Sunteți pe pagina 1din 13

Ceoehimica a Cosmockimica Ada Vol.50,pp.1619-1631 0016-7037/86/s3.

00
+ .oO
0 Peqmnoa Journals Ltd.1986.
Printed
inU.S.A.

Element redistribution during hydrothermal alteration of rhyolite in an active


geothermal system: Yellowstone drill cores Y-7 and Y-8
NEIL C. STIJRCHIO*, KARLIS MUEHLENBACHSP, and MARTIN G. SEITZ*
*Argonne National Laboratory, CMT-205, Argonne, IL 60439, U.S.A.
tDepartment of Geology, University of Alberta, Edmonton, Alberta T6G 2E3, Canada

(Received December 10, 1985; accepted in revisedform April 29, 1986)

Abstract-Hydrothermal alteration of the 0.54 Ma Biscuit Basin rhyolite flow in Yellowstone drill cores
Y-7 and Y-8 has involved significant changes in Ti-normalized whole rock concentrations of Li, Mg, Ca,
Na, K, Mn, Rb, Sr, Sb, Cs, Ba, and U, whereas such changes have generally not occurred for Al, Fe, SC, Co,
Y, Zr, REE, Hf, Ta, and Th. The relatively mobile elements have been redistributed over distances from
2-5 cm to more than lo- 100 m, and are associated mainly with zeolites and clays, in thoroughly altered
samples; average whole rock changes in Ti-normalized mass range from about OS-17 g/Kg (9-491) for
major elements and 0.1-280 mg/Kg (19-426OW) for minor and trace elements. Compositional changes
correlate with hydrothermal mineralogy. The relatively immobile elements have been redistributed over a
distance of at least lo-100 pm but less than 2-5 cm, and are associated mainly with clays, in thoroughly
altered samples. Addition of Si02 ranges from 1lo-890 g/Kg in samples where porosity has been thoroughly
sealedby silica deposition. Thermal water is not in oxygen isotopic equilibrium with any primary or secondary
phases analyzed except calcite. Modest depletions of “0 (l-2960) due to exchange with thermal water apparently
have occurred in Y-8 plagioclase and obsidian. Oxygen isotope ratios in obsidian-replacing smectite and in
veinlet celadonite and &cristobalite suggest that these phases precipitated from water locally enriched in
“0 (up to -810 heavier than present thermal water). The minimum integrated water/rock mass ratio in
Y-7 and Y-8 is lo’-10” (assuming advective transport).

INTRODUCTION However, relatively little attention has yet been given


to the geochemistry of the drill core samples, despite
THE BEHAVIOR OF the elements during hydrothermal
their unique value as a source of geochemical infor-
alteration of volcanic rock is a topic of considerable
mation.
interest to geochemists and petrologists. An under-
This paper presents data obtained in a multi-element
standing of mass transfer in active hydrothermal sys-
and oxygen isotopic study of rocks and minerals from
tems has important applications in exploration for en-
hydrothermally altered rhyolite in Yellowstone drill
ergy and mineral resources and radioactive waste iso-
cores Y-7 and Y-8. The samples selected for this study
lation (WHITE, 1981; HENLEY and ELLIS, 1983;
are from a single rhyolite flow, having relatively ho-
CONTINENTAL SCIENTIFK DRILLING COMMII-FEE,
mogeneous initial composition, that now va+s locally
1984). An excellent opportunity to study the details of
in in situ temperature, extent of alteration, and hydro-
element behavior during hydrothermal alteration of
thermal mineralogy (WHITE et al., 1975; KEITH et al.,
rhyolite in an active geothermal system has been pro-
1978). The data are interpreted in terms of the relations
vided by the U.S. Geological Survey research drilling
between compositional changes and hydrothermal
program at Yellowstone National Park (WHITE et al.,
mineralogy, the distance of element migration, the
1975). Mineralogical investigations of the 13 drill cores
amount of mass transfer, and the water/rock ratio.
are nearly complete (HONDA and MUFFLER, 1970;
BARGAR and BEESON, 1981; BARGAR et al., 1973;
KEITH and MUFFLER, 1978; KEITH et al., 1978; BAR- BACKGROUND INFORMATION
GAR MUFFLER, 1982; BARGAR and BEESON, 1984,
BEES~N and BARGAR, 1984). These studies have helped Location and strutigraphy. Drill holes Y-7 and Y-8 are lo-
to define the shallow physical environment and hy- cated about 130 m apart, adjacent to the Biscuit Basin parking
area in northern Upper Geyser Basin, southwestern Yellow-
drothermal mineralogy within thermal discharge areas stone National Park, Wyoming (Fig. 1). The holes were drilled
of the Yellowstone hydrothermal system. An extensive during September 1967 to depths of 73.8 and 153.3 m, re-
body of other geological, geophysical, and geochemical spectively. The stratigmphy with increasing depth in both holes
data has also been amassed through studies at Yellow- is similar, consisting of about 2 m of sinter and 50 m of silt,
stone in recent decades, providing a context for the sand, and gravel deposits, underlain by the 0.54 Ma Biscuit
Basin rhyolite flow (KEITH et al., 1978). In Y-7, the Biscuit
hydrothermal studies. These include the results of geo- Basin flow extends from 52.7 m depth to the hole bottom at
logic mapping (U.S. GEOLOGICAL SURVEY, 1972), re- 73.8 m, and consistsalmost entirely of vitrophyric flow breccia
gional heat flow and potential field studies (EATON et over this interval. In Y-8, the flow extends from 55.2 m depth
al., 1975; SMITH and BRAILE, 1984), volcanologic and to the hole bottom at 153.3 m: it consists mainly of vitrophyric
flow breccia resembling that of Y-7 to 62.9 m, and pumiceous
petrologic studies (CHRISTIANSEN, 1984, HILDRETH et tuff is found from there to the hole bottom. The pumiceous
al., 1984), and thermal water and gas analyses tuff is probably a pyroclastic precursor to the cogenetic lava
(THOMPSON and YADAV, 1979; KENNEDY et ul., 1985). flow (W. HILDRETH,~IX. commun., 1986).

1619
1620 N. C. Sturchio, K. Muehlenbachs and M. G. Seitz

and celadonite). Veinlets and other pore spaces contain hy-


drothermally precipitated clays (amorphous, smectite. mixed
smectite-celadonite, and celadonite), zeolites (mainly rlino-
ptilolite and mordenite), beta-cristobalite, and traces of calcite
and pyrite.
In Y-8, the flow breccia is more thoroughly altered than in
Y-7. although much obsidian remains in some vitrophyre
clasts. The hydrothermal mineralogy consists of zeolites (cli-
noptilolite, mordenite, analcime, and laumontite), ccladonite.
silica minerals (a- and fl-cristobalite, quartz, and opal), adu-
laria, and traces of calcite, chlorite, and pyrite.
The pumiceous tuff in Y-8 consists entirely of pumice fiag-
ments that are extensively altered. No obsidian remains. The
hydrothermal mineral assemblage of the tuff consists of mor-
denite, celadonite, quartz, adularia. analcime. and traces of
calcite and goethite.

ANALYTICAL METHODS
YELLOWSTONE Drill core samples were dissected using a rock saw with
distilled water as a lubricant. Portions of rock selected for
bulk analysis, weighing 50- IO0 g, were crushed in a hard steel
piston-sleeve mortar. Splits of the crushed samples were then
pulverized in a shatterbox using an alumina disk mortar.
FIG. 1. Map of Yellowstone National Park, Wyoming, Mineral separates for bulk analysis were obtained by several
showing location of drill holes Y-7 and Y-8 in Upper Geyser techniques. Vein- and pore-filling minerals were handpicked
Basin (UGB). Also shown for reference are Lower Geyer Basin under binocular microscope directly from crushed vein ma-
(LGB), and Norris Geyser Basin (NGB). terial or sliced core sections. Obsidian was handpicked from
size fractions in the +80-35 mesh range. Groundmass zeolites
were separated by centrifugation in heavy liquid of the - 230
mesh size fraction. After repeated rinsing with acetone to re-
Physicalconditions.The in situ temperature of the Biscuit move line adherent particles or residual heavy liquid, all sep-
Basin flow in Y-7 during drilling ranged from about 120- arates were inspected for impurities under a binocular micro-
14O”C,increasing with depth. In Y-8, the temperature ranged scope. Some X-ray powder diffraction patterns were also ob-
from about 16O’C at the upper contact of the flow to about tained for this purpose. Estimated purity of separates is >99%,.
170°C in the middle portion of the hole, then decreased to unless noted otherwise in tables.
about 167°C at the bottom of the hole. The Y-8 temperature Al, Ba, Ca, Fe, K, Li, Mg, Mn. Na, Sr, Ti, and Y were
data were interpreted by WHITEet al. (1975) as an indication determined by inductively coupled plasma atomic emission
that Y-S intersected “the edge of a convecting cell (of water) spectrometry (ICP-AES). Sample dissolution for ICP-AES was
that ‘mushrooms’ outward.” done using hydrofluoric-nitric-perchloric acid treatments in
Y-7 had a vertical water pressure gradient that was essen- open teflon beakers. Na, SC, Fe, Co, Rb, Sr, Zr, Sb. Cs, Ba,
tially hydrostatic, but that of Y-8 was much in excess of hy- La, Ce, Sm, Eu, Tb, Yb, Lu, Hf, Ta, Th, and IJ were deter-
drostatic, especially below about 55 m, due to a decrease in mined by instrumental neutron activation analysis (INAA)
permeabiIity by hydnxbermalself-sealing (WHITEet al., 1975). (JACOB!? et al., 1977; LINDSTROM and KOROTEV,1982). Rel-
Sampledescriptions.A detailed account of the petrography ative analytical uncertainties for ICP-AES and INAA methods
and mineralogy of the Biscuit Basin flow in Y-7 and Y-8 drill (20) are 2-5% for Li, Na, Mg, Al, K, Ca, SC, Ti, Fe, Co, Cs,
cores, emphasizing the effects of hydrothermal alteration, has REE, Hf, Ta, Th, and U; 5-10% for Mn, Rb, and Ba; IO-
been provided by KEITHet al. (1978). Therefore, only a brief 15% for Sr, Y, and Sb; and 25-30% for Zr. Certified values
summary is’given hem. These observations are consistent with of element concentrations in NBS standard reference materials
a single, continuing episode of hydrothermal alteration (T. 278 and 1633a, and recommended values for USGS standard
KEITH,pers. commun., 1985). basalt BCR-I, were routinely reproduced within analytical
In the flow breccia, subangular clasts of perlitic vitrophyre. uncertainty during the course of this study.
ranging from cm to m size, are contained in a finely crushed Electron microprobe analyses were done on polished thin
matrix of the same vitrophyre. The matrix is more altered sections using an automated JEOL JXA-733 Superprobe cal-
than the vitrophyre ciasts that it contains, probably because ibrated with natural glass and mineral standards. Accelerating
of its fine grain size and greater porosity and permeability. voltage for all analyses was 15 kV. For hydrated obsidian
Primary (igneous) mineralogy of nearly unaltered vitrophyre analyses, beam current was 20-30 nA and beam diameter was
is represented by a modal analysis of sample Y7-2 11V (vol.%): 30-50 pm. For clay, feldspar, and pyroxene analyses, beam
obsidian-82.3, plagioclase (An2J_~~b-Or,_~)-7.8, sani- current was 16-20 nA and beam diameter was 10 pm. Relative
dine-2.3, quark-2.1, clinopyroxene (C&,M&iFe+Mn&--- analytical uncertaipty (2~) is 3% or less for oxides having con-
2.8, orthopyroxene (CacrMuFe+Mn&-O.6, Fe-Ti oxides- centrations 28 wt%, and 6% or less for oxides having con-
2.0, zirco&-<O. 1, and-a&&e---iO.l-.’ centrations from I to 8 wt%.
Obsidian in the flow breccia in both Y-7 and Y-8 is thor- Oxygen was liberated from samples for isotopic analysis
oughly hydrated and laced with perlitic fractures. Further al- using the BrF, method of CLAYTONand MAYEDA f 19631.
teration of hydrated obsidian involves replacement by clay Isot&e ratios were measured using a dual inlet-dual collector
along hydration fractures, and progressive replacement of re- mass spectrometer. Data are reported in d notation relative
maining obsidian by zeolites. Total dissolution of obsidian to Standard Mean Ocean Water (SMOW), where S’sO_,
occurs IocalIy in extensively altered samples. FeIdspars and = [(‘*o/160 ,&‘*O/‘60-w) - I)] 1000. A 6’sO value of
quartz show neghgible visible alteration, whereas pyroxenes 9.6 was obtained for NBS28 quartz. Analytical precision is
are generally repIaced by cetadonite. Surfaces of primary Fe- *o. 14%0.
Ti oxides locally are oxidized. Total Hz0 content of hydrated obsidian was determined
In Y-7, the principal products of obsidian alteration are using a Perkin-Elmer Model 240 CHN analyzer, with an es-
clinoptilolite and clays (smectite, mixed smectite-celadonite, timated accuracy of *5%.
Yellowstone rhyolite alteration 1621

Fission track maps weremade on low-U muscovite detectors demonstrated by the generally small coefficients of
clamped to polished thin sections and irradiated for 248 sec- variation, being less than 5% for 22 of the 29 elements
onds in a thermal neutron flux of 1.6 X 10” n/cm2/sec. De-
tectors were etched in 48% HF at 22°C for 17-18 minutes in the RVC for which data are reported. Of the re-
prior to examination. maining seven elements, all but Co (c.v. = 7.6%) are
identified as being relatively mobile in this study, and
thus their relatively large variations suggest that they
MAJOR AND TRACE ELEMENT
GEOCHEMISTRY may have been mobile even in the least altered samples.
Also consistent with the assumption of local homo-
Whole rocks. Analyses of whole rock samples from geneity of the 5ow in northern Upper Geyser Basin is
Y-7 drill core are shown in Table 1. The first two col- an analysis (6YC-106, reported by HILDRETH et al.,
umns list the arithmetic mean and standard deviation 1984) of a sample collected from an outcrop located
of six samples of nearly unaltered vitrophyre, referred about 1 km south of Y-7 and Y-8, that is nearly in-
to hereafter as the “reference vitrophyre composite” distinguishable from the RVC when a slight correction
(RVC). Analyses of whole rock samples from Y-S drill is made for H20.
core are shown in Table 2. The RVC (Table 1) is used as a comparator to es-
Obsidian. Six vitrophyre samples were selected for tablish significant whole rock compositional differences
separation and analysis of obsidian. The analytical re- between samples. The comparison is done using the
sults are shown in Table 3. following normalization:
Secondary minerals. Table 4 gives analyses of smec-
tite, celadonite, clinoptilolite, calcite, and beta-cristo- Ri = (Ci/CTi)sample/(Ci/CTi)RVC (1)

balite from Y-7 drill core samples. Table 5 gives anal- where Ci is the concentration of element i and cri is the
yses of celadonite, clinoptilolite, mordenite, analcime, concentration of Ti. The data are normalized to Ti
and a silica-goethite veinlet from Y-S drill core samples. following the results of other studies where Ti was
found to be relatively immobile during hydrothermal
Compositional changes due to alteration of obsidian (e.g., DICKIN, 198 1).
hydrothermal alteration The normalized concentration (Ri) of an element in
an altered sample is considered to differ significantly
The initial composition of the Biscuit Basin flow in from that of the RVC only if it falls outside of a range
Y-7 and Y-8 is assumed to be adequately represented consisting of the larger a5 (1) three times the coefficient
by the six sample mean reference vitrophyre composite of variation of (ci/cri)Rvc or (2) three times the 1u an-
. .
(RVC) in Table 1. The homogeneity of the 5ow is

Table 1. Whole rock analyses*: V-7 dr,,, core

Sample' RVC" 0 181M 190T 199M 204M 211M 217M 2201.1 227H 234H 239i4

wt x no* 0.43 0.01 0.42 0.43 0.43 0.42 0.40 0.42 0.37 0.38 0. 37 0.38
AlnOs 11.8 0.1 12.2 12.3 12.1 12.6 11.5 11.3 10.7 10.6 10.0 10.3
FcO** 3.07 0.08 3.23 3.14 3.28 3.25 2.96 3.11 2.64 2.53 2.65 2.77
M90 0.42 a. 02 0.28 0.35 E 0.35 0.33 0.35 0.30 0.30 0.28 0.28
cao 1.53 0.05 1.34 1.55 1.46 1.47 1.39 1.39 1.39 1.34 1.54
Ha20 3.08 0.12 2.83 2.82 3: 40 2.98 3.47 3.22 3.34 3.52 3.19 3.24
KzO 4.25 0.13 3.78 3.81 2.52 4.00 3.23 4.04 3.69 3.82 3.79 3.94

PPm Lf 16 i.21 48 57 39 27 31 34 23 23 26 36
SC 6.91 1.01 7.03 7.11 6.95 6.71 6.94 6.13 6.38 6.52 6.00
Ma 641 35 380 630 840 700 490 640 430 480 470 430
Co 2.76 0.21 3.60 3.12 3.35 3.09 2.84 2.96 2. 58 2.46 2.69 2.68
165 5 155 140 163 150 132 140 131 136 133 120
s"! 103 5 107 122 109 112 110 107 106 101 103 106
50.8 0.4 49 47 46
:r 382 13 410 39500 2: 3:; 42500 3: 3:: 400 4:: 380
Sb 0.13 0.02 0.27 0.24 0.16 0.16 0.20 0.18 0.19 0.25 0.19 0.23
CS 5.13 0.42 6.17 5.04 4.65 4.68 3.60 3.56 2.94 2.99 2.82 2.77
8a 908 54 870 970 900 870 1010 1100 1060 940 1050 1230
La 68.8 0.8 66.9 68.4 71.6 68.8 66.5 66.2 63.1 64.5 63.7 62.1
Ce 139 134 141 143 136 135 133 126 132 128 126
11.4 :2 10.1 11.3 11.9 11.4 11.1 10.9 10.3 10.7 10.6 10.1
:zl 1.89 0: 06 1.86 2.04 1.94 1.89 1.89 1.80 1.72 1.84 1.76 1.69
1.78 0.06 1.63 1.72 1.90 1.69 1.76 1.72 1.62 1.73 1.66 1.64
;b" 5.89 0.10 5.42 5.9 6.21 5.62 5.32 5.30 5.11 5.42 5.26 5.21
0.87 0.01 0.86 0.88 0.91 0.90 0.81 0.78 0.76 0.82 0.80 0.76
i: 11.5 0.5 11.1 11.6 12.0 11.4 11.0 10.5 9.9 10.5 10.7 10.3
2.85 2.71 2.80 2.79 2.71 2.60 2.58 2.49 2.63 2. 53 2. 51
?l 20.6 E4 20.4 20.2 21.3 20.3 19.8 19.5 18.8 19.7 19.1 19.7
u 5.2 0:2 1.6 4.9 18.6 18.2 16.2 8.4 13.5 7.0 5.9 5.4

AI*** 0.02 0.02 0.2 0.4 0.2 0.2 0.9 0.8 0.9 0.9 0.8 0.9
"Ti, Al. Fe, Mg. Ca. K. Li, Mn, Sr. V. and Ba by ICP-AES. others by INAA
+Sanple number refers to depth (ft) in drill hole; V - vltrophyre, M = natrlx, T - pumlceous tuff
**Total Fe as Fe0
++RVC = Reference vitrophyrc conposlte comgrised of the mean of six analyses of nearly unaltered vitrophyre from samples
Y7-196. -205. -211. -217. -220. and -239. lxceot value for U that IS the orefcrrtd value chosen on basis of U-series
radioaitlw iquilibrlum ionridirations; .i Is &e standard dcvlatlon abouithe mean, except for U it IS analytical
uncertatnty.
***AI . Alteration index. Visual estlnate of fraction of obsidlsn replaced by secondary minerals.
1622 N. C. Sturchio, K. Muehlenbachs and M. G. Seitz

Table 2. Whole rock analyses': V-8 drill core

Sample+ 182D 19ov 195v 202v 2WM 1971 2267 273T 331T 377T 420T 452T 4BOT 499T

wt.%TiOz 0.20 0.43 0.42 0.43 0.42 0.43 0.40 0.45 0.23 0.45 0.43 0. 47 0.45 0.43
Al203 5.52 11.6 11.6 12.5 11.5 12.9 12.9 13.2 7.41 13. 2 12. 7 13. 1 12. 6 13.0
FeO** 1.27 2.83 2.84 2.78 3.10 3.15 2.75 3.94 1.62 3. 25 3. 26 3.58 3.14 3.16
N90 0.10 0.26 0.31 0.31 0.30 0.38 0.27 0.43 0.13 0.60 0.55 0.46 0.38 0.40
cao 0.38 1.76 1.60 2.17 1.93 1.55 1.01 2.42 0.53 1.82 1.62 1.59 1.73 1.55
0.51 3.86 4.18 3.78 3.80 2.50 1.91 2.48 0.67 2.52 2.44 2.62 2.89 2. 56
II":"
2 4.06 2.02 1.84 2.12 2.63 4.07 6.40 3.24 4.24 3.46 3.93 4.04 3. 25 4.23

ppm Li 270 63 62 60 59 30 124 50 210 74 52 46 42 43


SC 2.64 6.75 6.73 6.92 6.51 7.02 6.43 7.01 3.07 7.32 7.29 7.86 7. 29 6. 76
nn 160 530 580 490 440 410 360 560 170 430 440 530 620 390
CO 1.26 2.66 2.66 2.49 2.96 2.51 2.61 3.15 1.41 2.96 2.93 3.22 2. 77 2.63
Rb 120 134 120 114 133 160 230 126 113 157 136 148 135 156
sr 36 109 94 99 106 110 92 170 51 123 123 120 121 122
V 50 47 50 51 32 52 52 50
zr 2:; 444: 350 4:: 510 370 404: 350 270 4:: 3iA 280 370 475
Sb 0.36 0.25 0.36 0.32 0.15 1.74 0.63 3.49 0.67 0.53 0.26 0.31 0. 25
CS 67:A 228 223 250 185 8.02 131 198 5.34 49.4 15.0 6.53 13. 1 7.39
Ba 400 920 550 630 1040 900 780 1130 590 830 950 960 1010 940
La 30.1 69. 5 68.1 71.0 64.0 71.1 61.9 67.8 43.5 72.2 69.4 72.2 69.5 71.0
Ce 60.3 139 136.5 142 131 143 123 137 07.5 141 139 142 139 140
Sm 4.90 11.3 11.2 11.4 10.6 11.4 10.3 11.5 7.1 12. 1 11.5 12. 1 11.7 11.8
EN 0.78 1.92 1.90 1.90 1.81 1.91 1.86 1.94 1.15 2.04 2.12 2.12 1. 90 2.06
Tb 0.75 1.73 1.82 1.83 1.61 1. 73 1.49 1.70 1. 06 1.79 1.70 1. 72 1.70 1. 67
Yb 2.50 5.64 5.68 5.37 5. 62 5.00 5.48 3.41 5.91 5. 56 5.75 5.64 5.64
LU 0.37 0.85 %3 0.95 0.81 0.90 0.78 0.84 0.51 0.94 0.85 0.85 0.86 0.88
Hf 5.45 11.8 10.9 12.2 12.9 12.3 10.1 11.5 11.4 11.0 10.6 10.5 12.8
1.31 2.90 2. 76 3.30 2.66 2.80 2.53 2.72 X0 2.65 2. 74 2.74 2.82 2.64
:z 9.1 21.2 19.9 22.0 19.9 21.9 19.5 20.6 13:6 24.1 20.3 20.9 21.4 21.0
u 2.66 5.9 5.6 6.6 6.4 5.9 4.2 4.3 4.2 6.0 4.9 5.4 5. 1 5.5

Alit 1.0 0.8 0.5 1.0 1. 0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0

lTi, Al. Fe. Mg. Ca. K. Li, Mn, Sr, V. and Ba determined by ICP-AES: otters by IMUA
+SlaPle nuder refers to depth (ft) In drill hole; 0 = devltrifled. V * vitrophyre. M = mdtrix, T = pumfccous tuff
**Total Fe as Fe0
++A1 = Alteration index. VlsNal CStlnBte of fraction of obsidian replaced by secondary minerals.

Figure 2 shows a plot of Ri, by element, for ail sam-


ples that fall outside the 3a compositional “filter” des-
ignated for each element. Samples having Ri values
less than one have been depleted, and samples having Typed RE-S
wt.% sio,
Ri values greater than one have been enriched, respec- :IxP, 0.30
11.6
10.0 2.53
FeOe 4.93 Ii.1 18.7 0.36
n90 .46 1.46 1.65 0.07
cao 0.55 0.42 2. 03
Table 3. Obsidian analyses* NazO 1::: 0.26 0.45 0.0086 0.51
K?O 4.21 7.28 7.22 4.40

Sample Y7 174 Y7 196 17 205 Y7 217 17 239 Y8 195 ppm Li 148 <lo
SC 12.0 2.14 0.144
Mll 1390 80
CO 17. 3 -0.63 0.116
wt.x 5102 71.6 70.6 71.9 71.6 70.8 68.9 Rb 270 155 0.9
TiOl 0.23 0.22 0.23 0.21 0.22 0.22 Sr 47 46 75
11.3 11.1 11.2 11.1 11.2 11.1 Y
;:0ots 66 31
1.57 1.60 1.55 1.57 1.57 1.58 22” 145
H90 0;09 0.08 0.09 0.08 0.06 0.10 f,r 0.33 O.Q9 0.008 -0.28
CaO 0.57 0.54 0.57 0.52 0.51 0.58 8.6 2.84 0.011 3.01
:"'o" 2.41 2.45 2.57 2.64 2.93 4.41 ii: 480 100 1030 29 220
4.52 4.83 4.60 4.66 4.52 1.42 La 102 12.8 3.47 3.63
Hz0 6.75 6.93 6.93 7.38 7.65 10.2 Ce 207 28.1 7. 9 7.0
Sm 23.9 2.98 2.27 0.70
ppn Li Cl5 a5 Cl5 a5 <I5 63 2.26 0.36 0.440 0.063
SC 3.97 3.76 3.78 3.75 3.75 4.06 :;: 4.5 0.72 1.55 0.126
230 250 240 220 210 280 Yb 19.1 4.22 11.9 0.53
!", 0.71 0.68 0.72 0.80 0.74 1.36 LU 2.99 0.62 1.76 0.085
200 200 190 200 190 Hf 3.0 7. 1 0.44
!!," 38 40 36 39 30 ‘:: la 0.81 2.22 0.103
Th 7.7 11.9 0.381 0.96
V 53 53 48 48 39
" 3.4 5.2 0.14 1.42
Zr 370 360 3:: 360 350 360
Sb 0.11 0.13 0.12 0.12 0.11 0.16 Anlytlcalf 1 3 5 4.6 2 4
CS 4.96 4.48 4.64 5.15 5.74 336 PytNod n-4 n-5
630 850 MO 820 710 420 *C11Noptllolite scplr.tc contains cn cristoballte and trace of
:", 74.0 73.5 74.5 ‘14.5 73.6 62.4 celadonlte
C@ 145 I44 145 145 144 124
bN-cristobNl+te sep~rata contains 43X copnclpltited caladonite.
Sm 11.5 11.4 11.5 11.6 11.5 10.0 Plus dcWctNble .-trl~tob~HU. chaleCdony. and cllwoptllolitc
ELI 1.00 0.98 1.01 1.01 0.98 1.19
%ect1te SepwLte conulns-uD MIncrplndlng Clry (9.8 i basal
Tb 1.76 1.78 1.73 1.76 1.77 1.54
spacing) as detected by XRO folloufng 24 hr. glycolatlon at room
Yb 5.99 5.82 5.90 5.90 5.81 5.15 temperature
LU 0.91 0.90 0.90 0.90 0.91 0.82
Hf dRE - nplrcement of obsidian; VW - veinlet; S = smxtitc;
10.1 10.0 10.0 10.3 10.1 9.7
C = crladonlta
Ta 3.23 3.17 3.20 3.20 3.21 3.15
Th 24.9 24.8 24.9 24.9 24.8 23.9 *Total Fe 4s Fe0
U 5.8 5.4 6.8 5.9 5.5 4.B f"n" SigNiflNs nuder of spa: r~a;ys;f .;?a d
- 8elw drtrctiaf! limit - , 1 ,5. ca, u. L1.IIn. sr.
1 - hulysfr ky ICP-AES Y. in8 Ba by ICP-MS; others by
‘Si, Ti. Al, Fe. MB. and CN &wdnod k~ slcctr8e~croprolc INAA
2 - Anlysfs bL III
(6-11 s9ot.t iX?r NulySlst; 1. L1. Mn. Sr, Y. Nnd 8a by ICC-A%
3 - 1~lysls by @lcctron 5 - Si. Ti, Al, Fe. Ng. Ca. and K by
H,O by CHN analyzer: otters by INM
m1cr0proba electron mlcr0proW; otkerr by INAA
tTotal Fe as Fe0 6 - si titcltima ~r~v3rtrkaliy
Yellowstone rhyolite alteration 1623

Table 5. Hydrothermal mineral analyses: Y-8

Clinop- Red Chal-


Ccladonlte ti101ite Hordcnite Analcime cedony'
Sample Y8 19OV Y8 19OV Y8 19OV Y8 195V Y8 19OV Y8 273T Y8 273T Y8 226T Y8 331T

Typeb RE BUD VN VN RE RE PO VN VN
wt.%sio* 52.6 48.0 52.1 67.0 65. 7
TIO, 0.51 5.12 0.02 0.07 0.20 0.08 0.03
11.1 11.4 9.10 11.1 11.9 11.4
“,+p 13.8
13.3 17.4 19.2 13.2 1.80 1. 13 0.26 0.05 2.54
M90 1.30 0.81 1.97 1.44 0.15 0.17 0.03
cao 0.25 0.35 0.29 0.22 1.23 2.77 2.91
Nan0 0. 23 0.11 0.27 0.27 3.30 2.71 3.46 11.2 0.28
f.20 6.47 8.27 6.64 6.07 1.70 1.33 0.39

pm Li 139 35 19 27
SC 10.7 0.56 0.058 4.78
Mn 650 160 120 38
co 14.0 0.22 - 1.04
Rb 300 :i 53 123
sr 15 51 180 15
Y 150 28 13 <5
2r 130
Sb 0.31 5.17
CS 125 500 2900 9.09
Ba 110 970 1710 520 110 690
La 320 4.6 - 170
Ce 578 9.2 - 317
501 45.1 0.77 0.06 26.4
EU 2.86 0.12 0.32 2.31
Tb 6.0 3.13
Yb 16.8 0.34 - 7.28
LU 2.75 0.07 - 1.17
Hf 4.0 0.9 0.9 4.04
Ta 1.21 0.21 0.2 0.98
Th 9.4 2.0 - 5.67
U 8.9 0.20 - 10.7
Analytrcald 3 3 3 4 1.6 1.6 4 2 2
Method n-2 II=3 n=3
aRcd chalcedony separate contains several Dcrcent aoethite
-
bRE - replacing obsidian; VN = filling veinlct; W = filling pore space; BUD = nodular
within zeollte replacIng obsidian
cTota1 Fe as Fe0
dkc Table 4
- Below detection ,,nlt

tively, in element i relative to its normalized concen- netite surfaces (Fig. 3), where U may be absorbed on
tration in the RVC. Figure 2 demonstrates that either a thin ferric oxide or oxyhydroxide coating, and to a
very few or none of the samples exhibit significant dif- lesser extent with clays and Mn-rich material in perlitic
ferences relative to the RVC with respect to Al, Fe, SC, fractures. U-series isotopic data for Y..7 samples in-
Co, Y, Zr, rare earth elements (REE), Hf, Ta, and Th. dicate that U was added to these samples within the
This group of elements apparently had initial concen- last 30,000 years (STURCHIO et al., 1985).
trations that generally were not significantly different In matrix samples having high alteration indices,
from those in the RVC, and behaved coherently with behavior of several elements is observed to correlate
Ti. Therefore, this group of elements is considered to with the extensive replacement of obsidian by clino-
have been relatively immobile during hydrothermal ptilolite and clay; in these samples, normalized Na and
alteration. The data imply also that mechanical frac- Ba are high and normalized Cs is low (Table 1). This
tionation of phenocrysts and obsidian probably did suggests that Na and Ba are retained and then further
not occur to a significant extent during emplacement concentrated from thermal water when hydrated ob-
of the Biscuit Basin flow, because elements that would sidian is replaced by clinoptilolite and clay, whereas
show this effect (e.g., Eu in plagioclase or Zr, Hf, and there is a net loss of Cs during this process. The cli-
the heavy REE in zircon) are generally not deviant noptilolite analysis from Y7-234M (Table 4) shows that
relative to the RVC. Tables 4 and 5 show that the rel- this mineral has significantly more Ba and less Cs than
atively immobile elements released from obsidian be- the obsidian that it replaces (Table 3).
came concentrated in celadonite rather than in zeolites Normalized Mg and Mn tend to be low. The low
or other phases. Mg is probably a result of the replacement of pyroxenes
(having 10.5-l 1.9 wt% MgO) by clays (having <2 wt%
Relation of compositional changes to MgO). The low Mn also is probably a result of the
hydrothermal mineralogy replacement of pyroxenes (having 0.53-1.35 wt%
MnO) by clays (having x0.18 wt% MnO), as well as
Y-7 drill core. Li, Sb, and U have high normalized the replacement of obsidian (having -230 ppm Mn)
concentrations in nearly all altered Y7 samples. The by clinoptilolite (having -80 ppm Mn). One sample
high Li concentrations are probably due to the amount (Y7-199M) has high normalized Mn, and probably
of Li contained in clays (Table 4). Sb is apparently contains Mn that was lost from nearby rock.
concentrated in c&don&e and beta-cristobalite (Table Absolute concentrations of Ti and the other rela-
4). Excess U in Y-7 has been found by induced fission tively immobile elements in matrix samples below 2 17
radiography to be strongly associated with titanomag- ft are low relative to the RVC as a consequence of
1624 N. C. Sturchio, K. Muehlenhachs and M. G. Seitz

1 addition of silica. However, normalized concentra-


tions of Al, Fe, Co, Y, Zr, Ba, REE, Hf, Ta, and Th
show no significant deviation from the RVC.
Three altered samples of vitrophyre and one of ma-
trix from thejlow breccia in Y8 were analyzed. These
samples all have low normalized Mg and K and high
normalized Na and U, in addition to the ubiquitous
high Li, Sb, and Cs. Most of the Y8 flow breccia sam-
ples also have low Rb and high Ca; two have low Mn
and two have low Ba.
The loss of Mg is probably related to the replacement
of pyroxenes by celadonite as noted in Y-7. The loss
of Mn, also as in Y-7, is probably related to the re-
placement of obsidian by clinoptilohte, as well as the
replacement of pyroxenes by celadonite. The gain of
Na and Ca and the loss of K (and Rb?) are probably
a Y7 MATRIX related to the replacement of obsidian by clioptilolite
•I Y7 TUFF
and mordenite. For example, YS- 190 clinoptilolite
0 Y8 DEVITRIFIED
(Table 5) has substantially higher concentrations of Na
A Y8 VITROPHYRE
0 Y8 MATRIX
and Ca and a lower concentration of K than average
0 Y8 TUFF
Y-7 obsidian (Table 3).
The overall composition of the Y8 obsidian resem-
bles that of the Y7 obsidian, except for significant dif-
ferences in H20, the alkalis, and Ba (Table 3) that may
03 1.0 2.0 5.0
be due to obsidian-water ion exchange. Alkali ion ex-
change occurs commonly during low temperature hy-
(Cl/cTl),,,/(clm REFEI)L”CE VITROPIWRL COMPOSITE dration of obsidian (TRUESDELL,1966; NOBLE, 1967;
0.5 1.0 2.0 5.0 IO 20 !xJ ID0 JEZEK and NOBLE, 1978). The general trend observed
by previous workers is that Na+ from obsidian is ex-
Li
changed for K+ (and H+) from water. This contrasts
Sb 0
with the apparent exchange of K+ for Na+ observed in
CS
Y8-195 obsidian relative to Y7 obsidian (Table 3). In
FIG. 2. &@illll Shoed VdWSOf (Ci/CTi)-~/(Ci/C+i)Rvc addition, Y8-195 obsidian shows large gains of Cs and
(=R,) for all elements analyzed. Each symbol represents an Li, and losses of Ba and Rb, that may also be a result
individual whole rock sample; only samplesfor which Ri dB&s
significantly (>3~) from (c~/c&~~ are depicted. Horizontal of ion exchange. These effects may be related to a
bars represent 3u uncertainties in (Ci/c&vc as defined in text. structural change caused by the high (10.2 wt%) Hz0
Note different scale for Li, Sb, and Cs. content. However, the X-my diffraction pattern of Y8-
195 obsidian shows an amorphous structure.
Nine samples of pumiceous tuflfrom Y8 drill core,
dilution by precipitated silica (Table 1). The amount spaced at roughly 10 m intervals, were analyzed (Table
of dilution corresponds to addition of about 14-l 6 wt% 2). The most extreme compositional differences were
silica. found in samples Y8-226T and Y8-33 1T. These two
Y-8 drill core. All Y8 samples analyzed are high in samples are both cut by prominent fractures, indicating
normalized Li, Sb, and Cs (Table 2). As in Y-7, the that their compositional changes are related to the
high Li is probably related to the concentration of Li proximity of major thermal water flow channels. The
in celadonite, and the high Sb may be related to the fracture cutting YS-226T was still apparently open
concentration of Sb in celadonite and chalcedony (Ta- when drilled, and is lined with analcime and mordenite.
ble 5). The high Cs is related to the high concentrations The fracture cutting sample Y8-33 1T is filled with red
of Cs in analcime (KEITH et af., 1983), mordenite and chalcedony. The low absolute Ti concentration in
celadonite (Table 5), and in some vitrophyre samples, sample Y8-33 IT (0.53 X RVC) reflects silicification of
obsidian (Table 3). this sample, probably caused by infiltration of, and
Near the upper contact of the flow in Y-8 is a zone precipitation of silica from, silica-oversaturated water
of devitrified rhyolite about a meter thick. The obsidian from the crosscutting fracture. An analysis of the red
in this zone has been replaced mainly by quartz and chalcedony from sample Y8-33 1T (Table 5) shows high
alkali feldspar rather than clinoptilolite or mordenite. concentrations of Sb, Cs, Ba, U, and light REE, sug-
One sample from this zone was analyzed (Y8-182); gesting that the source of the high concentrations of
this exhibits the most extreme compositional changes these elements in the whole rock was the water flowing
of all samples analyzed. It also has the lowest absolute through the fracture. Both samples Y8-226T and Y8-
concentration of Ti (0.47 X RVC) and other relatively 33 1T have high Al, suggesting precipitation of an alu-
immobile elements, indicating dilution by roughly 1: minous phase such as aduiaria, which is present in
Yellowstone rhyolite alteration 1625

FIG. 3. Photomicrographs of: (a) titanomagnetite grain in slightly altered perlitic vitrophyre fragment,
sample Y7-205M (simultaneous reflected and transmitted light) and (b) fission track map of *% microdis-
tribution in same area showing extreme concentration of U at titanomagnetite surfaces. (Width of view
= 1.5 mm.)

these samples and could also account for the high K rienced minimal compositional changes (i.e., the RVC
and Rb. samples, Table 1) have obsidian containing about 7
The other pumiceous tuff samples exhibit various wt% Hz0 (Table 3).
normaliied compositional changes (Table 1). However, (2) In Y-7 matrix samples having low alteration in-
the only consistent changes are increases in Li, Sb, and dices, obsidian has been replaced by smectite adjacent
Cs and decreases in Na and Mn. to surfaces, and locally by clinoptilolite adjacent to
Sample Y8-452T has a high absolute concentration smectite. Such samples exhibit increased concentra-
of Ti relative to all other samples ( 1.09 X RVC). This tions of Li, Sb, and U (Table 1). Analyses of Y-7 smec-
may have been due to loss of silica; removal of about tite and celadonite (Table 4) indicate that the additional
8 wt% silica would account for the apparent gain in Li and Sb are associated with the clay that replaces
Ti. Secondary phases replacing obsidian in Y-8 samples obsidian. As determined by induced fission radiogra-
(Table 5) have lower Si/Al ratios than obsidian. How- phy, the additional U is associated mostly with titano-
ever, with the exception of Y8-452, all of the whole magnetite surfaces (Fig. 2), and to a lesser extent with
rock samples analyzed have either effectively retained clay and associated Mn-rich material.
or gained silica. This implies that the thermal water (3) In Y-7 matrix samples having high alteration
flowing through them has been consistently either sat- indices, most obsidian has been replaced by clinoptilo-
urated or supersaturated with silica. Removal of silica lite and celadonite, and pyroxenes have been replaced
probably requires infiltration by silica-undersaturated by celadonite. Such samples exhibit increased concen-
water such as would be expected if the flow path were trations of Ba and Na and decreased concentrations
up a thermal gradient. of Cs, Mn, and Mg (Table l), in addition to changes
acquired during stage (2).
The last two stages are inferred from Y-8 data:
Compositional evolution during alteration (4) Vitrophyre and matrix samples from the flow
breccia in Y-8 exhibit further increases in Na and Li,
On the basis of the observed compositional changes and a further decrease in Mg, along with increases in
and their relation to the extent of alteration and the Ca and Cs and decreases in K, Rb, and Ba (Table 2).
hydrothermal mineralogy, an inferred five-stage com- These changes occurred concomitantly with the re-
positional evolution (relative to the RVC) for an hy- placement of clinoptilolite by mordenite and the ap-
pothetical sample undergoing prograde hydrothermal pearance of analcime and adularia. Also observed in
alteration can be proposed. The first three stages are some Y-8 vitrophyre samples is obsidian containing
inferred from Y-7 data: near 10 wt% HZ0 that has apparently experienced sig-
nificant obsidian-water cation exchange (Table 3).
(1) The earliest stage of hydrothermal alteration re- (5) Pumiceous tuff (Table 2) exhibits further in-
sults in hydration of obsidian beyond the 1.8-3.5 wt% creases in Li and Sb, and a further decrease in Mn.
HZ0 observed in perlitic obsidian from outcrops of the Increases in K and Rb along with decreases in Na, Ca,
flow in the vicinity of Upper and Midway Geyer Basins. and Cs may correlate with the replacement of mor-
Vitrophyre samples in Y-7 that appear to have expe- denite by adularia and quartz.
1626 N. C. Sturchio, K. Muehlenbachs and M. G. Seitz

Table 6. Thermal water analyses. lmg/L)'

Y-l Y-E Y-7 Y-B

SiO2 379 340 La <0.0005 <0.0003


Ti <0.005 <0.005 Ce <o.Luu
Al 0.30 0.15 Sm ~0.0001 ~o.oow6
Fe 0.06 0.3 EU ~0.00004 ~O.OOLm3
ng 0.072 0.40 Yb <0.0005 <o.owz
CI 4.20 8.0 LU ~O.OoOO7 ~0.00002
Na 409 365 Hf <0.0004 ~.Oc@2
K 91.1 18 Ta ~0.00004 ~0.000+6
Li 0.95 0.8 Th <0.0001 0.00005
SC 0.00002 0.000009 " 0.00092 0.00007
Mll 0.02 0.1
CO 0.00007 ~0.0002 HCOs 565 533
Rb 0.18 0.1 SOS 11 14
Sr <0.005 qo.01 F 31.1 30
2r co.01 co.01 CT 336 233
Sb 0.021 0.030 Br 1.0 0.59
CS 0.0037 0.13 lab pH 7.64 6.62
Ba 0.012 <o.oi TOS 1829 1544
I
'Y-7 and Y-B thermal water collected by dcwnholc sampler at hole
bottoms in 1980 and 1977, rerpectlnly, by methods of Thompson and
Yadav (1979). Data for Si&, Fe, Ca, Na. K. Li. HCDj, SOt,.F, Cl,
and pH in both samples and M Mn, and Rb in Y-S from J. M. Thompson
(written co~nication. 19843,
. Oata for other elements in Y-B are
from analysis Of filtered acidified water collected from Rusty Geyser
in September 1984. because water from Y-8 is no longer available;
available dot4 show a close relation beteen Y-B water and tht of
Rusty Geyser. Analysis by ICP-AES: Ti, Al, Fe, Mg (Y-7 only), Hn
(Y-7 only), Sr. Zr, Ba; a~lysls by INAA: SC. Co. Rb (Y-7 only), Sb,
Cs. La. Ce, Sm. Eu. Yb. Lu. Hf. Ta: Th in Y-7 by INM; U in Y-7 by
laser fluo;iwry. Th and IIin Rusty Geyser by-isotope dilution
alpha spectrometry.

OXYGEN ISOTOPE GEOCHEMISTRY 6’800BsrDIAN= 0.5 in sample 6YC-101, These values


are similar to those for plagiclase and obsidian in Y-7
Oxygen isotope ratios were measured in 20 separates
samples (Table 7), suggesting that primary phases in
of primary and secondary phases and two whole rock
the Y-7 samples may not have experienced significant
powders. The data are shown in Table 7.
oxygen exchange with thermal water.
Primary phases. Plagioclase and obsidian separates
The equilibrium fractionation factors between pla-
from Y-8 samples are significantly depleted in “0 rel-
gioclase (mole fraction anorthite = 0.3) and water at
ative to those from Y-7 samples; the depletion of pla-
the in situ Biscuit Basin flow temperatures of 130°C
gioclase is about - l%, and that of obsidian is about
in Y-7 and 170°C in Y-8 would be approximately 13
-2%. The only available oxygen isotope data for feld-
and 10, respectively (O’NEIL and TAYLOR, 1967).
spar and obsidian from relatively unaltered vitrophyre
Present day thermal water in Biscuit Basin has d”0
from the Biscuit Basin flow are those of HILDRETH
near - 17 (C. JANIK, pers. commun., 1985). Plagioclase
et al. (1984), who report G”OsAmm = 0.6 and
in equilibrium with this water should thus have 6’*0
values of approximately -4 in Y-7 and -7 in Y-8.
Therefore, the piagioclase in both drill cores is not in
Table 7. oxygen isotope aata
equilibrium with the coexisting thermal water. Oxygen
diffusion in feldspar occurs at such a low rate that it
could require over a million years for isotopic exchange
equilibrium to occur in these mm-sized grains even at
Y7-174V Obtidjan -0.4 a temperature of 300°C (GILETTI et ul., 1978).
Y7-196V Obsidian +o.a
Y7-217V Obsfdian -0.1 Secondaryphases.Bladed calcite from an open frac-
Y7-239V Obsidlan +0.4 ture in sample Y7-209 has 6”O = -3.0. According to
Y7-174V Plagloclase +1.1
Y7-196V Plagioclase +1.1 the calcite-water fractionation curve of O’NEIL ct ui.
Y7-205V Plagfoclase +1.5 ( 1969), this calcite would be in equilibrium with pres-
Y7-217V Plagioclase +1.6
Y7-239V Plagfoclase +1.4 ent-day thermal water (a”0 = - 17) at a temperature
Y7-199M Swctlte (replacement) +O.S
Y7-209V Calcfte (vein) -3.0' of 13O”C, which is approximately the in situ temper-
Y7-239V 6-Crfstoballte(vein) t2.a ature in Y-7 at 209 ft. (WHITE et al., 1975). Therefore,
Y7-239V Celadonite (vein) +0.3
this calcite appears to be presently in equilibrium with
YE-195V Obsidian -2.1 thermal water.
Ye-195V Plagfoclase tO.5
YE-207T Plagfoclase The average S13CpDB of igneous rocks is -5. whereas
YE-226T Plagfoclase ::: that of sedimentary marine carbonates is 0 (OHMOTO
YE-273T Plagfoclase -0.3
YE-377T Plagfoclase +0.3 and RYE, 1979). The 613CpoHvalue of -0.8 in k’7-209
YE-452T Plagfoclase +0.4
YE-195V Celrdoftfte(vefn) -2.7 calcite indicates that its C may be derived from sedi-
YE-182U Whole Rock -1.7 mentary marine carbonate dissolved in the thermal
YE-331T Whole Rock -3.1
water. Possible sources for this C arc the pre-Cretaceous
*Y7-209V calcite has 6’3Cpm = -0.8 limestones and dolomites that underlie the Quaternary
Yellowstone rhyolite alteration 1627

volcanic units in the Yellowstone area (U.S.G.S., 1972). cedony-filled fracture; both were silicified to the extent
A marine carbonate source for C in Yellowstone ther- that the relatively immobile elements have been diluted
mal water has previously been suggested by isotopic (relative to the RVC) by about 50%. Oxygen isotope
studies of CRAIG (1963) and LEEMAN et al. ( 1977). ratios were measured in these two whole rocks (Table
This implies a km-scale transport distance for C. 7) to evaluate the hypothesis that their extreme com-
Celadonite and /3-cristobalite occur together in a vein positions may have been related to extreme tempera-
in sample Y7-239, where the celadonite forms a layer tures of alteration. Assuming that silicification occurred
about 2 mm thick on the fracture wall and is covered simultaneously with the other compositional changes
by a layer of fl-cristobalite up to 9 mm thick. The in in these samples during a single event, a range of values
situ temperature in Y-7 at 239 fi was near 140°C during for temperature and 6’80n,o can be estimated using
drilling (WHITE et al., 1975). Assuming that the cela- the quartz-water fractionation curve of CLAYTON et
donite and silica precipitated in equilibrium with water al. (1972) and reasonable initial values for 6’*ORock.
at this temperature, then 6’80n,0can be estimated from For example, if initial 6180Rocx had been -1 and
available mineral-water fractionation curves for illite 6’80n,o had been - 17, then 6”O for added silica would
(ESLINGERand SAVIN, 1973) and quartz (CLAYTONet have been -2.4 for YS-182D and -5.2 for YS-33 lT,
al., 1972). The resulting estimates of 6’80nfl are -9 and the respective temperatures of silicification about
from celadonite and - 14 from /3cristobalite. Similarly, 170°C and 2 10°C; if 6”Onfl had been - 13, then the
if the smectite that replaces obsidian in Y7- 199M pre- respective temperatures would have been about 230°C
cipitated at the present in situ temperature of 130°C and 280°C. Higher initial 6’80RocK or 6”Omo would
then 6’80n,o would have been - 11 (YEH and SAVIN, increase these estimated temperatures.
1977). Clearly, these minerals are not in isotopic equi- Several interpretations of the 6”O data are possible.
librium with present-day thermal water (6180 = - 17). For YS- 182D, the silicification may have occurred un-
Their S180 values may reflect a local 180 enrichment der present conditions (T = 170°C 6’80u,o = - 17).
in the water that was associated with their precipitation. Alternately, silicification may have occurred at a much
A transient I80 enrichment can be caused locally higher temperature, such as during interaction between
by the ingress of water into a newly opened fracture if ‘80-enriched water and a cooling lava flow (e.g., T
(A’80~c,cK_u20)is sufficiently large and the water/rock = 4oo”C, PO”,0 = -7). The second of these inter-
ratio is sufficiently small. In Y7-239, the fracture that pretations is favored, because the type of devitrification
is now occupied by celadonite and P-cristobalite cut observed in YS-182D (replacement of obsidian by
obsidian having aI80 probably near 0.5. This value of quartz and alkali feldspar) occurs under high temper-
6180 is sufficiently high to cause an “0 enrichment in ature conditions (KEITH and MUFFLER, 1978); also,
water having b180 = -17 at 140°C therefore the ef- there is similar devitrified rhyolite in surface outcrops
fective water/rock ratio in this fracture was apparently of the Biscuit Basin flow that are away from thermal
low when these phases precipitated. An additional discharge areas.
consequence of the ingress of water into this fracture For sample YS-331T, given the obvious source of
would have been dissolution of obsidian, causing sat- added silica represented by the chalecedony-filled frac-
uration of various solid phases in the water; 6180 of ture, it is reasonable that the silicification may have
those phases that precipitated might have reflected the been caused by relatively hotter, silica-saturated
transient I80 enrichment of the water. These effects or -supersaturated water flowing through the fracture
would have diminished with time as the surface equil- and infiltrating the surrounding porous and permeable
ibrated with the water or became coated with precip- pumiceous tuff. This water probably would have been
itates. out of chemical equilibrium with the tuff, therefore
Celadonite filling a 3 mm wide fracture in YS- 195 driving mineral reactions (e.g., replacement of zeolites
also reflects disequilibrium with present-day thermal by adularia) that would result in compositional changes
water. Assuming precipitation near the measured tem- in addition to the precipitation of silica.
perature at this depth in Y-8 (17O”Q this phase would
be in equilibrium with water having aI80 = - 10. Thus, DISCUSSION
a similar transient I80 enrichment in the water of pre-
Distance of element migration
cipitation is indicated by celadonite from YS-195 as
that from Y7-239. If the minerals formed during glacial Relatively immobile elements. The relatively im-
times, meteoric water would have had lower 6180 and mobile elements were clearly redistributed over a dis-
the transient “0 enrichment would have been even tance of at least lo-100 pm, which is the approximate
larger than the apparent 7-8%0. size range of obsidian fragments replaced by clays and
Whole rocks. The two samples examined in this zeolites in hydrothermally altered vitrophyre. The net
study that exhibit the most extreme bulk chemical dif- migration distance is small relative to the size of the
ferences relative to the RVC are YS-182D and YS- whole rock samples analyzed (2-5 cm), therefore sig-
33 1T. YS- 182D is from a layer of devitrified rhyolite nificant changes in the normalized concentrations of
about a meter thick that occurs at the upper contact these elements did not occur.
of the Biscuit Basin flow in Y-8, whereas YS-331T is Chondrite-normalized REE patterns for Y 8- 195
silicified pumiceous tuff that is cut by a cm thick chal- veinlet celadonite, YS-33 1 veinlet red chalcedony, and
1628 N. C. Sturchio, K. Muehlenbachs and M. G. Seitz

average obsidian from the Biscuit Basin flow in Y-7 regarding the migration distance of the relatively mo-
are essentially parallel (Fig. 4). This suggests that Biscuit bile elements can be gained by considering volumes of
Basin flow obsidian is the source of the REE in these rock much larger than those taken for whole rock
precipitates, and also that the REE were transported analysis. Table 8 contains average R, values for five
and precipitated without fractionation over a distance sample sets representing various larger scale rock vol-
that is small relative to the length of the flow path of umes within the Biscuit Basin flow. The only average
thermal water through the Biscuit Basin flow. However, Ri values shown are for elements that differ significantly
the REE pattern of Y7-239 veinlet celadonite is not from the RVC; these indicate the following salient fea-
quite parallel to that of the obsidian, but is relatively tures regarding the distances of element migration:
enriched in the heavy REE. This may indicate that the
REE were fractionated during transport or precipita- (1) The only analyzed elements that have migrated
tion. The Y7-239 veinlet ,&cristobalite separate has a significantly on the scale of the entire ffow breccia in
REE pattern parallel to that of Y7-239 veinlet cela- Y-7 (-20 m) are Mg, Li, Sb, and U. Mg was apparently
donite (Fig. 4). This most likely indicates the presence released mostly by destruction of pyroxenes and was
of about 3% of celadonite contaminant in the /&is- not fully retained in secondary phases, resulting in a
tobalite separate, consistent with the similar relative net loss to infiltrating thermal water that transported
concentrations of other relatively immobile elements the Mg outside of this rock volume. Concentrations of
in the two separates. The REE pattern of Y7-209 veinlet Li, Sb, and U have all increased within the flow breccia
calcite (Fig. 4) is distinct in that it has a steep positive matrix, and must therefore have had a source outside
slope. This possibly reflects the relative concentrations of this rock volume.
of REE-carbonate complexes in solution that may have (2) The flow breccia matrix in the 10 m section of
been destabilized and precipitated with calcite during core below Y7-210 is more thoroughly altered than
boiling. that above Y7-2 10, as indicated by the alteration index
Relatively mobile elements. The relatively mobile (Table 1). Significant migration of five elements, in
elements were identified by their differences in nor- addition to those four noted for the entire flow breccia,
mahzed concentration relative to the RVC. These dif- has occurred on the scale of this rock volume (- IO
ferences imply migration distances at least on the order m): Si, Na, and Ba are high; Mn and Cs are low.
of whole rock sample size (2-5 cm). Further insight (3) Only five elements (Li, Na, Mn. Sb, and Cs)
exhibit significant migration on the scale of the entire
Biscuit Basin flow in Y-8 (-98 m), despite the generally
IO00 large deviations (relative to the RVC) observed among
all of the relatively mobile elements in single whole
YE 195 CUADONITE rock samples. The average Ri values for Na and Mn
Y8 331 RED CHALCEDONY are low, and those for Li, Sb, and Cs are very high.
(4) The flow breccia in Y-8 has a thickness of about
10 m. Excluding sample YS-182D because of the un-
certainty associated with its alteration conditions, this
sample set exhibits significant average compositional
changes for all relatively mobile elements except Sr
and Ba. The average Ri values for Mg, K, Mn, and Rb
are low, while those for Ca, Na. Li. Sb. Cs, and U are
high.
(5) Although single whole rock samples of the
pumiceous tuff in Y-8 (thickness = 88 m) can deviate
significantly from the RVC for all but five of the 29
elements analyzed, only five of these elements have
average Ri values that differ significantly from the RVC
when the entire interval of tuff is considered. These
are Li, Na, Mn, Sb, and Cs. Na and Mn are low, while
Li, Sb, and Cs are high.
Although there is significant redistribution of many
elements on at least the scale of whole rock samples
\ (2-5 cm), most of these elements do not exhibit a net
I d
Sm f?u
I
Tb Yb Lu
flux on the scale of the entire volume of rock investi-
gated. Only two elements appear to have been input
ubiquitously in the flow in Y-7 and Y-8: Li and Sb. In
FIG. 4. Chondrite-non-a&cd (HASKIN et al.. 1968) REE
diagram for average Y-7 obsidian and various veinlet precip- Y-7, there appear also to have been a gain of LJ and a
itates from Y-7 and Y-8. Error bars are shown where 20 an- loss of Mg, and in Y-8, there appear also to have been
alytical uncertainty exceeds size ofdata points. a gain of Cs and losses of Na and Mn.
Yellowstone rhyolite alteration 1629

Table 8. Average normalized concentrations Table 9. Average normalized mass changes (AMi) for relatively
(Ri) for relatively mbile mbile elements*
elercents*

Element 1 2 3 4 5
Element 1 2 3 4 5

l4g (g/kg) -0.48 -0.49 -0.73


Hg 0.81 0.81 0.71 Ca (g/kg) 2.56
Ca 1.23 Ha (g/kg) 4.64 -2.06 6.47 -5.74
Na 1.20 0.91 1.28 0.75 K (g/kg1 -17.2
K 0.51 Li (&kg) 21 46 79
Li 2.32 2.07 5.28 3.86 5.91 14n (&kg) -::o -!880 -130 -210
Nn 0.81 0.72 0.80 0.68 Rb (nglkg) -38.3
Rb 0.77 Sb (mglkg) 0.09 0.10 0.82 0. 20 1.10
Sb 1.71 1.78 7.34 2.51 9.49 Cs (mglkg) -1.68 91.5 219 43.3
CS 0.67 20.0 43.6 8a (w/kg1 280
9.45
8a 1.31 U (&kg1 6.1 5.2 1.D
U 2.16 2.01 1.19
'Colum headings as in Table 8.
*Only values that differ significantly from
(C,/CTi )R”C at-2 shown

1 - All Y-7 samples other than RVC Ln=lO)


2 - V-7 samples from depths greater than g/kg. This is much greater than AM; determined for
210 ft (n=6) other major elements in Y-7 and Y-8. In the case of
3 - All Y-8 samples except YB-182D (n=131
4 - Y-8 vitroptiyre and matrix samples (n=4) sample Y833lT, AMsi - 414 g/kg, or AMsio, - 887
5 - Y-8 pumiceous tuff samples (n-9)
g/kg. The average porosity of the pumiceous tuff in Y-
8 is 39% (KEITH et al., 1978), thus the porosity in Y8-
331T has essentially been sealed with silica. Use of
Quantification of compositional changes other relatively immobile elements results in similar
estimates for AMsi.
Compositional changes in rock due to hydrothermal
alteration must be quantified to provide constraints Water/rock ratio
for accurate models of mass transfer in hydrothermal
Water/rock ratios in hydrothermal systems are often
systems. These data may also be used to interpret vari-
estimated by chemical or isotopic mass balance or heat
ations in thermal spring compositions in a given area
balance equations ranging from simple to complex
in terms of flow direction or aquifer mineralogy. The
(TAYLOR, 1974;CATHLES,1983). Most estimates of
normalized change in concentration of element i (AC)
water/rock ratios represent minimum values because
relative to its initial concentration (ciRVC)is defined as:
potential exchange of mass or heat between water and
AC, = (RI - l)~iRvc (2) rock may not actually proceed, because of kinetic fac-
tors or because local equilibrium has been reached.
recalling from Eqn. (1) that Ri = (Ci/CTJmmple/(Ci/
Table 10 shows estimates of minimum water/rock mass
CTi)RVC .
ratios derived by dividing average values of AMi (Table
The normalized change in mass of element i (AM,)
9) by (ci)wam(Table 6). These are uncertain to the extent
relative to its initial mass in a given mass of RVC (m),
that advective transport is assumed and thermal water
is defined as:
composition and mechanisms of element redistribution
AMi = mACi. (3) may have evolved with time, but indicate high ratios
These simple equations quantify the mass scale of ranging from 4.3 up to 23,000. Estimates tend to be
compositional change for an arbitrary volume of rock, higher for the less soluble elements, suggesting that
provided knowledge of the initial concentrations of the these may be more dependable indicators of minimum
elements, including at least one relatively immobile water/rock ratios because more water is required for
element (e.g., Ti) that can be used for a normalization their transport. Actual water/rock ratios required for
reference. advective transport of strongly sorbing species would
Using the average Ri values for the relatively mobile
elements listed in Table 8, a corresponding list of AMi
values (for m = 1 kg) has been constructed and is shown Table ID. Average minimum water/rock wss ratios*
in Table 9. The scale of compositional change for the
major alkalis and alkaline earths is on the order of OS- Element 1 2 3 4 5

17 g/kg (9-49%), whereas that for minor and trace ele-


M 6700 6800 1800
ments is on the order of 0.1-280 mg/kg (19-4260%). Ca 300
Although whole rock Si concentrations were not de- Na 11 5.6 18 16
K 960
termined, AMsi can be approximated from the dilution Li 22 18 85 60 100
Mn 6100 1800 1300 2100
of Ti (for m = 1 kg): Rb 380
Sb 4.3 4.8 27 6.7 33
CS 450 750 1700 330
(CTi)RVC
AMsi m -----1 m. Ba 23000
(4) U 6600 5700 14wo
( (CTihample 1

*Estimates based on data In Tables 6 and 9. Colum


Estimated AMsi for Y-7 matrix below 210 ft is 53 headings as in Table 8.
1630 N. C. Sturchio, K. Muehlenhachs and M. G. Seitz

probably be orders of magnitude higher (NERETNIEKS, Oxygen isotope ratios in obsidian-replacing smectite
1980). and in veinlet celadonite and &cristobalite indicate
Plow rates of hot springs in Biscuit Basin are gen- that the thermal waters from which these phases pre-
erally in the range of O-20 L/min (THOMPSONand cipitated underwent significant “0 enrichment relative
YADAV, 1979). A hot spring discharging 2 L/min can to present thermal water, if precipitation occurred at
discharge approximately 1O6L/yr. If this hot spring is temperatures similar to present temperatures at depth.
representative of the flux of water through lo6 kg of This suggests that a low effective water/rock ratio, pos-
underlying rock, then the water/rock ratio for that vol- sibly promoted by transient exposure of the water to
ume of rock increases at an annual rate of - 1 kg water/ fresh unexchanged obsidian, was associated with their
kg rock. If hot spring activity in Biscuit Basin has been precipitation.
continuous for thousands of years, then the high esti- (6) A very large water/rock ratio, although not re-
mated water/rock ratios in Table 10 are realistic. quired to explain oxygen isotope ratios, is necessary to
explain other data. If elements were redistributed pri-
CONCLUSIONS marily by advective transport in thermal water having
composition similar to present thermal water, the
(1) The Biscuit Basin flow in Y-7 and Y-8 was com- minimum integrated water/rock mass ratio was on the
positionally relatively homogeneous when emplaced, order of 103-lo4 in both Y-7 and Y-8.
with respect to the elements analyzed in this study.
Element mdistribution during hydrothermal alteration Acknowledgements-We thank Terry Keith (U.S. Geological
has been evaluated by comparing compositions of al- Survey, Menlo Park) for suggesting Y-7 and Y-8 drill cores
for this study, for assistance in obtaining samples, and for
tered samples to a reference composite of nearly un- providing much related information. We have also profited
altered samples. through discussions with J. M. Thompson, R. Christiansen.
(2) Elements that remained relatively immobile R. Fournier, and D. White. Members of the Analytical
during hydrothermal alteration are Ti, Al, Fe, Sc, Co, Chemistry Laboratory at Argonne assisted in obtaining the
data reported herein: E. Huff (ICP-AES), B. Tani (XRD), I.
Y, Zr, REE, Hf, Ta, and Th. These elements were gen-
Fox (CHN), and K. Jensen (gravimetric SiOz). R. Korotev
erally redistributed over a distance of at least lo-100 and M. Lindstrom assisted with INAA analyses, and D.
pm but less than 2-5 cm. The fraction of these elements Kremser assisted with electron microprobe operation, at
that was mobile during alteration is associated mostly Washington University. Mrs. E. Toth assisted with oxygen
with clays. isotope analyses at the University of Alberta. Helpful reviews
of early drafts of this paper were provided by R. Couture and
(3) Elements that were relatively mobile during hy- T. A. Abrajano, Jr. Lu Jensen endured numerous revisions
drothermal alteration are Si, Mg, Ca, Na, K, Li, Mn, of the manuscript with aplomb. Work supported by U.S. De-
Rb, Sr, Sb, Cs, Ba, and U. These elements migrated partment of Energy, O&e- of Basic Energy Sciences. under
over distances ranging from at least 2-5 cm to more Contract W-3 I- 109-ENG-38.
than lo-100 m. Estimated changes in normalized mass
range from about OS-17 g/kg (9-49%) for major ele- Editorial handling: J. R. O’Neil
ments to 0.1-280 mg/kg (19-4260%) for minor and
trace elements, and are generally larger in Y-8 because REFERENCES
of proximity to discharge conduits and the greater ex-
BARGARK. E. and BEESONM. H. (198 1) Hydrothermal al-
tent of alteration. Addition of silica, estimated from teration in research drill hole Y-2, Lower Geyser Basin,
dilution of Ti and other relatively immobile elements, Yellowstone National Park, Wyoming. Amer. Mineral. 66,
ranges up to - 160 B/kg in silicified flow breccia matrix 473-490.
and -890 g/kg in silicified pumiceous tuff. Some of BARGARK. E. and MUFFLERL. J. P. (1982) Hydrothermal
alteration in research drill hole Y-l I from a vapor-domi-
the relatively mobile elements are mainly associated
nated geothermal system at Mud Volcano, Yellowstone
with zeolites in altered samples (Na, Cs, Ca, Sr, and National Park, Wyoming. Wyoming Geol. Assoc. Guide-
Ba), but Li, Mg, K, Rb, Mn, and Sb are more strongly book, Thirty-Third Annual Field Conf, pp. 139- I5 1.
associated with clays, K and Rb are also associated BARGARK. E. and BEE.%N M. H. (1984) Hydrothermal al-
with adularia, Sb is also associated with silica, and U teration in research drill hole Y-6, Upper Firehole River,
Yellowstone National Park, Wyoming. U.S. Geol. Sun;.
is associated mostly with titanomagnetite surfaces. Pro/: Paper 1054-B.
(4) Elements that consistently tended to be gained BARGARK. E., BEESONM. H., FOURNIERR. 0. and MUFF’LER
during hydrothermal alteration are Li and Sb in both L. J. P. (1973) Present-day deposition of lepiddite from
Y-7 and Y-8, U in Y-7, and Cs in Y-8; those lost are thermal waters in Yellowstone National Park. Amer. Min-
eral. 58,901~904.
Mg in Y-7 and Na and Mn in Y-8. Those elements
BEE+SONM. H. and BARGARK. E. (1984) Major and trace
gained were probably derived from rock at greater element analyses of drill cores from thermal-areas in Yel-
depth; those lost were probably discharged by thermal lowstone National Park, Wyoming. U.S. Geol. Sure. Open-
springs. File Report 84-373.
(5) Oxygen isotope exchange between thermal water CATHLES L. M. (1983) An analysis of the hydrothermal system
responsible for massive sulfide denosition in the Hokuroku
and primary or secondary phases did not reach equi- Basin of Japan. In The Kuroko aid Related Volcunogenic
librium except for calcite. Minor I80 depletion oc- Massive Sulfide Deposits (eds. H. OHMOTOand B. J. SKIN-
curred in Y-8 plagioclase (- 1%~)and obsidian (-2L). NER),pp. 439-487. Econ. Geol. Monograph 3.
Yellowstone rhyolite alteration 1631

CHRISTIANSEN
R. L. (1984) Yellowstone magmatic evolution: National Park, Wyoming. U.S. Geol. Sunt. Pro&Paper 1054
Its bearing on understanding large-volume explosive vol- A, Al-A26.
canism. In Explosive Volcanism:Inception,Evolution,and KEITH T. E. C., THOMPSONJ. M. and MAYS R. E. (1983)
Hazards, pp. 84-95, Nat. Acad. Sci., Washington. Selective concentration of cesium in analcime during hy-
CLAYTONR. N. and MAYEDA T. K ( 1963)The use of bromine drothermal alteration, Yellowstone National Park, Wyo-
pentalluoride in the extraction of oxygen from oxides and ming. Geochim. Cosmochim. Acta 47, 795-804.
silicates for isotopic analyses. Geochim. Cosmochim. Acta KENNEDYB. M., LYNCHM. A., REYNOLDSJ. H. and SMITH
21,43-52. S. P. (1985) Intensive sampling of noble gases in 5uids at
CLAYTON R. N., O'NEIL J. R. and MAYEDA T. K. (1972) Yellowstone: I. Early overview of the data; regional patterns.
Oxygen isotope exchange between quartz and water. J. Geochim. Cosmoschim.Acta 49, 1251-1262.
Geophys.Res. 77, 3057-3067. LEEMANW. P., DOE B. R. and WHELANJ. (1977) Radiogenic
CONTINENTAL SCIENTIFICDRILLINGCOMMITTEE(1984) A and stable isotope studies of hot-spring deposits in Yellow-
NationalDrillingProgram to Study the RootsofActiveHy- stone National Park and their genetic implications. Geo-
drothermalSystemsRelatedto Young MagmaticIntrusions. them. J. l&65-74.
National Academy Press, Washington, D.C., 43~. LINDSTROMD. J. and KOROTEVR. L. (1982) TEABAGS:
CRAIGH. (1963) The isotopic geochemistry of water and car- Computer programs for instrumental neutron activation
bon in geothermal areas. In Nuclear Geologyon Geothermal analysis. J. RadioanalyticalChem. 70, No. l-2,439-458.
Areas (ed. E. TONGIORGI),pp. 17-53. Consiglio nationale NERETNIEKSI. (1980) Diffusion in the rock matrix: An im-
delle richerche, Piss. portant factor in radionuclide retardation? J. Geophys.Res.
DICKINA. P. (198 1) Hydrothermal leaching of rhyolite glass 85,4379-4397.
in the environment has implications for nuclear waste dis- NOBLE D. C. (1967) Sodium, potassium, and ferrous iron
posal. Nature 294, 342-347. contents of some secondarily hydrated natural silicic glasses.
EATONG. P., CHRISTIANSEN R. L., IYER H. M., P~-rr A. M., Amer. Mineral. 52,280-286.
MABEYD. R., BLNAKJR., H. R., ZEITZ I. and GETTINGS OHMOTOH. and RYE R. 0. (1979) Isotopes of sulfur and
M. E. (1975) Magma beneath Yellowstone National Park. carbon. In Geochemistryof HydrothermalOre Deposits(ed.
Science 188, 787-796. H. BARNES),pp. 509-567, Wiley-Interscience.
ESLINGERE. V. and SAVINS. M. (1973) Mineralogy and ox- O’NEILJ. R. and TAYLORJr. H. P. (1967) The oxygen isotope
ygen isotope geochemistry of the hydrothermally altered and cation exchange chemistry of feldspars. Amer. Mineral.
rocks of the Ohaki-Broadlands geothermal area. Amer. J. 52, 1414-1437.
Sci. 213, 240-267. O’NEIL J. R., CLAYTONR. N. and MAYEDAT. K. (1969)
Oxygen isotope fractionation in divalent metal carbonates.
GILETTIB. J., SEMETM. P. and YUND R. A. (1978) Studies
J. Chem. Phys. 51, 5547-5558.
in diffusion. III. Oxygen in feldspars, an ion microprobe
determination. Geochim. Cosmoschim. Acta 42,45-47. SMITHR. B. and BRAILEL. W. (1984) Crustal structure and
evolution of explosive silicic volcanic system at Yellowstone.
HASKINL., HASKINM., FREY F. and WILDEMANI. (1968)
In Explosive Volcanism:Inception,Evolution,and Hazards,
Relative and absolute abundances of the rare earths. In
National Academy of Sciences.
Origin and Distributionof the Elements (ed. L. AHRENS),
STURCHION. C., LEWISC. H. III and SEITZ M. G. (1985)
pp. 889-9 12. Pergamon Press, 1178~.
Uranium series disequilibrium in Y-7 drill core: Implica-
HENLEYR. W. and ELLISA. J. (1983) Geothermal systems tions for U mobility and the age of thermal activity in Biscuit
ancient and modem: A geochemical review. Earth Sci. Rev. Basin, Yellowstone National Park, Wyoming. EOS, Trans.
19, l-50. Amer. Geophys. Union 66-18,420.
HILDRETHW., CHRISTIANSEN R. L. and O’NEIL J. R. (1984) TAYLORJR. H. P. (1974) The application of oxygen and hy-
Catastrophic isotopic modification of rhyolitic magma at drogen isotope studies to problems of hydrothermal alter-
times of caldera subsidence, Yellowstone Plateau Volcanic ation and ore deposition. Econ. Geol. 69, 843-883.
Field. J. Geophys. Res. 89, no. BlO, 8339-8369. THOMPSONJ. M. and YADAVS. (1979) Chemical analyses of
HONDA S. and MUFFLER L. J. P. (1970) Hydrothermal al- waters from geysers, hot springs, and pools in Yellowstone
teration in core from research drill hole Y- 1, Upper Geyser National Park, Wyoming, from 1974 to 1978. U.S. Geol.
Basin, Yellowstone National Park, Wyoming. Amer. Min- Surv. Open-File Report 79-704, 49~.
eral. 55, 1714-1737. TRUESDELLA. H. (1966) Ion-exchange constants of natural
JACOBSJ. W., KOROTEVR. L., BLANCHARD D. P. and HASP glasses by the electrode method. Amer. Mineral. 51, 1 IO-
L. A. (1977) A well-tested procedure for instrumental neu- 122.
tron activation analysis of silicate rocks and minerals. J. U.S. GEOLOGICALSURVEY(1972) Geologic map of Yellow-
RadioanalyticalChem. 40,93-l 14. stone National Park. U.S. Geol. Surv. Misc. Geol. Znv.Map
JEZEK P. A. and NOBLED. C. (1978) Natural hydration and I-711.
ion exchange of obsidian: An electron microprobe study. WHITE D. E. (1981) Active geothermal systems and hydro-
Amer. Mineral. 63, 266-273. thermal ore deposits. Econ. Geol. 75, 392-423.
KEITH T. E. C. and MUFFLERL. J. P. (1978) Minerals pro- WHITED. E., FOURNIERR. O., MUFFLERL. J. P. and TRUES-
duced during cooling and hydrothermal alteration of ash- DELLA. H. (1975) Physical results of research drilling in
flow tuff from Yellowstone drill hole Y-5. J. Volcanol.Geo- thermal areas of Yellowstone National Park, Wyoming. U.S.
therm. Res. 3, 373-402. Geol. Surv. Prof Paper 892, 70~.
KEITH T. E. C., WHITE D. E. and BEESONM. H. (1978) Hy- YEH H. W. and SAVINS. M. (1977) The mechanism ofburial
drothermal alteration and self-sealing in Y-7 and Y-8 drill metamorphism of argillaceous sediments. 3. Oxygen iso-
holes in northern part of Upper Geyser Basin, Yellowstone topic evidence. Geol. Sot. Amer. Bull. 88, 132 l-1330.

S-ar putea să vă placă și