Introduction

Red mud is generated during beneficiation of bauxite ore by the aluminum industry and its valorization has attracted a lot of attention given the large volumes generated and the high alkalinity. Its application in different fields and its catalytic applications have been reviewed [13]. As an adsorbent, red mud has been studied for the removal of nitrate, phosphate, heavy metals, and organic substances from aqueous systems. The presence of iron compounds in red mud makes it a potential catalyst, and its application in hydrogenation, liquefaction, hydrodechlorination, hydrocarbon cracking, VOC oxidation, and pyrolysis has been studied. Different chemical and thermal treatments have been applied to improve the performance. In some processes, red mud has been applied directly where it generally functions as a pre-catalyst being transformed under reaction conditions to the active phase [4]. Other approaches have involved its pretreatment to enhance the performance or its application as a support for catalytically active phases [3]. When used directly, the spatial and temporal compositional variation of red mud sources must be taken into consideration since these may influence the catalytic performance. However, since it is such a cheap and abundant resource, it may be possible to accommodate such variations in behavior. By its very nature and given the large volumes of red mud produced on an annual basis, on its own, catalytic application of red mud does not offer a solution to its disposal and storage problems. However, it possibly does provide a component to the overall solution.

In our previous work, we have used unmodified red mud from different sources in India for cracking of methane. The hydrogen generation, the carbon deposition on red mud after exposure, and its application in heavy metal removal were discussed [46]. In this way, it was shown that two waste products—red mud and methane (a byproduct of refinery operations and of landfill dumping) could be combined to yield hydrogen and also a magnetic carbon-containing composite which has the potential for further application. In terms of the latter material, the form of carbon produced was generally graphitic, which limits further application. There is therefore potential importance in the production of more highly functionalized carbons, such as those containing structural nitrogen species, and also in the determination of the performance for the production of composite carbon products from other hydrocarbons. Furthermore, it is of interest to determine whether modifications of red mud can be made to both potentially enhance catalytic performance and also possibly mitigate variations from different red mud sources. Additional interest lies in the possible application of other iron-containing wastes for hydrocarbon decomposition. In this manuscript, some of these aspects are detailed including the application of toluene as an alternative carbon source, the use of acetonitrile as a reagent—either alone or in conjunction with other carbon-containing feeds—to introduce nitrogen functionality into the resultant carbon products, studies of the effect of modification on the red mud upon its efficacy and comparison with the historical legacy iron-containing waste from a former nailworks site. Preliminary characterization of carbon has been undertaken and its further characterization and application will form part of another study.

Experimental Procedure

The starting red mud was RM7 which has been characterized previously [4]. Toluene exposure was undertaken in a fixed bed reactor. Nitrogen at a flow rate of 100 ml min−1 was saturated with toluene at ambient temperature and passed over 1 ml of red mud in a fixed bed reactor maintained at the temperatures of 700 and 800 °C for different durations. The toluene conversion experiment was undertaken in a heating cycle and the reactor was maintained for about an hour at each temperature. The toluene conversion at different times was estimated by the difference in the toluene inlet and outlet peaks in a gas chromatograph (Nucon, India). The theoretical toluene concentration assuming attainment of equilibrium is 4.8 %. For obtaining nitrogen-doped carbon deposits, different nitrogen-containing precursors were used, i.e., acetonitrile in argon, acetonitrile in methane, and a mixture of acetonitrile and benzene in argon. The saturated gas was passed over 0.4 g of catalyst at a volumetric flow rate of 24 ml/min at 700, 800, and 900°C for different time durations till the back pressure buildup took place. The estimated acetonitrile concentration is around 200 ppm. In this set of experiments, the objective was to understand the effect of different hydrocarbons on carbon deposits.

Three types of modified red mud were used for methane cracking. Activated red mud (ARM) was obtained by boiling RM7 in dilute HCl, precipitating the iron-rich material from the solution by increasing the pH value to 8 and calcining the precipitate in air at 500 °C for 2 h. Reduced red mud (RRM) was obtained by passing a 1:3 mixture of H2 and N2 at 0.2 standard liters per minute (slpm) through the reactor at 800°C for 5 h. To obtain reduced activated red mud (RARM), the activated red mud was subjected to reduction in hydrogen. The modified red mud was subjected to methane flow of 0.2 slpm at 700, 800, and 900°C. A sample from a former nailwork factory site (at Lennoxtown in Scotland) was also used. The sample was tested and compared with unmodified red mud for methane decomposition using the procedure detailed elsewhere [4]. The H2 quantification was undertaken using online GC (Hewlett Packard 5890A) provided with a TCD and a Molecular Sieve 13X packed column. The sample after sieving and washing was oven dried to obtain a fine brown powder.

The samples after reaction were typically characterized by scanning electron microscopy (XL30 ESEM Phillips microscope), thermal analysis in an atmosphere of 20 % oxygen and balance nitrogen (TA Instruments, Q500), CHN analysis (CE-440 Elemental Analyzer), and powder X-ray diffraction (Siemens D5000 diffractometer with Cu Kα radiation). ICP optical emission spectroscopy (ICP-OES) was used to obtain information regarding the elemental composition of the samples. The ICP-OES instrument used in this work was an ICPS-7000 Shimadzu instrument. Approximately 0.3 g of the sample as well as certified reference material, Soil IAEA 7, was placed in a Teflon™ pressure vessel, and 6 ml nitric acid (65 % extra pure, Riedel-Deltaёn) and 4 ml hydrofluoric acid (40 %, Qualikems) were added. The vessels were capped, sealed, and heated following the digestion method cycles reported elsewhere with some modification [7], and the vessels were then cooled down for 2 h. Aqua regia solution was used to complete the digestion, where 12 ml of aqua regia was added to the samples. Each vessel was sealed and heated, which followed the above procedure. After digestion, the samples were transferred to a volumetric flask and diluted with deionized water to 100 ml. Surface areas were determined using a Micromeritics Gemini instrument by application of the BET method to nitrogen physisorption isotherms determined at −196 °C following sample outgassing at 110 °C overnight.

Results and Discussion

The results from RM7 exposed to toluene are shown in Figs. 1, 2, 3, and 4. XRD analysis undertaken on samples exposed to toluene typically corresponded to reduced oxide products of hematite (magnetite, wustite), iron, and iron carbide along with graphitic carbon depending on the temperature of exposure. At 700 °C, the main phases were magnetite and wustite, while at 800 °C, the main peaks were graphite, iron, and iron carbide (Fig. 1). The sample also showed cracking of toluene which is illustrated in the conversion curve in Fig. 2. The SEM images (Fig. 3) of the 800 °C sample shows the presence of some filaments corresponding to the presence of graphite in the form of CNTs in this sample. This was consistent with thermal analysis which revealed weight loss in the temperature range 450–650 °C due to oxidation of carbon (Fig. 4). The slight increase in weight at around 350 °C could be due to oxidation of low-valent iron oxide (wustite).

Fig. 1
figure 1

XRD pattern of red mud exposed to toluene at different temperatures (m-magnetite Fe3O4, w-wustite FeO, g-cliftonite C, q-quartz SiO2, c-iron carbide Fe3C, f-iron Fe) (from Ref. [8])

Fig. 2
figure 2

Toluene conversion as a function of temperature by red mud

Fig. 3
figure 3

SEM image of red mud exposed to toluene at 800 °C

Fig. 4
figure 4

TGA profile of red mud exposed to toluene at 800 °C (from Ref. [8])

The use of a nitrogen-containing precursor with RM7 led to the incorporation of N in the carbon deposit. In the case of acetonitrile, the maximum C (67.9 %) and N (1.8 %) deposition occurs at 800 °C (after 12 h on stream). When methane is used as the carrier for the acetonitrile, there is no nitrogen observed at higher temperatures (800 and 900 °C) and the amount of carbon deposited is less compared to that with only acetonitrile (in the range of 12 to ca. 40 % after 3 h on stream). This observation has been supported by the literature [9, 10] where it has been reported that the presence of hydrogen can lead to the formation of HCN which reduces the doping of nitrogen. With the addition of benzene to acetonitrile, there is a decrease in the overall amount of carbon deposited (carbon content in the range of 17 to ca. 48. % after 6 h on stream), but nitrogen content in the CNTs is observed at higher temperature (0.57 % at 900 °C). TEM analysis of samples prepared with acetonitrile showed a bamboo-like structure at 800 °C, while the samples prepared at 900 °C also comprised carbon spheres (Fig. 5). There appeared to be no amorphous carbon in these samples. XPS spectroscopy undertaken in the N1 s region and presented as Supplementary Information for the sample prepared at 800 °C suggests that there are possibly different nitrogen environments in the sample, demonstrating that there is a degree of heterogeneity in this respect which could impact upon its further utilization.

Fig. 5
figure 5

TEM images of nanocarbons forming from reaction with acetonitrile at a temperature of a 800 °C and b 900 °C and using a flow rate of 24 ml/min

The cracking of methane measured in terms of hydrogen formation rate showed different behaviors based on the pretreatment used (Fig. 6). Reduced activated red mud showed the highest activity compared to the other samples in terms of both hydrogen yield and reduction in activation time. The higher activity could be due to higher surface area (72 m2/g) compared to RRM (24 m2/g); ARM (139 m2/g) had an even higher surface area but would have needed activation time for reduction of iron oxide to iron and the lower activity could be due to some carbon deposition on the reduced iron. Furthermore, the activation process would have increased the iron content in the red mud and would also be affective in removing sodium and potassium. The hydrogen formation rate increased with increasing temperature. The carbon in the reduced activated red mud sample was a mixture of filaments and spheres and the XRD pattern is indicative of the presence of graphitic carbon (C) and iron (Fe) (Fig. 7). The hydrogen formation rate was also sensitive to flow rate of methane used as shown in Fig. 8 for unmodified red mud. This was similar at other temperatures as well and could due to greater availability of reactant and the kinetics of the reaction.

Fig. 6
figure 6

H2 formation rates for cracking of CH4 (0.2 slpm) over unmodified and modified red mud at 800 °C

Fig. 7
figure 7

a TEM image and b XRD pattern of RARM after methane cracking at 800 °C

Fig. 8
figure 8

H2 formation rates for cracking of CH4 over unmodified red mud at 800 °C

The performance of the waste from the former nailwork factory was compared in detail with that of unmodified red mud (RM7). Figure 9 shows the hydrogen formation rates for both the waste samples. On increasing the temperature from 600 to 800 °C, it can be seen that the nailwork factory sample showed higher hydrogen formation rate. However, beyond about an hour at 800 °C, the activity of RM7 was superior to that of nailwork factory sample. The difference is due to composition and surface area as explained further below. In addition, a decrease in hydrogen formation rate due to deactivation from carbon deposition was not observed in both samples and activity continued to increase for all samples up to the point at which reaction was stopped.

Fig. 9
figure 9

Hydrogen formation rates as a function of TOS for CH4 decomposition over the nailworks waste and RM7 samples in a temperature-programmed reaction from 600 to 800 °C. The CH4/N2 flow rate was 12 ml/min over 0.4 g sample

The difference in their performance can be attributed to their composition and surface area. As shown in Table 1, the iron content in both the samples is similar. However, the waste from the former nailworks has a higher carbon content and surface area. The higher surface area may have led to higher hydrogen formation rate in the initial stages, while the carbon content may have reduced the reactivity. XRD patterns of the post-reaction samples (Figs. 10 and 11) show the formation of Fe, Fe3C, SiO2, and graphite in the case of RM7 and FeO, Fe3C, FeFe4(PO4)4(OH)2.2H2O, SiO2, and graphite for the nailworks waste. The lower reactivity of the nailwork factory waste sample is also suggested by the fact that the iron may not fully reduced to Fe and FeO could be present in the post-reaction sample. As expected, the carbon content in the post-reaction RM7 (48.40 ± 0.10 wt%) was higher than that in the nailwork waste sample (41.60 ± 0.20 wt%). The TGA analysis of post-reaction samples (Fig. 12) shows an initial weight increase between 350 and 500 °C, possibly from the oxidation of reduced iron compounds followed by a decrease of up to 700 °C from carbon oxidation. Post-reaction RM7 with higher carbon content shows a larger weight loss. The first derivative weight change profiles show two mass loss peaks in the nailworks waste at around 480 and 600 °C, which indicate the presence of two different forms of carbon. In the case of the RM7 sample, there is only one broad peak, at approximately 600 °C, which could be attributed to a wide distribution in the nature of the carbonaceous species. Given that the carbon deposits obtained are mostly a mixture of graphite, iron, and iron carbide, they could have further application where magnetic properties are also of relevance such as a separable adsorbent. Similar magnetic material has been made using red mud and ethanol as the carbon source and used for dye removal. Modification by CO2 activation increased the surface area which made it suitable as a support for Pd catalyst [11]. Other studies have also reported magnetic carbon nanotubes for removal of dyes [12] and chromium [13]. Nitrogen-doped carbon nanotubes have been used for fuel cell applications [14] and have been prepared using iron catalysts [15].

Table 1 Composition and surface area of iron-containing waste samples prior to reaction
Fig. 10
figure 10

X-ray pattern for the post-reaction nailworks factory sample. 1: iron oxide (FeO), 2: iron carbide (Fe3C), 3: cohenite (Fe3C), 4: giniite (FeFe4(PO4)4(OH)2.2H2O), 5: graphite (C), and 6: quartz (SiO2)

Fig. 11
figure 11

X-ray diffraction pattern for post-reaction RM7 sample. 1: iron (Fe), 2: iron carbide (Fe3C), 3: quartz (SiO2), and 4: graphite (C)

Fig. 12
figure 12

TGA a weight loss and b first derivative weight change profiles for the post-reaction nailworks and RM7 samples

Conclusions

Red mud and other iron-containing waste samples show potential for use in hydrocarbon cracking leading to the formation of carbon-containing magnetic material in addition to hydrogen generation. Such magnetic composites may find use as easily separable adsorbents. Although such an approach is not likely to address the large amount of red mud being generated, it is an attractive option with potential for further modification of the carbon deposits for other catalytic and electrochemical applications. In terms of the latter aspect, it is of interest to note that nitrogen functionality can be introduced into the carbon-containing products through the use of acetonitrile as a reagent. Comparative studies have demonstrated a marked influence of the pretreatment of the red mud upon its activity for methane cracking with reduced activated red mud being significantly more active than its activated red mud parent and the red mud sample itself.