The following links will help you understand the effect of
varying the ratio of Al2O3 to Fe2O3 on liquid formation;-
[Link] (See page 22)
[Link]
(In particular the following paragraph;- "In a situation
where there is an AF ratio of 1.38, 2% MgO and 1.5%
clinker SO3 is it quite possible that the clinker may contain
well over 30% liquid well before the burning zone. In most
instances this will not be acceptable in terms of coating
tendency and potential ring formation within the kiln and a
move away from the traditional optimum alumina to iron
ratio will be necessary to optimise the kiln performance." )
To answer your other questions regarding AR and
viscosity;-
The main issue with clinker balling and kiln rings is not
liquid phase viscosity, but the sudden changes in liquid
phase quantity due to fluctuations in AR around a value of
1.38 at the point in the kiln corresponding to ~1338oC. The
amount of liquid phase present is at a maximum at this point
and is also very sensitive to changes in AR, whereas the
liquid phase viscosity is not increased significantly by just
moving from AR = 1.35 to 1.5. An extreme example of this is
white cement clinker, where the AR of the clinker is very
high (>10), yet this still produces a thin, stable coating with
good particle size distribution and almost never results in
clinker balling.
There also seems to be some differences of opinion between
various researchers and authors on the subject of liquid
phase viscosity and nodule formation.
In his book, "Cement Chemistry", Taylor claims that clinker
nodulisation is favoured by low liquid viscosity. (p74)
However, many other authors and articles suggest quite the
opposite. eg "Lea's Chemistry of Cement and Concrete", 4th
Edition p65, states that "lower viscosity flux may lead
to more dust being formed in the burning zone".
I agree with the view that, for a given burning temperature,
high C3A clinkers tend to nodulize better than low C3A
clinkers. However, better nodulisation is not the same as the
excessive nodulisation experienced with clinker balling,
which is almost invariably caused by too much liquid, too
early in the kiln. In addition, if the quantity of liquid phase
is also varying at this point (due to AR fluctuations
around the 1.38 point), this will only tend to further
promote excessive nodulisation and clinker balling.