Binder-Free Modification of a Glassy Carbon Electrode by Using Porous Carbon for Voltammetric Determination of Nitro Isomers.

In this study, Liquidambar formosana tree leaves have been used as a renewable biomass precursor for preparing porous carbons (PCs). The PCs were produced by pyrolysis of natural waste of leaves after 10% KOH activation under a nitrogen atmosphere and characterized by a variety of state-of-the-art techniques. The PCs possess a large surface area, micro-/mesoporosity, and functional groups on its surface. A glassy carbon electrode modified with high PCs was explored as an efficient binder-free electrocatalyst material for the voltammetric determination of nitro isomers such as 3-nitroaniline (3-NA) and 4-nitroaniline (4-NA). Under optimal experimental conditions, the electrochemical detection of 3-NA and 4-NA was found to have a wide linear range of 0.2-115.6 and 0.5-120 μM and a low detection limit of 0.0551 and 0.0326 μM, respectively, with appreciable selectivity. This route not only enhanced the benefit from biomass wastes but also reduced the cost of producing electrode materials for electrochemical sensors. Additionally, the sensor was successfully applied in the determination of nitro isomers even in the presence of other common electroactive interference and real samples analysis (beverage and pineapple jam solutions). Therefore, the proposed method is simple, rapid, stable, sensitive, specific, reproducible, and cost-effective and can be applicable for real sample detection.

Introduction

In recent decades, different allotropic forms of carbon materials such as amorphous carbon, diamond, graphite, nanotubes, fullerene, porous carbon (PC), graphene-templated carbons, and so forth have been used extensively in catalysis, sensors, and energy-related applications.[1−5] Among them, graphene and PC are chemically inert with high surface area. These physicochemical properties make them efficient to be used particularly in supercapacitors, solar cells, and electrochemical sensors.[6−8] PC owns unique properties, high surface areas, and large tunable pore channels, showing advantages in fast diffusion of molecules and the rapid electron transfer process, thus having an extensive application in electrochemical sensors and so forth.[9,10] Thus far, various synthetic methods such as chemical vapor deposition,[11] arc discharge,[12] laser ablation,[13] and chemical methods[14] have been developed to achieve high-quality carbon nanomaterials. More recently, the synthesis of high surface area carbon for electrochemical sensing applications are turned to natural biowaste materials.[15−18] In earlier dates, the activated coke and coal are used as the nonrenewable sources to generate high surface carbon,[19,20] but they do not draw much interest. Afterward, organic/inorganic moieties are adopted as the precursors to synthesize carbon materials,[21] but the treatment is costly and unfriendly to the environment.[22] Recently, sustainable PC derived from biomass or earth-abundant renewable resources has attracted much attention because it is relatively effective, environmentally friendly, and simple in the synthesis approach.[23−25] For instance, Liu et al. studied the nitrogen-doped carbon derived from biomass (grass), and the carbon-rich nanomaterial was exhibited as a promising fluorescent sensing platform for the label-free detection of Cu(II) ions.[26] Huang et al. developed the amorphous oxygen-doped carbon nanosheet (O@CN) from tannin which was applied as the electrocatalyst for efficient N2 fixation to NH3 under ambient conditions.[27] In addition, Wu et al. obtained the oxygen-doped hollow carbon microtubes (O-KFCNTs) derived from natural kapok fibers using high-temperature carbonization under argon flow. The O-KFCNTs were employed as an electrocatalyst for a metal-free N2 reduction reaction to convert N2 to NH3 with excellent selectivity and to achieve good electrochemical performances.[28,29]

Liquidambar formosana is commonly known as the Chinese sweetgum or formosan gum, and its crown is pyramid-shaped around 40–60 feet wide that becomes more oval as the tree ages.[30] Its leaves are star-shaped with a fragrant odor when crushed. Their colors change from green in the spring and summer to purple, red, or yellow in the fall. Thousands of tons of dead leaves are produced each year. However, it is easy to identify in the winter by the one-inch fruit balls called as Lu Lu Tong that hang from its branches, and its fruits have been used as traditional Chinese medicine.[31] It grows preferentially in East Asia, and its wood is second only to oak as commercial hardwood.[32] However, the tree resin was once made into gum that was used by confederate doctors to treat dysentery in the troops. Some organic substances were extracted from the sweetgum leaves to be used as medicines as well as essential oils were extracted to treat skin infections and other ailments.[33] Their dead leaves are also enriched in lignocellulosic materials.[34] The abundance of lignocelluloses in the sweetgum tree makes it a useful raw material for the fabrication of activated carbon (AC). Herein, we took the dead leaves of the tree as a precursor to derive PC.

Aromatic amines (AMs) such as aniline and its derivatives (ortho-, para-, and meta-nitroanilines) are commonly utilized in chemical, pharmaceutical, polymer, plastic, and dye industries in which the azo (−N=N−) components are useful in foods and textile applications.[35,36] In contrast, these AM compounds are suspected to cause potential diseases such as carcinogens, skin, and eye irritants. Herein, we developed electrochemical sensing based upon carbon materials for fast detection of AMs in drinking and environmental waters.[37−39] This electrochemical method was reported to be superior to traditional sensing and monitoring techniques such as capillary zone electrophoresis, high-performance liquid chromatography, planar electrochromatography, solid-phase extraction, and spectrophotometry methods.[40−44]

In this study, we aim to develop a binder-free, metal-free, high-sensitive electrochemical determination of 3-nitroaniline (3-NA) and 4-nitroaniline (4-NA) in an aqueous medium. The as-fabricated carbon materials showed high surface area, long-term stability, and easy substrate penetration structures for 3-NA and 4-NA molecules. The PC-modified electrodes were demonstrated to have remarkable sensitivity with an excellent detection limit for 3-NA and 4-NA. The selectivity, stability, reproducibility, interference effect, and analysis in real samples were also examined. Moreover, such prepared PC-modified glassy carbon electrode (GCE) materials show unique properties and remarkable electrochemical sensing performances.

Results and Discussion

XRD and Raman Analysis

The X-ray diffraction (XRD) patterns of as-prepared carbons are shown in Figure a. All samples exhibited two broad peaks with low intensity at 2θ = 25.3° and 43.4°, which can be indexed to (002) and (100) spacings of the graphitized turbostratic carbon structure, respectively. The resulting (002) crystalline plane of graphite indicates the presence of graphitization in all samples. A minor peak at 78.3° appeared corresponding to the (110) crystalline plane of graphite following the KOH activation process in the PC600, PC700, PC800, and PC900 samples. The graphitization of PC was apparently enhanced after the KOH treatment, and the enhanced crystallinity helped improve the conductivity of the carbon material.[45] The raw sample, PC (without activation), and KOH-activated (400 °C and 900 °C) PC samples were compared for their crystallinity and phase and characterized by XRD patterns (Figure S1 of the Supporting Information). In Figure b, two typical carbon peaks at 1366 and 1606 cm–1 were identified in the Raman spectra for the synthesized samples, corresponding to D (defect, sp3, A1g symmetry) and G (graphitic, sp2, E2g vibrations) bands, respectively.[46] Usually, an intensity ratio of ID/IG is used to characterize the extent of the disordered carbon presence in carbon samples,[47] yielding 1.01, 0.99, 0.99, and 0.98 for PC600, PC700, PC800, and PC900 samples, respectively, indicating that the synthesized samples have a less amount of defective sites in the structure as the carbonization temperature increased.[48] The resulting higher extent of the graphitic phase, associated with the conducting phase of the carbon, could facilitate electron transfer or enhance its electrocatalytic activity.

Figure 1

(a) XRD patterns and (b) Raman spectra of the PC samples.

N2 Adsorption–Desorption

The surface area and porosity of the PC samples were studied by nitrogen adsorption–desorption analysis. All samples displayed a typical type-IV isotherm,[49] as shown in Figure a–d. The hysteresis loop shown in a pressure range between 0.48 and 0.99 is categorized as type-H4, which can be attributed to the capillary condensation taking place in the mesopores.[50] More detailed pore parameters of PCs are shown in Table S1 (Supporting Information). When the activation temperature increased from 600 to 900 °C, the Brunauer–Emmett–Teller surface area (SBET) increased from 688.5 to 1210.4 m2 g–1, and the total pore volume (Vtot) rises from 0.82 to 0.98 cm3 g–1. While comparing these results, the KOH activation is found to be effective to increase the surface area by forming a porous structure inside the carbon framework. Figure a1–d1 shows the pore size distributions of as-prepared samples. Both micropores/mesopores persist in all carbon samples, consistent with the results of the N2 isotherms.[51] For comparison, the yield of different samples prepared by the method of chemical activation with KOH is summarized in Table S1 (Supporting Information). The results clearly show that as activation temperature increases, the yield of O, H, S, and N decreases, but the yield of C increases. At higher temperature, the aromatic condensation may take place among the adjacent molecules, thus generating gaseous products from the carbonized char to increase the C yield. Additionally, the elemental analysis (EA) evidenced the presence of the rich C and O elements in the L. formosana leaves collected from different places (see Table S2, Supporting Information).

Figure 2

(a–d) Nitrogen adsorption–desorption isotherms and (a1–d1) the pore size distributions for prepared PCs with variable activation temperatures.

Structural Properties

The filed emission-scanning electron microscopy (FE-SEM) micrographs of PCs are displayed in Figure . In Figure a,b, the formation of carbon particles with a large smooth surface could be seen in the PC600 sample, confirming that the carbonization process at 600 °C in an N2 atmosphere could convert the dried leaves to a valuable carbon product. At higher temperature, the carbon products (PC700, PC800, and PC900) showed a rough surface and highly porous sheet-like morphology with interconnected pores arranged in an irregular manner,[52] which are displayed in Figure c–h. The formation of porous PCs was attributed to the mixed treatment of 25 mL of 10% KOH solution with the PC powders. KOH was melted at the lower temperature, and then K2CO3 was produced from the reaction of KOH and PC powder at above 650 °C. Finally, K2CO3 was decomposed into K2O and CO2 at higher temperatures.[53] K2O was reduced by carbon to produce metallic potassium (K) at a temperature above 700 °C, which was then intercalated to the carbon matrix. After washing, this intercalated metallic K was eliminated, leading to the formation of pore structures; the treatment processes are provided in Figure S2 of the Supporting Information.

Figure 3

SEM micrographs of (a,b) porous PC600, (c,d) PC700, (e,f) PC800, and (g,h) PC900 samples.

Additionally, the microstructure of tree leaves-derived carbons was examined using a field emission transmission electron microscope (FE-TEM). Typical FE-TEM images of PC900 are presented in Figure . The PC900 sample exhibited a cottony structure, consisting of highly interconnected 3D hollow PC layers that overlap with each other (Figure a–c). The selected area electron diffraction (SAED) pattern (Figure d) further shows two clear rings, revealing the polycrystalline nature of PC900. The two weak diffraction rings for the (200) and (101) facet reflections of graphitized carbon reveal that the partial graphitization of PC900 may help increase electric conductivity.[50]

Figure 4

(a–c) FE-TEM images of PC900 at different magnifications and the (d) SAED pattern.

The elemental mappings of C, N, S, and O as well as the overlay images (Figure S3a–e, of the Supporting Information) were further examined to confirm that C, N, S, and O atoms are uniformly distributed on the PC900 sample, in agreement with the X-ray photoelectron spectroscopy (XPS) results (vide infra).

X-ray Photoelectron Spectroscopy

The surface states and components of the as-prepared PC samples were furthermore analyzed by XPS. Survey spectra scans (0–1200 eV) were used to quantify the surface elements, of which only oxygen and carbon were present as shown in Figure a. As shown in Figure b, the O 1s spectra of the PC600 is dominated by two components with binding energies (B.Es) at 531.8 and 533.2 eV, whereas PC700 and PC800 samples show three peaks at B.Es of 532.8, 533.6, and 535.3 eV corresponding to the carbonyl (C=O), hydroxyl (C–OH), and carboxyl (O=C–O), respectively, as displayed in Figure c,d. These ACs contain significant amounts of oxygen on their surfaces.[45] The O 1s peak in the PC900 sample can be deconvoluted into four contributions (Figure e). As reported,[45] the peak at 531.8 eV is related to the COO– group in carboxylate and the O=C group, and the peak at 532.6 eV corresponds to the C–OH and C=O functional groups, while the peaks at 534.0 and 538.1 eV are attributed to the oxygen single bond (O=C–O) in esters and carboxylic acids, and chemisorbed oxygen or water (H–O–H) molecules. The absorbed water or oxygen can be attributed to edge-dangling bonds induced by the high-temperature chemical treatment. The XPS analysis yielded the surface oxygen contents in both mass percentage (wt %) and atomic percentage (wt %) of all samples (see Table S3 of the Supporting Information). The higher O content is consistent with the formation of carbon–oxygen surface groups. In addition, the relative percentages of oxygenated functional groups increased with increasing activation temperature with KOH treatment. The C content by the XPS analysis (in Table S3) is smaller than that by EA, whereas the O contents show the opposite trend. These facts can be readily understood because the number of atomic layers in the samples amenable to the analysis is usually larger by the EA method than by XPS, and further, the oxygen surface groups should concentrate on the surface of ACs. In fact, XPS has been used for determination of the surface oxygen content in AC. These results are smaller than the O content obtained via EA analysis, suggesting that the O content on surface of the KOH ACs decreases with respect to the untreated carbon. Most of the surface of the ACs belongs to internal pores (micro/mesopores), which are thus not probed by XPS. This may explain the difference in oxygen content determination between XPS and CHN analysis. When the activation temperature increases, the reaction between carbon and KOH becomes more violent in PCs, resulting in carbon oxidation such that the carbon content is reduced and that is why the oxygen content increased after the activation at 900 °C (see Table S3).

Figure 5

(a) XPS survey spectra of PC600, PC700, PC800, and PC900, and (b–e) the corresponding spectra of O 1s.

FT-IR Study

The surface of carbon materials contains abundant functional groups among which the CO-type groups are predominant. These functional groups present on the surface are analyzed by Fourier transform infrared (FT-IR), showing the spectra of carbon activated at different temperatures (600–900 °C) in Figure S4, Supporting Information. The broad absorption band at 3400 cm–1 indicated the presence of both free and hydrogen-bonded −OH groups on the surface.[19] The definite bands at 1628–1710 cm–1 may be assigned to C=O stretching vibrations of aldehydes, ketones, lactones, or carboxyl groups.[24] However, the band intensity decreased after activation, especially at higher temperatures. It can be noticed that the peak at 1222 cm–1 is observed for all the carbons, as caused by the CO group (a stretch of C–O). The association/dissociation of surface functional groups would determine the density of the surface charge for electrostatic interactions and reactive sites for chemical interactions (e.g., ligand exchange) between the carbon surface and the chemical activation during pyrolysis.[46]

Electrochemical Studies

The PC-modified GCE was electrocatalytically analyzed by electrochemical impedance spectroscopy (EIS) using 5 mM [Fe(CN)6]3–/4– with 0.1 M KCl as the supporting electrolyte versus Ag/AgCl. The EIS profiles of the bare GCE and PC-modified GCEs are depicted in Figure S5a, Supporting Information. In addition, the charge transfer resistances depending on various electrocatalysts were measured. In general, semicircles at high and low frequencies may be attributed to the electron transfer and mass transfer-limited processes associated with ferricyanide ions, respectively.[51] While a bare GCE showed a sizable semicircle with higher charge transfer resistance (Rct), the PC600, PC700, PC800-modified GCE yielded a notable decrease in the semicircle and Rct value. This indicates that the bare GCE gave rise to an inferior electrocatalytic activity. On the other hand, the PC900-modified GCE, which exhibited the smallest semicircle and lowest Rct value compared to other electrodes showed excellent catalytic performance due to the rapid electron transfer process on the carbon surface layers (see Figure S5a, Supporting Information). Several works[25,39,46] have been reported that higher surface areas and larger pore structures are favorable for electrocatalytic applications because the mass transport of the electrolyte is facilitated to obtain higher catalytic currents. In this work, the PC900-modified GCE possesses larger pore structures that enable the electrolyte solution to flow into/out of PC more easily, thus resulting in better electrocatalytic performance. In addition, AC contains a variety of surface functionalities, especially oxygen-based functional groups, which may undergo electrochemically redox reactions. Hydrophilic functionalities (−OH, −COOH, −C=O, etc.) are likely to increase the wettability of the electrode and further facilitate the electrochemical reactions (Scheme ). The related redox reactions on the surface of PC were proposed as follows:[52,53]

Scheme 1

Proposed Redox Reactions on the Surface of PC

Cyclic voltammetry (CV) is a viable electrochemical technique which is used to measure the electrochemical properties of an analyte or system. The CV profiles of KOH AC samples (PC600, PC700, PC800, and PC900) were carried out at a scan rate of 100 mV s–1 using potassium ferricyanide as a redox probe, as presented in Figure S5b, Supporting Information. Owing to its low electrical conductivity, carbon restricted the electron transfer probe ([Fe(CN)6]3–/4–) toward the surface of the GCE. Furthermore, the scan rate influence of the high surface area PC900-modified GCE was studied from 20 to 300 mV s–1 by CV (Figure S5c, Supporting Information) showing that the process is not irreversible because there is no significant shift in the position of the peak potentials at different scan rates.

Notably, the total current obtained at PC900-modified GCE is larger than that at the bare GCE. It may arise from a larger surface area provided by PC to enhance the redox process probed by the electrode. However, the peak current obtained at the GCE is stronger than that at the PC900-modified GCE. The peak current as a function of the scan rate at the PC900-modified GCE is shown in Figure S5d, Supporting Information, yielding a linear relationship between oxidation peak current and the square root of the scan rate at the developed electrode.

Generally, most of the AC materials require the use of polymer binders, such as poly(vinylene fluoride) and Nafion to fabricate the working electrode.[54−56] However, the use of these binders can lead to structural disintegration under chemical attacks, such that reduced electrical conductivity and increased mass transfer resistance may happen. Besides, the involvement of these binders may increase the inner resistance and introduce unnecessary weight.[57,58] Herein, we facilely achieved a binder-free PC900/GCE electrode material and studied the possibility of its utilization in the voltammetric determination of nitro isomers. As such, the electrochemical performances in 50 μM of nitro isomers (3-NA and 4-NA) toward PC900 is directly applied without the binder and with the binder (Nafion) GCE in N2-saturated 0.05 M phosphate buffer solution (PBS) (pH 7.0) aqueous electrolyte (Figure S6, Supporting Information). Furthermore, the CV measurements were performed to study the electrocatalytic activity of as-prepared electrodes toward nitroaniline isomers such as 3-NA and 4-NA. Initially, the PC900-modified GCE sensing toward the individual analytes, 100 μM of 3-NA (Figure a), and 100 μM (Figure b) of 4-NA in N2-saturated 0.05 M PBS (pH 7.0) at a scan rate of 50 mV s–1 in the potential range of +0.4 to −1.0 V versus Ag/AgCl was examined. Then, CV measurements in the co-presence of 100 μM 3-NA and 4-NA were also performed (Figure c).

Figure 6

CV profiles for the bare GCE and different modified electrodes (PC600, PC700, PC800, and PC900) in the presence of 100 μM of (a) 3-NA, (b) 4-NA, and (c) 3-NA and 4-NA with the N2-saturated 0.05 M PBS (pH 7.0) electrolyte. Inset of the panels (a1–c1) indicate the corresponding histogram of the reduction current with different modified electrodes, and (d) CVs obtained for the PC900-modified GCE at varied pHs from 3.0 to 11.0 in the presence of 100 μM of the 3-NA and 4-NA mixture. The insets of the panels (d1 and d2) correspond to the 3-NA and 4-NA peak current (Ipc) and peak potential (Epc) at the PC900-modified GCE, respectively. All measurements were recorded at a scan rate of 50 mV s–1 under N2-saturated 0.05 M PBS solution.

In the case of individual detection of analytes (3-NA and 4-NA), a sharp reduction peak appeared at the bare GCE and PC-modified electrodes. Notably, the bare GCE showed an irreversible reduction, in contrast to the quasi-irreversible reduction with PC600, PC700, and PC800-modified GCE. On the other hand, a distinct reduction peak potential at −0.649 to −0.664 V versus Ag/AgCl was observed with the PC900-modified GCE to reduce 3-NA irreversibly to 3-hydroxylaminoaniline (3-HAA). An additional oxidation peak in the potential range 0.164–0.181 V was observed as attributed to the reversible redox of the 3-HAA to 3-nitrosobenzene (3-NSB).[59] The inset in panel Figure a1 shows the histogram of the reduction current of 3-NA with different modified electrodes.

Similar conclusions may be drawn for the detection of 4-NA (100 μM) in N2-saturated 0.05 M PBS (pH 7.0) at a scan rate of 50 mV s–1 for the bare GCE and PC-modified GCEs as shown in Figure b. Note that the modified electrodes provided a sharp reduction peak in the potential of −0.667 V, which is attributed to the irreversible reduction of the 4-NA to 4-hydroxylaminoaniline (4-HAA).[60−62] Various PC-modified electrodes and their corresponding reduction currents were examined as displayed in the inset of Figure b1. We found a redox peak pair detected at the potential range of +0.133 to +0.116 V, which is ascribed to the reversible redox of 4-HAA to 4-nitrosobenzene. Thus, the GCE and PC-modified GCEs exhibited profound electrocatalytic reduction current at lower peak potential, indeed serving as a superior electrocatalyst for individual detection of 3-NA and 4-NA.

Then, the capabilities of various modified electrodes for simultaneous detection of 3-NA and 4-NA were also assessed. Figure c shows the CVs obtained for the bare GCE and PC-modified electrodes in the presence of 100 μM concentration of 3-NA and 4-NA in N2-saturated PBS (0.05 M, pH 7.0) at a scan rate of 50 mV s–1. The bare GCE exhibited poor redox behavior, indicating the inefficient electrocatalytic capability when 3-NA and 4-NA were mixed for simultaneous detection. In contrast, the PC-modified electrodes exhibited two sharp reduction peaks around at −0.628 to −0.652 V and −0.718 to −0.741 V versus the Ag/AgCl reference electrode, corresponding to the irreversible reduction of 3-NA and 4-NA. Apart from this result, the reversible redox behavior of 3-NA and 4-NA was observed around +0.066 to −0.253 V, indicating that the PC-modified electrode is a suitable oxidation catalyst for simultaneous detection of 3-NA and 4-NA, in good agreement with previous reports.[63,64] The inset in panel Figure c1 shows the reduction current of 3-NA and 4-NA versus different modified electrodes with a trend similar to the individual detection of the sample. From the CV results, a small reduction peak current of 4-NA can be noticed at bare GCE due to the limited electrical conductivity or deficiency of active sites. The legible separate reduction peaks were observed for 3-NA and 4-NA in all the tested electrodes, of which the PC900 electrode yielded the highest reduction peak current for the two samples. The superior electrochemical activity for the PC-modified GCE may be attributed to the high surface area and porosity possessed. Especially, the metal-free PC900 with the functional moieties brings about fast electron transfer during the electrochemical process to provoke high electrocatalytic activity.

Because the electrocatalytic performance of the electrode depends on the pH value of the supporting electrolyte, CV curves were recorded as a function of pH from 3.0 to 11.0 in the presence of 100 μM of the 3-NA and 4-NA analyte in N2-saturated 0.05 M PBS, as shown in Figure d. With an increase in the pH values, the potential shifts toward the right direction. In addition, the reduction current of 3-NA and 4-NA increase from pH 3.0 to 7.0 and then decrease from pH 9.0 to 11.0 (Figure d1). The inset (Figure d2) displays a linear plot of the reduction potential (V) versus pH (3.0–11.0) for 3-NA and 4-NA. The linearity may be expressed by Ipa(μA) = −0.053 (μM) – 0.3148 (R2 = 0.9922) and Ipa (μA) = −0.0517 (μM) – 0.2297 (R2 = 0.9926), respectively. This indicates that a maximum oxidation peak current was observed with an electrolyte pH of 7.0 over the PC900-modified GCE. As such, pH 7.0 was chosen for the supporting electrolyte for subsequent electrochemical sensing studies.

Proposed Mechanism for 3-NA and 4-NA

The electrochemical mechanism of 3-NA and 4-NA at the PC-modified GCE is displayed in Scheme . It is well known that the pathway of nitro compounds detection involves successive four-electron transfer and protonation steps.[62,64]

Scheme 2

Proposed Electrochemical Mechanism of (a) 3-NA and (b) 4-NA at the PC-Modified Electrode

During the CV measurement, catalytic reduction of 3-NA tends to gain four electrons and four protons (4H+, 4e–) at first to form 3-HAA; this is an unstable intermediate, soon followed by releasing reversibly two electrons and two protons (2H+, 2e–) to form nitrosobenzene (3-NSB) and vice versa (Scheme a). As for the reduction reaction of 4-NA, the same processes associated with protons and/or electrons transfer are otherwise involved, as shown in Scheme b. Because the redox process depends largely on the position of the nitro group and pH of the electrolyte,[62,63] the nitro group (−NO2) at the ortho position in 3-NA is expected to have more rapid electron transfer reaction than the nitro group at the para position in 4-NA. The electrochemical process involves the formation of an intermediate compound (3-HAA or 4-HAA) at the 3- or 4-position of the benzene ring due to the transfer of two electrons and two protons. In short, the electrochemical reduction of 3-NA and 4-NA is associated with an equal number of protons and/or electrons transfer.

Feasibility for Individual and Simultaneous Detection

The electrocatalytic activity of the PC900-modified GCE toward the reduction of 3-NA and 4-NA at different concentrations was measured with CV performed under N2-saturated 0.05 M PBS (pH 7.0) at a scan rate of 50 mV s–1. The CV measurements were recorded (Figure a) in the presence of 3-NA at a concentration of 50–250 μM. Under these working conditions, 3-NA exhibited a reduction peak around −0.636 V versus Ag/AgCl, which is ascribed to the reduction of 3-NA to 3-HAA, as mentioned above. Also, during the reduction process, the cathodic current increased at reversible redox of 3-HAA peaking around +0.0661 V, which might correspond to the 3-HAA reduction to 3-NSB.

Figure 7

CV profiles for the PC900-modified GCE in the presence of (a) 50–250 μM [3-NA], (b) 50–150 μM of [4-NA], (c) fixed [3-NA] and [4-NA] in the range of 100–200 μM at a scan rate of 50 mV s–1, and (d) 100 μM of [3-NA] and [4-NA] at varied scan rates (20–300 mV s–1); the insets (a1–c1) are the linear plots of the reduction peak current vs concentration of 3-NA, 4-NA, and their mixture, respectively; the inset (d1) indicates a linear plot of the reduction peak current vs square root of the scan rate for a mixture of 3-NA and 4-NA. All measurements were recorded under the N2-saturated 0.05 M PBS (pH 7.0) electrolyte.

As shown in the inset (Figure a1), a linear plot of the reduction peak current at the potential −0.661 V versus concentration of 3-NA was obtained with a correlation coefficient (R2) better than 0.9969. This potential value was then selected for further detection experiments. Likewise, the reduction peak current as a function of 4-NA concentration (50–150 μM) was acquired with the PC900-modified GCE in N2-saturated PBS (Figure b). A linear plot with a correlation coefficient (R2) of 0.9971 is displayed in Figure b1. Obviously, the reduction peak current increases proportionally when the 3-NA and 4-NA concentration increases, suggesting that the PC900-modified GCE exhibits effective electrocatalytic ability toward the reduction of 3-NA and 4-NA, owing to high active surface area, excellent electronic feature, and good conductivity.

Figure c shows that the CVs obtained for the Ipc increases with increasing a mixed concentration of 3-NA and 4-NA. The corresponding CV results using the PC900-modified GCE in N2-saturated 0.05 M PBS (pH 7.0) and the mixture of 3-NA and 4-NA (100–200 μM) led to linear dependences of cathodic current (Ipc) versus their concentrations, as displayed in the inset (Figure c1). The linearity yielded R2 values of 3-NA (0.9913) and 4-NA (0.9885). These dependencies provide the basis for an amperometric method for the simultaneous determination of 3-NA and 4-NA by voltammetry.

Kinetic Studies

To examine the effect of the scan rate on the reduction peak current, the electrochemical performance over the PC900-modified GCE was conducted in the presence of analytes at different scan rates (20–300 mV s–1; Figure d). The inset of Figure d1 shows that the reduction peak currents increased linearly with the square root of the scan rates. The linear regression equations are found to be

This indicates that the reduction of 3-NA and 4-NA upon the PC900-modified GCE is a diffusion-controlled process.[62] As for an irreversible electrode process, the relationship between the slope of Ipa versus scan rate (V s–1)1/2 is expressed by the Laviron equation as follows:[63]where α is the cathodic electron transfer coefficient, ks is the rate constant of the surface reaction, n is the number of electron transfer, and R, T, and F have their usual notation. The slope of the above equation yields αn of 0.524 and 0.556 for 3-NA and 4-NA, respectively. Generally, α is assumed to be 0.5 for irreversible electrode process[64] and consequently, the values for the number of electrons (n) involved in the reduction of 3-NA and 4-NA is nearby 4.0. In addition, trace amounts of heteroatoms and surface functional groups may create the defects and accelerate the facile electron transfer at the electrode because the space between the electrode and electroactive species is reduced by the π–π stacking interaction between PC900 and an aromatic structure of 3-NA and 4-NA.

Concentration Effect of 3-NA and 4-NA

Apart from its electrocatalytic properties, the PC900-modified GCE demonstrates a higher capacity for analysis of 3-NA (Figure a) and 4-NA (Figure b). This electroanalytical study was conducted using differential pulse voltammetry (DPV). All electrochemical parameters were initially optimized in N2-saturated 0.05 M PBS (pH 7.0), containing 0.2–115.6 μM of 3-NA. The inset of Figure a shows the calibration curve for 3-NA and 4-NA reduction. The cathodic current peak height of 3-NA increases linearly with its concentration in the range of 0.2–115.6 μM of 3-NA. The minimum concentration calculated to be detected by the proposed electrode is 0.0551 μM (S/N = 3) in terms of limit of detection (LOD). Furthermore, we calculated the linear regression: Ipa (μA) = −0.099 (μM) – 1.402 with the sensitivity of 4.371 μA·μM–1 cm–2. Likewise, Figure a displays the DPV profiles of 4-NA, showing that the reduction current increases with increasing 4-NA concentrations from 0.5 to 120 μM. The inset in panel Figure b shows that the linear range was obtained for reduction current as a function of concentration of 4-NA following a linear regression: Ipa (μA) = −0.129 (μM) – 1.015 with a low detection limit (LOD) of 0.0326 μM (S/N = 3) and a sensitivity of 5.792 μA·μM–1 cm–2. The good analytical performance of the sensor is due to the fast electron mobility and strong electron donating nature of the high surface area PC matrix toward the reduction of nitro isomers.

Figure 8

DPV response of the PC900-modified GCE in N2-saturated 0.05 M PBS (pH 7.0) under varied analyte concentrations of (a) [3-NA] 0.2–115.6 μM and (b) [4-NA] 0.5–120 μM. The inset panels (a1 and b1) show the corresponding calibration plots of the reduction peak current vs concentration for 3-NA and 4-NA, respectively.

Selectivity and Stability of the Sensor

Furthermore, we evaluate the selectivity of the sensor for detection of 3-NA in the presence of 50 μM 4-NA (Figure a) or 4-NA in the presence of 50 μM 3-NA (Figure b); DPV measurements were performed using the PC900-modified GCE. Obviously, 3-NA increases its reduction current with concentration, while the concentration of 4-NA was fixed constant (50 μM). As shown in the Figure a1 inset, the reduction peak current of 3-NA is linear over the concentrations ranging from 2.0 to 69 μM with an R2 value of 0.9963. The linear regression equation is Ipa (μA) = −0.070 (μM) – 1.708, and LOD was estimated to be 0.53 μM (S/N = 3). Furthermore, the reduction current of 3-NA was directly proportional to its concentrations, while the reduction current of 4-NA barely changed. Likewise, Figure b reveals that the reduction current of 4-NA increased with its concentration, while the concentration of 3-NA was at constant (50 μM). In addition, as shown in Figure b1, the linear regression equation Ipa (μA) = −0.122 (μM) – 3.123 was obtained between the reduction current and concentration of 4-NA with an LOD of 0.7356 μM (S/N = 3). The analytical sensing performance of the proposed sensor was compared with previously reported 3-NA sensors, and the results are summarized in Table .

Figure 9

DPV profiles of the PC900-modified GCE in N2-saturated 0.05 M PBS (pH 7.0) under varied concentrations of (a) 3-NA and 50 μM 4-NA, (b) varied concentrations of 4-NA and 50 μM 3-NA, and (c) equal concentrations of 3-NA and 4-NA. The inset panels (a1–c1) show the corresponding calibration plots of the reduction peak current vs concentrations.

Table 1

Comparisons of Analytical Parameters for Detection of 3-NA among Various Modified Electrodes

electrode	method	electrolyte/pH	linear range (μM)	LOD (μM)	refs
poly-DHCBAQSa/Grb-NFc/GCE	DPVg	B-Rh/2.0	0.36–4.34	0.152	(62)
SWCNTsd-ILGe/GCEf	DPV	PBS/7.0	0.01–7	0.008	(65)
PC900/GCE	DPV	PBS/7.0	0.2–115.6	0.0551	this work

7-[(2,4-Dihydroxy-5-carboxybenzene)azo]-8-hydroxyquinoline-5-sulfonic acid.

Graphene.

Nafion.

Single-walled CNTs.

Ionic liquid gel.

GCEs.

Differential pulse voltammetry.

Britton–Robinson buffer solution.

Furthermore, we explore the sensor for practical applications, and the simultaneous determination of 3-NA and 4-NA of the PC900-modified GCE was evaluated in the presence of equal concentration of both analytes (3-NA and 4-NA). The DPV profile shows that the individual reduction peaks sharply increases when their concentrations were increased equally, as displayed in Figure c. The inset in the panel of Figure c1 displays two linear regression equations Ipa (μA) = −0.050 (μM) – 2.057 and Ipa (μA) = −0.077 (μM) – 3.635 for 3-NA and 4-NA, respectively. However, the LOD of 1.2371 μM (S/N = 3) for 3-NA and 1.7254 for 4-NA are calculated.

PC-based electrodes have been widely used in voltammetric studies because of low cost, availability, stability, and easy modification of the carbon morphology. There are a number of carbon-based electrodes such as GCE,[66] carbon nanotubes (CNTs),[67] and most recently, graphene.[73,74] Moreover, the proposed PC900-modified GCE-based sensor indeed exhibits superior analytical performances with a wide linear range and LOD compared to other previously reported electrode materials for electrochemical detection of 4-NA, as listed in Table .[61,65−75]

Table 2

Comparisons of Analytical Parameters for Detection of 4-NA among Various Modified Electrodes

modified electrodes	method	electrolyte/pH	linear range (μM)	LOD (μM)	refs
poly-DHCBAQSa/Grb-NFc/GCEd	DPVo	B-Rs/2.0	0.36–4.34	0.137	(62)
Ag electrode	DPV	B-R/2.0	0.008–1000	0.00474	(66)
GCE	DPV	B-R/2.0	2–100	0.2	(67)
BMIMPF6-SWNTe/GCE	LSVp	PBS/4.0	0.01–7	0.008	(68)
SWNT/GCE	DPV	PBS/4.0	0.5–10	0.2	(68)
DTDf/AgNPs/CPEg	DPV	PBS/4.0	1.0–100	0.23	(69)
AgNPs-POSSh/rGOi/GCE	DPV	PBS/7.2	0.7–551.6	0.36	(70)
Ag/CPE	DPV	B-R/2.0	0.08–100	0.0418	(71)
GC/TPDTj-SiO2k/AgNPs	SWVq	PBS/7.2		0.50	(72)
ZnO NRsl/FTOm	AMPr	PBS/7.0	1–80	0.5	(73)
Ag@Pd NRDs/L@ERGOn	DPV	0.5 M KCl	0.001–0.012	0.000018	(74)
Pt–Pd NPs/CNTs–rGO	DPV	0.5 M KCl	0.007–832.6	0.004	(75)
Ag/CPE	CV	B-R/2.0			(76)
PC900/GCE	DPV	PBS/7.0	0.5–120	0.0326	this work

7-[(2,4-Dihydroxy-5-carboxybenzene)azo]-8-hydroxyquinoline-5-sulfonic acid.

Graphene.

Nafion.

GCE.

1-Butyl-3-methylimidazolium hexafluorophosphate with single-walled CNT gel.

6,7,9,10,17,18,19,20,21,22-Decahydrodibenzo[h,r][1,4,7,11,15]trioxadiazacyclonanodecine-16,23-dione.

Carbon paste electrode.

Polyhedral oligomeric silsesquioxane.

Reduced graphene oxide.

N-[3-(Trimethoxysilyl)propyl]diethylenetriamine.

Silica.

Zinc oxide nanorods.

Fluorine-doped tin oxide.

Electrochemically reduced graphene oxide.

Differential pulse voltammetry.

Linear sweep voltammetry.

Square wave voltammetry.

Amperometry.

Britton–Robinson buffer solution.

Practical Analysis, Reproducibility, and Selectivity

The practicality of the PC900-modified GCE was tested for 3-NA and 4-NA contained in different real samples such as cold drinks (Figure a) and pineapple jam samples (Figure b). The experiments were carried out with the additions of 3-NA and 4-NA at the spiking level of 5, 10, and 20 μM in blank cold drinks and pineapple jam samples. As summarized in Table , the recoveries of the analyte in cold drinks and pineapple jam samples are from 94.80 to 98.87%. These results indicate that the electrochemical method proposed in this work should be suitable for sensing 3-NA and 4-NA in various food products.

Figure 10

DPV profiles for PC900-modified GCE (a) 3-NA- and 4-NA-spiked beverage drink sample, (b) pineapple jam sample, and (c) 0.5 mM of potential interference species and 50 μM and 75 μM of 3-NA and 4-NA. All measurements were recorded under the N2-saturated 0.05 M PBS (pH 7.0) electrolyte.

Table 3

Practical Analysis of PC900-Modified GCE for 3-NA + 4-NA-Spiked Samples

sample	added (μM)	found (μM)	recovery (%)
beverage (orange)	10	9.98	99.0
	20	20.03	101.5
	30	30.02	101.0
pineapple jam	10	10.01	100.5
	20	20.05	102.5
	30	30.01	100.5

The reproducibility of the PC900-modified GCE in presence of 3-NA and 4-NA was evaluated showing the relative standard deviation of 6.3% (n = 10) according to the current intensity for the concentration of 100 μM 3-NA and 4-NA. Such reproducibility is highly acceptable for an electrochemical detection of 3-NA and 4-NA. The stability was inspected by acquisition of the electrode response with the same concentration of 4-NA and 3-NA within 15 days interval. The current response was found to decrease to 93% after 30 days in the N2-saturated 0.05 M PBS (pH 7.0) electrolyte system, which was stored at 4 °C for each experiment. The unchanged curve shape and current after 30 days of successive scanning indicated that the fabricated carbon electrode material (see Figure S7, Supporting Information) can be used for nitro isomer long-term sensing applications with excellent electrochemical stability. Additionally, the selectivity of the sensor was examined using a PC900-modified GCE by DPV in N2-saturated PBS (0.05 M, pH 7.0) solution with 3-NA and 4-NA together in the co-presence of 0.5 mM in other possible electroactive interferences such as Cu2+, Hg2+, Pd2+, Fe2+, Cd2+, Cl–, and NO32–. Obviously, nearly no change in their respective peak potential was found with additions of 3-NA and 4-NA (Figure c); these results further reveal that the PC900-modified electrode may be exploited as a sensitive and selective electrochemical sensor for 3-NA and 4-NA.

Conclusions

In summary, activated tree leaves-derived PC has been obtained from L. formosana and applied as the binder-free electrode material for the detection of nitroaniline isomers (3-NA and 4-NA). The PC samples at varied carbonization temperatures have been characterized by XRD, Raman, BET, SEM, FE-TEM, EDX, CV, and DPV techniques. We found that mesoporous PC900 owns a very high surface area, resulting in better electrocatalytic activities for the reduction of nitroaniline isomers (3-NA and 4-NA) than the other PC-modified electrodes. The developed portable electrochemical sensor (PC900-modified GCE) shows attractive analytical features such as good sensitivity, widely operational linear range, LOD, and excellent selectivity. The surface area effect and functional moieties on their surface are responsible for the higher sensitivity with the ultratrace detection limit. Additionally, the PC900-modified GCE demonstrates good reproducibility and excellent long-term stability. Furthermore, the presence of 3-NA and 4-NA in the beverage and jam was determined, yielding satisfactory results. Consequently, our proposed electrode can be used for the sensitive and accurate determination of nitroaniline isomers in food samples.

Experimental Section

Chemicals

3-NA (96%, Sigma-Aldrich) and 4-NA (96%, Sigma-Aldrich) (KOH, Acros) were obtained commercially and used without further purification. The PBS at pH 7.0 was prepared by using 0.05 M Na2HPO4 and NaH2PO4 solutions. The pH of the solutions was adjusted with 0.5 M H2SO4 and 2.0 M NaOH. A conventional three-electrode cell system contained a modified GCE (working electrode), Ag/AgCl (in saturated KCl; reference electrode), and a platinum wire (counter electrode). All other chemicals used belonged to analytical grade, and all solutions were prepared by using ultrapure water (Millipore).

Preparation of Tree Leaves Carbon

Dry L. formosana (sweetgum) precursors were gathered during the fall season from the campus of National Taiwan University in Taiwan. In this work, hierarchical PCs were successfully prepared by using fallen dead tree leaves as the carbon source (Scheme ). The pretreatment of the dry leaves can be processed by the following three steps. (i) The leaves were washed with warm distilled water to remove impurities and undesirable materials and dried at room temperature for 2 days. The resulting leaf powders turn to brown color. (ii) Then, the brown powders (2.0 g) were dispersed in 25 mL of 10% KOH solution. The mixture was sonicated for 1 h to achieve homogeneity and then dried in atmospheric air conditions. The carbonization and activation processes were carried out at a single step. The dried leaf powder was heated at 600–900 °C with a heating rate of 10 °C min–1 in a tubular furnace under the N2 environment for 2 h. The samples were left to cool to ambient conditions and removed from the furnace prior to further analysis. Notably, the brown powder was totally turned into black. The obtained carbons were washed thoroughly by deionized water and then dried in an oven overnight. For comparison, some products were rinsed by a diluted hydrochloric solution (0.5 M). The resulting sample was placed in air and dried overnight at 110 °C. PC600, PC700, PC800, and PC900 were labeled for activated PC prepared from dead tree leaves carbonized at 600, 700, 800, and 900 °C.

Scheme 3

Schematic Diagram for the Preparation and Application of PCs

Electrochemical Measurements

A three-electrode system as mentioned above was employed for the electrochemical performance. The electrochemical analysis depended on a computerized electrochemical analyzer (CH Instruments; CHI 1205b and CHI410a) workstation. Prior to each measurement, N2 gas was purged through the electrolyte for 30 min. EIS (IM6ex ZAHNER, Kronach, Germany) was used to characterize the electrochemical properties of the as-prepared composites. To prepare PC-modified GCE, ∼2.0 mg of PC powder was dispersed in 1.0 mL of water and sonicated for 2 h. Then, 8.0 μL of the aliquot was dropped onto GCE to dry in air, followed by gentle heating in a hot air oven for 25 min without any binder used.