HAL Id: jpa-00208915
https://hal.archives-ouvertes.fr/jpa-00208915
Submitted on 1 Jan 1980
HAL
is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished 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, estdestinée 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.
β � γ-SnF2 phase transition : neutron diffraction and NMR study
J. Pannetier, G. Denes, M. Durand, J.L. Buevoz
To cite this version:
J. Pannetier, G. Denes, M. Durand, J.L. Buevoz.
β�
γ-SnF2 phase transition : neutron diffraction andNMR study. Journal de Physique, 1980, 41 (9), pp.1019-1024. �10.1051/jphys:019800041090101900�.
�jpa-00208915�
1019
03B2
~03B3-SnF2 phase transition : neutron diffraction and NMR study
J.
Pannetier (~),
G. Denes(~~),
M.Durand (~~)
and J. L.Buevoz (~) (*)
(~) Institut Laue-Langevin, 156 X, 38042 Grenoble Cedex, France
(~~) Laboratoire de Chimie Minérale D (**), Université de Rennes I, 35042 Rennes Cedex, France (Reçu le 27 février 1980, accepté le 6 mai 1980)
Résumé. 2014 Les paramètres de maille et les positions atomiques ont été déterminés par diffraction des neutrons pour les deux variétés orthorhombique (03B2) et quadratique (03B3) de SnF2. Les résultats de diffraction et de RMN
(19F)
indiquent une transition continue du 2e ordre à 66 °C. Laposition
des atomes de fluor varie avec la tempé-rature mais les
bipyramides
(SnF4) ne sont pas modifiées lors de la transition.Abstract. 2014 Lattice and
atomic-position
parameters of SnF2 have been measured by neutron diffraction for both orthorhombic (03B2) and tetragonal (03B3) phases. Both structural and 19F NMR results indicate a continuous second order transition at 66 °C. The fluorine atoms undergo large displacements as a function of temperature but there is no change of size of the (SnF4) bipyramids.J. Physique 41 (1980) 1019-1024 SEPTEMBRE 1980,
Classification Physics Abstracts 61.12 - 76.60 - 64.00
Introduction. - Stannous fluoride
SnF2
is knownto exhibit three different modifications : the stable
room-temperature phase
whichcrystallizes
fromaqueous solutions is monoclinic
a-SnF2
with space groupC2/c [1, 2].
In the range 140 to 180 °C it under- goes a first orderphase
transition[3]
toy-SnF2 (P41212,
Z =4) ;
uponcooling y-SnF2
transformsto
fi-SnF2 (P212121,
Z =4,
ferroelasticphase) through
a second order
phase
transition. The structure of these two new modifications(fl-
andy-SnF2)
has beenpreviously
determined[4]
fromX-ray powder
data.The purpose of this paper is to
present
a moreprecise study
of theP:;±
yphase
transition. We have made neutron diffraction measurements at differenttemperatures
in order to determine thetemperature dependence
of the cellparameters
and atomicposi-
tions ; it is concluded that the motion of fluorine atoms in the
(a, b) plane
isresponsible
for thephase
transition.
Changes
are observed in the19F
NMRline
shape
whengoing through
the transition but motionalnarrowing
occursonly
above 110 °C. ThisP y-SnF2 phase
transition is rather similar to thepressure-induced
transition inTe02 [5, 6].
1.
Experimental.
- Thesample
material was pow- dered monoclinica-SnF2
obtained from OSI.Samples
(*) Present address : CNET, BP 42, 38240 Meylan, France.
(**) Laboratoire associé au C.N.R.S., n° 254.
were sealed under vacuum in
high-purity
quartzampoules. Tetragonal y-SnF2
wasprepared in
situby heating
ana-SnF2 sample
up to 190 °C for half anhour and further
cooling
down to the measurementtemperature :
orthorhombicp-SnF 2
was obtainedby heating
ana-SnF2 sample (in
a quartzampoule)
up to 190 °C in a mufHe furnace andrapidly quenching
toroom temperature. No
a-SnF2
was observed in thep-SnF 2 samples,
even at thehighest temperature
we used(62 °C)
but ay-SnF2 sample
at 80 °Cpartly
transformed back to
oc-SnF2.
The neutron diffraction
experiments
wereperform-
ed at the Institut
Laue-Langevin
on two spectro-meters :
- D lA
(multiple
counter bank[7]) :
 = 1.908À,
T =
24, 42, 51,
62 and 177 °C.- D1B
(position
sensitive detector[8]) :
:/L = 1.28A,
T = 125 °C.
The
heating
device was a conventional furnace with vanadium heater and shieldsworking
underlow-pres-
sure
helium,
the temperature was measured with athermocouple
in contact with thesample.
D lA datawere collected from 160 to 150° in steps of 0.100
(2 0), taking
about 20 h for each temperature. The data from the 10 counters were summedusing
the ILLPOWDER program
[9] ;
the quartz pattern(measured
on an empty quartz
tube)
was substracted toproduce
the final diffraction
pattern.
DIB data were collectedArticle published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:019800041090101900
1020
Figs. 1a,16. - Neutron diffraction pattern of p-SnF 2 (24 °C) and y-SnF2 (177 °C). Crosses are experimental points ; the solid curve is the least-squares refined profile ; the difference profile is shown at the bottom.
1021
from 2° to 820
(2 0)
in 5 h and reducedusing
ILLprograms [ 10] .
The Rietveld program
[11]
as modifiedby
Hewat[12]
was used for all refinements.
Scattering lengths
of0.62
(Sn)
and 0.56(F),
all x10-12
cm, were used.Refinements used
only isotropic
thermal parameters ; three half-width parameters were included in the refinementtogether
with an overall scalefactor,
azero
point correction,
an asymmetry parameter and the cell parameters which led to21(fi-SnF2)
or14(y-SnF2)
refinedparameters. R
factors(nuclear)
range from 6.2 to 10
%
except for the pattern close to the transition(T
= 62°C)
for which R = 13.3%.
All
the
19F NMR spectra were taken on a BRUKER SXPpulse
spectrometeroperating
at 84.67 MHz.Ti
was measuredusing
a n - i -n/2 pulse
sequence.Spectra
were recorded from room temperature to190 °C,
thetemperature being
stabilized to better than ± 0.5 °C. Allsamples
were sealed under vacuumin
glass tubes ; they
wererelatively
free of paramagne- ticimpurities
as evidencedby
thelong spin-lattice
relaxation times
(Tl -
250 s atroom-temperature).
2. Cell déformation. - The
room-temperature
pat-tern of
fi-SnF2 (Fig.1) clearly
shows thesplitting
of theFigs. 2a, 2b. - Lattice parameters of fl- and y-SnF2 vs. temperature (neutron and X-ray data). Typical standard deviations are given for a
few X-ray points; for neutron results standard deviations are
smaller than the data points. For sake of clarity some X-ray points
have been omitted. Solid and broken lines are least-squares fitted
to the experimental data.
hkl and hkO lines
corresponding
to the orthorhombicdistortion ;
aspreviously
observedby high
tempera-ture
X-ray
diffraction[4],
thesesplittings disappear
at66 OC and the
corresponding
transition is a pure ferro-paraelastic phase
transition(422F222
in Aizu’s nota-tion).
The refined lattice parameters areplotted
infigures
2a and 2b.They
compare well withprevious X-ray
values[3].
The main difference is thehigher precision
of theparameters
from neutronprofile refinement, especially
in thefl-phase
below thetransition temperature
(i.e.
when a -b).
The transition temperature can be obtained accu-
rately
from aplot
of(b - alb
+a)2
vs. temperature.A
least-squares
fit to the four datayields
which goes to zero at T =
6b(1)
°C ingood
agreement withprevious
results.3. Atomic
displacements.
- Previous studies of thisphase
transition led us to assume space groupsP212121
andP41212 (or P43212) respectively
forfi-
and
y-SnF2.
Profilefitting
structure refinement from neutron diffraction data agree well with these space groups andrefinements, starting
fromX-ray
deter-mined
coordinates, quickly converged
tomeaningful
values of atomic coordinates and
isotropic
thermalparameters.
Figures
3 and 4 show the atomic coordi-Figs. 3a, 3b. - Temperature dependence of the x and y fluorine position coordinates in fi- and y-SnF2. Error bars on the data points
indicate the standard deviations as obtained from the profile fitting
structure refinement. Solid lines are only guides for the eye.
1022
Fig. 4. - Temperature dependence of the z fluorine position
coordinates in fl- and y-SnF2.
nates of fluorine atoms vs. temperature. These
posi-
tion parameters have been calculated in the same
coordinate system
(P41212)
in order todisplay
theirsplitting
in the(a, b) plane
and theirequivalence
inthe
tetragonal phase.
Theseplots clearly
show thatthe main feature of
the P -+
yphase
transition is the motion of fluorine atoms in the(a, b) plane.
Fluorinedisplacements along
the c axis are very small : asprofile
refinement is known to underestimate standard deviations[13],
thesedisplacements
as well as the dif-ference between
z(Fl)
andz(F2)
arehardly significant.
This also holds for the Sn
displacements (Fig. 5)
which are not
larger
than three standard deviations.Fig. 5. - Temperature dependence of the tin position coordinates in fi- and y-SnF2.
Worlton and
Beyerlein [5] pointed
out the increase of theDebye-Waller
factors of the anions in thevicinity
of theTe02 phase transition ;
no such effect is observed forSnF2
and theisotropic
thermal para- meters are almost constant in the range 24 OC to 62 OC with values(averaged
over fourtemperatures) :
They
increase up toin the
y-phase
at 177 °C.4. 19F NMR results. -
Figure
6 illustrates thechange
in the NMR spectra as a functionof tempera-
Fig. 6. - 19F powder spectra of fi- and y-SnF2. The measunng
frequency is 84.67 MHz.
ture. The low
temperature (fi-SnF2) spectrum
isstrongly asymmetric
and can beinterpreted
as dou-blets
coming
from two kinds of fluorine atoms(FI
and
F2)
with different chemicalshifts ; high
tempera-ture spectra
(y-SnF2)
suggestonly
one kind of fluorine atom withanisotropic
chemical shift.The width of the resonance line was measured for the three
phases
a-,13-
andy-SnF2 (Fig. 7).
For mono-clinic
a-SnF2,
motionalnarrowing
occursjust
aboveroom temperature whereas it is observed
only
above110 °C for
y-SnF2.
They-SnF2
lattice is thuseffectively rigid
and nochange
ordiscontinuity
is observed atthe fi
± yphase
transition.Fig. 7. - Temperature dependence of the width at half-maximum of the resonance line for a-, fl- and y-SnF2.
The results of the
T1
measurements arepresented
in
figure
8. Nochange
is observed at thefi
± yphase transition ;
this wasexpected since,
in this range of temperatures, thespin-lattice
relaxation isgoverned by paramagnetic impurities (extrinsic regime).
Thefit of
y-SnF2
data above 140 °C(intrinsic regime) yields
an activation energy for the relaxation processofE.(y)
=0.52(1)
eV much smaller than the activation energy obtained fromconductivity
measurements(0.74
eV[14])
which indicates that the relaxation1023
process which is
responsible
forTl
is digèrent than that forconductivity.
Fora-SnF2
one obtainsin close agreement with the
conductivity
data(0.52 eV [14]).
Fig. 8. - Spin-lattice (IIT,) relaxation rates as a function of
reciprocal temperature in a-, fl- and y-SnF2.
5. Discussion. - All the above observations indi-
cate that
the fi
+:ty-SnF2
transition is a continuous pure strain transition whose main structural feature is the motion of fluorine atoms in the(a, b) plane.
Moreover,
calculation of interatomic distances andangles
leads to thefollowing
results :- The shortest Sn-F distances in the
(SnF4)
tri-gonal bipyramid
are constant in the range 240 to 177 OC with thefollowing
values(Fig. 9) :
Fig. 9. - (SnF4) trigonal bipyramid ; black sphere is tin.
in
good
agreement with the valuesusually
observedfor this coordination
(A configuration according
toBrown
[15]).
- The Sn-F-Sn
angle
is constant iny-SnF2
buttwo different temperature
dependent angles
are observ-ed in the
low-temperature phase (Fig. 10).
Fig. 10. - Temperature dependence of the Sn-F-Sn angles.
Then we can conclude that the
(SnF4) trigonal bipyramids
retain theirshape
when the solid goesthrough
the transition and it isonly
the orientation of onebipyramid
with respect to theneighbouring bipyramids
which ismodified ;
in otherwords, fl-
andy-SnF2
can be considered asbeing
built up fromrigid (SnF4)
units and the soft feature[16]
of the structure isthe Sn-F-Sn
angle.
This is confirmedby looking
at theorientation of the cone of
zero-expansion (i.e.
thelocus of the directions
along
which the dilatation iszero
[17]);
as the thermalexpansion
isnegative along
the b axis in
fi-SnF2,
the axis of the cone coincides with b; the cone can be definedby
twosemi-opening angles 0(b, c)
and0(b, a) [3].
These twoangles
whichdefine the directions of
null-expansion
in theplanes (b, c)
and(b, a)
can be calculated from the thermalexpansion
coefficients[17] :
as shown infigure 11, they
are not constant(solid line, Fig. 11)
but theirvalues at the transition temperature
provide
informa-tion about the mechanism of the transition. Indeed,
Fig. 11. - p-SnF 2 : semi-opening angles of the zero-expansion
cone (angles between the directions of null-expansion and the
b axis). Solid lines : calculated from thermal expansion coefficients.
Broken line : calculated from the orientation of (SnF4) bipyramids
in the cell (F (axial)-F (axial) and F (equat.)-F (equat.) orientation within the cell). The values at the transition temperature are cal- culated from the y-SnF2 structure.
1024
the thermal contraction upon
cooling
in thetetragonal phase
takesplace
until somecritically
short distanceis reached and one can assume that the directions of
zero-expansion
at the transition coincide with the directions of the shortestbonds,
i.e. the structure is lockedalong
these bonds.Figure
11clearly
showsthat,
at the transition, the directionsof null-expansion
are
along
the F(axial)-F (axial) (e(a - b))
and F(equatorial)-F (equatorial) (O(b, c)) axis;
at lowertemperatures
(
40°C)
the cell distortion becomeslarge
and other interactions(F-F contacts)
occur inthe determination of the
zero-expansion
directionswhose variation with temperature becomes more
complex.
This
description
ofthe P
y transitionby
alocking
of the structure
along
the axis of the(SnF4) bipyra-
mids is
actually
in agreement with both the NMR results and theDebye-Waller
factors which do notshow any increase of the fluorine thermal motion at the transition. A recent
reinvestigation [6]
of thepressure induced ferroelastic
phase
transition inTe02 using
the Landautheory
led to similar conclu-sions about the structural mechanism of that transi- tion.
References
[1] MCDONALD, R. C., HO-KUEN HAU, H., ERIKS, K., Inorg.
Chem. 15 (1976) 762.
[2] DENES, G., PANNETIER, J., LUCAS, J., LE MAROUILLE, J. Y., J. Solid State Chem. 30 (1979) 335.
[3] DENES, G., Thèse d’Etat, Rennes (1978), unpublished.
[4] DENES, G., PANNETIER, J., LUCAS, J., J. Solid State Chem., in press.
[5] WORLTON, T. G., BEYERLEIN, R. A., Phys. Rev. B 12 (1975) 1899.
[6] UWE, H., TOKUMOTO, H., Phys. Rev. B 19 (1979) 3700.
[7] HEWAT, A. W., BAILEY, I., Nucl. Instrum. Methods 137 (1976) 463.
[8] ALLEMAND, R., BOURDEL, J., ROUDAUT, E., CONVERT, P., IBEL, K., JACOBE, J., COTTON, J. P., FARNOUX, B., Nucl.
Instrum. Methods 126 (1975) 29.
[9] HEWAT, A. W., Powder Retrieval and Refinement System, ILL Report (1978).
[10] WOLFERS, P., unpublished computer programs.
[11] RIETVELD, H. M., J. Appl. Crystallogr. 2 (1969) 65.
[12] HEWAT, A. W., UKAERE Harwell Report RRL 73/897 (1973).
[13] SAKATA, M., COOPER, M. J., J. Appl. Crystallogr. 12 (1979) 554.
[14] ANSEL, D., DEBUIGNE, J., DENES, G., PANNETIER, J., LUCAS, J., Ber. Bunsenges. Phys. Chem. 82 (1978) 376.
[15] BROWN, I. D., J. Solid State Chem. 11 (1974) 214.
[16] MEGAW, H. D., Crystal Structures : a Working Approach (Saunders Co., Philadelphia) 1973.
[17] WOOSTER, W. A., Tensors and Group Theory for Physical Properties of Crystals (Clarendon Press, London) 1973.