Journal of Petrology | Volume 40 | Number 5 | Pages 773-786 | 1999
© Oxford University Press 1999
Evolution of Silicic Magma through Assimilation and Subsequent Recharge: Evidence from Sr Isotopes in Sanidine Phenocrysts, Taylor Creek Rhyolite, NM
1 Department of Earth and Space Sciences UCLA, Los Angeles, CA 90095, USA
2 Us Geological Survey, 2255 North Gemini Drive Flagstaff, AZ 86001, USA, and Geology Department, Northern Arizona University Flagstaff, AZ 86011, USA
Received April 28, 1998; Revised typescript accepted November 11, 1998
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Isotopic fingerprinting of individual mineral phases, complemented by crystal size data, provides a unique avenue for elucidating the details of evolutionary histories of crustal magma systems. Here we report the first measurements of Sr isotopic compositions of single crystals as a function of size and Sr isotopic profiles constructed through microdrill sampling of sanidine crystals from a high-silica rhyolite lava from the Taylor Creek Rhyolite, NM. Whole-rock 87Sr/86Sr increases monotonically with modal abundance of sanidine phenocrysts, suggesting Taylor Creek magma evolved through a coupled process of assimilation and crystallization. In contrast, sanidine phenocrysts do not show simple monotonic increases in 87Sr/86Sr as a function of crystal size and core-to-rim stratigraphy. Instead, 87Sr/86Sr ratios and Sr concentrations of individual sanidines increase with crystal size to a maximum at
4 mm and then decrease with further increase in size. Microsampling of two crystals greater than 4 mm in length showed core-to-rim increase then decrease in 87Sr/86Sr, whereas a single sanidine crystal less than 4 mm in length displayed a simple core-to-rim decrease in 87Sr/86Sr. Furthermore, in contrast to measured size distributions of crystals in volcanic rocks, which commonly decrease exponentially with increasing size, crystal size frequency histograms are bell shaped, with decreasing numbers of crystals in the smallest size class. All these results are consistent with a model involving continuous phenocryst nucleation and growth in a crustally contaminated magma into which a lower-87Sr/86Sr, lower-Sr magma was injected. In such a scenario, it is argued that curved crystal size distributions mirror variations in nucleation rate in response to changes in undercooling as the magma body evolved from assimilation- to recharge-dominated regimes. KEY WORDS: assimilation; crystallization; isotopic microsampling; recharge; rhyolite
|
Introduction |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Radiogenic isotopic studies at many silicic volcanic centers indicate that the generation and differentiation of silicic magma typically involves open-system processes (e.g. Noble & Hedge, 1969
; Johnson & Fridrich, 1990; Tegtmeyer & Farmer, 1990; Grove et al., 1997). However, little agreement exists concerning the relative roles of assimilation, magma mixing and recharge in the petrogenesis of rhyolites. These uncertainties in part reflect the traditional study of the isotopic compositions of whole-rock and bulk-mineral samples, which only provide information concerning the state of the magma upon eruption and the sum of the compositional changes occurring throughout crystal growth. An alternative yet complementary approach is the determination of the isotopic compositions of individual mineral grains and isotopic variations within grains [for reviews of techniques and applications, see Neal et al. (1995) and Davidson et al. (1998)].
In situ microdrill sampling coupled with analysis using thermal ionization mass spectrometry shows great potential as a technique for measurement of Sr isotopes at the sub-millimeter scale in volcanic rocks and minerals. In a pilot study Davidson et al. (1990) demonstrated that microdrill sampling across the interface between a basaltic andesite inclusion and host dacite from the Purico–Chascon complex in northern Chile provided a means of investigating the time and length scales of mafic–silicic magma interactions. Subsequent microsampling studies by Feldstein et al. (1994) on rhyolites from San Vincenzo, Italy, and by Davidson & Tepley (1997) on arc volcanic rocks from Chile, Mexico, and California elegantly demonstrate that information regarding secular variation in magma isotopic compositions can be extracted from individual growth zones of single crystals—information that is generally unobtainable in studies utilizing bulk samples.
This paper reports results of a detailed Sr isotopic study of individual sanidine crystals from a single lava flow from the Taylor Creek Rhyolite, a mid-Tertiary lava dome field in New Mexico. Earlier work on whole-rock and bulk-mineral samples (Reece et al., 1990; Duffield & Ruiz, 1992) indicates that the rhyolite incorporated minor yet easily identifiable Precambrian country rock before eruption and that sanidine phenocrysts and host lavas show marked isotopic disequilibrium. We contend that the isotopic characteristics of individual sanidine crystals of different sizes and hence different residence times and isotopic variations within single crystals revealed by microdrill sampling further constrain the isotopic evolution of magma erupted to form the Taylor Creek Rhyolite. When complemented by study of grain size variations, the crystal isotope data provide fresh insights into the timing and dynamics of open-system differentiation processes involved in the genesis of silicic magmas.
|
Geologic Background |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
The Taylor Creek Rhyolite is composed of a group of late Oligocene rhyolite lava domes and flows that crop out in the northern Black Range of the Mogollon–Datil volcanic field in southwestern New Mexico (Fig. 1). The Taylor Creek Rhyolite has been the subject of numerous studies that delineation the eruptive chronology and physical volcanology (Duffield & Dalrymple, 1990
; Duffield et al., 1995), mineral chemistry (Duffield & du Bray, 1990; Wittke et al., 1996), volatile contents (Webster & Duffield, 1991), and major element, trace element and isotopic compositions of erupted lavas (Reece et al., 1990; Duffield & Ruiz, 1992). The lava field was emplaced during at least 20 eruptions, each beginning with a pyroclastic phase and ending with the effusion of magma to form a lava flow or dome. Individual lava volumes range from 0.15 km3 to 10.55 km3, totaling 55.83 km3 for the 20 map units. Nearly all pyroclastic deposits are poorly consolidated to unconsolidated and thus are highly eroded. High-precision 40Ar/39Ar ages for sanidine phenocrysts collected from each of the 20 lavas indicate that all eruptions occurred within <100 000 yr at 27.9 Ma (Duffield & Dalrymple, 1990).
|
The dome field consists of moderately to highly porphyritic, fluorine-rich, high-silica rhyolite lavas generally containing 10–35% phenocrysts of quartz and sanidine in subequal proportions with minor plagioclase (
1.5%) and trace hornblende and biotite. Electron microprobe studies of mineral chemistries (Duffield & du Bray, 1990; Wittke et al., 1996) and ion probe study of glass inclusion in quartz phenocrysts (Webster & Duffield, 1991) suggest phenocryst growth at temperatures between 775 and 805°C (two-feldspar and Fe–Ti-oxide geothermometry, respectively) in rhyolitic melt with a pre-eruptive H2O content of 2.7 wt % at 2 kbar. Comparison with experimental results in the granite system (e.g. Tuttle & Bowen, 1958; Whitney, 1988) indicates Taylor Creek Rhyolite magma was H2O undersaturated during growth of quartz and feldspar phenocrysts. Application of the temperature–XH2O phase diagram for the Cape Ann Granite (77.6 wt % SiO2) at 2 kbar confining pressure from Whitney (1988, fig. 4) indicates crystallinities between 20 and 35 vol. % over this temperature range at 2.7 wt % H2O, which within the uncertainty of the percent-melt contours (±10%) is compatible with that observed for the Taylor Creek Rhyolite.
|
) is equal to or less than symbol size except where denoted by error bars. Bulk isotopic compositions for microdrilled crystals (
) calculated assuming simple mixing among growth zones, which were defined by the number of drill steps (e.g. the 7 mm crystal was divided into nine zones). The contribution of each zone was weighted according to its volume.The rhyolite is metaluminous to weakly peraluminous with nearly constant major element compositions (Table 1). Feldspar phenocrysts show similar homogeneity in major element composition; for example, maximum variation in the albite molecular component for sanidine is 4.7 mol %, with most varying
1 mol % (Duffield & du Bray, 1990). Despite this compositional uniformity, the Taylor Creek Rhyolite is noteworthy for an extreme range of initial Sr isotopic compositions (ISr = 0.705–0.713 at 27.9 Ma) among individual lavas (Reece et al., 1990). Furthermore, lava ISr ratios correlate positively with Sr contents and negatively with Rb, Ta, and Th contents (Reece et al., 1990). The correlations with Ta and Th are particularly important because although Rb and Sr may be mobile during devitrification and post-emplacement alteration, both Ta and Th are generally considered immobile. Thus these correlations eliminate the possibility that the Sr isotopic variations among the lavas are solely the result of secondary processes. Likewise, a negative correlation between ISr and 87Rb/86Sr precludes the 87Sr/86Sr variations occurring by in situ decay of 87Rb to 87Sr in the magma before eruption.
|
The most straightforward explanation is that the Sr isotopic variations reflect variable late-stage assimilation of radiogenic and relatively Sr-rich upper-crustal wall and roof rocks by highly differentiated low-87Sr/86Sr, low-Sr rhyolitic magmas (Reece et al., 1990
). The low Sr contents of the rhyolite, as low as 2.5 ppm, require as little as 1% of local Proterozoic granitic basement to produce the entire range of lava Sr isotopic compositions through bulk mixing (Duffield & Ruiz, 1992).
Bearing on the nature of the contamination process is a positive correlation of whole-rock ISr ratios with sanidine contents, which indicates a coupling between crystallinity and degree of contamination (Fig. 2). On this basis and because bulk-sanidine separates are systematically more radiogenic than their host lavas (Fig. 2), Reece et al. (1990) suggested that crystal growth most probably occurred near the walls and roof of the magma chamber(s), where sanidine phenocrysts could easily incorporate radiogenic Sr derived from Precambrian wall rocks. Duffield & Ruiz (1992) further speculated that Sr isotopic disequilibrium between sanidines and host lavas reflected changes in the isotopic composition of the contaminant with time. They suggested that radiogenic Sr derived from initial contaminant melts produced by dehydration melting of biotite-bearing granitic roof and/or wall rock was incorporated in the early growth of sanidine phenocrysts. But as partial melting continued, contaminant melt with a greater budget of less radiogenic Sr derived from country-rock feldspars was assimilated and quenched into the rhyolite by eruption, before isotopic equilibration could be attained between sanidine and surrounding magma.
|
Independent evidence of Sr isotopic disequilibrium during wall-rock anatexis, such as that proposed by Duffield & Ruiz (1992)
, comes from field and experimental studies of xenolith melting and in situ melting along the contacts of shallow basic intrusions (e.g. Pushkar & Stoeser, 1975; Knesel & Davidson, 1996; Tommasini & Davies, 1997). Of particular relevance is the observation that in some cases melt Sr isotopic compositions evolve from high to low 87Sr/86Sr as anatexis of biotite-bearing granitic rock progresses (Tommasini & Davies, 1997; Knesel & Davidson, 1999).
As a test of their assimilation model, Duffield & Ruiz (1994) proposed that, in the absence of significant homogenization by diffusion, zoning within sanidine crystals should record a time-series change in magma 87Sr/86Sr caused by secular variation in contaminant melt compositions. Crystals that nucleated early in the contamination process should display a profile of first increasing then decreasing 87Sr/86Sr towards the rim, whereas crystals that nucleated and grew after the contaminant was dominated by wall-rock feldspar-derived Sr should display a simple core-to-rim decrease in 87Sr/86Sr. Microdrill sampling provides a unique tool for testing this model.
|
Methods |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Sample selection
Because of the time-intensive nature of isotopic study of individual crystals, we decided to concentrate our initial efforts on a single eruptive unit. The lava flow of White Water Canyon (map unit WHC), located within a cluster of lavas in the southwestern sector of the volcanic field (Fig. 1), was selected because of its high ISr (0.7122–0.7131) and high phenocryst content (31–36 vol. %).
The sample studied here was collected at site WHC/2 of Duffield & Ruiz (1992). The modal analysis by Duffield & Dalrymple (1990) showed that the rock consists of 17.7% quartz, 13.7% sanidine, 1.6% plagioclase, and 1.2% mafic minerals. The groundmass consists of devitrified intergrowths of silica and alkali feldspar. Quartz and sanidine phenocrysts are euhedral to subhedral and commonly display resorption embayments in thin section. Plagioclase grains are optically unzoned and commonly display lamellar twinning; some are partially overgrown by sanidine. Most of the mafic minerals are altered beyond identification. However, at least one grain of biotite can be identified per thin section.
Electron microprobe studies of sanidine and plagioclase from the lava of White Water Canyon by Duffield & du Bray (1990) demonstrate that the feldspars are compositionally homogeneous with regard to major and minor elements. Core–rim compositions vary by less than 5 mol % for Ab, Or and An molecular components (Duffield & du Bray, 1990, table 2).
|
Crystal isotope studies
Two approaches are taken here to recover the information concerning the origin of Sr isotopic diversity in the Taylor Creek Rhyolite as recorded by individual phenocrysts: (1) Sr isotopic characterization of individual sanidines as a function of crystal size; (2) core-to-rim Sr isotope profiles of individual sanidine crystals achieved through microdrill sampling. Single crystals covering most of the size range of sanidine phenocrysts in unit WHC were obtained by gently crushing
1 kg of rock. Individual whole crystals were then picked and inspected for freshness under a binocular microscope and cleaned in ethanol in an ultrasonic shaker. The long dimensions of individual crystals (i.e. along the c-axis) were measured to the nearest 0.1 mm with a binocular microscope. Strontium isotopic profiles of sanidine phenocrysts were constructed through incremental-depth drilling using a high-r.p.m. drill press and diamond-coated drill bits. Sanidine crystals were embedded in epoxy, mounted on a mechanical x–y stage, and positioned for drilling with the aid of a binocular microscope. The drill bit was lubricated with a drop of deionized water, which also served to suspend the powdered sample by creating a slurry. Starting at the surface, a single pit was excavated by sequential drilling; individual steps ranged between 0.3 and 0.1 mm in depth. Upon completion of each drill step the slurry was collected by pipette for Sr isotopic analysis (i.e. each depth step represents a single datum). Between each depth step, the crystal was cleaned by agitation in deionized water in an ultrasonic shaker followed by rinsing in a stream of water. The crystal was then dried and visually inspected for drilling residue with a binocular microscope. If residual powder was identified, the cleaning procedure was repeated as necessary. Sample slurries were dried to powders in aluminum foil cups on a hot plate and weighed to within 0.001 mg on a Metler balance.
Analytical procedures
Sr isotopic analysis for whole-rock, bulk-mineral separates, and large single-sanidine crystals (
5 mg) followed standard analytical procedures. Because of the low Sr contents of sanidines (10–50 ppm), chemical separations for microdrill samples (0.2–0.7 mg) and single sanidine crystals weighing <5 mg were performed using a modified cation exchange technique. Small-volume (250 µl), high-aspect-ratio (height/width 35) Pyrex columns were used to maximize chemical separation of Rb from Sr in high-Rb/Sr samples. Resin and acid volumes were reduced to minimize sample contamination. Total process blanks for the microsample chemistry were 30 pg and 6 pg for Sr and Rb, respectively. Blank corrections were negligible, as contributions were <1.0% of sample budget of Sr and Rb in all cases.
Strontium isotopic ratios of all samples were measured using a dynamic peak switching routine on a multicollector VG Sector magnetic mass spectrometer at UCLA. 87Sr/86Sr ratios for whole-rock and bulk-mineral separates and large single sanidine crystals were measured at 3 V (88Sr). Microdrill samples and small single sanidine crystals (5–50 ng total Sr) were run between 1 and 3 V (88Sr). All analyses were normalized to 86Sr/88Sr = 0.1194. Repeated analyses of a 6 ng NBS 987 standard solution used during microsample analyses yielded a value of 0.710225 ± 16 (2
SE, n = 7), whereas a 600 ng solution (run with whole-rock, bulk-mineral separates, and large single sanidine crystals) during the study period gave a value of 0.710223 ± 15 (n = 9). Rb isotopic compositions were measured at 500 mV (87Rb) with a static routine. Because of the small sample sizes, Rb and Sr concentrations were determined by isotope dilution on the same sample aliquots measured for isotope composition using a mixed 87Rb (>99%)–84Sr (>99%) spike and are accurate to 1% and 0.5%, respectively. Corrections for spike contributions to measured 87Sr/86Sr ratios were less than 5 x 10–5 in all cases and are generally within the 2 standard error for measured ratios.
|
Strontium Isotope Data |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Whole-rock and bulk-mineral isotopic compositions
Sr isotopic compositions of whole-rock and bulk-mineral separates for the lava flow of White Water Canyon are given in Table 2. Initial Sr isotopic ratios were calculated using an eruption age for the Taylor Creek Rhyolite of 27.92 Ma (Duffield & Dalrymple, 1990
). The uncertainties in the reported initial ratios were calculated by propagation of individual uncertainties of the eruption age (±4 x 104yr), 87Rb/86Sr ratio (±1%), and 2 standard error for the measured 87Sr/86Sr using a squared-sum partial differential expression (Bevington, 1969) of the decay equation.
Regression of the whole-rock and bulk-mineral data yields an ‘isochron’ with an apparent age of 31.0 ± 0.4 Ma (Fig. 3). Comparison of this apparent isochron with the eruption age suggests a residence time for the rhyolite magma of 3 x 106 yr! However, when adjusted to account for radiogenic ingrowth of 87Sr from 87Rb since 27.92 Ma, biotite has an ISr of 0.7218 ± 13, which is significantly greater than that of both sanidine at 0.7143 ± 2 and the whole-rock value of 0.7131 ± 5. The linear array on the Sr isochron diagram appears to reflect mixing of isotopically distinct components rather than dating a discrete differentiation or magma production event. Furthermore, the variation of ISr among minerals and whole rock indicates that individual mineral phases had not equilibrated isotopically with their host liquid at the time the magma erupted.
|
Isotopic variations as a function of crystal size
The dependence of Sr isotopic composition on crystal size was determined by measuring the 87Sr/86Sr of individual sanidine crystals (Table 3). The paucity of data for large (>3 mm) crystal sizes and the absence for small (<1.5 mm) crystal sizes are largely the results of decreasing crystal abundance about a maximum at intermediate sizes (see Fig. 10, below). The pool of large crystals was further depleted because of preferential selection of them for microdrill sampling.
|
|
[see also Cashman & Ferry (1988)Not withstanding these sampling biases, two domains are recognized as a function of crystal size (Fig. 4): sanidine initial 87Sr/86Sr ratios increase with increasing crystal size up to a maximum of 0.7166 ± 1, after which ISr appears to decrease with increasing size. Because of the decrease in crystal abundance and our inability to resolve whole crystals from crystal fragments below
1.5 mm, the variation of ISr as crystal size approached zero (i.e. nucleus-size crystals) was not measured directly. Linear regression of the data for crystals of 3 mm yields an intercept of 0.7089 (R2 = 0.92), which is compatible with rim compositions for two of the three crystals reported below. Calculated bulk isotopic compositions of the three sanidine crystals for which Sr isotope profiles were constructed are also shown in Fig. 4. Although the bulk isotopic compositions for the two large crystals are consistent with the single crystal data, the composition of the 3.8 mm crystal is not. Interestingly, the shape and absolute values for the profile for the 3.8 mm crystal are in excellent agreement with the profile for the 7 mm sanidine crystal.
Finally, we note that the Sr content of individual sanidines varies with crystal size in a fashion similar to Sr isotopic composition, albeit with a larger degree of scatter (Fig. 5). The important point is that a correlation between ISr and Sr content (Fig. 6) indicates that the process(s) affecting the Sr isotopic compositions also affected the Sr contents of the sanidine crystals.
|
|
Crystal isotope profiles
Results of microdrill sampling of individual sanidine crystals are given in Table 4. The form of ISr depth profiles varies as a function of crystal size (Fig. 7). In the largest crystal (7 mm), and therefore that with the potential for the greatest temporal record of changing magma compositions, ISr increases from the core (0.7054 ± 6) outward, reaching a maximum of 0.7144 ± 3 between the core and rim, then decreases towards the rim (0.7089 ± 1). In the 5.5 mm crystal, the ISr peak is closer to the core. It should be noted, however, that although the shape of the profile is similar to that of the 7 mm crystal, the profile is significantly displaced toward higher ISr values. We discuss this observation after evaluating potential origins for the Sr isotopic variations. Finally, as noted above, the 3.8 mm crystal displays a core-to-rim decrease in ISr that is remarkably similar to the outer portion of the 7 mm crystal. No systematic correlation between Sr isotopic composition and concentration was revealed by microdrilling (Table 4).
|
|
|
Discussion |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Possible causes of sanidine isotopic variations
The marked Sr isotopic disequilibrium between concentrates of sanidine and biotite and host lava indicates that the Taylor Creek magma evolved through interaction of isotopically diverse components. However, the bulk isotopic data provide limited insights into the identity of the endmembers and processes involved. The Sr isotopic variations with crystal size and within progressive growth zones of individual sanidine crystals provide an additional mechanism for investigating the open-system history of the Taylor Creek magma and a partial test of the contamination model proposed by Duffield & Ruiz
(1992, 1994). It is prudent, however, to first evaluate the potential influence of diffusive exchange between crystals and liquid of contrasting isotopic compositions.
Influence of diffusive equilibration on crystal isotope variations
The extent to which the sanidine isotope variations may be attributed to diffusive equilibration with a liquid of lower 87Sr/86Sr can be assessed using experimental constraints on Sr diffusion in alkali feldspars and analytical solutions for diffusion given by Crank (1975). Diffusive interaction between two isotopically distinct reservoirs may occur in the presence or absence of a bulk chemical gradient (chemical vs self or tracer diffusion). The observed correlation between Sr isotopic compositions and concentrations noted for the individual sanidines indicates, however, that if diffusion modified sanidine compositions it affected both Sr isotopic ratios and concentrations. Accordingly, we model diffusive equilibration of sanidine crystals using Sr chemical diffusivities given by Cherniak (1996).
Figure 8 shows the percent equilibration as a function of time for Sr chemical diffusion in sanidine crystals ranging in size from 1 mm to 7 mm. Over typical solidification times of crustal magma bodies (104–106 yr), a 1 mm sanidine could attain between 10 and 80% equilibration with its host liquid, whereas a 7 mm sanidine would achieve between 0 and 20% equilibration. Although these calculations provide an estimate of the time scales for crystal-melt equilibration, lack of constraint on residence times for the Taylor Creek magma system prevents critical evaluation of the extent of equilibration of individual crystals.
|
To evaluate further the potential influence of diffusive equilibration on sanidine compositions, we calculated diffusion profiles for the three microdrilled crystals. Assuming linear diffusion perpendicular to the crystal–melt interface (parallel to the depth profiles), between 106 and 107 yr are required to form the measured 87Sr/86Sr profiles by diffusion (Fig. 9). Clearly, these time scales are unrealistic even for thermally well-connected systems and therefore preclude a dominant role for diffusive equilibration in the production of the crystal isotope profiles.
|
Role of assimilation of incongruent country-rock melts
The observed Sr isotopic variations as a function of crystal size and core-to-rim stratigraphy are for the most part compatible with progressive assimilation of dehydration wall-rock melts; ISr first increases then decreases with distance from core to rim and crystal size. The variation of Sr content is problematic, however.
Assimilation of low-87Sr/86Sr country-rock melts with a large budget of feldspar-derived Sr should produce a marked increase in the Sr content of low-Sr rhyolite magma, but is not observed in either the crystal size data or core-to-rim profiles. After an initial increase, Sr contents of individual sanidine crystals in fact decrease with decreasing crystal size (Fig. 5). Thus a correlation of decreasing ISr with decreasing Sr (Fig. 6) argues against progression from biotite-dominated to feldspar-dominated contaminant melts.
Nevertheless, some dehydration melting of chamber wall rocks, which were presumably similar in composition to biotite-bearing Precambrian granites that crop out in southern New Mexico (e.g. Condie, 1978; White, 1978), seems likely, on the basis of the estimated magmatic temperature of 800°C. However, this temperature is certainly not sufficient to produce moderate to high-degree partial melts of granitic protoliths at shallow-crustal depths. We therefore conclude that, although the crystal isotope data indicate crystal growth in a melt of changing isotopic composition as a result in part of incorporation of crustal material, a secular change in the isotopic character of the contaminant because of progressive melting of wall rocks is not indicated.
Origin of Sr isotopic variations by assimilation followed by recharge
The Sr isotopic variations across sanidine crystals and with crystal size may alternatively reflect phenocryst growth in a crustally contaminated rhyolite magma to which a lower-87Sr/86Sr, lower-Sr rhyolite magma was added. This is arguably the most straightforward explanation, as back-mixing of a differentiated rhyolite with a rhyolite whose 87Sr/86Sr and Sr content have been raised by assimilation of wall-rock material would lead to a decrease in both the 87Sr/86Sr and Sr content of the resulting mixed magma.
Davidson & Tepley (1997) reported similar nonmonotonic variations in 87Sr/86Sr across plagioclase crystals from convergent margin volcanic rocks that appear to correlate with textural features (e.g. dissolution surfaces) and compositional data (e.g. An content). They interpreted these correlated features to reflect phenocryst growth in magma systems undergoing both crustal assimilation and repeated recharge events involving more mafic liquids. It is important to note that no such textural or compositional unconformities are expected in the Taylor Creek Rhyolite sanidines as it is likely that the contaminant (partial melts of Precambrian granitic basement) and recharge liquids differed only in their trace element and isotopic compositions.
Although assimilation followed by recharge is consistent with the systematic variation of 87Sr/86Sr and Sr concentration with crystal size and sanidine ISr profiles, the origin of the low bulk ISr for the 3.8 mm sanidine and the shift of the ISr profile towards higher values for the 5.5 mm sanidine relative to the other crystals is problematic. One interpretation is that these crystals are xenocrystic. However, this possibility seems unlikely as electron microprobe studies of randomly selected feldspar grains from all 20 eruptive units suggest that only 1% of the feldspar population may be xenocrystic (Duffield & du Bray, 1990). Nevertheless, it is conceivable that new growth on a xenocrystic core from an earlier magmatic phase could account for the low-87Sr/86Sr and low-Sr core for the 7 mm sanidine. A more compelling possibility is that the crystals were grown at different levels in a vertically (and potentially laterally) zoned magma reservoir and that compositional zones were partially mixed immediately before or during eruption (e.g. Spera, 1984). In such a scenario crystals grown closer to the roof and walls of the chamber display profiles displaced towards higher 87Sr/86Sr than crystals grown deeper in the interior of the magma body.
Origin of radiogenic biotite
As noted above, the high ISr of a biotite concentrate relative to the whole-rock ISr indicates that the biotite was out of equilibrium with surrounding melt at the time of eruption. A simple explanation for this disequilibrium, which is consistent with petrographic and compositional data, is that the biotite is xenocrystic. Arguments favoring a xenocrystic origin are the trace quantities, small size (0.5–1 mm width) and partially to completely altered nature of biotite grains, which are present only in lavas judged to be most contaminated on the basis of whole-rock trace element and Sr isotope compositions. In addition, electron microprobe data collected by Wittke et al. (1996) for the lava of Kemp Mesa (map unit KPM in Fig. 1) indicate that the biotite is too magnesian to be in equilibrium with the high-silica rhyolitic magma. They suggested the biotite grains were derived from Precambrian basement within which the magma(s) parental to the Taylor Creek Rhyolite was emplaced. If so, diffusion considerations suggest that biotite preserved in the lava of White Water Canyon may have been introduced into the magma immediately before or during eruption, as grains up to 0.5 mm thick are expected to equilibrate isotopically within a hundred years after immersion in rhyolite liquid at 800°C (Fig. 8).
Relationships among crystal nucleation and growth, cooling history and open-system processes
The size distributions of phenocrysts in volcanic rocks have great potential to provide insights into the cooling histories of natural magmas. Grain size distributions for the lava of White Water Canyon are bell shaped, with maximum numbers of crystals at intermediate grain sizes (Fig. 10a). Interestingly, the decrease in number of crystals in the smallest size class differs from typical size distributions in volcanic rocks, which show greater numbers of crystals skewed towards increasingly smaller sizes (e.g. Cashman, 1988; Cashman & Marsh, 1988; Armienti et al., 1994). The resulting curved crystal size distributions (Fig. 10b) might be argued as reflecting modification of initially log–linear crystal size distributions by growth of large crystals through resorption of smaller grains after nucleation had ceased (e.g. Cashman & Ferry, 1988; Marsh, 1998). However, this interpretation cannot easily account for the systematic decrease in 87Sr/86Sr with crystal size and core-to-rim stratigraphy. A more compelling interpretation, which is consistent with the sanidine isotopic data, is that the crystal size distributions mirror nucleation history (e.g. Spohn et al., 1988). In this case, magmatic cooling following initial injection of rhyolite into shallow crust gives rise to continuous nucleation and growth of crystals and hence an initial log–linear crystal size distribution. However, subsequent injections of fresh, relatively hot rhyolitic liquid from deeper levels of the system lead to decreasing effective undercoolings and therefore decreasing nucleation rates.
Comparison of calculated feldspar growth rates in natural rhyolite magmas, which range from 7 x 10–13 to 1 x 10–14 cm/s (Christensen & DePaolo, 1993; Davies et al., 1994), with feldspar growth rates determined experimentally (10–6–10–10 cm/s) at large undercooling (40–450°C) in granitic liquids (Fenn, 1977; Swanson, 1977) suggests that effective undercooling in natural magmas is remarkably small—probably less than a few degrees Celsius (Hort & Spohn, 1991; Cashman, 1993). This observation largely prohibits quantitative modeling of crystallization of rhyolitic magmas based on classic experimental studies of quartz and feldspar nucleation and growth, which, as noted above, have unfortunately been conducted at large undercoolings. Nevertheless, the recognition that crystallization appears to take place at undercoolings much less than that required to maximize nucleation indicates that nucleation and growth may occur simultaneously over much of the crystallization interval of natural magmas (Cashman, 1993).
A model for the shallow-crustal evolution of Taylor Creek Rhyolite magmas
Similar to the conclusion reached by Feldstein et al. (1994) to explain Sr isotopic disequilibrium among phenocrysts in rhyolites from San Vincenzo, Italy, we suggest that the isotopic systematics of sanidine crystals in the lava of White Water Canyon, Taylor Creek Rhyolite, NM, reflect crystal nucleation and growth in a liquid of changing Sr isotopic and chemical composition. The Sr isotopic compositions of whole-rock samples, which correlate with whole-rock Sr contents (Reece et al., 1990), clearly require addition of a high-87Sr/86Sr crustal component to the parental rhyolite magma to account for the range in ISr defined by the lavas. The crystal isotopic data suggest, however, that contamination occurred at an earlier stage of magmatic evolution and was followed by influx of low-87Sr/86Sr, low-Sr rhyolite.
Figure 11 presents a speculative model for the shallow-crustal evolution of Taylor Creek Rhyolite magma. We envision that, during early stages of emplacement of highly differentiated rhyolite into shallow crust, assimilation of radiogenic Precambrian country rock raised the 87Sr/86Sr of the magma reservoir (Fig. 11a). The initial increase in ISr with decreasing crystal size and immediately outward from sanidine cores for large crystals appears to record this process. Cooling rates and therefore effective undercoolings peaked during this stage, which resulted in the production of log–linear crystal size distributions. Renewed input of differentiated rhyolite from deeper levels of the magma system systematically lowered the 87Sr/86Sr and Sr content of the contaminated high-silica rhyolite magma (Fig. 11b). Growth of a chilled rind may have progressively isolated the magma from the surrounding country rock, thereby allowing for a somewhat broad transition from increasing to decreasing magma 87Sr/86Sr and Sr concentration. Influx of fresh rhyolitic magma and/or the buffering effect of the chilled rind or mush (e.g. Mahood, 1990) apparently reduced the effective undercooling, and gave rise to an episode of crystallization at reduced nucleation rates (Fig. 11b). Finally, continued input of fresh rhyolitic magma resulted in progressive decrease in 87Sr/86Sr and Sr concentration, effective undercoolings and nucleation rates of the mixed magma (Fig. 11c).
|
We emphasize that other interpretations may be possible and that details of our working model may require revision based on continuing studies of isotope systematics of sanidines from other lava domes that span the compositional spectrum of the volcanic field.
|
Summary and Conclusions |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Although whole-rock and bulk-mineral isotopic measurements are important in the study of the petrogenesis of volcanic rocks, they provide limited information on the temporal evolution of magmas because they dominantly reflect the time-integrated history of the system before quenching upon eruption. We show that the isotopic systematics recorded in individual crystals provide a more complete history of the compositional evolution of crustal magma systems.
Strontium isotopic studies of sanidine phenocrysts from a single lava dome of the Taylor Creek Rhyolite show that sanidine 87Sr/86Sr increases with crystal size and distance from crystal cores, reaching a maximum at 4 mm in size and a distance 2 mm from crystal rims, then decreases with decreasing size and distance towards crystal rims. Although no systematic variation in Sr content was observed for sanidine profiles, Sr concentrations increase then decrease with decreasing crystal size. Our data indicate that these variations are not solely the result of in situ decay, diffusive equilibration processes, or incorporation of assimilant melt whose composition varied with time as a result of progressive incongruent melting of chamber wall rocks.
The crystal isotopic data, complemented by grain size variations, lead us to propose a model in which assimilation is followed by recharge. Furthermore, crystal nucleation and growth appear to have occurred throughout much if not all of the open-system evolution of the magma. In contrast, it has been argued that assimilation and crystallization are generally separated in space and time in both small- and large-volume rhyolitic systems (e.g. Grove et al., 1988; Johnson, 1989; McCulloch et al., 1994). Decoupling of assimilation and crystallization in the silicic caps of ash-flow magma chambers is suggested by the fact that assimilation gradients are generally opposite to phenocryst abundance gradients (Johnson, 1989). However, assimilation gradients are typically defined by the bulk isotopic composition of phenocryst concentrates, which as shown here may provide limited information concerning the timing of contamination relative to crystallization. In this regard, detailed measurements of isotopic compositions of individual crystals and zoning within single crystals from ash-flow tuffs may prove a fertile area for future study of temporal and spatial relationships between crystallization and assimilation in large, silicic magma chambers.
|
Acknowledgements |
|---|
We thank Peter Bird, Wendy Bohrson, Craig Manning, and Frank Ramos for discussions and comments that helped clarify the ideas presented. Thoughtful and constructive reviews by George Bergantz, Nelia Dunbar, Clark Johnson, and Gail Mahood improved an early version of the manuscript. The work presented here benefited greatly from the combined efforts of Frank Ramos, Frank Tepley, and Matt Yeager in the set-up and development of microsampling at UCLA. This project was funded by National Science Foundation Grant EAR-9526903 to J.P.D. Writing of the manuscript was completed while the senior author was supported by a fellowship from the Cooperative Institute for Research in Environmental Sciences at the University of Colorado.
* Corresponding author. Present address: Cooperative Institute for Research in Environmental Sciences and Department of Geological Sciences, University of Colorado, Boulder, CO 80309, USA. Telephone: (303) 735-4723. Fax: (303) 492-2606. e-mail: knesel{at}colorado.edu
|
References |
|---|
|
TOPABSTRACTIntroductionGeologic BackgroundMethodsStrontium Isotope DataDiscussionSummary and ConclusionsReferences |
|---|
Armienti P., Pareschi M. T., Innocenti F., Pompilio M. Effects of magma storage and ascent on the kinetics of crystal growth. Contributions to Mineralogy and Petrology (1994) 115:402–414.[Web of Science]
Bevington P. R. Data Reduction and Error Analysis for the Physical Sciences (1969) New York: McGraw–Hill.
Cashman K. V. Crystallization of Mount St. Helens dacite, 1980–1986: a quantitative textural approach. Bulletin of Volcanology (1988) 50:194–209.[Web of Science]
Cashman K. V. Relationship between plagioclase crystallization and cooling rate in basaltic melts. Contributions to Mineralogy and Petrology (1993) 113:126–142.[Web of Science]
Cashman K. V., Ferry J. M. Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization III. Metamorphic crystallization. Contributions to Mineralogy and Petrology (1988) 99:401–415.[Web of Science]
Cashman K., Marsh B. D. Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization II. Makaopuhi lava lake. Contributions to Mineralogy and Petrology (1988) 99:292–305.[Web of Science]
Cherniak D. J. Strontium diffusion in sanidine and albite, and general comments on strontium diffusion in alkali feldspars. Geochimica et Cosmochimica Acta (1996) 60:5037–5043.[Web of Science]
Christensen J. N., DePaolo D. J. Time scales of large volume silicic magma systems: Sr isotopic systematics of phenocrysts and glass from the Bishop Tuff, Long Valley, California. Contributions to Mineralogy and Petrology (1993) 113:100–114.[Web of Science]
Condie K. C. Geochemistry of Proterozoic granitic plutons from New Mexico, U.S.A. Chemical Geology (1978) 21:131–149.[Web of Science]
Crank J. The Mathematics of Diffusion (1975) Oxford: Clarendon Press.
Davidson J. P., Tepley F. J. Recharge in volcanic systems: evidence from isotope profiles of phenocrysts. Science (1997) 275:826–829.
Davidson J. P., De Silva S. L., Holden P., Halliday A. N. Small-scale disequilibrium in a magmatic inclusion and its more silicic host. Journal of Geophysical Research (1990) 95:17661–17675.
Davidson J. P., Tepley F. J., Knesel K. M. Isotopic fingerprinting may provide insights into evolution of magmatic systems. Eos Transactions, American Geophysical Union (1998) 79(185):189–193.
Davies G. R., Halliday A. N., Mahood G. A., Hall C. M. Isotopic constraints on the production rates, crystallization histories and residence times of pre-caldera silicic magmas, Long Valley, California. Earth and Planetary Science Letters (1994) 125:17–37.[Web of Science]
Duffield W. A., Dalrymple G. B. The Taylor Creek Rhyolite of New Mexico: a rapidly emplaced field of lava domes and flows. Bulletin of Volcanology (1990) 52:475–487.[Web of Science]
Duffield W. A., du Bray E. A. Temperature, size, and depth of the magma reservoir for the Taylor Creek Rhyolite, NM. American Mineralogist (1990) 75:1059–1070.[Abstract]
Duffield W. A., Ruiz J. Compositional gradients in large reservoirs of silicic magma as evidenced by ignimbrites versus Taylor Creek Rhyolite lava domes. Contributions to Mineralogy and Petrology (1992) 110:192–210.[Web of Science]
Duffield W. A., Ruiz J. A model to help explain Sr-isotope disequilibrium between feldspar phenocrysts and melt in large-volume silicic magma systems. (1994) 94–423. US Geological Survey Open-File Report.
Duffield W. A., Richter D. H., Priest S. S. Physical volcanology of silicic lava domes as exemplified by the Taylor Creek Rhyolite, Catron and Sierra counties, New Mexico. (1995) Pamphlet accompanying United States Geological Survey Map I-2399. US Geological Survey.
Feldstein S. N., Halliday A. N., Davies G. R., Hall C. M. Isotope and chemical microsampling: constraints on the history of an S-type rhyolite, San Vincenzo, Tuscany, Italy. Geochimica et Cosmochimica Acta (1994) 58:943–958.[Web of Science]
Fenn P. M. The nucleation and growth of alkali feldspars from hydrous melts. Canadian Mineralogist (1977) 15:135–161.
Giletti B. J. Rb and Sr diffusion in alkali feldspars, with implications for cooling histories of rocks. Geochimica et Cosmochimica Acta (1991) 55:1331–1343.[Web of Science]
Grove T. L., Kinzler R. J., Baker M. B., Donnelly-Nolan J. M., Lesher C. E. Assimilation of granite by basaltic magma at Burnt Lava flow, Medicine Lake volcano, northern California: decoupling of heat and mass transfer. Contributions to Mineralogy and Petrology (1988) 99:320–343.[Web of Science]
Grove T. L., Donnelly-Nolan J. M., Housh T. Magmatic processes that generated the rhyolite of Glass Mountain, Medicine Lake volcano, N. California. Contributions to Mineralogy and Petrology (1997) 127:205–223.[Web of Science]
Hort M., Spohn T. Crystallization calculations for a binary melt cooling at constant rates of heat removal: implications for the crystallization of magma bodies. Earth and Planetary Science Letters (1991) 107:463–474.[Web of Science]
Johnson C. M. Isotopic zonations in silicic magma chambers. Geology (1989) 17:1136–1139.
Johnson C. M., Fridrich C. J. Non-monotonic chemical and O, Sr, Nd, and Pb isotope zonations and heterogeneity in the mafic- to silicic-composition magma chamber of the Grizzly Peak Tuff, Colorado. Contributions to Mineralogy and Petrology (1990) 105:677–690.[Web of Science]
Kirkpatrick R. J. Nucleation and growth of plagioclase, Makaopuhi and Alae lava lakes, Kilauea Volcano, Hawaii. Geological Society of America Bulletin (1977) 88:78–84.
Knesel K. M., Davidson J. P. Isotopic disequilibrium during melting of granite and implications for crustal contamination of magmas. Geology (1996) 24:243–246.
Knesel K. M., Davidson J. P. Sr isotope systematics during melt generation by intrusion of basalt into continental crust. Contributions to Mineralogy and Petrology (1999) in preparation.
Mahood G. A. Second reply to comment of Sparks, R. S. J. Huppert, H. E. & C. J. N. Wilson, 1990, on ‘Evidence for long residence times of rhyolitic magma in the Long Valley magmatic system: the isotopic record in precaldera lavas of Glass Mountain’ by Halliday, A. N. Mahood, G. A. Holden, P. Metz, J. M. Dempster, T. J. & Davidson, J. P. Earth and Planetary Science Letters (1990) 99:395–399.[Web of Science]
Marsh B. D. On the interpretation of crystal size distributions in magmatic systems. Journal of Petrology (1998) 39:553–599.
McCulloch M. T., Kyser T. K., Woodhead J. D., Kinsley L. Pb–Sr–Nd–O isotopic constraints on the origin of rhyolites from the Taupo Volcanic Zone of New Zealand: evidence for assimilation followed by fractional crystallizaton from basalt. Contributions to Mineralogy and Petrology (1994) 115:303–312.[Web of Science]
Neal C. R., Davidson J. P., McKeegan K. D. Geochemical analysis of small samples: micro-analytical techniques for the nineties and beyond. Reviews of Geophysics (1995) 25–32. Supplement.
Noble D. C., Hedge C. E. 87Sr/86Sr variations within individual ash-flow sheets. US Geological Survey Professional Paper (1969) 650-C:133–139.
Pushkar P., Stoeser D. B. 87Sr/86Sr ratios in some volcanic rocks and some semifused inclusions of the San Francisco volcanic field. Geology (1975) 3:669–671.
Reece C., Ruiz J., Duffield W. A., Patchett P. J. Origin of the Taylor Creek Rhyolite, Black Range, New Mexico, based on Nd–Sr isotope studies. Geological Society of America, Special Paper (1990) 246:263–273.
Spera F. J. Some numerical experiments on the withdrawal of magma from crustal reservoirs. Journal of Geophysical Research (1984) 89:8222–8236.
Spohn T., Hort M., Fischer H. Numerical simulation of the crystallization of multicomponent melts in thin dikes and sills 1. The liquidus phase. Journal of Geophysical Research (1988) 93:4880–4894.
Swanson S. E. Relation of nucleation and crystal-growth rate to the development of granitic textures. American Mineralogist (1977) 62:966–978.[Abstract]
Tegtmeyer K. J., Farmer G. L. Nd isotopic gradients in upper crustal magma chambers: evidence for in situ magma–wall-rock interaction. Geology (1990) 18:5–9.
Tommasini S., Davies G. R. Isotope disequilibrium during anatexis: a case study of contact melting, Sierra Nevada, California. Earth and Planetary Science Letters (1997) 148:273–285.[Web of Science]
Tuttle O. F., Bowen N. L. Origin of granite in the light of experimental studies in the system NaAlSi3O8–KAlSi3O8–SiO2–H2O. Geological Society of America, Memoir (1958) 74:153.
Webster J. D., Duffield W. A. Volatiles and lithophile elements in Taylor Creek Rhyolite: constraints from glass inclusion analysis. American Mineralogist (1991) 76:1628–1645.[Abstract]
White D. L. Rb–Sr isochron ages of some Precambrian plutons in south–central New Mexico. Isochron West (1978) 21:8–14.
Whitney J. A. The origin of granite: the role and source of water in the evolution of granitic magmas. Geological Society of America Bulletin (1988) 100:1886–1897.
Wittke J. H., Duffield W. A., Jones C. Roof-rock contamination of Taylor Creek Rhyolite, New Mexico, as recorded in hornblende phenocrysts and biotite xenocrysts. American Mineralogist (1996) 81:135–140.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Alves, V. de Assis Janasi, A. Simonetti, and L. Heaman Microgranitic Enclaves as Products of Self-mixing Events: a Study of Open-system Processes in the Maua Granite, Sao Paulo, Brazil, Based on in situ Isotopic and Trace Elements in Plagioclase J. Petrology, December 1, 2009; 50(12): 2221 - 2247. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Erdmann, R. A. Jamieson, and M. A. MacDonald Evaluating the Origin of Garnet, Cordierite, and Biotite in Granitic Rocks: a Case Study from the South Mountain Batholith, Nova Scotia J. Petrology, August 1, 2009; 50(8): 1477 - 1503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Beard, P. C. Ragland, and M. L. Crawford Reactive bulk assimilation: A model for crust-mantle mixing in silicic magmas Geology, August 1, 2005; 33(8): 681 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Davidson, B. Charlier, J. M. Hora, and R. Perlroth Mineral isochrons and isotopic fingerprinting: Pitfalls and promises Geology, January 1, 2005; 33(1): 29 - 32. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Eichelberger, D. G. Chertkoff, S. T. Dreher, and C. J. Nye Magmas in collision: Rethinking chemical zonation in silicic magmas Geology, July 1, 2000; 28(7): 603 - 606. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||