Materials Science Forum Vols. 467-470 (2004) pp. 1243-1250
online at http://www.scientific.net
© 2004 Trans Tech Publications, Switzerland
Dynamic recrystallisation of quartz
John Wheeler1, a, Zhenting Jiang1,2 , David Prior1 and Jan Tullis3
1
Dept. of Earth and Ocean Sciences, University of Liverpool, Liverpool, L69 3GP, UK
2
Department of Geology and Geophysics, Yale University, P.O. Box 208109, New Haven, CT
06520-8109, USA
3
Department of Geological Sciences, Brown University, Providence, RI 02912, USA
a
[email protected]
Keywords: dynamic recrystallisation, nucleation, quartz, subgrain rotation
Abstract. It is generally agreed that the driving force (plastic strain energy) is much too small to
allow "classical" nucleation during static and dynamic recrystallisation, and that rotation/growth of
subgrains is an alternative. The latter explanation predicts that new grains should begin at low
angles to old grains. We have used electron backscatter diffraction on an experimentally deformed
quartz polycrystal that has deformed by dislocation creep and partially recrystallised. In a region
shortened by about 30% new grains are at high angles (much greater than 15º) to adjacent parent
grains. A histogram of misorientation versus number of boundaries shows a gap at 15-20º. In its
simple form we expect the subgrain rotation model to predict a spectrum of misorientations but with
most of them being low angle. Instead, the histogram suggests that new boundaries began life as
high-angle structures, so current models for deformation-induced nucleation require refinement.
Introduction
Static and dynamic recrystallisation occur in rocks after or during deformation, for the same
general reasons that they occur in metals and ceramics. Earth scientists require an understanding of
these processes to gain insight into the strength of rocks and to enable interpretations of
deformation processes from microstructural observations. The initiation of new grains during
recrystallisation is not thought to occur by a classical nucleation process [1] but by a combination of
subgrain rotation (building up boundaries of progressively higher misorientation which eventually
become grain boundaries) and grain boundary migration. New grains should then, at low strains,
have a low angle misorientation relationship to their parent grain. Here we report the results of an
electron backscatter diffraction (EBSD) study on an experimentally deformed quartz (trigonal SiO2)
polycrystal. We show that the new grains formed by early stages of dynamic recrystallisation are at
high angles to relict parent grains, and infer that existing models for nucleation and growth require
moification.
Starting material
The starting material was a natural quartz polycrystal (Black Hills quartzite) which has a
well equilibrated undeformed microstructure. Grains show no internal plastic strain and the
polycrystal has no overall lattice preferred orientation. The rock contains polygonal grains of rather
constant size (80 – 100 µm) and consists of ~99% quartz with ~1% feldspars and oxides and an
average porosity up to ~1-2% by volume.
Experimental conditions
Sample BA-42 with 0.17 wt % water added was deformed by uniaxial compression to 60%
shortening at 900°C and 1.2 GPa confining pressure. It was deformed by dislocation creep (at
conditions classified as regime 3 in previous work [2]) and is characterised by grain boundary
migration recrystallisation. The added water causes softening, and leads to microstructures
comparable to those formed in dry quartz polycrystals at higher temperatures. The spatially
Licensed to John Wheeler ([email protected]) - University of Liverpool - UK
All rights reserved. No part of the contents of this paper may be reproduced or transmitted in any form or by any means without the
written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 138.253.112.107-05/11/04,13:29:28)
1244
Recrystallization and Grain Growth
averaged strain rate was 10-6 s-1 but there was a strain gradient between the two pistons over a
distance of 5 mm, from weakly deformed (partially recrystallised) near one piston to more strongly
deformed and completely recrystallised grains near the other. This contribution concerns only the
relatively low-strain part, which was shortened substantially less than 60%.
EBSD methodology
EBSD gives a method for determining the complete lattice orientation at a point in a
crystalline material, using a Scanning Electron Microscope [3, 4]. In the last 10 years EBSD has
become automated, in the sense that hardware and software are available to enable a regular array
of points to be measured and the results displayed as a map and analysed in a variety of ways [3].
Such maps would be appropriate for our investigation and could be used in future work; however at
the time our data were gathered, automatic EBSD analysis of quartz gave rise to substantial
misindexing problems. Instead, we analysed a discrete set of points from each subarea of the
sample. First, Orientation Contrast (OC) images from each subarea were collected using a Phillips
XL30 Scanning Electron Microscope. These show greyscale contrasts between domains of differing
P5
P1
P3
P2
P4
Figure 1. Map of deformed quartz sample showing all boundaries >15˚ (thick lines) and direction
of <0001> at each measurement point (double headed arrows). The length of the arrows indicates
how parallel each <0001> is to the map plane. Scale bar is 100 µm. Thin lines surround holes in the
slide induced during preparation.
orientation. An image analysis program [5] was used to pick out boundaries between orientation
domains. This initial boundary map was then used as a guide to where to put EBSD measurement
points, which were analysed using Channel 3.1 software. More EBSD points than domains were
used and this allowed a revision and refinement of the boundary map [5]. The result is a map with
domains that are assumed internally homogeneous (one analysis point per domain). Each boundary
Materials Science Forum Vols. 467-470
1245
(a)
(b)
P4
(c)
73º
P5
P3
64º
P3
P1
P1
P2
(e)
(d)
P5
P3
48º
P4
49º
P5
59º
90º
Figure 2: Pole figures for <0001> for (a) all measured points, (b) parent grains only. Pole figures
(c)-(e) show daughter grains lying between a pair of parent grains; a representative misorientation
between each cluster of parent orientations and the daughter region is shown by the arrows labelled
with angles.
1246
Recrystallization and Grain Growth
between domains has a misorientation that can be calculated from the EBSD measurements on
either side. Boundary maps can then be displayed (e.g. Fig. 1) showing any subset of boundaries
selected on the basis of their misorientation angles [3].
EBSD results
We present results for the low strain part of the sample. We find an abundance of low-angle
boundaries (illustrated in detail in [6]) decreasing in abundance with increasing misorientation to
15º. There is then a population of much higher angle boundaries. The misorientation at which a
quartz/quartz boundary may be classified as a grain boundary may be >10º [7] but we would expect
all boundaries above 15º to be grain boundaries. The pattern of boundaries with misorientation >15º
is shown in Fig. 1 (in fact many of those illustrated are much higher angle than that). The <0001>
axis orientation at each measured point is also shown which allows some visualisation of the local
lattice orientation, though the orientations of other crystal directions are not shown. Measured
points with subparallel <0001> axes do not necessarily have small relative misorientations. There is
a pattern of large, crudely elliptical grains separated by bands of smaller, irregular grains. The
larger grains are interpreted as the relict cores of original polygonal grains. The smaller and/or very
irregular grains have formed by recrystallisation from these “parent” grains. This can be described
as a “core and mantle” microstructure and has also been seen in naturally deformed quartz rocks
[8]. We have labelled grains we are fairly confident as parents as P1-P5. The axial ratio of P3 is
roughly 1.8; bearing in mind that uniaxial compression will have given rise to extension at right
angles to this map as well as in map view, this implies a shortening near 30%. After correction for
this strain we note that the inferred parent grains have sizes comparable to those in the undeformed
material.
Fig. 2(a) shows all the measured <0001> axes. Fig. 2(b) shows those axes from the five
parent grains, showing there is relatively little internal strain in these, although P3 has some spread
in orientations and has one small segment of internal boundary > 15º as shown in Fig. 1. Figures
2(c) to 2(e) illustrate the large differences in orientation found in the recrystallised mantles between
three pairs of parent grains. The arrows labelled with angles indicate a rough average
misorientation between each parent and the daughter orientations, bearing in mind that in detail this
angle varies along the parent/daughter interfaces. Fig. 3 illustrates, in black, a histogram of
misorientations (weighted according to boundary length, [6]), from >5º upwards. We omit lower
angle boundaries because they are so frequent that they subdue other parts of the histogram.
Superimposed on this, in white, is a histogram showing only those boundaries between parent and
daughter grains. Figs. 2 and 3 highlight the fact that most parent/daughter boundaries are much
greater than 15º, and that daughter orientations are far from those of either adjacent parent grain.
Discussion
In a model of subgrain rotation recrystallisation (e.g. [9, 10]) we expect the gradual buildup of
misorientation in subgrain walls, perhaps around the edges of original grains. Eventually the walls
become high enough angle that they lose much of their internal structure and become grain
boundaries. We would anticipate, then, that most new grain boundaries would be fairly low angle
with respect to one of the parent grains, as illustrated in e.g. dynamic recrystallisation of Al single
crystals [11]. A variation on this is the development of bulges in grain boundaries that may then
become misoriented [10, 12]. Neither model can explain the predominantly high angle relationships
of new grains to both adjacent parent grains. It is likely that higher angle boundaries will be more
mobile than low angle ones (e.g. [13]). This means that boundaries will become more mobile with
time, perhaps as they pass a “threshold” misorientation. The increased mobility means they could
eventually dominate the microstructure as dynamic recrystallisation proceeds (as proposed in [6],
by comparison with higher strain microstructures). The very elongate shape of some daughter
Materials Science Forum Vols. 467-470
1247
grains (e.g. that between P1 and P3, Fig. 1) suggests relatively fast grain boundary migration
subparallel to the original P1/P3 boundary. However, grain boundary migration does not itself
change the orientation of a growing grain. Even with grain boundary migration, then, we would
expect to see some bias towards low-angle misorientations between parent and daughter grain but
such bias is entirely lacking in Fig. 3 – in fact, it shows that parent/daughter boundaries develop, on
average, higher misorientations than parent/parent or daughter/daughter boundaries. We
hypothesise that the new grains nucleated with high-angle boundaries relative to adjacent material.
Viable nuclei originated on or near original grain boundaries, and the elongate shapes of new grains
suggest that their boundaries migrated subparallel to original boundaries faster than into the
interiors of parent grains. We cannot yet explain how these high angle nuclei originate. Computer
simulations of abnormal subgrain growth [13] may be relevant, since they show how a
microstructure can be drastically modified by such a process.
35
30
Line density (/mm)
Line density (/mm)
25
20
15
10
5
0
10
20
30
40
50
60
70
Misorientation
80
90
100
110
Figure 3: Boundary density histograms for all boundaries with misorientations in the range 5-105˚
(black), and for just those boundaries between parent and daughter grains (white superimposed). The
latter, by definition, are >15˚.
Acknowledgements
JW, ZJ and DJP were funded by NERC grant GR3/11768. JT was funded by NSF EAR 0106859.
1248
Recrystallization and Grain Growth
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
R. Cahn, in: Physical Metallurgy, R. Cahn, ed., pp. (1970), 1129-1198, Elsevier, New York.
G. Hirth and J. Tullis, Journal of Structural Geology 14, (1992), 145-159.
B.L. Adams, S.I. Wright and K. Kunze, Metallurgical Transactions a-Physical Metallurgy
and Materials Science 24, (1993), 819-831.
D.J. Prior, A.P. Boyle, F. Brenker, M.J. Cheadle, A. Day, G. Lopez, L. Peruzzo, G.J. Potts,
S.M. Reddy, R. Spiess, N.O. Timms, P.W. Trimby, J. Wheeler and L. Zetterström,
American Mineralogist 84, (1999), 1741-1759.
M. Bartozzi, A.P. Boyle and D.J. Prior, Journal of Structural Geology 22, (2000), 15691579.
J. Wheeler, Z. Jiang, D.J. Prior, J. Tullis, M. Drury and P.J. Trimby, Tectonophysics 376,
(2003), 19-35.
S.H. White, Tectonophysics 39, (1977), 143-170.
S. White, Philosophical Transactions of the Royal Society of London A 283, (1976), 69-86.
J.P. Poirier and A. Nicolas, Journal of Geology 83, (1975), 707-720.
J.L. Urai, W.D. Means and G.S. Lister, in: Mineral and Rock Deformation: Laboratory
Studies, B.E. Hobbs and H.C. Heard, eds., Geophysical Monograph 36, pp. (1986), 161-199,
AGU, Washington, D.C.
H. Yamagata, Y. Ohuchida, N. Saito and M. Otsuka, Scripta Materialia 45, (2001), 10551061.
J.E. Bailey and P.B. Hirsch, Proceedings of the Royal Society of London Series A 267,
(1961), 11-30.
E.A. Holm, M.A. Miodownik and A.D. Rollett, Acta Materialia 51, (2003), 2701-2716.