68934591-19_en

ENTHALPY Final Summary Report FINAL SUMMARY REPORT EUROPEAN NUCLEAR THERMODYNAMIC DATABASE (ENTHALPY Project) CO ORDINATOR Dr A DE BREMAECKER Detached from SCK CEN Mol at IRSN/DRS/SEMAR CEN de Cadarac[.]

FINAL SUMMARY REPORT

EUROPEAN NUCLEAR THERMODYNAMIC

DATABASE (ENTHALPY Project)

CO-ORDINATOR

Dr A DE BREMAECKER

Detached from SCK.CEN-Mol at IRSN/DRS/SEMAR

CEN de Cadarache Bât 702

3 AEA-Technology,Harwell, United Kingdom

4 THERMODATA, Grenoble, France

5 FRAMATOME-ANP, Erlangen, Germany

6 CEA-DRN-DTP, Cadarache, France

7 EdF, Clamart, France

8 AEKI-KFKI, Budapest, Hungary

9 SKODA-UJP, Praha, Czech republic

10 SCK.CEN, Mol, Belgium

11 ULB, Université Libre de Bruxelles, Brussels, Belgium

12 UCL, Université Catholique de Louvain, Louvain-la-Neuve, Belgium

CONTRACT N°: FI3S-CT1999-00001

B.1 Assembling of the two existing databases & extension to new elements

B.2 Separate Effect tests

B.3 Improvement and validation of the database

B.4 Evaluation of the consequences of the uncertainties

B.5 Methodologies of coupling the database with SA codes

B.6 Edition of the database

C WORK PERFORMED AND RESULTS

C.1 State of the Art Report

C.2 Assembling and extension of the database (WP 1)

C.2.1 Introduction

C.2.1.1 Selection of the elements to be included in NTD

C.2.1.2 Modelling and merging procedure

C.2.1.3 Selection of the systems to be included in NTD

C.2.1.4 Critical assessment work

C.2.2 Critical assessment and assembling (Task 1.1)

C.2.3 Extension of the database (Task 1.2)

C.2.3.1 Extension of NTDIV01 to Boron and Carbon and assembling of the final In-Vessel Nuclear Thermodynamic database : NTDIV02

C.2.3.2 Extension of NTDIV02 to concrete elements (Al, Ca, Mg, Si)

Assembling of the final In/ex Vessel Nuclear Thermodynamic

Database : NUCLEA

C.2.3.3 Comparison calculations

C.2.4 State of the validation

C.3 Separate Effect Tests (WP 2)

C.3.1 Solidus – Liquidus temperatures for (U,Zr)O2+x (Task 2.1)

C.3.4 Tliq in UO2-ZrO2-FeOx-CaO-Al2O3-SiO2 systems (Task 2.4)

C.3.5 Experimental study of (sub)system(s) in UO2-ZrO2-BaO-MoOx (Task 2.5)

C.3.5.1 Direct determination of the phase diagram (Task 2.5.2.1)

C.3.5.1.1 Experimental

C.3.5.1.2 Results in the ternary system BaO-ZrO2-MoO3

C.3.5.1.3 Results in the quaternary system UO2-BaO-ZrO2-MoO3 in reducing atmosphere at 1480°C and 1600°C

C.3.5.1.4 Results in the quaternary system UO2-BaO-ZrO2-MoO3 in oxidising atmosphere at 1600°C

C.3.5.3.1 Solid solubility and precipitation of FP Molybdenum

C.3.5.3.2 Solid solubility and precipitation of FP Baryum and Zirconium

C.3.5.3.3 Post-irradiation observations

C.3.5.3.4 Review of equilibria established at the Mo/MoOé equilibrium

C.3.5.3.5 Conclusions

C.4 Validation and Improvement of the database (WP 3)

C.4.1 Validation based on fuel experiments and empirical models (Task 3.1)

C.4.2 Validation based on experiments (task 3.2)

C.4.3 Validation based on VULCANO/CEA and COMETA/NRI experiments

(Task 3.3)

C.4.3.1 VULCANO VE-U3 experiment

C.4.3.2 VULCANO VE-U7

C.4.3.3 COMETA / NRI test

C.4.4 Improvement by literature review (task 3.4)

C.4.4.1 Boiling points and vapour pressures of the elements

C.4.4.2 Thermodynamic properties of UxOy gaseous species

C.4.5 Validation based on Tliq and Tsol measurements in UO2+x-ZrO2 (AEA-T) and Fe-Zr-O (SKODA), and global experiments (VERCORS, Hofmann) (Task 3.5)

C.4.5.1 UO2+x-ZrO liquidus and solidus measurements

C.4.5.2 Fe-Zr-O liquidus and solidus measurements

C.4.5.3 Fission Products release (Vercors tests)

C.4.5.4 Corium pool stratification involving Fe-O-U-Zr (+ Cr, Ni)

C.5 Influence of uncertainties

C.6 Coupling methodologies to SA codes

C.6.1 Coupling methodology of thermodynamic databases to SA codes (CROCO and

ICARE/CATHARE) (Task 5.1)C.6.1.1 Tabulation of the phase diagram

C.6.1.2 Interface with the thermochemical code

C.6.1.3 Generation of the library, research and interpolation in the library

C.6.1.4 Applications

C.6.1.5 Time consuming and coupling aspects

C.6.2 Recalculation of TMI2 with MAAP4 and the database (Task 5.2)

C.6.2.1 Thermo-chemistry of the U-Zr-O domain

C.6.2.2 Recalculation of the U-Zr-O diagram using the NUCLEA databaseC.6.2.3 Recalculation of TMI-2 using the new Tliq and Tsol

C.6.3 Simplification of the database and/or of the equilibrium code and adaptation to

SA codes (Task 5.3)C.7 Edition of the NUCLEA database

REFERENCES

TABLES

FIGURES

LIST OF ABBREVIATIONS & SYMBOLS

CIT Corium Interactions and Thermochemistry

COLOSS Core Loss

DTA Differential Thermal Analysis

EDX Energy Dispersive Microanalysis

EPMA Electron Probe Micro Analysis

FP Fission Product

ICP Inductively Coupled Plasma

LOCA Loss of Coolant Accident

MDB Material Data Bank ( module of the ASTEC System Code)

MCCI Molten Corium-Concrete Interaction

NTD Nuclear Thermodynamic Database

SA Severe Accident

SEM Scanning Electron Microscopy

SGTE Scientific Group Thermodata Europe

TDBCR-IV Thermodynamic Data Base Corium - In Vessel

THMO Thermo-chemical Modeling and data

TMI-2 Three Mile Island Unit 2

VPA Visual Polythermal Analysis

XRD X-Ray Diffraction

EXECUTIVE SUMMARY

ENTHALPY is a shared-cost action where 11 partners from 6 countries succeeded to pro- duce one unique well validated thermodynamic database for in- & ex-vessel applications

The first step was the critical assessment and merging of the two thermodynamic

databases (THERMODATA/Grenoble and AEA-Technology/Harwell) in one database called

"NTDiv" (for Nuclear Thermodynamic Database in-vessel) The second step was the

extension of the database to Boron and Carbon, leading to the edition of the NTDiv0201 version (14 elements) In a third step, four ex-vessel elements and four new oxides

components were added All together, the final database is based on 20 elements :

O-U-Zr-Ag-In-Fe-Cr-Ni-Ba-La-Ru-Sr-B-C-Al-Ca-Mg-Si + Ar-H

and includes in particular the 15 oxides system :

UO2-ZrO2-FeO-Fe2O3-Cr2O3-NiO-In2O3-BaO-La2O3-SrO-B2O3-Al2O3-CaO-MgO-SiO2

In view of the extension to Boron, AEKI determined experimentally 3 new phase

diagrams namely ZrO2-FeOx, ZrO2-B2O3 and B2O3 – UO2

About the open question of the solubility of zirconia in Zr-Fe, SKODA tests on Tliq and Tsol in the corner of low oxygen in the Zr-Fe-O system confirmed the large solubility

The influence of the hyperstoichiometry of UO2 on Tliq and Tsol of typical corium

(U,Zr)O2 was tested by AEA-T and the results confirmed the model used in the database.The direct construction of the complex phase diagram U-Zr-Ba-Mo-Owasmadein different atmospheres (reducing conditions, and in air) as an exploratory research and showed that in rather oxidizing atmospheres, the stability of the Ba zirconates decreases in favour of

Ba molybdates or scheelite (BaMoO4) This was confirmed by tests in thermal gradients at UCL and by PIE at SCK.CEN on irradiated fuel pins at high burn-up

Tliq of 10 ex-vessel mixtures (sub-systems in UO2-ZrO2-FeOx-CaO-Al2O3-SiO2)

measured by different techniques, confirmed or validated points of the ex-vessel database The database was further improved by THERMODATA The thermodynamic properties

of 11 gaseous elements and of the UxOy species were specially revisited and reassessed The database was validated by pre- and post-calculations of experimental tasks, and by (semi)global tests : FP release (VERCORS tests), in-vessel corium pool stratification

(SASCHA), and ex-vessel corium spreading (VULCANO)

The influence of uncertainties on the corium physical properties was studied on the point of view of the variability of key properties from different versions of the database

A strategy was developed by IRSN to couple the database to severe accident codes, through tabulation of the phase diagram, generation of a library by GEMINI2 and

interpolation This was proven to be efficient for the 4 elements system U-Zr-Fe-O

A reactor calculation (TMI2) on the new subsystem U-Zr-O did by EdF with MAAP4 concluded that the calculation must be enlarged to a subsystem including at least iron

The “NUCLEA” database, edited by Thermodata is now commercially available

A OBJECTIVES and SCOPE

For the prevention, mitigation and management of severe accidents, many problems related to core melt have to be solved : fuel degradation, melting and relocation, convection inthe core melt(s), coolability of the core melt(s), fission product release, hydrogen production, ex-vessel spreading of the core melt(s), etc

To solve these problems such properties like thermal conductivity, heat capacity, density,viscosity, evaporation or sublimation of melts, the solidification behavior, the tendency to trap

or to release the fission products, the stratification of melts, must be known

However most of these properties are delicate to measure directly at high temperature and/or in the radioactive environment produced by the fission products Therefore some of them must be derived by calculations from the physical-chemical description of the melt : number of phases, phase compositions, proportions of solids and liquids and their respective oxidation state, miscibility of the liquids, solubility of one phase in another, etc This

information are given by the phase diagrams of the materials in presence

The phase diagrams can be experimentally determined in specific tests, or they can be

constructed by calculations made on the basis of measurements of the Gibbs energies, binary

or ternary interaction parameters, and models (the « associated » model or the « ionic » model) and in a second step validated by application to global tests

Results of the “CIT” & “THMO” projects (4th R&D-FW Program) were previously obtained in this field but had to be confirmed and broadly enlarged

In this frame, the general objective of “ENTHALPY” was to obtain one unique

European commonly agreed thermodynamic database for in- and ex-vessel applications, well validated and with methodologies able to couple the database to Severe Accident codes used

by end-users like utilities, safety Authorities and nuclear designers

to re-in force the effort on the key Zr-Fe-O system (Task 2.2.1)

B.1 : Assembling of the two existing databases and extension to new elements (WP1)

The two existing nuclear thermodynamic databases had to be merged in one database, with agreed thermodynamic models, covering the entire field from metal to oxide

(borides/carbides) for a complex multi-component chemical system)

B.2 : Separate Effect Tests (WP2)

Tests were performed in the key U-Zr-Fe-O (sub)systems and in other in- & ex-vessel subsystems (B2O3, B, Pu, PuO2, Mo ; Si, SiO2, Ca, CaO, Al2O3 etc) including FPs with high

decay heat (Ba), in such a way that thermodynamic results (Tliq, Tsol, enthalpies, solubility

limits) were obtained

B.3 : Improvement and validation of the database (WP 3)

This WP included the improvement and validation of the new database against global tests

B.4 : Evaluation of the consequences of the uncertainties (WP 4)

This WP envisaged the consequences of uncertainties on Corium physical properties

B.5 : Methodologies of coupling the database with SA codes (WP 5)

Methodologies were developed to effectively couple the database to severe accidents codes, and a recalculation of TMI2 with MAAP4 was made

B.6 : Edition of the database (WP 6).

This WP included the documentation, the discussion of an agreement on property rights and the diffusion of the database

C WORK PERFORMED AND RESULTS

C.1 State of the Art Report

Before the start of ENTHALPY, THERMODATA, had developed the specific TDBCR thermodynamic database for nuclear safety applications It covered the entire field from metal

to oxide domains in the following multi-component system :

O-U-Zr-Fe-Cr-Ni-Ag-In-Ba-La-Ru-Sr-Al-Ca-Mg-Si + Ar-H

and was able to be used for both in- or ex- Vessel applications

AEA-T, through external collaborations, had developed a large oxide thermodynamic database, more oriented to Ex- Vessel applications, based on the following oxides :

UO2-ZrO2-Fe2O3-BaO-La2O3-CeO2-Ce2O3-SrO-Al2O3-CaO-MgO-SiO2

Specific databases were also available (metal domain : U-Fe-Zr-Si-Ba-Sr-Ce-La-Ru-Te where Cerium was introduced for simulating Plutonium ; metal oxide field : U-O-Zr-Si).These two databases needed to be improved, specially in some key-subsystems (U-O, Zr-Fe-

O for instance), to be extended to the absorber elements Boron and Carbon, and to be more largely validated The coupling of the database to SA codes was also weak Faster and more efficient methodologies of coupling were needed

C.2 Assembling of the database (WP 1)

to be adopted (thermodynamic data for unary systems, thermodynamic models)

C.2.1.1 Selection of the elements to be included in NTD

The following elements (18 + 2) were selected for being included in NTD :

O-U-Zr-Fe-Cr-Ni-Ag-In-Ba-La-Ru-Sr-B-C-Al-Ca-Mg-Si + Ar-H

These are the elements of the main materials involved in a severe accident : UO2 (fuel),

Zr (zircaloy cladding), Fe-Cr-Ni (steel of the structural components), SIC and B4C (control rods), Ba-La-Ru-Sr (selected fission products), Al2O3-CaO-MgO-SiO2 (concrete), H2O

(water), O (air, oxides) Argon was added only as a neutral species in the gas phase Hydrogenhas not been taken into account in non stoichiometric solution phases

As any metallic element can be oxidised at a given oxygen potential, the

multi-component (15) oxide system is also a subset of the whole database :

UO2-ZrO2-FeO-Fe2O3-Cr2O3-NiO-In2O3-BaO-La2O3-SrO-B2O3-Al2O3-CaO-MgO-SiO2.

(AgO and RuO2 are two oxides decomposed before melting.)

C.2.1.2 Modelling and merging procedure

The development of a thermodynamic database for multi-component system is based on the critical assessment of the most relevant sub-systems (binary, ternary, …) i.e the

compilation of the available experimental data (phase diagram and thermodynamic

properties), the inventory of all possible phases, the choice of suitable thermodynamic modelsfor each phase, and finally, the optimisation of Gibbs energy parameters

The whole subset of Gibbs energy parameters for a given sub-system is also called a

thermodynamic database, and allows the user to re-calculate the whole phase diagram

The list of all possible condensed solution phases was commonly built, and for each identified solution phase, a thermodynamic model was proposed and agreed

Common standards were also adopted for unary systems (pure elements and oxides)

C.2.1.3 Selection of the systems to be included in NTD

The thermodynamic modelling of the selected multi-component system was based on the critical assessment of all the possible binary (153) or pseudo-binary (105) systems

Only the most important ternary or pseudo-ternary systems (metal, metal-oxygen, oxide) werecritically assessed, due to the very high number of possible ternary systems

C.2.1.4 Critical assessment work

For each system, the following points were carefully treated :

- The list of bibliographic references has been up-dated

- The set of experimental data was re-checked

- The new experimental information was included in the optimisation process

- The heat capacity of stoichiometric compounds was estimated, in order to produce fundamental values for substances (H°298.15 K, S° 298.15 K, Cp(T), G - HSER)

- A new optimisation was performed in order to correct any evident disagreement between calculated and experimental values

- The calculated and experimental temperatures and compositions of the invariant reactions or specific points have been compared in numeric tables

- The phase diagram and specific thermodynamic properties were calculated and compared to experimental values taken from literature on figures

- A new set of optimised Gibbs energy parameters has been produced and included in the new nuclear thermodynamic databases (NTDIV01, NTDIV02, NTD)

- The lack of experimental knowledge was identified

- Finally, a new quality criterion was proposed for each sub-system

C.2.2 Critical assessment and assembling (Task 1.1)

(Assembling of a first partial In-Vessel Nuclear Thermodynamic Database : NTD IV01)

A partial thermodynamic database for In-Vessel applications, named NTDIV01, was firstassembled, based on 14 elements : O-U-Zr-Ag-In-Fe-Cr-Ni-Ba-La-Ru-Sr + Ar-H

It included in particular the 10 oxides system :

UO2-ZrO2-FeO-Fe2O3-Cr2O3-NiO-In2O3-BaO-La2O3-SrO

Then the Gibbs energy parameters of all possible phases were assembled by using the thermodynamic modelling of the Gibbs energy from the assessed binary and ternary sub-systems in the GEMINI2 code format (Ref [1] & [2])

NTDIV01 is composed of three different files :

1 The Gibbs energy parameters of the « lattice-stabilities », i.e the Gibbs energy difference of each element of the multi-component system in a given structure and a reference one (SER = Standard Element Reference)

2 The Gibbs energy parameters of all possible substances referred to any chosen reference state, provided that it is present in the first file

3 The Gibbs energy parameters of solution phases

C.2.3 Extension of the Database (Task 1.2)

C.2.3.1 Extension of NTD IV01 to Boron and Carbon and Assembling of the final

In-Vessel Nuclear Thermodynamic Database : NTD IV02

NTDIV01 was extended firstly to two new elements B and C and one oxide B2O3 to the database After assembling, the final thermodynamic database for In-Vessel applications, named NTDIV02,was thus based on 16 elements (O-U-Zr-Ag-In-Fe-Cr-Ni-Ba-La-Ru-Sr-B-C + Ar-H), and included in particular the following 11 oxides system (UO2-ZrO2-FeO-Fe2O3-

Cr2O3-NiO-In2O3-BaO-La2O3-SrO-B2O3)

This extension involved the thermodynamic modelling of new systems Some of them were already accepted : B-Fe, B-Ni, C-Cr, C-Fe, C-Ni, C-Zr, C-Cr-Fe, C-Cr-Ni, C-Fe-Ni, Cr-Fe-Ni, C-Cr-Fe-Ni Other ones, as B-Cr, B-O, B-U, C-U, B-Zr, B-C, B-C-Fe, B-C-U, B-C-Zr, B-Fe-U, B-Fe-Zr, C-O-U, C-O-Zr, C-U-Zr, C-O-U-Zr, were modelled by THERMODATA

35 new binary and 13 ternary systems were re-assessed, AEA-T being more involved in oxidic systems including B2O3, THERMODATA in metallic ones.

The oxide systems including B2O3 ,-BaO, -Cr2O3 , -FeO, -Fe2O 3 , -In2O 3, -La2 O 3 , -NiO, -SrO, -UO2, -ZrO2 and -FeO-Fe2O 3 , were made by AEA-T (Ref [3]) Experimental data for the pseudo-binary systems B2O3 with -FeO, -Fe2O 3 , -UO2, -ZrO2 came from Task 2.3 (Figure

1)

12 other binaries were assessed : Al-B, Al-C, B-Ba, Ba-C, B-In, In, B-La, La, B-Ru,

C-Ru, B-Sr, and C-Sr by THERMODATA (Ref [4]) The two important binary systems B-U andC-U were optimised (Ref [5]) and the model extended in the quaternary system C-O-U-Zr.NTDIV02 was assembled by the CALPHAD method on the basis of 91 binaries, 55 pseudo-binaries and 21 ternary systems It contains 35 solution phases,191 reference

substances, 203 substances and 148 gaseous species (Ref [6])

C.2.3.2 Extension of NTD IV02 to concrete elements (Al, Ca, Mg, Si) Assembling of

the final In/Ex Vessel Nuclear Thermodynamic Database : NUCLEAINTDIV02 was extended to the concrete elements for ex-vessel applications, adding fournew elements Al, Ca, Mg and Si, and four new oxide components, Al2O 3, CaO, MgO and SiO2 The final database named NTD and later “NUCLEA”, is based on 20 elements :

O-U-Zr-Ag-In-Fe-Cr-Ni-Ba-La-Ru-Sr-B-C-Al-Ca-Mg-Si + Ar-Hand includes in particular the 15 oxides system (Ref [7]) :

UO2-ZrO2-FeO-Fe2O 3-Cr2 O 3-NiO-In2O3-BaO-La2O3-SrO-B2 O 3-Al2 O 3-CaO-MgO-SiO2

23 binary or pseudo-binary systems were identified

- either to be checked : Al-Cr, B-Si, Ba-Si, Ca-La, Ca-Si, La-Si, Mg-O, Al2O3-BaO, Al2O3NiO, CaO-”FeO”, Fe2O3-O2Si, MgO-O2U, OSr-O2Si

or to be re assessed : O Si, Si Sr, Fe2O3-MgO, CaO-UO2

- or to be done : B-Ca, C-Ca, B-Mg, C-Mg, Al2O3-Fe2O3, B2O3-MgO

9 ternary systems were included, either already available as :

Al2O3-CaO-O2Si, Al2O3-O2Si-O2U, Al2O3-O2Si-O2Zr, Al2O3-O2U-O2Zr, O2Si-O2U-O2Zr

or recently made as : Al2O3-B2O3-CaO, Al2O3-B2O3-O2Si, B2O3-CaO-O2Si, B2O3-FeOx

9 ternary oxide systems were directly calculated from the binaries, but ternary

interaction parameters should be assessed in the future : Al2O3-CaO-FeO, Al2O3-CaO-Fe2O3,

Al2O3-Fe2O3-SiO2, CaO-FeO-Fe2O3,CaO-FeO-O2Si, CaO-Fe2O3-O2Si, FeO-Fe2O3-O2Si The following systems were evaluated or re-evaluated : Al2O3-B2O3, B2O3-CaO, B2O3-

O2Si, Al2O3-B2O3-MgO, B2O3-CaO-MgO, B2O3-MgO-O2S

The experimental state of the art and thermodynamic modelling of the O-U binary

system (Figure 2) was improved using a sub-lattice model to describe the UO2+x phase (Ref [8])

C.2.3.3 Comparison calculations

Calculations for selected oxide systems were performed and the results compared with those obtained using a commercial database The latter called MTOX has been produced as part of a collaborative project by the NPL (National Physical Laboratory, UK) and other

industrial sponsors The current database includes oxide systems for a range of components (e.g.K, Na, Fe, Ca, Si, Al, Mg) and has been well validated

Calculations of oxide systems using the MTOX database and the Gibbs energy sation code MTDATA were carried out (Ref [9]) The system FeO-Fe2O3-CaO-SiO2 was selected for the comparison exercise and 21 ternary and quaternary compositions proposed byIRSN The MTOX results showed that in some cases the change in the amount of liquid phase is gradual over the temperature range whereas in others the amount increases very rapidly over a temperature rise of a few degrees These results are to be compared with calculations performed by IRSN using NUCLEA and the GEMINI code Previous code comparison studies have shown that the results produced by the different equilibrium codes (MTDATA and GEMINI) using the same database are in good agreement

minimi-C.2.4 State of the validation

A quality criterion was established for each assessed sub-system:

* Estimated ; No experimental data available

** Perfectible ; Some domains need more experimental information (phase

diagram or thermodynamic properties)

*** Acceptable ; The system is well known and satisfactorily modelled

**** High quality ; The system is quite known and modelled

The complete list of binary systems based on pure elements and oxide pseudo-binary systems

based on pure oxides are presented in (Tables II and III) respectively

Due to the very high number of possible ternary (816) and pseudo-ternary (455) systems

only the most important ternary systems for practical applications (Tables IV and V) were

assessed at this time

C.3 Separate Effect Tests (WP 2)

C.3.1 Solidus – Liquidus temperatures for (U,Zr)O2+x (Task 2.1)

In a severe accident the fuel in the corium will oxidise in steam, in particular for the vessel sequence No thermodynamic data exists for the pseudo binary system (U, Zr)O2+x C.3.1.1 Experimental

ex-Tsol and Tliq measurements were measured by the thermal arrest The PTE included ceramography and SEM analysis

The experimental set-up and proposed compositions were assessed first by the CEA based on preliminary calculations performed using the TDBCR-IV 992 database

Based on this assessment the compositions of the three samples finally investigated were UO2.00-ZrO2, UO2.08-ZrO2 and UO2.15-ZrO2 with the ratio of UO2/ZrO2 set to 80/20 wt.%

C.3.1.2 Results

Tsol & Tliq of the three samples UO2.00-ZrO2, UO2.08-ZrO2 and UO2.15-ZrO2 were measured

(Table VI) (Ref [10]) The results indicate that the solidus-liquidus gap for the stoichiometric

and hyperstoichiometric compositions is slightly larger than predicted and some modifications

of the model for the U-Zr-O system are required These modifications to the NUCLEA

database have now been carried out

C.3.2 Tliq in Zr-Fe-O & Zr-Cr-O systems (Task 2.2)

The objective is the experimental determination of Tsol (and Tliq) of one selected alloy in

the Zr-rich corner of each of the Zr-Fe-O and Zr-Cr-O systems (Zr-Fe-O alloy near 17.3-0.4 wt% and one Zr-Cr-O alloy near 79.6-20.0-0.4 wt%)

82.3-C.3.2.1 Experimental

6 Zr-Fe-O and 5 Zr-Cr-O alloys were prepared from pure metals (+ ZrO2) Their

chemical composition was analysed, and their hardness (HV10) measured

The resistance furnace of the Balzers Exhalograph was adapted to the video-recording This resulted in a successful off-line magnified viewing of the melting process on a computerscreen The camera & pyrometer were coupled with a PC-based image analyser

Tsol & Tliq could be approximately determined from the off-line sequential set of images.C.3.2.2 Results

Tsol and Tliq were measured by both the first appearances of the liquid phase (solidus)

and the complete melting manifested by formation of a drop (liquidus) (Figure 3 a, b) This

method seems to locate the solidus and liquidus with a precision of ~  15-25 C

Tsol agreed relatively well with the values calculated using the NTDiv01 database The measured liquidus temperatures TLIQ for both systems were relatively close to Tsol A question arose : does a solid phase exist in the interior of the presumed “molten” drop or not? To answer this question the Zr-Fe-O specimens were heated to various temperatures beyond TLIQ, i.e between TLIQ and ~1700 C and cooled down as rapidly as possible (~11 C/s)

The microstructure of the formed drops was investigated by both the optical metallography,

by X-ray phase analysis and by microprobe analysis The two main phases had an appearance of

a liquid phase, the third one being cracks.A small black phase (less than 1% at TLIQ-CALC) is

~(Zr,Y)O2 originating from the Y2O3 support pad In both specimens no ZrO2 was found The oxygen was found only in solid solution, either within -Zr(O) containing small amount of iron orwithin the Zr-Fe matrix

Analysis of both the cooling curves (from ~1700 C) and the recorded images of the melting process also indicated that the main phases were liquid at TLIQ SKODA concluded (Ref [11]) that the specimens were completely molten at the liquidus temperature TLIQ

(formation of spheres) This is probably valid for both systems Zr-Fe-O and Zr-Cr-O

C.3.3 Tliq in B2O3 + FeOx/ZrO2/ UO2 systems (Task 2.3)

Boron trioxide has relatively low melting point (~450 oC), density and viscosity Fused B2O3 readily dissolves many metal oxides may influence fuel degradation, melting and

relocation processes To model this influence, phase diagrams of B2O3 with Fe2O3, ZrO2 and UO2 are needed but were lacking

C.3.3.1 Experimental

High purity powdered Fe2O3, ZrO2, UO2 and amorphous granulated B2O3 were used as starting materials Due to the hygroscopic nature, boron trioxide was grinded to powder and dehydrated

Temperature history measurements with the test mixtures were carried out in a resistant furnace under a protecting gas The liquidus temperatures are signalled by a characteristic brake on the temperature difference (T) curves of the cooling cycles

C.3.3.2 Results

C.3.3.2.1 Fe 2 O 3 -B 2 O 3 system

The mean values of Tliq are summarised in Table VII Reproducibility of the data was estimated as ±10 oC The phase diagram for the Fe2O3-B2O3 system could

then be constructed (Figure 4), the phase relations bellow liquidus temperatures being

taken from Makram (Ref [12])

C.3.3.2.2 ZrO 2 – B 2 O 3 system

The liquidus temperatures for the ZrO2-B2O3 system were determined by temperature

history and as well as from solubility measurements (Ref [13])

In the former experiments, test mixtures, containing 2.5 and 5.0 wt % ZrO2 were investigated

The results are given in Table VII In the solubility experiments, the test mixtures containing

ZrO2 in excess as compared with the saturated solution at a given temperature, were fused in afurnace at constant temperature The original composition of the test mixture for both

temperatures was chosen as 20 wt% ZrO2 – 80wt% B2O3

Because the density of ZrO2 (~5,6 g/cm3) is significantly higher than that of B2O3 (~1,5 g/cm3) , during the isothermal heating at the selected liquidus temperature there is

stratification between the dense solid phase (a layer of unsolved ZrO2 deposited on the

bottom) and the liquid phase (a saturated solution of ZrO2 in B2O3)

Quenching allows to preserve the composition of the saturated liquid solution This

composition and the isothermal heating temperature (1800° and 2000°C) determine the

position of the liquidus temperature on the phase diagram They are given in Table VII.

The ZrO2 – B2O3 phase diagram (in the rich B2O3 side) is shown in Figure 5

C.3.3.2.3 UO 2 – B 2 O 3 system

The liquidus temperatures for the UO2-B2O3 system were determined by temperature

history measurements (Ref [14]) and are summarised in Table VII Combining these results with previous experimental results, the phase diagram could be constructed (Figure 5).

C.3.4 Tliq in UO2 – ZrO2 – FeOx – CaO – Al2O3 – SiO2 systems (Task 2.4)

The objective was to measure Tliq and possibly Tsol in systems relevant to SA ex-vessel sequences

10 temperatures (Tliq and/or Tsol)were investigated by : Visual Polythermal Alalysis in

an induction-melting furnace with a cold crucible (VPA IMCC), by VPA in the Galakhov micro-furnace, and by classical DTA (for solidus temperature determination)

The compositions were discussed and defined together with the modelers (Table VIII).

The knowledge of Tliq allows to compare the dissolution ability of different concretes to dissolve corium or to evaluate the viscosities of different concretes (Ref [15])

C.3.5 Experimental study of (sub)system(s) in UO2-ZrO2-BaO-MoOx (Task 2.5)

NB : Task 2.5.1 was merged with Task 2.4

C.3.5.1 Direct determination of the UO2–MoO2–BaO-ZrO2 phase diagram (Task 2.5.2) The study is subdivided into 3 subtasks, performed resp by ULB, SCK.CEN, UCL

C.3.5.1.1 Direct determination of the phase diagram ( Task 2.5.2.1)

Two main steps were proposed in order to understand the mechanisms of the fuel degradation and the formation of the new phases as a function of oxygen potential :

A The study of the pseudo-ternary system BaO-ZrO2-MoOx ( with x = 0 ,2 ,3 ) in oxidising and reducing atmosphere giving a preliminary knowledge of the behaviour of the Mo0, Mo4+and Mo6+ compounds in presence of BaO and ZrO2

B The extension of the ternary system to the quaternary system UO2-BaO-ZrO2-MoOx at 1600°C at 3 different oxygen pressures, i.e.10-16 atm, 10-9 atm and 10-5 atm

C.3.5.1.1.1 Experimental

A major part of anneal experiments were performed in electrically heated furnaces.Preliminary works were performed in oxidising atmosphere (air and nitrogen Tests under controlled atmospheres were performed in three tubular furnaces equipped either with a zirconia oxygen sensor, either with an H2O sensor and a classic air reference zirconia gauge operating at 700°C) Quenching was possible in two of these furnaces

Phase identification and composition analysis were performed by optical microscope, X-RD, X-Ray fluorescence, SEM with EDX and EPMA

Calculations based on H2/H2O equilibria in order to check whether the effective pO2 over samples heated at 1400°C-1600°C correspond to those measured in the outlet part of thefurnace (700°C) with the zirconia sensor, demonstrated that H2/H2O equilibria are

immediately reached in all parts of the furnace so that the measured pO2 have to be corrected

at 1400°C-1600°C, owing to higher water dissociation at those temperatures Thus, the pO2 values measured at 700°C(10–24 atm, 10–17 atm and 10–13 atm )* correspond actually to 10–16atm, 10–9 atm and 10–5 atm at 1600°C

C.3.5.1.1.2 Results in the ternary system BaO - ZrO 2 – MoO 3

a) In oxidising conditions (pO2> 10 -4 atm, T<= 1600°C), BaO reacts preferably with MoO3 to yield BaMoO4 (scheelite) For Ba/Mo ratios >1, BaO reacts with the phases BaZrO3 and BaMoO4 to form other Ba-rich zirconates and molybdates of higher Ba order like Ba3Zr2O7, Ba2ZrO4, Ba2MoO5 and Ba3MoO6 For the identified Mo-compounds, the stable Mo valence is

6+; therefore solid solutions between Zr4+ and Mo6+ are limited : no Mo detected in BaZrO3; and a maximum of 5,5 mol% of Zr in substitution for Mo in sheelite BaMoO4

b) In reducing conditions (10-20 <= pO2 <=10-17 atm, 1200°C <= T <= 1500°C), in Mo-rich compositions, the Mo compounds are reduced into Mo0 so that BaO reacts preferably with ZrO2 The perovskites compounds BaZrO3 and BaMoO3 form solid solutions BaMo1-yZryO3 with y = 0,1 and BaZr1-xMoOxO3 with x = 0,08 at 1200°C

The thermal evolution shows that the major part of BaMoO3ss disappears according to a disproportionate reaction yielding Mo0 in a dispersed phase and Mo6+ in Ba3MoO6 while the remainder Mo+4 is stabilised only in BaZrO3 ss As a function of composition and temperature,

Mo0, Mo4+ and Mo6+ compounds can thus co-exist in strongly reducing conditions

BaO volatilisation in Ba-rich compositions C and D was not observed at 1400°C

C.3.5.1.1.3 Results in the quaternary system UO 2 - BaO - ZrO 2 – MoO 3 in

reducing atmosphere : 1480°C/ 10 -17 atm 2 pO /8 & 32 h ; and 1600°C/ 10 -16 atm pO 2 /8h

The main stability domains evidenced from the identified compounds UO2ss, ZrO2ss, Mo0and the perovskite phase BaZrO3ss or BaUO3ss are depicted in Figure 6

BaO reacts preferably with ZrO2 to form the perovskites BaZrO3ss when BaO/ZrO2 2 and BaUO3ss when BaO/ZrO2  2

Mo : appears as a dispersed Mo0 metallic phase in equilibrium with Mo4+ present mainly

in the BaZrO3ss (max 6.8 % in composition A) and BaUO3ss perovskite phases Mo0 can dissolve up to 1 at%Ba, 1.5 at%Zr and 0.5 at%U Its oxygen content (up to 14%) is probably related to oxidation upon cooling There is no stable phase of molybdate

UO 2 matrix forms a solid solution UO2-ZrO2 which may contain up to 14 % ZrO2 at 1480°C and 26.6% ZrO2 at 1600°C When the concentration of Ba increases, the solubility of ZrO2 in UO2 gets reduced because of the formation of BaZrO3 ss

The UO2 matrix dissolves a very low quantity of Ba (<1 at%) and of Mo (<2 at%)

ZrO 2 : exists as a solid solution ZrO2-UO2 up to 12.8 mol% UO2 in sample F at 1600°C.

Perovskite phases BaZrO3 “or BaUO 3 ” : The coexistence of BaZrO3ss and BaUO3ss established at 1480°C is no longer observed at 1600°C where there is a complete miscibility

of barium uranates and zirconates in each other

At 1600°C in sample A (with Ba/Zr = 1), the only BaZrO3 phase dissolving up to 5,8 mol% U and 6.8 mol% Mo is observed For much higher Ba content (samples C and D), perovskites ofuranates are found Therefore, as in the samples as well as in irradiated nuclear fuel BaO is never exceeding the ZrO2 nor UO2 content, the volatility of BaO in reducing conditions is limited by its interactions with UO2 and ZrO2 (contrarily with the ternary observations).Thus, the major effect of the UO2 matrix is to reduce the stability of the Ba molybdates :these are reduced in Mo0 and Mo4+ with an enrichment of the Ba zirconates in U and

formation of Ba uranates

C 3.5.1.1.4 At 1600°C in oxidizing atmosphere (pO 2 10 –9 atm)

Stability domains evidenced from the identified compounds UO2ss, ZrO2ss, BaMoO4 (scheelite), BaZrO3ss, the MoO2-rich phase and a new uranate (Ba2U3Ox by SEM) are drawn in

Figure 6B

BaO reacts preferably with oxidised Mo to give BaMoO4 (already observed in the ternary system at pO2 >10–4 atm) which is present in all the nine compositions studied Owing

to its congruent melting at 1457°C

UO 2 dissolves a greater amount of ZrO2 (32% in sample F) than in strongly reducing

medium The content of Mo and Ba in this phase is low (1% and 0.3% resp.)

ZrO 2 in excess of its solubility limit in UO2 exists as a free phase of zirconia (dissolving12.3% of UO2) in equilibrium with the matrix

Mo : All Mo compounds are oxidised at this oxygen pressure.

The perovskite phase is limited to a Zr-rich solid solution occuring for Ba/Mo ratios

>1, in a domain including compositions B and G In this domain, BaZrO3ss contains 18.4 at% BaUO3 and 1.7 at% BaMoO3

The uranate phase appears in a large domain where Ba/Mo >1 (compositions B, G, C

and D) This relatively homogeneous uranate whose composition is close to

Ba 2 (U 3-x Zr x )O 9+y is unknown in the more recent JCPDS files but has been assigned to a fluorite – type structure

C.3.5.1.5 Conclusion

The major effect of increasing the pO2 from reducing (10-16 atm) to oxidizing conditions (10-9atm) is a reduction of the stability of the Ba perovskites in favour of the formation of stable phases scheelite (BaMoO4) and MoO2 when Ba/Mo < 1 or scheelite (BaMoO4) and a

new uranate Ba 2 (U 3-x Zr x )O 9+y when Ba/Mo > 1 (Ref [16])

C.3.5.2 Release in thermal gradient (Task 2.5.2.3)

C.3.5.2.1 Experimental

The investigations covered the study of oxides of U, Zr, Ba and in strong temperatures gradients and in different atmospheres.Three different powders mixtures (A, D, F) were prepared : (mol%) : A : 80% UO2 + 7% Ba + 7% Zr + 7% Mo

D: 80% UO2 + 18% Ba + 1% Zr + 1% MoF: 80% UO2 + 1% Ba + 18% Zr + 1% MoTwo types of tests were performed :

A) release of condensation particles by homogeneous heating, and

B) differential heating (in strong temperature gradients)

A) The pellets were heated up to 2400°C Condensation particles were sampled by an

impactor Compositions and test conditions are given in Table IX.

B) Tests in a strong temperature gradient were also performed in order to reveal possible

migrations or segregation of the different elements inside the pellet matrices Pellets A, D and F were submitted to a strong temperature gradient of approximately 1000-2000°C from side to side

XRD, EDX analyses,SEM and neutron activation were performed on pellets before and after heating as well as on the condensed particles

C.3.5.2.2 Results

A) Release from heated pellets

Ba and Mo prevail in particles condensed from pellet A (rich in Ba and Mo) In the case

of pellet D, Ba, is most abundant in condensed particles No Zr has been found even in the case of pellet F, where Zr is abundant

In the released and condensed particles, Ba and Mo(O) are closely associated suggestingmolybdates and as in the pellets, U segregates from Ba/MO, but no Zr was observed at all

B) Behaviour in temperature gradients (7,5 mm pellets)Grains grow up with temperature as a consequence of fusion-recristalisation

Starting from an homogeneous distribution of U, Zr, Ba and Mo before heating, Ba and Mo migrate to the hot zone and conversely, U migrates to (or concentrates in) the colder zone

(Figure 8) There is a marked segregation between U and Ba-Mo Mo and Ba are strongly

associated suggesting the presence of molybdates (observed by XRD as BaMoO4, BaMo4O13, Ba3Mo18O28) Migration of Zr from the cold to the hot zones in pellet F is compensated by U

In hot zones of pellet F (2000°C), in opposition to the cold ones, U and Zr seem to be in close association, probably as a solid solution or as urinates (Ref [17])

C.3.5.3 Review of the U-Zr-Mo-Ba-O system including in irradiated fuel

pins(Task 2.5.2.4)The objective is to study the phase formation and the stability of the phases formed by the FPs Ba, Zr and Mo in nuclear fuel, focused on the influence of the oxygen potential

C.3.5.3.1 Solid solubility and precipitation of FP Molybdenum

In the actual fabrication of LWR fuels, UO2 is slightly hyperstoichiometric, and MOX is slightly hypostoichiometric But during and after irradiation, the initial O/M ratio increases Different methods to measure the real oxygen potential (or O/M ratio) were developed, notably through the analysis of the repartition of Mo as metallic precipitates and in oxides Since the free energy of formation of MoO2 is slightly above that of nuclear fuel, the measurements of the activities of Mo and MoO2 could be a method to estimate the local oxygen potential of irradiated fuel although the basic assumption of the elevated solubility of MoO2 in the UO2 matrix was never confirmed Instead subsequent measurements and

theoretical work confirmed that Mo solubility limit is very low The quantity of Mo dissolved

in irradiated UO2 is overestimated because of the contribution of submicron metallic or oxide precipitates that are not resolved in SEM or TEM microscopes

Nevertheless the mechanism remains partly valid : when pO raises up, Mo disappears from metallic inclusions and is incorporated in the grey phases which are perovskite ABO3 precipitates with Ba and Zr for the A and B sites respectively Therefore the equilibrium Mo +O2  MoO2 is established between the Mo activity in metallic precipitates to the Mo activity

in the ABO3 grey phase But these ABO3 particles remain too small to be analyzed directly

C.3.5.3.2 Solid solubility and precipitation of FP Ba and Zr

The solid solubility of Ba in the oxide matrix of irradiated FBR fuel and of simulation specimens is lower than 1,6 wt% In LWR fuels, the ceramic precipitates are too small to be quantitatively measured The solubility limit of 0.2wt% for barium in oxide fuel is only

reached at a BU in excess of 45GWd/tHM only recently reached in LWR (see Table X)

At temperatures below 1200°C, the solubility of ZrO2 in UO2 is limited (<0.2wt%), but

it increases above this temperature At 1600°C, a maximum solubility of 27.5 mol% is reported A continuous solid solution of ZrO2 – UO2 is obtained at temperatures >2285°C

C.3.5.3.3 Post-irradiation observations

Given a Mo production rate of about 100ppm per GWd/tHM, Mo tends at any relevant

BU (Table 10), to precipitate from the UO2 matrix The pO in normal LWR conditions being

lower than the Mo/MoO2 equilibrium, Mo is normally incorporated in metallic precipitates For the FPs Ba and Zr, the solid solubility in UO2 of both elements is limited to about 2000ppm under normal LWR operating temperatures The solubility limit of both elements

will be reached at intermediate burn-up level (Table X) Moreover PIE of fast reactor fuels

have shown that precipitation of Ba is enhanced in the presence of dissolved Zr and that the complex perovskite phase (Ba,Sr,Cs)(Zr,Mo,U)O3 is formed

As long as the cladding did not fail, the conditions inside the fuel pin remain sufficientlyreducing to maintain Mo in the metallic precipitates; the ceramic precipitates are of the general compositions (Ba,Sr,Cs)(Zr,U)O3

C.3.5.3.4 Review of equilibria established at the Mo/MoO 2 equilibrium

The experiments on the quasi-ternary system UO2-ZrO2-BaO showed complete mixing

of BaUO3 and BaZrO3 in any proportion In the UO2-ZrO2 rich side a broad two phase field exists, with either UO2 in equilibrium with Ba(U1-xZrx)O3 or U1-xZrxO2 in equilibrium with BaZrO3 Since the production rate of the fission products Ba and Zr is roughly 2:1, it is most likely that BaZrO3 is formed and is in equilibrium with (U1-xZrx)O2 In fuels with low oxygen potentials, molybdenum will not be part of the ceramic precipitates

At 1700°C and at the oxygen potential of the Mo/MoO2 equilibrium , in presence of Mo,Ba(U1-xZrx)O3 solid solution can incorporate limited amounts of Mo and forms a

Ba(U,Zr,Mo)O3 perovskite phase, and the solubility of BaMoO3 in Ba(U1-xZrx)O3 depends on

the U:Zr ratio (see ( Task 2.5.2.1)

Therefore at moderate oxygen potentials, when only part of the produced (Table X) Mo

oxidizes, Mo is incorporated in the perovskite precipitates When more Mo is oxidized, several ceramic phases will co-exist and the formation of Scheelite type BaMoO4 becomes possible, even when the global oxygen potential is still buffered by the Mo/MoO2 couple.Experiments of Pascoal showed that the sole increase of pO induces a major change in the ceramic precipitate stability : while at lower oxygen potentials Mo is in equilibrium with BaZrO3, at more elevated oxygen potentials, it is ZrO2 that is stable against BaMoO4

C.3.5.3.5 Conclusions (Ref [18])

The oxygen potential controls the phase formation of the ceramic precipitates

At low to intermediate BU and at low pO, Ba remains dissolved in the fuel matrix

At more elevated BU and at low pO, because the fission yield of Zr is larger than that of Ba, the perovskyte phase BaZrO3 is the preferred phase for precipitation of the fission products in

a UO2 matrix and molybdenum precipitates as a metal in metallic precipitates

Mo-Pd-Rh-Tc-Ru The perovskyte BaZrO3 can dissolve important amounts of UO2 and various FPs

When the pO rises, molybdenum oxidizes into MoO2 which is primarily incorporated in ceramic precipitates BaZrO3 ; at a pO above that of the couple Mo/MoO2 scheelite BaMoO4 is formed This was observed in a LWR MOX irradiated fuel pin (at BU = 56GWj/tM)

The solubility limit of MoO2 in BaZrO3 also is modified : from 25mol% substitution of ZrO2

at low oxygen potentials, down to below the detection limit in more oxidizing conditions

C.4 Validation and Improvement (WP 3)

C.4.1 Validation based on fuel experiments and empirical models (Task 3.1)

The objective of the task is to compare the predictions of fuel hyperstoichiometry using the model in NUCLEA with the experimental data and other empirical models and thermody-namic databases adopted within the nuclear industry The validation of the model adopted for UO2+x in the database was carried out over the temperature range 973 to 1973 K

Results

Agreement of the calculated values of oxygen potential for UO2+x obtained using NUCLEA,

in conjunction with GEMINI2, and the experimental data is very good across the temperature

range (Ref [19]) For the single phase fluorite region, the calculations confirm the use of some of the empirical models of Lindemer and Besmann, and de Franco and Gatesoupe and show reasonable agreement between the different methods The agreement with the Green and Leibowitz data (only intended for use with stoichiometric and hypostoichiometric

compositions (UO2-x) at high temperatures) is logically poor for O/U ratios greater than ~2.05and temperatures less than 1373 K

C.4.2 Validation based on experiments (Task 3.2)

This task was dropped by a decision at the mid term assessment meeting (Ref [20]).C.4.3 Validation based on Vulcano/CEA & COMETA/NRI experiments (Task 3.3) The VULCANO facility allows to heat up to 3000K and to melt roughly 100 kg of corium composed of representative materials : UO2, ZrO2, FexOy, etc in various proportions, and to pour it on representative materials of ex-vessel core catchers or spreading areas

C.4.3.1 VULCANO VE-U3 experimentThe VULCANO VE-U3 (final melt composition in wt% : 60 UO2 – 24,1 ZrO2 – 6 Fe3O4– 0,8 Fe2O3– 8,2 SiO2- 0,7 CaO – 0,2 Al2O3) experiment has been analysed and compared withsuccessive versions of the nuclear thermodynamic databases (Ref [21])

One of the major qualitative differences between the experiments and the computations regards the computation of the so called chernobylite phase i.e the zircon-coffinite, which is asilicate of mixed UO2 and ZrO2 : (ZrxU1-x)SiO4 These compounds were calculated by

different versions of the THERMODATA database coupled with the GEMINI 2 code but not

found by material analysis in the VE-U3 corium For different quaternary systems of the vessel corium, the uncertainties are important For instance, in the quaternary system U-Fe-Si-

ex-O, no dissolution of iron in urania-zirconia phases is considered in the databases whereas in fact iron was observed in the urania-zirconia phase Similarly, in the (U-Zr-Si-O) and (U-Fe-Si-O) systems, there are uncertainties on the miscibility limits, particularly for the

"chernobylite" phase (ZrxU1-x)SiO4 where the solubility limit, x, is not well known Some authors report a value up to 5% molar

Up to version 981 included, dissolution of “USiO4” in zircon (ZrSiO4) was not

considered in the TDBCR database When the post-test analyses of VE-U1 showed evidences

of the dissolution of uranium in zircon, i.e the formation of chernobilyte, this dissolution was

introduced in the version 992 of TDBCR with a large solubility limit : up to 30 %mol of

uranium could be dissolved in zircon (Zr0.7U0.3 SiO4 compound Such a high solubility seems incompatible with the observed absence of chernobylite in the VE-U3 samples

Now, for the reactor case a precise knowledge of the remaining free silica content is

needed (Figure 9) Testing the TDBCR versions (98,99,00), differences can go up to 26% for

the solid fraction at a given temperature The last version is probably the most exact

C.4.3.2 VULCANO VE-U7 experimentVULCANO VE-U7 consisted in pouring corium over two parallel paths : one in

zirconia, the second in concrete After spreading the global composition of the melts are very similar : 54 UO2 – 32 ZrO2 – 7 FeO – 3 SiO2 – 2 CaO – 0,3 MgO – 0,4 Al2O3 (in wt%) FeO ispartly reduced in Fe This presence of some iron balls gives a valuable information on pO2.Thermodynamic calculations with the GEMINI2 code and the TDBCR001 database predicted correctly the major phases of corium, but some discrepancies were found with the minor phases (Fe-concrete mixtures) and the evolution of the mortar (Ref [22])

C.4.3.3 COMETA / NRI test

In ECOSTAR, mixtures of simulant corium of ZrO2 - Fe2O3 – concrete and prototypic corium systems (UO2-ZrO2-Fe2O3 –concrete) were melted above 2000 K in air

The main important results are (Ref [23]):

 For the ZrO2-Fe2O3 pseudo-binary system, the existence of the monotectic point M

implies a difference of 700 K for the “liquidus” temperature

 In the ZrO2-FeOx-SiO2 system, a miscibility gap is observed between 2423 K and 2073

K whereas the thermodynamic data calculates a full liquid domain (Figure 10).

 In the UO2-ZrO2-Fe2O3 –concrete systems, a monotectic temperature was experimentally identified at 2223K whereas the calculated Tliq is 2200K Between 2623K and 2223K, two liquids were observed whereas GEMINI 2 computed only one liquid phase

C.4.4 Improvement by literature review (Task 3.4)

C.4.4.1 Boiling points and vapour pressures of the elementThermodynamic properties of all gaseous metallic elements were revisited and

reassessed, mainly at high temperature A survey of the boiling points of 11 metallic elements

(Table XI) with the related vapour pressures was made (Ref [24])

A critical assessment was used to optimise of the thermo-chemical properties of gases in consistency with SGTE data for condensed species

C.4.4.2 Thermodynamic properties of UxOy gaseous speciesThe gaseous binary O-U system is complex Tests to determine pvap over UO2 are very difficult because the non-congruent vaporisation of the ceramic and its non-stoichiometry The specific heats were taken from the last work made in Glushko THERMOCENTER

After the analysis of more than 150 scientific papers concerning the composition of the gaseous phase in equilibrium with urania, a c

ritical assessment and optimisation of the thermodynamic properties led to a new set of thermodynamic data The validation of that new data was made by c

omparison between calculations and experiments (Figures 11, 12)

A new set of thermodynamic data are available for the OxUy gaseous species (Ref [25]) but experimental information are still needed

C.4.5 Validation based on Tliq & Tsol measurements in UO2+x–ZrO2 (AEA-T) and Zr-O (SKODA), & on global experiments (VERCORS, Hofmann) (Task 3.5) (Ref [26])

Fe-C.4.5.1 UO2+x – ZrO2 liquidus and solidus measurements