Solid state 27Al NMR and X-ray diffraction study of alumina–carbon composites

CARBON

9 3 ( 2 0 1 5 ) 7 5 1 –7 6 1

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/carbon

Solid state 27Al NMR and X-ray diffraction study of alumina–carbon composites Thierry R. Lopes, Gustavo R. Gonc¸alves, Ewerton de Barcellos Jr., Miguel A. Schettino Jr., Alfredo G. Cunha, Francisco G. Emmerich, Jair C.C. Freitas * Laboratory of Carbon and Ceramic Materials, Department of Physics, Federal University of Espı´rito Santo, 29075-910 Vito´ria, ES, Brazil

A R T I C L E I N F O

A B S T R A C T

Article history:

The structural and chemical transformations occurring in alumina–carbon composites

Received 14 January 2015

upon heat treatment were investigated by using a combination of X-ray diffraction (XRD)

Accepted 1 June 2015

and solid-state

Available online 4 June 2015

precursors were employed: a commercial activated carbon and a char obtained by car-

27

Al nuclear magnetic resonance (NMR) spectroscopy. Two different carbon

bonization of the endocarp of babassu coconut at 700 °C. The alumina–carbon composites were prepared by aqueous impregnation of the carbon supports with aluminum nitrate and, after filtering and drying, the as-synthesized powders were heat-treated under argon flow at temperatures up to 1000 °C. The Al compounds present in the as-synthesized samples were identified by XRD and solid-state

27

Al NMR as nanocrystalline aluminum oxyhy-

droxides or hydroxides, depending on the synthesis conditions. All Al-containing phases were XRD-amorphous in the char-derived nanocomposites, with the presence of a distribution of AlO6 (octahedral Al site), AlO5 (pentacoordinated Al) and AlO4 (tetrahedral Al site) units revealed by solid-state

27

Al NMR spectroscopy. The heat treatments caused the for-

mation of transition aluminas dispersed over the carbon supports, with the occurrence of different amounts of AlO6, AlO5 and AlO4 units depending on the heat treatment temperature and on the type of carbon precursor used for the preparation of the composites. Ó 2015 Elsevier Ltd. All rights reserved.

1.

Introduction

Composite materials containing metallic or oxidic nanoparticles supported on porous carbon materials have many technological applications in fields such as catalysis [1,2], magnetic separation [3], water cleaning [4] and others [1–4]. Depending on the characteristics of the carbon materials

used as supports, carbon-supported nanocomposites may exhibit high surface area and good structural and thermal stability, precluding the agglomeration of the supported nanoparticles. These nanocomposites can be prepared by many different routes. Wet chemical methods are among the most used ones, allowing the synthesis of products with varied chemical and physical properties [5,6].

* Corresponding author. Fax: +55 27 4009 2823. E-mail address: [email protected] (J.C.C. Freitas). Abbreviations: AC, commercial charcoal activated carbon; BC, char obtained by carbonization of the endocarp of babassu coconut; CT, central transition; DSC, differential scanning calorimetry; EDX, energy dispersive X-ray spectroscopy; EFG, electric field gradient; FID, free induction decay; MAS, magic-angle spinning; MQ-MAS, multiple-quantum magic-angle spinning; NMR, nuclear magnetic resonance; SEM, scanning electron microscopy; SPE, single pulse excitation; TG, thermogravimetry; XRD, X-ray diffraction. http://dx.doi.org/10.1016/j.carbon.2015.06.001 0008-6223/Ó 2015 Elsevier Ltd. All rights reserved.

752

CARBON

9 3 ( 2 0 1 5 ) 7 5 1 –7 6 1

Porous carbon materials containing dispersed aluminum oxides are of interest for several applications, including the removal of fluoride from water [7,8] and the adsorption of ammonia from gaseous mixtures [9]. Moreover, alumina–carbon composites can also be used to support metallic catalysts, combining the positive aspects of both alumina and carbon as porous supports. For instance, an aluminaactivated carbon composite was used to prepare supported NiMo catalysts, which showed better stability and enhanced activity for the reaction of hydrodesulfurization of dibenzothiophene, when compared to similar catalysts prepared on either activated carbon or alumina supports [10]. This effect was attributed to the mesoporous nature of the composite support and to the high dispersion achieved for the Ni and Mo species. A similar result of higher activity of NiMo/alumina–carbon black composite in the hydroconversion of a vacuum residue (hydrodesulfurization and hydrodenitrogenation reactions) when compared to an aluminabased commercial catalyst was attributed to the occurrence of macropores in the composite material (with 11–20% of pore volume) [11]. The properties of alumina–carbon composites depend on the physical, chemical and structural characteristics of the Al-containing phases and of the carbon support, as well on the possible interactions between them. To elucidate these features, a number of characterization methods are routinely applied, including X-ray diffraction (XRD), scanning electron microscopy (SEM), textural and thermal analyses [8,10]. Solid-state nuclear magnetic resonance (NMR) spectroscopy is a very useful tool for this kind of characterization, since it allows the assessment of information about the chemical environment of the chosen probe nuclei in a nondestructive way. Moreover, contrary to XRD and other diffraction-based techniques, solid-state NMR spectra can be successfully recorded for crystalline, nanocrystalline or amorphous phases, making possible the characterization of atomic environments even in structurally disordered and heterogeneous materials [12,13]. In this way, solid-state 27Al NMR has been used for many years in studies of aluminas and related materials, including nanostructured templates, mesoporous supports, layered compounds, amorphous alumina nanoparticles, etc. [14–19]. One of the main findings derived from solid-state 27Al NMR studies was the establishment of the thermal pathways leading from aluminum hydroxides or oxyhydroxides up to the most stable alumina phase (a-Al2O3, or corundum), passing through different transition aluminas (many of them presenting low degree of crystallinity) [20–25]. These investigations have been possible because 27Al is a very favorable nucleus from the point of view of NMR spectroscopy, with a natural abundance of 100% and often quite fast spin–lattice relaxation processes that enable short delay times to be used and, thus, allow the achievement of spectra with good signal-to-noise ratio in relatively short measurement times [12]. However, with a nuclear spin of 5/2, 27Al is a quadrupolar nucleus, with a sizeable electric quadrupole moment that interacts with any local electric field gradient (EFG) that exists at the nuclear site. This so-called quadrupole interaction contributes to large, anisotropic broadening of solid-state 27Al NMR spectra in aluminum oxides and other

compounds. On the other hand, it also expands the amount of information that can be gathered from the NMR spectra, since the quadrupole interaction is highly sensitive to the local structure around the nuclear sites [26]. In most typical cases, the 27Al NMR spectra recorded with magic angle spinning (MAS) exhibit strong sets of spinning sidebands, associated with the satellite transitions, which span a very large frequency range. On the other hand, the resonance lines due to the central transition (CT) – i.e., between the energy levels with ±1/2 spin quantum numbers – are much narrower and thus these are the resonances most commonly observed and used for the analysis of Al compounds [12,26]. The shape and the frequency shift of these resonances are determined by the following spectral parameters: the isotropic chemical shift (diso), the quadrupole coupling constant (Cq) and the asymmetry parameter of the EFG tensor (gq). For crystalline materials, it is possible to determine these parameters for each Al site by spectral fitting of the observed CT lineshapes [12]. However, when there is some disorder in the Al environment (of either structural or chemical origin), the patterns are broadened and the features and singularities typical of the CT pattern are not directly observed. Particularly, the occurrence of variations in bond lengths and bond angles cause a distribution in the EFG parameters (Cq, gq), which broadens the CT pattern in a quite typical fashion, leading to a pronounced asymmetry and a long tail developing towards the lower frequency side [12,15,17,22]. These broad patterns can also be fitted assuming a continuous distribution of the spectral parameters, thus revealing details about the type of disorder present in the material [15,18,22,27]. When the quadrupole interaction is present, the isotropic chemical shift of a peak in 27Al MAS NMR spectra does not correspond to the peak position. Even so, it is possible in many cases to determine the diso values associated with different sites, either by spectral fitting or by other methods such as the use of multiple magnetic fields [28] or multiplequantum (MQ) MAS experiments [12]. In the case of materials containing Al–O linkages, the Al coordination number results in a number of characteristic chemical shift ranges [12,26]. In most alumina forms, Al sites are distributed among 4coordinated (tetrahedral symmetry) and 6-coordinated (octahedral symmetry) sites. Typically, AlO6 units are characterized by shifts in the range 10–15 ppm, whereas the shifts due to AlO4 units fall in the range 50–80 ppm (with all shifts measured with respect to the 27Al NMR peak of 6-coordinated Al3+ ions in aqueous solution) [12,20–24]. In transition aluminas presenting reduced crystallinity, AlO5 units (corresponding to 5-coordinated Al in trigonal bipyramidal geometry) are frequently present, giving rise to broad resonances with chemical shifts falling somewhat between the above mentioned ranges, at about 30–40 ppm [12,19,20]. The relative contributions associated with these species can be computed from the integrated areas corresponding with each resonance, provided a reasonable fitting model is established and the correction factors associated with the distribution of intensity between the centerband and the spinning sidebands are taken into account [12,15,19,27,29,30]. In this work, the structural and chemical transformations occurring upon heat treatments in alumina–carbon

CARBON

9 3 (2 0 1 5) 7 5 1–76 1

composites prepared using porous carbons are investigated by a combination of thermal analysis, XRD, SEM and solidstate 27Al NMR spectroscopy. Although there have been many previous investigations on the thermal conversion of aluminum hydroxides or oxyhydroxides [20–25], no thorough study of this type has been reported dealing with such processes for aluminum compounds dispersed in carbon supports, as it is the case of alumina–carbon composites here described. Also, the previous reports on the preparation and applications of carbon–alumina composites [7–11] did not delve into the details of the structural and chemical aspects of these materials. Taking advantage of the suitability of solid-state NMR methods for studying atomic environments and considering the amorphous/nanocrystalline nature of the Al-containing phases formed in most cases, the use of 27 Al NMR is decisive here for the identification and quantification of the AlOn units, as well as for the assessment of how these contributions change as a function of the heattreatment temperature. Thus, the present results offer a detailed perspective about the chemical nature of the Alcontaining phases present in alumina–carbon nanocomposites and allow a better perception of the role played by the carbon support during the synthesis and the thermal treatments of these nanocomposites.

2.

Experimental methods

2.1.

Samples preparation

The composites were prepared starting from two different carbon materials: a commercial charcoal activated carbon (denoted as AC), obtained from Merck Chemicals (product number 102186), and a char prepared by carbonization of the endocarp of babassu coconut at 700 °C under N2 flow (denoted as BC), with specific surface areas of 1300 and 340 m2/g, respectively. These carbon supports were chosen in order to compare the consequences of changing the porosity of the carbon matrix on the type of synthesized Alcontaining phases. The endocarp of babassu coconut is an abundant solid residue of the production of babassu oil in Brazil; as it presents relatively high lignin content, this precursor is useful for the production of a charcoal with good mechanical properties, high mass yield and moderate porosity [31,32]. The alumina–carbon composites were synthesized following a precipitation method [5,33], using Al(NO3)3Æ9H2O (from Ecibra) as the aluminum source and NH4OH (in 7.1 M aqueous solution, from Dinaˆmica) as the precipitating agent. During all syntheses, the pH of the reaction medium was carefully monitored and remained fixed at the value pH = 11. The impregnation of each carbon precursor was conducted in aqueous suspension in a recipient with 350 mL of distilled water under constant stirring at room temperature for 3 h, followed by vacuum filtration and drying at 100 °C in a stove. Four different series of samples were synthesized, with different types of carbon precursor (AC or BC or none) and/or different amounts of reactants: (i) AC_Al_1: a mixture of 2.0 g of AC and 2.7 g of Al(NO3)3Æ9H2O was suspended in water and 100 mL of NH4OH were added to the suspension. (ii) BC_Al: the same amounts of reactants were used, but with BC as the carbon support. (iii) AC_Al_2: in this case, the

753

mixture involved 10.0 g of AC, 17.5 g of Al(NO3)3Æ9H2O and 350 mL of NH4OH. (iv) Al_OxHydrox: in order to study the effect of the carbon support, this sample was prepared with 35.0 g of Al(NO3)3Æ9H2O and 350 mL of NH4OH, without the addition of any carbon material to the reaction medium. These products were named ‘‘as-synthesized samples’’; portions of these samples were selected for heat treatments conducted under argon flow up to 1000 °C, so as to investigate the thermal transformations occurring in the prepared alumina– carbon nanocomposites.

2.2.

Characterization

The as-synthesized and heat-treated samples were characterized by XRD at room temperature, using a Shimadzu XRD6000 powder diffractometer with Cu-Ka radiation (k = 1.5418 ). The diffraction angle (2h) was varied from 10° to 90° at steps of 0.05°. The Scherrer equation [34] was used to estimate the average crystallite sizes of the Al-containing crystalline or nanocrystalline phases. Thermogravimetry (TG) curves were recorded using a Shimadzu TGA-50H instrument, with constant heating-rate of 30 °C/min up to 1000 °C under air flow (50 mL/min). Differential scanning calorimetry (DSC) curves were recorded using a Shimadzu DSC-50Q instrument, with constant heating-rate of 30 °C/min up to 600 °C under argon flow (50 mL/min). SEM images were obtained using a Shimadzu SSX-550 instrument equipped with an accessory for energy dispersive X-ray spectroscopy (EDX) analysis. Solid-state 27Al NMR experiments were performed at room temperature using a Varian/Agilent V NMR 400 MHz spectrometer, operating with a magnetic field of 9.4 T (27Al NMR frequency of 104.16 MHz). Single pulse excitation (SPE) experiments were performed using a p/6 pulse with duration of 1.6 ls and a recycle delay of 1.0 s (which was verified to be long enough to avoid any saturation effect). The spectra were obtained by Fourier transform of the free induction decays (FIDs), after accumulation of 200 transients. The frequency shifts, expressed in parts per million (ppm), were measured relative to the single resonance peak observed in the 27Al NMR spectrum of an aqueous solution of aluminum nitrate (1.02 M). SPE experiments were conducted with magic angle spinning (MAS) at a frequency of 14 kHz. The 27Al NMR spectra were fitted to a set of second-order quadrupolar lineshapes broadened by a distribution of quadrupole couplings and chemical shifts associated with disorder in the 27Al environments. In this way, it was then possible to determine the mean values of the quadrupole coupling constant (arbitrarily assuming zero asymmetry parameter) and of the isotropic chemical shift associated with each site [22,27]. The relative fractions of each component present in the spectra were also estimated from this fitting process, assuming that the intensity of each simulated peak matches the whole intensity of the corresponding CT; the correction factors that should be considered to account for the small portion of the CT intensity not present in the centerband in the spectra recorded with MAS [26,30] were found to be negligible in most cases and were then disregarded in this estimate of the relative fractions of each spectral component. All spectral simulations were performed using the DMFIT software [35].

754

CARBON

3.

Results and discussion

3.1.

As-synthesized samples

9 3 ( 2 0 1 5 ) 7 5 1 –7 6 1

A typical SEM image recorded for the as-synthesized AC_Al_1 sample is shown in Fig. 1, together with EDX maps showing the distribution of the elements oxygen, aluminum and carbon. From these maps, a homogeneous dispersion of aluminum oxides throughout the carbon matrix can be inferred. The Al contents present in the as-synthesized alumina–carbon composites were estimated by measuring the residual weights in the TG curves after the complete oxidation of the carbon supports, as shown in Fig. 2 for some representative samples. After the complete oxidation of carbon, the remaining Al-containing product is a-Al2O3, as identified by XRD. Therefore, the increase in the residual weight observed in the TG curves recorded for the Al-containing materials as compared to the Al-free carbon supports allows the calculation of the Al content in each case. From these data, the Al contents in the as-synthesized samples are found to be 8 wt.% for samples AC_Al_1 and AC_Al_2 and 10 wt.% for sample BC_Al (see Table 1). Fig. 3 shows de XRD patterns recorded for the assynthesized samples. The X-ray diffractograms of AC_Al_1 and AC_Al_2 samples show the presence of the broad maxima associated with the turbostratic structure of the char [31,36],

Fig. 2 – TG curves recorded under oxidizing atmosphere for the as-synthesized samples and carbon supports. (A color version of this figure can be viewed online.)

together with some broad peaks at the angular positions expected for boehmite (for AC_Al_1 sample) and bayerite (dominant in the case of AC_Al_2 sample) [24]. These results agree with previous reports on the products of the reaction

Fig. 1 – SEM image (top, left) and the corresponding EDX maps recorded for the as-synthesized AC_Al_1 sample. The color scales in the EDX maps correspond to the elements carbon (green), aluminum (blue) and oxygen (red). (A color version of this figure can be viewed online.)

CARBON

755

9 3 (2 0 1 5) 7 5 1–76 1

Table 1 – Summary of TG, XRD and NMR results obtained for the Al-containing phases in the as-synthesized samples. Sample

Al content Al major (wt.%) phase

AC_Al_1 8 BC_Al 10 AC_Al_2 8 Al_OxHydrox – a b

Da AlO5 AlO6 AlO4 (nm) d Cq Ab (%) Ab (%) diso (ppm) Cq (MHz) Ab (%) diso (ppm) Cq iso (ppm) (MHz) (MHz)

Boehmite 4 – – Bayerite 19 Bayerite 52

– 74.8 – –

4.3 –

– 18 – –

– 38.8 – –

2.6 –

– 9 – –

9.5 10.8 9.6 12.0

2.8 4.8 4.9 2.6

100 73 100 100

D = average crystallite size. A = relative area.

Fig. 3 – XRD patterns recorded at room temperature for the as-synthesized samples. (A color version of this figure can be viewed online.) between Al(NO3)3 and NH4OH in aqueous solution [33]; the differences between the AC_Al_1 and AC_Al_2 samples are clearly related to the different amounts of Al salt and precipitating agent used in their syntheses. In contrast, the XRD pattern of BC_Al sample shows only the typical maxima due to turbostratic carbon, evidencing the complete absence of crystalline Al-containing phases in this material. The XRD pattern of the carbon-free Al_OxHydrox sample exhibits a set of narrow peaks, associated with bayerite and also with ammonium nitrate; this latter compound is formed as a byproduct of the reaction between aluminum nitrate and ammonium hydroxide. From these results, it is possible to assess the effect of the presence of the carbon supports in the reaction conditions here employed, with the formation of nanocrystalline or amorphous Al-containing phases in AC_Al_1, BC_Al and AC_Al_2 nanocomposites as opposed to the crystalline phases observed in the carbon-free Al_OxHydrox sample. This finding is also confirmed by the

analysis of the average crystallite sizes of the dominant Alcontaining phases in the as-synthesized nanocomposites, as given in Table 1, which range from 4 nm (for boehmite in AC_Al_1 sample) to 19 nm (for bayerite in AC_Al_2 sample). For BC_Al sample it is not possible to estimate the average crystallite size, since no diffraction peak associated with any Al-containing phase is detected in the XRD pattern shown in Fig. 3, thus pointing to an XRD-amorphous phase in this material. The thermal behavior of the synthesized nanocomposites was investigated by TG (Fig. 2) and DSC (Fig. 4). For all analyzed samples (including the carbon supports, AC and BC), the first weight loss in TG curves shown in Fig. 2, around 100 °C, is due to the release of moisture; this gives rise to the strong endothermic peak observed in the corresponding DSC curves in Fig. 4. The second weight-loss, which begins at an onset temperature in the range 450–550 °C depending on the sample, is associated with the oxidation of carbon

756

CARBON

9 3 ( 2 0 1 5 ) 7 5 1 –7 6 1

Fig. 4 – DSC curves recorded under argon atmosphere for the as-synthesized samples. (A color version of this figure can be viewed online.) matrix. A continuous weight decrease is also observed in the TG curves recorded for the Al-containing samples before the main oxidation reaction, which is associated with the decomposition of the Al hydroxides and oxyhydroxides present in these materials. Correspondingly, endothermic events are also detected in the same temperature range in the DSC curves of these samples in Fig. 4. It is particularly noticeable the endothermic peak at 266 °C in the DSC curve of the AC_Al_2, coinciding with the weight-loss observed close to the same temperature in the TG curve of this sample and

indicating the decomposition of bayerite [24,37]. In the case of the carbon-free Al_OxHydrox sample, a similar peak associated with the bayerite decomposition is detected at 253 °C and a set of narrow endothermic peaks are further observed, which are due to the decomposition of ammonium nitrate [38]. Given the structurally disordered nature of many of the Alcontaining phases present in the studied materials, solidstate 27Al NMR spectroscopy was used to achieve more detailed information about the local atomic arrangement in the Al-containing species. Fig. 5 shows the 27Al SPE/MAS NMR spectra recorded for the as-synthesized samples, with the simulated components indicated as dashed lines. The spectra of as-synthesized AC_Al_1, AC_Al_2 and Al_OxHydrox samples are similar, showing single peaks with diso = 9.5 ppm (for AC_Al_1), 9.6 ppm (for AC_Al_2) and 12.0 ppm (for Al_OxHydrox), which are associated with Al3+ ions bonded to oxygen in octahedral coordination (AlO6 sites) [12,21,39]. This is consistent with the above mentioned detection of boehmite and bayerite in the XRD patterns of these samples. The 27Al NMR spectrum of the as-synthesized BC_Al sample is also dominated by a strong peak with diso = 10.8 ppm, associated with AlO6 sites [12]; however, this peak is much broader than the ones observed with similar chemical shifts in the spectra recorded for the other samples shown in Fig. 5. At the same time, broad peaks associated with AlO4 (around 66 ppm) and AlO5 (around 35 ppm) units are observed in the 27Al NMR spectrum of BC_Al sample. The detection of these broad and asymmetric peaks due to AlO6, AlO5 and AlO4 units is characteristic of materials containing Al–O linkages in structurally disordered environments, such as it is

Fig. 5 – Solid-state 27Al SPE/MAS NMR spectra for the as-synthesized samples. The dashed lines indicate the simulated spectral components. (A color version of this figure can be viewed online.)

CARBON

9 3 (2 0 1 5) 7 5 1–76 1

found in amorphous aluminas [19,30], in many transition aluminas [12,22,23] and in Al-containing glasses [27]. These findings are thus consistent with the amorphous nature of the Alcontaining phases present in BC_Al sample revealed by XRD (Fig. 3). The relative fractions of AlO4 and AlO5 units, determined by fitting the 27Al NMR spectrum to a set of secondorder quadrupolar lineshapes including a distribution of quadrupole couplings and chemical shifts, are 18% and 9% for BC_Al sample, respectively (see Table 1). The isotropic chemical shifts obtained from the fittings (also given in Table 1) are clear indications that the three peaks detected in the NMR spectrum of this sample are indeed associated with AlO6, AlO5 and AlO4 units, in increasing order of chemical shifts [12]. The mean Cq values determined for the AlO5 units are found to be smaller than the values corresponding to AlO6 and AlO4 units, which is consistent with previous reports on amorphous aluminas [30]. However, these Cq values are to be viewed as useful only for comparisons between spectra analyzed with the same fitting model, since, in the case of featureless resonances as the ones shown here, the recording of spectra at more than one magnetic field strength would be required for an accurate determination of the quadrupolar parameters [12,26–28].

3.2.

Heat-treated samples

In the sequence, the chemical and structural changes caused by heat treatments performed on the AC_Al_2 and BC_Al samples will be discussed in detail. These two series of

757

heat-treated samples were chosen for a detailed analysis since, as described above, the corresponding as-synthesized materials presents either only AlO6 units (for AC_Al_1 and AC_Al_2 samples) or a distribution of AlO6, AlO5 and AlO4 units (for BC_Al sample); it is thus important to understand how these different types of carbon-supported species are altered when the material is heated up. Fig. 6 shows the XRD patterns of as-synthesized and heattreated AC_Al_2 samples. As previously discussed, the endothermic events detected in the range 200–300 °C in the TG and DSC curves recorded for the as-synthesized AC_Al_2 sample (Figs. 2 and 4) indicate the decomposition of the hydroxides and oxyhydroxides formed within the porous carbon matrix. Accordingly, the broad peaks present in the XRD patterns recorded for the samples heat-treated at 400–1000 °C (Fig. 6) evidences the presence of c-Al2O3 and d-Al2O3 with poor crystallinity. These transition aluminas are well known to be formed as intermediate structures during the thermal conversion of Al-based hydroxides and oxyhydroxides [22,23,40,41]. The main effect of the heat treatments in the 400–1000 °C range in the AC_Al_2 series is the progressive narrowing of the XRD peaks associated with the transition aluminas, pointing to thermally-induced crystallite growth [34]. It is important to note that similar findings are also observed for the carbon-free Al_OxHydrox sample (XRD patterns not shown), with the important difference that the transition aluminas present improved crystallinity (i.e., larger average crystallite sizes) in the absence of the carbon support.

Fig. 6 – XRD patterns recorded at room temperature for as-synthesized and heat-treated AC_Al_2 samples. (A color version of this figure can be viewed online.)

758

CARBON

9 3 ( 2 0 1 5 ) 7 5 1 –7 6 1

Fig. 7 – XRD patterns recorded at room temperature for as-synthesized and heat-treated BC_Al samples.

The most extreme effect caused by the presence of the carbon support on the crystallinity of the formed aluminas is observed for the BC_Al samples. As it can be seen in the XRD patterns shown in Fig. 7, no crystalline phase associated with Al compounds is detected for any heat treatment temperature; the XRD patterns show just the presence of the broad maxima associated with the turbostratic structure of the char [31,36]. Thus, in spite of the sizeable Al content of these samples (10 wt.% for the as-synthesized nanocomposite), no indication of the presence of crystalline Al oxides is obtained up to a heat treatment temperature of 1000 °C, evidencing the role played by the carbon support in preventing the full crystallization of these species. The chemical/structural changes caused by the heat treatments performed on the nanocomposites were also studied by solid-state 27Al NMR spectroscopy. Fig. 8 shows the spectra obtained for the as-synthesized and heat-treated AC_Al_2 samples, with the simulated components indicated as dashed lines. The relative contribution from AlO4 units increases as the heat treatments cause the conversion of Al hydroxides and oxyhydroxides into transition aluminas [22,29,30], which is consistent with the above discussed XRD data. The sample heat-treated at 600 °C contains AlO6 and AlO4 units with relative fractions of 65% and 35%, respectively (determined by spectral fitting, as discussed in the previous section); these values are comparable with the occupancies of 70% and 30% for AlO6 and AlO4 units, respectively, reported for a sample of c-Al2O3 [29]. In contrast, all 27Al NMR spectra recorded for the heattreated BC_Al samples (shown in Fig. 9, with the simulated components indicated as dashed lines) exhibit the typical features expected for disordered aluminas, with the presence of

three asymmetric, broad and partially superimposed peaks indicating a distribution of AlO6, AlO5 and AlO4 units. This is consistent with the XRD-amorphous nature of the Albearing phases in these samples revealed by XRD (Fig. 7). Thus, the results of solid-state 27Al NMR spectroscopy confirm that the aluminas formed in the BC_Al samples are structurally less ordered in comparison to the other series, evidencing the role played by the carbon support. It is particularly noticeable the quite high relative intensity associated with AlO5 units, which ranges from 9% (for the assynthesized sample) to 1% for the sample heat-treated at 1000 °C. This finding shows that the presence of the BC carbon support contributes to the preservation of the disordered Al sites, which is quite different from the AC-supported samples (where poorly crystalline transition aluminas with no AlO5 units were detected after heat treatments from 400 °C upward). The reduction in the intensity associated with AlO5 units is accompanied by an increase in the contribution from AlO6 units, which goes from 73% to 80% with the increase in the heat treatment temperature (whereas the relative fraction associated with AlO4 units remains nearly stable at 19%). This behavior is clearly a consequence of the progressive structural ordering caused by the heat treatments. It is interesting to observe how solid-state 27Al NMR (with the appropriate spectral fitting procedure here employed) is capable of catching this short-range ordering effect, even though no indication of bulk crystalline order is revealed by XRD. The XRD-amorphous or nanocrystalline nature of the Alcontaining phases present in the as-synthesized and heattreated nanocomposites, with distributions of AlOn units revealed by solid-state 27Al NMR, indicates that these

CARBON

9 3 (2 0 1 5) 7 5 1–76 1

759

Fig. 8 – Solid-state 27Al SPE/MAS NMR spectra recorded at room temperature for as-synthesized and heat-treated AC_Al_2 samples. The dashed lines indicate the simulated spectral components and the asterisks indicate the spinning sidebands. (A color version of this figure can be viewed online.)

Fig. 9 – Solid-state 27Al SPE/MAS NMR spectra recorded at room temperature for as-synthesized and heat-treated BC_Al samples. The dashed lines indicate the simulated spectral components. (A color version of this figure can be viewed online.)

760

CARBON

9 3 ( 2 0 1 5 ) 7 5 1 –7 6 1

materials may have interesting applications in fields where amorphous alumina nanoparticles are also promising [19]. For example, it has been reported that amorphous alumina nanoparticles are useful for the adsorption of chemical warfare agents [42]. The nanoscale dispersion of amorphous alumina and other inorganic oxides into charcoal has also been suggested as an efficient method for fluoride removal from drinking water [8]. By supporting the amorphous or nanocrystalline alumina nanoparticles on carbon materials that preserve their structural characteristics, it is then possible to obtain nanocomposites with outstanding thermal stability, high surface area and elevated dispersion of the active species.

4.

Conclusions

The use of solid-state 27Al NMR along with XRD allowed the assessment of the chemical nature and the degree of structural disorder in Al-containing compounds present in alumina–carbon composites synthesized by a wet chemical route followed by heat treatments. These results show that the nature of the carbon material used as support is of great relevance for the definition of the type and crystallinity of the alumina phases formed after heat treatments. Depending on the type of support and on the heat treatment temperature, alumina–carbon composites containing either nanocrystalline or amorphous alumina phases can be produced. It is generally observed that the presence of the carbon support results in the formation of more disordered Alcontaining phases, with reduced average crystallite sizes and significant contributions from Al sites in tetrahedral and pentacoordinated sites, as revealed by solid-state 27Al NMR spectroscopy.

Acknowledgments The support from Brazilian agencies CNPq, CAPES, FINEP and FAPES is gratefully acknowledged.

R E F E R E N C E S

[1] He Y, Zhang N, Gong Q, Li Z, Gao J, Qiu H. Metal nanoparticles supported graphene oxide 3D porous monoliths and their excellent catalytic activity. Mater Chem Phys 2012;134:585–9. [2] Zhao G, Shi J, Liu G, Liu Y, Wang Z, Zhang W, et al. Efficient porous carbon-supported MgO catalysts for the transesterifications of dimethyl carbonate with diethyl carbonate. J Mol Catal A: Chem 2010;327:32–7. [3] Safarik I, Horska K, Pospiskova K, Safarikova M. Magnetically responsive activated carbons for bio- and environmental applications. Int Rev Chem Eng 2012;4(3):346–52. [4] Oliveira LCA, Rios RVRA, Fabris JD, Garg V, Sapag K, Lago RM. Activated carbon/iron oxide magnetic composites for the adsorption of contaminants in water. Carbon 2002;40:2177–83. [5] Schwarz JA, Contescu C, Contescu A. Methods for preparation of catalytic materials. Chem Rev 1995;95(3):447–510.

[6] Schettino Jr MA, Gonc¸alves GR, Morigaki MK, Nunes E, Cunha AG, Passamani EC, et al. Synthesis, characterization and study of thermal behavior of nanostructured iron oxides embedded in porous carbon prepared from the decomposition of iron pentacarbonyl. In: Martinez AI, editor. Iron oxides. Hauppage: Nova Science Publishers; 2012. p. 19–51 [chapter 2]. [7] Ramos RL, Ovalle-Turrubiartes J, Sanchez-Castillo MA. Adsorption of fluoride from aqueous solution on aluminumimpregnated carbon. Carbon 1999;37:606–17. [8] Tchomgui-Kamga E, Alonzo V, Nanseu-Njiki CP, Audebrand N, Ngameni E, Darchen A. Preparation and characterization of charcoals that contain dispersed aluminum oxide as adsorbents for removal of fluoride from drinking water. Carbon 2010;48:333–43. [9] Petit C, Bandosz TJ. Role of aluminum oxycations in retention of ammonia on modified activated carbons. J Phys Chem C 2007;111:16445–52. [10] Liu F, Xu S, Chi Y, Xue D. A novel alumina-activated carbon composite supported NiMo catalyst for hydrodesulfurization of dibenzothiophene. Catal Commun 2011;12:521–4. [11] Lo´pez-Salinas E, Espinosa JG, Herna´ndez-Cortez JG, Sa´nchezvatente J, Nagira J. Long-term evaluation of NiMo/alumina– carbon black composite catalysts in hydroconversion of Mexican 538 C+ vacuum residue. Catal Today 2005;109:69–75. [12] Mackenzie KJD, Smith ME. Multinuclear solid-state NMR of inorganic materials. Oxford: Pergamon; 2002. [13] Freitas JCC, Cunha AG, Emmerich FG. Solid-state nuclear magnetic resonance (NMR) methods applied to the study of carbon materials. In: Radovic LR, editor. Chemistry and physics of carbon, vol. 31. Boca Raton: CRC Press; 2012. p. 85–170. [14] Brown IWM, Bowden ME, Kemmitt T, MacKenzie KJD. Structural and thermal characterisation of nanostructured templates. Curr Appl Phys 2006;6:557–61. [15] O’Dell LA, Savin SLP, Chadwick AV, Smith ME. A 27Al MAS NMR study of a sol–gel produced alumina: Identification of the NMR parameters of the h-Al2O3 transition alumina phase. Solid State Nucl Magn Reson 2007;31:169–73. [16] Rashidi F, Kharat AN, Rashidi AM, Lima E, Lara V, Valente JS. Fractal geometry approach to describe mesostructured boehmite and gamma-alumina nanorods. Eur J Inorg Chem 2010:1544–51. [17] Boissie`re C, Nicole L, Gervais C, Babonneau F, Antonietti M, Amenitsch H, et al. Nanocrystalline mesoporous c-alumina Powders ‘‘UPMC1 material’’ gathers thermal and chemical stability with high surface area. Chem Mater 2006;18:5238–43. [18] Vyalikh A, Massiot D, Scheler U. Structural characterization of aluminium layered double hydroxides by 27Al solid-state NMR. Solid State Nucl Magn Reson 2009;36:19–23. [19] Tavakoli AH, Maram PS, Widgeon SJ, Rufner J, van Benthem K, Ushakov S, et al. Amorphous alumina nanoparticles: structure, surface energy, and thermodynamic phase stability. J Phys Chem C 2013;117:17123–30. [20] Slade RCT, Southern JC, Thompson IM. 27Al nuclear magnetic resonance spectroscopy investigation of thermal transformation sequences of alumina hydrates. Part 1. – Gibbsite, c-Al(OH)3. J Mater Chem 1991;1:563–8. [21] Slade RCT, Southern JC, Thompson IM. 27Al nuclear magnetic resonance spectroscopy investigation of thermal transformation sequences of alumina hydrates. Part 2. – Boehmite, c-AlOOH. J Mater Chem 1991;1:875–9. [22] Meinhold RH, Slade RCT, Newman RH. High field MAS NMR, with simulations of the effects of disorder on lineshape, applied to thermal transformations of alumina hydrates. Appl Magn Reson 1993;4:121–40. [23] Pecharroma´n C, Sobrados I, Iglesias JE, Gonza´les-Carren˜o T, Sanz J. Thermal evolution of transitional aluminas followed

CARBON

[24]

[25]

[26] [27]

[28]

[29]

[30]

[31]

[32]

9 3 (2 0 1 5) 7 5 1–76 1

by NMR and IR spectroscopies. J Phys Chem B 1999;103(30):6160–70. MacKenzie KJD, Temuujin J, Okada K. Thermal decomposition of mechanically activated gibbsite. Thermochim Acta 1999;327:103–8. Malki A, Mekhalif Z, Detriche S, Fonder G, Boumaza A, Djelloul A. Calcination products of gibbsite studied by X-ray diffraction, XPS and solid-state NMR. J Solid State Chem 2014;215:8–15. Smith ME. Application of 27Al NMR techniques to structure determination in solids. Appl Magn Reson 1993;4:1–64. Neuville DR, Cormier L, Massiot D. Al environment in tectosilicate and peraluminous glasses: a 27Al MQ-MAS NMR, Raman, and XANES investigation. Geochim Cosmochim Acta 2004;68:5071–9. Freitas JCC, Schettino Jr MA, Emmerich FG, Wong A, Smith ME. A multiple-field 23Na NMR study of sodium species in porous carbons. Solid State Nucl Magn Reson 2007;32:109–17. Lee M-H, Cheng C-F, Heine V, Klinowski J. Distribution of tetrahedral and octahedral Al sites in gamma alumina. Chem Phys Lett 1997;265:673–6. Kunath-Fandrei G, Bastow TJ, Hall JS, Ja¨ger C, Smith ME. Quantification of aluminum coordinations in amorphous aluminas by combined central and satellite transition magic angle spinning NMR spectroscopy. J Phys Chem 1995;99:15138–41. Emmerich FG, De Sousa JC, Torriani IL, Luengo CA. Applications of a granular model and percolation theory to the electrical resistivity of heat treated endocarp of babassu nut. Carbon 1987;25:417–24. Emmerich FG, Luengo CA. Babassu charcoal: a sulfurless renewable thermo-reducing feedstock for steelmaking. Biomass Bioenergy 1996;10(1):41–4.

761

[33] Shin DC, Park SS, Kim JH, Hong SS, Park JM, Lee SH, et al. Study on a-alumina precursors prepared using different ammonium salt precipitants. J Ind Eng Chem 2014;20:1269–75. [34] Cullity B, Stock S. Elements of X-ray diffraction. Upper Saddle River, Massachusetts: Prentice Hall; 2001. [35] Massiot D, Fayon F, Capron M, King I, Le Calve S, Alonso B, et al. Modelling one- and two-dimensional solid-state NMR spectra. Magnetic Reson Chem 2002;40:70–6. [36] Li ZQ, Lu CJ, Xia ZP, Zhou Y, Luo Z. X-ray diffraction patterns of graphite and turbostratic carbon. Carbon 2007;45:1686–95. [37] Sato T, Sato K. Preparation of gelatinous aluminium hydroxide from aqueous solutions of aluminium salts containing sulphate group with alkali. J Ceram Soc Jpn 1996;104:377–82. [38] Gunawan R, Zhang D. Thermal stability and kinetics of decomposition of ammonium nitrate in the presence of pyrite. J Hazard Mater 2009;165:751–8. [39] Isobe T, Watanabe T, d’Espinose de la Caillerie J, Legrand A, Massiot D. Solid state 1H and 27Al NMR studies of amorphous aluminum hydroxides. J Colloid Interface Sci 2003;261:320–4. [40] Levin I, Brandon D. Metastable alumina polymorphs: crystal structures and transition sequences. J Am Ceram Soc 1998;81:1995–2012. [41] MacKenzie KJD, Temuujin J, Smith ME, Angerer P, Kameshima Y. Effect of mechanochemical activation on the thermal reactions of boehmite (c-AlOOH) and c-Al2O3. Thermochim Acta 2000;359:87–94. [42] Saxena A, Srivastava AK, Sharma A, Singh B. Kinetics of adsorption of 2-chloroethylethylsulphide on Al2O3 nanoparticles with and without impregnants. J Hazard Mater 2009;169:419–27.