Pyrolysis is a critical initial step in the conversion of coal to other forms, whether by combustion, gasification or liquefaction  . On subjecting coal particles to heat treatment, the coal particles undergo thermo-chemical decomposition with liberation of volatiles and the formation of a resultant solid residue, char. The reactivity of the char is dependent on both the physical and chemical changes that results from heat treatment of the coal particles. The role played by char in coal utilization processes has created a need for high quality; hence it is important to determine its chemical properties to ensure smooth operation  and high productivity of such processes as in fixed bed gasification  . The physical properties of coal which impact on coal conversion processes relate to those processes which results in charge of particle size  and surface morphology (density, pore size, pore size distribution)  . The chemical properties of coal which impact on the conversion processes are those which effect a change in the chemical constitution and these in turn may lead to changes in the physical properties such as the surface area  . It was generally found that increasing devolatilization temperature increases the structural orderliness of the resultant char carbon crystallite, which affects the lattice properties with increasing aromaticity  -  . Increase in aromaticity in coals indicates substantial chemical and spatial rearrangement    . It has generally been found that gasification and combustion reactivities decrease with increasing aromaticity  -  . Thus, thermal treatment leads to a more ordered and unreactive char matrix  .
Increased volatile matter release creates fine pores  , thus increasing the micropore porosity  and surface area  . All the changes indicate increased graphitization which can be expected to affect the intrinsic reactivity. The nature and extent of microstructural changes of a coal particle upon heat treatment depend on the coal composition  -   . Low rank coals have the features of high concentration of oxygen-containing functional groups and high proportion of mesopores and macropores, and thus play a dual role of active and exchanging sites during coal utilization processes. At lower temperatures, the heterogeneous nature of coal is still evident in the char, having higher proportions of oxygen, hydrogen and disorganized carbon, while at higher coalification temperatures, atomic O/C and H/C ratios of char fall. This is due to progressive dehydrogenation and deoxygenation reactions as polycondensed aromatic structures are formed and polyaromatizaton (i.e. growth in the size of the aromatic sheets becomes dominant)  -  .
Coal chars are comprised of non-uniform structures containing micro-, meso-, and marco-pores  . It is also reported that chars obtained under rapid heating conditions have rather porous structures  . It is reported that very low porosity may be in the chars obtained from coals with high inertinite and high mineral matter contents  . The porosity of chars is affected not only by intrinsic coal properties but also by pyrolysis conditions such as heating rate  , temperature  , and residence time  . Different behaviour of coals could originate from differences in their chemical structures: high rank coals consist of larger aromatic units with fewer ether and methylene bridges. The lower rank coals consist of smaller units with more numerous reactive ether and methylene linkages   . Organic oxygen is an important component of the overall structure and plays a key role in determining reactivity   . Hence, a decrease in oxygen content in coal leads to energy content increasing automatically  .
The pore structure of coal has long been the subject of intense interest and investigation and as such has been extensively researched  -  . This is in large part due to the importance of the pore structure to the passage of fluids in and out of coal, which naturally occur when coal is processed or burnt  . While the chemical structure of the coal determines the kinetics of the fluid-coal reaction at the surface, the porous structure dictates how much surface is available for reaction, and what role mass transfer will play in the overall rate of reaction   . Conventional analytical techniques such as gas adsorption and mercury porosimetry has been employed in evaluating coal pore structures, however, these techniques comes with some limitations as data from these techniques often lead to incomplete and contradictory conclusions. Generally, adsorption works best when measuring surface areas of nonporous or macroporous substances with chemically homogenous surfaces. Mercury porosimetry equally works best with macroporosity   .
In this study, the internal and external structures of chars obtained at six pyrolysis temperature were characterized by small angle X-ray scattering (SAXS) and X-ray diffraction (XRD) respectively. The internal structure was characterized by determining the basic physical properties, as specific surface area, specific volume of pores, pore size distribution and porosity. The external structure was characterized by determining the basic chemical properties, as aromaticity, interlayer spacing and crystalline height. Both SAXS and XRD serve as attractive alternative characterization techniques because, above all, they are both nonintrusive (requiring little material and sample preparation). The SAXS readily penetrates the entire sample making it possible for pore size distribution (micropore, mesopore and macropore) to be evaluated in a single experimental run: the capability of measuring both closed and open porosity of carbonaceous materials such as coal   . When X-rays penetrate materials, they are scattered off density contrasts within the material at a range of scattering angles  . The resulting scattering pattern, which is specific to the structure of the material can be analyzed to estimate surface areas and porosity features of the material  . Because SAXS scatters off density contrasts in coals, it not only scatters off density differences between the material and empty pores, but also scatters off density differences between their organic and inorganic components   . To correct this anomaly, the coal could be demineralized to minimize the inorganic components.
While the XRD gives detail account of the crystallite features (aromaticity, interlayer spacing, and crystallite height) in an experimental run, SAXS accounts for the morphology of the carbonaceous material   . Six coals of different ranks were used to investigate the coal to char transition and insights from the findings will be discussed. These coals covered a wide rank interval (lignite to anthracite) and have different maceral compositions. To achieve this, the selected coals were demineralized in order to study the effect of the organic part only and to establish links between the physical and chemical properties using spectroscopy technique.
2.1. Sample Preparation
Six coals of varying rank were used: a lignite coal from Germany, coded as LIG; a sub-bituminous coal from Nigeria, coded as SUB; two bituminous coals from South Africa (one is low volatile bituminous, coded as BIT-LV and the other, high volatile bituminous coal, coded as BIT-HV); a semi-anthracite from South Africa, coded as SA; an anthracite from South Africa, coded as ANT. The coal samples were subjected to coal preparation and pulverized to coal particle size of −75 µm by employing a mechanical size reduction jaw crusher (Samuel Osborne (SA) LTD, model: 66YROLL) and a Fritsch P-14 rotary mill containing ceramic balls(Model number: 46 - 126). The required particle size of −75 µm was finally obtained from screening the particles from the rotary mill using a 75 µm screen All the samples were stored under argon in sealed bags.
The prepared coal samples were demineralised by sequential leaching with hydrofluoric acid (HF) and hydrochloric acid (HCl) The HF (48%) and HCl (32%) were obtained from Associated Chemical Enterprise (ACE), South Africa.
2.2. Apparatus and Procedure
The char production sequence from the parent coal samples are as follows: The weighed coal samples (40 g) were placed in a ceramic boat and put in a horizontal tube furnace at atmospheric conditions Initially, the samples were left for fifteen minutes and flushed with nitrogen (AFROX, ultra high pure grade) at atmospheric conditions, to remove oxygen from the oven. A flow rate of 1 liter/min of nitrogen was used. The furnace was then heated at 20˚C/min to the target temperature, and held isothermal for 60 minutes. The target temperature ranged from 450˚C to 700˚C, while keeping the samples under a nitrogen atmosphere.
The XRD experiments were done at XRD Analytical and Consulting, Pretoria, South Africa following the procedure and experimental method detailed in Everson and co-workers  . All the XRD experiments were repeated three times and the final results are averages of three independent results from which experimental errors were calculated at 95% confidence level. The surface areas of the various samples were determined using the carbon dioxide adsorption CO2 DR method on a Micromeritics ASAP2020 surface area analyser  . Prior to CO2 adsorption, the samples (about 0.20 gram) were degassed under vacuum (10.0 µmHg), at 380˚C for 48 hours for chars. The evacuated sample was analysed at 0˚C in an ice bath. The results were processed using the Accelerated Surface Area and Porosimetry System (ASAP) 2020 software linked to the Surface Area Analyzer. All the adsorption experiments were repeated three times and the final results are averages of three independent results from which experimental errors were calculated at 95% confidence level.
The char morphology were performed at Council of Scientific and Industrial Research (CSIR), South Africa, using the analytical technique of small angle X-ray scattering (SAXS), following the procedure and equipment at CSIR, Nano-Structured Materials Centre detailed in Calo and Hall  . All the SAXS experiments were repeated three times and the final results are averages of three independent results from which experimental errors were calculated at 95% confidence level.
3. Results and Discussion
Detailed characterization of the demineralized coal samples used in this study is presented in Table 1. As expected and as can be seen from the petrographic analysis in Table 1, the least value of 0.35 was obtained for the lignite coal (low rank) while the value of 2.94 was obtained for the anthracite coal (high rank) indicating that all the coal ranks were covered in this investigation. Detailed characterization of all chars consisting of results obtained from well-known standard analytical techniques such as CO2 adsorption DR, SAXS and XRD are given in Tables 2-4, where chars are listed in the order of increasing rank. The chemical rank parameters as can be seen in Table 2 and Figure 1 followed the expected trend with increasing rank, that is, an increase in the crystallite height and aromaticity and a decrease in interlayer spacing. The crystallite height was determined to be in the range of 14 to 13 Å for the lignite char from 450˚C - 700˚C; 14 to 13 Å for sub-bituminous; 18 to 13 Å for light volatile bituminous; 16 to 13 Å for high volatile bituminous; 20 to 14 Å for semi-anthracite and 20 to 13 Å for anthracite.
The two low rank coals (lignite and sub-bituminous) seems to exhibit the same trend of slight decrease (a transformational change of 1 Å) of crystallite height with increasing pyrolysis temperature. The same story could be said of the other ranks; the two
Table 1. Properties of demineralized coal.
Table 2. Determined chemical properties of heat-treated coal (chars).
medium rank coals exhibited same trend of gradual decrease (5 and 3 Å respectively) of crystallite height with increasing pyrolysis temperature while the high rank coals (SA and ANT) exhibited similar trend of gradual decrease (6 and 7 Å respectively) with increasing temperature. The interesting observation on the crystallite height is that, at the final pyrolysis temperature of 700˚C, there was a convergence to a value of 13 Å for all coal ranks. The decrease in crystallite height and the convergence to same value at higher pyrolysis temperature could be attributed to an increase in the number of aromatic layers as reported in Dangyu and co-workers  and an indication of no more effect of heat treatment on the coal  . Lu et al.  has reported that coal crystallites consist of 2 - 4 aromatic layers on average (basic structural units (BSU) in stacks of aromatic molecules). The interlayer spacing was determined to be in the range of 3.67 to 3.78 Å for LIG from 450˚C to 700˚C; 3.66 to 3.78 Å for SUB; 3.54 to 3.64 Å for BIT-LV; 3.55 to 3.60 Å for BIT-HV; 3.47 to 3.52 Å for SA and 3.47 to 3.52 Å for ANT.
The interlayer spacing demonstrated an increasing trend with increasing pyrolysis
Table 3. Determined physical properties of heat-treated coal (chars).
temperature for all coal chars regardless of the rank. A transformational change of 0.11 Å was observed for LIG; 0.12 Å for SUB; 0.10 for BIT-LV; 0.05 Å for BIT-HV; 0.05 Å for SA and 0.05Å for ANT. The lateral adjustment in the value of the interlayer spacing for all the coal chars can be due to re-alignment of the basic structural units of the coal matrix which leads to the chars becoming more ordered and condensed      . Lu et al. has reported that the interlayer spacing is strongly affected by heat treatment temperature of 900˚C and above, and this result into a decrease in the interlayer spacing which invariably amounts to a more condensed char structure  . The condensed structure of the char is attributed to the removal of aliphatic side chains from the coal matrix and the reduction of hydrogen content which makes room for hydrogen radicals to re-combine forming higher aromatic molecules at higher temperatures  .
The aromaticity was determined to be in range of 0.66 to 0.76 for LIG; 0.67 to 0.80 for SUB; 0.91 to 0.99 for BIT-LV; 0.93 to 0.99 for BIT-HV; 0.96 to 0.99 for SA and 0.96
Table 4. Pore size distribution, porosity and DR surface area for the different char suites.
to 0.99 for ANT. The transformational change in aromaticity was determined to be 0.1 for LIG; 0.13 for SUB; 0.08 for BIT-LV; 0.06 for BIT-HV; 0.03 for the anthracitic group (SA and ANT) respectively. The aromaticity was found to increase with increasing
Figure 1. Variation of char’s chemical properties with temperature.
pyrolysis temperature for all coal char as a result of the breaking off of the side aliphatic chain carbons  . The gradual increase in aromaticity with increasing temperature was more pronounced for the low rank coals due to the fact that the aliphatic side chains are not strongly bonded to the coal matrix. Similar trend was also observed for the medium rank coals as the bond to the coal matrix is not as firm as with the high rank coals.
As the data in Table 3 and Figure 2 show, the specific surface area did not demonstrate any given trend with coal rank and increasing pyrolysis temperature, rather it was a dramatic display of values from the lowest to the highest pyrolysis temperature. However, it can be observed that the semi-anthracite coal (SA) has the highest value of 348 m2/cm3 at the final pyrolysis temperature of 700˚C. This implies that this coal sample is the most reactive of the coals examined, since it has the largest volume of active sites for reaction to take place at high temperatures   . The trend of increasing reactivity was followed by the two medium rank coals, having specific surface area value of 242 m2/cm3 and 248 m2/cm3 for BIT-LV and BIT-HV. The surface area was determined to be 260 m2/cm3 for LIG; 239 m2/cm3 for SUB and 125 m2/cm3 for ANT. From a quantitative point of view and considering the highest pyrolysis temperature of 700˚C, the data suggests that the reactivity order of the examined coals may be represented as:
In contrast to this observation at 700˚C, at lower heating temperature of 450˚C, and as expected from literatures due to the fact that coals were acid washed, the two medium rank coals were more reactive than the low and high rank coals. This could be
Figure 2. Variation of char’s physical properties with temperature.
attributed to the washing off of the alkaline and alkaline earth metals in the low rank coals that aids coal reaction  and the inactivity of the active minerals present in the high rank coals  . Though considerable variation was seen in the values of the specific surface area between coals from low to high pyrolysis temperature, the results fit well with published rank trends  , and when the conventional properties (proximate and ultimate analyses data) are put into consideration  .
As can also be seen in Table 3 and Figure 2, the radius of gyration (a parameter proportional to the surface area) increased gradually with increasing pyrolysis temperature and was more pronounced for the low rank coals. Though the pattern of increase did not exhibit any defined trend but of interest, it was observed that the two low rank coals converged to the same value of 15.50 nm (155 Å) at the final pyrolysis temperature of 700˚C. The two medium rank coals converged to the same value of 13.86 nm (139 Å) while the two high rank coals converged to the same value of 13.00 nm (130 Å) at the highest pyrolysis temperature of 700˚C. It is evident that the convergence to the same value by coals of same rank is a clear display of the mechanism of devolatilization and a demonstration of a decreasing anisotrophy tendency of the coals with increasing pyrolysis temperature as expected due to the complex character in coal and coal chars    . Coals of same rank seem to exhibit same devolatilization mechanisms that lead to more aromatic structure of the coal matrix where the size and the thickness of the carbon sheets increase with the heating temperature   . Thus, as the rank of the coal increases, the radius of gyration decreases.
Figure 3. Variation of crystallite features with micro- + meso-pore concentration.
char suites. As can be seen, there seems to be no contribution from the macropores and little contribution from the micropores, hence, the discussion in this paper will be focused on the mesopores (the major contributor to the reaction). Similar reference to the importance of the mesopore group during coal conversion processes has been reported in the work of Jagtoyen  . Table 3 also reveals that there is no common pattern of the contribution of the mesopores to the pyrolysis process among the coals of different rank. The mesopores was determined to be 68% at the lowest pyrolysis temperature for LIG and 58% at the highest pyrolysis temperature, implying the 10% of the pores were consumed in the process. The mesopores was determined to be 59% at 450˚C and 47% at 700˚C for SUB, with a transformational change of 12% pore consumption. The transformational change was determined to be 8% pore growth for BIT-LV; 7% pore growth for BIT-HV; 14% pore growth for SA and 4% pore growth for ANT. Pores in coals have a broad size distribution and form a constricted, interconnected network which defines the channels and pathways for fluids to move into and interact with the macromolecular structure of coals. Hence, porosity along with the surface area is considered as key parameters in the technical performance evaluation of coal during its utilization processes such as in fixed bed gasification. As can be seen in Table 4, the porosity in the lower ranked lignite and sub-bituminous coals is primarily in the form of macropores, which is considered as primary porosity. For the intermediate ranked bituminous coals, loss of primary porosity sets in and the evolution of secondary porosity (micro- and meso-porosity) starts such that 80% of the accessible porosity is contributed by micro- and meso-pores  . While in the higher ranked semi-anthracite and anthracite coals, micro-porosity predominates. Therefore, it could be inferred that, as the rank of the coal increases, the importance of macro-porosity decreases and micro-porosity becomes increasingly significant. Figure 3 reveals that coal porosity and pore size distribution varies with the degree of maturity of coals as measured by the indictors: aromaticity, interlayer spacing and crystallite height.
Again, aligning this result of pore size distribution with the analysis of specific surface area confirms the SA coal as the most reactive of the coals investigated. Previous work by Mare and co-workers has reported that porosity is predominantly composed of microporosity, which contributed the majority of the available specific surface area to the pyrolysis process   , but this is contrary to what was obtained in this study as revealed in Table 4, inferring that majority of the contribution to specific surface area comes from the mesopores. As reported in Gibaud et al.  and experienced in this study, the macropore surface area does not change significantly as a function of the heat-treatment temperature  , while the micropore surface area does not change significantly upon further heating, above 700˚C. The lower mesopore-micropore range (daimeter of 5 to 175 Å)  for some carbonaceous materials such as coal have been associated with the inter-layer spacing in clay minerals  , based on this fact, a plot of combined mesopore-micropore was constructed against the crystallite properties of the coals examined in this investigation to established some links as can be seen in Figure 3. No correlation was established, and this can be attributed to the complicated nature of coal and coal char  .
As the data in Table 4 show, the porosity increases with increasing pyrolysis temperature to a maximum value. Apart from the LIG and BIT-LV coals that had a fall after 600˚C, the other coals experienced the decrease after 650˚C. The maximum porosity was determined to be 30% for LIG, 20% for SUB; 21% for BIT-LV; 23% for BIT-HV; 23% for SA and 21% for ANT. If porosity would have been used for coal reactivity assessment as obtained in this work, LIG would have been assumed to be the more reactive, which is not the case in this investigation. This reinforces the notion that a range of rank-related factors (of which porosity is an important one) contributes to the high relative reactivity of low-rank coals such as lignite  .
Char micropore surface area was measured using adsorption of CO2 at 273 K and data analysis using the Dubinin-Radushkevich (DR) method is presented in Table 4 and as can be seen in Figure 4; a plot of the DR micropore against the char crystallite features reveals an almost linear relationship, consistent with what one might expect based on the relative ranks of the parent coals (Table 1). The low ranked lignite and sub-bituminous coals exhibited the highest volumes of micropores among the coal samples used in this study. The transformational change in the volume of the micropores was also more pronounced for the low rank coals than the higher rank coals. Thus, the micropores available for coal reaction during utilization processes increases with coal maturity as reflected in Figure 4. However, the limitation of CO2 adsorption technique in measuring coal and coal char surface areas, is that it only probes the micropore size range and thus cannot be used independently to describe the overall porosity or pore size distribution of coal and coal char. Results obtained from CO2, N2 and Mercury porosimetry are usually combined to obtain a complete description of the pore structure, surface area and porosity of coal  . SAXS has the advantage of probing a wider range of pore sizes in a single experiment than the gas adsorption (CO2 and N2) or Mercury porosimetry methods.
As can be seen in Figure 5, the relationship between the chemical and physical properties of coal during pyrolysis was established only for aromaticity and the physical properties that gave definite good correlation. Evidence from other works reported this demonstration of the transition from amorphous aliphatic fraction to a more condensed aromatic fraction to be attributed to the loss of components having higher proportions of oxygen, hydrogen and disorganized carbon as the transition progress from low to high temperatures  -     -    . Moreover, the structural significance of this relationship cannot be over-emphasized in the interpretation of the reactivity of coal char, even when the fundamental physical parameters does not change in a straightforward way as graphitization proceeds (increase in aromaticity). As depicted in Figure 5, the aromaticity: which is one of the parameters used in measuring the chemical stability of coal transformation stands as a good indicator of coal maturity because of the realignment of the carbon molecules gave definite relationship with the coal physical properties of specific surface area, mesopore and porosity, which was more pronounced for the low rank coals   .
Figure 4. Variation of crystallite features with DR surface area.
4. Concluding Remarks
Investigations of six different coals of different rank were performed in order to evaluate the relationship between char morphology and its chemical properties with non-invasive techniques of SAXS and XRD respectively. Although both analytical techniques offer interesting perspective for the thermal analysis of coal (internal and external
Figure 5. Plot of aromaticity against char’s physical properties.
features), it is realized that fractal dimension of the surface of a porous solid does not depend, theoretically on the size of the pores or the amount of surface but rather is an intrinsic characteristic of the surface itself but rather depends on other coal properties such as petrographic composition. Physically, the specific surface area obtained using SAXS role in determining the reactivity of a coal cannot be undermined as it reveals the accessible surface available for reaction. The crucial nature of SAXS analysis reveals the possibility of evaluating char reactivity based on its morphology from a single analytical technique without relying on the combination of other analytical techniques such as adsorption method (CO2 and N2) and mercury porosimetry. As far as the author could assess, this is a novel finding, the use of the SAXS analytical technique to assess the reactivity of coal char. Demineralization produced a material whose pore size distribution as determined by SAXS and aromaticity as determined by XRD is that of the organic component of the coal. Of the chemical properties obtained from the XRD, only the aromaticity gave a definite relationship with the physical properties obtained by SAXS. This is generally attributed to the increase in aromaticity as a result of loss of aliphatic, carboxyl, and carbonyl groups from the carbonaceous material and the fact that SAXS probes a wider range of pores, including pores that are closed to, or restricted in access by other techniques such as gas adsorption or mercury intrusion.
The work presented in this paper is based on the research supported by the South African Research Chairs Initiative of the Department of Science and Technology and National Research Foundation of South Africa (Coal Research Chair Grant No. 86880). Any opinion, finding or conclusion or recommendation expressed in this material is that of the author and the NRF does not accept any liability in this regard.