A Cobalt (II) Oxide Carbon Nanotube Composite to Assay Dopamine

A cobalt (II) oxide/carboxylic acid functionalized multiwalled carbon nanotube (CoO/COOH-MWNT) composite was fabricated for the biochemical detection of dopamine (DA). CoO particles were tethered to COOH-MWNTs by sonication. The current response versus different concentration was measured using cyclic voltammetry. Various parameters, including sonication time, pH, and loading were varied for the best current response. The composite with optimum current response was formed using a 30-min sonication time, at pH 5.0 and a 0.89 µg/mm2 loading onto the glassy carbon electrode surface. Good sensitivity with a limit of detection of 0.61 ± 0.03 μM, and dynamic range of 10–100 µM for DA is shown, applicable for neuroblastoma screening. The sensor was selective against ascorbic and uric acids.


Introduction
Neuroblastoma is a type of pediatric cancer that forms in the adrenal glands, which can be exhibited in the spine, chest, neck and abdomen. This condition is characterized by a high dopamine (DA) content in urinary excretion. Multiple analytical approaches, such as UV-vis spectrophotometry, high-performance liquid chromatography, fluorescence spectrophotometry and electrochemical methods, have been applied to detect dopamine in environmental and biological samples [1][2][3][4][5]. However, sample preparation prior to analysis by applying these methods can be time-consuming and complex.
Among the various techniques used for the detection of DA, electrochemical methods provide a simple and rapid means of detection without the need for costly instrumentation [6][7][8][9][10][11][12][13][14][15][16]. Normal urine pH can be slightly acidic or basic, ranging from pH = 4.5 to 8.0 [17]. Electrochemical analysis provides a rapid, cost-effective means of disease monitoring. The presence of ascorbic acid (AA) and uric acid (UA), in particular, interferes with DA during electrochemical detection and the electrochemical approach becomes difficult [18][19][20][21][22]. To this end, Khamlichi et al. [23] developed an L-leucine modified sol-gel carbon electrode for DA detection in the 10-to-120 µM concentration range in urine. We have re-designed the working electrode by incorporating solely non-perishable materials for the electrocatalyst surface to detect dopamine, making it more amenable for industrial scale-up, while also overcoming selectivity issues with UA and AA.
Carbon nanotubes are attractive materials for use in electrochemical electrodes applicable for DA [24]. Cobalt (II)-containing complexes are known to be effective electrocatalyst materials for the detection of DA [25]. In this study, we explore the utilization of the cobalt (II) oxide (CoO) particles tethered to COOH-functionalized multiwalled carbon nanotubes (COOH-MWNTs) to produce a composite to modify glassy carbon electrodes (GCE) for assaying DA using cyclic voltammetry (CV).

Results and Discussion
Scheme 1 shows the electrochemical mechanism of DA on the working electrode surface corresponding to the two observed peaks in the CV data (vide infra). XPS binding energy core level shifts of the synthesized CoO/COOH-MWNTs verified that the tethered particles were CoO, and not from Co3O4 ( Figure S1 in the Supplemental Materials).

Scheme 1. Reduction products of dopamine.
A pH dependence on signal sensitivity was observed with maximum current occurring at pH 5.0 for 100 mM DA ( Figure S2); pH values below and above 5.0 had a diminished current. We attribute the variations in signal to Coulombic effects from protonation/deprotonation of the CoO/COOH-MWNT surface contributing to electrostatic attraction of DA to the electrode surface. In addition to increased sensitivity to DA, surface charge can also explain the electrode's selectivity. AA, UA and DA have known Ip values (based on their pKas) of 4.2, 5.4 and 8.93, respectively. Under pH 5.0 solution conditions, AA and UA are near electrostatically neutral conditions as compared to DA, which, on the other hand, would be significantly more protonated according to Gouy-Chapman theory, and hence adopt a positive charge. The measured Ip of the CoO/COOH-MWNT composite itself was found to be 7.86 ( Figure S3 in the Supplemental Materials). The lower Ip value of the CoO/COOH-MWNTs as compared to that of DA (8.93) denotes a greater electron rich density for the composite at pH 5.0. The combination of the positively charged DA and the electron-rich character of the electrocatalyst serves as the driving force for the electrode's pronounced Coulombic attraction to DA as compared to AA and UA, providing a basis for the sensor's selectivity.
The CV of 100 µM DA shows a sharp anodic peak at oxidation potential +0.300V ( Figure 1A). The asymmetric shape of CV indicates a quasi-reversible redox process. We observe an electrocatalytic activity dependence with sonication time ( Figure 1B). Sonication was used to attach the particles to the underlying COOH-MWNT surface, as demonstrated previously in our laboratory [26][27][28]. Sonication periods that are too brief are known to result in a sparse density of attached particles while excessively prolonged times damages the stable carboxylate -COO-structures needed to create a fine dispersion of them, often resulting in agglomeration and reduced electroactive surface area [31]. Hence, the sonication time needed to produce the optimized dispersion of the composite needs to be experimentally determined prior to its use as a sensor. Figure 1B shows the ideal sonication period to be 30 minutes at the +0.300 V anodic voltage for 100 µM DA in order to achieve maximum current output. Furthermore, no signal at +0.300V was observed for either 100 µM AA or 100 µM UA, common interferents in urinalysis. Additional control experiments involving composite sub-components and various composite loadings on the GCE are shown in Figure S4 in the Supplementary Materials. A pH dependence on signal sensitivity was observed with maximum current occurring at pH 5.0 for 100 mM DA ( Figure S2); pH values below and above 5.0 had a diminished current. We attribute the variations in signal to Coulombic effects from protonation/deprotonation of the CoO/COOH-MWNT surface contributing to electrostatic attraction of DA to the electrode surface. In addition to increased sensitivity to DA, surface charge can also explain the electrode's selectivity. AA, UA and DA have known Ip values (based on their pK a s) of 4.2, 5.4 and 8.93, respectively. Under pH 5.0 solution conditions, AA and UA are near electrostatically neutral conditions as compared to DA, which, on the other hand, would be significantly more protonated according to Gouy-Chapman theory, and hence adopt a positive charge. The measured Ip of the CoO/COOH-MWNT composite itself was found to be 7.86 ( Figure S3 in the Supplemental Materials). The lower Ip value of the CoO/COOH-MWNTs as compared to that of DA (8.93) denotes a greater electron rich density for the composite at pH 5.0. The combination of the positively charged DA and the electron-rich character of the electrocatalyst serves as the driving force for the electrode's pronounced Coulombic attraction to DA as compared to AA and UA, providing a basis for the sensor's selectivity.
The CV of 100 µM DA shows a sharp anodic peak at oxidation potential +0.300V ( Figure 1A). The asymmetric shape of CV indicates a quasi-reversible redox process. We observe an electrocatalytic activity dependence with sonication time ( Figure 1B). Sonication was used to attach the particles to the underlying COOH-MWNT surface, as demonstrated previously in our laboratory [26][27][28]. Sonication periods that are too brief are known to result in a sparse density of attached particles while excessively prolonged times damages the stable carboxylate -COO-structures needed to create a fine dispersion of them, often resulting in agglomeration and reduced electroactive surface area [31]. Hence, the sonication time needed to produce the optimized dispersion of the composite needs to be experimentally determined prior to its use as a sensor. We have recently discovered that sonication can disentangle the COOH-MWNT support such that more of the particulate surface area can be exposed and thereby increase electrocatalytic activity of ZnO/COOH-MWNTs [28]. However, this same effect was not observed for CoO/COOH-MWNTs for the detection of DA; the degree of disentanglement of the underlying carbon nanotube surface had no effect on the sensitivity of the composite. Dissimilar to the findings pertaining to UA detection with ZnO/COOH-MWNTs as the working electrode material, measuring the integrated peak area intensity of the G band alone was insufficient to identify the optimum sonication time for DA detection. Rather, the dominant factor affecting sensitivity was the CoO particle surface density on the COOH-MWNT underlying substrate, which can be indirectly measured by noting the sp 3 -to-sp 2 hybridized carbon ratio of the electrocatalyst surface. A correlation between the relative sp 3hybridized carbon with increased DA signal at the maximum Raman D/G band ratio (relative integrated peak area at 2.59) in occurring at 30 minutes of sonication was observed ( Figure 2A). The peaks at 1325 and 1625 cm -1 correspond to the D and G bands, respectively. The D bands at 1325 cm -1 denote the disordered sp 3 -hybridized carbon and the G bands at 1625 cm -1 are related to ordered sp 2 -hybridized carbon [32].   Figure 1B shows the ideal sonication period to be 30 min at the +0.300 V anodic voltage for 100 µM DA in order to achieve maximum current output. Furthermore, no signal at +0.300 V was observed for either 100 µM AA or 100 µM UA, common interferents in urinalysis. Additional control experiments involving composite sub-components and various composite loadings on the GCE are shown in Figure  S4 in the Supplementary Materials.
We have recently discovered that sonication can disentangle the COOH-MWNT support such that more of the particulate surface area can be exposed and thereby increase electrocatalytic activity of ZnO/COOH-MWNTs [28]. However, this same effect was not observed for CoO/COOH-MWNTs for the detection of DA; the degree of disentanglement of the underlying carbon nanotube surface had no effect on the sensitivity of the composite. Dissimilar to the findings pertaining to UA detection with ZnO/COOH-MWNTs as the working electrode material, measuring the integrated peak area intensity of the G band alone was insufficient to identify the optimum sonication time for DA detection. Rather, the dominant factor affecting sensitivity was the CoO particle surface density on the COOH-MWNT underlying substrate, which can be indirectly measured by noting the sp 3 -to-sp 2 hybridized carbon ratio of the electrocatalyst surface. A correlation between the relative sp 3 -hybridized carbon with increased DA signal at the maximum Raman D/G band ratio (relative integrated peak area at 2.59) in occurring at 30 min of sonication was observed (Figure 2A). The peaks at 1325 and 1625 cm −1 correspond to the D and G bands, respectively. The D bands at 1325 cm −1 denote the disordered sp 3 -hybridized carbon and the G bands at 1625 cm −1 are related to ordered sp 2 -hybridized carbon [32].
Unlike the electrocatalytic activity of the ZnO/COOH-MWNTs, the relative degree of exposed sp 2 -hybridized carbon as exhibited by the Raman G bands (Figure 2A) did not correlate with the magnitudes if the CV intensities observed. Rather, it was the relative degree of disordered sp 3 -hybridized carbon on the MWNTs that correlated with the CV peak maximum current, coinciding with the largest D-to-G band ratio ( Figure 2B). It should be noted that the CoO particles used in this composite had a markedly larger diameter (~80 nm) (Figure 3) as compared to the previously tethered metal oxide nanoparticles (~15 nm diameter). Both CoO/COOH-MWNT (in this experiment) and ZnO/COOH-MWNT (in our previous experiment [29]) composites were tethered to carbon nanotube supports with the same length and diameter. We postulate that disentanglement plays a larger role in affecting the electroactive surface area for the particles (tethered to the MWNT surface) that are smaller in diameter, suggesting that the aspect ratio between the tethered particulate diameter vs. the carbon nanotube length and diameter is an important variable governing the exposed electroactive surface area. Beyond a certain particle-to-COOH-MWNT size ratio threshold (i.e., if the diameter is too large), disentanglement has little, if any, effect. In the case of the CoO/COOH-MWNT composite, the degree of sp 3 -hybridized carbon from the MWNT sidewalls plays a dominant role, linking to the density of tethered particles tethered to the sidewalls corresponding to the relative electroactive surface area; this phenomenon has also been observed for other particles attached to COOH-MWNTs, also verified by Raman spectroscopy [29]. that more of the particulate surface area can be exposed and thereby increase electrocatalytic activity of ZnO/COOH-MWNTs [28]. However, this same effect was not observed for CoO/COOH-MWNTs for the detection of DA; the degree of disentanglement of the underlying carbon nanotube surface had no effect on the sensitivity of the composite. Dissimilar to the findings pertaining to UA detection with ZnO/COOH-MWNTs as the working electrode material, measuring the integrated peak area intensity of the G band alone was insufficient to identify the optimum sonication time for DA detection. Rather, the dominant factor affecting sensitivity was the CoO particle surface density on the COOH-MWNT underlying substrate, which can be indirectly measured by noting the sp 3 -to-sp 2 hybridized carbon ratio of the electrocatalyst surface. A correlation between the relative sp 3hybridized carbon with increased DA signal at the maximum Raman D/G band ratio (relative integrated peak area at 2.59) in occurring at 30 minutes of sonication was observed (Figure 2A). The peaks at 1325 and 1625 cm -1 correspond to the D and G bands, respectively. The D bands at 1325 cm -1 denote the disordered sp 3 -hybridized carbon and the G bands at 1625 cm -1 are related to ordered sp 2 -hybridized carbon [32].  Unlike the electrocatalytic activity of the ZnO/COOH-MWNTs, the relative degree of exposed sp 2 -hybridized carbon as exhibited by the Raman G bands (Figure 2A) did not correlate with the magnitudes if the CV intensities observed. Rather, it was the relative degree of disordered sp 3hybridized carbon on the MWNTs that correlated with the CV peak maximum current, coinciding with the largest D-to-G band ratio ( Figure 2B). It should be noted that the CoO particles used in this composite had a markedly larger diameter (~80 nm) (Figure 3) as compared to the previously tethered metal oxide nanoparticles (~15 nm diameter). Both CoO/COOH-MWNT (in this experiment) and ZnO/COOH-MWNT (in our previous experiment [29]) composites were tethered to carbon nanotube supports with the same length and diameter. We postulate that disentanglement plays a larger role in affecting the electroactive surface area for the particles (tethered to the MWNT surface) that are smaller in diameter, suggesting that the aspect ratio between the tethered particulate diameter vs. the carbon nanotube length and diameter is an important variable governing the exposed electroactive surface area. Beyond a certain particle-to-COOH-MWNT size ratio threshold (i.e., if the diameter is too large), disentanglement has little, if any, effect. In the case of the CoO/COOH-MWNT composite, the degree of sp 3 -hybridized carbon from the MWNT sidewalls plays a dominant role, linking to the density of tethered particles tethered to the sidewalls corresponding to the relative electroactive surface area; this phenomenon has also been observed for other particles attached to COOH-MWNTs, also verified by Raman spectroscopy [29].  Figure 3A shows a representative TEM image of CoO/COOH-MWNTs at 30-minute sonication time. The composite becomes disentangled with an increase in sonication time. The highest amount of disordered sp 3 -hybridized carbon occurred at 30 minutes of sonication of the composite as observed from Raman analysis. As sonication proceeded beyond the 30-minute point, the composite becomes more entangled with the increase in sonication time. The TEM histogram ( Figure 3B) shows an average CoO particle size of 315 ± 12 nm (n = 74). Close-up TEM images revealing the underlying carbon nanotube support are shown in Figures S6 and S7 in the Supplemental Materials.
XPS analysis ( Figure S1) showed that the atomic percent composition of the CoO/COOH-MWNT composite was 97.03% carbon, 2.83% oxygen and 0.14% cobalt from the normalized integrated peak areas of the C 1s, O 1s and Co 2p core levels, respectively. Core level binding energy (BE) shifts confirm that the attached particles were those of CoO, not Co3O4 (see Figure S1 and narrative in the Supplemental Materials).
The asymmetric line shape of the CV shows greater signal for reduction than for oxidation ( Figure 1A). The measured voltage for oxidation remained constant, while that for reduction shifted  XPS analysis ( Figure S1) showed that the atomic percent composition of the CoO/COOH-MWNT composite was 97.03% carbon, 2.83% oxygen and 0.14% cobalt from the normalized integrated peak areas of the C 1s, O 1s and Co 2p core levels, respectively. Core level binding energy (BE) shifts confirm that the attached particles were those of CoO, not Co 3 O 4 (see Figure S1 and narrative in the Supplementary Materials).
The asymmetric line shape of the CV shows greater signal for reduction than for oxidation ( Figure 1A). The measured voltage for oxidation remained constant, while that for reduction shifted to higher voltages with increase in DA concentration. For this reason, the CV oxidation current at +0.300 V was selected over that for reduction (at~0.4V) to quantify DA concentration in aqueous solution. Figure 4 shows the CV current vs DA concentration plot. Asymmetry in the CV indicating irreversibility suggests the possible decomposition of DA on the electrode surface. Surface adsorption of the residue could impede DA oxidation and result in deviation from linear current vs concentration measurements. However, the saturation of surface sites was minimal since Randles-Sevcik analysis also showed a diffusion controlled process for the amperometric detection of DA ( Figure S5   In summary, the CoO/COOH-MWNT/GCE sensor, fabricated in a facile fashion using entirely non-perishable components, showed fast electrochemical responses with good sensitivity and selectivity. DA was detected selectively in the presence of AA and UA at pH 5.0, relevant for urinalysis to screen for neuroblastoma.  In summary, the CoO/COOH-MWNT/GCE sensor, fabricated in a facile fashion using entirely non-perishable components, showed fast electrochemical responses with good sensitivity and selectivity. DA was detected selectively in the presence of AA and UA at pH 5.0, relevant for urinalysis to screen for neuroblastoma.