15512916, 1990, 10, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/j.1151-2916.1990.tb06682.

x by NRI for Material Physics, INCDFM Bucharest, Wiley Online Library on [17/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
J Am Ceram SOC 73 (101 2828-33 (1990)

Neutron Diffraction Studies of Phase Transformations between

Tetragonal and Orthorhombic Zirconia in
Magnesia-Partially-Stabilized Zirconia
Christopher J. Howard and Erich H . Kisi*
Australian Nuclear Science and Technology Organization, Menai, New South Wales 2234, Australia

Ronald B. Roberts
CSlRO Division of Applied Physics, Lindfield, New South Wales 2070, Australia

Roderick J. Hill*
Mineralogisches Institut, Universitat Wurzburg,
Am Hubland, D-8700 Wurzburg, Federal Republic of Germany

Neutron powder diffraction and conventional dilatometry c-related structure with ordered anion vacancies2) and an or-
,have been used to investigate the tetragonal-to-orthorhombic thorhombic (0) z i r ~ o n i a . This
~ , ~ paper reports new studies of
phase transformation and the orthorhombic-to-tetragonal the o phase; the occurrence of the S phase and its influence
reversion in a high-toughness magnesia-partially-stabilized on properties arc the subjcct of another study which will be
zirconia. For this material, the onset temperature on cooling described elsewhere.'
for the tetragonal-to-orthorhombic transformation (deter- The occurrence of an o phase in zirconia-toughened ce-
mined by dilatometry) was 192 K, and the reversion on sub- ramics was first noted6” in transmission electron microscopy
sequent heating occurred between 500 and 620 K. Neutron (TEM) studies of thin foils of partially stabilized zirconia
diffraction patterns were recorded at temperatures down to (PSZ). There have been further T E M observations of o-ZrOz
19 K then up to 664 K, and analyzed by the multiphase in a number of different materials, and these are conveniently
Rietveld method to determine the amounts of different summarized in a recent article by Heucr et a1.* It was sug-
phases as well as their lattice parameters and unit-cell vol- gested8 that the occurrence of the o phase in T E M thin foils
umes. It is notable that, at its maximum, the orthorhombic was artifactual. The first evidence that the o phase occurrcd
phase amounted to 45% of the sample by weight. Length in substantial quantities was provided by Marshall et d 3from
changes were measured, using pushrod dilatometers, in the an X-ray diffraction examination of high-toughness magnesia-
temperature range 80 to 700 K. Length changes calculated partially-stabilized zirconia (Mg-PSZ) samples which had
from the neutron diffraction determinations of the propor- been cooled below about 170 K. Marshall et al. concluded
tions and unit-cell volumes of the different phases are in that a t-to-o transformation occurred throughout the bulk with
very good agreement with the directly measured values. most of the t (estimated at 30% by volume) becoming o. Shortly
[Key words: zirconia, partially stabilized zirconia, neutrons, afterwards, thc present authors completed a neutron powder
dilatometry, phase transformations.] diffraction investigation of high-toughness Mg-PSZ cooled to
30 K.4 It was confirmed that, upon cooling, an o phase is
I. Introduction produced in considerable quantities at the expense of the
t phase. The crystal structure of this o phase was determined

I N THE majority of studies of zirconia-toughened ceramics, it

has been considered that the cubic (c), tetragonal ( t ) , and
monoclinic (m) polymorphs of zirconia are the most impor-
from the neutron diffraction pattern.
In this paper we report the results from neutron powder
diffraction analyses of high-toughness Mg-PSZ at different
tant and abundant phases. Thcse phases are indeed impor- temperatures. The use of neutrons, which penetrate zirconia
tant, with the martensitic transformation from metastable t ceramics to depths of several centimeters, ensures that the
to rn (accompanied by a volume increase of almost 5%) diffraction patterns recorded are representative of the bulk.
making a critical contribution to the toughening.’ However, Neutron diffraction patterns were recorded at temperatures
it has recently become apparent that other phases may be ranging from 19 to 664 K, allowing the t-to-u transformation
present in substantial amounts in certain zirconia-toughened on cooling and the reversion of o to t on subsequent heating
ceramics. These other phases are thc 6 phase (Mg2Zr5012, a to be followcd in detail. The analysis of the diffraction pat-
terns yields not only the amount of the different phases
(c, t , m, 0 , 6 ) present at any temperature, but also the lattice
Kobert Roth-contrihuting editor parameters of the major phases as a function of temperature.
The data are sufficient to allow a calculation of the variation
of sample volume with temperature, taking into account both
Manuscript No. 197831. Received January 12, 1990; approved May 14, phase transformations and the temperature dependence of
1990.
Presented at the Fourth International Conference on the Science and lattice parameters. Precise measurements of t h e sample
Technology of Zirconia, Los Angcles, CA, November 1-3, 1989. length as a function of temperature were made on an Mg-PSZ
The work of R. .I.Hill was supported by a Ludwig Leichhardt Memorial from the same batch as the sample examined with neutrons,
Fellowship and the work of E. H. Kisi was supported by the National Re-
search Fellowships scheme. and the agreement of the measured length changes with the
*Member, American Ceramic Society. calculations from neutron data is very good. T h e crystal
*On leave f r o m CSlRO Mineral Products, Port Melhourne, Victoria 3207,
Australia. Recorded neutron diffraction patterns at the Institut Laue- structure of o - Z r 0 2 in PSZ and its relationship to the other
Langcvin, Grenoble, France. phases of zirconia are described briefly below.
2828
 15512916, 1990, 10, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/j.1151-2916.1990.tb06682.x by NRI for Material Physics, INCDFM Bucharest, Wiley Online Library on [17/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
October 1990 Neutron Diffraction Studies of Phase Transformations 2829

11. Crystal Structure of Orthorhombic Zirconia -

a
1

The crystal structure of o-ZrO,, as determined from the

c
neutron powder diffraction, is taken from our earlier r e p ~ r t . ~ I

The details are reproduced in Table I. The structure shows

a sevenfold coordination of the Zr atom (Fig. 1) similar to
that found in m-ZrOz.' One of the oxygen atoms is three-
coordinated, while the coordination of the other oxygen is
nearly tetrahedral.
The t-to-o phase transformation is accompanied by a vol-
ume expansion of 1.25%, less than the 5% for the t-to-m
transformation. The o and t structures can be compared by
W
setting [OlO], parallel to [OOl],, and [lOO], parallel to [OlO],
where the t structure is described in its usual face-centered
setting. With this comparison, the atomic displacements re- Fig. 1. Sevenfold coordination of Zr in o-ZrOz (Fig. 3 from
sponsible for the transformation can be readily identified (see Ref. 4). Mean Zr-0 bond length is 0.2158 nm.
Fig. 6 in Ref. 4).
It is also interesting to compare the o and m structures.
There are considerable similarities which become evident when urement, and each diffraction pattern took about 120 min to
the structures are set with [OlO], I( [OlO], and [lOO], 11 [look. collect. A typical diffraction pattern appears in Fig. 3.
In fact, the differences are substantial only in the coordination The neutron diffraction patterns were analyzed by the
net of the three-coordinated oxygen atoms (01).These differ- Rietveld method.I5 In this method, a least-squares fit is made
ences are illustrated in Fig. 2, where the net of 01 atoms in of a diffraction pattern calculated from the crystal structure
the o structure is superimposed on the corresponding net in the models to the entire observed pattern obtained (in this appli-
m. Since the t-to-o transformation is expected to relate closely cation) from the multiphase mixture. The analysis involves
to the t-to-m transformation, this suggests the lattice corre- the refinement of crystal structure parameters and the least-
spondence in the t-to-m transformation might be [OlO], 11 [OOl], squares determination of a scale factor for each phase. In
and [look 11 [OlO],. In current terminology,"'~" this is lattice every case the sample was analyzed as a five-phase mixture of
correspondence "B' and orientation relationship "1". In fact the c, t, m, and o polymorphs with the 6 phase. The computer
for the low-temperature stress-induced transformation only program used was an extensively modified version of that of
correspondence C has been reported.'* However, lattice cor- Wiles and Young," which has been described by Hill and
respondence B has been observed (along with correspondence Howard." For the o phase we incorporated the crystal struc-
C) in the t-to-m transformation at high ternperature.l3 ture obtained in our earlier work,4and described in the section
above. The crystal structures of the c, t, and m polymorphs
and the 6 phase are taken from other sources,IR although for
111. Experimental Procedure and Data Analysis m the lattice parameters were amended in accord with morc
The samples were of high-toughness 9.7-mol%-MgO- recent unpublished work. The details of the c - , t-, m-, and
partially-stabilized zirconiat corresponding to the commer- &phase structures are given in Table 11. In the five-phase
cially produced thermal-shock-resistant (TS) grade and were refinements, a refinement of the o structure (scale, lattice
obtained in the form of cylinders approximately 14 mm in parameters, individual isotropic thermal parameters) was
diameter by 30 mm high. They were fabricated by a propri- undertaken along with refinements of the scale factors, lattice
etary process, the details of which are not available; it is be- parameters, and appropriate structural parameters of the c, t ,
lieved to involve sintering at 170O"C, controlled cooling to and 6 phases and only the scale factor of the m phase. Differ-
room temperature, then a subeutectoid heat treatment.I4 The ent peak shapes and widths were allowed in the different
two samples, one for neutron diffraction and one for dila- phases, although in the m phase they were not refined. The
tometry, were from the same batch, but this was not the same
batch which provided the sample used in our earlier s t ~ d y . ~
Diffraction patterns were recorded using the high-resolution
neutron powder diffractometer D2B at the Institut Laue-
Langevin (ILL), Grenoble, France. The patterns were re-
corded, at a neutron wavelength of 0.1594 nm, over the 20
range from 0" to 1hO"with a step size of 0.025". The sample was
loaded in an ILL "orange" cryostat for the low-temperature
measurements, and a 900°C vanadium furnace was used for
the measurements above room temperature. The sample was
equilibrated at temperature for about 30 min before each meas-

'From Nilcra Ceramics Pty Ltd., Northcote, Vlctoria 3070, Australia

Table I. Crystal Structure Parameters for o-Zr02* Found

in Toughened Mg-PSZ After Cooling to 30 K
Atom Site X Y z B nm')'
Zr 4a 0.267(1)* 0.030(1) 0.250(1) 0.4(1)
01 4a 0.068(1) 0.361(1) 0.106(1) O.8(1) C
\ I

02 4a o m ( i j 0.229iij 0.0 0.2(1j

Fig. 2. Comparison of the coordination of the 01 atoms i n o-ZrOz
*Orthorhombic, space group Pbc2, (No. 29), Z = 4. Cell constants are with that in rn-ZrOz. Layer of 01 atoms in o-ZrOz with the coordi-
a = 0..5068(1), b = 0.5260(1), and c = 0.5077(1) nm. +Isotropic thermal pa-
rameter. 'Figures in parentheses here and elsewhere in the tables indicate nating Zr atoms is shown bold, the 0 1 and Zr atoms being repre-
the standard deviations (referred to the least significant digit) as estimated sented by the large and small circles, respectively (Fig. 4 from
in the refinement. Ref. 4). Equivalent atoms and bonds in rn-ZrOz are shown dashed.
 15512916, 1990, 10, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/j.1151-2916.1990.tb06682.x by NRI for Material Physics, INCDFM Bucharest, Wiley Online Library on [17/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2830 Journal of the American Ceramic Society -Howard et al. Vol. 73, No. 10

3000

I I I II II II II III I I I I IIIIII I IIUIIII Inil I UNIIIIII 1111I II I I I I I m i I

I 1 I I 1 I 1 1 I I I I
I I1 I1 I II I I I I l l 1 1 1 1 N I I I N I I I I I I I I I I I I I I I I I 11111 1111l1111111111II1HIIIIIII111111111111
I I1 I I1 I I1 I I 1 I I I IIIII 1 I II I I I1 I I I I I 1 I I1 I
I II I I I N I IRIII 111 IIIII IIUU I IIIIMII I I I I I N I I I I R H ~ I I ~ I II II I M I I I I R I ~ I ~ I NRII IIIII IIIIII ~~~UIIIIIIIIIII
I LLLyl
I 1 1 I I I I I I I I 1 1 1
1 1 LLy ,
I I t I I I. uul.ul&
I I 1 I 1 . I I I.I u >-
I 1 I B . 1 I I I .. I 6 1 I I I I I I .. I 1 I I

Two Theta (degrees)

Fig. 3. Neutron powder diffraction pattern from the toughened Mg-PSZ at 307 K, and the Rietveld method fit to this pattern. Observed
data are indicated by crosses and the calculated pattern by the solid line overlying them. Rows of markers underneath the pattern show
( f r o m t h e t o p down) t h e calculated peak p o s i t i o n s for o - Z r 0 2 , c - Z r O 2 , t h e 6 p h a s e ( r e l a t e d t o cubic), a n d t-ZrOz and
m-ZrOz. The line above the pattern is the difference between the observed and calculated patterns. This shows the agreement between the
two to be good; the weighted profile R-factor (the usual measure of fit) is R, = 6.77% with an expected value R , = 4.65%.

background was fitted by a simple polynomial. At 20 and where xp is the mass fraction of the pth phase, Z , is the num-
307 K, the atomic position parameters for o-Zr02 were also ber of formula units each of mass M Pin the unit-cell volume
refined. Figure 3 shows the plot output from the Rietveld V,, and S, is the Rietveld scale factor for thepth phase. The
refinement of the 307 K diffraction pattern, and shows that values for 2, M , and I/ for the different phases are included
the fit of the calculated to the observed diffraction is good. in Table 111.
The amounts of the different polymorphs at each temperature Precise measurements of the changes of length with tem-
were estimated from the scale factors obtained in the final perature were made on the second sample using a silica
refinements, using the equation" pushrod dilatometer (similar to that described previouslyz1)
below 20°C and a commercial silica pushrod dilatometer
(Adamel-Lhomargy, France) above 20°C. The dilatometers
were calibrated against an NBS standard copper rod of the

Table 11. Crystal Structures of c-Zr02,t-ZrOz, m-Zr02, and the 6

Phase Used in the Analysis*
Atom Site X Y z B (lo-? nm')'
Cubic Zr086Mg~.1401.8(1;
Fmm; a = 0.5073(1) nm
Zr Mg 4a 0 0 0 0.9(3)
0 32f 0.284(1) 0.284(1) 0.284(1) 1.2(6)
Tetrdgonal ZrOz; P4z/nmc; a = 0.35925(3) and c = 0.51837(5) nm
ZJ 2a !A 3/4 0.41(4)
0 4d Y4 ?4 0.4546(6) 0.36(4j
Monoclinic Zr02; P2,/c; a = 0.5144(1), b = 0.5133(1), c = 0.5347(1) nm, and
p = 98.88(2)"
Zr 4e 0.2734(2) 0.0395(2) 0.2083(2) 0.33(2)
01 4e 0.0700(3) 0.3317(3) 0.3447(3) 0.55(2)
02 4e 0.4496(3) 0.7569(3) 0.4792(3) 0.46(2)
Rhombohedral' Mg2ZrSOlZ;RT; a = 0.9427(1) and c = 0.8780(2) nm
Zr/Mgl 3a 0 0 0 1.5(4)
Zr/Mg2 l8f 0.295(1) 0.412(1) 0.018(1) 0.1(1)
01 18f 0.295(1) 0.457(1) -0.221(2) 0.7(3)
02 18f 0.308(2) 0.453(1) 0.273(1) OS(3)
*Room-tem erdture structures. Parameters such as cell constants and thermal param-
eters vary wit\ temperature. 'Isotropic thermal parameter. ' 0 atoms are disordered over
sites displaced from the ideal fluorite positions. See Ref. 18 for detail. 'Hexagonal setting.
 15512916, 1990, 10, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/j.1151-2916.1990.tb06682.x by NRI for Material Physics, INCDFM Bucharest, Wiley Online Library on [17/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
October 1990 Neutron Diffraction Studies of Phase Transformations 2831

Table 111. Mass, Volume, and Thermal Expansion Data on the Phases
in a Toughened
- Mg-PSZ
-
Phase Z M (amu) V293(A3)* pi K-')* p2 (10.' K-')*
c-ZrOz 4 111.61t 130.93 0.230 0.243
t-Zr07 2' 123.22 66.84' 0.231 0.285
o-ZrO2 4 123.22 135.35 0.123 0.106
m-Zr02 4 123.22 139.56§ 0.263" -0.1657
M ~ 7 Z r ~ 03~ ~ 696.72 675.95 0.226 0.293
Tonstants obtained from fitting the neutron data on unit-cell volumes with V ( T )=
V293[1 + P I ( T 293) + &(T - 293)2]. V293is the unit-cell volume at 293 K, and PI is the
volume expansivity at the same temperature. 'Calculated for the composition ZrOr14%
frlg0. 'The primitive unit cell is used here. *Unrefined estimate from the neutron data.
These values, obtained from the literature? correspond to the expansion of pure m-Zr02
above room temperature. They were not used in the length calculations.

same length as the sample. For the measurements below room Fig. 4. The t-to-o transformation commenced just below
temperature, the sample was cooled to 80 K in nitrogen gas at 200 K, and at 20 K the sample contained about 38% by weight
about 2 to 4 K/min, held for 15 min in liquid nitrogen at 77 K, o-ZrOz. The o content increased further, to 45%, when the
then warmed in nitrogen gas at about 1 to 2 K/min. The sample was warmed just above room temperature (307 K). In
sample was sliced in half, down the diameter, to fit the high- comparison with the likely errors, this further increase is sig-
temperature rig. The sample was taken to 723 K at 2.5 K/min, nificant, and a similar increase has been observed in a meas-
then cooled at rates diminishing from an initial 2 K/min. urement on a closely related sample. We cannot yet offer any
convincing explanation for this effect. The o-to-t reversion
IV. Results and Discussion upon heating occurred between 500 and 620 K. However, a
small amount of the o phase remained in the sample and did
Results from the neutron diffraction are in accord with not revert even at 664 K, the highest temperature at which a
those from our previous study of a similar ample.^ There was neutron pattern was recorded. The amounts of the c, m, and 6
little or no evidence for the presence of the o phase in the phases changed little through the series of analyses and, in
sample as received. The diffraction patterns showed signifi- the absence of strong evidence to the contrary, are (for the
cant changes when the sample was cooled below 200 K, con- purposes of later calculations) assumed to remain constant. In
sistent with a substantial t-to-o transformation. The patterns summary, the sample comprises 4.7(0.9)% by weight c-ZrOz,
indicated that the o phase was retained on heating, until the 7.1(0.5)% m-ZrOZ, 24.0(0.9)% 6 phase, with the remainder
onset of o-to-t reversion just above 500 K. (64.2(1.2)%by weight) being t- and o-ZrO,, in the proportions
Detailed results from the quantitative phase analysis are indicated in Fig. 4.
presented in Table IV. They are listed in the order in which Lattice parameters for the o phase were determined, in the
the measurements were made; therefore, the sequence of the course of the Rietveld analysis, at temperatures of 19 to
temperatures summarizes the thermal history of the neutron 597 K, inclusive. The variation of these lattice parameters
diffraction sample. The table includes an indication of stan- and the unit-cell volume with temperature is shown in Fig. 5.'
dard deviations based on measurement statistics, but these Though the precision of the lattice parameter determinations
are possibly exceeded by the systematic errors which are ex- is not high, it is sufficient to show at least that the thermal
pected in the analysis of such a complex mixture of phases. expansion of o-Zr02 is anisotropic, the expansion of the
This is taken into consideration in the account we give below. b parameter (the long axis) being considerably less than the
Although the analyses for the first two patterns, recorded expansion of the a and c. The data on unit-cell volume versus
at 220 and 200 K, indicate an amount of o-Zr02 in the sample temperature have been fitted with a quadratic function, the
at these temperatures, this amount is really too small to say details of which are recorded in Table 111. The entries there
that the presence of the o phase is definitely established. It is confirm that the volume expansion associated with the t-to-
concluded that the sample as received comprised mostly
t-ZrO,, with a substantial amount of 6 phase, and minor
amounts of the c and rn phases. The remaining analyses con-
firm the development of the o at the expense of t on cooling 'Refinement of the 496 K pattern gave anomalous values for the lattice
to lower temperatures, and the reversion of the o to t on sub- parameters in all phases, and these have been ommitted durinj, the suh-
sequent curve fitting. The anomalies are associated with certain ifficulties
sequent heating. These transformations are highlighted in which were enountered in recording the neutron data at this temperature.

Table IV. Phase Composition in a Toughened Mg-PSZ at

Different Temperatures
Composition (wt%)*
Temperature (K) c-ZrO2 z-ZrO2 0-Zr02 m-Zr02 Mg2ZrsO12
220 6.0(5) 61.5(4) 2.2(4) 7.0(5) 23.3(4)
200 6.0(4) 61.5(4) 2.4(4) 7.0(5) 23.2(4)
180 6.1(4) 42.7(10) 19.1(9) 6.9(4) 25.116)
157 5.5(4) 33.9(8) 28.9(7) 6.8(4) 25.0(6)
141 5.0(4) 30.8(8) 33.2(7) 6.4(4) 24.5(6)
100 4.6(3) 26.5(7) 38.3(7) 6.3(4) 24.3(6)
-19 4.0(4) 26.2(10) 38.2(9) 7.1(4) 24.5(9)
307 3.8(2) 19.4(3) 4544) 6.7(2) 24.6(3)
496 3.8(7) 20.9(7) 44.2(10) 7.0(9) 24.2(7)
597 3.5(5) 49.3(11) 14.6(9) 8.0(5) 24.6(10)
620 4.1(4) 59.6(10) 5.0(4) 7.9(5) 23.3(10)
649 4.4(4) 59.6(10) 4.9(5) 7.6(5) 23.5(11)
664 4.1(4) 61.6(11) 5.1(5) 73.5) 21.8(10)
*Entries calculated using Eq. (1).
 15512916, 1990, 10, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/j.1151-2916.1990.tb06682.x by NRI for Material Physics, INCDFM Bucharest, Wiley Online Library on [17/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2832 Journal of the American Ceramic Society - Howard et al. Vol. 73, No. 10

' 023 I----

0
0
.,1,
LL..-..
100
,. .-.
200
I . =

300
-
A
, , , L

400 503 600 700 1019 [ ,

, , , , , , , , , , , , , , , , , -,,, ,
, , , , , , , , , , , , , , , , , , , , ,I , > , , , , , , , l,,,A
0 200 400 600
Temperature (K)

Fig. 4. Plot indicating the progress of the t-to-o phase transfor- Temperature (K)
mation on cooling a high-toughness Mg-PSZ, and the reverse Fig. 6 . Temperature variation of the c/a ratio for t-ZrOz in a
transformation on subsequent heating. Points are taken from the high-toughness Mg-PSZ. Line is a least-squares fit to the data, ex-
analyses (Table IV) of the neutron diffraction patterns recorded at cluding the 496 K point.
different temperatures. In this and other figures, the error bars are
drawn to plus and minus one standard deviation. Line here is
simply a guide for the eye.
anisotropy, and we believe that tensile stresses on the t pre-
cipitates, resulting from the more rapid contraction of the
o transformation is about 1.25% at 293 K, and show that the (tetragonal) c on cooling, play a major role in the t-to-o phase
volume expansivity of the o is only about half that of the transformation. For all phases (but m) the lattice parameters
other phases. At 19 and 307 K the Rietveld analysis has in- have been used to calculate the unit-cell volumes, and these
cluded a full refinement of the structure of o-Zr02. At each data on cell volumes and their variation with temperature
temperature, the atomic positions are essentially identical (no are summarized in Table 111. The m cell volume entered
differences outside two combined standard deviations) with in Table I11 is an unrefined estimate from the neutron pat-
the results obtained in our earlier study4 and shown in Table I terns, and the data on its temperature variation are from
of this paper. other sources.22
Lattice parameters at different temperatures were also ob- Now, the volume, V', of the entire sample is governed by
tained for the other phases, except for them, which is a minor the amounts of the different phases, and the densities of
phase with a weak diffraction pattern. These will not be dis- these phases, through
cussed in detail, although it is interesting to note that the
thermal expansion of the t is also markedly anisotropic, as
evidenced by the increase of the ratio c/a with increasing tem- V' = A X (qJ/Pp)
P
perature (Fig. 6). This anisotropy is opposite in sign to the
thermal expansion anisotropy of o-Zr02 (recalling the corre- where 4 is the total mass of the sample, pp (=ZpMp/V,) is
spondence between the b and c axes of the o and t structures). the density of t h e p t h phase, and the other symbols are as
There is thermal expansion mismatch associated with this before. The sample length 2 is taken to vary as the cube root
of the volume, so for this we have

5?S F
t

The neutron diffraction analyses provide the data required

for a calculation of the variation of 2 as the temperature
changes and the transformation progresses. The values for
the mass fraction xp are taken from the neutron diffraction,
assuming constant amounts of c, m , and S phases (see above),
and apportioning the remainder between the t and o phases
5251
5 :1
as indicated in Table IV and Fig. 4. The appropriate values
for Z, M, and V are obtained from Table 111. For the c, t , 0,
and S phases, this involves a calculation of V from the fitted
functions, For the m, due to of the lack of data on its expan-
sion in Mg-PSZ and doubts about the relevance of literature
data on pure ZrOz, the 293 K estimate of V has been used at
'37L all temperatures. Since the amount of m in the sample is
small, this neglect of m thermal expansion does not seriously
affect the results.
Figure 7 presents the results from dilatometry and com-
pares these with calculations based on the neutron diffraction
analyses of the similar sample. The dilatometry indicates that
0 1M) 200 300 4(D 5M) 600 700
the t-to-o transformation commenced quite suddenly at a tem-
Temperature (K) perature of 192 K and that the room-temperature expansion
of the sample due to the t-to-o transformation was 0.15%. It
Fig. 5. Temperature variation of the lattice parameters and unit- confirms that the o-to-t reversion commenced on heating
cell volume of o - Z r 0 2 in a high-toughness Mg-PSZ. Results
from the neutron diffraction (points) are fitted by linear (forb) and above 500 K. The dilatometry suggests that the reversion was
quadratic (for a , c, and V ) functions of temperature. Results at nearing completion at 550 K, whereas the neutron diffraction
496 K have been excluded from the fits. results (Fig. 4) show a reducing o content to temperatures as
 15512916, 1990, 10, Downloaded from https://ceramics.onlinelibrary.wiley.com/doi/10.1111/j.1151-2916.1990.tb06682.x by NRI for Material Physics, INCDFM Bucharest, Wiley Online Library on [17/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
October 1990 Neutron Diffraction Studies of Phase Transformations 2833

Acknowledgments: The authors thank Nilcra Ceramics Pty Ltd., and

particularly Mr. Michael Marmach, for providing the samples. R. J. Hill
thanks the Mineralogisches Institut, Universitat Wurzburg, for its hospital-
$0
Y
ity. The lnstitut Laue-Langevin is thanked for the provision of the neutron
facilities used in this work, as are Dr. Alan Hewat and Mr. John Davies at
a,
that institute for their excellent support.
0,

6
6 References
f ‘R. C. Garvie, R. H. Hannink, and R.T. Pascoe, “Ceramic Steel?,” Nature
0
(London), 258, 703-704 (1975).
a,
_1 ZH.J. Rossell and R. H . J. Hannink, “The Phase MgzZr5012in MgO Par-
tially Stabilized Zirconia”; pp. 139-51 in Advances in Ceramics, Vol. 12,
Science and Technology of Zirconia 11. Edited by N. Claussen, M. Riihle,
and A. H. Heuer. American Ceramic Society, Columbus, OH, 1985.
’D. B. Marshall, M. R. James, and J. R. Porter, “Structural and Mechani-
0 203 400 600 cal Property Changes in Toughened Mg-PSZ at Low Temperatures,” J. Am.
Cerum. Soc., 72, 218-27 (1989).
Temperature (K) ‘E. H . Kisi, C. J. Howard, and R. J. Hill, “The Crystal Structure of
Orthorhombic Zirconia in Partially Stabilized Zirconia,” J. Am. Cerum.
Fig. 7. Length changes during the cooling and subsequent heat- Soc., 72, 1757-60 (1989).
ing of a high-toughnessMg-PSZ. Points are derived (using Eq. (2)) 5R. H . J. Hannink, C. J. Howard, E. H. Kisi, and M.V. Swain, “Relation-
from neutron data on the amounts and densities of the different ship between Fracture Toughness and Phase Assemblage in Mg-PSZ; un-
phases. Solid lines represent length changes measured directly by published work.
dilatometry. 6A. H. Heuer, L. H. Schoenlein, and S. Farmer, “New Microstructural
Features in Magnesia-Partially-Stabilized Zirconia (Mg-PSZ)”; pp. 257-66
in Science of Ceramics, Vol. 12. Edited by P. Vincenzini. Ceramurgica,
Faenza, Italy, 1984.
’B. C. Muddle and R. H. J. Hannink, “Phase Transformations Involving
an Orthorhombic Phase in MgO-Partially-Stabilized Zirconia”; pp. 89-102
in Advances in Ceramics, Vol. 24, Science and Technology of Zirconia 111.
high as 597 K. The discrepancy may be due to sample varia- Edited by S. SGmya, N. Yamamoto, and H . Yanagida. American Ceramic
tion and/or the different thermal histories of the samples ex- Society, Westerville, OH, 1988.
amined. The net expansion at room temperature, after *A. H . Heuer, V. Lanteri, S. C . Farmer, R. Chaim, R.-R. Lee, B.W.
completing the temperature cycle, was about 0.03%. The Kibbel, and R. M. Dickerson, “On the Orthorhombic Phase in ZrOz-Based
Alloys,” J. Muter. Sci., 24, 124-32 (1989).
comparison of the neutron-based calculation with the 9D. K. Smith and H.W. Newkirk, “The Crystal Structure of Baddeleyite
dilatometry involves an arbitrary vertical displacement of (Monoclinic ZrOz) and its Relation to the Polymorphism of ZrOl,” Acru
the calculation, which effectively matches the calculated ref- Crystullogr., 18,983-91 (1965).
erence length to that used in dilatometry. No other adjustable ‘OW. M. Kriven, W. L. Fraser, and S.W. Kennedy, “The Martensite Crys-
tallography of Tetragonal Zirconia”; pp. 82-97 in Advances in Ceramics,
parameters are involved. The agreement between the length Vol. 3, Science and Technology of Zirconia 1. Edited by A. H . Heuer and
changes calculated following diffraction and the measured L.W. Hobbs. American Ceramic Society, Columbus, OH, 1981.
length changes is remarkably good. IlP. M. Kelly, “Martensitic Transformations in Ceramics,” Muter. Sci. Fo-
rum, 56-58, 335-46 (1990).
12P.M. Kelly and C. J. Ball, “Crystallography of Stress-Induced Marten-
sitic Transformations in Partially Stabilized Zirconia,” J. Am. Cerum. Soc.,
V. Summary and Conclusions 69 [3] 259-64 (1986).
I3G.K. Bansal and A. H. Heuer, “On a Martensitic Phase Transformation
The results of this study confirm earlier observation^^,^ that in Zirconia (Zr02)-II. Crystallographic Aspects,” Actu Metall., 22, 409-17
cooling high-toughness magnesia-partially-stabilized zirconia (1974).
14R.H. J. Hannink, “Significance of Microstructure in Transformation
(Mg-PSZ) results in the transformation of substantial amounts Toughening Ceramics,” Muter. Forum, 11, 43-60 (1988).
of tetragonal material to an orthorhombic (0)phase. The vol- 15H.M. Rietveld, ‘A Profile Refinement Method for Nuclear and Mag-
ume of the o phase at room temperature exceeds that of the netic Structures,” J. Appl. Crystullogr., 2, 65-71 (1969).
tetragonal ( t ) by 1.25%. I6D.B. Wiles and R. A . Young, ‘A New Computer Program for Rietveld
Analysis of X-ray Powder Diffraction Patterns,” J. Appl. Crystullogr., 14,
The combination of neutron diffraction and dilatometry 149-51 (1981).
has established that the t-to-o transformation onset tempera- I’R. J. Hill and C. J. Howard, ‘X Computer Program for Rietveld Analysis
ture on cooling was 192 K, that the transformation produced of Fixed-Wavelength X-ray and Neutron Powder Diffraction Patterns,” Aus-
up to 45% by weight of o-Zr02 (sample expansion 0.15% at tralian Atomic Energy Commission Research Establishment, Report No.
AAEC/MllZ, 1986.
room temperature), and that the o reverted to t between I8C. J. Howard, R. J. Hill, and B. E. Reichert, “Structures of the Z r 0 2
500 K and 620 K. Most, but not all, of the o reverted to t, so Polymorphs at Room Temperature by High-Resolution Neutron Powder Dif-
that the sample was largely but not fully restored to its origi- fraction,” Acfu Crystullogr., B44, 116-20 (1988).
I9E.H. Kisi, T. R. Finlayson, and J. R. Griffiths, “Phase Proportions in
nal condition. PSZ Creep Specimens,” Muter. Sci. Forum, 56-58, 351-56 (1990).
The length changes calculated after neutron diffraction are ZO(a)R. J. Hill and C. J. Howard, “Quantitative Phase Analysis from Neu-
based on an elaborate five-phase model (including crystal tron Powder Diffraction Data Using the Rietveld Method,” J. Appl. Crystal-
structure models) of the Mg-PSZ. Dilatometry measures logr., 20, 467-74 (1987). (b) C. J. Howard, R . J. Hill, and M.A. M. Sufi,
length changes directly. The excellent agreement indicates “Quantitative Phase Analysis by Rietveld Analysis of X-ray and Neutron
Diffraction Patterns,” Chem. Aust., 55, 367-69 (1988).
that the model (including the recently determined o crystal 21R.B. Roberts, C. Andrikidis, R. J . Tainsh, and G . K. White, “Thermal
structure) is generally correct. Expansion of an Ice Clathrate”; pp. 499-502 in Proceedings of the Tenth
We believe that the combination of dilatometry and se- International Cryogenic Engineering Conference. Edited by H. Collan,
P. Berglund, and M. Krusius. Butterworth, Boston, MA, 1984.
lected measurements using neutron diffraction (to check bulk 22Y.S. Touloukian, R. K. Kirby, R. E. Taylor, and T.Y. R. Lee, Therrno-
phase composition at critical stages) provides a powerful physical Properties of Matter, Vol. 13, Thermal Expansion-Nonmetallic
means for the study of phase transformations of this type. Solids; p . 451. IPI/Plenum, New York, 1977.