The adsorptive properties of oxidized activated carbons and their applications as carbon paste electrode modifiers

The 4-chlorophenol (≥ 99%) was purchased from Sigma-Aldrich (USA). All other chemicals used were of analytical grade and were obtained from Avantor Performance (Poland). Carbon paste electrode components—the graphite powder (< 45 µm) was from Sigma-Aldrich (USA), while the mineral oil was received from Fluka (Switzerland). The adsorbent samples were prepared from the Organosorb-10 activated carbon (Desotec, Belgium) which was demineralized with concentrated hydrofluoric and hydrochloric acids. So treated activated carbon (OS-NM) was oxidized with ozone at 25 °C in a fixed bed reactor loaded with 2 g of carbon under a constant ozone flow of 15 dm³/min (Hailea HLO-820A ozone generator, China). Two different exposure times were used—15 min (OS-O3-15) and 45 min (OS-O3-45), respectively.

The structural properties of the activated carbon samples were characterized using low-temperature N₂ adsorption–desorption isotherms (ASAP 2010, Micromeritics, USA). The specific surface area (S_BET) was calculated by using BET method, the total pore volume (V_t) was obtained from the amount of nitrogen adsorbed at P/P₀ = 0.99, the pore volume of micropores (V_mi) was estimated from t-plot method. The pore volume of mesopores (V_me) was calculated from the difference of both values. For characterization of carbon materials the other techniques were also used: scanning electron microscope (SEM-EDS, LEO 1430VP, Electron Microscopy Ltd., UK), thermogravimetry (TG-DTA/DSC, STA 449 Jupiter F1, Netzsch, Germany) and the X-ray photoelectron spectroscopy (XPS, Multi-chamber UHV Analytical System, Prevac (2009) with the hemispherical analyzer Scienta R4000). During XPS analysis the monochromatic 450 W Al Kα radiation from high-intensity source MX-650 (1486.6 eV Gammadata Scienta) was applied. The average depth of analysis was approximately 5 nm. The data processing and high-resolution deconvolution of C1s and O1s peaks were done by CasaXPS software.

All the adsorption experiments were carried out at 25 °C in Erlenmeyer flasks filled with 0.04 L of 4-chlorophenol solutions (the initial concentration of 4-CP was in the range of 0.25–2.0 mmol/L). Appropriate amounts of adsorbent (20 mg) were added and the solutions were agitated at 100 rpm for 8 h. Then the mixtures were filtered through a filter paper and the obtained filtrates were analyzed for 4-CP concentration (C_e) using UV–Vis spectrophotometry (Varian Carry 3E, USA). The amount of 4-CP adsorbed at equilibrium per unit mass of activated carbon (q_e) was calculated using the following formula: where: C₀ (mmol/L) is the initial 4-CP concentration, C_e (mmol/L) is the equilibrium 4-CP concentration, m (g) is the mass of the adsorbent used and V (L) is the solution volume.

The experiments were carried out in duplicate, and the data obtained were used for analysis.

The electrochemical measurements were conducted using an Auto-Lab apparatus (PGSTAT 20, EcoChemie B.V., The Netherlands) in a three-electrode configuration system: carbon paste electrode (as a working electrode), SCE—saturated calomel electrode (as a reference electrode) and platinum wire (as a counter electrode). The carbon paste electrodes were prepared according to the previously described procedure (Skrzypczyńska et al. 2016). The bare (graphite) electrode as well as the CPEs containing 2.5%, 5% and 10% by mass of all of the activated carbon samples were prepared and tested.

The specific surface areas of the activated carbon samples were determined from the nitrogen adsorption–desorption isotherms at 77.4 K (Fig. 1). The values of S_BET as well as micropore and mesopore volumes are given in Table 1. In order to characterize the surface chemistry of the activated carbon samples (oxygen content), analyses were performed by SEM/EDS, TG and by standard neutralization with 0.1 mol/L NaOH solution (NaOH uptake indicating changes in acid properties) (Table 2). Results of samples surface analyses by XPS are presented in Fig. 2 as well in Tables 3 and 4.

The XPS analysis was applied for obtaining quantitative and chemical state information from the surface of the carbon materials. The major signals observed in XPS spectra were identified as C1s and O1s signals with the peak position at 284.7 eV and 532–533 eV respectively. The general atomic concentration (%at) of oxygen increases from 3.6 to 5.8% and 7.13% for OS-NM, OS-15 and OS-45 samples respectively. Thus, the atomic concentration of carbon is 95.9%, 93.7% and 92.3% for OS-NM, OS-15 and OS-45 samples, respectively. The increase of oxygen content reflects changes of the carbon surface during the oxidation process.

The high resolution XPS spectra illustrate the chemical states of carbon and oxygen atoms as C1S and O1s respectively. The C1s and O1s regions from three carbon materials are shown in Fig. 2a, b. The C1s peak has been deconvoluted into four Lorentzian peaks with binding energies at approximately 284.7 eV, 285.4 eV, 286.0 eV and 286.5 eV and marked as C 1sA, C 1sB, C 1sC, C 1sD features. In this case, the observed binding energies for the extracted components depend on the specific functional groups located on the carbon surface. Moreover, the intensity of the core-level signals can be proportional to the atoms density. The first C1sA signal with the lowest binding energies (284.73 eV, 284.12 eV and 284.69 eV for OS-NM, OS-15 and OS-45 samples respectively) can be assigned as carbon–carbon interactions. In details the binding energy values from 284 eV to 285.0 eV can specify the sp3 and sp2 bonded carbons state typical for graphitic carbons and C=C, C–C and C–H species (Polovina et al. 1997; Gai et al. 1989). In all cases, the C1sA signals are the strongest what suggest the C–C bonds are dominant in the investigated carbon layer. The C1sB signals with the position of binding energy of 285.56 eV, 285.46 eV and 285.42 eV for investigated samples were shifted by more than 0.7 eV from the position of first C1sA signal. The C1sB components are assigned to the C atoms directly bonded with the O atoms from hydroxyl groups (C–OH) groups (Stankovich et al. 2007; Mattevi et al. 2009). The amount of carbon-hydroxyl interactions decrease significantly during the oxidation process. The atomic concentration of C1sB peaks is the greatest for OS-NM sample (5.6%) and the lowest for the OS-45 sample (1.6%). At the same time, the amount of carbonyl C=O groups increases significantly. The atomic concentration (%at) of C1sD peaks (typical for carbons in carbonyl groups) increases from 6.3 to 10.5% and finally to 11.1% for OS-NM, OS-15 an OS-45 samples, respectively. The amount of the components at ~ 286.0 eV [attributed to epoxide or ether groups (C–O–C)] are very similar for all samples.

The O1s peak is a result of the imposition of five component peaks. First O1sA signal is typical for oxygen in oxide form. Due to the chemistry of the carbon surface, this type of oxygen interaction is not considered, it is not meaningful in these studies. The C1sB signals with the position of 531.01 eV, 531.09 eV, and 531.02 eV can be assigned as oxygen-carbon interaction according to carbonyl O=C and carboxylic species. The atomic concentration of this type of functional groups increases considerably for oxidized samples. The atomic concentration of carbonyl groups increases from 34.12 to 49.0% and finally to 59.7% for OS-NM, OS-15 and OS-45 samples, respectively. Moreover, decreasing of the hydroxyl groups (C1sC and C1sD signals) was observed. Such phenomenon was confirmed previously by high-resolution core level analysis of carbon atoms.

The results show only small changes in the porous structure of the oxidized activated carbons but significant changes in their surface chemistry. The ozone treatment decreased the specific surface area, S_BET, of the carbons; this reduction as well as the increase in oxidation efficiency was dependent on the duration of the modification. Modification in a shorter time (15 min) changes slightly the porous structure of activated carbon, while increase of time to 45 min gives a pronounced effect (greater decrease of porous structure parameters). However, these changes are not very large. In the case of surface chemistry, the situation is different. Oxidation even for a short time (15 min) gives significant changes in the chemistry of the carbon surface, while the extension of the process to 45 min already causes a smaller change, although still quite significant. Similar effects were observed in other papers (Valdes et al. 2002).

The adsorption isotherms of 4-CP from aqueous solutions onto the activated carbons are presented in Fig. 3. As can be seen, the oxidation process influences strongly the adsorption effectiveness; the higher number of oxygen surface groups the lower adsorption value. For better interpretation of the results, the experimental data for 4-CP adsorption were analyzed using the Langmuir (2), Freundlich (3) and the three-parameter Sips (4) isotherm equations (Hamdaoui and Naffrechoux 2007):

where: q_m (mmol/L) and b (L/mmol) are the Langmuir constants related to the maximum adsorption capacity and adsorption energy, respectively; K_F ((mmol/g)(L/mmol)^1/n) and n are the Freundlich constants related to the adsorption capacity and adsorption intensity, respectively; q_mS (mmol/g) is the Sips adsorption capacity, while K_S (L/mg) and m are the Sips model constants. All theoretical isotherm parameters were evaluated by non-linear regression (OriginPro 7.5 software) and are listed in Table 5. A high value of the correlation coefficient indicates better agreement between experimental and predicted data using the Langmuir (R² ≥ 0.990) equation than the Freundlich (R² ≥ 0.948) and Sips (R² ≥ 0.981) isotherm models.

The adsorption of phenol and its derivatives (including 4-CP) on activated carbons from aqueous solutions was the subject of many research (Dąbrowski et al. 2005; Hamdaoui and Naffrechoux 2007; Hameed et al. 2008; Lorenc-Grabowska et al. 2010; Deryło-Marczewska et al. 2011; Kuśmierek and Świątkowski 2015; Kuśmierek et al. 2017). This process occurs as a result of π–π type dispersive adsorbate-adsorbent interactions, the formation of hydrogen bonds and/or acceptor–donor complexes (Salame and Bandosz 2003; Lorenc-Grabowska et al. 2010; Deryło-Marczewska et al. 2011). It is well known, that the adsorption of organic compounds on the activated carbons depends on their porosity including surface area, pore volume, and pore size distribution as well as on their surface chemistry. The calculated parameters related to the adsorption capacity (q_m,K_F, and q_mS) decrease in the order OS-NM > OS-O3-15 > OS-O3-45. 4-CP was adsorbed better on the non-oxidized activated carbon than on the OS-O3-15 and OS-O3-45 materials. This fact is closely correlated with the reduction of the specific surface area of activated carbons as well as mainly with the increase in the content of acidic oxygen groups on the adsorbent surface. The oxygen surface groups decrease the surface hydrophobicity and the π-electron density in the graphene layers and consequently reduce the dispersive adsorption potential of the activated carbon (Salame and Bandosz 2003; Álvarez et al. 2005; Deryło-Marczewska et al. 2011).

The electrochemical behavior of the prepared CPEs was investigated using cyclic voltammetry (CV). In the first stage, optimization studies have been conducted. These investigations were performed using the electrodes containing the highest modifier additive (10%).

For the purpose of determining the electroactive surface area of all electrodes the electrochemical behavior of potassium ferrocyanide in 1.0 mol/L KCl supporting electrolyte was studied. The cyclic voltammograms of 2.0 mmol/L Fe(CN)₆³⁻/Fe (CN)₆⁴⁻ in 0.1 mol/L KCl are presented in Fig. 4 (scan rate 100 mV/s). The peak current for a reversible process is described by the Randles–Sevcik Eq. (5):

where: A—the area of the electrode (cm²), n—the number of electrons participating in the reaction (equal to 1), D—the diffusion coefficient of the molecule in solution, C—the concentration of the probe molecule in the solution (2 mmol/L) and ν—the scan rate (V/s).

The redox couple of Fe(CN)₆³/Fe(CN)₆⁴⁻ appeared to vary with the use of different electrodes. Based on the results presented in Fig. 4 the value of the active electrode area was found to be 0.145 cm², 0.102 cm² and 0.083 cm² for OS-NM, OS-O3-15 and OS-O3-45, respectively.

The influence of the accumulation time on the peak current was also determined. Accumulation time was varied from 1 min to 10 min and the corresponding current value was measured using one concentration of 4-CP (0.1 mmol/L). The peak current increased with the increase of accumulation time up to about 4 min and then becomes stable. For further research, 5 min were selected as the optimal accumulation time. The possible memory effect of the CPEs was also examined. To evaluate the memory effect of the electrode the measurements were performed in the sequence of high-to-low 4-CP concentration (0.5 and 0.05 mmol/L). The signal response showed good repeatability. No memory effect was observed.

The effect of scan rate (10–100 mV/s) on the peak current and peak potential of 4-CP was evaluated. The scan rate increased with increasing anodic peak current. The anode current recorded at a potential of about + 0.75 V vs. SCE increased linearly with an increase of square root of scan rate as shown in Fig. 5. The values of the R² for the anodic peaks were 0.989; 0.993 and 0.994 for OS-NM, OS-O3-15 and OS-O3-45, respectively. The logarithm of the peak current vs. logarithm of the scan rate suggested that the oxidation process is predominantly diffusion-controlled.

The oxidation process of the 4-CP on the CPE occurred at the potential value of approximately + 0.82 V vs. SCE, and only one peak appeared. The current corresponding to the peak oxidation of 4-CP increased with increase in the active area of the electrode and/or with increase in the adsorption capacity of the modifier used (CPE/OS-O3-45 < CPE/OS-O3-15 < CPE/OS-NM). The measured peak currents were strongly dependent on oxidation of activated carbons and correlated with the 4-CP adsorption efficiency. The results are presented in Fig. 6.

Figure 7 shows the effect of the amount of modifier added (2.5, 5.0 and 10% m/m) on the 4-CP peak current in the studied concentration range of 0.05–0.5 mmol/L. The linearity of the methods was tested, the calibration plots were constructed by plotting peak current against 4-CP concentration and the plots were fitted by least-squares linear regression analysis. On the basis of the obtained results, it can be concluded that increase in the addition of the material (from 2.5 to 10% by weight) results in an increase in the oxidation peak height. The oxidation potential occurs at a value of approximately 0.81 V. For comparison, linearity of the bare CPE/graphite was investigated under the same experimental conditions and the equation for regression line was y = 0.12x − 0.004 (R² = 0.987). Compared to unmodified (CPE/graphite) electrode, it can be stated that even a relatively small addition of a modifier to CPE gives significant effects and increases the sensitivity of electroanalytical methods. The limit of detection (LOD) was determined using the following equation:

where SD_blank is the standard deviation of the blank and a is the slope of the calibration curve (y = ax + b). The best sensitivity was observed for the carbon paste electrode modified with OS-NM; the LOD was found to be 2.38 µmol/L and was over 30 times lower than for the bare (graphite) electrode (Table 6). Compared with other electrodes and electrochemical sensors (Table 6), the cheap and facile fabricated CPEs exhibited relatively good sensitivity. The sensitivity of these methods is comparable to, or slightly worse than those of electrochemical methods for 4-CP detection described by others. However, the limit of detection for 4-CP shows worse sensitivity compared with advanced techniques, such as HPLC but that requires time-consuming sample preparation and using organics. The LOD of 4-CP in HPLC was 0.006 mg/L (Higashi 2017), while in Headspace-SPME was 0.23 µg/mL (Djozan and Bahar 2003).