MEASUREMENT OF -OH GROUPS IN COALS OF DIFFERENT RANK USING MICROWAVE METHODOLOGY - AND THE DEVELOPMENT OF QUANTITATIVE SOLID-STATE NMR METHODS FOR IN-SITU ANALYSIS
P. Monsefmirzai et al., MEASUREMENT OF -OH GROUPS IN COALS OF DIFFERENT RANK USING MICROWAVE METHODOLOGY - AND THE DEVELOPMENT OF QUANTITATIVE SOLID-STATE NMR METHODS FOR IN-SITU ANALYSIS, Fuel, 74(5), 1995, pp. 674-683
Experiments with both model compounds (substituted phenols) and with 1
1 coals (nine British and two American) have established that microwav
e heating will greatly accelerate silylation reactions of the phenolic
-OH groups, e.g. for Creswell coal complete silylation of -OH groups
occurs in 35 min in the microwave oven, whereas 24 h is required using
a bench reflux technique. Microwave reaction times for coals vary fro
m 35 min to 3 h for more dense coals such as Cortonwood. The above obs
ervations have allowed the development of a 'one pot' silylation of co
al, followed by an in situ analysis of the added Me(3)Si- groups by qu
antitative Si-29 magic angle spinning nuclear magnetic resonance (MAS
n.m.r.) spectroscopy. The development of a quantitative n.m.r. method
required the determination of Si-29 spin lattice relaxation times, T-1
, e.g. for silylated coals T-1 similar to 8 s; for silylated phenols,
T-1 similar to 25 s; for the synthetic smectite clay laponite, T-1 sim
ilar to 25 s; and for Ph(3)SiH, T-1 similar to 64 s. Inert laponite wa
s selected as the standard. The requirement to wait for five T-1max; b
etween pulses, together with the relatively low natural abundance of S
i-29 (4.71%), results in rather long accumulation times to obtain spec
tra of analytical quality (8-48 h). However, in comparison with other
methods, even in the most unfavourable case, the total time from comme
ncement of analysis to result may be described as 'rapid'. The results
for O-OH/O-total obtained are compared with other literature data. Co
mparison with ketene data, for example, shows agreement to vary from e
xcellent (Creswell) through satisfactory (Cortonwood) to poor (Pittsbu
rgh). Even in cases where agreement with ketene data is less good, the
silylation result may be close to estimates made via other acetylatio
n methods. Possible reasons for the variations observed are discussed.