Textural, surface, thermal and sorption properties of the functionalized activated carbons and carbon nanotubes

Abstract Two series of functionalised carbonaceous adsorbents were prepared by means of oxidation and nitrogenation of commercially available activated carbon and multi-walled carbon nanotubes. The effect of nitrogen and oxygen incorporation on the textural, surface, thermal and sorption properties of the adsorbents prepared was tested. The materials were characterized by elemental analysis, low-temperature nitrogen sorption, thermogravimetric study and determination of the surface oxygen groups content. Sorptive properties of the materials obtained were characterized by the adsorption of methylene and alkali blue 6B as well as copper(II) ions. The final products were nitrogen- and oxygen-enriched mesoporous adsorbents of medium-developed surface area, showing highly diverse N and O-heteroatom contents and acidic-basic character of the surface. The results obtained in our study have proved that through a suitable choice of the modification procedure of commercial adsorbents it is possible to produce materials with high sorption capacity towards organic dyes as well as copper(II) ions.


INTRODUCTION
Adsorption processes are widely applied in many branches of industry, both in technologically developed and developing countries. Thus, it is understood that the search for new, more selective and effective adsorbents and optimisation of those already known have been of continuous interest all over the world. Much attention is paid to the search for new precursors of adsorbents, in particular new precursors of activated carbon adsorbents, which is evidenced by a large number of papers whose authors have been testing all kinds of industrial and agricultural waste products 1-8 . Also much interest has been paid to the effects of different modifi cations of adsorbents, both used on laboratory and industrial scale. The modifi cation is usually realised as functionalization of the adsorbent surface with various functional groups (organic and inorganic). Then, the effects of such modifi cation on the textural parameters, thermal stability and mechanical strength are optimised 9- 13 .
Although the effects of introduction of heteroatoms (e.g. O, N, S, P, Cl, F) or metal ions (e.g. Pt, Ag, Au, V, Fe) into the structure of adsorbents have been reported in many papers 14-20 , there is still much to be learnt in the fi eld and the problems are still of interest for many research groups, including our group. Earlier 21-23 we have studied the effect of nitrogen and oxygen functional groups on the acid-base properties, porous structure and sorption ability from gas phase of the activated carbons of our interest. Results of our studies have shown benefi cial effect of such modifi cations, which prompted the attempts at modifi cation of commercial products.
The main aim of this study was to obtain and characterise new carbonaceous adsorbents obtained by introduction of nitrogen and oxygen functional groups into the structure of commercially available activated carbon and multi-walled carbon nanotubes. The suitability of these new materials as adsorbents of organic dyes and copper(II) ions from liquid phase was also tested.

Material
The investigation presented in this work was carried out on commercial micro/mesoporous activated carbon -Norit ® SX2 (denoted as AC) and multi-walled carbon nanotubes (O.D.× L 6-9 nm × 5 μm, >95%C, Sigma-Aldrich ® , denoted as NT. The starting materials were subjected to three different treatments: (1) oxidation with 20% HNO 3 (OX), (2) reaction with urea (U) and (3) oxidation followed by reaction with urea (OXU), as presented in Scheme 1. These modifi cations were performed according to the procedures described below.

Scheme 1. Preparation of modifi ed activated carbons and carbon nanotubes
Oxidation with 20% solution of HNO 3 Portions of 10 g of activated carbon or carbon nanotubes were placed in two-necked fl ask, equipped with a refl ux condenser and dropping funnel. Next, 150 ml of 20% HNO 3 was added and the reaction mixture was heated up to boiling point. After that, the remaining portion of the acid (150 ml) was slowly added to the fl ask.
After addition of the last portion of HNO 3 , the reaction was continued for 3 h. The fi nal product was recovered by fi ltration, washed repeatedly with deionized water (until the pH of the effl uent water was constant) and dried at 110 o C (samples ACOX and NTOX, respectively).

Reaction with urea at high temperature
Portions of 10 g of carbonaceous materials were impregnated with urea at the weight ratio of 1:1, dried at 110 o C to constant mass and then subjected to thermal treatment at 350 o C, for 3 h, in air fl ow (100 ml/min). The nitrogen-enriched materials were washed with hot deionized water in order to remove the unreacted part of urea and dried at 110 o C. Incorporation of nitrogen was applied to starting activated carbon (ACU) and nanotubes (NTU) as well as to products of their oxidation (ACOXU and NTOXU samples).

Samples characterization
The elemental analysis of the all samples under investigation was performed on an Elementar Analysensysteme instrument, model Vario EL III. The textural characterisation of the samples was based on the nitrogen-adsorption isotherms determined at -196 o C with a Quantachrome Autosorb iQ surface area analyser. Before the isotherm measurements the samples were outgassed at 150 o C for 8 h. Surface area and pore size distribution were calculated by BET (Brunauer-Emmett-Teller) and BJH (Barrett-Joyner-Halenda) methods, respectively. Average pore diameter and total pore volume were determined as well. Additionally, micropores surface area and volume were calculated using t-plot method. The content of surface functional groups of acidic and basic character was determined by the Boehm method 24 , whereas the pH of the materials was evaluated using the procedure described in detail in our previous study 25 . Thermogravimetric analysis was performed on an SETSYS 12 made by Setaram, according to the procedure described earlier 26 .

Adsorption tests
Determination of the adsorption of methylene blue, alkali blue 6B and copper(II) ions was performed using the following procedure: samples of the prepared adsorbents of equal masses of 0.2 g were added to 100 ml of appropriate adsorbate solution with initial concentrations of 1000 mg/l and the suspension was stirred for 24 h to reach equilibrium. The concentrations of organic dyes as well as copper(II) ions in the solution after adsorption were determined using a double beam UV-Visible spectrophotometer (Cary Bio 100, Varian) at 660, 593 and 620 nm wavelength, respectively.

Elemental composition of adsorbents prepared
The chemical compositions of the starting and modifi ed samples are given in Table 1.
As follows from these data, each of the modifi cations performed resulted in signifi cant changes in the content of mineral substances (ash) and in the contribution of particular elements in the carbon structure. The character of the changes depends on the variant of modifi cation and on the type of material subjected to the thermo--chemical treatment. Oxidation with 20% HNO 3 leads not only to introduction of signifi cant amounts of oxygen into the adsorbent structure, but also to a considerable increase in the content of ash and to small changes in the contribution of the other heteroatoms, nitrogen and hydrogen. Introduction of the oxygen functional groups is accompanied by a signifi cant decrease in the content of carbon, particularly pronounced for the activated carbon ACOX. Exposure of the activated carbon ACOX and multi-walled nanotubes NTOX to urea under oxidising conditions results in the introduction of signifi cant amount of nitrogen in their structures and rather insignifi cant changes in the contents of the other elements. The activated carbon sample proved much more susceptible to modifi cation, as the content of nitrogen in ACOXU was 9.1 wt% which was almost twice greater than in NTOXU modifi ed in the same way which was 4.7 wt% N daf . The greater reactivity of activated carbon is also indicated by a different character of changes in the content of oxygen as a result of the reaction with urea. The content of O daf in NTOXU is by 1.7 wt% lower than in NTOX, while ACOXU shows a small increase in the content of oxygen with respect to that in ACOX.
The decrease in the content of O daf in nanotubes can be a consequence of the fact that part of oxygen groups was engaged in oxygen-nitrogen connections as it has been established 21, 27 , that the presence of oxygen functional groups on the surface of carbon favours incorporation of nitrogen groups. As the structure of nanotubes is less susceptible to modifi cation than that of activated carbon, the conditions used for the reaction with urea (temp. 350 o C, air atmosphere) could be too mild to lead to further oxidation of their surface. This supposition is to some degree confi rmed by changes in the content of oxygen in the samples ACU and NTU subjected exclusively to the reaction with urea. According to our results, the content of oxygen in NTU is at the same level as in the initial nanotubes material NT, while the content of oxygen in ACU is by about 1% greater than volume and pore diameter are most probably caused by blocking of some part of pores by the nitrogen and oxygen functional groups introduced upon modifi cation 22 . Different character of the changes caused by different modifi cations suggests that in the process of oxidation more important are the pores of greater diameters and inside them the majority of carbon-oxygen connections are generated, whereas in the reaction with urea more important are smaller pores. However, verifi cation of this supposition needs further studies.
The mesoporous character of the initial materials and the adsorbents obtained by their modifi cation has been confi rmed by the shapes of the adsorption/desorption isotherms presented in Figures 1-2. in AC, which points to a small oxidation of the activated carbon upon the reaction of urea.
The materials obtained as a result of the reaction with urea ACU and NTU have a lower content of nitrogen (nanotubes almost twice lower) than the oxidised samples treated with urea ACOXU and NTOXU, which confi rms the earlier reports of a benefi cial effect of the oxygen functional groups on the effi ciency of nitrogenation of carbon materials.

Textural structure of adsorbents prepared
The three variants of thermo-chemical treatment were applied to activated carbon and nanotubes to provoke not only chemical but also textural changes. As follows from the data collected in Table 2, the initial materials show much pronounced textural differences.
The multi-walled carbon nanotubes show not much developed surface area (271 m 2 /g) and remarkably mesoporous structure, while the activated carbon used has twice larger surface area S BET and a considerable contribution of micropores and mesopores of small diameter, as its mean pore diameter is 5.02 nm. Each of the modifi cations applied brings about signifi cant changes in the textural parameters of both materials, but the scale of changes is much greater in the samples obtained from activated carbon ACOX, ACOXU and ACU.
Modifi cations of nanotubes lead to an insignifi cant change in their surface area, by 34-49 m 2 /g, while the same modifi cations of activated carbon cause a drastic decrease in its surface area, from 136 m 2 /g for ACOX to 328 m 2 /g for ACU. The decrease in the surface area is accompanied by a signifi cant decrease in the micropores area, while the change in the area of external surface is much smaller. The modifi cations by oxidation and enrichment in nitrogen lead also to a considerable decrease in the total pore volume. It is particularly pronounced for nanotubes whose total pore volume becomes about twice smaller than that of the initial samples. A similarly drastic decrease also takes place in the mean pore diameter; the mean pore diameter in the modifi ed nanotubes is by 1.73-2.49 nm smaller than that of the unmodifi ed material. For ACOX and ACOXU a decrease in the mean pore size is also observed but it is much smaller, by 0.48-0.73 nm. For ACU (activated carbon subjected to the reaction with urea only) the mean pore diameter increased by 1.51 nm. It is most probably related to the fact that in ACOX and ACOXU the contribution of micropores slightly increased as a result of modifi cations, while ACU shows almost twice smaller V m /V t , ratio than the initial material AC. So pronounced changes in the total pore Table 2. Textural parameters of adsorbents prepared According to the IUPAC classifi cation, the carbon nanotubes show type IV isotherm (Fig. 1), characteristic of solid state materials on which the adsorption proceeds via multilayer adsorption followed by capillary condensation 28 . The activated carbon isotherm is close to type I, (Fig. 2), characteristic of microporous materials and mesoporous materials with pore size close to the micropores range. The isotherms show clear hysteresis loops. For carbon nanotubes the hysteresis loop is close to H1 type, characteristic of materials with cylindrical pore geometry and high degree of pore size uniformity, while for activated carbon the hysteresis loop looks like H3 type, characteristic of materials with slit-like pores.

Acid-base properties of adsorbents prepared
As follows from the data presented in Table 3, the carbon materials differ signifi cantly in the number of acid and basic functional groups and in the surface pH.
As far as nanotubes are concerned, the modifi cations lead to insignifi cant changes in pH, from weakly acidic for NTOX to weakly basic for NTU. The lowest pH for NTOX is related to the presence of considerable amounts of acidic oxygen functional groups forming as a result of exposure to HNO 3 . NTOXU, which was oxidised and then subjected to enrichment in nitrogen, shows an intermediate acid-base character and its pH is slightly above 7. A slightly basic character of NTU (pH 8.0) most probably follows from the domination of basic oxygen functional groups on the surface of this sample.
As to the activated carbon sample, its oxidation caused much greater changes in the acid-base surface character. ACOX surface showed acidic character of pH 4.6 which was a consequence of introduction of much greater amount of acidic functional groups (up to 1.86 mmol/g) than upon oxidation of nanotubes. The values of pH of the other samples obtained as a result of AC modifi cation are similar to that of the initial carbon, but the content of acidic and basic groups on the surface of ACOXU and ACU is much greater than on the unmodifi ed carbon.

Thermal properties of adsorbents prepared
Thermal properties of the samples studied were characterised by thermogravimetric measurements in the helium atmosphere. As follows from the character of TG and DTG curves, presented in Figures 3-4, the activated    Table 3. Acid-base properties of adsorbents prepared carbon and carbon nanotubes are thermally stable up to about 400 and 800°C, respectively.
Unfortunately, each modifi cation affects the thermal stability of the materials studied and the changes are much more pronounced for the samples obtained from activated carbon. This observation confi rms the earlier suppositions on greater reactivity and susceptibility to modifi cation of activated carbon. The least thermally stable was ACOX, obtained by oxidation with HNO 3 , showing the fi rst symptoms of thermal decomposition already above 200 o C. For NTOX also obtained by oxidation with HNO 3 , the loss of mass is over twice smaller, which confi rms greater thermal stability of nanotubes. Thermal stability of the samples oxidised and exposed to urea, in particular ACOXU, is much higher. A possible explanation is that the rather unstable oxygen groups have been removed or underwent transformation in the process of enrichment in nitrogen, so the carbon after such treatment was more thermally stable. The most thermally stable were ACU and NTU, subjected only to the reaction with urea. It may mean that some part of urea introduced upon the modifi cation was built into the external graphene layers, which improved the thermal stability of these samples 29 .
The DTG curves of the samples studied (Figs. 5-6), except the initial nanotubes, show a characteristic minimum at 80-90 o C, corresponding to the desorption of physically adsorbed water. reaction with nitric(V) acid, mainly such as carboxyl and lactone groups or acidic anhydrides. The second peak can be attributed to the decomposition of more stable oxygen groups such as phenol, carbonyl, quinone and ether ones. The DTG curve recorded for the analogous sample NTOX does not show the fi rst of the above mentioned minima, but an additional peak appears with a minimum close to 900 o C, probably corresponding to the thermal decomposition of the most stable carbonoxygen connections or pyrone structures 32-33 .
The character of DTG curves recorded for the nitrogenenriched samples, both with and without preceding oxidation. The DTG curve of ACOXU above 300 o C shows a very broad peak with not much distinct minimum at about 500 o C, smoothly passing into a peak with a clear minimum at about 700 o C. The fi rst minimum can be assigned to the decomposition of poorly thermally stable nitrogen groups such as amines, amides, imines and lactames, generated in the reaction of oxidised carbon with urea. This peak can also indicate the decomposition of nitrogen-oxygen functional groups formed upon nitrogenation. The second peak can be assigned to the decomposition of nitrogen groups in which nitrogen is built into the aromatic ring, so pyrrole, pyridone ones or pyridine N-oxide. The DTG curve of ACU also shows these two minima, but the former is much more pronounced.
The character of DTG curves recorded for nitrogenenriched nanotubes above 200 o C is different than that for the analogous activated carbon samples. For NTOXU, two clear minima appear at about 400 and 580 o C together with a poorer marked one at about 700 o C, corresponding to the decomposition of nitrogen and oxygen groups generated in the reaction with urea in the oxidising conditions. For NTU an additional peak is observed with a minimum close to 380 o C. The appearance of these minima at much lower temperatures means that the groups introduced upon nitrogenation of nanotubes not only show lower thermal stability but are also much weaker bound to the carbon structure so their decomposition begins earlier. The above observations confi rm that the same type modifi cation of activated carbon and carbon nanotubes gives materials of drastically different thermal and surface properties.

Sorption properties of adsorbents prepared
To characterise the sorption abilities of activated carbon and carbon nanotube samples towards removal of organic and inorganic pollutants from liquid phase, the samples studied were tested as adsorbents of methylene blue, alkali blue 6B and copper(II) ions at room temperature. Analysis of the results presented in Table 4 proves that the modifi cations applied brought about different changes in the sorption properties of the materials studied.
Unfortunately, the majority of modifi ed samples show worse sorption properties towards methylene blue and alkali blue 6B than the initial materials, which may be related to the changes in their textural parameters (Table 2). Only the sample obtained by oxidation of nanotubes with HNO 3 revealed improvement in the sorption properties towards the above two dyes. This result can indicate a benefi cial effect of the oxygen groups generated in this process that built into the porous structure  It is the most marked for the samples oxidised with HNO 3 and for those oxidised and subjected to the reaction with urea. For ACOX and ACOXU, the depth of this minimum is much greater, fi vefold and threefold, respectively, than for NTOX and NTOXU, obtained by modifi cation of nanotubes. The presence of this minimum means that incorporation of signifi cant amount of oxygen and to a less degree also nitrogen, causes a signifi cant increase in the hydrophilic properties of the samples' surfaces which usually is hydrophobic 30 .
As follows from the DTG curves obtained for ACU and NTU, the process of enrichment in nitrogen (without preliminary oxidation) brings about much smaller changes in hydrophilicity. Above 200 o C the DTG curves become more diverse. The curve for ACOX reveals two broad peaks with minima at about 250-270 o C and at 700 o C. In consistence with the earlier literature reports 31 , the fi rst most probably corresponds to decomposition of poorly thermally stable oxygen groups formed as a result of the of nanotubes leading to signifi cant decrease in the pore diameter and an increase in acidity of the surface. The latter improves sorption ability towards organic dyes, especially towards the cationic methylene blue.
Unfortunately, for ACOX (after the same modifi cation) the infl uence of oxygen groups was negative as the sorption properties of this sample towards both dyes were worse than those of the initial AC. The possible reasons for the different effect of the same modifi cation on the sorption properties of nanotubes and activated carbon are the different textural parameters of the materials and fi rst of all the dramatically different effectiveness of oxidation. The oxidation of activated carbon permits the introduction of over 4 times more oxygen groups than that of nanotubes (Table 2) and as the pores in AC are much smaller the oxygen groups could block the access to part of them by the dye molecules.
The situation is different for copper(II) ions adsorption. As follows from Table 4 data, introduction of signifi cant amounts of oxygen and nitrogen functional groups into the carbon structure has considerably improved the effectiveness of copper(II) ions removal from water solutions. The effect is particularly pronounced for AC samples and the best result was an over sevenfold increase in the sorption capacity of ACOXU sample with respect to that of unmodifi ed AC. This improvement is most probably related to the fact that copper(II) ions readily react with nitrogen and oxygen functional groups occurring together on the surface of ACOXU, which has been shown in earlier works 34-36 .
The sorption capacity of ACU and ACOX (subjected either only to the reaction with urea or only to oxidation with HNO 3 ) is by 25-30 mg/g smaller than that of ACO-XU but still high. These results imply that introduction of such heteroatoms as nitrogen and oxygen into the structure of adsorbents leads to an increase in the polarity of their surface and improvement of ion-exchange properties, which gives the benefi cial effect of increased sorption capacity of metal cations. Modifi ed nanotubes also revealed signifi cant improvement in the sorption capacity but not so spectacular as that of activated carbons. The probable reason is much less amount of nitrogen and oxygen incorporated into the nanotubes structure upon their modifi cation. Detail explanation of this problem needs further thorough studies, preceded by optimisation of the conditions of adsorption tests, the choice of the best pH, temperature and contact time 37-39 .
A comparison of our experimental data with the earlier literature reports on removal of Cu 2+ (Table 5) [40][41][42][43][44][45][46][47][48][49] shows that the modifi ed carbon sorbents obtained by us (in particular those based on activated carbon) are characterised by very good sorptive properties. It is expected that careful optimisation of the modifi cation procedures of carbon materials will lead to materials of sorption capacities and selectivity on a level close to that obtained for mesoporous silica as reported by Awual et al. 46,49 .

CONCLUSIONS
As follows from the above presented results, nitrogenation and oxidation of commercial adsorbents give a wide gamut of new carbon materials, differing in acid-base properties, textural parameters and at the same time showing good sorption properties towards organic dyes and copper(II) ions. The character of changes induced by the reaction with urea or by oxidation with HNO 3 signifi cantly depends on the type of material subjected to the thermo-chemical treatment. The material much more susceptible to the modifi cations applied was the commercially available activated carbon, as the amounts of nitrogen and oxygen functional groups incorporated in it were much greater than those incorporated to carbon nanotubes modifi ed in the same way. The most probable reason is the greater thermal and chemical stability of nanotubes confi rmed by the results of thermogravimetric measurements. The thermal analysis also proved that the functional groups introduced into the structure of the carbon materials studied, signifi cantly reduce their thermal stability, which is particularly pronounced for activated carbons.

ACKNOWLEDGEMENTS
Financial support received from the Polish Ministry of Science and Higher Education (Project Iuventus Plus No. IP2012 004072) is gratefully acknowledged.