Studies on fuel-cladding chemical interaction between U−10 wt%Zr alloy and T91 steel

Accepted Manuscript Studies on fuel-cladding chemical interaction between U−10�wt%Zr alloy and T91 steel Santu Kaity, Joydipta Banerjee, S.C. Parida, Arijit Laik, C.B. Basak, Vivek Bhasin PII:

S0022-3115(18)31051-1

DOI:

https://doi.org/10.1016/j.jnucmat.2018.10.041

Reference:

NUMA 51280

To appear in:

Journal of Nuclear Materials

Received Date: 31 July 2018 Revised Date:

12 October 2018

Accepted Date: 25 October 2018

Please cite this article as: S. Kaity, J. Banerjee, S.C. Parida, A. Laik, C.B. Basak, V. Bhasin, Studies on fuel-cladding chemical interaction between U−10�wt%Zr alloy and T91 steel, Journal of Nuclear Materials (2018), doi: https://doi.org/10.1016/j.jnucmat.2018.10.041. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Studies on fuel-cladding chemical interaction between U− −10wt%Zr alloy and T91 steel Santu Kaitya,f,*, Joydipta Banerjeea, S.C. Paridab,f, Arijit Laikc,f, C.B. Basakd,f, Vivek Bhasina,e a b

RI PT

Radiometallurgy Division, Bhabha Atomic Research Centre, Mumbai 400085, India

Product Development Section, Bhabha Atomic Research Centre, Mumbai 400085, India c

Materials Science Division, Bhabha Atomic Research Centre, Mumbai 400085, India

d

Glass and Advanced Materials Division, Bhabha Atomic Research Centre, Mumbai 400085, India e

Nuclear Fuels Group, Bhabha Atomic Research Centre, Mumbai 400085, India Homi Bhabha National Institute, Mumbai 400094, India

M AN U

SC

f

Abstract

The chemical interaction of U−Zr based fast reactor fuel with T91 grade steel cladding was investigated by employing DTA (differential thermal analysis), diffusion couple experiment and SEM/EDS, SEM/EBSD analyses. The DTA results revealed a strong exothermic reaction just below the solidus temperature of the U−Zr alloy. The eutectic liquefaction temperature between

TE D

U−Zr alloy and T91 steel was found to be at 995 K. The interdiffusion between U–10Zr and T91 resulted in the formation of UFe2-type phase, a Zr-rich layer and a Zr-depleted layer at the interface. The growth kinetics of the reaction zone was also determined using the total width of UFe2 and Zr-rich reaction layers. The activation energy for the reaction was found to be 170 kJ

EP

mol-1. The Zr-rich layer at the interface acts as diffusion barrier which enhances the fuelcladding chemical compatibility at reactor operating temperature. However, the U−Zr fuel is no

AC C

more compatible with the cladding above the eutectic liquefaction temperature (995 K).

Key words: Fuel-cladding chemical interaction, U−Zr alloy, T91 steel, Differential thermal analysis, Diffusion couple experiment

*Corresponding author Radiometallurgy Division, Bhabha Atomic Research Centre, Mumbai 400085, India. E-mail address: [email protected]; [email protected] (Santu Kaity)

1

ACCEPTED MANUSCRIPT

1. Introduction: Metallic alloys are considered as candidate fuels for fast breeder reactors (FBRs) due to high breeding potential, high thermal conductivity, high atom density, low doubling time and ease of fabrication compared to other ceramic fuels [1]. Under the Integrated Fast Reactor (IFR)

RI PT

program, irradiation experiments of U−Pu−Zr and U−Zr based metal fuel pins were carried out in the Experimental Breeder Reactor (EBR-II) and the Fast Flux Test Facility (FFTF) [2, 3]. The cladding materials used in the irradiation test were austenitic (SS 316, D9) and tempered martensitic alloys (HT9, HT9M). However, a few shortcomings of metallic fuels like low solidus

SC

temperature, high swelling rate and fission gas release, and susceptibility to chemical and

M AN U

mechanical interaction with cladding materials suppress their full potential.

Recently, increased attention has been paid to develop metallic fuel for future Fast Breeder Reactors (FBRs) in India. As a part of that, two design concepts being explored for the metallic fuel are (a) sodium bonded ternary U−Pu−Zr alloy as fuel and (b) mechanically bonded binary U−Pu alloy as fuel with a Zr liner inside cladding [4, 5]. T91 grade steel is chosen as cladding material, while binary U–Zr alloy is proposed as the blanket materials, which is a sub-system of

TE D

ternary U–Pu–Zr alloy. T91 grade steel is 9Cr–1MoVNb type steel having ferritic-martensitic structure. The ferritic steels have excellent irradiation resistance to void swelling compared to austenitic stainless steels. In addition to that, the steels exhibit higher thermal conductivity and lower expansion coefficient than austenitic stainless steels [4, 6]. However, a rise of the ductile-

AC C

steels.

EP

brittle transition temperature (DBTT) due to fast neutron irradiation is a concern in these types of

The chemical compatibility between the fuel and cladding materials also known as fuel-cladding chemical interaction (FCCI) is considered as a potential problem area in the application of the metallic fuel in liquid-metal cooled fast reactors. During irradiation, the fuel material swells and comes in direct contact with the cladding which in turn promotes interdiffusion between the fuel and the cladding. This results in the development of intermetallic phases that can adversely affect the structural integrity of the cladding [7]. The interaction layer formed at the interface may cause a thinning of cladding which may reduce the load bearing capability of the cladding tube. FCCI is of prime importance because of the formation of low melting eutectic which may often 2

ACCEPTED MANUSCRIPT

limit the life of the fuel pin in a reactor. For reliable operation of a fast breeder reactor, the fuel elements must be resistant to breaching even in case of overpower transients. From the standpoint of accident transients, the most important factor is the accelerated rate of cladding attack, once eutectic liquefaction forms at the interface. Hence, liquefaction temperature due to

RI PT

FCCI is considered as an important parameter in the assessment of cladding temperature limit for the use of metallic fuels in fast reactors.

FCCI, a complex phenomena, depends on many factors like fuel composition, type of cladding

SC

materials with their composition, linear power rating, burnup, fuel-cladding interfacial temperature, fission products etc. [8, 9]. In general, FCCI can be avoided in two ways i.e. either

M AN U

by alloying of uranium or by using a barrier layer between fuel and cladding. Zr was chosen as an alloying element as well as a barrier layer since it increases the solidus temperature and improves the fuel-cladding chemical compatibility by suppressing the interdiffusion between the fuel and cladding components. The best way to investigate the fuel-cladding chemical interaction is to study the chemical compatibility test by carrying out diffusion couple experiment at different isothermal region close to reactor operating temperature. Another way to study the

TE D

chemical compatibility is by differential thermal analysis where the reaction temperature can be determined very accurately.

The interdiffusion reactions of U−Zr alloy with Fe, Fe−Cr, Fe−Ni−Cr, austenitic (SS316 and

EP

D9), ferritic-martensitic (HT9) were investigated by many authors [7, 10-15]. However, void swelling limits the use of high swelling austenitic steels cladding at high burnup. Hence, FCCI

AC C

problem between metallic fuel and austenitic steel cladding might be largely academic [1]. The ferritic/martensitic steels such as HT9 and T91 have been considered as promising candidate materials for high burn-up operation due to their excellent irradiation swelling resistance. The literature data on the interdiffusion reaction between U−Zr alloy and feritic/martensitic steels mainly emphasizes on the U–10Zr alloy fuel and HT9 cladding steel. The interdiffusion behaviour between U–10Zr alloy and HT9 at 973 K was investigated by Keiser et al. [13, 14] and Lee et al. [15]. The diffusion behaviour between U–10Zr with T91 was studied at 1013 K by Ryu et al. [16]. However, it appears that the systematic investigations on the chemical compatibility between U−10Zr alloy and T91 grade steel are limited in the literature. This paper 3

ACCEPTED MANUSCRIPT

aims to fill the existing gaps in the understanding of the chemical compatibility between U−Zr alloy and T91 steel. Hence, a detail investigation on FCCI between these two materials was carried out in the present investigation.

RI PT

The main objectives of the present study are:

(i) investigation on chemical reaction between U−10Zr alloy and T91 grade steel by differential thermal analysis (DTA)

(ii) studies on the FCCI between U–10Zr and T91 by diffusion couple experiments

SC

(iii) determination of growth kinetics of reaction layers at U−10Zr/T91 interface.

M AN U

The outcome of these studies will give feedback about fuel-cladding chemical interaction during nominal steady-state operating conditions and further about its behaviour at elevated temperatures during off-normal reactor events. 2. Experimental Procedure:

U−10Zr alloy was prepared using nuclear grade uranium and 99.95% pure crystal bar of

TE D

zirconium by following injection casting route. The typical diameter of the alloy slug was around 5 mm. The structural characterization of the as-cast sample was carried out with X-ray diffraction (XRD) analysis with Cu Kα radiation. The microstructural analysis was carried out using a scanning electron microscope (SEM) equipped with energy dispersive spectroscope

EP

(EDS). T91 grade steel was used in the standard normalized and tempered condition with hardness of 220 kg mm-2. The heat treatment of T91 follows austenitization at 1323 K and air

AC C

quenching, followed by tempering at 1023 K for 1h. The chemical composition of T91 steel is given in Table 1. The chemical reaction/interaction between U−10Zr and T91 steel was investigated by two methods (i) differential thermal analysis and (ii) diffusion couple experiment.

2.1 Differential thermal analysis (DTA) To investigate the chemical interaction, DTA experiments were carried out by placing small discs of U−10Zr and T91 samples one on top of the other inside sample crucible. For these experiments, total sample of ∼200 mg was used by keeping the ratio of U−10Zr and T91 of 4

ACCEPTED MANUSCRIPT

approximately 5:1 to simulate fuel rich part of fuel-cladding system. Here, Al2O3 crucible with an inner coating layer of Y2O3 was used to avoid any interaction between sample and crucible material. The instrument was evacuated with a standard mechanical pump and backfilled with high purity argon several times prior to the tests. During the experiment, the furnace and sample

RI PT

chambers were purged with highly purified argon gas flowing at 20 ml min-1. In addition to this, commercially available traps for oxygen and moisture were used to avoid oxidation. The temperature calibration was accomplished by measuring the melting points of high purity metals, such as Zn, Al, Ag, Au etc. The rates of heating and cooling were programmed at 10 K min-1.

SC

The thermogram of both the heating and cooling cycles were recorded and analyzed. A similar DTA experiment was also carried out to investigate the chemical interaction between U metal

M AN U

and T91 steel. The DTA data may help to explore the role of Zr in the interaction between Ubased fuel and T91 steel cladding. After the DTA experiment, the samples were characterized by energy dispersive spectroscope (EDS) and electron back scattered diffraction (EBSD) attached to SEM to identify the reaction products.

2.2 Diffusion couple experiment

TE D

U−10Zr alloy was cut into 3 mm thick discs and T91 steel rod of the same diameter was cut into discs about 0.5 mm thick. The surfaces of all these discs were metallographically polished to 1 µm surface finish. To investigate the diffusion reaction between U−10Zr and T91, diffusion couples of a disc of U−10Zr alloy between two T91 steel discs were prepared. These couples

EP

would be referred to as U−10Zr/T91 couples. The components of the diffusion couples were kept in fixtures made of Inconel 600, to ensure intimate contact during annealing. Tantalum foil was

AC C

used to prevent any chemical reaction between couples and fixture. The diffusion couple-fixture assembly is shown schematically in Fig. 1. The fixtures containing these couples were encapsulated in quartz tube in helium atmosphere after repeatedly flushing by helium gas and then annealed in a resistance heating furnaces maintained at 823, 923, 973 and 1003 K for duration up to 1500 h. The heat treatment schedule of the couples is given in Table 2. Subsequent to annealing, the couples were sectioned using a slow speed diamond cutting wheel and then the exposed cross sections were metallographically polished. The extent of reaction and phases formed at the interface were characterized using SEM/EDS and SEM/EBSD. The X-ray

5

ACCEPTED MANUSCRIPT

line scans of U, Zr, Fe, and Cr were acquired across the interfaces of the diffusion couples to determine the distribution of each element.

3. Results:

RI PT

3.1 Structural and microstructural characterization of U−10Zr alloy

The X-ray diffraction pattern of as-cast U−10Zr alloy is shown in Fig. 2. The XRD peaks correspond to orthorhombic α-U phase (space group Cmcm). The equilibrium phase diagram of U−Zr system [17] indicates that the equilibrium phases of U−10Zr alloy are α-U and δ-UZr2 at

SC

room temperature. Most of the peaks of δ-UZr2 phase overlap with that of α-U phase except few low intensity peaks at 46.58° and 71.82° which correspond to (111) and (112) planes of δ-phase,

M AN U

respectively. The XRD data shows a low intensity peak at ~47°. On the otherhand, the peak at ~72° is not clearly revealed. Furthermore, the surface oxidation of the polished sample might have given rise to peak at around ~47° due to formation of UO2. Hence, the formation of δ-phase cannot be stated conclusively from the XRD result of as cast alloy. The lattice parameters of αphase calculated by Rietveld refinement method using ‘MAUD’ program are a = 2.860Å, b =

TE D

5.865Å, c = 4.980Å.

The microstructure of as-cast alloy (Fig. 3a) shows that the sample is having Widmanstatten structure. Furthermore, the magnified micrographs (Fig. 3(b-c)) show prior γ-grain boundaries with very fine lamellar structure. The lamellar structure is formed due to invariant reactions

EP

(monotectoid, eutectoid) during cooling. However, δ-UZr2 phase could not be resolved by SEM.

AC C

One of the representative EDS spectra with the compositional analysis is shown in Fig. 3d.

3.2 Chemical interaction by differential thermal analysis The phase transformation temperatures of U metal, U−10Zr alloy and T91 steel were examined with DTA and the corresponding DTA curves are collectively shown in Fig. 4. The DTA peaks for different phase transformation are presented in Table 3. For pure uranium metal, DTA peaks at 941, 1046 and 1393 K follow the transformation sequence of α → β → γ → melting, respectively. The phase transformation characteristic of U−10Zr alloy can be best evaluated with the help of U−Zr phase diagram [17]. For U−10Zr alloy, the peak observed at 860 K corresponds 6

ACCEPTED MANUSCRIPT

to the α + δ → α + γ2 reaction. The second peak at 958 K corresponds to the α + γ2 → γ transformation (i.e., α + γ2 → β + γ2 → γ1 + γ2 → γ) and the peak at 1510 K is due to the solidus temperature. It may be noted here that the intermediate transitions are not well resolved at the heating rate of 10 K min-1. For T91, the peak observed at 1015 K is attributed to the

RI PT

ferromagnetic to paramagnetic Curie temperature whereas the peak at 1111 K is due to the ferrite (α) to austenite (γ) transition.

The DTA run of U+T91 composite (small discs of U and T91 rods) was carried out up to 1473 K

SC

and both heating as well as cooling curves were presented collectively in Fig. 5(a-b). During the first heating cycle (Fig. 5a), DTA curve shows endothermic peaks at 941, 1015, 1046 and 1111

M AN U

K which are due to the phase transformations of uranium and T91 as explained above. Apart from these, the U+T91 composite exhibits one very intense exothermic peak at 1383 K which is followed by an endothermic peak at 1387 K. The intense exothermic peak is indicative of strong chemical reaction between the two whereas the peak at 1387 K indicates the melting of unreacted uranium metal. However, only two peaks were observed at 1313 and 988 K during cooling cycle (Fig. 5a). The DTA run was then repeated with the same sample and the resulting

two peaks.

TE D

curves are shown in the Fig. 5b. Both the heating and cooling curves of the repeat run show only

The DTA curves of U−10Zr+T91 composite were shown in Fig. 5(c-d). The endothermic peaks

EP

at 861 and 958 K correspond to the phase transformation of U−10Zr alloy whereas the Curie temperature and ferrite to austenite transition of T91 are observed in the U−10Zr+T91 composite

AC C

at 1015 and 1111 K, respectively. In this case, the exothermic reaction peak occurs at 1495 K which is significantly higher than that observed in U+T91 composite. The exothermic peak was followed by an endothermic peak at 1503 K. During cooling, two peaks were identified at 1383 and 978 K. In the repeat run, the first peak appeared at 995 K during heating. The peak observed at 1383 K during cooling was not resolved properly in the heating curve of the repeat run. The cooling curves of both the first and repeat runs are almost identical in nature.

The BSE micrographs of the reaction products of the U+T91 and U−10Zr+T91 composites are shown in Fig. 6(a-d). The microstructural investigation indicates that the reaction products do not 7

ACCEPTED MANUSCRIPT

contain any unreacted starting materials/components. The reaction product of U+T91 composite (Fig. 6(a-b)) contained two phases i.e., bright and grey coloured phases. The larger grey coloured phase are uniformly distributed in the lamellar structure of grey and bright phases. On the other hand, the reaction product of U−10Zr+T91 composite had three distinct phases which appear as

RI PT

bright, grey and dark in Fig. 6(c-d). It was revealed in the magnified micrographs (Fig. 6d) that the dark phases are surrounded by grey phases and the matrix contained bright and grey phases.

The different phases were further characterized EBSD and EDS attached to SEM. For EBSD,

SC

standard metallographic processes of sectioning, grinding and polishing were carried out upto 1 µm surface finish and then the final polishing step was done with 0.025 µm colloidal silica for

M AN U

about an hour to ensure that the sample surface is free of any sectioning damage. In the SEM, the sample was tilted approximately 70º with respect to the incident beam and the accelerating voltage of 20 kV was used. The electron forescatter images of the U+T91 reaction product at different magnifications were shown in Fig. 7(a-b). The EBSD patterns (Kikuchi patterns) of the different phases were acquired by point mode analysis and then matched with the crystal structure for the possible phases of interest (i.e., UFe2, U6Fe, ZrFe2, Zr(Fe,Cr)2 etc.). The

TE D

experimental and indexed EBSD patterns as shown in Fig. 7(c-f) revealed two types of phases i.e., cubic UFe2-type (space group Fd-3m) and tetragonal U6Fe (space group I4/mcm). The EDS spectra of two phases are shown in Fig. 7(g-h). The presence of Cr in the U6Fe phase is negligible and hence could not be detected in the EDS. On the other hand, the UFe2 phase with

EP

soluble Cr can be represented as U(Fe,Cr)2. It may be noted here that both the UFe2 and

AC C

U(Fe,Cr)2 compounds have cubic Laves structure (space group Fd-3m) [18].

The electron forescatter images of the U−10Zr+T91 reaction product at different magnifications were shown in Fig. 8(a-b). The experimental and indexed EBSD patterns (Fig. 8(c-h)) indicate that the sample has three types of phases i.e., hexagonal Zr(Fe,Cr)2 (space group P63/mmc), UFe2 (space group Fd-3m) and U6Fe (space group I4/mcm). The EDS spectra of the three phases are shown in Fig. 8(i-k). The intermetallic compounds of Zr(Fe,Cr)2 type with hexagonal Laves structure (space group P63/mmc) were reported by many authors in the literature [19-21]. The solubility of Cr in the U6Fe phase is negligible. Like U+T91 reaction product, the UFe2 phase with small fraction of soluble Cr can be represented here as U(Fe,Cr)2. The Cr content in the 8

ACCEPTED MANUSCRIPT

Zr(Fe,Cr)2 phase is higher than that of the U(Fe,Cr)2 phase. The Zr(Fe,Cr)2 phase has significant amount of soluble U and the U(Fe,Cr)2 phase has significant amount of soluble Zr. It may be noted here that the significant amount of solubilities of Zr in UFe2 and U in ZrFe2 are also reported in the U−Zr–Fe phase diagram [22]. The results indicate that the bright, grey and dark

3.3 Diffusion couple experiment 3.3.1 Interfacial reaction in U−10Zr/T91 couple at 823 K

SC

UFe2/U(Fe,Cr)2 and Zr(Fe,Cr)2, respectively.

RI PT

phases as observed in the BSE micrographs of the reaction product (Fig. 6(a-d)) are U6Fe

The microstructure of the interdiffusion layer formed at the interface of the U−10Zr/T91

M AN U

diffusion couple after annealing at 823 K for 1500 h is shown in Fig. 9a. The magnified microstructure of the interface is shown in Fig. 9b. The intensity profiles of U-Mα, Zr-Lα, FeKα, Cr-Kα X-ray lines, as shown in Fig. 9c, were recorded along a line AB marked in Fig. 9b. The interdiffusion between the two specimens leads to the formation of different reaction layers at the interface namely UFe2-type intermetallic, Zr-rich layer, and Zr-depleted layer. The interdiffusion layer adjacent to the T91 contained approximately 29 at.% U, 8 at.% Zr, 57 at.%

TE D

Fe and 6 at.% Cr and hence it is most likely to be of UFe2-type intermetallic with soluble Zr and Cr. This layer is followed by a thin Zr-rich layer. The Zr-rich layer contains approximately 65 at.% Zr, 12 at.% U, 20 at.% Fe and 3 at.% Cr. The total thickness of the UFe2-type and Zr-rich layers as measured from the X-ray line profile was ∼ 1.5 µm. However, it was difficult to find

EP

the exact width of the Zr-depleted zone in this case. The Zr depleted zone formed at the U−Zr alloy side contains approximately 95 at.% U, 5 at.% Fe. In U−Zr alloy side, one could see the

AC C

fine precipitates of δ-UZr2 phase which are uniformly distributed in α-U matrix. The Zr-content of the matrix α-phase is negligibly small and therefore could not be detected in the EDS spectrum. The detail microstructural investigation particularly on the U−Zr side of the interfacial region indicates the formation of intermetallic phases τ2 (6 at.% Fe, 67 at.% U, 27 at.% Zr) in a discrete manner as indicated by arrow (Fig. 9b). It may be noted here that the formation of intermetallic phase τ2 was reported in the ternary U−Zr−Fe system [22]. 3.3.2 Interfacial reaction in U−10Zr/T91 couple at 923 K 9

ACCEPTED MANUSCRIPT

Fig. 10(a-b) show the microstructures of the interdiffusion region formed at the interface of U−10Zr/T91 couple after heating at 923 K for 1000 h. The microstructural investigation revealed the formation of similar types of UFe2-type intermetallic, Zr-rich layer, and Zr-depleted layer in the interdiffusion zone. Fig. 10a shows that Zr-rich layer is not uniform throughout the interfacial

RI PT

region which leads to the formation of UFe2-type layer of different thickness. Fig. 10c shows the intensity profiles of U-Mα, Zr-Lα, Fe-Kα, Cr-Kα X-ray lines recorded along the line CD marked in Fig. 10a (with non-uniform Zr-rich layer). Similarly, Fig. 10d shows the intensity profiles of U-Mα, Zr-Lα, Fe-Kα, Cr-Kα X-ray lines recorded along the line EF marked in Fig. 10b (with

SC

uniform Zr-rich layer). These data indicate that the Zr rich layer plays an important role in the overall interdiffusion behaviour. The interfacial region with lower thickness of Zr-rich layer

M AN U

resulted in formation of thicker UFe2-type layer and the combined thickness of these two layers was ∼ 4.5 µm. In this case, the non-uniform/thinner Zr-rich layer contained approximately 81 at.% Zr, 10 at.% U, 8 at.% Fe and 1 at.% Cr and the uniform/thicker Zr-rich layer had approximately 91 at.% Zr, 3 at.% U, 5 at.% Fe and 1 at.% Cr. The Zr depleted zone on the U−Zr alloy side contains approximately 98 at.% U, 2 at.% Fe. In addition to that the formation of intermetallic phase τ2 (6 at.% Fe, 69 at.% U, 25 at.% Zr) was also revealed in the Zr-depleted

TE D

layer as shown in Fig 10a. The extent of formation of τ2 phase is reduced when the Zr-rich layer is uniform (Fig. 10b). The microstructure of U−10Zr alloy contains δ-UZr2 phase in the α-U matrix. In this case, the δ-phase precipitates are much larger in size as compared to that of the

EP

couple annealed at 823 K.

The microstructural investigations revealed that the width of UFe2-type layer is close to the depth

AC C

of X-ray production from interaction volume of electron and therefore, it may not be accurate to assign that phase based on the EDS result. In this context, the interfacial region of the diffusion couple was further characterized using electron back scattered diffraction (EBSD) technique. It may be noted here that the information depth of EBSD is of several tenths of nm, depending upon the material’s atomic number and density [23]. The diffraction volume in EBSD is significantly smaller than that of the interdiffusion layer and hence the layer can easily be characterized by EBSD. The electron forescatter image of the diffusion couple annealed at 923 K for 1000 h was shown in Fig. 11a. The EBSD patterns (Kikuchi patterns) were taken from the different points across the diffusion layer/interface in point mode analysis and then matched with 10

ACCEPTED MANUSCRIPT

the crystal structure for the possible phases of interest (e.g., UFe2, U6Fe, ZrFe2, Zr, Fe etc.). The experimental and indexed EBSD patterns of interdiffusion layers as shown in Fig. 11(b-g) revealed that the Zr-rich layer is α-Zr phase (space group P63/mmc) and the thin diffusion layer corresponds to UFe2-type phase (space group Fd-3m). The Kikuchi pattern on the T91 steel is

RI PT

indexed with bcc Fe (space group Im-3m). All these indexing were carried out with a mean angle deviation (from the ideal Euler Angle) less than 0.5. The combined EDS and EBSD results

3m) with Zr and Cr present in the phase as solute. 3.3.3 Interfacial reaction in U−10Zr/T91 couple at 973 K

SC

confirmed that the crystal structure of the interface layer is of cubic UFe2-type (space group Fd-

M AN U

The microstructures and intensity profiles of U-Mα, Zr-Lα, Fe-Kα, Cr-Kα X-ray lines of the U−10Zr/T91 diffusion couple annealed at 973 K for 500 h are shown in Fig. 12(a-c). The scheme of the microstructure in the interdiffusion zone was found to be similar to that of the couples annealed at 823 and 923 K. However, the Zr-rich layer formed at 973 K is wider and uniform and as a result of that a thin reaction layer of UFe2-type was formed at the interface. The Zr-rich layer (containing 97 at.% Zr, 2 at.% Fe, 1 at.% U) acts as a barrier layer for the interdiffusion

TE D

between U and Fe/Cr. The width of the UFe2 reaction layer is strongly dependent on the Zr-rich layer. The combined width of these two layers is ∼ 6 µm. Furthermore, the formation of any U−Zr−Fe intermetallic compound on the U−Zr alloy side was not revealed in the microstructural investigation. Rather, a lamellar structure was found to be formed in the U−Zr alloy (Fig. 12b).

EP

The Zr-depleted layer formed on the U−Zr alloy side has no sharp boundary and Zr concentration in this layer is gradually increasing towards the bulk U−Zr alloy. However, liquid

AC C

phase formation was not observed at the interface. 3.3.4 Interfacial reaction in U−10Zr/T91 couple at 1003 K Two types of interfacial reaction zone were observed at the interface of the U−10Zr/T91 diffusion couple heated at 1003 K for 200 h. These are described as (a) interface with stable Zrrich layer and (b) interface with broken Zr-rich layer.

(a) Interface with stable Zr-rich layer

11

ACCEPTED MANUSCRIPT

The microstructure and intensity profiles of U-Mα, Zr-Lα, Fe-Kα, Cr-Kα X-ray lines of this interface are shown in Fig. 13(a-c). The interfacial region revealed the formation of UFe2type, Zr-rich and Zr-depleted layers like other diffusion couples. Here, the Zr-rich layer showed variation in Zr-content across its width. The thickness of the combined UFe2, Zr-rich

RI PT

layers is ∼ 4 µm.

(b) Interface with broken Zr-rich layer

In few places of the interfacial region, the Zr-rich layer was found to be broken and

SC

rounded/curved shape of reaction layer with larger depth of penetration of the constituent elements was observed (Fig. 14a). A magnified view of the region is shown in Fig. 14b. This

M AN U

accelerated rate of diffusion is due to the formation of the liquid phase in those interfacial regions. It appears from the microstructure in Fig. 14a that solidification of the liquid phase formed during the heat treatment leads to formation of a porous layer. The interdiffusion region contains number of reaction layers namely:

I. U-rich layer (80 at.% U, 17 at.% Fe, 3 at.% Cr) with few black precipitates (69 at.% Fe, 18 at.% Cr, 1 at.% Mo and 12 at.% U) which are relatively rich in Cr (on the T91 side),

TE D

II. UFe2 layer with few tiny bright spots of U6Fe phase,

III. Zr-rich layer (87 at.% Zr, 9 at.% U, 3 at.% Fe, 1 at.% Cr), IV. a reaction layer containing U-rich (80 at.% U, 18 at.% Zr, 2 at.% Fe) and Zr-rich (78 at.% Zr, 20 at.% U, 2 at.% Fe) layers,

EP

V. Zr-depleted layer on the U−Zr alloy side.

All the above reaction layers are marked in Fig. 14b. The intensity profiles of U-Mα, Zr-Lα,

AC C

Fe-Kα, Cr-Kα x-ray lines are shown in Fig. 14c.

4. Discussion:

The purpose of this study is to investigate the chemical compatibility between U−Zr based fuel and T91 cladding for future fast breeder reactor. Fuel-cladding chemical interaction (FCCI) is considered as a complex multicomponent diffusion problem. The interdiffusion between fuel and cladding components may result in the formation of relatively low-melting-point compositions in the fuel and reduce mechanical strength of the cladding. In this background, the results of the present study are discussed below. 12

ACCEPTED MANUSCRIPT

4.1 Differential thermal analysis The chemical interaction between U-based metallic fuel and T91 steel cladding could be best evaluated with the help of U−Fe system. In U−Fe system [24], there are two intermetallic

RI PT

compounds i.e., UFe2 which melts congruently at 1503 K and U6Fe which is formed via a peritectic reaction at 1083 K (i.e., L + (γU) → U6Fe). UFe2 has a cubic Laves structure (space group Fd-3m) while U6Fe exhibits a body centered tetragonal structure (space group I4/mcm). The U−Fe system is comprised of two eutectics i.e., L → (γFe) + UFe2 at 1353 K and L → UFe2

SC

+ U6Fe at 998 K; having eutectic compositions of U−52wt%Fe and U−10.2wt%Fe, respectively.

M AN U

In the present study, the U+T91 composite for DTA experiment was comprised of U and T91 in the ratio of around 5:1. This ratio was purposely chosen to focus on the uranium rich part of the system, since the U-rich part of U−Fe system exhibits eutectic reaction only at 998 K. Hence, chemical interaction between U and T91 particularly on the U-rich part of the system may give an important output which will be really useful for the purpose of reactor safety. The liquefaction reaction temperature between fuel and cladding is considered as very critical data. A similar

TE D

weight ratio of U−10Zr and T91 (i.e., 5:1) was also chosen for DTA experiment of U−10Zr+T91 composite. The DTA curve of U+T91 composite contains the transformation peaks for both U and T91 during first heating cycle. The intense exothermic peak observed at 1383 K is attributed to strong chemical reaction between two. In the present case, the U−T91 system is considered as

EP

pseudo binary system. The chemical interaction U and T91 may lead to the formation of UFe2 and/or U6Fe types of intermetallic compounds. The Gibbs free energies of formation (∆G° ) of the

AC C

UFe2 and U6Fe intermetallics were measured by Gardie et al. [25]. According to them, the ∆G° (UFe2) values are -54.1 kJ mol-1 at 998 K and -52.8 kJ mol-1 at 1500 K, whereas the ∆G° (U6Fe) data are -31.3 kJ mol-1 at 998 K and -28.5 kJ mol-1 at 1068 K. These data indicate that the formation of UFe2 is thermodynamically more favourable as compared to U6Fe in the temperature range of our interest. Hence, the intense exothermic peak in the DTA curve of U+T91 composite at 1383 K may be attributed to the formation of UFe2-type compound. The exothermic peak was observed just below melting temperature of uranium. After the melting of unreacted uranium metal, the heating of the composite sample was continued up to 1473 K. The

13

ACCEPTED MANUSCRIPT

whole sample got converted to liquid phase. In the repeat heating run, the first peak was observed at 995 K and the second peak (~1384 K) was not sharp (Fig. 5b). Any peaks correspond to pure U and T91 was not observed in the repeat run. By comparing with the reference phase diagram of U−Fe system [24], it can be concluded that the peak at ~1384 K is

RI PT

due to the liquidus temperature and the peak at 995 K represents the eutectic reaction (i.e., L ↔ UFe2 + U6Fe). The above explanation is further supported by the microstructural investigation. The microstructure of U+T91 reaction product (Fig. 6(a-b)) revealed large equiaxed grains of pro-eutectic UFe2 phase and lamellar eutectic structure of UFe2 and U6Fe phases. As already

SC

mentioned that UFe2 phase has soluble Cr and hence it can also be represented as U(Fe,Cr)2. The sequence of phase transformation during heating and cooling cycles of U+T91 composite is

M AN U

presented in Table. 4.

Similarly, the DTA graph of U−10Zr+T91 composite (Fig. 5c) contains the transformation peaks for both U−10Zr and T91 during first heating cycle and in this case the exothermic reaction peak was identified at 1495 K which is approximately 110 K higher than that of U+T91 composite. U−10Zr alloy exists as single bcc γ-phase from 995 K up to its solidus temperature [17]. The

TE D

intense exothermic peak is most likely due to the formation of UFe2-type compound. The higher reaction temperature is due the higher solidus temperature of U−10Zr alloy than the melting temperature of uranium. The exothermic reaction was followed by an endothermic reaction of melting of unreacted U−10Zr alloy at 1503 K. After that, the DTA curves repeatedly showed

EP

only two peaks (i.e., 995 K and 1384 K). The microstructure of the reaction product exhibited three phases namely U6Fe, UFe2/U(Fe,Cr)2 and Zr(Fe,Cr)2. Since the present study was carried

AC C

out by taking U−10Zr and T91 in the ratio of 5:1, the resultant U−10Zr+T91 composite sample contained approx 75wt%U, 8wt%Zr and 17wt%Fe (where the sum of the concentrations of Fe, Cr, Mo etc. of the steel is taken as one concentration variable). According to the ternary phase diagram of U−Zr−Fe system at 853 K [22], the equilibrium phases of U−Zr−Fe alloy of similar composition are U6Fe, UFe2 and ZrFe2. In fact, similar types of phases were also revealed in the present study. By combining all the results, it can be concluded here that the peak at 995 K corresponds to the liquefaction temperature due to the eutectic type of reaction and the peak at 1384 K indicates the liquidus temperature. The details of all the reactions are shown in Table 4. The equilibrium phase relations in the U−Zr−Fe ternary alloys were investigated by Nakamura et 14

ACCEPTED MANUSCRIPT

al. [26] using DTA and electron probe microanalysis. According to them, the phase transformation of U−7.6Zr−16.3Fe alloy followed the transformations of U6Fe + UFe2 + ZrFe2 → ZrFe2 + L → L at 995 and 1273 K, respectively. This alloy composition is similar to that of our composite sample and therefore, the phase transformation sequence is consistent with the

RI PT

present results. Another important point was evident from their study that the minimum temperature at which the liquid phase appeared i.e., liquefaction temperature was around 994995 K which is almost identical to the eutectic reaction temperature of present study. Kutty et al. [9] investigated the chemical interaction behaviour of metal fuels such as U, U–Zr alloy with

SC

T91 steel using differential thermal analysis (DTA) and differential scanning calorimetry (DSC). In their study, the intense exothermic reactions in the U+T91 and U−10Zr+T91 composites were

M AN U

identified at 1371 K and 1465 K, respectively, which are lower than the present results (i.e., 1383 and 1495 K). The differences may be due to the variation in the initial ratio of the reactants (i.e., U, U−Zr and T91), impurity level etc.

Several important features of the fuel-cladding chemical reaction between U-based fuel and T91 cladding have been revealed from the DTA results of the present investigation. The addition of

TE D

Zr increases the temperature of the exothermic reaction between U-based fuel and T91 cladding. This could be due to the increase in solidus temperature which may indirectly affect the diffusivity. In fact, it was reported that the diffusivity decreases with addition of Zr in U−Zr system [9, 27]. However, the liquefaction temperature has not been altered with the addition of

EP

10wt%Zr. Similarly, alloying elements of T91 grade steel i.e. Cr, Mo, V, Nb etc. have no remarkable effect on the eutectic temperature. In both the U-rich U+T91 and U−10Zr+T91

AC C

composites, the initial liquefaction temperatures also known as eutectic temperatures are found to be identical (i.e., 995 K) and the measured eutectic temperatures are in good agreement with the literature data. Hence, one can conclude here that the presence of Zr (~ 10wt%) in the fuel or Cr, Mo etc. in the cladding steels will not have much effect on eutectic reaction temperature of fuel-cladding interaction since the main contributors to the fuel-cladding eutectic reaction are U of the U-based fuel and Fe of the cladding. 4.2 Interfacial reactions in U−10Zr/T91 couples

15

ACCEPTED MANUSCRIPT

Based on the feedback from the DTA results on chemical reaction between U−10Zr alloy and T91 cladding steel, the isothermal diffusion couple experiments were carried out in the temperature range of 823-1003 K. The temperature and time schedules were purposely chosen so that there is a measurable amount of diffusion after reasonable length of time. The results of

condition as well as in the transient accidental situation.

RI PT

these extensive studies will provide useful output on chemical compatibility at reactor operating

The results of the present study indicated that the interdiffusion between U−10Zr alloy and T91

SC

leads to formation of three layers namely UFe2-type intermetallic, Zr-rich layer and Zr-depleted layer at the interface. Similar type of interdiffusion layers at the interface of diffusion couple

M AN U

between U−19Pu−10Zr alloy fuel and HT9 ferritic steel cladding at 973 K was reported by Hofman and Walters [1]. According to them, a Zr-rich layer containing interstitial nitrogen is first formed at the interface of U–19Pu–10Zr and HT9. Then U and Pu of the fuel alloy diffuse through the Zr-rich layer and interact with the cladding components. This results in formation of U6Fe and UFe2 type phases in a single zone on the cladding side of the Zr-rich layer. Keiser and Dayananda [13] reported that the interdiffusion between U−10Zr alloy and HT9 steel at 973 K

TE D

resulted in formation of similar types of reaction layers i.e., U(Fe,Cr)2, Zr-rich phases at the interface. The diffusion behaviour of U−10Zr alloy and HT9 steel was also investigated by Lee et al. at 973 and 1003 K [15]. According to them, the δ-UZr2 phases in the vicinity of the interface would decompose into U and Zr in the first step. The U in the matrix would interact

EP

with cladding components and form (U,Zr)(Fe,Cr)2 phase. At the same time, the decomposed Zr is saturated to form a Zr-rich layer which acts as an interdiffusion barrier layer. These studies

AC C

indicate that the interdiffusion behaviour of U−Zr and U−Pu−Zr alloys with HT9 steel are consistent with the present observation using T91 steel. Both the HT9 and T91 steels are ferritic in nature and hence, they follow similar interdiffusion behaviour with U−Zr alloy.

All the results indicate that there is a propensity of Zr-segregation towards the interfacial region. The phenomenon of Zr-segregation plays a vital role in the overall interdiffusion mechanism. In a recent work, the segregation in the U−Zr solid solution and the interaction of cladding elements (Fe, Ni, Cr) with U–Zr fuel have been interpreted by atomistic simulations using the concepts of strain and chemical energy underlying the Bozzolo–Ferrante–Smith (BFS) method [28, 29]. The 16

ACCEPTED MANUSCRIPT

segregation in U–Zr solid solution is strongly dependent on lattice parameter, temperature, concentration of Zr and surface energy. According to Bozzolo et al. [28, 29], at low Zr concentration, (0 < xZr < 22.45 (at.%) (U–10 wt% Zr)), a slightly favorable Zr surface energy and small negative lattice deviation from their average value lead to enhanced Zr segregation to the

RI PT

surface as the low value of the bulk lattice parameter increases the BFS strain of Zr atoms. On the other hand, as the concentration of Zr increases the lattice parameter expands above average values, fast approaching bulk Zr values, thus reducing the BFS strain energy of Zr atom and U finds itself in a high BFS strain environment which favours U segregation to the surface. The

SC

enrichment of Zr at the surface of U−8.74Zr (U−20at.%Zr) alloy during vacuum annealing above 550 K has been reported by Paukov et al. [30]. As per their study, a Zr-depleted zone was formed

M AN U

just below the Zr-rich overlayer at surface. The Zr content in the Zr-rich surface layer of was ∼ 80at% in the temperature range of 673-973 K. They have assumed that the Zr-rich phase belongs to the intermediate phase of UZr2+x type. There may be even a tendency to have more Zr-rich material at the very surface, as the atomic diameter of Zr is larger than that of U [30].

Furthermore, U−Zr alloy has different phases at different annealing temperature which may also

TE D

affect the interdiffusion characteristics. Now, careful investigation on the U−10Zr part of the U−10Zr/T91 diffusion couples revealed significant variations in the microstructures depending on the annealed temperature. The microstructural characteristics and phases of U−10Zr alloy are highlighted here based on the present results. The equilibrium phases of U−10Zr alloy at room

EP

temperature are α-U (space group Cmcm) and δ-UZr2 (space group P6/mmm) [17]. The structural and microstructural analysis (Fig. 2 and Fig. 3) of the as-cast U−10Zr alloy revealed

AC C

the α-U phase with very fine lamellar structure and the δ-UZr2 phase could not be resolved. McKeown et al. [31] also reported that the presence of the δ-UZr2 phase in the as-cast U−10Zr alloy cannot be determined from XRD and the coexistence of fine-scale lamellar structure of the α and δ phases was revealed using TEM analysis. However, the U−10Zr side of the diffusion couple annealed at 823 K contains α-U and fine precipitates δ-UZr2 phases (Fig. 9b), which are actually the equilibrium phases at that temperature. Unlike as-cast microstructure, the δ-phase was clearly resolved in the annealed sample due to longer diffusion time. A similar type of microstructure containing α-U and δ-UZr2 phases was also revealed in the microstructure of the 17

ACCEPTED MANUSCRIPT

diffusion couple annealed at 923 K. Although, the equilibrium phases of U−10Zr alloy at 923 K are orthorhombic α-U and bcc γ2 as evident from the U−Zr phase diagram, the δ-phase was formed here via a peritectoid reaction i.e., α + γ2 → δ during cooling from 923 K. A lamellar structure was revealed in the U−10Zr part of diffusion couple at 973 K. The U−10Zr alloy exists

RI PT

as bcc γ phase with some miscibility gap (i.e., γ1 + γ2) at 973 K. The lamellar structure is formed due to invariant reactions (monotectoid, eutectoid) during cooling from γ-phase. An interfacial Zr-depleted layer on U−Zr side is common for all the diffusion couples. Hence, it can be

SC

concluded here that the microstructures of bulk U−10Zr part correspond well with the equilibrium phase diagram and the Zr segregation of U−10Zr alloy was observed only at the peripheral region which is adjacent to the interface. In addition to that the formation of

M AN U

intermetallic phase τ2 was observed in the Zr-depleted layer at 823 and 923 K. No τ2 phase was observed in the diffusion couple heated at 973, 1003 K. The phase diagrams for the U−Fe−Zr ternary system [22] indicate the solubility range of the τ2 phase decreases with increasing temperature and finally decomposes above 999 K. The absence of τ2 phase in the diffusion

TE D

couple at 973 K is because of the uniform Zr-rich diffusion barrier layer.

In this context, an attempt has been made to explain the typical interdiffusional behaviour between U−10Zr and T91 steel by combining the reported and present observations. For low Zr containing U−Zr alloy/U-rich U−Zr alloy, there is a tendency of Zr segregation towards the

EP

surface to minimize the lattice strain of Zr atoms as well as favorable Zr surface energy [28, 29]. The segregation of Zr toward the surface is further accelerated in the presence of interstitial

AC C

impurity atom in the steel, since Zr has strong affinity towards the interstitial impurity atoms. The segregation of Zr towards the surface leads to the formation of Zr-rich layer at the interface. As a consequence of that, a Zr-depleted layer on the U−10Zr side was formed to maintain mass balance of Zr element. As the temperature increases from 823 to 973 K the diffusivity of Zr increases which may result in formation of very strong and thicker Zr-rich layer. It may be noted that the Zr-rich layer contained approximately 97 at.% Zr, 2 at.% Fe, 1 at.% U at 973 K and the EBSD analysis confirmed that the Zr-rich layer is α-Zr phase (space group P63/mmc). It acts as a diffusion barrier by reducing the penetration of Fe and Cr elements towards fuel lattice. The thickness of the interfacial Zr-rich layer at 823 and 923 K is not uniform and as a result, the 18

ACCEPTED MANUSCRIPT

penetration of Fe through it towards fuel lattice leads to the formation of low Fe-containing intermetallic phase τ2 (6 at.% Fe, 67 at.% U, 27 at.% Zr) on the U−Zr side. Due to strong interaction of U with Fe of T91 steel, U diffuses through the Zr rich layer and comes in contact with Fe of the steel. This results in formation of an UFe2-type layer on the T91 side. The

RI PT

thickness of UFe2 layer gradually decreases as the thickness of Zr-rich layer increases with temperature. Another important aspect is that any notable interaction between Zr-rich layer and T91 has not been revealed even though there are many intermetallic compounds, viz. ZrFe2, Zr2Fe and Zr3Fe in the Zr–Fe system [32]. The unusual nature denotes/demonstrates the sluggish

SC

nature of diffusion reaction between Zr and Fe. The diffusion reaction experiments by Bhanumurthy et al. [33] also showed that none of these compounds formed at the Zr/Fe interface

M AN U

up to 1073 K.

When the annealing temperature was increased from 973 to 1003 K, the Zr-rich layer could not effectively prevent the interdiffusion. In some interfacial region, the Zr-rich layer was broken and the interaction between U and Fe leads to the formation liquid phase. In fact, the annealing temperature (1003 K) is slightly above the liquefaction temperature (i.e., 995 K) of U−10Zr+T91

TE D

composite. This particular temperature (i.e., 1003 K) was purposely chosen to investigate whether the Zr-rich layer is able to prevent the liquefaction or not. This result indicates that in any case, if the temperature at fuel-cladding interface reaches above the eutectic liquefaction temperature, the direct contact of U–10Zr and T91 cladding would cause eutectic melting in spite

EP

of the formation of Zr-rich layer. In an earlier study, Kaity et al. [4] have shown that at 1023 K, U–6Zr/T91 couple reacted with each other causing a complete melt down of the cladding in 100

AC C

h. The complete melting of the couple was not observed in the present case. The diffusion couple experiment between U and T91 steel with a Zr foil of thickness ~200 µm between them was also carried out by Kaity et al. [4]. It was observed that the chemical compatibility between U and T91 is significantly improved by the Zr liner and liquefaction due the eutectic reaction was not observed even at 1023 K. Ryu et al. [16, 34] reported that the interdiffusion between U–10Zr alloy and ferritic martensitic steels (FMS) was inhibited effectively by the presence of thin Zr-foil (∼20-30 µm) or thin Zr-layer (∼6 µm) even above the eutectic melting temperature of U–Zr alloys and FMS. Eutectic melting between the fuel and 19

ACCEPTED MANUSCRIPT

steel was not observed in the diffusion couples even after annealing at 1073 K for 25 h. These are the important observations in support of the concept of using Zr foil as a diffusion barrier between fuel and cladding.

RI PT

Several important features have been revealed from the results of diffusion couple experiment of the present study. The Zr plays most crucial role in the interdiffusion between U−Zr based fuel and T91 steel cladding. There are few direct and indirect effects of Zr-segregation at the interface. The Zr-segregation leads to the formation of Zr-depleted layer. Since Zr increases the

SC

solidus temperature of U−Zr alloy, it is expected that the solidus temperature may decrease at the Zr-depleted layer. At the same time, the thermal conductivity may increase/improve at the Zr-

M AN U

depleted layer. In fact, these effects will be further complicated during actual reactor operation by the presence of other fission products which may also play some role in the interdiffusional behaviour. Hence, it can be concluded based on the present results that the compatibility of U−Zr alloy and T91 steel cladding seems to be good at reactor operating temperature due to the formation of Zr-rich barrier layer at the interface. However, at any accidental condition if the fuel-cladding interfacial temperature reaches above the eutectic liquefaction temperature (995

TE D

K), the U−Zr fuel is no more compatible with T91 cladding.

4.2.1 Kinetics of the interfacial reaction

The interaction between U−Zr alloy and T91 steel is a multi-component diffusion problem. As a

EP

consequence of Zr segregation towards the surface, there is a gradient of Zr concentration in the Zr-depleted layer from interfacial region to the bulk U−Zr alloy. Furthermore, there is no sharp

AC C

boundary for the Zr-depleted layer. With this in mind, an attempt has been made to calculate the growth kinetics of reaction layers based on the total thickness of UFe2-type and Zr-rich layers. Here, the combined thickness of the two layers is considered as width of reaction layer. The measured width of reaction layers (i.e., UFe2 + Zr-rich layers) are 1.5, 4.5 and 6.0 µm at 823, 923 and 973 K, respectively. The growth of the reaction layers at the interface of diffusion couples are generally diffusion controlled and hence follow a parabolic relationship [12, 35]. The growth of the reaction layer at the interface can therefore be expressed as [36]: 20

ACCEPTED MANUSCRIPT

w2 = kt

(1)

where, w is width of the reaction layer, t is time of the reaction, k is diffusion controlled rate constant. The temperature dependence of the rate constant k can be expressed as: k = k0 exp (-Qk/RT)

(2)

RI PT

where, k0 is the pre-exponential factor, Qk is the activation energy for growth, R is the universal gas constant and T is the temperature in absolute scale. Combining Eqn (1) and (2), w2/t = k0 exp (-Qk/RT)

(3)

SC

The activation energy Q can be determined from the slope of log (w2/t) versus 1/T plot. The preexponential factor k0 can be calculated from the intercept of the plot. The plot of log (w/t1/2) versus 1/T is presented in Fig. 15. The activation energy Q for the reaction was found to be 170

M AN U

kJ mole-1 and the pre-exponential factor k0 was found to be 2.698 × 10-8 m2 s-1. Hence, the growth of the interdiffusion layers at the interface can be represented as: w2 = 2.698 × 10-8 t exp (-170 kJ mol-1/RT)

(823 ≤ T/K ≤ 973) (4)

Park et al. [12] reported the diffusional interaction between U−10Zr and Fe in the temperature

TE D

range of 903−953 K where the interdiffusion zone had various intermetallic phases and the activation energy of the entire diffusion zone was 296.6 kJ mol-1. Huang et al. [37] reported the growth rate constants and the activation energies for U/Fe, U/Fe−15Cr and U/Fe–15Cr–15Ni (wt%) diffusion couples where the interdiffusion zone had two reaction layers of U6Fe and UFe2

EP

intermetallic phases. According to them, the activation energy for total interdiffusion zone (i.e., U6Fe + UFe2 layers) of U/Fe, U/Fe−15Cr and U/Fe–15Cr–15Ni couple are 137, 231 and 135 kJ

AC C

mol-1, respectively. The interdiffusion and reactions in the U/Fe and U/Fe–Cr–Ni showed similar temperature dependence, while the U/Fe–Cr diffusion couples exhibited a larger magnitude of activation energy for growth constants. The activation energy of the present study is relatively closer to that of U/Fe, U/Fe–15Cr–15Ni and U/Fe−15Cr couples. However, it is difficult to compare the activation energy of the present study with the above literature data, since the diffusion behaviour between U−Zr and T91 is different from those couples due to the formation Zr-rich diffusion barrier layer at the interface and no separate reaction zone of U6Fe and UFe2 phases were formed here. Moreover, the diffusion kinetics in the present study was calculated based on the diffusion layer of the UFe2 phase and Zr-rich layer. In this context, it may be noted 21

ACCEPTED MANUSCRIPT

that the data on the diffusion kinetics between U−Zr and ferritic steel are limited in the literature. Based on the literature and present studies, it was observed that the interdiffusion zone at the interface of diffusion couple between U−Zr alloy and ferritic steels (T91, HT9) contained fewer

U−Zr alloy and Fe, Fe−Cr, Fe−Ni, Fe−Ni−Cr, 316SS and D9.

5. Summary and conclusions:

RI PT

numbers of phases with smaller width than the reaction layers formed at the interfaces between

The present investigation provides useful output on the chemical compatibility between U−Zr

SC

based metallic fuels and T91 steel cladding. The following important observations could be made from the results of the present investigation:

A strong exothermic reaction occurred between U/U−Zr alloy and T91 steel just below

M AN U

•

the melting/solidus temperature of U/U−Zr alloy could be attributed to the formation of UFe2-type of compound. The presence of Zr shifts the exothermic reaction temperature to higher value. •

The eutectic liquefaction temperatures between U/U−Zr alloy and T91 steel are found to be identical (i.e., 995 K).

Microstructural evaluation of the reaction product reveals the presence of U6Fe,

TE D

•

UFe2/U(Fe,Cr)2 and Zr(Fe,Cr)2 phases. •

The interdiffusion between U–10Zr and T91 resulted in the formation of three different layers at the interface: UFe2-type intermetallic, a Zr-rich layer and a Zr-depleted layer. Zr as an alloying element plays most important role in fuel-cladding interdiffusion by

EP

•

forming of Zr-rich diffusion barrier layer at the interface. The growth kinetics of reaction zone was determined using the total width of UFe2 and

AC C

•

Zr-rich reaction layers and the activation energy Q for the reaction was found to be 170 kJ mol-1.

•

The chemical compatibility of U−Zr alloy and T91 steel cladding seems to be good at

reactor operating temperature due to the formation of Zr-rich barrier layer at the interface. However, at any accidental condition if the fuel-cladding interfacial temperature reaches above the eutectic liquefaction temperature (995 K), the U−Zr fuel is no more compatible

with T91 cladding due to formation of liquid phase.

22

ACCEPTED MANUSCRIPT

References [1] G.L. Hofman, L.C. Walters, Metallic fast reactor fuels, in: B.R.T. Frost (Ed.), Vol. 10A, Nuclear Materials, Materials Science and Technology, VCH Publishers Inc., New York, 1994, pp. 1–43.

RI PT

[2] R.G. Pahl, D.L. Porter, C.E. Lahm, G.L. Hofman, Experimental studies of U-Pu-Zr fast reactor fuel pins in the experimental breeder reactor-ll, Metall. Mater. Trans. A 21A (1990) 1863-1870.

perspective, J. Nucl. Mater. 371 (2007) 202-231.

SC

[3] D.C. Crawford, D.L. Porter, S.L. Hayes, Fuels for sodium-cooled fast reactors: US

[4] S. Kaity, T.R.G. Kutty, R. Agarwal, A. Laik, A. Kumar, Chemical compatibility of uranium

M AN U

based metallic fuels with T91 cladding, Nucl. Eng. Des. 250 (2012) 267-276. [5] S. Kaity, J. Banerjee, M.R. Nair, K. Ravi, S. Dash, T.R.G. Kutty, A. Kumar, R.P. Singh, Microstructural and thermophysical properties of U–6wt.% Zr alloy for fast reactor application, J. Nucl. Mater. 427 (2012) 1-11.

[6] R.L. Klueh, A.T. Nelson, Ferritic/martensitic steels for next-generation reactors, J. Nucl. Mater. 371 (2007) 37-52.

TE D

[7] D.D. Keiser Jr., M.A. Dayananda, Interdiffusion between U-Zr fuel and selected Fe-Ni-Cr alloys, J. Nucl. Mater. 200 (1993) 229-243.

[8] A.B. Cohen, H. Tsai, L.A. Neimark, Fuel/cladding compatibility in U-19Pu-10Zr/HT9-clad fuel at elevated temperatures, J. Nucl. Mater. 204 (1993) 244-251.

EP

[9] T.R.G. Kutty, S.K. Rakshit, A. Laik, A. Kumar, H.S. Kamath, C.B. Basak, Studies of the reaction between metal fuels and T91 cladding, J. Nucl. Mater. 412 (2011) 53-61.

AC C

[10] T. Ogata, M. Kurata, K. Nakamura, A. Itoh, M. Akabori, Reactions between U–Zr alloys and Fe at 923 K, J. Nucl. Mater. 250 (1997) 171-175. [11] K. Nakamura, T. Ogata, M. Kurata, A. Itoh, M. Akabori, Reactions of U–Zr alloy with Fe and Fe–Cr alloy, J. Nucl. Mater. 275 (1999) 246-254. [12] Y. Park, K. Huang, A. Paz Y Puente, H.S. Lee, B.H. Sencer, J.R. Kennedy, Y.H. Sohn, Diffusional Interaction Between U-10 wt pct Zr and Fe at 903 K, 923 K, and 953 K (630° C, 650° C, and 680° C), Metall. Mater. Trans. A 46A (2015) 72-82. [13] D.D. Keiser Jr., M.A. Dayananda, Interdiffusion between U-Zr fuel vs selected cladding steels, Metall. Mater. Trans. A 25A(8) (1994) 1649-1653. 23

ACCEPTED MANUSCRIPT

[14] D.D. Keiser Jr., M.C. Petri, Interdiffusion behavior in U-Pu-Zr fuel versus stainless steel couples, J. Nucl. Mater. 240 (1996) 51-61. [15] C.T. Lee, H. Kim, T.K. Kim, C.B. Lee, Diffusion behavior in an interface between U–10Zr alloy and HT-9 steel, J. Nucl. Mater. 395 (2009) 140-144.

RI PT

[16] H.J. Ryu, B.O. Lee, S.J. Oh, J.H. Kim, C.B. Lee, Performance of FCCI barrier foils for U– Zr–X metallic fuel, J. Nucl. Mater. 392 (2009) 206-212.

[17] H. Okamoto, U-Zr (uranium-zirconium), J. Phase Equil. 13(1) (1992) 109-110.

[18] E. Beahm, C. Culpepper, O. Cavin, Laves phases of uranium and 3d transition metals, J.

SC

Less Common Met. 50 (1976) 57-71.

[19] S.S. Makridis, C. Christodoulou, M. Konstantakou, T.A. Steriotis, M. Daniil, A. Ioannidou,

M AN U

E. Kikkinidis, A.K. Stubos, Intermetallic hydrides based on (Zr-Ti)(Fe-Cr)2 type of compounds, Mater. Sci. Forum 514-516 (2006) 666-671.

[20] O. Canet, M. Latroche, F. Bourée-Vigneron, A. Percheron-Guégan, Structural study of Zr (Cr1− xFex)2Dy (0.4⩽ x⩽ 0.75; 2< y< 3) by means of neutron powder diffraction, J. Alloys Compd. 210 (1994) 129-134.

[21] J. Coaquira, H. Rechenberg, J. Mestnik Filho, Structural and Mössbauer spectroscopic study

(1999) 42-49.

TE D

of hexagonal Laves-phase Zr (FexCr1− x) 2 alloys and their hydrides, J. Alloys compd. 288

[22] O. Fabrichnaya, Iron–Uranium–Zirconium, in: G. Effenberg, S. Ilyenko (Eds.), Vol. 11, Ternary Alloy Systems Phase Diagrams, Crystallographic and Thermodynamic Data, Subvolume

EP

C, Non-Ferrous Metal Systems, Springer Berlin Heidelberg, New York, 2007, pp. 320-327. [23] A.J. Schwartz, M. Kumar, B.L. Adams, D.P. Field, Electron backscatter diffraction in

AC C

materials science, Second ed., Springer 2000. [24] H. Okamoto, Fe-U (Iron-Uranium), Phase Diagrams of Binary Iron Alloys (1993) 429-432. [25] P. Gardie, G. Bordier, J. Poupeau, J. Le Ny, Thermodynamic activity measurements of U-Fe and U-Ga alloys by mass spectrometry, J. Nucl. Mater. 189 (1992) 85-96. [26] K. Nakamura, M. Kurata, T. Ogata, A. Itoh, M. Akabori, Equilibrium phase relations in the U–Zr–Fe ternary system, J. Nucl. Mater. 275 (1999) 151-157. [27] Y. Adda, J. Philibert, H. Faraggi, Étude des phénomènes de diffusion intermétallique dans le système uranium-zirconium, Revue de metallurgie 54(8) (1957) 597-610.

24

ACCEPTED MANUSCRIPT

[28] G. Bozzolo, H. Mosca, A. Yacout, G. Hofman, Y. Kim, Surface properties, thermal expansion, and segregation in the U–Zr solid solution, Comput. Mater. Sci. 50 (2010) 447-453. [29] G. Bozzolo, H. Mosca, A. Yacout, G. Hofman, Atomistic modeling of the interaction of cladding elements (Fe, Ni, Cr) with U–Zr fuel, J. Nucl. Mater. 414 (2011) 101-108.

RI PT

[30] M. Paukov, I. Tkach, F. Huber, T. Gouder, M. Cieslar, D. Drozdenko, P. Minarik, L. Havela, U-Zr alloy: XPS and TEM study of surface passivation, Appl. Surf. Sci. 441 (2018) 113119.

[31] J.T. McKeown, S. Irukuvarghula, S. Ahn, M.A. Wall, L.L. Hsiung, S. McDeavitt, P.E.A.

SC

Turchi, Coexistence of the α and δ phases in an as-cast uranium-rich U–Zr alloy, J. Nucl. Mater. 436 (2013) 100-104.

(1988) 597-604.

M AN U

[32] D. Arias, J. Abriata, The Fe−Zr (Iron-Zirconium) system, Bull. Alloy Phase Diagrams 9(5)

[33] K. Bhanumurthy, G. Kale, S. Khera, Reaction diffusion in the zirconium-iron system, J. Nucl. Mater. 185 (1991) 208-213.

[34] J.H. Kim, H.J. Ryu, J.H. Baek, S.J. Oh, B.O. Lee, C.B. Lee, Y.S. Yoon, Performance of a diffusion barrier under a fuel–clad chemical interaction (FCCI), J. Nucl. Mater. 394 (2009) 144-

TE D

150.

[35] C. Matthews, C. Unal, J. Galloway, D.D. Keiser Jr, S.L. Hayes, Fuel-Cladding Chemical Interaction in U-Pu-Zr Metallic Fuels: A Critical Review, Nucl. Technol. 198 (2017) 231-259. [36] H.J. Ryu, S.J. Oh, C.T. Lee, K.H. Kim, Y.M. Ko, Y.M. Woo, S.H. Kim, C.H. Hahn, B.O.

EP

Lee, C.B. Lee, Fuel-Cladding Chemical Interaction between U-Zr-X Metallic Fuel and Ferritic Martensitic Steels, Transactions of the Korean Nuclear Society Autumn Meeting (2007) 245-

AC C

246.

[37] K. Huang, Y. Park, L. Zhou, K. Coffey, Y. Sohn, B. Sencer, J. Kennedy, Effects of Cr and Ni on interdiffusion and reaction between U and Fe–Cr–Ni alloys, J. Nucl. Mater. 451 (2014) 372-378.

25

ACCEPTED MANUSCRIPT

List of Figures Fig. 1 Schematic diagram of the diffusion couple-fixture assembly for U−10Zr/T91diffusion couple.

orthorhombic α-U phase.

RI PT

Fig. 2 XRD pattern of as-cast U–10Zr alloy sample. All the peaks correspond to the

representative EDS spectrum with chemical composition.

SC

Fig. 3(a, b & c) SEM micrographs of as-cast U−10Zr alloy at different magnifications, (d)

M AN U

Fig. 4 DTA curves of U metal, T91 steel, U−10Zr alloy during heating cycle.

Fig. 5(a & b) DTA curves of U+T91 composite, and (c & d) DTA curves of U−10Zr+T91 composite during both heating and cooling cycles with a scanning rate of 10 K min-1. Fig. 6(a & b) Microstructures of the reaction product of U+T91 composite at different

TE D

magnification and (c & d) Microstructures of the reaction product of U−10Zr+T91 composite at different magnification showing different phases (i.e., bright, grey and dark).

Fig. 7(a-b) Electron forescatter images of the U+T91 after DTA run; (c-d) experimental EBSD

EP

patterns of phases as indicated by arrows; (e-f) indexed EBSD patterns on c and d correspond to UFe2 (Fd-3m) and U6Fe (I4/mcm) phases, respectively; (g-h) EDS spectra of respective phases

AC C

with their chemical composition.

Fig. 8(a-b) Electron forescatter images of U−10Zr+T91 after DTA run; (c-e) experimental EBSD patterns of phases as indicated by arrows; (f-h) indexed EBSD patterns on c, d and e correspond to Zr(Fe,Cr)2 (P63/mmc), UFe2 (Fd-3m) and U6Fe (I4/mcm) phases, respectively; (i-K) EDS spectra of respective phases with their chemical composition showing solubilities of U in Zr(Fe,Cr)2 and Zr in U(Fe,Cr)2.

26

ACCEPTED MANUSCRIPT

Fig. 9 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 823 K for 1500 h, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line AB marked in Fig. 9b.

RI PT

Fig. 10 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 923 K for 1000 h, (c & d) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line CD and EF marked in Fig. 10a and Fig. 10b, respectively.

SC

Fig. 11(a) The electron forescatter image of the diffusion couple annealed at 923 K for 1000 h; (b-d) the experimental EBSD patterns of the diffusion layers as indicated by arrows; (e-g) the

M AN U

indexed EBSD patterns on b, c and d correspond to Zr (P63/mmc), UFe2-type phase (Fd-3m) and Fe (Im-3m), respectively.

Fig. 12 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 973 K for 500 h, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line GH marked in

TE D

Fig. 12b.

Fig. 13 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 1003 K for 200 h, showing the interface with stable Zr-rich layer, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and

EP

Cr-Kα X-ray lines along the line IJ marked in Fig. 13b.

Fig. 14 (a) SEM image of U−10Zr/T91 diffusion couple annealed at 1003 K for 200 h, showing

AC C

the interface with broken Zr-rich layer, (b) magnified image marked with different reaction layers, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line KL marked in Fig. 14b.

Fig. 15 Plot of w2/t versus 1/T for the combined width of UFe2-type and Zr-rich reaction layers.

27

ACCEPTED MANUSCRIPT

Table 1 Chemical composition of T91 cladding steel (wt %)

RI PT

List of Tables

Mn

P

S

Si

Cu

Ni

Cr

Mo

Al

0.099

0.368

0.015

0.0013

0.209

0.068

0.197

9.161

0.882

0.008

Nb

V

Ti

0.079

0.207

0.003

AC C

EP

TE D

M AN U

SC

C

28

N

Fe

0.0457 balance

ACCEPTED MANUSCRIPT

Table 2 Annealing temperature and time of the diffusion couple experiments

Time (h) 1500 1000 500 200

RI PT

Couple Temperature (K) U−10Zr/T91 823 923 973 1003

AC C

EP

TE D

M AN U

SC

No. 1. 2. 3. 4.

29

ACCEPTED MANUSCRIPT

Table 3 The transformation temperatures obtained from DTA curves for uranium metal, T91

RI PT

steel and U−10Zr alloy samples during heating

DTA peak temp (K)

Origin of Peak

U

941 1046 1393 1015

α-U → β-U β-U → γ-U Melting point Curie temperature, (ferromagnetic → paramagnetic) ferrite (α) → austenite (γ)

T91

SC

Sample

1111

AC C

EP

TE D

1510

α + δ → α + γ2 α + γ2 → γ (i.e., α + γ2 → β + γ2 → γ1 + γ2 → γ) Solidus temperature

M AN U

860 958

U−10Zr

30

ACCEPTED MANUSCRIPT

Table 4 The sequences of phase transformations for U+T91 and U−10Zr+T91 composites samples during heating and cooling at scanning rate of 10 K min-1

Cooling cycle Heating cycle

U+T91* (Repeat)

U−10Zr+T91**

Cooling cycle Heating cycle

Nature of peak

Endothermic Exothermic Endothermic Exothermic

Endothermic

Exothermic

Endothermic

TE D

Heating cycle

AC C

U−10Zr+T91** (Repeat)

EP

Cooling cycle

1015 1111 1495 1503 1383 978

Cooling cycle

Exothermic Endothermic Exothermic Exothermic

995

∼1384 (not sharp) 1384 978

Origin of Peak

RI PT

U+T91*

DTA peak temp (K) 941 1015 1046 1111 1383 1387 1313 988 995 ~1314 (not sharp) 1314 988 861 958

α-U → β-U Curie temperature (T91) β-U → γ-U ferrite (α) → austenite (γ) (T91) Reaction peak for UFe2 formation Melting peak Liquidus (L → UFe2 + L) UFe2 + L → UFe2 + U6Fe UFe2 + U6Fe → UFe2 + L UFe2 + L → L

SC

Heating sequence Heating cycle

M AN U

Sample

Endothermic

Exothermic

Liquidus (L → UFe2 + L) UFe2 + L → UFe2 + U6Fe α + δ → α + γ2 (U−10Zr) α + γ2 → γ (U−10Zr) (i.e., α + γ2 → β + γ2 → γ1 + γ2 → γ) Curie temperature (T91) Ferrite (α) → austenite (γ) (T91) Reaction peak for UFe2 formation Melting peak Liquidus (L → Zr(Fe,Cr)2 + L) L + Zr(Fe,Cr)2 → UFe2 + U6Fe + Zr(Fe,Cr)2 UFe2 + U6Fe + Zr(Fe,Cr)2 → L + Zr(Fe,Cr)2 L + Zr(Fe,Cr)2 → L Liquidus (L → Zr(Fe,Cr)2 + L) L + Zr(Fe,Cr)2 → UFe2 + U6Fe + Zr(Fe,Cr)2

* For U+T91: UFe2 phase contains soluble Cr, ** For U−10Zr+T91: UFe2 phase contains soluble Zr and Cr; Zr(Fe,Cr)2 contains soluble U

31

ACCEPTED MANUSCRIPT

Tantalum foil (0.1 mm) T91 (0.5 mm)

M AN U

SC

RI PT

List of Figures

SS packing

TE D

U−10Zr (3 mm)

T91 (0.5 mm)

Fixture (Inconel 600)

1

Schematic

AC C

Fig.

EP

Hole for evacuation & flushing

diagram

of

the

U−10Zr/T91diffusion couple.

32

diffusion

couple-fixture

assembly

for

ACCEPTED MANUSCRIPT

3000

45

50

65

75

(040)

60

70

80

85

(114)

SC (131)

55

(042) (221)

40

TE D

35

(023) (113) (132)

0 30

(130)

(022)

500

(020)

1000

M AN U

(112)

1500

(111)

(110) (002)

Counts (au)

2000

(133)

2500

RI PT

(021)

Orthorhombic (α) phase

90

2θ (degree)

EP

Fig. 2 XRD pattern of as-cast U–10Zr alloy sample. All the peaks correspond to the

AC C

orthorhombic α-U phase.

33

ACCEPTED MANUSCRIPT

(a)

M AN U

SC

RI PT

(b)

9.95%Zr,

(d)

90.05%U

AC C

EP

TE D

(c)

Fig. 3(a, b & c) SEM micrographs of as-cast U−10Zr alloy at different magnifications, (d) representative EDS spectrum with chemical composition.

34

ACCEPTED MANUSCRIPT

20

1510 K

860 K 958 K

5

1046 K

U 0

1393 K

941 K

-5

-10

SC

Heat Flow (µV)

RI PT

U-10Zr 10

T91

M AN U

Exo

15

1015 K 1111 K

-15 -20 500

700

900

1100

1300

1500

TE D

Temperature (K)

AC C

EP

Fig. 4 DTA curves of U metal, T91 steel, U−10Zr alloy during heating cycle.

35

ACCEPTED MANUSCRIPT

160

20

140

(a)

Heating Cooling

U + T91

15

Heating Cooling

U + T91 (Repeat run)

(b)

-10

995 K

5

988 K

0

-5

-10

-20

-15 800

900

1000

1100

1200

1300

Temperature (K)

1400

700

800

900

M AN U

700

1314 K

RI PT

1387 K

988 K

1313 K

0

10

SC

1046 K

10

1111 K

1015 K

941 K

Heat Flow (µV)

20

Heat Flow (µV)

1383 K

Exo

Exo

30

1000

1100

1200

1300

1400

Temperature (K)

30

Heating Cooling

U-10Zr + T91 30

(c)

25

Heating Cooling

U-10Zr + T91 (Repeat run)

(d)

-10 700

800

900

EP

Temperature (K)

15

995 K

Exo

10 5

1384 K

0

978 K

1503 K

1000 1100 1200 1300 1400 1500 1600

Heat Flow (µV)

1495 K 1383 K

TE D

978 K

0

1111 K

1015 K

10

958 K

20

861 K

Heat Flow (µV)

Exo

20

-5 -10 -15 -20 700

800

900

1000

1100

1200

1300

1400

1500

1600

Temperature (K)

Fig. 5(a & b) DTA curves of U+T91 composite, and (c & d) DTA curves of U−10Zr+T91

AC C

composite during both heating and cooling cycles with a scanning rate of 10 K min-1.

36

ACCEPTED MANUSCRIPT

(b)

M AN U

SC

RI PT

(a)

(d)

AC C

EP

TE D

(c)

Fig. 6(a & b) Microstructures of the reaction product of U+T91 composite at different magnification, (c & d) microstructures of the reaction product of U−10Zr+T91 composite at different magnification showing different phases (i.e., bright, grey and dark).

37

ACCEPTED MANUSCRIPT

(b)

(d)

M AN U

SC

(c)

RI PT

(a)

(f)

TE D

(e)

UFe2/U(Fe,Cr)2 (g)

(at.%) 84%U 14%Fe

(h)

AC C

EP

(at.%) 35%U 57%Fe 8%Cr

U6Fe

Fig. 7(a-b) Electron forescatter images of the U+T91 after DTA run; (c-d) experimental EBSD patterns of phases as indicated by arrows; (e-f) indexed EBSD patterns on c and d correspond to UFe2 (Fd-3m) and U6Fe (I4/mcm) phases, respectively; (g-h) EDS spectra of respective phases with their chemical composition. 38

ACCEPTED MANUSCRIPT

(b)

SC

RI PT

(a)

(c)

Zr(Fe,Cr)2

(i)

AC C

EP

(at.%) 5%U 32%Zr 48%Fe 15%Cr

M AN U

(h)

(g)

TE D

(f)

(e)

(d)

U6Fe

UFe2/U(Fe,Cr)2 (at.%)

28%U 7%Zr 61%Fe 4%Cr

(j)

(at.%)

(k)

84%U 14%Fe

Fig. 8(a-b) Electron forescatter images of U−10Zr+T91 after DTA run; (c-e) experimental EBSD patterns of phases as indicated by arrows; (f-h) indexed EBSD patterns on c, d and e correspond to Zr(Fe,Cr)2 (P63/mmc), UFe2 (Fd-3m) and U6Fe (I4/mcm) phases, respectively; (i-K) EDS spectra of respective phases with their chemical composition showing solubilities of U in Zr(Fe,Cr)2 and Zr in U(Fe,Cr)2. 39

ACCEPTED MANUSCRIPT

(a)

(b)

RI PT

UFe2 Zr-rich layer

δ-UZr2

B

SC

A

U-Zr

U Zr Fe Cr

(c)

EP

TE D

Counts (a u)

τ2

Zr-depleted

M AN U

T91

AC C

0

2

4

6

8

10

12

14

16

18

Distance (µm)

Fig. 9 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 823 K for 1500 h, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line AB marked in Fig. 9b.

40

ACCEPTED MANUSCRIPT

τ2

(b) δ-UZr2

Zr-rich layer UFe2

Zr-rich layer

Zr-depleted C

D E

T91

SC

Zr-depleted U-Zr

F

U-Zr

M AN U

T91

U Zr Fe Cr

(c)

(d)

0

5

10

15

EP

TE D

Coutns (a u)

Counts (a u)

U Zr Fe Cr

RI PT

UFe2

(a)

δ-UZr2

20

25

30

35

0

Distance (µm)

5

10

15

20

25

30

35

AC C

Distance (µm)

Fig. 10 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 923 K for 1000 h, (c & d) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line CD and EF marked in Fig. 10a and Fig. 10b, respectively.

41

ACCEPTED MANUSCRIPT

Zr-rich layer

Reaction layer

RI PT

(a) T91

M AN U

SC

U-Zr

TE D

(b)

(d)

(f)

(g)

Zr

AC C

EP

(e)

(c)

UFe2

Fe

Fig. 11(a) The electron forescatter image of the diffusion couple annealed at 923 K for 1000 h; (b-d) the experimental EBSD patterns of the different layers as indicated by arrows; (e-g) the indexed EBSD patterns on b, c and d correspond to Zr (P63/mmc), UFe2-type phase (Fd-3m) and Fe (Im-3m), respectively.

42

ACCEPTED MANUSCRIPT

(a)

(b)

RI PT

UFe2 Zr-rich layer

Lamellar structure

SC

Zr-depleted G U-Zr

T91

M AN U

T91

U-Zr

(c)

TE D

EP

T91

Counts (a u)

U Zr Fe Cr

H

2

4

6

8

10

12

14

16

18

20

22

Distance (µm)

AC C

0

Fig. 12 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 973 K for 500 h, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line GH marked in Fig. 12b.

43

RI PT

ACCEPTED MANUSCRIPT

(a)

(b)

UFe2 Zr-rich layer

II

SC

J

M AN U

Zr-depleted

U-Zr

T91

U-Zr

T91 T91

U Zr Fe Cr

AC C

EP

Counts (a u)

TE D

(c)

0

2

4

6

8

10

12

14

16

Distance (µm)

Fig. 13 (a & b) SEM image of U−10Zr/T91 diffusion couple annealed at 1003 K for 200 h, showing the interface with stable Zr-rich layer, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line IJ marked in Fig. 13b.

44

ACCEPTED MANUSCRIPT

I

II

T91

IV

V (b)

RI PT

(a)

III

K

SC

L

M AN U

Liquefaction

U-Zr

T91

(c)

EP

TE D

Counnts (a u)

U Zr Fe Cr

2

4

6

8

10

12

14

16

18

20

22

Distance (µm)

AC C

0

Fig. 14 (a) SEM image of U−10Zr/T91 diffusion couple annealed at 1003 K for 200 h, showing the interface with broken Zr-rich layer, (b) magnified image marked with different reaction layers, (c) intensity profiles of U-Mα, Zr-Lα, Fe-Kα and Cr-Kα X-ray lines along the line KL marked in Fig. 14b.

45

10

-17

10

-18

10

-19

RI PT

-16

SC

10

1.00

TE D

2

M AN U

2

w / t (m /s)

ACCEPTED MANUSCRIPT

1.05

1.10

1.15

1.20

1.25

3

EP

1/T (x10 ) (1/K)

AC C

Fig. 15 Plot of w2/t versus 1/T for the combined width of UFe2-type and Zr-rich reaction layers.

46