Interactions between Pyridine and Coal at the Molecular Level

Energy & Fuels 2002, 16, 497-509

497

Interactions between Pyridine and Coal at the Molecular Level: Insights from Variable-Temperature 1H NMR Studies of Pyridine-Saturated Coal Jincheng Xiong† and Gary E. Maciel* Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523 Received October 11, 2001. Revised Manuscript Received November 9, 2001

The first systematic, in-situ, variable-temperature 1H NMR study, based on the CRAMPS (combined rotation and multiple-pulse spectroscopy) technique, is reported on pyridine-saturated Argonne premium coals. This work focused on how noncovalent associative interactions are affected at the molecular level by thermal treatment and solvent saturation. With improved proton dipolar-dephasing experiments based on CRAMPS detection, we were able to monitor quantitatively changes of molecular mobility of various molecular structural moieties upon pyridine saturation and temperature variation. It was found that pyridine saturation can dramatically promote fast, random molecular motion with correlation times in the range of 1 µs to 30 ns, even at room temperature, while thermal treatment alone is much less effective in promoting molecular motion even at 230 °C. For pyridine-saturated coal, the molecular mobility of coal can be enhanced further with just a moderate increase in temperature. Correlations between molecular structure and molecular mobility have been established through comparative variable-temperature 1H CRAMPS studies of coals of different ranks. A mechanism for explaining the enhancement of molecular mobility induced by thermal treatment and solvent saturation is proposed and discussed in terms of these experimental results. Dramatic expansion of void space in coal as a result of disruption of noncovalent associative interactions by pyridine is the key feature of our understanding of these new NMR results, as well as previous work by others. The nature of the molecular phase and macromolecular phase in coal and the relationships between the molecular/ macromolecular (M/M) structural model and molecular mobility, as detected experimentally by NMR, have been shown to be much more complicated than previously thought. There is no simple one-to-one correspondence between molecular mobility and M/M phases in coal, but we have developed a coherent view of the molecular structure and dynamics of coal on the basis of the new experimental results. Many discrepancies in the literature are well resolved by this view, and the resolution enhancement due to solvent saturation is satisfactorily explained.

Introduction Solvent extraction as a means for studying coal compositions dates from early work in the 1850s,1,2 and has become of a major research topic since Bedson discovered that bituminous coals were substantially soluble in hot pyridine in 1902.3 A variety of technical processes for coal conversion, including liquefaction of coal, have also been based on interactions between solvents and coals.4 The understanding of such interactions at the molecular level is still very limited, despite great efforts that have been directed to this area of research.4 Both liquid-state and solid-state NMR techniques have played a very important role in studying coal * Author to whom correspondence should be addressed. Tel: (970) 491-6480. Fax: (970) 491-1801. E-mail: [email protected]. † Current Address: Dow Chemical Technical Center, 3200 Kanawha Turnpike, South Charleston, WV 25303. (1) Schrotter, A. Sitzungsber. Kaiser. Akad. Wissensch. Wien 1849, 3, 240. (2) de Marsilly, C. Ann. Chim. Phys. Ser. 3 1862, 66, 167. (3) Bedson, P. P. J. Soc. Chem. Ind. 1902, 21, 241. (4) Berkowitz, N. In An Introduction to Coal Technology, 2nd ed.; Academic Press: San Diego, 1994.

structures by solvent swelling and extraction.5,6 As coal is partially soluble in various organic solvents, some information about coal structure has been obtained from liquid-state 1H and 13C NMR experiments, but there are two problems with such studies. First, it is uncertain how well the structural information derived from the liquid extract of a coal corresponds to the actual structure of the whole, solid coal. Second, the structure of the insoluble residue of coal extraction cannot be determined from liquid-state NMR techniques. The development of high-resolution solid-state 1H and 13C NMR techniques from the late 1970s to the early 1980s has made it possible to extract structural information directly from untreated, as well as treated, solid coal samples.5,6 Both 13C CP/MAS and 1H CRAMPS techniques have been used to study untreated coals and solvent-extraction products of coal.5-15 (5) Mriler, W.; Meusinger, R. In Annual Reports on NMR Spectroscopy; Webb, G. A., Ed.; Academic Press: London, 1991; Vol. 23, p 375. (6) Mriler, W.; Meusinger, R. In Annual Reports on NMR Spectroscopy; Webb, G. A., Ed.; Academic Press: London, 1992; Vol. 24, p 331. (7) Bartuska, V. J.; Maciel, G. E.; Miknis, F. P. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1978, 23, 19.

10.1021/ef010248p CCC: $22.00 © 2002 American Chemical Society Published on Web 01/08/2002

498

Energy & Fuels, Vol. 16, No. 2, 2002

Previous studies in this group found that pyridine saturation can dramatically increase the apparent spectral resolution of 1H CRAMPS spectra of some types of coals at room temperature,10-13 although the exact mechanism of the resolution enhancement was not fully determined. Dipolar-dephasing experiments based on 1H CRAMPS detection were used to confirm the fact that molecular mobility is enhanced significantly in pyridine-saturated coals.10-12 The mobilized structure in pyridine-saturated coal was assigned as the molecular phase, which is the relatively mobile structure determined from experiments of the dipolar-dephasing type on either untreated coals or pyridine-saturated coals.10-17 However, significant differences in chemical structure (e.g., aromaticity) between the mobile component of untreated coal and solvent-saturated coal were reported in both 13C CP/MAS experiments and 1H CRAMPS experiments.9,10,18-20 The mobile component in untreated coal contains more aromatic hydrogen than does the immobile component, and the situation is just reversed in pyridine-saturated coal.9,18-20 A 1H CRAMPS study of pyridine-saturated Polish coal shows similar aromaticity in both mobile and rigid phases.10 Apparent discrepancies among different experiments (or for different coals) have existed for years. A reconciliation of such discrepancies would ideally be an important consequence of a better understanding of solvent-coal interactions at the molecular level. In previous papers,21,22 we presented results of systematic 1H CRAMPS studies of untreated coals at temperatures between 25 °C and 230 °C. In contrast to the great resolution enhancement observed for pyridinesaturated coal at room temperature, we did not observe such resolution enhancement at temperatures up to 230 °C, although we confirmed that there is a significant increase of molecular mobility at high temperatures via various CRAMPS-based time-domain experiments. Our results suggest that thermally induced molecular motion is very different from the motion enhanced by solvent saturation. Variable-temperature time-domain 1H CRAMPS experiments on pyridine-saturated coal can unveil the nature of the molecular motion that is (8) Gerstein, B. C.; Chou, C.; Pembleton, R. G.; Wilson, R. C. J. Phys. Chem. 1977, 66, 361. (9) Davis, M. F.; Quinting, G. R.; Bronnimann, C. E.; Maciel, G. E. Fuel 1989, 68, 763. (10) Jurkiewicz, A.; Bronnimann, C. E.; Maciel, G. E. Fuel 1989, 68, 872. (11) Jurkiewicz, A.; Bronnimann, C. E.; Maciel, G. E. Fuel 1990, 69, 804. (12) Jurkiewicz, A.; Bronnimann, C. E.; Maciel, G. E. In Magnetic Resonance of Carbonaceous Solids; Botto, R. E., Sanada, Y., Eds.; Advances in Chemistry 229; American Chemical Society: Washington, DC, 1993; p 401. (13) Jurkiewicz, A.; Bronnimann, C. E.; Maciel, G. E. Fuel 1994, 73, 823. (14) Jurkiewicz, A.; Marzec, A.; Idziak, S. Fuel 1981, 60, 1167. (15) Stock, L. M.; Muntean, J. V. Energy Fuels 1993, 7, 704. (16) Given, P. H.; Marzec, A.; Barton, W. A.; Lynch, L. J.; Gerstein, B. C. Fuel 1986, 65, 155. (17) Derbyshire, F.; Marzec, A.; Schulten, H.-R.; Wilson, M. A.; Davis, A.; Tekely, P.; Delpuech, J.-J.; Jurkiewicz, A.; Bronnimann, C. E.; Wind, R. A.; Maciel, G. E.; Narayan, R.; Bartle, K.; Snape, C. Fuel 1989, 68, 1091. (18) Tekely, P.; Nicole, D.; Brondeau, J.; Delpuech, J. J. J. Phys. Chem. 1986, 90, 5608. (19) Tekely, P.; Nocole, D.; Delpuech, J. J.; Julien, L.; Bertha, C. Energy Fuels 1987, 1, 1. (20) Kamienski, B.; Pruski, M.; Gerstein, B. C. Energy Fuels 1988, 1, 45. (21) Xiong, J.; Maciel, G. E. Energy Fuels 1997, 11, 856. (22) Xiong, J.; Maciel, G. E. Energy Fuels 1997, 11, 866.

Xiong and Maciel

enhanced by pyridine saturation. The present study not only enhances our understanding of solvent-coal interactions, but also provides insight into the mechanism of 1H CRAMPS spectral resolution enhancement induced by solvent swelling. In this paper, we present experimental results of a variable-temperature (VT) 1H CRAMPS study of pyridine-saturated coals. Improved dipolar-dephasing experiments, which were described for untreated coal samples in a previous paper,21 were applied to the corresponding pyridine-saturated coals. These results provide new insights into the molecular dynamical and structural changes induced in coal by thermal treatment and solvent saturation. The new results are discussed in terms of the molecular/macromolecular (M/M) structural model of coal,16 leading to a coherent view of the molecular structure and dynamics of coal. As a result, discrepancies in the literature on these issues are explained. Experimental Section 1. Sample Preparation. The premium coal bank of the Argonne National Laboratory provided the untreated original coal samples. Most of the work discussed in this paper concentrates on the Blind Canyon coal (Argonne premium coal 601), which is a high-volatile bituminous (HVB) coal and is substantially soluble in hot pyridine. Comparative studies on the low-volatile bituminous (LVB) coal, Pocahontas No. 3 (premium coal 501), which is not very soluble in pyridine, are also presented. The original coal samples were evacuated to about 10-3 Torr at room temperature for 20 h to reduce the residual water content. The coal samples were then saturated with perdeuterated pyridine-d5 (Cambridge Isotope Laboratory, 99.94% 2H) with a weight ratio of 1:2 (one part coal:2 parts pyridine). Each sample was sealed in a homemade glass MAS rotor with epoxy resin. Details of the homemade MAS rotor are described in the previous paper.21 The sealed glass spinner with the pyridine-saturated coal sample was put in an oven at 80 °C for one week to allow the establishment of equilibrium between pyridine and coal. 2. NMR Experiments. 1H CRAMPS experiments were performed on a severely modified Nicolet NT-200 NMR spectrometer, operating at a proton Larmor frequency of 187 MHz.23 Modification and redesign on both hardware and software have been made in this work to enhance the performance and stability of the spectrometer.24 Details of the experimental setup were described elsewhere.24 A Chemagnetics VT CRAMPS probe was used in this work. The experimental setup and conditions of 1H CRAMPS experiments were the same as those described in previous papers concerned with untreated coal samples.21,22 As the normal boiling point of pyridine is 115 °C, the highest temperature used in these experiments was 90 °C. The 1H CRAMPS dipolar-dephasing pulse sequence employed in this study was the improved version in which appropriate composite pulse(s) with proper phase cycling are inserted prior to the multiple-pulse trains in order to minimize baseline distortions caused by magnetization that is spinlocked to the effective field of the average Hamiltonian of the multiple-pulse sequence employed (BR-24).21,24 Without these improvements, reliably quantitative analysis of time-domain experimental data would be very difficult. (23) Maciel, G. E.; Bronnimann, C. E.; Hawkins, B. L. In Advances in Magnetic Resonance: The Waugh Symposium; Warren, W. S., Ed.; Academic: San Diego, CA, 1990; Vol. 14, p 125. (24) Xiong, J. Ph.D. Dissertation, Colorado State University, Fort Collins, CO, 1996.

Interactions between Pyridine and Coal

Energy & Fuels, Vol. 16, No. 2, 2002 499

Figure 1. 1H CRAMPS spectra of C5D5N-saturated premium coal 601 obtained at (b) 25 °C, (c) 60 °C, and (d) 90 °C. For comparison, the 1H CRAMPS spectrum of the untreated coal at 25 °C is shown in (a). The BR-24 pulse sequence was used, with a cycle time of 108 µs and a 90° pulse width of 1.25 µs. Each spectrum was acquired with 200 scans and a 3 s recycle delay. The MAS speed was between 1.5 and 1.8 kHz.

Figure 2. 1H CRAMPS spectra of C5D5N-saturated premium coal 501 obtained at (b) 25 °C, (c) 60 °C, and (d) 90 °C. For comparison, the 1H CRAMPS spectrum of the original coal at 25 °C is shown in (a). The BR-24 pulse sequence was used, with a cycle time of 108 µs and 90° pulse width of 1.2-1.3 µs. Each spectrum was acquired with 500 scans and a 3 s recycle delay. The MAS speed was between 1.5 and 1.8 kHz.

3. Data Analysis. A computer program for automatic peak deconvolution was developed in this work to ensure reliable and efficient quantitation based on spectral deconvolution. Another computer program was also developed for fitting different models that describe time-evolution of NMR signals in time-domain experiments. The details of these programs were presented elsewhere.24 The accuracy of peak deconvolution for a relatively broad peak can deteriorate significantly with slight baseline distortions or misphasing of a 1H CRAMPS spectrum. This problem is especially serious for the spectra with relatively low signalto-noise ratios (S/N), such as a 1H CRAMPS spectrum obtained after a long period of dipolar dephasing. To ensure reliable deconvolution of a series of 1H CRAMPS spectra acquired in a time-domain experiment, the spectra with good S/N were deconvoluted first, with all the peak parameters optimized by the computer deconvolution program. Some of the optimized parameters, such as peak shape, full width at half-height (fwhh) and peak position, were then averaged and used as fixed values for peak deconvolution of the spectra with low S/N. This procedure appeared to work very well for quantitative analysis of time-domain data. Standard deviations for the percentages of areas of the various chemical functionalities range from 2% to 5% (of the total spectral area) from the deconvolution procedure, depending on the sample and the specific peak area.

(premium coal 501) is also substantial (Figure 2). These observations are in agreement with the work of Jurkiewicz et al.12 When the temperature is increased from 25 °C to 90 °C, there are no dramatic changes in the 1H CRAMPS spectra of C5D5N-saturated coals. For C5D5N-saturated premium coal 601, the 1H CRAMPS spectrum obtained at 60 °C has better spectral resolution than the spectra obtained at 25 °C and 90 °C. For C5D5N-saturated premium coal 501, the fine structure in the 1H CRAMPS spectra is seen most distinctly at 25 °C. These small spectral resolution changes that occur with temperature variation are probably due to small changes in the spectral densities of random motion with correlation times around the cycle-time of the multiple-pulse sequence (10-100 µs), which may interfere with coherent averaging of dipolar interactions by the multiple-pulse sequence.21,25,26 2. Variable Temperature Proton Dipolar-Dephasing Experiments with CRAMPS Detection. In contrast to the large resolution enhancement of 1H CRAMPS spectra of coal brought about by pyridine-saturation at room temperature, we did not observe in the previous work21,22 such a large enhancement with untreated coal (no pyridine) with thermal activation up to 230 °C. This difference suggests that thermal treatment and solvent saturation have quite different effects on the molecular dynamics of coals. 1H-1H dipolar-dephasing experiments were thus performed on pyridine-saturated coals in this study to explore the molecular dynamical and structural changes induced by pyridine saturation. Although dipolar-dephasing experiments based on 1H CRAMPS detection were used in previous studies on

Results 1. Variable Temperature 1H CRAMPS Experiments. Figures 1 and 2 show 1H CRAMPS spectra of pyridine-saturated premium coals 601 and 501 obtained at three temperatures, along with the 1H CRAMPS spectra of the untreated coals obtained at 25 °C. Compared with 1H CRAMPS spectra of the untreated coal, large resolution enhancement can be seen in the 1H CRAMPS spectra of the C D N-saturated Blind 5 5 Canyon coal (premium coal 601) (Figure 1). The spectral change for C5D5N-saturated Pocahontas No. 3 coal

(25) Burum, D. P. Concepts Magn. Reson. 1990, 2, 213. (26) Kinney, D. R.; Chuang, I.-S.; Maciel, G. E. J. Am. Chem. Soc. 1993, 115, 6785.

500


Xiong and Maciel

Figure 3. Stack plots of 1H VT CRAMPS spectra of C5D5Nsaturated premium coal 601 obtained at (a) 25 °C and (b) 90 °C. The BR-24 pulse sequence was used, with a cycle time of 108 µs and a 90° pulse width of 1.2-1.3 µs. Each spectrum was acquired with 200 scans and a 3 s recycle delay. The MAS speed was between 1.5 and 1.8 kHz.

Figure 4. Stack plots of 1H VT CRAMPS spectra of C5D5Nsaturated premium coal 501 obtained at (a) 25 °C and (b) 90 °C. The BR-24 pulse sequence was used, with a cycle time of 108 µs and a 90° pulse width of 1.2-1.3 µs. Each spectrum was acquired with 500 scans and a 3 s recycle delay. The MAS speed was between 1.5 and 1.8 kHz.

coal in this group,10-12 those earlier studies were all carried out at room temperature. As demonstrated in this work, in-situ variable-temperature CRAMPS experiments are necessary for correlating molecular structure and mobility in coal. Figure 3 illustrates stack plots of 1H CRAMPS spectra of C5D5N-saturated premium coal 601 obtained in 1H1H dipolar-dephasing experiments at temperatures of 25 °C and 90 °C. One can see that, for both temperatures, a significant fraction of the 1H NMR signal still remains even with a dipolar-dephasing time as long as 6 ms. For comparison, the 1H NMR signals of the original Blind Canyon coal (premium coal 601) at 180 °C almost completely disappeared with a dipolardephasing time of 185 µs.21 It is therefore clear that pyridine saturation can promote molecular motion more dramatically even at room temperature than thermal treatment alone, even at 180 °C. As shown in Figure 3, the fraction of proton signals that survive a long dipolar-dephasing period at 90 °C is clearly larger than that at 25 °C. This result shows that raising the temperature from 25 to 90 °C enhances molecular mobility significantly, although the 1H CRAMPS spectra at 90 °C are not as well resolved as at 25 °C. This again demonstrates the point that higher mobility does not necessarily lead to higher spectral resolution, because of interference between random molecular motion and the coherent averaging of CRAMPS.21 High rank coal Pocahontas No. 3 (premium coal 501) has much lower pyridine extractability than Blind Canyon premium coal 601. It is therefore interesting to compare the effects of pyridine saturation on molec-

ular mobility of premium coal 501 with that of premium coal 601. Figure 4 presents the stack plots of 1H CRAMPS spectra of C5D5N-saturated premium coal 501 obtained in dipolar-dephasing experiments at 25 °C and 90 °C. One can see that a small, but significant, fraction of the 1H NMR signal persists up to 6 ms, after a fast initial decay of the signal. In dipolar-dephasing experiments on the original (untreated) premium coal 501, a detectable 1H NMR signal was not observed with a dephasing time of 180 µs, even at 180 °C.24 This proves, as was seen above for coal 601, that a small but significant fraction of coal protons is effectively mobilized by pyridine saturation in coal 501 even at room temperature. However, the proton fraction in pyridine-mobilizable components of premium coal 501 is smaller than that in premium coal 601. This is qualitatively in agreement with the fact that pyridine extractability of premium coal 501 is much lower than that of premium coal 601. Discussion 1. Chemical Functionality in Premium Coals 601 and 501. Determining chemical structure and functionality is one of the most difficult challenges in coal science. For bituminous coals, 13C CP/MAS and 1H CRAMPS techniques can usually resolve only two broad peaks, corresponding to aliphatic and aromatic structures in coal.7,8 The enhanced spectral resolution of coal by pyridine-saturation makes it possible to elucidate additional structural details. Figure 5 shows the experimental 1H CRAMPS spectrum of pyridine-saturated premium coal 601 and a



Figure 5. (a) 1H CRAMPS spectrum of C5D5N-saturated premium coal 601 obtained at 25 °C. (b) Spectrum simulated by computer peak deconvolution. In the simulated spectrum, the dashed lines represent individual peaks from the deconvolution, and the solid line represents the sum of all the deconvoluted peaks. Table 1. The Relative Concentrations (%) of the Various Types of Protons Observed in the 1H CRAMPS Spectra of C5D5N-Saturated Premium Coal 601 (Blind Canyon) chemical shift (ppm)a percentage (%) fwhh (ppm) a

1.0 6.5 0.8

1.7 35 2.5

3.4 20 3.4

6.9 19 1.9

8.4 15 2.3

11.5 3.8 5.9

Relative to tetramethylsilane (TMS).

spectrum simulated by computer deconvolution. The following peaks and shoulders can be identified tentatively via deconvolution: 1.0 ( 0.1 ppm, 1.7 ( 0.2 ppm, 3.4 ( 0.9 ppm, 6.9 ( 0.3 ppm, and 8.4 ( 0.6 ppm. The standard error of the position and area of the 3.4 ppm peak are larger than those of other peaks, because there are severe overlaps of aliphatic and aromatic peaks in this region, and only a shoulder is seen at about 3.4 ppm in the experimental 1H CRAMPS spectrum. To get a good fit in the spectrum simulated by peak deconvolution, a broad weak band centered at 11.5 ( 1.5 ppm must be included. The position and area of this broad band cannot be determined as accurately as those of other peaks due to inherent uncertainties about the nature of the baseline of typical 1H CRAMPS spectra.9 The fractions of each of the six types of protons are listed in Table 1. The 1.0 ( 0.1 ppm chemical shift is characteristic of CH3 groups on aliphatic chains or attached at a position γ or further away from an aromatic ring.27 This assignment is also confirmed by the dipolar-dephasing experiments, which are discussed later in this paper. The largest fraction of protons (35%) has a chemical shift of 1.7 ( 0.2 ppm, which can be attributed to two possible structures: alicyclic CH2, and CH2 in a position β to an aromatic ring or in tetalin-like structures.13 As the full width at half-height (fwhh) of this peak is 2.5 ppm, we cannot exclude the existence of a small fraction of other chemical functionalities with chemical shifts different from, yet still close to, 1.7 ppm. For example, a methyl group that is attached to an aromatic ring in the β position will have a chemical shift around 1.4 ppm; methines other than those R to an aromatic ring will (27) Bartle, K. D.; Jones, D. W. In Analytical Methods for Coal and Coal Products; Karr, C., Jr., Ed.; Academic Press: London, 1979; p 104.

have chemical shifts from 1.6 to 2.0 ppm.13,27 These functional groups have been identified previously in solvent extracts of coal.27,28 The aliphatic peak at 3.4 ( 0.9 ppm contains about 19% of the protons in the C5D5N-saturated coal 601. The chemical shift of this peak matches well with methylene protons of aliphatic ethers (R-CH2-O-). The oxygen content of premium coal 601 is about 11% by weight. The molar fraction of oxygen is only 5% based on the chemical analysis. According to various molecular models of coal structure, oxygen exists in coal mainly as hydroxyl groups and carbonyl groups. So, it is unlikely that 19% of the protons are all in aliphatic ether structures. In addition, the peak at 3.4 ppm is broad, with a fwhh of 3.4 ppm, and this peak position is not determined as accurately as those of other aliphatic peaks. A variety of chemical structures with chemical shifts between 2.0 and 5.0 ppm could contribute to this peak. Such structures include CH2 groups of acenaphthenes and indenes (3.0-3.3 ppm), CH groups R to an aromatic ring (2.0-3.3 ppm), and CH2 groups R to two aromatic rings (3.3-4.7 ppm).27 The peak at 6.9 ( 0.3 ppm can be attributed to protons attached to monoaromatic rings, rather than condensed-ring systems such as naphthalene, anthracene or pyrene. The peak at 8.4 ( 0.6 ppm can be assigned to protons in sterically hindered aromatic structures,27 to phenolic OH protons, to protons at positions R to nitrogen in an aromatic (pyridine-type) ring, and to protons in benzothiophene-type structures.29 The fractions of these two peaks are similar in C5D5Nsaturated premium coal 601, as can be seen in Table 1. The total percentage of aromatic protons is 34%. The broad band around 11.5 ( 1.5 ppm contains only 4% of the total proton population. As this band is broad and low in intensity, the peak position and area calculated from spectral deconvolution are very sensitive to distortions in the baseline of 1H CRAMPS spectra. The chemical shift of this peak corresponds to protons in carboxyl groups, or perhaps hydroxyl protons in structures of the hydroxypyridine type.13 An analogous deconvolution of the 1H CRAMPS spectrum of C5D5N-saturated premium coal 501 is difficult due to limited spectral resolution. However, we found that the dipolar-dephasing experiment is very helpful for assigning features of spectral fine structure that are dominated by broad peaks due to protons in rigid chemical structures. As shown in Figure 6b, 1H CRAMPS spectra of C5D5N-saturated premium coal 501 obtained with a dipolar-dephasing time of 184.6 µs has substantial high-resolution features of individual chemical functionalities in mobile components. All of the sharp features of fine structure in the un-dephased 1H CRAMPS spectrum (Figure 6a) are preserved in the dipolar-dephased spectrum. As will be discussed below, the sharp features in the 1H CRAMPS spectra of C5D5Nsaturated coals are largely due to protons in mobile components that show Lorentzian dipolar-dephasing behavior with long dephasing time constants (0.2-10 ms). The total 1H CRAMPS spectrum of a pyridine(28) Stefanova, M.; Simoneit, B. R. T.; Stojanova, G.; Nosyrev, I. E.; Goranova, M. Fuel 1995, 74, 768. (29) Khan, M. R. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1988, 33, 253.

502


Figure 6. 1H CRAMPS spectra of C5D5N-saturated premium coal 501 obtained at 25 °C. (a) No dipolar dephasing. (b) Obtained with a dipolar-dephasing time of 184.6 µs. (c) Difference spectrum obtained by subtracting the spectrum in (b) from the spectrum in (a).

saturated coal can be considered as a superposition of a spectrum of “mobile” protons and a spectrum of “rigid” protons. The former generate a much higher spectral resolution, while the later render a resolution that is comparable to that of the 1H CRAMPS spectrum of the untreated coal. As can be seen from Figure 6c, the difference spectrum, obtained by subtraction of the 1H CRAMPS spectrum of C5D5N-saturated premium coal 501 obtained with a 184.6 µs dipolar-dephasing period from the spectrum obtained without dipolar dephasing, looks very much like the 1H CRAMPS spectrum of the untreated coal without pyridine saturation (Figure 2a). Most of the fine structure features have disappeared in the difference spectrum, except for small humps around 4.7 ppm due to incomplete cancellation of the 4.7 ppm peak in the difference spectrum. As can be seen from the deconvolutions shown in Figures 7a and 7b, the difference spectrum can be well fitted with two broad peaks, just as in the deconvolution of the 1H CRAMPS spectrum of the untreated premium coal 501.21 Six peaks can be identified in the deconvolution of the 1H CRAMPS spectrum of C D N-saturated premium 5 5 coal 501 obtained with a dipolar-dephasing period of 184.6 µs (Figure 7 c,d). The chemical shifts of the six peaks are 1.0 ( 0.1 ppm, 2.0 ( 0.1 ppm, 3.7 ( 0.1 ppm, 4.8 ( 0.1 ppm, 7.0 ( 0.3 ppm, and 8.6 ( 0.2 ppm, respectively. The relative fractions and fwhhs of the six peaks are listed in Table 2. We can see that the chemical shifts of these peaks are close to those of the C5D5Nsaturated premium coal 601, except for the peak at 4.8 ppm. The peak assignments given above for coal 601 for spectral contributions at 1.0, 2.0, 3.7, 7.0, and 8.6 ppm should still apply to the 1H CRAMPS spectrum of C5D5N-saturated premium coal 501. The narrow peak at 4.8 ppm in the coal 501 sample may be due to water, which could be released from the very tight internal voids of coal when the coal structure is swollen by pyridine.12

Xiong and Maciel

Figure 7. (a) The difference spectrum of Figure 6(c) and (b) its simulated spectrum from peak deconvolution. (c) Dipolardephasing 1H CRAMPS spectrum of Figure 6(b) and (d) its simulated spectrum from peak deconvolution. In the simulated spectra, dashed lines represent individual peaks from the deconvolution; the solid line represents the sum of all the deconvoluted peaks. Table 2. Relative Concentrations (%) of the Various Types of Protons in the 1H CRAMPS Spectrum of C5D5N-Saturated Premium Coal 501 (Pocahontas No. 3) obtained with a Dipolar-Dephasing Time of 184.6 µs chemical shift (ppm)a percentage (%) fwhh (ppm) a

1.0 23 1.1

2.0 5.3 0.6

3.7 35 1.9

4.8 15 0.5

7.0 19 1.3

8.2 2.5 0.6

Relative to tetramethylsilane (TMS).

The reader should note that the deconvolution results listed in Table 2 are for mobile protons in the Pocahontas No. 3 coal, which, from analysis of the results shown in Figure 7, accounts for only 7% of the total protons in the coal. All the numbers in Table 2 should be multiplied by 7% to convert the percentages of various types of mobile protons to the percentages relative to the total number of protons in the whole coal. For example, the percentage of mobile water protons (4.8 ppm peak) in the whole coal should be 7% × (15% ( 4%), or 1.1% ( 0.3% (68% confidence limit). The percentage of aromatic protons in the Pocahontas coal was derived from peak deconvolution on both rigid and mobile protons as 65% ( 5%. Since there are 58.6 total protons per 100 carbons in the Pocahontas coal from elementary analysis,15 we can then determine the value of 38 ( 3 for the number of aromatic protons per 100 carbons, which is in reasonable agreement with previous work by others.15 2. Quantitative Analysis of 1H-1H DipolarDephasing Behavior of Pyridine-Saturated Premium Coal 601 (Blind Canyon). 1H-1H dipolardephasing experiments are very useful for exploring structural-dynamical details of coals. To quantify the dipolar-dephasing behavior of the C5D5N-saturated coal, computer peak deconvolution was employed to obtain the areas of deconvoluted peaks in each spectrum shown



in Figure 3. As spectral resolution between the peak at 1.7 ppm and the one at 2.5 ppm was not good enough to ensure reliable deconvolution of these two peaks over the entire range of dephasing periods, we combined the areas of these two peaks together to model their dipolardephasing behavior. As these two peaks represent nonmethyl aliphatic protons, we will use the term “nonmethyl protons” for later discussion. For the same reason, we combined the areas of the 6.9 and 8.4 ppm peaks as the area for aromatic protons. The peak at 1.0 ppm represents protons in methyl groups, and is analyzed separately from nonmethyl protons. As the very weak and broad peak at 11.5 ppm is too sensitive to variations/distortions in the baseline of 1H CRAMPS spectra, we did not include this peak in the quantitative analysis of dipolar dephasing. Both Gaussian and Lorentzian decay components were used to model the dipolar-dephasing behavior of C5D5N-saturated premium coal 601. Gaussian dipolardephasing behavior is a characteristic of rigid-lattice structures, while Lorentzian dephasing behavior is a characteristic of structures with substantial rotational mobility.15 A general equation for describing the dipolardephasing curve is as follows: 2

2

M(t) ) MGF(0)e-t /(2T GF) + MGS(0)e-t /(2T GS) + 2

2

MLF(0)e-t/TLF + MLS(0)e-t/TLS (1) where MGF(0) and MGS(0) are the initial amplitudes of the fast and slow Gaussian dephasing components (GDC), respectively; MLF(0) and MLS(0) are the initial amplitudes of fast and slow Lorentzian dephasing components (LDC), respectively; TGF, TGS, TLF, and TLS represent corresponding dipolar-dephasing time constants. The fraction of each dephasing component can be calculated from its initial amplitude as:

fi )

Mi(0) MGF(0) + MGS(0) + MLF(0) + MLS(0)

(2)

where i ) GF, GS, LF, LS Nonlinear least-squares analysis, based on eq 1, was used to determine how many dephasing components should be included in the model that fits best the dipolar-dephasing curves of methyl protons, nonmethyl protons, and aromatic protons. The dipolar-dephasing time constants were then calculated from the best fit of the model for each type of protons. As an example, Figure 8 shows the experimental dipolar-dephasing curves of the methyl protons at 25 °C and 90 °C and the simulated curves corresponding to the best fit. The results from such data analysis are listed in Table 3. The 1H NMR line width of coal in the absence of multiple-pulse decoupling should correspond roughly to (the inverse of) the dipolar-dephasing time constant, Tdd. Random reorientational motion, with a correlation time of τc, will generally reduce the line width and extend the dipolar-dephasing behavior. For isotropic random motion of many equivalent dipolar coupled spins, the reduced line width can be analyzed quantitatively on the basis of the spectral density function of random motion under magic-angle spinning (MAS). In the regime where the line is substantially narrowed motionally from its rigid lattice value ∆, the dependence

Figure 8. Experimental dipolar-dephasing data (open circles) of methyl protons in C5D5N-saturated premium coal 601 obtained at (a) 25 °C and (b) 90 °C. The data were fitted with eq 1. Dashed curves represent individual dephasing components; the solid curve is the sum of all the dephasing components. Table 3. Dipolar-Dephasing Parameters of the Various Types of Protons in C5D5N-Saturated Premium Coal 601a,b 25 °C

90 °C

methyl nonmethyl aromatic methyl nonmethyl aromatic fGF (%)a TGF (ms)b fGS (%)a TGS (ms)b fLF (%)a TLF (ms)b fLS (%)a TLS (ms)b

11 0.031 41 0.25 48 9.9

63 0.011 11 0.061 11 0.24 15 8.6

55 0.020 33 0.17 13 4.3

27 0.18 73 10

34 0.012 6.0 0.032 34 0.13 26 9.3

38 0.021 49 0.13 13 11

a The fractions (%) of the four dephasing components are denoted as fGF, fGS, fLF, and fLS. b TGF, TGS, TLF, and TLS denote the dipolar dephasing time constants of fast Gaussian, slow Gaussian, fast Lorentzian, and slow Lorentzian dephasing components, respectively.

of 1/Tdd on the correlation time of random motion τc and magic angle spinning speed ωr can be expressed as,30

{

1 2 1 ) ∆2τc + 2 2 Tdd 1 + ωr τc 1 + 4ω2r τ2c

}

(3)

In the case that ωr is much smaller than 1/τc, the expression can be simplified as follows: (30) Waugh, J. S. In Nuclear Magnetic Resonance in Modern Technology; Maciel, G. E., Ed.; NATO ASI Series; Kluwer: Dordretcht, The Netherlands, 1994; p 359.

504


1 ) 3∆2τc Tdd

Xiong and Maciel

(4)

From eq 4, we see that a dipolar-dephasing time constant of 0.2 ms for a fast Lorentzian dephasing component corresponds to a random motion with a correlation time of about 1.4 µs, assuming that the line width of protons in a rigid lattice is about 35 kHz. A dipolar-dephasing time constant of 10 ms for a slow Lorentzian dephasing component corresponds to a random motion with a correlation time of about 27 ns. For comparison, we note that the correlation time of random motion responsible for generating the slow Gaussian dephasing component in coal was derived from our previous work to be on the order of 16 µs.21 Thus, it is clear that the Lorentzian dephasing components are produced by random motions with correlation times that are orders of magnitude shorter than that of the slow Gaussian dephasing component. A methyl group has a very low energy barrier for internal rotation around its C3 axis. Moreover, since a methyl group is often a terminating group of a long aliphatic chain, it can also rotate easily relative to the rest of the molecule. Inspection of Table 3 reveals, as one might have expected, that most of the methyl protons (89%) belong to Lorentzian dephasing components at 25 °C. The remaining 11% of the methyl protons, which show slow Gaussian dephasing behavior at 25 °C, are all converted to more mobile Lorentzian decay components at 90 °C. This suggests that the energy barriers that restrict rotation of the methyl protons are not much higher than kT at room temperature (i.e., 0.6 kcal/mol). The dipolar-dephasing curve of nonmethyl aliphatic protons consists of contributions from all four types of dephasing components in eq 1. As shown in Table 3, time constants of the GDCs in the C5D5N-saturated coal are close to those of aliphatic protons in untreated premium coals.21 Most of the nonmethyl aliphatic protons show Gaussian dephasing behaviors at 25 °C. This means that most nonmethyl aliphatic protons are distributed among very rigid structures, which cannot be easily mobilized by pyridine saturation. When the temperature is increased to 90 °C, a large fraction of the GDCs at 25 °C are thermally activated to show fast or slow Lorentzian dephasing behaviors. The fractions of fast and slow LDCs are increased by 23% and 11%, respectively, at 90 °C. The time constants of nonmethyl aliphatic LDCs are close to those of methyl protons at both 25 °C and 90 °C. This suggests that mobile methyl and nonmethyl aliphatic protons may exist in the same domain and that 1H-1H spin-diffusion effectively averages out differences of dephasing time constants on a time scale of 0.1 to 10 ms. The existence of a large aliphatic-rich domain in premium coal 601, probably a network of cross-linked aliphatic chains with domain sizes larger than about 100 nm, was demonstrated by spin-lattice relaxation measurements in our previous work.22 The dipolar-dephasing curve of aromatic protons in C5D5N-saturated coal 601 consists of three components: fast Gaussian, fast Lorentzian, and slow Lorentzian. A single GDC with a dephasing time constant of

16 µs was observed for the original (untreated) premium coal 601.21 We can still identify one GDC for rigid protons in the C5D5N-saturated coal, although the dephasing constant is a little larger (20 µs) than that in the original coal (16 µs). This suggests that some of the rigid aromatic structures in the C5D5N-saturated coal are essentially unperturbed by the addition of pyridine. However, a large fraction (45-60%) of the aromatic protons in the untreated coal are mobilized by pyridine saturation to show either fast or slow Lorentzian dephasing behaviors. When the temperature is increased from 25 °C to 90 °C, the dephasing time constant of the slow LDC of aromatic protons is significantly prolonged, from 4.3 to 11.0 ms, but the fraction of this component is essentially unchanged. The rigid Gaussian dephasing aromatic protons are apparently converted (partially) only to the fast LDC at 90 °C, a behavior that is discussed in a later section. 3. Quantitative Analysis of 1H-1H DipolarDephasing Behavior of Pyridine-Saturated Premium Coal 501 (Pocahontas No. 3). Quantifying the dipolar-dephasing behavior of the C5D5N-saturated premium coal 501 is difficult because of very poor spectral resolution. As discussed above, the 1H CRAMPS spectrum of C5D5N-saturated premium coal 501 can be considered as a superposition of a spectrum of mobile protons and a spectrum of rigid protons. In the 1H CRAMPS spectrum obtained with a dipolar dephasing time of 184.6 µs, almost all the fast dephasing components due to rigid protons have decayed away, while the signals due to mobile protons are hardly affected. Thus, peak deconvolution of the well-resolved spectra obtained with a dephasing period of 184.5 µs or longer is straightforward. The spectrum due to rigid protons can be obtained from the difference spectrum derived by subtraction of 1H CRAMPS spectra obtained with a very short dipolar-dephasing period from one obtained with a dipolar dephasing time of 184.6 µs. This kind of strategy was also successfully used for deconvoluting 1H CRAMPS spectra obtained with short dipolardephasing periods. As only two peaks can be resolved in the 1H CRAMPS spectra of rigid protons, we used the total areas of aliphatic components and the total areas of aromatic components to analyze the dipolardephasing data over the entire range of dephasing periods. The resulting dipolar-dephasing curves of aliphatic and aromatic protons were analyzed using the model described in eq 1 and the results are presented in Table 4. At both temperatures used in this work, the dipolardephasing curve of aromatic protons in pyridinesaturated coal 501 consists of only two dephasing components, a fast GDC and a slow LDC. The time constants for the GDCs in the C5D5N-saturated coal at both temperatures are essentially the same as measured for the original coal at 25 °C.21 This suggests that the rigid aromatic structures are preserved under pyridine saturation. At either temperature, only a very small fraction (4-5%) of the aromatic protons show slow Lorentzian dephasing behavior in the pyridine-saturated coal 501. The absence of other dephasing components with intermediate rotational mobility suggests that the aromatic structures are very rigid, with a high degree of covalent cross-links, which cannot be broken by



Table 4. Dipolar-Dephasing Time Constants of Aliphatic and Aromatic Protons in C5D5N-Saturated Premium Coal 501a,b 25 °C (%)a

fGF TGF (ms)b fGS (%)a TGS (ms)b fLF (%)a TLF (ms)b fLS (%)a TLS (ms)b

90 °C

aliphatic

aromatic

aliphatic

aromatic

81 0.0092 8.0 0.042

95 0.014

67 0.0091

96 0.013

4.8 6.2

18 0.18 14 12

4.3 5.9

11 9.9

a The fractions (%) of the four dephasing components are denoted as fGF, fGS, fLF, and fLS. b TGF, TGS, TLF, and TLS denote the dipolar dephasing time constants of fast Gaussian, slow Gaussian, fast Lorentzian, and slow Lorentzian dephasing components, respectively.

pyridine. This is in agreement with the expected structure of a high rank LVB coal. When the temperature is increased from 25 °C to 90 °C, no significant changes in either the dephasing time constants or the fractions of the two dephasing components are observed. This further confirms the view that pyridine is not effective in mobilizing the aromatic structures in premium coal 501. The aliphatic protons in C5D5N-saturated premium coal 501 exhibit very different behaviors. At 25 °C, the dephasing curve of aliphatic protons consists of three types of dephasing components: fast Gaussian, slow Gaussian, and slow Lorentzian. In contrast to the very small changes discussed above for the aromatic structures, a significant fraction (11%) of the aliphatic protons are mobilized to show Lorentzian dephasing behavior at 25 °C when the coal is saturated with C5D5N. This implies that a significant fraction of the cross-links or “bridges” that constrain aliphatic structures in the untreated coal may be based on hydrogen bonds, which could be disrupted by the addition of pyridine. At 90 °C, the dephasing curve of aliphatic protons in pyridine-saturated coal 501 still consists of three types of dephasing components. However, the slow GDC observed at 25 °C has disappeared at 90 °C, and a fast LDC is observed at 90 °C. These results again demonstrate that aliphatic structures are much easier to mobilize than aromatic structures in premium coal 501. One should note that, although only 10.5% of the carbons in the Pocahontas coal are aliphatic,15 35% of the total protons are aliphatic protons. That means that protons are concentrated in aliphatic structures. Note also that 11% of the slow LDC at 25 °C (Table 4) is the percentage relative to the total aliphatic protons, not the total number of protons. 11-14% of the aliphatic protons would be equivalent to 3.9-4.9% of the total protons in the Pocahontas No. 3 coal. Although only less than 3% of the Pocahontas No. 3 coal can be extracted by pyridine, the protons in the extract could account for close to 10% of the total protons in the coal, if the extract were dominated by small aliphatic molecules. Thus, the 11-14% slow LDC (relative to total aliphatic protons) could be mainly due to aliphatic structures in mobile small molecules. However, both aliphatic and aromatic slow LDCs could also be due to flexible side chains. Further work to compare

mobility of the Pocahontas No. 3 coal and its pyridineextraction residue would shed light on this issue. The significant change of molecular mobility in aliphatic structures with an increase in temperature suggests that hydrogen bonding may play a more important role in the Pocahontas No. 3 coal than previously thought. 4. Mechanism for Promoted Molecular Motion in C5D5N-Saturated Coal. Our results have established that a significant fraction of protons in C5D5Nsaturated premium coals 601 and 501 show Lorentzian dephasing behaviors that are governed by random motion with correlation times in the range of 1 µs to 30 ns at 25 °C. In contrast, our previous work has shown that the protons in the corresponding original coals show only Gaussian dephasing behavior under thermal treatment at temperatures up to 180 °C.21 The correlation time of promoted random motion in the original coals under thermal treatment is on the order of 16 µs. The presence of very mobile, Lorentzian dipolar-dephasing components in the C5D5N-saturated coal demonstrates that pyridine-saturation is much more effective in promoting molecular motion than is thermal treatment. This result correlates well with differences between the macroscopic properties of coal with solventsaturation or thermal treatment. For example, the uniaxial optical anisotropy of raw bituminous coal disappears within a few seconds when the coal sample is immersed in pyridine at room temperature; but the anisotropy is not destroyed thermally until the temperature reaches around 350 °C.31 Brenner concluded, on the basis of a thermodynamics perspective in his study on the optical anisotropy of solvent-swollen coal, that pyridine-swollen coal behaves like a cross-linked rubber, while untreated coal behaves like a macromolecular glassy plastic.32 In fact, the Lorentzian dephasing time constants obtained for C5D5N-saturated coals in the present study are much smaller than those expected for small molecules in liquid phases, but comparable with typical values expected for the mobile, amorphous phases of cross-linked semicrystalline polymers above the glass-rubber transition temperatures.16 Our results, obtained from direct detection of molecular dynamics, support Brenner’s thermodynamic view of solvent-swollen coal. The rigid macromolecular structures in coals are stabilized by intra- and interchain cross-links. Secondary interactions, such as hydrogen bonds and electrostatic bonds in charge-transfer complexes, are believed to play a very important role in stabilizing coal structure.31 Green studied bituminous coals in which hydroxyl groups had been acelylated.27 He concluded that there are about three times as many hydrogen-bonding cross-links as covalent cross-links in the untreated coal samples studied. In exhaustively Soxhlet-extracted coal samples, the cross-link density is less than half of the density in the corresponding unextracted coals.33 The dramatically enhanced molecular mobility in coals brought about by pyridine saturation can be understood in terms of the following solvent/coal interaction mechanism. Pyridine molecules can effectively penetrate into the macromolecular network of the (31) Brenner, D. Nature 1983, 306, 772. (32) Barton, W. A.; Lynch, L. J.; Webster, D. S. Fuel 1984, 63, 1262. (33) Green, T. K. Ph.D. Thesis, University of Tennessee, 1983.

506


untreated coal by disrupting most of the hydrogen bonds and electrostatic interactions that cross-link and stabilize the macromolecular structure of coal. This causes significant swelling or expansion of the macromolecular network of coal. Small molecules and long side chains that are tightly confined in small voids in the coal structure are released and become a major component of the very mobile phase (slowly dephasing Lorentzian component) in pyridine-saturated coal. It is also possible that some small molecules are “attached” to the macromolecular structure of coal through hydrogen bonds or electrostatic interactions; these molecules can be detached from the macromolecular structure by pyridine. When the temperature is increased from 25 °C to 90 °C, the kinetic energy and mobility of pyridine molecules are significantly increased; pyridine molecules can break apart additional interactions. Therefore, more trapped small molecules are released at higher temperature and the void space is further enlarged. In this manner the fraction of the very mobile phase in pyridine-saturated premium coal 601 is significantly increased at higher temperature. The key to understanding the enhanced mobility brought about by pyridine saturation is the expansion of void spaces in pyridine-swollen coals. It was reported that the average pore volume could increase 2-fold after pyridine saturation of bituminous coals.34 The increase in void space has several kinds of impact on molecular mobility in coal and proton dipolar-dephasing behavior. First, the increase of void space can dramatically lower the energy barrier that restricts the rotational mobility of small molecules and long side chains originally trapped in tight voids; the increased void space will substantially reduce steric hindrance of rotational motion. Enlarged void space also lowers intermolecular interactions between small molecules and the macromolecular matrix, as van der Waals interactions and electrostatic interactions are strongly distance dependent. The increased rotational mobility will effectively average out both intramolecular and intermolecular 1H-1H magnetic dipolar interactions, reducing the rate of dipolar dephasing. Second, the larger void space permits more isotropic character in the motion of small molecules and side chains, also yielding more effective averaging of 1H-1H dipolar couplings. Third, a larger void space means a larger average distance between protons in small molecules and protons in the rigid macromolecular matrix around the molecules. As the strength of 1H-1H dipolar interactions is strongly distance-dependent, the increased distance can effectively reduce dipolar couplings between protons in small molecules and protons in the surrounding macromolecular matrix and reduce the rate of dipolar dephasing. The point of view that the dramatically increased void space leads to a dramatic increase in molecular mobility in pyridine-saturated coal also explains satisfactorily why much lower molecular mobility is induced by thermal treatments than by pyridine saturation. The thermal energy at temperatures up to 230 °C is not large enough to break either multiple hydrogen-bonding (34) Mahajan, O. P. In Sample Selection, Aging, and Reactivity of Coal; Klein, R., Wellek, R., Eds.; John Wiley & Sons: New York, 1989; p 157.

Xiong and Maciel

“bridges” or covalent-bonding cross-links in macromolecular structures of coal. In the absence of pyridine treatment, the macromolecular structure and void spaces are essentially unchanged when the temperature is increased from 25 °C to 230 °C. Thus, the energy barrier that restricts the motion of small molecules and side chains trapped in tight voids is essentially unchanged at temperatures up to 230 °C. When the temperature is increased, the fraction of small molecules and side chains that are able to overcome the energy barrier increases according to the Boltzmann distribution. That explains why the fraction of the slow Gaussian dephasing component in untreated coals increases as the temperature is increased. However, these thermally induced motions are very restricted and anisotropic in nature, as the very tight void spaces in the untreated coal structure prohibit isotropic rapid rotational motion. That is why we do not observe a mobile Lorentzian dephasing behavior in an untreated coal at temperatures as high as 180 °C. It is interesting to see how our measures of molecular mobility correlate with the size distribution of void spaces measured in pyridine-saturated coal. Spears et al. studied changes in pore structures of premium coals using an electron paramagnetic resonance (EPR) spin probe method.35 Using nitroxide radicals of different sizes to determine the pore size distribution in solventswollen coals, they found that the shape of the pore in coals is cylindrical and a significant pore elongation occurs with pyridine saturation, even at room temperature. The pore size distribution of pyridine-swollen lower rank coals, such as premium coal 301 and 601, were found to be clearly bimodal. At 25 °C, one maximum was found for a spin probe skeletal chain length of 11 carbon atoms, while another maximum was found for a chain length of 15 atoms. When the temperature was increased to 60 °C, the pore size distribution still showed a bimodal distribution, with one maximum unchanged at a chain length of 11 atoms and another maximum shifted to a chain length of 24 atoms. Those results correlate well with our finding of two Lorentzian dephasing components. The slow LDC could correspond to small molecules and side chains in large pores created by pyridine swelling. It is possible that the fast LDC corresponds to small molecules and side chains in smaller pores that are larger than the voids in the original coal, but not large enough to allow a more isotropic rotational motion. The further enlargement of large pores at higher temperatures is in agreement with the slight increase in the time constants of the slow LDC (Table 3). For pyridine-saturated premium coal 601, there is actually a slight decrease in the dephasing time constants of the fast LDCs when the temperature is increased from 25 °C to 90 °C, but the fractions of both slow and fast LDCs increase with increasing temperature. These two observations may be explained as manifestations of the enlargement of existing pores and the creation of new pores at higher temperature. Enlargement of the small pores associated with the fast LDC will lead to an increase of the fraction of the slow LDC. The creation of new, small pores will increase the (35) Spears, D. R.; Kispert, L. D.; Piekara-Sady, L. Fuel 1992, 71, 1008.


fraction of the fast LDC. Pores that can be created only at higher temperature are more likely in a rigid environment with a significant population of covalent cross-links, which cannot be broken by pyridine. It is conceivable that such pores are smaller in size or less flexible than pores that are created in an environment that is rich in hydrogen bonding. This could be the reason a smaller dephasing time constant is observed at higher temperature for the fast Lorentzian dephasing component. As mentioned above, when the temperature of C5D5Nsaturated premium coal 601 is increased from 25 °C to 90 °C, the fraction of protons in the slow aromatic LDC is unchanged, but its time constant is increased, as seen in Table 3. This change in time constant is understandable on the basis of the above discussion. Aromatic structures are less flexible and perhaps typically larger than aliphatic structures in bituminous coals, so isotropic motion of aromatic structures requires a much larger free space. This explains why we observe a pronounced increase in the time constant for aromatic protons with slow Lorentzian dephasing characteristic, but only a slight increase in the time constant for aliphatic protons in C5D5N-saturated coal 601 as its temperature is increased. The unchanged fraction of the slow aromatic LDC as the temperature is increased from 25 °C to 90 °C could be due to the difficulty associated with further enlargement of small pores created at 25 °C by pyridine saturation. Another possible explanation is that the fast LDC of aromatic protons is mainly associated with segmental motion of mobile chains of the macromolecular structure. Such motion is not directly related to the sizes of pores. In contrast to premium coal 601, there is no identifiable fast LDC for aromatic protons of C5D5N-saturated premium coal 501 (Table 4). This difference between the coals is related to fundamental structural differences between HVB and LVB coals. The macromolecular structure of a HVB coal is believed to consist of relatively small aromatic clusters connected by flexible aliphatic cross-links.32 Hydrogen-bonding bridges further stabilize the macromolecular network and make it rather rigid. When the hydrogen-bonding bridges are disrupted by pyridine, the aromatic structure in the macromolecular chain could gain substantial segmental mobility in HVB coals. In LVB coals, such as 501, the aromatic clusters are much larger, and connected by shorter and less flexible cross-links; this makes the aromatic structure much more rigid, so it cannot be effectively mobilized by pyridine. 5. Resolution Enhancement by Solvent Saturation. We have seen that resolution enhancement in 1H CRAMPS spectra of coal can be realized by pyridine saturation. A comparison between room temperature 1H CRAMPS spectra of untreated and C5D5N-saturated samples for all eight Argonne premium coals can be found in a paper by Jurkiewicz et al.12 The extent of resolution enhancement depends very much on the structure of an individual coal. In general, HVB and MVB coals show more dramatic resolution enhancement than coals of other ranks. The origin of this difference was not fully understood in the previous work, but the change in spectral resolution with variation of the dipolar-dephasing period provides a hint on this issue.


For 1H CRAMPS spectra of C5D5N-saturated premium coals 501 and 601, the following observations were established: (1) The spectral resolution of 1H CRAMPS spectra obtained with dipolar dephasing increases with the dipolar-dephasing time until it reaches about 180 µs. 1H CRAMPS spectra obtained with a dipolar-dephasing time of 185 µs show very good spectral resolution, irrespective of the resolution of 1H CRAMPS spectra obtained without a dephasing period. With a dephasing time longer than 185 µs, no appreciable change of spectral resolution can be observed beyond that achieved with a 185 µs dephasing period. (2) All of the fine structure (such as identifiable sharp peaks and shoulders) in the 1H CRAMPS spectrum obtained without a dipolar dephasing period is preserved in the 1H CRAMPS spectra obtained with a dipolar dephasing period longer than 185 µs. In fact, these spectral fine-structure features are much easier to recognize in 1H CRAMPS spectra obtained with dipolar-dephasing periods longer than 180 µs, because of much better resolution. (3) Almost all the spectral fine structure seen in the 1H CRAMPS spectrum of a coal disappears in the difference spectrum obtained by subtracting the 1H CRAMPS spectrum obtained with a dipolar-dephasing period of 185 µs from the 1H CRAMPS spectrum obtained without a dephasing period. The spectral resolution of the difference spectrum is comparable to that obtained on the untreated coal. Only two broad peaks, due to aliphatic and aromatic protons, are clearly recognizable in the difference spectrum. The spectrum obtained with a dephasing period of 185 µs represents the spectrum of mobile protons with Lorentzian dephasing characteristics. Thus, the difference spectrum represents the spectrum of rigid protons in C5D5N-saturated coals. The 1H CRAMPS spectrum of C5D5N-saturated coal obtained without a dipolardephasing period can be considered as a superposition of a spectrum of rigid protons and a spectrum of mobile protons. This clean separation of the 1H CRAMPS spectrum into two distinct component types is understandable on the basis of results from 1H spin-exchange experiments,24 which have shown that there is no observable spin exchange and/or spin diffusion between the rigid protons with Gaussian dephasing characteristics and the mobile protons with Lorentzian dephasing characteristics. Based on these observations, we can conclude that the enhanced resolution of the 1H CRAMPS spectrum of a C5D5N-saturated coal has its origin in the high-resolution 1H CRAMPS spectrum of the mobile protons showing Lorentzian dephasing characteristics. This conclusion explains satisfactorily differences in the 1H CRAMPS resolution observed previously among C5D5Nsaturated coals of various ranks.12,24 In HVB coals, pyridine can effectively break the hydrogen-bonding bridges that stabilize the macromolecular structures. The resulting severe swelling of coal structure converts a substantial portion of the proton containing moieties into very mobile states with Lorentzian dephasing characteristics. As the fraction of Lorentzian dephasing component is large, the overall 1H CRAMPS spectral resolution of a C5D5N-saturated HVB

508


coal appears much higher than that of the corresponding untreated coal. In a high rank coal, such as a LVB coal, the bridging density due to hydrogen bonding is very low. Therefore, pyridine cannot effectively swell the macromolecular structure of a high rank coal, and only a very small fraction (∼10%) of protons show Lorentzian dephasing behavior. The 1H CRAMPS spectrum of a C5D5N-saturated high rank coal is thus a superposition of a large, relatively low-resolution signal and a small, high-resolution signal. The overall resolution will not be much improved from that of the original coal. 6. Correlation of Molecular Mobility with Structure in Coals. From 1H NMR experiments with 1H CRAMPS detection and pyridine saturation, we have shown that molecular motion in coal, as detected from proton dipolar-dephasing experiments, depends very much on the rank of the coal, molecular structures of the various components in coals and the sample treatment conditions (e.g., solvent, temperature). We have shown in this and previous work21,22 that temperature has a marked effect on the mobility of different molecular moieties in coal, even for pyridine-saturated coal. Nevertheless, most of the previous NMR work on coal was carried out only at room temperature. It is no surprise to see many discrepancies regarding correlation between molecular mobility and coal structure in the literature. Some papers have made a direct comparison of the aromaticity of the mobile component between an untreated coal and a pyridine-saturated coal and attributed the difference in measured aromaticity to the idiosyncrasies of certain NMR techniques used.18-20 However, our work clearly shows that the mobile component in an untreated coal is not directly related to the mobile component in the corresponding pyridinesaturated coal, so the measured difference in aromaticity is a natural, intrinsic coal characteristic and is not based primarily on the characteristics of the NMR techniques employed. For example, in untreated premium coal 601 at 25 °C, all the aromatic protons show rigid-lattice behavior (i.e., 100% fast GDC) and no mobile aromatic proton component was detected,21 but in pyridine-saturated premium coal at 25 °C there are 55% aromatic protons in fast GDC, 33% in fast LDC, and 13% in slow LDC (Table 3). Clearly, there is no direct correlation of aromatic protons of different mobilities in pyridine-saturated premium coal 601 with those in the untreated premium coal 601 at 25 °C. The structural moieties that can be mobilized effectively by pyridine addition are structures with high densities of hydrogen bonding and other secondary interactions that can be disrupted by pyridine. This type of structure could be very difficult to mobilize with thermal treatment alone, as hydrogen-bonding networks consisting of multiple hydrogen bonds cannot be broken up thermally at temperatures up to 200 °C. A common approach used in numerous published papers10-12,16,26 was simply to assign the rigid components (fast GDC), determined from proton NMR experiments of the dipolar-dephasing type on either untreated coal or pyridine-saturated coal, as the rigid macromolecular phase in coal. However, this may not be entirely appropriate, based on our results. As shown elsewhere,24,36 29% of aromatic protons in premium coal 601 (36) Xiong, J.; Maciel, G. E. Energy Fuels, submitted .

Xiong and Maciel

can be extracted by pyridine, and thus belong to the molecular phase; yet, all the aromatic protons show rigid-lattice behavior (i.e., 100% fast GDC) in untreated premium coal 601 at 25 °C. Thus, assigning the rigid component in untreated coal simply as the macromolecular phase is not correct. Even in pyridine-saturated coal, small molecules can still show rigid-lattice behavior, appearing in that sense like rigid macromolecular networks. As shown in Table 3, about half of the fast Gaussian dephasing component of nonmethyl protons in pyridine-saturated premium coal 601 was converted to very mobile states with either slow or fast Lorentzian dephasing characteristics when the temperature was increased merely from 25 °C to 90 °C. As shown elsewhere,24,36 most of mobile LDCs in pyridine-saturated premium coal 601 at 90 °C are extractable, thus belonging to the molecular phase. Therefore, a significant fraction of nonmethyl protons in the fast GDC at 25 °C still belong to the molecular phase, instead of the macromolecular phase. This leads us to conclude that protons constituting the fast Gaussian dephasing characteristic are not entirely located in the rigid macromolecular structure. A substantial fraction of the fast Gaussian dephasing components in the original untreated coal corresponds to protons in small molecules and/or long side chains of the macromolecular structure. These small molecules and side chains are either severely restricted by very tight-fitting voids in the rigid macromolecular matrix of coal, or strongly associated with the macromolecular structures through noncovalent attractive forces. Summary and Conclusions We have carried out the first systematic in-situ variable-temperature proton NMR study on pyridine saturated high-volatile and low-volatile bituminous coals from the Argonne premium coal bank, using CRAMPS detection. With improved proton dipolardephasing experiments, we were able quantitatively to monitor changes of molecular mobility of various molecular structural moieties upon pyridine saturation and temperature variation. The correlation between molecular structure and mobility was thus established. We have identified four types of dipolar dephasing behavior in pyridine-saturated coals. The fast Gaussian dephasing component with a dephasing time constant of 10-20 µs corresponds to protons in the rigid lattice of coal structure. Graphite-like large polycyclic aromatic structures in LVB coals are expected to show such dephasing behavior. However, small molecules trapped in very small voids in macromolecular networks of coal could also show such fast Gaussian dephasing behavior. The slow Gaussian dephasing component, with a dephasing time constant of 30-70 µs, represents protons in structures with very limited anisotropic rotational mobility and is produced by random molecular motion with a correlation time on the order of 16 µs. If the noncovalent attractive forces that stabilize the macromolecular network of coal are not strong enough, protons in such structures could show slow Gaussian dephasing behavior. The slow Lorentzian dephasing component, with a dephasing time constant of 4-12 ms, corresponds to protons in structures with very high rotational mobility and are produced by random molec-


ular motion with a correlation time on the order of 30 ns. Such structures could conceivably be assigned to small molecules, and to very flexible, long side chains on the macromolecular network of coal. The fast Lorentzian dephasing component, with a dephasing time constant of 0.1-0.3 ms, could also be assigned to protons in small molecules and side chains of macromolecular structures; however, for fast Lorentzian dephasing components, the rotational mobility of small molecules and side chains are limited to a certain extent by their environments, such as relatively tight-fitting voids around the substantially mobile structures. Random motion with a correlation time on the order of 1 µs is responsible for the fast Lorentian dephasing behavior. Another possible contributor to the fast Lorentzian dephasing component is from protons in macromolecular chain segments that undergo fast segmental reorientation. The presence of such segmental chain motion was confirmed by rapid reappearance of the optical anisotropy with pressures exerted on pyridine-saturated coal.31 This work was focused on how noncovalent associative bonds are affected at the molecular level by thermal treatment and solvent saturation. We have found that pyridine saturation can dramatically promote molecular motion, even at room temperature, while thermal treatment alone is much less effective in promoting molecular motion even at 230 °C. However, for pyridine-saturated coal, the molecular mobility of coal can be further enhanced with just a moderate increase in temperature. A model for explaining the enhancement of molecular


mobility induced by thermal treatment and solvent saturation has been proposed and discussed in terms of our new experimental results. A key element of this model is the dramatic expansion of void space in coal as a result of disruption of noncovalent associative interactions by pyridine. The CRAMPS resolution enhancement due to solvent-saturation is also satisfactorily explained. In this work, we have, for the first time, proposed and proved that small molecules can be buried in tight voids of the macromolecular matrix of coal and show dynamical behavior that is similar to that of rigid macromolecules, even when the coal is saturated with pyridine. Our variable temperature studies on both untreated coal and C5D5N-saturated coal support the basic idea of the M/M model. However, our new results clearly demonstrate that the correlation between molecular mobility and M/M structures is much more complicated than previously thought. There is no simple one-to-one correspondence between molecular mobility and molecular/ macromolecular phases in coal. This point was further confirmed in a separate study of residues of pyridineextracted coal, which will be presented in a separate paper.24,36 Acknowledgment. This work was supported by U.S. Department of Energy Grant No. DE-FG2293PC93206. EF010248P

Interactions between Pyridine and Coal at the Molecular Level

Recommend Documents