Effect of heat reflux extraction on the structure and composition of a high-volatile bituminous coal

Applied Thermal Engineering 109 (2016) 560–568

Contents lists available at ScienceDirect

Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Effect of heat reflux extraction on the structure and composition of a high-volatile bituminous coal Bin Tian a, Ying-yun Qiao a, Yuan-yu Tian a,b,⇑, Ke-chang Xie c, Da-wei Li a a

State Key Laboratory of Heavy Oil Processing, China University of Petroleum (East China), Qingdao 266580, China Key Laboratory of Low Carbon Energy and Chemical Engineering, Shandong University of Science and Technology, Qingdao, Shandong 266590, China c State Key Laboratory Breeding Base of Coal Science and Technology Co-founded by Shanxi Province and the Ministry of Science and Technology, Taiyuan University of Technology, Taiyuan 030024, China b

h i g h l i g h t s
A novel HRE process with CYC is proposed to dissolve coal.
Most of the aliphatic compounds in coal are extracted during HRE process.
The carbon crystallite structure of coal changes after HRE process with CYC.
The thermal degradation behavior of ER is significantly different from that of the SFHB.

a r t i c l e

i n f o

Article history: Received 24 May 2016 Revised 25 July 2016 Accepted 18 August 2016 Available online 20 August 2016 Keywords: High-volatile bituminous coal Heat reflux extraction Structure evolution Functional groups TG-FTIR

a b s t r a c t Heat reflux extraction (HRE) process with cyclohexanone (CYC) in a high-performance mass transfer extractor was applied to dissolve Shenmu-Fugu high-volatile bituminous (SFHB) coal for the first time to afford extract (E) and extract residue (ER) from the extraction. SFHB, E, and ER were characterized by elemental analysis, solid-state 13C NMR spectrometry, FTIR spectrometry, XRD, SEM, and TG-FTIR to elucidate the effect of HRE on the evolution of functional groups and macromolecular structure of coal during extraction. The soluble portion in SFHB was 24.37% in the course of HRE with CYC. The aromaticity of SFHB derived from both curve-fitting of 13C NMR and FTIR spectra was obviously increased after extraction suggesting that most of the aliphatic fractions were extracted during HRE process. It was clarified that the substituted degree of aromatic ring in SFHB became low but the substituents on aromatics were larger after extraction. Due to irreversibly swelling crystal structure of SFHB, its interlayer spacing became larger and the stacking height of crystallite decreased after extraction. Moreover, significant amounts of volatile matters were extracted, which caused relatively lower mass loss rate and contents of gaseous products (CO2, aliphatic moieties, CH4, and CO) of ER than SFHB during main pyrolysis stage. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Coal is a biological sedimentary rock created from the remains of plant debris that lived and died around 100–400 million years ago when parts of the earth were covered with huge swampy forests [1]. The nature of carbonaceous-enrichment macerals in coal in addition to inorganic minerals makes this abundant resource widely apply in electricity generation, steel production, cement manufacturing, and chemical engineering [2–4]. An accurate and deep understanding the organic structures of coals are crucial for

⇑ Corresponding author at: State Key Laboratory of Heavy Oil Processing, China University of Petroleum (East China), Qingdao 266580, China. E-mail address: [email protected] (Y.-y. Tian). http://dx.doi.org/10.1016/j.applthermaleng.2016.08.104 1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.

optimizing the above processes and efficient utilization of coals, especially for obtaining value-added chemicals from coals. Solvent extraction of coal under mild conditions is an essential approach for studying chemical composition, distribution of functional groups, and molecular skeleton structures of coals as well as their extracts and residues [5,6]. Solvent extraction methods mainly include low-temperature extraction, thermal dissolution, and ionic liquids extraction, among which low-temperature extraction is believed to be an effective approach for preliminarily separating organic portion from coals and is also the most commonly used method [7,8]. This extraction procedure is usually performed in a container reactor (Bunsen beaker and Erlenmeyer flask) or a Soxhlet extractor with single or mixed organic solvents such as alkanes, alkanols (e.g., methanol, ethanol, and isopropanol),

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568

561

Nomenclature A ad CYC d daf Di d002 DTG E EBF EOM ER FTIR FWHM HRE Lc M NMP PAHs

ash air dry base cyclohexanone dry base dry and ash-free base devolatilization index interlayer spacing of aromatic layers derivative mass loss extract extraction bag filter extractable organic matter extract residue fourier transform infrared full width at half maximum heat reflux extraction mean height of crystallite in c-direction moisture N-methyl-2-pyrrolidone polycyclic aromatic hydrocarbons

amines, benzene, chloroform, CS2, dichloromethane, ketones, N-methyl-2-pyrrolidone (NMP), N,N-dimethylacetamide, pyridine, quinolone, tetrahydrafuran (THF), tetralin, toluene, CS2/NMP mixture, and CS2/pyridine mixture [9]. Maceral fractions derived from six coals were extracted using CS2/NMP mixed solvent in stainless centrifuge tube by Dyrkacz and Bloomquist [10] and they found that vitrinites showed higher extract yields than both liptinites and inertinites. In another example, Xue et al. [11] detected the polycyclic aromatic hydrocarbons (PAHs) in extracts obtained from CS2 and CH2Cl2 extractions of coals both in Soxhlet and ultrasonicassisted extractors. They concluded that 24 h was suitable for extraction of PAHs from coals by Soxhlet extraction method and both Soxhlet and ultrasonic-assisted extraction methods showed similar PAHs concentration patterns. Three extraction solvents involving CS2, CH2Cl2, and supercritical carbon dioxide were utilized to dissolve three high volatile bituminous coals, Kolak and Burruss [12] found that Soxhlet-CH2Cl2 and Soxhlet-CS2 extractions yielded similar amounts of extractable organic matter (EOM) and distributions of individual hydrocarbons, whereas supercritical CO2 extractions (40 °C, 100 bar) yielded approximately an order of magnitude less EOM. In addition, a resent work by Zou et al. [13] has tried to use THF for dissolving the mobile phase of a lignite in a 1000 mL conical flask and the obtained 5.1% percentage of mobile phase was further pyrolyzed in a new in-situ pyrolysis-time of flight mass spectrometry to evaluate the initial products species. Although a high extraction performance can be achieved by low-temperature extraction of coal in the container and Soxhlet reactors with single or mixed solvents, there are defects in further enhancement of the extraction yields when the two reactors are involved. As we known, sample is often immersed into the solvent and forms reaction mixture together with extract and solvent during extraction process in the container reactors. This limits the release of soluble portions from the pores of inner coal matrix to the external solvent phase and results in a poor diffusion effect during extraction. For Soxhelt extraction process, sample is placed inside a glass microfiber thimble and further loaded into the main chamber of the Soxhlet extractor. The solvent vapor rises from the distillation flask to the condenser through a distillation arm and then droplets of the solvent drip down to the chamber as long as the temperature reaches to its boiling point. After several minutes of soaking, solvent dissolving the extract will be led into the distillation flask again under the action of siphon

Rmax maximum decomposition rate R1 mass loss rate for DTG peak 1 mass loss rate for DTG peak 2 R2 SEM scanning electron microscope SFHB shenmu-fugu high-volatile bituminous coal SS 13C NMR solid-state 13C nuclear magnetic resonance terminated decomposition temperature Tf TGA thermogravimetric analysis THF tetrahydrafuran Tin initial devolatilisation temperature maximum mass loss temperature Tmax T1 temperature for DTG peak 1 T2 temperature for DTG peak 2 V volatile matter XRD X-ray diffraction h diffraction angle

mechanism when the chamber is full enough. It is worthwhile to note that the solvent is not heated (normally slightly higher than room temperature) as it extracts the organic portion in the sample during Soxhelt extraction. This greatly limits the solvent extraction efficiency because the effect of temperature-induced mass transfer disappears under the ambient temperature [14]. However, heat reflux extraction (HRE) process using reflux extractor as reactor will not only enhance the contact efficiency but also promote the effect of temperature-induced mass transfer between solvent and coal on account of the fact that hot fresh solvent vapor generated from the solvent container directly passes through sample via a vapor hole and then flows down continuously from the condenser, which minimizes the secondary interaction between extract and coal sample [15]. In addition, cyclohexanone (CYC) is known as an important industrial solvent possessing characteristics of electron donor and condensed plane molecular structure, which was proved to have excellent performance during extraction such as, good compatibility with coals, high swelling power, and easily penetrating into the inner region of coal matrix [16]. Therefore, extraction of coal using CYC in a heat reflux extractor may further maximize the extraction yields compared with other lowtemperature extractor such as container reactor or Soxhlet extractor. Generally, the amounts of solvent insoluble portions in coals are higher than that of the soluble portions and the insoluble portions can more accurately reflect the structural characteristics of the macromolecular skeleton of coals. Furthermore, identical structures (e.g., functional groups and aromatic skeletons) are normally distributed in both extract and residue during extraction. Hence, a deep insight into changes of chemical structures between raw coal and its solvent insoluble portions is of significant importance for efficient use of coals as well as directly revealing of solvent extraction mechanism during extraction. However, so far, there are few related studies on the structural changes of coals during HRE process with CYC and some essential issues such as the extraction performance of the solvent and the evolution of functional groups and macromolecular structure have not yet been fully understood [17]. Recently, numerous efforts have been focused on the identification of composition and structure of coal and its extract and residue from low-temperature extraction by several direct characterization technologies including solid-state 13C nuclear magnetic resonance (SS 13C NMR) spectrometry [18–20] elemental

562

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568

analysis, Fourier transform infrared (FTIR) spectrometry [21], X-ray diffraction (XRD) [22,23] and scanning electron microscope (SEM) [24,25]. Thermogravimetric analysis (TGA) coupled with on-line FTIR spectrometer is a well-established pyrolysis method, which is usually employed to study the thermal degradation profiles (e.g., devolatilization index (Di), initial and terminated decomposition temperatures) and the time-dependent evolution of the gaseous products of solid fossil fuels, such as coal and biomass samples [26,27]. With this method, Ding et al. [28] investigated the pyrolysis parameters of extraction residues from the sequential extraction and thermal dissolution of a lignite. The release characteristics of O-containing gases (H2O, CO and CO2) from the O-containing functional groups in coal during rapid pyrolysis was also determined by TG-FTIR technique [29]. However, the relatively low extract yields of coals using common solvents in container and Soxhlet extractors has restricted comprehensive exploration of coal structures to a great extent. Extraction of coals with CYC under heat reflux state for improving dissolution performances of organic matters in coals via accelerating diffusion of both solvent and extract has not been perceived. Furthermore, detailed investigations on the structure features of the obtained soluble and insoluble fractions from HRE process as well as their corresponding parent coal were less explored and discussed. In this work, we have reported a novel extraction method of HRE with CYC for obtaining higher amounts of soluble organic matters from coal by extracting under boiling point of solvent. The effects of HRE process on structural and compositional features of Shenmu-Fugu high-volatile bituminous coal (SFHB) and the chemical properties of its resulting extract were evaluated by a wide array of analytical techniques. This work will provide a new idea for solvent extraction of coals for obtaining value-added chemicals and it can also shed light on some essential issues such as, the extraction performance of CYC, the structure changes of coal, and the evolution of functional groups and macromolecular structure during HRE process with CYC.

Fig. 1. The schematic representation of heat reflux extractor.

Extract yields ðwt% dafÞ ¼

½1  ðresidue ðgÞ=coal feed ðgÞÞ ½1  ðash ðwt% d:b:Þ=100Þ

ð1Þ

In order to ensure the accuracy and reproducibility of the extract yield, SFHB was extracted using three parallel samples simultaneously and the mean value was reported in this study. The relative standard deviation of extract yield was found be less than 2.31%.

2. Experiment

2.2. Analytical methods

2.1. SFHB and its HRE process

Proximate analysis of SFHB was measured following the Chinese National Standard GB/T 212-2008 and Ultimate analyses of SFHB, E, and ER were performed on an elemental analyzer (Vario Macro cube, Elementar Corp., Germany). The molecular structure of SFHB and ER was determined using a Bruker Avance III 400 SS 13 C NMR spectrometer at a carbon frequency of 100.63 MHz at room temperature. About 200 mg dry samples were packed into a 4 mm diameter zirconia rotor and spun at 14 kHz. A contact time of 2 ms and recycle delay time of 3 s were used in the crosspolarization (CP) experiments. A software PeakFit V4.12 (SeaSolve Software Inc.) was used for curve fitting of the 13C NMR spectra to quantify the relative proportion of different carbon types in SFHB and ER [30–32]. SFHB, ER, and E were analyzed on a FTIR spectrometer (Bruker Tensor 27, Bruker, Germany) by collecting 32 scans at a resolution of 4 cm1 in the wavenumber range of 4000–400 cm1 to record the FTIR spectra. Sample discs for FTIR analysis were prepared by diluting accurately 1 mg of samples in 100 mg of KBr under infrared heat lamp. Curve fitting of absorption bands from the FTIR spectra of the samples in the region of 3000–2700 cm1 and 1800–1000 cm1 were performed to quantify the changes of functional groups during extraction. The positions and number of bands were established from the second derivative of the spectrum and according to relevant literatures [21,33,34]. PeakFit software was used to conduct the curve fitting procedure, during which selected regions of the FTIR spectra were baseline-linearized by connecting the left and right points of the interval with a straight

SFHB was obtained from Shenmu-Fugu coal mine located in the north of Shaanxi province, China and was ground to pass through a 200 mesh sieve, followed by desiccation in vacuum oven at 80 °C for 12 h prior to use. SFHB was extracted in a heat reflux extractor with CYC under 156 °C. The heat reflux extractor was connected to a water-cooled condenser to drop the evaporated solvent back, as illustrated in Fig. 1. Coal sample (5 g) was added into a tapered filter cartridge to form extraction bag filter (EBF) and the EBF was further suspended in the heat reflux extractor. Then, 150 mL of CYC (analytical reagents) was added into the Erlenmeyer flasks and some zeolites were used to avoid bumping due to they can provide cores for forming bubbles. The HRE process was conducted at boiling point of CYC for three hours, during which exhausted extraction was achieved when the droplets of the reflux solvent became almost colorless. After cooling to room temperature, the extraction mixture was filtered through a microporous membrane (porosity = 0.8 lm) and the filtrate was evaporated under vacuum at 70 °C to remove most of the CYC. The wet extract obtained was washed with 1:4 (volume ratio) acetone-water solution to get rid of residual solvents. The residue was washed with deionized water, and then rinsed 3 times with acetone under ultrasonic to remove residual solvents. Afterwards, the extract (E) and extract residue (ER) were dried in vacuum at 80 °C for 12 h. The extract yield was determined from the mass of the residue and given by Eq. (1).

563

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568 Table 1 Proximate and ultimate analyses (wt%) for SFHB, E, and ER. Sample

SFHB ER E a

Proximate analysis

Ultimate analysis (daf) a

Mad

Ad

Vdaf

C

H

O

12.45

7.96

39.72

81.19 79.58 75.24

4.82 4.54 7.46

11.73 13.48 16.52

N

S

1.15 1.21 0.28

1.11 1.19 0.51

H/C

O/C

0.712 0.685 1.189

0.108 0.127 0.164

By difference; d: dry base; Mad: moisture (air dried base); Ad: ash (dry base, i.e., moisture-free base); Vdaf: volatile matter (dry and ash-free base).

Table 2 Chemical shift values and molar percentage of different carbon types in SFHB and ER.

Fig. 2. Solid-state 13C CP/MAS NMR spectra and their fitting curves of SFHB and ER.

line. All the band shapes, heights, and widths were allowed to adjust. The Gauss/Lorentz combination was selected for the band shape of fitting peaks. The X-ray Diffraction method was used to determine the crystal structure of SFHB and ER. The X’Pert Pro MPD Philips diffractometer using Cu Ka (1.542 Å) radiation at room temperature was applied. The interlayer spacing (d002) of aromatic layers in samples was calculated from the (0 0 2) peak using Bragg equation. The mean height of crystallite in c-direction (Lc) was determined using full-width at half peak height of (0 0 2) peak by the Scherrer formula as defined in Eq. (2) [35]:

Peak

Symbol

Carbon type

SFHB

ER

1

1 f al

14.4

12.84

4.43

2

f al

a

19.9

0.32

3.82

3

f al

2

25.3

0.90

4.41

4

f al

3

31.9

5.59

4.58

5

f al

4

38.0

3.71

1.55

6

5 f al

44.0

10.41

0.89

7

f al

O

79.8

3.18

0.78

8

fa

O1

108.0

4.07

2.25

9

fa

O2

119.4

10.29

8.90

10

fa

H

124.0

2.86

13.24

11

fa

B

128.7

10.29

36.33

12

fa

S

139.0

11.13

13.11

13

fa

O3

145.0

2.19

2.85

14

fa

O4

155.0

15.14

1.92

15

fa

CC1

176.2

1.98

0.15

16

fa

CC2

209.5

5.09

0.87

Table 3 Carbon structural parameters in SFHB and ER determined by SS Structural parameter

K k Lc ¼ b2h  cos h

Chemical shift (ppm)

Symbol

Definition

where Lc is the mean crystallite height, K is the shape factor with a constant value of 0.89 for Lc, k is the X-ray wavelength (1.5606 Å), b is the full width at half maximum intensity (FWHM) in radians, and h is the Bragg angle. The surface morphology of the SFHB and ER was examined using SEM (JSM-7001F, Japan Electron Optics Laboratory Ltd. Corp., Japan) at an acceleration voltage of 15 kV. Before observation, the samples were coated with gold in E-1010 ion sputter.

fa

f a ¼ fa

Ratio of aliphatic carbon Ratio of carbonyl carbon Mole percent of aromatic bridgehead carbon Substituted degree of aromatic ring Average methylene chain length

f al

f al ¼ f al

CC fa

CC fa

xb

r Cn

O1—O4

¼

H

B

þ f al þ f al

a

CC1 fa

CC2 fa

þ

O

.

xb ¼ f a f a 

S

þ fa þ fa þ fa

1—5

B

13

C NMR analysis. Value

ð2Þ Aromaticity

Mole percent (%)

S r ¼ f O1O4 þ fa a

. fa

 . 2 3 S C n ¼ f al þ f al fa

SFHB

ER

55.98

78.50

36.95

20.47

7.07

1.02

0.18

0.46

0.76

0.16

0.58

0.69

564

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568

samples were placed in a platinum crucible and temperature was ramped from room temperature to 1050 °C at a heating rate of 10 °C min1 in high purity N2 atmosphere with a flow rate of 100 mL min1. The transfer line between the TG and FTIR apparatuses was made by polytetrafluoroethylene with an internal diameter of 2 mm and length of 1.0 m. In order to ensure that all the emitted gas phase products could be fed into the infrared gas cell, the transfer line was heated and maintained at a constant temperature of 200 °C. The spectral region of the FTIR was 4000–500 cm1 with the resolution 4 cm1 and 16 scans per sampling. In this work, devolatilization index (Di) of the samples was calculated for quantifying the performance of volatile matters releasing during thermal degradation, as defined in Eq. (3) [36,37]:

Di ¼ Rmax =Tin Tmax DT1=2

Fig. 3. FTIR spectra of the SFHB and its extract and residue.

ð3Þ

where Rmax, Tin, and Tmax are the maximum decomposition rate, initial devolatilisation temperature, and maximum mass loss temperature, respectively. The three parameters can be obtained from the TG and DTG curves [38]. DT1/2 is temperature interval when Rd/Rmax is equal to 0.5. Rd is decomposition rate, which is defined as Eq. (4):

Rd ¼ dmt =dt

ð4Þ

where mt is mass of the raw sample at time t. Rmax and Rd can be obtained from the derivative thermogravimetric curves. 3. Results and discussion 3.1. Extract yield and elemental composition The extractable organic portion in SFHB was 24.37 wt% during HRE process with CYC. As listed in Table 1, the contents of C and H in ER were significantly lower than those in SFHB, but the contents of O, N, and S in ER were obvious higher than those in SFHB. The differences in the distribution of C, H, and O simultaneously led to a minor increase in O/C atomic ratio, and a major decrease in H/C atomic ratio of SFHB after HRE process. These results imply that the possible enrichment of heavy macromolecular aromatics, including N, S-containing aromatics in ER and that most of the light components should be extracted and concentrated in E, which is in accordance with the elemental composition of the E. 3.2. Structural features of SFHB and ER from SS

Fig. 4. Curve fitting analysis of 3000–2800 cm1 and 1800–900 cm1 regions, use spectrum of E as an example.

13

C NMR

The CP/MAS 13C NMR provides detailed information about the carbon types in SFHB and ER. The major characteristics of the spectra were the two relatively broad bands (Fig. 2), in which the more intense one was ascribable to aliphatic carbons (0–90 ppm) and

Table 4 FTIR derived structural parameters for SFHB, E, and ER. Sample

AaroCH/AaliCH

CH2/CH3

(CAO + C@O)/Car

SFHB ER E

3.565E2 6.274E2 7.834E3

3.389 4.373 1.785

2.068 1.491 3.080

2.3. Pyrolytic analysis The pyrolysis experiments of the SFHB, E, and ER were carried out on a TG analyzer (STA449F3, NETZSCH-Gerätebau GmbH, Germany) coupled with an online FTIR Spectrometer (Bruker Tensor 27, Bruker, Germany) to evaluate thermal stability and gaseous products species during thermal degradation. About 10 mg

Fig. 5. XRD patterns of the SFHB and its extract residue.

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568 Table 5 Structural parameters of SFHB and its extract residue from XRD. Sample

2h002 (°)

FWHM (°)

d002

Lc

SFHB ER

22.131 22.066

8.054 8.428

0.3689 0.3699

1.0548 1.0078

the broader one was attributed to aromatic carbons (90–220 ppm, including carbonyl/carboxyl and phenolic groups). The spectra can be further divided into 16 individual peaks via curve fitting (Fig. 2 and Table 2), which are ascribed to different types of aliphatic, aromatic, and carbonyl carbon species in SFHB and ER [22,32]. As listed in Table 3, the aromaticity (fa) of SFHB and ER were 55.98% and 78.50%, respectively, suggesting that ER contained much more aromatic carbon atoms than SFHB. In other words, most of the aliphatic species were extracted during HRE process with CYC. The molar content of aromatic bridgehead carbon (xb) is an important parameter because it can be used to estimate the aromatic cluster size [39]. The xb value in ER (0.46) was significantly higher than the one in SFHB (0.18), which showed the same trend with fa. This result meant that the aromatic cluster size of SFHB became larger after extraction. Moreover, there were 6 methylene carbons and about 13 alkyl branched aromatic carbons per 100 carbon atoms in SFHB, while the methylene carbons increased to 9 and the alkyl branched aromatic carbons reduced to 8 per 100 carbon atoms in ER. It is considered that methylene carbons exist as both a portion of straight chains and some saturated alicyclic structures [40]. Thereby, the results quoted above indicate that alkyl side chains attached to the aromatic clusters are relatively shorter in SFHB than those in ER and the aromatic species containing short alkyl side chains are easier to be dissolved than those having long alkyl side one during HRE with CYC. The reason may be that more nucleophilic reagent (CYC) can attack the carbon atoms on aromatic nuclei at a low steric hindrance of substituent groups, which leads to dissolve more portion of soluble fragments. The substituted degree (r) is determined by dividing the percentage of oxyaromatic groups (faO) plus alkyl-substituted aromatic (faS) by the percentage of aromatic carbons and usually used for reflecting the

565

metamorphic degree of organic matters in coal. According to the results listed in Table 3, the value of r was considerably high for SFHB and this value decreased from 0.76 to 0.16 in SFHB after extraction, meaning that the number of substituents on the aromatic nuclei in SFHB decreased after extraction and the ER may be more mature in metamorphism degree than its parent coal. 3.3. Structural features of SFHB, E, and ER from FTIR Significant differences in the FTIR spectra among SFHB and its extract and residue can be observed from Fig. 3. The absorptions of aliphatic moieties (AM) in SFHB around 2961, 2920, and 2856 cm1 were stronger than those in ER. Furthermore, the extract exhibited strongest absorptions in these bands, implying that AM-rich species in SFHB were easier to be extracted. The absorbance near 3418 cm1 was attributed to hydroxyl groups. Strong absorbance in this region was observed in the spectrum of extract compared with that of SFHB and ER, suggesting strong interactions between the CYC and bound -OH-containing species in SFHB. In addition, the characteristic absorptions of the carbonyl group from aldehydes, ketones and carboxylic acids, and CAO bonds from ethers, alcohols, and phenols were around 1708 cm1 and located in 1294–1048 cm1 region, respectively. These absorptions in the spectrum of E were very strong, but only very low absorbance and overlapped peaks can be found in SFHB and ER. This result may suggest that the CYC extract contains large numbers of the oxygen-containing species. Due to the influence of mineral matters in coal and overlapped peaks referring to C@C and C@O bonds in E, the absorbance peaks of aromatic hydrogens at 3056, 874, 812, and 752 cm1 were relatively obvious in extract, while that of the aromatic C@C at 1604 cm1 was quite apparent in SFHB and ER. In this work, three main structural parameters derived from the curve-fitting analysis of FTIR spectra were used to evaluate the changes of chemical characteristics in SFHB and its extract and residue, including (1) aromaticity (AaroCH/AaliCH), which was determined by calculating the integrated areas ratio of 3080–3000 cm1 and 3000–2780 cm1, respectively, (2) chain length (CH2/CH3),

Fig. 6. Surface morphology of the SFHB and its extract residue.

566

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568

matrix and irreversibly swell crystal structure of the coal [16]. Once the solvent molecules continuously penetrate into the coal aromatic layers, swelling effect makes aromatic lamellar spacing increase. However, the aromatic spacing cannot return to its original state after solvent molecules leave away from the coal matrix. This leads the increase in the interlayer spacing of aromatic layers in ER. Furthermore, owing to the concentration of mineral matters in coal during HRE process, the intensities of diffraction peaks for the mineral matters in SFHB became more prominent after extraction and the mineral matters identified in SFHB were mainly in the form of calcite and ankerite.

Fig. 7. TG and DTG curves of SFHB, E and ER.

which was determined by calculating the integrated areas ratio of peaks (2932 + 2856 cm1) and peaks (2964 + 2875 cm1), and (3) maturation level ((CAO + C@O)/Car), which was determined by calculating the integrated areas ratio of the non-hydroxyl oxygencontaining compounds (1780–1660 cm1 region and 1260– 1040 cm1 region) and aromatic C@C (1610 cm1 band). An example of curve fitting of 3000–2800 cm1 and 1800–900 cm1 bands for E was illustrated in Fig. 4. As listed in Table 4, the (CAO + C@O)/ Car value in SFHB decreased from 2.068 to 1.491, whereas both the CH2/CH3 and AaroCH/AaliCH ratios for SFHB increased after extraction. These results indicated that ER obtained from the HRE of SFHB with CYC contained more arenes but less oxygen-containing compounds, and the aliphatic chains attached to aromatic clusters were larger in ER than those in SFHB. The trends of these parameters were consistent with the structural parameters derived from the SS 13C NMR. 3.4. Structural features of SFHB and ER from XRD and SEM Structural parameters about the carbon clusters revealed by the XRD can help us indirectly understand the structure of aromatic carbons in SFHB and its extraction residue [41]. As shown in Fig. 5, the (0 0 2) peak around 22° comes from the diffraction of X-rays of the stacks of aromatic molecules, while the (1 0 0) peak at about 43° is attributed to the in-plane structure of aromatics [42]. However, the (1 0 0) peak, with a considerable weak intensity, was presented in the XRD patterns of both SFHB and ER, indicating that few aromatic carbons in SFHB and ER give contribution to orderly clustering and most of them are separated by various bridge bonds, such as ACH2A and AOA. The interlayer spacing of aromatic layers (d002) and mean height of crystallite in c-direction (Lc) are commonly used to evaluate the structure of aromatic carbon crystallite, which has been proved to be suitable for graphite-like materials, such as coal and its derivatives. Since the (1 0 0) peaks in XRD spectra of both SFHB and ER were not significant and also influenced by the minerals in coal, it is difficult to analyze the diameter of aromatic layer (La). Thus, only the d002 and the Lc of the carbon crystal were studied in this work. As listed in Table 5, the d002 value in ER was slightly higher than that in SFHB and the parameter of Lc in SFHB decreased from 1.0548 to 1.0078. For one respect, these results indicated that the crystallographic texture of SFHB was slightly destroyed during HRE with CYC. Meanwhile, these results revealed that the interlayer spacing of aromatic layers in SFHB became larger, whereas the mean height of crystallite decreased after extraction. This may be ascribed to the characteristic of condensed plane molecular structure of CYC, which could let it easily penetrate into the coal

Fig. 8. FTIR spectra for gaseous products emitted when SFHB, ER, and E were heated to some typical temperatures.

567

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568

Compared to SFHB, the particle sizes of the ER were substantially reduced, as demonstrated in Fig. 6, indicating that the HRE with CYC under the boiling point of solvent led to significant destruction of SFHB particles. In addition, due to effects of swelling, dissolution, and erosion by hot CYC, the surface morphology of SFHB became rough and porous after extraction.

Table 6 Pyrolysis parameters of the SFHB, E, and ER with heating rate of 10 °C min1.

3.5. Pyrolysis of SFHB, E, and ER As Fig. 7 displays, the variations of derivative mass losses (DTG) of SFHB and its extract and residue could be divided into several stages according to the degradation temperature. Prior to 350 °C, the mass loss was in the order of E > ER > SFHB at the same temperature, thus two obvious mass loss rate peaks were observed in ER around 251.5 °C and E at 304.9 °C. From the gaseous products produced during thermal degradation of SFHB, ER, and E (Fig. 8), the emitted gaseous species at 251 °C were mainly CO2 and aliphatic C-H moieties and the content of these species was in the order of E > ER > SFHB. In fact, the pyrolysis reaction of coal does not take place below 300 °C [43], so an obvious DTG peak in ER rather than SFHB prior to 350 °C may be ascribed to the release and evaporation of little amounts of extracts having been disaggregated from the coal matrix but are still retained in the micropores during thermal degradation. Compared with the first pyrolysis stage, the mass losses trend became opposite between ER and SFHB as the temperatures were in the range of 350–650 °C and sharp peaks were observed near 450 °C in DTG curves of all samples, which represented the main pyrolysis reaction stage of coal [44,45]. Further, as listed in Table 6, the absolute value of mass loss rate for SFHB was 1.27% min1 in this region and this value was much higher than that for ER (0.78% min1), suggesting significant amounts of volatile matters were extracted during HRE process with CYC. The gaseous products released during thermal degradation at about 450 °C were mainly composed of CO2, aliphatic moieties, CH4, and CO, which was related to the cleavage of alkyl side chains and decomposition of functional groups in coal such as methoxy and carbonyl groups [46,47]. As expected, the contents of these gaseous products released displayed the order of E > SFHB > ER during the main thermal degradation stage. After 650 °C, the mass losses trend of SFHB, ER, and E was the same as that the first stage again and the total mass losses of SFHB, ER, and E recorded at 1000 °C were 34.07, 36.16, and 82.56%, respectively. Furthermore, absorption peaks in IR spectra at about 735 °C were attributed to the condensation of aromatic clusters and a large quantity of CO and few CH4 were detected in the gaseous products of SFHB and E rather than ER. It is considered that the formation of CO and CH4 up to 700 °C originated from the decomposition of ether and hydroxyl groups with high bonding energies and cleavage of strongly bonded aryl-methyl groups in coal [48,49]. As listed in Table 3 from the Section 3.2, both the content of O-containing species and the r value decreased in coal after extraction, which can account for the absence of peaks for CO and CH4 in the infrared spectrum of ER. Since Di was closely related to Rmax and the Rmax in the temperature range of 350–650 °C was in the order of E > SFHB > ER, thus the values of Di also showed the same trend with Rmax and achieved 6.490, 18.158, and 2.951 for SFHB, E, and ER, respectively. These results suggested that E showed the highest releasing rate of volatile-matter during thermal degradation prior to 650 °C whereas ER exhibited the lowest one. 4. Conclusion The soluble organic matter in SFHB was 24.37% in the course of HRE with CYC and ER showed relatively lower concentrations of C

a

Parameters

SFHB

E

ER

T1 (°C) R1 (% min1) Tin (°C) Tmax (°C) Rmax (% min1) Tf (°C) T2 (°C) R2 (% min1) DT1/2(°C) Di (108% min1 °C3) Mass changea (%)

– – 400.7 452.2 1.27 579.5 737.7 0.55 108 6.490 34.07

– – 230.2 304.9 2.94 487.7 439.2 2.62 235.4 18.158 82.56

251.5 0.34 344.5 451.3 0.78 638.6 732.3 0.60 170 2.951 36.15

At end temperature of 1000 °C.

and H elements than the corresponding parent coal. The fa of SFHB derived from curve-fitting of SS 13C NMR increased from 55.98 to 78.50 after extraction suggesting that most of the aliphatic fractions were extracted during HRE process with CYC. Furthermore, the substituted degree of aromatic nuclei in SFHB reduced by nearly five times but the substituents on aromatics became larger after extraction. Due to irreversibly swell crystal structure of the coal and dissolution of the organic matters in coal, interlayer spacing of aromatic layers in SFHB became larger and the mean height of crystallite decreased from 1.0548 to 1.0078 after HRE process with CYC. TG-FTIR experiments showed that thermal degradation of SFHB, E, and ER mainly contained three stages, among which mass loss rate and contents of gaseous products (CO2, aliphatic moieties, CH4, and CO) for ER were much lower than those of SFHB during main thermal degradation stage (350–650 °C). Furthermore, characteristic parameter of Di presented the same trend with Rmax and achieved 6.490, 18.158, and 2.951 for SFHB, E, and ER, respectively during thermal degradation prior to 650 °C, which suggested that E had the highest releasing rate of volatile-matter and ER exhibited lower pyrolysis reactivity than its corresponding parent coal in the course of main thermal degradation stage. This work will provide an alternative approach for obtaining value-added chemicals from coals as well as give deep understanding of the evolutions of macromolecular skeletons and functional groups in coal during HRE process with CYC.

Acknowledgements The authors gratefully acknowledge the support from the National Natural Science Foundation of China (Grant Nos. 21576293 and 21576294).

References [1] X. Querol, X. Zhuang, O. Font, M. Izquierdo, A. Alastuey, I. Castro, B.L. van Drooge, T. Moreno, J.O. Grimalt, J. Elvira, M. Cabañas, R. Bartroli, J.C. Hower, C. Ayora, F. Plana, A. López-Soler, Influence of soil cover on reducing the environmental impact of spontaneous coal combustion in coal waste gobs: a review and new experimental data, Int. J. Coal Geol. 85 (2011) 2–22. [2] J. Blondeau, R. Kock, J. Mertens, A.J. Eley, L. Holub, Online monitoring of coal particle size and flow distribution in coal-fired power plants: dynamic effects of a varying mill classifier speed, Appl. Therm. Eng. 98 (2016) 449–454. [3] W. Duan, Q. Yu, J. Liu, L. Hou, H. Xie, K. Wang, Q. Qin, Characterizations of the hot blast furnace slag on coal gasification reaction, Appl. Therm. Eng. 98 (2016) 936–943. [4] W. Lin, Y. Feng, X. Zhang, Numerical study of volatiles production, fluid flow and heat transfer in coke ovens, Appl. Therm. Eng. 81 (2015) 353–358. [5] H.R. Schulten, A. Marzec, Time- and temperature-resolved field ionization mass spectrometry of coals and their extracts, Fuel 65 (1986) 855–860. [6] B. Chen, X.Y. Wei, Z.M. Zong, Z.S. Yang, Y. Qing, C. Liu, Difference in chemical composition of supercritical methanolysis products between two lignites, Appl. Energy 88 (2011) 4570–4576. [7] F.R. Mayo, N.A. Kirshen, Comparison of fractions of pyridine extract and solvent-refined coal from Illinois no. 6 coal, Fuel 57 (1978) 405–408.

568

B. Tian et al. / Applied Thermal Engineering 109 (2016) 560–568

[8] N.S. Vorob’yeva, Z.K. Zemskova, D. Bodzek, V.I. Kiselev, A.A. Petrov, Hydrocarbons of the soluble part of coal, Petrol. Chem. U.S.S.R. 23 (1983) 232–239. [9] F.J. Liu, X.Y. Wei, M. Fan, Z.M. Zong, Separation and structural characterization of the value-added chemicals from mild degradation of lignites: a review, Appl. Energy 170 (2016) 415–436. [10] G.R. Dyrkacz, C.A.A. Bloomquist, Solvent extraction of separated macerals in carbon disulfide/N-methylpyrrolidone, Energy Fuels 15 (2001) 1403–1408. [11] J. Xue, Liu, Niu, C.-L. Chou, Qi, Zheng, Zhang, Factors that influence the extraction of polycyclic aromatic hydrocarbons from coal, Energy Fuels 21 (2007) 881–890. [12] J.J. Kolak, R.C. Burruss, The use of solvent extractions and solubility theory to discern hydrocarbon associations in coal, with application to the coal– supercritical CO2 system, Org. Geochem. 73 (2014) 56–69. [13] L. Zou, L. Jin, X. Wang, H. Hu, Pyrolysis of Huolinhe lignite extract by in-situ pyrolysis-time of flight mass spectrometry, Fuel Process. Technol. 135 (2015) 52–59. [14] T. Yoshii, F. Yoshimura, Mechanism of solvent extraction from coal and its chars with chloroform, Fuel 50 (1971) 113–121. [15] K. Miura, M. Shimada, K. Mae, H.Y. Sock, Extraction of coal below 350 °C in flowing non-polar solvent, Fuel 80 (2001) 1573–1582. [16] Y.J. Shin, Y.H. Shen, Preparation of coal slurry with organic solvents, Chemosphere 68 (2007) 389–393. [17] C. Chen, J. Gao, Y. Yan, Observation of the type of hydrogen bonds in coal by FTIR, Energy Fuels 12 (1998) 446–449. [18] Z.K. Li, X.Y. Wei, H.L. Yan, Z.M. Zong, Insight into the structural features of Zhaotong lignite using multiple techniques, Fuel 153 (2015) 176–182. [19] L. Yan, Y. Bai, R. Zhao, F. Li, K. Xie, Correlation between coal structure and release of the two organic compounds during pyrolysis, Fuel 145 (2015) 12–17. [20] Y.G. Wang, X.Y. Wei, S.K. Wang, Z.K. Li, P. Li, F.J. Liu, Z.M. Zong, Structural evaluation of Xiaolongtan lignite by direct characterization and pyrolytic analysis, Fuel Process. Technol. 144 (2016) 248–254. [21] K.H. Michaelian, W.I. Friesen, Photoacoustic FT-i.r. spectra of separated western Canadian coal macerals: analysis of the CH stretching region by curve-fitting and deconvolution, Fuel 69 (1990) 1271–1275. [22] Y.G. Wang, X.Y. Wei, R.L. Xie, F.J. Liu, P. Li, Z.M. Zong, Structural characterization of typical organic species in Jincheng no. 15 anthracite, Energy Fuels 29 (2015) 595–601. [23] J.M. Lee, Y.J. Kim, S.D. Kim, Catalytic coal gasification in an internally circulating fluidized bed reactor with draft tube, Appl. Therm. Eng. 18 (1998) 1013–1024. [24] D.L. Shi, X.Y. Wei, X. Fan, Z.M. Zong, B. Chen, Y.P. Zhao, Y.G. Wang, J.P. Cao, Characterizations of the extracts from getting bituminous coal by spectrometries, Energy Fuels 27 (2013) 3709–3717. [25] S. Datta, P. Sarkar, P.D. Chavan, S. Saha, G. Sahu, A.K. Sinha, V.K. Saxena, Agglomeration behaviour of high ash Indian coals in fluidized bed gasification pilot plant, Appl. Therm. Eng. 86 (2015) 222–228. [26] L. Ge, Y. Zhang, C. Xu, Z. Wang, J. Zhou, K. Cen, Influence of the hydrothermal dewatering on the combustion characteristics of Chinese low-rank coals, Appl. Therm. Eng. 90 (2015) 174–181. [27] J.S. Bae, D.W. Lee, Y.J. Lee, S.J. Park, J.H. Park, J.G. Kim, C. Han, Y.C. Choi, An investigation of the evaporation behavior of bioliquid in the pores and its application to hybrid coal combining biomass with coal, Appl. Therm. Eng. 90 (2015) 199–206. [28] M. Ding, Y.P. Zhao, Y.Q. Dou, X.Y. Wei, X. Fan, J.P. Cao, Y.L. Wang, Z.M. Zong, Sequential extraction and thermal dissolution of Shengli lignite, Fuel Process. Technol. 135 (2015) 20–24. [29] J.A. MacPhee, J.P. Charland, L. Giroux, Application of TG–FTIR to the determination of organic oxygen and its speciation in the Argonne premium coal samples, Fuel Process. Technol. 87 (2006) 335–341.

[30] M.J. Trewhella, I.J.F. Poplett, A. Grint, Structure of Green River oil shale kerogen: determination using solid state 13C NMR spectroscopy, Fuel 65 (1986) 541–546. [31] Z. Ẑujovic´, R. Srejic´, D. Vucˇelic´, D. Vitorovic´, B. Jovancˇic´evic´, Structural analysis of Aleksinac oil shale kerogen by high-resolution solid-state 13C NMR spectroscopy, Fuel 74 (1995) 1903–1909. [32] P.G. Hatcher, Chemical structural models for coalified wood (vitrinite) in low rank coal, Org. Geochem. 16 (1990) 959–968. [33] J. Ibarra, E. Muñoz, R. Moliner, FTIR study of the evolution of coal structure during the coalification process, Org. Geochem. 24 (1996) 725–735. [34] V. Zubkova, M. Czaplicka, Changes in the structure of plasticized coals caused by extraction with dichloromethane, Fuel 96 (2012) 298–305. [35] M.S. Nyathi, C. Burgess Clifford, H.H. Schobert, Effect of petroleum feedstock and reaction conditions on the structure of coal-petroleum co-cokes and heattreated products, Energy Fuels 26 (2012) 4413–4419. [36] Z. Wu, S. Wang, J. Zhao, L. Chen, H. Meng, Synergistic effect on thermal behavior during co-pyrolysis of lignocellulosic biomass model components blend with bituminous coal, Bioresour. Technol. 169 (2014) 220–228. [37] Z. Wu, S. Wang, J. Zhao, L. Chen, H. Meng, Thermal behavior and char structure evolution of bituminous coal blends with edible fungi residue during Copyrolysis, Energy Fuels 28 (2014) 1792–1801. [38] C.A. Wang, Y. Liu, X. Zhang, D. Che, A study on coal properties and combustion characteristics of blended coals in northwestern China, Energy Fuels 25 (2011) 3634–3645. [39] M.S. Solum, R.J. Pugmire, D.M. Grant, Carbon-13 solid-state NMR of Argonnepremium coals, Energy Fuels 3 (1989) 187–193. [40] A. Tahmasebi, J. Yu, S. Bhattacharya, Chemical structure changes accompanying fluidized-bed drying of Victorian brown coals in superheated steam, nitrogen, and hot air, Energy Fuels 27 (2013) 154–166. [41] H. Takagi, K. Maruyama, N. Yoshizawa, Y. Yamada, Y. Sato, XRD analysis of carbon stacking structure in coal during heat treatment, Fuel 83 (2004) 2427– 2433. [42] I. Watanabe, K. Sakanishi, I. Mochida, Changes in coal aggregate structure by heat treatment and their coal rank dependency, Energy Fuels 16 (2002) 18–22. [43] P. Sarkar, S.G. Sahu, A. Mukherjee, M. Kumar, A.K. Adak, N. Chakraborty, S. Biswas, Co-combustion studies for potential application of sawdust or its low temperature char as co-fuel with coal, Appl. Therm. Eng. 63 (2014) 616–623. [44] X. Lin, C. Wang, K. Ideta, J. Miyawaki, Y. Nishiyama, Y. Wang, S. Yoon, I. Mochida, Insights into the functional group transformation of a Chinese brown coal during slow pyrolysis by combining various experiments, Fuel 118 (2014) 257–264. [45] K. Jayaraman, I. Gokalp, Gasification characteristics of petcoke and coal blended petcoke using thermogravimetry and mass spectrometry analysis, Appl. Therm. Eng. 80 (2015) 10–19. [46] G. Domazetis, M. Raoarun, B.D. James, Low-temperature pyrolysis of brown coal and brown coal containing iron hydroxyl complexes, Energy Fuels 20 (2006) 1997–2007. [47] X. Li, B.M. Krooss, P. Weniger, R. Littke, Liberation of molecular hydrogen (H2) and methane (CH4) during non-isothermal pyrolysis of shales and coals: systematics and quantification, Int. J. Coal Geol. 137 (2015) 152–164. [48] J. Liu, X. Jiang, J. Shen, H. Zhang, Pyrolysis of superfine pulverized coal. Part 1. Mechanisms of methane formation, Energy Convers. Manage. 87 (2014) 1027– 1038. [49] J. Liu, X. Jiang, J. Shen, H. Zhang, Pyrolysis of superfine pulverized coal. Part 2. Mechanisms of carbon monoxide formation, Energy Convers. Manage. 87 (2014) 1039–1049.