YBa2Cu3Oz was the first high-temperature supraconductor in which the c
ritical temperature T-c has reached the liquid nitrogen temperature. T
he oxygen non-stoichiometry is limited to the CuOx basal planes, the c
omposition of which being related to z by z = 6 + 2 x. It is by now cl
ear that not only the oxygen content but also the ordering of the oxyg
en in the CuOx plane play an important role on the T-c in this compoun
d. YBa2Cu3Oz exhibits three different crystalline forms according to t
he oxygen content and the temperature : the disordered tetragonal phas
e (T) at high temperature and low oxygen concentrations, the ordered o
rthorhombic phase (O-\) centered on z = 7 (x = 0.5) and the double-cel
l orthorhombic (O-parallel to) centered on z = 6.5 (x = 0.25) at low t
emperature. Experimental studies of non-stoichiometric YBa2Cu3Oz have
been performed with a dual purpose : the preparation of samples with a
n accurate controlled oxygen content and the determination of thermody
namic properties as functions of temperature and composition. On the o
ther hand, statistical thermodynamic studies have been performed in or
der to account for the experimental data. Ten years ago, thermogravime
tric and microcalorimetric technics have been used to measure Delta G(
O-2) and Delta H(O-2), the relative partial molar free enthalpy and en
thalpy of oxygen, respectively. A statistical thermodynamic study base
d on Monte Carlo and CVM calculations has been performed in the framew
ork of an asymmetric two-dimensional lattice gas model with three inte
ractions between nearest and next-nearest neighbours (ASYNNNI model).
A set of effective pair interactions (EPI) obtained by first principle
calculations has been used. This model can account for the essential
features of the phase diagram, but large discrepancies have been obser
ved between experimental and calculated functions, for both Delta G(O-
2) and Delta H(O-2). In this context both experiments and the model ma
y be questioned. Therefore, recently, a complete study including exper
imental measurements and theoretical calculation has been performed ag
ain. With regard to the previous study; we have improved the experimen
tal set-up by coupling the thermobalance and the calorimeter on the sa
me gas-line. This new method ensures a real direct determination of De
lta H(O-2). On the other hand, heat treatments have been applied to hi
ghly improve the purity of the samples by minimising the quantity of s
econdary phases. The new results on Delta G(O-2) = f(z) are very close
d to the previous one but those on Delta H(O-2) = f(z) are very differ
ent close to the structural transition T - O-\. The presence of the se
condary phases is probably the main cause of this behaviour which is n
ot still completely understood and is being investigated at the presen
t time. As a consequence, large discrepancies between calculated and e
xperimental Delta G(O-2) are still observed, whereas the agreement bet
ween experimental and calculated Delta H(O-2) has been improved. So, t
he original model has been modified in order to take into account the
degrees of freedom of electronic origin which are generally neglected.
As a matter of fact, when an oxygen is added to the basal plane, the
number of electron holes is modified either as a change of the formal
copper valence or as the creation of oxygen 2p holes which are assumed
to hop in an extreme narrow band limit as shown by Schleger et al. Th
ese processes generate two additional entropy terms. In the same way,
we have distinguished several oxygen site energies according to the di
fferent cases. These modifications have allowed us to largely improve
the agreement between calculated and experimental partial molar functi
ons as well as for the phase diagram. In this paper, a new experimenta
l method for an accurate determination of the relative molar free enth
alpy and enthalpy of oxygen is presented We have shown that Delta H(O-
2) = f(z) is very sensitive to the samples purity and microstructure.
We also presented a modified lattice gas model for the oxygen ordering
thermodynamics in YBa2Cu3Oz. This modification which consists to taki
ng into account the electronic degrees of freedom, improves drasticall
y the agreement between calculations and experimental data.