1 Introduction

Bitumens are thermoplastic, dark and sticky solid, semisolid, or viscous liquid materials that are soluble in carbon disulfide and are mainly made of high molecular weight hydrocarbons [1,2,3]. Bitumen from the refining of crude oil is routinely utilized as a roofing felt material [4] in the construction and architecture industry [5] and in road construction and waterproofing/insulating applications due to its rheological, physical, and chemical properties and excellent adhesive and water-barrier properties [5,6,7]. In parallel, naturally occurring bitumen, e.g., gilsonite (GLS), has been extensively applied as an asphalt modifier, additive to drilling mud, and printing ink material [8, 9].

Polyfurfuryl alcohol (PFA) is a liquid resin with a viscosity in the range of 100–30000 cp (at 25 °C) and can be cross-linked in the presence of Bronsted or Lewis acid catalysts to yield black solid networks with high mechanical, chemical, and thermal stability [10, 11]. Due to the toxicity of formaldehyde residues leaving cured compositions, PFA resins are preferred over conventional phenol/formaldehyde resins [12]. PFA resins are also characterized by their ability to leave a high percentage of carbon char ca. 63 wt% after pyrolysis, which still benefits from a high strength-to-density ratio [13]. The authors have been active in the field of investigating the synthesis, curing kinetics, and application of furan resins to develop intumescent coatings for several years [14]. Increasing the char residue percentage and improving the thermal stability of PFA has attracted our attention in recent years. Previously, furan resin/bitumen (50/70 penetration-grade bitumen) mixtures were investigated by Bostancioglu et al. with the aim of modifying the mechanical behavior, thermal and moisture resistance of bitumen, and asphalt properties [15, 16]. Accordingly, the modification of the bitumen with furan resin increased the aging resistance and consistency of the bitumen and decreased its temperature and moisture susceptibility. In other words, the permanent deformation resistance of bitumen at high temperatures improved in this way. The studies also revealed that the modification increased the stiffness and elasticity of the bitumen and developed rutting (permanent deformation) resistance [15, 16]. The composite can be commonly used as a binder in refractory materials [17], but it can also be used as an intumescent coating, as previously reported by the same authors [18]. The effect of adding carbon black or mesophase pitch to novolac/furfuryl alcohol resins on the carbonization performance of the resulting composite material was also investigated by Zhang et al. [18]. According to their findings, mesophase pitch participated in the carbonization of the resin binder, resulting in an amorphous carbon structure with a homogeneous morphology, porous structure, and dimensional expansion. Thermal analyses showed that a major proportion of the weight loss occurred below 700 °C. The results also showed that the cured resin (NHF-R) released the greatest amount of volatiles during the heating process, followed by the resin containing mesophase pitch (NHF-M) and then the resin containing carbon black (NHF-C). Furthermore, the total weight losses for the NHF-R, NHF-M, and NHF-C specimens were 57 wt.%, 47 wt.%, and 30 wt.%, respectively. Machnikowski et al. [19] also reported on a blend composed of novolac- and resole-type phenolic resins filled with coal tar pitch aimed at preparing homogeneous carbon materials. The results showed that the addition of phenolic resins to coal tar pitch is effective in preventing structural ordering during pyrolysis. Compared with those of the resin-derived carbons, the activated carbons remaining from the blends were microporous materials with reduced surface areas and wider pore size distributions. The literature discussed above raises the question of whether the char-leaving characteristics of the resin can be improved by adding GLS as a particulate filler with high thermal stability and inherent char-leaving properties. This postulate can be justified considering the special structural features of GLS.

The complex chemical structure of GLS mainly consists of a bulk of aliphatic and aromatic compounds. Some of these compounds, including porphyrins, have attracted the attention of many researchers as biomarkers. However, the major constituents, including asphaltenes and polycyclic aromatic compounds, are well known for having a high ratio of carbon in their molecular structure compared to other atoms, including hydrogen and heteroatoms [20, 21]. Davis et al. classified GLS as a carbonaceous material. These structural features are expected to provide high thermal stability for GLS. The interesting thermal properties of GLS, including its high thermal stability and char yield, must be interpreted in the context of its low cost, high availability in some countries, including Iran, and excellent processability. Despite the previously described applicability of GLS and its well-foreseen ability to find new applications, very few attempts have been made in the literature to develop our knowledge of its thermal properties, whether modified or nonmodified, considering its complex chemical structure. With respect to the use of GLS as a filler in polymer compounding, Vellez et al. added GLS in the micrometer particle size range to carbon black-reinforced natural rubber compounds [22]. Rheometry and mechanical tests indicated a synergistic effect of GLS by facilitating carbon black incorporation during mixing and improving the dielectric strength. GLS also increased the dielectric strength of carbon black-reinforced natural rubber compounds. The results showed that adding GLS to carbon black-reinforced natural rubber resulted in a compound with viscoelastic properties, compressive modulus, and wear resistance similar to those of carbon black-reinforced natural rubber, suggesting that GLS is a new filler in natural rubber compounding, especially considering its low cost and high availability.

As a naturally occurring mineral, GLS contains some accompanying impurities, such as sulfur, mineral ash, and volatile organic compounds, leading to a decrease in its asphaltenes and carbon content [3]. It can be expected that these impurities not only negatively affect the filler char yield but also may affect the carbon char structure. Furthermore, sulfur produces harmful gases during pyrolysis via oxidation, which adversely affects the carbon char yield and stability [23]. Mineral ash is also expected to decrease the thermal stability of bitumen and play a catalytic role due to its metal ion content. In addition, ash impurities in GLS damage furnaces and the environment [24]. Nonhydrocarbon impurities present in GLS play a limiting role in developing other applications, such as printing inks and clean fuels [3]. Therefore, the elimination of the two sulfur and mineral ash impurities was highly important. Bitumen purification is possible physically via heavy media separation and froth flotation or chemically, e.g., acid leaching, which dissolves the ash-forming constituents in a liquid and consequently separates the pure coal by filtration [25]. Although physical methods are less expensive than chemical methods, they cannot meet the purity requirements alone according to many studies reporting on chemical purification methods, including alkaline or acid leaching of bitumen [26,27,28].

Considering the high need for fillers applicable in developing thermosetting composite materials with high thermal stability and char-leaving characteristics, modified, natural bitumen seems to be an interesting choice. However, this choice is highly neglected except for a very different previous art in the field of rubber technology by Vélez et al. who reported on the vulcanization of natural rubber in the presence of GLS [22]. Hence, this research tries to add a new class of fillers, i.e., modified GLS particles, to the current choices available for particulate composite developers and scientists. To this end, here, we report on a PFA resin/modified GLS composite after the chemical modification of natural bitumen particles via an acid leaching methodology to modify its sulfur and mineral ash content. The modified GLS particles were then added to PFA resin and cured to make a particulate composite. The thermal properties of the composite samples were compared to those of the neat PFA networks to investigate the thermal stability of the composites as a function of the filler weight percentage. In contrast to the old interest in GLS as an asphalt modifier or drilling mud additive, modified GLS particles have not been previously reported elsewhere as a filler for thermosetting resins except in a report on developing natural rubber compounds [22]. To the best of our knowledge, no previous studies have investigated the carbonization behavior of the resulting composite materials.

2 Experimental

2.1 Materials

Crude natural bitumen (Gilsonite, IM 105) was purchased from Sormak Mining Company (Tehran, Iran) and treated extensively before any further use. Nitric acid (65%, extra pure) was purchased from Merck KGaA (Darmstadt, Germany). Furfuryl alcohol (Daejung Chemicals and Metals Co., South Korea) and maleic anhydride (Merck KGaA) were used for the synthesis of polyfurfuryl alcohol resin (PFA) according to a previously reported procedure [29]. PFA resin characteristics were as follows: Mw = 1064.6 g mol−1, Mn = 579.57 g mol−1, PDI = 1.8, and dynamic viscosity = 150 Pa.s. p-Toluene sulfonic acid (Merck KGaA) was used as a curing agent for the fabrication of PFA-natural bitumen composites. Deionized water was prepared in-house by reverse osmosis using an aquaMAX™-ultra water purification system (Younglin Instrument Co., Ultra370 Series, South Korea).

2.2 Methods

2.2.1 Acid leaching

Crude natural bitumen (250–300 g) was first ground using an in-house developed horizontal ball mill filled with stainless steel balls. The ball diameters are 2.54 cm and 4.3 cm, and the mill length and diameter are 30.5 cm and 15.5 cm, respectively. The ball-milled material was sieved after 24 h of milling using an American standard test sieve series (Damavand Screen Co., Tehran, Iran). A major fraction of the ground material (80 mesh eq. to 177 μm) was separated and ball milled for an additional 48 h when sieved in the same way, and particles in 100 mesh eq. to 149 μm were separated. Grinding was continued in several steps using an ultrasound mill (Glen Creston Ltd., Stanmore, UK) to achieve final particles (300 mesh eq. to 44–53 μm) to provide a higher ratio of area to volume for leaching studies. The prepared fine powders were stored in plastic bags and kept desiccated at room temperature until they were used for acid leaching.

Natural bitumen particles were modified by the acid leaching technique according to a D-optimal experimental design approach using Minitab software (Ver. 16.2.0.0, Minitab Inc., State College PA, USA). The nitric acid concentration (10, 20, and 30 wt.%), leaching time (30, 75, and 120 min) and temperature (30, 50, and 70 °C) were selected as the independent input variables. The levels for the input variables were selected considering both previous [27] and preliminary experiments (data not shown here) and were coded as shown in Table 1. Fifteen experimental runs (see Table 1) were randomized and performed according to their run order in triplicate. The mean values for the output responses were included in the experimental design analysis when applicable.

Table 1 Run order, levels examined for input parameters in the acid leaching process of GLS particles according to a D-optimal design-of-experiments approach
Full size table

The acid leaching reaction was carried out in a 1000 mL three-necked reaction flask equipped with a water-cooled reflux condenser, a thermometer, and a mixer (Heidolph RZR 2020, Germany). A Teflon half-moon-shaped impeller operating at 300 rpm was used for mixing. First, natural bitumen particles (10 g) were charged into the reactor, and a definite volume of nitric acid solution was added to reach a certain concentration for an experimental run according to Table 1.

The reaction temperature was adjusted using an oil bath (IKA®-Werke GmbH & Co. KG, Staufen, Germany) preheated to a certain temperature for the same experimental run. To ensure the accuracy and precision of the reaction conditions, the oil bath temperature was validated using a PT 1000.80 (IKA®-Werke GmbH & Co. KG, Staufen, Germany) stainless steel external temperature sensor. At a certain time, the reaction was quenched by separating the particles from the reaction medium by vacuum filtration. The resulting particles were washed extensively with distilled water until a neutral pH was achieved in the filtrate. The sample was then placed in a forced-air convection oven (Fan Azma Gostar, 3M55, Iran) operating at 80 °C and completely dried for 8 h.

2.2.2 Sample preparation

PFA resin was filled with previously modified natural bitumen particles. To this end, modified natural bitumen particles (sample #15 in Table 1) were added to the resin at 5, 10, 15, and 20 wt.% and mixed homogeneously by a high-speed centrifugal mixing technique using a SpeedMixer™ DAC instrument (Hauschild Eng., Germany) operating at 3000 rpm for 2 min to ensure a uniform mixture in terms of both dispersion and distribution of the particles. Then, p-toluene sulfonic acid (3 wt%, relative to the PFA resin weight) was added to the sample as a curing catalyst and mixed uniformly again. The resulting powders (≈ 1.5 g, n = 5) were transferred into silicon molds and cured in a convection oven operating at 120 °C for 12 h.

2.3 Characterizations

2.3.1 Fourier transform infrared spectroscopy (FTIR)

The chemical structure of the GLS particles was characterized before and after acid treatment by FTIR spectroscopy using an Equinox 55 instrument (Bruker Optics, Germany). To this end, 5–10 wt.% of the sample was milled and mixed with dried KBr powder and pressed into a disk in a mold using a hydraulic press. FTIR spectra were acquired by collecting 32 scans in the 500–4000 cm−1 range with 4 cm−1 resolution.

2.3.2 Scanning electron microscopy (SEM)

The morphology, particle size, and particle size distribution of GLS particles before and after acid treatment (sample #15, Table 1) were determined using SEM (VEGA3, TESCAN, Brno, Czech Republic) in combination with energy-dispersive X-ray spectroscopy (EDXA, INCA250, Oxford Instruments, UK). The samples were spread uniformly over a conductive double-sided silver adhesive on a brass stub and coated with a thin layer of gold before SEM analysis. SEM micrographs were obtained at different magnifications using a secondary electron detector operating at 20 kV.

The particle size and morphological characteristics of the particles were analyzed by image analysis using ImageJ software (version 1.43b, NIH, USA). To this end, particles (n = 400) were selected randomly from the SEM micrographs, and their Fret diameter was determined. The sphericity index (roundness) of the particles was determined by measuring the ratio of the diameters for each of the 400 particles in both the horizontal and vertical directions [30]. In this way, a ratio of 1 indicates an ideal spherical particle. The more the ratio deviates from 1, the less the morphology is spherical and the more stretched the morphology is.

2.3.3 Particle size distribution

The particle size distribution of the natural bitumen particles (neat samples) was determined according to the ASTM B214-16 standard test method using a series of five sieves with different aperture sizes [31]. To this end, a definite weight of the material was placed on the top sieve and shaken for 15 min at 300 rpm in an orbital shaking pattern. The particle size distribution was determined by the weight percent of the particles retained on each screen according to Eq. (1):

$$\% {\text{Retained }} = \, \left( {W_{s} /W_{t} } \right) \times {1}00$$
(1)

where Ws and Wt denote the weight of the aggregate retained over a certain screen and the total sample weight, respectively. Assuming a log-normal distribution, the cumulative weight percentage of the particles retained on each screen was plotted on the probability scale against the logarithm of the arithmetic mean size of the sieve aperture for each of the two successive screens. The average particle size (dmean) was determined by the Edmundson equation:

$${d}_{mean}={\left(\frac{\sum {nd}^{p+f}}{\sum {nd}^{f}}\right)}^\frac{1}{p}$$
(2)

In this equation, n denotes the number of particles in a certain size range; d is one of the equivalent diameters; and p is an index related to the size of the individual particles, for which 1, 2, or 3 numerical values can be taken as an expression of the particle length, surface or volume, respectively. The frequency index, f, is the frequency with which a particle in a certain size range occurs. The frequency numbers 0, 1, 2, and 3 are related to the number, length, surface, and volume, respectively. Thus, the volume-surface or surface-weighted mean, dvs, (p = 1 and f = 2) can be calculated according to the following equation:

$${d}_{vs}=\frac{\sum {nd}^{3}}{\sum {nd}^{2}}$$
(3)

Span, another parameter used to describe the width (breadth) of a particle size distribution, was calculated according to Eq. (4):

$${\text{Span:}} \, \left( {{\text{X}}_{{{9}0}} - {\text{X}}_{{{1}0}} } \right)/\left( {{\text{X}}_{{{1}0}} } \right)$$
(4)

2.3.4 Moisture content

The moisture content was determined for all GLS samples (neat or acid-leached) according to the ASTM-D3173 standard test method. To this end, a definite weight of the sample (n ≥ 3) was weighed using a four-decimal place balance (Model 770, Kern, Germany) and heated in a convection oven at 110 °C for 1 h. Afterward, the sample was transferred to a desiccator and weighed again after cooling to ambient temperature. The moisture content was calculated according to Eq. (5) as follows:

$$Moisture \, content\, \left( {\% {\text{wt}}} \right) \, = \, \left[ {\left( {A - B} \right)/A} \right] \times {1}00$$
(5)

where A and B are the sample weights before and after drying, respectively.

2.3.5 Ash content

The ash content was determined according to the ASTM-D3174 standard operating procedure for all of the investigated natural bitumen samples. To this end, a definite weight of the sample (n ≥ 3) was transferred to a porcelain crucible and placed in a furnace (EHRET, DIPL.-ING. W.EHRET GmbH, Germany) operating at 700 °C for 4 h. The sample was then cooled to ambient temperature in a desiccator and weighed again. The ash content was calculated according to Eq. (6):

$$Ash \, content\, \left( {{\text{wt}}\% } \right) = \left[ {\left( {A - B} \right)/C} \right] \times {1}00$$
(6)

where A, B, and C are the weights of the crucible containing GLS after heat treatment, the empty crucible, and the initial GLS weight, respectively.

2.3.6 Sulfur content

The sulfur content of the samples was determined using the LECO method according to the ASTM-E1915 test method. The samples (n = 3) were first dried at 80 °C in a convection oven until reaching a constant weight and then ground until 100% passed through a No. 100 (150 µm) sieve. The samples were weighed (10.0 ± 0.1 g) and combined with 40.0 ± 0.1 g of dry silica (SiO2). The mixture was ground in a ring and puck mill until 100% passed through a No. 100 (150 µm) sieve and stored in a 250 mL glass bottle. The sulfur in the sample was converted to sulfur dioxide by combustion in a stream of oxygen, and its amount was measured according to its infrared absorption.

2.3.7 Thermogravimetric analysis (TGA)

Thermogravimetric analysis (TGA) of the PFA-GLS composite samples (n ≥ 3) was performed using a TGA1 instrument (Mettler-Toledo, Greifensee, Switzerland) operating at 10 °C min−1 heating rate under inert gas (N2) purge flowing at 50 mL min−1 rate in a temperature range of 25–800 °C. At 800 °C, the purging gas was changed to O2 to determine the ash content of the samples. The weight precisions were better than 0.01 wt% according to the calibration findings. To further analyze the interactions between GLS and PFA resin, the difference between the residual weight determined experimentally and the residual weight predicted theoretically were calculated for all of the composite specimens. The weight difference was calculated as [32, 33]:

$$\Delta \left( W \right) = M_{exp(wt)} - M_{{th\left( {wt} \right) }}$$
(7)
$$M_{{th(wt)}} = \sum\nolimits_{{i = 1}}^{n} {\alpha _{i} M_{{i\left( {wt} \right)}} }$$
(8)

where αi is the weight ratio of Mi and Mi(wt) is the weight of each component. Mexp(wt) and Mth(wt) are the weights of the whole component determined experimentally and theoretically, respectively.

2.3.8 Statistical analysis

The samples were prepared and analyzed in sufficient quantities and repetition numbers as mentioned appropriately for each sample preparation and test procedure. The results are reported as the mean ± standard deviation and were statistically analyzed using one-way analysis of variance when applicable.

3 Results and discussion

3.1 Particle size and distribution

The naturally occurring bitumen GLS is a rich source of aliphatic and aromatic hydrocarbons with a few olefinic groups [34]. Despite its long history in the scientific literature beginning in the 1860s, there are many unknown issues about the chemical structure and composition of GLS. However, there are certain data available indicating the presence of polycyclic aromatic compounds and asphaltenes that have partially transformed into carbonaceous materials after several millions of years of geochemical changes. The abundance of aromatic compounds and carbonaceous materials suggests that GLS is a thermally stable and highly char-leaving material and is thus a potential candidate for filling furan resins and making particulate composites capable of leaving a high percentage of char after pyrolysis. Following this postulate, the char residue percentage that remained after pyrolysis of a particulate composite made of polyfurfuryl alcohol filled by the modified GLS particles was investigated.

Due to its natural origin, GLS from mines located in Iran suffers from the presence of impurities such as sand, soil, and sulfur, which negatively affect the thermal properties of any composite material. To this end, the acid-leaching process was examined to at least partially eliminate these impurities. The sulfur, moisture, and ash contents of the GLS samples after acid leaching treatment are tabulated in Table 1.

SEM micrographs taken from GLS particles before and after acid treatment (sample 15, as an example) are depicted in Fig. 1. Before acid treatment, the surface of the GLS particles was smooth with signs of brittle fracture, including mirror-like surfaces, as shown by the orange-colored arrows; however, after surface treatment, the GLS particles became rough due to the presence of adsorbed smaller particles and not the intrinsic changes in the surface morphology of the particles themselves. The morphological changes observed may be due to the dissolution of minerals such as ash and pyrite in the bitumen during the acid treatment process or the oxidization of the organic functional groups and subsequent increase in the free surface energy of the particles in favor of the adsorption of the smaller particles provided by rupturing of the bitumen particles [25]. The oxidization of the organic functional groups on the GLS samples confirmed by EDXA was due to an increase in the oxygen content after acid treatment (Table 2). On the other hand, the particle size and particle size distribution of GLS changed drastically after the acid treatment process, which can be observed in the SEM micrographs before and after the acid washing process. The quantified results after image analysis using ImageJ software are provided in Fig. 2. As illustrated in Fig. 2a, the frequency of particles before acid treatment was greater than that after acid treatment at sizes greater than 8 µm, as expected. The results are also shown as a cumulative frequency profile (Fig. 2b), which indicates the degradation of particles through the acid-leaching process. The same change in the particle size distribution was found for the other samples examined here. The volume-surface or simply surface-weighted mean diameter of the particles (dvs) before and after acid treatment were calculated to be 11.21 and 12.26, respectively, according to Eq. (3).

Fig. 1
figure 1

SEM micrographs of GLS particles before (a showing mirror-like fracture surfaces by orange colored arrows) and after (b showing adsorbed, smaller particles using orange-colored arrows) acid treatment (sample 15) at two magnification ratios of 1 and 20 kx

Full size image
Table 2 EDXA analysis of GLS particles before and after acid treatment
Full size table
Fig. 2
figure 2

Plots of the frequency (a) and cumulative frequency (b) distributions of the GLS particle size before and after acid treatment for at least 400 particles (sample 15)

Full size image

When the logarithm of the particle size was plotted against the cumulative percent frequency on a probability scale, a linear plot was obtained in which the slope and the reference point were the two determining parameters. The reference point of the particle size equivalent to 50% (X50) on the probability scale is known as the geometric mean diameter, dg. The diameters of the particles before and after acid treatment were 8.09 µm and 8.11 µm, respectively. Zhang et al. attributed the size increase to the oxidation of structures resulting in a corresponding increase in the aromatic lamellar nucleus (La) and stack height of the aromatic lamellar nucleus (Lc) [35]. According to their findings, the acid treatment process significantly affects the microcrystalline structure of coal and enhances the aromaticity, which ultimately makes the carbon atom arrangement more orderly. Calahorro et al. reported on the coal demineralization process through HNO3 or HF acid treatment, resulting in a decrease in the surface area of the micropores and mesopores present inside the particles as a result of a reorganization of the material remaining after the acid treatment process [36]. The slope of the plot is the geometric standard deviation, σg, which is the quotient of the ratio (50% size)/(84% oversize or 16% undersize), which is equal to 1.43 and 1.5 for the number distribution of GLS particles before and after the acid treatment process, respectively. The diameter (d′g) and geometric standard deviation (σ′g) of the weight distribution of the particles were 9.98 µm, 10.97 µm, 1.51, and 1.53 µm before and after acid treatment, respectively. Zhang et al. reported greater adsorption of oxygen molecules on the surface of coal particles after acid treatment due to their greater specific surface area [35]. We are interested in adding a comment on their postulated mechanism and suggesting a corresponding change in chemical surface properties of the particles after the acid treatment process that ends with more affinity to adsorb polar oxygen molecules to reduce surface free energy thermodynamically. This increase in the surface energy provides a thermodynamic tendency to adsorb smaller, GLS particles, as previously observed in SEM micrographs (Fig. 1). The increase observed in d′g could be due to oxidation of the GLS structure and a corresponding increase in the oxygen content and moisture adsorption capability of the structure through acid leaching, which was confirmed by EDX analysis. According to the results, the geometric mean diameter increased partially through the acid leaching process. However, the σg and σ′g for the samples after acid treatment were greater, which indicates a wider particle size distribution after the treatment for both the number and weight distribution profiles.

The span values for the number distribution of the particles before and after acid treatment were 0.046 and 0.998, respectively. For the weight distribution of the particles, the span values were 0.04 and 1.04 before and after acid treatment, respectively. Both the number and weight span and the σg value increased for the samples after acid treatment, which confirms the wider particle size distribution after acid treatment.

The sphericity of the particles significantly affects the flow properties of solid materials, with the ultimate importance of efficiently mixing the filler particles with the resin and the final properties of the composite samples because it affects the aspect ratio of the particles. The abundance of particles with different horizontal/vertical diameter ratios before and after acid treatment is depicted in Fig. 3. According to the results, most of the raw particles (approximately 70.71%) possessed a sphericity ratio in the range of 0.5–1 to 1–1.5 (Fig. 3a). In contrast, the percentage of acid-treated particles in the same diameter range decreased to 54.54%. The number of particles in the other ranges shown in the same figure is indicative of an increase in the number of stretched particles (Fig. 3b). Thus, the sphericity ratio of the GLS particles decreased after the acid-leaching process due to the destruction of their bodies during the elimination of some impurities.

Fig. 3
figure 3

The ratio of horizontal to vertical diameter for at least 400 GLS particles before (a) and after (b) the acid treatment process

Full size image

Elemental X-ray mapping of the neat and acid-washed GLS particles is shown in Fig. 4, which shows the content and distribution of Si and sulfur. Accordingly, the content of carbon, Si, and sulfur atoms decreased significantly after the acid treatment process (see Table 2), but the oxygen content increased due to the oxidation of mainly the aliphatic structures by nitric acid [35, 37, 38]. Therefore, a greater surface area may be provided after treatment, resulting in a greater affinity of oxygen molecules to adsorb on GLS bodies [35].

Fig. 4
figure 4

Si- (green) and S-maps (red) of GLS particles before (a) and after (b) the acid treatment process

Full size image

FTIR spectra representing the functional group analysis of the raw and acid-washed GLS particles (sample 15 according to Table 1) are shown in Fig. 5. The wide and strong signal at 3400 cm−1 is due to the stretching vibration of the hydroxyl functional groups present in the sample. The high intensity of the bands appearing at 2922 and 2850 cm−1 were attributed to the aliphatic C–H stretching vibrations of the GLS particles, which indicates the presence of a high proportion of aliphatic hydrocarbons inside the chemical structure of the material.

Fig. 5
figure 5

FTIR spectra of neat and acid-treated GLS specimens (#15 in Table 2)

Full size image

The aromatic = C–H vibrations appearing between 3000 cm−1 and 3100 cm−1 (Fig. 5) are usually small and not well defined due to overlap with the O–H band vibrations reported by various investigators, but the peak at 795 cm−1 can be assigned to aromatic = C–H groups [38]. The presence of stretching vibrations of O–H functional groups of the silicate/clay mineral kaolinite is evident according to the signal that appeared at 3750 cm−1 [38,39,40]. On the other hand, the signal observed at 1650 cm−1 can be assigned to carbonyl bonds in amide functional groups, which improved in terms of intensity after the acid treatment process due to the oxidation of GLS [38]. Aromatic ring stretching vibrations are observed as a shoulder in the vicinity of the 1620 cm−1 signal [38]. The peak observed at 1551 cm−1 in the spectrum of the acid-treated sample was attributed to metal carboxylate groups [35]. In addition, Si–O–Si bands appeared at approximately 447 and 1036 cm−1, implying the presence of inorganic components in the gilsonite chemical structure [41, 42]. The vibrations of Si–O–Al cause bands at approximately 800 cm−1 and 720 cm−1, and the Si–O vibrations occur at approximately 667 cm−1. The strong band between 950 and 1100 cm−1 is due to anti-symmetric Si–O–Al or Si–O–Si stretching vibrations. The other strong band between 400 and 500 cm−1 is caused by Al–O or Si–O bending vibrations. The sulfur content was detected by S=O stretching vibrations at ~ 1178 and 1312 cm−1, which are persistent in acid-treated specimens. This band can be attributed to organic sulfur compounds such as sulfone and sulfonate. Notably, inorganic sulfur such as pyrite is removed mainly by nitric acid treatment [35]. In addition, this phenomenon can also be attributed to a reduction in the concentration of the other impurities and the consequent concentration of the acid-treated samples in terms of the sulfur content.

3.2 Moisture content

The effects of time, temperature, and acid concentration on the moisture content of GLS are depicted in Fig. 6. The results show that the moisture content decreases with increasing acid treatment time and temperature. According to the results, the time slope is more severe; hence, it is a more effective parameter than the temperature for decreasing moisture. However, the acid concentration harms the moisture content and causes an increase in the moisture content at higher levels. The increase observed in the moisture content of the GLS sample means that the sample weight increases after acid treatment in addition to the specific surface area, as confirmed in Sect. 2.1 [35].

Fig. 6
figure 6

Effects of time–temperature (Temp.) (a), time-acid concentration (b), and acid concentration-temperature (c) effects on the moisture content of the GLS during the acid treatment process

Full size image

3.3 Ash content

As shown in Fig. 7a and b, the acid treatment duration and temperature are the most effective parameters for decreasing the ash content, but the acid concentration has no considerable influence. The dependence of the acid treatment reaction time and temperature on sulfur and ash removal was also reported in previous studies [43]. According to Fig. 7c, the acid treatment time plays a more significant role in reducing the ash content than does the reaction temperature.

Fig. 7
figure 7

Effects of time-acid concentration (a), temperature-acid concentration (b), and time–temperature (Temp) (c) parameters on the ash content of the GLS during the acid treatment

Full size image

In coal, demineralization is attributed to the dissolution of water- and acid-soluble materials generated during the reaction of mineral impurities. This process also occurs due to the conversion of nonwater soluble impurities into soluble impurities resulting from the oxidation of mineral matter such as pyrite. Pyrite (FeS2) demineralization is also accompanied by desulfurization [44]. Nitric acid treatment of GLS removes those components of the mineral matter, such as pyrites, which need to be oxidized before solubilization [36].

3.4 Sulfur content determination

Figure 8 illustrates the effect of time, temperature, and acid concentration on the changes in the sulfur content during the acid-leaching process. The results imply that the sulfur content decreases dramatically with increasing acid concentration and reaction temperature but increases with increasing acid treatment reaction time. This behavior can be attributed to the changes in the sulfur content with time due to the elimination of the other impurities. According to previous studies, nitric acid can remove pyrite from coal [35].

Fig. 8
figure 8

Effects of time-acid concentration (a), temperature-acid concentration (C) (b), and time–temperature (Temp) (c) parameters on the sulfur content of the GLS specimens during the acid treatment process

Full size image

3.5 Thermal properties of PFA/GLS composites

Thermograms obtained by TGA analysis of neat, acid-treated GLS (sample 15), PFA resin, and their composites are illustrated in Fig. 9a. The same results were obtained and are abbreviated in Table 3 for clarity. The results show that Tmax, the temperature at the maximum mass loss rate, is 464 °C for neat GLS particles, and the step continues up to 500 °C. Afterward, the slope of the thermal degradation curve decreased but continued to increase to 640 °C. The weight loss in the first step is approximately 3.34 wt.% at 172 °C, which can be attributed to the evaporation of low-molecular-weight hydrocarbons in GLS. For PFA resin, thermal degradation starts at 220 °C and continues at 800 °C.

Fig. 9
figure 9

TGA thermograms (a) and ∆W curves (b) of the PFA resin, neat acid-treated GLS particles, and PFA resin filled with 0, 5, 10, 15, or 20 phr of acid-treated GLS particles

Full size image
Table 3 Thermal properties of PFA, GLS, and PFA/GLS composites according to TGA thermograms obtained under purge N2 and O2 atmospheres
Full size table

The temperature and weight loss percentage for the first step of weight loss are shown in Fig. 10. The temperature of this step can be attributed to the decreased evaporation of low-molecular-weight components in samples filled with GLS particles. Furthermore, the amount of low-molecular-weight compounds evaporated from the composites increased, as shown in Fig. 10. In other words, the more GLS filler present in the composite, the more low-molecular-weight components evaporate with increasing temperature. Nonetheless, the trend is absent in the composite samples containing 5 wt.% GLS, which have less loss of low-molecular-weight components. This behavior may be described according to percolation theory. In other words, as the filler content increases in the composite specimens, some pores are created that facilitate the loss of low-molecular-weight components more easily. Therefore, the weight loss due to the low-molecular-weight fraction loss is greater than expected in the composite specimens filled with GLS after reaching the 5 wt.% percolation threshold [45].

Fig. 10
figure 10

Effect of the GLS filler content on the thermal degradation of composites based on the PFA resin matrix against temperature

Full size image

The terminal temperature of evaporation for the low-molecular-weight components decreased with increasing filler content up to 5 wt.%. However, for the composite specimens containing more than 5 wt.% filler, the terminal temperature increased to 411 °C due to the porous structure remaining after evaporation. Nevertheless, the temperature never reached 425 °C, i.e., the final temperature of monomer evaporation in PFA resin in the absence of any filler. According to Fig. 9b, there is an impressive negative interaction between GLS and PFA resin (∆W < 0) in the composite specimens containing 5–15 wt.% GLS in the temperature range of 400–800 °C. This phenomenon can be attributed to two main reasons. The first is a reduction in the degree of crosslinking due to the interaction of GLS with the PFA resin structure. On the other hand, some impurities may exist in the GLS composition, affecting the degree of crosslinking and thermal stability [46]. The greater the GLS content is, the less negative the effect can be observed, resulting in a high char yield and improved thermal stability in the presence of GLS. As shown in Fig. 9b, for the composite specimen based on the PFA resin filled with 20 phr of the GLS filler, no considerable negative ∆W was observed after the 400 °C temperature reached a positive value at temperatures higher than 570 °C.

Tmax for the PFA resin was approximately 445 °C and decreased for the composite specimens to 421–429 °C except for the samples filled with 20 wt.% GLS. In addition, the char residue of the PFA resin under N2 and O2 atmospheres, especially under O2 atmosphere, was greater than that of the composite resin (Table 3). These results support that GLS unexpectedly reduced the thermal stability of the PFA resin due to oxidation reactions. In other words, the filler disturbed the integrity of the resin structure and provided a route for oxygen diffusion.

4 Conclusion

The effects of different parameters, including time, temperature and acid concentration, on the moisture, ash, and sulfur content of the GLS particles during the acid treatment of natural bitumen and gilsonite (GLS) were investigated as response parameters. According to the findings, the lowest ash content for GLS was obtained after treatment with 10 wt.% acid at 75 °C for 75 min. In addition, the acid concentration had a less significant effect on reducing the ash content in GLS particles compared to the two other input parameters. Sulfur, especially in its inorganic form of pyrite, was separated through the acid leaching process. The increase in the sulfur content as a function of the acid treatment time interval was due to the elimination of other impurities, which resulted in more concentrated sulfur residue in the GLS particles. The TGA results showed a negative interaction between GLS and PFA resin (∆W < 0) in the composite specimens containing 5–15 wt.% GLS in the temperature range of 400–800°C. The results also indicate that the structural integrity of the resin was disturbed in the presence of the GLS filler, resulting in facilitated oxygen diffusion and oxidative degradation reactions. All of these investigations aimed to prepare PFA/gilsonite particulate composites filled with GLS particles at low filler contents that do not negatively affect the overall thermal stability of the composite. There are ongoing studies by the same authors investigating the thermal, physical and mechanical properties of the same composites that are highly filled by thermally treated GLS particles.