Employing Triphenylene-based, Layered, Conductive Metal–Organic Framework Materials as Electrochemical Sensors for Nitric Oxide in Aqueous Media

Abstract

This paper describes the first use of conductive metal–organic frameworks as the active material in the electrochemical detection of nitric oxide in aqueous solution. Four hexahydroxytriphenylene (HHTP)-based MOFs linked with first-row transition metal nodes (M = Co, Ni, Cu, Zn) were compared as thin-film working electrodes for promoting oxidation of NO using voltametric and amperometric techniques. Cu and Ni- linked MOF analogs provided signal enhancement of 5–7 fold over a control glassy carbon electrode (SA_NO = 6.7 ± 1.2 and 5.7 ± 1.1 for Ni₃(HHTP)₂ and Cu₃(HHTP)₂, respectively) for detecting micromolar concentrations of NO. Zinc-based MOF electrodes offered more limited enhancement (SA_NO = 3.1 ± 0.5), while the Cobalt-based MOF analog had intrinsic redox activity at potentials close to NO oxidation, which interfered with sensing. Combining MOFs with a conductive polymer improved electrode stability under repeated electrochemical scanning (14 ± 3% decrease in signal over 10 scans). The stabilized Ni₃(HHTP)₂@polymer-coated electrodes were able to detect NO at physiologically relevant concentrations (LOD = 9.0 ± 4.8 nM) in amperometric sensing experiments, and exhibited moderate selectivity against ascorbic acid and nitrite (log k_j,NO = −1.3 ± 0.3 and −0.83 ± 0.68 for ascorbic acid and nitrite, respectively). This study demonstrates that layered, conductive 2D MOFs have promising applicability for NO detection in aqueous environments.

Graphical Abstract

As a gaseous signaling molecule, also known as a gasotransmitter,^1,2 nitric oxide participates in many signaling pathways in the human body, some of which are well characterized,^3–5 while others are less well-understood.^6,7 NO is implicated in many current health challenges, including gut health and digestion,⁸ neurodegenerative diseases,⁹ depression and anxiety,¹⁰ and the immune response to viral infections.^11,12 Electrochemical sensing of NO has established utility in rapid detection, providing dynamic information about the role nitric oxide is playing in various physiological processes.^2,13,14 Miniaturized sensors for nitric oxide incorporated into wearable or insertable devices have the potential to offer highly time- and location- specific information about nitric oxide production and signaling in animal models as well as for humans in next generation medical devices.^15–18

The detection of nitric oxide has greatly benefitted from the development of nanomaterials capable of enhancing the sensitivity and selectivity of NO detection in aqueous solution and in single cell environments.^19–21 Electrocatalysts, such as metal nanoparticles²² and metallophthalocyanines²³ can selectively promote NO oxidation, leading to enhanced sensor response. Nickel-based coordination complexes have previously shown particular sensitivity in electrochemical nitric oxide detection.^21,23–25 The implementation of polymer coatings and semipermeable membranes, such as poly(o-phenylenediamine), poly(eugenol) and Nafion, has been established as a promising strategy to enhance sensor selectivity.^14,26,27 While significant progress in the detection of NO has been achieved, providing insight into the production and transport of NO in biological samples,^21,28 the established materials for sensitive and selective detection of NO in complex environments have key limitations, including complex top-down fabrication strategies,²⁹ architectures that are not amenable to miniaturization,³⁰ brittleness or lack of flexibility,²⁰ and lack of long-term stability.¹⁴

Layered conductive metal–organic frameworks hold promise as electrode materials in electrochemical sensors³¹ that make them good candidates as electrocatalytically-active working electrodes in electrochemical NO sensors. These materials have been previously employed for the detection of NO in the gas phase,^32,33 and the detection of neurochemicals, glucose, and other biologically relevant organic analytes in the liquid phase,^31,34–37 but have not been previously examined for the detection of NO in aqueous media. These self-assembling, nanostructured materials with high porosity possess a large electrochemically active surface area for interactions with NO.^31,38,39 A unique feature of these materials is that their regularly spaced metal center active sites embed the functionality of transition-metal complex electrocatalysts directly into conductive porous scaffolds, bridging the gap between molecular components and functional devices.^33,40,41

This study compares the use of four layered conductive MOFs based on the 2,3,6,7,10,11-hexahydroxytriphenylene (HHTP) ligand (M₃(HHTP)₂, M = Co, Ni, Cu, Zn) as working electrode materials for the liquid-phase electrochemical detection of nitric oxide. We employ HHTP-based MOFs as electrochemical NO sensors for the first time, and seek to understand the role of the metal node in electrochemical sensing interactions. The triphenylene-based MOFs used in this study were selected based both on synthetic accessibility and previous use as working electrode materials for electrochemical sensing.^{34,35,41–43} Nickel-based macrocycle materials in particular have demonstrated electrocatalytic activity in nitric oxide sensors,^21,24 and in comparisons of metallophthalocyanines containing first-row transition metals, the metal center identity has been shown to have a large influence on NO binding and oxidation.²⁵ The series of M₃(HHTP)₂ MOFs featuring cobalt, nickel copper and zinc offers a similar opportunity for comparison of the electronic contributions to NO electrocatalytic activity, and also allows for comparisons between two possible MOF structures, one with intercalated monomers between layers, and one without (see Fig. 1a). We used cyclic voltammetry to qualitatively asses the NO sensing performance of the series of HHTP-based MOFs, similar to previous reports employing layered conductive MOFs for electrochemical sensing.^34,36,42 We also employed CV to compare the efficacy of MOF-polymer combinations, using the decrease in NO oxidation peak current as a metric for assessing the stability of the MOF film on the electrode. To quantitatively evaluate NO sensitivity of the optimized Ni₃(HHTP)₂-functionalized electrode, we utilized constant potential amperometry (CPA) in a dual electrode architecture. This two-electrode approach allowed for simultaneous monitoring of the same solution with a MOF electrode and a control electrode, ideal for assessing the enhancements to sensitivity and selectivity offered by the MOF.

Figure 1:

Structure and characterization of HHTP MOFs. (a) Scheme showing the two reported packing modes for layered, conductive MOFs based on the HHTP ligand coordinated with Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ ions. (b-c) Characterization of HHTP MOFs with powder X-ray diffraction. Diffraction patterns for each MOF are shown in reference to simulated patterns (generated from crystal structures using Mercury software) of with intercalated layer structure (ABAB stacking) and without intercalated layered structure (AAAA stacking). Corresponding Bragg planes are labeled on the plots. (d-g) SEM images of the MOFs used in this study.

This study focuses on a fundamental proof-of-concept demonstration using 2D layered conductive MOFs as voltammetric sensors for the detection of NO in aqueous solution. The comparison of four structurally similar MOFs enabled the comparison of structure–property relationships, and correlation of the sensing performance with the electronic and structural contributions of these materials to their electrochemical behavior. By varying the metal node of the MOF, we demonstrated its role as the putative binding site for MOF-NO interactions, and shown that the nickel-based analog has particular sensitivity to nitric oxide, enhancing NO oxidation signal 5–7 fold over control materials. While the stability of MOF films was limited at the high potentials required for NO oxidation (> 0.9 V vs. Ag/AgCl), MOF films were successfully stabilized on glassy carbon electrodes with the use of a PEDOT:PSS adhesive layer. Ni₃(HHTP)₂@PEDOT:PSS-coated electrodes detected NO at nanomolar concentrations in aqueous solution (LOD = 9.0 ± 4.8 nM), and exhibiting modest selectivity against ascorbic acid (k_AA,NO = −1.3 ± 0.3) pointing the way toward MOF-based devices for NO detection in real samples.

EXPERIMENTAL METHODS

Synthesis of Triphenylene-based Metal–Organic Frameworks:

Co₃(HHTP)₂, Ni₃(HHTP)₂, and Cu₃(HHTP)₂ MOF powders were synthesized hydrothermally according to previously reported methods^42,44 (See Figure S1 and Section II of the Supporting information for additional details. Zn₃(HHTP)₂ was synthesized hydrothermally as follows: Hexahydroxytriphenylene (0.03 mmol, 9.7 mg) was suspended in water with 150 mg (1.8 mmol) sodium acetate. Zinc chloride (0.045 mmol, 6.1 mg) was added, and the mixture was heated to 70 °C for 24 hours. The MOF was collected via vacuum filtration and washed with water and ethanol. The MOF was activated by soaking in ethanol followed by acetone.

General Preparation of HHTP MOF Electrodes:

3 mm. dia. glassy carbon working electrodes were polished using 0.05 μm alumina slurry and rinsed with water and ethanol prior to each use. HHTP MOF suspensions were prepared by mixing 5 mg of MOF powder with 1 mL of water, and sonicating for at least 5 hours. Electrodes were prepared by drop-casting 2– 2.5 μL drops of suspension onto a 3 mm dia. glassy carbon electrode. Electrodes were air dried for 30 min – 1 hr after each drop, and used on the same day as prepared. See Section IV of the Supporting Information for the preparation of electrodes with polymer coatings, binders, and adhesives, as well as the preparation of Ni₃(HHTP)₂-coated screen-printed electrodes.

Preparation of Nitric Oxide Solutions:

Dilute nitric oxide solutions were generated from a 1% Nitric oxide (nitrogen balance gas) tank. 1% NO was bubbled first through a vial containing 100 mL 1 M NaOH, then through a sealed vial containing degassed 0.1 M phosphate-buffered saline (PBS, pH 7.4). Gas was bubbled from the tank (4 psi outlet pressure) for 1 hour, after which the solution was stored in the fridge for a maximum of 1 day. The theoretical concentration of this solution was calculated based on Henry’s law, and was determined to be 17 μM (see Equations. S1–S3). For concentrations of 1.7 μM and lower, aliquots of this stock solution were added to degassed 0.1 M PBS in the electrochemical cell. See Section VI of the Supporting Information for the preparation of saturated nitric oxide solutions, and Section VIII of the Supporting Information for the calculation of stock solution concentrations.

Electrochemical Experiments:

All electrochemical experiments with glassy carbon electrodes used Pt wire counter electrodes and Ag/AgCl electrodes filled with 1 M KCl. In a typical electrochemical experiment, 0.1 M PBS was deoxygenated by bubbling N₂ through the electrochemical cell, then preconditioning scans were run. Following preconditioning scans, analyte was added to the desired concentration, after which more scans were run. All cyclic voltammetry scans were run at 50 mV/s unless otherwise noted. All CV experiments began with a 60s equilibration period at 0 V vs. Ag/AgCl prior to scanning.

Amperometry experiments with glassy carbon electrodes were set up as above, but with an additional working electrode inserted into the solution. Typical amperometry experiments were preceded by CV preconditioning scans at 50 mV/s (10 scans each working electrode) in 0.1 M PBS. After that, electrodes were set to a constant potential and equilibrated while stirring for at least 1 hr. Then, successive additions of NO stock solution were added. See Section VII of the Supporting information for additional details about electrochemical experiments, including experiments with screen printed electrodes.

RESULTS AND DISCUSSION

Characterization of HHTP MOFs:

We used established methods of MOF preparation to generate microcrystalline powders of M₃(HHTP)₂ materials (M = Co, Ni, Cu, Zn). Powder X-ray diffraction (PXRD) confirmed the crystalline structure of the powdered HHTP MOF samples (Figure 1b–c), which matched previous reports.^33,45,46 Cobalt- and nickel- HHTP MOFs exhibited a diffraction pattern consistent with the ABAB stacking pattern of alternating MOF and intercalated layers.⁴⁵

Copper- and zinc- HHTP MOFs exhibited a diffraction pattern consistent with the stacking of MOF layers without the presence of the intercalated layer (Figure 1a).⁴² The similarity between the experimental PXRD patterns of Cu₃(HHTP)₂ and Zn₃(HHTP)₂ (Figure 1c) suggests that the stacking of Zn₃(HHTP)₂ layers more closely resembles the slipped parallel packing mode previously reported by Chen et al.⁴⁷ rather than the staggered packing previously reported by Choi et al.⁴⁸ Additionally, the measured BET surface area for Zn₃(HHTP)₂ (360 m²/g, see Figure S2) was similar to those measured for Cu₃(HHTP)₂ and Ni₃(HHTP)₂ (405 and 360 m²/g, respectively), suggesting that the packing of Zn₃(HHTP)₂ is similar to that of Cu₃(HHTP)₂. As illustrated by scanning electron microscopy (SEM), all of the MOFs in this study showed a generally rod-like morphology, although the Cu₃(HHTP)₂ and Co₃(HHTP)₂ nanorods were larger in diameter (~0.5 μm) than the Zn₃(HHTP)₂ and Ni₃(HHTP)₂ nanorods (50–100 nm) (Figure 1d–g, Figures S3–S6). PXRD spectra and SEM images of dropcast HHTP MOF films confirmed that the MOFs retained their crystalline structure and rod-like morphology after 5 hrs of sonication (Figures S7–11). Synthetic details for the M₃(HHTP)₂ MOF series and details of the preparation of MOF-functionalized electrodes can be found in Sections II and IV of the Supporting Information.

Comparison of Nitric Oxide Detection Performance of HHTP-based MOFs:

In order to evaluate the NO sensing capabilities of layered, conductive MOFs, we prepared MOF-modified-glassy carbon electrodes by dropcasting thin films of MOFs onto the surface of glassy carbon electrodes (see Supporting Information, Section IV, for more details). We then used cyclic voltammetry to observe the oxidation of nitric oxide at the electrode surface in a saturated NO solution (1.9 mM at 25 °C), as well as a dilute solutions (17 μM & 1.7 μM). All MOFs in this study showed an oxidation peak for nitric oxide in the region of 1.0–1.1 V (vs. Ag/AgCl reference electrode, the electrode setup for CV studies can be seen in Fig. S11). The reduction of NO is expected to occur below −0.7 V vs. Ag/AgCl and was not observed in CV scans.⁴⁹ Figures 2a–d show representative cyclic voltammograms for all MOFs in 17 μM NO solution (additional replicate CVs are shown in Figure S14).

Figure 2:

Representative cyclic voltammograms of MOF-modified electrodes in 17 μM nitric oxide solution (electrolyte = 0.1 M PBS buffer, pH 7.4). Nitric oxide sensing using (a) Co₃(HHTP)₂ (b) Ni₃(HHTP)₂, (c) Cu₃(HHTP)₂, and (d) Zn₃(HHTP)₂ dropcast onto glassy carbon electrodes. Solid lines represent the 2^nd scan in the presence of 17 μM NO solution, dashed lines represent the 2^nd scan in 0.1 M PBS. Peak potentials and oxidation peak currents are indicated on each plot. All scans were run at 50 mV/s. (e-f) NO oxidation signal enhancement over bare glassy carbon for MOF electrodes in (e) 1.7 μM NO solution and (f) 17 μM NO solution. Error barsrepresent the standard deviation of (e) three trials and (f) six trials.

To compare the sensing performance of the MOFs tested, we calculated the signal enhancement for each MOF-electrode in reference to bare glassy carbon (see Equation S4) based on the 2^nd scan in NO solution. In saturated solution, all four MOFs offered a modest (less than 2x) enhancement to the peak current for NO oxidation over bare glassy carbon (Figure S15). Cu₃(HHTP)_2, Ni₃(HHTP)₂, and Zn₃(HHTP)₂ showed marginally greater enhancement than Co₃(HHTP)₂, but the difference was not statistically significant (p<0.05). At lower (17 μM) concentrations of NO, Co₃(HHTP)₂ electrodes showed the onset of water splitting around 1.1 V, as well as intrinsic redox activity in the 0.9 V region, indicating that this MOF is not well suited for the electrochemical nitric oxide detection. By contrast, the intrinsic redox activity of Cu₃(HHTP)₂ between −0.2 V and 0 V (vs. Ag/AgCl ref.), was outside the electrochemical window relevant to NO detection, thus making it a feasible option for electrochemical NO sensing. In dilute (17 μM) NO solution, Cu₃(HHTP)₂ and Ni₃(HHTP)₂ offered significantly higher signal enhancement for nitric oxide (SA_NO = 6.7 ± 1.2 and 5.7 ± 1.1 for Ni₃(HHTP)₂ and Cu₃(HHTP)₂, respectively) than that of Co₃(HHTP)₂ or Zn₃(HHTP)₂ (only 2–3x increase over GCE, Figure 2f). Similar trends in signal enhancement were observed when the HHTP MOF electrodes were used to detect lower concentrations of NO (1.7 μM), with Cu₃(HHTP)₂ and Ni₃(HHTP)₂ offering 5–6x increase in signal over glassy carbon and Co₃(HHTP)₂ or Zn₃(HHTP)₂ offering a 1–2x increase (Figures 2e, S16). The signal enhancement in dilute NO solution suggest that these materials are well suited for detecting NO at physiologically relevant (nanomolar-micromolar) concentrations. We also conducted scan-rate dependent studies to assess the relative electrochemical surface areas (ECSA) of the MOF electrodes (Figure S17, Table S1). The relative trend in ECSA for the four MOFs (Zn>Co=Ni>Cu) did not match the trend in NO sensing signal enhancement (Ni=Cu>Zn>Co). This suggests that the observed enhancements to NO sensing performance is related to the specific metal center of the MOF, rather than the increased surface area of the MOF-coated electrodes. Details of electrochemical experiments can be found in Section VII of the Supporting Information.

While the pure MOF samples showed initially high responses to NO, the signal intensity in electrochemical measurements showed a decrease in NO oxidation signal over the course of 25 scans (Figure S18). After cyclic voltammetry experiments, delamination of these MOF films was evident upon visual inspection of the electrodes. This decrease in signal was most pronounced for Ni₃(HHTP)₂, but was also observed with Zn₃(HHTP)₂ and Cu₃(HHTP)₂, both in NO sensing and through decreases in the intrinsic redox peaks for the MOFs (see Figure S19). This delamination was consistent across electrodes prepared with the same suspension, and across multiple batches of MOF suspension generated from the same sample of MOF powder (Figures S20–S21). The sonication time and aging of the suspensions also impacted this delamination, with films cast from aged (3 months - 1 year) or insufficiently sonicated (<4 hrs sonication time) MOF suspensions visibly flaking into the electrochemical cell upon insertion of the electrode into the cell (see Figures S22–S23), suggesting that MOF adhesion to the electrode surface was linked to the homogeneity of the suspension. We attribute differences in delamination between MOF samples to size and morphology differences in the dropcast films. We also attribute this delamination to the challenging electrochemical requirements associated with the detection nitric oxide oxidation at potential of around 1.0 V. While HHTP-based MOFs have proven to be stable on glassy carbon electrodes for the detection of analytes such as dopamine (0.3 V),^35,42 serotonin (0.5 V)⁴² and urea (0.45 V),⁵⁰ extending the adhesion of this class of MOFs to withstand high applied potentials over repeated cycles in aqueous solution proved to be challenging. This limited stability may be due in part to the changing surface chemistry of the glassy carbon electrode itself at potentials above 1.0 V,⁵¹ which in turn impacts the adhesion of the MOF to the electrode surface.

Stabilization of MOF films with polymers:

In an attempt to stabilize the MOF films, we combined the most promising MOF material, Ni₃(HHTP)₂, with various polymer coatings, binders, and adhesives (see Figure 3a, Table S2), and evaluated the NO sensing performance of the composite Ni₃(HHTP)₂/polymer electrodes with cyclic voltammetry. We assessed the film stability by comparing NO signal enhancement for the 1^st and 10^th scan from −0.7 − +1.2 V vs. Ag/AgCl (Figures 3b, S24–S25). We attempted to enhance MOF-film stability by coating pure MOF films with NO selective polymers Nafion and poly(o-phenylenediamine) (PPD). These polymers were selected for their precedent in electrochemical sensors for nitric oxide.¹⁴ We compared the NO sensing performance of PPD and Nafion coated both on Ni₃(HHTP)₂ electrodes and bare glassy carbon electrodes (Figure 3b). Coating Ni₃(HHTP)₂ in Nafion (Figures 3b, S24) decreased the NO oxidation signal slightly, but resulted in a stable film. By contrast, coating the electrode with PPD (Figures 3b, S24) decreased the signal below that of a bare glassy carbon electrode (SA_NO < 1). We sought to further optimize the Nafion@Ni₃(HHTP)₂ electrode by comparing the performance of Ni₃(HHTP)₂ electrodes dropcast with varying amounts of Nafion. Nafion solutions of 0.25 – 1.0 wt % were prepared and dropcast onto Ni₃(HHTP)₂-functionalized electrodes, which were tested in 17 μM NO solution (Figure S26). Higher concentrations of Nafion (1%, 0.75%, 0.5%) decreased the sensitivity of the electrode to close to that of a bare glassy carbon electrode (SA_NO ≈ 1), but did result in a stable film. By contrast, the 0.25 % coating resulted in much less signal reduction, but did not stabilize the MOF film (Figure S26).

Figure 3:

The impact of polymer coatings, binders and adhesives on the NO oxidation peak current of Ni₃(HHTP)₂. (a) Schematic of MOF-functionalized electrode showing the use of polymers either as a coating dropcast on MOF layer, as a binder mixed with MOF layer, or as an adhesive underneath the dropcast MOF layer. (b-d) NO signal enhancement and film stability for (b) polymer coating@Ni₃(HHTP)₂ electrodes, (c) Ni₃(HHTP)₂/polymer binder electrodes and (d) Ni₃(HHTP)₂@polymer adhesive electrodes in 17 μM NO solution. Signal enhancements are shown for 1^st and 10^th scans, as a metric for film stability. Bars represent average signal enhancement + SD (n = 3) Abbreviations: PPD = poly(o-phenylenediamine), PVF = polyvinyl fluoride, PMMA = poly(methyl methacrylate), PVP = polyvinylpyrrolidone PDA = polydopamine.

In an alternate approach, we combined Ni₃(HHTP)₂, with various polymer binders (see Table S1), and evaluated the NO sensing performance of the Ni₃(HHTP)₂ /polymer electrodes. In general, the polymer binders failed to enhance Ni₃(HHTP)₂ film stability (Figures 3c, S27–S28). Additionally, all but one of the polymer binders decreased the overall NO sensitivity of the electrode. We attribute these decreases primarily to a reduction in conductivity of the MOF/polymer composite film, which was also observed in sample charging in SEM images of the MOF/polymer films (Figure S28). The exception to this was the conductive polymer PEDOT:PSS, which did not decrease the overall signal, but did greatly increase background current (Figure S27a).

We also tested PEDOT:PSS and polydopamine (PDA) as adhesive layers underneath the MOF film (Figures 3d, S29–S30). Polydopamine is a bioinspired polymer developed as an adhesive coating^52–54 and PEDOT:PSS is commonly used as a base electrode layer in flexible sensors.⁵⁵ These layers were successful in stabilizing the MOF film while maintaining moderate signal enhancement for nitric oxide. The success of this approach suggests the role of surface roughness in the adhesion of dropcast MOF films to electrode surfaces. The use of PEDOT:PSS as an adhesive layer under Ni₃(HHTP)₂ decreased the NO oxidation peak current relative to pristine Ni₃(HHTP)₂ or Ni₃(HHTP)₂ with PEDOT:PSS as a binder (Figure 4a–d). However, this approach was still a modest enhancement over bare glassy carbon, and the adhesive layer minimized the signal decay over the course of 10 scans (Figure 4e). A single electrode was also reused in the same NO solution three times, and showed a slight decrease in sensitivity between each re-use (Figure S31)

Figure 4:

The stabilization of Ni₃(HHTP)₂ films on glassy carbon electrodes by a PEDOT:PSS adhesive layer. (a-d) Cyclic voltammograms of 17 μM nitric oxide solution (in 0.1 M PBS, pH 7.4) with (a) Ni₃(HHTP)₂@GCE working electrode, (b) Ni₃(HHTP)₂@PEDOT:PSS@GCE working electrode, (c) Ni₃(HHTP)₂+PEDOT:PSS@GCE working electrode, and (d) bare glassy carbon working electrode. Each plot shows the 1^st scan in NO solution (solid line) the 10^th scan in NO solution (dotted line) and the 10^th scan in PBS (dashed line). All scans were run at 50 mV/s. (e) Average peak currents (n=3) for each working electrode in 17 μM NO solution for the 1^st, 2^nd, 5^th, and 10^th scans. Error bars represent the standard deviation from the mean of 3 replicates. Average signal decay values (% decrease in peak current between the 1^st and 10^th scans) are shown above each bar. Error indicates the standard deviation from the mean of three replicates.

Amperometric detection of nitric oxide with Ni₃(HHTP)₂@PEDOT:PSS electrodes:

Having stabilized the MOF film, we sought to use the optimized Ni₃(HHTP)₂@PEDOT:PSS electrode to detect low concentrations of nitric oxide in amperometric sensing experiments. These experiments used a dual-working electrode set-up, with a Ni₃(HHTP)₂@PEDOT:PSS-functionalized electrode (WE1) and a bare glassy carbon control electrode (WE2) inserted into the cell (See Figure 5a inset and Figure S12). A multiplex-capable potentiostat was used to monitor the current at both working electrodes simultaneously, while aliquots of nitric oxide stock solution (17 μM) were added to the electrochemical cell (Figures 5a, S32). The MOF-functionalized electrode demonstrated the ability to detect low nanomolar concentrations of NO, with a calculated limit of detection (S/N =3) of 9 ± 5 nM in 0.1 M PBS buffer (vs. 26 ± 23 nM for the glassy carbon control, Figure 5b–c). The MOF electrode displayed a moderate enhancement in sensitivity to NO over the control electrode (408 ± 203 pA/nM vs. 297 ± 146 pA/nM for WE1 and WE2, respectively, Figure 5d). The signal enhancement of the Ni₃(HHTP)₂@PEDOT:PSS electrode over glassy carbon control was 1.3 ± 0.1, similar to the enhancement observed for this electrode in CV studies. The enhanced limit of detection for the MOF electrode is primarily driven by the low amount of noise in the signal from this electrode compared with the control electrode.

Figure 5:

Amperometric detection of nitric oxide (8.5 nM – 1.5 μM) with optimized Ni₃(HHTP)₂@PEDOT:PSS deposited on a glassy carbon electrode (WE1) and glassy carbon control electrode (WE2). (a-b) Amperometric sensing traces showing current response of each electrode to the addition of aliquots of 17 μM NO stock solution. (a) sensing trace of all additions from 8.5 nM to 1.5 μM. Inset shows dual-electrode cell setup. (b) Sensing trace of low concentration additions of NO solution from 8.5 – 85 nM. (c) Response vs. concentration curves for WE1 (Ni₃(HHTP)₂@PEDOT:PSS@GCE) and WE2 (Bare GCE). Calculated limits of detection for each electrode are shown on the plot. (d) Sensitivity (S_NO) pA/nM for each electrode, based on amperometry experiments. (e) Selectivity for each electrode, calculated from amperometric sensing data (1 – 50 μM) according to Equation S6. Error bars represent the standard deviation from the mean for three replicates.

We also evaluated the selectivity of the MOF electrode against ascorbic acid (AA) and nitrite (NO₂⁻), two common interferents in NO detection, often present at micromolar concentrations in physiological environments.^14,19 We tested the response of the MOF electrode (and glassy carbon control electrode) to concentrations of AA and NO₂⁻ ranging from 1 – 50 μM (see Figures S33–S34). We then used the sensitivity values obtained in these experiments to calculate the selectivity coefficients (log k_j,_NO) for each electrode and interferent (See Equation S6). The MOF-functionalized electrode offers a moderate enhancement in selectivity for NO against ascorbic acid over bare glassy carbon, and a similar level of selectivity against NO₂⁻ as the control electrode (See Figure 5e, Table S3). The selectivity coefficient of −1.3 ± 0.3 for WE1 to NO over AA indicates that the MOF-functionalized electrode is approximately 10-fold more sensitive to NO than to ascorbic acid. Although modest, the enhancements to sensitivity and selectivity for NO offered by the Ni₃(HHTP)₂ electrode show the promise of framework materials to serve as effective electrode materials for the detection of nitric oxide in physiological settings.

Nitric Oxide Sensing with Ni₃(HHTP)₂-functionalized screen-printed electrodes:

The enhanced stability of Ni₃(HHTP)₂ deposited on top of conductive polymers indicated the potential for stable Ni₃(HHTP)₂ sensors on top of rough surfaces such as porous polymers. To that end, we deposited Ni₃(HHTP)₂ onto carbon-paste working electrodes screen-printed on a polyethylene terephthalate (PET) substrate (see Supporting Information, Section XI, for more details). This miniature electrode design included two working electrodes (a primary electrode and control/blank electrode) a counter electrode (also carbon paste) and a silver-paste reference electrode (see Figure 6a inset, and Figure S35). This compact design, in which all four electrodes fit into a circular area approximately 3.5 mm in diameter, allows for NO detection in smaller volumes of solution (2 – 4 mL electrolyte, compared with 7 – 10 mL for glassy carbon electrode cells). The Ni₃(HHTP)₂ film demonstrated stability on the carbon paste electrode surface without a PEDOT:PSS adhesive layer (see Figure S36), indicating the suitability of the Ni₃(HHTP)₂@SPE electrode for amperometric detection of NO. In amperometric experiments with a constant potential of 0.95 V, the MOF-functionalized electrode responded to μL additions of nitric oxide solution (Figures 6a, S37) and had a limit of detection of 9 ± 11 nM in PBS buffer solution (the control electrode had an LOD of 20 ± 13 nM, Figure 6b). The calculated sensitivity of the screen-printed electrode for NO was lower than that of the Ni₃(HHTP)₂@PEDOT:PSS@glassy carbon electrodes, due to the smaller surface area. However, the signal amplification was similar (1.4 ± 0.2 -fold increase in sensitivity). The limits of detection were also comparable to those observed with the larger area glassy carbon electrodes.

Figure 6:

Amperometric detection of nitric oxide with a Ni₃(HHTP)₂-functionalized screen-printed electrode. (a-b) Amperometric sensing traces showing current response of each electrode to the addition of aliquots of 17 μM NO stock solution. (a) sensing trace of all additions from 8.5 nM to 1.5 μM. Inset shows layout of screen-printed electrode. (b) Sensing trace of low concentration additions of NO solution from 8.5 – 85 nM. (c) Response vs. concentration curves for WE1 (Ni₃(HHTP)₂@SPE) and WE2 (Bare SPE). Calculated limits of detection for each electrode are shown on the plot. (d) Sensitivity (S_NO) pA/nM for each electrode, based on amperometry experiments. (e) Selectivity for each electrode, calculated from amperometric sensing data (1 – 50 μM) according to eq. S6. Error bars represent the standard deviation from the mean for three replicates.

We also evaluated the selectivity of the screen-printed electrodes against NO₂⁻ and ascorbic acid (Figures 6e, S38, Table S4); the calculated selectivity coefficients were similar to those for the Ni₃(HHTP)₂@PEDOT:PSS@GCE electrodes. The detection of nanomolar concentrations of nitric oxide with this miniaturized sensing architecture demonstrated the potential of MOF-based electrodes to serve as electrochemical sensors in physiological conditions. The screen-printed electrode device is similar in size and form-factor to recent reports of NO sensor devices employed in in vitro settings.^55,56 The Ni₃(HHTP)₂@SPE electrodes exhibit comparable analytical performance (similar sensitivity and limits of detection) with recently reported electrode materials for electrochemical NO detection (See Table S5).

To demonstrate the potential for HHTP MOF-based electrodes to detect NO in a biomedically relevant setting, we evaluated the NO sensing performance of the Ni₃(HHTP)₂@SPE electrodes in simulated wound fluid (SWF). The fluid was generated by diluting fetal bovine serum (FBS) with PBS buffer according to previously reported procedures (see Section XII of the Supporting Information for additional details).^57,58 The Ni₃(HHTP)₂@SPE electrodes showed signs of biofouling in the presence of SWF, as evidenced by high current and irregularly-shaped CV scans of the electrode in the presence of SWF (Figure S39a,b). This phenomenon was lessened by further dilution of the SWF with PBS (Figure S39c,g). In dilute SWF (100x dilution of FBS in PBS), the screen-printed electrodes responded to micromolar concentrations of NO when aliquots of saturated NO solution were added to the SWF in amperometry studies (Figure S40, S41). The sensitivity of the electrodes was decreased relative to the NO sensing performance of the electrodes in PBS (S_NO = 18 ± 6 pA/nM in SWF vs. 86 ± 28 pA/nM in PBS for the Ni₃(HHTP)₂@SPE electrode). However, the signal enhancement of the MOF-coated electrode over the bare electrode was similar in SWF vs. PBS (SA_NO = 1.4 ± 0.4 in SWF vs. 1.4 ± 0.2 in PBS, se Table S6). As a consequence of the decreased sensitivity, the electrodes displayed micromolar LODs for NO in SWF (1.1 ± 0.7 μM and 0.7 ± 0.5 μM for WE1 and WE2, respectively). The electrodes displayed minimal response to NO in more concentrated preparations of SWF (Figure S42). The decreased responses in the presence of SWF indicate the need for further development of MOF-based electrodes to protect against biofouling in clinical or physiological settings.

CONCLUSIONS

This work represents the first demonstration of MOFs as electrochemical sensors for nitric oxide in the liquid phase. In this study, we have expanded the classes of analytes detected with layered, conductive framework materials in voltammetric sensing to include one of the dissolved gasotransmitters: NO. A comparison of a series of HHTP MOFs based on first-row transition metal ions Co²⁺, Ni²⁺, Cu²⁺, and Zn²⁺ provides fundamental insight into the role of the metal node as the binding site that facilitates NO oxidation. The identity of the metal center is the primary driver of the observed nitric oxide sensitivity of the HHTP-based MOF materials, with the nickel- and copper-based analogs showing greater response than the zinc- and cobalt analogs. By contrast, the layered structure of the MOF does not appear to significantly impact the response to NO, with Cu₃(HHTP)₂ (AAAA structure) and Ni₃(HHTP)₂ (ABAB structure) showing similar NO oxidation peak currents. This fundamental comparison supports the hypothesis that the metal centers in these materials are serving as the electrocatalytic sites for the oxidation of nitric oxide, with the sensitivity of Ni₃(HHTP)₂ and Cu₃(HHTP)₂ MOFs likely arising from a favorable binding interaction between NO and the metal-bis(dioxolene) host sites in the material. The general enhancement of sensitivity over glassy carbon offered by all the MOFs suggests that the increased surface area of the MOF-coated electrodes may be beneficial to electrochemical detection of analytes at low concentrations.

Our study also highlighted a limitation of HHTP-MOF films — their limited stability on a smooth glassy carbon electrode surface at high applied potentials (above 1 V vs. Ag/AgCl) under repeated electrochemical scanning. This limited stability presented challenges in evaluating the sensing performance of these materials towards NO. However, we have shown that a stable MOF film capable of NO detection can be fabricated by combining MOFs with polymer films, with a layer of PEDOT:PSS successfully acting as a conductive adhesive between the MOF film and the electrode surface. This optimized electrode was used in amperometric sensing experiments to detect NO at nanomolar concentrations at 1.0 V vs Ag/AgCl, with a limit of detection of 8.7 ± 4.9 nM for NO in PBS. The electrode also showed moderate selectivity against micromolar levels of two common interferents, ascorbic acid and nitrite. The limitations of MOF adhesion to electrode substrates could likely be overcome with direct growth and deposition of MOFs onto electrode substrates. Such methods have been shown to generate highly robust coatings with a high degree of control over particle size and morphology.^32,46,59 The optimization of these methods to deposition MOFs onto electrodes for liquid-phase detection is a ripe area for continued study.

Nevertheless, the incorporation of Ni₃(HHTP)₂ into a screen-printed electrode enabled the detection of NO at nanomolar concentrations, with similar limits of detection to the MOF on glassy carbon electrodes (9 ± 11 nM vs. 9 ± 5 nM) in a highly miniaturized format. These flexible devices are similar in size to recently reported NO sensors employed in NO detection of real samples both in vitro⁵⁶ and in vivo.^55,60 The electrodes maintained the ability to detect nitric oxide in the presence of simulated wound fluid, though the sensitivity of the electrodes was notably decreased. This demonstration suggests that with further development, layered conductive MOFs are promising electrode materials for NO detection that could be incorporated into a variety of flexible wearable or insertable devices for real-time nitric oxide sensing.

Supplementary Material

The Supporting Information is available free of charge on the ACS Publications website. The following files are available free of charge:

Additional experimental details, additional SEM and PXRD characterization for MOF samples and electrode films, and additional data from electrochemical sensing experiments. (PDF)