CHAPTER 12 . Modelling of Natural Fluids: Are the Available Databases Adequate for this Purpose?
299
of minimizing the time needed for the experimental part is particularly risky in
polynuclear or multi-component systems, although extremely versatile computer
programmes will suggest a possible interpretation of the data. But the probability
that the interpretation might not be unique increases. We, conversely, believe that
the utmost care must be reserved to the acquisition of data, bearing a dowry of information on the system investigated that is as rich as possible. As pointed out earlier, good data may always be reinterpreted by more powerful methods of calculation, as the TDB project has shown. Bad data cannot become good data, no matter
how sophisticated the method of interpretation is. Bad data, since the first compilation of stability constants have always plagued the literature. However, we continue
to produce bad data, whereas the utmost care should be used to design the experimental approach of an investigation, in order to measure as many parameters as
possible with the highest accuracy. Time cannot be saved in this stage of the investigation. The amount of labour must instead be decreased by using fully automated
acquisition systems that operate around the clock. In our laboratory, we have developed one which is capable of a precision of 10 microvolts, in the course of coulometric
or volumetric titrations, performed into stainless steel air boxes, where the temperature is kept constant to within 0.02 DC, with the possibility of holding a free concentration (usually pH or pE) constant.
With the contributions of many solution chemists all over the world, EA has grown
into a precise methodology for the study of equilibrium reaction with solid bases of
thermodynamics. It is not obsolete because it thus far has not been replaced by any
approach of comparable efficiency. Yet its role within the established chemical sectors
is not well defined. It is not rare to find people who question the relevance of our work,
on the basis of the contradictory and discrepant results reported in the literature, practic ally on all the systems studied. The project of realizing a thermodynamic database
has even incited comments such as the following: "Metal speciation is the analytical
chemist's answer for eternal employment" (Grenthe and Puigdomenech 1997, p. 132).
However, the discipline EA is not to blame for this distrust. We think instead that to
some extent, we the users of EA deserve this sarcastic attitude for the reasons discussed
above. As pointed out earlier, most of the methods adopted in EA, because of their
indirect nature, only produce probabilistic results that have often been questioned.
There have been in the past decades people who did not believe in the existence of
complexes in solutions, and they ascribed the deviations from ideal behaviour to variations of the activity coefficients. The development of the X-ray diffraction technique
in concentrated solutions demonstrated the existence of complexes whose stoichiometric composition had been obtained by the methods of EA. Later on, other spectroscopic methods, such as NMR, Raman and, more recently, Mass Spectrometry and the
possibility to carry out measurements in relatively dilute solutions using light of Synchrotron, offer means to corroborate the results of EA by producing, sometimes, even
the structure of the complexes. But one must be careful, because witl10ut a prior knowledge of the speciation of the solution, spectroscopic data would be very difficult to
interpret, for instance in a polynuclear system. Thus, spectroscopic measurements can
provide essential information, particularly on the structure of the complexes, provided
the speciation of the solution is known, and they should be carried out every time this
is possible. In some way, a few scientific journals are already encouraging this trend
299
of minimizing the time needed for the experimental part is particularly risky in
polynuclear or multi-component systems, although extremely versatile computer
programmes will suggest a possible interpretation of the data. But the probability
that the interpretation might not be unique increases. We, conversely, believe that
the utmost care must be reserved to the acquisition of data, bearing a dowry of information on the system investigated that is as rich as possible. As pointed out earlier, good data may always be reinterpreted by more powerful methods of calculation, as the TDB project has shown. Bad data cannot become good data, no matter
how sophisticated the method of interpretation is. Bad data, since the first compilation of stability constants have always plagued the literature. However, we continue
to produce bad data, whereas the utmost care should be used to design the experimental approach of an investigation, in order to measure as many parameters as
possible with the highest accuracy. Time cannot be saved in this stage of the investigation. The amount of labour must instead be decreased by using fully automated
acquisition systems that operate around the clock. In our laboratory, we have developed one which is capable of a precision of 10 microvolts, in the course of coulometric
or volumetric titrations, performed into stainless steel air boxes, where the temperature is kept constant to within 0.02 DC, with the possibility of holding a free concentration (usually pH or pE) constant.
With the contributions of many solution chemists all over the world, EA has grown
into a precise methodology for the study of equilibrium reaction with solid bases of
thermodynamics. It is not obsolete because it thus far has not been replaced by any
approach of comparable efficiency. Yet its role within the established chemical sectors
is not well defined. It is not rare to find people who question the relevance of our work,
on the basis of the contradictory and discrepant results reported in the literature, practic ally on all the systems studied. The project of realizing a thermodynamic database
has even incited comments such as the following: "Metal speciation is the analytical
chemist's answer for eternal employment" (Grenthe and Puigdomenech 1997, p. 132).
However, the discipline EA is not to blame for this distrust. We think instead that to
some extent, we the users of EA deserve this sarcastic attitude for the reasons discussed
above. As pointed out earlier, most of the methods adopted in EA, because of their
indirect nature, only produce probabilistic results that have often been questioned.
There have been in the past decades people who did not believe in the existence of
complexes in solutions, and they ascribed the deviations from ideal behaviour to variations of the activity coefficients. The development of the X-ray diffraction technique
in concentrated solutions demonstrated the existence of complexes whose stoichiometric composition had been obtained by the methods of EA. Later on, other spectroscopic methods, such as NMR, Raman and, more recently, Mass Spectrometry and the
possibility to carry out measurements in relatively dilute solutions using light of Synchrotron, offer means to corroborate the results of EA by producing, sometimes, even
the structure of the complexes. But one must be careful, because witl10ut a prior knowledge of the speciation of the solution, spectroscopic data would be very difficult to
interpret, for instance in a polynuclear system. Thus, spectroscopic measurements can
provide essential information, particularly on the structure of the complexes, provided
the speciation of the solution is known, and they should be carried out every time this
is possible. In some way, a few scientific journals are already encouraging this trend