Detection of Hydroxyl Radicals Using Cerium Oxide/Graphene Oxide Composite on Prussian Blue

A composite sensor consisting of two separate inorganic layers of Prussian blue (PB) and a composite of cerium oxide nanoparticles (CeNPs) and graphene oxide (GO), is tested with •OH radicals. The signals from the interaction between the composite layers and •OH radicals are characterized using cyclic voltammetry (CV). The degradation of PB in the presence of H2O2 and •OH radicals is observed and its impact on the sensor efficiency is investigated. The results show that the composite sensor differentiates between the solutions with and without •OH radicals by the increase of electrochemical redox current in the presence of •OH radicals. The redox response shows a linear relation with the concentration of •OH radicals where the limit of detection, LOD, is found at 60 µM (100 µM without the PB layer). When additional composite layers are applied on the composite sensor to prevent the degradation of PB layer, the PB layer is still observed to be degraded. Furthermore, the sensor conductivity is found to decrease with the additional layers of composite. Although the CeNP/GO/PB composite sensor demonstrates high sensitivity with •OH radicals at low concentrations, it can only be used once due to the degradation of PB.


Introduction
Hydroxyl radicals (•OH radicals) are one of the most reactive free radicals among reactive oxygen species (ROS). In a human body, •OH radicals are produced as a by-product of cellular respiration primarily in the mitochondria [1,2], the oxidation burst in phagocytic cells [3,4], and enzyme reactions [5,6] for various cellular functions such as restoration of damaged DNA [7], activating vital proteins [8,9], signaling pathways [10], and responding to external impacts [11]. The imbalance between production and elimination of •OH radicals occurs due to the overproduction of ROS or oxidants beyond the capability of cells to facilitate an effective antioxidant response [12,13]. The excess of •OH radicals could develop the oxidative stress condition in a human body leading to interference of the normal function of cells [14] and damage of cellular components including DNAs [15,16], and lipids [17,18]. Acceleration of aging, cancer, cardiovascular diseases, and neurodegenerative diseases, such as Alzheimer's disease and Parkinson's disease, are a few examples of the negative impacts from the oxidative stress [19][20][21]. The detection of •OH radicals, as a biomarker, therefore, is a crucial step in the diagnosis of those severe diseases at initial stages.

Synthesis of CeNP/GO Composite
Both CeNPs and GO (50 mg each) were added into 100 mL of deionized water. The mixing solution was then placed into an ultra-sonication bath for one hour. Following sonication, the mixing solution was stirred for two hours to form a composite. The homogeneous mixing solution was then transferred to a centrifuge tube and centrifuged at 12,000 rpm for 30 min to receive the precipitated solid from the liquid portion of the mixing solution. The composite sample was then collected and dried at 60 • C for 12 h [47]. Once dried, the solid was grounded to a fine powder and kept in a desiccator at room temperature. The final CeNP/GO composite was confirmed by SAXS and STEM.

Deposition of PB on a GCE
The deposition of PB on a GCE was reported in previous literatures [48,49]. Briefly, before any PB was immobilized on a working electrode, a GCE was cleaned with 0.1 N sulfuric acid using CV to eliminate impurities on the surface of electrode. After that, two solutions were prepared to deposit PB on the working electrode. The first solution was made of 2 mM potassium ferricyanide, K 3

Preparation of the CeNP/GO Composite on the PB-Modified GCE
10 mg of the CeNP/GO composite powder were suspended in 10 mL of deionized water. The solution was then sonicated for one hour to obtain a homogenous solution. The CeNP/GO composite solution was applied to the PB-modified glassy carbon working electrode by delivering 10 µL with a pipette and dried in an oven at 60 • C for one hour. After drying, the CeNP/GO composite was reduced by CV through electrochemical reduction with the potential range from 1.7 to −1.7 V at 40 mV/s for 12 cycles to improve the overall composite conductivity [45]. Then, the CeNP/GO composite layer on top of the PB layer was rinsed with deionized water and dried again under nitrogen gas. CV was used to confirm the presence of composite layer on top of the PB-modified glassy carbon working electrode with the potential range between −0.8 V to 0.8 V at a scan rate of 100 mV/s in the same CV solution used in 2.3.

Detection of •OH Radicals by CeNP/GO/PB on a GCE
To test the composite sensor, •OH radicals were generated using the Fenton reaction. 10 mM of H 2 O 2 solution was mixed with 10 mM solution of FeSO 4 ·7H 2 O with an equal volume to perform the Fenton reaction. The H 2 O 2 solution was covered with aluminum foil to prevent the oxidation from UV light exposure for the duration of the experiment. CV was implemented to test the sensor during the Fenton reaction. The first cycle in CV was run in the H 2 O 2 solution. After that, the test was paused and an equal volume of the FeSO 4 ·7H 2 O solution was added to the H 2 O 2 solution to begin the Fenton reaction. CV was continuously used to detect the current change of the sensor during the Fenton reaction with the potential range of −0.6-0.4 V at 100 mV/s. After the Fenton reaction terminated within 15 min, the sensor was transferred to the same CV solution used in 2.3 and CV was run to check for the degradation of PB and composite layers on the surface of electrode.
After testing, the sensors were washed with deionized water and dried under nitrogen gas for next tests. The same test procedure was repeated for a sensor multiple times to investigate the reusability of sensor. Both the reduction and oxidation responses (i.e., redox responses) in the cyclic voltammogram were used to calculate the redox response (∆A) of the sensor due to the redox reaction between the CeNP/GO composite and •OH radicals. The redox response in terms of the current change (∆A) was calculated using the procedure described in Figure 1, in which ∆A is taken from the difference between the currents at the oxidation and reduction peaks. The CV curve for H 2 O 2 shows no significant redox peaks, which proves that there is no considerable redox reaction between the CeNP/GO modified electrode and H 2 O 2 . Figure 2a summarizes the synthesis of the CeNP/GO/PB modified electrode and the detection of •OH radicals in the Fenton solution. The design concept of the sensor is also shown in Figure 2b. UV light exposure for the duration of the experiment. CV was implemented to test the sensor during the Fenton reaction. The first cycle in CV was run in the H2O2 solution. After that, the test was paused and an equal volume of the FeSO4·7H2O solution was added to the H2O2 solution to begin the Fenton reaction. CV was continuously used to detect the current change of the sensor during the Fenton reaction with the potential range of −0.6-0.4 V at 100 mV/s. After the Fenton reaction terminated within 15 min, the sensor was transferred to the same CV solution used in 2.3 and CV was run to check for the degradation of PB and composite layers on the surface of electrode. After testing, the sensors are washed with distilled water and dried under nitrogen gas for next tests. The same test procedure was repeated for a sensor multiple times to investigate the reusability of sensor. Both the reduction and oxidation responses (i.e., redox response) in the cyclic voltammogram were used to calculate the redox response (ΔA) of the sensor due to the redox reaction between the CeNP/GO composite and •OH radicals. The redox response in terms of the current change (ΔA) was calculated using the procedure described in Figure 1, in which ΔA is taken from the difference between the currents at the oxidation and reduction peaks. The CV curve for H2O2 shows no significant redox peaks, which proves that there is no considerable redox reaction between the CeNP/GO modified electrode and H2O2. Figure 2a summarizes the synthesis of the CeNP/GO/PB modified electrode and the detection of •OH radicals in the Fenton solution. The design concept of the sensor is also shown in Figure 2b.

Synthesis and Characterization of the CeNP/GO Composite
The composite was synthesized by a low-temperature solution process. The XRD patterns of GO, CeNPs, and the CeNP/GO composite are showed in Figure 3a-c, respectively. Figure [50,51]. As for the XRD pattern of the CeNP/GO composite, Figure 3c demonstrates the crystalline structure of CeNPs which confirms the presence of CeNPs in the composite. It is worth mentioning that the refractive index of the CeNP/GO composite spikes with a sharper peak in comparison to that of CeNPs, which is attributed to a highly ordered CeNP crystallinity in the composite. On the other hand, it is observed that the characteristic XRD pattern of GO around 25° significantly reduces in the CeNP/GO composite, which is thought to be due to the disorder of stacking of graphene oxide sheets in the composite.
The morphologies of CeNPs and the CeNP/GO composite were investigated using STEM. Figures 3d,e show the bright field TEM images of CeNPs and CeNP/GO composite, respectively. In Figure 3d, the CeNPs have an average size from 15 nm to 60 nm with a consistent cubic shape. For the CeNP/GO composite, which is exhibited in Figure 3e, the CeNPs are homogeneously dispersed all over the GO sheets. Thus, it is confirmed that the low-temperature solution process can be successfully used to prepare the CeNP/GO composite.

Characterization of the PB Layer Deposited on a GCE
The CV results for a bare GCE and the PB modified GCE are shown in Figure 4a,b. Once the electrochemical deposition was performed, two distinct redox peaks appear in the cyclic voltammogram for the PB modified electrode as shown in Figure 4b. These two redox peaks, which are found at 0.1 V and 0.6 V, represent the reduced form (Prussian white) and the oxidized form (Berlin green) of PB, respectively. Furthermore, the PB modified GCE shows a higher conductivity in comparison to the bare GCE. The increase of sensor conductivity is explained with an intrinsic characteristic of PB as an electrocatalyst. PB is well-known for its redox catalysis that increases a rate of electron transfer in a redox reaction between an electrode surface and electrolyte in a solution [52,53]. The addition of a PB layer on the electrode surface as an interlayer between the electrode and

Synthesis and Characterization of the CeNP/GO Composite
The composite was synthesized by a low-temperature solution process. The XRD patterns of GO, CeNPs, and the CeNP/GO composite are showed in Figure 3a-c, respectively. Figure [50,51]. As for the XRD pattern of the CeNP/GO composite, Figure 3c demonstrates the crystalline structure of CeNPs which confirms the presence of CeNPs in the composite. It is worth mentioning that the refractive index of the CeNP/GO composite spikes with a sharper peak in comparison to that of CeNPs, which is attributed to a highly ordered CeNP crystallinity in the composite. On the other hand, it is observed that the characteristic XRD pattern of GO around 25 • significantly reduces in the CeNP/GO composite, which is thought to be due to the disorder of stacking of graphene oxide sheets in the composite.
The morphologies of CeNPs and the CeNP/GO composite were investigated using STEM. Figure 3d,e show the bright field TEM images of CeNPs and CeNP/GO composite, respectively. In Figure 3d, CeNPs have an average size from 15 nm to 60 nm with a consistent cubic shape. For the CeNP/GO composite, which is exhibited in Figure 3e, CeNPs are homogeneously dispersed all over the GO sheets. Thus, it is confirmed that the low-temperature solution process can be successfully used to prepare the CeNP/GO composite.
Additionally, SEM was used to investigate the morphologies of the deposited PB layer on a GCE. Figure 4c,d are SEM images of a bare GCE and the PB modified GCE, respectively. Figure 4c shows an uneven surface of glassy carbon electrode. After the electrochemical deposition of PB, a homogenous PB layer across the electrode surface was formed as shown in Figure 4d. Thus, it is confirmed that, from the CV and SEM results, the electrochemical deposition is successfully used to deposit a PB layer on the electrode surface.

Characterization of CeNP/GO/PB on a GCE
The composite layer was deposited on an electrode surface using the drop casting method. The chemisorption interaction is responsible for the attachment of the CeNP/GO composite with the PB modified electrode. CV was employed to verify the deposition of CeNP/GO composite on top of the PB modified electrode. As shown in Figure 5, two redox peaks of PB turn into one redox peak of the CeNP/GO composite modified sensor. Furthermore, the electrode conductivity increases after applying the CeNP/GO composite layer on top of the PB modified electrode, which is attributed to the highly conductive GO in the composite. The potential change (ΔEp) of the oxidation and reduction peaks also decreases for the composite modified sensor. The shift of redox peaks either to positive or negative potential indicates the reversibility of redox reaction at the electrode surface as a peak-topeak separation (ΔEp). The ΔEp's of a bare and the composite on the PB modified electrode are 980 mV and 170 mV, respectively. This result indicates that PB in the composite tremendously enhances the

Characterization of the PB Layer Deposited on a GCE
The CV results for a bare GCE and the PB modified GCE are shown in Figure 4a,b. Once the electrochemical deposition was performed, two distinct redox peaks appear in the cyclic voltammogram for the PB modified electrode as shown in Figure 4b. These two redox peaks, which are found at 0.1 V and 0.6 V, represent the reduced form (Prussian white) and the oxidized form (Berlin green) of PB, respectively. Furthermore, the PB modified GCE shows a higher conductivity in comparison to the bare GCE. The increase of sensor conductivity is explained with an intrinsic characteristic of PB as an electrocatalyst. PB is well-known for its redox catalysis that increases a rate of electron transfer in a redox reaction between an electrode surface and electrolyte in a solution [52,53]. The addition of a PB layer on the electrode surface as an interlayer between the electrode and the CeNP/GO composite layer can facilitate the electron transfer resulting in an increase in the sensor conductivity [54,55].
Additionally, SEM was used to investigate the morphologies of the deposited PB layer on a GCE. Figure 4c,d are SEM images of a bare GCE and the PB modified GCE, respectively. Figure 4c shows an uneven surface of GCE. After the electrochemical deposition of PB, a homogenous PB layer across the electrode surface was formed as shown in Figure 4d. Thus, it is confirmed that, from the CV and SEM results, the electrochemical deposition is successfully used to deposit a PB layer on the electrode surface.
composite layer on top of the PB modified electrode. As demonstrated in Figure 5d,e, the surface morphology of PB modified GCE is completely different from the image taken after depositing CeNP/GO composite on the PB layer. Figure 5e shows the homogeneous dispersion of CeNP/GO composite on top of the PB modified GCE. Therefore, it is concluded that the CeNP/GO composite layer was successfully deposited on the PB modified electrode, and it showed a higher conductivity and required a lower potential to operate than the bare and PB modified electrodes.

Characterization of CeNP/GO/PB on a GCE
The composite layer was deposited on an electrode surface using the drop casting method. The chemisorption interaction is responsible for the attachment of the CeNP/GO composite with the PB modified electrode. CV was employed to verify the deposition of CeNP/GO composite on top of the PB modified electrode. As shown in Figure 5, two redox peaks of PB turn into one redox peak of the CeNP/GO composite modified sensor. Furthermore, the electrode conductivity increases after applying the CeNP/GO composite layer on top of the PB modified electrode, which is attributed to the highly conductive GO in the composite. The potential change (∆E p ) of the oxidation and reduction peaks also decreases for the composite modified sensor. The shift of redox peaks either to positive or negative potential indicates the reversibility of redox reaction at the electrode surface as a peak-to-peak separation (∆E p ). The ∆E p 's of a bare and the composite on the PB modified electrode are 980 mV and 170 mV, respectively. This result indicates that PB in the composite tremendously enhances the electron transfer for the redox reaction at the surface of electrode, which results in the significant reduction of ∆E p . Furthermore, SEM images were used to confirm the presence of CeNP/GO composite layer on top of the PB modified electrode. As demonstrated in Figure 5d,e, the surface morphology of PB modified GCE is completely different from the image taken after depositing CeNP/GO composite on the PB layer. Figure 5e shows the homogeneous dispersion of CeNP/GO composite on top of the PB modified GCE. Therefore, it is concluded that the CeNP/GO composite layer was successfully deposited on the PB modified electrode, and it showed a higher conductivity and required a lower potential to operate than the bare and PB modified electrodes.

Electrochemical Reduction of the CeNP/GO Composite
As mentioned earlier, the electrochemical reduction can improve the intrinsic conductivity of GO. Figure 6 shows the cyclic voltammogram for the CeNP/GO composite modified electrode before and after the electrochemical reduction. It is found that, the conductivity of CeNP/GO composite modified electrode significantly increases after treatment with the electrochemical reduction. The increase in the conductivity of the CeNP/GO composite modified electrode is due to the elimination of oxygen groups on GO by electrochemical reduction.

Electrochemical Reduction of the CeNP/GO Composite
As mentioned earlier, electrochemical reduction can improve the intrinsic conductivity of GO. Figure 6 shows the cyclic voltammogram for the CeNP/GO composite modified electrode before and after the electrochemical reduction step. It is found that, the conductivity of CeNP/GO composite modified electrode significantly increases after treatment with electrochemical reduction. The increase in the conductivity of the CeNP/GO composite modified electrode is due to the elimination of oxygen groups on GO by electrochemical reduction. Nanomaterials 2020, 10, x FOR PEER REVIEW 9 of 17

CV for •OH Radical Detection
As mention before, a CeNP has the dual oxidation states as Ce 3+ and Ce 4+ on the surface of particle. Several works have verified that the Ce 3+ oxidation state on the surface of CeNP is responsible for the oxidation reaction with high selectivity toward •OH radicals [40,41]. Our hypothesis is that CeNPs possessing the Ce 3+ oxidation state can be used as sensing element for •OH radicals via the oxidation reaction. Figure 7 shows the cyclic voltammograms of three different layers of the CeNP/GO composite sensor with (7a, b, and c) and without the PB deposition (7d, e, and f) in the presence of H2O2 and •OH radicals. Regardless of PB layer and additional composite layer(s), the CeNP/GO composite sensor shows the increase of oxidation current peak around 0.2 V in the presence of •OH radicals; in contrast, there is no oxidation current peak from the bare electrode. The composite shows greater reactivity with •OH than with H2O2 as Figure 7a shows, for example, that the redox response (∆A) for •OH is 87 ± 6.2 µA while the ∆A for H2O2 is 37 ± 0.5 µA. Therefore, it proves our hypothesis that CeNPs can be used as a sensing element and the Ce 3+ oxidation state on the surface of CeNP is the reactive site for •OH radicals.
The CeNP/GO composite was catalyzed with PB to improve the conductivity and sensitivity of the sensor with low detection limits. The redox response (∆A) of three different layers of a composite with and without PB to •OH radicals is presented in Figure 8. As expected, the PB modified composite sensor delivers a significant increase in the ∆A to •OH radicals compared to the composite sensor without the PB modification. Therefore, this experimental result confirms that the PB layer can be used as an electrocatalyst in this composite sensor configuration.
It was found, however, that the PB layer degraded after contacting with H2O2 or •OH radicals. In order to prevent the degradation of PB layer, additional layers of the CeNP/GO composite were deposited on top of the PB layer. It was thought that the extra layers of the composite deposited on

CV for •OH Radical Detection
As mention before, a CeNP has the dual oxidation states as Ce 3+ and Ce 4+ on the surface of particle. Several works have verified that the Ce 3+ oxidation state on the surface of CeNP is responsible for the oxidation reaction with high selectivity toward •OH radicals [40,41]. Our hypothesis is that CeNPs possessing the Ce 3+ oxidation state can be used as a sensing element for •OH radicals via the oxidation reaction. Figure 7 shows the cyclic voltammograms of three different layers of the CeNP/GO composite sensor with (7a, b, and c) and without the PB deposition (7d, e, and f) in the presence of H 2 O 2 and •OH radicals. Regardless of the PB layer and additional composite layer(s), the CeNP/GO composite sensor shows the increase of oxidation current peak around 0.2 V in the presence of •OH radicals; in contrast, there is no oxidation current peak from the bare electrode. The composite shows greater reactivity with •OH than with H 2 O 2 as Figure 7a shows, for example, that the redox response (∆A) for •OH radicals is 87 ± 6.2 µA while the ∆A for H 2 O 2 is 37 ± 0.5 µA. Therefore, it proves our hypothesis that CeNPs can be used as a sensing element and the Ce 3+ oxidation state on the surface of CeNP is the reactive site for •OH radicals.
The CeNP/GO composite was catalyzed with PB to improve the conductivity and sensitivity of the sensor with low detection limits. The redox response (∆A) of three different layers of the composite with and without PB to •OH radicals is presented in Figure 8. As expected, the PB modified composite sensor delivers a significant increase in the ∆A to •OH radicals compared to the composite sensor without the PB modification. Therefore, this experimental result confirms that the PB layer can be used as an electrocatalyst in this composite sensor configuration.
top of the PB layer would prevent the degradation of PB layer. As shown in Figure 8, the addition of composite layers is found to reduce the ∆A of the composite sensor in the presence of •OH radicals. This could be due to the additional layer(s) enhances agglomeration of the nanoparticles that results in the reduction of active sites and the decrease in the ∆A. Moreover, the increased layer thickness with the additional composite layer(s) results in a longer distance for electrons to transfer from active sites at the composite surface to the PB layer, leading to the reduction of the ∆A.  It was found, however, that the PB layer degraded after contacting with H 2 O 2 or •OH radicals. In order to prevent the degradation of PB layer, additional layers of the CeNP/GO composite were deposited on top of the PB layer. It was thought that the extra layers of the composite deposited on top of the PB layer would prevent the degradation of PB layer. As shown in Figure 8, the addition of composite layers is found to reduce ∆A of the composite sensor in the presence of •OH radicals. This could be due to the additional layer(s) enhances agglomeration of the nanoparticles that results in the reduction of active sites and the decrease in ∆A. Moreover, the increased layer thickness with the additional composite layer(s) results in a longer distance for electrons to transfer from active sites at the composite surface to the PB layer, leading to the reduction of ∆A. Nanomaterials 2020, 10, x FOR PEER REVIEW 11 of 17

Composite Sensor Response to Different •OH Radical Concentrations
The single layer of composite modified sensors with and without the PB deposition were used to detect •OH radicals in the concentration range from 0.1 to 10 mM as shown in Figure 9. Both the modified composite sensors show linear relationships between the ∆A and different concentrations of •OH radicals with R-square (R 2 ) values equal to 0.93 and 0.89 for with and without the PB deposition, respectively. A higher R 2 value of composite sensor with the PB deposition could be yielded from the electrocatalytic property of PB, which improves both conductivity and sensitivity as hypothesized before. Furthermore, the CeNP/GO composite modified sensors with the PB deposition shows a higher ∆A for all tested •OH radical concentrations than that without a PB layer in Figures 7 and 8. The limits of detection (LOD) of the composite sensor, calculated by the equation, (3.3 × SD)/b [56], where SD and b represent the standard deviation and a slope of the regression line, are 60 and 100 µM with and without the PB modification, respectively. The electrocatalytic effect of PB is the main factor contributing to a better sensor performance in terms of ∆A and LOD of the composite sensor. The LOD of this CeNP/GO composite sensor with the PB deposition are found to be comparable to other sensors, which are in the range of 1-100 µM [37,[57][58][59].

Composite Sensor Response to Different •OH Radical Concentrations
The single layer of composite modified sensors with and without the PB deposition were used to detect •OH radicals in the concentration range from 0.1 to 10 mM as shown in Figure 9. Both the modified composite sensors show linear relationships between the ∆A and different concentrations of •OH radicals with R-square (R 2 ) values equal to 0.93 and 0.89 for with and without the PB deposition, respectively. A higher R 2 value of composite sensor with the PB deposition could be yielded from the electrocatalytic property of PB, which improves both conductivity and sensitivity as hypothesized before. Furthermore, the CeNP/GO composite modified sensors with the PB deposition shows a higher ∆A for all the tested •OH radical concentrations than that without a PB layer in Figures 7 and 8. The limits of detection (LOD) for the composite sensor, calculated by the equation, (3.3 × SD)/b [56], where SD and b represent the standard deviation and a slope of the regression line, are 60 and 100 µM with and without the PB modification, respectively. The electrocatalytic effect of PB is the main factor contributing to a better sensor performance in terms of ∆A and LOD of the composite sensor. The LOD of this CeNP/GO composite sensor with the PB deposition are found to be comparable to other sensors, which are in the range of 1-100 µM [37,[57][58][59].

Effects of PB Degradation on Sensor Performance
PB turns out to be an important layer to improve the sensor conductivity and sensitivity. As mentioned before, however, PB is found to be degraded by oxidizing species, H2O2 and •OH radicals. Since PB is used as the electrocatalyst to improve the electron transfer for redox reactions, the degradation of PB surely impacts the ∆A of this composite sensor. Cyclic voltammograms of three different composite layers with the PB deposition before and after running in the Fenton reaction are showed in Figure 10. The ∆A of all composites with single, double, and triple layers are observed to decrease after performing the detection of •OH radicals regardless of the thickness of layer. To confirm the reduction of ∆A in Figure 10 results from the PB degradation, SEM images of the PB layers before and after the Fenton reaction are shown in Figure 11. Figure 11a shows the homogenous structure of PB layer, whereas a damaged rough surface of PB layer is shown after exposure to •OH radicals in the Fenton reaction in Figure 11b.
In Figure 12, the percent decreases of the sensor conductivities are estimated as 22.1%, 19.4%, and 23.2% for the single, double, and triple composite sensors with the PB deposition, respectively. On the other hand, the composite sensors without the PB deposition show the 7.2%, 7.8%, and 8.8% decreases in sensor conductivity for the single, double, and triple composite layers. From Figure 12, all the composite sensors of three different layers with the PB deposition show approximately three times more degradation compared to those without the PB deposition. From experimental results in Figures 10-12, it is concluded that the decrease of ∆A mainly results from the degradation of PB layer on the composite sensor. In addition, the different thicknesses of composite layer(s) (single, double, and triple) show no effect on the protection of PB from degradation.

Effects of PB Degradation on Sensor Performance
PB turns out to be an important layer to improve the sensor conductivity and sensitivity. As mentioned before, however, PB is found to be degraded by oxidizing species, H 2 O 2 and •OH radicals. Since PB is used as the electrocatalyst to improve the electron transfer for redox reactions, the degradation of PB surely impacts ∆A of this composite sensor. Cyclic voltammograms of three different composite layers with the PB deposition before and after running in the Fenton reaction are showed in Figure 10. ∆As of all composites with single, double, and triple layers are observed to decrease after performing the detection of •OH radicals regardless of the thickness of layer. To confirm the reduction of ∆A in Figure 10 resulting from the PB degradation, SEM images of the PB layers before and after the Fenton reaction are shown in Figure 11. Figure 11a shows the homogenous structure of PB layer, whereas a damaged rough surface of PB layer is shown after exposure to •OH radicals in the Fenton reaction in Figure 11b.
In Figure 12, the percent decreases of the sensor conductivities are estimated as 22.1%, 19.4%, and 23.2% for the single, double, and triple composite sensors with the PB deposition, respectively. On the other hand, the composite sensors without the PB deposition show the 7.2%, 7.8%, and 8.8% decreases in sensor conductivity for the single, double, and triple composite layers. From Figure 12, all the composite sensors of three different layers with the PB deposition show approximately three times more degradation compared to those without the PB deposition. From the experimental results in Figures 10-12, it is concluded that the decrease of ∆A mainly results from the degradation of PB layer on the composite sensor. In addition, the different thicknesses of composite layer(s) (single, double, and triple) show no effect on the protection of PB from degradation. Nanomaterials 2020, 10, x FOR PEER REVIEW 13 of 17

Conclusions
The CeNP/GO composite deposited on the PB modified GCE is successfully synthesized by the electrochemical deposition and the drop casting method. The one layer of CeNP/GO composite sensor shows its sensitivity with •OH radicals as it produces the current increase of 87 ± 6.2 µA in CV when contacts with •OH radicals, whereas the current increases by 37 ± 0.5 µA with H2O2. The composite sensors with and without PB modification show the linear relationships of redox response with •OH radical concentrations from 0.1 to 10 mM with the LOD as 60 and 100 µM, respectively. The PB layer is found to be a crucial factor as an electrocatalyst to improve the sensor efficiency in terms of the redox response and the LOD. Unfortunately, PB layer is found to degrade when exposed to •OH radicals or H2O2. The thicker composite layers show no effect on protecting the degradation of PB. Moreover, the thicker composite layers produce lower current responses. The optimum sensor configuration for •OH radical detection is the PB modified electrode with one layer of CeNP/GO composite. This work presents the promising results on the integration of PB and CeNP to develop the electrochemical sensor for the detection of •OH radicals. Moreover, the PB degradation by •OH radicals is confirmed in this study.

Single layer
Dobble layer Triple layer 0

Conclusions
The CeNP/GO composite deposited on the PB modified GCE is successfully synthesized by the electrochemical deposition and the drop casting method. The single layer of CeNP/GO composite sensor shows its sensitivity with •OH radicals as it produces the current increase of 87 ± 6.2 µA in CV when contacts with •OH radicals, whereas the current increases by 37 ± 0.5 µA with H 2 O 2 . The composite sensors with and without the PB modification show the linear relationships of redox response with •OH radical concentrations from 0.1 to 10 mM with the LOD as 60 and 100 µM, respectively. The PB layer is found to be a crucial factor as an electrocatalyst to improve the sensor efficiency in terms of the redox response and the LOD. Unfortunately, the PB layer is found to degrade when exposed to •OH radicals or H 2 O 2 . The double and triple composite layers show no effect on preventing the degradation of PB. Moreover, the double and triple composite layers produce lower current responses than the single composite layer. The optimum sensor configuration for •OH radical detection is the PB modified electrode with one layer of CeNP/GO composite. This work presents the promising results on the integration of PB and CeNPs to develop the electrochemical sensor for the detection of •OH radicals. Moreover, the PB degradation by •OH radicals is confirmed in this study.