Professor Roger Powell and Dr. Tim
Featured Scientist Interview
Reuters, the paper, "An internally consistent
thermodynamic data set for phases of petrological
interest," (Holland TJB and Powell R, J.
Metamorph. Geol. 16: 309-43, May 1998), ranks at
#6 among Geosciences Highly Cited Papers, with 911
The papers authors are Dr. Tim Holland and Professor Roger Powell. Dr.
Holland is a Reader in Petrology in the Department of Earth Sciences at the
University of Cambridge. Professor Powell is a Professorial Fellow in the
School of Earth Sciences at the University of Melbourne.
In the interview below,
ScienceWatch.com talks with Professor Powell about this
Would you please sum up your 1998 Journal of
Metamorphic Geology paper, "An internally consistent thermodynamic
data set for phases of petrological interest," for our readers?
First to say that the "I" here is Roger Powell, and I am speaking for both
myself and my friend and collaborator, Tim Holland. The work, of which this
is the most recent published part, is the result of more than 25 years of
ongoing research collaboration, most of which has taken place around the
world, with me at the University of Melbourne, Australia, and Tim at the
University of Cambridge in the UK.
What this paper provides is a dataset of thermodynamic properties of the
constituents (end-members) of the minerals, fluids, and melt that are
needed in order to perform calculations on the conditions of formation of
rocks. It is an internally consistent dataset because all of the available
information has been appraised and combined to establish consistency. This
is done statistically (in a least squares sense), giving uncertainties and
correlations on the dataset. This then allows calculations with the dataset
to be done such that uncertainties on the results can be obtained, an
important aim in any science.
This data set is an update of a prior data set. How do
the sets differ from each other?
The research started in 1982, and the first dataset papers came out in
1985, explaining the philosophy and providing the first actual dataset. The
first published update was in 1990, the second in 1998. With each of these
updates, the evolution is firstly in the methodology to process and
represent the data, particularly the pressure and temperature dependence of
the thermodynamic properties. Secondly it is in the scope and quality of
the data themselves. The dataset is calculated from a large body of
experimental data (on individual minerals and reactions between them) that
encode the thermodynamic properties, and data from new experiments continue
to be published. As new data become available the dataset changes and
Has the data been updated since your 1998 paper, or are
there plans to do so?
Although the 1998 paper is the second published upgrade of the dataset, we
have continued to release the current version. The dataset that is
currently being used is the fifth update of the one in that paper. In the
same way our software that uses the dataset, THERMOCALC, was presented in a
1988 paper, and updated in a 1998 paper, but the software has evolved more
or less continuously through to the present. The software and the dataset
are free downloads on the web.
The third update of the dataset, which is near completion, should be
submitted for publication later this year (2008), also in the Journal
of Metamorphic Geology.
Who uses this data set, and for what tasks?
Petrologists, mainly metamorphic petrologists working on the rocks that
form in the bowels of ancient mountain belts, use the dataset primarily to
determine the pressure and temperature of formation of rocks now found at
the Earth's surface. This information gives critical clues to the processes
that form and evolve mountain belts. This is a critical part of geology, as
much of the basement of the continents is the stitched-together remains of
such ancient mountain belts.
What initially sparked your interest in this particular
"As new data become available the
dataset changes and hopefully
Both Tim and I were supervised for our D.Phil. theses at the University of
Oxford by a brilliant and far-sighted geologist, Steve Richardson, who saw
that the application of equilibrium thermodynamics to rocks could be a
supremely powerful tool. He pushed us in that direction. I produced a very
rudimentary thermodynamic dataset for use in my thesis work in the early
'70s, and when Tim started his work soon after I had left Oxford, he took
up a similar approach. The meeting of minds that started our collaboration
was the result of Tim, who was keen to take the dataset idea further,
wanting to have someone involved who was interested in the
maths/stats/computing side of the research, someone who also could see the
relevance geologically. The someone was me.
What should the "take-away lesson" about your work be
The take-away lesson is that it would be very difficult for work of this
scope to be done by one person: it requires a collaborative effort, and the
effort ideally needs to involve people with different abilities and skills.
I have brought to this in some sense the simpler part, the algorithmic and
program design, and the writing and upgrading of general software to
process the available data to make the dataset (as well as the software
THERMOCALC to actually use the dataset). Tim has done the actual appraisal
and assembly of all the disparate data and information to run through the
software to make the dataset itself.
Professor Roger Powell
School of Earth Sciences
University of Melbourne
Dr. Tim Holland
Department of Earth Sciences
University of Cambridge
Cambridge, United Kingdom
Keywords: thermodynamic properties, mineral constituents,
fluid constituents, melt constituents, rock formation, data set
updates, THERMOCALC software, mountain belts,