Baldwin and Pearce-1982
Baldwin and Pearce-1982
Baldwin and Pearce-1982
664-674
A geochemical comparison betweenproductiveintrusions (thoseassociated with mineralization)and nonproductive intrusions (thoseunrelatedto mineralization) has been made on the basis of over 100 freshgranitoidrocksfrom the Chileanporphyrycopperbelt. The study wascarried out on the scaleof a mining district (El Salvador-Potrerillos), then on the scale of a singletraverse(El Salvador-Chilean coast),and finally on the scaleof the Andesitself. The productiveintrusions are bestcharacterized by large negativeanomalies in Y, Mn, Th, andthe heavyrare earthelements. A discriminant diagramof Y against MnO is recommended for predictingwhetheran apparentlyfreshintrusion hasassociated peripheralmineralization. The low valuesof Y and Mn in productiveintrusions can be partly explainedby the early crystallization of Y- and Mn-rich hydrousphases suchas hornblendeand by the loss of Mnrich fluidsfrom the magma.The sameelements make usefuldiscriminants in other porphyry copperprovinces, althoughthe exactposition of the discriminantboundaryvariesfrom island arcsthroughcontinental margin to continental settings.
Mason and Feiss (1979) that productive intrusions were characterizedby low ratios of A1203/(K20 a significant numerousplutonic rocksof intermediate,calc-alka- q- Na20 q- CaO). There is, nonetheless, overlap in this ratio between known productive and line character, but only a tiny proportion of theseare nonproductive intrusions. Other major elements and associated with porphyry copper deposits.In such selected trace elements (mainly Rb, Sr, and the tranarid terrain, streamsedimentand soil geochemistry havea limited application in exploring for thesede- sitionmetals)(Oladeand Fletcher,1975, 1976;Oyarposits(Coope, 1975). On the other hand, the good zun, 1975; Armbrus et al., 1977; Chaffee, 1976) have on a exposuremeans that lithogeochemical prospecting provided useful meansof defining orebodies mine scale, but their applicability on a regional scale may provefruitful, providedsuitable techniques can be developed. This paperinvestigates onepossibility: has not yet been tested. This particularstudyis basedon analyses for over the useof whole-rock geochemistry to evaluatethe $0 major and trace elements in about 100 productive exploration potentialof an intrusion or intrusive complex in the absence of obvious hydrothermalaltera- and nonproductiveintrusions.The discriminationis examinedon three scales (Fig. 1): (1) on the scaleof tion or mineralization. a single miningdistrict, thatof E1Salvador-Potrerillos Several attempts havebeenmadeto do thisin other in central and northern Chile; (2) on the scaleof a porphyry belts, with mixed success. Halogensand singleeast-west transect from E1Salvador to the Chilwater, the obvious discriminants betweenproductive ean coast; and ($) on the scale of a single tectonic (ore-forming) and nonproductiveintrusions,have province,the Chilean Andes.The techniques develbeenshownby Kesleret al. (1975a,b) to be generally oped are then applied to some other porphyry copper unsuccessful, for bothbiotiteand whole-rock analyses.
THE Andes of central and northern Chile contain
Introduction
The most comprehensive study basedon other elements is from the Papua-New Guinea-Solomon Islandsbelt, where Masonand McDonald (1978) have comparedthe compositions of productiveand nonproductive intrusionsfrom island-arc, continental margin,and continental settings. They found no single elementcapable of discriminating the two groups but notedthat, whendifferentrocktypesare present in a district,the porphyrycopperdeposits are associatedwith late, low potassium intrusions. An interestingextension of this work was the suggestion by
0561-0128/82/29/664-1152.50
provinces.
The El Salvador-Potrerillos
District
Selectionof samples
ern Chile some800 km north of Santiago(Fig. '1). The geologyof the E1 Salvador mine area hasbeen described by Gustarson andHunt (1975)andthe geologyof the districtby Tobar (1978).The mainvolcanic sequences are the Upper CretaceousCerillos For-
664
PORPHYRITIC
INTRUSIONS
IN CHILEAN
ANDES
i I i
665
mation, which is predominantlyandesitic,and the lower Tertiary Hornitosand La PeinetaFormations, which largely comprise silicictuffs and ignimbrites.
The intrusions can be divided into an older monzo-
ARICA
- 20S
El Abra
, .huquicamata
blagein places but elsewhere is a barrengranodiorite containing hornblende, biotite, and ilmenite as the
main mafic minerals. It contains internal intrusive
Potrerillos [L
t,/r
!
/
F,g E'S{],vador
ular to strongly porphyritic with a fine-grained aplitic groundmass. Samples of thisL porphyry werechosen
'asexamplesof a productive intrusion;they are associated with mineralizationbut are not usuallythemselves visibly mineralized,and are fresh enoughto
/'
100km
- 30S
Anda11o
I
to $4 m.y. (Zentilli, 1974). Recentmappingby Codelcogeologists showed that, like E1 Salvador, it is associated with an intrusive complex, known as the Cobreporphyry,the coreof which consists of an apparentlyfreshhornblende diorite (L-type) porphyry. Sampleswere taken of this hornblendediorite porphyry, and of a very weaklymineralized satellite, the North porphyry. Thesesamples were alsoincluded in the productiveintrusiongroup. The distributionof nonproductive intrusions from thisdistrictis shownin Figure 2; thosethat are numbered are referred to in the text or havetheir analyses recorded in Table 1. The samplingprogramcovered both the monzoniteand the granodiorite suitesand includedrock typesrangingfrom dioritesto dacites. The sampleswere otherwiselimited to thosewhose age, mineralogy, and texture were comparable with the productiveintrusions. Thus their age range is lowerTertiary (55-$4 m.y.), their mineralogy isdominated by feldspar,hornblende,and biotite phenocrysts,and their texture is porphyritic. Alsoincludedin the samples analyzedwere A porphyriesand latite dikes,two minor intrusives from E1 Salvador.The A porphyriesare characterized by an abundanceof mafic phenocrysts and were intruded during or after crystallization of the L porphyry. The latitescontainplagioclase and hornblende
I
as insets.
'lnient e(
Fie. 1. Locationmap of the Chilean porphyry copperdeposits mentionedin the text. Areasof more detailed study are denoted
phenocrysts and postdate the main intrusivecomplex by up to 2 m.y. (Gustarson and Hunt, 1975). Neither classifies clearlyinto the productive or nonproductive categories. The A porphyry is itself weakly mineralizedandthe genetic relationship between the latites and the productiveL porphyry is uncertain. Choiceof geochemical discriminants
Our attempts to discriminate betweenthe productive and nonproductive intrusions using whole-rock abundances of the transitionmetalstypically associated with porphyrycopperdeposits met with limited success. Figure $ summarizes the results. Of the elements, copper,gold, and molybdenum, which are concentrated in the primary central zones of porphyrycopperdeposits (Jerome,1966; Levinson,
666
4km I R29
Tertiary Intrusives,
Granodicrite samples
Monzonite samples
Rhyolite samples
R10
-I-
EL SALVADOR
town
.Mine
POTRERILLOS
667
Representative Geochemical Analyses for Some Productive and Nonproductive Intrusions from the E1Salvador District
Productive Nonproductive
24L Pot9B R 10 R29 R 17 R 11 R 13 RY19
Sample
no.
ES
$$$8
IL
6BL
L porphyries (Gd)
SiO2 TiO2 AI2Oa Fe2Oa
MnO
Gd
62.08 0.74 18.26 $.85
0.019
DP
55.77 0.74 14.84 7.$$
0.151
Gd
60.95 0.64 15.95 5.54
0.091
Gd
6:3.05 0.55 15.99 4.$9
0.071
QEP
67.86 0.23 16.07 2.$2
0.065
QMP
65.09 0.63 15.98 $.75
0.065
RYP
75.60 0.17 12.90 1.14
0.15
MgO
CaO
2.25
4.41
1.91
4.14
2.08
4.54
1.50
$.55
2.15
$.68
5.12
5.99
$.41
5.24
2.68
$.93
0.85
2.51
0.87
2.01
0.14
0.28
4.51 $. 17 0.01
2.79
Subtotal
99.17
99.25
98.$6
98.75
97.87
XRF
99.96
99.55
99.99
100.$8
100.59
100.45
Rb Ba Sr Th La
Ce Y Zr Nb Zn Cu Ni Pb
56 541 602 4 16
35 6 121 5 $7 $65 0 4
58 589 676 5 17
$5 7 128 7 24 655 1 7
$4 598 690 2 19
42 6 128 6 70 35 4 6
65 652 558 6 15
27 20 112 6 74 100 10 11
67 752 565 $ 15
$$ 15 131 2 45 81 1 11
INAA
Yb Th
Ta
0.55 $.66
0.56
0.55 3.55
0.54
0.52 1.45
0.$5
1.72 6.22
0.43
1.79 13.41
0.98
Hf
Sc
$.59
5.0
$.6
4.9
$.$2
4.8
4.22
12.0
7.73
5.2
Co
7.2
10.4
$.8
9.0
2.7
Figure 1 shows sample locations; Figure2 shows field localities; notethat sample R29 is the typicalbarrengranodiorite (T.B.G.)used asthe normalizing composition in Figure4 and that the Th dataanalyzed by XRF are subiect to muchgreatererrors (___2 ppm at 10 ppm) thandataanalyzed by instrumental neutron activation analysis (INAA) Abbreviations: Gd -- granodiorite, QMP = quartzmonzonite porphyry,DP -- daciteporphyry, QEP = quartz-eye porphyry,RYP = rhyoliteporphyry
whereas copper(Fig. :3)had on average higherconcentrations in the productive intrusions. However,it was still possible to find productiveintrusions containinglow copperand apparentlynonproductive intrusionscontaining high copper. Preliminary data suggest that molybdenum may givea betterdiscrimi-
nationthan copper(<1 ppm in nonproductive intrusions and >2 ppm in productiveintrusions), but these data are closeto the detectionlimit of the technique used and have therefore not been plotted. The elementslead and zinc are typically found in the peripheral zonesof porphyry copperdeposits (Jerome,
Fc. 2. Generalized geologic map of the E1Salvador-Potrerillos district(from Tobar, 1977). Symbols showthe locations of the intrusivesamples usedin this paper and are numberedwhere referred to in the text. All samples are nonproductive exceptthose from the E1Salvador and Potrerillos mines.Sample locations for the freshest L porphyries from the E1Salvador mineare shown on a 2,600-mlevelprojection
(see inset).
668
AM t'
0.0 lppm
Cu
10ppm
FI4 -t'. t-N -,.t--N--,,- I,-,,. ,,= -N, --,-,-{ Non-produchve 1 mentin K, Rb, Ba, Th, Zr, Hf, Ta, andthe light rare
Produchve
I I
from left to right as large ion lithophile elements, highfield strength elements, andtransition metals but with somemodification to highlightanomalies. Patternscharacteristic of nonproductive intrusions arethe fiat patternof the barrengranodiorite (for the granodioritesuite) and the pattern of R18 (for the
monzonite suite); the monzonite exhibits an enrich-
100
500
1000
Non-producbve
Zn
__
::;:; :: :
',',',Y,;', ',',
1O0
',', Produchve
5 O0 1000
earth elementsand a depletionin Ca and Sr relative to the granodiorites. Two L porphyries,chosenas characteristic of productiveintrusions, have alsobeen plotted. Some of the hornblendein sample 1L has beenalteredto biotite,but 8888 shows negligible signs
of alteration.
10ppm
1 pp m
10
I-ltll-IN44.4--tNon-produchv i 50 100
FIG. 8. Bar graphsof the abundances (ppm) of Zn, Cu, Pb, and Au in productive and nonproductiveintrusions in the El SalvadorPotrerillos district.The significant overlapin all eases makesthese elementsunsuitablediscriminants.(Zn and Pb analyzed by XRF
The principalfeatures of the productive intrusions are the strongnegativeanomalies in the elements, Mn, Y, Yb, and Th. Sr shows a smallpositive anomaly, but otherelements appearto be of little useby themselvesin carrying out the discrimination.It should alsobe noted that K, Rb, and Ba, which proved effective in Pacificarcs(Masonand McDonald, 1978), are ineffectivein this setting.Theseobservations can now be used to construct a discriminantdiagram
based on all available data.
1966). We might thereforeexpecta productiveintrusionto be depleted in Pb or Zn relative to a nonproductive intrusion. Zn shows no discrimination (Fig. 8), but Pb shows, on average,lower valuesin productive compared to nonproductiveintrusions (Fig. 8). Again, however,there is too much overlap to provide a completely reliable discriminant.The halogenelements,fluorine and chlorine, were also plottedin this way, but theseprovedpoor discriminants,as Kesleret al. (1975a) showedfor other areas. The likely valuesof other elements in carryingout the discriminationbetween productive and nonproductive intrusions can be deducedby constructing geochemical patternssuchas thoseshownin Figure 4. The patterns are constructedfrom ratios of the
element abundances in the rock to the element abun-
Discriminationdiagramfor productiveand nonproductive intrusions Of the four elements that showed the largest anomaliesin productiveintrusions, Mn and Y are in greatest absoluteabundanceand can easily be analyzed to the required level of precisionby X-ray fluorescence.Yb (which behaves in a similar way to Y) and Th can both be analyzedaccuratelyby neutron activation analysis, but this is not a practical method for most explorationprograms.Accordingly,a covariationdiagramof MnO againstY hasbeendrawn
/. .,__ /
ROCK
//
x, /% R13
\ Nonroducbve
dancesin a typical barren granodiorite.Any barren rockcouldhavebeenusedasthe normalizing factor; R29 (seeTable 2) was in fact chosenbecauseit superficiallyresembled mostof the productivesamples in its granodiorite composition and intermediateSiO2 (61%) and because its main component minerals(plagioclase, hornblende,biotite, and quartz) were apparentlyunaffectedby weatheringor alteration.The elementsin the pattern include mosttrace elements that are readily analyzedby X-ray fluorescence and FIC. 4. Geochemical patternsfor productiveand nonproductive instrumentalneutron activationanalysis techniques. intrusions from the El Salvador-Potrerillos district. Data normalized The major elements K and Ca have also been in- to typicalbarrengranodiorite (T. B. G., R29). The arrows highlight in the productiveintrusions. cluded. The elementsare approximatelyarranged importantnegativeanomalies
IL
_/i. G (R2
3338
Productwe
Ca
Sr
Rb
Ba
La
Ce
Zr
HI
Ta
Yb
Ti
Mn
Zn
Co
Sc
PORPHYRITIC
INTRUSIONS
IN CHILEAN
ANDES
I
669
asFigure5 to highlight the combined effectof these anomalies, although the limitedavailable data indicate that diagrams suchas MnO againstTh or Yb, andTh against Yb, are likelyto be equallysuccessful. The points plotted in this diagram have been restrictedto samples that are fresh or exhibit only incipientalteration, although the samefeatures are apparent when all samples from the area are plotted. On the basis of the relatively fresh samples, a single
line could be drawn which discriminates between
711
.
Siena El Jardin
El Salvador
X Jurassic
? 26 I
50kin
INTRUSIONS
Basement
-F
Cretaceous
Tertiary
unambiguous productive and nonproductive intrusions with a classification success of about92 percent. A problemwith any diagramof thistype is the possibilitythat the discrimination may be masked by the effectsof fractional crystallization. In calc-alkaline intrusions MnO usually decreases from diorites to graniteswhereasY can increaseslightly, decrease
-Calde, .
27-
FIG. 6. Locationof samples and intrusions in the Copiapo-E1 Salvadortransect(much of coastal data from J. Tarney).
rentlyminable deposits but whichareassociated with significant mineralization. For example,the A porphyriesat E1 Salvador and the North porphyryat Potrerillos, both shownin Figure 5, can containup to maskthe large negativeanomalies of theseele- to 0.8 percentCu. It might have been thoughtthat the latite dikes at E1 Salvador could be classified as mentsin productiveintrusions. in the sameway as L porphyIt shouldbe noted, in addition,that the boundary productiveintrusions and betweenproductiveand nonproductive intrusions is ries. However, they plot in the nonproductive fieldsin Figure 5. Moreover,micronot fixed but will dependon what sizeand gradeof subproductive of amphiboles in the latites are very deposit can be mined at any giventime. It wasthus probe analyses in L porphyries, andplagioclases thought necessary alsoto definea field of subproduc- differentfrom those than L portive intrusions, which are not associatedwith cur- are significantlymore calcic (An47_44) phyry plagioclases (Ana_4). Thesefeatures,and the 2-m.y. age differencebetweenthe 89-m.y. latite and the 41-m.y. L porphyry(Gustarson and Hunt, 1975), suggest that the latitesmay be part of an independent intrusiveevent. If true, this impliesthat it may not be justifiedto uselate dikesas freshsamples repreNon-productive sentativeof a porphyry suite. Only one anomalous barren sample,R6, plots in ,o.N e 750 75076 855 the productive field. Its age is 86 m.y. (Tobar,1977), ,"' /t. Yppm and its chemistry closely resembles that of a typical L porphyry except for higherBa and a higherinitial 64 e59' o60 D58 SrS7/SrS6 ratioIt appears to be unrelatedto miner_Productve 60 60 62 alization,althoughit couldbe the root of an eroded, El Salvador 'L'-porphyry 6261 4 Potrerdlos Cobre porphyry ' poorlydeveloped porphyrycopperdeposit. e60 Sub-productive
uesof the samples usedcan be seento vary from 55 to 76 percent.There is, however,no obvious correlation betweenSiO2and either MnO or Y, indicating that fractionalcrystallization variations are too small
10
e64
latltes
El Salvador-Copiapo Transect Selectionof samples The regionfrom E1Salvador and Potrerillos to the coast is characterized by a generalinland migration of igneous activity from Jurassic to Recent (Farrar eta!., 1970) (Fig. 6) and an increase in erosion level
toward the coast. Two sets of data are used: one from
El Salvador-Potreriltos
005
01
MnO%
05
FIG. 5. Covariation diagramof Y againstMnO, limited to the freshest and mostporphyritic intrusions in the E1Salvador-Potrerilles district. SiO2 values for eachsample areshown. A discriminant line hasbeendrawn which separates mostproductive from most nonproductive intrusions.
670
El Salvador Transec!
Non-produchve
Yppm
10
Sub-
productsveX
. /
. . .'/%
,.'
'
s,,dor
Sub-productive
most porphyriticrocksin the nonproductive field have lowestY and MnO values,but they do nonetheless plot outside the area occupied by the productive intrusions. R6, discussed earlier, remainsthe only really anomalous point. The data from the submineralizedintrusions at CopperHill andSierraE1Jardin provide further confirmationthat it is also possible to define a subproductive field. Compositional fields, for productive, subproductive, and nonproductive intrusions have beendrawn by eye onto this diagram. It is alsoworth notingthat, althoughonly intrusions areconsidered in Figure7, some 30 andesites sampled from the transectwould alsoplot within the nonproductive field. The Chilean Andes
'
Sample collection
In addition to the deposits of E1 Salvadorand Potrerillos and the subeconomic prospectat Sierra E1 Jardin, a limited number of samples were analyzed Flg. 7. Covariation diagramof Y against MnO for productive, shownin subproductive, and nonproductiveintrusionsin the Copiapo-E1 from the other porphyrycopperdeposits
I
001
i I I I 1011
0 5
I 015 I I III
1
MnO%
El Abra, AndoSalvadortransect. The samples shownare from porphyriticand Figure 1, namely, Chuquicamata, nonporphyritic, fresh and slightlyaltered intrusions. Fields have collo, and E1 Teniente. Y and MnO data for these beendrawn for eachintrusive type and showproductive intrusions intrusions are plottedin Figure 8. to the lower left (solidline), nonproductive intrusions to the upper Chuquicamata: One of the world'slargestcopper right (fine-dashed line),and subproductive intrusions in the center. knownas The fieldfor nonproductive intrusions hasbeenextended to include mines,it is locatedalonga majorstructure data from 30 andesitesamples from the El Salvadordistrict (not the West fissure and related to a complexof granoplotted in this diagram). diorite porphyries. Mineralization is thoughtto have
beencaused by the Chuquicamata porphyries, which have been dated at between33.6 and 30.3 m.y. and XRF spectrometer at BirminghamUniversityusing which have been subdivided texturallyinto the East, the methodof Tarhey et al. (1978). West, and Bancoporphyries (Ambrus,1979). These
No restriction was made on the texture and com-
position of the samples; both equigranularand porphyriticsamples havingan SiO2rangefrom 48 to 75 percent were included. The vast majority of these samples were barren and had no obvious association with porphyrycoppermineralization, although some couldhave underlainnow-eroded deposits. The only intrusions obviouslyassociated with mineralization are from Sierra E1 Jardin, a subeconomic porphyry Yppm copperprospect marked in Figure 1 (Oyarzun, 1973; Neumann, 1972). Samplesfrom Copper Hill, a satellite depositof E1 Salvadorand not plotted earlier because of slightalteration, havealsobeenincluded.
Geochemical discrimination
\ \
I
i
Non-productive
\
I I
i
/,/
Relabvely fresh sample with Amphlbole
Productive
10
El Abra
not fresh
Figure 7 showsthe covariationdiagram of Y againstMnO for all available fresh to moderately freshrocksfrom the El Salvador traverse,including thoseplotted in Figure 5. It is apparent that the boundary betweenproductive and nonproductive intrusions, asdefinedin Figure 5, applies to the whole
traverse as well as to the E1 Salvador district; in fact
Produchve
o
1
El Temente Chuqulcamata
samples
Non-productive
Tablalume batholith
Problemahc
El Abra dorlte
01
| 005 I I I I I011 .
MnO%
I 05I ! I I I 1 I
the classification success rate for productiveand nonproductive intrusions is greater(95%).In generalthe
FIG. 8. Covariation diagramof Y against MnO for someother Chilean porphyrycopperdeposits. The fieldsfor productivesubproductive and nonproductiveintrusionsare taken from the Copiapo-E1 Salvadortransect(Fig. 7).
PORPHYRITIC
IN TR USINS
IN CHILEAN
ANDES
671
rocksare exposed in the central area of the open pit. Samples of East and West porphyries were collected although the wholezonehad suffered pervasive wallrock alteration. A further granodiorite intrusion, thoughtby Ambrus(1979) to be of similarage and relatedto the samecomplex,the Elena granodiorite, was also sampledfrom the east side of the pit, althoughthis, too, had been stronglyaltered. An additional nonproductive Tertiary intrusionfrom the area, E1 Mineral, was alsosampled. El Abra:This deposit lies42 km northof Chuquicamata,is of similarage ($5.4-$$.5 m.y.) and is also closeto the West fissure(Ambrus, 1977). Samples
The positions of the data points from Andocollo are alsoproblematical.The Tablalaluma batholith plots correctly asnonproductive, but the three samples of porphyries plot as subproductive. Andocollois not,
however, typicalof otherChileanporphyries, being of UpperCretaceous ageandnearthecoast, andhaving characteristics of manto,aswell asporphyry copper, deposits. It is possible that the preciseposition of the discriminant boundary is differentfor intrusions of thisageor for intrusions in thissegment of
the Andes. However, further work is needed in the
were taken from each of the intrusions making up In summary,the Y-MnO discriminantdiagram the complex: the Southern granodiorite, the E1Abra givesmeaningfulresults for plutonicrocksrelated to diorite, the quartz monzonite, and the dacite por- the variousChilean porphyry copper deposits, alphyry. All theseintrusions are hostsfor the miner- thoughit is much more successful when limited to alization,but the two mostimportantare the diorite a single transect. Eachtransect may havea particular and dacite. Most samples collectedwere, therefore, background Y and MnO value that needsto be catstrongly altered,an importantexception being the egorized.
diorite, which is exposed outsidethe zone of mineralization.
porphyry. The finalintrusions are lamprophyre dikes the Pacific arcs and the New Guinea mobile belt. The areshown in Figure9A and B. Discrimination whichhave a similarrelationship to the orebody at results boundaries E1 Tenienteas the latite dikeshave to the orebody in eachcaseis good,but the discriminant
at E1 Salvador.Samples were collectedfrom all three
El Teniente: In central Chile this is the youngest Usingdata from Mason and McDonald(1978)and of the large Tertiary porphyries, dated at 4.$ m.y. and Feiss (1979),Y-MnO diagrams havebeen (Camus, 1975). The important intrusions associated Mason plotted for nonproductive and productive samples in with the orebodyare a quartz diorite and a dacite
rocktypes,but only the lamprophyre wasunaltered. Andacollo:This depositis much nearer the coast and is thoughtto be older than the other porphyry copperdeposits, probablyUpper Cretaceous (Llaumet et al., 1975; Ambrus, 1979). A representative samplewastakenfrom a nonproductive intrusive, the
Tablalaluma batholith, in the northeastof the district.
are different from each other and from that determined for the E1 Salvador transect. Mason and
McDonald (1978) recognized that Y valuesin their plutonicrocksincreased (at a given SiO2 content) from island-arc through continental marginto continentalsettings. This shift in the position of the discriminantaxisis thereforeto be expected.Of course,
variable factors such as the thickness of crust and the
Three samples weretakenfrom a porphyritictonalitc depth of the underlying Benioff zone, have also stock anditsrelateddikes andsills, collectively known caused variations within the E1 Salvador-coast traas the Andacollo porphyries, with which the miner- verse(Zentilli, 1974). In this casehowever,the effect on the discrimination wasrelativelysmall. alizationis geneticallyrelated.
This difference between Chile and the southwest
Pacificemphasizes that the ideal unit for consideraAt E1 Abra, the fresh E1 Abra diorite (EA4) from tion is a single tectonic province and that the exact of discrimination betweennonproductive and southof the E1 Abra alterationzone(Ambrus,1977) details
Geochemical
discrimination
However,there are alternativepossibilities: the dior- productive and nonproductive intrusions ite is less fractionated (58%SiO2)than the rocksused Depletion of Y and MnO couldtheoretically have in the original discriminationand should contain taken place at various stages in the history of the more MnO and Y than a granodiorite and it is not intrusion. The four main possibilities are: certain at presentwhether this intrusion causedmin-
samples (especially for Y) will vary achashigh Y and Mn and consequently plotsas non- productive cording to the tectonic province in question. productive(Fig. 8). It is possible that this intrusion does not conform with the discrimination presented. Explanations for the Y-MnO discrimination of
eralization or wasmerelythepassive host to thedacite 1. During partial melting: the productive intruporphyry. Furtherstudies should distinguish between sions couldhavebeenderivedfrom a source region thesepossibilities. depleted in these elements or phases that accomodate
672
lOO
Pacihc
Arcs
50
[ I
I
Non-produchve
a /
//
ppm
//
o/////
greater distributioncoefficients for Y and Mn than phases crystallizing from nonproductive intrusions. 8. During expulsion of a volatilephase: a CO-rich phaseexpelled at depth (Gustarson, 1978, 1979) or an HO-rich phase expelledat high levelsin the crust couldhave removedY and MnO from the magma. 4. During hydrothermal alteration:passage of hot fluidsof magmaticor meteoricderivationcouldhave reactedwith the productiveintrusions after crystallization,removingY and Mn in solution.
Productive
Non-produchve
The effectsof hydrothermalalterationare easiest to assess. None of the samples plottedhaveundergone sericitic,argillic, or intenseK silicatealteration.However, the samples do showa variable degree of subsolidusreplacementof hornblende,sphene,and ilmenite by biotite and rutile. The changesin MnO
and Y that result from these reactions can be assessed
I I I I I 1
1
100
001
I I I I I I IoI
005
50
by comparing Figure 5 (whichwasrestricted to samples with negligibleto slight replacementof hornblende)with Figure 7 (which alsocontainedsamples showing significant to total replacement). When the morealteredsamples were included,the field of productive intrusions expandedto lower MnO, but similar Y, values.Thus, it is apparent that replacement
of hornblende causes MnO to decrease but does not
[]
.
Y ppm
10
I
/
///
x /
'
"...
"".. ////
5
Productive
Non-productive
the productivefield were almostfresh and that nonproductiveintrusionswith comparabledegreesof weak alterationhave higher concentrations of both MnO and Y. We therefore have to look to precrystallization processes for the causes of the Y anomaly and at least part of the MnO anomaly. Of the precrystallization processes, loss of a volatile phasemay be the mostimportantin the caseof man-
ganese, sincemosthydrothermal fluidscontainsignificant concentrations of manganese, and sincemost I I 0 g5 I I I I01I I I I I I I II1 I porphyrycopperdeposits are knownto be surrounded by manganese-rich halos.However, hornblende and FI(;. 9. Covariation diagrams of Y against MnO for productive ilmenite can contain significantconcentrations of and nonproductive intrusions in Pacific arcs (A) and continental manganese and, if thesephases crystallized from the margins(B) (data from Masonand McDonald, 1978, and Mason in the and Feiss,1979). "A" givesthe discriminant boundaryfor Pacific magmaor if thesemineralswere concentrated arcs(heavyline) and the fieldsfor the Chileantransect (from Fig. residuefrom partial melting, the resultingmagma 7). "B" showshow discriminant boundariesfor Pacific continental couldbe depletedin MnO withoutthe needto invoke margins(heavy line), Pacific arcs (heavy dashedline), and the volatile transfer.The depletionin Y and heavy rare
1
MnO%
earth elementsis lesseasyto explain. Theseelements are not strongly partitioned into volatile phases (Flynn and Burnham,1978).They are, however,parthese elements could have been left in the melt restitioned into hornblende,and into minor phases such idue. as spheneand apatite. Detailed petrogenetic modeling,to be published 2. During fractional crystallization: phases cryssuggests that the manganese depletionis tallizing from productive intrusions could have elsewhere,
675
best explained by lossof Mn in the volatile phase whereas Y depletion is best explained by crystallization of phases(hornblendeand sphene)that are stableunder hydrousconditions or by the presence of suchphases in a melt residue.It appears therefore that the two axesof the Y-MnO covariation diagram complementeach other. Low Y in a magma may indicate the involvementof hydrousphases during the earlier historyof the magma, whereaslow MnO may indicateextensive loss of magmaticfluidsfrom the magmaand thus be directly relatedto the mineralization event.
useful in identifying porphyry copper productive suites. No single element isan adequate discriminant,
but productiveand nonproductivesamplesare well separated on Y and MnO covariationdiagrams,productivesamples being anomalously low in both elements. The discrimination works best when limited
Exploration Application
or Explorationat the reconnaissance stage for por- to a particulartransect(e.g.,E1 Salvador-Copiapo) tectonic belt (e.g., a Pacific arc), because Y abunphyry copperdeposits in arid or glaciatedterrains would be improved by any techniqueof lithogeo- dancesmay vary from provinceto province. Low Y may reflect the distinctivemagmatic prochemicalsampling that couldfingerprintapparently cesses necessary for porphyrycoppergenesis whereas fresh intrusions which were associated with hidden low MnO may reflect the size and strength of decopper mineralization.In this study of the E1 Salvelopment of a hydrothermal system. Y and MnO vador-Copiapo transect,Y and MnO were found to
stageof explorationand could give evibothbeingtypicalof productive suites. In a less well- naissance dence of which intrusive suites crystallized from exploredmetallogenic provinceor transect, a similar magma that generatedporphyry coppermineralizaprogramof whole-rock sampling of the accessible intion. trusions, say 200 to 300 samples, and XRF analysis Acknowledgments for Y and MnO, would be relatively cheap.A more detailed follow-up investigation or extra sampling We are gratefulto JohnP. Hunt and AlvaroTobar, could be carried out around intrusions with unusually and to Enrique Tidy, JorgeQuiroga,and other Codlow Y and MnO. elcogeologists, for their support and help in thisproAsa cautionary notethe following limitations ought ject. We would alsolike to thank Peter Francis, Chris to be taken into account: Hawkesworth,Richard Holt, and Andy Tindie for JohnTarney and Nick Marshfor 1. The possible SiO2rangeof samples is around48 helpful discussions, to use their Y and MnO data, and Juan to 75 percentbut preferably55 to 70 percent(quartz permission from Andiorites,granodiorites, and quartz monzonites). The Carlos Marquardt for providing samples fresh productiverocksconsidered in this study con- docollo.The authorswere both supportedin their Council tain significant proportions of modal amphiboleand work by the Natural EnvironmentResearch of the United Kingdom. ilmenite but are still depleted in Y and MnO. Nevertheless, it should be noted that it is theoretically July 30, November20, 1981 possible for rockscontainingabundantcumulateilREFERENCES menite to have anomalously high MnO and rocks Ambrus, J., 1977, Geology of the El Abra porphyrycopperdeposit, containingabundantcumulate hornblendeto have Chile: ECON. GEOL., v. 72, p. 1062-1085. anomalously high MnO and Y. -1979, Emplazamiento y mineralizacion de losporfidos cupri2. Sampling should be restricted to fresh and ferosde Chile: Unpub. Ph.D. thesis, Univ. Salamanca, 313 p. slightly altered or slightly weatheredsamples.In- Armbrus,G. A., Oyarzum, J., and Arias, J., 1977, Rubidium as a guideto ore in Chileanporphyrycopperdeposits: ECON. GEOL., tenselyaltered samplesmay have lost both Y and v, 72, p. 1086-1100. MnO and may thereforegive misleading results on Camus, F., 1975,Geology of the E1Teniente orebody with emphasis the discrimination diagram. on wall-rockalteration: ECON.GEOL.,v, 70, p. 1341-1372. 3. Sampling should be limited to a particulartec- Chaffee, M. A., 1976, The zonal distribution of selectedelements tonic provinceor transect,as background Y values abovethe Kalamazooporphyrycopperdeposit,SanManuel district, Pinal County, Arizona: Jour. Geochem Explor., v 5, p. vary with the tectonicprovincein question. 145-165. 4. The Y and MnO discrimination may only be Coope, J. A., 1973,Geochemical prospecting for porphyrycopperusefulin tectonicprovinces situatedat Phanerozoic type mineralization--a review: Jour. Geochem Explor., v. 2, p. convergent plateboundaries, which,anyway,is where 81-102. the main porphyrycopperprovinces are located(Sil- Fartar, E., Clark, A. H., Haynes,S. J., Quirt, G. S., Conn, H., and
discrimination
674
between Zentilli, M., 1970, K-Ar evidencefor post-Paleozoic migration Mason,D. R., and Feiss,P. G., 1979, On the relationship whole rock chemistry and porphyry copper mineralization: of graniticintrusion loci in the Andesof northernChile: Earth Eco. GEOL., v. 74, p. 1506-1510. Planet. Sci. Letters, v. 10, p. 60-66. Flynn, R. T., and Burnham,C. W., 1978, An experimenthi deter- Mason, D. R., and McDonald,J. A., 1978, Intrusiverocks and porminationof rare earth partitioncoefficients betweena chloride phyry copperoccurrences of the PapuaNew Guinea-Solomon containing vapourphase and silicate melts,in Drake, M. J., and Islands region:A reconnaissance study:ECON. GEOL.,v. 75, p. Holloway,J. R., eds.,Experimentaltraceelementgeochemistry: 857-877. Proceedings of an InternationalConference,held in Sedona, Neumann,H. J., 1972,SierraDel Jardin,Provincia de Atacama, Arizona,12-16September, 1977:Oxford,Pergamon, v. $, p. 685-
Investigaciones Geologicas, unpub.companyrept., 19 p. Gustarson, L. B., 1978, Some major factorsof porphyry copper Olade, M. A., and Fletcher, W. K., 1975, Primary dispersion of genesis: Ecoa. GEOL.,v. 75, p. 600-607. rubidiumandstrontium aroundporphyry copper deposits, High-1979,Porphyrycopperdeposits and calc-alkaline volcanism, land Valley, BritishColumbia:Ecoa. GEOL.,v. 70, p. 15-21. in McElhinny,M. W., ed., Earth: its origin,structure and evo-- 1976,Traceelement geochemistry of theHighland Valleyand lution: New York, Academic Press,p. 427-468. Guichon Creek batholith in relation to porphyry copper minGustarson, L. B., and Hunt, J.P., 1975, The porphyrycopperdeeralization: Ecoa. GEOL., v. 71, p. 755-748. posit at El Salvador, Chile:Ecoa. GEOL., v, 70, p. 857-912. Jerome, S. E., 1966,Some features pertinent in exploration of por- Oyarzun, J., 1975,Distribucion de Cobre,molibdeno, plomoy rubphyrycopper deposits, in Titley, S. R., and Hicks,C. L., eds., idio en el areade SierraJardin,Provincia de Atacama: Santiago, Geology of the porphyrycopperdeposits, southwestern North Chile, Instituto de Investigaciones Geologicas, unpub.company
701.
rept., 15 p.
Kesler, S. E., Issigonis, M. J.,and Van Loon, J. C., 1975a, An evaluationof the useof halogen and waterabundances in effortsto distinguish mineralized and barrenintrusive rocks: Jour.Geo-
chem. Explor., v. 4, p. 255-245. Potrerillos region, Atacama, Chile:Unpub.Ph.D.thesis, Berkeley Kesler,S. E., Issigonis, M. J., Brownlow,A. H., Damon, P. E., Univ., 117 p. Moore,W. J., Northcote, K. E., and Preto,V. A., 1975b,GeoA.D., Weaver,S. D., Donnellan, N. C. B., chemistry of biotites from mineralized and barrenintrusive sys- Tarney,J., Saunders, and Hendry, G. L., 1978, Minor elementchemistry of basalt tems: Ecoa. GEOL., v. 70, p. 559-567. from Leg 49, North AtlanticOcean:Deep SeaDrilling Project, Levinson,A. A., 1974, Introductionto explorationgeochemistry: Initial repts.,v. 49, p. 567-691. Wilmette, Illinois,Applied Pub., p. 614. evolution and metallogenic relationLlaumet,C., Olcay, L., Marin, C., Marquardt,J. C., and Reyes, Zentilli, M., 1974,Geological ships in theAndes of northern Chile between 26 and29 south: E., 1975,E1yacimiento cobreporfidico "Andacollo", Provincia Unpub. Ph.D. thesis,Queen'sUniv., Kingston,410 p. de Coquimbo,Chile: Rev. Geol. Chile, v. 2, p. 56-66.
Sillitoe,R. H., 1972, A plate tectonicmodel for the origin of porphyry copperdeposits: Ecoa. GEOL.,v. 5, p. 184-197. Tobar,A. B., 1977,Stratigraphy and structure of the El Salvador-