Aquotris(benzotriazole)sulfatocopper(II).benzotriazole Framework Assembled on Multiwalled Carbon Nanotubes through π-π Interaction for H₂O₂ Sensing in pH 7 Buffer Solution

1,2,3-benzotriazole (BTAH) and multiwalled carbon nanotube (MWCNT, >90% carbon basis, diameter: 110–170 nm; length: 5–9 μm) were purchased from Sigma Aldrich. Graphite fine powder was obtained from Lobachemie Pvt. Ltd., Mumbai. CuSO₄.5H₂O, hydrogen peroxide and tetra sodium pyrophosphate were procured from Rankem, India. L-Ascorbic acid, uric acid, dopamine and sodium nitrite were purchased from SRL, India and used without further purification. Glycerol was obtained from Fisher Scientific. Aqueous solutions were prepared using Millipore water (18 MΩ). Nitrogen saturated phosphate buffer solution (pH 7 PBS) of ionic strength, I = 0.1 mol L⁻¹ was used as supporting electrolyte in this study.

Cyclic voltammetric measurements were carried out with Auto Lab Potentiostat/Galvanostat 502 N Netherland. A three electrode setup involving either glassy carbon electrode (GCE) or its chemically modified electrode as a working electrode (5 mm diameter), Ag/AgCl as a reference electrode and platinum wire as a counter electrode was employed. The bioanalytical system (BAS, U.S.) polishing kit was used to polish the GCE surface. 5mm diameter GCE embedded rotating disk electrode (RDE, pine instruments) coated with BTAH-Cu²⁺ was also used to estimate the number of electrons involved during H₂O₂ reduction through Koutecky-Levich analysis.⁴² Quartz crystal microbalance (QCM) experiments were carried out using a gold coated quartz piezo electric crystal of geometric surface area of 0.38 cm²_. Piezoelectric crystal was controlled using Auto Lab Potentiostat/Galvanostat 502 N Netherland. Single crystal XRD study (Bruker Apex II) was performed to understand the structure and lattice parameters of BTAH-Cu²⁺ MOF. Hitachi S4800 high resolution field emission scanning electron microscope instrument was employed to understand the morphology and elemental distribution of the samples. FT-IR and Raman spectroscopy (1064 nm) analysis were carried out using JASCO 4100 and BRUKER RFS 27, respectively. Sample mixed KBr pellet was used in the FT-IR and Raman spectroscopy. Powder X-ray diffraction (PXRD) was performed on the samples using Bruker 8 fitted with Cu Kα X-ray source at a continuous scan rate of 1° 2Θ/min and step size of 0.02° 2Θ.

An aqueous solution (1 mL) of CuSO₄.5H₂O (0.001 mol) was added to an ethanol solution (1 mL) of benzotriazole (0.002 mol). After stirring for 30 min. at 25°C, a blue solution was obtained, which was filtered and the filtrate was evaporated slowly for one week. Blue colored crystals of BTAH-Cu²⁺ was obtained. A single crystal was chosen for crystal structure analysis.

The surface of the GCE electrode was cleaned both by mechanically and electrochemically. GCE was polished with 0.5 μm alumina powder, followed by washing with Millipore water and sonicating for 5 min in water. Cyclic voltammograms (CVs) of the polished electrode are obtained in the potential window of −0.2 to 1 V vs. Ag/AgCl at a scan rate of 50 mV s⁻¹ in pH 7 PBS. MWCNTs coated GCE was prepared by drop casting 5 μL of colloidal dispersion containing 2 mg MWCNTs in 500 μL ethanol, followed by air drying for 3 min at 25°C.^43–45 Thus obtained GCE/MWCNTs electrode was immersed for 1000 seconds in a PBS solution containing 3 mM of BTAH-Cu²⁺, for adsorbing the BTAH-Cu²⁺ on to the surface of the MWCNTs, to afford a surface modified electrode, GCE/MWCNTs@BTAH-Cu²⁺. To understand the redox activity of GCE/MWCNTs@BTAH-Cu²⁺ electrode, CV was done in pH 7 PBS, until a stable CV was realized, thereafter the electrode is ready for sensing the H₂O₂. Gold surface of the QCM crystal was coated with the 9.7 μL of colloidal dispersion of MWCNTs followed by air-drying, and is called as Au/MWCNTs electrode.

For the FT-IR, Raman and SEM analyses, as prepared GCE/MWCNTs@BTAH-Cu²⁺electrode was dried in a vacuum desiccator for overnight. MWCNTs@BTAH-Cu²⁺ film on the modified GCE electrode carefully peeled off from the GCE surface using a doctor's syringe and examined further.

QCM experiments were performed to understand the amount of BTAH-Cu²⁺ adsorbed on the MWCNTs, as the frequency change of the quartz crystal is related to the mass change as per Sauerbrey equation (Equation 1).⁴⁶

Where A is the area of the gold coated quartz crystal (0.38 cm²), ρ is the density of the crystal (2.648 g cm⁻³), k is the shear modulus of the quartz AT-cut crystal (2.947 × 10¹¹ g cm⁻¹ s⁻²), Δf is the measured frequency change, and f₀ is the oscillation frequency of the crystal (8 MHz).

Theoretical calculations were performed at the B97D level of theory with the 6–31G(d) basis set by employing the Gaussian 09 suit.⁴⁷ Furthermore the accuracy of B97D (D stands for dispersion) for describing the weak interactions has been recently established.⁴⁸ On the basis of previous theoretical calculations and considering the computational cost, we used C₅₄H₁₈ as a model of graphene sheet to predict π-π interactions between BTAH and C₅₄H₁₈.

Single crystal XRD study of BTAH-Cu²⁺ reveals formation of monoclinic crystal, space group p 2₁/c₁, with lattice parameters of a = 7.2927 4 Å, b = 19.7429 (14) Å, c = 19.2265 (12) Å, α = 90°, β = 97.054 3° and γ = 90°. Table I shows the detailed crystal structure data of the BTAH-Cu²⁺ [Cu(SO₄)(H₂O)(BTAH)₃]. Fig. 1 shows the structure of BTAH-Cu²⁺ complex. Figs. 2a and 2b show the a-axis and c-axis view of the unit cell of the BTAH-Cu²⁺. The crystal contains 5-coordinated BTAH-Cu²⁺ molecules having a tetragonal pyramidal symmetry, wherein Cu²⁺ is co-ordinated to the N₍₁₎ nitrogen of three benzotriazoles, a sulfate and a water molecule. A BTAH molecule acts as a bridge between two BTAH-Cu²⁺ molecules. Fig. 2c shows the framework of the H-bonded network present in the crystal viewed along 'a' axis and hence BTAH-Cu²⁺ is a MOF. Crystal structure of the BTAH-Cu²⁺ framework is reported earlier.⁴⁹

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Fig. 3A shows the change in frequency of the QCM crystal as a function of time while the Au/MWCNTs was let immersed in PBS solution containing 3 mM of BTAH-Cu²⁺. At the end of 1000 seconds, a frequency change of 120 Hz was obtained, which corresponds to a mass change of 1.475 μg of BTAH-Cu²⁺ complex. This experiment confirms adsorption of BTAH-Cu²⁺ onto MWCNTs, possibly due to strong π-π stacking. After adsorbing the complex from the PBS on to the Au/MWCNTs, the electrode was cycled 40 times in deaerated PBS solution. The last 20 cycles are shown in Fig. 3B. As one can see, the voltamograms are stable and reproducible from the 20^th cycle onwards (Fig. 3B). From the 40^th cycle data, the area under the anodic and cathodic peaks were calculated to be 5.41 × 10⁻⁴ (±0.10) C cm⁻² and 5.83 × 10⁻⁴ (±0.10) C cm⁻² respectively. The areas under the peaks were nearly equal suggesting that the peaks were belonging to the same redox couple, Cu²⁺/Cu⁺. The steep reduction current observed at around −1.0 V vs Ag/AgCl is due to H₂ evolution. From the area under the peaks, the amount Cu²⁺ present in the complex is estimated to be 2.299 × 10⁻⁹ moles (0.146 μg). Subtracting the mass of Cu, SO₄²⁻ and H₂O from the weight observed in the QCM measurement, the amount of BTAH calculated to be 9.285 × 10⁻⁹ moles (1.106 μg). The molar ratio of Cu²⁺ to the BTAH suggests that there are four BTAH ligands associated with one Cu²⁺. This result is in conformity with the single crystal data of BTAH-Cu²⁺.

Zoom In Zoom Out Reset image size

GCE@BTAH-Cu²⁺ electrode was prepared by dipping GCE for 1000 seconds in PBS containing 3 mM BTAH-Cu²⁺. CV of GCE@BTAH-Cu²⁺ is devoid of any redox signature in the potential range between −0.4 to 1.0 V vs. Ag/AgCl (Fig. 4A). Hence, presence of MWCNTs are essential for immobilizing the BTAH-Cu²⁺ in appreciable quantity. Reduction current observed below −0.4 V is due to H₂ evolution originating from the electrolysis of water. CVs (20^th-40^th cycles) of GCE/MWCNTs@BTAH-Cu²⁺ shows a quasi-reversible redox peak due to Cu²⁺/Cu⁺ redox couple (Fig. 4B) with ΔEp (peak to peak separation) of −311 mV and E° value of −182 mV vs Ag/AgCl. E° value was obtained from the peak potentials of the cathodic (E_pc) and anodic peaks (E_pa) as shown in Equation 2.

Zoom In Zoom Out Reset image size

Integration of the anodic and cathodic peak applying the polynomial correction (5^th order) to the baseline yielded charge values of 5.32 × 10⁻⁴ (±0.12) C cm⁻² and 5.70 × 10⁻⁴(±0.13) C cm⁻², respectively. Baseline corrected CV is shown as inset to Fig. 4B. As the loading of MWCNTs on the Au and GCE electrodes were same, the amount of charge density estimated from the peak area for Au/MWCNT@BTAH-Cu²⁺ and GCE/MWCNT@BTAH-Cu²⁺were same. This clearly indicates that the substrate used for this study (Au or GCE) do not have any influence on the amount of the BTAH-Cu²⁺ adsorbed on MWCNTs. Fig. 4C shows the CV of GCE/MWCNTs@Cu²⁺ and GCE/MWCNTs@BTAH controls. As one can see, there is no redox behavior in absence of either BTAH or Cu²⁺, indicating that the redox property exhibited in Fig. 4B is solely due to the BTAH-Cu²⁺ complex. The surface excess (Γ) calculated from the anodic peak area of the 40^th cycle of CV shown in Fig. 2B is 5.52 × 10⁻⁹ mol cm⁻² of Cu²⁺/Cu⁺-BTAH. The relative standard deviation (RSD) calculated from the 20^th–40^th CV is 2.7%, which is below the allowed limit of 5%. In order to confirm adsorption of complex by π-π interactions, we have explored the graphite (as it contains graphitic carbon) in place of MWCNTs and generated a GCE/GC@BTAH-Cu²⁺ electrode. It's CV in pH 7 PBS is shown in Fig. 4D. The calculated Γ value is (from the anodic peak area) 3.4 × 10⁻⁹ mol cm⁻². Γ value of GCE/GC@BTAH-Cu²⁺ is about 1.6 times lesser than that of the GCE/MWCNTs@BTAH-Cu²⁺cogener. This indicates that, presence of π-π interaction between graphitic carbon containing materials (graphite and MWCNTs) and BTAH-Cu²⁺. Surface confined CV response of the adsorbed species is considered as a proof of π-π interaction in the literature.⁵⁰

A linear increase in the peak currents (i_pa/i_pc) with the scan rate was observed (Figs. 5A and 5B) and it suggests presence of surface confined electron transfer reaction of Cu²⁺/Cu⁺ redox couple. Effect of solution pH (pH 3 to 11) on the peak potential of GCE/MWCNTs@BTAH-Cu²⁺ is shown in Fig. 5C. As BTAH-Cu²⁺ MOF is unstable in pH < 3 it is not explored at low pH solutions. A plot of E_pc versus solution pH (Fig. 5C) shows a linear line with a slope of −54 mV pH⁻¹ indicating a Nernstian type mechanism with the involvement of equal number of e⁻ and H⁺ during the reduction of Cu²⁺ into Cu⁺. Similarly, with the plot obtained for anodic process between pH 7 and 11 a slope of −64 mV pH⁻¹ was realized. However at pH < 7, the slope was about 21 mV pH⁻¹ indicating a deviation from the Nernstian type mechanism. This could be the reason behind the quasi-reversible CV obtained for GCE/MWCNTs@BTAH-Cu²⁺ in pH 7 PBS. BTAH is a base, it can attract proton from the solution at its N₍₁₎ position, hence a proton and electron mediated Cu²⁺/Cu⁺ redox reaction is justifiable. The single crystal XRD study confirms the bonding between N₍₁₎ nitrogen and Cu²⁺, and hence N₍₁₎ nitrogen is the best electron donor among the three nitrogen's in BTAH.

Zoom In Zoom Out Reset image size

Fig. 6 shows the PXRD pattern of graphite (curve a), graphite dipped in PBS (curve b) and graphite dipped in the BTAH-Cu²⁺ solution (GC/BTAH-Cu²⁺) (curve c). The peak at 2Θ value of 26.9° corresponds to (002) plane of graphite. The extra peaks in curve c at ∼7.56°, ∼19.83° and ∼31.52° are due to BTAH-Cu²⁺ complex adsorbed on graphitic carbon. Inset to Fig. 6c shows the indexed PXRD pattern of the BTAH-Cu²⁺ complex alone. The peaks observed in curve c are also present in the inset as well confirming the presence of BTAH-Cu²⁺ adsorbed on graphitic carbon in GC/BTAH-Cu²⁺.

Zoom In Zoom Out Reset image size

SEM examined surface morphology of GCE/MWCNTs@BTAH-Cu²⁺ system. As can be seen in SEM micrographs, both MWCNTs and MWCNTs@BTAH-Cu²⁺ (which is scrapped from the GCE/MWCNTs@BTAH-Cu²⁺ electrode) shows tube like structures with diameter ranging between 91–180 nm (Figs. 7A and 7B). EDS mapping (Fig. 7C) of MWCNTs@BTAH-Cu²⁺ shows the elements N, S, O and Cu all over the sample, indicating adsorption of BTAH-Cu²⁺ complex on the MWCNTs. The Raman spectra of MWCNTs and MWCNTs@BTAH-Cu²⁺ show D (∼1330 cm⁻¹) and G bands (∼1570 cm⁻¹) due to sp³ and sp² hybridized carbon respectively (Fig. 8).⁵¹ The intensity ratio of D and G bands (I_D/I_G) can be taken as a quantity of the disorder in the graphitic structure. I_D/I_G ratio of MWCNTs@BTAH-Cu²⁺ (1.02) was lower than that of the MWCNTs (1.39) as BTAH does not contain any sp³carbon leading to decrease in I_D/I_G ratio in MWCNTs/BTAH-Cu²⁺sample. CV, QCM, Raman and EDS analyses clearly confirm the π-π interaction between the BTAH-Cu²⁺ complex and MWCNTs leading to the adsorption of the former on MWCNTs. Geng et al. have modeled π-π interaction energies of indole (which contain NH group as in BTAH)-benzene complex, and found that the extension of aromaticity and the highly positive H atom of the N-H bond, both exhibited by Indole, enhance the strength of non-bonded interaction with benzene compared to those in the benzene dimer or in the pyridine-benzene complex.⁵²

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The interaction energies between the graphene and the BTAH were calculated using the following Equation 3,⁵⁰ wherein the energy was calculated using density functional theory (DFT).

Where E_complex, and E_BTAH correspond to the energies of the optimized C₅₄H₁₈ (graphene analogue)-BTAH complex, C₅₄H₁₈ and BTAH respectively.

The optimized geometry of the complex shows (Figs. 9A and 9B) that BTAH is at a vertical distance of 3.3 Å from the graphitic carbon analogue (C₅₄H₁₈). This distance is reasonable for the π-π stacking while comparing with previously reported distances of 3.0 to 3.6 Å for benzene derivatives and nucleobases adsorbed on graphene sheets.^53–55 The calculated E_int value is −18.76 kcal mol⁻¹, indicating π- π interaction. The predicted binding energy for benzene and naphthalene on graphene is −11.4 kcal mol⁻¹⁵³ and −17.6 kcal mol⁻¹,⁵³ respectively. For nucleobases the value ranges from −16 to −25 kcal mol⁻¹.^54,55 Possibly nitrogen atoms of the BTAH ring are as well contributing to the π-π interactions due to their close proximity (in comparison to carbon) towards graphene sheet.

Zoom In Zoom Out Reset image size

FT-IR spectra of MWCNTs@BTAH-Cu²⁺ and the controls (MWCNT, BTAH, BTAH-Cu²⁺ and MWCNTs@BTAH) are given in Fig. 10A. Fig. 10B shows the lower frequency range of FT-IR spectra. MWCNT shows –O-H, C=C and C-H stretching peak at 3500, 2878 and 1597 cm⁻¹ respectively. At low frequency region, low intensity peaks were observed which are possibly due to the radial breathing modes originating from the narrow inner most tubes of MWCNTs.⁵⁶ BTAH shows C-H and N-H stretching peaks between 2700–3500 cm⁻¹, C=C stretching at 1620 cm⁻¹ and N=N stretching at 1600 cm⁻¹. The peak at around 1200 cm⁻¹ is due to C-N stretching frequency. The peak at 750 cm⁻¹ possibly due to out of plane =C-H bending.⁵⁷ MWCNTs@BTAH, shows peak due to N=N stretching at 1621 cm⁻¹ and N-H stretching peak at 3436 cm⁻¹ which is overlapping with the O-H stretching observed with MWCNTs. FT-IR of MWCNTs@BTAH-Cu²⁺ shows a downshift in the N=N frequency from 1621 to 1518 cm⁻¹ possibly due to BTAH N₍₁₎ nitrogen coordinated with Cu²⁺. In addition, the Cu-N stretching frequency is clearly seen at around 440 cm⁻¹ confirming the interaction between BTAH and Cu²⁺ (Fig. 10B).⁵⁸

Zoom In Zoom Out Reset image size

Electrocatalytic activity of GCE/MWCNTs@BTAH-Cu²⁺ was examined with 500 μM H₂O₂ in pH 7 PBS. Fig. 11A shows the CV responses of GCE/MWCNTs@BTAH-Cu²⁺ in absence (curve a, N₂ gas purged) and presence of H₂O₂ (curve b) measured at a scan rate of 10 mV s⁻¹. Curve a shows the quasi reversible redox couple pertain to the BTAH-Cu²⁺ complex. Curve b, shows a huge cathodic peak with onset potential of 0.1 V vs Ag/AgCl corresponding to the H₂O₂ reduction. During the anodic scan, at around 0.4 V a sudden increase in current observed due to oxidation of H₂O₂. GCE/MWCNTs@Cu²⁺ (curve c) and GCE/BTAH-Cu²⁺ (curve d) show H₂O₂ reduction peak at a higher over potential than GCE/MWCNTs@BTAH-Cu²⁺. In case of GCE/MWCNTs@BTAH-Cu²⁺, BTAH-Cu²⁺ complex mediated the H₂O₂ reduction effectively at a lower overpotential than controls (GCE/BTAH-Cu²⁺ and GCE/MWCNT@Cu²⁺). MWCNTs provide the required surface area to adsorb BTAH-Cu²⁺ and electrical conductivity to the electrode. Ten repetitive measurement with H₂O₂, resulted in a relative standard deviation (RSD) of 3.01%, which indicated the good stability of GCE/MWCNTs@BTAH-Cu²⁺ towards H₂O₂ reduction reaction (data not shown).

Zoom In Zoom Out Reset image size

The electrocatalytic function of MWCNTs@BTAH-Cu²⁺ is further analysed using RDE technique, where a glassy carbon electrode of RDE was modified with MWCNTs@BTAH-Cu²⁺. Using the background corrected current responses obtained at −0.1, −0.2, −0.3 and −0.4 V. Koutecky-Levich analyses were performed in pH 7 PBS medium (Fig. 11B). Equations 4 and 5 show the Levich and Koutecky-Levich (KL) equations.

Where i, i_k and i_L are the net current, kinetic current and mass transfer limited current respectively. ω = rotation per minute (RPM), C_o* = H₂O₂ concentration, n = number of electron transferred during the analyte reduction, F = Faraday constant, A = Area surface of the electrode, D = diffusion coefficient (1.43 × 10⁻⁵ cm² s⁻¹)^5,9 ν = kinematic viscosity (0.9 × 10⁻² cm² s⁻¹).⁵⁹ The KL lines measured at −0.1 V, −0.2 V, −0.3 V and −0.4 V vs. Ag/AgCl were parallel to the theoretical 2e⁻ KL line confirming 2e⁻ reduction of H₂O₂ at these potentials. As these lines are parallel, the mechanism of H₂O₂ reduction on the electrode is same at these potentials. This result indicates that the overall electrocatalytic reduction of H₂O₂ follows two electron transfer process on GCE/MWCNTs@BTAH-Cu²⁺ modified electrode. The overall electrocatalytic reduction mechanism of H₂O₂ on BTAH-Cu²⁺ MOF is shown in Equation 7.

Further, GCE/MWCNTs@BTAH-Cu²⁺ electrode was subjected to chronoamperometry sensing of continuous spiking of 20 μM of H₂O₂ at fixed potential of −0.1 V vs Ag/AgCl while stirring the solution at 200 rpm with rice pellet (Fig. 12A, curve a). A systematic increase in the current signals against H₂O₂ spikes were noticed. Whereas the BTAH free electrode (GCE/MWCNTs@Cu²⁺) (Fig. 12A, curve b) under the similar working condition showed tiny increment in current signals with the addition of H₂O₂. Constructed calibration plot (inset plot in Fig. 12A) for GCE/MWCNTs@BTAH-Cu²⁺ was linear in the concentration range 20–200 μM of H₂O₂ with regression coefficient and sensitivity values of 0.9971 and 0.03 μA μM⁻¹ (0.153 A M⁻¹ cm⁻²), respectively. The error bars are introduced using the responses from the three freshly prepared electrodes. The calculated detection limit (Signal to noise ratio = 3) is 3.3 μM. The obtained detection limit value is compared with other reported H₂O₂ sensors in pH 7 medium (Table II). Interference due to various chemicals such as ascorbic acid (AA), uric acid (UA), dopamine (DA), and nitrite (NO₂⁻) was tested on the working electrode at a fixed potential of −0.1 V vs Ag/AgCl (Fig. 12B). Except DA, which showed weak current signal, no other chemicals interfere with the H₂O₂ sensing. These observations indicate the selective H₂O₂ sensing of GCE/MWCNTs@BTAH-Cu²⁺ system. Table II compares the applied potential, linear range, sensitivity and detection limit of various metal oxides and carbon based H₂O₂ sensors in pH 7 medium.^41,60–68 Among these, GCE/MWCNTs@BTAH-Cu²⁺ is able to detect H₂O₂ with less over potential. The sensitivity and detection limit obtained are comparable to that of the reported systems. To demonstrate the suitability of our H₂O₂ sensor for commercial samples, we have tested the sensor in the solution made from Natures' fruit bleach (Fig. 12C) and developer milk of new Garnier color natural cream (Fig. 12D). These cosmetic samples obtained from local super market and they were subjected to the real sample analysis by standard addition approach method. 1g of the cosmetic sample was mixed with 10 mL of pH 7 PBS under sonication and the filterate was used for the analysis. Real sample analyses were carried out using chronoamperometry method with 50 μL spikes of the analyte (first spike by commercial sample in PBS and subsequent spikes by 50 μmoles of H₂O₂) in to the 10 mL electrochemical bath. The original detected values of H₂O₂ are 24.7 μM and 15.38 μM in the Nature's fruit bleach and Garnier color natural respectively. The calculated recovery values are nearly equal to 105 and 102% for Nature's fruit bleach and Garnier color natural respectively, which indicates the high sensitivity and selectivity towards H₂O₂ while using GCE/MWCNTs@BTAH-Cu²⁺ as sensor. This electrode shows appreciable analytical performance. Table III shows the detailed information about the H₂O₂ detection in real sample analysis. As Nature's fruit bleach contain grape seed oil, its interference during H₂O₂ sensing is tested adding 50 μL of grape seed oil (Fig. 12E) and found that it does not interfere with H₂O₂ sensing. Similarly, Garnier color natural cream contain glycerol, tetrasodium pyrophosphate, sodium stannate and phosphoric acid. The above chemicals were added (50 μL) to understand their interference and they did not interfere with H₂O₂ sensing. (Fig. 12F).

Zoom In Zoom Out Reset image size