Introduction

Carbon forms various allotropes that exhibit valuable characteristics, which has led to a wealth of studies of these allotropes to create new functional materials in a wide range of scientific fields. The number of reports on functional carbon materials has increased dramatically since nanocarbons, fullerene, graphene, and carbon nanotubes, have been developed. Functional carbon materials exhibit excellent electrical conductivity, thereby providing added value. Furthermore, carbon, which is an inexpensive conductive filler, can be produced on a large scale.

In the last two decades, woody biomass has drawn public attention as a sustainable energy source. Although the biomass is considered as an important source of carbon, its practical uses have been hindered by its disadvantages. The carbon content of wood is much lower than that of fossil fuels (petroleum, coal, etc.), and the biomass can be difficult to handle, because of its extremely limited solubility in any solvents.

T. Suzuki and K. Suzuki have studied the carbonization and gasification of wood to enlarge the versatility of woody biomass, revealing that the formation of crystallized mesoporous carbon with high electrical conductivity is accelerated at relatively low temperatures by coexisting iron or nickel salts in the carbonization procedures [1,2,3,4,5,6,7,8]. They claimed that such a high electrical quality of the crystalized carbon should be related to a novel carbon structure, “graphite-shell-chains” [9], induced by chemical reactions with metal-salt species. Moreover, Suzuki and Suzuki have promoted the practical application of the crystallized charcoal made from iron-loaded wood using a two-step carbonization method [5, 8], because iron salts can generally be purchased much inexpensively compared with nickel salts. In addition, iron-loaded charcoal possesses a potential to be utilized as carbon-based magnetic materials. Two-step carbonization is ordinarily used to synthesize activated charcoal. The first step is heat-treatment at relatively low temperatures to remove tar and pyroligneous acids, and then the second step is real carbonization at higher temperatures to obtain the desired charcoal.

The chemical roles of metal salts are, however, still controversial as to the electrical properties and formation mechanism of the nanoshells. A deeper understanding of changes in the chemical state of metals is important to elucidate the whole carbonization mechanism; however, there have been few studies focusing on the chemical changes of metal during the carbonization processes. Mössbauer spectroscopy is a powerful tool to determine the electron states of iron in solid: valences, spin states, internal magnetic fields, etc. Nevertheless, there have been no Mössbauer studies on iron-loaded charcoal, and therefore, it is expected that new findings of the physicochemical state of iron will be obtained from Mössbauer characterization. Raman spectroscopy is recognized to provide important information about graphite and graphite-like structures. Recently, the detection of G´(2D)-band due to sp2-carbon has been frequently adopted as a criterion of graphitization in the carbon-material sector. To obtain clues as to mechanisms of carbonization, it seems useful to investigate a dependence of products on the reaction process.

The aim of the present study is to understand the physicochemical changes of iron and carbon during carbonization. The characterization of iron and carbon in iron-loaded charcoal carbonized at various temperatures was performed using Mössbauer and Raman spectroscopy. Here, the one-step carbonization procedure was adapted in addition to the two-step one which Suzuki and Suzuki developed [8].

Experimental

Preparation of wood powder impregnated with iron(III) nitrate

Wood powder with Fe3+ ions was prepared essentially in accordance with the previous work [8]. Japanese cypress (Chamaecyparis obtusa) wood powder, which passed through a mesh opening of 1.4 mm but not through a mesh opening of 0.5 mm, was used as the starting material for the production of iron-loaded charcoal. The raw wood powder was impregnated with Fe3+ ions as aqueous iron(III) nitrate solution with a concentration of 3 wt% of an iron amount. Excess solvent water was removed using a rotary evaporator at about 60 °C under reduced pressure, and subsequently, the wood powder containing Fe3+ ions was dried under vacuum using an oil rotary pump at ambient temperature.

Carbonization of iron-impregnated wood powder

The schematic of the carbonization apparatus used in this study has been described in detail previously [2]. Two kinds of carbonization procedure were adopted in this study. The first procedure was one-step carbonization. Vacuum-dried wood powder (50 g) was added into a stainless-steel vessel, and the vessel was placed in a vertical stainless-steel tube reactor. The reactor was heated to a carbonization temperature (CT; 750 °C, 800 °C, 850 °C, or 900 °C) in a downstream of N2 gas (1 mL STP cm−2 min−1) using an electric furnace; the temperature was increased at a rate of 10 °C/min. The reactor was maintained at the desired temperature for 1.0 h, and it was cooled down rapidly to room temperature by removal from the furnace and blowing air over it. All carbonization processes, including the cooling process, were carried out under N2 gas flow.

The second procedure was two-step carbonization. Charcoal carbonized at 500 °C by the one-step procedure was used as the starting material and re-carbonized at 850 °C using the one-step carbonization method described above.

The iron-loaded charcoal samples carbonized at 750 °C, 800 °C, 850 °C, and 900 °C using the one-step method are denoted as Fe-C750-1, Fe-C800-1, Fe-C850-1, and Fe-C900-1, respectively, and the charcoal sample produced at 850 °C using the two-step method is abbreviated as Fe-C850-2.

Raman spectroscopy

Raman spectra were obtained using a Raman spectrometer (inVia Raman Microscope, Renishaw PLC, UK) with 532 nm laser excitation (0.75 or 1.5 mW). Back-scattered Raman signals were collected through a microscope (50 × objective lens) with a spectral resolution of approximately 3 cm−1, and the exposure time was 50 s. Raman measurements were repeated at more than 30 different points for each charcoal sample because the measurement scope (~ 2 μm diameter) was too small for a sample size, in which one measurement might lead to a possible incorrect result for a sample with inhomogeneity. The calibration of Raman shifts was carried out using the 520 and 1332 cm−1 for a silicon wafer and diamond, respectively.

Mössbauer spectroscopy

57Fe Mössbauer spectra were obtained in a conventional transmission mode on a Mössbauer spectrometer (Model-222, Topologic System Co.) with a 57Co (Rh) source (925 MBq). All measurements were carried out at 298 K. The Mössbauer spectra were analyzed using a curve-fitting processing with the MossWinn 4.0Pre. Program, assuming that all the spectra were composed of Lorentzian-shaped peaks. The isomer shift (IS) and Doppler velocity scale were calibrated with respect to the sextet of α-Fe at room temperature. The sample thicknesses of the iron-loaded charcoal samples ranged from 21 to 24 mg/cm2.

Results and discussion

Raman spectroscopy

Figure 1 shows a typical Raman spectra of the iron-loaded charcoal samples. Although there were not found significantly different spectral-shapes among the 30 measurement points for Fe-C750-1, Fe-C800-1 and Fe-C850-2, it was proved that the spectra showed a measurement-point dependence for Fe-C850-1 and Fe-C900-1 due to the sample inhomogeneity. The different Raman spectral-shapes due to the inhomogeneity are illustrated for Fe-C850-1 in Fig. 2, where a, b, and c set for the three different points. These results are described in details in the preceding paragraph. The whole line-shape of the Raman spectrum changes remarkably as CT increases from 800 °C to 850 °C, as shown in Fig. 1. The drastic spectral changes are caused by the sharpening of the D-band and the appearance of the G´-band, indicating crystallization of carbon (i.e., graphitization).

Fig. 1
figure 1

Typical Raman spectra of iron-loaded charcoal samples: a Fe-C750-1; b Fe-C800-1; c Fe-850–1; d Fe-C850-2; e Fe-C900-1

Full size image
Fig. 2
figure 2

Measurement-point dependence of Raman spectral line-shape observed on Fe-C850-1 sample: a, Raman spectrum with little or no G’-band; b and c, Raman spectrum with G’-band

Full size image

As is well known, D-, G-, and G´-bands are due to sp2-carbon. Unlike G-band, D- and G´-bands are not ordinary Raman bands and their shifts are dependent on the wavelength of an excitation laser. Furthermore, the intensity of G´-band is not proportional to that of D-band, despite the fact that G´-band is the overtone of D-band. Nearly perfect graphite that has few lattice defects shows a strong G´-band but an extremely weak or no D-band. Additionally, D-band intensity increases as a lattice defect increases [10]. Accordingly, the CT-dependence of the D-band line-shape and G´-band intensity suggest that graphitization progresses substantially between 800 °C and 850 °C.

Figure 2 demonstrates three Raman spectra for different three positions on the Fe-C850-1 sample showing different G´-band intensities and D-band line-shapes. The top spectrum exhibits no G´-band and a collapsed D-band as well as those of Fe-C750-1 and Fe-C800-1. Such spectra were obtained for approximately 15% of all the measurements points of Fe-C850-1 in this study. This is an important result, which indicates that Fe-C850-1 contains a certain amount of amorphous carbon inhomogeneously. The two other spectra exhibit G´-bands; however, the relative G´-band intensity in the bottom spectrum (Fig. 2c) is stronger than that in the middle spectrum (Fig. 2b). The intensity of G´-band relative to that of the G-band was dependent on the measured point on the Fe-C850-1 sample. The same results were obtained from Raman spectroscopic measurements of Fe-C900-1.

Conversely, even though being repeated 30 times, the Raman measurement of Fe-C850-2 did not show any spectral types of Fig. 2a, suggesting that Fe-C850-2 might contain little or no amorphous carbon; graphitization probably progresses in Fe-C850-2 more actively and homogeneously than in Fe-C850-1 and Fe-C900-1. In other words, the two-step procedure would be superior to the one-step procedure in graphitizing carbon in iron-loaded charcoal under the conditions set in this study when CT is set to 850 °C.

Moreover, the significant differences in Raman line-shape depending on measurement positions on Fe-C850-1 and Fe-C900-1 suggest that the rapid development of graphite structure in the iron-loaded charcoal might not be observed between 800 and 850 °C under different carbonization conditions containing the preparation of iron-impregnated wood powder. Further Raman studies on carbonization of iron-impregnated wood under different conditions are required to obtain more detail information about the formation of graphite structures.

Mössbauer spectroscopy

Figure 3 depicts the 57Fe-Mössbauer spectra of the iron-loaded charcoal samples. The resultant Mössbauer parameters and so on are summarized in Table 1. As shown in the figure and the table, the following four iron-species were deconvoluted in the present system: high-spin Fe3+ species, Fe3C, γ-Fe, and α-Fe, highlighted by brown, blue, orange, and green, respectively. Fe3C is an iron carbide, and γ-Fe and α-Fe are metallic iron. The room temperature IS values of iron in the carbide are in the range of 0.17–0.19 mm/s, and the values agree closely with those reported previously [11,12,13].

Fig. 3
figure 3

Mössbauer spectra at 298 K of iron-loaded charcoal samples: a Fe-C750-1; b Fe-C800-1; c Fe-C850-1; d Fe-C850-2; e Fe-C900-1; Brown solid curve (Doublet): high-spin Fe3+ species, Blue solid curve (Sextet 1): Fe3C, Green solid curve (Sextet 2): α-Fe, Orange solid curve (Singlet): γ-Fe, Red solid curve (Overall spectral line), Vertical black dotted line: IS of γ-Fe

Full size image
Table 1 Mössbauer parameters at 298 K of iron-loaded charcoal samples
Full size table

IS value is an important Mӧssbauer parameter and depends strongly on the oxidation state of iron. In this study, it is assumed that the oxidation number of Fe in Fe3C is zero; in other words, the carbide is assigned to Fe0 species, because the IS values of Fe0 species range from −0.25 to 0.63 mm/s according to Ref. [14]. The CT-dependence of the relative yields for the four species are illustrated in Fig. 4, where plot symbols of circles and triangles set for the one-step and two-step carbonization procedures, respectively, and the colors of the symbol for the respective iron-species.

Fig. 4
figure 4

CT-dependence of iron species yield in charcoal samples. Circle: iron-loaded charcoal synthesized by the one-step procedure, Triangle: iron-loaded charcoal synthesized by a two-step procedure, Brown: high-spin Fe3+, Blue: Fe3C, Green: α-Fe, Orange: γ-Fe

Full size image

Iron-loaded charcoal synthesized by one-step carbonization

In the Mössbauer spectra of the iron-loaded charcoal samples carbonized at 750 °C and 800 °C, two iron species were observed; a doublet absorption due to paramagnetic high-spin Fe3+ species and a sextet due to Fe3C. The absorption lines attributed to α-Fe (sextet) and γ-Fe (singlet) additionally appear in the Mössbauer spectra of Fe-C850-1 and Fe-C900-1. As can be seen both in Figs. 3 and 4, there are a huge decrease in the Fe3+ species and the corresponding increases in the reduced iron species (Fe3C, α-Fe, and γ-Fe) with the rise of CT from 800 to 850 °C, suggesting that the graphitization of carbon is closely related to the generation of metallic iron and/or Fe3C in the iron-loaded charcoal. The Fe3+ species disappears completely at 900 °C. These findings should presume that some reductive chemical reactions related to iron progress successively over the wide CT region.

Iron-loaded charcoal synthesized by two-step carbonization

It is interesting to note that the Mössbauer line-shape of Fe-C850-2 is entirely different from that of Fe-C850-1, which is caused by that the iron-species yields generated at 850 °C depend on the two carbonization methods: one-step and two-step. Furthermore, precisely comparing the results depicted in Fig. 4, it can be seen that the amount of reduction from Fe3+ to Fe0 species was almost same, however, the reduction products are greatly different; it seems that a substantial part of γ-Fe detected in the one-step carbonization procedure should transform to Fe3C and α-Fe in the two-step procedure.

Comprehensive explanation for Mössbauer spectra of iron-loaded charcoal

With respect to the Mössbauer spectra of Fe-C750-1 and Fe-C800-1, the doublet line shows the IS and quadrupole splitting (QS) value assignable to paramagnetic high-spin Fe3+ species. Although it is difficult to identify a precise chemical species for the paramagnetic Fe3+ specie with a doublet line-shape, it is a key factor for understanding the iron-speciation and reactions in the carbonization process to discuss what are such Fe3+ species in the Mössbauer spectra of Fe-C750-1 and Fe-C800-1. Two possible explanations are proposed for the doublet lines as follows.

  1. (1)

    The Fe3+ species are iron oxides (α-Fe2O3 and/or γ-Fe2O3); however, superparamagnetic relaxation [15] appears owing to the nano-sized oxide particles.

  2. (2)

    The Fe3+ ions form Werner-type multinuclear complexes.

The second explanation is probably not appropriate for the iron-loaded charcoal in this study, because Fe(NO3)3 hydrates will decompose completely and changed to Fe2O3 up to ~ 400 °C [16, 17]. Moreover, charcoal carbonized at temperatures exceeding 650 °C would contain neither H2O molecules nor hydroxy groups, and almost all the functional groups originated from wood constituents were thermally decomposed before the carbonization temperature was reached [18, 19]. Consequently, the superparamagnetic relaxation effect seems to provide a powerful explanation for the doublet lines.

The IS vales of the doublet lines due to Fe3+ species are compatible with those of hematite (α-Fe2O3) and maghemite (γ-Fe2O3) reported previously [20, 21]. The absolute QS values are much larger than that of ordinary hematite or maghemite [20, 21], suggesting that the ratio of iron atoms in a sphere with a low electrical symmetry is considerably larger. In other words, the large QS values seem likely to be evidence that a large portion of iron in the particles is present in the surface layers of Fe2O3 particles because of their nm-scale diameter. For example, if the diameter of α-Fe2O3 particle is 3 nm, only ten iron atoms are present in this diameter. Hence, the outermost and second surface layers contained approximately 30% and 20% of iron atoms, respectively. Similarly, this is applicable to γ-Fe2O3 and, therefore, it is very difficult to distinguish between α-Fe2O3 and γ-Fe2O3 on the basis of the Mössbauer parameters.

The Mössbauer spectra of Fe-C850-1 and Fe-C900-1 show that almost all Fe3+ ions were reduced to metallic iron and Fe3C from 800 to 850 °C. The primary iron species in these charcoal samples are γ-Fe and Fe3C, and their contents are compatible with each other. Metallic iron in α-form is only a minor component at ~ 5%. The yields of iron species in Fe-C850-2 are significantly different from those of Fe-C850-1, as shown in Fig. 4; it is found that there are a large increase (32%) in Fe3C and a comparable decrease (35%) in γ-Fe. It seems that a large portion of the γ-Fe component yielded in the one-step procedure might transform to Fe3C (major part) and α-Fe (minor part) in the two-step process as illustrated by a downward arrow (orange) and upward ones (blue and green) in Fig. 4. This finding is tentatively explainable in terms of the thermodynamic property of γ-Fe, which is in a metastable phase, forming only in rapidly cooling from a higher temperature than 727 °C, which is the eutectoid transformation temperature [22]; γ-Fe cannot exist as a thermodynamically stable species below this temperature.

Our possible explanation for the formation of Fe3C, γ-Fe, and α-Fe during the carbonization of iron-impregnated wood is described as follows. First of all, we postulate an iron-species precursor that could be generated around 500 °C and transforms directly to Fe3C from a higher temperature than 500 °C. Thus, the precursor is abundantly produced during the first stage in the two-step procedure and subsequently changes to Fe3C in the second stage. The Fe3C portion might not transform into any other iron species. On the other hand, the iron-impregnated wood yields only a small amount of the precursor during the one-step procedure, because the temperature rising rate of 10 °C/min is too fast to generate sufficiently the precursor while the reaction media was kept for an hour at 500 °C in the two-step procedure. Iron other than the precursor might change to γ-Fe in the temperature range above 727 °C. Most of the metallic iron maintains γ-form in room temperature by quick cooling; however, a small portion of metallic iron would transform from γ-form to α-form. No observation of metallic iron in Fe-C750-1 and Fe-C800-1 may suggest that the reduction to Fe0 does not sufficiently progress. In other words, the metallic iron may be detected by Mössbauer spectroscopy if charcoal is held at 750 °C or 800 °C more than 1.0 h.

Conclusion

The characterization of iron-loaded charcoal using Mössbauer and Raman spectroscopy provided information on the physicochemical changes in iron and carbon during carbonization, as mentioned below.

The yields of the iron species reduced in the iron-loaded charcoal have a positive correlation with the progress of graphitization; moreover, the carbonization conditions (e.g., iron content in wood powder, increasing rate of temperature, retention time at CT, etc.) other than CT should affect the chemical changes of iron and carbon. The two-step carbonization is expected to have an advantage over the one-step carbonization to obtain iron-loaded charcoal applicable to magnetic materials, since the total yield of ferromagnetic iron species, Fe3C and α-Fe, in Fe-C850-2 is approximately 80%.

A key outcome in this study is the suggestion of a chemical reaction path from Fe3+ species to Fe3C not through γ-Fe in the carbonization of iron-impregnated wood. Further Mössbauer studies on the effects of carbonization conditions are required to elucidate the formation mechanism of reduced iron species.