Heteroatom-doped mesoporous carbons as efficient adsorbents for removal of dimethoate and omethoate from water

Abstract

Extensive use of organophosphate pesticides (OPs) invokes development of efficient procedures for their removal from the environment. By introducing low levels (<1 at%) of B, N or P into the structure of mesoporous carbons, we have produced a series of materials with different surface chemical composition, textural properties and level of structural disorder. These adsorbents were applied for removal of dimethoate and its oxo-analogue omethoate from aqueous solutions under batch adsorption conditions and by filtration using modified nylon membrane filters. Adsorption capacities up to 164 mg g⁻¹ were measured, with OPs uptake typically above 80% for dimethoate concentration as high as 5 × 10⁻³ mol dm⁻³. After the adsorption, neurotoxic effects of OP-containing water samples were significantly reduced or completely removed. The level of structural disorder was identified as a key parameter for efficient removal of dimethoate and omethoate while in the filtration experiments surface area of adsorbents also played an important role. While the presented research appeals to new fundamental studies of OP–carbon surface interactions, it also indicates a possible strategy in designing new efficient adsorbents for OPs removal from water.

---

1. Introduction

Organophosphorus pesticides (OPs) are extensively used in developed countries as they play an important role in the improvement of agricultural production by controlling various diseases.^1,2 Their common application in agriculture also leads to high-level OP residues in food and the environment,^3,4 which is highly undesirable due to the dangerous and harmful nature of these compounds.⁵ Their toxic effects are related to the irreversible inhibition of the enzyme acetylcholinesterase⁶ (AChE). Toxic potential is even more pronounced upon chemical transformations of thio-OPs during which sulfur atom from the thionate group is replaced by oxygen atom. In this way OPs are transformed to their oxo-forms (oxons), which have higher inhibitory power towards AChE compared to the parental thio-forms.^7–9 Dimethoate (Scheme 1A), as one of the major organophosphorus pesticides, is widely used in the fruit and field crops to promote the development of agricultural production and satisfy the need for farming products due to its low persistence and biodegradation.¹⁰ Its oxo-form, omethoate (Scheme 1B), could also be found in the environment, due to the presence of various oxidizing agents.⁵

Scheme 1 The structures of dimethoate (A) and its oxo-analogue omethoate (B).

Extreme use of dimethoate could lead to excessive residues accumulating in the human body through the food chain, which disrupt AChE function and cause cholinergic dysfunction or even death.^11,12

For the reasons stated above, rapid and accurate monitoring emerged together with the necessity of the efficient removal of OPs in order to control the levels of these compounds in food and the environment.^13–15 One of the main strategies for removal of pesticides from water is the adsorption on different types of materials. Numerous studies can be found in the literature reporting pesticide adsorption on mineral surfaces,^16–18 organohydrotalcite,¹⁹ activated carbon and zeolites,²⁰ carbon-based materials,^21,22 materials from graphene family²³ and others. Considering the interaction of pesticides with carbonaceous materials, the interaction with π conjugated system and functional groups present on the surface of carbon material play dominant role.^21,24 In specific, Vukčević et al.²⁴ recently considered batch adsorption of several pesticides on carbon monoliths surfaces and suggested that the π–π interactions are the main driving force for the adsorption of pesticides with an aromatic structure, while acidic groups play the crucial role in the adsorption of pesticides with no aromatic ring in their chemical structures. In general, the performance of carbonaceous material as an adsorbent is determined by the surface chemical composition and textural properties. This is in fact a general observation when it comes to the adsorption-based processes.²⁵ Mesoporous materials (MMs) stand as a particularly interesting class of adsorbents owing to the intrinsic high specific surface areas, tunable pore sizes, large pore volumes, as well as stable and interconnected frameworks with active pore surfaces for modification or functionalization.²⁵ In this sense, mesoporous carbons (MCs) stand as one of the most promising materials in adsorption-based technologies, but also in many others, such as energy related applications.^26,27

In this contribution, we consider mesoporous carbons doped with low concentrations of heteroatoms (B, N or P) as adsorbents for the removal of dimethoate and omethoate from water. Here, the focus is on the crucial link between physical and chemical properties of presented adsorbents and their performance, and not the analytical aspect and the details regarding the adsorption process itself. By introducing low concentration of heteroatoms into the MCs carbon network, surface functional groups, textural properties and the structural disorder of MCs are tuned and the effects of these properties of the investigated materials on their performances as adsorbents are discussed. We have analyzed the physiological effects of purified water samples, and, finally, demonstrated direct use of investigated materials in water filter for removal of investigated pesticides. Although adsorption kinetics can be rather important, in this contribution we have focused on the link between physicochemical properties of investigated samples and their performance as adsorbents under equilibrium conditions and in filtration experiments.

2. Experimental

2.1. Synthesis of heteroatom-doped MC

The synthesis of X–MC was described in contribution given by Pasti et al.²⁷ using evaporation-induced self-assembly (EISA) method²⁸ and for the sake of completeness we present details here as well. The synthesis was done as follows. Firstly, 4.52 g of Pluronic F127 (EO₁₀₆PO₇₀EO₁₀₆) (Sigma-Aldrich) was dissolved in the mixture of 18 cm³ of deionized water and 18 cm³ of ethanol (95 wt%) and strongly stirred for 15 min at room temperature. After that, 3 g of resorcinol was added and stirred for the next 30 min when the mixture was acidified with 0.35 cm³ of HCl (37 wt%). Subsequently, P, B and N sources were introduced. Various amounts of phosphoric acid (Centrohem),²⁹ boric acid (Merck)³⁰ and urea (Merck)³¹ were used in order to obtain different P, B and N doping levels in the final material. After 2 h, 4.5 cm³ of formaldehyde (37 wt%) was slowly added dropwise, stirred for another hour, aged, covered at room temperature for 3 days and dried at 85 °C for two days. Obtained polymeric cakes were carbonized under nitrogen atmosphere at 800 °C for 3 h at a ramping rate of 5 °C min⁻¹ to achieve complete carbonization, cooled down in the same atmosphere and used for further examination. The investigated materials are denoted hereafter as X–MC-EISA-800–Y, where X denotes heteroatom (B, N or P), EISA stands for the synthetic approach, 800 stands for the temperature of carbonization of the precursor and Y makes distinction between different MCs with the same heteroatom.

2.2. Adsorbents characterization

X-ray photoelectron spectroscopy (XPS) measurements were performed using a spectrometer PHI Quantum 2000 (Physical Electronics, USA) with a source analyzer angle of 45°. The base pressure in the spectrometer was 2 × 10⁻⁸ mbar. X-rays were obtained using a monochromatic AlK_α source (incident energy of 1486.6 eV). The position of the carbon peak C 1s was at 284.1 eV. Data analysis was performed using CasaXPS employing Kratos' relative sensitivity factors.

Total dopant content in the investigated materials was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, for B- and P-doped MCs) or elemental microanalysis (EA, for N-doped MCs). ICP-OES measurements were done using Thermo Scientific iCAP 6500 Duo ICP (Thermo Fisher Scientific, Cambridge, United Kingdom) equipped with concentric nebulizer. Radiofrequency power was 1150 W. Phosphorus was determined using P I signal at 178.766 nm, while boron was determined using the line at 208.9 nm. Samples were prepared for ICP-OES analysis employing microwave digestion method. Microwave digestion was performed using ETHOS 1 Advanced Microwave Digestion System (MILESTONE, Italy). Nitrogen content in N-doped MCs was determined using the Elemental Analyzer Vario EL III (Elementar).

Nitrogen adsorption/desorption was measured at −196 °C using the gravimetric McBain method.³² From the obtained isotherms, the specific surface area (S_BET), mesopore surface (including external surface area, S_meso) and micropore volume (V_mic) of the samples were calculated. S_meso and V_mic were estimated using the high-resolution α_s-plot method.³³ Micropore surface (S_mic) was calculated by subtracting S_meso from S_BET.

Doped MC samples were characterized at room temperature by X-ray powder diffraction (XRD) using Ultima IV Rigaku diffractometer, equipped with Cu Kα 1,2 radiation source, using a generator voltage of 40.0 kV and a generator current of 40.0 mA. The range of 10–90° 2θ was used for all powders in a continuous scan mode with a scanning step size of 0.02° and at a scan rate of 2° min⁻¹.

Raman spectra excited with a HeNe gas laser (excitation wavelength 633 nm) were collected on a DXR Raman microscope (Thermo Scientific, USA) equipped with an Olympus optical microscope and a CCD detector. The laser beam was focused on the sample using objective magnification 50×. The scattered light was analyzed by the spectrograph with a 600 lines mm⁻¹ grating. Laser power on the sample was kept at 1 mW.

The surface morphology of the samples was observed using a field emission scanning electron microscope (FESEM) TESCAN Mira3 XMU at 20 kV.

2.3. Adsorption studies

2.3.1. Adsorption of OPs in batch system. Investigated X–MC-EISA-800–Y samples were dispersed in water (deionized or tap water) and desired amount of pesticide stock solution (Pestanal, Sigma Aldrich, Denmark) was added to give the final concentration of the carbon material amounting to 10 mg cm⁻³. The concentration of pesticides was set to either 5 × 10⁻³ mol dm⁻³ (dimethoate) or 5 × 10⁻⁶ mol dm⁻³ (omethoate). The choice of pesticide concentrations is rationalized below. Then, the vessel containing the mixture of X–MC-EISA-800–Y + OP was placed on a laboratory shaker (Orbital Shaker-Incubator ES-20, Grant-Bio) and left overnight at room temperature to ensure reaching the equilibrium, similarly to previous studies.¹⁸ After that, the mixture was centrifuged for 10 min at 14 500 rpm. Supernatant was filtered though the nylon filter membrane and the concentration of OP was determined using UPLC analysis, while its physiological effect were determined as explained in Section 2.4. Control experiments were performed in identical way but without X–MC-EISA-800–Y adsorbents. The performance of each of adsorbent was quantified as the OP uptake, defined as:where C₀ and C stand for the OP concentration (dimethoate or omethoate) before and after the adsorption experiment, respectively. This actually gives the percentage of adsorbed OP from the solution. In the preliminary experiments with varying OPs concentration and fixed concentration of the adsorbent, the uptake, as defined here, was found to increase as the concentration of OP is decreasing. Another way of quantifying adsorption efficiency is to use the adsorption capacity (AC) defined as the mass of the adsorbate (OP) per unit mass of the adsorbent (X–MC-EISA-800–Y), (usually expressed in mg g⁻¹). If the mass of adsorbent is kept fixed, AC decreases rapidly as the initial concentration of adsorbate is decreasing as there are less molecules to attach on the adsorbent.³⁴ To get the insight into the dependence of the uptake and AC on the initial OP concentration the reader is referred to Fig. 9.

2.3.2. Water purification via X–MC-modified nylon membrane filter. In order to investigate the possibility of the use of X–MC-EISA-800–Y adsorbents in water filters, we modified commercial nylon membrane filter to include the layer of active material (Scheme 2). Modification was done as follows. The total amount of 10 mg of each sample was dispersed in 1.5 cm³ of deionized water and injected into the commercial filter (KX Syringe Filter, Kinesis, pore size 220 nm). Excess water was removed from the modified filter by compressed air. After that desired volume of tap water spiked with dimethoate or omethoate (the same initial concentrations as for the batch adsorption experiments) was injected through the modified filter. The flow rate was 10 cm³ min⁻¹. In these experiments OP-to-(X–MC-EISA-800–Y) mass ratio was kept the same as in the batch adsorption experiments in order to have direct comparison. Filtrate was subjected to the UPLC analysis to determine the concentration of dimethoate or omethoate. Also, the physiological effects of filtrates were determined as described below. Control experiments were performed using the non-modified filters. Also, by comparing the OP concentrations before and after the filtration through the non-modified filter we confirmed that removal of OP from the filtrate, as observed in the experiments with modified filters, is not due to the nylon membrane. The efficiency of a modified filter towards OP removal was also quantified as the OP uptake, eqn (1).

Scheme 2 Cross section of a modified nylon membrane filter.

After the experiments, modified filters were disassembled to check for the uniformity of X–MC-EISA-800–Y layer. In each case we observed uniform distribution of the adsorbent over the nylon membrane. Moreover, the integrities of nylon membranes were not compromised during the experiments.

2.3.3. UPLC analysis. Waters ACQUITY Ultra Performance Liquid Chromatography (UPLC) system coupled with a tunable UV detector controlled by the Empower software was used. Chromatographic separations were run on an ACQUITY UPLC™ BEH C₁₈ column with the dimensions 1.7 μm, 100 mm × 2.1 mm (Waters). The analyses of dimethoate and omethoate solutions were done under isocratic conditions with mobile phase consisting of 10% acetonitrile and 90% water (v/v). The eluent flow rate was 0.2 mL min⁻¹ and the injection volume was 10 μL. Under these experimental conditions the retention time of dimethoate is 2.6 min, while omethoate is found at the retention time of 1.7 min.

2.4. Physiological effects – measurement of AChE activity

As mentioned, physiological effects of dimethoate and omethoate, in terms of their (neuro)toxicity, differ to a great extent as omethoate is much more potent inhibitor of AChE. By inspecting the AChE inhibition curves of dimethoate and omethoate (Fig. 10) one can see that omethoate is approx. 1000 time more potent inhibitor compared to the corresponding thio-form. As we intended to compare the adsorption of dimethoate and omethoate on each X–MC-EISA-800–Y at the concentrations which induce similar physiological effects, we have chosen the concentration of 5 × 10⁻³ mol dm⁻³ in the case of dimethoate and 5 × 10⁻⁶ mol dm⁻³ in the case of omethoate.

AChE activity was assayed according to Ellman's procedure.³⁵ The in vitro experiments were performed by exposure of 2.5 IU commercially purified AChE from electric eel to OP solutions obtained in adsorption experiments (batch adsorption or filtration) at 37 °C in 50 mM PB pH 8.0 (final volume 0.650 mL). The enzymatic reaction was started by addition of acetylcholine-iodide (AChI) in combination with DTNB as a chromogenic reagent, and allowed to proceed for 8 min until stopped by 10% sodium dodecyl sulfate (SDS). The product of enzymatic reaction, thiocholine, reacts with DTNB and forms 5-thio-2-nitrobenzoate, whose optical adsorption was measured at 412 nm. It should be noted that in these measurements the enzyme concentration was constant and set to give an optimal spectrophotometric signal. Physiological effects were quantified as AChE inhibition given as:

where A₀ and A stand for the AChE activity in the absence of OP and the one measured after the exposure to a given OP.

3. Results

3.1. Materials characterization

3.1.1. Elemental and surface composition. XPS analysis showed that approx. 90 at% of each X–MC-EISA-800–Y surface is consisted of carbon while relatively high concentration of oxygen (approx. 10 at%) is also confirmed (Table 1). Both surface and bulk concentrations of dopants are rather low and are always below 1 at% (Table 1). Further information about surface functional groups is obtained by deconvolution of high resolution XPS spectra (Fig. 1 and 2). The fitting of the C 1s spectra showed the main carbon lattice peak observed around 284.6 eV in all samples. The broadening of the C 1s signals on the high energy side (286–291 eV) is attributed to the carbon atoms forming single and double bonds with different atoms.³⁶ Two shoulder peaks are fitted and attributed to carbon species in alcohol and/or ether groups³⁷ (binding energy 285 eV), and to carbon atoms in carbonyl, carboxyl and/or ester groups³⁷ (286–291 eV range). The O 1s XPS response was split into two components located at binding energies of 529.8 and 531.7 eV. These are recognized as C=O quinone type groups, C–OH phenol groups or ether C–O–C groups, respectively.³⁸ The deconvolution of B 1s indicated two components at 188.7 and 190.5 eV, ascribed to BC₃ and BC₂O species, respectively.^39,40 The chemical state of nitrogen was ascribed solely to pyrrolic/pyridone type moiety with the XPS peak binding energy of 399.8 eV.⁴¹ Phosphorus P 2p signal was below the detection limit of the instrument, suggesting that P is not present in the external surface layer above the photoelectron escape depth. Nevertheless, the incorporated phosphorus was confirmed by ICP-OES measurements (Table 1). The absence of P in the surface layers of P-doped MCs was discussed previously²⁷ and it is suspected to be the consequence of the removal of phosphorous during the carbonization of polymeric precursor.

Table 1 Elemental composition of X–MC-EISA-800–Y adsorbents determined by ICP-OES, elemental analysis and XPS. The results for X concentration obtained by XPS were converted to wt% to have direct comparison with the results of ICP-OES and elemental microanalysis

Sample	ICP/elemental analysis	XPS analysis
	X/wt%	C/at%	O/at%	X/at%	X/wt%
B–MC-EISA-800–1	0.33	88.68	11.2	0.13	0.11
B–MC-EISA-800–2 (ref. 27)	0.72	89.52	10.0	0.48	0.42
B–MC-EISA-800–3 (ref. 27)	0.94	87.39	11.77	0.84	0.73
N–MC-EISA-800–1 (ref. 27)	0.39	93.20	6.51	0.29	0.33
N–MC-EISA-800–2 (ref. 27)	0.69	90.9	8.64	0.46	0.51
N–MC-EISA-800–3 (ref. 27)	0.94	90.97	8.44	0.59	0.67
P–MC-EISA-800–1	0.14	87.96	12.04	—	—
P–MC-EISA-800–2 (ref. 27)	0.30	89.16	10.84	—	—
P–MC-EISA-800–3 (ref. 27)	0.37	90.48	9.52	—	—

---

Fig. 1 O 1s, C 1s and B 1s high resolution XPS spectra of investigated B-doped carbons.

Fig. 2 O 1s, N 1s and C 1s high resolution XPS spectra of investigated N-doped carbons.

By analyzing the effects of doping on the presence of different types of oxygen and carbon moieties in investigated samples it can be seen that type of dopant affects more than its concentration (Fig. 1 and 2). For example, doping with nitrogen results with lower relative fraction of C=O groups compared to the other types of O-functional groups than in the case of B-doping. Also, relative content of O-functional groups is lower in the case of N-doped MCs. Nevertheless, a solid conclusion regarding the connection between the doping level and the concentration of O-functional groups cannot be derived. In fact, it seems that a delicate interplay between chemical species present in MCs' precursors exists, so that the final chemical composition depends in a non-linear fashion on the concentration of dopant. This conclusion, in fact, holds for other physical properties of investigated carbons, as well, which are described below.

3.1.2. Textural properties of X–MC-EISA-800–Y materials. Investigated adsorbents have significant S_BET, which ranges between 500 and 600 m² g⁻¹, except for the case of N–MC-EISA-800–3, which has S_BET of only 103 m² g⁻¹ (Table 2). Moreover, it can be seen that the S_BET for N-doped carbons decays with the dopant concentration. We suspect that this behavior can due to the interaction of N-source (urea) with Pluronic F127 and resorcinol during the preparation of polymeric precursor. This interaction affects self-assembly of the precursor components and results with significantly reduced porosity of the resulting carbons when urea concentration is high (see also Fig. 3, middle). When considering specific surface, in the case of B-doped MCs mesopores (radius between 2 and 50 nm) dominate over micropores (radius below 2 nm). In all the other cases, S_micro is larger than S_meso. Also, when comparing V_micro for the samples with similar S_BET, one can see that B-doped samples have micropore volume smaller by a factor 3–4 compared other samples (Table 2).

Table 2 Textural properties of investigated P-, B- and N-doped carbons: BET surface areas (S_BET), mesopore surface (S_meso), micropore surface (S_micro), total pore volume (V_tot), mesopore volume (V_meso) and micropore volume (V_micro)

Sample	SBET/m^2 g^−1	Smeso/m^2 g^−1	Smicro/m^2 g^−1	Vtot/m^3 g^−1	Vmeso/m^3 g^−1	Vmicro/m^3 g^−1
B–MC-EISA-800–1	509	324	185	0.332	0.281	0.051
B–MC-EISA-800–2 (ref. 27)	532	312	220	0.371	0.284	0.087
B–MC-EISA-800–3 (ref. 27)	533	200	333	0.350	0.179	0.171
N–MC-EISA-800–1 (ref. 27)	607	246	361	0.362	0.185	0.177
N–MC-EISA-800–2 (ref. 27)	517	146	371	0.277	0.075	0.202
N–MC-EISA-800–3 (ref. 27)	103	42	61	0.090	0.028	0.062
P–MC-EISA-800–1	590	268	322	0.428	0.271	0.157
P–MC-EISA-800–2 (ref. 27)	585	228	357	0.390	0.181	0.209
P–MC-EISA-800–3 (ref. 27)	550	199	351	0.357	0.149	0.208

---

Fig. 3 Pore size distribution (PSD) of investigated B-, N- and P-doped carbon materials. Insets give N₂ adsorption isotherms.

According to the evaluated pore size distributions (PSD, Fig. 3) it is possible to conclude that in most cases relatively narrow PSD exists, but there are some outliers such as B–MC-EISA-800–1 sample.

3.1.3. XRD analysis. As a common feature of XRD patterns of all the investigated samples (Fig. 4), one finds wide and shallow graphite (002) (ICSD no. 617290) reflection at low angles characteristic for amorphous carbons with small regions of crystallinity. At higher angles, diffuse (100) and (101) reflections of graphite are also observed. Hence, XRD suggests low crystallinity and high structural disorder of investigated doped MCs.

Fig. 4 XRD patterns of investigated X–MC-EISA-800–Y carbons.

3.1.4. Raman spectroscopy. Collected Raman spectra of the investigated adsorbents display characteristic features of carbonaceous materials. The spectra are dominated by the two major bands designated as the G band (the graphitic band) and the D band (the disorder-induced band), located around 1592 cm⁻¹ and 1332 cm⁻¹, respectively.^42–44 Upon deconvolution of Raman spectra, two additional bands with significantly lower intensities were observed. A shoulder found around 1200 cm⁻¹ was associated previously with the double resonance process.⁴⁵ Additional band located around 1520 cm⁻¹ is associated with the presence of the amorphous carbon.⁴⁶

In order to obtain the insight into the structural disorder of the investigated doped MCs we have calculated corresponding I_D/I_G ratios, where I_D and I_G are the intensities of the D and the G band, found upon the deconvolution of each Raman spectra (Table 3). Although there is no clear correlation between the I_D/I_G ratio and the doping level, it can be concluded that significant degree of structural disorder is present in these materials. The same conclusion was derived previously, along with the observation that P-doped samples display the lowest degree of structural disorder.²⁷ This was rationalized in terms of the ordering of surface layers during the carbonization processes once phosphorus is being burnt off from the surface (as indicated by XPS).

Table 3 The I_D/I_G ratio of investigated X–MCs

Sample	ID/IG ratio
B–MC-EISA-800–1	3.26
B–MC-EISA-800–2 (ref. 27)	3.72
B–MC-EISA-800–3 (ref. 27)	3.46
N–MC-EISA-800–1 (ref. 27)	2.63
N–MC-EISA-800–2 (ref. 27)	3.61
N–MC-EISA-800–3 (ref. 27)	3.31
P–MC-EISA-800–1	1.42
P–MC-EISA-800–2 (ref. 27)	2.18
P–MC-EISA-800–3 (ref. 27)	2.78

---

3.1.5. Morphology of X–MC-EISA-800–Y. FESEM micrographs of the investigated carbons (Fig. 5) show that these materials are rather porous, but there are no significant differences between surface morphologies among different materials. Also, the inspection of the obtained results has showed that there is no formation of any specific structures during carbonization process.

Fig. 5 FESEM micrographs of investigated X–MCs.

3.2. Dimethoate and omethoate removal via adsorption on X–MCs

In batch adsorption experiments we observed that most of the investigated adsorbents display similar performance with the uptake typically above 60% (Fig. 6). This holds for both dimethoate and omethoate adsorption, with the exception of P–MC-EISA-800–1 sample which showed rather poor performance, particularly in the case of omethoate adsorption. Considering rather high initial concentration of dimethoate, adsorption capacities of X–MC-EISA-800–Y materials are of the order of 100 mg g⁻¹ and the highest one of 164 mg g⁻¹ was found in the case of the B–MC-EISA-800–1 sample. The comparison of different adsorbents, presented so far in the literature, in terms of the adsorption capacity is difficult as the value depends on the concentrations of the adsorbate and the adsorbent (Fig. 9), unless the saturation coverage by adsorbate is precisely determined, which is usually not the case. For example, recently reported adsorption capacities of graphene-oxide (GO) and reduced GO towards chlorpyrifos, endosulfan and malation adsorption are of the order of 1000 mg g⁻¹ for the OPs stock solutions with the concentration of 2 mg dm⁻³ and GO/reduced GO dose of approx. 3 mg dm⁻³. On the other hand, the capacity decays abruptly upon increasing the concentration of adsorbent (for 10 mg dm⁻³ it is below 100 mg g⁻¹).²³ With 1000 times smaller initial concentration of omethoate, adsorption capacities determined here are scaled by the same factor, as expected.³⁴ It seems that, in most of the cases, adsorption performance of doped MCs is not affected by the complexity of the matrix as rather similar results are obtained for the adsorption of OPs from deionized water solutions and spiked tap water samples. However, in the case of P-doped MCs there is rather large difference between OPs uptake, especially dimethoate, from spiked tap water and deionized water solutions. Moreover, while dimethoate uptake is large from spiked tap water, omethoate uptake is higher from deionized water solutions (Fig. 6). Considering the fact that both dimethoate and omethoate are neutral species in the solution surface charge of X–MCs particles can have only indirect effects on the OPs adsorption through the adsorption of other charged species present in the matrix. Moreover, we have measured pH of the solutions after the adsorption and for all the samples it was between 4 and 4.5, which is in agreement with the fact that the same type of oxygen functional groups are present on the surface of all X–MC samples, as found by XPS analysis. The explanation of this behavior of P–MCs samples is not available at the moment but could potentially lead to the guidelines for the design of selective adsorbents for OPs and is currently under the investigation. By looking at the overall performance, B-doped MCs stand out as the best choice for the removal of dimethoate and omethoate under equilibrium conditions.

Fig. 6 Dimethoate (left) and omethoate uptake by X–MC-EISA-800–Y adsorbents in batch adsorption experiments. Initial concentrations of dimethoate and omethoate were 5 × 10⁻³ mol dm⁻³ and 5 × 10⁻⁶ mol dm⁻³, respectively. For brevity X–MC-EISA-800–Y was shorted as “X–Y”.

In the experiments with the nylon membrane filters modified with doped MCs (Scheme 2) and OPs-spiked tap water samples we observed a bit different behavior compared to batch adsorption experiments. First, dimethoate uptake is, in general, significantly lower compared to the batch adsorption experiments (Fig. 7, left). On the other hand, omethoate uptake is significantly higher compared to batch adsorption experiments and it is being above 90% in all cases. As we kept OP-to-(X–MC-EISA-800–Y) mass ratio the same for the batch and filtration experiments, these results led us to conclude that different modes of action of doped MCs hold for OPs removal under batch (equilibrium) adsorption conditions and in the filtration experiments. The sample denoted as B–MC-EISA-800–2 stands out as the adsorbent of choice for the removal of dimethoate and omethoate, having the best overall performance in filtration experiments.

Fig. 7 Dimethoate (left) and omethoate uptake by X–MC-modified nylon membrane filters. For brevity X–MC-EISA-800–Y was shorted as “X–Y”.

Although dimethoate removal by doped MCs in filtration experiments is not as efficient as in the case of batch adsorption experiments, there is no doubt that these materials can be used for the purification of water. This is especially true having in mind rather high concentration of dimethoate used in these experiments. This was further demonstrated by measuring physiological effects of purified water samples.

3.3. Physiological effects

As explained in the Section 2.4 we have analyzed the physiological effects of purified water samples, considering the ones obtained in (i) batch experiments with OP solutions in deionized water, (ii) batch experiments with spiked tap water and (iii) filtration experiments with spiked tap water. As shown (Table 4), for the initial concentrations of OPs used in the adsorption experiments AChE inhibition is induced, which translates into neurotoxic effects.

Table 4 AChE inhibition by purified water samples

Doped MC	AChE inhibition (%)
	Dimethoate (5 × 10^−3 mol dm^−3)			Omethoate (5 × 10^−6 mol dm^−3)
	Deionized water	Tap water	Filter	Deionized water	Tap water	Filter
B–MC-EISA-800–1	0	0	0	0	0	0
B–MC-EISA-800–2 (ref. 27)	0	0	0	0	0	0
B–MC-EISA-800–3 (ref. 27)	0	0	0	0	0	0
N–MC-EISA-800–1 (ref. 27)	0	0	0	0	0	0
N–MC-EISA-800–2 (ref. 27)	0	0	0	0	0	0
N–MC-EISA-800–3 (ref. 27)	0	0	21 ± 2	0	0	0
P–MC-EISA-800–1	18 ± 2	0	9 ± 2	9 ± 2	15 ± 2	0
P–MC-EISA-800–2 (ref. 27)	12 ± 2	0	9 ± 2	0	0	0
P–MC-EISA-800–3 (ref. 27)	0	0	0	0	0	0
Control	37 ± 3	33 ± 3	36 ± 2	15 ± 3	12 ± 3	13 ± 3

---

However, after the adsorption of OPs on doped MCs, purified water samples in most cases show no AChE inhibition. This especially holds for the filtration experiments with omethoate-spiked tap water: neurotoxic effects are absent in all the cases, which is what is counted for when it comes to the practical application. On the other hand, presented results clearly outline two samples with the lowest OPs removal performance: the one denoted as N–MC-EISA-800–3 and the one named P–MC-EISA-800–1 (Table 1). Based on the results of physical and chemical characterization of the investigated materials, the first one has the lowest specific surface area among all the samples, while the second one has the lowest level of structural disorder. In the following section we discuss the results of adsorption experiments in terms of the physical and chemical properties of doped MCs.

4. Discussion

As we have seen from the above presented data, structural disorder and specific surface area have the most pronounced impact on the adsorption properties of investigated materials. Moreover, we suspect that different mechanisms of OPs removal are operative in the case of batch adsorption and filtration.

First of all, based on the concentration of heteroatoms in X–MCs determined by XPS their direct role in OPs removal (for example, serving as the adsorption sites) can be excluded. This conclusion is also supported by the fact that P–MCs do not contain heteroatom moieties in the surface layers. On the other hand, the presence of heteroatoms affects other physical and chemical properties of X–MC. In order to discuss the roles of these properties, in the forthcoming paragraphs we shall focus on the results obtained for the cases of spiked tap water samples.

Under batch adsorption conditions we expect that an equilibrium distribution is achieved between the fraction of OPs in the solution and the fraction being adsorbed on the surface of doped MCs. According to the results of Vukčević et al.,²⁴ the presence of the surface functional groups, which serve as adsorption sites for OPs, affects the adsorption to a greater extent compared to the specific surface area. We see that our samples posses large amounts of surface functional groups and high levels of structural disorder. Knowing that defects in carbon materials present the sites of the enhanced reactivity⁴⁴ we have correlated OPs uptake to the I_D/I_G ratio and obtained good overall correlation (Fig. 8, left). As we see an agreement with the work of Vukčević et al.²⁴ in terms of the secondary role of the specific surface area on the adsorption capacity, we suggest that under batch conditions OPs adsorb on the external surface of X–MCs. An additional point related to specific interaction of OPs with carbon surfaces must be discussed here. First of all, dimethoate and omethoate do not have π electron system so dispersion interactions with the π electronic system of carbon²¹ can be excluded. However, although the concentration of omethoate is 1000 times smaller than the concentration of dimethoate, we see similar or even smaller uptake of omethoate (Fig. 8, left), in spite the fact that the OP uptake increases as the initial concentration of OP is decreasing (Fig. 9). As mentioned, this translates to 1000 time smaller adsorption capacities. This is clear indication that the interaction of omethoate with the doped MCs surface is much weaker compared to dimethoate. As the only difference in the structures of thio- and oxo-form is the sulfur atom in thiophosphate group (Scheme 1), we conclude that precisely this group is responsible for the adsorption of dimethoate on X–MCs. This seems contrary to the results of Maliyekkal et al.²³ who suggested that the adsorption of OPs on GO and reduced GO is mediated through water. This conclusion was based on the density functional theory calculations which showed that direct interactions between pristine graphene and the OPs are unfavorable.²³ This indeed seems plausible if one works with ideal defect-free graphene as set in the theoretical model of abovementioned authors.²³ However, materials investigated here, just like the graphene-based adsorbents presented in the work of Maliyekkal et al.²³ posses large number of surface function groups (as seen by XPS) and high level of structural disorder (as confirmed by Raman spectroscopy) and these defective site present the centers enhanced reactivity⁴⁴ which have to be considered when analyzing adsorption on carbon materials. Hence, we believe that the atomic level picture of the OPs interactions with carbon surfaces should be revisited and that the presence of different surface hetero-functions and defects (such as single and double vacancies and Stone–Wales defects) should be taken into account.

Fig. 8 Correlation of measured dimethoate and omethoate uptakes in batch adsorption experiments (left) and filtration experiments (right) to the properties of the doped MCs. The results for dimethoate and omethoate removal from spiked tap water solutions are presented. The error bars indicating the uncertainties of data points are removed for clarity; these are indicated in Fig. 6 and 7, while X–MC-EISA-800–Y was shorted as “X–Y”.

Fig. 9 The dependence of the adsorption capacity (AC) and the corresponding dimethoate uptake for the case of the B–MC-EISA-800–2 sample given as the function of dimethoate initial concentration in the batch adsorption experiments.

Fig. 10 AChE inhibition curve of dimethoate and omethoate.

Now we turn to the OPs removal by doped MC-modified filters. Here we see completely opposite situation where omethoate uptake is significantly higher than dimethoate uptake. We suspect that in this case the equilibrium cannot be achieved and that OPs get physically entrapped inside the pores and the interparticle voids in the doped MC layer (Scheme 2). Of course, the interaction between the OP and the X–MC surface also contributes removal, so we correlated BET surface area multiplied by I_D/I_G ratio to the OPs uptake and observed rather good correlation (Fig. 8, right). If these assumptions are taken as a starting premise, higher omethoate uptake can be explained. Namely, due to much higher concentration of dimethoate, pores and interparticle voids get saturated (saturation corresponding to non-equilibrium flow conditions) which prevents additional dimethoate uptake during the filtration. However, even if such situation occurs, in most of the cases we see the absence of the neurotoxic effect upon the filtration (Table 4). Moreover, under realistic conditions, the concentration of dimethoate is unlikely to be as high as 5 × 10⁻³ mol dm⁻³, the one we use in described experiments.

5. Conclusion

In the present work we demonstrate the use of heteroatom-doped mesoporous carbons for efficient removal of dimethoate and omethoate from aqueous solutions. By introducing low level (<1 at%) of B, N or P into the carbon structure, materials with different surface chemical composition, textural properties and the levels of structural disorder were obtained. These adsorbents were able to remove dimethoate and omethoate from deionized water solutions as well as spiked tap water samples under batch adsorption conditions. We have also demonstrated the principle of the modified nylon membrane filter which can be used for water purification in the case of contamination with OPs. In spite the fact that after filtration certain amount of OPs remains in the spiked tap water samples, most of the investigated materials are capable to completely remove neurotoxic effect of water samples upon the filtration. The best adsorption performance was observed for the materials with high surface area and high level of structural disorder. By analyzing the results for dimethoate and omethoate adsorption we have concluded that specific interactions exist between OPs molecule and carbon surface and that it is mediated by thiophosphate group. Although presented results appeal to new fundamental studies of OP–carbon interactions, it is clear that presented mesoporous carbons possess high potential for practical application in water purification and OPs removal.

Acknowledgements

Authors would like to thank to the Ministry of Education, Science and Technological Development of Republic of Serbia for their financial support (Project No. 172023).

References

1. T. Perez-Ruiz, C. Martinez-Lozano, A. Sanz and E. Bravo, Chromatographia, 2005, 61, 493–498.
2. T. Lazarevic-Pasti, T. Momic, M. M. Radojevic and V. Vasic, Pestic. Biochem. Physiol., 2013, 107, 55–60.
3. J. E. Casida and G. B. Quistad, Chem. Res. Toxicol., 2004, 17, 983–998.
4. K. Taira, Y. Aoyama and M. Kawamata, Environ. Toxicol. Pharmacol., 2006, 22, 40–45.
5. T. D. Lazarevic-Pasti, A. M. Bondzic, I. A. Pasti, S. V. Mentus and V. M. Vasic, J. Electroanal. Chem., 2013, 692, 40–45.
6. J. S. Van Dyk and B. Pletschke, Chemosphere, 2011, 82, 291–307.
7. Q. Zhang and S. O. Pehkonen, J. Agric. Food Chem., 1999, 47, 1760–1766.
8. T. D. Lazarevic-Pasti, A. M. Bondzic, I. A. Pasti and V. M. Vasic, Pestic. Biochem. Physiol., 2012, 104, 236–242.
9. R. Schwartz, O. Lahav and A. Ostfeld, Water Res., 2014, 63, 271–284.
10. Y. Lv, Z. Lin, W. Feng, X. Zhou and T. Tan, Biochem. Eng. J., 2007, 36, 221–229.
11. K. Kanagaraj, A. Affrose, S. Sivakolunthu and K. Pitchumani, Biosens. Bioelectron., 2012, 35, 452–455.
12. J.-J. Du, R.-X. Gao, H. Yu, X.-J. Li and H. Mu, J. Pharm. Anal., 2015, 5, 200–206.
13. J. He, L. Song, S. Chen, Y. Li, H. Wei, D. Zhao, K. Gu and S. Zhang, Food Chem., 2015, 187, 331–337.
14. J. J. Rasmussen, P. Wiberg-Larsen, A. Baattrup-Pedersen, N. Cedergreen, U. S. McKnight, J. Kreuger, D. Jacobsen, E. A. Kristensen and N. Friberg, Water Res., 2015, 84, 25–32.
15. Q. Wang, X. Zhang, Z. Xu and H. Gao, Food Anal. Method., 2015, 8, 2044–2051.
16. L. Clausen, I. Fabricius and L. Madsen, J. Environ. Qual., 2001, 30, 846–857.
17. A. Leovac, E. Vasyukova, I. Ivancev-Tumbas, W. Uhl, M. Kragulj, J. Trickovic, D. Kerkez and B. Dalmacija, RSC Adv., 2015, 5, 8122–8133.
18. M. M. Mirkovic, T. D. L. Pasti, A. M. Dosen, M. Z. Cebela, A. A. Rosic, B. Z. Matovic and B. M. Babic, RSC Adv., 2016, 6, 12219–12225.
19. F. Bruna, I. Pavlovic, C. Barriga, J. Cornejo and M. A. Ulibarri, Appl. Clay Sci., 2006, 33, 116–124.
20. M. Valickova, J. Derco and K. Simovicova, Acta Chim. Slovaca, 2013, 6, 25–28.
21. E. Ayranci and N. Hoda, Chemosphere, 2005, 60, 1600–1607.
22. P. Wang, Y. Yin, Y. Guo and C. Wang, RSC Adv., 2016, 6, 10615–10624.
23. S. M. Maliyekkal, T. S. Sreeprasad, D. Krishnan, S. Kouser, A. K. Mishra, U. V. Waghmare and T. Pradeep, Small, 2013, 9, 273–283.
24. M. Vukčević, A. Kalijadis, B. Babic, Z. Lausevic and M. Lausevic, J. Serb. Chem. Soc., 2013, 78, 1617–1632.
25. Z. Wu and D. Zhao, Chem. Commun., 2011, 47, 3332–3338.
26. M. Momcilovic, M. Stojmenovic, N. Gavrilov, I. Pasti, S. Mentus and B. Babic, Electrochim. Acta, 2014, 125, 606–614.
27. I. Pasti, N. Gavrilov, A. Dobrota, M. Momcilovic, M. Stojmenovic, A. Topalov, D. Stankovic, B. Babic, G. Ciric-Marjanovic and S. Mentus, Electrocatalysis, 2015, 1–14.
28. C. J. Brinker, Y. Lu, A. Sellinger and H. Fan, Adv. Mater., 1999, 11, 579–585.
29. C. H. Choi, M. W. Chung, S. H. Park and S. I. Woo, Phys. Chem. Chem. Phys., 2013, 15, 1802–1805.
30. S. Ding, S. Zheng, M. Xie, L. Peng, X. Guo and W. Ding, Microporous Mesoporous Mater., 2011, 142, 609–613.
31. S. Ratso, I. Kruusenberg, M. Vikkisk, U. Joost, E. Shulga, I. Kink, T. Kallio and K. Tammeveski, Carbon, 2014, 73, 361–370.
32. M. Peters, Angew. Chem., 1932, 45, 365.
33. K. Kaneko, C. Ishii, M. Ruike and H. Kuwabara, Carbon, 1992, 30, 1075–1088.
34. J. M. Salman, J. Chem., 2013, 2013, 5.
35. G. L. Ellman, K. D. Courtney, V. Andres Jr and R. M. Featherstone, Biochem. Pharmacol., 1961, 7, 88–95.
36. D.-W. Wang, F. Li, Z.-G. Chen, G. Q. Lu and H.-M. Cheng, Chem. Mater., 2008, 20, 7195–7200.
37. A. M. Puziy, O. I. Poddubnaya, R. P. Socha, J. Gurgul and M. Wisniewski, Carbon, 2008, 46, 2113–2123.
38. N. Gavrilov, I. Pasti, M. Vujkovic, J. Travas-Sejdic, G. Ciric-Marjanovic and S. Mentus, Carbon, 2012, 50, 3915–3927.
39. T. Kwon, H. Nishihara, H. Itoi, Q.-H. Yang and T. Kyotani, Langmuir, 2009, 25, 11961–11968.
40. Z.-H. Sheng, H.-L. Gao, W.-J. Bao, F.-B. Wang and X.-H. Xia, J. Mater. Chem., 2012, 22, 390–395.
41. S. Biniak, G. Szymanski, J. Siedlewski and A. Switkowski, Carbon, 1997, 35, 1799–1810.
42. Q.-H. Yang, P.-X. Hou, M. Unno, S. Yamauchi, R. Saito and T. Kyotani, Nano Lett., 2005, 5, 2465–2469.
43. G. Ciric-Marjanovic, I. Pasti, N. Gavrilov, A. Janosevic and S. Mentus, in Chemical Papers, 2013, vol. 67, pp. 781–813.
44. G. Ciric-Marjanovic, I. Pasti and S. Mentus, Prog. Mater. Sci., 2015, 69, 61–182.
45. E. F. Antunes, A. O. Lobo, E. J. Corat, V. J. Trava-Airoldi, A. A. Martin and C. Verissimo, Carbon, 2006, 44, 2202–2211.
46. T. Ungar, J. Gubicza, G. Ribarik, C. Pantea and T. W. Zerda, Carbon, 2002, 40, 929–937.

---