Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Dec 9.
Published in final edited form as: Angew Chem Int Ed Engl. 2024 Dec 2;64(7):e202419869. doi: 10.1002/anie.202419869

Reversible and Ultrasensitive Detection of Nitric Oxide Using a Conductive Two-Dimensional Metal–Organic Framework

Hyuk-Jun Noh a, Doran L Pennington b, Jeong-Min Seo c, Evan Cline a, Georganna Benedetto a, Jong-Beom Baek c, Christopher H Hendon b, Katherine A Mirica a
PMCID: PMC12685488  NIHMSID: NIHMS2126851  PMID: 39581884

Abstract

This paper describes the use of a highly crystalline conductive 2D copper3(hexaiminobenzene)2 (Cu3(HIB)2) as an ultrasensitive (limit of detection of 1.8 part-per-billion), highly selective, reversible, and low power chemiresistive sensor for nitric oxide (NO) at room temperature. The Cu3(HIB)2-based sensors retain their sensing performance in the presence of humidity, and exhibit strong signal enhancement towards NO over other highly toxic reactive gases, such as NO2, H2S, SO2, NH3, CO, as well as CO2. Mechanistic investigations of the Cu3(HIB)2-NO interaction through spectroscopic analyses and density functional theory revealed that the Cu-bis(iminobenzosemiquinoid) moieties serve as the binding sites for NO sensing, while the Ni-bis(iminobenzosemiquinoid) MOF analog shows no noticeable response to NO. Overall, these findings provide a significant leap in the development of crystalline metal-bis(iminobenzosemiquinoid)-based conductive 2D MOFs as highly sensitive, selective, and reversible sensing materials for the low-power detection of highly toxic gases.

Keywords: Two Dimensional, Metal Organic Framework, Conductivity, Chemiresistive Sensor, Nitric Oxide

Graphical Abstract

Highly crystalline two-dimensional conductive metal–organic framework, copper3(hexaiminobenzene)2 (Cu3(HIB)2), has been obtained through strategic synthetic modification. The Cu3(HIB)2 device shows ultrasensitive, highly selective, reversible, and low-power chemiresistive sensor for nitric oxide. Mechanistic investigation reveals the contribution of Cu-bis(iminobenzosemiquinoid) functionality of Cu3(HIB)2) for highly sensitive and reversible NO binding sites.

graphic file with name nihms-2126851-f0006.jpg

Introduction

Two-dimensional (2D) conductive metal–organic frameworks (cMOFs) are an emerging class of well-ordered nanoporous materials constructed through the coordination of redox-active organic ligands with transition metal nodes.[1] With their intrinsic electrical conductivity, permanent porosity, and integrated functionality, 2D cMOFs have been extensively explored for wide applications, including as electrocatalysts,[2] energy storage materials,[3] and chemical sensors.[4] In particular, utilization of 2D cMOFs as chemiresistive chemical sensors offers significant advantages, including reticularly designed framework for selective material–analyte interactions, low dimensionality for enhancing the sensitivity of detection, and high conductivity for low-power electronically transduced sensing. Due to these beneficial properties, 2D cMOFs possess tremendous potential for achieving selective, sensitive, and low-power detection of gases and vapors over other established sensing materials, such as metal oxides,[5] carbon-based nanomaterials,[6] and hybrid materials.[7] However, with few exceptions,[8] most reported 2D cMOF-based sensor materials so far have showed significant cross-reactivity to a range of reactive gases and dosimetric responses with limited reversibility, which together limit selectivity and long term reusability of these materials in sensing applications.[4b, 9] Thus, development of cMOFs capable of ultrasensitive, selective, yet reversible detection of toxic gases and understanding their interaction toward analytes remains an ongoing challenge.

Nitric oxide (NO), a product of fossil fuel combustion, is a well-known hazardous environmental pollutant with strong contributions to acid rain and photochemical smog.[10] Thus, detection of NO is of interest for air quality monitoring. Furthermore, NO detection is essential in medical applications for its biological functions as a messenger molecule, and as an asthma biomarker.[11] Accordingly, developments of sensitive gas sensors for the detection of sub-ppm NO gas molecules at room temperature are highly demanded not only for environmental, but also for biomedical applications.

Since their first report of M3(hexaiminobenzene)2-based 2D cMOFs (M3(HIB)2, M: Ni, Cu, Co),[12] these materials have shown promising performance as semiconducting chemiresistive sensors[13] due to their dense assemblies of M-N4 groups, highly oriented single atoms of transition metal in the structure, as well as their high intrinsic conductivity.[14] However, the previously reported M3(HIB)2 MOFs exhibit aggregation of small crystallites (< 40 nm). This low crystallinity with large structural defects of HIB-based MOFs hinders the fundamental understanding their intrinsic structural properties and sensing mechanism. Hence, synthetic development for achieving high quality (less defective) crystalline HIB-based MOF is highly desirable to harness their full potential as chemiresistive sensors and other electronic devices.

This paper describes the significant improvement in crystallinity for the synthesis of Cu3(HIB)2 MOF with achievement of submicrometer size (~500 nm of length) of crystallites through strategic synthetic modifications. We further demonstrate that the intrinsically dense Cu-bis(iminobenzosemiquinoid) functionality of Cu3(HIB)2 with significantly improved crystallinity enables ultrasensitive, reversible, and low power detection of the nitric oxide (NO). The prepared Cu3(HIB)2 devices exhibit rapid and reversible NO detection at room temperature and low driving voltage of 0.1 V as well as 1.8 parts per billion (ppb) level of ultralow limit of detection (LOD). Additionally, Cu3(HIB)2 show outstanding reusability for at least 7 cycles of sequential NO detection at 10 ppm without any treatment, which has not been observed by previously reported hexahydroxytriphenylene-[15] and metallophthalocyanine-[9c, 16] based 2D cMOFs sensors. Moreover, Cu3(HIB)2 devices demonstrate strong signal enhancement toward detection of NO at 25 ppm, the 8 hr time weighted average permissible exposure limit (PEL) required by Occupational Safety and Health Administration (OSHA), in comparison to sensing other toxic and highly reactive gases, such as NO2, H2S, SO2, NH3, CO, and CO2 at the same or much higher concentrations. Our spectroscopic studies using diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), X-ray photoelectron spectroscopy (XPS), and electron paramagnetic resonance (EPR) spectroscopy suggest that the sensing mechanism of Cu-bis(iminobenzosemiquinoid) towards NO originates from strong material-analyte interactions, which are not observed between the structurally analogous Ni-bis(iminobenzosemiquinoid) MOF and NO. These findings are supported by density functional theory (DFT) models of the two MOF materials, which demonstrate strong NO binding and electron transfer in Cu3(HIB)2 and inactivity for the Ni-bis(iminobenzosemiquinoid) MOF. Furthermore, significant structural distortion is observed upon NO binding to Cu3(HIB)2, which destabilizes the bound state and leads to reversible NO adsorption at room temperature. Taken together, our study suggests huge potential of Cu3(HIB)2 for practical NO detection while providing the insights for design of next-generation sensitive, selective and reversible chemiresistive 2D cMOFs sensors.

Results and Discussion

Although Cu3(HIB)2 MOF has been previously reported with various synthetic methods,[13b, 14, 17] the relatively low crystallinity and small crystallite size (< 40 nm) of these previously reported materials hindered the understanding of the intrinsic properties and their utility in electronic devices. Generally, the formation of Cu3(HIB)2 requires deprotonation of aromatic amines by a base (typically by ammonium hydroxide solution), coordination to Cu ions, as well as oxidation of the ligand to yield a charge neutral MOF. We found that the addition of organic linker in the form of undissolved powder into the solution of Cu2+ ions and use of CH3COOK, instead of ammonium hydroxide, led to significant improvements in the crystallinity for Cu3(HIB)2 (Figures 1a and S1), compared to previous reports (see Supporting Information, Section S2, Figure S2). The increased crystallinity may be attributed to the fact that CH3COOK is weaker base than ammonium hydroxide, which could slow down the rate of deprotonation of HAB ligands, while also serving as the source of the acetate ion, which can act as coordination modulator (Figures S2 and S3).[18]

Figure 1.

Figure 1.

Open in a new tab

Synthesis of Cu3(HIB)2 crystals. (a) The synthetic scheme for the formation of Cu3(HIB)2 and structural illustration of three layers for slipped parallel interlayer stacking mode of Cu3(HIB)2 MOF. (b) Experimental and Pawley refined PXRD patterns of Cu3(HIB)2. (c) SEM image of Cu3(HIB)2. (d) HR-TEM image of a Cu3(HIB)2 plate along the [001] direction. Inset: FFT of the image from the selected cyan square area. Scale bar: 2 1/nm. (e) HR-TEM image of a Cu3(HIB)2 rod crystal along the [010] direction. Inset: corresponding FFT of the image. Scale bar: 2 1/nm. (f) Schematic illustration of Cu3(HIB)2 structure for imaging directions. Bottom: Line intensity profile of the lattice planes from the cyan rectangle in (e).

After reaction optimization (Table S2 and Figure S1), the Cu3(HIB)2 MOF was obtained as crystalline powder with submicrometer length (~ 500 nm) of hexagon-shaped rods, as observed by scanning electron microscopy (SEM, Figure S4) and transmission electron microscopy (TEM, Figure S5). While previous methods of accessing Cu3(HIB)2 MOFs produced only small aggregated crystallites (< 40 nm), our modified synthetic conditions significantly improved the crystallinity and crystallite size (~ 500 nm length and ~100 nm width of rod crystals) of Cu3(HIB)2 MOF (Figure S2 and S3).

To evaluate the crystallinity of the Cu3(HIB)2, powder X-ray diffraction (PXRD) was measured (Figure 1b). The PXRD patterns of the Cu3(HIB)2 exhibited sharp peaks at 2θ = 7.6°, 15.2°, 10.3° and 27.3°, corresponding to the (100), (110), (200) and (001) facets. The Pawley refinement fitting of the PXRD pattern of Cu3(HIB)2 generated best fit for slipped parallel interlayer stacking model with the space group Cmcm with unit cell parameters a= 13.380 Å, b= 23.175 Å and c= 6.520 Å, indicating near-eclipsed interlayer with slight displacement of layers (Figure 1a), which was not observed by previously reported Cu3(HIB)2 MOFs (See Section S4, Figure S6).[13b, 14, 17]

To further investigate the rod-shaped crystallites of Cu3(HIB)2 as shown in SEM (Figure 1c), high-resolution transmission electron microscopy (HR-TEM) was also examined (Figure 1d). The HR-TEM image oriented along the [001] direction (Figure 1d) and its corresponding fast-Fourier transformation (FFT) (inset of Figure 1d) of thin flake of Cu3(HIB)2 displayed a hexagonal crystalline structure with honeycomb diffraction pattern, which is consistent with expected crystal structure. Imaging perpendicular to the long axis of a Cu3(HIB)2 rod crystal revealed an ordered structure oriented along the [010] direction and well-resolved extended pore channels with 1.1 nm of pore diameter (Figures 1e and S5b). The FFTs for this rod crystal (inset of Figure 1e) and line intensity profile of the lattice (Figure 1f) demonstrated the fringe periodicity to be 1.13 and 0.57 nm, in accordance with the PXRD result. Along with two clear diffraction fringes in FFT, an additional periodic feature (diffuse spot) with a spacing of 0.326 nm could be observed owing to high crystallinity of material, indicating that the 2D layers stack in a near-eclipsed configuration as illustrated in Figure 1f.

To confirm the bonding nature of Cu3(HIB)2, X-ray photoelectron spectroscopy (XPS) was performed (Figure S7). The XPS survey spectrum of Cu3(HIB)2 displayed the presence of Cl−, which was likely introduced from the monomer, HAB·3HCl (Table S3). The existence of Cl− indicates that overall charge of Cu3(HIB)2 could be positive, which was verified in deconvoluted N 1s XPS spectrum. The deconvoluted N 1s XPS spectrum revealed two major peaks at 398.2 and 399.8 eV, which were attributed to the quinoid imine (C=N) and the benzenoid amine (C−NH),[19] corresponding to the observation of C=N and C–N stretching in Fourier-transform infrared spectroscopy (FT-IR) and Raman spectroscopy (Figures S8 and S9). Along with these two dominant peaks, an additional minor peak at 402.6 eV could be deconvoluted, which was assigned to oxidized amine (C=NH+).[20] This result supports the existence of Cl− in the Cu3(HIB)2 for charge neutrality of material. The high-resolution Cu 2p spectrum of Cu3(HIB)2 exhibited a dominant Cu2+ peak at the binding energy of 935.0 eV and a minor Cu+ peak at 932.3 eV, indicating that Cu linkage coexisted as mixed valence state of Cu+ and Cu2+ with a ratio of Cu2+/Cu+ of 5.5.[14] This result agreed well with the electron paramagnetic resonance (EPR) spectra of Cu3(HIB)2 (Figure S10). The EPR spectra of Cu3(HIB)2 exhibited a strong broad peak at g=2.067, implying the presence of paramagnetic Cu2+. Thermogravimetric analysis (TGA) of Cu3(HIB)2 revealed two gradual weight losses before 187 °C and after 285 °C (Figure S11), which could be ascribed to loss of coordinated water/physically adsorbed moisture and decomposition of material, respectively. The porosity of Cu3(HIB)2 was assessed using nitrogen (N2) isotherm measured at 77 K (Figure S12). The sample was activated at 100 °C for 12 h before measurement to remove residual solvents trapped inside the pores. Brunauer-Emmett-Teller (BET) surface area of material was observed to be 193.1 m2 g−1, which is comparable to previously reported Cu3(HIB)2.[14, 17]

UV-vis-NIR spectrum of Cu3(HIB)2 exhibited broad absorption at 1000 nm extending to the near-infrared (NIR) region, which was ascribed to a strong d-π conjugation between the copper node and HAB linker (Figure 2a). The direct optical band-gap of the Cu3(HIB)2 was evaluated from a Tauc plot, which is found to be 0.80 eV, indicating the semiconducting nature of this material, and aligning closely with previous DFT predictions for the out-of-plane gap.[14]

Figure 2.

Figure 2.

Open in a new tab

Electrical properties of Cu3(HIB)2. (a) UV-vis-NIR spectrum of Cu3(HIB)2 deposited on quartz substrate. Inset: Tauc plot derived from UV-vis-NIR spectrum of Cu3(HIB)2. (b) Arrhenius fitting of electrical conductivity of Cu3(HIB)2 as a function of temperature.

Given the improved crystallinity and crystallite size of Cu3(HIB)2, we sought to investigate the electrical property of the material. The electrical conductivity was evaluated by the four-point probe method on a pressed pellet under ambient conditions (Supporting Information, Section S11 and Figure S13). The electrical conductivity of Cu3(HIB)2 was found to be 3.23 × 10−2 S cm−1, which is 3 orders of magnitude higher than previously reported electrical conductivity of Cu3(HIB)2 measured via four-point probe method under ambient conditions.[17a, 21] We attributed this significant increase of conductivity to the improved crystallinity achieved in this study. Temperature-dependent conductivity measurement of bulk Cu3(HIB)2 MOF revealed the activation energy (Ea) of 0.19 eV for the charge carrier transport by fitting the data with the Arrhenius equation, σ = σ0exp(–Ea/(kBT), where σ0 is the prefactror, Ea is the activation energy, kB is the Boltzmann constant, and T is the absolute temperature (Figure 2b). The small Ea value with narrow optical band-gap of Cu3(HIB)2 support the semiconducting properties of the bulk Cu3(HIB)2.

To further study the overall charge of Cu3(HIB)2, dye adsorption study with cationic methylene blue (MB+) and anionic methyl orange (MO−) were performed (see Section S12. Figure S14). The material showed complete adsorption of MO− after 6 h in aqueous solution, whereas negligible adsorption of MB+ was observed. Combining the XPS result and selective dye adsorption of MO− over MB+, we conclude that Cu3(HIB)2 is positively charged in the solid state and in aqueous environment. The Cu3(HIB)2 MOF was found to be mixed valent with a Cu2+/Cu+ ratio of 5.5 in the XPS (Figure S7), which implies the presence of corresponding H+.[22] Despite the presence of Cu+, the dye uptake experiments indicate that the material likely possesses an overall positive charge, suggesting either missing linkers in the framework or the presence of two protons per Cu+ center. This can be explained by the linkers’ tendency to aromatize, with two protons (i.e., one per linker per unit cell) closely matching the thermodynamic concentration predicted by DFT.[23] The presence of protons has also been shown to result in band gap openings, further attributing the enhanced semiconducting behavior to their presence.

The significant improvements in electrical conductivity and the intrinsic semiconducting property of this material prompted us to examine the gas sensing properties of Cu3(HIB)2 at a low driving voltage of 0.1 V. Additionally, we prepared an isomorphic material with different transition metal (nickel), Ni3(HIB)2, to investigate the effect of the metal on its linker within the MOF sensing performance (see Supporting information, Section S13 and Figure S15–S19). The chemiresistive devices were fabricated by dropcasting 10 μL of MOF suspensions in water (1 mg mL−1) onto glass substrates equipped with interdigitated gold electrodes separated by 10 μm gaps. The Cu3(HIB)2 suspension exhibited excellent dispersibility, which enabled us to fabricate highly uniform and reproducible MOF films (Figure S20).

The Cu3(HIB)2 devices exhibited a substantial normalized sensing response (−ΔG/G0; Supporting Information, Section S14) of −2978 and −1959 % after the exposure to 40 ppm and OSHA required permissible exposure limit (PEL, 25 ppm) concentrations of NO for 10 min at 0.1 V (Figures 3a, S21a and S22a). Cu3(HIB)2 showed complete recovery after 20 min exposure to N2 (Figure S21), in contrast to reported metallophthalocyanine (MPc)[9c, 16] and HHTP-based 2D conductive MOFs.[15] The sensing performance of Cu3(HIB)2 with lower crystallinity and smaller crystallite sizes (synthesized under conditions specified in entry 12, Table S2) was evaluated at 25 ppm NO. The results indicate a notably weaker response and slightly reduced reversibility, compared to the highly crystalline MOF (Figure S22). This behavior can be attributed to an increased presence of defects and edge sites in the less crystalline MOF, which may interfere with the adsorption and desorption dynamics of NO, ultimately diminishing its sensing capabilities. Importantly, structurally analogous MOF, Ni3(HIB)2, displayed negligible and irreversible response of −1 % to the PEL concentration and −6 % to 40 ppm of NO after 10 min of exposure (Figures 3b and S23). Notably, Cu3(HIB)2 MOF retained its sensitivity towards NO even at 0.01 V driving voltage (Figure S24), thus showcasing strong potential for ultra-low power detection.

Figure 3.

Figure 3.

Open in a new tab

Chemiresistive sensing performances. (a) Saturation sensing traces of Cu3(HIB)2 after 10 min exposure to 40, 20, 10, 5, 2,1, and 0.5 ppm of NO. (b) Saturated responses of Cu3(HIB)2 and isoreticular structure of Ni3(HIB)2 to 25 ppm of NO in N2. (c) Responses (−ΔG/G0) of Cu3(HIB)2 after 1.5 (sky blue symbol) and 10 min exposure (purple symbol) versus concentration of NO and the observed linear relationship between the response and concentration (solid line). (d) Sensing traces of 7 sequential exposure-recovery to 10 ppm NO. Each cycle comprised a 3 min exposure and 15 min recovery. (e) Saturated responses of Cu3(HIB)2 to 25 ppm of NO in N2 and humid N2 with 5000 ppm of H2O (18 % relative humidity, RH). (f) Saturated responses of Cu3(HIB)2 to 25 ppm of NO, NO2, H2S, SO2, CO and NH3, and 5000 ppm of CO2. All error bars represent standard deviation from the average response based on at least four and three devices for Cu3(HIB)2 and Ni3(HIB)2, respectively.

Plotting saturated responses at 1.5 min and 10 min exposure against the concentrations of NO revealed strong linear relationships with coefficients of determination of R2=0.99 and 0.97 in the range of 0.5-20 and 0.5-10 ppm, respectively (Figures 3c and S25). The LOD value was derived from the linear relationship in the linear range (see the Supporting Information for detailed information). The theoretical LOD value of Cu3(HIB)2 for NO was found to be 1.8 ppb, based on responses after 1.5 min exposure, which is significantly below OSHA PEL (25 ppm) for a time-weighted average of 8 hours for NO.[24] Furthermore, the LOD for NO obtained by Cu3(HIB)2 was also lower than metal oxides,[25] conductive MOFs[15b, 26] and composite materials (Table S5).[27]

In addition to high response and low LOD of Cu3(HIB)2 to NO exposures, the Cu3(HIB)2 displayed remarkably rapid increase of current upon exposure of NO (Figure 3a). We hypothesize that the observed continuous decrease in conductance following the initial rapid response to NO exposure can likely be attributed to the gradual formation of NO2 due to the oxidation of NO by residual O2 molecules trapped within the pores of Cu3(HIB)2.[9b] Plotting the rate of initial response during less than 1 min of exposure versus the concentration of NO provided a linear relationship (Figure S26), suggesting Cu3(HIB)2 could detect NO in a concentration-dependent manner within only 1 min of initial NO exposure. Encouraged by its reversible sensing response to NO, we tested the capability of the Cu3(HIB)2 to respond to 7 consecutive exposures of NO under 10 ppm. Although gradual decrease of responses was observed with the sequential exposure of 10 ppm of NO, Cu3(HIB)2 still exhibited remarkable reusability even after 7 times exposure-recovery cycles without any further activation (Figure 3d), which suggested the high potential of Cu3(HIB)2 for real-time monitoring of NO.

The NO sensing performance under humid conditions (5000 ppm in N2 18% relative humidity, RH) was evaluated to further investigate potential suitability of Cu3(HIB)2 for practical applications. Cu3(HIB)2 devices maintained highly reversible responses of −1469 % to 25 ppm of PEL concentration of NO in humid condition (Figure 3e). However, the initial increase of normalized sensing response was observed during the first few seconds of NO exposure which was not observed in the NO sensing under dry N2 condition (Figure S27). This initial increase of normalized response and decrease in total saturated response observed under humid condition suggest possible competitive effect from the interfering species, such as H2O molecules, which may partially occupy the NO binding sites.[13b]

Selectivity is one of the essential parameters of chemiresistive sensors for practical applications. Accordingly, we compared the magnitude of response of Cu3(HIB)2 at 25 ppm (PEL concentration of NO) toward NO, versus 25 ppm of nitrogen dioxide (NO2), hydrogen sulfide (H2S), sulfur dioxide (SO2), carbon monoxide (CO) and ammonia (NH3), and 5000 ppm of carbon dioxide (CO2) after 10 min of exposure (Figure 3f). Cu3(HIB)2 exhibited significantly enhanced response to NO (−1959 %), compared to one to three orders of magnitude lower responses toward other toxic gases, suggesting that Cu3(HIB)2 device could potentially distinguish NO from conventional interference (Figure S28). Notably, the normalized sensing response (−ΔG/G0) of Cu3(HIB)2 decreased with H2S exposure (Figure S28c), which is in contrast to the sensing behavior to reported p-type semiconductor materials, such as MPc[9c, 16] and HHTP-based 2D conductive MOFs.[15] Based solely on the sensing responses, which show decreased resistance in response to the oxidizing NO and reducing H2S gas analytes, Cu3(HIB)2 might behaves as a mixed-type (p/n) semiconductor, as observed in reported Ni3(HITP)2.[15b] Taken together, remarkable reversibility of response, high NO sensing performance under humid condition and substantial signal enhancement of Cu3(HIB)2 toward NO over other reactive gases indicated strong potential of the Cu3(HIB)2 MOF for practical use as chemiresistive NO sensor.

To gain insight into the sensing mechanisms and interactions between NO gas and MOFs, we performed spectroscopic analysis using DRIFTS, XPS, and EPR upon continuous 40 ppm of NO exposure for DRIFTS and after 2 h exposure of 40 ppm NO for XPS and EPR, respectively. Cu3(HIB)2 MOF maintained its stability through evaluation of PXRD and SEM of Cu3(HIB)2 after NO exposure. Given that PXRD measurements were performed ex situ, capturing real-time changes in diffraction patterns under NO exposure was not feasible. However, crystallinity and morphology of Cu3(HIB)2 crystallites were retained after NO exposure without degradation of crystallinity of material (Figures S29 and S30).

Exposure to NO caused a substantial decrease in baseline of the Cu3(HIB)2 spectrum (Figure 4a), indicating the exposure of NO generated the change of electronic structure of MOF.[28] The new adsorption peaks at 1648 and 665 cm−1 emerged upon total 12 min of continuous 2 min exposure of NO (Figure 4a), which can be assigned to the vNO and vCu-N of Cu-NO interaction, respectively.[29] The FT-IR spectra of Cu3(HIB)2 MOF before and after DRIFTS upon continuous 40 ppm of NO exposure are indistinguishable (Figure S31), indicating the high reversibility of Cu-NO interaction. The XPS comparative analysis (carried out at 10−9 Torr) was used to further investigate the bonding chemistry of Cu3(HIB)2 in its pristine state and after exposure to NO. High-resolution XPS N 1s spectrum of Cu3(HIB)2 after NO exposure (named as Cu3(HIB)2-NO) displayed the additional weak peak at the binding energy of 405.2 eV, which could be ascribed to that NO is weakly bound to Cu in the Cu-bis(iminobenzosemiquinoid) moieties (Figures 4b and S32),[30] this spectroscopic feature was absent in the pristine MOF. Slight oxidation of Cu+ to Cu2+ was observable in high-resolution Cu 2p XPS spectrum (Figure 4c) after NO exposure. The NO exposure to Cu3(HIB)2 led to observable increase of unsymmetrical EPR signal, which may be due to anisotropic binding NO to Cu. The slightly increased the Cu-centered radical EPR signal with shifted g factor was observed after NO exposure (Figure 4d). This corresponds to the small increase of paramagnetic Cu2+ to Cu+ ratio, as evidenced by high-resolution Cu 2p XPS spectrum (Figures 4c and S32).

Figure 4.

Figure 4.

Open in a new tab

Mechanistic investigation of Cu3(HIB)2 toward NO exposure. (a) FT-IR spectrum of Cu3(HIB)2 MOF mixed with KBr and DRIFTS of Cu3(HIB)2 after continuous exposure to 40 ppm of NO for 12 min. Pristine Cu3(HIB)2 MOF peaks are marked (*). Comparison of the high-resolution (b) N 1s and (c) Cu 2p XPS spectra of the pristine Cu3(HIB)2 and Cu3(HIB)2 after 2 h exposure to 40 ppm of NO named as Cu3(HIB)2-NO. (d) Comparison of the EPR of the pristine Cu3(HIB)2 and Cu3(HIB)2-NO.

To further study sensing mechanisms, we examined the DRIFTS, XPS, and EPR for the Ni3(HIB)2 before and after NO (Ni3(HIB)2-NO) exposure as control experiments. Exposure to NO created only a comparatively negligible increase in baseline of the Ni3(HIB)2 spectrum compared to DRIFTS spectrum of Cu3(HIB)2 without any additional peak appearances (Figure S33), suggesting a weak interaction with NO. In high-resolution Ni 2p XPS spectra, no peak shifts were detected after 40 ppm of NO exposure to Ni3(HIB)2 (Figure S34), indicating oxidation state of Ni maintained as +2,[31] and interaction between gas and MOF might be weak. Additionally, N 1s XPS spectra of pristine Ni3(HIB)2 and Ni3(HIB)2-NO showed identical deconvoluted peaks, implying absence of NO species in Ni3(HIB)2-NO even after 2h exposure of NO. No detectable EPR signal was observed in Ni3(HIB)2-NO (Figure S35), indicating absence of species with unpaired spins, which agreed with the XPS spectra of Ni 2p (Figure S34b). Taken together, no significant interactions between Ni3(HIB)2 and NO were observed, which might result in negligible sensing responses to NO compared to substantial responses of Cu3(HIB)2 to NO.

To understand the highly sensitive and reversible NO sensing response of the Cu3(HIB)2, as well as the differing responses of Cu3(HIB)2 and Ni3(HIB)2 to NO exposure, density functional theory was used to model the MOFs and examine thenature of NO binding.[32] The MOFs structure were first geometrically optimized as bulk periodic crystals, which recovered the Cmcm point group observed by PXRD (Figure 1b). An inspection of the hybrid functional band structure for Cu3(HIB)2 showed semiconducting behavior in-plane with a band-gap of 0.91 eV, which is in good agreement with the experimental optical gap of 0.80 eV (Figure 5a). The band structure displayed metallic bands crossing the Fermi level along the stacking direction. However, the electrical conductivity measurement of pelletized Cu3(HIB)2 MOF can be influenced by significant thermally activated hopping across the grain boundaries (i.e., interparticle transport),[33] which can help rationalized the semiconducting property of Cu3(HIB)2 MOF, observed experimentally. Additionally, the 3 orders of magnitude enhancement in bulk electrical conductivity of Cu3(HIB)2 MOF observed in this work, compared to the previously reported Cu3(HIB)2 MOF, can be attributed to highly improved interlayer stacking.

Figure 5.

Figure 5.

Open in a new tab

Hybrid functional (HSEsol) band structures for Cu3(HIB)2. (a) The ground state all-Cu2+ material displays in-plane semiconducting behavior and a direct gap of 0.91 eV. (b) Singly reducing the unit cell to 1/6 Cu+ to resemble the configuration responsible for NO binding raises the Fermi level and contracts the spacing between the in-plane bands, yielding an in-plane conductor with a direct gap of 0.38 eV. (c) Structural distortion upon NO binding for a bilayer of Cu3(HIB)2, which imparts reversibility at room temperature.

To model NO binding, bilayer slabs of the relaxed frameworks were constructed incorporating a 15 Å vacuum layer. The metal-NO geometry was initialized using reported bond distances and angles for Cu and Ni nitrosyls.[34] For ground-state configurations of both Cu3(HIB)2 and Ni3(HIB)2, NO desorption occurred (Figure S36). However, when one electron was added to the Cu3(HIB)2 unit cell reducing a single Cu2+ site to Cu+, a stable Cu-NO interaction was formed with an adsorption energy of 18 meV, defined as the difference in energy between the Cu3(HIB)2-NO complex and the sum of the clean slab and free NO. The binding event oxidizes the Cu+ site back to Cu2+ according to the d-orbital occupancies of the Cu atoms, which supports experimental evidence for oxidation upon NO exposure. Although the direct reduction of the unit cell does not explicitly model the presence of protonic defects inferred by the dye uptake experiment, direct reduction captures the resulting effect of reduced Cu sites accurately in the absence of definitive evidence of structural defects.

Hybrid band structures obtained by reducing a single pristine unit cell by 1 electron to 1/6 Cu+ raises the Fermi level by one band (Figure 5b). Although the general character of the bands is conserved, structural distortions induced by the reduction cause the in-plane band-gap to contract to 0.38 eV, resulting in the in-plane bands crossing the Fermi level and leads to increase of conductivity as observed by sensing responses of decreased resistance in response to the NO. This further points to the function of Cu+ sites distributed through the crystal as being responsible for NO sensing activity and contributing enhanced conductivity to Cu3(HIB)2, offering a new handle by which to tune the properties of this framework material. The Cu-N bond distance of 1.83 Å corresponds to a typical bent Cu nitrosyl and indicates a stable bonding interaction, yet there are significant structural distortions that accompany NO binding which destabilize the bound state and leads to reversible NO adsorption at room temperature (Figure 5c).

The Cu-NO binding interaction was further examined by modeling the change in N-O stretching frequency upon binding. The vibrational frequency of adsorbed NO is highly sensitive to its bonding environment and can range from 1600 to 1900 cm−1 depending on the electronic configuration of the Cu-N bond.[29b] By modeling the NO stretching frequency using a finite differences approach to solve for the force constant of the harmonic oscillator, a vibrational frequency of 1655 cm−1 was calculated, which agrees with the experimental IR peak measured at 1648 cm−1 in DRIFTS analysis (Figure 4a), and is dissimilar to unbound NO (1716 cm−1), strongly suggesting that the NO interacts with Cu open metal-sites.

Conclusion

In this work, we achieved the Cu3(HIB)2 MOF-based chemiresistive sensor, capable of realizing low-power, sensitive, reversible, and selective detection of nitic oxide at room temperature. The significant improvement of crystallinity and size of crystallites (~ 500 nm length and ~ 100 nm width of rod crystals) of Cu3(HIB)2 for this work overcame the relatively low crystallinity and small crystallites (< 40 nm) of previously reported Cu3(HIB)2 MOFs, which enabled the use of abundant Cu(iminobenzosemiquinoid) moieties in the MOF as sensing sites.

In the implementation of highly crystalline hexaiminobenzene-based 2D MOF as chemiresistive devices for detection of toxic gases, Cu3(HIB)2 showed remarkable response toward NO with ultralow LOD of 1.8 ppb level after only 1.5 min exposure of NO. The fabricated Cu3(HIB)2 devices exhibited rapid and reversible NO response, as demonstrated by at least 7 sequential exposure-recovery sensing experiments to 10 ppm of NO, showing good recyclability without any additional treatment. The sensing traces of several interfering toxic gases of NO2, H2S, SO2, CO, and NH3 at PEL concentration of NO (25 ppm) and 5000 ppm of CO2 demonstrated outstanding signal enhancement for detection of NO. The LOD value for NO is comparable to reported MPc-based 2D MOFs, but notable reversibility and high selectivity for NO detection at room temperature constitute a significant novel advance for this class of materials. Combining spectroscopic analyses and DFT calculation results of Cu3(HIB)2, along with comparative studies using Ni3(HIB)2, we found that Cu(iminobenzosemiquinoid) moieties showed the slight oxidation of Cu+ to Cu2+ and structural distortion upon NO binding to Cu3(HIB)2, which provide a powerful platform for highly sensitive and reversible chemiresistive NO sensing. Taken together, our findings pave the way for the design and implementation of new generation of chemiresistor materials for practical NO detection. Moreover, these results provide improved access and fundamental insight into highly crystalline bis(iminobenzosemiquinoid)-based conductive 2D MOF, which has promising potential for various applications that range beyond chemiresistive sensors, and into (opto)electronic multifunctional devices.

Supplementary Material

Accepted Supporting Information

The authors have cited additional references within the Supporting Information. [9c, 14, 15b, 17b, 25b, 26–27, 31, 35]

Acknowledgements

We acknowledge support from NSF CAREER Award (#1945218) and NIH MIRA Award (R35GM138318). KAM also acknowledges the partial support under PE 0603734A, “Energy and Technology Research in Cold and Arctic Regions,” Task 2 under Contract W913E519C0008 and under PE 0633119, and under Contract W913E520C0010, both managed by the US Army Engineer Research and Development Center (ERDC). H.-J.N. acknowledges support by the Postdoctoral Fellowship from the National Research Foundation of Korea under grant no. RS-2023-00238918. The authors thank the University Instrumentation Center at the University of New Hampshire (Durham, NH) for the access to XPS. Computational aspects of the paper were supported by National Science Foundation through the Division of Materials Research under Grant DMR-1956403 and support from the Camille and Henry Dreyfus Foundation. This work used Expanse at SDSC through allocation CHE160003 from the Advanced Cyberin-frastructure Coordination Ecosystem: Services and Support (ACCESS) program, which is supported by the National Science Foundation grants #2138259, #2138286, #2138307, #2137603, and #2138296.

References

  • [1].Xie LS, Skorupskii G, Dincă M, Chem. Rev 2020, 120, 8536–8580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2] a).Miner EM, Fukushima T, Sheberla D, Sun L, Surendranath Y, Dincă M, Nat. Commun 2016, 7, 10942; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Liu L, Xu Q, Zhu Q-L, Adv. Energy Sustainability Res 2021, 2, 2100100. [Google Scholar]
  • [3] a).Liu J, Song X, Zhang T, Liu S, Wen H, Chen L, Angew. Chem. Int. Ed 2021, 60, 5612–5624; [DOI] [PubMed] [Google Scholar]; b) Sheberla D, Bachman JC, Elias JS, Sun C-J, Shao-Horn Y, Dincă M, Nat. Mater 2017, 16, 220–224. [DOI] [PubMed] [Google Scholar]
  • [4] a).Park C, Baek JW, Shin E, Kim I-D, ACS Nanosci. Au 2023, 3, 353–374; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Campbell MG, Liu SF, Swager TM, Dincă M, J. Am. Chem. Soc 2015, 137, 13780–13783. [DOI] [PubMed] [Google Scholar]
  • [5] a).Kim D-H, Kim JK, Oh D, Park S, Kim YB, Ko J, Jung W, Kim I-D, ACS Nano 2023, 17, 5842–5851; [DOI] [PubMed] [Google Scholar]; b) Kumar R, Al-Dossary O, Kumar G, Umar A, Nanomicro Lett 2015, 7, 97–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6] a).Salehi-Khojin A, Khalili-Araghi F, Kuroda MA, Lin KY, Leburton J-P, Masel RI, ACS Nano 2011, 5, 153–158; [DOI] [PubMed] [Google Scholar]; b) Schedin F, Geim AK, Morozov SV, Hill EW, Blake P, Katsnelson MI, Novoselov KS, Nat. Mater 2007, 6, 652–655. [DOI] [PubMed] [Google Scholar]
  • [7] a).Koo W-T, Cha J-H, Jung J-W, Choi S-J, Jang J-S, Kim D-H, Kim I-D, Adv. Funct. Mater 2018, 28, 1802575; [Google Scholar]; b) Hatchett DW, Josowicz M, Chem. Rev 2008, 108, 746–769. [DOI] [PubMed] [Google Scholar]
  • [8] a).Campbell MG, Sheberla D, Liu SF, Swager TM, Dincă M, Angew. Chem. Int. Ed 2015, 54, 4349–4352; [DOI] [PubMed] [Google Scholar]; b) Chen P, Su X, Wang C, Zhang G, Zhang T, Xu G, Chen L, Angew. Chem. Int. Ed 2023, 62, e202306224; [DOI] [PubMed] [Google Scholar]; c) Yan X, Chen J, Su X, Zhang J, Wang C, Zhang H, Liu Y, Wang L, Xu G, Chen L, Angew. Chem. Int. Ed 2024, n/a, e202408189. [DOI] [PubMed] [Google Scholar]
  • [9] a).Roh H, Kim D-H, Cho Y, Jo Y-M, del Alamo JA, Kulik HJ, Dincă M, Gumyusenge A, Adv. Mater 2024, n/a, 2312382; [DOI] [PubMed] [Google Scholar]; b) Meng Z, Stolz RM, De Moraes LS, Jones CG, Eagleton AM, Nelson HM, Mirica KA, Angew. Chem. Int. Ed 2024, n/a, e202404290; [DOI] [PMC free article] [PubMed] [Google Scholar]; c) Meng Z, Aykanat A, Mirica KA, J. Am. Chem. Soc 2019, 141, 2046–2053. [DOI] [PubMed] [Google Scholar]
  • [10].Tsujita W, Yoshino A, Ishida H, Moriizumi T, Sens. Actuators, B 2005, 110, 304–311. [Google Scholar]
  • [11] a).Kuzmych O, Allen BL, Star A, Nanotechnology 2007, 18, 375502; [Google Scholar]; b) Ratnawati, Morton J, Henry RL, Thomas PS, Pediatr. Pulmonol 2006, 41, 929–936. [DOI] [PubMed] [Google Scholar]
  • [12].Lahiri N, Lotfizadeh N, Tsuchikawa R, Deshpande VV, Louie J, J. Am. Chem. Soc 2017, 139, 19–22. [DOI] [PubMed] [Google Scholar]
  • [13] a).Stassen I, Dou J-H, Hendon C, Dincă M, ACS Cent. Sci 2019, 5, 1425–1431; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Park S, Zhang Z, Qi H, Liang B, Mahmood J, Noh H-J, Hambsch M, Wang M, Wang M, Ly KH, Wang Z, Weidinger IM, Zhou S, Baek J-B, Kaiser U, Mannsfeld SCB, Feng X, Dong R, ACS Materials Lett. 2022, 4, 1146–1153. [Google Scholar]
  • [14].Dou J-H, Sun L, Ge Y, Li W, Hendon CH, Li J, Gul S, Yano J, Stach EA, Dincă M, J. Am. Chem. Soc 2017, 139, 13608–13611. [DOI] [PubMed] [Google Scholar]
  • [15] a).Eagleton AM, Ko M, Stolz RM, Vereshchuk N, Meng Z, Mendecki L, Levenson AM, Huang C, MacVeagh KC, Mahdavi-Shakib A, Mahle JJ, Peterson GW, Frederick BG, Mirica KA, J. Am. Chem. Soc 2022, 144, 23297–23312; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Smith MK, Mirica KA, J. Am. Chem. Soc 2017, 139, 16759–16767. [DOI] [PubMed] [Google Scholar]
  • [16].Aykanat A, Meng Z, Stolz RM, Morrell CT, Mirica KA, Angew. Chem. Int. Ed 2022, 61, e202113665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17] a).Choi JY, Park J, ACS Appl. Electron. Mater 2021, 3, 4197–4202; [Google Scholar]; b) Feng D, Lei T, Lukatskaya MR, Park J, Huang Z, Lee M, Shaw L, Chen S, Yakovenko AA, Kulkarni A, Xiao J, Fredrickson K, Tok JB, Zou X, Cui Y, Bao Z, Nat. Energy 2018, 3, 30–36. [Google Scholar]
  • [18].Chen T, Dou J-H, Yang L, Sun C, Libretto NJ, Skorupskii G, Miller JT, Dincă M, J. Am. Chem. Soc 2020, 142, 12367–12373. [DOI] [PubMed] [Google Scholar]
  • [19].Park J, Lee M, Feng D, Huang Z, Hinckley AC, Yakovenko A, Zou X, Cui Y, Bao Z, J. Am. Chem. Soc 2018, 140, 10315–10323. [DOI] [PubMed] [Google Scholar]
  • [20].Shinde SS, Lee CH, Jung J-Y, Wagh NK, Kim S-H, Kim D-H, Lin C, Lee SU, Lee J-H, Energy Environ. Sci 2019, 12, 727–738. [Google Scholar]
  • [21].Choi JY, Wang M, Check B, Stodolka M, Tayman K, Sharma S, Park J, Small 2023, 19, 2206988. [DOI] [PubMed] [Google Scholar]
  • [22] a).Debela TT, Hendon CH, ACS Materials Lett. 2024, 6, 2698–2702; [Google Scholar]; b) Debela TT, Yang MC, Hendon CH, J. Am. Chem. Soc 2023, 145, 11387–11391. [DOI] [PubMed] [Google Scholar]
  • [23].Demuth MC, Hendon CH, ACS Materials Lett. 2023, 5, 1476–1480. [Google Scholar]
  • [24].Permissible Exposure Limits—Annotated Tables, Technical Report for United States Department of Labor, OccupationalSafety and Health Administration, https://www.osha.gov/anno-tated-pels. [Google Scholar]
  • [25] a).Liu H, He H, Chen L, Pan Q, Zhang G, J. Alloys Compd 2021, 868, 159180; [Google Scholar]; b) Sun C, Maduraiveeran G, Dutta P, Sens. Actuators, B 2013, 186, 117–125. [Google Scholar]
  • [26].Xu S, Liu X, Wu J, Wu J, ACS Sens. 2023, 8, 2348–2358. [DOI] [PubMed] [Google Scholar]
  • [27] a).Gao Y, Wang J, Feng Y, Cao N, Li H, de Rooij NF, Umar A, French PJ, Wang Y, Zhou G, Small 2022, 18, 2103259; [DOI] [PubMed] [Google Scholar]; b) Hou S, Pang R, Chang S, Ye L, Xu J, Wang X, Zhang Y, Shang Y, Cao A, ACS Appl. Mater. Interfaces 2020, 12, 29778–29786. [DOI] [PubMed] [Google Scholar]
  • [28].Stolz RM, Mahdavi-Shakib A, Frederick BG, Mirica KA, Chem. Mater 2020, 32, 7639–7652. [Google Scholar]
  • [29] a).Rout KC, Mondal B, Dalton Trans. 2015, 44, 1829–1835; [DOI] [PubMed] [Google Scholar]; b) Sarma M, Mondal B, Inorg. Chem 2011, 50, 3206–3212; [DOI] [PubMed] [Google Scholar]; c) Fallberg A, Ottosson M, Carlsson J-O, J. Cryst. Growth 2010, 312, 1779–1784. [Google Scholar]
  • [30] a).Křepelová A, Newberg J, Huthwelker T, Bluhm H, Ammann M, Phys. Chem. Chem. Phys 2010, 12, 8870–8880; [DOI] [PubMed] [Google Scholar]; b) Dubey M, Bernasek SL, Schwartz J, J. Am. Chem. Soc 2007, 129, 6980–6981. [DOI] [PubMed] [Google Scholar]
  • [31].Chen T, Dou J-H, Yang L, Sun C, Oppenheim JJ, Li J, Dincă M, J. Am. Chem. Soc 2022, 144, 5583–5593. [DOI] [PubMed] [Google Scholar]
  • [32] a).Kresse G, Furthmüller J, Comput. Mater. Sci 1996, 6, 15–50; [Google Scholar]; b) Schimka L, Harl J, Kresse G, J. Chem. Phys 2011, 134, 024116; [DOI] [PubMed] [Google Scholar]; c) Kresse G, Furthmüller J, Phys. Rev. B 1996, 54, 11169–11186; [DOI] [PubMed] [Google Scholar]; d) Kresse G, Hafner J, Phys. Rev. B 1993, 47, 558–561; [DOI] [PubMed] [Google Scholar]; e) Kresse G, Joubert D, Phys. Rev. B 1999, 59, 1758–1775; [Google Scholar]; f) Perdew JP, Ruzsinszky A, Csonka GI, Vydrov OA, Scuseria GE, Constantin LA, Zhou X, Burke K, Phys. Rev. Lett 2008, 100, 136406. [DOI] [PubMed] [Google Scholar]
  • [33] a).Day RW, Bediako DK, Rezaee M, Parent LR, Skorupskii G, Arguilla MQ, Hendon CH, Stassen I, Gianneschi NC, Kim P, Dincă M, ACS Cent. Sci 2019, 5, 1959–1964; [DOI] [PMC free article] [PubMed] [Google Scholar]; b) Zhang J, Zhou G, Un H-I, Zheng F, Jastrzembski K, Wang M, Guo Q, Mücke D, Qi H, Lu Y, Wang Z, Liang Y, Löffler M, Kaiser U, Frauenheim T, Mateo-Alonso A, Huang Z, Sirringhaus H, Feng X, Dong R, J. Am. Chem. Soc 2023, 145, 23630–23638. [DOI] [PubMed] [Google Scholar]
  • [34] a).Uzunova EL, Göltl F, Kresse G, Hafner J, J. Phys. Chem. C 2009, 113, 5274–5291; [Google Scholar]; b) Landry VK, Parkin G, Polyhedron 2007, 26, 4751–4757. [Google Scholar]
  • [35] a).Smits FM, The Bell System Technical Journal 1958, 37, 711–718; [Google Scholar]; b) Dudarev SL, Botton GA, Savrasov SY, Humphreys CJ, Sutton AP, Phys. Rev. B 1998, 57, 1505–1509; [Google Scholar]; c) Li J, Lu Y, Ye Q, Cinke M, Han J, Meyyappan M, Nano Lett. 2003, 3, 929–933; [Google Scholar]; d) Wang L, Maxisch T, Ceder G, Phys. Rev. B 2006, 73, 195107; [DOI] [PubMed] [Google Scholar]; e) Grimme S, Antony J, Ehrlich S, Krieg H, J. Chem. Phys 2010, 132, 154104; [DOI] [PubMed] [Google Scholar]; f) Ammu S, Dua V, Agnihotra SR, Surwade SP, Phulgirkar A, Patel S, Manohar SK, J. Am. Chem. Soc 2012, 134, 4553–4556; [DOI] [PubMed] [Google Scholar]; g) Mahmood J, Kim D, Jeon I-Y, Lah MS, Baek J-B, Synlett 2013, 24, 246–248; [Google Scholar]; h) Akamatsu T, Itoh T, Izu N, Shin W, Sato K, in Sensors, Vol. 15, 2015, pp. 8109–8120; [DOI] [PMC free article] [PubMed] [Google Scholar]; i) Cheng L-C, Wu M-R, Huang C-Y, Juang T-K, Liu P-L, Horng R-H, ACS Appl. Electron. Mater 2019, 1, 253–259; [Google Scholar]; j) Lukatskaya MR, Feng D, Bak S-M, To JWF, Yang X-Q, Cui Y, Feldblyum JI, Bao Z, ACS Nano 2020, 14, 15919–15925; [DOI] [PubMed] [Google Scholar]; k) Qiu J, Hu X, Min X, Quan W, Tian R, Ji P, Zheng H, Qin W, Wang H, Pan T, Cheng S, Chen X, Zhang W, Wang X, ACS Appl. Mater. Interfaces 2020, 12, 19755–19767. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Accepted Supporting Information
NIHMS2126851-supplement-Accepted_Supporting_Information.pdf (5.5MB, pdf)

RESOURCES