Titanium Evolution and Nickel Restoration Under
Neutron Irradiation in Ni/Ti Multilayers
B. Ballot, A. Menelle, J. Mimault, T. Girardeau, F. Samuel, K. Al Usta
To cite this version:
B. Ballot, A. Menelle, J. Mimault, T. Girardeau, F. Samuel, et al.. Titanium Evolution and
Nickel Restoration Under Neutron Irradiation in Ni/Ti Multilayers. J. Phys. IV, 1995, 05 (C3),
pp.C3-305-C3-310. <10.1051/jp4:1995331>. <jpa-00253697>
HAL Id: jpa-00253697
https://hal.archives-ouvertes.fr/jpa-00253697
Submitted on 1 Jan 1995
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
JOURNAL DE PHYSIQUE IV
Colloque C3, supplBment au Journal de Physique 111, Volume 5, avril 1995
Titanium Evolution and Nickel Restoration Under Neutron Irradiation in
NiITi Multilayers
B. Ballot, A. Menelle, J. Mimault*, T. Girardeau*, F. Samuel** and K. A1 Usta**
Luboratoire Lkon Brillouin, CEA-CNRS, Bht. 563, CE Saclay, 91191 Gif-sur-Yvette cedex, France
* Luboratoire de Me'tallurgie Physique, URA 131 du CNRS, 40 avenue du Recteur Pineau, 86022 Poitiers
cedex, France
** Compagnie Zndustrielle des Lasers, B.P. 27, route de Nozay, 91460 Marcoussis, France
Abstract : N m i multilayers are used as supermirrors for neutron guides. These multilayers are
submitted to neutron irradiation. In order to determine the effect of irradiation on supermirrors
performances we studied Ni/'Ti multilayers irradiated with thermal neutrons. We present here the
results obtained by neutron reflectivity, X-ray diffraction and EXAFS on one sample. It has been
found that the supermirrors performances are not reduced. Nevertheless, Ti shows an evolution
from the hcp structure to another crystalline state while Ni layers show a restoration of the fcc
structure.
1. INTRODUCTION
The development of the use of neutrons in solid state leads to an important need of neutron beams. The
number of these latter disposed around an experimental reactor can be markedly increased by using neutron
guides. Neutron guides are glass tubes of which the inside surface is coated with a neutron reflecting
material. The best coating used at the moment are supermirrors which are made of aperiodic layer of
The two materials we used for
reflector and spacer materials alternately deposited on the glass sub~trate'.~.
our neutron guides are Ni and Ti because of their very important index
As the coating stack are
submitted to neutron irradiation, we are interested to determine the effects of such radiation on supermirrors
performances. Reference and irradiated NiRi multilayers samples have thus been studied by neutron
reflectivity, X-ray diffraction and EXAFS.
2. SAMPLES
The samples used in this study are periodic stacks of 10 bilayers (monochromators) of Titanium and Nil.,C,
(, is of the order of 0.05) deposited on Silicon wafers of 3 mrn thick. The use of Nil.,C, instead of pure Ni is
due to the N m i interface antidiffusion barrier property of the alloy5; in the following text, the Nil.,C, layers
will be called NiC. The thickness of one bilayer (also called period) is 200A (100W Ti and 100W NiC). The
fist layer deposited on the substrate is Ti. The last layer before air is NiC.
One of the samples has been irradiated with thermal neutrons of which energy is included between 5 and
80 meV. The instant flux was 4 ~ 1 0 ~ ~ n . c m The
~ ~ . irradiation
s~'.
time is 173 hours, that means a fluence of
2 . 5 ~ 1 0n' ~. ~ m -The
~ . other sample is kept as grown and taken as a reference.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1995331
C3-306
JOURNAL DE PHYSIQUE IV
3. RESULTS
3.1. Neutron reflectivity measurements
Reference and irradiated samples have been investigated by neutron reflectivity on the reflectometer EROS
at the Laboratoire U o n Brillouin - Saclay - France. The time of flight technique is used where the incident
angle is fmed and the neutron wavelength varies from 2 to 23 A (see reference 6). The experimental neutron
reflectivity data processing is done by numerical simulation7where the fit parameters are : the bilayer period
(d), the ratio (y) of nickel thickness on bilayer thickness, the neutron coherent scattering length density of
each material Nb where N is the density and b the neutron coherent scattering length, the roughness at the
interface substratelfirst layer (o,), the roughness at the interfaces NiClTi or Ti/NiC which is supposed to
keep constant on the all stack (o~i~m,)
and the roughness at the interface last layerlair (ok).
On Figure 1 are presented the experimental neutron reflectivity curves of both reference and irradiated
samples. This superposition shows that the first Bragg peak keeps a reflectivity value equals to unity after
the irradiation. The main difference between the two experimental curves is the non coecidence between
to 0.015A-I).
the first order fringes (placed between the cut off and the first Bragg peak, from q=0.0065~-1
Figure 1 also displays the simulated curve of each sample. Figure 2 presents the coherent length density
profiles of the two sarnpIes from which are calculated the reflectivity simulations of Figure 1. Note that for
clarity of Figure 2, the profiles do not display the roughness. The very good agreement between the
experimental points and the simulation for the reference sample shows that the initial state is very well
,
known and defined. The parameters values of the reference state are : d=204A, y=0.5, o ~ 0 A~iv;mi=10A,
od,-=~A,Nb(~i)=-2.2~10'~A-'
and N ~ ( N ~ c ) = ~ . ~ x ~ oThe
- ~ Aprofile
- ' . used to simulate the reflectivity curve
of the irradiated sample allows a reasonabl~fit, except on the second order fringes (placed between the first
and the second Bragg peak, from q=0.019~-'to 0.03A").~hefit parameters values for the irradiated sample
are equals to the reference sample ones for the first nine layers of the sample. The two last layers have
and
- ' N b ( N i ~ ) = l x l ~ - ~ This
A " . change on the two last layers
different Nb values that are : ~ b ( ~ i ) = l x l o ' ~ A
Nb values allowed to fit the first order fringes shifted position. The profile shown on Figure 2b, without
being exactly the one corresponding to the irradiated sample, is close to it. The influence of roughness
on a calculated &flectivity curve can be seen essentially at high q values. In our case, one can
note the calculated curve does not fit very well the experimental irradiated sample curve at high q, and that
is why a change on roughness parameters does not increase the fit quality. Thus, as the fit stands, it is not
possible to analyze the irradiation effect on interfaces roughness.
0
-
Ref. : exp
Irr. : exp
Figure 1 : Neutron reflectivity versus momentum transfer q=2x sin0 lil for the reference and the irradiated samples.
,
oNimi=108.,
The simulation parameters are : reference sample : 10 identical bilayers with d=204A, ~ 0 . 5 oS=0A,
o e 0 A , Nb(~i)=-2.2xlo~A-'
and N ~ ( N ~ C ) = ~ . ~ X; irradiated
I O ~ A ' sample : all the parameters are the same than for
the reference sample except for the Nb values of the last two layers before air that are : N b ( ~ i ) = l x 1 0 ~and
8;~
~ b ( ~ i ~ ) = l x l(see
~ ~Figure
A - ' 2).
x 1 0 - 9 ) Irradiated sample
~ 1 0 'a)~ Reference sample
Figure 2 : Neutron coherent length density profiles used for the simulation of the experimental neutron reflectivity curves of
Figure 1. a) reference sample ; b) irradiated sample, The only difference is the Nb values of the last two layers. Note that these
representations do not take in account the layers roughness.
3.2 X-ray diffraction
The two samples have been measured by X-ray diffraction on the diffractometer of Cilas Company - OrlCans
- France. The incident wavelength is &(CU)=I.S~~A.
Figure 3 shows the reference and irradiated samples
spectra. The initial state shows a textured sample with crystallographic plans Ti hcp (002) and Ni fcc (111)
oriented parallel to the surface samples. The irradiated sample shows the same peaks position than the
reference but a supplementary peak at 17.4" that corresponds to an intereticular distance dhkl equals to 2.6A.
We can also note the rise in intensity of the Ni fcc (111) peak from the reference to the irradiated spectra,
meaning an increase of the texture. From the Ni peak width, the Schemer formula8 allows an estimation of
the grain dimension along the layers growth axis : the grains diameter is 68A and 72A before and after
irradiation respectively. (The Scherrer formula usually yields underestimated values compared to other
methods like the integrated peak surface one for example.) The values obtained here are very close one to
the others and are of the order of the layer thickness. This shows no evolution in the grain size along the
layers growth axis after irradiation.
p
L " ' I " ' I " ' I " ' I ' " -
-
Ti
-
hcp
002
-
:
a)
Ni
fcc
111
l
-
200
€I (degree)
Figure 3 : X-ray diffraction spectra for a) reference sample and b) irradiated one.
€I (degree)
JOURNAL DE PHYSIQUE IV
C3-308
3.2 EXAFS
The two samples have been investigated by EXAFS at low temperature (80K) at the synchrotron radiation
in the LURE laboratory - Orsay - France. EXAFS investigation have been carried out in order to obtain
complementary information on the crystallographic structure obtained by X-ray diffraction. Measurements
have been done at the Ni as well as at the Ti threshold K-edge energies. Bulk materials of hcp Ti and fcc
pure Ni have also been investigated and taken as a reference. One can note that with EXAFS the
investigated sample thickness is lower than with X-rays diffraction or neutrons (200 to 300A against the all
sample). The first neighbors EXAFS experimental and calculated signals versus k for bulk material, thin
layer reference sample and thin layer irradiated sample are displayed on Figure 4 and Figure 5 for Ni and Ti
threshold energies respectively. k is the wave vector of the photoelectron emitted by the excited Ni or Ti
-E,)
atoms (k =
), Ei is the incident X-ray photon energy and E, the Ni (or Ti) K-edge absorption
energy.
A
Figure 4 shows a quasi perfect in phase relation between the three signals indicating the Ni is crystallized
with fcc structure. The simulations confm this result : the three signals are simulated with a constant
distance Ni-Ni equals to 2.48& this value corresponds to the lattice parameter of fcc Ni a=3.52A. These
results are in agreement with X-ray diffraction. Figure 4 shows an intensity decrease between the bulk
sample and the thin layers reference signals. This phenomena is revealed on the calculated curves by a
decrease of the first neighbors number, which is going from 12 to 7.9. This decrement is attributed to a
diminution of the grain size : for small dimensions grains (as nanograins in thin layers samples), the ratio of
atom number at the grain surfaces on atom number in the grain volume increases. As the number of first
neighbors is roughly proportional to the number of atoms in the grain volume, its decrease implies the ratio
"surface on volume" increase and the grain size diminution. The simulation of the irradiated sample signal
shows an increase of first neighbors number (9.5) compared to the thin layers reference sample (7.9),
indicating an increase of the grain size.
Note that Ni K-edge thin layers samples signals can also be fitted including a small amount of Ni-Ti bounds
at 2.55A (about 10%).The Ni-Ti bounds amount does not increase between the reference and the irradiated
samples while the number of Ni-Ni does. It means the Ni-Ti bounds can be due to the interface and the Ni
grain increase is also confirmed with this model.
BulkNi
Thin Layers Reference
Thin Layers Irradiated
-
3
5
9
7
k
(A-
11
h
13
')
Figure 4 : Experimental (dots) and calculated (solid lines) first neighbors EXAFS signal versus k at the Ni K-edge absorption
energy for bulk fcc Ni, thin layers reference sample and irradiated thin layers sample. The simulations are obtained for : bulk
sample : 12 atoms at 2.48A ; thin layers reference : 7.9 atoms at 2.48A ; irradiated sample : 9.5 atoms at 2.48A. For clarity,
the thin layers reference signal is shifted of -0.5 from the origin, the thin layers irradiated signal is shifted of -1.
The comparison of the Ti experimental signals and the results of the simulations between bulk material and
thin layers reference sample confirm the X-ray diffraction results : the thin layer reference sample is
crystallized with hcp structure and the lattice parameters are very close to the theoretical ones : a=2.95A
and c=4.68A. The intensity decrease due to smaller grains can be seen here as it has been seen on the Ni
signals.
The intensity of the irradiated signal is comparable to the thin layers reference one, the grains have thus
similar sizes in the two samples. But the irradiated signal shows an important shift on the first oscillations
(low k) and a decrease of the oscillations period around k=10W-'. The period decrease requires a large
distance to be simulated. The fit can be achieved with the following model : 4 Ti atoms at 2.90A and 6.4 Ti
atoms at 3.08W. Titanium is the only chemical species that allows to fit the signal. Xn particular, Ni-Ti
bounds are not possible as Ni K-edge signal cannot be fitted with 3.08A Ni-Ti distances. Moreover, Ti-C
bounds cannot fit the period decrease at high q values.
Bulk Ti
Thin layers Reference
0.4
0.3
0.2
8
0.1
3
0
2
-0.1
-0.2
-0.3
-0.4
3
5
9
7
k
(A-
11
13
')
Figure 5 : Experimental (dots) and calculated (solid lines) first neighbors EXAFS signal versus k at the Ti K-edge absorption
energy for bulk hcp Ti, thin layers reference sample and irradiated thin layer sample. The simulations are obtained for : bulk
sample : 6 atoms at 2.88A and 6 atoms at 2.92A ; thin layers reference : 3.9 atoms at 2.88A and 3.9 atoms at 2.90A ;
irradiated thin layers : 4 atoms at 2.90A and 6.4 atoms at 3.08A.
4. DISCUSSION
Neutron reflectivity experimental results show the intensity of the first Bragg peak is not modified with
irradiation treatment. Supermirrors are stack of which thickness are calculated so that a succession of first
order Bragg peaks are situated just after the natural cut off. The intensity of the first Bragg peak is not
lowered under neutron irradiation. This shows that supermirrors performances are kept constant after
irradiation.
Nevertheless, neutron reflectivity measurements show that even if the fist Bragg peak intensity is kept
constant under neutron irradiation, the positions of the fringes situated between the cut off and the Bragg
peak are modified. This variation is attributed to a variation of the coherent length density profile.
The complementary studies done by X-ray diffraction and EXAFS first show the conservation of the two
initial phases : hcp Ti and fcc Ni.
EXAFS showed a Ni grain size increase. Note that EXAFS investigates grain size in all directions, while Xray diffraction peak width leads to grain size along the growth axis. X-ray diffraction results also showed an
increase of the texture (increase of the surface of (1 11) plans oriented parallel to the sample surface), while
the grain size did not markedly increase along the growth axis. It thus means the grain size increase takes
place along the direction parallel to the layers plan.
A new crystalline phase containing Ti has appeared while the hcp Ti structure is conserved. This result can
be seen by both EXAFS and X-ray diffraction methods. This phase is different from the initial hcp one, but
C3-310
JOURNAL D E PHYSIQUE IV
it does not correspond to the two other known Ti phases (P and w Ti), and it has yet not been possible to
determine it. It is characterized by an interatomic distance Ti-Ti equals to 3.08A and an intereticular
distance dhklequals to 2.6W.
5. CONCLUSION
Thermal neutron irradiation has been done on NiC/Ti multilayers in order to determine the effect of these
radiations on neutron optical properties of the stack when they are used as supermirrors. It has been found
that the performances of the superminors are not diminished under thermal neutron irradiation.
Nevertheless, NiC layers show an increase of the grain size, while Ti seems to show an evolution from the
hcp phase to another crystalline state that has to be determined. Further investigations are now in progress
in order to determine this Ti phase.
REFERENCES
'. V.F. Turchin, Soviet At. Energy (Trans. from Atomnaya Energiya), 22, (2), p119, 1967
2
. F. Mezei, Cornmun. on Phys. 1,81-85, 1976
F. Samuel, B. Farnoux, B. Ballot, B. Vidal, SPIE Proc., 1738,54-66, 1992
B. Ballot, F. Samuel, B. Farnoux, SPIE Proc., 1738, 159-165, 1992
*.
M. Maaza, Thesis, 215-217, University Paris VI, 1991 (available at the Laboratoire U o n Brillouin)
6
.B. Farnoux, Neutron Scattering in the 'nineties, Conf. Proc. LAEA, 205-9, Vienna, 1985
7. 0. Guiselin, L.T. Lee, B. Farnoux, A. Lapp, J. Chem. Phys., 95, (6), 4632-4640, 1991
. A.J.C. Wilson, X-rays Optics,Metheum, London, 1949
3.
4.