Full text of DOI:10.1007/s004100000217

Contrib Mineral Petrol ꢀ2001) 141: 174±185
DOI 10.1007/s004100000217
W. J. Dunlap á A. K. Kronenberg
Argon loss during deformation of micas: constraints
from laboratory deformation experiments
Received: 12 October 1999 / Accepted: 3 October 2000 / Published online: 16 March 2001
Ó Springer-Verlag 2001
Abstract The role of internal deformation in resetting
Introduction
argon ages of micas has been investigated by measuring
Ar/ Ar ratios of biotite and muscovite, before and
4
0
39
Internal deformation of micas under greenschist facies
conditions is suspected of inducing or enhancing argon
loss, with the e€ect of erasing or obscuring age informa-
tion for events prior to deformation, and providing a
means, in specialized cases, of dating the deformation
after experimentally deforming them. Neither mica cru-
shed cataclastically at room temperature on-line with a
mass spectrometer showed any measurable change in
4
0
39
Ar/ Ar age. Muscovite crystals either sheared along
the K-interlayer and/or kinked at 400 °C and 100±
00 MPa con®ning pressure, exhibit small ꢀ0.7±1.0%)
ꢀ
e.g., Dunlap et al. 1991). Positive correlations between
2
physical grain size and bulk age observed for micas from
mylonites are commonly attributed to either combined
syntectonic grain size reduction and subsequent volume
di€usional loss of argon, or to the growth of ®ne neo-
crystallized grains during deformation and the develop-
ment of two generations of mica in the same rock.
However, it is dicult to rule out that the smallest grains
in mylonites are younger because they have experienced
more internal deformation, and that this deformation has
resulted in greater loss of accumulated radiogenic argon.
reductions in age and marked increases in bulk di€usion
coecients, as determined from argon release during the
initial stages of step-heating between 550 and 810 °C. We
conclude that the relatively young ages of ®ne-grained
mica populations in naturally deformed mylonites result
primarily from grain size re®nement and reductions in
length scale for volume di€usion and/or by syntectonic
neocrystallization. Internal deformation involving dis-
location slip and kinking may contribute to some argon
loss by pipe di€usion, but reductions in closure temper-
ature associated with multipath di€usion are small.
4
0
39
Ar/ Ar ages of mica aggregates from highly de-
formed, greenschist facies rocks are dicult to interpret
because such mica populations ꢀe.g., Fig. 1) are not
generally homogeneous ꢀe.g., Klig®eld et al. 1986).
Micas are weak at low temperatures relative to the
strengths of other rock-forming silicates and they may
be actively deformed while they cool through the closure
window for argon retention ꢀKelley 1988; Dunlap 1997).
Thus, although micas undergo closure, they may also be
undergoing internal strain by dislocation slip, some
grains may be dissolving ꢀperhaps completely) while new
mica grains may be growing ꢀDunlap et al. 1991; West
and Lux 1993), grain sizes may be decreasing because of
strain-related segmentation ꢀe.g., Goodwin and Renne
Supplementary material for this paper can be obtained electroni-
cally by using the Springer LINK server located at http://
dx.doi.org/10.1007/s004100000217.
1
W. J. Dunlap ꢀ&)
UCLA, Department of Earth and Space Sciences,
Los Angeles, CA 90024, USA
E-mail: jim.dunlap@anu.edu.au
1991), and the stoichiometry of grains may be changing
A. K. Kronenberg
through time ꢀe.g., Scaillet et al. 1992; Hames and
Cheney 1997). Some or all of these processes may con-
spire to determine the ®nal argon concentration that we
measure in the laboratory.
One of the most poorly understood of the above
e€ects is internal deformation of micas. Ages of ®ne-
grained micas are commonly younger in deformed rocks
Center for Tectonophysics, Department of Geology
and Geophysics, Texas A&M University,
College Station, TX 77843-3113, USA
Present address:
Research School of Earth Sciences, The Australian
National University, Canberra, ACT, 0200 Australia
1
Editorial responsibility: K.V. Hodges

1
75
nenberg 1993). Mechanisms ꢀ1) and ꢀ3) shorten di€usion
length scales, thereby enhancing di€usive argon loss by
providing shorter pathways, whereas mechanism ꢀ2)
should not have an e€ect if di€usion of argon is parallel
to cleavage. Rather than changing length scales, dislo-
cations introduced during deformation ꢀmechanism 4)
may provide rapid, short-circuit di€usional paths ꢀin
parallel with lattice di€usion) between grain interiors and
grain boundaries ꢀe.g., Lee 1995).
In this paper the relationship between deformation of
mica and argon loss is examined through experiments in
which micas with large radiogenic argon contents are
deformed and mass spectrometry is used to determine
the amount of argon loss, either during or after the
deformation experiments. Two cases are examined in
detail, ꢀ1) cataclastic ꢀbrittle) deformation of biotite,
Fig. 1 Photomicrograph of porphyroclast of muscovite sheared involving cracking and brittle grain re®nement, in ad-
along the basal plane under greenschist facies conditions ꢀsample
dition to some dislocation glide, by crushing specimens
on-line with a mass spectrometer, and ꢀ2) intracrystalline
glide and kinking of muscovite with little or no cracking
7
96 of Dunlap 1997), long dimension ꢀ1.5 mm. Deformation of the
porphyroclast is predominantly by dislocation glide within the
cleavage plane, resulting in the undulatory extinction, but there is
some evidence for the operation of additional deformation induced at elevated pressure and temperature in a tri-
mechanisms, including kinking and cracking ꢀresulting in the ®ner
grains that appear to be torn away from the porphyroclast), and
Mares and Kronenberg 1993).
axial gas-medium deformation apparatus ꢀsamples of
possibly cataclasis and chemical alteration between cleavages
ꢀresulting in the darker amorphous zones). Neocrystallized mica
has also formed in the ®ne grained groundmass ꢀDunlap 1997)
Background
than those of coarse micas, and syntectonic grain size
reduction by dynamic recrystallization and neocrystal- Basal dislocation slip within the K-interlayer of micas
lization provide clear means of obtaining young argon ꢀe.g., Banos et al. 1983) has been activated experimen-
ages by reducing length scales for volume di€usion; tally at temperatures as low as 20 °C and as high as
argon may be released to mobile grain boundaries and it 800 °C ꢀBorg and Handin 1966; Etheridge et al. 1973;
may be excluded from new grains, respectively, during Meike 1989; Kronenberg et al. 1990, 1992; Mares and
recrystallization and neocrystallization ꢀKlig®eld et al. Kronenberg 1993; Noe et al. 1999). Dislocations have
6
±2
1
986; Kelley 1988; Dunlap et al. 1991; Goodwin and been imaged ꢀat densities of ꢀ10 mm ) within the basal
Renne 1991; Hames and Bowring 1994; Dunlap 1997; plane in micas deformed in nature ꢀe.g., Bell et al. 1986)
Reddy and Potts 1999). In contrast, we know little about with the same slip vectors as activated in experiments.
the direct e€ects of deformation. We do not know Optical undulatory extinction of micas ꢀWilson and Bell
whether deformation decreases the ability of a mica 1979; Wilson 1980; Behrmann 1984; Lister and Snoke
grain to retain radiogenic argon, nor do we know any- 1984; Dunlap et al. 1995) is evident in rocks deformed
thing about argon partitioning at dislocations or rates of over a range of pressure and temperature conditions,
pipe di€usion along dislocations.
from the very low grade environments that generally
Micas deform by a variety of mechanisms under lead to cataclasis, to the high-grade environments ꢀi.e.,
greenschist facies conditions ꢀe.g., Wilson and Bell 1979). amphibolite facies) that lead to ductile shear and crystal
Common deformation mechanisms include ꢀ1) kinking plasticity. Recrystallization tends to remove the evidence
during cleavage-parallel shortening and associated seg- for dislocation slip in micas of high grade rocks, yet mica
mentation by the introduction of dilatant kink band porphyroclasts that have escaped recrystallization dur-
boundaries ꢀEtheridge et al. 1973; Bell et al. 1986; Shea ing greenschist facies deformations ꢀFig. 1) o€er dra-
and Kronenberg 1992; Christo€ersen and Kronenberg matic examples of external rotations caused by basal
1993), ꢀ2) cracking and separation along cleavage sur- slip. If argon loss from micas is signi®cantly enhanced by
faces under a wide range of low normal stress states the introduction and glide of dislocations, then the de-
ꢀ
1
1
Obreimo€ 1930; Leonesio 1968; Goodwin and Wenk formation process could have a profound, direct e€ect
990), ꢀ3) cracking across cleavage ꢀGoodwin and Wenk on Ar/ Ar geochronology because it operates below
990; Mares and Kronenberg 1993) during bending or temperatures where closure to volume di€usion of argon
4
0
39
cleavage-parallel extension, and ꢀ4) intracrystalline takes place ꢀ<ꢀ300 °C; Dodson 1973).
plastic deformation via dislocation slip ꢀor glide) within
A case has been made that di€usive loss of argon
the K-interlayer ꢀi.e., within the cleavage) in response to from micas is predominantly parallel to the cleavage
shear stresses resolved on cleavage ꢀDemny 1963; Bell ꢀi.e., in the K-interlayer), and that the characteristic
and Wilson 1981; Banos et al. 1983; Meike 1989; di€usion dimension is the physical radius of the crystal
Christo€ersen and Kronenberg 1993; Mares and Kro- measured within the basal plane ꢀHames and Bowring

176
1994), or some signi®cant fraction of this distance argon pipe di€usion in any silicate. Nevertheless, there
ꢀ
Dunlap 1997). It follows that fragmentation of micas are good reasons to suspect that argon loss may be af-
during deformation, with consequent decreases in fected by the introduction and glide of dislocations.
cleavage-parallel di€usion distances, results in decreased Attractive interactions in a wide variety of crystalline
retention of radiogenic argon within grain interiors. materials between point defects and dislocations lead to
Detailed studies of mica populations in mylonite zones preferential partitioning of solute impurities within dis-
have generally supported this interpretation ꢀe.g., location cores and to di€use local solute concentrations,
Goodwin and Renne 1991).
known as Cottrell atmospheres, surrounding the dislo-
Deformed and recrystallized micas of the Arltunga cations ꢀCottrell and Jaswon 1949; Friedel 1964; Ho and
nappe complex of central Australia ꢀFig. 1) dated by Pratt 1983; Chen 1984; Lee 1995; Gukasyan et al. 1996;
40
39
Dunlap et al. ꢀ1995) using the Ar/ Ar method ꢀtheir Varschavsky and Donoso 1997). Once localized at dis-
Fig. 4) show a remarkably wide range of apparent ages. locations, solutes may be transported parallel to the
The age spectra for aggregates of the course porphyro- dislocation lines by pipe di€usion at rates that exceed
clasts generally rise over the course of step-heating ex- those of solute di€usion through the lattice by many
periments by as much as 1 Ga. Dunlap et al. ꢀ1995) orders of magnitude ꢀe.g., Hart 1957; Suprun 1982;
interpreted these age gradients to re¯ect partial argon Le Claire and Rabinovitch 1984; Arabczyk et al. 1988;
loss from Proterozoic igneous micas brought about Badrour et al. 1989; Shiryaev et al. 1989; Hara and Endo
mostly by late Paleozoic reheating, but also by minor 1997; Estrin and Rabkin 1998). Net transport by pipe
recrystallization. However, the relict igneous grains di€usion is particularly e€ective when dislocations are
ꢀ
large grain in Fig. 1) were pervasively deformed by mobile because exchange of the di€using species is en-
dislocation slip in the Paleozoic, leading to the charac- hanced between moving dislocations and the surround-
teristic textures shown in Fig. 1, and the e€ect of this ing crystal, and because point defect concentration
internal deformation on argon retention remains gradients along the lengths of dislocations are greater
unclear.
during slip than can be maintained within dislocations
Hames and Cheney ꢀ1997) considered the possible that are stationary ꢀRuo€ and Ballu 1963). Transport
causes of Paleozoic deformation-related argon loss from of solutes perpendicular to mobile dislocations may be
Proterozoic muscovite of the Green Mountain massif, enhanced in the direction of dislocation motion as sol-
Vermont. They suggested that, for brief Paleozoic utes are swept up and segregated at the moving dislo-
overprinting events at ꢀ425 °C, di€usive loss of argon cations and as di€use Cottrell atmospheres follow the
played a subordinate role to a combination of chemi- mobile dislocations ꢀCottrell and Jaswon 1949; Friedel
cally-induced and deformation-induced loss of argon. 1964).
They concluded that chemical alteration and deforma-
tion can result in loss of accumulated Ar, but they had tions and other high-di€usivity pathways to the loss of
no means of addressing whether deformation alone radiogenic isotopes from minerals and the implications
Potential contributions of di€usion along disloca-
4
0
might enhance argon loss.
for geochronology have been considered by Lee ꢀ1995).
Klig®eld et al. ꢀ1986), Kelley ꢀ1988), and Goodwin In the case of argon loss from micas undergoing dislo-
4
0
and Renne ꢀ1991) noted dramatic grain size e€ects in cation slip, daughter Ar impurities that initially occupy
micas deformed over the sub-greenschist to amphibolite parent potassium sites may concentrate at dislocations
facies range of conditions. Smaller grain sizes generally by di€usion or by being swept up by mobile dislocations
40
39
recorded younger K±Ar and Ar/ Ar ages and the that glide within the potassium interlayer. If, as Lee
preferred interpretations were that the smaller grains ꢀ1995) suggests, pipe di€usion of argon is three to four
had lost more argon via volume di€usion. These authors orders of magnitude ꢀor more) faster than volume dif-
have also speculated that intracrystalline deformation fusion through the lattice and argon partitions prefer-
may have had an a€ect on the argon systematics, but entially into dislocations, then we might expect greater
they were unable to address this issue in any detail. In argon loss from deformed micas than from undeformed
contrast, Itaya and Takasugi ꢀ1988) noted a lack of micas. Alternatively, if pipe di€usion of argon is rapid
correlation between K±Ar ages and grain sizes of mus- compared with lattice di€usion and dislocations form a
covites from greenschist to amphibolite facies Sanbagwa signi®cant volume fraction of the mica ꢀi.e., dislocations
schists. In their case there is a clear correlation between occur in high densities and/or their e€ective radii
the youngest ages and the lowest metamorphic grades. are large, corresponding to signi®cant Cottrell
Their preferred interpretation was that the argon was atmospheres), then we may obtain the same result.
depleted from the lower grade schists during late syn- However, until now there has been no experimental
metamorphic deformation at ꢀ100±250 °C, the extent study that bears directly on this question.
of the depletion depending on the severity of the
deformation.
Dating studies of experimentally deformed micas are
lacking and the results that are available are incon-
Only few studies have examined quantitatively the clusive. Gusev et al. ꢀ1969) found that dry grinding of
contributions of pipe di€usion along dislocations to net ꢀ250 lm muscovite and phlogopite in a planetary ball
transport in silicates ꢀYund et al. 1981, 1989; Yurimoto mill induced an exponential decline in argon content of
et al. 1989, 1992), and we know of no measurements of the micas to essentially zero concentration after ꢀ10 min

1
77
of grinding. In the initial 1±2 min of grinding, argon loss 11% under similar conditions to those imposed on M-
was pronounced ꢀ>50%), at the same time that infrared 11, loading at 45° to ꢀ001), but with shear parallel to
absorption spectra indicated little change in the struc- [110]. Sample M-48 was deformed by 11% at 400 °C and
tural state of the micas. In the initial stage of grinding, a con®ning pressure of 100 MPa, this time loading the
the micas were reduced to 1-to 10- lm fragments and specimen parallel to ꢀ001) and to [010] to promote kink
they concluded that the argon loss was purely a result of band formation. None of these deformed muscovite
``mechanical destruction''. However, prolonged grinding samples contain recrystallized grains, subgrains, or ®nely
appears to have resulted in transformation of the micas crushed products of cataclasis.
into other phases, most likely as a result of the genera-
Preparation of the muscovite samples for triaxial
tion of locally high temperatures during grinding. Thus, compression was done with a slow-speed wire saw using
the signi®cance of these results is unclear because ther- low loads to minimize damage. Prior to deformation,
mal and mechanical processes may have been competing samples were heat-treated at 400 °C in a mu‚e furnace for
to release the argon. We are aware of only one other 24±36 h and they were further subjected to variable in-
published attempt to release argon by mechanical dis- tervals ꢀof up to 24 h) at <300 °C inside the gas-medium
integration, in which Hanson and Gast ꢀ1967) reduced deformation apparatus. Dry commercial argon gas was
the grain size of ꢀ410 lm biotite using a mortar and used as the high pressure con®ning ¯uid in these defor-
pestle. In that study, grinding did not release any argon mation experiments. Although the isotopic content of this
from the resultant 180-lm fragments ꢀwithin a reported gas is not known, thick silver jackets isolated the mica
uncertainty of 4%).
samples from the con®ning ¯uid, and any contamination
of the samples by this ¯uid would have been detected with
the mass spectrometer. Jacket leaks during the experi-
ments were not detected. Deformation experiments lasted
for about 1.5 h. Further details of these deformation ex-
Methods
A biotite, designated Fe-mica ꢀcf. Grove and Harrison periments are given in Mares and Kronenberg ꢀ1993).
1996), was irradiated to produce Ar ꢀa proxy for K)
39
Two splits of the undeformed Methuen muscovite
and an undeformed split was step-heated for comparison single crystal were irradiated with neutrons and then
with the deformed material. A second split of the Fe- step-heated to provide control on the amount of argon
mica ꢀ5.50 mg; ®ve grains) was crushed on-line with a lost during deformation. Another split of the unde-
40
39
mass spectrometer to measure the Ar/ Ar age of the formed muscovite was irradiated and then crushed on-
released gas. The driving force on the crusher ꢀa modi- line with the mass spectrometer. The undeformed splits
®
ed Varian minivalve; Harrison et al. 1993), applied that were used for step-heating and crushing were not
through a torque wrench, was increased in steps up to heat-treated in a mu‚e furnace prior to analysis.
4
The general procedures followed in Ar/ Ar dating of
0
39
65 inch-pounds and the isotopic content of the gas
evolved during each step was analyzed. Individual micas in this study are described in detail by Dunlap
crushing steps took place ꢀ1 h apart. The crushing ex- ꢀ1997); additional speci®cs are given here. The samples
periments ended when the copper gasket on the assem- were irradiated for 120 h in a Ford reactor, University of
bly was overcome by the driving force ꢀsigni®ed by a Michigan, without cadmium shielding. Samples were
major air component in the analyzed gas). The temper- stacked in single ®le in a silica tube ꢀ5 mm i.d.), with the
ature during slow crushing was probably within a few neutron ¯uence monitor Fish Canyon Sanidine ꢀ27.8 Ma;
degrees of room temperature. We did not attempt to Cebula et al. 1985) interspersed between samples at ꢀ5-
install a thermocouple, although we suspect that the mm intervals. J-values for samples were determined by
thermal inertia of the plunger and base ꢀseveral cubic interpolation to be in the range 0.01957±0.01967‹0.3%
centimeters of stainless steel) would have maintained the for the Methuen muscovite and 0.01524‹0.3% for the
tiny micas to within a few degrees of room temperature. single Fe-mica sample, with total variation along the
After crushing, the fragments were sieved into four size length of the tube occupied by the samples ꢀ1 cm) of <1%.
fractions, <25, 25±75, 75±150, and >150 lm for step- During step-heating, sample temperature was monitored
heating. The 75±150 lm fraction was not analyzed be- by a thermocouple at the base of a tantalum crucible
cause no signi®cant change in age was observed for the within a double-vacuum resistance furnace. Experiments
other three fractions.
were started in the 400±550 °C range and concluded at
Three muscovite samples cut from a single crystal 1,500 °C. Heating schedules are listed in the supplemen-
from Methuen Township, Ontario; Hurlbut 1956) and tary data tables; each heating step was for 10±25 min de-
ꢀ
deformed in a triaxial deformation apparatus ꢀMares pending on the amount of gas desired for analysis.
and Kronenberg 1993) were subsequently irradiated At each extraction temperature, the gas released was
with neutrons and then analyzed for their argon isotopic exposed to Zr±Al getter pumps to remove active gases,
content. Muscovite Sample M11 was deformed to 14% and the argon was subsequently isotopically analyzed
strain at 400 °C and a lateral con®ning pressure of using a VG1200 gas-source mass spectrometer operated
2
4
00 MPa by applying a di€erential compressive load at in the static mode. Measurements were made with an
5° to ꢀ001) to promote dislocation slip with shear tak- electron multiplier ꢀwith an Ar sensitivity of about
±
17
±1
ing place parallel to [310]. Sample M22 was shortened by 5´10 mol mV ). Getter section blanks were normally

1
78
±
16
below 5´10 mol on mass 40 for static collection over
0 min. Blanks were atmospheric, as well as could be
determined, and the mass 39 blank ꢀ10 min) was nor-
1
±
18
mally below 2´10 mol. Furnace blanks below 1,000 °C
±
15
40
40
were generally well below 5´10 mol of Ar. The
K
a )
±
10 ±1
abundance and decay constants ꢀk=5.543´10
recommended by the IUGS Subcommission on Geo-
chronology were used ꢀSteiger and Jager 1977). Data
were corrected for mass spectrometer discrimination,
37
39
line blanks, and for the decay of Ar and Ar during
39
and after irradiation. Amounts of Ar listed in the
supplementary data tables are derived from the mea-
sured sensitivity of the mass spectrometer ꢀuncertainties
of ꢀ5%). Totals are the percent Ar weighted means of
the analyses. Correction factors for Methuen muscovite
determined from irradiated salts are the following: ꢀ40/
39
39) =0.025; ꢀ36/37) =0.00026; ꢀ39/37) =0.00072;
K Ca Ca
ꢀ
38/39) =0.012. For the Fe-mica these values are ꢀ40/
K
39) =0.021; ꢀ36/37) =0.00024; ꢀ39/37) =0.00074;
K Ca Ca
ꢀ
38/39) =0.012. These factors have measured errors
K
eo xf p <e riꢀm 5e %n t ,s .and thus little a€ect the outcome of the
Results
Fe-mica
The age spectrum for the undeformed split of the Fe-
mica is somewhat erratic, not forming an ideal plateau,
but it is essentially ¯at over the entire gas release, yielding
an integrated weighted mean age of 307.2 ‹ 0.8 Ma
ꢀ
ꢀ
Fig. 2a); all ages quoted with one sigma uncertainty
results for all experiments are tabulated in supplemen-
tary database). This result is in agreement with the
07.6‹0.4 Ma weighted mean age determined by Grove
3
and Harrison ꢀ1996) from eight total fusion measure-
ments on aliquots of several milligrams of the Fe-mica.
A limited amount of gas was extracted from the ali-
quot of Fe-mica crushed on-line with the mass spec-
trometer. Using the measured sensitivity of the mass
spectrometer, the weight of the crushed aliquot, and the
argon concentration of the mica, we estimate that
3
9
ꢀ
2.5% of the total available Ar was released during
crushing. The age spectrum of the crush release experi-
ment ꢀFig. 2b) exhibits individual ages that range from
2
89 to 328 Ma. Although the age spectrum is markedly
4
0
39
erratic, the integrated gas age is 307.3‹4.3 Ma, which is Fig. 2 Ar/ Ar age spectra for aliquots of Fe-mica. a Unde-
formed starting material, same as used by Grove and Harrison
ꢀ
identical within uncertainty to that of the step-heat
measurements of the undeformed material.
The four sieve fractions of the crushed Fe-mica were
1996). b Crush release from undeformed aliquot of Fe-mica
39
comprising only ꢀ2.5% of the available Ar. c Age spectra for
resized fractions of crushed Fe-mica: <25, 25±75, and >150 lm.
roughly equal in weight, with a total weight of 4.45 mg: The 75-to 150- lm-size fraction was not analyzed
<
ꢀ
25 lm ꢀ1.02 mg), 25±75 lm ꢀ1.22 mg), 75±150 lm
1.18 mg), and >150 lm ꢀ1.03 mg). Some of the crushed
material ꢀ1.05 mg) was ``lost'' in the sieving process. determined for the undeformed material. The age spec-
Step-heating of the <25-, 25±75- and >150-lm crushed tra ꢀFig. 2c), however, are more variable than that of the
fractions yielded age spectra with integrated weighted undeformed material, with the two ®nest fractions
mean ages of 309.5‹1.3, 308.4‹0.9, and 305.9‹1.1 Ma, yielding distinctly higher ages in the latter portion of gas
respectively. These integrated ages are within 1% of that release.

1
79
Methuen muscovite
The Methuen muscovite specimen had an original size of
at least 50 mm measured perpendicular to cleavage, and
4
1
5´90 mm within the cleavage ꢀMares and Kronenberg
993). The cleavage section has the general shape of a
rhombus. The two undeformed splits, designated M-A
and M-B, were cut at ꢀ1 and 2 cm from the margin of
the crystal. M-B is a rectangular section, in the cleavage,
measuring ꢀ2.5´5mm ꢀꢀ120 lm thick), with minimum
dimension in the cleavage of ꢀ1,200 lm; M-A is an
isosceles triangular section with 3 mm short dimension
and 5 mm long dimensions in the cleavage, with mini-
mum distance to geometric center in the cleavage of
ꢀ
1,200 lm and a thickness of ꢀ200 lm. The splits ex-
tracted for the deformation experiments of Mares and
Kronenberg ꢀ1993) were extracted from the core of the
crystal and the results below show that possible inho-
mogeneity in argon concentration is not an issue.
The step-heating analyses of undeformed splits M-A
ꢀ
ꢀ
4.69 mg) and M-B ꢀ4.25 mg) yield similar age spectra
Fig. 3a) with integrated ages of 1,043‹1 and
1,044‹1 Ma, respectively. Anomalously high ages are
evolved in the ®rst 10% of gas release in both cases, with
a large fraction at ꢀ1,386 Ma for M-A and one at
ꢀ
1,098 Ma for M-B. The remainder of the gas release
for both samples is essentially ¯at. For sample M-A a
plateau age of 1,031‹1 Ma is formed by the last 77% of
gas release ꢀweighted mean age; i.e., all steps overlap
with the weighted mean at the 2r level). The result for
sample M-B forms a plateau at 1,040‹1 Ma for the
interval of 27±97% of gas release. These two age spectra
are non-ideal in light of the observation that micas do
not generally exhibit strongly elevated ages in the initial
ꢀ
10% of gas release.
One split ꢀ4.54 mg) of the M-B undeformed musco-
vite was crushed on-line with the mass spectrometer
Fig. 3b). Using the measured sensitivity, the weight of
the split, and the argon concentration of the step-heated
ꢀ
3
9
split, we estimate that ꢀ0.53% of the total available Ar
was released during crushing. The ages evolved during
crushing declined continuously from 2,470±1,221 Ma
and the amount of gas in each step increased as the
torque on the crusher driver was increased to ꢀ65 inch-
pounds. The integrated age of the gas is ꢀ1,381 Ma,
which is ꢀ33% older than the integrated age of the mica.
The modes of deformation and strain heterogeneity
varied markedly in Methuen muscovite samples de-
4
0
39
Fig. 3 Ar/ Ar age spectra for aliquots of Methuen muscovite. a
Age spectra for two undeformed aliquots of the starting material;
not heat treated. b Crush release from split of undeformed aliquot
3
9
formed in the triaxial deformation apparatus, depending M-B; gas comprises only ꢀ0.54% of the available Ar. c Age
on the crystallographic orientation of loading ꢀMares spectra for sheared aliquots of Methuen muscovite deformed by
Mares and Kronenberg ꢀ1993); samples are M-11, M-11b ꢀa second
split), M-22, and the kinked portion of M-48
and Kronenberg 1993). Sample M-11, loaded at 45° to
cleavage ꢀand to [310]), was sheared by basal dislocation
glide resulting in shear strains that are relatively ho-
mogeneous ꢀFig. 4a of Mares and Kronenberg 1993). direction of [110]), and optical deformation micro-
Aliquots of 4.16 and 6.03 mg ꢀmeasuring about structures are similar to those of M-11; some displace-
ꢀ
1.5´4 mm rectangular in the cleavage; thicknesses ment may also have occurred on several cleavage
250 and ꢀ350 lm respectively) were extracted from the surfaces, but pervasive dislocation slip is still believed to
ꢀ
margin of the charge for step-heating. Sample M-22 was have been the predominant deformation mechanism at
deformed in a similar geometry ꢀthough with a shear the con®ning pressure ꢀ200 MPa) imposed. A 5.35-mg

1
80
portion of the sample is intense with neighboring kink
band boundaries de®ning domains of 2 lm to 400 lm in
scale ꢀFig. 5b of Mares and Kronenberg 1993). A 4.80-
mg aliquot of this kinked material was excavated for
step-heating; the analyzed fragments are generally rect-
angular in the cleavage and they exhibit a grain size
distribution similar to that shown in Mares and Kro-
nenberg ꢀ1993, their Fig. 5b).
The four step-heating analyses of deformed Methuen
muscovite samples ± two from M-11 and one each from
M-22 and M-48 ± yield remarkably similar age spectra
ꢀ
Fig. 3c). These age spectra are also essentially indis-
tinguishable from the step-heating measurements of the
undeformed material; if they were plotted at the scale of
Fig. 3a the age spectra would exhibit nearly continuous
overlap. Anomalously high ages are exhibited in the ®rst
ꢀ
10% of gas release from the deformed samples. In
Fig. 4 Arrhenius plot of lnD/a versus reciprocal temperature for addition, the M-48 spectrum exhibits a rise in age to
the step heat results for the six aliquots of Methuen muscovite
2
1,086 Ma in the last step ꢀand we do not understand
why). The two step-heating measurements of the M-11
muscovite are slightly di€erent, the plateau-like portions
di€ering by as much as 20 million years ꢀM-11b plateau
is younger by ꢀ2%), and the initial elevated ages dif-
fering signi®cantly in volume ꢀthat of M-11b is much
larger). Nevertheless, the integrated ages of the two
spectra measured for deformed sample M-11 are re-
markably similar at 1,033‹1 ꢀM-11b) and 1,036‹1 Ma
aliquot of the sample, measuring ꢀ2´3 mm was ex-
tracted for step-heating ꢀrectangular in the cleavage with
one 2-mm edge forming the margin of the charge;
thickness ꢀ300 lm). The aliquot was extracted from a
region of the charge away from any displaced cleavages
ꢀ
displaced segments indented the jacket). Sample M-48,
loaded parallel to ꢀ001), experienced intense kinking at
one end of the sample ꢀFig. 5a of Mares and Kronenberg
ꢀ
1
M-11), as are the M-22 and M-48 integrated ages of
,036‹1 and 1,037‹1 Ma, respectively. In general, the
integrated ages of the deformed Methuen muscovites are
.7±1.0% below those of the undeformed samples.
1993). Strains within much of this sample are negligible,
whereas shear strains within the kinked material are
much larger than implied by total sample shortening ꢀof
0
Radiogenic argon contents generally >90% indicate
that exchange of argon between the mica starting ma-
terial and atmospheric argon ꢀor the argon con®ning
11%). Deformation within the kinked regions involved
dislocation slip parallel to cleavage and at least two
generations of kink band formation accompanied by
localized opening of cleavage cracks and development of
sub-micron rhombohedral voids at kink band bound-
aries ꢀChristo€ersen and Kronenberg 1993; Mares and
Kronenberg 1993). Kink band segmentation within this
¯
uid) was negligible during the deformation experi-
ments.
Step-heating data are known to exhibit excess scatter
on Arrhenius plots ꢀMcDougall and Harrison 1989) in
comparison with isothermal hydrothermal data. Never-
theless, Arrhenius plots for the six heating experiments
on Methuen muscovite exhibit excellent linear ®ts as
di€usivity increases with increasing furnace temperature
from ꢀ500 to 810 °C ꢀFig. 4); D is the apparent di€u-
o
sivity and a is the characteristic di€usion length scale.
Errors in temperature are within the width of the sym-
bols and a one sigma error in the log of di€usivity is less
than one half of a log unit ꢀe.g., Grove and Harrison
1
996). At temperatures >ꢀ810 °C, the apparent di€u-
sivities depart from the linear relationship, most likely
because of the progressive breakdown of the samples by
dehydroxylation and dehydration ꢀVedder and Wilkins
1
969). For temperatures <810 °C, the samples yield
±1
activation energies ranging from 37.0±50.3 kcal mol ,
which is about the range previously observed for mus-
covite, although higher activation energies seem to be
preferred ꢀe.g., Hames and Bowring 1994).
Samples were step-heated by similar, although not
identical, temperature±time schedules. By 750 °C, the
undeformed M-B muscovite had released only 0.56% of
2
2
Fig. 5 Plot of predicted Dt/a versus observed Dt/a for the step
heat results on Methuen muscovite at 650 °C

1
81
3
9
its total Ar whereas by ꢀ750 °C the sheared muscovite whereas that for the observed values is in the ability to
aliquots M-22, M-11, and M-11b had released 1.7, 1.1, measure fractional loss ꢀ‹<0.5%, drift in machine
and 0.5%, respectively, and by 730 °C the heavily de- sensitivity was <1% during the course of the step-
formed sample M-48 had released ꢀ2.6% of its total heating experiments, and cannot account for the
3
9
Ar ꢀby 780 °C M-48 had released 5.1%).
consistent di€erences between measured bulk ages of
4
0
39
Although Ar/ Ar ages of the deformed Methuan deformed and undeformed aliquots).
muscovite samples are only slightly a€ected ꢀ0.7±1.0% While the measured aliquots vary considerably in
level), apparent di€usivities of Ar determined from shape, making the estimate of A problematic, we con-
quantities of argon released during step-heating seem to sider that the shortest distance within the cleavage is a
be greater than determined for the undeformed starting reasonable estimate of the e€ective di€usion distance for
material for three of the four deformed sample analyses argon escape, within ꢀ10%, and the error bars for
ꢀ
Fig. 4). The scatter expected from the step-heating ``predicted'' values on Fig. 5 re¯ect this uncertainty. Our
method may account for the di€erences between the preferred values for A are 1,200, 750, 1,000, and 40 lm
M-A, M-B, M-11, and M11b results. However, increases for samples M-A, M-11, M-22, and M-48, respectively.
in apparent di€usivity are seen for the M-22 muscovite, The value of A for the M-48 kinked material was esti-
of about two orders of magnitude, and the intensely mated petrographically from 100 measurements on a
2
kinked region of M-48 yields calculated D/a values split of the step-heated material, although it is clear that
some 2.7 orders of magnitude greater than determined this very ®ne material represents a minor amount by
2
for the starting material ꢀlnD/a =±18.1 for M-48 as volume and the ꢀ400 lm material dominates the vol-
2
compared with lnD/a =±20.8 for Methuan muscovite ume. The 1:1 line on Fig. 5 is where the data would plot
starting material).
if our fractional loss data corresponded exactly to loss
from an in®nite cylinder using the above values for A.
Although observed values are equivalent to or exceed
predicted values for M-11 and M-22, that for the kinked
M-48 material is signi®cantly lower than predicted, as-
Estimates of fractional loss
The most important consideration in this study is suming the mean distance of 40 lm between kink band
whether the bulk di€usivity of argon is increased by the boundaries. If 400 lm is taken for the M-48 calculation,
introduction of dislocations and whether this can ac- the observed and predicted values agree, plotting close
count for the results above. The simplest explanation of to the 1:1 line. Thus, the fractional loss observed for the
the apparent increase in di€usivity for M-48 would be intensely kinked material of M-48 is not clearly en-
that the di€usional distance has been severely reduced hanced over what we would expect from single domain
by kinking. One way to assess the possibility of increased di€usion. Note also that M-11b would plot to the left of
di€usivity for the M-48 kinked material is to plot the the y-axis on this plot ꢀsee Discussion section where we
data as log D versus 1/T, using the minimum physical speculate that this is because of re®lling of dislocations,
dimension in the cleavage between kink band bound- rather than di€usion directly out of the grain).
aries as a. However, as the M-48 aliquot is composed of
By projecting the linear ®ts on the Arrhenius plot
a wide range of kink band widths ꢀꢀ2±400 lm), such a back to 400 °C we can estimate the amount of argon that
plot is dicult to interpret. Run times vary as well, so should have been lost via volume di€usion from samples
times for di€usion are not exactly equal.
during pretreatment in the mu‚e furnace and the sub-
Another way of examining argon loss is to plot the sequent deformation experiments. Using the approxi-
2
observed Dt/a versus predicted Dt/a for an in®nite mation for fractional loss from an in®nite cylinder
2
2
cylinder, expressed as a ratio against that of the starting ꢀReichenberg 1953), the projected D/a from the Ar-
material at one temperature ꢀFig. 5). Grain size e€ects rhenius plots, and the times of pretreatment and defor-
and di€erences in run times cancel out on this plot and mation, we estimate that 0.04% of the total argon
the departure from theory should be obvious if the dif- should have been lost during pretreatment ꢀusing Ar-
fusivity does not vary systematically with the square of rhenius data for the starting material), and that <0.01%
2
di€usion dimension. A reference value of Dt/a was should have been lost during the deformation experi-
taken for undeformed sample M-A at T=650 °C and ments ꢀusing the Arrhenius relation ®t to M-48 data), for
ratios of Dt/a ꢀPredicted/reference and Observed/ref- a total of 0.05%. These estimates stand in contrast with
2
erence) are shown for samples M-11, M-22, and M-48. the total gas ages of the deformed aliquots of Methuen
The ``observed'' value of Dt/a for each deformed sam- muscovite, which are up to 1% younger than the un-
2
ple was calculated from the measured fractional losses treated and undeformed starting material. The 1% re-
using the approximation of Reichenberg ꢀ1953), and the duction in age would seem to require ꢀ20 times the loss
predicted value was found from the relationship ꢀDt/ expected from the above calculations.
2
2
2
a ) =ꢀDt/a ) ´ꢀA /A ) , where A is taken as the mini-
Lee ꢀ1995) suggested that available hydrothermal
d
s
s
d
mum physical crystal dimension in the cleavage and the di€usion data on micas is consistent with simultaneous
subscripts s and d refer to starting material and de- multipath di€usion through both the mica lattice and
formed charge, respectively. The largest source of un- extended defects such as dislocations, and there are some
certainty for the predicted values is in the estimate of A, interesting arguments for multiple length scales of

182
di€usion in micas ꢀe.g., Grove and Harrison 1996; results of our experiments may be applicable to some
Dunlap 1997). Lee ꢀ1995) estimates that di€usion of natural situations and provide constraints on the e€ect
argon through dislocations could be faster than that that deformation might have on the argon age of micas.
through a mica lattice by more than 3±4 orders of The crushing experiments performed on biotite most
magnitude. If we assume that the e€ective radius of a closely approximate cataclastic deformations, with the
dislocation ꢀincluding its core and a di€use Cottrell at- production of numerous cleavage-perpendicular cracks,
Ê
mosphere) is 100 A and we take the observed maximum accompanied by the glide motion of dislocations and
6
±2
dislocation density of ꢀ5´10 mm ꢀMares and Kro- propagation of cracks along the cleavage at conditions
nenberg 1993), we obtain a volume fraction for dislo- that preclude any thermally activated argon loss through
Ê
cations in deformed muscovite samples of 0.16% ꢀ50 A the lattice. The triaxial compression of muscovite single
Ê
gives 0.04% and 200 A gives 0.63%), which is of the crystals has resulted in undulatory extinction and the
order of the amount of argon lost during the deforma- development of kink bands reminiscent of microstruc-
tion experiments. Although the dislocation densities of tures observed in micas deformed under greenschist
naturally deformed micas are known for only a few lo- facies conditions, though in the absence of recrystalli-
calities ꢀe.g., Wilson and Bell 1979), we do not expect zation. We anticipate that the deformation in these
densities of naturally deformed micas to exceed those of experiments provides a close analog to deformation of
experimentally deformed samples in many cases, and muscovite under greenschist facies conditions and can be
volume fraction of anomalously high di€usivity sur- used to separate the importance of slip-induced argon
rounding dislocations is unlikely to be more than a few loss from changes in argon content associated with re-
percent. We have modeled the argon loss from Methuen crystallization and resultant changes in di€usion length
muscovite as a multipath di€usion problem, employing scales. The principal shortcoming of these experiments
the Fortran program of Lee and Aldama ꢀ1992). Dif- involves the short time scale of the laboratory defor-
fusional transport out of an in®nite cylinder was exam- mations.
±
1
ined using a frequency factor D =10 s and activation
The results of crushing Fe-mica are not dicult to
energy 43 kcal mol for volume di€usion; these values interpret. Although the crush release spectrum is mark-
0
±1
40
are suggested by the Arrhenius relation ®t to the data for edly erratic ꢀconsistent with di€erential siting of Ar
undeformed Methuen muscovite. For di€usion through and Ar), and the step-heating of the crushed material
3
9
the dislocations we chose the same value for D as for yields non-ideal age spectra, the integrated ages of these
0
volume di€usion, but assumed a smaller activation en- experiments are essentially indistinguishable from that
ergy of 38 kcal mol ; these values predict a di€erence in of the undeformed material. The slightly elevated ages
di€usivity between the dislocations and the lattice of near the end of step-heating of the two ®nest crushed
±1
2
almost four orders of magnitude at 400 °C ꢀlnD /a =± fractions is worth noting, but we would prefer to run
0
21.5 for dislocations and ±25.2 for the lattice). For the further experiments rather than speculate on their sig-
multipath calculations, a number of additional param- ni®cance. The broad conclusion from the crushing of Fe-
eters needed to be speci®ed ꢀsee Lee and Aldama 1992; mica is that no signi®cant argon has been lost in the
±
15 ±1
±15 ±1
j =2´10 s ; j =1´10 s ; c =c or c =1, c =0) crushing process. The implication for cataclasis of micas
1
2
1
2
2
1
depending on whether dislocations are ®lled with or in low temperature natural environments is that argon
empty of the di€using species. For the run conditions loss during mechanical grain size reduction, resulting
and the parameters cited above, the multipath model directly from the deformation process, is not signi®cant.
predicts that the loss of Ar during the relatively brief Therefore, the mechanisms responsible for positive cor-
deformation experiments at 400 °C, added to the loss by relations between grain size and age in natural, catac-
volume di€usion during heat pretreatment, should be lastically deformed mica populations are likely to be
4
0
ꢀ
1%. Although the model parameters are only our best either ꢀ1) di€usive loss of argon over shorter distances
guesses, and we might have used a more retentive model following mechanical grain size reduction, and/or ꢀ2)
for volume di€usion ꢀe.g., Hames and Bowring 1994), it chemically driven neocrystallization, in which new ®ne
is clear from our calculations that the 1% loss of argon mica grains are introduced, leading to increased mica
that we observe for the deformed Methuen muscovite contents and populations with mixed ages.
requires the existence in our samples of fast di€usion The results of the triaxial compression experiments
pathways where argon di€usion is several orders of on muscovite are somewhat more dicult to interpret.
magnitude faster than volume di€usion at 400 °C.
The bulk ages of the undeformed and deformed mus-
covites that were step-heated are distinguishable at the
1% level and the plateau-like portions of the six exper-
Discussion
iments range <2%. This indicates that the level of argon
lost during the course of the deformation experiments is
Not all processes involved in natural deformations of at the limit of resolution of typical dating studies.
micas are replicated in our experiments, nor can we ex- However, the deformed muscovite samples yield
pect to simulate di€usive transport in experimental time systematically lower bulk ages, suggesting a slight
scales with the same relative rates of volume and pipe enhancement of argon loss during deformation. The
di€usion as are important in the Earth. Nevertheless, the slightly lower argon concentration in the deformed

1
83
samples might be attributed to some contribution of seen in micas, but in other minerals such as K-feldspars
relatively rapid pipe di€usion along dislocation net- and amphiboles it is indicative of excess radiogenic ar-
works. This possibility is supported by our estimate that gon contamination. The anomalously high ages evolved
40
only ꢀ0.05% of the total Ar should have been lost by during the crushing experiment on the undeformed split
volume di€usion during the heat treatments prior to ꢀat of the muscovite would seem to con®rm the presence of
4
00 °C for 24±36 h in a mu‚e furnace and ꢀ300 °C for excess argon. One test for excess argon contamination is
ꢀ
24 h) and during deformation ꢀ400 °C for 1.5 h). In to check for positive correlation between elevated ages
38 40
contrast, the multipath model, which combines volume and chlorine-derived Ar. However, plots of Ar*/K
di€usion and fast pathway di€usion in parallel, predicts versus Cl/K ꢀTurner and Wang 1992) for the Methuen
4
1% loss of Ar for the run conditions.
0
ꢀ
muscovite samples do not reveal any correlation. It may
Is it possible to draw conclusions from our data re- be that the elevated ages in the initial gas release result
garding segregation of argon impurities at dislocations from lattice damage during sample preparation; how-
and enhanced mobilities in Methuen muscovite? The ever, the results for crushing and step-heating of the Fe-
challenge is to reconcile the Arrhenius data, the form of mica do not support this possibility; the initial stages of
40
the age spectra, the 1% loss of Ar prior to the step gas release during Fe-mica step-heating experiments re-
heating experiments, the similarity of the Arrhenius re- veal only slightly elevated ages relative to the plateau.
40
lations for deformed and undeformed aliquots, and the An alternative possibility is that loosely held Ar in
observation that loss was enhanced over volume di€u- Methuen muscovite is released from defects during both
sion by a factor of up to ꢀ20.
crushing and the early stages of step-heating, from sites
Activation energies for pipe di€usion are generally that are otherwise devoid of K±Ca±Cl-bearing phases
smaller than those for volume di€usion ꢀBallu 1970; from which argon isotopes might be derived during ir-
Yund et al. 1981; Ho and Pratt 1983; Wolfenstine 1990; radiation. We speculate that this loosely held component
4
0
Sakaguchi et al. 1992) and an increase in the contribu- of Ar resides at defects and that it was accessed pref-
tion of pipe di€usion should lead to reduced values for erentially during the deformation experiments. During
bulk multipath di€usion. However, unless the net step-heating of the undeformed aliquots this component
transport along dislocations becomes very large, ap- was accessed in excess over Ar that had recoiled into
parent activation energies of bulk di€usion rates are not more retentive sites.
3
9
expected to be reduced very much below that of lattice
The age spectra for deformed Methuen muscovite
4
0
di€usion ꢀShewmon 1963; Yund et al. 1981). No such suggest that the loosely held component of Ar ꢀage
change is evident from our Arrhenius plots of argon spikes, ®rst ꢀ15% of gas) is markedly reduced relative
released during step-heating ꢀFig. 4); instead, activation to that in the undeformed mica. The loss of this com-
±
1
energies for the deformed samples ꢀ41.0±50.3 kcal mol ) ponent during deformation and heat pretreatment can
would appear to be the same as or greater than values account for the observed 0.7±1.0% reduction in age of
±
1
for the undeformed samples ꢀ37.0 and 40.2 kcal mol ). the deformed aliquots. The possibility that dislocations
Thus, under conditions of the step-heating analyses have accessed this component and acted to remove it
ꢀ
tribute signi®cantly to the overall argon loss, at least as do we explain the dramatic age spike for M-11b? This
T>450 °C), pipe di€usion does not appear to con- from the deformed aliquots is appealing, but how then
4
0
loosely held component of Ar in sample M-11b is the
re¯ected in the form of the Arrhenius data.
This conclusion may or may not hold at the condi- most dramatic age spike of the deformed samples. Note
tions of the deformation experiments ꢀT=400 °C). The also, however, that the plateau-like segment of M-11b is
4
0
ꢀ
1% reduction in Ar prior to step-heating analyses distinctly lower ꢀꢀ1.5%) than that of the other de-
suggests that a small fraction of the total argon may formed aliquots and that is also has the lowest inte-
have been lost by way of mobile dislocations. At lower grated age. If a loosely held component of argon was
temperatures, the di€erence between di€usion coe- released from dislocations and other defects and lost
cients along dislocations and within the lattice is ex- during the heat pretreatment and deformation of M-
pected to increase, which may lead to extensive drainage 11b, we may speculate that these sites became re®lled
of argon from dislocation networks while argon con- with argon ꢀa second loosely held component) from the
centrations of the lattice remain little changed. In ad- lattice during the initial stages of step-heating. When the
dition, argon loss may have been facilitated during the mica dehydroxylated at 880 °C the lattice-derived argon
deformation experiments by the active slip of disloca- that re®lled the defective sites was lost in a burst. This
tions, increasing the e€ective volume fraction of the lost fraction of lattice-derived argon was then not able to
mica grains that dislocations are able to deplete. Dislo- contribute to the plateau-like segment of the age spec-
cation motions during hydrostatic step-heating, associ- trum, thus, the ꢀ1.5% reduction in plateau-like age.
ated with any recovery of dislocations into stable
con®gurations, were probably very limited.
The Arrhenius data and the age spectra for the de-
formed and undeformed aliquots of Methuen muscovite
The elevated ages of the initial steps during heating of indicate that deformation-related argon loss is minor, in
both the deformed and undeformed Methuen muscovite dating terms. Yet, the consistency between observed
40
are derived from gas that is loosely held relative to the ꢀ0.7±1.0%) and expected ꢀmultipath model) loss of Ar
bulk of the radiogenic argon. This behavior is not often due to heat pretreatment and deformation, and the

1
84
possibility that argon loss was enhanced over volume speculate that the wide range of deformation features
di€usion by a factor of ꢀ20, are all consistent with the observed within individual porphyroclasts may a€ect the
possibility that argon has been partitioned into dislo- ability to retain argon, such that the porphyroclasts are
cations and other extended defects during deformation, likely to contain age gradients that correlate spatially
and then lost from the grains via rapid pipe di€usion. In with their deformation features. This may provide
all fairness, however, we must emphasize that this evi- an explanation for the markedly rising age spectra of
dence, although compelling, is still circumstantial.
aggregates of internally deformed micas.
Acknowledgements Laboratory facilities were freely provided by
Prof. Mark Harrison ꢀUCLA). This study bene®ted signi®cantly
from discussions with Matt Heizler and Marty Grove. Dunlap
wishes to thank the National Science Foundation for generous
salary support while at the University of California, Los Angeles as
an NSF Postdoctoral Fellow, and Kronenberg thanks the National
Conclusions
4
0
39
Ar/ Ar dating of micas deformed ꢀ1) by crushing
cataclastically at room temperature on-line ꢀunder vac-
uum) in a mass spectrometer and ꢀ2) by loading in tri- Science Foundation for support under grant EAR-8816283.
axial compression at T=400 °C ꢀand con®ning pressures
of 100 and 200 MPa) in orientations favoring dislocation
slip and kinking failed to reveal any major changes in
age from those determined for the mica starting mate-
References
rials. Although time scales for deformation di€er in the
laboratory and in nature, these results provide con-
straints on direct e€ects of deformation processes on age
resetting of micas. Argon loss by purely mechanical
processes of cataclasis in the natural environment is
Arabczyk W, Militzer M, Mussig H-J, Wieting J ꢀ1988) Contri-
bution of pipe di€usion to surface segregation kinetics. Surface
Sci 198:167±179
Badrour L, Moya EG, Bernardini J, Moya F ꢀ1989) Fast di€usion
of silver in single and polycrystals of a-alumina. J Phys Chem
Solids 50:551±561
likely to be insigni®cant, as it is in the laboratory. Thus, Ballu RW ꢀ1970) On measurements of self-di€usion rates along
positive correlations observed between grain size and
age in natural cataclastically deformed mica populations
are most likely caused by ꢀ1) di€usive loss of argon after
dislocations in FCC metals. Phys State Solid 42:11±34
Banos JO, Amouric M, De Fouquet C, Baronnet A ꢀ1983) Inter-
layering and interlayer slip in biotite as seen by HRTEM. Am
Mineral 68:754±758
mechanical grain size reduction, and/or ꢀ2) growth of Behrmann JH ꢀ1984) A study of white mica microstructure and
new grains of mica driven by chemical driving forces
microchemistry in a low grade mylonite. J Struct Geol 6:283±
92
2
ꢀ
e.g., neocrystallization), augmenting the overall mica
Bell IA, Wilson CJL ꢀ1981) Deformation of biotite and muscovite:
TEM microstructure and deformation model. Tectonophysics
population and leading to mixed ages.
The percent level argon loss observed in our experi-
7
8:201±228
ments, and the fact that this component is loosely held Bell IA, Wilson CJL, McLaren AC, Etheridge MA ꢀ1986) Kinks in
4
0
39
Ar ꢀin excess over Ar) suggests that argon was par-
mica: role of dislocations and ꢀ001) cleavage. Tectonophysics
27:49±65
Borg I, Handin J ꢀ1966) Experimental deformation of crystalline
rocks. Tectonophysics 3:249±368
1
titioned into defective sites in the natural environment,
accessed preferentially by moving dislocations during
the deformation experiments and lost from the grains via Cebula GT, Kunk MJ, Mehnert HH, Naeser CW, Obradovich JD,
rapid pipe di€usion. However, argon loss from white
micas undergoing dislocation slip within the K-inter-
layer and/or kinking under greenschist facies conditions,
either by processes of pipe di€usion or solute segrega-
Sutter JF ꢀ1985) The Fish Canyon Tu€, a potential standard for
40 39
the Ar/ Ar and ®ssion track dating methods. Terra Cognita
:139
6
Chen I-W ꢀ1984) Anisotropic di€usion of point defects to edge
dislocations. J. Nucl Mater 125:52±63
tion at mobile dislocations ꢀand resultant solute sweep- Christo€ersen R, Kronenberg AK ꢀ1993) Dislocation interactions
in experimentally deformed biotite. J Struct Geol 15:1077±1095
Cottrell AH, Jaswon MA ꢀ1949) Distribution of solute atoms
round a slow dislocation. Proc R Soc Lond 199:104±114
Demny J ꢀ1963) Elektronenmikroskopische Untersuchungen an
sehr dunnen glimmer Folien. A. Naturforsch A18a:1088±1096,
ing), is likely to be limited, and signi®cant changes in
argon ages are not expected. Di€usional argon loss
during the initial stages of step-heating analyses fol-
lowing the triaxial compression experiments suggest that
closure temperatures of intensely sheared and kinked
micas may be somewhat lower than those of unde-
formed, but otherwise identical micas. The ability of
1
097±1101
Dodson MH ꢀ1973) Closure temperature in cooling geochrono-
logical and petrological systems. Contrib Mineral Petrol
4
0:259±274
4
0
39
experimentally sheared and kinked micas to retain argon Dunlap WJ ꢀ1997) Neocrystallization or Cooling?: Ar/ Ar ages
is reduced primarily by reductions in the characteristic
di€usion dimensions associated with the introduction of
kink band boundaries, and to a lesser extent by modest
contributions of pipe di€usion to net transport. We
of white micas from low grade mylonites. Chem Geol 143:181±
03
Dunlap WJ, Teyssier C, McDougall I, Baldwin S ꢀ1991) Ages of
2
40
39
deformation from Ar/ Ar dating of white micas. Geology
19:1213±1216
conclude that the most signi®cant role of deformation to Dunlap WJ, Teyssier C, McDougall I, Baldwin S ꢀ1995) Thermal
and structural evolution of the intracratonic Arltunga Nappe
Complex, central Australia. Tectonics 14:1182±1204
Estrin Y, Rabkin E ꢀ1998) Pipe di€usion along curved dislocations:
an application to mechanical alloying. Scripta Mater 39:1731±
1736
the argon ages of micas in natural mylonites and shear
zones is in grain segmentation and grain size re®nement,
reducing the length scales for volume di€usion, and/or
contributing to the nucleation of new mica grains. We

1
85
Etheridge MA, Hobbs BE, Paterson MS ꢀ1973) Experimental de- McDougall I, Harrison TM ꢀ1988) Geochronology and thermo-
4
0
39
formation of single crystals of biotite. Contrib Mineral Petrol
8:21±36
Friedel J ꢀ1964) Dislocations. Pergamon Press, Oxford
chronology by the Ar/ Ar method. Oxford University Press,
New York
3
Meike A ꢀ1989) In situ deformation of micas: a high voltage elec-
tron microscope study. Am Mineral 74:780±796
tion on Ar retention in biotites from the Santa Rosa Mylonite Noe DC, LaVan DA, Veblen DR ꢀ1999) Lap shear testing of
Goodwin LB, Renne PR ꢀ1991) E€ects of progressive mylonitiza-
Zone, California, and thermochronologic implications. Contrib
Mineral Petrol 108:283±297
Goodwin LB, Wenk H-R ꢀ1990) Intracrystalline folding and ca-
biotite and phlogopite crystals and the application of inter-
ferometric strain/displacement gages to mineralogy. J Geophys
Res 104:17811±17822
taclasis in biotite of the Santa Rosa mylonite zone: HVEM and Obreimo€ JW ꢀ1930) The splitting strength of mica. Proc R Soc
TEM observations. Tectonophysics 172:201±214 Lond Ser A 127:290±297
Grove M, Harrison TM ꢀ1996) Ar di€usion in Fe-rich biotite. Reddy SM, Potts GJ ꢀ1999) Constraining absolute deformation
4
0
*
Am Minerol 81:940±951
Gukasyan A, Kvit A, Klevkov Y, Oktyabrsky S ꢀ1996) High-res-
ages: the relationship between deformation mechanisms and
isotope systematics. J Struct Geol 21:1255±1265
olution PL characterization of impurity segregation and their Reichenberg D ꢀ1953) Properties of ion-exchange resins in relation
complex formation on extended defects in CdTe. Solid State to their structure: III. Kinetics of exchange. Am Chem Soc J
Commun 97:897±902 75:589±597
Gusev GM, Arkhipenko DK, Doil'nitsyn EF, Klyarovskiy VM Ruo€ AL, Ballu RW ꢀ1963) Strain-enhanced di€usion in metals.
ꢀ1969) E€ect of ultra®ne grinding on micas and their argon II. Dislocations and grain-boundary short-circuiting models.
content. Geochem Int 6:80±88 J Appl Phys 34:1848±1853
Hames WE, Bowring SA ꢀ1994) An empirical evaluation of the Sakaguchi I, Yurimoto H, Sueno S ꢀ1992) Self-di€usion along
argon di€usion geometry in muscovite. Earth Planet Sci Lett
24:161±167
dislocations in single-crystal MgO. Solid State Commun
84:889±893
1
4
0
Hames WE, Cheney JT ꢀ1997) On the loss of Ar* from muscovite Scaillet S, Feraud G, Ballevre M, Amouric M ꢀ1992) Mg/Fe and
during polymetamorphism. Geochim Cosmochim Acta [ꢀMg,Fe) Si±Al ] compositional control on argon behaviour in
1:3863±3872
2
4
0
39
6
high-pressure white micas: A Ar/ Ar continuous laser-probe
study from the Dora±Maira nappe of the internal western Alps,
Italy. Geochim Cosmochim Acta 56:2851±2872
Hanson GN, Gast PW ꢀ1967) Kinetic studies in contact meta-
morphic zones. Geochim Cosmochim Acta 31:1119±1153
Hara K, Endo T ꢀ1997) Numerical approach to the I-S e€ect in Shea WT, Kronenberg AK ꢀ1992) Rheology and deformation
mechanisms of an isotropic mica schist. J Geophys Res
97:15201±15237
ferritic heat-resisting steels. Mater Sci Eng A234:234±236
Harrison TM, Heizler MT, Lovera OM ꢀ1993) In vacuo crushing
experiments and K-feldspar thermochronology. Earth Planet Shewmon PG ꢀ1963) Di€usion in solids. McGraw-Hill, New York,
Sci Lett 117L:169±180 p 178
Hart EW ꢀ1957) On the role of dislocations in bulk di€usion. Acta Shiryaev SY, Nylandsted Larsen A, Safronov N ꢀ1989) Rapid
Metall 5:597
thermal annealing of indium-implanted silicon single crystals.
J Appl Phys 65:4220±4224
Steiger R, Jaeger E ꢀ1977) Subcommission on geochronology:
convention on the use of decay constants in geo-and cosmo-
chronology. Earth Planet Sci Lett 36:359±362
Ho YK, Pratt PL ꢀ1983) Dislocation pipe di€usion in sodium
chloride crystals. Radiat E€ects 75:183±192
Hurlbut CS Jr ꢀ1956) Muscovite from the Methuen Township,
Ontario. Am Mineral 41:892±898
Itaya T, Takasugi H ꢀ1988) Muscovite K±Ar ages of the Sanbag- Suprun IT ꢀ1982) Forced pipe di€usion of mobile pinning points.
awa schists, Japan and argon depletion during cooling and
deformation. Contrib Mineral Petrol 100:281±290
Phys State Solid A 74:411±419
Turner G, Wang S ꢀ1992) Excess argon, crustal ¯uids and apparent
isochrons from crushing of K-feldspars. Earth Planet Sci Lett
110:193±211
Varschavsky A, Donoso E ꢀ1997) Modelling the kinetics of solute
segregation to partial dislocations for isothermal microcalori-
metric evaluations. J Thermal Anal 50:533±545
Kelley S ꢀ1988) The relationship between K/Ar mineral ages, mica
grain sizes and movement on the Moine thrust zone, NW
Highlands, Scotland. J Geol Soc Lond 145:1±10
Klig®eld R, Hunziker J, Dallmeyer RD, Schamel S ꢀ1986) Dat-
ing of deformation phases using K±Ar and Ar/ Ar tech-
4
0
39
niques: results from the Northern Appenines. J Struct Geol Vedder W, Wilkins RWT ꢀ1969) Dehydroxylation and rehydr-
8
Kronenberg AK, Kirby SH, Pinkston J ꢀ1990) Basal slip and
:781±798
oxylation, oxidation and reduction of micas. Am Miner 54:482±
509
mechanical anisotropy of biotite. J Geophys Res 95:19257± West DP, Lux DR ꢀ1993) Dating mylonitic deformation by the
4
0
39
1
9278
Kronenberg AK, Mares VM, Christo€ersen RG ꢀ1992) Deforma-
tion of micas: easy glide and e€ects of crystal chemistry and Wilson CJL ꢀ1980) Shear zones in a pegmatite: a study of albite±
Ar- Ar method: an example from the Norumbega Fault
Zone, Maine. Earth Planet Sci Lett 120:221±237
structure on ꢀ001) dislocation motion. EOS Trans Am Geophys
Union 73:528
Le Claire AD, Rabinovitch A ꢀ1984) The mathematical analysis of
mica±quartz deformation. J Struct Geol 2:203±209
Wilson CJL, Bell IA ꢀ1979) Deformation of biotite and muscovite:
optical microstructure. Tectonophysics 58:179±200
di€usion in dislocations. In: Murch GE, Nowick AS ꢀeds) Wolfenstine J ꢀ1990) E€ects of pipe di€usion on the creep behavior
of dry olivine. Philos Mag 62:233±238
Di€usion in crystalline solids. Academic Press, London, pp 257±
18
Lee JKW ꢀ1995) Multipath di€usion in geochronology. Contrib
Mineral Petrol 120:60±82
Lee JKW, Aldama AA ꢀ1992) Multipath di€usion: a general
numerical model. Comput Geosci 18:531±555
3
Yund RA, Smith BM, Tullis J ꢀ1981) Dislocation-assisted di€usion
of oxygen in albite. Phys Chem Mineral 7:185±189
Yund RA, Quigley J, Tullis J ꢀ1989) The e€ect of dislocations on
bulk di€usion in feldspars during metamorphism. J Metamorph
Geol 7:337±341
Leonesio RB ꢀ1968) An observation made on the fracture prop- Yurimoto H, Morioka M, Nagasawa H ꢀ1989) Di€usion in single
crystals of melilite: I. oxygen. Geochim Cosmochim Acta
53:2387±2394
erties of mica. Eng Fract Mech 1:237±238
Lister GS, Snoke AW ꢀ1984) S±C mylonites. J Struct Geol 6:617±
6
38
Yurimoto H, Morioka M, Nagasawa H ꢀ1992) Oxygen self-di€u-
sion along high-di€usivity paths in forsterite. Geochem J
26:181±188
Mares VM, Kronenberg AK ꢀ1993) Experimental deformation of
muscovite. J Struct Geol 15:1061±1075