Ionomers and the Structure of Coal - Energy & Fuels (ACS Publications)

---

Energy & Fuels 2000, 14, 1115-1118

1115

Ionomers and the Structure of Coal P. C. Painter,* P. Opaprakasit, and A. Scaroni The Energy Institute, The Pennsylvania State University, University Park, Pennsylvania 16802 Received May 9, 2000. Revised Manuscript Received July 17, 2000

The decrease in pyridine-soluble material observed after soaking coals in solvents is examined. It is demonstrated that this is due to an increase in cross-linking associated with the formation of ionic domains or clusters, similar to those observed in a class of synthetic polymers known as ionomers.

Introduction The nature of cross-linking in coal remains controversial. On one hand, the “standard model” continues to treat coal as a covalently cross-linked macromolecular network. There are also those that hold the hydrogen bonds or other secondary forces provide additional crosslinks or junction zones, even though there is now a wealth of evidence in the polymer literature demonstrating that such point interactions (with strengths of the order of 5 kcal/mol) cannot act in this manner in the time frame of experiments such as swelling. In contrast to this approach, Iino and co-workers1,2 and Nishioka3-6 propose that coal is largely an associated structure, held together by secondary forces. Hydrogen bonds are again invoked, as are charge-transfer complexes, that must, given the strength of such interactions, act in some so far undefined cooperative manner in order to provide the necessary cohesion. The arguments and counter-arguments surrounding this issue have been laid out in the pages of this journal and at various ACS meetings, so it is unnecessary to reproduce them here. Nevertheless, they provided the impetus for this work, where we have been seeking to understand how various solvents interact with coal. One result that we found particularly intriguing was obtained by Larsen et al.,7 who determined that simply soaking various coals in chlorobenzene (at 115 °C) for a few days resulted in a significant decrease in the amount of pyridine-extractable material. The swelling ratio of the gel also decreased. Soaking in pyridine for just 1 day had the same effect. In work that preceded this, Takanohashi and Iino8 also reported that refluxing with pyridine greatly decreased the NMP/CS2 extractability of certain coals. * Author to whom correspondence should be addressed. (1) Takanohashi, T.; Iino, M.; Nishioka, M. Energy Fuels 1995, 9, 788. (2) Iino, M.; Takanohashi, T.; Ohsuga, H.; Toda, K. Fuel 1988, 67, 1639. (3) Nishioka, M.; Gebhard, L. A.; Silbernagel, B. G. Fuel 1991, 70, 341. (4) Nishioka, M. Fuel 1991, 70, 1413. (5) Nishioka, M. Fuel 1992, 71, 941. (6) Nishioka, M. Energy Fuels 1991, 5, 487. (7) Larsen, J. W.; Flowers, R. A.; Hall, P. J.; Carlson, G. Energy Fuels 1997, 11, 998. (8) Takanohashi, T.; Iino, M. Energy Fuels 1991, 5, 708.

Larsen et al.7 observed that this decrease in swelling and extractability must be a consequence of an increase in cross-link density, but there was no evidence for the formation of new covalent bonds which, as they pointed out, would also be unlikely under the conditions of their experiments. Furthermore, the controversy concerning the role of hydrogen bonds does not enter into this discussion, given that pyridine would be expected to “break” all such coal inter-segment interactions. Intriguingly, these authors also found an increase in the intensity of an X-ray diffraction peak corresponding to structures approximately 20 Å in size. In work that we will report here we have repeated these experiments and examined the infrared spectra of the coal before and after various solvent treatments. Our results demonstrate that the increased cross-link density is due to the formation of ionic complexes involving carboxylic acid groups. Given the largely hydrocarbon nature of the organic component of coal, it is likely that these complexes resemble the multiplets or clusters found in ionomers, a name used to describe synthetic polymers that contain a low concentration of functional groups such as carboxylic acids9 that are capable of ion-exchange to form ionic complexes. Although the concentration of such groups is usually small, they have a disproportionate effect on behavior and we believe this would explain a number of recent observations of coal properties. Experimental Section A Pittsburgh No. 8 coal sample was supplied by The Argonne National Laboratory. The sample was stored under dry nitrogen and was not dried before use. Preparation of Refluxed Coal Samples. The reflux procedure was equivalent to that described by Larsen et al.7 The coal sample was placed in excess chlorobenzene and refluxed at 115 °C under a nitrogen atmosphere for 4 days. Chlorobenzene was then removed by rotary evaporation followed by high vacuum at room temperature. Preparation of Acid and Alkali-Treated Coal Samples. A 2.5 g of coal sample was placed in 250 mL of 1 N HCl in THF and stirred at room temperature under a nitrogen (9) Eisenberg, A.; Kim, J. S. Introduction to Ionomers; John Wiley and Sons: New York, 1998.

10.1021/ef000092d CCC: $19.00 © 2000 American Chemical Society Published on Web 08/22/2000

1116

Energy & Fuels, Vol. 14, No. 5, 2000

Painter et al.

Table 1: Pyridine (Soxhlet) Extraction Yields (wt % of extracted material) of Original and Treated Coal Samples Pittsburgh No. 8 sample

extraction yield (%)

original coal sample Pittsburgh No. 8 coal sample refluxed in chlorobenzene acetylated coal sample Pittsburgh No. 8 acetylated sample refluxed in chlorobenzene refluxed sample treated with 1 N HCl acid-treated coal sample alkali-treated coal sample

38 30 38 28 46 51 29

atmosphere for 2 days. The 1 N HCl in THF was prepared by diluting an ACS reagent 37% HCl (12.1 mol/L in water). The THF was then removed from the coal using a rotary evaporator followed by high vacuum treatment at room temperature until it reached a constant weight. The alkali-treated sample was prepared by placing the acid-treated sample in 1 N NaOH in THF. The mixture was stirred at room temperature under a nitrogen atmosphere for 2 days and the THF was removed, as above. Preparation of Acetylated Samples. The acetylated coals were prepared following a procedure used previously in our laboratory.10 Determination of the Pyridine Extraction Yield. Extraction experiments were performed on the original coal sample, Pittsburgh No. 8, and the refluxed, acetylated-refluxed, and acid- and alkali-treated coal samples using the usual Soxhlet technique. The solvent used was pyridine. FTIR Spectra. FTIR spectra were recorded on a Digilab model FTS-45 at a resolution of 2 cm-1 by co-adding 200 scans. Samples for FTIR experiment were prepared as standard KBr pellets using approximately 2.0 mg of coal sample in 250 mg of KBr.

Results and Discussion Larsen et al.7 focused most of their attention on Pittsburgh No. 8 coal, so we will do the same. The weight % yield of pyridine (Soxhlet)- extracted material that we obtained from the starting coal was 38%, as listed in Table 1, somewhat less than the 44% obtained by Larsen and co-workers. This concerned us at first, but our results were repeatable and, as we will show below, would depend on variables such as moisture content of the samples, hence ion mobility during storage, and so on. After heating in chlorobenzene at 115 °C for 4 days the yield of extractable material decreased from 38 to 30 wt %, a 21% decrease relative to the extraction yield of the original coal, mirroring the changes observed by Larsen et al. It is well-known that methylol groups in phenolic resins (resols) can form cross-links at temperatures >100 °C and, given that coal has some alkyl OH groups, we were concerned that similar reactions could occur. However, the alkyl (and phenolic) OH group content of the starting and treated coals were within error the same, as measured by the acetylation/FTIR method.9 Furthermore, the acetylated starting coal gave an identical extraction yield to the parent sample (38%), while the acetylated sample subjected to heat treatment at 115 °C in chlorobenzene also displayed a marked reduction in the amount of pyridine (Soxhlet)-extractable material, this time to 28 wt % (see Table 1). We believe this result rules out any possible involvement (10) Snyder, R. W.; Painter, P. C.; Havens, J. R.; Koenig, J. L. Appl. Spectrosc. 1983, 37, 497.

Figure 1. FTIR spectra of a chlorobenzene-treated coal Pittsburgh No. 8 (a), an original coal sample (b), and the difference (c).

of phenolic and alkyl OH groups in cross-link formation and the decrease in extraction yields. Although, we did not expect to see any major changes, we decided to examine the infrared spectrum of the treated and original coal. The 1850-1450 cm-1 region of the spectra of the original and heat-treated coal are shown in Figure 1, which also shows a difference spectrum. Essentially, this difference spectrum shows a slight decrease in a broad band envelope between 1650 and 1735 cm-1, characteristic of carboxylic acid groups and perhaps some acid salts, and an increase due to a broad mode near 1600 cm-1, which overlaps the aromatic ring stretching mode of the coal, as a result of heat treatment in chlorobenzene. Carboxylate and acid salts have bands in the range 1500-1650 cm-1, the precise position of the bands depending on the nature of the counterion and the degree of hydration of the complexes, as we showed many years ago in studies of the ionomer, ethylene-co-methacrylic acid.11-14 This suggested to us that heat treating in chlorobenzene may mobilize ions, presumably from the mineral matter component of the coal, forming what are often called “clusters” in the ionomer literature. The strongly polar carboxylate salts essentially phase separate from the nonpolar part of the material to form ionic microdomains that act as cross-links or, more accurately, junction zones. Even though the concentration of car(11) Painter, P. C.; Brozoski, B. A.; Coleman, M. M. J. Polym. Sci., Polym. Phys. Ed. 1982, 20, 1069. (12) Brozoski, B. A.; Coleman, M. M.; Painter, P. C. Macromolecules 1984, 17, 230. (13) Brozoski, B. A.; Painter, P. C.; Coleman, M. M. Macromolecules 1984, 17, 1591. (14) Coleman, M. M.; Lee, J. Y.; Painter, P. C. Macromolecules 1990, 23, 2339.

Ionomers and the Structure of Coal

Figure 2. FTIR spectra of a chlorobenzene-treated coal sample (a), a chlorobenzene-treated coal after washing with acid (b), and the difference (c).

boxylic acid groups is often small (∼5%) in commercial resins, they have a profound effect on properties. The formation of ionic clusters is also accompanied by the appearance of an X-ray scattering peak corresponding to an average cluster spacing of ∼20 Å in size,9 an intriguing parallel given the X-ray diffraction results obtained by Larsen et al. that we mentioned in the Introduction to this paper.7 This hypothesis is easily checked. If ionic clusters are the cross-linking entity that is responsible for the decrease in swelling and extraction yields, their effect should be reversed by simply washing with 1 N HCl in THF. And it is, as also shown in Table 1. After washing the refluxed coal with 1 N HCl the amount of soluble material obtained increased from about 30% to 46%. Washing the original untreated coal with 1 N HCl in THF increased its extraction yield from 38% to 51%. Rewashing the acid-washed sample with 1 N NaOH (again in THF), which would form sodium salts, results in the extraction yield again decreasing, to 29%. The FTIR spectra of these samples confirm these conclusions, but also reveal some added complexity. First, if the spectrum of the chlorobenzene-treated coal is compared to that of the same sample after washing with 1 N HCL and a difference spectrum obtained, as shown in Figure 2, then it can be seen that modes near 1711 and 1650 cm-1 become more prominent upon acid washing (they appear below the baseline in the treatedacid washed difference spectrum). Second, if the original coal is washed with 1 N HCl, a -COOH band, now centered near 1723 cm-1, increases significantly in intensity, but the difference spectrum reveals an additional mode near 1540 cm-1, as shown in Figure 3. The appearance of the 1723 cm-1 band indicates that there are ionic clusters in the original coal which are

Energy & Fuels, Vol. 14, No. 5, 2000 1117

Figure 3. FTIR spectra of an acid-treated coal sample Pittsburgh No. 8 (a), an original coal sample (b), and the difference (c).

converted to acid groups upon washing. Furthermore, the appearance of the 1540 cm-1 band suggests that we also form some ionic clusters of a more specific character than those that give rise to the broad band near 1600 cm-1 in the chlorobenzene-soaked coal. We believe that this is probably a result of a complex equilibrium that would be set up between the mineral matter, coal functional groups, and the acid in the confined volume of the swollen coal particles. The strength of the ionic junction zone depends on how the metal ion is coordinated by the carboxylate groups, which in turn depends on the size of this counterion, the degree of hydration, and the presence of unexchanged acid groups, which can form acid salts with the cluster. Some band assignments we made in our previous studies of a zinc ionomer are shown in Figure 4, to give the interested reader a “feel” for the spectroscopy. Similar spectroscopic results are obtained for sodium and potassium ionomers, with the bands shifted and broadened somewhat, depending upon the structure and its degree of hydration. Consistent with this are the results we obtained by treating the acid-washed coal with 1 N NaOH, where there is a marked decrease in the 1720 cm-1 mode and the appearance of bands near 1580 and 1667 cm-1, as shown in Figure 5. The spectroscopy of these treated samples is complex and needs a more detailed study, but these preliminary results clearly support the conclusion that noncovalent cross-linking in this coal is a reversible process probably involving ionic clusters. Conclusions and Ramifications Given the recent results of Aida et al.,15 that coals have a much higher carboxylic acid content than previ(15) Aida, T.; Nishisu, A.; Yamanishi, I. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1999, 44 (3), 571.

1118

Energy & Fuels, Vol. 14, No. 5, 2000

Painter et al.

Figure 5. FTIR spectra of an alkali-treated coal sample Pittsburgh No. 8 (a), an acid-treated sample (b), and the difference (c).

Figure 4. Schematic representation of local structures of zinc ionomers: carboxylic acid “monomer” (a), carboxylic acid dimer (b), tetracoordinated zinc carboxylate (c), hexacoordinated zinc carboxylate (d), and zinc acid salt (e).

ously thought, we believe our conclusion that Pittsburgh No. 8 coal has an ionomeric structure may have broader ramifications. We would note that 1 mmol of COOH groups per gram of coal corresponds to 5 wt %, about what is found in ionomers such as surlyn. Even coals with 25%-50% of this value, corresponding to those with a carbon content of up to 87%, according to a plot

presented by Aida et al.,15 would be affected by the formation of ionic clusters. Accordingly, coals in general may have a higher content of pyridine-soluble material than previously thought. Also, unlike single hydrogen bonds, ionic complexes do have sufficient cohesive strength to act as cross-links, and it will prove interesting to investigate the extent to which this affects coal properties. Acknowledgment. The authors gratefully acknowledge the support of the Office of Chemical Sciences, U.S. Department of Energy, under Grant No. DE-FG0286ER13537. EF000092D