H2O2 Detection at Carbon Nanotubes and Nitrogen-Doped Carbon

Article pubs.acs.org/ac

H2O2 Detection at Carbon Nanotubes and Nitrogen-Doped Carbon Nanotubes: Oxidation, Reduction, or Disproportionation? Jacob M. Goran, Ethan N. H. Phan, Carlos A. Favela, and Keith J. Stevenson* Department of Chemistry, Center for Nano- and Molecular Science and Technology, The University of Texas at Austin, 105 E. 24th St. Stop A5300, Austin, Texas 78712-1224, United States S Supporting Information *

ABSTRACT: The electrochemical behavior of hydrogen peroxide (H2O2) at carbon nanotubes (CNTs) and nitrogen-doped carbon nanotubes (N-CNTs) was investigated over a wide potential window. At CNTs, H2O2 will be oxidized or reduced at large overpotentials, with a large potential region between these two processes where electrochemical activity is negligible. At N-CNTs, the overpotential for both H2O2 oxidation and reduction is significantly reduced; however, the reduction current from H2O2, especially at low overpotentials, is attributed to increased oxygen reduction rather than the direct reduction of H2O2, due to a fast chemical disproportionation of H2O2 at the N-CNT surface. Additionally, N-CNTs do not display separation between observable oxidation and reduction currents from H2O2. Overall, the analytical sensitivity of N-CNTs to H2O2, either by oxidation or reduction, is considerably higher than CNTs, and obtained at significantly lower overpotentials. N-CNTs display an anodic sensitivity and limit of detection of 830 mA M−1 cm−2 and 0.5 μM at 0.05 V, and a cathodic sensitivity and limit of detection of 270 mA M−1 cm−2 and 10 μM at −0.25 V (V vs Hg/Hg2SO4). N-CNTs are also a superior platform for the creation of bioelectrodes from the spontaneous adsorption of enzyme, compared to CNTs. Glucose oxidase (GOx) was allowed to adsorb onto N-CNTs, producing a bioelectrode with a sensitivity and limit of detection to glucose of 80 mA M−1 cm−2 and 7 μM after only 30 s of adsorption time from a 81.3 μM GOx solution.

C

arbon nanotubes (CNTs) have been identified as an electrocatalytic material for the electrochemical detection of hydrogen peroxide (H2O2), an important biogenic analyte for enzyme based applications.1−9 While some attribute the electrocatalytic activity of CNTs to metallic impurities (residuals from the catalysts for CNT growth),10−12 or the graphitic edge plane character of the CNT ends and defects13,14 (similar to the edge plane of pyrolytic graphite which has been shown to promote facile electron transfer compared to the basal plane15−19), it is clear that CNTs are an attractive material, as numerous reports continue to incorporate CNTs into enzyme-coupled biosensors,20−22 biofuel cells,23−25 and bioelectrodes.26,27 Heteroatom-doped CNTs, such as the ntype nitrogen-doped CNTs (N-CNTs), have been shown to further increase the electrocatalytic properties for H2O2 detection, either by oxidation or reduction at the electrode surface.28−31 For H2O2 reduction at N-CNTs, the oxygen reduction reaction (ORR) is intricately tied with H2O2 detection because of a fast chemical disproportionation of H2O2 back into oxygen.29,32−34 The electrocatalytic detection of H2O2 at reducing potentials, especially at low overpotentials, is actually due to increased oxygen reduction, rather than the direct reduction of H2O2, which occurs at nondoped CNTs at higher overpotentials. Thus, H2O2 can be electrochemically detected at CNTs/N-CNTs by three different methods: oxidation, reduction, or disproportionation. In addition to their enhanced electrocatalytic activity, N-CNTs have also shown better biocompatibility,35,36 increased hydrophilicity,37,38 © XXXX American Chemical Society

and increased surface area for protein loading compared to nondoped CNTs.39,40 CNT mesh electrodes, which are good scaffolds for enzyme adsorption,41,42 further increase enzyme loading effective for enzyme-based biosensing applications, since an increased current response to enzymatically generated H2O2 will be observed at lower overpotentials.43 Herein, we report a thorough assessment of the electrochemical behavior of H2O2 at CNT and N-CNT mesh electrodes over a wide potential window, by cyclic voltammetry and rotating disk electrode amperometry. Additionally, we allow glucose oxidase (GOx) to spontaneously adsorb to the CNT/N-CNT surface, demonstrating a simple platform to create enzyme-based biosensors. In most studies, CNTs are modified by acid oxidation, solubilizing/dispersion agents, or immobilized with binders, polymers, biopolymers, or other matrixes like sol−gels, oil pastes, and hydrogels. These treatments can inhibit the inherent reactivity of CNTs, since the surface will be partially utilized by the binder/solubilizing agent. Here, we investigate the natural reactivity of the as-synthesized CNTs/N-CNTs without modifying the surface, allowing a more accurate depiction of the CNT/N-CNT electroactivity to H2O2, with and without adsorbed enzyme. Received: January 7, 2015 Accepted: May 26, 2015

A

DOI: 10.1021/acs.analchem.5b00059 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. Oxidation (A, C, E) and reduction (B, D, F) of H2O2 at GC (A, B), CNT (C, D) and N-CNT (E, F) electrodes (Ar saturated solution for CVs between 0.00 V and −1.30 V, air saturated solution for CVs between 0.00 V to 0.80 V, 0.1 M SPB, pH 7.0, scan rate 25 mV/s).

■

EXPERIMENTAL SECTION Enzyme and Chemicals. Glucose Oxidase (Type X-S from Aspergillus niger, E.C. 1.1.3.4, lyophilized powder, 100−250 U/ mg, MW 160 kDa), α-D-glucose, and m-xylene (anhydrous) were obtained from Sigma-Aldrich. Sodium phosphate monobasic (NaH2PO4, monohydrate), sodium phosphate dibasic (Na2HPO4, anhydrous), hydrogen peroxide 30% (w/ w), pyridine, pH calbration buffers (4.00, 7.00, 10.00), and sodium hydroxide were purchased from Fisher. Bis(cyclopentadienyl)iron (ferrocene) was obtained from Alfa Aesar. CNT/N-CNT Synthesis. CNTs and N-CNTs were synthesized in a quartz tube by a floating catalyst chemical vapor deposition (CVD) process. Ferrocene, which served as the catalyst, was dissolved in either m-xylene (CNTs) or pyridine (N-CNTs) at 20 mg mL−1 and injected into the quartz tube via a glass syringe (Hamilton 81320) and an automated syringe pump (New Era Pump Systems NE-1000) at 0.1 mL min−1. The quartz tube was laid lengthwise across two identical tube

furnaces (Carbolite Model HST 12/35/200/2416CG), set at different temperatures. The temperature of the first furnace was set to cause the injected ferrocene solution to vaporize (150 °C for CNTs or 130 °C for N-CNTs). Argon gas was coinjected with the ferrocene solution to carry the vaporized solution along the quartz tube into the second tube furnace. The argon flow rate was 500 sccm for CNTs, and 532 sccm for N-CNTs. In addition to argon, hydrogen (75 sccm, for CNTs) or ammonia (43 sccm, for N-CNTs) was also injected to provide a total gas flow of 575 sccm, which was controlled by two gas mass flow controllers (MKS type 1179A). The second furnace was set at 700 °C (CNTs) or 800 °C (N-CNTs) to cause the chemical vapor to deposit multiwalled CNTs/N-CNTs along the inside of quartz tube. N-CNTs made by this process displayed 7.4 ± 0.5 atom % N, as measured by X-ray photoelectron spectroscopy (XPS). Electrochemistry and Electrode Preparation. Cyclic voltammograms and amperograms are presented with a positive cathodic current and a negative anodic current. Amperometry was performed by poising the rotating disk electrode at a B


Article


unpretreated GC used here is essentially an inert electrical support to evaluate the electrochemical differences between CNT and N-CNT electrodes. For both GC and CNT electrodes, there is a “dead zone” where electrochemical activity to H2O2 is negligible. This zone is from 0.30 V to −0.60 V for GC and from 0.20 V to −0.50 V for CNTs. In contrast to CNTs, N-CNTs do not display a “dead zone”, but rather, a single potential where anodic current from H2O2 will turn into cathodic current (Figure 1F). This fulcrum point, which resides around −0.16 V to −0.19 V, is in the same potential region as the open circuit potential for NCNTs. Note that in the high positive potential region for NCNTs (starting around 0.65 V of Figure 1E), the background voltammogram (0 mM H2O2) has the highest current response, while increasing H2O2 concentrations display concurrently lower current responses. This indicates that the oxygen evolution reaction and/or oxidation of the N-CNTs themselves occurs at lower potentials than at CNTs. It should also be noted that around 200 mV more positive or negative of the fulcrum point for N-CNTs, the observed current is fairly constant as a function of potential, while CNTs and GC generally display a continually increasing current response as a function of applied potential. Figure 2 presents the unique behavior of H2O2 at N-CNTs using lower concentrations of H2O2 in an argon saturated

constant potential, while aliquots of either H2O2 or glucose were injected into solution. Electrochemical measurements were acquired with an Autolab PGSTAT30 potentiostat (GPES software version 4.9), a Hg/Hg2SO4 reference electrode (CH Instruments, +0.64 V vs SHE; +0.44 V vs Ag/AgCl; +0.40 V vs SCE), and a coiled Au counter electrode. Synthesized CNTs/ N-CNTs were sonicated in absolute ethanol for 2 h to homogenize the solution (0.4 mg mL−1 for CNTs and 2 mg mL−1 for N-CNTs) before use. CNT/N-CNT electrodes were created by drop casting 24 μg of nanotubes (one 12 μL aliquot of the 2 mg mL−1 solution of N-CNTs or six 10 μL aliquots of the 0.4 mg mL−1 for CNTs) onto a 0.5 cm diameter glassy carbon (GC) rotating disk electrode (Pine Instruments AFE2M050GC). Prior to drop casting, GC electrodes which were polished with a 0.05 μm alumina slurry on microcloth (Buehler) and briefly sonicated in 18 MΩ cm water to remove remaining alumina. Drop cast CNT/N-CNT electrodes were placed in a dilute mixture of ethanol and 0.1 M sodium phosphate buffer (SPB), where a visible phase boundary between the ethanol/0.1 M SPB top layer and a pure 0.1 M SPB bottom layer could be distinguished. This “wetting” procedure ensured solution entered into the porous CNT/NCNT mesh electrode. After CNT/N-CNT electrodes were “wet”, they were cycled between 0.00 V and −1.20 V (vs Hg/ Hg2SO4) to passivate electroactive iron remaining from the CVD synthesis.44 Rotating disk electrode (RDE) experiments were performed on a Pine Instruments AFMSRX rotator. Adsorption of GOx was performed at room temperature in a 81.3 μM GOx solution (13 mg GOx in 1 mL of 0.1 M SPB, pH 7.0).

■

RESULTS AND DISCUSSION Cyclic Voltammograms of CNT/N-CNT Electrodes in the Presence of H2O2. Figure 1 presents cyclic voltammograms (CVs) of glassy carbon (GC), CNTs, and N-CNTs cycled from 0.00 V to 0.80 V and from 0.00 V to −1.30 V (V vs Hg/Hg2SO4) in the presence of increasing amounts of H2O2 in 0.1 M sodium phosphate buffer (SPB) at a pH of 7.0 (scan rate 25 mV/s). For GC and CNTs, these two potential windows represent the regions where H2O2 is oxidized (0.00 V to 0.80 V) and reduced (0.00 V to −1.30 V). CNTs display a slightly lower onset potential for both oxidation and reduction, but a significantly increased current at any given potential compared to GC. The GC response shown here is representative of the electrode preparation procedure, consisting of polishing the GC surface with a 0.05 μm alumina slurry on microcloth (Buehler), followed by sonication in 18 MΩ cm water to remove adsorbed alumina (as outlined in the Experimental Section). The GC CV response to H2O2 can be drastically improved by electrochemical pretreatment, as shown in the Supporting Information Figure S-1, where the GC was cycled between 0.00 V and 1.30 V (10 times at 100 mV/s).18,45−47 The GC response can be further improved by increasing the anodic potential limit during the electrochemical pretreatment. Pushing the potential more positive to 1.80 V (cycled 10 times at 100 mV/s from 0.00 V to 1.80 V) causes electrochemically active surface functional groups to appear in the potential window 0.00 V to −1.20 V (shown in the Supporting Information Figure S-3), while simultaneously decreasing the peak potential for H 2 O 2 oxidation and the current response for H2O2 reduction (shown in the Supporting Information Figure S-2). Although GC is not the focus of this report, we note that pretreatment of GC is necessary to ensure a clean surface, while the

Figure 2. A N-CNT electrode in the absence (black) and presence of 0.1 mM (red), 0.2 mM (orange), 0.3 mM (yellow), 0.5 mM (green), 0.75 mM (light blue), and 1.0 mM (blue) H2O2 in an argon saturated solution, and in an oxygen saturated solution without H2O2 (purple) (0.1 M SPB, pH 7.0, scan rate 25 mV/s).

solution, along with the oxygen reduction reaction (ORR) from an oxygen saturated solution, between 0.00 V and −0.50 V. We have previously reported on the chemical disproportionation that occurs to H2O2 on the surface of N-CNTs, including mechanistic details.32,33 Figure 2 corroborates earlier reports, where in an argon saturated solution, H2O2 will display a peak identical to the ORR, indicative of H2O2 disproportionation into O2. This chemical disproportionation can also be observed through its effects on the open circuit potential, when H2O2 is introduced into solution. Figure 3 presents the open circuit potential (OCP) of a CNT (A) and a N-CNT (B) rotating disk electrode as 25 μM injections of H2O2 are introduced into solution every 60 s starting at 360 s, identical to the sensitivity measurements discussed in the next section. The OCP dramatically shifts, toward more negative potentials, when H2O2 interacts with N-CNTs, indicating a chemical reaction is taking place (H2O2 disproportionation). The OCP for CNTs C


Article


Figure 3. Open circuit potential of (A) a CNT and (B) a N-CNT rotating disk electrode as 25 μM injections of H2O2 are introduced into solution at 360 s (oxygen saturated solution, 0.1 M SPB, pH 7.0, 1000 rpm).

sensitivity and limit of detection of CNTs to H2O2. The sensitivity was calculated from the slope of the current response as a function of H2O2 concentration, using all 20 aliquots (all 20 points remained within the linear range). The limit of detection was calculated as three times the standard deviation of the background current (before H2O2 was introduced into solution, shown in the first 360 s of Figures 4−6), divided by the sensitivity. In the case of H2O2 oxidation, 25 μM aliquots were added to the solution, while 100 μM aliquots were used for the less responsive reduction of H2O2. The reduction of oxygen occurs at a lower potential than H2O2 reduction at CNTs,30−32 evidenced by the increasing cathodic background current as a function of potential, shown in Figure 4B. The ORR causes difficulty in obtaining a reproducible and analytically interpretable amperometric signal from the reduction of H2O2, since the background current obscures the H2O2 signal. Additionally, at the highest reducing potential of −0.70 V, the signal tends to tilt toward increasing current responses, rather than displaying a steady-state current at a given H2O2 concentration. This tilt can also be observed in an argon saturated solution, shown in the Supporting Information Figure S-4, where 25 μM H2O2 aliquots were added to solution (which extends the potential needed to observe tilt to −1.00 V). This tilt may be due to Au dissolution from the counter electrode, and redeposition onto the CNTs. Overall, the reduction of H2O2 mimics the CVs in Figure 1D, where the current response to H2O2 increases concurrently with potential, but is not very sensitive. The Supporting Information Figures S5 and S-6 present the individual amperograms (since some of the amperograms are difficult to see in Figure 4) for the selected potentials for H2O 2 oxidation and reduction, respectively. The oxidation of H2O2 at CNTs is considerably more sensitive and reproducible, since oxygen does not contribute to the background current. The initial oxidation sensitivity, at 0.10 V, is nearly identical to the highest reduction sensitivity, at −0.70 V. Take note that a tilt toward increased current occurs at 0.70 V. At this potential, the CNTs will reproducibly desorb from the GC surface due to oxidation of both the CNTs and the supporting GC electrode, causing the CNTs to disengage from the GC surface. The carbon oxidation current can be seen in the increased background current at higher potentials, before H2O2 is introduced into solution. The Supporting Information Figure S-7 displays the reproducible separation of CNTs from the GC surface at 0.70 V. Figure S-8 in the Supporting Information presents a close-up of the background carbon oxidation current, before H2O2 is introduced into solution. As

does not noticeably change (from the original trajectory) when H2O2 is introduced into solution. Sensitivity Measurements of CNT/N-CNT Rotating Disk Electrodes to H2O2. Amperometry at a poised rotating disk electrode can provide a more precise assessment of the reactivity of CNT and N-CNT electrodes to H2O2. Electrodes were rotated at 1000 rpm as aliquots of H2O2 were injected into the solution. Since these electrodes are intended to be coupled with H2O2 producing enzymes, which require O2 as a cofactor, analysis was carried out in an oxygen saturated solution. CNT electrodes were evaluated from 0.70 V to −0.70 V in 100 mV increments. Figure 4 presents representative amperograms of both the oxidation (A, 0.00 V to 0.70 V) and reduction (B, −0.10 V to −0.70 V) of H2O2, while Tables 1 and 2 present the

Figure 4. Current response of CNT rotating disk electrodes to aliquots of (A) 25 μM H2O2 at oxidizing potentials and aliquots of (B) 100 μM H2O2 at reducing potentials (O2 constantly bubbled 0.1 M SPB, pH 7.0, 1000 rpm). D


Article

Analytical Chemistry Table 1. Sensitivity and Limit of Detection of H2O2 at Oxidizing Potentials at CNTs potential (V vs Hg/Hg2SO4) sensitivity (mA M−1 cm−2) limit of detection (μM) no. of trials a

0.00 NDa NDa 5

0.10 1.8 ± 0.6 80 ± 30 6

0.20 6±2 20 ± 10 7

0.30 14 ± 6 20 ± 20 8

0.40 22 ± 5 9±6 12

0.50 70 ± 10 5±4 8

0.60 250 ± 50 1 ± 0.8 20

0.70 460 ± 60 0.6 ± 0.3 9

ND = Not Detected.

Table 2. Sensitivity and Limit of Detection of H2O2 at Reducing Potentials at CNTs potential (V vs Hg/Hg2SO4) sensitivity (mA M−1 cm−2) limit of detection (μM) no. of trials a

−0.10 NDa NDa 4

−0.20 NDa NDa 7

−0.30 NDa NDa 6

−0.40 0.07 ± 0.03 1800 ± 700 8

−0.50 0.27 ± 0.09 400 ± 100 8

−0.60 1.0 ± 0.6 200 ± 200 12

−0.70 2±2 300 ± 300 9

ND = Not Detected.

Figure 6. Current response to glucose at (A) CNT bioelectrodes poised at 0.60 V and (B) N-CNT bioelectrodes poised at 0.05 V after allowing GOx to spontaneously adsorb onto the electrode surface for 5 s (black), 30 s (red), 5 min (yellow), 30 min (green), 4 h (light blue), and 24 h (blue) at room temperature from a 81.3 μM GOx solution. (O2 constantly bubbled, 0.1 M SPB, pH 7.0, 1000 rpm, 1 mM glucose aliquots for CNT electrodes poised at 0.60 V, 50 μM glucose aliquots for N-CNT electrodes poised at 0.05 V).

Figure 5. Current response of N-CNT rotating disk electrodes to 25 μM H2O2 aliquots at (A) oxidizing potentials and (B) reducing potentials (O2 constantly bubbled, 0.1 M SPB, pH 7.0, 1000 rpm).

expected from the CVs in Figure 1C, the sensitivity increases as a function of potential, while the limit of detection decreases. Overall, oxidation of H2O2 provides significantly more current than H2O2 reduction at CNTs. Since N-CNTs were observed to be electrocatalytic compared to CNTs in Figure 1, the sensitivity of N-CNTs to H2O2 was investigated in a narrower potential range, from 0.60 V to −0.60 V. The potential regions where H2O2 is oxidized or reduced at N-CNTs are 0.60 V to −0.15 V and −0.20 V to −0.60 V, respectively. Figure 5 displays representative amperograms of the N-CNT rotating disk electrodes, while the electrodes are poised at oxidizing potentials (Figure 5A) or reducing potentials (Figure 5B). Table 3 and 4 present the sensitivity and limit of detection of N-CNT electrodes to H2O2

as a function of potential. The oxidation of H2O2 occurs at significantly lower overpotentials at N-CNTs than at CNTs, but more importantly, the initial sensitivity to H2O2 observed at N-CNTs, poised at −0.15 V, is not met by CNTs until they are poised at 0.60 V, a 750 mV difference. As the potential is pushed more positive, beyond −0.15 V, the sensitivity of NCNTs to H2O2 oxidation increases until 0.05 V, where the sensitivity and limit of detection seem to remain constant at about 840 mA M−1 cm−2 and 0.5 μM, respectively. At 0.50 V, the sensitivity starts to decrease while the limit of detection increases, a trend which continues at 0.60 V. Thus, 0.05 V represents the lowest potential with the highest current E


Article

Analytical Chemistry Table 3. Sensitivity and Limit of Detection to H2O2 at Oxidizing Potentials at N-CNTs potential (V vs Hg/ Hg2SO4) sensitivity (mA M−1 cm−2) limit of detection (μM) no. of trials

−0.15

−0.10

0.00

0.05

0.10

0.20

0.30

0.40

0.50

0.60

170 ± 40

410 ± 40

700 ± 60

830 ± 60

800 ± 100

870 ± 40

860 ± 60

830 ± 50

730 ± 30

700 ± 100

4±1

1.6 ± 0.2

0.8 ± 0.3

0.5 ± 0.2

0.5 ± 0.2

0.5 ± 0.2

0.5 ± 0.2

0.8 ± 0.3

2±1

4±2

6

6

6

17

9

7

6

6

7

7

Table 4. Sensitivity and Limit of Detection to H2O2 at Reducing Potentials at N-CNTs potential (V vs Hg/Hg2SO4) sensitivity (mA M−1 cm−2) limit of detection (μM) no. of trials

−0.20 100 ± 20 20 ± 20 7

−0.25 270 ± 20 10 ± 10 23

−0.30 260 ± 30 30 ± 10 8

−0.40 290 ± 40 60 ± 30 8

−0.50 430 ± 50 50 ± 20 7

−0.60 570 ± 60 40 ± 10 7

Table 5. Sensitivity, Limit of Detection, and Linear Range to Glucose after Adsorption of GOx onto CNTs adsorption time sensitivity (mA M−1 cm−2) limit of detection (μM) linear range (mM) no. of trials

5s 3±1 200 ± 200 6±1 6

30 s 2±1 300 ± 200 8±1 6

5 min 1.7 ± 0.6 400 ± 600 11 ± 5 6

30 min 2±1 300 ± 300 10 ± 2 6

4h 4±2 200 ± 300 13 ± 4 6

24 h 2±1 300 ± 200 15 ± 3 6

M−1 cm−2). For CNTs, the oxidation sensitivity at 0.60 V (250 mA M−1 cm−2) is only one-third the oxidation sensitivity for NCNTs poised at 0.05 V. Overall, N-CNTs display higher H2O2 sensitivities, at lower overpotentials. Spontaneous Adsorption of Glucose Oxidase onto CNT/N-CNT Electrodes. The spontaneous adsorption of enzyme onto CNTs/N-CNTs provides a simple platform to create bioelectrodes for biosensing applications. The sensitivity of the bioelectrode to an enzymatic substrate is a function of the surface coverage of the enzyme, since the CNTs/N-CNTs are needed to both immobilize the enzyme and detect the enzymatic turnover. By adjusting the time a CNT/N-CNT electrode is allowed to adsorb enzyme, a sensitivity peak is observed, indicating a balance between the blocked sites where enzyme has adsorbed to the surface, and unblocked sites where an enzymatic byproduct can be electrochemically detected.49 Glucose oxidase (GOx) is a model enzyme for bioelectrode studies,50 which will oxidize glucose into β-D-glucono-1,5lactone in the presence of oxygen. Enzymatic turnover of glucose from GOx adsorbed onto CNTs or N-CNTs is not observed without the presence of oxygen, or an appropriate mediator.51 Since the oxidation of H2O2 provides significantly more current than reduction for CNTs or disproportionation for N-CNTs (Tables 1−4), we have chosen to poise the CNT bioelectrodes at 0.60 V, and the N-CNT bioelectrodes at 0.05 V, after adsorption of GOx, in order to determine the sensitivity of the bioelectrodes to glucose as a function of adsorption time. Figure 6 displays representative amperograms of CNT (Figure 6A) and N-CNT (Figure 6B) rotating disk bioelectrodes to glucose after GOx has been allowed to adsorb for 5 s, 30 s, 5 min, 30 min, 4 h, and 24 h at room temperature from a 81.3 μM GOx solution (0.1 M SPB, pH 7.0). Aliquots of 50 μM glucose were injected into solution for the N-CNT bioelectrodes, but the less sensitive CNT bioelectrodes required 1 mM glucose injections in order to obtain a similar current response. Consequently, the CNT bioelectrode measurements were within the linear range (R2 ≥ 0.995). Table 5 presents the sensitivity, limit of detection, and linear range of glucose at each adsorption time point for CNT bioelectrodes, while Table 6 presents only the sensitivity and

response to H2O2. The reason the sensitivity decreases at 0.50 V is due to the oxidation of N-CNTs themselves, as indicated by the increased background current at 0.40 V, 0.50 V, 0.60 V in Figure 5A, and as reported elsewhere.48 An interesting feature of the N-CNT electrodes held at open circuit (between −0.16 V to −0.19 V), is that an initial anodic current response to aliquots of H2O2 will switch into cathodic current responses during the course of the measurement, due to the influence of H2O2 disproportionation on the N-CNT electrode (Figure 3B). Figure S-9 in the Supporting Information presents the current response switch during the course of a typical measurement, while the N-CNT electrode is poised at −0.18 V. The reduction of H2O2 also occurs at significantly lower overpotentials at N-CNTs than at CNTs. The initial reduction sensitivity to H2O2 observed at N-CNTs, poised at −0.20 V, is never met by CNTs. The highest reduction potential employed at CNTs, −0.70 V, is one-fiftieth the initial reduction sensitivity at N-CNTs (poised at −0.20 V). Furthermore, the reduction current response at N-CNTs from H2O2 is well behaved, even through the background current from the ORR, which increases as a function of potential, is over 100 fold larger than the background current observed at CNTs. Like the oxidation of H2O2 at N-CNTs, the reduction sensitivity seems to plateau, at around 270 mA M−1 cm−2, from −0.25 V to −0.40 V. This plateau is due to the fast disproportionation of H2O2 and subsequent current from the ORR, compared to the small or negligible contribution from the direct reduction of H2O2. Although the disproportion rate does show a potential dependence,33 the increase in the reduction sensitivity at larger overpotentials (−0.50 V and −0.60 V) is most likely due to an increased contribution of the direct reduction of H2O2. Figure S-10 in the Supporting Information presents the individual amperograms from Figure 5B, since the high background current obscures the stair-step current response. For both CNTs and N-CNTs, the oxidation of H2O2 provides significantly higher sensitivities than the reduction or disproportionation of H2O2. For N-CNTs, the sensitivity at the oxidation plateau (starting at 0.05 V, plateau at about 840 mA M−1 cm−2) is over three times higher than the sensitivity at the reduction plateau (starting at −0.25 V, plateau at about 270 mA F


Article

Analytical Chemistry Table 6. Sensitivity and Limit of Detection to Glucose after Adsorption of GOx onto N-CNTs adsorption time sensitivity (mA M−1 cm−2) limit of detection (μM) no. of trials

5s 60 ± 20 5±3 7

30 s 80 ± 10 7±7 14

5 min 50 ± 10 6±3 6

■

limit of detection of glucose at the same adsorption time points for N-CNT bioelectrodes. The CNT bioelectrodes do not display a sensitivity peak as a function of GOx adsorption time. In fact, the sensitivity and limit of detection are nearly identical at all adsorption time points, around 2.5 mA M−1 cm−2 and 300 μM. The lack of a pattern between the sensitivity and the adsorption time is most likely due to the potential chosen to poise the bioelectrode, 0.60 V. N-CNTs poised at a lower potential of 0.50 V oxidize and cause adsorbed enzyme to disengage from the surface.48 Although CNTs do not display as high of an anodic background current as N-CNTs at 0.60 V, oxidation of CNTs and the subsequent desorption of GOx would display a nearly uniform sensitivity to glucose as a function of adsorption time. N-CNTs display a clear pattern between the GOx adsorption time and the sensitivity of the resulting bioelectrode to glucose. The sensitivity increases from 5 to 30 s, then decreases concurrent with adsorption time as the adsorbed enzyme blocks more and more of the sites available for electrochemical detection. The highest sensitivity, 80 mA M−1 cm−2 at 30 s of GOx adsorption, is 40 times more sensitive to glucose than the identical adsorption time for CNT bioelectrodes, and obtained at a 550 mV lower potential.

■

CONCLUSION

■

ASSOCIATED CONTENT

30 min 30 ± 10 30 ± 20 6

4h 21 ± 9 22 ± 9 6

24 h 16 ± 5 30 ± 20 6

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

Financial support of this work was provided by the R.A. Welch Foundation (grant F-1529). E.N.H.P. and C.A.F. acknowledge support from NSF-REU program (grant CHE-1003947).

(1) Lin, Y.; Yantasee, W.; Wang, J. Front. Biosci. 2005, 2, 492−505. (2) Valentini, F.; Amine, A.; Orlanducci, S.; Terranova, M. L.; Palleschi, G. Anal. Chem. 2003, 75, 5413−5421. (3) Wang, J.; Musameh, M.; Lin, Y. J. Am. Chem. Soc. 2003, 125, 2408−2409. (4) Wang, J.; Musameh, M. Anal. Chem. 2003, 75, 2075−2079. (5) Rubianes, M. D.; Rivas, G. A. Electrochem. Commun. 2003, 5, 689−694. (6) Wang, J.; Deo, R. P.; Poulin, P.; Mangey, M. J. Am. Chem. Soc. 2003, 125, 14706−14707. (7) Wang, J. Electroanalysis 2005, 17, 7−14. (8) Rubianes, M. D.; Rivas, G. A. Electroanalysis 2005, 17, 73−78. (9) Chen, W.; Cai, S.; Ren, Q.-Q.; Wen, W.; Zhao, Y.-D. Analyst 2012, 137, 2602−2605. (10) Sljukić, B.; Banks, C. E.; Compton, R. G. Nano Lett. 2006, 6, 1556−1558. (11) Banks, C. E.; Crossley, A.; Salter, C.; Wilkins, S. J.; Compton, R. G. Angew. Chem., Int. Ed. Engl. 2006, 45, 2533−2537. (12) Pumera, M. Langmuir 2007, 23, 6453−6458. (13) Banks, C. E.; Davies, T. J.; Wildgoose, G. G.; Compton, R. G. Chem. Commun. 2005, 829−841. (14) Banks, C. E.; Compton, R. G. Analyst 2006, 131, 15−21. (15) Bowling, R.; Packard, R.; McCreery, R. J. Am. Chem. Soc. 1989, 111, 1217−1223. (16) Rice, R. J.; McCreery, R. L. Anal. Chem. 1989, 61, 1637−1641. (17) Cline, K.; McDermott, M.; McCreery, R. J. Phys. Chem. 1994, 98, 5314−5319. (18) McCreery, R. L. Chem. Rev. 2008, 108, 2646−2687. (19) McCreery, R. L.; McDermott, M. T. Anal. Chem. 2012, 84, 2602−2605. (20) Balasubramanian, K.; Burghard, M. Anal. Bioanal. Chem. 2006, 385, 452−468. (21) Jacobs, C. B.; Peairs, M. J.; Venton, B. J. Anal. Chim. Acta 2010, 662, 105−127. (22) Vashist, S. K.; Zheng, D.; Al-Rubeaan, K. Biotechnol. Adv. 2011, 29, 169−188. (23) Willner, I.; Yan, Y.-M.; Willner, B.; Tel-Vered, R. Fuel Cells (Weinheim, Ger.) 2009, 9, 7−24. (24) Meredith, M. T.; Minteer, S. D. Annu. Rev. Anal. Chem. 2012, 5, 157−179. (25) Holzinger, M.; Le Goff, A.; Cosnier, S. Electrochim. Acta 2012, 82, 179−190. (26) Walcarius, A.; Minteer, S. D.; Wang, J.; Lin, Y.; Merkoçi, A. J. Mater. Chem. B 2013, 1, 4878−4908. (27) Sarma, A. K.; Vatsyayan, P.; Goswami, P.; Minteer, S. D. Biosens. Bioelectron. 2009, 24, 2313−2322. (28) Xu, X.; Jiang, S.; Hu, Z.; Liu, S. ACS Nano 2010, 4, 4292−4298.

CNTs and N-CNTs were directly compared as electrode materials for H2O2 detection. At CNTs, H2O2 is oxidized or reduced at high overpotentials, with oxidation being significantly more sensitive than reduction. At N-CNTs, the overpotentials needed for H2O2 oxidation and reduction are substantially lower than at CNTs, with the reduction current at low overpotentials due to increased oxygen reduction, since H2O2 will rapidly disproportionate at the surface of N-CNTs. Similar to CNTs, oxidation of H2O2 at N-CNTs is more sensitive than reduction/disproportionation. Unlike CNTs, both the anodic and cathodic current responses at N-CNTs are considerably more sensitive than those observed at CNTs, and obtained at lower potentials. N-CNTs were also shown to be more effective for the creation of bioelectrodes, by allowing enzyme to spontaneously adsorb onto the electrode surface. The sensitivity of the N-CNT bioelectrode to glucose after GOx was allowed to adsorb onto the surface displayed a sensitivity peak, associated with the balance between sites available for enzyme adsorption, and sites available for the subsequent detection of the enzymatically generated H2O2.

* Supporting Information S

Supporting Information includes ten figures as noted in the text. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b00059. G


Article

Analytical Chemistry (29) Tang, Y.; Allen, B. L.; Kauffman, D. R.; Star, A. J. Am. Chem. Soc. 2009, 131, 13200−13201. (30) Lyon, J. L.; Stevenson, K. J. ECS Trans. 2009, 16, 1−12. (31) Goran, J. M.; Lyon, J. L.; Stevenson, K. J. Anal. Chem. 2011, 83, 8123−8129. (32) Maldonado, S.; Stevenson, K. J. J. Phys. Chem. B 2005, 109, 4707−4716. (33) Wiggins-Camacho, J. D.; Stevenson, K. J. J. Phys. Chem. C 2011, 115, 20002−20010. (34) Gong, K.; Du, F.; Xia, Z.; Durstock, M.; Dai, L. Science 2009, 323, 760−764. (35) Carrero-Sanchez, J. C.; Elías, a L.; Mancilla, R.; Arrellín, G.; Terrones, H.; Laclette, J. P.; Terrones, M. Nano Lett. 2006, 6, 1609− 1616. (36) Zhao, M. L.; Li, D. J.; Yuan, L.; Yue, Y. C.; Liu, H.; Sun, X. Carbon N. Y. 2011, 49, 3125−3133. (37) Goran, J. M.; Stevenson, K. J. Langmuir 2013, 29, 13605− 13613. (38) Majeed, S.; Zhao, J.; Zhang, L.; Anjum, S.; Liu, Z.; Xu, G. Nanotechnol. Rev. 2013, 2, 615−635. (39) Wiggins-Camacho, J. D.; Stevenson, K. J. J. Phys. Chem. C 2009, 113, 19082−19090. (40) Burch, H. J.; Contera, S. A.; de Planque, M. R. R.; Grobert, N.; Ryan, J. F. Nanotechnology 2008, 19, 384001. (41) Felhofer, J. L.; Caranto, J. D.; Garcia, C. D. Langmuir 2010, 26, 17178−17183. (42) Nejadnik, M. R.; Deepak, F. L.; Garcia, C. D. Electroanalysis 2011, 23, 1462−1469. (43) Dey, R. S.; Bera, R. K.; Raj, C. R. Anal. Bioanal. Chem. 2013, 405, 3431−3448. (44) Lyon, J. L.; Stevenson, K. J. Langmuir 2007, 23, 11311−11318. (45) Engstrom, R. Anal. Chem. 1982, 2314, 2310−2314. (46) Engstrom, R. C.; Strasser, V. a. Anal. Chem. 1984, 56, 136−141. (47) McCreery, R. In Electroanalytical Chemistry; Bard, A. J., Ed.; Dekker: New York, 1991; pp 221−374. (48) Goran, J. M.; Favela, C. A.; Stevenson, K. J. ACS Catal. 2014, 4, 2969−2976. (49) Goran, J. M.; Favela, C. A.; Stevenson, K. J. Anal. Chem. 2013, 85, 9135−9141. (50) Wilson, R.; Turner, A. Biosens. Bioelectron. 1992, 7, 165−185. (51) Goran, J. M.; Mantilla, S. M.; Stevenson, K. J. Anal. Chem. 2013, 85, 1571−1581.

H


H2O2 Detection at Carbon Nanotubes and Nitrogen-Doped Carbon

Recommend Documents