditions.Nano-FTIR measurements were conducted using a commercial s-SNOM (neaSCOPE, Attocube Systems AG) with a broad-bandinfrared laser covering 8−16 μm. An AFM tip (NSG30Pt, TipsNano)was operated in tapping mode with an initial tapping amplitude of 90nm and a tip resonance frequency of 297 kHz for backgroundsuppression. FTIR-based detection was used to determine theamplitude (sn) and phase (φn) of the complex scattering coefficientσ = sneiφn at the second demodulation order (n = 2). All of thecollected spectra were normalized against a silicon substratecommonly used for calibration (TGQ1, TipsNano) due to silicon’shigh reflectance in the mid-IR range. The top-view sample amplitudespectra were collected at three different positions and then averaged.The cross-sectional sample was analyzed by hyperspectral imagingwith a scan area of 61 nm × 138 nm with 10 pixels in both directions.The interferometer center and distance were set to 400 and 800 μm,respectively. Each spectrum was averaged 6 times with an integrationtime of 50 ms.The reflection was measured using an integrated sphere (Mid-IRIntegratIR, PIKE Technologies, Inc.) combined with an FTIRspectrometer (Nicolet iS50, Thermo Scientific).The sheet resistance was measured with a noncontact sheetresistance measurement instrument (LEI-1510, Semilab Inc.).■ RESULTS AND DISCUSSIONFigure 1 shows the XRD patterns and AFM images of the threeAlGaN/GaN samples with variable AlGaN thicknesses of 15,30, and 45 nm. In the following, the three samples are labeled,Figure 1. (a−c) Measured (in black) and fitted (in red) XRD patterns and (d−f) AFM images of the (a, d) 15 nm AlGaN, (b, e) 30 nm AlGaN,and (c, f) 45 nm AlGaN samples where AFM scanned the AlGaN surfaces.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450078https://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig1&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig1&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig1&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig1&ref=pdfwww.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asdepending on the AlGaN thickness. The fitting of the XRDpatterns determined the thickness and the Al contents, whichwere 0.286 for all of the samples. The AFM images show thatall of the samples are smooth, having an RMS smaller than 1nm. A commercial software, Nextnano, was used to simulatethe 2DEG layers. The simulations in Figure 2a show that thecarrier concentration increases with increasing AlGaN thick-ness. This thickness dependence aligns with the measuredsheet resistance, which decreases with increasing AlGaNthickness. The carrier concentration is dependent on theAlGaN thickness:36 as the thickness increases, the triangularpotential becomes deeper, resulting in a more confined 2DEG.In other words, increasing the film thickness allows a highernumber of free electrons to accumulate at the interface. Whenthis thickness is increased beyond a critical value, however,lattice relaxation occurs, decreasing the polarization charge andeventually leading to a decrease in the carrier concentration.On the other hand, when the AlGaN thickness is decreasedfrom the optimum value, even though the polarization chargeis present, the potential is too shallow for sufficient confining,leading to a decrease in carrier concentration, and in the caseof extremely thin layers, there is a risk of increased leakage dueto tunneling effects.Figure 2b shows the simulated energy level and electrondensity, which show the existence of the 2DEGs in the GaNsides of the AlGaN/GaN interfaces. Even though the electrondensity increases with increasing AlGaN thickness, the effective2DEG width remains nearly identical.The second-order optical amplitude (O2A) spectra weremeasured for the three AlGaN/GaN samples and the GaNsample and compared to the modeled O2A utilizing a Pythonpackage for modeling scanning near-field optical microscopymeasurements.37 The fundamental equation governing theSNOM scattering, eq 1, defines the scattering coefficient as= = +EEc r(1 )scatincr2eff(1)This equation contains information regarding both the near-field interaction (αeff) and the far-field radiation reflected fromthe substrate onto the tip (1 + crr).To compute the scattering coefficient, the FDM wasimplemented in the multilayer mode, which resulted in amore accurate description of the experimental data comparedto the Q average model, as shown in the second-orderharmonics data in Figure S1. The second-order harmonicsresponse is used for background suppression, which isconsidered in the model by simulating lock-in amplifierdemodulation at the second harmonics.37 The O2A multilayermodel better reproduces the features found in the measureddata, particularly from 600 to 700 cm−1 of GaN phononresonances, while for the second-order optical phase (O2P), itreproduces the same features at 720 cm−1, giving comparableresults. Completely opposite is the simulated result from thecharge average model, which fails to reproduce any of thepreviously mentioned distinctive attributes in both O2A andO2P.By dividing the scattering coefficient of the modeled samplewith the scattering coefficient of a reference sample (in ourcase, silicon), the near-field contrast is obtained and expressedas = = s en ninnnSi . All of the presented simulation results plotO2A and O2P of the complex-valued near-field contrast as afunction of wavenumber. Prior to that, the sample multilayercharacteristic was taken into account by defining a multilayeredsubstrate with the multilayer method proposed by Hauer etal.38 An essential parameter of the model is the AFM tipcurvature radius rtip = 10 nm, which was obtained bycomparing experimentally obtained approach curves on siliconfor demodulation orders n = 2 and n = 3 with the simulatedones of varying tip radii, shown in Figure S2a−c. In Figure S2,the normalized second and third harmonic scatteringamplitudes of a silicon substrate for different tip height valuesare compared to the simulated tip effective polarizabilitydemodulated at the second and third harmonics for the sametip heights. The comparison shows that the optimized tipradius is 10 nm, and larger (Figure S2a−c) and smaller values(Figure S2d) lead to curves that deviate far from theexperimental ones. The approach curve on silicon is commonlyused to validate simulation parameters, especially geometricalparameters, by comparing analytical results with measuredones.39 Further confirmation is found when calculating theoptical response of the 45 nm thick sample at different tip radii,as shown in Figure S2e,f. When the tip radius is too large, a redshift in the O2A dip caused by the GaN phonon resonance isobserved (Figure S2e), which contrasts with the measuredspectrum. Also, in the O2P spectrum, at higher tip radii, asecond peak appears at 735 cm−1 (Figure S2f), which contrastswith the experimentally observed data (Figure S2f in violetsolid line).Other crucial parameters include the finite dipole length L =300 nm, tapping amplitude Atip = 72 nm obtained from theexperimental conditions, light incidence angle θin = 60°, andthe weighting factor cr = 0.9. The top-view sample wasmodeled as a semi-infinite substrate of GaN with permittivity(eq 2) described as an isotropic single Lorentzian oscillatorFigure 2. (a) Calculated carrier concentration (in black) and measured sheet resistance (in red) plotted against AlGaN layer thicknesses. (b)Calculated band edges and electron density by Nextnano. In the x-axis, the GaN and AlGaN layers are located in negative and positive regions,respectively.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450079https://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig2&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig2&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig2&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig2&ref=pdfwww.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-as= +ikjjjjjjy{zzzzzzi1GaN ,GaNLO,GaN2TO,GaN2TO,GaN2 2GaN (2)where the wurtzite GaN phonon frequencies utilized weretaken from the experimental literature,40 which describes theAlN phonon red shift well due to the lattice mismatch.41 Ontop of the GaN semi-infinite layer, a 5 nm thick virtual 2DEGlayer was modeled (eq 3) with permittivity42= + id1( )2DEG2DEG(3)The 2DEG layer was capped with an AlGaN layer, where thepermittivity was described as a combination of the GaN andAlN permittivities as ϵAlGaN = xϵAlN + (1 − x)ϵGaN withconcentration x = 28.6% obtained from the XRD measure-ments. Lastly, the multilayer model ended with a top semi-infinite air layer described by constant permittivity ϵAir = 1. Indetail, the surface conductivity of the 2DEG layer wascomputed using eq 4 in the Drude model approximation= *e Nm i( )112DEG2s(4)where Ns is the carrier concentration, e is the electron charge,m* is the effective electron mass, and τ is the scattering timedefined as = *me, with μ being the electron mobility, all ofwhich were obtained using the typical experimental data foundin the literature.43,44 The carrier concentration Ns is dependenton the AlGaN thickness and was derived for each sample byutilizing the measured average sheet resistance Rsh as an inputin eq 545=eN R1s sh (5)While for the AlN permittivity, an identical single Lorentzianoscillator model used for the GaN layer was adopted, wherethe strained AlN phonon frequencies were taken from theliterature.40 For the silicon reference model, a semi-infinitelayer of silicon with constant permittivity ϵSi = 11.7 in the mid-IR (MIR) region was adopted, topped by the semi-infinite airlayer.A schematic representation of the top-view samples used forthe measurements and the detection method is shown inFigure 3a. The measured second-order optical amplitudes ofthe O2A spectra for the three AlGaN/GaN samples and theGaN sample are compared to the modeled O2A in Figure 3b,c.The measured GaN spectrum was obtained from the AlGaN/GaN cross section by exclusively scanning the GaN part, whilethe simulated GaN is the modeled semi-infinite GaN. Also, thedips in the measured AlGaN/GaN spectra at around 730 cm−1,which corresponds to the GaN LO phonon resonance, agreeconsiderably with the simulated spectra for all three samplesinvestigated, showing a similar blue shift as the AlGaNthickness increases. A similar blue shift was reported in SrTiO3nano-FTIR measured amplitude spectra and attributed to achange in the carrier concentration of the 2DEG.46 However,our simulated O2A spectra without implementing a 2DEGlayer still show a blue shift at increasing AlGaN thickness(Figure S3). Thus, the effect observed on the O2A spectracannot be fully attributed to the influence of the carrierconcentration variation in the 2DEG and suggests that theinteraction with the surrounding layers has a more prominentrole for the O2A.When analyzing the O2P spectra, significant deviations fromthe measured spectra are observed when the 2DEG layers wereFigure 3. (a) Schematic example of the top-view sample along with the detection methods summarized as an AFM tip above the sample, theincident (Einc), and scattered (Escat) electric fields. (b) Simulated and (c) measured nano-FTIR O2A for the AlGaN/GaN samples with differentAlGaN thicknesses. Bare GaN in the simulated data is the optical response of a semi-infinite GaN. (d−f) Measured and (g−i) simulated O2Pspectra for the 15, 30, and 45 nm AlGaN samples.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450080https://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig3&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig3&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig3&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig3&ref=pdfwww.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asnot implemented in the simulated model, as shown in Figure3d−i. The measured spectra (Figure 3d−f) are characterizedby soft features between 690 and 720 cm−1 or wider. Incontrast, the simulated spectra without the implementation ofthe 2DEG layers (dashed line in Figure 3g−i) have sharper andfast-varying features. Simulated spectra can reproduce similarfeatures only when the 2DEG layers are included in the model.In fact, the measured and modeled phases show a relativelybetter agreement when the AlGaN thickness is 30 and 45 nmcompared to when the thickness is 15 nm, reflecting that forthicker AlGaN layers, the effect of the increased carrierconcentration in the simulated 2DEG layer is paramount forthe correct interpretation of the measurements.In essence, the free electrons at the 2DEG interface actpartly as a mirror, reflecting the near-field, thus screening thescattering amplitude signal of the underlying GaN phononswhile enhancing the peak at 760 cm−1 of AlN at the overlyingAlGaN layer. Figure S3 shows the simulated O2A spectrawithout considering a virtual 2DEG layer; here, we can seehow the peak at 760 cm−1 increases in intensity with increasingAlGaN thickness. Additional simulations where the AlNcomposition was increased to unity showed an increase ofthe same peak, providing additional evidence supporting theassignment of the peak to AlN. Previous studies of 2DEGheterostructures via nano-FTIR considered solely theamplitude spectra while disregarding the information on thephase spectra. The zero-crossing of the real part of thedielectric function is particularly sensitive to changes in theoptical properties of the system, making it a powerful tool forinvestigating the properties of 2DEG. In our case, this region islocated around the 720 cm−1 mark, where the phase spectra areparticularly sensitive to the presence (or absence) of 2DEG,explaining why our study focuses on that region.However, certain discrepancies in both amplitude and phaseare observed when the measured and simulated spectra arecompared, highlighting the limitations of the proposedapproach. For instance, in the O2A spectrum, the AlN peakat 760 cm−1 predicted by the model is absent in theexperimental data, where a corresponding feature appearsblue-shifted to 800 cm−1 with reduced intensity, as shown inFigure S1a. Additionally, the simulated O2P spectrum exhibitsa compression in frequency, whereas the experimental resultsdisplay the same trend but over a significantly broader range.These discrepancies can be mitigated by adjusting the AlNlongitudinal optical (LO) phonon frequency from approx-imately 880 to 700 cm−1, while keeping all other parameters ofour model the same, hinting that the FDM model can widenthe calculated phase by modifying the parameters already inthe model, but as no valid physical reason has been found toadapt such an adjustment, we kept the utilized opticalparameters as from the cited literature for all of the resultsshown in this work.Apart from single-point measurements from the top,hyperspectral imaging was conducted on the cross section ofthe 45 nm AlGaN sample (Figure 4a), where the SEM imageof the cross section is shown in Figure S4. The integratedamplitudes of O2A and O2P in the range of 700 and 737 cm−1are shown in Figure 4b,c, respectively. The amplitude andphase decrease at around 75 nm in the y-axis, where thepositions match the interface of AlGaN and GaN layers. Theresults clearly show that both amplitude and phase images canresolve the 2DEG layer. Together with the measuredhyperspectral imaging, an FDM-based hyperspectral simulationwas performed, and the integrated O2A and O2P images areshown in Figure 4d,e, respectively. The simulations show asimilar trend to the experimental data when the modeled2DEG layer thickness is 5 nm or above. If the 2DEG thicknessis set to 2 nm, which is the typical 2DEG thickness for theAlGaN/GaN,47 the simulated O2P image shows a similarposition dependence; however, the simulated O2A image has adifferent position dependence compared to the measuredimage (see Figure S5a,b). Thus, to reproduce the measuredimage, setting the 2DEG thickness to 5 nm or higher isnecessary. Similarly, for AlGaN/GaN heterostructures, the2DEG thickness has been arbitrarily adjusted to even 10 nmafter considering electron density and wave function.48−50 Thenecessity to add a thicker thickness may come from a larger tipradius than the 2DEG thickness, and the reason behind this isthat an increased tip radius reduces the spatial resolution,leading to a convolution of the signals due to the presence ofthe 2DEG and either or both of the neighboring layers (GaNor AlGaN). This change in the 2DEG layer thickness, however,does not affect the top-view spectra significantly, as shown inFigure S6.Figure 4. (a) Schematic of the cross-sectional sample with a representation of the detection method. Hyperspectral imaging of the integrated (b)O2A and (c) O2P of the 45 nm thick AlGaN sample for the range between 700 and 737 cm−1. Simulated hyperspectral images of (d) and (e) O2P.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450081https://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig4&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig4&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig4&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig4&ref=pdfwww.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asIn the following, far-field optical measurements, includingreflectance and ellipsometry, are presented, which explored thepossibility of detecting traces of the 2DEG layer. Thereflectance of the three samples by the FTIR measurementand numerical simulations based on the finite element methodare shown, respectively, in Figure 5a−c. The similarity betweenthe measured and simulated spectra validates the measuredresults. They all possess similar features; the reflectance is notsignificantly different between the samples, and a negligiblechange is observed when adding the 2DEG layer in thesimulation. Further validation comes from the measuredellipsometry data (Ψ and Δ) shown in Figure S7. Themeasured values of Ψ and Δ among the three samples show nosignificant variation. Hence, obtaining notable features on the2DEG from the far-field measurements is impossible. This iscaused by the fact that the incident MIR light at the far-fieldmeasurements propagated much deeper than the 2DEG layer,such that the reflected light was not sensitive to the existenceof the 2DEG. In contrast, the limited probing depth of nano-FTIR allows the detection of the 2DEG, highlighting theadvantage of near-field measurement.■ CONCLUSIONSIn summary, we successfully detected 2DEGs in AlGaN/GaNsamples with AlGaN thicknesses from 15 to 45 nm throughnano-FTIR spectroscopy. Although the major cause of theshifts in the O2A spectra was not due to the existence of 2DEGlayers, these spectra clearly differentiate the AlGaN/GaNsamples. Hyperspectral imaging directly visualized the spatialdistribution and optical properties of the 2DEG layers,showing a stark contrast between the 2DEG layer and thesurrounding GaN and AlGaN. This highlights the novelty ofour work, as this is, to the best of our knowledge, the firstnano-FTIR study on GaN/AlGaN 2DEGs that also includescross-sectional sample analysis. The nanometer resolution andambient operation offer a significant advantage over traditional2DEG visualization methods, enhancing current character-ization methodologies. We anticipate that nano-FTIR will beapplied to characterize 2DEG layers in other samples, andfurther advances in analysis and modeling offer the potential toenhance the benefits of nano-FTIR.■ ASSOCIATED CONTENT*sı Supporting InformationThe Supporting Information is available free of charge athttps://pubs.acs.org/doi/10.1021/acsami.5c12417.Difference in simulation results between the chargeaverage method and multilayer method (Figure S1);simulated approach curve, O2A and O2P spectra fordifferent tip radii considered in the model (Figure S2);simulated O2A response without including the 2DEGlayer in the model (Figure S3); SEM image of theAlGaN cross-sectional sample after Ar milling (FigureS4); simulated integrated O2A and O2P of the 45 nmthick AlGaN/GaN sample (Figure S5); differences ofsimulated spectra between 2 and 5 nm thick 2DEG layer(Figure S6); and measured ellipsometric data for thethree AlGaN/GaN top-view samples (Figure S7) (PDF)■ AUTHOR INFORMATIONCorresponding AuthorSatoshi Ishii − International Center for MaterialsNanoarchitectonics (MANA), National Institute forMaterials Science (NIMS), Tsukuba, Ibaraki 305-0044,Japan; Graduate School of Science and Technology,University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan;orcid.org/0000-0003-0731-8428; Email: sishii@nims.go.jpAuthorsIlario Bisignano − International Center for MaterialsNanoarchitectonics (MANA), National Institute forMaterials Science (NIMS), Tsukuba, Ibaraki 305-0044,Japan; Graduate School of Science and Technology,University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan;orcid.org/0009-0004-9964-2106Masataka Imura − Research Center for Functional Materials,National Institute for Material Science (NIMS), Tsukuba,Ibaraki 305-0047, Japan; orcid.org/0000-0002-4236-9549Nicholaus Kevin Tanjaya − International Center forMaterials Nanoarchitectonics (MANA), National Institutefor Materials Science (NIMS), Tsukuba, Ibaraki 305-0044,Japan; Graduate School of Science and Technology,University of Tsukuba, Tsukuba, Ibaraki 305-8577, JapanMing-Jyun Ye − International Center for MaterialsNanoarchitectonics (MANA), National Institute forMaterials Science (NIMS), Tsukuba, Ibaraki 305-0044,Japan; College of Photonics, National Yang Ming ChiaoTung University, Tainan 711010, TaiwanNoriyuki Okada − Research Network and Facility ServicesDivision, National Institute for Materials Science (NIMS),Tsukuba, Ibaraki 305-0047, JapanFigure 5. (a) Measured and (b, c) simulated reflectance of the AlGaN/GaN top-view samples with different AlGaN thicknesses. The simulationswere conducted (b) without and (c) with the 2DEG layers.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450082https://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?goto=supporting-infohttps://pubs.acs.org/doi/suppl/10.1021/acsami.5c12417/suppl_file/am5c12417_si_001.pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Satoshi+Ishii"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://orcid.org/0000-0003-0731-8428https://orcid.org/0000-0003-0731-8428mailto:sishii@nims.go.jpmailto:sishii@nims.go.jphttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Ilario+Bisignano"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://orcid.org/0009-0004-9964-2106https://orcid.org/0009-0004-9964-2106https://pubs.acs.org/action/doSearch?field1=Contrib&text1="Masataka+Imura"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://orcid.org/0000-0002-4236-9549https://orcid.org/0000-0002-4236-9549https://pubs.acs.org/action/doSearch?field1=Contrib&text1="Nicholaus+Kevin+Tanjaya"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Ming-Jyun+Ye"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Noriyuki+Okada"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig5&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig5&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig5&ref=pdfhttps://pubs.acs.org/doi/10.1021/acsami.5c12417?fig=fig5&ref=pdfwww.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asComplete contact information is available at:https://pubs.acs.org/10.1021/acsami.5c12417NotesThe authors declare no competing financial interest.■ ACKNOWLEDGMENTSS.I. acknowledges the support from JST FOREST(JPMJFR2139) and JSPS Kakenhi (24K21723, 22H01917),Japan.■ REFERENCES(1) Alivov, Y. I.; Kalinina, E. V.; Cherenkov, A. E.; Look, D. C.;Ataev, B. M.; Omaev, A. K.; Chukichev, M. V.; Bagnall, D. M.Fabrication and Characterization of N-ZnO/p-AlGaN HeterojunctionLight-Emitting Diodes on 6H-SiC Substrates. Appl. Phys. Lett. 2003,83 (23), 4719−4721.(2) Bernardini, F.; Fiorentini, V. Macroscopic Polarization and BandOffsets at Nitride Heterojunctions. Phys. Rev. B 1998, 57, R9427−R9430.(3) Cimalla, I.; Will, F.; Tonisch, K.; Niebelschütz, M.; Cimalla, V.;Lebedev, V.; Kittler, G.; Himmerlich, M.; Krischok, S.; Schaefer, J. A.;Gebinoga, M.; Schober, A.; Friedrich, T.; Ambacher, O. AlGaN/GaNBiosensor-Effect of Device Processing Steps on the Surface Propertiesand Biocompatibility. Sens. Actuators B 2007, 123 (2), 740−748.(4) Neuberger, R.; Müller, G.; Ambacher, O.; Stutzmann, M. High-Electron-Mobility AlGaN/GaN Transistors (HEMTs) for FluidMonitoring Applications. Phys. Status Solidi A 2001, 185 (1), 85−89.(5) Steinhoff, G.; Hermann, M.; Schaff, W. J.; Eastman, L. F.;Stutzmann, M.; Eickhoff, M. PH Response of GaN Surfaces and ItsApplication for PH-Sensitive Field-Effect Transistors. Appl. Phys. Lett.2003, 83 (1), 177−179.(6) Kumar, A.; Paliwal, S.; Kalra, D.; Varghese, A.; Tripathy, S.;Ghoshal, S. K. Development of AlGaN/GaN MOSHEMT Biosensors:State-of-the-Art Review and Future Directions. Mater. Sci. Semicond.Process. 2024, 174, No. 108225.(7) Boppel, S.; Ragauskas, M.; Hajo, A.; Bauer, M.; Lisauskas, A.;Chevtchenko, S.; Ramer, A.; Kasalynas, I.; Valusis, G.; Wurfl, H. J.;Heinrich, W.; Trankle, G.; Krozer, V.; Roskos, H. G. 0.25-Mm GaNTeraFETs Optimized as THz Power Detectors and Intensity-GradientSensors. IEEE Trans. Terahertz Sci. Technol. 2016, 6 (2), 348−350.(8) Bauer, M.; Ramer, A.; Chevtchenko, S. A.; Osipov, K. Y.;Cibiraite, D.; Pralgauskaite, S.; Ikamas, K.; Lisauskas, A.; Heinrich,W.; Krozer, V.; Roskos, H. G. A High-Sensitivity AlGaN/GaN HEMTTerahertz Detector with Integrated Broadband Bow-Tie Antenna.IEEE Trans. Terahertz Sci. Technol. 2019, 9 (4), 430−444.(9) Baliga, B. J. Gallium Nitride Devices for Power ElectronicApplications. Semicond. Sci. Technol. 2013, 28 (7), No. 074011.(10) Hu, B.; Wang, Z.; Xu, K.; Tang, D. Hotspot and Non-equilibrium Thermal Transport in AlGaN/GaN FinFET: A CoupledElectron-Phonon Monte Carlo Simulation Study. Int. J. Heat MassTransfer 2025, 241, No. 126679.(11) Russo, S.; Di Carlo, A.; Ruythooren, W.; Derluyn, J.; Germain,M. Scaling Effects in AlGaN/GaN HEMTs: Comparison betweenMonte Carlo Simulations and Experimental Data. J. Comput. Electron.2006, 5 (2−3), 109−113.(12) Nifa, I.; Leroux, C.; Torres, A.; Charles, M.; Blachier, D.;Reimbold, G.; Ghibaudo, G.; Bano, E. Characterization of 2DEG inAlGaN/GaN Heterostructure by Hall Effect. Microelectron. Eng. 2017,178, 128−131.(13) Yamada, S.; Fujimoto, A.; Yagi, S.; Narui, H.; Yamaguchi, E.;Imanaka, Y. Magnetoresistance Analysis of Two-Dimensional HoleGases in GaN/AlGaN/GaN Double Heterostructures. Appl. Phys.Lett. 2024, 124 (26), No. 262102.(14) Nifa, I.; Leroux, C.; Torres, A.; Charles, M.; Reimbold, G.;Ghibaudo, G.; Bano, E. Characterization and Modeling of 2DEGMobility in AlGaN/AlN/GaN MIS-HEMT. Microelectron. Eng. 2019,215, No. 110976.(15) Nakagami, K.; Ohno, Y.; Kishimoto, S.; Maezawa, K.; Mizutani,T. Surface Potential Measurements of AlGaNGaN High-Electron-Mobility Transistors by Kelvin Probe Force Microscopy. Appl. Phys.Lett. 2004, 85 (24), 6028−6029.(16) Lo, I.; Gau, M. H.; Tsai, J. K.; Chen, Y. L.; Chang, Z. J.; Wang,W. T.; Chiang, J. C.; Aggerstam, T.; Lourdudoss, S. Anomalous k-Dependent Spin Splitting in Wurtzite Alx Ga1-x N GaNHeterostructures. Phys. Rev. B 2007, 75 (24), No. 245307.(17) Lisesivdin, S. B.; Balkan, N.; Makarovsky, O.; Patan, A.; Yildiz,A.; Caliskan, M. D.; Kasap, M.; Ozcelik, S.; Ozbay, E. Large Zero-Field Spin Splitting in AlGaN/AlN/GaN/AlN Heterostructures. J.Appl. Phys. 2009, 105 (9), No. 093701.(18) Jana, D.; Sharma, T. K. An Unambiguous Identification of 2DElectron Gas Features in the Photoluminescence Spectrum ofAlGaN/GaN Heterostructures. J. Phys. D: Appl. Phys. 2016, 49(26), No. 265107.(19) Fang, C. Y.; Lin, C. F.; Chang, E. Y.; Feng, M. S. A Study ofSubbands in AlGaN/GaN High-Electron-Mobility Transistor Struc-tures Using Low-Temperature Photoluminescence Spectroscopy.Appl. Phys. Lett. 2002, 80 (24), 4558−4560.(20) Santander-Syro, A. F.; Copie, O.; Kondo, T.; Fortuna, F.;Pailhés, S.; Weht, R.; Qiu, X. G.; Bertran, F.; Nicolaou, A.; Taleb-Ibrahimi, A.; Le F́vre, P.; Herranz, G.; Bibes, M.; Reyren, N.; Apertet,Y.; Lecoeur, P.; Barthélémy, A.; Rozenberg, M. J. Two-DimensionalElectron Gas with Universal Subbands at the Surface of SrTiO 3.Nature 2011, 469 (7329), 189−194.(21) Toyama, S.; Seki, T.; Kanitani, Y.; Kudo, Y.; Tomiya, S.;Ikuhara, Y.; Shibata, N. Real-Space Observation of a Two-Dimen-sional Electron Gas at Semiconductor Heterointerfaces. Nat. Nano-technol. 2023, 18 (5), 521−528.(22) Di Capua, R.; Radovic, M.; De Luca, G. M.; Maggio-Aprile, I.;Miletto Granozio, F.; Plumb, N. C.; Ristic, Z.; Scotti Di Uccio, U.;Vaglio, R.; Salluzzo, M. Observation of a Two-Dimensional ElectronGas at the Surface of Annealed SrTiO3 Single Crystals by ScanningTunneling Spectroscopy. Phys. Rev. B 2012, 86 (15), No. 155425.(23) Zenhausern, F.; Martin, Y.; Wickramasinghe, H. K. ScanningInterferometric Apertureless Microscopy: Optical Imaging at 10Angstrom Resolution. Science 1995, 269 (5227), 1083−1085.(24) Mester, L.; Govyadinov, A. A.; Hillenbrand, R. High-FidelityNano-FTIR Spectroscopy by on-Pixel Normalization of SignalHarmonics. Nanophotonics 2022, 11 (2), 377−390.(25) Taubner, T.; Hillenbrand, R.; Keilmann, F. Nanoscale PolymerRecognition by Spectral Signature in Scattering Infrared Near-FieldMicroscopy. Appl. Phys. Lett. 2004, 85 (21), 5064−5066.(26) Nishida, J.; Otomo, A.; Koitaya, T.; Shiotari, A.; Minato, T.;Iino, R.; Kumagai, T. Sub-Tip-Radius Near-Field Interactions inNano-FTIR Vibrational Spectroscopy on Single Proteins. Nano Lett.2024, 24 (3), 836−843.(27) Govyadinov, A. A.; Mastel, S.; Golmar, F.; Chuvilin, A.; Carney,P. S.; Hillenbrand, R. Recovery of Permittivity and Depth from Near-Field Data as a Step toward Infrared Nanotomography. ACS Nano2014, 8 (7), 6911−6921.(28) Mester, L.; Govyadinov, A. A.; Chen, S.; Goikoetxea, M.;Hillenbrand, R. Subsurface Chemical Nanoidentification by Nano-FTIR Spectroscopy. Nat. Commun. 2020, 11 (1), No. 3359.(29) Caldiroli, A.; Cappelletti, S.; Birarda, G.; Redaelli, A.; AdeleRiboldi, S.; Stani, C.; Vaccari, L.; Piccirilli, F. Infrared Nano-spectroscopy Depth-Dependent Study of Modern Materials: Morpho-Chemical Analysis of Polyurethane/Fibroin Binary Meshes. Analyst2023, 148 (15), 3584−3593.(30) Barnett, J.; Wirth, K. G.; Hentrich, R.; Durmaz, Y. C.; Rose, M.A.; Gunkel, F.; Taubner, T. Low Temperature Near-Field FingerprintSpectroscopy of 2D Electron Systems in Oxide Heterostructures andBeyond. Nat. Commun. 2025, 16 (1), No. 4417.(31) Rose, M. A.; Barnett, J.; Wendland, D.; Hensling, F. V. E.;Boergers, J. M.; Moors, M.; Dittmann, R.; Taubner, T.; Gunkel, F.Local Inhomogeneities Resolved by Scanning Probe Techniques andTheir Impact on Local 2DEG Formation in Oxide Heterostructures.Nanoscale Adv. 2021, 3 (14), 4145−4155.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450083https://pubs.acs.org/doi/10.1021/acsami.5c12417?ref=pdfhttps://doi.org/10.1063/1.1632537https://doi.org/10.1063/1.1632537https://doi.org/10.1103/PhysRevB.57.R9427https://doi.org/10.1103/PhysRevB.57.R9427https://doi.org/10.1016/j.snb.2006.10.030https://doi.org/10.1016/j.snb.2006.10.030https://doi.org/10.1016/j.snb.2006.10.030https://doi.org/10.1002/1521-396X(200105)185:1<85::AID-PSSA85>3.0.CO;2-Uhttps://doi.org/10.1002/1521-396X(200105)185:1<85::AID-PSSA85>3.0.CO;2-Uhttps://doi.org/10.1002/1521-396X(200105)185:1<85::AID-PSSA85>3.0.CO;2-Uhttps://doi.org/10.1063/1.1589188https://doi.org/10.1063/1.1589188https://doi.org/10.1016/j.mssp.2024.108225https://doi.org/10.1016/j.mssp.2024.108225https://doi.org/10.1109/TTHZ.2016.2520202https://doi.org/10.1109/TTHZ.2016.2520202https://doi.org/10.1109/TTHZ.2016.2520202https://doi.org/10.1109/TTHZ.2019.2917782https://doi.org/10.1109/TTHZ.2019.2917782https://doi.org/10.1088/0268-1242/28/7/074011https://doi.org/10.1088/0268-1242/28/7/074011https://doi.org/10.1016/j.ijheatmasstransfer.2025.126679https://doi.org/10.1016/j.ijheatmasstransfer.2025.126679https://doi.org/10.1016/j.ijheatmasstransfer.2025.126679https://doi.org/10.1007/s10825-006-8828-zhttps://doi.org/10.1007/s10825-006-8828-zhttps://doi.org/10.1016/j.mee.2017.05.009https://doi.org/10.1016/j.mee.2017.05.009https://doi.org/10.1063/5.0208784https://doi.org/10.1063/5.0208784https://doi.org/10.1016/j.mee.2019.05.003https://doi.org/10.1016/j.mee.2019.05.003https://doi.org/10.1063/1.1835551https://doi.org/10.1063/1.1835551https://doi.org/10.1103/PhysRevB.75.245307https://doi.org/10.1103/PhysRevB.75.245307https://doi.org/10.1103/PhysRevB.75.245307https://doi.org/10.1063/1.3120782https://doi.org/10.1063/1.3120782https://doi.org/10.1088/0022-3727/49/26/265107https://doi.org/10.1088/0022-3727/49/26/265107https://doi.org/10.1088/0022-3727/49/26/265107https://doi.org/10.1063/1.1485310https://doi.org/10.1063/1.1485310https://doi.org/10.1063/1.1485310https://doi.org/10.1038/nature09720https://doi.org/10.1038/nature09720https://doi.org/10.1038/s41565-023-01349-8https://doi.org/10.1038/s41565-023-01349-8https://doi.org/10.1103/PhysRevB.86.155425https://doi.org/10.1103/PhysRevB.86.155425https://doi.org/10.1103/PhysRevB.86.155425https://doi.org/10.1126/science.269.5227.1083https://doi.org/10.1126/science.269.5227.1083https://doi.org/10.1126/science.269.5227.1083https://doi.org/10.1515/nanoph-2021-0565https://doi.org/10.1515/nanoph-2021-0565https://doi.org/10.1515/nanoph-2021-0565https://doi.org/10.1063/1.1827334https://doi.org/10.1063/1.1827334https://doi.org/10.1063/1.1827334https://doi.org/10.1021/acs.nanolett.3c03479?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.nanolett.3c03479?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/nn5016314?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/nn5016314?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1038/s41467-020-17034-6https://doi.org/10.1038/s41467-020-17034-6https://doi.org/10.1039/D3AN00336Ahttps://doi.org/10.1039/D3AN00336Ahttps://doi.org/10.1039/D3AN00336Ahttps://doi.org/10.1038/s41467-025-59633-1https://doi.org/10.1038/s41467-025-59633-1https://doi.org/10.1038/s41467-025-59633-1https://doi.org/10.1039/D1NA00190Fhttps://doi.org/10.1039/D1NA00190Fwww.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-as(32) Barnett, J.; Rose, M. A.; Ulrich, G.; Lewin, M.; Kästner, B.;Hoehl, A.; Dittmann, R.; Gunkel, F.; Taubner, T. Phonon-EnhancedNear-Field Spectroscopy to Extract the Local Electronic Properties ofBuried 2D Electron Systems in Oxide Heterostructures. Adv. Funct.Mater. 2020, 30 (46), No. 2004767.(33) Soltani, A.; Kuschewski, F.; Bonmann, M.; Generalov, A.;Vorobiev, A.; Ludwig, F.; Wiecha, M. M.; Čibiraite,̇ D.; Walla, F.;Winnerl, S.; Kehr, S. C.; Eng, L. M.; Stake, J.; Roskos, H. G. DirectNanoscopic Observation of Plasma Waves in the Channel of aGraphene Field-Effect Transistor. Light: Sci. Appl. 2020, 9, 97DOI: 10.1038/s41377-020-0321-0.(34) Kusnetz, B.; Belhassen, J.; Tranca, D. E.; Stanciu, S. G.; Anton,S. R.; Zalevsky, Z.; Stanciu, G. A.; Karsenty, A. Generic Arrays ofSurface-Positioned and Shallow-Buried Gold Multi-Shapes asReference Samples to Benchmark near-Field Microscopes. Part 1:Applications in s-SNOM Depth Imaging. Results Phys. 2024, 56,No. 107318.(35) Vurgaftman, I.; Meyer, J. R. Band Parameters for Nitrogen-Containing Semiconductors. J. Appl. Phys. 2003, 94, 3675−3696.(36) Ambacher, O.; Foutz, B.; Smart, J.; Shealy, J. R.; Weimann, N.G.; Chu, K.; Murphy, M.; Sierakowski, A. J.; Schaff, W. J.; Eastman, L.F.; Dimitrov, R.; Mitchell, A.; Stutzmann, M. Two DimensionalElectron Gases Induced by Spontaneous and Piezoelectric Polar-ization in Undoped and Doped AlGaN/GaN Heterostructures. J.Appl. Phys. 2000, 87 (1), 334−344.(37) Vincent, T.; Liu, X.; Johnson, D.; Mester, L.; Huang, N.;Kazakova, O.; Hillenbrand, R.; Boland, J. L. Snompy: A Package forModelling Scattering-Type Scanning near-Field Optical Microscopy,2024. DOI: 10.48550/arXiv.2405.20948.(38) Hauer, B.; Engelhardt, A. P.; Taubner, T. Quasi-analyticalmodel for scattering infrared near-field microscopy on layeredsystems. Opt. Express 2012, 20, 13173−13188.(39) Zhang, L. M.; Andreev, G. O.; Fei, Z.; McLeod, A. S.;Dominguez, G.; Thiemens, M.; Castro-Neto, A. H.; Basov, D. N.;Fogler, M. M. Near-Field Spectroscopy of Silicon Dioxide Thin Films.Phys. Rev. B 2012, 85, No. 075419.(40) Shi, J. J. Interface Optical-Phonon Modes and Electron−Interface-Phonon Interactions in Wurtzite GaN/AlN Quantum Wells.Phys. Rev. B 2003, 68, No. 165335.(41) Khan, I.; Fang, Z.; Palei, M.; Lu, J.; Nordin, L.; Simmons, E. L.;Dominguez, O.; Islam, S. M.; Xing, H. G.; Jena, D.; Podolskiy, V. A.;Wasserman, D.; Hoffman, A. J. Engineering the Berreman Mode inMid-Infrared Polar Materials. Opt. Express 2020, 28 (19), 28590.(42) Ando, T.; Fowler, A. B.; Stern, F. Electronic Properties of Two-Dimensional Systems. Rev. Mod. Phys. 1982, 54, 437.(43) Bile, A.; Centini, M.; Ceneda, D.; Sibilia, C.; Passaseo, A.;Tasco, V.; Larciprete, M. C. Tuning of the Berreman Mode of GaN/AlxGa1-XN Heterostructures on Sapphire: The Role of the 2D-Electron Gas in the Mid-Infrared. Opt. Mater. 2024, 147, No. 114708.(44) Adamov, R. B.; Pashnev, D.; Shalygin, V. A.; Moldavskaya, M.D.; Vinnichenko, M. Y.; Janonis, V.; Jorudas, J.; Tumeṅas, S.;Prystawko, P.; Krysko, M.; Sakowicz, M.; Kasǎlynas, I. OpticalPerformance of Two Dimensional Electron Gas and Gan:C BufferLayers in Algan/Aln/Gan Heterostructures on Sic Substrate. Appl. Sci.2021, 11 (13), 6053.(45) Wilson, M.; Marinskiy, D.; Lagowski, J.; Almeida, C.;Savtchouk, A.; Nguyen, D.; Benjamin, M. Method for NoncontactMeasurement of 2DEG Mobility and Carrier Density in AlGaN/GaNon Semi-Insulating Wafers. Jpn. J. Appl. Phys. 2022, 61 (2),No. 020903.(46) Lewin, M.; Baeumer, C.; Gunkel, F.; Schwedt, A.; Gaussmann,F.; Wueppen, J.; Meuffels, P.; Jungbluth, B.; Mayer, J.; Dittmann, R.;Waser, R.; Taubner, T. Nanospectroscopy of Infrared PhononResonance Enables Local Quantification of Electronic Properties inDoped SrTiO3 Ceramics. Adv. Funct. Mater. 2018, 28, No. 1802834.(47) Kong, Y. C.; Zheng, Y. D.; Zhou, C. H.; Deng, Y. Z.; Gu, S. L.;Shen, B.; Zhang, R.; Jiang, R. L.; Shi, Y.; Han, P. Study of Two-Dimensional Electron Gas in AlN/GaN Heterostructure by a Self-Consistent Method. Phys. Status Solidi B 2004, 241 (4), 840−844.(48) Zhang, Y.; Qiao, S.; Liang, S.; Wu, Z.; Yang, Z.; Feng, Z.; Sun,H.; Zhou, Y.; Sun, L.; Chen, Z.; Zou, X.; Zhang, B.; Hu, J.; Li, S.;Chen, Q.; Li, L.; Xu, G.; Zhao, Y.; Liu, S. Gbps Terahertz ExternalModulator Based on a Composite Metamaterial with a Double-Channel Heterostructure. Nano Lett. 2015, 15 (5), 3501−3506.(49) Wang, L.; Hu, W.; Wang, J.; Wang, X.; Wang, S.; Chen, X.; Lu,W. Plasmon Resonant Excitation in Grating-Gated AlN BarrierTransistors at Terahertz Frequency. Appl. Phys. Lett. 2012, 100 (12),No. 123501.(50) Cai, M.; Liu, H.; Wang, S.; Wang, Y.; Wang, D.; Su, M.Transmission Properties of Grating-Gate GaN-Based HEMTs underDifferent Incident Angles in the Mid-Infrared Region. Opt. Express2022, 30 (25), 45599.ACS Applied Materials & Interfaces www.acsami.org Research Articlehttps://doi.org/10.1021/acsami.5c12417ACS Appl. Mater. Interfaces 2025, 17, 50077−5008450084https://doi.org/10.1002/adfm.202004767https://doi.org/10.1002/adfm.202004767https://doi.org/10.1002/adfm.202004767https://doi.org/10.1038/s41377-020-0321-0https://doi.org/10.1038/s41377-020-0321-0https://doi.org/10.1038/s41377-020-0321-0https://doi.org/10.1038/s41377-020-0321-0?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1016/j.rinp.2023.107318https://doi.org/10.1016/j.rinp.2023.107318https://doi.org/10.1016/j.rinp.2023.107318https://doi.org/10.1016/j.rinp.2023.107318https://doi.org/10.1063/1.1600519https://doi.org/10.1063/1.1600519https://doi.org/10.1063/1.371866https://doi.org/10.1063/1.371866https://doi.org/10.1063/1.371866https://doi.org/10.48550/arXiv.2405.20948https://doi.org/10.48550/arXiv.2405.20948https://doi.org/10.48550/arXiv.2405.20948?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1364/OE.20.013173https://doi.org/10.1364/OE.20.013173https://doi.org/10.1364/OE.20.013173https://doi.org/10.1103/PhysRevB.85.075419https://doi.org/10.1103/PhysRevB.68.165335https://doi.org/10.1103/PhysRevB.68.165335https://doi.org/10.1364/OE.401733https://doi.org/10.1364/OE.401733https://doi.org/10.1103/RevModPhys.54.437https://doi.org/10.1103/RevModPhys.54.437https://doi.org/10.1016/j.optmat.2023.114708https://doi.org/10.1016/j.optmat.2023.114708https://doi.org/10.1016/j.optmat.2023.114708https://doi.org/10.3390/app11136053https://doi.org/10.3390/app11136053https://doi.org/10.3390/app11136053https://doi.org/10.35848/1347-4065/ac41cahttps://doi.org/10.35848/1347-4065/ac41cahttps://doi.org/10.35848/1347-4065/ac41cahttps://doi.org/10.1002/adfm.201802834https://doi.org/10.1002/adfm.201802834https://doi.org/10.1002/adfm.201802834https://doi.org/10.1002/pssb.200301946https://doi.org/10.1002/pssb.200301946https://doi.org/10.1002/pssb.200301946https://doi.org/10.1021/acs.nanolett.5b00869?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.nanolett.5b00869?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.nanolett.5b00869?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1063/1.3695154https://doi.org/10.1063/1.3695154https://doi.org/10.1364/OE.477742https://doi.org/10.1364/OE.477742www.acsami.org?ref=pdfhttps://doi.org/10.1021/acsami.5c12417?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://www.cas.org/solutions/biofinder-discovery-platform?utm_campaign=GLO_ACD_STH_BDP_AWS&utm_medium=DSP_CAS_PAD&utm_source=Publication_ACSPubs