Phonon-enhanced mid-infrared CO2 gas sensing using boron nitride nanoresonators

Phonon-Enhanced Mid-Infrared CO2Gas Sensing Using Boron Nitride Nanoresonators Nestor Jr. Bareza, Bruno Paulillo,*Tetiana M. Slipchenko, Marta Autore, Irene Dolado, Song Liu, James H. Edgar, SaulVélez, Luis Martín-Moreno, Rainer Hillenbrand, and Valerio Pruneri Cite This: ACS Photonics 2022, 9, 34−42 Read Online ACCESS Metrics & More Article Recommendations * sıSupporting Information ABSTRACT: Hexagonal boron nitride (hBN) hosts long-lived phonon polaritons, yielding a strong mid-infrared (mid-IR) electric field enhancement and concentration on the nanometer scale. It is thus a promising material for highly sensitive mid-IR sensing and spectroscopy. In addition, hBN possesses high chemical and thermal stability as well as mechanical durability, making it suitable for operationindemandingenvironments.Inthiswork,we demonstrate a mid-IR CO2gas sensor exploiting phonon polariton (PhP) modes in hBN nanoresonators functionalized by a thin CO2adsorbing polyethylenimine (PEI) layer. We find that the PhP resonance shifts to lower frequency, weakens, and broadens for increasing CO2concentrations, which are related to the change of the permittivity of PEI upon CO2adsorption. Moreover, the PhP resonance exhibits a high signal-to-noise ratio even for small ribbon arrays of 30 ×30 μm2. Our results show the potential of hBN nanoresonators to become a novel platform for miniaturized phonon-enhanced SEIRA gas sensors. KEYWORDS: boron nitride, phonon-polaritons, gas sensor, SEIRA Mid-infrared (mid-IR) spectroscopy is a powerful technique that directly probes vibrational fingerprints of molecules (e.g., gas, biomarkers, etc.), providing specificity and structural information on molecular bonds in nondestructive fashion. This is the basis of spectroscopic gas detection devices, among which the nondispersive infrared (NDIR) sensor is arguably the most popular. 1 It consists of a gas cell, IR source and detector, and allows quantification of gas concentration through characteristic IR absorption. However, such a simple setup suffers from cross-sensitivity to different gases, not allowing selectivity since their vibrations can overlap in frequency. 2 Also, the gas chamber of NDIR is typically bulky to counteract low cross-section IR interaction with gas molecules. Alternative techniques have been developed to enhance the light−gas interaction such as cavity-enhanced absorption spectroscopy (CEAS), 3 cavity ring-down spectroscopy (CRDS), 4 and quartz-enhanced photoacoustic spectroscopy (QEPAS). 5 For spectroscopic sensing, the optical setup can be miniaturized by implementing surface enhanced infrared absorption (SEIRA) schemes, where resonant nanostructures (e.g., antennas, 6 split ring resonators, 7 nanoholes 8 ) create highly confined optical fields that greatly enhance light−matter interaction. SEIRA exceptionally improves the sensitivity of detection, allowing analysis of even minute amounts of analytes. Most established materials for SEIRA sensors are noble metal (e.g., Au) thin films. However, they exhibit lossy plasmon modes (typical quality factor Q≈10) 9 with a relatively large evanescent field decay. Newly contending materials include nanostructured graphene that supports highly confined localized surface plasmon resonances (LSPRs) that can be tuned by either electrostatic 10 or chemical doping. 11 Graphene LSPR modes, however, have low extinction, typically a few %, and poor Qfactors (typical Q≈4), 12 , which limit the strength of light-molecular interaction and signal-to-noise ratio. Emerging alternatives are van der Waals crystals supporting volumeand surface-confined phonon polaritons (PhPs) modes with long lifetimes. 13−18 PhPs arise from the coupling between electromagnetic radiation and crystal lattice vibrations in the material. In particular, nanostructured hexagonal boron nitride (hBN) exhibits resonant PhP modes with high quality factors. 13,19 High-Qresonances (Q∼102) of natural 13 and monoisotopic hBN 15 nanoresonators were recently employed Received: August 18, 2021 Published: January 5, 2022 Letter pubs.acs.org/journal/apchd5 © 2022 The Authors. Published by American Chemical Society 34 https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 Downloaded via CSIC on October 6, 2022 at 10:08:43 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. to demonstrate molecular sensing at the strong coupling limit. 19,20 The main challenge for using hBN and other 2D materials as gas sensors is the inherently poor physisorption of gas molecules on 2D nanostructures; the near-field interaction of the gases with the surface modes is weak. Functionalizing the nanostructures using a gas-adsorbing thin layer can solve this issue by concentrating the gas molecules inside the polariton field. For instance, thin films of polyethylenimine (PEI) polymer were shown to selectively adsorb CO2and could be regenerated through thermal desorption. 21 Efficient optical CO2detection was demonstrated with PEI deposited on resonant nanostructures surfaces such as metal metasurfaces, 22,23 graphene nanostructures, 24 and all-dielectric photonic crystal slab. 25 In this work, we demonstrate phonon-enhanced mid-IR gas sensing using monoisotopic hBN nanoribbon arrays functionalized with a thin CO2-adsorbing PEI layer. By recording the far-field transmission spectra of the PEI-coated ribbon arrays, we show PhP resonance modulation dependent on CO2 concentration. We also demonstrate reversible optical response by thermally desorbing CO2molecules from PEI. The main added value of hBN with respect to contending materials (e.g., metals, 22 graphene 24 ) is the high Q(∼100) of the PhP resonances within the Reststrahlen region. This is advantageous for gas selectivity because one specific gas-sensitive PEI vibrational mode can be targeted, unlike metal or graphene plasmonic nanostructures (Q≲10), where multiple vibrations overlap with the resonant mode. The proposed hybrid material sensing platform, leveraging the high Q, and high extinction PhP modes can potentially become a miniaturized sensor component for indoor air quality (IAQ) monitoring and smart ventilation systems. 26,27 The gas sensing setup and proposed hybrid-material sensor chip are sketched in Figure 1a. The gas cell with IRtransmitting windows is mounted in a Fourier Transform IR (FTIR) microscope operated in transmission mode. Inside the gas cell, the sensor chip is placed on a heating stage, and CO2 gas is introduced through gas valves. The sensor chip consists of monoisotopic hBN nanoribbons on a CaF2substrate, which exhibit transversal volume-confined PhP Fabry−Perot resonances. 19,20 They are covered with a PEI thin layer. The sketch conceptualizes the chemisorption of CO2molecules in the PEI layer which interact with amines groups producing carbamates. 28 The CO2−amines interaction and adsorption dynamics in the PEI matrix were thoroughly studied in literature and consists in a rapid interface adsorption followed by a slower diffusion into the bulk of the layer. 29,30 When the sensor chip is heated to >85 °C, CO2molecules are released and the amine sorbents in PEI are regenerated. 21 In Figure Figure 1. (a) Schematic diagram of the IR gas sensing setup consisting of a sensor chip placed on a heating stage embedded in a gas cell with IRtransmitting windows. The sensor chip consists of hBN nanoribbons fabricated on CaF2substrate with a 75 nm PEI layer coating. The cross-section sketch shows reversible chemisorption and thermal desorption of CO2molecules in the amine-rich PEI layer. (b) Experimental extinction spectra of hBN nanoribbons (without PEI layer) of two different widths and 400 nm period. For comparison, the extinction spectrum for 75 nm PEI on CaF2 is shown (solid gray) and rescaled (dashed gray) to highlight vibrational overlap with PhP modes. (c) Scanning electron microscope (SEM) and (d) atomic force microscopy (AFM) characterizations of fabricated hBN nanoribbons. ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 35 1c,d, we report the structural characterizations (scanning electron microscopy SEM and atomic force microscopy AFM) of fabricated hBN nanoribbon array with 400 nm period and 30 nm thickness. Measured IR extinction spectra (see Methods for measurement details) for two hBN nanoribbon arrays in Figure 1b (different ribbon width, same period) show sharp and intense (∼30%) PhP resonances. Increasing ribbon width shifts the PhP resonance toward lower wavenumbers, as expected from geometrical tuning above the transverse optical (TO) phonon position. 19 For comparison, the spectrum of a PEI thin film (∼75 nm) alone is displayed (gray full line) and rescaled (gray dashed line) to highlight the overlap of the PEI vibrational band at ∼1470 cm−1with the PhP resonances. In the following, we show how the addition of a PEI thin film on top of hBN nanoribbons affects their optical response. We spin coated prepared PEI solution that forms a planar layer (see Methods for preparation and SI.1 for characterizations) entirely covering the hBN nanoribbons (see sensor chip sketch in Figure 1a). The PEI layer is 75 nm thick. Since hBN is approximately 30 nm thick (Figure 1d), we estimate that the PEI layer’s thickness on top of hBN is ∼45 nm, this value conveniently matching the field decay length of the highly confined hBN PhP modes. 19 Figure 2a shows that addition of PEI on top of hBN nanoribbon arrays (Figure 1b) redshifts and broadens the PhP resonances. Similar effects are experimentally observed for other ribbon widths (see Figure 2b), which reports the extracted values of resonant peak positions and Qfactors for several ribbon widths without (blue curve) and with (red curve) PEI coating. The PhP resonances damping can be explained by its coupling to the much broader vibrational resonance of the PEI molecules. Finally, the resonance of the hybrid hBN+PEI system spectrally moves closer to PEI vibration mode at ∼1470 cm−1, corresponding to the NCOO skeletal vibration of carbamate, 31 which emerges upon CO2adsorption. 22 Using numerical simulations, we first analyzed the effect of the damping on the PhP modes, that also depends on defects induced by nanofabrication. In Figure 2c, we report the simulated extinction spectra for a representative hBN ribbon array (w= 160 nm) for different hBN damping γhBN values without (full lines) and with (dashed lines) a planar 75 nm PEI coating. A noticeable drop in the extinction and Q-factor of the PhP mode occurs upon adding PEI layer for low damping values. Our experiments can be reproduced with γhBN close to 15 cm−1. For such relatively high value of damping, the addition of PEI has negligible effect on the extinction but significantly reduces the Q-factor. In the hBN permittivity model (see Methods) and numerical simulations (lower panel of Figure 2a) for the nanoribbons in combination with a planar 75 nm PEI coating, we used γhBN =15cm −1. The next part tackles the sensitivity of the proposed sensor to CO2gas exposure. To evaluate the sensor response, the surface was exposed to varying concentrations of CO2ranging from ambient atmosphere (390 ppm) to higher levels, beyond the classified harmful value (>1000 ppm) in IAQ safety monitoring. 32 The experimental spectra in Figure 3a show that the PhP resonance redshifts, reduces its intensity, and broadens with increasing gas concentration. These changes Figure 2. Characterization of hBN nanoribbons modes with functional PEI coating. (a) Experimental (top) and corresponding simulated (bottom) extinction spectra of hBN nanoribbons without (blue) and with (red) PEI layer for two ribbon widths. PEI spectrum is also displayed for comparison (dashed gray). (b) Resonant peak positions and Qfactors extracted from experiments for various fabricated ribbon widths without (blue curves) and with (red curves) PEI layer. (c) Simulated PhP resonances of hBN nanoribbons of width 160 nm for different hBN damping values before (solid curves) and after (dashed curves) coating with PEI layer. ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 36 are not seen in the control experiment where hBN nanoribbons without PEI layer are exposed to varying CO2 concentration (see SI.3). Increasing the CO2concentration modifies the intensity of the PEI vibrational bands in the range from 1300 to 1700 cm−1, thus, altering its permittivity. Such a change in the dielectric environment induced by the adsorbed CO2is then reflected in the modulated PhP resonance. This trend agrees with simulations in Figure 3b (refer to SI.4 for the PEI permittivity fit as a function of CO2concentration). The simulated extinction and Q-factor of PhP resonances are higher than those from the experiment. This may be due to nonuniformity and defects of fabricated nanoribbons (e.g., trapezoidal shape, edge sharpness, and ribbon size deviations). Similar trends in the simulated response are obtained for different ribbon widths (see SI.5): (i) PhP resonance redshifts and broadens upon adding PEI, and (ii) PhP resonance further redshifts, reduces in intensity, and broadens with increasing CO2concentration. Moreover, SI.2B elucidates the effect of hBN damping and ribbon geometry with the optical response of PhP against the CO2concentration. Particularly, lower hBN damping and optimal hBN ribbon width yield high sensitivity against gas exposure, that is, the highest mode extinction difference between atmospheric CO2level and high CO2 concentration. The limit of detection (LOD) of our experiment can be estimated to be equivalent to atmospheric CO2 concentration (∼390 ppm), limited by our gas cell, which is not hermetically sealed. The hBN+PEI surface allows to transduce gas concentration into PhP resonance modulation with large signal-to-noise ratio (SNR), even under a very small footprint size (30 ×30 μm2). In contrast, spectra of PEI only measured with same footprint size (Figure 3c) under same CO2level variations yield a poor SNR. The spectra noise floor can be estimated as N≈0.3%, which gives S/N≥100 for the hBN+PEI surface as compared to S/Nof ≤10 for the bare PEI, that is, more than 10×SNR improvement. This highlights the advantage of utilizing highly confined and low loss hBN PhP modes. Figure 3d shows the extracted peak positions and extinctions of PhP mode in (a) as a function of average PEI signal in (c), representing nominal CO2levels. In the explored concentration range we have spectral shift of 4 cm−1and relative extinction reduction of ∼10% with respect to the peak values. Note that, for our hBN +PEI surface, monitoring the PhP mode intensity changes upon gas exposure is the most effective way to detect variations in gas concentration. This is because extinction changes are 1 order of magnitude larger than the noise floor (determined as ∼0.3%), whereas the overall spectral peak (resonance) shifts are rather small, approaching the spectral resolution (4 cm−1) of the FTIR spectrometer used in the experiments. Additional experiments performed with a thinner (∼30 nm) PEI layer show similar but weaker modulation effect upon CO2 exposure, because less PEI material overlaps with the PhP field (see SI.6). Regarding the gas selectivity, note that the PEI coating prevents direct interaction of hBN with water and other molecules present in the atmosphere. The PhP mode Figure 3. Gas sensor response of hBN+PEI against CO2gas concentration. (a) Experimental changes of extinction spectra against increasing CO2 concentration (direction of gray arrow) expressed in parts per million (ppm). Inset is a zoom-in of peak resonances. (b) Simulated extinction spectra vs increasing CO2concentration (see SI for fitted PEI permittivity values). (c) Extinction spectra of bare PEI on CaF2against same CO2 concentration levels as in (a). Inset is a zoom-in of the shaded region. (d) Extracted experimental resonant peak position kres (left axis, solid markers) and peak extinction (right axis, open markers) of PhP modes in (a) as a function of average PEI signal in (c). ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 37 thus only senses dielectric variations within the PEI layer. Those close to the vibrational mode at ∼1470 cm−1are mostly due to CO2−amine interactions. PEI interaction with water molecules is negligible around this frequency and mostly observed in the spectral range from 3000 to 3800 cm−1. 33 Beyond CO2molecules, PEI can also adsorb and react in a specific way with other gases such as volatile organic compounds (VOCs, e.g., acetone and ethanol) and NH3. 22,34 However, the PEI vibrational bands that respond to VOCs or NH3are also far from the spectral region where PhP modes can be excited. The sensor response to CO2adsorption and desorption has been investigated for several cycles at different days as shown in Figure 4a. For each cycle, the PEI is regenerated by heating the device at 95 °C for 2 min with N2gas flow to thermally desorb the CO2previously injected. In Figure 1a, we represent the reaction through which the carbamate dissociates to regenerate the amine-sorbents while releasing the CO2upon heating. 21,35 Figure 4a shows repeatable response over several days of cycles, consisting of subsequent thermal desorption and re-exposure to high CO2levels. Similar behavior occurs upon CO2exposure wherein the PhP reduces in intensity and redshifts. We estimate the response time to CO2concentration changes of 75 nm PEI to be less than 2 min, in agreement with work of Hasan et al. 22 that reported ∼2 min for 300 nm PEI. Also, in previous work, 24 it was shown that the PEI layer starts to undergo degradation after ∼3 weeks. Thus, stabilizing the PEI coating formulation (e.g., adding cross-linkers during solution preparation) is required to increase the device lifetime. 36,37 For example, the same high molecular weight branched PEI used in this work was combined with mesocelullar foam which resulted in very stable response even after 100 adsorption−desorption cyclic runs. 38 Furthermore, Figure 4b shows the reusability of the sensor. The old PEI layer is removed by oxygen plasma etching after 8 months and the bare hBN nanoribbons regains its PhP resonance at higher wavenumber. Upon addition of new 75 nm PEI coating, the PhP resonance redshifts and broadens, that is consistent with what was described previously. In our proof-of-concept experiment, a simple hBN nanoribbons geometry was used and the coupling of the PhP mode with the gas-sensitive PEI vibrational band was not optimized to fully exploit their high Q. We anticipate that further work employing advanced photonic design (e.g., embedding the hBN+PEI surface in a cavity design 22 ) will lead to improved performance with respect to contending systems. The proposed sensing surface can be further improved by reducing fabrication-induced defects, using a larger footprint area (hundreds of μm), and emerging large scale and monolayer hBN (for example by avoiding overlapping of fundamental PhP modes with higher PhP modes 20 ). Other gases could be targeted by choosing different gas-adsorbing layer (polymers, metal−organic frameworks, etc.) showing IR-active modes that are affected by the presence of a specific gas within the hBN Reststrahlen band. 39,40 We believe that our work paves the way to the use of high-QhBN resonant surfaces for applications that go Figure 4. Repeatability and reusability of the CO2sensor. (a) Cycles of response measurements upon PEI thermal desorption (gray) and CO2gas injection at high concentration (red). (b) Extinction measurements of hBN nanoribbons with and without PEI layer for (left) first time use and (right) removal of old PEI coating and application of new PEI coating after 8 months. The old PEI layer was fully removed with plasma etching. ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 38 beyond the context of a lab experiment with film model molecules such as oxides 41 and CBP 19 layers toward industrial applications, such as environmental sensing. In summary, we demonstrated a proof-of-concept experiment of phonon polariton-enhanced mid-IR gas sensing using hBN nanoribbons functionalized with CO2-adsorbing PEI polymer layer. In our system, the PhP resonance is frequencyshifted and weakened with increasing gas concentration due to the change of the local refractive index of the gas absorbing polymer. Moreover, the low loss and high extinction hBN nanoribbon PhP modes allow to sense different CO2levels with a large SNR even with a small sensing area (30 ×30 μm2). Our results confirm that low-loss phononic materials such as hBN are efficient for surface-enhanced molecular gas sensors. However, further work is needed for a comprehensive comparison with existing SEIRA materials (e.g., metals and graphene), in this way assessing their full potential for applications. ■METHODS Monoisotopic hBN Crystal Growth Method. The 10Benriched hBN crystals were grown from a metal flux method as described previously. 42 A Ni-Cr-10B powder mixture at respective 48, 48, and 4 wt % was first loaded into an alumina crucible and placed in a single-zone furnace. The furnace was evacuated and then filled with N2and H2gases to a constant pressure of 850 Torr. During the reaction process, the N2and H2gases continuously flowed through the system at rates of 125 and 5 sccm, respectively. All the nitrogen in the hBN crystal originated from the flowing N2gas. H2gas was used to minimize oxygen and carbon impurities in the hBN crystal. After a dwell time of 24 h at 1550 °C, the hBN crystals were precipitated on the metal surface by cooling at a rate of 0.5 °C/ h to 1525 °C, and then the system was quickly quenched to room temperature. EBL Fabrication of hBN Nanoribbons. Mechanically exfoliated monoisotopic hBN flakes were dry transferred onto CaF2substrate. High-resolution electron beam lithography was performed with spin-coated PMMA on top of hBN flakes. Different ribbon widths with 400 nm period were patterned on 30 ×30 μm2array elements. PMMA was developed in MIBK/ IPA (3:1), then exposed hBN areas were subsequently removed with reactive ion etcher in a SF6/Ar 1:1 plasma mixture at 20 sccm flow, 100 mTorr pressure and 100 W power. Finally, the PMMA mask was removed by immersing the sample overnight in acetone, rinsing it in IPA and drying it using a N2gun. Ultrathin PEI Coating. Branched PEI (Mw∼25000 from Sigma-Aldrich) was diluted in ethanol by magnetic stirring to obtain 1.58 wt % solution. Prepared PEI solution was spincoated on top of fabricated hBN nanoribbons at 5000 rpm for 1 min, then baked at 100 °C for 2 min. The PEI characterizations are further detailed in SI. Infrared Spectrum Measurement. Fourier transmission IR spectrometer (Bruker Tensor II) coupled with an IR microscope (Bruker Hyperion 2000) was used to collect transmission spectra. The incident IR light was polarized perpendicular to long-axis plane of hBN nanoribbons. The beam was focused using a Cassegrain objective (NA = 0.4, 15×) and passed through an aperture window of 30 ×30 μm2 that matched the array element size of fabricated regions. The transmitted IR light was collected by liquid nitrogen-cooled mercury−cadmium−tellurium (MCT) detector. Subsequent measurements were taken on the hBN+PEI surface (T, sample) and bare substrate surface (T0, background) to plot the extinction spectra (Ext = 1 −T/T0). The spectral resolution used was 4 cm−1. All the experimental spectra are the average of 100 acquisitions (number of FTIR scans) for a total acquisition time of 1.5 min (signal + background) per displayed spectrum. Subsequent measurements performed on the hBN+PEI and PEI-only surface under the same gas concentration and with the same parameters showed negligible deviations in the observed peak positions and intensities, below the spectral point spacing (Δk= 1.4 cm−1) and the intensity noise floor (ΔI= 0.3%). The volume around the optical path and of the gas cell are nitrogen-purged to remove atmospheric gases prior to CO2injection. IR fingerprints due to residual atmosphere along the beam path are eliminated when spectra are normalized to calculate extinction. EM Simulations. Full wave electromagnetic simulations were performed using the COMSOL software based on finiteelement methods in frequency domain. In order to achieve convergence, the mesh element size in the vicinity of hBN ribbon was much smaller than the wavelength of excited phonon−polariton. Dielectric Function Model of PEI and hBN. The dielectric permittivity of CO2-adsorbed PEI was obtained so that the experimentally measured transmission through 75 nm PEI layer on CaF2substrate is reproduced by the analytically computed transmission. For that, the dielectric permittivity of PEI can be expressed as a Lorentz model with three coupled oscillators: ∑ ε ε γ =+ −− ∞ = S kk i kk1(/) (/) j j jjj PEI 1 3 0 2 0 Here, ε∞ arepresents the high-frequency dielectric constant, Sjis the oscillator strength, k0jis the resonance wavenumber, and γj represents the damping factor of the Lorentzian line shapes. The values were obtained through the free parameters when fitting the transmission spectrum. The fitting procedure was made using the MatLab implemented function “nlinfit”. This function obtains the fitting parameters using iterative leastsquares estimation starting from initial values. The initial values for k0jwere those of the resonance wavenumbers in the experimental transmission spectrum and the initial values for Sj and γjwere ones and zeros, respectively. The fitted parameters for different CO2concentrations are summarized in Table S1 in the SI. The dielectric function of hBN was extracted by fitting various transmission spectra through nonpatterned hBN flakes with several thicknesses on CaF2substrates. The collection of spectra were fitted using analogous approach as the one described above for extracting the dielectric function of PEI. The in-plane ε⊥and out-of-plane ε∥dielectric permittivity of hBN, can be expressed in the single-Lorentzian form: ε ε ωω ωωωγ =+ []−[] []−− ∞i 1 a a aa a a LO 2 TO 2 TO 22 i k j j j j j j y { z z z z z z where arefers to either the transverse (⊥or a,b crystal plane) or z(∥or c crystal axis) axes, ε∞ arepresents the high-frequency dielectric constant, ωTO aand ωLO arefer to the transverse (TO) and longitudinal (LO) phonon−polariton frequencies, respectively; and γarepresents the damping factor. The free parameters ε∞ a,ωTO a,ωLO a, and γawere fitted to reproduce ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 39 the experimentally obtained transmission spectrum through hBN on CaF2substrate. As before, the fitting procedure was made using Matlab implemented function “nlinfit”. The initial guess for the parameters ε∞ a,ωTO a,ωLO a, and γawas taken from ref 13. With the procedure we obtain ε∞ ⊥= 6.1, ωTO ⊥= 1395 cm−1,ωLO ⊥= 1630 cm−1,γ⊥=8cm −1,ε∞ ∥= 2.8, ωTO ∥= 785 cm−1,ωLO ∥= 845 cm−1, and γ∥=1cm −1. The calculations of the transmission spectra through patterned hBN on CaF2 systems produce resonant values that are larger than the experimental ones, which we associate to imperfections created in the actual sample during fabrication. We have found that this effect can be taken into account by adjusting γ⊥close to 15 cm−1. ■ASSOCIATED CONTENT * sıSupporting Information The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsphotonics.1c01254. (S1) Thin PEI film characterizations such as thickness, surface morphology and deposited layer profile. (S2) Differential resonant extinction contour plots as functions of hBN damping and ribbon geometry. (S3) Sensing of bare hBN nanoribbons with varying CO2 concentration. (S4) Fit of PEI permittivity model from experimental transmission spectrum as a function of CO2concentration. (S5) Simulated response of different hBN ribbon geometry in varying CO2concentration. (S6) Sensing of hBN nanoribbons and 30 nm PEI coating with varying CO2concentration (PDF) ■AUTHOR INFORMATION Corresponding Author Bruno Paulillo −ICFO-Institut de Ciencies Fotoniques, The Barcelona Institute of Science and Technology, Castelldefels, Barcelona 08860, Spain; orcid.org/0000-0002-66750141; Email: [email protected] Authors Nestor Jr. Bareza −ICFO-Institut de Ciencies Fotoniques, The Barcelona Institute of Science and Technology, Castelldefels, Barcelona 08860, Spain Tetiana M. Slipchenko −INMA-Instituto de Nanociencia y Materiales de Aragón, CSIC-Universidad de Zaragoza, Zaragoza 50009, Spain; Departamento de Física de la Materia Condensada, Universidad de Zaragoza, Zaragoza 50009, Spain; orcid.org/0000-0003-3918-0275 Marta Autore −CIC nanoGUNE BRTA, Donostia-San Sebastián 20018, Spain Irene Dolado −CIC nanoGUNE BRTA, Donostia-San Sebastián 20018, Spain Song Liu −Tim Taylor Department of Chemical Engineering, Kansas State University, Manhattan, Kansas 66506, United States James H. Edgar −Tim Taylor Department of Chemical Engineering, Kansas State University, Manhattan, Kansas 66506, United States; orcid.org/0000-0003-0918-5964 Saul Vélez −CIC nanoGUNE BRTA, Donostia-San Sebastián 20018, Spain; IFIMAC-Condensed Matter Physics Center and Departamento de Física de la Materia Condensada, Universidad Autónoma de Madrid, Madrid E-28049, Spain Luis Martín-Moreno −INMA-Instituto de Nanociencia y Materiales de Aragón, CSIC-Universidad de Zaragoza, Zaragoza 50009, Spain; Departamento de Física de la Materia Condensada, Universidad de Zaragoza, Zaragoza 50009, Spain; orcid.org/0000-0001-9273-8165 Rainer Hillenbrand −CIC nanoGUNE BRTA and Department of Electricity and Electronics, UPV/EHU, Donostia-San Sebastián 20018, Spain; IKERBASQUE, Basque Foundation for Science, Bilbao 48009, Spain; orcid.org/0000-0002-1904-4551 Valerio Pruneri −ICFO-Institut de Ciencies Fotoniques, The Barcelona Institute of Science and Technology, Castelldefels, Barcelona 08860, Spain; ICREA-InstitucióCatalana de Recerca i Estudis Avancats, Barcelona 08010, Spain Complete contact information is available at: https://pubs.acs.org/10.1021/acsphotonics.1c01254 Funding The research leading to these results has received funding from the H2020 Programme under Grant Agreement No. 881603 (Graphene Flagship). This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie Grant Agreement No. 754510. This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie Grant Agreement No. 665884. This work was partially funded by CEX2019000910-S [MICINN/AEI/10.13039/501100011033] and Project TUNA-SURF (PID2019-106892RB-I00), Fundacio Cellex, FundacioMir-Puig, and Generalitat de Catalunya through CERCA. We acknowledge financial support from the Spanish Ministry of Science, Innovation and Universities (RTI2018-094830-B-100 and the Project MDM-2016-0618 of the Maria de Maeztu Units of Excellence Program) and the Basque Government (Grant Number IT1164-19). We acknowledge the Ministry of Science, Innovation and Universities through the ‘Maria de Maezt’Programme for Units of Excellence in R&D (CEX2018-000805-M). Further, support from the Materials Engineering and Processing program of the National Science Foundation, Award Number CMMI 1538127 for h-BN crystal growth is greatly appreciated. The hBN crystals growth is also supported by an Office of Naval Research Award No. N00014-20-1-2474. I.D. acknowledges the Basque Government (Grant No. PRE_2019_2_0164). We acknowledge Project PID2020115221GB-C41 financed by MCIN/AEI/10.13039/ 501100011033 and Aragon Government through Project QMAD. Notes The authors declare no competing financial interest. ■ACKNOWLEDGMENTS We thank Daniel Martinez for help with AFM measurements. ■REFERENCES (1) Wong, J. Y. NDIR Gas Sensor. US5747808A, 1998. (2) Stolberg-Rohr, T.; Buchner, R.; Clausen, S.; Jensen, J. M.; Skouboe, A.; Hawkins, G.; Hansen, R. S. In Optics Humidity Compensation in NDIR Exhaust Gas Measurements of NO2.Optical Sensors, 2014; Optical Society of America, 2014; paper SeTh1C.3, DOI: 10.1364/SENSORS.2014.SeTh1C.3. (3) Wang, H.; Chen, J.; Lu, K. Development of a Portable CavityEnhanced Absorption Spectrometer for the Measurement of Ambient NO3and N2O5: Experimental Setup, Lab Characterizations, and Field ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 40 Applications in a Polluted Urban Environment. Atmos. Meas. Tech. 2017,10, 1465. (4) Brown, S. S.; Stark, H.; Ravishankara, A. R. Cavity Ring-down Spectroscopy for Atmospheric Trace Gas Detection: Application to the Nitrate Radical (NO3). Appl. Phys. B: Lasers Opt. 2002,75, 173. (5) Patimisco, P.; Scamarcio, G.; Tittel, F. K.; Spagnolo, V. QuartzEnhanced Photoacoustic Spectroscopy: A Review. Sensors 2014,14 (4), 6165−6206. (6) Neubrech, F.; Huck, C.; Weber, K.; Pucci, A.; Giessen, H. Surface-Enhanced Infrared Spectroscopy Using Resonant Nanoantennas. Chem. Rev. 2017,117, 5110. (7) Cubukcu, E.; Zhang, S.; Park, Y. S.; Bartal, G.; Zhang, X. Split Ring Resonator Sensors for Infrared Detection of Single Molecular Monolayers. Appl. Phys. Lett. 2009,95 (4), 043113. (8) Gopalan, K. K.; Paulillo, B.; Mackenzie, D. M. A.; Rodrigo, D.; Bareza, N.; Whelan, P. R.; Shivayogimath, A.; Pruneri, V. Scalable and Tunable Periodic Graphene Nanohole Arrays for Mid-Infrared Plasmonics. Nano Lett. 2018,18 (9), 5913−5918. (9) Lilley, G.; Messner, M.; Unterrainer, K. Improving the Quality Factor of the Localized Surface Plasmon Resonance. Opt. Mater. Express 2015,5(10), 2112. (10) Rodrigo, D.; Limaj, O.; Janner, D.; Etezadi, D.; García De Abajo, F. J.; Pruneri, V.; Altug, H. Mid-Infrared Plasmonic Biosensing with Graphene. Science 2015,349 (6244), 165−168. (11) Paulillo, B.; Bareza, N. J.; Pruneri, V. Controlling Mid-Infrared Plasmons in Graphene Nanostructures through Post-Fabrication Chemical Doping. J. Phys.: Photonics 2021,3(3), 034001. (12) Apalkov, V.; Stockman, M. I. Proposed Graphene Nanospaser. Light: Sci. Appl. 2014,3, e191. (13) Caldwell, J. D.; Kretinin, A. v.; Chen, Y.; Giannini, V.; Fogler, M. M.; Francescato, Y.; Ellis, C. T.; Tischler, J. G.; Woods, C. R.; Giles, A. J.; Hong, M.; Watanabe, K.; Taniguchi, T.; Maier, S. A.; Novoselov, K. S. Sub-Diffractional Volume-Confined Polaritons in the Natural Hyperbolic Material Hexagonal Boron Nitride. Nat. Commun. 2014,5, 5221 DOI: 10.1038/ncomms6221. (14) Basov, D. N.; Fogler, M. M.; García De Abajo, F. J. Polaritons in van Der Waals Materials. Science 2016,DOI: 10.1126/science.aag1992. (15) Giles, A. J.; Dai, S.; Vurgaftman, I.; Hoffman, T.; Liu, S.; Lindsay, L.; Ellis, C. T.; Assefa, N.; Chatzakis, I.; Reinecke, T. L.; Tischler, J. G.; Fogler, M. M.; Edgar, J. H.; Basov, D. N.; Caldwell, J. D. Ultralow-Loss Polaritons in Isotopically Pure Boron Nitride. Nat. Mater. 2018,17 (2), 134. (16) Ma, W.; Alonso-González, P.; Li, S.; Nikitin, A. Y.; Yuan, J.; Martín-Sánchez, J.; Taboada-Gutiérrez, J.; Amenabar, I.; Li, P.; Vélez, S.; Tollan, C.; Dai, Z.; Zhang, Y.; Sriram, S.; Kalantar-Zadeh, K.; Lee, S. T.; Hillenbrand, R.; Bao, Q. In-Plane Anisotropic and Ultra-LowLoss Polaritons in a Natural van Der Waals Crystal. Nature 2018,562 (7728), 557. (17) Taboada-Gutiérrez, J.; Alvarez-Pérez, G.; Duan, J.; Ma, W.; Crowley, K.; Prieto, I.; Bylinkin, A.; Autore, M.; Volkova, H.; Kimura, K.; Kimura, T.; Berger, M. H.; Li, S.; Bao, Q.; Gao, X. P. A.; Errea, I.; Nikitin, A. Y.; Hillenbrand, R.; Martín-Sánchez, J.; Alonso-González, P. Broad Spectral Tuning of Ultra-Low-Loss Polaritons in a van Der Waals Crystal by Intercalation. Nat. Mater. 2020,19 (9), 964. (18) Li, P.; Dolado, I.; Alfaro-Mozaz, F. J.; Nikitin, A. Y.; Casanova, F.; Hueso, L. E.; Vélez, S.; Hillenbrand, R. Optical Nanoimaging of Hyperbolic Surface Polaritons at the Edges of van Der Waals Materials. Nano Lett. 2017,17 (1), 228. (19) Autore, M.; Li, P.; Dolado, I.; Alfaro-Mozaz, F. J.; Esteban, R.; Atxabal, A.; Casanova, F.; Hueso, L. E.; Alonso-González, P.; Aizpurua, J.; Nikitin, A. Y.; Vélez, S.; Hillenbrand, R. Boron Nitride Nanoresonators for Phonon-Enhanced Molecular Vibrational Spectroscopy at the Strong Coupling Limit. Light: Sci. Appl. 2018,7, 17172. (20) Autore, M.; Dolado, I.; Li, P.; Esteban, R.; Alfaro-Mozaz, F. J.; Atxabal, A.; Liu, S.; Edgar, J. H.; Vélez, S.; Casanova, F.; Hueso, L. E.; Aizpurua, J.; Hillenbrand, R. Enhanced Light−Matter Interaction in 10B Monoisotopic Boron Nitride Infrared Nanoresonators. Adv. Opt. Mater. 2021,9(5), 2001958. (21) Drage, T. C.; Smith, K. M.; Arenillas, A.; Snape, C. E. Developing Strategies for the Regeneration of Polyethylenimine Based CO2Adsorbents. Energy Procedia 2009,1, 875−880. (22) Hasan, D.; Lee, C. Hybrid Metamaterial Absorber Platform for Sensing of CO2Gas at Mid-IR. Advanced Science 2018,5(5), 1700581. (23) Pohl, T.; Sterl, F.; Strohfeldt, N.; Giessen, H. Optical Carbon Dioxide Detection in the Visible down to the Single Digit Ppm Range Using Plasmonic Perfect Absorbers. ACS Sensors 2020,5(8), 2628. (24) Bareza, N. J.; Gopalan, K. K.; Alani, R.; Paulillo, B.; Pruneri, V. Mid-Infrared Gas Sensing Using Graphene Plasmons Tuned by Reversible Chemical Doping. ACS Photonics 2020,7(4), 879. (25) Chang, Y.; Hasan, D.; Dong, B.; Wei, J.; Ma, Y.; Zhou, G.; Ang, K. W.; Lee, C. All-Dielectric Surface-Enhanced Infrared AbsorptionBased Gas Sensor Using Guided Resonance. ACS Appl. Mater. Interfaces 2018,10 (44), 38272−38279. (26) Kwon, J.; Ahn, G.; Kim, G.; Kim, J. C.; Kim, H. A Study on NDIR-Based CO2Sensor to Apply Remote Air Quality Monitoring System. ICCAS-SICE 2009 - ICROS-SICE International Joint Conference 2009 Proceedings, Fukuoka, Japan, Aug 18−21, 2009, IEEE, 2009. (27) Pitarma, R.; Marques, G.; Ferreira, B. R. Monitoring Indoor Air Quality for Enhanced Occupational Health. Journal of Medical Systems 2017,DOI: 10.1007/s10916-016-0667-2. (28) Wang, X.; Schwartz, V.; Clark, J. C.; Ma, X.; Overbury, S. H.; Xu, X.; Song, C. Infrared Study of CO 2 Sorption over “Molecular Basket”Sorbent Consisting of Polyethylenimine-Modified Mesoporous Molecular Sieve. J. Phys. Chem. C 2009,113, 7260. (29) Sharma, P.; Chakrabarty, S.; Roy, S.; Kumar, R. Molecular View of CO2 Capture by Polyethylenimine: Role of Structural and Dynamical Heterogeneity. Langmuir 2018,34 (17), 5138. (30) Al-Marri, M. J.; Khader, M. M.; Tawfik, M.; Qi, G.; Giannelis, E. P. CO2Sorption Kinetics of Scaled-up Polyethylenimine-Functionalized Mesoporous Silica Sorbent. Langmuir 2015,31 (12), 3569. (31) Hiyoshi, N.; Yogo, K.; Yashima, T. Adsorption Characteristics of Carbon Dioxide on Organically Functionalized SBA-15. Microporous Mesoporous Mater. 2005,84, 357. (32) Mumma, S. a. Transient Occupancy Ventilation By Monitoring CO2.IAQ Applications 2004,5(1), 21−23. (33) Liu, J. Multilayered PEI-Based Films for CO2Adsorption and Diffusion. Master’s thesis, University of Akron, 2013. http://rave. ohiolink.edu/etdc/view?acc_num=akron1367839488. (34) Lin, Y.; Kan, K.; Song, W.; Zhang, G.; Dang, L.; Xie, Y.; Shen, P.; Li, L.; Shi, K. Controllable Synthesis of Co3O4/Polyethyleneimine-Carbon Nanotubes Nanocomposites for CO and NH3Gas Sensing at Room Temperature. J. Alloys Compd. 2015,639, 187. (35) Chowdhury, F. A.; Yamada, H.; Higashii, T.; Goto, K.; Onoda, M. CO2 Capture by Tertiary Amine Absorbents: A Performance Comparison Study. Ind. Eng. Chem. Res. 2013,52, 8323. (36) Zhang, S.; Kang, P.; Ubnoske, S.; Brennaman, M. K.; Song, N.; House, R. L.; Glass, J. T.; Meyer, T. J. Polyethylenimine-Enhanced Electrocatalytic Reduction of CO2 to Formate at Nitrogen-Doped Carbon Nanomaterials. J. Am. Chem. Soc. 2014,136, 7845. (37) Zhai, Y.; Chuang, S. S. C. Enhancing Degradation Resistance of Polyethylenimine for CO2 Capture with Cross-Linked Poly(Vinyl Alcohol). Ind. Eng. Chem. Res. 2017,56 (46), 13766. (38) Zhang, H.; Goeppert, A.; Czaun, M.; Prakash, G. K. S.; Olah, G. A. CO2Capture on Easily Regenerable Hybrid Adsorbents Based on Polyamines and Mesocellular Silica Foam. Effect of Pore Volume of the Support and Polyamine Molecular Weight. RSC Adv. 2014,4 (37), 19403. (39) Zoshki, A.; Rahmani, M. B.; Masdarolomoor, F.; Pilehrood, S. H. Room Temperature Gas Sensing Properties of Polyaniline/ZnO Nanocomposite Thin Films. J. Nanoelectron. Optoelectron. 2017,12 (5), 465. (40) Zhou, H.; Hui, X.; Li, D.; Hu, D.; Chen, X.; He, X.; Gao, L.; Huang, H.; Lee, C.; Mu, X. Metal−Organic Framework-SurfaceEnhanced Infrared Absorption Platform Enables Simultaneous OnACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 41 Chip Sensing of Greenhouse Gases. Advanced Science 2020,7(20), 2001173. (41) Berte, R.; Gubbin, C. R.; Wheeler, V. D.; Giles, A. J.; Giannini, V.; Maier, S. A.; de Liberato, S.; Caldwell, J. D. Sub-Nanometer Thin Oxide Film Sensing with Localized Surface Phonon Polaritons. ACS Photonics 2018,5(7), 2807. (42) Liu, S.; He, R.; Xue, L.; Li, J.; Liu, B.; Edgar, J. H. Single Crystal Growth of Millimeter-Sized Monoisotopic Hexagonal Boron Nitride. Chem. Mater. 2018,30 (18), 6222. ACS Photonics pubs.acs.org/journal/apchd5 Letter https://doi.org/10.1021/acsphotonics.1c01254 ACS Photonics 2022, 9, 34−42 42 Recommended by ACS RF Sputtered Nb-Doped MoS2 Thin Film for Effective Detection of NO2 Gas Molecules: Theoretical and Experimental Studies Sankar Ganesh Ramaraj, Sang-Woo Kim, et al. MARCH 15, 2022 ACS OMEGA READ Organic Gas Sensing Performance of the Borophene van der Waals Heterostructure Jiale Shen, Xuguang Liu, et al. DECEMBER 28, 2020 THE JOURNAL OF PHYSICAL CHEMISTRY C READ First-Principles Study on Metal-Modified N2P6 Nanoscale Layers for Adsorption Performance and Sensing Capability Wenyu Guo, Xiang Ye, et al. JANUARY 05, 2022 ACS APPLIED NANO MATERIALS READ Ultrasensitive Boron–Nitrogen-Codoped CVD GrapheneDerived NO2 Gas Sensor Shubhda Srivastava, Bipin Kumar Gupta, et al. FEBRUARY 21, 2022 ACS MATERIALS AU READ Get More Suggestions >