JOURNAL Or GF.OPHYSICAL RESEARCH, VOL. 93, NO. B6, PAGES 6103-6107, JUNE 10, 1988

LIGHI' UiliOPifil.E ELEMENT (BORON, UI1iIUM) ABUNDANCES IN TiiE V ALLES CALDERA, NEW MEXICO, VC-1 CORE HOLE

Michael D. Higgins 1

Depanmcnt of Geology, McMaster University. Hamilton, Ontario, Canada

~- Boron and lithium abundancc distributions are vcry diffcrent in the sedimcnrary and volcanic sections of the VC-1 corc bote, Valles caldcra, New Mexico. ln the sedimcnrary section the widc variations in boron and lithium abundanccs probably reflcct both diagcncsis and introduction of thcsc clcmcnts in early hydrolhcrmal fluids. Late, higho'-icmpcratun: fluids appcar ro have lcachcd boron and lithium from somc rocks adjacent lO fault zones and aquifers. The com:lation of boron with lithium suggcslS lhat lhcsc clements behavc in similar ways during diagcncsis and hydrothcrmal altcration. The small variation of boron and lithium abundanccs in mœt of the vitrcous volcanic rocks from the uppcr part of the corc implics lhat their source was relativcly homogeneous. Following primary dcvitrification, boron appcars lO have bccn stable, but lithium was mobile. Boron may have bccn stabilizcd by the dcvelopmcnt of boron-bcaring submicl'OSCQpic phcnocrystS in the Javas, or boron may be lcss soluble lhan lithium in relalivcly cool groundwaJ.Cr.

Inlroduction

The geochcmistry of the light, lithophilc clcmcnts lithium and boron has bccn somcwhat ncglcclCd until rcccnlly. in part duc IO

difficultics in the analysis of boron at the low lcvcls commonly encountcred in many rocks. ln this study an unusual nuclear icchniquc, Prompt Gamma Neutron Activation, bas bccn uscd for the analysis of boron, white lithium has bcen analyzed by atomic absorbtion spcctroscopy.

V alles caldcra has bccn a major wget of the U .S. continental scicntific drilling program. and core holc VC-1 is the first bote complclCd undcr this program (Figun: 1). One of the major objectives of VC-1 was lO pencuaic a high-icmperature hydrothennal outflow plume ncar its source [Goff et al., 1986]. The core hole thus providcs a unique opportunity ro examine the relationship of boron and lithium in diffcrent rock types in an active hydrothcrmal environmcnL

The rop 330 m of the holc penctratcs intracaldera flows, tuffs. and interstratified scdimcnts deposited in the moat of the caldera (Figun: 2). Ali the volcanic rocks penctralCd by VC-1 are lcss lhan 0.6 m.y. old [Geissman, this issue). The volcanic rocks are relativcly fresh, except for patchy dcvilrification and minor, low-t.emperawre alteration in the middlc VC-1 tuff. associalCd with a fault zone. High-temperature h ydrothcnnal alteration is abscnL

Two clay-rich units occur ncar the centcr of the core holc: a volcaniclastic trcccia at the bottom of the volcanic scqucnce and in the underlying Permian Abo formation (Figun: 2). Thesc zones appcar ro associalCd with major fractun:s.

ln contrast to the volcanic rocks, the underlying scdimcnts show abondant evidcncc of faulting and shcaring. They have bccn lightly ro scverely altcred and mincralized by high-1Cmpera1ure hydrothcrmal fluids, cspccially along fractures.

Sarnplcs wcre talccn at more or lcss regular inicrvals throughout the corc, except whcre a unit would have bccn misscd. No aticmpt was made lO sarnplc altcred zones more intensively lhan othcr parts of the section.

Now at Sciences de la icrre. Universiic du Qucbcc a Chicoutimi, Chicoutimi, Qucbcc, Canada.

Copyright 1988 by the Amcrican Gcophysical Union.

Paper numbcr 7B7046 0148-0227 /8MXl7B-7046$05.00.. .

Mcdlods

Boron was analyzcd using Prompt Gamma Neutron Activation. This mcthod utilizes the nuclcar reactions

10s(n,alpha} 7Lt 7Li •(gamma) 7Li

Two grams of powdercd sarnplc arc ~kcd in a 1-cm-diametcr tcflon tube and i.rradialCd in a flux of 2x 10 nculroflS per second at the McMastcr Nuclcar Rcactor. The characteristic gamma rays (478 kcV) cmitlCd by thcsc reactions during the irradiation are counlCd using an inlrinsic Ge detector. Boron was detcnnined with rcferencc National Bureau of Standards standard 1571, for which boron • 33 ppm. Prccision as deicrmined by rcplicaic analyses of samples is estimated to be ±3%. U.S.Gcological Survey standard rock RGM-1 gives a value of boron = 27.8 ppm, which is closcly comparable with that obtaincd by othcr rescarchcrs [Higgins, 1984). Fa- furthcr dctails of the mcthod sec Higgins Cl al., [1984].

Lithium was dctermincd using atomic absorption Spectroscopy with refercncc IO chemical standards. One gram of cach sarnple was dissolved in HF+HN03+HCI04 , dilutcd IO 100 mL and run in an acctylcnc-air flamc. Prccision is estimated IO be ±5%.

Rcsults and Discussion

Preca.tdera Sedjmeriwv Rocks

1llc lowcr 524 m of the core comprises principally Palacozoic scdimcntary rocks: The Pennsylvanian Madera limestonc and Sandia formations; and the Pcrmi.an Abo formation (Figun: 2) and possibly the Mississippi.an Esprito Santo and Log springs formations [Hulen and Niclson, this issue] . A thin inicrval of Precambrain gneiss was also intcrscclCd by the ca-e holc.

Diagenctic aheration is pervasive throughout the scdimentary section. Smcctiic-illiic is an important diagcnetic minerai [Keith, this issue) especially as it is known lO host boron and lithiwn (reviewed by Bergeron, [1985]). The füst stage of hydrothermal alicration was confined to fractures and inicrgranular space in the more permcablc laycrs where chloriic, calcite, somc pyriic and somc additional smectilC· illitc wcre dcposilCd from fluids at approximatcly 200°C [Keith, this issue). L.aicr stages of alicration wcrc confincd to zones of high permeability in the Sandia fonnation and the upper part of the Madera limcstonc. The fluids wcre at 275°-300°C [Keith, this issue] and did not dcposit any minerais liltely to host boron and lithium.

Within the scdimenrary section, bcron and lithium abondances vary widcly (0.7-56 ppm and 6.5-249 ppm respcctively, Table 1 and Figure 3). Boron abundances are positively corrclaled to lithium abundances with a correlation coefficient of 0.60 (Figure 4). A correlation coefficient for a group of 13 sarnplcs greater than 0.53 is considered significanL The correlation bctwecn boron and lithium abundances suggcsts that their bchavior is geochemically similar in low- and mcdium-icmperatun: environmcnts. Lia.Je wori: bas bccn donc on bcron and lithium correlations in scdimcntary rocks elscwhcre. Howcvcr, thesc elemcnts are well correlated in e:ittremely altcred scafloor basalts [Thompson and Melson, 1970]. ln contrast, Bergeron [1985] found no correlation in basalts with lowcr degrccs of altcration.

Thcre is no simple relationship betwcen depth and boron, lithium abundanccs, as is lO be expeclCd in a complcx system. The most likely host for boron and lithium is smecti1c-illi1c, which is prescnt

6103

6104 Higgins: Boron and Lithium in YC-1 C«e Ho1e

- ....

c

' ' ' ' ' '

+ N

4 Km

Fig. 1. Location map oC Valles caldera and YC-1 core hole siie.

throughout the scdimcncary section. Samplcs low in boron and lithium at 546 m and 307.8 mare adjacent to an aquifcr at 539 m [Goff et al., 1986] and fault zone ab318 m [Goff et al., 1986]. Circulaùng high tempcraturc lluids (275 -300°C) may have parùally broken down the

clay mincrals and lcachcd boron and lithium from the system. lt is diCficult to compare the overall abundances oC boron and

lithium found in this study with that similar scdimcncary rocks elscwhcrc bccausc oC the scarcity oC data and Jack oC control over its quality. Howcvcr. a literaturc survcy [Harder, 1974; Steinncs et al. , 1974] indicaies that boron abundanccs arc similar to thosc ofnonmarinc

scdimcnts elscwhcre, whcreas lithium appcars to be enrichcd (Table 1 and Figure 4). The mobility oC lithium alonc in this section is not consisrent with the obscrvcd cortelation of boron and lithium: thcrcforc it is likcly lhat the "bcnch man• data are incorrect. Boron was formcrly analyzcd by cmission spcctroscopy. Work in our laboratory suggcsts that this technique tcndcd to overestimatc boron abundances. If boron abundanccs in typical scdimcncary rocks arc indccd lowcr, then both boron and lithium must have bcccn introduccd by circulating hydrothcrmal lluids.

The minimum ovcrall lluid/rock ratio Cor the scdimcncary section rnay be calculaied as follows: The conccnttation of boron and lithium in the lluids is assumcd to be the samc as that of prescnt~y hydrothcnnal lluids in the area (B = 15 ppm, Li '"' 25 ppm [Keith, 1986]). The original boron and lithium content of the rocks is assumcd to be the sarnc as that oC the continental crust (B = 9 .2 ppm, Li = 22 ppm) as estirnatcd from the composition oC the Canadian shicld by Shaw et al., [1986] . Thcsc data werc uscd in prcfcrencc to the abundanccs in scdimcntary rocks as the boron data was known to be rcliablc. The mcan boron and lithium conteai,.of the scdimcntary rocks in this

section, 22.4 and 107 ppm (Table 1) thcn givc minimum lluid/ro ratios oC 0.9 and 3.4 respcctivcly. lbcsc values are much lower th thosc indicated for the Yellowstone system [Sturchio et al., 198i lbcy are probably underestirnatcs as it is unrcasonablc to assume û soluble clcmcnts such u boron and lithium could be entirc precipitated out oC hydrothcnnal lluids into a rock.

Posglde@ Volcanic Rocks

The scdimcncary formations arc overlain by a thin volcanic brecc the clay-rich parts of which arc rich in boron and lithium. In 1

remaindcr of the volcanic scquencc, boron and lithium show a mu smallcr range of abundanccs as comparcd with the precaldc scdimcncary rocks (Figures 3 and 5). Four samplcs from the Mid( VC-1 tuff, which ovcrlics the brcccia, have almost constant bor abundanccs (10.0-11.4 ppm). In contrast. lithium is more abundant the dcviaificd sarnplcs (80, 81 ppm) lhan the glassy samplcs (61, ppm, Figures 5 and 6). Within the ncxt threc ovcrlying formatic (Upper VC-1 tuff, YC-1 rhyolite, and Banlcship Rock lllff), boron a lithium show somc variation, but the nwnbcn of samplcs analyz Crom cach unit arc insufficient to indicaie any trends.

The uppcrmost eight sarnplcs of the core arc Crom the Ban Bonito obsidian, the youngcst volcanic unit in the Valles caldcra (0. Ma). Yitrcous sarnplcs show relatively littlc variation in boron a lithium (9 .9-13.0 ppm and 59-68 ppm, respcctivcly), which may accounted for by magmatic proccsscs, such as fractional cryslalliz.atic Dcviaificd samplcs have similar boron contents to the vill"COUS samp, but much lowcr lithium abundanccs (20-42 ppm). This behavi

Higins: Boron and Lithium in YC-1 Core Hole 6105

OEPTH

VC·1 STRATIGRAPHY

ABOFM

MADERA ~IMESTONE

SANOIA FM

- BRECCIA ZONE OF MIXEO PRECAMllRIAN GNEISS ANO SANOIA FM

Fig. 2. Stratigraphie section of core hale VC-1.

10 •

-101

s Q., Q.,

- 10

.....

...J

Volconics Il

Sediments

10·•+,-.,...,....,...,.......,.......,.......,.,.,,...,......,......,...,....,...,.......,.......,.........,.,., ........ ..,......,...,.... ...... ...,.......,........ ...... 0 200 400 8~0 800

Depth (mJ Fig. 3. Variation of boron (bottom linc, asterisks) and lithium (!Op line, circles) with depth in the YC-1 hole. In the volcanic section, solid symbols indicate devittified samples, and open symbols are vitreous samples. There is a volcanic brcccia unit in between the volcanic and sedimentary sections of the core hole.

, ..

10.

-8 l?-to • Q., -

• e ee .. Shaln

fiJJ '" Sandstones

'" Cont. crust

• Carbonates

1 -r-....... -.-.............. ..--.-..................... .--............................. ..--....... "T"" ...........

10 -• 1 10 ~o • 10 •

Boron (ppmJ Fig. 4. Boron and lithium variations in the 9edimentary rocks.

100

- 80 8 Q.,

Q., 80 -8 40 =' .....

..c: == 20 ...J

4 Boron

0 0

* .. ~ •* *

* * *

(ppm) 12 18

Fig. S. Boron and lithium variations in the volcanic rocks (excluding the volcanic breccia). Asterisk.s are samples from the Banco bonito fonnation, Diamond.s are from the Middle VC-1 tuff, and lriangles are olher samples. Solid symbols are vilreous samples, open symbols are deviaified.

0.8

50 100 150 2~0

Depth (mJ 250 300

Fig. 6. Variation of boron/lilhium ratios wilh deplh in the volcanic section of the con: hole. Open symbols are vitrcous, solid symbols are devittified.

6106 Higgins: Boron and Lithium in VC-1 Core Ho1e

TABLE l. Analytic.a.I 0ara.. VC-1 Samples

Unit Samplc

Banco Benito obsidian 17-R~l.5

Banco Benito obsidian 35-R-119.8 Banco Benito obsidian 49-R-180.5 Banco Benito obsidian 62-R-241 Banco Bonito obsidian 75-R-300 Banco Bonito obsidian 87-2-1 Banco Benito obsidian 98-R-421.S Banco Benito obsidian l 10-R-481.S Battlcship Rock tuff ll~R-510

VC-1 rhyolite 120-R-528 VC-1 rhyolite 127-R-586 Upper VC-1 tuff l~R~7

Middle VC-1 tuff 148-R-770 Middle VC-1 tuff 154-10 Middle VC-1 tuff 160-R-891 Middle VC-1 tuff l~R-953.5

Mean (volcanics)

Volcanic brcccia 173-R-1010 Volcanic brcccia 181-R-1070

Abo fonnation 187-R-1125 Abo formation 194-R-1188 Abo formation 211-R-1341 Madera limcstonc 228-R-JSOl Madera limcstonc 245-R-1647 Madera limcstone 265-SB-l Madera limcstonc 277-6-2 Madcr.> lirncstone 291-R-2101 Madcra lirncstonc ~R-2250 Madera limcslOOC 322-40 Madera limcstonc 338-R-2550 Sandia formation 353-R-2700 Sandia formation 365-R-2800 Mean (scdimcn!S)

Argill.accous scdimen!S and shalcs Sand.stoncs LimC5loncs and dolomilCS Continental crust

constraslS with that found in the Middle VC-1 tuff whcrc dcviaificd samplcs are enrichcd in lithium.

The differcnccs in baron and lithium abundanccs bctwecn dcvitrificd and vitteous samplcs in both the Banco Bonito obsidian and Middle VC-1 tuff indicate that lithium may be mobilizcd by shallow groundwatcr at tempcraturcs lcss than 100°C following primary dcvitrification but that boron appcars to be stable undcr thcse conditions. Possibly, lithium was removed from the Banco Bonito obsidian and addcd to the Middle VC-1 tuff by t.hc action of circulating low-tempcrature groundwater in t.hc caldcra moat zone. Similarly, in the Y cllowstonc caldcra, lithium is one of t.hc most mobile clcmcn!S in t.hc altcrcd rhyolites [Sturchio et al., 1986].

The behavior of boron and lithium in this system contrasts with that of t.hc rhyolilCS in the Inyo volcanic chain, whcrc both boron and lithium wcrc mobilizcd following dcvitrification [Higgins, 1988). Perhaps boron has bccn immobilizcd in t.hc V alles caldcra rhyolites by the dcvclopmcnt of submicroscopic aystals of a boron-bcaring phase, such as tourmaline. Howcvcr, •bic minerais have JlO( bccn obstncd

'"

B. Li, Depth,m ppm ppm Lilhology

18.2 36.5 55.0 73.4 91.4

109.7 128.5 146.8 155.4 160.9 178.6 197.2 234.7 253.0 271.6 290.6

307.8 326.1

342.9 362.1 4()8.7 457.5 502.0 546.8 594.4 640.4 658.8 731.5 777.2 823.0 853.4

11.3 61 GWs}' rhyolite 10.1 59 Glas.sy obsidian 11.1 20 Devittified obsidian 11.1 32 Dcviaificd obsidian 11.7 41 Devittificd obsidian 11.8 30 Devittificd obsidian i3.0 68 Glas.!)' obsidian 9.9 61 Glas.!)' obsidian

10.9 60 Glas.sy wcldod tuff 8.2 60 Deviaificd rhyolite

15.8 89 Glas.!)' obsidian 2.9 49 Deviaificd wcldod tuff

10.0 61 Glas.sy wcldod tuff 10. J 65 Gl.my wcldod tuff 10.6 81 Deviaificd wcldcd tuff 11.4 80 Deviaificd wcldod tuff 10.6 57 ±2.6 ±19

1.2 88 Volcalic brcccia (clay) 28.5 267 Volanic brcccia (clay)

49.J 249 Muœmc (clay) 49.8 140 sanœaooe (clay) 55.8 148 Sanmtonc (clay) 13.2 32 Limcstonc 15.5 65 Altmd, sheaml limestonc 0.7 6 Limcs1onC

11.S 29 Limcslone 21.2 176 Limcs1onC 7.7 208 Mudstonc 4 .0 27 LimcSIOOC

22.1 57 Limcsronc 18.7 88 SillStonC 22.6 ~68 Brca:ialod shalc. 22.4 107

±18.0 ±79

130 66 Harder [1974), (B). 30 38 Sccinncs et al., 20 8 [1974), (Li)

9.2 22 Shaw et al., [ 1986]

using a microscope, XRD or alpha-track mapping technique: [Carpcnter, 1972). The diffcrenccs bctwccn t.hc behavior of boron al il t.hc lnyo volcanic chain and V alles caldcra could aJso reflect diffcrence: in t.hc chemisa-y of t.hc circulating fluids.

Boron/lithium ratios in ail t.hc vitrcous samplcs al.J vcry simila (Figure 6). This suggcsts that the magma source for ail t.hc moa volcanics pcncttared by VC-1 was homogcncous in lithium and baron as wcll as othcr clcmcn!S, as suggcsred by Kirchcr et al. [this issue].

Conclusions

Within t.hc lowcr, scdimcncary part of t.hc oorc holc t.hc correlatior betwccn boron and lithium suggcS!S that boron and lithium behavc in 1

similar way during low- and mcdium-tempcraturc geochcmica proccsses. Ovcrall boron and lithium have probably bccn inll'Oduœ( into this section by carly hydrochermal fluids. Late, highcr-ternpcrarun fluids have lcachcd boron and lithium Crom f.ault wnes and aquifen.

ln t.hc volcanic section lhete is linlc variation in baron and lithiwt

Higgins: Boroo and Lithium in VC-1 Core Hole 6107

abundances in the vitreous samples. However, lithimn appean to have been mobilized following primary devitrification. Lithium was probably uansponed in the sysiem by ci.n:ulating, cool groundwaier (Jess than 100°C) in the caldera moat zone. Boron may have been stabilized in the volcanic rocks by the developrnent of boron-bearing submicr09Copic phenocrysu. Aliematively, boron may be much less soluble lhan lithium in the relatively cool fluids lhat circu.lated in the moatzone.

Acknowledgements 1 would like to lhank my collegues in the "Bcron project" Marilyn Truscou, Mario Bergeron, and Denis Shaw, for much help. F~ Goff carefully read the manuscr:ipL This Jrojcct was funded by an NSERC (Canada) grant to D. M. Shaw. McMasier Nuclear, Isotopie and Geochem.ical S111dies group contribution 156.

Bergeron, M., Le comportement du bore dans la lithosphè'e octanique, Ph.D. thesis, McMasler Univ.,. Hamilton, OnL, Canada, 1985.

Carpenlel', B. S .• Dciennination of trace concentrations of boron and uranium in glass by the nuclear traclc t.echnique, AD31 Cbem., ~ 600-602, 1972.

Geissman, J ., Paleomagnctism and rock magnetism of quatemary volcanic rocks and laie palcozoic strata, VC-1 core hole, Valles caldera, New Mexico, with cmphasis on remagnetisation of laie Paleozoic sttata, 1 Geophys Res., this issue.

Goff, F., J. Rowley, J. N. Gardner, W. Hawlrins, S. Goff, R. Charles, D. Wacbs, L. Maasscn, and G. Heiken, Initial resullS from VC-1, first continental scientific drilling prograrn hole in V alles caldera, New Mexico. I Geopbys. Res .. il(B2}, 1742-1752, 1986.

Harder, H .• Handbo9k of Geos;hemjstry. Vol. II-1, ediled by K. A. Wcdepohl, chapter 5, Springer-Verlag, New Yorlc, 1974.

Higgins, M. D., Abundance of boron in international geochemical standards by Prompt-gamma neutron activation analysis, Geostand. News!., B.. 31-34, 1984.

Higgins, M. D., Trace element geochemistry of the lnyo volcanic chain, Califomia: Evidence for multiple magma sources. magma mixing and post-eruption loss of boron, J Vo!canol Geotherm ~.in press, 1988.

Higgins, M. D., M. G. Truscou. D. M. Shaw, M. Bergemon, G. H. Buffet, J. R. D. Copley, and W. V. Prestwick, Prompt-gamma neutron activation analysis al McMaslCr University Rcactor, in ~ and Development of Law and Medium Flux Nuc!ear Reactors Procecdjngs of an International Symposjum, AL energie Kemtech • ediled by O. K. Harling, L. Clarie and P. von der Hardt. vol. 44 supplement, pp 690-697, Karl Thiemig, Munchen, Germany, 1984.

Hulen, 1. B. and D. L. NeWon, Hydrothcrmal allel'21iœ in the Baca geothermal sysiem, Redonde dome, Valles caldera, New Mexico. L. Gr.oplm Res il. 1867-1886. 1986.

Hulen, J. B .• and D. L. Neilson, Hydrodlennal breccialioo in the Jemez fault zone, Valles caldera, New Mexico: Rcsults from CSDP core hole VC-1, J Gcoohys Res, this issue.

Keith, T. E. C., Hydrothennal alieration in the Madera limcslOllC and Sandia formation from core hole VC-1, Valles caldera, New Mexico (abstract), E°' Trans AGU 66. 1080, 1985.

Keith, T. E. C., Alteration in the Madera limestone and Sandica formation front corc hole VC-1, V alles caldera, New Mexico. 1. Geopbys Res. this is.tue.

Kircher, D .• S. Self, and J. A. Wolff, Volcanology and preliminary petrology of the El Cajeie series, Valles caldera, New Mexico. l.. Geopbys Res this issue.

Shaw, D. M., 1. J. Cramer, M. D. Higgins, and M. G. Truscou, Composition of lhe Canadian Prccambrain shield and the continental crust of the canh, The Nawrc of the Lower Continental Crust, ediled by J. B. Dawson, D. A. Carswell, 1. Hall and K. H. Wedepobl, Gcol Soc London Spcc Publ ~ 275-282, 1986.

Sieinnes, E., K. S. Heier and G. K. Billings, Handboolc of Geocbemjstry, Vol.II-!, edited by K. A. Wcdcpohl, chap. 3, Springer-Verlag, New Yodt. 1974.

Sturchio, N. C., K. Mueblenbachs and M. G. SeilZ, Element rcdisaibution dllling hydrothermal allel'ation of .rhyolile in an active geolhermal sysiem. : Yellowstone drill cores y.7 and Y-8, Ocochim Cosmochjm Acta. .m, 1619-1632, 1986.

Thompson, G .• and W. G. Melson, Boron conienlS of sapentiniies and mctabasalts in the oceanic cru.se Implications for the boron cycle in the oceans. Eanh Planet Sei l.etL, 8. 61-65, 1970.

Trucsdell, A. H. and C. J. J ani.lc, Reservoir processes and fluid origins in the Baca geolhcrmal sysiem, Valles caldera. New Mexico, 1. Geophys. Res .. il. 1817-1833, 1986.

Whiie, A. F., Chcmical and isotopie characteristics of fluids within the Baca geothermal resevoir, Valles caldera, New Mexico. J Geopbys .lill. .• il. 1855-1866. 1986.

M. D. Higgins, Sciences de la Tene, Universite du Quebcc a Chicoutimi, 555, boulavard de l'universite, Chicoutimi, Quebcc, Canada, G7J 291.

(Rcceived March 2, 1987; revised October 29. 1987; accep<ed November 23, 1987.)