Macromolecular structure of coals. 9. Molecular structure and glass

---

Energy & Fuels 1987,1, 56-58

56

Macromolecular Structure of Coals. 9. Molecular Structure and Glass Transition Temperature Lucy M. Lucht,? John M. Lamon,$ and Nikolaos A. Peppas* School of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907 Received August 18, 1986. Revised Manuscript Received October 10, 1986

The glass transition temperatures, Tg, of coal samples with a wide range of carbon content were determined by differential scanning calorimetry of untreated and pyridine-extracted coal particles. Their values varied from 580 to 632 K depending on the coal structure studied. During dynamic pyridine transport in coal samples, the value of Tgdecreased as a function of solvent weight fraction. These results indicate a coal network structure behaving similarly to that of cross-linked synthetic polymer networks.

Introduction The chemical structure of coal consists of aromatic and hydroaromatic condensed ring clusters linked by etheric, methylenic, and/or sulfidic bonds. The physicochemical structure of coal, the size of each macromolecule thus formed, and the forces holding one to another are not known. However, there is evidence that coal behaves as a cross-linked macromolecular structure, where the cross-links could be due to covalent bonds, hydrogen bonds, or entanglements between the macromo1ecules.l" The fact that coals swell in thermodynamically compatible solvents without complete dissolution suggests that cross-links may be extensive enough to form cross-linked network structures. A network structure would account for many aspects of coal behavior, including its tendency to swell in organic vapor and its limited solubility except under thermally degradative conditions. Macromolecular structures are characterized in part by their glass transition temperature, T , i.e. temperature of transition from the glassy to the rubbery state. In the glassy state, macromolecular mobility is impeded and only short segmental motion may be observed. In the rubbery state, the macromolecular chains exhibit a significant increase in their mobility.6 Various macromolecular structural properties such as the elastic modulus, diffusivity, and dielectric properties pass through an abrupt change at the glass transition temperature of the macromolecular system. Thus, property changes by 2-4 orders of magnitude may be detected. Glassy macromolecular network structures are transformed into rubbery materials by an increase of the experimentation temperature above Tr A similar effect is observed when a thermodynamically compatible penetrant enters the macromolecular structure, thus leading to increased macromolecular mobility and significant decrease of the

TB'

If the basic premise of a cross-linked network structure in coals is correct, then coal samples should show similar behavior as polymer networks. Thus, they should exhibit a glass transition temperature. Also, a substantial reduction of its value should be observed in the presence of swelling agents. Such information would be important in analysis of solvent (penetrant) transport at high temperatures and could lead to important information concerning

* Author to whom correspondence should be addressed. 'Present address: Lawrence Livermore Laboratory, Livermore, CA 94550.

*Present address: US Naval Base, Norfolk, VA. 0887-0624/87/2501-0056$01.50/0

the design and operation of liquefaction processes. In fact, some preliminary evaluation of the glass transition of coals has been reviewed re~ently.~ Several years ago7we offered a preliminary analysis of the glassy behavior of coals. Here we present a more complete evaluation of the characteristics of the glassy state.

Experimental Section Coal particles of 20-30 mesh were obtained from coal samples of The Pennsylvania State University coal bank. Some of the samples were used as received while others were first floated in a benzene and carbon tetrachloride solution of density 1.3 g/cm3, Soxhlet-extracted with pyridine at its atmospheric boiling point (115.5 "C), dried, and resieved to 20-30 mesh. Henceforth, these latter samples are referred to as the "pyridine-extracted" samples. Unexpected coal samples of approximately 1 g were placed in Coors crucibles. The crucibles were placed in a sealed dessicator, which contained a layer of liquid pyridine in the bottom. At specified time intervals, the crucibles were removed and weighed to determine the amount of penetrant uptake. At the same time, small portions of each sample were removed, weighed, and analyzed by differential scanning calorimetry (DSC-2, Perkin Elmer, Norwalk, CT). The crucibles were also reweighed upon removal of the samples for thermal analysis to ensure accurate penetrant swelling data. This procedure was continued for 1-4 weeks for each coal, until no further appreciable weight gain was observed. During these studies, no pyridine condensation in the crucibles was observed. Because previous work had indicated that thermal measurements depended on sample weight,8 consistent sample weights were used in the DSC analyses. The coal samples were promptly sealed in the DSC sample chamber and purged with nitrogen for 30 min to provide an inert atmosphere. The samples were heated at the rate of 10 OC/min, and the constant-pressure heat capacity curve was obtained as a function of temperature and time. Data

(1)Peppas, N. A.; Lucht, L. M. Chem. Eng. Commun. 1984, 30, 291-310. (2)Lucht, L. M.;Peppas, N. A. Prepr. Pap.-Am. Chem. SOC.Diu. Fuel. Chem. 1984,29(1),213-217. Kovac, J.; Brenner, D.; Larsen, J. W. In Coal Structure; (3)Green, T.; Meyers, R. A,, Ed.; Academic: New York, 1982;pp 199-282. (4)Larsen, J. W. In Organic Chemistry of Coal; Larsen, J. W., Ed.; ACS Symposium Series 71; American Chemical Society: Washington, DC, 1978;-pp 36-53. (5) Larsen, J.W. In Chemistry and Physics of Coal Utilization; Coouer, B. R., Petrakis, L., Eds.; AIP: New York, 1981;AIP Conference Proceedings Vol. 70, pp 2-27. (6) Haward, R. N. The Physics of Glassy Polymers; Wiley: New York, 1973. (7)Lucht, L. M.; Larson, J. M.; Peppas, N. A. Bull. Am. Phys. SOC. 1983,28,566. (8)Peppas, N. A. J. Appl. Polym. Sci. 1976,20, 1715-1716.

0 1987 American Chemical Society

Energy & Fuels, Vol. 1, No. 1, 1987 57

Macromolecular Structure of Coals

Table 111. Thermal Analysis Data of Dynamically Swelling, Untreated 20-30-mesh Coal Particles of PSOC-212 with 79.3% C (dmmf) Exposed to Pyridine at 35 OC pyridine uptake glass trans M p / M c ,g of sample wt for swelling time t , h py/g of coal DSC m,, mg temp T,,K 0.0 0.000 5.1 607 4.2 0.072 5.3 502 21.2 0.219 5.6 424 90.9 0.654 6.9 410 189.0 0.732 5.1 395 I 620

I

I 640

I

I 660

I

I 680

Temperature, T P K I

Figure 1. Typical DSC thermogram of a coal sample of (2-384.

PSO-

Table I. Thermal Analysis Data of Dynamically Swelling, Unextracted 20-30-mesh Coal Particles of PSOC-418 with 69.9% C (dmmf) Exposed to Pyridine at 35 O C pyridine uptake M p / M c ,g of sample w t for glass trans swelling DSC m,,mg temp TB, K py/g of coal time t , h 6.1 580 0.0 0.000 4.7 498 40.3 0.265 0.339 4.8 476 95.3 5.4 439 382.0 0.461 5.1 423 664.0 0.516 Table 11. Thermal Analysis Data of Dynamically Swelling, Untreated 20-30-mesh Coal Particles of PSOC-416 with 74.7% C (dmmf) Exposed to Pyridine at 35 OC pyridine uptake M p / M , , of sample wt for glass trans swelling temp Tg, K time t , h py/g of coal DSC m,,mg 604 0.0 0.000 5.9 485 44.8 0.140 4.4 446 93.8 0.263 4.3 436 359.0 0.345 5.7 406 662.0 0.414 5.7

were obtained from 320 to 650 K.

Results and Discussion At a scanning rate of 10 OC/min and under continuous purge of nitrogen, minimal thermal degradation of the coal samples was observed as a function of temperature. A slight, continuous increase of the heat capacity was noted as a function of temperature until a temperature range was reached where a significant shift of the thermogram was observed. In Figure 1 we present a typical thermogram obtained with coal particles of PSOC-384 with 94.2% C on a dmmf basis. The slope shift (A) was identified an4 its onset temperature was reported as the initial glass of the coal network structure. transition temperature, Tg, The exact value of Tgwas best identified from the derivative curve of the thermogram of Figure 1. Studies with unextracted coal samples indicated values of Tgranging from 580 K for a coal with 69.9% C (dmmf) to 632 K for a coal with 94.2% C (dmmf) as seen in Tables I-VI for the samples corresponding to a swelling time of t = 0 min. The reproducibility of these values was rather good (*8 K) considering the heterogeneity of the samples studied. To minimize any effects from sample weight and scanning speed, all studies where performed with samples weighing less than 8 mg and at a constant scanning speed of 10 OC/min. A slight trend toward increased values of Tgas a function of coal carbon content was observed in unextracted coal networks. This is an indication of an increased degree of cross-linking, and therefore of de-

Table IV. Thermal Analysis Data of Dynamically Swelling, Untreated 20-30-mesh Coal Particles of PSOC-772 with 79.8% C (dmmf) Exposed to Pyridine at 35 "C pyridine uptake glass trans M p / M , , g of sample wt for swelling DSC m,,mg temp Tg, K time t , h py/g of coal 0.0 0.000 5.3 602 6.0 0.068 5.8 494 23.3 0.196 5.5 492 92.2 0.695 5.1 405 190.0 0.819 3.6 423 Table V. Thermal AEalysis Data of Dynamically Swelling, Untreated 20-30-mesh Coal Particles of PSOC-341 with 86.0% C (dmmf) Exposed to Pyridine at 35 OC pyridine uptake glass trans M p / M , , g of sample wt for swelling temp Tg, K py/g of coal DSC m,, mg time t , h 3.3 600 0.0 0.000 0.096 7.0 502 42.8 142.0 0.377 7.3 491 8.3 413 383.0 0.611 0.650 8.2 401 665.0 Table VI. Thermal Analysis Data of Dynamically Swelling Untreated 20-30-mesh Coal Particles of PSOC-384 with 94.2% C (dmmf) Exposed to Pyridine at 35 OC pyridine uptake swelling M p / M , , g of sample wt for glass trans time t , h py/g of coal DSC m,, mg temp Tg, K 0.0 0.000 5.2 632 26.1 0.024 8.4 547 82.7 0.042 6.7 516

creased chain mobility, as the carbon content increased. We do not wish to draw any firm conclusions on this point due to the heterogeneity of unextracted coal samples, but we note this interesting observation. The possibility of preextraction of coal samples with pyridine as a means of preparation of more uniform samples was examined. In all cases the pyridine-extracted samples showed a slightly lower value of Tgwith respect to the unextracted samples. This observation was an indication that a small quantity of pyridine remained permanently sorbed in the coal structures and lowered their Tgvalues. For example, whereas an unextracted PSOC-384 samples gave Tg= 632 K, a pyridine-extracted and redried sample gave Tg= 609 K. For this reason, all subsequent studies were performed only on unextracted coal samples. The mineral matter of these coals did not interfere in the determination of Tg, since any important thermal transitions of the mineral matter occur at temperatures higher than the Tr An additional reason for the glass transition decrease after preextraction may be some cleavage of hydrogen bonds due to pyridine absorption. Figure 1and similar thermograms for other coal samples indicated a clear first-order thermodynamic peak (B)which is probably indicative of major coal degradation. In no case was this peak observed before the baseline shift A, an indication that major thermal degradation of coals oc-

Lucht et al.

58 Energy & Fuels, Vol. 1, No. 1, 1987

Mp/Mc = (Mp/Mc)+ 1 The general correlative curve shown in Figure 2 indicates that all coals achieved a long-term equilibrium swelling value of about w = 0.45 (or 45% pyridine), with an approximate "equilibrium glass transition" of 400 K as shown by eq 2.

T g = 6121 - 9 9 8 . 5 ~+~1260~:

0 10 020 030 040 SOLVENT WEIGHT FRACTION

Figure 2. Dependence of coal glass transition temperature, Tg (in K), on the penetrant weight fraction, up.

curred only in the rubbery state and above Tr Finally, a small peak was observed around 373 K, an indication of moisture evaporation. To study the effect of pyridine transport on the glass transition temperature, 20-30-mesh particles of six coals were swollen by pyridine vapor at 35 "C for long periods of times, usually 30 days. Such a process would lead to an increase of macromolecular m ~ b i l i t y . ~ In J ~all cases, the glass transition temperature of the partially swollen samples was recorded as a function of penetrant uptake, in g of pyridinelg of coal and reported in Tables I-VI. Examination of the data of Table I indicates that there was a very large decrease of Tgfor coal PSOC-418 (with 69.9% C) in the early portion of the swelling process, e.g. from 580 to 498 K, since the pyridine uptake increased to 0.265 g/g of coal (48% of its equilibrium value). The value of Tgcontinued decreasing until thermocynamic equilibrium was achieved at 0.516 g of pyridinelg of coal after 664 h of swelling at 35 "C. At that point, the value of T g was 423 K, i.e. 150 "C vs. an experimental swelling temperature of 35 "C. Similar results were obtained with the other coal samples. The mechanistic implications of this significant decrease of Tg,especially when swelling samples at temperatures around 100 "C, has been discussed in other publications by BrennergJoand us.l1 The proximity of glass transition and swelling temperatures favors macromolecular relaxations during swelling+'l and leads to non-Fickian pyridine transport much in the same way as for transport in glassy polymers.l'J2 A general dependence of the glass transition temperature on the penetrant (solvent) weight fraction was established for all coals as shown in Figure 2. This graph is very similar to those obtained for penetrant effects on the T g of glassy crosslinked polymers.6 Transformation of the pyridine uptake data, Mp/Mc,to pyridine weight fractions, up, was done by the simple eq 1. (9) Brenner, D. Fuel 1984,63, 1324-1327. (10) Brenner, D. Fuel 1985,64, 167-172. (11) Peppas, N. A.; Lucht, L. M. Chem. Eng. Commun. 1985, 37, 334-354. (12) Hsieh, S. T.;Duda, J. L. Polym. Mater. Sci. Eng. 1984, 51, 703-706.

(2)

To achieve lower T values in the presence of a penetrant, one must use a tLermodynamical1y better "solvent". This can be proven by calculating the theroretically required amount of pyridine, M */Mc*,in order for the coal network to become barely rubbery at the experimentation temperature. This calculation can be done by the following equation, proposed by Ferry,13which is applied here to coal by using approximate values: Mp*

Tg - Texptl

Mc*

P/af

-- -

(3)

Here, the threshold concentration of pyridine in the coal, Mp*/Mc*,for the rubbery transition to occur at the experimentation temperature, Texptl= 308 K, is related to the glass transition temperature of the penetrant-free coal network, e.g. T g = 600 K for PSOC-341, the free-volume constant of the penetrant-free coal network, af,and the free-volume contribution constant of the pyridine in coal, p. An essentially constant value of af 3.7 X K can be used as determined by Ferry13 for most polymers, whereas p was approximated as 0.15 by Fujita and Kishimoto.14 Thus, for PSOC-341 the value of Mp*/Mc*was calculated as 0.725 g of pyridinelg of coal. Clearly, at least 0.125 g of pyridinelg of coal are required for the glass transition of PSOC 341 to occur at 35 "C, whereas the maximum (equilibrium) value attained by this coal was only 0.650 g of pyridinelg of coal. Similar calculations can be performed for the other coal samples. In all cases, the threshold concentration, Mp*/ Mc*, is higher than the maximum (equilibrium) value attained.

Conclusions In conclusion, coal networks exhibit a glass transition temperature that seems to be somewhat dependent on carbon content and is a strong function of the pyridine uptake in the coal structure. This analysis is instrumental in our understanding of the significant increase of macromolecular mobility as the coal structure is exposed to thermodynamically compatibility solvents.

Acknowledgment. This work was supported by Department of Energy Grants DE-FG22-78ET13379 and DE-FG22-80PC30222. (13) Ferry, J. D. Viscoelastic Properties of Polymers; Wiley: New York, 1980. (14) Fujita, H.; Kishimoto, A. J. Polym. Sci. 1958,28, 547-569.