The Influence of Mineral Matter on X-Ray Photoelectron Spectroscopy Characterization of Surface Oxides on Carbon

Abstract

The chemical structure of coal is very composite, consisting of a heterogeneous carbonaceous matrix with variable degrees of “turbostratic” order and the inclusion and/or exclusion of mineral matter (ash). The formation of surface oxides on carbon has long been recognized as a key to understanding many chemical and physical properties of carbon materials relevant to their consolidated or emerging applications. The extent and nature of surface oxides can effectively be assessed by high-resolution X-ray photoelectron spectroscopy (XPS), which provides excellent insight into the functional nature of C-O moieties. However, the XPS analysis of ash-bearing carbons may be biased by the interfering effects of inorganics with the most relevant spectral ranges, namely the core levels O_1s and C_1s. The effect of ash components on the spectroscopic characterization of carbon is scrutinized here with reference to a sub-bituminous coal characterized by a fairly large ash content. The coal is subjected to different treatments, including devolatilization, milling, and oxidation. A synthetic carbon (Carboxen) is used as a reference sample for the correct assignment of the carbon–oxygen functionalities in the core-level XPS spectra (C_1s and O_1s) in the absence of mineral matter. On the opposite side, fly ash from an industrial coal boiler is analyzed to investigate the effects of mineral matter. It is shown that the establishment of non-uniform charging of the sample induced by ash provides a key to the interpretation of the XPS spectra of ash-bearing carbon samples. The positive charge on the surface, referred to as the charging effect, brings about a shift of the core-level binding energies towards higher values. Grinding of the samples or partial combustion emphasizes the charging effect. XPS analysis of the fly ash, where carbon is largely consumed and dispersed in the inorganic matter, confirms that charging arises from non-conductive aluminosilicates. These effects may induce remarkable changes in carbon and oxygen peak shapes and need to be accounted for to obtain correct interpretations of the XPS spectra of ash-rich carbonaceous fuels.

1. Introduction

The surface and microstructural properties of carbons are very relevant to their use in consolidated and emerging applications as fuels, adsorbents/catalysts for chemical and environmental applications, or electrodes/supercapacitors for energy storage [1,2,3,4,5,6]. The characterization of the surface chemistry of carbons is of primary importance for the proper design and tailoring of smart carbon materials. Among the most widely used and effective surface characterization methods are X-ray photoelectron spectroscopy (XPS) [7,8,9,10,11,12,13,14,15,16,17,18], Fourier-transformed infrared spectroscopy (FTIR) [19,20,21], temperature-programmed desorption (TPD) [7,22,23,24,25,26,27,28,29,30], and attenuated total reflectance (ATR) [31,32]. Due to its sensitivity to the chemical composition and environment of atomic species, XPS has been extensively utilized to determine the carbon structure features on the surface of graphite and graphene oxides, providing valuable insights into the mechanical and electronic properties of these materials [9,12,13,14]. Senneca et al. [7,15] used XPS to scrutinize the nature of carbon–oxygen functionalities as intermediates in the heterogeneous oxidation of different types of coals and synthetic carbons. Surface chemistry has been correlated with the exothermicity or endothermicity of reactions occurring during oxidation and desorption investigated by thermoanalytical methods (DTG, DSC, TPD). Studies on heterogeneous carbon–oxygen chemistry are relevant to the combustion and gasification of carbon from either fossil or biogenic renewable sources to improve efficiency and reduce CO₂ emissions to the atmosphere [33]. The nature of carbon–oxygen complexes and the mechanisms underlying the desorption of CO and CO₂ under steady and dynamic conditions have been further investigated in [34,35]. Results from oxidation/desorption experiments are fully consistent with the characterization of the nature and extent of surface oxides formed on the carbon substrate, assessed via XPS spectral intensities. In particular, the CO/CO₂ ratio in the off gas turns out to be a meaningful fingerprint of the relative abundance of “edge” oxides (ether, carbonyl, lactone) versus epoxy moieties assessed via XPS.

However, the XPS data interpretation of heterogeneous samples, characterized by non-uniform electrical conductivity, may not be straightforward. It is widely recognized that the positive charge generated on the sample surface during the photoemission process is neutralized by an electric current to the grounded sample when the sample’s conductivity is prominent, as well as by metals and semiconductors. In contrast, for insulating materials, neutralization is only partial, leading to the accumulation of a significant net positive charge on the surface [36,37,38,39]. As a result, the sample surface develops a positive potential, typically ranging from a few volts to several tens of volts, which results in a lower measured kinetic energy of the photoelectrons and, as a consequence, a higher apparent binding energy. If the charging is very strong, the shift in binding energy may be large to the point of hiding the photoelectron peak in the secondary electron tail. In intermediate cases, the BE shifts lead to difficulties in interpreting the chemical states of the elements or even in the identification of the observed chemical species.

The situation is further complicated for composite samples, in which case effects like differential charging [40], where regions with different conductivities are present, can entail different retarding potentials and lead to peak broadening and even distortions [41]. The differential charging effect depends on many factors, such as the size and specific resistivity of the sample, the lateral and vertical inhomogeneity of the structure, the chemical constitution of the surface phases within the sample, the total photoionization cross-section of the materials being analyzed, and the features of the incoming X-ray beam (size, shape, intensity, homogeneity, etc.) [42,43,44]. For such inhomogeneous samples, the use of a low-energy electron flood gun to compensate for the positive charge on the sample surface [42,43,44,45]—a common procedure with fully insulating samples—may not be resolute.

The interpretation of carbon XPS analysis has been debated in the literature [46,47]. The XPS analysis of the core-level spectra of carbon materials has been discussed in [46], underlining the ill-informed approach to analyzing the spectroscopic data. Indeed, the carbon species and banding energies for several carbon materials (graphite, carbon black, graphene, carbon nanotubes, carbides, and polymers) are summarized in [46], underlining that carbon materials with different preparation and analytical methods or storage time differ in carbon species and banding energies. The effect on the surface properties of carbon samples exposed to solar radiation was also investigated in [47].

Surface charging during XPS analysis has been accurately scrutinized in recent literature [48,49,50,51,52]. Baer et al. [48] summarize methods to control surface charging during the XPS analysis of insulating samples and different possible approaches for extracting useful information on binding energy. Because of the variety of approaches followed to extract binding energies for insulating materials, each with its own limitations, it is critical for researchers to identify the best procedure.

Gorham et al. [49] presented a method which employs XPS imaging to chemically detect spatially segregated multiwalled carbon nanotubes (MWCNT)-rich regions on an composite surface by exploiting differential charging methods. The results demonstrated that the XPS imaging of differentially charging MWCNT composite samples was an effective means for assessing dispersion quality.

Greczynski and Hultman [50,51,52] revised essential concepts and key experiments related to BE referencing. They discussed appropriate energy reference levels for conducting and insulating samples with and without electrical contact with the spectrometer, and defined criteria for the ultimate charge-reference method, based on adventitious carbon (AdC), which turned out to be the least reliable. They suggested easy-to-perform control experiments that refute the notion that the C_1s peak has a constant BE.

In this study, we investigate non-homogeneous carbons in which a conductive matrix (near-graphitic carbonaceous domains) coexists with dispersed dielectric domains (the mineral fraction of the coal, mainly aluminosilicates). The spectral features from the dielectric areas are expected to show significant shifts towards higher binding energy because of the net positive charge, whereas the spectral features of the conducting domains are expected to show no (or limited) binding energy shift by means of XPS.

The bias deriving from the influence of ash material is assessed by a comparative analysis of spectroscopy data obtained with different samples:

• Carboxen has been selected as an essentially ash-free reference carbon. It is an ideal model carbon to ensure the reliable peak assignment of XPS carbon–oxygen functionalities, ruling out interference due to the presence of inorganics.
• Char from a sub-bituminous coal, characterized by a large content of both inertinite and vitrinite macerals, has been investigated as a sample of an ash-bearing carbon.
• Fly ash from entrained flow combustion of the same sub-bituminous coal has been selected as an ideal sample of ash-enriched carbon. The sample contains 68.0% carbon and 15.7% ash. The latter is predominantly constituted by an amorphous alumina silica glass phase (62.1%) and by two crystalline phases, mullite (31.8%) and quartz (6.1%).
• Milled char samples have been investigated within the scope of elucidating how the exposure of metal oxides from the bulk to the surface of carbon, which is indeed enhanced by grinding, affects the characterization of the sample surfaces by XPS.

2. Materials and Methods

Experiments were carried out on (a) a synthetic carbon (commercial name Carboxen), (b) char from a bituminous coal from South Africa, and (c) fly ash from suspension firing of the same coal.

The synthetic carbon was supplied by Sigma–Aldrich, under the commercial brand of Carboxen 1000. This is a spherically shaped 180–250 μm carbon molecular sieve. It is synthesized by the pyrolysis of an organic polymeric precursor. The properties of Carboxen are reported in

Table 1. Proximate and ultimate analysis of the Carboxen samples [15].

Ash a	Volatiles a	C a	H a	N a
(wt%)	(wt%)	(wt%)	(wt%)	(wt%)
3.0	5.0	92.0	0.7	0.1

^a dry basis.

.

The South African coal is a medium-rank bituminous coal with a high content of inertinite macerals. Chars were produced in a lab-scale fluidized bed reactor (ID = 10 cm) at 1123 K under a nitrogen flow. The proximate and ultimate analysis results are provided in

Table 2. Properties of South African coal, char, and ash.

	Ash a	Volatiles a	Cfix a	C a	H a	N a	S a	Vitrinite Rr
	(wt%)	(wt%)	(wt%)	(wt%)	(wt%)	(wt%)	(wt%)	(%)
Coal	15.7	23.1	61.2	68.0	3.8	1.2	0.6	0.72
Char	20.4	4.6	75.0	75.4	1.2	1.8	n.d.	n.d.
Ash	SiO2	44.1	Al2O2	34.0	CaO	8.1	MgO	2.2
	K2O	0.62	Na2O	0.15	FeO	1.53	MnO	0.01
	TiO2	1.41	P2O5	2.35	SO3	2.08	Others	3.45

^a dry basis.

. The microstructure of South African char has been investigated by a combination of techniques, including porosimetric analysis, XRD, Raman, and EPR, reported in [53]. This sample can be considered a “turbostratic” carbon, with a structural order intermediate between that of amorphous carbon and crystalline graphite.

The South African char was ground by means of a laboratory-scale planetary centrifugal ball mill (RETSCH PM 100). The milling balls and drum lining were composed of agate. Approximately 60 g of char samples was loaded into the drum before each milling session; the milling duration ranged from 5 to 30 min, with a rotational speed of 500 rpm. Additional details on the milling process can be found in [54].

Fly ash samples came from a 240 MWe front-fired PC boiler with low-NOx burners burning pulverized South African coal. The ash samples were collected from the hoppers below the first and second rows of electrostatic precipitators and then mechanically sieved using mesh sizes of 75 and 150 μm. The results of the ash analysis are presented in Table 2. Additional details regarding their features can be found in an earlier study [30].

Oxidation treatments were carried out by heating the samples in an electric oven with an airflow. The specific time and temperature conditions for the oxidation treatment of each sample are provided in

Table 3. Carbon samples and their respective thermal treatments.

Sample	Heat Treatment Air
Fly ash	None
Carboxen	773 K 30 min
SA unmilled	573 K 2 h
SA milled 5 min	573 K 2 h
SA milled 30 min	573 K 2 h

.

SEM/EDX results on char particles and raw and milled fly ash are reported in Appendix A.

The XPS experiments were performed in the ultra-high vacuum chamber (UHV) of the SuperESCA beamline at the Elettra synchrotron radiation facility (Trieste, Italy).

A thin layer of powdered sample was deposited on circular adhesive foils (6 mm in diameter). A Au foil (3 mm × 6 mm) was located in good electrical contact with the sample, and it was used as a reference to calibrate the binding energy position.

Electrons were detected at a 20° angle relative to the surface normal using a Phoibos electron energy analyzer (SPECS GmbH, Berlin, Germany), equipped with a custom delay-line detector. The measurements were taken with the X-ray beam impinging at a 50° angle relative to the normal of the sample surface. The size of the X-ray beam on the sample was approximately 100 μm × 10 μm.

Survey spectra were acquired at photon energies of 650 eV to confirm alignment and examine potential interference from inorganic substances and contaminants. The selected photon energy allowed for the inclusion of both O_1s and C_1s peaks in the survey spectrum.

Core-level spectra for C_1s and O_1s were measured at photon energies of 400 eV and 650 eV, respectively, with overall energy resolutions of 80 meV and 150 meV.

After the removal of a Shirley background [8], C_1s and O_1s spectra were fitted with Doniach–Sunjic functions convoluted with Gaussians [16]. The XPS curve fitting was performed in agreement with the procedure developed by Levi et al. [7,15] and compared with the results of [9,13]. The following

Table 4. C_1s and O_1s binding energies (BEs) of various carbon–oxygen groups in electron volts (eV) [15].

C1s	C Sp2	C Sp3 C-C(O) a	C Vacancy b	Ether, Carbonyl	Epoxy, Hydroxyl	Carboxyl, Lacton
	284.3–284.5	284.9–285.0	283.8–284.7	285.7–285.6	286.3–286.0	287.7–288.2
O1s	epoxy	ether, hydroxyl		carbonyl, carboxyl, lacton
	532.6–532.0	533.5–533.2		531.1–530.7

^a nearest and next-nearest neighbours to O-bonded C atoms. ^b C atoms surrounding a single C surface vacancy.

reports the binding energies used by Levi et al. [15].

Microstructural investigations of raw materials were carried out by means of X-ray powder diffraction (XRD) analysis in the 2Θ range of 3–90° using a Rigaku Miniflex 600 automated diffractometer (Rigaku, Tokyo, Japan) equipped with a CuKα radiation source. Phases were identified by using the PDF-5 + 2024 database (ICDD—International Centre for Diffraction Data^®, Newtown Square, PA, USA) and the Rigaku Smart Lab II software v4.5.162.0. The determination of Reference Intensity Ratios (RIRs) was performed without considering the amorphous phase. The mass fraction of each crystalline phase can be determined using the ratio of diffraction peak intensities between crystalline phases, allowing quantitative analysis. The equation of the RIR method is shown as follows:

$$w_i=\propto {\displaystyle \frac{I_i^{max}}{R_i}}$$

(1)

where w_i is the mass fraction of an analytical phase i, $I_i^{max}$ is the highest peak intensity, and R_i is the RIR value [55].

Microstructural investigation of raw samples, SA milled, fly ash, and Carboxen, have been performed by XRD and the results are reported in Figure 1.

The XRD pattern of fly ash shows diffraction peaks (labelled as 1) corresponding to crystalline phases of Al₂(Al₂.Si)O₁₀, with an RIR of 82.2 wt%, and a minor diffraction peak attributable to αSiO₂ (labelled as 2), with an RIR of 17.8 wt% Figure 1a. Conversely, the XRD pattern of SA milled shows the presence of an amorphous phase and the coexistence of two different phases of silicon dioxide, which were αSiO₂, with an RIR of 90 wt%, and SiO₂, with an RIR of 10 wt%, (labelled as 3). Instead, the XRD pattern of Carboxen (Figure 1b) shows the presence of a single amorphous phase.

3. Results

3.1. Characterization of Carboxen

The details of the spectral components obtained by deconvolution of C_1s and O_1s spectra are shown in Figure 2. The peak assignments are reported in the figure legend, in which black dots and red lines represent the experimental data and the best-fit curves, respectively.

The C_1s spectrum exhibits a prominent peak around 284.51 eV, which corresponds to sp² carbons in C=C bonds within aromatic regions of unoxidized carbon rings and in aliphatic chains [7,9,15]. Additional components associated with sp³ and oxidized carbons appear on the higher-binding-energy side: the feature at 288.29 eV is attributed to HO-C=O [7,9,13,15], the peak at 286.12 eV is linked to epoxides and C-OH groups [7,9,13,15], and the peak at 284.93 eV corresponds to C sp³ and C-C(O) [7,13,15]. On the lower-binding-energy side, a component at 283.89 eV is associated with carbon vacancies [7,9,15].

The O_1s spectrum shows three prominent peaks: The main peak at 532.28 eV is related to photoemitted electrons from epoxy oxygen [7,9,15]. A peak at 533.39 eV on the higher-binding-energy side is attributed to ether and hydroxyl groups [7,9,15], while the component at 531.02 eV on the lower-binding-energy side is ascribed to carbonyl, carboxyl, and lactone groups [7,9,15]. Finally, a very weak component at higher binding energy is assigned to H₂O, likely resulting from the adsorption of hydroxyl groups in the pore structure [25].

3.2. Characterization of Milled and Unmilled Oxidized Coal Char Samples

In this section, the results obtained from milled (5 and 30 min) and unmilled coal char samples oxidized at 573 K for 2 h are reported. From the O_1s high-resolution XPS spectra shown in Figure 3a, it is clear that, while the O_1s core level is broadly comparable with the one detected for Carboxen for the unmilled char (even if the detailed features and contribution of spectral components are locally different), the O_1s spectra of the milled chars show very intense extra features shifted at ~538 eV for the 30 min milled sample, and at ~542 eV for the 5 min milled sample.

A similar, although less pronounced, result is found when C_1s spectra are analyzed. It is remarkable that even though all the three samples studied in this section underwent the same oxidative treatment, only the C_1s spectra of the two milled samples exhibit higher-binding-energy carbon components. In particular, for the 5 min milled sample, a clearly distinguishable peak emerges at around 288.31 eV, while for the 30 min milled sample, two distinguishable peaks emerge at around 288.19 eV and 289.39 eV. The identification of this last component is indeed rather controversial in the literature:

• Some authors reported that higher-binding-energy carbon components should reflect more oxidized forms of surface carbon [44,56];
• According to other authors, higher-binding-energy peaks might originate from the inhomogeneous charging artefacts that occur in dielectric compounds [41,45,50].

We believe that such large shifts and associated energy differences cannot be explained on the basis of the changing nature of surface oxides of carbon, as binding energy chemical shifts related to organic oxygen moieties lie mostly in the 527–538 eV range [57]. Instead, we interpret these strong shifts as the perturbation of O_1s electrons photoemitted by oxides due to differential charging effects induced by the dielectric nature of mineral components of the char. Char milling causes fragmentation and an increase in the surface area, and favours the emergence of ash inclusions on the surface (see Appendix A) where they are exposed to and detected by XPS. Within this picture, the peak centred at around 532 eV is due to the oxygen signal from the conductive carbonaceous matrix of the char, which is not affected (or only moderately affected) by the charging of the neighbouring mineral particles.

A confirmation of our interpretation comes from the analysis of the survey spectra, acquired at the same photon energy of 650 eV. As shown in Figure 3b, photoemission intensities related to the Si_2p and Al_2p core levels are detected, in agreement with the expectations as the inorganic fraction of the coal char is mainly composed of silica and alumina (see Table 2). Although the XPS signal is fairly low and noisy and the disorder of the system under investigation prevents thorough resolution of the spin–orbit splitting, for both core levels, we can recognize two features:

• A weaker intensity at the binding energy at which the signals from silica and alumina are expected (about 103–104 eV for Si_2p and about 74–75 eV for Al_2p [57]) that can be assigned to small mineral particles dispersed within the organic carbonaceous matrix that experiences weak or no charging effects.
• A more intense signal shifted by 6–10 eV towards higher binding energy that can be due to larger inorganic grains where photo-induced charging effects are stronger.

It is worth noting the correspondence among the charging-induced shifts in the photoemission spectra of the same sample measured at the same photon energy (650 eV): all three core levels under consideration (O_1s, Si_2p, and Al_2p) show a shift of about 9.5 eV for the 5 min milled sample and a shift of about 6.5 eV for the 30 min milled sample. This observation is consistent with the common origin of the spectral shifts, namely differential charging.

Relating spectroscopic features to the milling time is not straightforward. The emergence of a pronounced peak shifted by about 9 eV toward higher energies is observed after 5 min of milling. Further milling for 30 min is associated with the appearance of a peak of comparable intensity but with a smaller shift, nearly 6 eV, in the resulting spectrum. It may be speculated that early milling discloses ash inclusions of fairly coarse sizes, whereas prolonged milling reduces the size of disclosed mineral inclusions so that the related shift is attenuated.

Figure 4a–c report the C_1s spectra of the three coal char samples acquired at 400 eV photon energy. The spectral patterns associated with differential charging are still detected, but they are less pronounced than in spectral analysis performed at a 650 eV photon energy. This finding may be attributed to the larger kinetic energy for the C_1s photoelectrons (about 370 eV kinetic energy at 650 eV vs. 120 eV kinetic energy for 400 eV photons) that allows us to probe a deeper fraction of the insulating particles, thereby increasing the amount of signal shifted to higher binding energy. On the other hand, the largest part of the C_1s signal is still due to the conductive carbonaceous matrix of the char and does not experience charging-related shifts: only a small fraction of carbon is localized in ash and gives rise to the structure at high binding energy, which is much smaller compared to what is observed in the O_1s spectra.

Deconvolution of the C_1s spectra measured at 400 eV is also reported in Figure 4a–c.

The unmilled sample shows a main component peak at about 284.54 eV. This main component represents sp² carbons in C=C bonds in aromatic domains of unoxidized carbon rings and in aliphatic chains [7,9,15]. Other components are identified at other binding energy values: 284.98 eV related to C sp³ [7,13,15], 283.94 related to carbon vacancies [7,9,15], 285.46 eV related to ether [7,9,13,15], and 286.26 eV related to epoxides [7,9,13,15]. The C_1s signal of milled samples shows a main peak ascribed to sp² carbons (284.31 eV), a much smaller contribution of the sp³ component (284.92 eV), and an even smaller contribution peak at 283.69 eV related to carbon vacancies [7,9,15]. Oxidized carbons are identified on the high-binding-energy side: at 285.61 eV related to the ether component [7,9,13,15], and at 286.19 eV related to the epoxy–hydroxyl component [7,9,13,15].

The effect of milling on spectral patterns is remarkable, with a pronounced reduction in the sp³ component and an increase in the vacancy component as compared with the sp² component (see also Appendix B). This finding may be interpreted in light of the mechanochemical effects documented in previous work [54], where a remarkable increase in the intrinsic combustion reactivity of residual carbon in fly ash was observed upon milling. The mechanochemical effect in [54] was assessed by non-isothermal thermogravimetric analysis and temperature programmed desorption [54], even though it was observed that milling did not induce relevant changes in the porosity of the samples, nor in the ash content and composition. It was in fact postulated that mechanochemical activation was due to the increase in strong electron-acceptor sites that could be identified with carbon atoms with unsaturated bonds located at the edges of carbon anisotropic layered structures.

While we can be confident about the interpretation of the spectral components reported above, the situation is less clear for the smaller features at high binding energies, where disentangling structures due to different carbon species from the effects of differential charging appears to be a hard task.

3.3. Characterization of Residual Carbon in Fly Ash

The C_1s and O_1s high-resolution XPS spectra measured on coal fly ash samples are reported in Figure 5. As it is possible to notice from the O_1s signal, the high content of ash, in particular non-conductive aluminosilicates, generates more intense and more shifted features in the O_1s spectrum compared to what was observed for the previous set of oxidized char.

The sample shows a very intense C sp² component at 284.57 eV [7,9,15], because of the thermal annealing [54]. Two components at 285.41 and 286.28 eV were also identified. They could be assigned to ether [7,9,13,15] and epoxy–hydroxyl components, respectively, [7,9,13,15] while the component at 288.14 could be, in principle, assigned to the carboxyl and lactone group [7,9,13,15]. The most remarkable feature of this sample is the presence of a large amount of non-conducting metal oxide particles that cause a positive charge on the surface, shifting the core-level binding energies towards higher values. As for the spectra reported in Figure 4, the identification of the component at 289.27 eV is uncertain, since it can be attributed either to oxidized carbon forms or to differential charging effects.

4. Conclusions

The present study confirms that the spectroscopic characterization of surface oxides on carbon by high-resolution XPS analysis of ash-bearing carbons may be biased by interfering effects of inorganics with the most relevant spectral ranges, namely the core levels O_1s and C_1s. The comparison of the XPS spectra of unmilled and milled oxidized coal char samples with those of a synthetic ash-free sample (Carboxen) and of fly ash from suspension firing of the same coal sheds light on the effect of mineral matter on photoelectron spectral patterns. The O_1s and, to a lesser extent, the C_1s high-resolution XPS spectra of milled and oxidized coal char samples display high-energy spectral components that cannot simply be related to modifications of carbon–oxygen moieties. Similar shifts are observed when the Si_2p and Al_2p core levels are scrutinized. The similarity of spectral shifts observed with all the core levels is interpreted in light of a common mechanism, that is, differential positive charging of dielectric regions of the composite sample surface, most likely aluminosilicate grains, reflected by the emergence of high-binding-energy spectral components. Consistent with this framework, the O_1s and C_1s spectra of coal fly ash show qualitatively similar, but even more pronounced, spectroscopic patterns. It is remarkable that unmilled coal char samples do not display the same high-binding-energy spectral components observed in milled char and fly ash samples. It may be speculated that the milling of “as-prepared” and oxidized coal char emphasizes the effect of mineral matter by disclosing ash inclusions, whose crystalline core is eventually exposed to photoemission phenomena, from the carbon matrix. The C_1s spectra at the lower photon energy of 400 eV is less affected by the “differential charging effects”. Deconvolution of these spectra has been accomplished with some success and suggests a pronounced change in the carbon sp² vs. sp³ and vacancy components upon milling that may reflect the mechanochemical activation of carbon. The purpose of the present investigation is not to solve the charging effect issue but to provide evidence for it in the case of ash-rich carbonaceous fuels. Future efforts will be required by the scientific community to establish a quantitative relationship model between mineral components (such as the SiO₂/Al₂O₃ ratio) and charge effect intensity so as to provide a basis for XPS data correction.