Energy & Fuels 1993, 7,479-481
479
13C Solid-state NMR Spin-Lattice Relaxation Time Measurements of Coals P. J. Barrie* Department of Chemistry, University College London, 20 Gordon Street, London WClH OAJ, U.K.
E. S . Madrali and R. Kandiyoti Department of Chemical Engineering, Imperial College, Prince Consort Road, London SW7 2BY, U.K. Received January 14, 1993. Revised Manuscript Received March 29, 1993
Spin-lattice (2'1) relaxation times of 13Cnuclei have been measured on eight different coals in order to determine the best experimental conditions for recording 13Csingle-pulse-excitation NMR spectra. It is found that a recycle delay of approximately 100 s is necessary to obtain quantitatively reliable spectra of coals using single-puhe-excitation methods. Introduction
13C solid-state NMR spectroscopy has now become a routine tool in the investigation of the chemical composition of coal and other fossil fuel materials.ls2 The most common technique used is cross-polarization (CP) with magic-angle spinning (MAS). However, it is widely known that the quantitative reliability of such spectra is highly dependent on the CP process, and experiments need to be performed at a variety of different mixing times in order to attempt to correct for the CP dynamics effects. The standard procedure for this correction is to fit the observed peak intensities as a function of mixing time to a model which assumes two rates of CP growth (corresponding to 13C with and without directly bonded hydrogen atoms) and one 2'lP relaxation rate.lP2 This is, however, an oversimplificationof the true situation in coals, which have a wide variety of different 13Cenvironments due to their inhomogeneous composition. In view of these problems a number of workers have suggested that single-pulseexcitation (SPE) methods, using high-power proton decoupling during the acquisition time only, give more reliable resulta.3-s This will only be the case, however, if a sufficiently long time is left between the scans for all 13C environmenta to have relaxed completely back to Boltzmann equilibrium. In general, recycle delays of at least 52'1 after a pulse flip angle of 90° are required, where 21' is the 13C spin-lattice relaxation time. Thus, a knowledge of l3C 2'1's in coals is crucial for recording SPE spectra. 13C SPE NMR studies of coals have used recycle delays anywhere between 20 and 200 5.8-8 There have so far been (1) Axelson, D. E.Solid State Nuclear Magnetic Resonance of Fossil Fuels; Multiscience Publications Ltd.:Canada, 1985. (2) Wilson,M.A.NMR Techniques and Applicationsin Geochemistry and Soil Chemistry; Pergamon Press: New York, 1987. (3) Dudley, R. E.; Fyfe, C. A. Fuel 1982,61, 651. (4)Botto, R.E.; Wileon, R.; Winma, R. E. Energy Fuels 1987,1,173. (6) Snape, C. E.; Axelson, D. E.; Botto, R. E.; Delpuech,J. J.; Tekely, P.; Gerstein, B. C.; Pruski, M.; Maciel, G.E.; Wilson, M. A. Fuel 1989, 68,547. (6) Muntean, J. V.; Stock, L. M. Energy Fuels 1991,5,765. (7) Zhang, M.; Maciel, G. E. Fuel 1990, 69, 557. (8)Franz, J. A.; Garcia, R.; Linehan, J. C.; Love, G.D.; Snape, C. E. Energy Fuels 1992,6,598.
only a few isolated examples of 21' measurements in coal^.^-^^ In this paper, we report spin-lattice relaxation times for eight coal samples, five of which form a series from the European Bank for coal specimens (SBN), one of which is a member of the Argonne Premium coal sample program, and two maceral concentrates. Experimental Section 13C solid-state NMR spectra were recorded on a Bruker MSL300 spectrometer at room temperature. Relaxation time measurements were obtained using the method of Torchia (the socalled TlCP method),I2which measures W spin-lattice relaxation times while taking advantage of the intensity enhancement obtained using cross polarization. The pulse sequence measures the decay from the enhanced CP signal back to that which would result from normal Bloch magnetization alone; residual signal from Bloch magnetization is eliminated by the phase cycling scheme employed. This method allows accurate T{s to be obtained quicker than alternative methods such as inversion recovery. It does,however,only detectthose carbonenvironments which cross polarize. Thus, it is possible that a small amount of carbon with no nearby protons is overlooked by this technique. Similarly, environments close to paramagnetic centers may not be observed, but these are likely to have fast relaxation times and so not affect the purpose of this work. Thirteen spectra were acquired in all experiments using a l-ms contact time with delay times of 50 pa, 10 ma, 200 me, 18, 3 8, 6 8, 10 s,17 s,25 s,40s,606,110 s, and 200 8. The longer delay times were felt to be particularly important as it is only these times which can detect any slowly relaxing components. The order in which the experiments were performed was randomized in order to avoid any systematic errors due to variation in spectrometer performance during the course of the measurement. Typically, 160 scans were acquired at each delay, which meant that the total accumulation time was approximately 21 h. Spinning sidebands were suppressed using the TOSS pulse (9) Miknis, F.P.; Sullivan, M. J.; Bartuska, V. J.; Maciel, G. E. Org. Geochem. 1981,3, 19. (10) Botto, R. E.;Axelson, D. E. R e p . Pap-Am. Chem. SOC.,Diu. Fuel Chem. 1988,33,50. (11) Pruski, M.; dela Roea, L.;Gerstein, B. C. Energy Fuels l990,4, 160. (12) Torchia, D. A. J. Magn. Reson. 1978,30,613.
0887-0624/93/2507-0479$04.00/00 1993 American Chemical Society
Barrie et al.
480 Energy & Fuels, Vol. 7, No.4, 1993 Table I. Elemental Analyeie of Coal Samples sample % asha % Cb % Hb 4.40 Thanstone 7.2 91.4 4.48 Emil Mayrisch 6.6 89.7 4.90 Heinrich Robert 5.3 87.4 5.20 5.3 81.5 Linby 4.89 Illinois No. 6 SBN 12.9 73.3 5.00 Illinois No. 6 APCS 15.5 77.7 Point of Ayr vitrinite 2.3 84.9 5.00 5.00 DNC 50% inertinite 15.6 85.2 a Dry basis. Dry ash-free basis.
*
-
110s
Delay time / s
4
17s
100 0 -100 PPm Figure 1. 18C TlCP spectra of coal sample DNC 50%-I, with the relaxation delays indicated.
300
200
sequence of Dixon et al.;l*while TOSS spectra are not necessarily fully quantitative, this error is independent of relaxation delay and thus has no effect on the measured relaxation times. During processing a line-broadening factor of 125 Hz was applied in order to reduce noise to an acceptable level. The aromatic and aliphatic peak areas were then integrated, and plotted as a function of relaxation delay. The resulting curves were fitted to a biexponential decay by a least-squares fitting procedure. The measurements were carried out on a set of rank ordered coale: Tilmanstone, Emil Mayrisch, Heinrich Robert, Linby, and Illinois No. 6 SBN. These were selected from a set presently being examined as part of an ongoing project funded by the European Community to investigate relationships between the fundamental properties of coals and their pyrolytic behavior." Linby was included as the standard sample for Imperial College, London. Illinois No. 6 from the Argonne Premium coal sample program (APCS)l6was also studied-we have previously found significant differences in yields of tar and volatile8 on pyrolysis between the distinct Illinois No. 6 samples (SBN and APCS).16 The two other coals studied were maceral-enriched concentrates: a Point of Ayr vitrinite concentrate, and Durban Navigation (13)Dixon,W. T.; Schaefer, J.; Sefcik, M. D.;Stejskal,E. 0.;McKay, R. A. J. Magn. Reeon. 1982,49,341. (14) Li, C.4.; Madrali, E. S.; Wu, F.; Xu, B.; Cai, H.-Y.; Guell, A. J.; Kandiyoti, R., submitted for publication in Fuel. (15)Vorres, K. 5.Energy Fuels 1990,4, 420. (16)Cai, H.-Y.; Guell, A. J.; Dugwell, D.R.; Kandiyoti, R. Fuel, in preas.
Figure2. Experimental points and simulation of the relaxation of the aromatic peak of coal sample DNC 50% inertinite Table 11. Summary of Meaeured Relaxation Times, Tp aromatic peak aliphatic peak sample TlIs fraction Tl/s fraction t&s Tilmanstone 3.2 0.45 1.8 0.57 23.7 0.55 19.5 0.43 95 Emil Mayrisch 3.4 0.35 1.7 0.46 20.6 0.65 16.8 0.54 86 Heinrich Robert 3.6 0.30 0.88 0.45 25.0 0.70 16.6 0.55 106 Linby 2.8 0.40 0.61 0.44 20.3 0.60 10.9 0.56 83 Illinois No. 6 SBN 1.8 0.44 0.61 0.51 17.8 0.56 11.0 0.49 72 Illinois No. 6 APCS 2.0 0.27 0.50 0.39 22.3 0.73 12.9 0.61 96 Point of Ayr vitrinite 3.5 0.25 0.67 0.52 23.1 0.75 13.6 0.48 100 DNC 50% inertinite 4.8 0.36 0.66 0.44 33.6 0.64 17.7 0.56 140 a The &, column gives the calculated minimum recycle delay necessary in SPE experiments to allow relaxation to within 1 % of the Boltzmann equilibrium value. Colliery 50% inertinite concentrate (DNC 50% -I, South Africa). Elemental analysis of the eight samples studied is shown in Table I. All the samples were stored under nitrogen, but no precautions were taken to exclude air/oxygen during sample packing into the NMR rotor, and dry air was used as the spinning gas. The samples were not treated in any way to remove paramagnetic impurities. Methods to do this have been suggested in the literature as lsC environments in close proximity to unpaired electrons may give broad NMR resonances that might not be observed.6~6 In this study it was felt, however, that any paramagnetic material would benefit the system as it would have the effect of reducing relaxation times. Thus TI'Sin samples which have been treated to remove paramagnetic species are expected to be as long, or possibly longer, than the relaxation times observed here.
Results and Discussion Figure 1shows six of the spectra obtained for one of the coal samples. The full signal intensity is present at short delay times. At longer delay times there is t h e complete loss of aliphatic signal (usually after about 40 s), while there remains detectable aromatic intensity even after a delay of 110s in the case of the DNC 50% -Isample (Figure 1). None of t h e coals studiedshowed any detectable signal after a 200-5 delay for relaxation. Following integration of the aromatic and aliphatic peak areas, the resulting
Spin-Lattice Relaxation Time of Coals curves were fitted to biexponential decays. The fit for the DNC 50%-I sample is shown in Figure 2, and results for all samples are presented in Table 11. The relaxation of the aromatic peak for all samples investigated is best fitted by a significant proportion of fairly fast relaxing sites (with T