an evaluation of rb-sr dating of pseudotachylyte...

Geochemical Journal, Vol. 37, pp. 21 to 33, 2003

21

*e-mail: [email protected]

An evaluation of Rb-Sr dating of pseudotachylyte:Structural-chemical models and the role of fluids

JERRY F. MAGLOUGHLIN*

Department of Earth Resources, Colorado State University, Fort Collins, CO 80523, U.S.A.

(Received February 12, 2001; Accepted July 14, 2002)

Rb-Sr geochronometry has been used in several studies in an effort to date pseudotachylyte and there-fore the fault hosting the pseudotachylyte and the tectonism that produced the faulting. The usual proce-dure has been the “thin-slab” technique, which in metamorphic rocks involves construction of an isochronusing whole rock analyses of adjacent but compositionally different slices of rock, and is based on theassumption of isotopic equilibrium between the slices. In attempts to date pseudotachylyte, analyses havebeen obtained using a pseudotachylyte vein and the immediately adjacent host rock. Such studies com-monly yield: 1) apparent ages with high errors owing to small a small range of Rb/Sr values or to“scatterchron” data; 2) host rocks not plotting on the same isochron or scatterchron defined by thepseudotachylyte; 3) apparent dates that are inconsistent with other estimates of the time of faulting; or4) apparent ages that are not independently constrained. These studies commonly fail to record the type ofpseudotachylyte vein utilized; injection veins are commonly larger and more obvious targets for collectingbut are likely to yield geologically meaningless dates.

A Rb-Sr geochronometric study of pseudotachylyte from the North Cascade Mountains, Washington,reveals: 1) an apparent age older than the likely age of the pseudotachylyte; 2) pseudotachylyte and hostrock chemical and isotopic characteristics indicating that the pseudotachylyte was not the result of bulkmelting of normal schist; and 3) metasomatism prior to pseudotachylyte formation likely affected both theprotolith and what is now the host rock.

In light of previous studies, the new data, and related information regarding fault-related metasomatism,three closed system models and one open system model for pseudotachylyte formation are proposed. Thesemodels make specific predictions regarding apparent whole rock-pseudotachylyte dates, microstructuralfeatures, and the relationship between the whole-rock pseudotachylyte date and the whole rock-biotite Rb-Sr dates in the same rock. The models suggests pseudotachylyte-host rock dating is unlikely to producereal ages of pseudotachylyte formation. Worse, because many pseudotachylytes apparently form duringcooling and uplift, the technique will instead yield a seemingly reasonable but geologically meaninglessdate. Certain approaches to the problem, including a microstructural analysis coupled with a whole rock-biotite age on the same rock, may yield an actual age of pseudotachylyte formation, but the assumptionsmade heretofore about the behavior of the Rb-Sr system at the time of pseudotachylyte formation arequestionable.

1999), and therefore provides direct physical evi-dence of discrete structural events that likely oc-curred during protracted tectonic processes. Un-like other fault rocks that evolve during ongoing,progressive deformation, potentially over tens ofmillions of years or more, pseudotachylyte formsin a geologically instantaneous event. Melting on

INTRODUCTION

Pseudotachylyte is a rock locally presentwithin fault and shear zones, formed as a result offriction-induced melting (Magloughlin and Spray,1992). At present, it is the only definite evidenceof paleoseismicity in such rocks (e.g., Cowan,

22 J. F. Magloughlin

the fault surface occurs in seconds, and quench-ing of the melt typically takes seconds to hours.Consequently, successful geochronologic analy-sis of pseudotachylyte has the potential to date notmerely a tectonic episode but a single, discretefault slip episode within an extended tectonicevent, and thus is a valuable exercise.

Several workers have attempted to datepseudotachylyte using the Rb-Sr “thin-slab” tech-nique (Maddock, 1986; Thoeni, 1988; Petermanand Day, 1989; Kamineni et al., 1990; Boeckeler,1994; Adams and Su, 1996). This technique hasbeen used in high grade metamorphic rocks (andalso in certain fault rocks), where small-scalelithologic heterogeneities between layers havingdifferent Rb/Sr ratios offer the possibility of ob-taining a “metamorphic isochron”, if Sr achievesisotopic homogeneity between layers at the timeof metamorphism (e.g., Hofmann, 1979; Peucatand Martin, 1985; Thoeni and Hoinkes, 1987;Adams and Su, 1996). In most cases where thethin slab technique has been applied topseudotachylyte, it has been concluded that geo-logically reasonable dates have been obtained. Insome cases, fault movement and therefore regionaltectonic events have been constrained solely onthis evidence. However, most previous studieshave not considered in detail the basis of suchdating, nor the relationship of such dates to realgeologic events. Furthermore, there has been lit-tle consideration of the geologic mechanisms op-erative within fault zones and shear zones suchthat the requisite conditions for obtaining a validRb-Sr date might be established. The purpose ofthis paper to explore these questions and to con-sider whether Rb-Sr apparent dates onpseudotachylyte are meaningful.

For ease of discussion, I use the term“protolith” to refer to the volume of rock, includ-ing all pre-melting metasomatic modifications ofwhatever origin, in which extensive frictionalmelting occurs. From field and microscale obser-vations and chemical considerations, it appearsthat the protolith can undergo moderate melting,where much material survives as clasts in thepseudotachylyte, to near total melting. A difficulty

exists in that the protolith is by definition exten-sively destroyed during melting, and owing toprobable variations in the protolith along the fault,and flow of and mixing within the melt, the re-sultant pseudotachylyte cannot automatically beassumed chemically equivalent to any particularprotolith. The li thic clasts within thepseudotachylyte might in some cases yield a bet-ter approximation of the protolith than would theimmediately adjacent rock.

In the thin slab technique, in order to obtainan isochron, Rb-Sr analyses of pseudotachylyteare commonly paired with a “whole rock” analy-sis, normally obtained through analysis of rockimmediately adjacent to the pseudotachylyte vein.For better distinction, I will refer to this as the“host rock”. For a Rb/Sr spread to exist in orderfor an isochron to be defined, the host rock andpseudotachylyte must be chemically different(e.g., different mica-plagioclase ratios). The va-lidity of the thin slab approach therefore dependsupon Sr isotopic equilibrium between the host rockand pseudotachylyte at the time of melting.

PREVIOUS Rb-Sr DATES

ON PSEUDOTACHYLYTE

Maddock (1986) analyzed pseudotachylytefrom a concordant fault vein, along with the hostrock from the Outer Hebrides Fault in Scotland.He found that the pseudotachylytes were enrichedapproximately 10–40% in 87Rb/86Sr relative to thehost rock, yielding a scatterchron and apparentdate of 2640 ± 210 Ma (2σ). This apparent date isconsistent with the probable age of the host rocks(Maddock, 1986), but is far older than other esti-mates of the time of faulting, including mid-Paleozoic (Kelley et al . , 1994), and lateProterozoic (Magloughlin et al., 2001).

Thoeni (1988) applied the Rb-Sr thin slab tech-nique to two samples of pseudotachylyte from theeastern Alps. He used 9 thin slabs across each veinto generate scatterchrons (MSWD = 5.24 and2.26); the adjacent host rock in each case did notfall on the scatterchron. Thin slabs within thepseudotachylyte veins ranged considerably in

Evaluation of Rb-Sr dating of pseudotachylyte 23

87Rb/86Sr values (ca. 4–16 and 4–10), possiblyreflecting the mm-cm scale color banding, but didfall within the range of the host rocks (Thoeni,1988). Thoeni interpreted the results to indicatethat at the time the pseudotachylyte formed, vari-ations in Rb/Sr were preserved across the thinslabs and color bands within the pseudotachylyte,but at the same time, Sr achieved isotopic homo-geneity across the same thin slabs and color bands,and thus the apparent dates of 73 ± 3 and 79 ± 5Ma (2σ) represent the time of pseudotachylyte for-mation and faulting. Significantly, these apparentdates fall within the range of biotite-whole rockRb-Sr dates (see discussion below), and alsoamong K-Ar and 40Ar/39Ar white mica dates fromthe same region. Analyses of the adjacent rock(host rock) did not plot along the isochron, andwere not included in the regression. Thoeni wasunable to explain the large range in 87Rb/86Sr val-ues yet apparent Sr isotopic homogenization at thesame scale.

Peterman and Day (1989) analyzedpseudotachylyte, the immediately adjacent hostrock, and “far-field” country rock from two loca-tions near Rainy Lake, Minnesota (U.S.A.) andOntario (Canada). In both locations, they foundpseudotachylyte enriched in Rb (strongly) and inSr relative to the immediately adjacent rock. Basedon the 87Rb/86Sr enrichment of thepseudotachylyte relative to the host, the apparentlack of thermal overprinting since about 2.5 Ga,and the assumption of (1) Sr isotopic equilibriumbetween the pseudotachylyte and immediatelyadjacent rock at the time of formation and (2) sub-sequent closed system behavior, Peterman and Day(1989) obtained a pooled apparent date of 1947 ±23 Ma. They interpreted this date as indicating thetime of pseudotachylyte formation, and attemptedto correlate this apparent age with other tectonicevents in the North American craton.

In discussing their results, Peterman and Day(1989) invoked significant mobility of Rb and Sr,and suggested that Sr achieved isotopic equilib-rium between the host schist and pseudotachylytevein immediately following pseudotachylyteformation, but the system remained closed to later

re-equilibration. They used three observations tosupport elemental mobility: 1) the presence ofconcordant quartz lenses in the fault zone(s);2) small-scale chemical variations in apseudotachylyte vein present between a biotiteschist and a quartz-plagioclase vein; and 3) higherSr and Rb concentrations in the pseudotachylytethan in the host rock. Objections can be raised toall three points, including the timing of the quartzvein formation, and the fact that several of the el-ements involved (their figure 3) showed relation-ships opposite to those expected if diffusion hadbeen extensive. An additional objection is thatwithin foliated rocks, faults are likely to nucleatealong structural anisotropies, such as contacts be-tween mica-rich and feldspar-rich domains. Mica-rich domains are inherently weaker and moreprone to melting. Thus, the protolith, and conse-quently the friction melt and pseudotachylyte,commonly will not be chemically representativeof the bulk rock on a regional or even outcropscale, and this provides a non-diffusion based ex-planation for point (3). Likewise, for mechanicalreasons, the rock immediately adjacent to apseudotachylyte vein is likely to be quartz and/orfeldspar rich, and chemical comparison of this hostrock to the pseudotachylyte is likely to show anapparent depletion of Rb and possibly Sr of thehost rock, as observed in the Peterman and Daysamples. A similar effect may be described byO’Hara (1992) and discussed further in Adams andSu (1996).

Kamineni et al. (1990) report a Rb-Sr hostrock-pseudotachylyte apparent isochron from theSydney Lake Fault Zone in the western SuperiorProvince of North America. Two pseudotachylytesand host rocks were used (no specific informa-tion on the pseudotachylytes or the methodologyis given), yielding an apparent age of 2183 ± 74Ma. The pseudotachylytes yielded 87Rb/86Sr val-ues (4–11 times higher than those of the host rocks.Kamineni et al. (1990) used these data to supporta model of widespread transcurrent faulting alongsubprovince boundaries in the Superior Province,postulating a tectonic link between separateorogens, and to support the timing of tectonism


for which little direct geologic evidence exists.Boeckeler (1994) analyzed pseudotachylytes

from the Great Common Fault Zone in NewHampshire and Maine, where fault movementlikely occurred between the early Devonian andTriassic. He obtained a Rb-Sr scatterchron of298 ± 31 Ma, interpreted as reflecting tectonismpostdating the Acadian orogenic event.

Adams and Su (1996) analyzed thin slabs froma pseudotachylyte (although some uncertaintyexists regarding the nature of these particular faultrocks), and obtained 87Rb/86Sr values on the slabsof approximately 7–10, yielding an apparent ageconsistent with thin slab ages on ultramylonites.However, these samples have higher Rb and lowerSr than nearby samples of gneiss (O’Hara, 1992),a relationship that could have resulted from re-moval of lithic clasts from the pseudotachylyteprior to analysis. In other words, one possible ori-gin for this apparent isochron is through mixingbetween quartzofeldspathic and mica-rich mate-rials.

In summary, most previous published studiesproduced geologically plausible apparent ages thatare generally post-tectonic with respect to the mostrecent major tectonometamorphic event. In gen-eral, most studies appear to assume that co-lin-earity of pseudotachylyte analyses on the isochrondiagram is produced by Sr isotopic equilibrationacross domains in the pseudotachylyte havingvariable Rb/Sr ratios, the latter produced by un-explained mechanisms, even though linearity canbe produced by mixing and other mechanisms(e.g., Field and Råheim, 1979; Wendt, 1993). Mix-ing as a mechanism has generally not been con-sidered or has been discounted, despite the factthat pseudotachylytes commonly show muchhigher Rb/Sr values than their host rocks, valueslikely intermediate between the values for thewhole rock and a high Rb/Sr mineral such as mica.Host rock analyses were either not obtained, orwere included or not included in the isochronsdepending upon whether they were co-linear withthe pseudotachylyte analyses. Finally, the agesobtained through these methods commonly havebeen used in an attempt to establish or constrain

fault movement, and even tectonism affectingmajor orogens.

Rb-Sr ANALYSIS OF PSEUDOTACHYLYTE

FROM THE NASON TERRANE, WASHINGTON

To provide an additional test of this method,Rb-Sr analysis was carried out on twopseudotachylyte veins and their host rocks in asetting where independent knowledge of the ageof the pseudotachylyte is available. The samplesare from the Nason Terrane (Washington, U.S.A.;Tabor et al., 1987), where pseudotachylytes havebeen described previously (Magloughlin, 1989,1992, 1993). Pseudotachylyte occurs in a varietyof host rocks and is widely distributed, formingfault veins and injection veins from less than 0.5mm to 25 cm thick, and extending from as littleas a few mm to over one hundred meters alongstrike. The most common host rocks are graphiticgarnet biotite schist, including the two samplesdescribed below. 40Ar/39Ar analyses indicate anage of about 55 Ma for these pseudotachylytes(Magloughlin et al., 2001).

AnalysesBoth samples (Fig. 1) were obtained from a

locality a few km northwest of Lake Wenatchee,

Fig. 1. Two samples of pseudotachylyte and host rockfrom the Nason Terrane, Washington. Sample 1 is afault vein breccia, whereas sample 2 is a fault vein;both are hosted by pelitic schist.


in Chelan County, Washington. Sample 1 is apseudotachylyte breccia forming a zone severalcm wide, consisting of a black pseudotachylytematrix and about 50% angular to rounded clastsof the host schist . Sample 2 is a blackpseudotachylyte vein 0.5–1 cm thick, with a waxyto vitreous luster, and with inclusions smaller than~0.5 mm. Both samples appear internally homo-geneous and thus a single pseudotachylyte analy-sis of the vein material was obtained for each. Theimmediately adjacent host schist was used for the“host rock” analysis.

The same locality yielded a Rb-Sr biotite-whole rock date of 65.5 ± 2.2 Ma, and a zirconfission track date of 60.5 ± 5 Ma (Magloughlin,1993, 1995). The biotite-whole rock date is some-what younger than other whole rock-mica agesfrom the area; two three-point isochrons(Magloughlin, 1995) and two precise two-pointisochrons (Magloughlin, 1993) all yield ages inthe 79.5–89 Ma range. These ages are interpretedto yield the approximate time of cooling through~300°C. A hornblende separate from thepseudotachylyte locality gave a 40Ar/39Ar age of88.1 ± 0.8 Ma (Magloughlin and Dunlap, unpub-lished data).

The Rb-Sr analyses for this study were madeat the University of British Columbia; the tech-niques were similar to those given in Magloughlin(1995) and owing to instrumental limitations,yielded relatively low-precision 87Sr/86Sr analy-ses (Table 1). In sample 1, the 87Rb/86Sr of thepseudotachylyte is 43% higher than the 87Rb/86Srof the host schist, and the apparent age has a large2σ error, 43 ± 86 Ma (Fig. 2a). Thepseudotachylyte from sample 2 has a 87Rb/86Srabout five times higher than the host and morethan twice the value of any whole rock sample(25 analyses) of Chiwaukum Schist analyzed(Magloughlin, 1993, 1995). The pseudotachylyte-host pair defines an apparent age of 64 ± 5 Ma(2σ) (Fig. 2a). The pseudotachylyte contains 263ppm Rb. Whole-rock samples of ChiwaukumSchist contain 17–96 ppm Rb (25 analyses),whereas biotite in the schist (the only modally sig-

nificant Rb-bearing mineral) contains 167–310ppm Rb (6 analyses) (Magloughlin, 1993, 1995).

Discussion of Nason Terrane pseudotachylytesLittle interpretation is possible from sample 1.

The pseudotachylyte in sample 2 has a much

Fig. 2. a) Rb-Sr isochron diagram for samples 1 and2. Apparent ages are shown, along with the range of25 whole rock analyses (the range the host analyseswould be expected to fall within). The pseudotachylyteanalysis for sample 2 does not fall within the range ofthe 25 schist whole rock analyses, indicating that thepseudotachylyte was not derived by bulk melting ofnormal schist. b) Comparison of host-pseudotachylyteisochron for sample 2 with a whole rock-biotiteisochron on a sample from the same outcrop. Abbre-viations: pst = pseudotachylyte, wr = whole rock.


higher Rb/Sr ratio than its host schist, and thusthe elevated Rb/Sr value cannot be explained bytotal melting of typical schist . Thepseudotachylyte displays no evidence forrecrystallization or alteration, and thus later infil-tration of Rb-rich fluids is unlikely. The onlymodally significant Rb-rich mineral in the hostrock is biotite. Chiwaukum Schist biotite has 87Rb/86Sr values of 6 to 32 (Magloughlin, 1993, 1995).

Two models may explain both the enrichmentof Rb in the pseudotachylyte and the similaritybetween the host rock-pseudotachylyte date andthe host rock cooling dates. In the first, either thepseudotachylyte or its protolith became enrichedin Rb as a result of pre-, syn-, or post-melting proc-esses. Following pseudotachylyte formation, thepseudotachylyte remained open to Sr exchangewith the immediately adjacent host rock, closingto further exchange at approximately 64 Ma. How-ever, there is little evidence for alteration of thevein and therefore for an intergranular fluid ei-ther to serve as a means of post-melting Rb en-richment, or to aid Sr diffusion. Whether or notthere is a fluid phase to allow relatively fastintergranular diffusion rather than volume diffu-sion, at temperatures near 300°C, it is unreason-able that the pseudotachylyte and the host rockshould remain open to Sr exchange on the scaleof several cm, only to close to further Sr exchangeat about the same time as sub-mm-scale biotitewithin the schist. A variation on this model is thatthe pseudotachylyte coincidentally formed at thetime of biotite Rb-Sr closure, with Rb enrichmentand Sr isotopic equilibrium established within theprotolith or at the time of pseudotachylyte forma-tion. This variation is similar to the scenario im-plied by Peterman and Day (1989). In the presentcase, however, mineralogic evidence suggests that

the temperature was considerably below 300°C(Magloughlin, 1993); moreover, thepseudotachylyte is likely 10 million years youngerthan the Rb-Sr biotite-whole rock age(Magloughlin et al., 2001).

An alternative explanation more consistentwith textural and mineralogic observations(Magloughlin, 1992) follows. During the devel-opment of cataclasite zones as a precursor topseudotachylyte-generation, biotite is preferen-tially mechanically and chemically degraded andforms a disproportionately high fraction of thecataclasite and ultracataclasite matrix. Duringpseudotachylyte generation, it is also more sus-ceptible to melting than, for example, the frame-work silicates. Biotite is almost completely ab-sent as crystal fragments in pseudotachylyte fromthe Cascades and elsewhere (Magloughlin, 1992,1993; Spray 1992) whereas plagioclase is muchmore commonly present. It is thus very likely thatthe lithologic precursor and melt are more enrichedin biotite, or more depleted in the main Sr-bear-ing mineral plagioclase, than the bulk schist. Uponformation, the melt should therefore fall on a mix-ing line between the bulk schist and biotite. Ow-ing to the typically high Rb/Sr value of biotite,even a slight enrichment in “biotite component”of the melt would allow a fairly precise apparentdate to be obtained on the host rock-pseudotachylyte (matrix) pair. The nearly identi-cal host rock-pseudotachylyte apparent date andwhole rock-biotite date from the same outcropsupports this (Fig. 2b). This model and the modelof selective biotite contribution to the melt arefurther supported by the composition of apseudotachylyte vein analyzed by Spray (1993).Spray analyzed the matrix to a Nason Terranepseudotachylyte (his sample number NT-5), and

Sample Sr, ppm Rb, ppm 87Rb/86Sr 87Sr/86Sr

1, host 298 39.6 0.385 ± 0.015 0.70621 ± 0.000261, pseudotachylyte 272 51.6 0.549 ± 0.022 0.70631 ± 0.000122, host 241 57.0 0.685 ± 0.027 0.70749 ± 0.000182, pseudotachylyte 220 263 3.464 ± 0.139 0.71003 ± 0.00022

Table 1. Rb-Sr analyses of pseudotachylyte and host rock. Errors are 2σ.


the chemical composition, including water con-tent, is close to that of a typical metamorphicbiotite. Observations of pseudotachylyte gradinginto biotite-rich regions in the host rocks and lithicclasts (Magloughlin, 1993, 1998) also suggests apseudotachylyte composition similar to biotite.

However, in a pseudotachylyte vein formedwithin chemically homogeneous rock, analyses ofthe bulk pseudotachylyte (no mechanical removalof clasts) paired with host rock will not in thisscenario yield a Rb-Sr separation. A Rb-Sr sepa-ration would be obtained through any of the fol-lowing processes: 1) the pseudotachylyte or itsprecursory cataclasite zone forms within a rela-tively higher Rb/Sr zone, which is likely becausebiotite-rich layers are mechanically weaker andfavor fault nucleation and slip; 2) low Rb/Sr,plagioclase-rich clasts are filtered out during mi-gration of the melt; 3) a clast-poor portion of thepseudotachylyte is selected for analysis; 4) clastsrelatively enriched in plagioclase are removedprior to chemical analysis, producing apseudotachylyte (melt) enriched separate, as iscommonly the practice (e.g., Adams and Su,1996).

In summary, the second model requires:1) enrichment of the melt in Rb/Sr via precursorydeformational processes or melting processes; and2) either an initial difference in Rb/Sr between thepseudotachylyte-generating layer and the adjacentlayers, or some process to remove some fractionof the plagioclase-bearing clasts in the analyzed“pseudotachylyte”. We can evaluate this latterprocess by considering the Rb and Sr concentra-tions in biotite and plagioclase, in a model rockin which all of the Rb and Sr are contained in thesetwo minerals. From available Rb-Sr analyses(Magloughlin, 1993, 1995) on biotite, and micro-probe data on SrO and CaO in plagioclase(Magloughlin, unpublished data), Rb concentra-tions of approximately 200 and 20, and Sr con-centrations of 20 and 100 are reasonable valuesfor biotite and plagioclase, respectively. For thesevalues (and more extreme values are possible) re-moval of only 20% of the plagioclase can yieldapproximately 20% 87Rb/86Sr separation between

the protolith and the clast-depleted melt; pairingprotolith-melt would yield a moderately preciseapparent date. As the protolith is a mixture ofplagioclase and biotite, the protolith-clast-de-pleted-melt apparent date will be identical to theplagioclase-protolith-biotite date. With increasingremoval of clasts, the composition of thepseudotachylyte migrates away from the protolithalong the isochron toward biotite, allowing an in-creasingly precise apparent date to be obtained.

MECHANICAL, MELT, AND FLUID EFFECTS

ON THE Rb/Sr VALUE OF PSEUDOTACHYLYTE

Several mechanical and melting processescould have a significant effect on the Rb/Sr andSr isotopic characteristics of pseudotachylyte, andthus are important in evaluating whether Rb-Srdates are valid. As described above, the melt tendsto be enriched in Rb/Sr owing to easier mechani-cal and chemical destruction of the biotite duringcataclasis, relative to plagioclase, and also itsfaster melting following melt generation. This al-lows mechanical separation of the lithic clasts andplagioclase crystal fragments and produces a bulkpseudotachylyte with an elevated Rb/Sr. But sev-eral other processes may influence the composi-tion of pseudotachylytes.

Lithologic contacts—Two different lithologies,simple mixing

Commonly, pseudotachylyte-bearing faultzones develop along lithologic contacts withinregionally metamorphosed and foliated rocks. Inthe North Cascade Mountains, for example, theycommonly develop along schist-metatonalite andschist-quartz vein contacts (Magloughlin, 1989).Assume a simple case where Sr isotopic equilib-rium is achieved between two lithologies havingdifferent Rb/Sr values, separated by a contact (thequestion of the scale of probable isotopic exchangeis another issue). At some point, isotopic exchangeceases and the two lithologies define two pointson an isochron diagram with a slope that increasesover time. When friction melting occurs along thecontact at some time after the cessation of iso-


topic exchange, and in a simple case where thepseudotachylyte is formed from contributions byboth li thologies, then the pairing of thepseudotachylyte with either lithology, in an effortto obtain an isochron, yields an apparent age (aninherited isochron) that is older than the actualage of the pseudotachylyte.

Fluids and contacts/fault zonesLithologic contacts, fault zones, and shear

zones are good pathways for fluids, and extensivemetasomatism could occur prior topseudotachylyte generation, especially if thepseudotachylyte is preceded by cataclasite forma-tion (Magloughlin, 1992). Many examples of suchzones acting as fluid conduits have been described(Winchester and Max, 1984; Smalley et al., 1988;Dipple et al., 1990; Gates and Speer, 1991). Re-moval or addition of alkalis and Sr, potentiallyaided by recrystallization, neomineralization, andmineral dissolution, has the potential to changegreatly the Rb/Sr of the protolith relative to theunaltered rock. Indeed, Rb (or K) gain and Sr (orCa) loss appear to be likely at least in certainlithologies based on field and experimental stud-ies (Orville, 1963; Winchester and Max, 1984;Smalley et al., 1988; Gates and Speer, 1991;Hamamoto et al., 1994). Sr isotopic exchangebetween the fluid and the protolith can also leadto initial slopes of the pseudotachylyte-host rockisochron. For example, prior to melting, if aprotolith (such as a fault rock or a permeablelithologic layer) is initially in Sr isotopic equilib-rium with host rock having a different Rb/Sr, fluidflow along the fault zone exchanging Sr with theprotolith could change the 87Sr/86Sr of theprotolith. Upon pseudotachylyte formation, thepseudotachylyte would already have a higher orlower 87Sr/86Sr than the immediately adjacent hostrock, and thus an isochron pairing the two wouldhave an initial non-zero slope thus yielding a geo-logically meaningless apparent date.

Post-faulting fluidsFluids migrating along the fault zone after

pseudotachylyte generation could also affect Rb-

Sr dates, but this may be less likely than pre-gen-eration alteration. Pseudotachylyte-bearing faultzones do not seem to be commonly reactivated byextensive later faulting. This may be the result ofstrain hardening owing to the solidification of themelt along the main fault and within all connectedopen fractures. Consequently, within and close tothe generation zone (Grocott, 1981; Magloughlinand Spray, 1992), there may be relatively few con-nected fractures to serve as fluid pathways. Inaddition, alteration of the pseudotachylyte by ex-tensive fluid flow should be associated with cross-cutting veins containing low-temperature miner-als; alteration of minerals stable at high tempera-ture to minerals appropriate to lower temperature;and growth of new low-temperature minerals inthe pseudotachylyte groundmass (e.g., Thoeni,1988; Magloughlin, 1989). Where carbonates haveprecipitated, for example to form amygdules(Maddock et al., 1987; Thoeni, 1988; Kropf et al.,1993), significant shifts in Rb/Sr or 87Sr/86Sr orboth can be expected.

Volume diffusionOne notion put forth to defend the validity of

Rb-Sr pseudotachylyte-host rock dates is that Srcould, through volume diffusion, equilibrateisotopically between the pseudotachylyte and hostrock, presumably just after pseudotachylyte for-mation. However, volume diffusion is unlikely tocontribute to isotopic homogeneity owing to theminimal distance over which it is likely to oper-ate. Using literature values (Margaritz andHofmann, 1978) of D for K and Na (as proxiesfor Rb) and for Sr in obsidian and basalt at tem-peratures from 800–1300°C, and the simple equa-tion

x D tcm sec( ) = ∗ ( )[ ]

we find that even for unusually thick veins andthe improbable situation of a melt remaining athigh temperature for more than an hour or so, dif-fusion would only allow mobility of 20–300 µmfor Sr, and up to 1 mm for Na. Thus, even within


the melt itself, diffusion is very unlikely to pro-duce isotopic homogenization.

Models and discussionA generalized model for the behavior of the

Rb-Sr system during pseudotachylyte formationin biotite-rich rocks is shown in Fig. 3. This is forsituations where: 1) the initial rock is homogene-ous; 2) the Rb is concentrated in a single high Rb/Sr mineral, or several mineral species if their clo-sure temperatures are very similar; 3) there is a

low Rb/Sr, Sr-rich mineral resistant to assimila-tion by the melt. These criteria are likely to becommonly met owing to the presence of biotiteand plagioclase in many pseudotachylyte hostrocks. The first three (Figs. 3a–c) are consistentwith microstructural observations and differ in thetime of pseudotachylyte formation relative tobiotite closure in the Rb-Sr system. These are alsoclosed system models, consistent with the com-mon observation that pseudotachylytes containhigh temperature dendritic crystalli tes or

Fig. 3. Models for the formation of pseudotachylyte veins in biotite-bearing rocks for the Rb-Sr system. Threeclosed-system scenarios occur if the pseudotachylyte is generated at ambient temperatures exceeding the closuretemperature of biotite (a), at the closure temperature of biotite (b), and below the closure temperature of biotite(c). An open system model (d) occurs after biotite closure, and can produce a date younger than the biotite-wholerock date. Left side figures show the isochron situation at the time of pseudotachylyte formation; right side figuresshow the isochron situation at some later time. Abbreviations: wr = whole rock, bio = biotite, pst = pseudotachylyte.


microlites that would be easily alteredmineralogically and texturally as a result of opensystem behavior.

In (a), the pseudotachylyte forms while thetemperature of the country rock exceeds the biotiteclosure temperature, and because thepseudotachylyte behaves as a new whole rockclosed system, it begins to evolve on the isochrondiagram. The 87Sr/86Sr ratio of biotite simply in-creases along with the host rock with which it re-mains in isotopic equilibrium until the closuretemperature is reached, at which time the wholerock-biotite isochron begins to attain a positiveslope. This results in a date from the host rock-pseudotachylyte pair that exceeds that of the wholerock-biotite pair, and could result if thepseudotachylyte were relatively deep in origin,forming at temperatures above about 300–350°C.Such a situation might be demonstrated by evi-dence showing plastic deformation of quartz (andpossibly feldspar) accompanying pseudotachylyteformation, and possibly pseudotachylytes that aresubsequently plastically deformed (Passchier,1982). However, at very elevated temperature, thepossibility develops of ongoing isotopic exchangebetween the pseudotachylyte and host, and thusthe host rock-pseudotachylyte age would under-estimate the true age of formation.

In (b), pseudotachylyte generation occurs at thesame time as biotite closure, and thus the wholerock, biotite, and pseudotachylyte are collinear.This coincidence of pseudotachylyte formation atthe approximate time of biotite closure might berecorded texturally by both brittle and plasticmicrostructures intimately associated with thepseudotachylyte; but these textures are both strainrate and fluid dependent.

In (c), pseudotachylyte formation occurs afterbiotite closure, and again, the three are both ini-tially and remain collinear. Texturally, this situa-tion might be represented by evidence ofpseudotachylyte formation accompanyingcataclastic behavior of quartz, accompanied bylow temperature metamorphic retrogression.

In (a), the whole rock-pseudotachylyte pair-ing would produce the actual age of the

pseudotachylyte if the “whole rock” analyzed werechemically identical, with respect to the Rb-Srsystem, to that which formed the pseudotachylyte.The difficulty is that the “whole rock” is destroyedin the pseudotachylyte formation process, and acommon observation is that pseudotachylyte faultveins are sharply bounded and appear to haveformed from a very specific lithologic layer. Ingeochemically non-homogeneous rocks such asmost schistose metasedimentary rocks, the rockpresently adjacent to the pseudotachylyte cannotbe assumed to be representative of the whole rockfrom which the pseudotachylyte formed; in highlyhomogeneous rocks, this assumption might bevalid. The pseudotachylyte paired with lithic clastscontained within the pseudotachylyte would yielda valid date if the clasts are chemically representa-tive of the 87Sr/86Sr of the whole rock that formedthe pseudotachylyte.

In (b), the whole rock-pseudotachylyte pairproduces a true geologic age, but in (c), the ap-parent date is a maximum age of thepseudotachylyte formation. Unfortunately, (b) and(c) are chemically indistinguishable, and so if thewhole rock-pseudotachylyte and whole rock-biotite dates agree, only a maximum time of for-mation can be determined. None of these threescenarios explain cases where the whole rock-pseudotachylyte date is younger than the wholerock-biotite date. This relationship requires eitherlocal, closed system rehomogenization or inter-action with a fluid. In a local equilibrium model(Fig. 3d), time has elapsed since biotite closure.As a narrow shear zone or cataclasite zone devel-ops, the central part of the zone becomes enrichedin mica as quartz and feldspar are removed andthen precipitated, possibly in the adjacent, lessdeformed host rock. This results in a decrease inthe Rb/Sr value of the adjacent rock (fault rockhost, Fig. 3d), whereas the central part of the shearzone is a new mica-rich fault-rock that wouldlikely have a lower Rb/Sr value than biotite, ow-ing to incomplete removal of low Rb/Sr phases(e.g., apatite, epidote, or plagioclase). Aided byfluids, Sr isotopic homogenization occurs betweenthe fault rock and the adjacent host rock, result-


ing in a slight elevation of the 87Sr/86Sr of the host.When melting occurs, the melt may have the sameRb/Sr as the fault rock, or it may have a ratio be-tween the fault rock and the host rock if some ofthe latter contributes to the melt. If thepseudotachylyte forms soon after the fault rock,rather than by reactivation in a later event, the nowaltered “fault rock host”, pseudotachylyte, and anyvestiges of the fault rock will define an trueisochron yielding the age of the pseudotachylyte,an age younger than the whole rock-biotite age.Whether or not the fault rock equilibrates with Srin the infiltrating fluid is irrelevant to the age de-termination, as long as the fault rock and fault rockhost remain in Sr isotopic equilibrium. However,if the infiltrating fluid has a lower or higher 87Sr/86Sr than the fault rock and only equilibrates orpartly equilibrates with the fault rock, then ayounger or older, geologically meaningless datemay result. If the 87Sr/86Sr of the fluid is muchlower than the host rock and fault rock, negativeapparent dates are possible.

These models allow general predictions of therelationships among Rb/Sr and Sr isotopic valuesin host rocks and pseudotachylytes formed in vari-ous lithologies. The most common situation ap-pears to be mica-bearing rocks withpseudotachylyte formed in the upper part of thecrust without significant fluid involvement. In thissetting, Rb-Sr pseudotachylyte-host rock appar-ent dates should be similar to whole rock-micadates, or at the oldest, they will reflect whole rock-scale Sr isotopic closure (likely peak metamor-phism in medium to high grade rocks). In a mica-bearing rock lacking a modally significant low Rb/Sr mineral, no significant spread in Rb/Sr valuesbetween the pseudotachylyte and host rock islikely. For rocks lacking a modally important highRb/Sr mineral, such as may be the case ingranulites, ultrabasic rocks, many amphibolites,and certain monomineralic rocks such asanorthosite, no significant difference in Rb/Srbetween the host rock and pseudotachylyte islikely. Fluid involvement prior to or afterpseudotachylyte generation could change theseanticipated relations.

CONCLUSIONS

For the Nason Terrane samples, the wholerock-biotite and host rock-pseudotachylyte appar-ent dates agree within error. Thus, either scenarios(b) or (c) could apply. A situation equivalent toscenario (d) could also play a role; there is evi-dence for fluids accompanying pseudotachylyteformation (Magloughlin, 1992), and the host rocksin samples 1 and 2 have the among the lowest Rb/Sr ratios (and therefore the oldest model ages andlowest slopes on a Sr evolution diagram) of 25whole rock schist samples (Magloughlin, 1993,1995). Overall, the apparent age from sample 2must represent the maximum possible time ofpseudotachylyte formation, and the most probabletrue age for the pseudotachylyte (55 Ma,Magloughlin et al., 2001) suggests scenario (c) isthe most likely.

The Peterman and Day (1989) example couldbe a model (d) situation, because the whole rock-pseudotachylyte apparent date is drasticallyyounger than the whole rock-biotite apparent ageand Peterman and Day present evidence for fluidinvolvement. In addition, the host rocks hadanomalously old model ages compared to manyother samples from the same unit. Microstructuralstudy of the host rocks and the Rb-Sr analysis ofany unmelted fault rock host would test whethermodel (d) is appropriate.

Previous Rb-Sr studies of pseudotachylytehave typically assumed that Sr attains isotopicequilibrium with the host rocks as a result of themelting process or through some other mechanismimmediately after melting, but this attainment ofSr isotopic equilibrium between thepseudotachylyte and host rock does not seem verylikely. In contrast, this study has described sev-eral possible structural-chemical models forpseudotachylyte formation within lithologieswhere many pseudotachylytes are found. Sce-narios where a host rock-pseudotachylyte age rep-resents the actual time of pseudotachylyte forma-tion do exist theoretically, either through coinci-dence of fault slip at the time of isotopic closurein the dominant Rb-bearing mineral, or through


the pervasive influence of fluids prior topseudotachylyte formation. The latter scenariomight be possible to demonstrate. Otherwise, theresults of Rb-Sr host rock-pseudotachylyte datingought to be considered tentative at best.

Acknowledgments—I acknowledge the late R. L.Armstrong for help obtaining the Rb-Sr analyses, R.H. Maddock for sharing the Rb-Sr information on theOuter Hebrides Fault Zone from his Ph.D., Ben vander Pluijm and Chris Hall for reviewing an earlier draftof this paper, and an anonymous reviewer. This workwas supported by a United States Geological Surveygrant through the National Earthquake Hazards Reduc-tion Program, #1434-HQ-97-GR-03007, by NationalScience Foundation Grant EAR-9805138, and by aColorado State University Faculty Research Grant.

REFERENCES

Adams, M. G. and Su, Q. (1996) The nature and timingof deformation in the Beech Mountain Thrust Sheetbetween the Grandfather Mountain and MountainCity Windows in the Blue Ridge of NorthwesternNorth Carolina. Journal of Geology 104, 197–213.

Boeckeler, A. J. (1994) Isotopic ages of pseudotachylyteveins from coastal New Hampshire and SW Maine:evidence for post-Acadian strike-slip motion. Geo-logical Society of America Abstracts with Programs26, 7.

Cowan, D. S. (1999) Do faults preserve a record of seis-mic slip? A field geologist’s opinion. Journal ofStructural Geology 21, 995–1002.

Dipple, G. M., Wintsch, R. P. and Andrews, M. S.(1990) Identification of the scales of differential el-ement mobility in a ductile shear zone. Journal ofMetamorphic Geology 8, 645–661.

Field, D. and Råheim, A. (1979) A geologically mean-ingless Rb-Sr total rock isochron. Nature 282, 497–499.

Gates, A. E. and Speer, J. A. (1991) Allochemical ret-rograde metamorphism in shear zones: an examplein metapelites, Virginia, USA. Journal of Metamor-phic Geology 9, 581–604.

Grocott , J . (1981) Fracture geometry ofpseudotachylyte generation zones: a study of shearfractures formed during seismic events. Journal ofStructural Geology 3, 169–179.

Hamamoto, R., Ohta, T. and Yanagi, T. (1994) Sr andRb extraction by low-temperature hydrothermal so-lution from a granite gneiss sample and its implica-tion for Rb-Sr geochronology. Science Reports ofKyushu University 18, 97–102.

Hofmann, A. W. (1979) Rb-Sr dating of thin slabs: animperfect method to determine the age of metamor-phism. Lectures in Isotope Geology (Jäger, E. andHunziker, J. C., eds.), 27–29, Springer-Verlag, NewYork.

Kamineni, D. C., Stone, D. and Peterman, Z. E. (1990)Early Proterozoic deformation in the western Supe-rior province, Canadian Shield. Geological Societyof America Bulletin 102, 1623–1634.

Kelley, S. P., Reddy, S. M. and Maddock, R. (1994)Laser-probe 40Ar/39Ar investigation of apseudotachylyte and its host rock from the Outer Islesthrust, Scotland. Geology 22, 443–446.

Kropf, E. P., Craddock, J. P. and Wirth, K. R. (1993)Calcite-bearing pseudotachylyte in Archean Granu-lite grade rocks of the Minnesota River ValleyTerrane. EOS 74, 577.

Maddock, R. H. (1986) The petrology and geologicalsignificance of some fault-generatedpseudotachylytes. Ph.D. Dissertation, University ofLondon.

Maddock, R. H., Grocott, J. and Van Ness, M. (1987)Vesicles, amygdules and similar structures in fault-generated pseudotachylytes. Lithos 20, 419–432.

Magloughlin, J. F. (1989) Nature and significance ofpseudotachylyte from the Nason Terrane, North Cas-cade Mountains, Washington. Journal of StructuralGeology 11, 907–917.

Magloughlin, J. F. (1992) Microstructural and chemi-cal changes associated with cataclasis and frictionmelting at shallow crustal levels: the cataclasite-pseudotachylyte connection. Tectonophysics 204,243–260.

Magloughlin, J. F. (1993) A Nason Terrane Trilogy: I.Nature and significance of pseudotachylyte, II. Sum-mary of the structural and tectonic history, III. Ma-jor and trace element geochemistry, and strontiumand neodymium isotope geochemistry of theChiwaukum Schist, amphibolite, and meta-tonalitegneiss. Ph.D. Dissertation, University of Minnesota,308 pp.

Magloughlin, J . F. (1995) Rb-Sr dating ofmetasedimentary, metavolcanic, and metaplutonicigneous rocks from the Nason and Mad Riverterranes, North Cascade Mountains, Washington.Isochron/West 62, 6–15.

Magloughlin, J. F. (1998) I. Flow features andcataclasite-pseudotachylyte relations; II. Melting re-lations within clasts in pseudotachylyte; III .Amygdules within pseudotachylyte. IV. Amygdulesand microbreccia collapse structures in apseudotachylyte. Fault-Related Rocks—A Photo-graphic Atlas (Snoke, A., Tullis, J. and Todd, V.,eds.), Princeton University Press, 617 pp.


Magloughlin, J. F. and Spray, J. G. (1992) Frictionalmelting processes and products in geological mate-rials: introduction and discussion. Tectonophysics204, 197–204.

Magloughlin, J. F., Hall, C. M. and van der Pluijm, B.A. (2001) Dating of ancient earthquakes through40Ar-39Ar geochronometry of pseudotachylyte. Ge-ology 29, 51–54.

Margaritz, M. and Hofmann, A. W. (1978) Diffusionof Sr, Ba and Na in obsidian. Geochimica etCosmochimica Acta 42, 595–606.

O’Hara, K. (1992) Major- and trace-element constraintson the petrogenesis of a fault-relatedpseudotachylyte, western Blue Ridge province, NorthCarolina. Tectonophysics 204, 279–288.

Orville, P. M. (1963) Alkali ion exchange betweenvapor and feldspar phases. American Journal of Sci-ence 261, 201–237.

Passchier, C. W. (1982) Pseudotachylyte and the de-velopment of ultramylonite bands in the Saint-Barthelemy Massif, French Pyrenees. Journal ofStructural Geology 4, 69–79.

Peterman, Z. E. and Day, W. C. (1989) EarlyProterozoic activity on Archean faults in the westernSuperior Province: evidence from pseudotachylyte.Geology 17, 1089–1092.

Peucat, J.-J. and Martin, H. (1985) Ar Rb-Sr thin slabsmigmatite ages always meaningful? Neues JahrbuchMiner. Abh. 152, 1–21.

Smalley, P. C., Field, D. and Råheim, A. (1988) Rb-Srsystematics of a Gardar-age layered alkalinemonazite suite in southern Norway. Journal of Ge-

ology 96, 17–29.Spray, J. G. (1992) A physical basis for the frictional

melting of some rock-forming minerals.Tectonophysics 204, 205–221.

Spray, J. G. (1993) Viscosity Determinations of somefrictionally generated silicate melts: Implications forfault zone rheology at high strain rates. Journal ofGeophysical Research 98, 8053–8068.

Tabor, R. W., Zartman, R. E. and Frizzel, V. A., Jr.(1987) Possible tectonostratigraphic terranes in theNorth Cascades Crystalline Core, Washington. Se-lected Papers on the Geology of Washington(Schuster, J. E., ed.), Washington Division of Geol.and Earth Res. Bull. 77, 107–127.

Thoeni, M. (1988) Rb-Sr isotopic resetting in mylonitesand pseudotachylytes: implications for the detach-ment and thrusting of the Austroalpine basementnappes in the eastern Alps. Jahrbuch der GeologischeBundesanstalt (Austria) 131, 169–201.

Thoeni, M. and Hoinkes, G. (1987) The southernOetztal basement; geochronological and petrologi-cal consequences of Eoalpine metamorphic overprint-ing. Geodynamics of the Eastern Alps (Fluegel, H.W. and Faupl, P., eds.), 200–213, Franz Deuticke,Vienna, Austria.

Wendt, I. (1993) Isochron or mixing line? ChemicalGeology 104, 301–305.

Winchester, J. A. and Max, M. D. (1984) Element mo-bility associated with syn-metamorphic shear zonesnear Scotchport, NW Mayo, Ireland. Journal of Meta-morphic Geology 2, 1–11.

an evaluation of rb-sr dating of pseudotachylyte...

Documents