In order to investigate the influence on the textural properties of the surface of activated carbons, isotherms of adsorption–desorption of N₂ at 77 K were determined and are shown in Figure 2.

The isotherms were type I according to the IUPAC classification. The major consumption of N₂ in the adsorption–desorption isotherms of the activated carbons occurred at relatively low relative pressure (<0.2) and reached a plateau at high relative pressure. These results indicate that the activated carbons obtained by the various treatments were microporous. In the case of activated carbon treated with acid, the amounts of N₂ adsorbed decreased with an increasing concentration of HNO₃, as in those treated with CuO, probably because this treatment produced highly blocked pores in the activated carbon. With regards CAN83, there was a decrease in its area but not to the same magnitude as the other carbons. In the case of treatment with HNO₃ for 8 h, the surface area of the activated carbon was still considerable.

The CuO/CAC catalysts that were prepared by impregnation of the pre-treated CAC with an aqueous solution of the precursor (copper nitrate, Cu(NO₃)₂ · 3H₂O) showed a decrease in surface area, but the volume of micro and meso pores was of the same order of magnitude as the other porous materials that were synthesized.

The pore characteristics of activated carbons based on N₂ adsorption–desorption isotherms are shown in Table 1.

The CAC was highly microporous and had a high BET surface area of more than 1400 m²/g. In the case of acid-treated activated carbons, the BET surface area decreased following treatment with HNO₃ solutions. It is assumed that the porous structure of activated carbon is destroyed partially by the strong oxidative effect of HNO₃ solutions and that micropores are not detected. Table 1 reports the surface area calculated by the BET method, the total volume, and the mesoporous and microporous volume of each activated carbon.

The numbers of acidic groups on the activated carbons as estimated by a titration method [13] are shown in Table 2. The treatment of a carbon surface with HNO₃ solutions results in an increase in the amount of oxygented groups, both acidic and basic, on that surface. A higher number of oxygenated groups were found on CAC compared with the other activated carbons obtained in this work, which shows the effect of the oxidizing agent HNO₃ as reported in the literature.

Table 2 shows that in the activated carbons, the amounts of oxygenated groups resulting from the HNO₃ treatment increases with time, and a similar behaviour is observed after treatment with copper ions as the amount of oxygenated groups also increases. This amount of copper deposited on the carbon surface is greater, and a corresponding enhanced catalytic effect of these ions in the decomposition of propanethiol may be expected.

The number of oxygenated groups in each of the samples of activated carbon decreased, as shown in Table 2. The basicity of the activated carbons modified with nitric acid was less than that in CAC in all cases. The activated carbons modified with copper nitrate showed higher acidity since copper has a higher affinity with the base, consuming more moles of sodium hydroxide. The total acidity increased with the increasing number of oxygenated groups.

The results on the immersion enthalpy of the activated carbons in water and propanethiol are shown in Table 3. The surface chemistry is one of the variables that most affects the magnitude of the heat generated by carbon immersed in the test liquids [13]. High values of immersion enthalpies of activated carbon in propanethiol were obtained for CACuO (0.25) and CACuO (0.5). The modified activated carbons did not show a significant interaction because the thermal effect that develops when the carbons are immersed in propanethiol is low; the enthalpy of the samples increased without increasing the concentration of the activating agent or the duration of modification. Considering the order of magnitude of the enthalpic values of the adsorbent–adsorbate interaction, this may be considered a physiorption [18].

This is because in this case there is an interaction of a chemical, when comparing the orders of magnitude between the enthalpy values is obtained in propanethiol values are higher compared to water. In addition, we see the effects the different treatments have on the activated carbon, which can be characterized by immersion microcalorimetry.

Figure 3 shows the XRD patterns of the support catalysts CACCu(0.25) and CACCu(0.5). The XRD pattern for copper oxide has peaks at angles of 35.7 and 38.55 of 2θ; these angles of reflection indicate the formation and presence of copper oxide particles on activated carbon and are consistent with the literature [13]. Figure 2 shows these peaks for CACuO(0.25) and CACuO(0.50).

Analysis of the spent adsorbent material by powder X-ray diffraction revealed very strong reflections with the preferred orientation of copper(I) propanethiolate crystals (Figure 2). The strongest intensity was measured in CACCu(0.5) before and after top adsorption and decomposition of propanethiol, indicating the formation of relatively large crystals. Similar diffraction patterns are shown for other thiolate crystals in reference [20], but with what seem to be much lower intensities, confirming the redox chemistry between Cu²⁺ and the propanethiol in the solutions in hexane.

Based on Bragg’s law of diffraction and the peak positions for the crystals, the interlayer spacing for the layered copper(I) propanethiolate crystals was 35.7° and 38.55°, which is in close agreement with the values reported in the literature.

Therefore, the conclusion is that in the bulk of the copper is utilized in this process. As already shown, the dispersion of the copper particles enhances the total capacity for sulphur adsorption. In light of the above observations and comparing these data with the chemical mechanisms that were reviewed in the Introduction, Figure 4 illustrates the proposed primary mechanism for the adsorption-decomposition of mercaptans using the activated carbon-copper as catalyst, based on the proposal by Turbeville et al. [24].

Figure 5 is provided as an illustration of a proposed primary mechanism for adsorption of mercaptans onto copper-containing adsorbents in these conditions. First, the well-known redox chemistry occurs between the CuO and mercaptan, resulting in a disulphide and a Cu₂O site on the surface. This is followed by a displacement reaction of the Cu₂O with the mercaptan, in this case, forming copper(I) butanethiolate on the surface. It should be noted that at this point, a SAM has effectively been formed on the surface. Indeed, SAM formation on the surface of CuO has been reported to occur [2,9]. Since copper(I) propanethiolate crystals were formed separately from the adsorbent materials, it is necessary for the surface species to dissolve into the feed liquid stream and migrate to crystallize at another site. This mechanism is therefore quite effective in uncovering bulk CuO in the material, making it available to participate in the adsorption of the mercaptan.

The reducibility of the synthesized catalysts was evaluated by TPR measurements in a H₂ flow, and the results are shown in Figure 5. The main reactions showed two peaks in the temperature range from 150 to 300 °C. These peaks were interpreted as corresponding to the stepwise reduction of CuO [15]. The peak at the lower temperature corresponds to the reduction of Cu²⁺ to Cu⁺ and the peak at the relatively high temperature corresponds to the reduction of Cu⁺ to Cu⁰ [17]. In the case of CACCu(0.5), the first peak was observed at a lower temperature (185 °C, TR₁) than that of the CACCuO(0.50) catalyst (197 °C), because the change in the oxygenated content promotes hydrogen consumption at a lower temperature range. Similarly, the introduction of more oxygenated groups induces a shift in the first and second peaks towards a lower temperature. These peak shifts indicate that the reduction of CuO to Cu occurred more easily in the CACuO(0.25) and CANCu(0.50) catalysts because of the development of acid and basic groups that increase oxygen migration and permit the oxygen emitted from CuO to migrate more actively to the catalyst surface.

Figure 5 shows that the activated carbon with the highest quantity of Cu is the CACuO(0.5), which shows a higher intensity than the activated carbon modified with 0.25 M copper nitrate.

Figure 6 shows the adsorption of propanethiol on the activated carbons as a function of initial concentration.

The amounts of propanethiol adsorbed on the original activated carbon (AC) increased with time and reached a plateau, as shown. Propanethiol adsorbs physically, e.g., by a dispersive force, on the surface of acid-metal-treated activated carbon. The amounts of propanethiol adsorbed on treated activated carbon also increased with time. In the case of acid-treated activated carbons, the adsorption of propanethiol increased with the concentration of HNO₃ solutions used for treatment, more so than in those treated with copper.Since the number of acidic groups on activated carbons increased with the concentration of HNO₃ solutions, the amount of propanethiol adsorbed must be influenced by the acidic groups. With regards the effect of the carboxyl group, the catalytic effect on the oxidation of methyl mercaptan to dimethyl disulphide by oxygen in air under dry and wet conditions has previously been reported, with values consistent with those obtained in this work. Therefore, other factors, e.g., hydrogen bonding between carboxyl groups and thiol groups of methyl mercaptan rather than oxidation of methyl mercaptan, are thought to play a role in the increase in the adsorption of propanethiol.

Table 4 shows the adsorption constants obtained when applying the Freundlich model to the experimental data [8,9]; the K_F values for the CAN86 and CACuO(0.25) samples (6.63 and 7.21, respectively) display the highest values compared to the other samples, indicating that the adsorption capacity higher for these samples. When the correlation coefficients are analyzed, it is observed that they fall in a range between 0.993–0.879, suggesting a high dispersion and indicating that this model does not adequately describe the adsorption of propanethiol from the hexane solution onto the solids. Figure 7 shows the dotted lines of the Freundlich model.

Table 5 shows adjustment of the experimental data to fit the Langmuir model [8,9]. The adsorption constants and R² show a reduced range of 0.996–0.982, which indicate that this model was a better fit for the experimental data obtained in the isotherms.

The higher values for the maximum adsorbed quantity found for the samples of activated carbon impregnated with copper agree with the results obtained for the immersion enthalpies; the K_L constant values indicated the high adsorption of the propanethiol in these samples. With regards the samples modified by surface oxidation, CAN26 showed the highest adsorption value, being the sample that was modified with 6 M nitric acid solution, and thus showing the influence of the superficial chemistry on propanethiol adsorption. Figure 7 shows the fit to this model with solid lines; it was found that this model was a better fit and therefore adequately describes the adsorption of propanethiol on activated carbon.

Two kinetic models were studied, the pseudo first and pseudo second order [8,9,14]. Figure 8 shows the experimental results obtained for the solute adsorbed as a function of time on the surface of the activated carbon, which becomes saturated after about 800 minutes. The CACuO(0.25) sample showed the highest adsorption capacity for propanethiol and the original activated carbon the lowest (36%).

Table 6 shows the quantity of propanethiol adsorbed at equilibrium for each of the activated carbon samples synthesized, using the pseudo first order model, the constant kinetics and a factor of correlation R² of between 0.982–0.963. These values suggest dispersion of the experimental data.

The pseudo second order model has a better fit for the CAN than the CACuO. The correlation coefficients, between 0.997–0.983 are better than those of the Freundlich model. Table 7 shows the amount of propanethiol adsorbed in equilibrium, the velocity constant of the model and the correlation factors; these values are taken from the linearization of the model.

By comparing with the previous kinetic model it is possible to establish that the kinetics of propanethiol removal to a second-rate kinetics, since this model adjusts the experimental data with less dispersion.

N₂ adsorption isotherms at 77 K were measured in a volumetric system (Autosorb-3B, Quantachrome). For the adsorption data analysis the Dubinin–Radushkevich equation (DR) was used. The activated carbon samples, about 0.100 g, were degassed at 523 K for 3 h in an Autosorb 3B (Quantachrome Co.). The corresponding adsorption nitrogen isotherms were obtained in this apparatus at 77 K [5].

The basic and acid sites of the activated carbons were determined by means of the titration method proposed by Boehm using 0.1 M HCl, NaHCO_3, Na₂CO₃ and NaOH solutions [6]. In this method, 50 mL of the solution is placed in contact with 0.500 g of activated carbon in a glass flask for 48 h at a constant temperature of 298 K. A sample of 10 mL was titrated with the corresponding 0.1 M solution. Titration was carried out using a Trito Line Alpha Plus (SCHOTT Instruments, Woburn, MA, USA) [4].

Determinations of the immersion enthalpy of the activated carbons were carried out in water and propanethiol. A Calvet-type heat conduction microcalorimeter of local construction [23] was used in order to determine the immersion heat. The cell was filled with 8 mL of the liquid. A sample of activated carbon weighing approximately 0.100 g was placed in the calorimetric cell inside a glass ampoule. When the equipment reached an equilibrium temperature the glass ampoule was broken; the thermal effect generated was recorded and electrically calibrated.

The activated carbon samples were macerated to obtain a fine powder prior to measurements made using this technique. The powder was placed in a support to measure the X-ray spectrum, with data taken from 2 to 180° at intervals of 0.02°/min. Rigaku Mini Flex equipment was used to take the measurements.

The activated carbon samples were submitted to a hydrogen flow of 50 mL per minute. The temperature was increased gradually, in order to determine the temperature at which reduction took place. A MICROMERITICS 2720 chemisortometer with a mixture of He/H₂ (10:90) was used.