Lee Et Al - 2017 - Magmatic Evolution at El Salvador Porphyry - Domeyko Faults

©2017 Society of Economic Geologists, Inc.
Economic Geology, v. 112, pp. 245–273
Magmatic Evolution of Granodiorite Intrusions at the El Salvador Porphyry Copper

Deposit, Chile, Based on Trace Element Composition and U/Pb Age of Zircons‡
Robert G. Lee,1,†,* John H. Dilles,1,† Richard M. Tosdal,2,** Joseph L. Wooden,3*** and Frank K. Mazdab3,****
1College of Earth, Ocean and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 97331-5506
2University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z4
3U.S. Geological Survey, Menlo Park, California 94025
Abstract
Uranium-lead ages and trace element compositions of zircon from a series of shallow porphyry intrusions docu-
ment the temporal, chemical, and thermal magmatic evolution of magmatic-hydrothermal porphyry Cu (Mo-
Au) ores in the El Salvador district, Chile. Zircons (n = 240) from 15 Eocene age diorite, granodiorite, and
granite porphyry intrusions were analyzed by SHRIMP-RG ion microprobe. The weighted means of 207Pb-
corrected 206Pb/238U zircon ages span 3 m.y. from about 44 to 41 Ma, with peak magmatic flux at 44 to 43 Ma.
The granodiorite porphyries at the Turquoise Gulch copper deposit record waning stages of magmatism at 42.5
to 42.0 Ma and were followed by postmineral latite dikes at about 41.6 Ma. Porphyry copper ores formed con-
temporaneously with porphyry intrusion centers that progressed temporally from north to south, from the small
deposits at Cerro Pelado (~44.2 Ma), Old Camp (~43.6 Ma), and at M Gulch-Copper Hill (~43.5–43.1 Ma) to
the main ore deposit at Turquoise Gulch (~42 Ma). The Eocene porphyry intrusions contain a few Mesozoic
(n = 9) inherited zircons and numerous (n ≥19) antecrystic zircons about 1 to 2 m.y. older than the host intru-
sion that provide evidence of extensive Eocene magmatic recycling.
The Ti-in-zircon geothermometer provides estimates of 890° to 620°C for zircon crystallization and records
both core to rim cooling and locally high-temperature rim overgrowths. Most zircon in ore-related K, L, and
R porphyries yields near-solidus temperatures of 750° to 650°C and crystallized from compositionally diverse
granodiorite porphyries that are a product of crystal fractionation of hornblende, apatite, and titanite with
lesser crustal contamination and mixing with high-temperature deep-sourced mafic magma. During a 3-m.y.
period, porphyry intrusions tapped an evolving granodioritic magma chamber that was periodically heated,
locally remelted, and mixed with mafic magma during recharge events but cooled between recharge events to
evolve ore fluids.
Europium anomalies (chondrite-normalized EuN/EuN*) in zircons become more pronounced with increased
Hf content and cooling but display two distinct evolutionary paths: EuN/EuN* of early quartz porphyry evolves
from 0.8 to 0.3, whereas the late synmineralization porphyries evolve from 0.8 to 0.65. The EuN/EuN* ratio of
zircon reflects the Eu3+/Eu2+ ratio of the melt, and therefore the granodiorite porphyries at Turquoise Gulch
were the most strongly oxidized of the El Salvador magmas. The strongly oxidized trend porphyry magmas at
Turquoise Gulch are apparently directly linked to magmatic degassing at ~700°C to produce large amounts of
ore-forming copper, sulfur, and chlorine-enriched magmatic-hydrothermal aqueous fluids.
Introduction with overlying rock to form hydrothermal ore minerals, veins,

Hydrous intermediate to silicic calc-alkaline magmas may and wall-rock alteration (Gustafson and Hunt, 1975). These
crystallize in the upper crust to form granitoid plutons, some magma chambers typically reside at depths of 3 to more than
of which are associated with large sulfur- and metal-rich mag- 10 km and are sampled by small volumes of magma that
matic-hydrothermal mineral deposits.These systems include sequentially intrude upward as porphyry dikes to form por-
economic porphyry deposits and associated epithermal phyry-type (Cu ± Mo ± Au) deposits mainly at depths of 2
deposits that annually contribute billions of dollars of gold, to 5 km (Seedorff et al., 2005). Isotopic and fluid inclusions
silver, copper, molybdenum, lead, and zinc to the global studies of porphyry deposits support a magmatic source of
economy. Porphyry deposits result when hydrothermal fluids water, metals, and the metal ligands required for ore transport
are released from upper crustal magma chambers of inter- and deposition (Ulrich et al., 1999; Harris and Golding, 2002;
mediate to silicic composition (cf. Burnham, 1979) and react Rusk et al., 2004, 2008; Halter et al., 2005).
Ore-related magmatism may be short- or long-lived and
may be related to one or more discrete periods of mineral-
† Corresponding authors: e-mail, [email protected]; [email protected]. ization. Short-lived magmatic systems include differentiated
edu batholiths with multiple porphyry systems of approximately
*Present address: University of British Columbia, Vancouver, British Colum-
bia, Canada V6T 1Z4. the same age such as Yerington, Nevada (1 m.y., Dilles and
**Present address: PicachoEx LLC, Folly Beach, South Carolina 29439. Wright, 1988; Dilles and Proffett, 1995) and the El Abra-For-
***Present address: 785 Nob Ridge SW, Marietta, Georgia 30064. tuna granodiorite complex, Chile (1–2 m.y., Dilles et al., 1997;
****Present address: Department of Geosciences, University of Arizona, Campbell et al., 2006; Barra et al., 2013), and long-lived mag-
Tucson, Arizona 85721.
‡ Appendix 1 follows the Reference section, below. A digital supplement con- matic systems with one or more temporally discrete hydro-
taining electronic Appendices 2–5 is available at http://economicgeology. thermal events such as Butte, Montana (4 m.y., Dilles et al.,
org/ and at http://econgeol.geoscienceworld.org/. 2003), Tampakan, Philippines (~8 m.y., Rohrlach and Loucks,
Submitted: May 10, 2016
0361-0128/17/4468/245-29 245 Accepted: September 1, 2016
Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

by Juan P. Jaimes Bermudez
246 LEE ET AL.
2005), Chuquicamata, Chile (2–3 m.y., Reynolds et al., 1998; composition of the coexisting melt or fluid (Hanchar and van
Ballard et al., 2001), the Potrerillos district, Chile (8 m.y., Westrenen, 2007). Therefore, the zircon compositional range
Marsh et al., 1997), and Yanacocha, Peru (~6 m.y., Longo and and zonation may be used to estimate the temporal change of
Teal, 2005; Longo et al., 2010). Hydrothermal ores associated trace element composition of magma within the context of the
with porphyry intrusions predominately form late in the lifes- U/Pb age data (e.g., Pettke et al., 2005; Pelleter et al., 2007).
pan of the magmatic systems (cf. Sillitoe, 1988; Seedorff et al., These changes can include magmatic cooling during assim-
2005, 2008). ilation-fractionational crystallization or heating during mafic
The Indio Muerto (El Salvador) district of northern Chile recharge events (Miller et al., 2003). Furthermore, changes
is characterized by a series of late Eocene porphyry intru- of oxidation state of the magmas can be monitored via the
sions spanning several million years and including several Eu and Ce anomalies of zircon (cf. Ballard et al., 2002; Dilles
spatially separate and apparently temporally discrete centers et al., 2015) and changes in temperature via the Ti-in-zircon
of hydrothermal copper sulfide deposition (Gustafson and geothermometer (Watson et al., 2006).
Hunt, 1975; Cornejo et al., 1997; Gustafson et al., 2001). In Our goal in this study is to provide improved understanding
aggregate, the Indio Muerto district represents a moderate- of the magmatic processes related to porphyry copper depos-
sized porphyry Cu-(Mo-Au) district containing approximately its by adding new age and trace element data of one of the
15 million metric tons (Mt) of resources, principally at the best known porphyry copper districts. To this end we analyzed
Turquoise Gulch center (the main El Salvador mine) and also zircons for U/Pb age and trace element composition from
in smaller centers at Old Camp, M Gulch-Copper Hill, Sector 15 samples of the complete intrusive suite at El Salvador. The
Granito, and Cerro Pelado (cf. Gustafson et al., 2001). Exist- data presented below are one of the largest data sets of zircon
ing age data suggest that Eocene magmatism commenced age and trace element data for a single porphyry system (cf.
at approximately 44 to 45 Ma with intrusion of a series of Ballard et al., 2002; Campbell et al., 2006).
quartz-feldspar and feldspar porphyries and was sustained for
about 3 m.y. until the emplacement of a series of granodioritic Tectonic Setting of the Andean Precordillera
stocks, plugs, and dikes associated with the formation of the Subduction tectonics and related arc magmatism have formed
main porphyry copper deposit at Turquoise Gulch at approxi- arc-parallel belts in northern Chile that migrated eastward in
mately 42 Ma (Gustafson and Hunt, 1975; Cornejo et al., time from the coastal belt to the active volcanoes along the
1997; Gustafson et al., 2001; Lee, 2008; Zimmerman et al., Chile-Bolivia and Chile-Argentina borders. Transpressional
2014). The relative chronology of the intrusions is well known arc-parallel fault systems along the Andean margin locally
within individual centers based on the seminal early work by focused magma ascent and pluton emplacement (Taylor et al.,
mine geologists (Gustafson and Hunt, 1975; Gustafson et al., 1998). These fault systems also migrated eastward with time,
2001); however, geochronological studies of the El Salvador from the Late Triassic to Early Cretaceous of the Atacama
porphyry Cu-Mo district have not resolved absolute ages well fault zone in the Cordillera de la Costa; to the Late Cretaceous
except for the 41.8 to 41.2 Ma Re-Os ages of molybdenite at of the Central Valley fault zone; and to the Eocene Domeyko
Turquoise Gulch (Watanabe and Hedenquist, 2001; Zimmer- fault zone in the precordillera of Cordillera Domeyko (Fig.
man et al., 2014). 1). Jurassic back-arc rifting and associated basaltic magmatism
In this study, the SUMAC SHRIMP-RG was employed to occurred in the Central Valley. The Domeyko fault zone is
analyze zircon for U/Pb ages with weighted mean errors 1 to a north-south network of late Eocene reverse and strike-slip
1.5% (at 2s or 95% confidence) as well as to identify inher- faults that define the precordillera of the Andean arc (Elderry
ited zircons and those with postcrystallization Pb loss (e.g., et al., 1996).
Miller and Wooden, 2004). The SHRIMP-RG simultaneously The late Eocene to early Oligocene magmatic arc belt of
collected the abundances of rare earth elements, Hf, Y, Th, the Cordillera Domeyko hosts a series of porphyry copper
U, and Ti that robustly record magmatic processes related to deposits which extend from Peru to central Chile (Camus
production of porphyry ore fluids. This study tests the alterna- and Dilles, 2001; Sillitoe and Perelló, 2005). These porphyry
tive hypotheses that these ore-forming intermediate compo- deposits are time transgressive in step with the eastward
sition magmas may have formed via (1) crystal fractionation migration of the arc. The El Salvador porphyry Cu (Mo-Au)
from a water-rich mantle-derived basaltic melt in the lower or deposit is located in the Indio Muerto district part of the Cor-
middle crust (Kay and Mpodozis, 2001; Rohrlach and Loucks, dillera Domeyko in northern Chile approximately 330 km
2005), or (2) crustal assimilation, fractional crystallization, south of Antofagasta and 1,100 km north of Santiago. The
mafic recharge with open-system volatile loss within an upper deposit consists of multiple Eocene age porphyry intrusions
crustal magma chamber (cf. Field et al., 2005; Chambefort along a 6-km north-northeast trend (Fig. 2).
et al., 2008).
Accessory zircon crystallizes in nonalkaline intermediate Sampling Method and Zircon Analyses
to silicic magmas at up to 950°C and continually crystallizes Zircons were obtained from 15 samples representing all the
to 750°C or less (Finch and Hanchar, 2003; Hanchar and Eocene El Salvador porphyry intrusions and four samples
Watson, 2003; Miller et al., 2003). It typically forms ~100- to of older Paleocene and Permian igneous rock in the district.
200-mm-long grain with complex internal oscillatory zonation Seven zircon separates (IT-3, IT-6, IT-8, IT-9, IT-10, IT-14,
and in some cases has an inherited core. Zircon preferen- and IT-17) were obtained from Cornejo et al. (1997). Zir-
tially incorporates a suite of lithophile elements, including cons were separated using standard procedures from 12 new
middle and heavy rare earth elements, Hf, U, and Th in con- rock samples collected from the least altered and weathered
centrations dependent upon the pressure, temperature, and localities available (Table 1; App. 1). The U-Th-Pb and trace

EL SALVADOR ZIRCONS 247
70 Cerro Colorado [52-50]

O
O
20
Asuncion
Quebrada Blanca
DOMEYKO FAULT ZONE

[37-35]
Buenos
Santiago Aires
Rosario [34-32] Montevideo
Ujina [35]
AULT ZONE
El Abra [38-37]
O
22
Sierra Radomiro Tomic [34]
Gorda Chuquicamata [35-32]
ATACAMA F
Mansa Mina
Spence Toki [38] Sierra Miranda [36]
[58-57] Opache [37] Bolivia Argentina
Chile
Esperanza Salar de
[41] Atacama
Antofagosta Lomas
TA
Bayas
LA COS
A DOMEYKO
Gaby Sur [43-42] O

24
O
24
LLEY
Zaldivar
ERA DE
La Escondida [39-36]
CENTRAL VA
CORDILLER
L
CORDIL
Exploradora [33-31]
O
26
O
26
El Salvador
[44-41]
Potrerillos N
[37-35]
100km
Country Boundary
Copiapo Mineral deposit [Age Ma]
Fault
O Thrust Fault
70 Early to mid-Paleogene
Fig. 1. Tectonic map of northern Chile outlining major fault zones, cordilleras, and porphyry copper deposits. Major por-
phyry Cu-(Mo) deposits denoted by solid black squares with deposit/mineralization ages listed in brackets. Modified from
Cornejo et al. (1997), Taylor et al. (1998), and Camus and Dilles (2001).

248 LEE ET AL.
El Salvador Lithology Cerro Pelado
Eocene El Salvador Intrusions & Breccias

latite dikes
breccias (includes tourmaline,
andesitic, & contact breccias)
7,100,000
R porphyry
U/Pb Samples
A porphyry
QG porphyry Copper Ore Zones
Chalcopyrite
L porphyry (Cp)-Pyrite
Chalcopyrite
K porphyry -Bornite IT-3 - Rhy
Cerro Pelado 44.2 ± 0.6 Ma
aplite anomalous Mo
(+Pyrite+Cp)
X porphyry Old Camp
late quartz porphyry (late Qtz) ES-12808 - Qtz
43.6 ± 0.4 Ma
quartz porphyry (Qtz)
Paleocene-Eocene rhyolite vol- 7,099,000
canics, porphyries, & breccias
Cretaceous andesitic volcanic
& sedimentary rocks; local
Paleocene volcanic rocks ES-12791 - Late Qtz
43.6 ± 0.3 Ma
M Gulch
ES-12800 - L
43.8 ± 0.5 Ma
7,098,000
ES-12782 - R Copper Hill

43.3 ± 0.4 Ma
ES-12783 - A
42.8 ± 0.5 Ma
Red Hill
IT-14 - Sierra Castillo
290.9 ± 3.2 Ma
O Nose
7,097,000
ES-3239 - Qtz 43.2 ± 0.3 Ma
ES-12787 - L 42.4 ± 0.6 Ma
ES-12785a -K 42.9* ± 0.4 Ma

ES-12792 Latite
41.6* ± 0.5 Ma Turquoise Gulch
IT-9 - L 42.2 ± 0.5 Ma 7,096,000
IT-10 - X 43.6 ± 0.5 Ma
ES-12811 - X
43.9 ± 0.6 Ma
Sector Granito
ES-12789a - L IT-6 - Basal Tuff
43.9 ± 0.5 Ma 61.3 ± 0.8 Ma
ES-12807 -
K(G ppy?) Meters
42.3 ± 0.5 Ma
Granite Gulch 0 500 1000 7,095,000
IT-17 - qtz mo
444,000
446,000
445,000
61.0 ± 0.9 Ma IT-8 - Rhyolite

60.5 ± 0.5 Ma
Fig. 2. Geologic map of the Indio Muerto district, Chile, at the 2,600- to 2,660-m elevation delineating the rhyolite porphyry,
quartz porphyry, and granodiorite. Sample location, type, and interpreted U/Pb zircon ages (±2s, 95% confidence) from this
study denoted by solid black squares. Qtz = quartz porphyry, qtz mo = quartz monzonite, ppy = porphyry, and Rhy = rhyolite
porphyry. Modified from Gustafson et al. (2001) and compilations by CODELCO geologists. Grid coordinates in UTM WGS
84 zone 19.

Table 1. Interpreted Zircon 206Pb/238 U Weighted Mean Ages for El Salvador Porphyry Samples by SHRIMP-RG
UTM coordinates
Elevation Standard Quality Total Spots Age ±95%
Sample no. Rock type Location N E (m) plug (run) std spots used (Ma)1 2s CL MSWD
ES-12782 R porphyry M Gulch-Copper Hill, in pit 7097266 444306 2624 OSU-1 G 16 13 43.3 0.4 1.3
ES-12783 A porphyry M Gulch-Copper Hill, in pit 7097125 444568 2661 OSU-2(1) OK 18 15 42.8 0.5 1.9
ES-12785a K porphyry Turquoise Gulch, Inca adit 7096410 444410 2440 OSU-2(2) OK 8 6 42.9^ 0.4 1.0
ES-12787 L porphyry Turquoise Gulch, Inca adit 7096420 444390 2440 OSU-6 G 18 7 42.4 0.6 0.3
11 40.3 0.5 0.7
ES-12789a L porphyry Granite Gulch, surface 7094339 443412 2850 OSU-1 G 15 13 43.9 0.5 1.2
ES-12791 Late quartz porphyry Old Camp, in pit 7098730 444950 2711 OSU-2(1) OK 18 13 43.6 0.3 1.2
ES-12792 Latite dike Turquoise Gulch, Inca adit 7096245 443800 2476 OSU-2(2) OK 12 9 41.6^ 0.5 3.0
ES-12800 L porphyry M Gulch-Copper Hill, in pit 7097301 444806 2700 OSU-3 OK 12 10 43.8 0.5 1.9
ES-12807 K porphyry Turquoise Gulch DD-1367 269-286 m 7095770 443780 2384 OSU-3 G 20 14 41.8 0.5 0.7
12 42.3 0.5 1.0
ES-12808 Quartz porphyry Old Camp, in pit 7098940 445110 2650 OSU-1 G 13 11 43.6 0.4 1.8
ES-12811 X porphyry Turquoise Gulch DD-8480 168.1 m 7095950 443930 2574 OSU-3 G 13 9 43.9 0.6 1.3
ES-3239 Quartz porphyry Turquoise Gulch, underground 7096745 444610 2935 OSU-2(1) OK 15 8 43.2 0.3 0.6

5 44.7 0.3 0.5
IT-3 Rhyolite porphyry Cerro Pelado, surface (*) 7099339 446129 -- OSU-8 P 17 12 44.2 0.6 1.5
IT-9 L porphyry Turquoise Gulch, UG (*) 7096410 444240 2445 OSU-1 G 14 9 41.7 0.4 0.5
EL SALVADOR ZIRCONS
10 42.2 0.5 1.6

IT-10 X porphyry Turquoise Gulch, UG (*) 7096350 444250 2445 OSU-1 G 18 10 43.6 0.5 0.5

IT-6 Paleocene basal tuff Southeast of Turquoise Gulch (*) 7095200 445920 -- OSU-24 G 20 19 61.3 0.7 2.0
IT-8 Co Indio Muerto
Rhyolite dome West of Granite Gulch (*) 7094930 443790 -- OSU-24 G 15 15 60.4 0.6 1.2
IT-17 Paleocene quartz El Salvador caldera,
monzon La Canterra area (*) 7087610 442425 -- OSU-24 G 15 13 60.5 0.9 1.4
IT-14 Sierra Castillo batholith East of Turquoise Gulch and
Sierra Castillo fault (*) 7097134 454932 -- OSU-24 G 14 10 290.9 31 0.76
Notes: Ages using the same standards are directly comparible; comparison of ages between standards requires inclusion of errors in standards (0.5-1%); standard quality: G = good; OK = just accept-
able; P = poor (IT-3: OSU-8 use of 12 of 18 standards has 1% error, MSWD = 2.14); (*) = sample from Cornejo et al. (1997); others this study
1 Preferred mean age in bold; alternative mean age in italics (see text); ages marked with ^ are less robust and accurate than indicated by error (see text)
249
250 LEE ET AL.
element concentrations of zircons were analyzed using the A. IT-10 X porphyry

sensitive high-resolution ion microprobe-reverse geometry 171.2 +- 1.7 187.2 +- 1.7
141.8 +- 1.3
(SHRIMP-RG) instrument at the Stanford-U.S. Geological
Survey Micro Analysis Center (SUMAC). Zircon rims display-
ing regular concentric (oscillatory) growth zones were pref-
erentially analyzed and are interpreted to represent normal
41.4
magmatic crystallization from zircon-saturated melt (Vavra, +-
1994; Hoskin, 2000; Hoskin and Schaltegger, 2003). Areas of C
1.2 C
truncated or resorbed cores were avoided if possible as early
analyses indicated inherited grains some of which gave Juras-
sic ages (Fig. 3).
Major and trace element concentrations including Ti, REE, 43.1
+-
Y, and Hf were analyzed simultaneously with U-Th-Pb isotope 44.3 +- 0.5
1.2
analysis. A full description of analytical proecedure is given
43.2 +- 1.2 100 µm
in Appendix 1, with complete analytical data contained in
Appendices 2 to 6. B. ES-12782 R porphyry
Interpretation of U/Pb data 180.7
43.8 +-
Eight to 20 zircons were analyzed from each of the 15 42.6 +-
C 1.6
Eocene samples for a total of 244 spots from El Salvador. +- 43.0 0.3
0.7 +-
One run (20 spots) was discarded because of poor stan-
dardization, and nine additional zircons yielded Mesozoic 0.4
inherited ages. The remaining 215 individual 207Pb-cor- OG
rected 206Pb/238U ages have relatively uncomplicated spectra
of which >90% fall in the range of 46 to 40 Ma. For each
sample, we selected a preferred weighted mean age with
a 2σ error assigned from the standard error of the mean.
Terra-Wasserburg concordia plots and age probability plots
were used to identify age populations of zircons within a
43.7 +- 0.3 100 µm
single sample, and a preferred age was chosen on the basis
of (1) exclusion of samples with Pb loss, (2) exclusion of old C. IT-9 L porphyry
age populations, (3) acceptable MSWD, (4) consistency with 41.6 41.1 +- 0.6
40Ar-39 Ar and Re-Os ages, and (5) consistency with geologic
42.4 +-
crosscutting age relationships (Table 1). The Eocene age +- OG
0.6
data are presented in Terra-Wasserburg Concordia diagrams 0.7 OG
together with the preferred weighted mean average of the S
207Pb-corrected 206Pb/238U age with 2σ error (Fig. 4A-O). In S
these plots, filled circles were used to calculate the preferred
mean age. Some of these grains are slightly discordant, and S S
in these cases a 207Pb correction was applied using the plot-
41.3
ted regression lines (e.g., Fig. 4F-H, K-L) to calculate the
+-
intercept age on concordia. 0.6
In some samples, a few zircons yield ages that are younger
than the main population, for example, younger than the 42.8 +- 0.3
41 Ma age of the youngest intrusion of latite (Fig. 4N, O). 100 µm
These younger ages were not used in the mean age calcula-
Fig. 3. Cathodoluminescence images of selected zircon grains from three
tion as they potentially represent postcrystallization Pb loss, samples that illustrate oscillatory growth zoning (OG), sector zoning (S), and
most likely due to a district-wide event of supergene weather- older inherited cores (C). A. IT-10: X porphyry from Turquoise Gulch. B.
ing and acidic groundwater leaching between ~35 and 13 Ma ES-12782 R: porphyry from M Gulch. C. IT-9: L porphyry from Turquoise
(Mote et al., 2001). Gulch. Each circle marks the location of an individual ion microprobe analy-
In a few cases, there is a small population of zircons that sis with a spot diameter of 30 mm, and age in Ma with 1s error.
have ages slightly older than the preferred mean age. Such
older zircons may be antecrysts (Miller et al., 2007) that crys-
tallized from earlier Eocene intrusions and were incorporated porphyry samples we consider the older age population to be
into later intrusions. These zircons are indicated with the antecrystic (Fig. 4D).
star symbol in Figure 4. For one sample (IT-9, L porphyry), The data used to calculate the preferred weighted mean
four grains form an older population that may be antecrys- ages also produce acceptable mean sum of the weighted devi-
tic (Fig. 4M). For sample ES-3239 of quartz porphyry there ates (MSWD; Wendt and Carl, 1991). The MSWD are sta-
are two age populations of zircon (n = 5, n = 8), and because tistical measures of the ratio between observed and expected
the younger age is similar to the ages of the two other quartz scatter of the data. Where MSWD are near or equal to one

0.09
A. IT-3 Rhyolite porphyry CP B. ES-12808 Quartz porphyry OC
44.2 ± 0.6 Ma 0.12 43.6 ± 0.4 Ma
0.08 (MSWD = 1.5) (MSWD = 1.8)
207Pb/206Pb
0.10
0.07
0.08
0.06
0.06
0.05
48
56 52 44 40 36
0.04 50 48 46 44 42 40 38
0.04
110 130 150 170 190 125 135 145 155 165
0.08
C. ES-12791 Late quartz porphyry OC D. ES-3239 Quartz porphyry TG
0.07 43.6 ± 0.3 Ma 43.2 ± 0.3 Ma (n=8, MSWD = 0.6)
(MSWD = 1.3) 0.07
44.7 ± 0.3 Ma (n=5, MSWD = 0.5)
207Pb/206Pb
0.06
0.06
0.05
0.05
50 48 46 44 42 40 38
48 46 44 42 40 38
0.04
0.04
130 140 150 160 170 125 135 145 155 165 175
0.11
E. ES-12811 X porphyry 43.9 ± 0.6 Ma 0.22 F. IT-10 X porphyry TG
TG-SG (MSWD = 1.3) 43.6 ± 0.5 Ma
(MSWD = 0.5)
0.09
0.18
207Pb/206Pb
0.14
0.07
0.10
0.05
0.06
48 46 44 42 40 38
52 48 44 40
0.03 0.02
130 140 150 160 170 115 125 135 145 155 165
0.12
0.14 G. ES-12807 K porphyry SG H. ES-12785a K porphyry TG
41.8 ± 0.5 Ma (n = 14,MSWD = 0.7) 42.9 ± 0.4 Ma (MSWD = 1.0)
42.3 ± 0.5 Ma (n = 12,MSWD = 1.0)
0.12 0.10
207Pb/206Pb
0.10
0.08
0.08
0.06
0.06
0.04 50 38
46
42
48 46 44 42
0.02 0.04
120 130 140 150 160 170 180 132 136 140 144 148 152 156 160
238U/206Pb 238U/206Pb
Fig. 4. Terra-Wasserburg concordia diagrams showing U/Pb geochronologic data for all El Salvador samples (A-O) with
interpreted weighted mean age based on sample populations marked with filled circles (and in one case partially filled circles).
Open circles, including many with Pb loss, and stars, which likely represent antecrysts, are not included in mean age calcula-
tions. Dashed regression line assumes intercept with common Pb (207Pb/206Pb = 0.8383 ± 0.0001) with interpreted 206Pb/238U
intercept age given (cf. Schmitt et al., 2003). Error of ages and error ellipses all at 2s level.

252 LEE ET AL.
0.070
0.064 I. ES-12800 L porphyry MG
0.066
J. ES-12783 A porphyry MG
43.8 ± 0.5 Ma (MSWD = 1.9) 42.8 ± 0.5 Ma (MSWD = 1.9)
0.062
0.060
0.058
207Pb/206Pb
0.056 0.054
0.050
0.052
0.046 40 38
48 46 44
0.048 0.042 42
48 46 44 42 0.038
0.044
0.034
0.040 0.030
130 134 138 142 146 150 154 158 130 140 150 160 170
0.26
K. ES-12782 R porphyry MG
0.22
L. ES-12789a L porphyry GG
43.3 ± 0.4 Ma (MSWD = 1.3) 43.9 ± 0.5 Ma (MSWD = 1.2)
0.22
0.18
0.18
207Pb/206Pb
0.14
0.14
0.10
0.10
0.06 0.06
56 52 48 44 40 56 52 44 40
48
0.02 0.02
110 120 130 140 150 160 170 110 120 130 140 150 160 170
0.12
M. IT-9 L porphyry TG
0.08
N. ES-12787 L porphyry TG
42.2 ± 0.5 Ma (n = 10, MSWD = 1.6) 42.4 ± 0.6 Ma (n = 7, MSWD = 0.3)
0.10 41.7 ± 0.4 Ma (n = 9, MSWD = 0.5) 40.3 ± 0.5 Ma (n = 11, MSWD = 0.7)
207Pb/206Pb
0.07
0.08
0.06
0.06
0.05
46 42 36
48 46 44 42 40 44 38
0.04 40
0.04
0.02 0.03
130 140 150 160 135 145 155 165 175
ES-12792 Latite dike, TG 238U/206Pb

0.16
O. 41.6 ± 0.5 Ma
0.14 (MSWD = 3.0) Geographic Location of Samples (see Fig. 2)
CP = Cerro Pelado
207Pb/206Pb
0.12
OC = Old Camp
0.10 MG = M Gulch - Copper Hill
TG = Turquoise Gulch
0.08 SG = Sector Granito
0.06
GG = Granite Gulch
Classification of zircon geodata
0.04 48 46 44 42 40 38
Concordant age groups, used for mean age
0.02 Discordant ages, excluded from mean age
130 140 150 160 170
Likely antecrysts, excluded from mean age
238U/206Pb
Fig. 4. (Cont.)

the assigned errors are the only cause of scatter and one age caldera (Cornejo et al., 1997). The Eocene porphyries form
population may be assumed, whereas deviations greater or intrusive centers, from north to south, at Cerro Pelado, Old
less than one indicate under- or overestimation of analyti- Camp, M Gulch-Copper Hill, Turquoise Gulch-Red Hill,
cal errors (Ludwig, 2001, 2003). For 13 of 15 samples, the Sector Granito, and Granite Gulch. Based on the presence
selected population yields a relatively robust MSWD <2 (two of pyroclastic textures and breccias in the porphyry intrusions
samples yield MSWD of 2.3 and 3). Standardization ranges at Cerro Pelado, Gustafson et al. (2001) concluded that the
from fair to good for all samples except IT-3 (MSWD of stan- porphyries were emplaced at no more than a few hundred
dard = 2.14; Table 1). We note that the assigned error is only meters depth. These extremely shallow depths are consistent
robust for comparison within a single standardization, and with the emplacement of porphyries into Paleocene volca-
additional error must be assumed when comparing different nic rocks in the Indio Muerto district, and the presence of
standardizations (Table 1). low-temperature advanced argillic minerals such as alunite,
dickite, and pyrophyllite in the upper exposures at Turquoise
Geology and Age Interpretation of the Gulch (Watanabe and Hedenquist, 2001).
Indio Muerto District
Porphyry intrusions
Geology Rhyolite porphyry: Based on field relationships the oldest
The Indio Muerto district is centered on a series of late Eocene intrusions are rhyolite porphyry and quartz porphyry, the
(~44–41 Ma) granite (rhyolite) and granodiorite porphyry latter is inferred to be younger based on age dating despite
intrusions that were emplaced into a sequence of Mesozoic lack of relative age information from contacts (Cornejo et
to early Cenozoic volcanic and sedimentary rocks. The Sierra al., 1997; Gustafson et al., 2001). The quartz-sanidine rhyo-
Castillo fault of the Domeyko fault system juxtaposes an east- lite porphyry and breccia is exposed at Cerro Pelado (Fig. 2)
ern basement of upper Paleozoic granitoids against a west- and is referred to as quartz rhyolite porphyry after Gustafson
ern basement of Lower Jurassic to Lower Cretaceous basaltic and Hunt (1975). Hydrothermal alteration and sulfides are
and andesitic sequences with minor sedimentary interbeds, associated with the intrusion. Previous age determinations
Upper Cretaceous andesitic and sedimentary sequences, and on quartz rhyolite range from 45 to 43 Ma based on a U-Pb
Upper Cretaceous granitoid intrusions. Upper Cretaceous zircon age 43 ± 2 Ma and a whole-rock K-Ar age of 45.3 ±
igneous rocks are transitional into Paleocene (64–56 Ma) high 2.0 Ma (samples IT-3, ES-7458; Cornejo et al., 1997), and an
K calc-alkaline volcanic rocks that include ignimbrites and Rb-Sr isochron of six whole-rock specimens of 45.4 ± 1.4 Ma
rhyolitic domes related to the El Salvador caldera, andesite (Gustafson and Hunt, 1975).
to trachyandesitic lavas, and associated diorite to monzo- We reanalyzed 17 zircons from sample IT-3 of rhyolite
nite intrusions (Cornejo et al., 1994, 1997). After porphyry porphyry and obtained a weighted mean 206Pb/238U age of
magmatism and ore formation in the Indio Muerto district, 44.2 ± 0.6 Ma (n = 12; MSWD = 1.5; Fig. 5A). Although the
several younger porphyry intrusions, mostly 41 to 33 Ma, are 44.2 ± 0.6 Ma age is relatively robust, this age is difficult to
temporally associated with porphyry and skarn Cu-Mo and compare with other ages because of the poor R33 standard
younger epithermal Au mineralization in the Potrerillos dis- reproducibility.
trict located approximately 3 km to the south of El Salvador Quartz porphyry: Quartz-plagioclase granite porphyry,
(Marsh et al., 1997). Regional shortening and transpressive referred to here as quartz porphyry after Gustafson and Hunt
faults were active in the Late Cretaceous-Paleocene and again (1975), crops out as a series of dikes and plugs throughout
during the late Eocene-early Oligocene porphyry magmatism the district. The largest exposure occurs at Old Camp where
in both the Indio Muerto and Potrerillos districts (Cornejo the porphyry is associated with low-grade pyrite-chalcopyrite
et al., 1997). mineralization and sericitic alteration. The quartz porphyry
The geology of the Indio Muerto district is well under- is characterized by large phenocrysts of quartz (1- to 5-mm
stood on the basis of detailed mine-scale work by geologists of diam, Fig. 5A), plagioclase (1–3 mm), and altered biotite
the Anaconda Company at the El Salvador mine (Gustafson set in a fine-grained, aplitic quartzo-feldspathic groundmass
and Hunt, 1975), as well as district studies by Sernageomin (Gustafson and Hunt, 1975; Cornejo et al., 1997). Two ages
(Cornejo et al., 1994, 1997) and the Compania Minera de from intrusions at Old Camp yielded 43.8 ± 0.4 Ma (biotite
Cobre de Chile (CODELCO; Gustafson et al., 2001). The 40Ar-39Ar; Gustafson et al., 2001) and 42.3 ± 2.6 Ma (zircon
geologic map of the Indio Muerto district at the 2,600- to U-Pb; Cornejo et al., 1997).
2,660-m elevation (Fig. 2) shows porphyry rock units cor- Three samples of quartz porphyry were analyzed via
related by lithology and crosscutting field relationships SHRIMP-RG and have preferred ages of 43.6 ± 0.4, 43.6 ±
and named by Gustafson and Hunt (1975) and Gustafson 0.3, and 43.2 ± 0.3 Ma similar to the 40Ar-39Ar age. Two sam-
et al. (2001). It may be noted that these lithologic correla- ples were collected from the Old Camp copper mine. Eleven
tions may or may not be robust, for example, Cornejo et al. of 13 zircons were analyzed from hydrothermally altered and
(1997) interpreted, on the basis of radiometric ages, that the copper-mineralized sample ES-12808 and have a mean age of
granodiorite porphyry intrusions at M Gulch-Copper Hill 43.6 ± 0.4 Ma (MSWD = 1.8; Fig. 5B). Eighteen zircons were
were older than the texturally similar porphyry intrusion at analyzed from a younger quartz porphyry and associated brec-
Turquoise Gulch. In the map area of Figure 2, Eocene por- cia that cuts the altered quartz porphyry (ES-12808) at Old
phyries intruded Upper Cretaceous andesite lavas, breccias, Camp. Sample ES-12791 of the younger quartz porphyry has
and epiclastic rocks as well as overlying Paleocene rhyolite a mean age of 43.6 ± 0.3 Ma (n = 13, MSWD = 1.2; Fig. 5C).
domes and pyroclastic rocks associated with the El Salvador Fifteen zircon grains were analyzed from a quartz porphyry

254 LEE ET AL.
A. B.
Qtz Plg Qtz
Mfe
apl Bi
500 µm 500 µm
C. D.
Plg
Hbl
Bi Plg Plg
500 µm 500 µm
E. F. Plg
Plg
Bi
Hbl
Hbl
500µm 500 µm
Fig. 5. Photomicrographs of main porphyry types analyzed: apl = aplite, Bi = biotite, Hbl = hornblende, Mfe = mafic enclave,
Plg = plagioclase, Qtz = quartz. A. Quartz porphyry from Old Camp with quartz phenocrysts in a fine-crystalline aplitic
groundmass. Localized phenocrysts of biotite and plagioclase are present. Highly altered mafic enclaves were observed but
are rare. B. K porphyry from Turquoise Gulch, coarser grained groundmass of quartz and plagioclase with phenocrysts of
plagioclase, biotite, and hornblende. C. L porphyry from M Gulch, with coarse plagioclase phenocrysts in a fine-grained
groundmass. D. L porphyry from Turquoise Gulch composed of equigranular plagioclase, biotite, and hornblende in coarse
groundmass rich in quartz and alkali feldspar. E. A porphyry from M Gulch with phenocrysts of biotite and plagioclase in a
fine-grained biotite, hornblende, and plagioclase rich groundmass. F. Latite dike with large sieved plagioclase, hornblende,
and biotite phenocrysts in a pilotaxitic groundmass.
from Turquoise Gulch (ES-3239) of which 13 zircon ages ages of the rhyolite porphyry and quartz porphyry are analyti-
have a bimodal population distribution (Fig. 5D). We prefer cally indistinguishable within error, the latter have a younger
the larger population as the intrusion age of 43.2 ± 0.3 Ma (n age range.
= 8; MSWD = 0.6). The smaller population has an older mean Granodiorite porphyries at Turquoise Gulch: These younger
age of 44.7 ± 0.3 Ma (n = 5; MSWD = 0.44) that is older than granodiorite intrusions are found in the southern two-thirds
the ages of the two quartz porphyry ages from Old Camp, and of the district and are associated with copper deposits extend-
is indistinguishable from the 44.2 ± 0.6 Ma age of the rhyolite ing from Copper Hill on the north to Granite Gulch on the
porphyry. This older population of zircons in quartz porphyry south. The granodiorites are characterized by phenocrysts of
sample ES-3239 may be antecrysts inherited from older El plagioclase, hornblende, biotite, and locally minor amounts
Salvador magma such as rhyolite porphyry intrusions. Regard- of quartz. Granodiorites vary from nearly equigranular in
less, the new U/Pb zircon ages provide robust estimates of the center of stocklike intrusions to porphyries with up to
~43.6 to 43.2 Ma for quartz porphyry that are consistent with 50 vol % fine-grained (<0.5 mm) quartzo-feldspathic aplitic
previous but less precise ages. Furthermore, although the groundmass (e.g., L porphyry of Turquoise Gulch, Gustafson

and Hunt, 1975). Individual porphyries are distinguishable 127 Ma, similar to those reported for the X porphyry reported
based on distinctive textures and crosscutting relationships by Cornejo et al. (1997) and Tosdal et al. (2000). Nine of the
(Gustafson and Hunt, 1975). Based on crosscutting relation- 13 zircons from sample ES-12811 of the X porphyry from
ships in Turquoise Gulch, these porphyries include in order Sector Granito have a mean age of 43.9 ± 0.6 Ma (MSWD =
of decreasing relative age: the X porphyry (weakly porphy- 1.3; Fig. 5E). Previous analyses of sample IT-10 of X porphyry
ritic to equigranular granodiorite); K porphyry (porphyry with from the 2,445-m level of the Turquoise Gulch underground
aplitic quartzo-feldspathic groundmass, Fig. 5B); L porphyry mine yielded a 41.6 ± 1.2 Ma K-Ar biotite age and 41.8 ±
(texturally variable heterogeneous intrusions of equigranu- 2.3 Ma U/Pb zircon age (Cornejo et al., 1997). In our results,
lar granodiorite to granodiorite porphyry with large 0.5- to most zircons are slightly discordant (n = 9), but application of
6-mm-long plagioclase, Fig. 5C, D); and the A porphyry with a common lead correction yields a mean age of 43.6 ± 0.5 Ma
an abundant, dark, fine-grained 0.1- to 0.4-mm groundmass (n = 10; MSWD = 0.5) or alternatively, 43.4 ± 0.5 Ma (n = 12,
and dioritic composition (Fig. 5E). The A porphyry intrusions MSWD = 1.2; Fig. 5F). Our new ages suggest that the X por-
are closely associated with the L porphyry and locally occur phyry is ~43.9 to 43.4 Ma, or similar to the age of the quartz
as the late, outer, and more mafic margin of the L porphyry porphyry it intruded.
intrusion in the northwest part of Turquoise Gulch (Gustafson The two K porphyry samples from Turquoise Gulch yielded
and Hunt, 1975). Narrow W- to NW-striking latite dikes are slightly different apparent ages of ~43 and ~42 Ma. Twenty
the youngest Eocene intrusions at El Salvador. They postdate zircons were analyzed from sample ES-12807 collected from
all Cu (Mo-Au) mineralization and localize pebble dikes and a drill core between Turquoise Gulch and Granite Gulch (Fig.
argillic alteration. The latites contain about ~64 wt % silica 2) and have a mean age of 41.8 ± 0.5 Ma (n = 14; MSWD =
and phenocrysts of sieved and spongy plagioclase with abun- 0.7) or alternatively a mean age of 42.3 ± 0.5 Ma (n = 12,
dant melt inclusions, two amphibole populations (edenite and MSWD = 1.0) that is more compatible with other ages (Fig.
Ti pargasite), and biotite set in a pilotaxitic fine-crystalline 5G, Table 2). Sample ES-12785a was collected underground
groundmass (Fig. 5F). Based on petrography and lithochem- on the 2,440-m level of the Turquoise Gulch mine and yielded
istry the latites appear to be the product of mixing between a common lead-corrected intercept age of 42.9 ± 0.4 Ma (n =
mafic and silicic magmas (Lee, 2008). 6; MSWD = 1.0) after discarding the youngest two ages (Fig.
Granodiorite porphyries in other copper centers: The por- 5H).
phyries elsewhere in the district have been tentatively cor- L Porphyry: Six samples of L porphyries and the related A
related with the porphyries in the Turquoise Gulch center on and R porphyries were analyzed to test geographic and local
the basis of texture, petrology, and associated hydrothermal age span: one sample each of L, A, and R porphyry from M
features (Gustafson et al., 2001). At the M Gulch-Copper Hill Gulch-Copper Hill, two samples of L porphyry from Tur-
porphyry Cu(Mo) center (Fig. 2), the quartz, K, L, and A por- quoise Gulch, and one sample of L porphyry from Granite
phyries are similar in texture, composition, and relative ages Gulch (Fig. 2). Twelve zircons were analyzed from an L por-
to those found elsewhere in the district, but here Cornejo et phyry sample at M Gulch-Copper Hill (ES-12800), and the
al. (1997) reported an age range of ~44 to 43 Ma for L por- main population yielded an age of 43.8 ± 0.5 Ma (n = 10;
phyry that is older than the ~42 Ma ages reported for the L MSWD = 1.9) that is similar to the three ages obtained for
and A porphyries in Turquoise Gulch. On this basis, Cornejo quartz porphyries. Eighteen zircons were analyzed from the
et al. (1997) interpreted M Gulch-Copper Hill to represent A porphyry from M Gulch (ES-12783), and provide a mean
a porphyry intrusion center slightly older than the Turquoise age of 42.8 ± 0.5 Ma (n = 14; MSWD = 1.9; Fig. 5J). Sixteen
Gulch center. At M Gulch, copper mineralization is hosted zircons from the R porphyry at M Gulch (ES-12782) include
in quartz porphyry and adjacent andesite and cut by L and several discordant analyses but contain a main population
A porphyry (Gustafson et al., 2001). The A porphyry here is with a common lead-corrected weighted mean age of 43.3 ±
found on the western contact of the L porphyry and is more 0.4 Ma (n = 13; MSWD = 1.3; Fig. 5K). These new zircon
weakly mineralized than the L porphyry. The A porphyry ages from L, A, and R porphyries from M Gulch-Copper Hill
appears to represent a quenched, mafic margin of the L por- suggest intrusions are approximately 43.8 to 42.8 Ma, similar
phyry or a related and slightly younger intrusion. In the west to three previous 40Ar-39Ar and K-Ar ages on micas of 42.6,
part of the pit west of the L and A porphyry exposures, a 5- 43.1, 43.8 Ma but older than five ages of 41 to 42 Ma reported
to 10-m-wide feldspar porphyry dike termed the R porphyry by Cornejo et al. (1997) and Gustafson et al. (2001) for this
cuts the M Gulch breccia that is related to the L porphyry site.
(Gustafson et al., 2001). The R porphyry also appears to be The 15 zircons from L porphyry of Granite Gulch (ES-
closely associated with the L and A porphyries here. Although 12789a) mostly yield discordant ages. Following the removal
the contact is not exposed, the lower degree of alteration and of two younger ages, presumably due to Pb loss, common
finer grained quenched groundmass distinguishes the R por- lead-corrected intercepts have a mean age of 43.9 ± 0.4 Ma (n
phyry from the A porphyry. = 13; MSWD = 1.2; Fig. 5L). This age is nearly identical to the
X and K porphyries: New zircon ages of all four samples age of the L porphyry from M Gulch-Copper Hill.
of the granodioritic X and K porphyries are difficult to inter- Two samples of L porphyry have been analyzed from Tur-
pret due to the large range of individual zircon ages and large quoise Gulch. We have reanalyzed 14 zircons from L porphyry
analytical errors. Zircon ages from two samples of X porphyry sample IT-9 (Cornejo et al., 1997) from the 2,445-m level of
are mostly 44 to 43 Ma, but also include older Eocene ages the Turquoise Gulch underground mine. Zircons include both
(45–46 Ma) and Mesozoic ages. Mesozoic ages (n = 7) include concordant and discordant data with a bimodal grouping.
four Jurassic 187 to 171 Ma and three Cretaceous 141 to Excluding four zircons with ages of 44 to 45 Ma that may be

256 LEE ET AL.
Table 2. Zircon Saturation Temperatures Calculated from Whole-Rock Major Element Concentrations1
Sample ES-12782 ES-12783 ES-12785a ES-12787 ES-12789a ES-12791 ES-12792 ES-12800 ES-12807 ES-12795
Rock type R (ppy) A (ppy) K (ppy) L (ppy) L (ppy GG) Lt (qtz ppy) Latite dike L (ppy MG) K (ppy) X (ppy)
Major elements (wt %)

SiO2 65.90 60.31 69.47 66.88 63.41 71.71 63.71 65.35 68.30 60.79
TiO2 0.57 1.04 0.58 0.67 0.76 0.64 0.93 0.67 0.62 0.89
Al2O3 17.01 19.78 14.98 16.64 17.47 20.62 17.06 17.78 15.81 19.38
FeO* 4.14 6.70 2.42 3.00 5.67 1.05 4.08 3.85 3.30 2.82
MnO 0.15 0.03 0.01 0.03 0.03 0.00 0.07 0.03 0.01 0.01
MgO 1.44 2.55 1.32 1.31 1.77 0.12 1.72 1.73 1.56 2.00
CaO 0.44 1.69 2.77 3.87 3.54 0.31 5.42 1.86 4.26 5.79
Na2O 2.65 5.66 4.42 4.97 5.09 4.20 4.59 6.06 4.63 5.96
K2O 7.50 1.92 3.84 2.42 1.96 1.22 2.13 2.49 1.33 2.04
P2O5 0.21 0.34 0.18 0.21 0.28 0.13 0.28 0.19 0.18 0.32
LOI (%) 2.63 3.51 2.58 0.82 1.13 6.38 4.96 2.05 3.99 2.83
SO3 (≥) 0.35 0.18 1.01 0.20 0.10 0.17 0.09 0.12 1.45 2.09
Sum 99.23 98.63 99.47 99.45 98.99 99.76 99.36 98.93 99.43 99.23
Zr 111 139 104 121 133 119 127 125 113 152
Zr/Hf 33.4 38.0 34.2 32.4 34.6 33.6 34.8 35.3 32.4 39.7
M2 1.27 1.32 1.71 1.96 1.66 1.74 0.644 1.52 1.68 2.11
T (°C3) 764 779 729 727 752 739 820 757 737 722
1 Whole-rock major element concentrations determined by XRF, oxides normalized to volatile free
2 M value equals the cation ratio (Na + K + 2Ca)/(Al*Si) after the Watson and Harrison (1983) model
3 Zircon saturation temperature assuming Zr concentration and M value of whole rock as melt
4 Latite M value falls outside accepted range for calculation by model
antecrysts, the remaining zircons yield a weighted mean age long-lived, with episodic intrusions spanning at least 3 m.y.
of 42.2 ± 0.5 (n = 10; MSWD = 1.6; Fig. 5M). Alternatively, by from ~44 to ~41 Ma (Table 2). A histogram of all 224 ages of
removing a high U (1,031 ppm) zircon, we obtain a younger Eocene zircons analyzed in this study forms a distinct Gauss-
weighted mean age of 41.7 ± 0.5 Ma (n = 9; MSWD = 0.4) ian distribution (Fig. 6). The bin sampling in this diagram
that is similar to the 41.8 to 41.2 Ma Re-Os ages on molybde- (0.4 m.y.) is representative of the standard error of mean for
nite from Turquoise Gulch (Zimmerman et al., 2014). Sample the preferred mean ages (errors of 0.3–0.6 m.y., Table 2), and
ES-12787 was collected 150 m from sample IT-9 from the therefore the counts per bin should be relatively robust.This
same L porphyry on the 2,440-m level. Eighteen zircon ages distribution of zircon crystallization ages suggests that magma-
from ES-12787 are bimodal with mean ages of (1) 40.3 ± tism in the district may have initially begun at ~46 to 45 Ma,
0.5 Ma (n = 11; MSWD = 0.7, and (2) 42.4 ± 0.6 Ma (n = ramped up at about 44 Ma, reached a peak between 43.5 and
7; MSWD = 0.3; Fig. 5N). The younger population includes 43 Ma, then declined, and ended by about 41.6 Ma. One pos-
several discordant ages and has a mean age younger than the sibility is that deep-sourced mafic magma intruded into the
11 previous K-Ar and 40Ar-39Ar ages from Turquoise Gulch shallow crust periodically and initiated partial melting, and
(cf. Gustafson et al., 2001, fig. 8), so we here interpret these also reheated and locally mixed with previously emplaced dif-
ages as being too young because of Pb loss.The preferred 42.4 ferentiated dacitic magmas.After each intrusion, subsequent
± 0.6 Ma age is in agreement with the previous determined cooling of both mafic and silicic melts caused crystallization
ages and our U/Pb zircon age from L porphyry sample IT-9. of new zircon to produce the observed age spectrum (Fig. 6).
Latite dike: Twelve zircons were analyzed from latite dike Cerro Pelado: Rhyolite porphyry dikes, with marginal intru-
sample (ES-12792) from the Inca Oeste area of the Turquoise sive and hydrothermal breccias are spatially associated with
Gulch underground mine, and most are discordant ages to pyroclastic rocks in the district and are dominant at Cerro
which a common lead correction has been applied. Three zir- Pelado (Fig. 2). Gustafson et al. (2001) suggested rhyolite
cons form a young population with an age of 39 Ma that likely breccias with pyroclastic textures were emplaced no more
results from Pb loss. The remaining zircons have mean age than a few hundred meters of the Eocene surface, and prob-
of 41.6 ± 0.5 Ma (n = 9; MSWD = 3.0; Fig. 5O). Although ably represent ignimbrite vents such as those documented
this age is not robust (MSWD >2), it is consistent with previ- at Butte, Montana, and at the Gabbs Valley Range, Nevada
ous age determinations on igneous biotite from latite dikes: a (Proffett, 1973; Ekren et al., 1980; Houston and Dilles, 2013).
K-Ar age of 42.0 ± 1.0 Ma (Gustafson and Hunt, 1975) and Rhyolite porphyry and breccia here is associated with strong
a 40Ar-39Ar isochron age of 41.16 ± 0.48 Ma (Gustafson et al., sericitic alteration and pyrite-chalcopyrite-molybdenite min-
2001). eralization that contains <0.1 wt % Cu and ~200 ppm Mo.
The U/Pb zircon age for rhyolite of 44.2 ± 0.5 Ma reported
Age of the El Salvador Porphyry Centers and here is analytically identical to the 40Ar-39Ar age of 44.13 ±
Timing of Mineralization 0.40 Ma on hydrothermal sericite reported by Gustafson et al.
The new age data from the El Salvador porphyry Cu-Mo (2001), and we interpret that the weak Mo-Cu mineraliza-
deposit suggest that the magmatic system was relatively tion center is related to early rhyolite porphyry. So far, no

Probability density plot for El Salvador samples
N = 224 Peak magmatism:

M Gulch-Copper Hill, Granite Gulch
L, A, & R porphyries,
X porphyries & minor Cu ores (age uncertain)
25
Quartz porphyry & Old Camp,
minor Cu ores Waning magmatism:
K & L porphyries at Turquoise
Gulch with main Cu(Mo) ores
Early magmatism:
20 Rhyolite porphyry
Relative probability
& minor Mo at Cerro Pelado
Final magmatism:
Latite dikes
15
Number
Pb-loss ages
Cryptic earliest magmatism(?) related to
recorded as antecrysts supergene acids
10
0
48 47 46 45 44 43 42 41 40 39 38
206 238
Pb/ U Age
Fig. 6. Probability density plot with a bin of 0.4 m.y. of all 224 U/Pb zircon ages from the 15 El Salvador porphyry samples,
excluding Mesozoic inherited ages. Individual spots have 1 to 4% 1s errors (0.4–1.6 Ma). Labeled events represent the best
interpretations of all age data from this study and are consistent with Figure 7.
intrusions older than the rhyolite porphyry have been found, and contact breccia, L porphyry, and associated A porphyry
but six individual zircons have Eocene ages 46 or older Ma dikes were interpreted as late intramineralized at M Gulch-
may be antecrysts derived from earlier magmatism (~46 Ma) Copper Hill. They have truncated A veins containing copper
in the district (Figs. 6, 7). These ages may not be distinctly sulfides (<0.2 wt % Cu) and biotite-K-feldspar alteration in
older on the basis of analytical uncertainty. older quartz porphyry and are in turn cut by B quartz-molyb-
Old Camp: Located ~750 m south of Cerro Pelado, the denite veins and D veins with sericitic alteration (Gustafson
Old Camp open pit contains a quartz porphyry associated and Hunt, 1975; Gustafson et al., 2001). R porphyry postdates
with pyrite-chalcopyrite sulfides (>0.7 wt % Cu with a diam- Cu-(Mo) ores and does not contain Cu-Mo sulfides, but is
eter of 250 m at the 2,600-m level) associated with early weakly sericitically altered. Both Gustafson et al. (2001) and
quartz-sulfide A veins and K-feldspar-biotite alteration at Cornejo et al. (1997) suggested that M Gulch-Copper Hill
depth that grades upward to sericite-chlorite alteration.The is a pre-Turquoise Gulch center of mineralization, perhaps
quartz porphyry sampled for this study (ES-12808) contains related to the quartz porphyry and granodioritic L porphyry.
quartz veins, sericitic alteration, and pyrite-chalcopyrite min- The 40Ar/39Ar ages of 42.60 ± 0.18 Ma (igneous biotite from
eralization, and is clearly crosscut by a late quartz porphyry quartz porphyry) and 42.85 ± 0.52 Ma (sericite) suggest these
(ES-12791) and associated breccia that truncates the quartz rocks cooled through ~325°C at ~42.6 Ma (Gustafson et al.,
veins and contains only weak supergene clay alteration. We 2001). Our U/Pb zircon ages of igneous intrusions from M
obtained identical ages of 43.6 ± 0.4 and 43.6 ± 0.3 Ma from Gulch are consistent with the 42.6 Ma 40Ar/39Ar age. The ages
the main mineralized quartz porphyry and the late quartz por- of L porphyry (43.8 ± 0.5 Ma) and R porphyry (43.3 ± 0.4 Ma)
phyry, respectively.Thus, the age of Cu mineralization at Old are slightly, but distinctly, older whereas A porphyry (42.8 ±
Camp is ~43.6 Ma (Fig. 7). 0.5 Ma) has an identical age.
M Gulch-Copper Hill: The granodiorite L porphyry is asso- Granite Gulch-Sector Granito: In the south part of the
ciated with weak Cu mineralization at M Gulch. Hydrothermal district, weak Cu sulfide mineralization has been found in

258 LEE ET AL.
Relative age based on cross- U/Pb zircon mean age (this study) Ar-39Ar biotite (Gustafson et al., 2001)
40
cutting geologic relations preferred (2σ) U/Pb zircon age (Cornejo et al., 1997)
alternative (2σ) CT & MD Re-Os age (Zimmerman et al., 2014)
IT-3 Rhyolite ppy CP Pb-loss 8-
Old Camp
ES-12808 Quartz ppy OC Pb-loss 1**

Errors on individual spots (1σ) 2(1)*
ES-12791 Late quartz ppy OC
ES-3239 Quartz ppy TG 2(1)*
IT-10 X ppy TG 1**

ES-12811 X ppy TG 3**
M Gulch (MG)
ES-12800 L ppy MG ? 3**

ES-12783 A ppy MG 2(1)*
ES-12782 R ppy MG 1**
ES-12789a L ppy Granite Gulch 1**
?
2(2)*
ES-12785a K ppy TG
3**
Turquoise Gulch (TG) center
ES-12807 K ppy TG
B vein (Re-Os age) TG Age of antecrysts(?)

D vein (Re-Os age) TG
IT-9 L ppy TG 1**
ES-12787 L ppy TG U/Pb age range con- 6**
sistent with geology
ES-12792 Latite dike TG Porphyry Cu-Mo Pb-loss 2(2)*
ES-12152 Latite dike TG mineralization
2(2) Plug(run). Standard quality: 47 46 45 44 43 42 41 40 39 38

** = good,* = OK; - = poor Isotopic Age (Ma) with error
Fig. 7. Summary of El Salvador geochronology, based on the best isotopic ages of igneous rocks. Shaded bars show an inter-
nally consistent set of preferred ages based on geochronology and geologic relative age relationships as discussed in text. CP
= Cerro Pelado, GG = Granite Gulch, MG = M Gulch, OC = Old Camp, ppy = porphyry, TG = Turquoise Gulch.
granodioritic X and K porphyries, but largely predates the rel- Several interpretations of the new data are possible. The
atively unaltered L porphyry that has 40Ar/39Ar ages of 42.26 ± L porphyries from M Gulch and Granite Gulch are younger
0.24 and 42.88 ± 0.18 Ma for hornblende and biotite, respec- than the quartz and X porphyries, and represent evolution of
tively (Gustafson et al., 2001). The granodioritic X porphyry these magmas. Thus, they were likely emplaced slightly after
sampled from Sector Granito yielded an U/Pb zircon age of ~43.6 Ma and are probably older than the L porphyry from
43.9 ± 0.6 Ma but given the field relationship that X porphyry Turquoise Gulch (cf. Cornejo et al., 1997). Alternatively, the
cuts the ~43.6 Ma quartz porphyry throughout the district, ages of 42.8 ± 0.5 and 43.3 ± 0.4 Ma, for the A and R porphy-
we interpret that X porphyry is slightly younger than 43.6 Ma. ries, respectively, which on a geologic basis are likely cogenetic
Our U/Pb zircon age of 43.9 ± 0.4 Ma for the L porphyry at with the L porphyry at M Gulch-Copper Hill, may provide a
Granite Gulch is also slightly older than the Ar ages and field best estimate of ~43.1 Ma for porphyry intrusion at this local-
relationships, but is somewhat less robust because most of the ity. In this case, the older zircon ages from the L porphyries
zircons in this sample are discordant (Fig. 5K). could be due to either analytical error or long residence time
Although the ages of the L porphyries and X porphyry of the L porphyry magmas at low temperature (~700°C).
overlap within uncertainty, we conclude that our 43.9 ± 0.4 The ~43.1 Ma age agrees with the slightly younger 40Ar/39Ar
and 43.8 ± 0.5 Ma U/Pb age estimates for the L porphyry cooling ages of ~42.8 to 42.6 Ma (Gustafson et al., 2001). In
from both M Gulch and Granite Gulch are probably too old this case, the L, A, and R porphyries and Cu mineralization
for emplacement ages. These intrusions have cut both the at M Gulch-Copper Hill, and perhaps Granite Gulch-Sector
~43.6 Ma quartz porphyry and the younger X porphyry. The Granito, are likely to be intermediate in age between the min-
trace element composition of zircons for both L porphyries eralization ages at Old Camp and Turquoise Gulch.
differ significantly from the zircon compositions from quartz Turquoise Gulch: Gustafson and Hunt (1975) described the
and X porphyry samples, so the zircons from the L porphyries main orebody at El Salvador, where the quartz and X porphy-
are not antecrysts. ries were emplaced prior to mineralization. About two-thirds

of the copper was introduced together with K feldspar-bio- zircon saturation temperatures. Hanchar and Watson (2003)
tite-anhydrite alteration and granular quartz-K feldspar A suggested these factors provide a good first-order tempera-
veinlets during intrusion of the K porphyry complex. The ture approximation although other factors including oxygen
intramineralized L porphyry was emplaced after most quartz- fugacity, halogen content, and Fe and Mg concentration
molybdenite B veins, but is weakly altered and cut by some may affect zircon solubility (e.g., Keppler, 1993; Baker et
quartz-molybdenite veins and polymetallic D veins with ser- al., 2002). Indeed, the M value for the latite dike sample lies
icitic selvages. Latite dikes postdate all mineralization and outside of the calibration range for the saturation model and
are associated with tourmaline-bearing pebble dikes and clay yielded a calculated temperature of 820°C that is higher than
alteration. would be expected based on the experimental calibrations.
A sample of premineral quartz porphyry from Turquoise This and other latite samples are interpreted to be the prod-
Gulch contains a population of zircons with an age of 43.2 uct of mixing of hot mafic magma with cooler granodiorite
± 0.3 Ma (n = 8) and represents the oldest exposed Eocene magma on the basis of multiple amphibole populations (Lee,
intrusion here. The sample also contains a smaller popula- 2008), sieved plagioclase feldspars (e.g., Coombs et al., 2000;
tion of zircons with an age of 44.7 ± 0.3 Ma (n = 5) similar to Fig. 5F), and a large range of Ti-in-zircon temperature esti-
the age of rhyolite porphyry (IT-3), suggesting these zircons mates, and so we have discarded the zircon saturation tem-
are antecrysts derived from older unexposed intrusions. The perature estimates for these rocks. Saturation temperatures
X porphyry which cuts quartz porphyry at Turquoise Gulch for the granodiorite samples varies from 780° to 730°C, with
yielded a U/Pb age of 43.6 ± 0.4 Ma. The similar ages of the a median temperature of 748°C, and suggests that most El
quartz and X porphyries at Sector Granito and Turquoise Salvador granodiorites were zircon undersaturated when
Gulch, and field relationships at M Gulch-Copper Hill, sug- emplaced into the upper crust at >800°C.
gest a close genetic and temporal relationship of these two
porphyries. The age of the quartz and X porphyry correspond Titanium-in-zircon temperature estimates
to maximum occurrence of U/Pb zircon ages in the district For comparison of trace elements of Eocene zircons, we culled
illustrated in Figure 6, which potentially defines the peak grains with apatite inclusions as determined by enriched
magmatic fluxes. LREE contents (Hinton and Upton, 1991; Sano et al., 2002)
Zimmerman et al. (2014) obtained Re-Os ages of hydrother- and grains where 206Pb/238U ages were not robust (i.e., Pb loss,
mal molybdenite that suggests mineralization at Turquiose discordance, and error >3s).
Gulch occurred from 41.8 to 41.2 Ma. Our U/Pb determina- Titanium and full trace element concentrations were deter-
tions for two K porphyry samples, although not very robust, mined for zircons from samples of K porphyry, A porphyry, L
yield interpreted ages of 42.9 ± 0.4 and 42.3 ± 0.5 Ma, and porphyry, and latite dike from Turquoise Gulch (App. 3). We
our two L porphyry samples yield interpreted ages of 42.4 ± used the Ti-in-zircon thermometer of Watson and Harrison
0.6 and 42.2 ± 0.5 Ma. The inclusion of antecrystic zircons in (2005) and Watson et al. (2006), corrected to an activity of
the populations used to calculate the mean ages, as noted for TiO2 ≈ 0.7, to reflect titanite and titanomagnetite saturation
porphyry intrusions at Bingham, Utah, and Bajo Alumbrera, (Claiborne et al., 2006; Ferry and Watson, 2007; Walker et al.,
Argentina (von Quadt et al., 2011), implies that our mean ages 2013). Uncertainty in the activity of TiO2 by ± 0.1 introduces
are maximum estimates. All four U/Pb zircon ages are within an error ±25° to 30°C over the calculated temperature range.
error of one another and (except for the 42.9 Ma age) of the
Re-Os sulfide age, and strongly support emplacement of the
K and L porphyries and associated sulfide mineralization dur- 1.2
A. Late quartz ppy
ing a relatively short interval of <1 m.y. (Fig. 7). As noted
TiO2 wt. %
1.0
above, the U/Pb age on the postmineral latite dike is not very X ppy
robust (41.6 ± 0.5 Ma) but is corroberated by an independent 0.8
K ppy
cooling age provided by the 40Ar-39Ar age of 41.16 ± 0.48 Ma 0.6
on biotite (Gustafson et al., 2001). L ppy MG
0.4
A ppy
Zircon Trace Element Composition 42
B. R ppy
40
Zircon saturation for El Salvador porphyries L ppy GG
38
Zircon saturation temperatures from each porphyry intru- L ppy TG
sion (Watson and Harrison, 1983; Hanchar and Watson, 2003)
Zr/Hf
36
were calculated from selected samples based on whole-rock Latite dike
major element compositions determined by XRF analyses 34 ppy = porphyry
MG = M Gulch
(Table 2; Lee, 2008). Titanium content generally decreases 32 GG = Granite Gulch
with increasing silica content in porphyries (Fig. 8A). Zr/Hf TG = Turquoise Gulch
ratios vary between 30 and 40 for porphyries (Fig. 8B), simi- 30
lar to cumulate rocks that represent the low-temperatue end 58 60 62 64 66 68 70 72 74
stages of repeated magma injection followed by crystallization SiO2 wt. %
(i.e., Clairborne et al., 2010). Fig. 8. Whole-rock variation diagram for El Salvador porphyry samples plot-
The silica content of whole rock and peraluminousity “M ting SiO2 wt % versus A. TiO2 wt %. B. Zr/Hf ratio. Major oxides normalized
factor” of Harrison and Watson (1983) were used to estimate to volatile free. Data referenced from Table 2.

260 LEE ET AL.
Melts saturated in one or more Ti bearing phases will have the preserved due to the slow diffusion rate of Ti in zircon (Cher-
activity of TiO2 buffered, but nonetheless the activity cannot niak and Watson, 2007). Whereas the temperatures deter-
be estimated precisely based on present data (Claiborne et al., mined here are estimates for crystallization temperatures, our
2006). The L porphyry and latite samples contain both titanite results suggest the choice of spot location must be taken into
and ilmenite, and we infer that both the K and A porphyries consideration when analyzing zircon samples.
that are hydrothermally altered, and now lack titanite and Samples ES-12787 (L porphyry) and ES-12792 (latite dike)
ilmenite, originally contained these phases. The calculated have large temperature variations from 877° to 657° and 850°
temperatures for the El Salvador zircons range from 877° to to 631° C, respectively. About half of the Ti-in-zircon temper-
630°C (App. 3). atures for the L porphyry exceed the zircon saturation tem-
Sample ES-12785a (K porphyry) yielded temperatures from perature of 727°C, suggesting the high-temperature zircons
750° to 671°C in agreement with zircon saturation tempera- did not form from granodiorite melt and instead are ante-
tures for these rocks. Temperatures of cores were slightly but crysts derived from an earlier melt. Most Ti-in-zircon tem-
distinctly hotter (750°–711°C) than the rims (728°–671°C), peratures decrease from core to rim consistent with cooling
suggesting zircon growth during cooling as has been recorded during crystallization, but some zircons have reversals or high-
in other studies of granites (Miller et al., 2006). Zircons display temperature rims that indicate magma reheating, also noted
sector zoning in CL images, and for each of the three zircons by Clairborne et al. (2006). The large range of Ti-in-zircon
we analyzed the core as well as spots from identical growth temperatures, exceeding 200°C, together with the presence
zones but different crystallographic faces (Fig. 9). The Ti and of sieved plagioclase and two populations of amphiboles in
Hf contents are lower within the bright CL (100) sector zones the latite (Lee, 2008), strongly support origin of these rocks
and yield temperatures 18° to 12°C, lower than that of the via injection of high-temperature (>800°C), possibly mafic
corresponding (101) growth zones. Nonetheless, REE con- magma within a low-temperature (~700°–750°C) granodio-
tents are similar in both the (100) and (101) crystallographic rite magma.
zones. Watson and Liang (1995) suggested that sector zoning The Ti-in-zircon temperature allows quantification of the
is dependent on growth rate and diffusivity of elements along trace element variations of zircon as a function of tempera-
different crystallographic lattices. If Ti is enriched along a par- ture. The Hf content in zircon is an effective proxy for the
ticular growth surface during crystallization, this difference is crystallization of zircon in a melt and increases with cooling
(Pupin, 2000; Claiborne et al., 2006; Watson et al., 2006,
Wooden et al., 2006). For El Salvador zircon, as Ti content
A. 4 15 ES-12785a and temperature decrease the rare earth element (REE) and
K porphyry Y contents also decrease but the Hf content increases (Fig.
5
10A). Furthermore, as Ti-in-zircon temperature decreases,
1 2
the Th/U ratio decreases (Fig. 10B), and the ratio of heavy
REE to middle REE, as exemplified by the Yb/Gd ratio,
7
increases (Fig. 10C).
11 Trace element geochemistry of zircon
The trace element content of zircon from the El Salvador dis-
16
10 trict displays distinct differences between older and younger
12 8 intrusions (App. 4). The simultaneous collection of the age
and trace element composition of zircon allows for the assess-
B. ES-12792
2 3 Latite dike 11
ment of the geochemical evolution of magmas by recognizing
1 those xenocrystic or antecrystic zircons that are present within
4 5
the host intrusion. Zircons from three samples of Paleocene
volcanic rocks and one sample of Paleozoic granite from the
Indio Muerto district were additionally analyzed for trace
6 8
element composition to compare with the Eocene intrusions
7 9 of the El Salvador district and antecrystic zircons contained
12 therein (App. 4).
The REE concentrations were normalized to chondrite
values of Anders and Grevesse (1989) multiplied by 1.3596
(Korotev, 1996; Mazdab and Wooden, 2006; Lee, 2008).
Fig. 9. Cathodoluminescence images of El Salvador zircon grains with cor- The REE diagrams for Eocene zircons have a typical con-
rected temperatures based on Ti content (±2°C). Numbers correspond to
zircon grain sample number (App. 2). A. Zircons from sample ES-12785a
vex-upward pattern for zircon that includes enriched heavy
show sector zoning and rounded cores. Core temperatures vary from 750° to REE (HREE), depleted light REE (LREE), a large positive
711°C, whereas rim temperatures vary from 728° to 671°C. The (100) sector Ce anomaly (CeN/CeN*), and a modest to significant nega-
zones contain less Ti and are 12° to 18°C cooler than corresponding non- tive Eu anomaly (EuN/EuN*; where “N” indicates chondrite-
(100) sector rims. Hf and Th concentration are also lower within these (100) normalized values; Fig. 11; i.e., Ballard et al., 2001; Hoskin
sector zones, however, no other trace element is depleted in these zones.
B. Sample ES-12792 shows a temperature range from 850° to 631°C, with and Schaltegger, 2003; Hanchar and van Westrenen, 2007).
a higher temperature range of 850° to 750°C occurring within the cores of Individual samples have large ranges of REE contents, for
the grains. example, middle REE (MREE) contents varying by factors

900 early rhyolite porphyry and quartz porphyry samples have a

A. K ppy cores
REE pattern with a relatively large negative Eu anomaly and
K ppy (100) sectors
850 a strongly curved and convex upward MREE to HREE pat-
K ppy rims
A ppy cores tern (Fig. 11B, C). The rhyolite zircons have the highest REE
800
Temp corr. C
A ppy rims contents and quartz porphyry zircons range to concentrations

o
L ppy cores as little as 1% of those from the rhyolite. In a few samples, this
750 L ppy rims
same pattern is evident in inherited zircons, for example, the
Latite cores
Latite rims 43 to 44 Ma L porphyry from M Gulch-Copper Hill contains
700
an older 46 Ma zircon that has a similar REE pattern to the
650 quartz porphyry samples (pale gray line, Fig. 11D).
The second REE pattern characterizes the younger K por-
600 phyry and L porphyries, the intrusions most closely associated
6000 8000 10000 12000 14000 16000 18000 with the bulk of copper mineralization. These zircons have a
Hf ppm relatively low REE content that varies little, a small negative
900 Eu anomaly, and a relatively straight MREE to HREE pat-
B. tern with only a slight curvature (Fig. 11F, G).
850 The third style of REE patterns is observed for zircons from
the X porphyry (ES-12811), latite dike (ES-12792), and R
800 porphyry. These samples show a large range of REE contents
Temp corr. C
o
(up to a factor of 100). Furthermore, a bimodal compositional

750
distribution is evident in the X porphyry in which the popu-
700
lation with low REE contents has a similar pattern to the L
El Salvador porphyry zircons of group 2 (Fig. 11E). The zircon population
650
porphyry (ppy) with high REE contents has a similar pattern to zircons from
& latite intrusions the rhyolite porphyry, but has higher contents of LREE and
600 MREE. The REE-rich zircons of the X porphyry and latite
0.0 0.5 1.0 1.5 2.0 2.5 3.0 also have the highest Th/U and lowest Yb/Gd ratios relative to
Th/U the other samples (Fig. 12E), as well as high concentrations
900
C. of Th, U, Eu, and Yb (Fig. 12A-C). As noted above, the latite
850
shows evidence for magma mixing, as does the R porphyry,
and the bimodal zircon populations are supportive of crystal-
800 lization from two magma reservoirs, one REE-poor similar to
L porphyry, and one with much higher REE contents.
Temp corr. oC
750 As noted previously, Hf content increases with decreasing

temperature and magma crystallizaton and ranges from 7,000
700 to 14,500 ppm for all the samples (Fig. 13, App. 4). As Hf
increases, the magnitute of the Ce anomaly (CeN/CeN* >1)
650 increases and Eu anomaly (EuN/EuN* <1) increases slightly
(Fig. 13E, F). These reflect highly oxidized magmatic condi-
600 tions (Ballard et al., 2001; Dilles et al., 2015).The 15 samples
0 5 10 15 20 25 30 35
Yb/Gd of the Eocene intrusions display some distinctive REE, Th,
and U compositions but also vary considerably as illustrated
Fig. 10. Trace element content vs. Ti-in-zircon temperature (°C) illustrating
core, rim, and (100) sector zonation from the K porphyry (ES-12807), A por- in Figures 12 and 13.
phyry (ES-12783), L porphyry (ES-12787), and latite dike (ES-12792), ppy
= porphyry. Note large temperature and trace element range for latite and L Discussion: Temporal Evolution of Magmas
porphyry. A. Hf ppm vs. T. B. Th/U vs. T. C. Yb/Gd vs. T.
Sources of porphyry magmas
The diversity of zircon compositions attest to multiple pro-
of <10 to 100. The REE patterns of Eocene zircon antecrysts cesses that contribute to the El Salvador porphyry magmas.
in the El Salvador samples are similar to the patterns from The ages of the inherited zircons (n = 9 of 233, or 4%) are
Paleozoic and Paleocene zircons, which show a wide range of Jurassic (187–173 Ma, n = 5) and Cretaceous (141–68 Ma,
REE concentration and significant Eu anomalies relative to n = 4). Paleozoic inherited zircons were not analyzed dur-
the Eocene age grains (Fig. 11A). ing this study but were reported in previous analyses of the
Three REE patterns are evident within the Eocene age X porphyry IT-10 (Tosdal et al., 2000; App. 2). A Paleozoic
zircons. The zircons from samples of early rhyolite porphyry age (290.9 ± 31 Ma) granite of the Sierra Castillo batholith
(IT-3, Fig. 11C), quartz porphyry (Fig. 11B), X porphyry (Fig. located immediately east of the Sierra Castillo fault is a poten-
11E), and the late latite dike (Fig. 11H) have a large range tial source of Paleozoic zircon (App. 5). The inherited zircons
of REE abundances (by factors of 20–100) compared to the are particularly abundant in the X porphyry (7 of 31 zircons).
small range of REE content (less than a factor of 10) for zir- Therefore, Mesozoic crust was a principal contaminant of
cons from the L porphyry samples (Fig. 11). Grains from the Eocene magmas, consistent with the position of El Salvador

262 LEE ET AL.
10000
A. B.
1000 60-62 Ma tuff & granites
100
zircon/chondrite
10
292 Ma granite
1 Zircon Antecrysts
190 - 170 Ma
140 Ma
0.1
127 Ma ES-12808
68 Ma Quartz porphyry
0.01
C. IT-3 Rhyolite D.
1000 porphyry
zircon/chondrite
100
10
0.1
ES-12791 ES-12800
Late quartz porphyry L porphyry M Gulch
0.01
E. F.
1000
zircon/chondrite
100
10
0.1
ES-12811 ES-12785a
X porphyry K porphyry
0.01
G. H.
1000
zircon/chondrite
100
10
0.1 IT-9
L porphyry ES-12792
Turquiose Gulch Latite dike
0.01
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
REE REE
Fig. 11. Rare earth element (REE) plots for selected samples analyzed by SHRIMP-RG. A. Fields of Paleocene volcanics
and Paleozoic Sierra Castillo batholith zircons with Mesozoic/Paleocene antecrysts from El Salvador intrusions. B. Quartz
porphyry sample with abundant REE and with moderate negative Eu anomaly. C. Late quartz porphyry sample shows large
negative Eu anomaly; shaded pattern is range of rhyolite porphyry (IT-3). D. X porphyry with bimodal REE patterns similar
to quartz porphyry and L porphyry. E. L porphyry from M Gulch. F. and G. K porphyry and L porphyry from Turquiose
Gulch with characteristic low REE abundances and small negative Eu anomalies. The two outlying patterns in the L porphyry
IT-9 represent zircon cores (antecrysts) with older Eocene ages of 42.9 ± 1.1 and 43.9 ± 1.2 Ma. H. Latite dike sample shows
the highest variation between individual grains.The rhyolite porphyry, K porphyry and latite samples were analyzed with the
full REE suite (App. 3), whereas the other samples were analyzed during the U/Pb isotope analyses (App. 4; see text).

1600 1.0
A. B.
0.9
1400
=1 0.8
1200 /U
Th 0.7
EuN/EuN*
Th (ppm)
1000 antecrysts
0.6
800 0.5
.5 0.4
Paleocene
600 Th/ U=0 volc & gran field
0.3
400
0.2 Paleozoic
granite field
200 0.1
0 0.0
0 300 600 900 1200 1500 0.1 1 10 100
U (ppm) Eu (ppm)
0.9 40
C. D.
0.8 35
30
0.7
EuN/EuN*
25
Yb/Gd
0.6
20
0.5
15 ap
tit
0.4
10
hbl (>px)
0.3 5 zrc,
ga crystal
vectors
0.2 0
0 500 1000 1500 2000 2500 0 5 10 15 20 25 30 35 40
Yb (ppm) Ce/Sm
70
Mesozoic inherited zircons
E. Mixing least evolved IT-3
60 ES-12808 Quartz
& crustal melts
ES-3239 porphyries
Crystal fractionation ES-12791
50 IT-10
of hornblende, X porphyries
Evolved titanite, & apatite ES-12811
Yb/Gd
crustal melts & ES-12807

40 fractionated melts K porphyries
ES-12785a
ES-12800
MG-CH L, A,
30 ES-12783
R porphyries
ES-12782
ES-12789a GG L porphyry
20 Least evolved melts
contaminated with IT-9 TG L
high U, Th, & REEs ES-12787 porphyries
10 ES-12792 Latite dike
Simple Mixtures zircon antecrysts
0
0.0 0.5 1.0 Th/U 1.5 2.0 2.5
Fig. 12. Trace element plots from individual zircon grains. A. Th/U vs. Yb/Gd. B. U ppm vs. Th ppm. C. Eu ppm vs. EuN/
EuN*. D. Yb ppm vs. EuN/EuN*. E. Ce/Sm vs. Yb/Gd. See text for discussion. GG = Granite Gulch, MG-CH = M Gulch-
Copper Hill, TG = Turquoise Gulch. In (A), mixing trends are apparent for zircons from X porphyry and latite dike from high
to low Th/U ratios at low Yb/Gd ratios, as well as for quartz porphyry and M Gulch-Copper Hill porphyry at slightly higher
Yb/Gd. The curve outlines the fractionation trend of Turquoise Gulch K and L porphyries from Turquoise Gulch. Shaded
regions in (B) and (D) represent fields for the three samples of Paleocene age volcanics (volc) and the Paleozoic age Sierra
Castillo batholith granite (gran).

264 LEE ET AL.
2.5 1600
A. Inferred magma processes B.
Magmatic crystallization/
fractionation trend 1400
Latite
2.0 Magma mixing (&
assimilation) trend 1200
X ppy Qtz ppy

1000 Qtz ppy
1.5
U (ppm)
Th/U
L, A, & 800
R ppy L, A, &
1.0
R ppy
600 Latite
400
0.5
X ppy
200
0.0 0
1600 70
C. D.
hbl, tit, & ap
1400 60 crystallization
1200 Qzt ppy

L, A, & 50 zrc
R ppy
1000 L, A, &
Th (ppm)
40 R ppy
Yb/Gd
800 Latite
30 Qtz ppy
600
X ppy
20
400
10
200 Latite
X ppy
1.0 1000
E. F. 3390
0.9 900 Strongly
X ppy Strongly oxidized
L, A, & oxidized Qtz ppy
0.8 R ppy 800 conditions
conditions
Latite
0.7 700
0.6 600 L, A, &

CeN/CeN*
R ppy
EuN/EuN*
0.5 500 X ppy
0.4 400
0.3 Qtz ppy 300

Latite
0.2 200
Less
0.1 oxidized 100 Less
oxidized
0.0 0
7000 9000 11000 13000 15000 17000 7000 9000 11000 13000 15000 17000
Hf (ppm) Hf (ppm)
Fig. 13. Trace elements in zircon plotted as a function of Hf variation, where Hf content increases with magmatic evolution
and decreased crystallization temperature of zircon. A. Hf vs. Th/U. B. Hf vs. U. C. Hf vs Th. D. Hf vs. Yb/Gd. E. Hf (ppm)
vs. EuN/EuN*. F. Hf (ppm) vs. CeN/CeN*. Explanation, symbols, and shaded regions in (D) and (F) as in Figure 12 (see App.
3 for CeN/CeN* and EuN/EuN* calculation). Light pink shaded region highlights L, A, and R porphyries associated with min-
eralization; ap = apatite, hbl = hornblende, ppy = porphyry, qtz ppy = quartz porphyry, tit = titantite.

west of the Sierra Castillo fault in the area with little Paleo- estimated on the basis of the presence of high Al amphiboles
zoic crust as a result of Jurassic-Cretaceous rifting and related that the mafic component of latite dikes was derived from the
magmatism (Tosdal et al., 2000). mid-crust (~6–8 kbar pressure and ~900°C). In contrast, the
At many as 26 of the 224 Eocene zircons analyzed (12%) low-temperature and Hf-rich zircons in these samples contain
are slightly older than the host magmas, and of these 19 (8%) less U and Th, suggesting that they crystallized following mix-
have ages >2σ errors greater than the host-rock mean age ing of high- and low-temperature melts.
(Figs. 6, 7). We attribute these zircons to be antecrysts (cf. REE contents and ratios also strongly indicate roles of crys-
Miller et al., 2007). Nine antecrysts have ages >45 Ma, the tal fractionation of parental basalt or andesite to produce silicic
maximum possible age of El Salvador porphyry magmatism, melts. This hypothesis is supported by a decrease of Th/U, an
based on the new U/Pb zircon data. Some of the antecryst zir- increase of incompatible Th and U, and increase in both the
cons can be readily distinguished by trace element character- Ce/Sm and Yb/Gd ratios (Fig. 12). The increase in Ce/Sm and
istics. For example, three antecrysts in L porphyry have high Yb/Gd results from preferential removal of MREE (Sm, Gd)
U and Th and large Ce and Eu anomalies that are dissimilar from the melt compared to LREE and HREE, and produces
to L porphyry zircons, but similar to zircons from the early the decreased convexity of the REE pattern characteristic of
quartz and rhyolite porphyries (Figs. 12, 13). These antecryst the mineralizing L porphyry at Turquoise Gulch (Fig. 11). The
zircons provide evidence of xenolith contributions to younger X porphyry and latite display linear arrays in Ce/Sm versus
magmas from earlier crystalline intrusions of the El Salvador Yb/Gd plots that have steep and shallow slopes, respectively,
magmatic system, or by a long-lived middle to upper crustal that reflect different initial compositions and crystallization
magma chamber containing both crystals and melt. Such a (Fig. 12D). The diversity of zircon compositions along the
magma mush could have continuously evolved over a period observed evolutionary paths apparently reflects different pro-
of several million years if periodically replenished by new, hot portions of hornblende, apatite, and titanite crystallization
magma. (Figs. 12D, E, 13D). Using available mineral-melt partition
The early rhyolite and quartz porphyries have variable but coefficients for basalt and andesite, the vectors in Figure 12D
generally higher Th and U contents and Th/U ratios com- illustrate the effect of crystallization of hornblende, apatite,
pared to the late K and L porphyries (Th/U ~1, 0.5, respec- titanite, zircon and garnet. These vectors suggest that Ce/Sm
tively; Fig. 12B, C). The zircons from these porphyry samples and Yb/Gd increased initially as a result of high-temperature
all evolved to high Th and U and low Th/U, which is the nor- apatite and hornblende crystallization, and with further cool-
mal evolution due to crystallization of the magma or crystal ing by titanite crystallization (≤780°C). The steep X porphyry
fractionation to produce residual melts with relatively high U array is suggestive of apatite control, and the shallow latite
and Th contents (Miller and Wooden, 2004). Nonetheless, we array indicates hornblende control. Zircon and garnet control
observe extreme enrichment of U and Th in zircon by factors is lacking. Nonetheless, latite zircons have depressed Yb/Gd
of 40 and 60, respectively, in low versus high Hf zircons. Clair- ratios and low HREE contents, supportive of a role for garnet
borne et al. (2006) proposed that the high Th and U of zircon in melt generation, and with similar to garnet-rich porphy-
is a result of sharply increased zircon-melt partition coef- ries elsewhere (Thomas Bissig, writ. commun., 2016). This is
ficient for Th and U due to increased crystallization rate as consistent with whole-rock compositions that suggest garnet
the magma approaches eutectic-like conditions, or as the melt fractionation in deep and high-temperature source areas for
structure changed with increased water content. This expla- the latite melt (Lee, 2008).
nation is considered to be unlikely for the El Salvador zircons The plot of Th/U versus Yb/Gd illustrates a curved path
because quartz porphyry evolved to high Th and U contents, interpreted to be the result of crystal fractionation (Fig. 12E).
whereas most samples of K and L porphyries did not, despite Most zircons in the X porphyry and latite follow this path from
all having similar Th and U contents at low Hf contents (Fig. high to low temperature (high to low Th/U), and most low-
13B, C). We conclude that low-temperature quartz porphyry temperature zircons (~750°–650°C) from the K and L por-
melts were contaminated by assimilation of U- and Th-rich phyries at the strongly mineralized Turquoise Gulch center
crustal materials. also lie along the curve at low Th/U and high Yb/Gd (Fig.
Zircons from the latite dike and the X porphyry, which are 12E). In contrast, many of the zircons from poorly mineral-
the lowest silica intrusions, have U and Th contents consistent ized intrusions, including the quartz porphyry and the L, A,
with high-temperature melt contamination. These zircons and R porphyries from M Gulch-Copper Hill, form linear
with the lowest Hf content and highest temperature (800°– gently dipping mixing arrays. The end members of mixtures
900°C; Fig. 10A) have the highest U, Th, Y, and REE con- are a large proportion of the most evolved K and L porphyry
tents of all porphyry sampled in this study. With increased Hf granodiorites and a smaller proportion of the least evolved of
content and decreased temperature, the Th and U contents the X porphyry and latite (Fig. 12E). These results imply that
of zircon remain relatively constant and do not show the typi- evolution of mafic magma to granodiorite dominantly via frac-
cal increase as expected. One explanation is that the magmas tional crystallization was favorable to ore generation in the K
were enriched in aqueous fluids and fluid mobile elements and L porphyries, whereas the zircons from other porphyries
derived from the subducting slab (cf. Arculus, 1994; Bailey that display effects of wall-rock contamination and mixing
and Ragnarsdottir, 1994, Noll et al., 1996; Vigouroux et al., paths produced much less copper ore.
2008).
We propose that these andesitic or basaltic melts assimilated Changes of oxidation state of magmas
at high-temperature small amounts of U-, Th-, Y-, and REE- Porphyry copper intrusions are typically derived from oxi-
enriched crustal materials, for example, shale. Lee (2008) dized magmas with inclined rare earth patterns (LREE >

266 LEE ET AL.
HREE) and small or negligible negative Eu anomalies (Bal- EuN* <0.2 at Hf >10,000 ppm as observed in other arc suites
lard et al., 2002). The El Salvador porphyries contain the (Dilles et al., 2015). For the Eocene porphyries, several high-
mineral assemblage magnetite-titanite-quartz, consistent temperature zircons (<9,000 ppm Hf) from X porphyry and
with strongly oxidized conditions near the NNO + 2 buffer latite dike are similar and have high REE contents and small
(Dilles, 1987), and the granodiorite porphyry samples lack a negative Eu anomalies with EuN/EuN* ~0.8 (Figs. 11D, E, H,
significant Eu anomaly (Gustafson, 1979). The magnitude and 13E). At slightly higher Hf contents of ~9,000 ppm, all of the
relative change of the Ce and Eu anomalies has been pro- Eocene El Salvador porphyries have small negative Eu anom-
posed to reflect oxidation state (Ballard et al., 2002; Dilles alies (EuN/EuN* = 0.6–0.8), but as Hf increases two trends are
et al., 2015). Both Ce and Eu have two common oxidation apparent (Fig. 13E). The rhyolite and quartz porphyry as well
states in nature, whereas other REE almost invariably have as the L, A, and R porphyries from M Gulch-Copper Hill show
+3 valence. More oxidized conditions increase the Ce content a decrease of EuN/EuN* to ~0.4 at Hf of 12,000 ppm, whereas
of zircon because the proportion of Ce4+ compared to Ce3+ is zircon from the K and L porphyries from Turquoise Gulch,
increased and +4 valence cations are preferred in the zircon some X porphyries, and latite dikes evolve to EuN/EuN* to
crystal structure. Europium exists mainly in the +2 valence ~0.6 to 0.7. The two trends are accentuated by grouping zir-
where it may be incorporated into crystallizing plagioclase to cons by age (Fig. 14), where the older Eocene intrusions show
produce a Eu deficiency or negative Eu anomaly in both melt evolution to more negative Eu anomalies (EuN/EuN* ~0.4),
and zircon. The Eu anomaly may be reduced by suppression whereas the younger K and L porphyry intrusions associated
of plagioclase crystallization at high magmatic water content, with Cu-Mo ores evolve with little increase in the Eu anomaly.
or under oxidized conditions which increase the proportion These three trends are also observed in the plot of Eu versus
of Eu3+ that cannot be incorporated into plagiocase but does EuN/EuN* (Fig. 12B), where crystallization and cooling pro-
get incorporated into zircon (Wilke and Behrens, 1999; Bal- duces zircon with lower Eu contents. The evolution to large
lard et al., 2002). Dilles et al. (2015) proposed that oxidation negative Eu anomalies (EuN/EuN* <0.3) is typical of most arc
of Eu2+ to Eu 3+ is produced in magma as a result of sulfate magmas, whereas less negative Eu anomalies characterize the
reduction to sulfur dioxide during separation of the sulfur-rich mineralized El Salvador porphyries and are here proposed to
magmatic-hydrothermal fluid. result from suppression of plagioclase during crystallization at
Hf versus EuN/EuN* plots of the Paleozoic and Paleocene high water pressures and high Eu3+/Eu2+ ratios as a result of
igneous zircons yield typical nonmineralizing trends to EuN/ strongly oxidized conditions, as outlined above. Ballard et al.
0.9 Age < 41.5 Ma Age 46.0 - 43.5 Ma

source? Age 42.5 - 41.5 Ma Age 62-60 Ma
0.8 Age 43.5 - 42.5 Ma Age 291 Ma
0.7
0.6 Late, well mineralized

trend
EuN/EuN*
0.5
0.4
0.3
0.2 Early, poorly

mineralized trend
0.1 pre-El Salvador porphyry
non-mineralized trend
0.0
6000 8000 10000 12000 14000 16000
Hf ppm
Fig. 14. Plot of Hf ppm vs. EuN/EuN* for zircons grouped by age; modified after Dilles et al. (2015). Using Hf increase as a
proxy for crystallization and cooling; pre-Eocene age zircons define a trend of low EuN/EuN* values suggestive of a reduced
magmatic melt.The oldest Eocene age zircon grains (46–43.5 Ma) from quartz and rhyolite porphyry as well as inherited
zircons in younger porphyries evolve to low EuN/EuN* consistent with Eu2+ incorporation into crystallizing plagioclase (lower
arrow).Younger zircons (<43.5 Ma) evolve to a wide variety of EuN/EuN* values. Ruled region denotes the field for all M
Gulch-Copper Hill porphyries, whereas the shaded region denotes the K and L porphyry zircons (~42.5–42 Ma) from the
Turquoise Gulch ore deposit.The latter follow an evolution path to high Hf content and high EuN/EuN*, consistent with lesser
amounts of plagioclase crystallization or more strongly oxidized melt conditions in which the Eu3+/Eu2+ ratio increases and
little Eu is incorporated into plagioclase (see text).

(2002) reported that EuN/EuN* >0.4 characterize zircons Evidence for late and low-temperature crustal contamination
from intrusions associated with porphyry copper mineraliza- by U- and Th-rich rocks is provided by zircons from the early
tion in the El Abra district, whereas porphyry copper ores at quartz porphyry that evolved from high to low temperature
El Salvador are associated with EuN/EuN* >0.55, suggesting and U- and Th-poor to -rich compositions. Many of the early
higher oxidizing conditions relative to El Abra. magmas (quartz and X porphyry) and late latite dikes contain
Ballard et al. (2002) also proposed that the large positive Ce zircons with trace element compositions reflecting mixing
anomalies (e.g., as recorded by CeN/CeN*) characterize por- of the mafic and silicic melt components. Inherited Meso-
phyry intrusions that are highly oxidized. Although we do not zoic zircons are only abundant in the ~43.6 Ma X porphyries
calculate the CeIV/CeIII in zircon here, the plot of CeN/CeN* as and reflect assimilation of the relatively primitive composi-
a function of Hf content illustrates no systematic differences tion basaltic middle to upper crust so that Sr and Pb isotope
between the K and L porphyries associated with Turquoise compositions of all porphyries are relatively nonradiogenic
Gulch copper mineralization and the weakly mineralized early (Gustafson, 1979; Tosdal et al., 1999, 2000). Eocene zircon
rhyolite and quartz porphyries to the north (Fig. 13F). At El antecrysts older than the host porphyry record the recycling
Salvador, high CeN/CeN* ratios characterize nonmineralized of older porphyry magmas within the El Salvador district.
Paleocene and Paleozoic igneous rocks and poorly mineral- These antecrysts can only be identified robustly for ~8% of
ized early rhyolite and quartz porphyry with large negative all zircons analyzed due the relatively large analytical errors in
EuN/EuN* anomalies. Although the theory outlined in Ballard age (±2% or ~1 m.y.), but we infer that the presence of these
et al. (2002) is robust, there are large errors in the estimation antecrysts reflect magmatic recycling as a dominant process in
of Ce* from the measured concentrations of La, which is com- the middle to upper crustal magma chambers.
monly 0.01 ppm in zircon and may be contaminated by min- Zircons from the Paleozoic-Paleocene magmas and the
ute La-rich inclusions of apatite that are ubiquitous in zircon. earliest Eocene rhyolite and quartz porphyry magmas dis-
Similarly, Pr concentrations in zircon are also low and direct play a large range of REE contents and small negative Eu
measurement of Pr by mass spectroscopic methods are ham- anomalies, the latter increasing with decreasing temperature
pered in more Ce rich materials by the 140Ce1H+ mass interfer- and attesting to shallow crystallization of plagioclase. In the
ence. Our CeN/CeN*estimates derived via the SHRIMP-RG Eocene zircons, Hf increases in zircon as Ti-in-zircon tem-
method may be more robust due to the small sample volume, perature decreases, and both REE and Y decrease by factors
greater sensitivity, and smaller degree of 140Ce1H+ mass inter- of up to 100. The MREE content of zircon also decreases with
ference on 141Pr compared to the LA-ICP-MS method used respect to LREE and HREE, consistent with the crystalliza-
by Ballard et al. (2002). Apatite inclusions were avoided in our tion of an assemblage rich in hornblende, apatite, and titanite
analyses, but nonetheless we monitored Ca and P contents (Figs. 11, 13D; cf. Ballard, 2001). This crystallization probably
and found them to be insignificant where Ti and the full trace generated the late granodiorite porphyries from Turquoise
element suite was analyzed (App. 4). We conclude that EuN/ Gulch that are closely associated with the bulk of copper-
EuN* tracks late oxidation of the El Salvador magmas associ- molybdenum mineralization in the district and contain zir-
ated with mineralization, but that CeN/CeN* does not. cons with small ranges of Th, U, Y, and REE contents. These
zircons have small negative Eu anomalies at high temperature
Model for the Formation of the El Salvador Deposit and only slightly more negative Eu anomalies at low tempera-
Based on the zircon geochemistry and the U/Pb ages presented tures. Thus, the late mineralizing porphyries have smaller Eu
above, we propose that the porphyry intrusions emplaced at anomalies (EuN/EuN* >0.55) compared to earlier intrusions
<3-km depth in the Indio Muerto district progressivly tapped and apparently crystallized under more strongly oxidizing and
one (or more) complexly evolving crustal granodioritic magma hydrous conditions that suppressed plagioclase crystallization
chamber(s) at an indeterminant depth over at least a 3-m.y. and decreased the Eu2+/Eu3+ ratio. The mineralizing grano-
period (Fig. 15). Magmatism became voluminous by 44 Ma, diorite porphyries appear to have crystallized under relatively
reached a peak at 43.5 to 43 Ma, declined during most cop- closed-system conditions compared to earlier magmas, as
per introduction at 42.2 to 41.8 Ma, and terminated by about indicated by the narrow ranges of Ti-in-zircon temperature
41.6 Ma (Fig. 6). Hot, hydrous, and relatively mafic (andes- (750°–650°C) and trace element content as well as the curvi-
itic?) magma derived from the lower crust underpinned linear trace element evolution paths characteristic of crystal-
and fed the magma chambers. These magmas were rich in lization rather than mixing (Fig. 12E).
incompatible trace elements (Y, REE) and had high MREE/ The long-lived magmatism in the Indio Muerto district has
HREE ratios, and we speculate potentially may have sup- a complex association with ores. Oxidized and hydrous grano-
plied significant S, Cu, and Au (i.e., Rohrlach and Loucks, diorite magmas formed through the entire magmatic cycle
2005; Chamberfort et al., 2013). Locally, high Th and U in from ~44 to 41 Ma, but where zircon geochemistry provides
highest temperature zircon in the most mafic samples pro- evidence for either substantial late mafic magma input and
vides evidence that mafic melts were locally contaminated by contamination by silicic crustal melt, the intrusions are rela-
incompatible element-enriched crustal melts. The diversity tively poorly mineralized. In contrast, the K and L porphy-
of Ce/Sm, Yb/Gd, and Th/U ratios observed in zircon crys- ries associated with the main Cu(Mo) orebody at Turquoise
tallized at low temperature provides evidence of both crystal Gulch evolved at ~750° to 650°C via fractionation with little
fractionation and crustal contamination of low-temperature late input of mafic magma or crustal melt. The El Salvador
melts (cf. Miller and Wooden, 2004). The evolution to high magmatic evolution and ore metals (Cu-Mo and minor Au)
REE ratios (e.g., Ce/Sm and Yb/Gd) was probably produced differs from Bingham, Utah, where mafic alkaline magmas
by crystallization of hornblende, titanite, and minor apatite. containing magmatic sulfides are observed and associated

268 LEE ET AL.
A. Ramp Up B. Peak Magma C. Waning Stage D. End

Rhyolite ppy Qtz & X ppy K & L ppy at TG Latite Dikes
≥44 Ma ~44-43 Ma ~42.5-42 Ma ~41 Ma
CP OC CP OC MG GG CP OC MG TG GG CP OC MG TG GG
N Paleosurface? S N S N S N S
0 x
xx x
x xx
x
xxx x
x x
X XX
XX oX X
x x XX X X X
o XX X X X X
o X X X XX X
? o
o X X XX X
o XX
o o
o oo o
o o o
Upper-crustal o
plutons oo
10
o o
Mid-crustal
magma zone?
Depth (km)
20
30
MASH zone
40
base of crust
?
Upper mantle
Silicic porphyry Mafic to intermediate, Zoned silicic to inter- Granodiorite Ppy Mesozoic
intrusion volatite & U-Th rich intrusions mediate magma intrusion basement
xxxx Porphyry-Cu oo
o oo o Vapor Mesozoic zircons Eocene zircons
x
Fig. 15. Conceptual north-south cross-sectional model of upper and middle crustal magmatic processes related to porphyry
intrusion and porphyry Cu-Mo ore formation in the Indio Muerto district with age progression. Age progression is shown
by four panels from left to right, and makes use of modern crustal structure and active magmatism from Volcan Aucanquil-
cha, northern Chile (Grunder et al., 2008). A. Magmatism was underpinned by mafic to intermediate melts that originated
in the middle crust or deeper and produced crustal melting. Initial rhyolite porphyry melts mark the beginning of crustal
melting and magma chamber formation in the shallow crust. B. Mafic to intermediate composition Mesozoic basement rock
was assimilated during peak magmatic fluxes marked by the X porphyries. C. A large granodiorite magma chamber formed
in the upper crust, and as mafic input declined it cooled, crystallized, and saturated with aqueous fluids to form magmatic-
hydrothermal fluids and mineralization that accompanied the K and L porphyry intrusions at Turquoise Gulch. D. The final
intrusion of small volumes of postore latite dikes records a mixing of deep-seated high-temperature mafic magma with the
nearly solid granodiorite body. CP = Cerro Pelado, GG = Granite Gulch, MG = M Gulch, OC = Old Camp, TG = Turquoise
Gulch; ppy = porphyry, qtz ppy = quartz porphyry, MASH zone = melting, assimilation, storage, and homogenization after
Hildreth and Moorbath (1988). See text for further discussion of model.
with Cu-Mo-Au ores with elevated PGE, Ni, and Cr contents expanded via crustal melting, but during periods of low flux
(Keith et al., 1997). The available age data are not sufficiently cooling and crystallization led to water saturation in an apical
precise to be conclusive, but the best interpretations provided cupola of the magma chamber from which magmatic-hydro-
by the new U/Pb data suggest that porphyry Cu-Mo miner- thermal fluids were extracted (cf. Burnham, 1979; Dilles,
alization moved generally southward in time, progressively 1987). At high oxygen fugacity, characteristic of these mag-
forming the Cerro Pelado (~44.2 Ma), Old Camp (~43.6 Ma), mas, sulfur would be dominated by sulfate (SO42–; Field et al.,
M Gulch-Copper Hill (~43.1–43.5 Ma), and Turquoise Gulch 2005; Chambefort et al., 2008) and copper would remain in
(~42 Ma) ore centers, however, >80% of Cu-Mo was depos- the melt rather than be incorporated into sulfide (Halter et
ited during the final event at Turquoise Gulch. al., 2001; Stavast et al., 2006) or silicates (Dilles and Proffett,
We infer that mafic magma flux was not constant with time, 1995). Upon vapor saturation, sulfur species, chlorine, copper,
and that during periods of high flux the magma chambers and other chalcophile elements would partition into the vapor

phase, separate from the magma chamber, rise, and react with Baker, D.R., Conte, A.M., Freda, C., and Ottolini, L., 2002, The effect of hal-
wall rocks to form porphyry Cu-Mo ore minerals, veins and ogens on Zr diffusion and zircon dissolution in hydrous metaluminous gra-
nitic melts: Contributions to Mineralogy and Petrology, v. 142, p. 666–678.
wall-rock alteration (cf. Candela and Holland, 1985; Ulrich Ballard, J.R., Palin, J.M., Williams, I.S., Campbell, I.H., and Faunes, A.,
et al., 1999). All sulfide mineralization so far known in the 2001, Two ages of porphyry intrusion resolved for the super-giant Chuqui-
Indio Muerto district formed at shallow depths of <0.5 to ~ 3 camata copper deposit of northern Chile by ELA-ICP-MS and SHRIMP:
km. During the late-stage solidification of the main granodio- Geology, v. 29, p. 383–386.
Ballard, J.R., Palin, J.M., and Campbell, I.H., 2002, Relative oxidation states
rite chamber at ~42.2 Ma, the aqueous fluids, Cl, S, Cu, and of magmas inferred from Ce(IV)/Ce(III) in zircon: Application to porphyry
Mo were quantitatively extracted from the magma and this copper deposits of northern Chile: Contributions to Mineralogy and Petrol-
process appears to be closely linked to the strongly oxidized ogy, v. 144, p. 347–364.
trend recorded in the EuN/EuN* of zircons from the K and L Barra, F., Alcota, H., Rivera, S., Valencia, V., Munizaga, F., and Mak-
porphyries. At ~41.6 Ma, the final intrusion of latite dikes with saev, V., 2013, Timing and formation of porphyry Cu-Mo mineraliza-
tion in the Chuquicamata district, northern Chile: New constraints from
mixed magma textures and trace elements apparently resulted the Toki cluster: Mineralium Deposita, v. 48, p. 629–651. doi 10.1007/
from a deep-sourced mafic melt that propogated along frac- s00126-012-0452-1.
tures through the largely solidified upper crustal granodiorite Black, L.P., Kamo, S.L., Allen, C.M., Davis, D.W., Aleinikoff, J.N., Valley,
magma chamber. J.W., Mundilf, R., Campbell, I.H., Korsch, R.J., Williams, I.S., and Fou-
doulis, C., 2004, Improved 206Pb/238U microprobe geochronology by the
Conclusions monitoring of a trace element-related matrix effect: SHRIMP, ID-TIMS,
ELA-ICP-MS and oxygen isotope documentation for a series of zircon stan-
The El Salvador porphyry Cu (Mo-Au) district formed over a dards: Chemical Geology, v. 205, p. 115–140.
period of at least 3 m.y., with at least one mineralization event Burnham, C.W., 1979, Magmas and hydrothermal fluids [2nd edition], in
associated with peak magmatism at 44 to 43 Ma followed by Barnes, H.L., ed., Geochemistry of hydrothermal ore deposits: Wiley,
p. 71–136.
the main mineralization event at ~42 Ma. Zircon trace ele- Campbell, I.H., Ballard, J.R., Palin, J.M., Allen, C., and Faunes, A., 2006,
ment chemistry record an evolving mid- to upper-crustal U-Pb zircon geochronology of granitic rocks from the Chuquicamata-El
magma chamber(s) with decreasing U, Th, and REE content Abra porphyry copper belt of northern Chile: Excimer laser ablation ICP-
as well as higher EuN/EuN* values, reflecting increasing oxi- MS analysis: Economic Geology, v. 101, p. 1327–1344.
dation through time. Periodic pulses of magmatism increased Camus, F., and Dilles, J.H., 2001, A special issue devoted to porphyry copper
deposits of northern Chile: Economic Geology, v. 96, p. 233–237.
the metal and sulfur content in the melt followed by increas- Candela, P.A., and Holland, H.D., 1986, A mass transfer model for copper
ing water conent due to cooling and crystallization allowing and molybdenum in magmatic hydrothermal system: The origin of por-
for the extraction of the metals in an SO2-rich hydrothermal phyry-type ore deposits: Economic Geology, v. 81, p. 1–19.
fluid. The variations observed in the zircon geochemistry of Chambefort, I., Dilles, J.H., and Kent, J.R., 2008, Anhydrite-bearing andesite
and dacite as a source for sulfur in magmatic-hydrothermal mineral depos-
nonore-forming intrusions versus ore-related intrusions from its: Geology, v. 36, p. 719–722.
El Salvador and noted by Dilles et al. (2015) present a poten- Chambefort, I., Dilles, J.H., and Longo, A.A., 2013, Amphibole geochemistry
tial method for determining fertility of intrusions associated of the Yanacocha Volcanics, Peru: Evidence for diverse sources of magmatic
with economic porphyry copper deposits. volatiles related to gold ores: Journal of Petrology, v. 54, p. 1017−1046.
Cherniak, D.J., and Watson, E.B., 2007, Ti diffusion in zircon: Chemical
Acknowledgments Geology, v. 242, p. 470–483.
Clairborne, L.L., Miller, C.F., Walker, B.A., Wooden, J.L., Mazdab, F.K., and
This project was funded in part by the Exploraciones Min- Bea, F., 2006, Tracking magmatic processes through Zr/Hf ratios in rocks
eras Andinas (EMA) division of CODELCO, a 2007 Society and Hf and Ti zoning in zircons: An example from the Spirit Mountain
of Economic Geologists student research grant, and a gener- batholith, Nevada: Mineralogical Magazine, v. 70, p. 517–543.
ous contribution from Freeport McMoran, Inc. The authors Clairborne, L.L., Miller, C.F., and Wooden, J.L., 2010, Trace element com-
position of igneous zircon: a thermal and compositional record of the accu-
thank Lew Gustafson and Enrique Tidy for providing guid- mulation and evolution of a large silicic batholith, Spirit Mountain, Nevada:
ance, help, and the initial samples for analyses. The National Contributions to Mineralogy and Petrology, v. 160, p. 511–531.
Science Foundation, the U.S. Geological Survey, and Stanford Coombs, M.L., Eichelberger, J.C., and Rutherford, M.J., 2000, Magma stor-
University provided funds and access to the SHRIMP-RG, and age and mixing conditions for the 1953–1974 eruptions of Southwest Tri-
dent volcano, Katmai National Park, Alaska: Contributions to Mineralogy
Barry A. Walker and Mark Ford assisted with SHRIMP-RG and Petrology, v. 140, p. 99–118.
data acquisition. Walter Orquera, Ricard Santelices, Christian Cornejo, P., Mpodozis, C., Kay, S.M., and Tomlinson, A.J., 1994, Volcanismo
Rojas, and Eduardo Gonzalez of EMA at the El Salvador mine bimodal potásico en regimen extensional del Cretácio superior-Eoceno
assisted in the collection and initial processing of the samples. en la región de El Salvador (26°-27° S), Chile: VII Congreso Geológico
The Oregon State University Volcanology, Igneous Petrology, Chileno, Concepción, Actas, v. 2, p. 1306–1310.
Cornejo, P., Tosdal, R.M., Mpodozis, C., Tomlinson, A.J., Rivera, O., and
and Economic Resources (VIPER) group, particularly Anita Fanning, M., 1997, El Salvador, Chile porphyry copper deposit revisited:
Grunder and Adam Kent, provided valuable critical input. We Geologic and geochronologic framework: International Geology Review,
also thank David Cooke, Sebastien Meffre, and Pete Hollings v. 39, p. 22–54.
for their helpful reviews of this manuscript as well as an anony- Dilles, J.H., 1987, The petrology of the Yerington batholith, Nevada: Evi-
mous reviewer from an earlier version of this study. dence for the evolution of porphyry copper ore fluids: Economic Geology,
v. 82, p. 1750–1789.
Dilles, J.H., and Proffett, J.M., 1995, Metallogenesis of the Yerington batho-
REFERENCES lith, Nevada: Arizona Geological Society Digest 20, p. 306–315.
Anders, E., and Grevesse, N., 1989, Abundances of the elements: Meteoritic Dilles, J.H., and Wright, J.E., 1988, Chronology of early Mesozoic arc mag-
and solar: Geochimica et Cosmochimica Acta, v. 53, p. 197–214. matism in the Yerington district of western Nevada, and its regional impli-
Arculus, R.J., 1994, Aspects of magma genesis in arcs: Lithos, v. 33, p. 189–208. cations: Geological Society of America Bulletin, v. 100, p. 644–652.
Bailey, E.H., and Ragnarsdottir, K.V., 1994, Uranium and thorium solubili- Dilles, J.H., Tomlinson, A., Martin, M., and Blanco, N., 1997, El Abra and
ties in subduction zone fluids: Earth and Planetary Science Letters, v. 124, Fortuna complexes: A porphyry copper batholith sinstrally displaced by the
p. 119–129. Falla Oeste: Congreso Geologico Chileno, v. 3, p. 1883–1887.

270 LEE ET AL.
Dilles, J.H., Martin, M.W., Stein, H., and Rusk, B., 2003, Re-Os and U-Pb Korotev, R.L., 1996, A self-consistent compilation of elemental concentration
ages for the Butte copper district, Montana: A short- or long-lived hydro- data for 93 geochemical reference samples: Geostandards Newsletter, v. 20,
thermal system? [abs.]: Geological Society of America Abstract with Pro- p. 217–245.
grams, v. 35, p. 400. Lee, R.G., 2008, Genesis of the El Salvador porphyry copper deposit, Chile
Dilles, J.H., Kent, A.J.R., Wooden, J.L., Tosdal, R.M., Kolezar, A., Lee, R.G., and distribution of epithermal alteration at Lassen Peak, California: Ph.D.
and Farmer, L.P., 2015, Zircon compositional evidence for sulfur-degassing thesis, Corvallis, OR, Oregon State University, 344 p.
from ore-forming arc magmas: Economic Geology, v. 110, p. 241–251. Longo, A.A., and Teal, L.B., 2005, A summary of the volcanic stratigraphy
Ekren, E.B., Byers, F.M., Jr., Hardyman, R.F., Marvin, R.F., and Silberman, and the geochronology of magmatism and hydrothermal activity in the
M.L., 1980, Stratigraphy, preliminary petrology, and some structural fea- Yanacocha gold district, northern Perú: Geological Society of Nevada, Sym-
tures of Tertiary volcanic rocks in the Gabbs Valley and Gillis Ranges, Min- posium, 2005, Proceedings, v. 12, p. 797–808.
eral County, Nevada: U.S. Geological Survey Bulletin 1464, 54 p. Longo, A.A., Dilles, J.H., Grunder, A.L., and Duncan, R., 2010, Evolution of
Elderry, S.M., Chong Diaz, G., Prior, D.J., and Flint, S.S., 1996, Struc- calc-alkaline volcanism and associated hydrothermal gold deposits at Yana-
tural styles in the Domeyko Range, northern Chile: Third International cocha, Peru: Economic Geology, v. 105, p. 1191–1241.
Symposium on Andean Geodynamics, L’Institut de Recherche pour le Ludwig, K.R., 2001, Squid version 1.02: A user’s manual: Berkeley Geochro-
De´veloppement en Coop., St. Malo, France, September 17–19, p. 439–442. nology Center Special Publication 2, p. 1–22.
Ferry, J.M., and Watson, E.B., 2007, New thermodynamic models and revised ——2003, Isoplot version 3.0: A geochronology toolkit for Microsoft Excel:
calibrations for the Ti-in-zircon and Zr-in-rutile thermometers: Contribu- Berkeley Geochronology Center Special Publication 4, v. 71, p. 2.
tions to Mineralology and Petrology, v. 154, p. 429–437 ——2009, Squid 2, A user’s manual: Berkeley Geochronology Center Special
Field, C.W., Zhang, L., Dilles, J.H., Rye, R.O., and Reed, M., H.2005, Sulfur Publication 5, p. 1–110.
and oxygen isotopic record in sulfate and sulfide minerals of early, deep, Marsh, T.M., Einaudi, M.T., and McWilliams, M.O., 1997, 40Ar/39Ar chronol-
pre-main stage porphyry Cu-Mo and late main stage base-metal mineral ogy of Cu-Au and Au-Ag mineralization in the Potrerillos district, Chile:
deposits, Butte district, Montana: Chemical Geology, v. 215, p. 61–93. Economic Geology, v. 92, p. 784–806.
Finch, R.J., and Hanchar, J.M., 2003, Structure and chemistry of zircon and Mazdab, F.K., and Wooden, J.L., 2006, Trace element analysis in zircon by
zircon-group minerals: Reviews in Mineralogy and Geochemistry, v. 53, ion microprobe (SHRIMPRG): Technique and applications: Geochimica et
p. 1–25. Cosmochimica Acta Supplement, v. 70, p. 405.
Grunder, A.G., Klemetti, E.W., Feeley, T.C., and McKee, C.M., 2008, Eleven Miller, C.F., McDowell, S.M., and Mapes, R.W., 2003, Hot and cold granites?
million years of arc volcanism at the Aucanquilcha volcanic cluster, north- Implications of zircon saturation temperatures and preservation of inheri-
ern Chilean Andes: Implications for the life span and emplacement of plu- tance: Geology, v. 31, p. 529–532.
tons: Transactions of the Royal Society of Edinburgh, v. 97, p. 415–436. Miller, J.S., and Wooden, J.L., 2004, Residence, resorption and recycling of
Gustafson, L.B., 1979, Porphyry copper deposits and calc-alkaline volcanism, zircons in Devils Kitchen Rhyolite, Coso volcanic field, California: Journal
in McElhinny, M.W., ed., The Earth: Its origin, structure, and evolution: of Petrology, v. 45, p. 2155–2170.
London, Academic Press, p. 427–468. Miller, J.S., Matzel, J.E.P. , Miller, C.F., Burgess, S.D., and Miller, R.B., 2007,
Zircon growth and recycling during the assembly of large, composite arc plu-
Gustafson, L.B., and Hunt, L.B., 1975, The porphyry copper deposit at El
tons: Journal of Volcanology and Geothermal Research, v. 167, p. 282–299.
Salvador, Chile: Economic Geology, v. 70, p. 857–912.
Mote, T.I., Brimhall, G.H., Tidy, E.F., Müller, G.H., and Carrasco, P. , 2001,
Gustafson, L.B., Orquera, W., McWilliams, M., Castro, M., Olivares, O.,
Application of mass-balance modeling of sources, pathways, and sinks
Rojas, G., Maluenda, J., and Mendez, M., 2001, Multiple centers of miner-
of supergene enrichment to exploration and discovery of the quebrada
alization in the Indio Muerto district, El Salvador, Chile: Economic Geol-
Turquesa exotic copper orebody, El Salvador district, Chile: Economic
ogy, v. 96, p. 325–350.
Geology, v. 92, p. 367–386.
Halter, W.E., Heinrich, C.A., and Pettke, T., 2005, Magma evolution and the
Noll, P.D., Jr., Newsom, H.E., Leeman, W.P., and Ryan, J.G., 1996, The role
formation of porphyry CU-Au ore fluids: Evidence from silicate and sulfide
of hydrothermal fluids in the production of subduction zone magmas: Evi-
melt inclusions: Mineralium Depositia, v. 39, p. 845–863. dence from siderophile and chalcophile trace elements and boron: Geochi-
Hanchar, J.M., and van Westrenen, W., 2007, Rare earth element behavior in mica et Cosmochimica Acta, v. 60, p. 587–611.
zircon-melt systems: Elements, v. 3, p. 37–42. Pelleter, E., Cheilletz, A., Gasquet, D., Mouttaqi, A., Annich, M., El Hakour,
Hanchar, J.M., and Watson, E.B., 2003, Structure and chemistry of zircon A., Deloule, E., and Féraud, G., 2007, Hydrothermal zircons: A tool for ion
and zircon-group minerals: Zircon: Reviews in Mineralogy and Geochem- microprobe U-Pb dating of gold mineralization (Tamlalt-Menhouhou gold
istry, v. 53, p. 89–112. deposit, Morocco): Chemical Geology, v. 245, p. 135–161.
Harris, A.C., and Golding, S.D., 2002, New evidence of magmatic-fluid- Pettke, T., Audetat, A., Schaltegger, U., and Heinrich, C.A., 2005, Magmatic-
related phyllic alteration: Implications for the genesis of porphyry Cu to-hydrothermal crystallization in the W-Sn mineralized Mole Granite
deposits: Geology, v. 30, p. 335–338. (NSW, Australia). Part II: Evolving zircon and thorite trace element chem-
Hildreth, W., and Moorbath, S., 1988, Crustal contributions to arc magma- istry: Chemical Geology, v. 220, p. 191–213.
tism in the Andes of central Chile: Contributions to Mineralogy and Petrol- Proffett, J.M., 1973, Structure of the Butte district, Montana, in Miller, R.,
ogy, v. 98, p. 455–48. ed., Guidebook for the Butte field meeting of the Society of Economic
Hinton, R.W., and Upton, B.G.J., 1991, The chemistry of zircon: Variations Geologists: Butte, Montana, Anaconda Company, p. G-1–G-12.
within and between large crystals from syenite and alkali basalt xenoliths: Pupin, J.P., 2000, Granite genesis related to geodynamics from Hf-Y in zir-
Geochimica et Cosmochimica Acta, v. 55, p. 3287–3302. con: Transactions of the Royal Society of Edinburgh, Earth Sciences, v. 91,
Hoskin, P.W.O., 2000, Patterns of chaos: Fractal statistics and the oscillatory p. 245–256.
chemistry of zircon: Geochimica et Cosmochimica Acta, v. 64, p. 1905–1923. Reynolds, P., Ravenhurst, C., Zentilli, M., and Lindsay, D., 1998, High-
Hoskin, P.W.O., and Schaltegger, U., 2003, The composition of zircon and precision 40Ar/39Ar dating of two consecutive hydrothermal events in the
igneous and metamorphic petrogenesis: Reviews in Mineralogy and Geo- Chuquicamata porphyry copper system, Chile: Chemical Geology, v. 148,
chemistry, v. 53, p. 27–62. p. 45–60.
Houston, R.A, and Dilles, J.H., 2013, Structural geologic evolution of the Rohrlach, B.D., and Loucks R.R., 2005, Multi-million-year cyclic ramp-up of
Butte district, Montana; Economic Geology, v. 108, p. 1397–1424. volatiles in a lower crustal magma reservoir trapped below the Tampakan
Kay, S.M., and Mpodozis, C., 2001, Central Andean ore deposits linked to Cu-Au deposit by Mio-Pliocene crustal compression in the southern Philip-
evolving shallow subduction systems and thickening crust: GSA Today, pines, in Porter, T.M., ed., Super porphyry copper and gold deposits, Aus-
v. 11, p. 4–9. tralia, PCG Publishing, v. 2, p. 369–407.
Keith, J.D., Whitney, J.A., Hattori, K., Ballantyne, G.H., Christiansen, E.H., Rusk, B.G., Reed, M.H., Dilles, J.H., Klemm, L.M., and Heinrich, C.A.,
Barr, D.L., Cannan, T.M., and Hook, C.J., 1997, The role of magmatic sul- 2004, Compositions of magmatic hydrothermal fluids determined by LA-
fides and mafic alkaline magmas in the Bingham and Tintic mining districts, ICP-MS of fluid inclusions from the porphyry copper-molybdenum deposit
Utah: Journal of Petrology, v. 38, p. 1679–1690. at Butte, MT: Chemical Geology, v. 210, p. 173–199.
Keppler, H., 1993, Influence of fluorine on the enrichment of high field Rusk, B., Reed, M., and Dilles, J.H., 2008, Fluid inclusion evidence for mag-
strength trace elements in granitic rocks: Contributions to Mineralogy and matic-hydrothermal fluid evolution in the porphyry copper-molybdenum
Petrology, v. 114, p. 479–488. deposit, Butte, Montana: Economic Geology, v. 103, p. 307–334.

Sano, Y., Terada, K., and Fukuoka, T., 2002, High mass resolution ion micro- Von Quadt, A., Erni, M., Martinek, K., Moll, M., Peytcheva, I., and Hein-
probe analysis of rare earth elements in silicate glass, apatite and zircon: rich, C.A., 2011, Zircon crystallization and the lifetimes of ore-forming
Lack of matrix dependency: Chemical Geology, v. 182, p. 217–230. magmatic-hydrothermal systems: Geology, v. 39, p. 731–734.
Schmitt, A.K., Grove, M., Harrison, T.M., Lovera, O., Hulen, J., and Walters, Walker, B.A., Klemetti, E., Grunder, A., Dilles, J., Tepley, F., and Giles, D.,
M., 2003, The Geysers-Cobb Mountain magma system, California (Part 2013, Crystal reaming during the assembly, maturation and waning of an
1): U-Pb zircon ages of volcanic rocks, conditions of zircon crystallization eleven-million-year crustal magma cycle: Thermobarometry of the Aucan-
and magma residence times: Geochimica et Cosmochimica Acta, v. 67, quilcha volcanic cluster: Contributions to Mineralogy and Petrology, v. 165,
p. 3423–3442. p. 663–682. DOI 10.1007/s00410-012-0829-2
Seedorff, E., Dilles, J.H., Proffett, J.M., Jr., Einaudi, M.T., Zurcher, L., Wil- Watanabe, Y., and Hedenquist, J.W., 2001, Mineralogic and stable isotope
liam J.A.Stavast, W.J.A., Johnson, D.A., Barton, M.D., 2005, Porphyry- zonation at surface over the El Salvador porphyry copper deposit, Chile:
related deposits: Characteristics and origin of hypogene features: Economic Economic Geology, v. 96, p. 1775–1798.
Geology 100th Anniversary Volume, p. 251–298. Watanabe, Y., Stein, H.J., Morgan, J.W., and Markey, R.J., 1999, Re-Os
Seedorff, E., Barton, M.D., Stavast, W.J., and Maher, D., 2008, Root zones geochronology brackets the timing and duration of mineralization for the
of porphyry systems: Extending the porphyry model to depth: Economic El Salvador porphyry Co-Mo deposit, Chile [abs.]: Geological Society of
Geology, v. 103, p. 939–956. America Annual Meeting, Abstracts with Program, v. 30, p. A30.
Sillitoe, R.H., 1988, Epochs of intrusion-related copper mineralization in the Watson, E.B., and Harrison, T.M., 1983, Zircon saturation revisited: Tem-
Andes: Journal of South America Earth Science, v. 1, p. 89–108. perature and composition effects in a variety of crustal magma types: Earth
Sillitoe, R.H., and Perelló, J., 2005, Andean copper province: Tectonomag- and Planetary Science Letters, v. 64, p. 295–304.
matic settings, deposit types, metallogeny, exploration, and discovery: ——2005, Zircon thermometer reveals minimum melting conditions on ear-
Porphyry-related deposits: Characteristics and origin of hypogene features: liest Earth: Science, v. 308, p. 841–844.
Economic Geology 100th Anniversary Volume, p. 845–890. Watson, E.B., and Liang, Y., 1995, A simple model for sector zoning in slowly
Stavast, W.J., Keith, J.D., Christiansen, E.H., Dorais, M.J., Tingey, D., grown crystals: Implications for growth rate and lattice diffusion, with
Larocque, A., and Evans, N., 2006, The fate of magmatic sulfides during emphasis on accessory minerals in crustal rocks: American Mineralogist,
intrusion or eruion, Bingham and Tintic districts, Utah: Economic Geology, v. 80, p. 1179–1187.
v. 101, p. 329–345. Watson, E.B., Wark, D.A., and Thomas, J.B., 2006, Crystallization thermom-
Taylor, G.K., Grocott, J., Pope, A., and Randall, D.E., 1998, Mesozoic fault eters for zircon and rutile: Contributions to Mineralogy and Petrology,
systems, deformation and fault block rotation in the Andean forearc: A v. 151, p. 413–433.
crustal scale strike-slip duplex in the Coastal Cordillera of northern Chile: Wendt, I., and Carl, C., 1991, The statistical distribution of the mean squared
Tectonophysics, v. 299, p. 93–109. weighted deviation: Chemical Geology, v. 86, p. 275–285.
Tosdal, R.M., Wooden, J.L., and Bouse, R.M., 1999, Pb isotopes, ore depos- Wilke, M., and Behrens, H., 1999, The dependence of the partitioning of iron
its, and metallogenic terranes: Reviews in Economic Geology, v. 12, p. 1–28. and europium between plagioclase and hydrous tonalitic melt on oxygen
Tosdal, R.M., Cornejo, P. , Fanning, C.M., Mortensen, J.K., and Wooden, fugacity: Contributions to Mineralogy and Petrology, v. 137, p. 102–114.
J.L., 2000, Crustal interaction of central Andean porphyry-Cu related Wooden, J.L., Mazdab, F.K., Barth, A.P., Miller, C.F., and Lowery, L.E.,
granodioritic magmas [abs.]: Geological Society of America Abstracts with 2006, Temperatures (Ti) and compositional characteristics of zircon: Early
Programs, v. 32, p. 72. observations using high mass resolution on the USGS Stanford SHRIMP-
Ulrich, T., Gunther, D., and Heinrich, C.A., 1999, Gold concentrations RG: Geochimica et Cosmochimica Acta Supplement, v. 70, p. 707.
of magmatic brines and the metal budget of porphyry copper deposits: Zimmerman, A., Stein, H.J., Morgan, J.W., Markey, R.J., and Watanabe, Y.,
Nature, v. 399, p. 676–679. 2014, Re-Os geochronology of the El Salvador porphyry Cu-Mo deposit,
Vavra, G., 1994, Systematics of internal zircon morphology in major Variscan Chile: Tracking analytical improvements in accuracy and precision over the
granitoid types: Contributions to Mineralogy and Petrology, v. 117, past decade: Geochimica et Cosmochimica Acta, v. 131, p. 13–32.
p. 331–334.
Vigouroux, N., Wallace, P.J., and Kent, A.J.R., 2008, Volatiles in high-K mag-
mas from the western Trans-Mexican volcanic belt: Evidence for fluid flux-
ing and extreme enrichment of the mantle wedge by subduction processes:
Journal of Petrology, v. 49, p. 1589–1618.

272 LEE ET AL.
APPENDIX 1
Sample Descriptions, Preparation, and Methodology SHRIMP-RG Zircon Methodology

El Salvador porphyry samples Sample preparation
Nineteen samples were collected from the El Salvador district Twelve rock samples were collected from nonweathered open
and surrounding Indio Muerto region for zircon analysis via pit, underground, and drill core materials. The samples were
SHRIMP-RG technique. Sample ES-3239, quartz porphyry prepared in a steel jaw crusher and steel disk grinder to a
from the underground mine at Turquoise Gulch was collected powder at the El Salvador mine and shipped to Oregon State
by L. B. Gustafson. Samples IT-3 (rhyolite porphyry), IT-9 (L University. A Wilfley table was used on the <500-μm-size frac-
porphyry), IT-10 (X porphyry), IT-6 (basal tuff), IT-8 (Indio tion to produce a heavy mineral concentrate, from which the
Muerto rhyolite), IT-17 (quartz monzonite), and IT-14 (Sierra nonmagnetic fraction was isolated using a Frantz isodynamic
Castillo batholith) were originally collected by Corenjo et al. magnetic separator. Euhedral, well-formed zircons were
(1997). Previous analyses of zircons from the rhyolite por- handpicked under a binocular microscope. Lab procedures
phyry yielded a U/Pb age of 43 ± 1 Ma, although all zircon emphasize careful cleaning (with air, water, soap, alcohol),
fractions separated and leached from this sample were discor- between samples to mininize possibilities of cross sample
dant. The L and the K porphyry were analyzed by both Ar-Ar contamination.
and zircon U/b methods and yielded Ar-Ar ages of 41.2 ± 1.1
and 41.6 ± 1.2 Ma for the two samples, respectively, and U/ Zircon separation procedure
Pb zircon ages of 41 ± 2 and 41.8 ± 2.3 Ma, respectively. The Zircons from 15 samples representing all the major porphyry
identical zircon mineral separates from all three samples were intrusions and mineralized centers were prepared for this
reanalyzed in this study for comparison. Ten samples were study. Three zircon separates (IT-3, IT-9, and IT-10) were
collected from surface exposures, underground mine work- obtained from Cornejo et al. (1997), and 12 new rock samples
ings, and from drill core and represent a survey of the several were collected in this work (Table 1). Rock samples were pre-
distinct intrusive centers within the district (Fig. 2). Samples pared in a steel jaw crusher and steel disk grinder to a powder
with the least hydrothermal alteration were collected insofar at the El Salvador mine and shipped to Oregon State Uni-
as possible, but this was not always possible due to the exten- versity. A Wilfley table was used on the <500-μm-size frac-
sive alteration within the deposit (Gustafson and Hunt, 1975; tion to produce a heavy mineral concentrate, from which the
Gustafson et al., 2001; Watanabe and Hedenquist, 2001). nonmagnetic fraction was isolated using a Frantz isodynamic
Two samples of quartz porphyry were collected at the Old magnetic separator. Euhedral, well-formed zircons were
Camp pit. Sample ES-12808 is a quartz porphyry with quartz handpicked under a binocular microscope. Lab procedures
phenocrysts set within a fine-grained aplitic groundmass emphasize careful cleaning (with air, water, soap, alcohol),
with 2- to 3 vol % pyrite and chalcopyrite (Fig. 5A). Sample between samples to mininize possibilities of cross sample
ES-12791 is a late and brecciated quartz porphyry dike that contamination.
was affected only by low-temperature or supergene kaolinite
alteration and lacks quartz veins and chalcopyrite. The late Cathodolumiscence
quartz porphyry dike cuts the quartz veined and pyrite-chal- Zircons were mounted in a 2.5-cm-diam epoxy plug together
copyrite-bearing quartz porphyry sampled by ES-12808 and with a zircon age standard R33 (419 ± 0.4 Ma; Black et al.,
was sampled from a bench on the south-southwest side of the 2004) and a rare earth element zircon standard MAD green
main pit. (Frank Mazdab, pers. commun., 2007). The plugs were pol-
Three samples were collected from M Gulch-Copper Hill: ished to expose grain centers and photographed in reflected
L porphyry, A porphyry, and R porphyry. Gustafson and Hunt light and using a cathodoluminescence (CL) detection sys-
(1975) reported that the A porphyry is a dike that is associated tem at Stanford University with a JEOL JSM 5600 scan-
with, and emplaced at, the end stages of the L porphyry intru- ning electron microscope operating at 15-keV accelerating
sions. The L porphyry was collected on the east edge of the M potential. The CL images were used to screen the zircons;
Gulch open pit, and in this locality the A porphyry appears to “CL-dark” areas of high uranium content were selected for
represent a mafic border phase of the L porphyry intrusion. analysis due to the low uranium content in the El Salvador
The R porphyry was collected in the western part of the M samples. The most uranium-rich zones contain <20 to >3000
Gulch pit and is weakly propyliticaly alterated. ppm U, but average ~300 ppm U. CL images of the zircon
Three samples were collected from the Turquoise Gulch in each sample display a range of textures including oscil-
underground mine: a K porphyry (ES-12785a) and an L por- latory growth zones, sector zoning, and rounded cores. In
phyry (ES-12787) from the 2440 level of Inca Norte and a the current study, zircon rims displaying regular concentric
latite dike from the 2476 level of Inca Oeste.The latite sample (oscillatory) growth zones were sampled and interpreted to
was relatively fresh with minor smectite alteration as com- represent normal magmatic crystallization from zircon-satu-
pared to surface samples that are completely altered to mont- rated melt (Vavra, 1994; Hoskin, 2000; Hoskin and Schalteg-
morillonite (Gustafson et al., 2006). ger, 2003). Identifiable truncated or resorbed cores were
South of Turquoise Gulch, two samples were collected from typically avoided although they were sometimes sampled in
drill core: an X porphyry (ES-12811) and a K porphyry (ES- cases where the core represented the only “dark” portion
12807). One sample of the L porphyry (ES-12789a) was col- of the grain. Inherited zircons with ages >45 Ma constitute
lected from Granite Gulch. <5% of all spots analyzed.

SHRIMP-RG procedure (although Pr concentrations were calculated as above), 146Nd+,

147Sm+, 153Eu+, 165Ho+, 157Gd16O+, 159Tb16O+, 163Dy16O+,
The U-Th-Pb and trace element concentrations of zircons 166Er16O+, 169Tm16O+, 172Yb16O+, 175Lu16O+, 180Hf16O+, 206Pb+,
were analyzed using the ensitive high-resolution ion micro- 232Th16O+, and 238U16O+. Both 48Ti+ and 49Ti+ were measured;
probe-reverse geometry (SHRIMP-RG) instrument at the
Stanford-U.S. Geological Survey Micro Analysis Center however, titanium concentration was typically calculated from
49Ti+. Although the 48Ti+ peak is almost 13 times larger, in
(SUMAC). Samples were analyzed during four sessions from
February 2007 to October 2008. The zircon mount was Au low Ti samples it was prone to potential interference from
coated and placed in the sample chamber where a 5-nA 16O2– ubiquitous 96Zr2+ (and also occasionally 48Ca+ in altered spots;
primary ion beam removed surface contamination and the Claiborne et al., 1996). Each mass peak was normalized to
gold coat prior to collection of positive secondary ions (cf. the 30Si+ count rate to minimize instrumental drift and sput-
Miller and Wooden, 2004). The U-Th-Pb ratios were cor- tering effects and standardized to the MAD green zircon.
rected using the 419 Ma standard zircon R33, which was ana- Aluminum, calcium, and iron count rates were monitored
lyzed after every four unknown analyses. Average count rates for any contamination by inclusions or altered zones in zir-
were determined from five to six scans of masses 90Zr216O+, con, and scans with anomalously high Al, Ca, and/or Fe were
204Pb+, 206Pb+, 207Pb+, 208Pb+, 238U+, 232Th16O+, and 238U16O+ eliminated.
with beam tuning and centering conducted on the 238U16O+ Geochronology data were reduced using the Squid ver-
peak (cf. Miller and Wooden, 2004). Count times of 8 and 24 sion 1.13 and Isoplot version 3.41 software of Ludwig (2001,
s were used for 206Pb and 207Pb, respectively. 2003); composition data were processed in Excel. Isotopic
Rare earth element (REE) concentrations were analyzed data are presented in Appendix 2 and interpreted weighted
simultaneously with U-Th-Pb as described by Mazdab and mean ages are summarized in Table 1. Uranium content for
Wooden (2006), and standardized against MAD green zircon, the samples varied widely (see above) but 30% of the samples
which was analyzed each 4 h. Elemental concentrations were analyzed had relatively low U concentrations of less than 100
derived from the average of six 1- to 2-s counting sequences ppm and consequently relatively low amounts of 207Pb and
206Pb. The El Salvador samples are relatively young, there-
of masses 139La+, 140Ce+, 146Nd+, 147Sm+, 153Eu+, 155Gd+,
163Dy16O+, 166Er16O+, 172Yb16O+, and 180Hf16O+. Analysis of Pr fore, the amount of 235U is low and the 207Pb/235U ratio and age
is considered unreliable due to an unresolvable interference has a large error that is typically 3 to 6% (1s). For this reason
from 140Ce1H+, so Pr concentration is estimated by normaliz- the 206Pb/238U ratio with its smaller error (1–5%, 1s) is used
ing to 1/3 La concentration and 2/3 Nd concentration (Wooden for the age calculation. The raw 206Pb/238U age was corrected
et al., 2006; Claiborne et al., 2006). for common 206Pb contributions using the measured common
207Pb/206Pb ratio and the assumption of concordancy (Ludwig,
During the final two sessions (February and October
2008) four samples (ES-12785a, ES-12783, ES-12792, and 2001, 2003; Schmitt et al., 2003; Miller and Wooden, 2004).
ES-12787) were reanalyzed for the entire trace element suite Because 204Pb is low in abundance, it does not yield precise
currently available for analysis on the SHRIMP-RG using estimates of common 206Pb and is not used. The ages herein
masses: 7Li+, 9Be+, 11B+, 19F+, 23Na+, 24Mg+, 27Al+, 30Si+, 31P+, are therefore the 207Pb-corrected 206Pb/238U ages and are
32S+, 35Cl+, 39K+, 40Ca+, 45Sc+, 48Ti+, 49Ti+, 51V+, 52Cr+, 55Mn+, given at 95% (2s) confidence level.
56Fe+, 74Ge+, 89Y+, 93Nb+, 93Zr1H+, 96Zr+, 139La+, 140Ce+, 141Pr+


Lee Et Al - 2017 - Magmatic Evolution at El Salvador Porphyry - Domeyko Faults

Uploaded by

Copyright:

Available Formats

Lee Et Al - 2017 - Magmatic Evolution at El Salvador Porphyry - Domeyko Faults

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lee Et Al - 2017 - Magmatic Evolution at El Salvador Porphyry - Domeyko Faults

Uploaded by

Copyright:

Available Formats

©2017 Society of Economic Geologists, Inc.

Economic Geology, v. 112, pp. 245–273

Magmatic Evolution of Granodiorite Intrusions at the El Salvador Porphyry Copper

Introduction with overlying rock to form hydrothermal ore minerals, veins,

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

70 Cerro Colorado [52-50]

DOMEYKO FAULT ZONE

Gaby Sur [43-42] O

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

El Salvador Lithology Cerro Pelado

Eocene El Salvador Intrusions & Breccias

ES-12782 - R Copper Hill

ES-3239 - Qtz 43.2 ± 0.3 Ma

ES-12787 - L 42.4 ± 0.6 Ma

ES-12785a -K 42.9* ± 0.4 Ma

61.0 ± 0.9 Ma IT-8 - Rhyolite

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

10 42.2 0.5 1.6

element concentrations of zircons were analyzed using the A. IT-10 X porphyry

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

ES-12792 Latite dike, TG 238U/206Pb

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Qtz Plg Qtz

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Major elements (wt %)

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Probability density plot for El Salvador samples

N = 224 Peak magmatism:

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

ES-12808 Quartz ppy OC Pb-loss 1**

IT-10 X ppy TG 1**

ES-12800 L ppy MG ? 3**

B vein (Re-Os age) TG Age of antecrysts(?)

2(2) Plug(run). Standard quality: 47 46 45 44 43 42 41 40 39 38

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

900 early rhyolite porphyry and quartz porphyry samples have a

A ppy rims contents and quartz porphyry zircons range to concentrations

(up to a factor of 100). Furthermore, a bimodal compositional

750 As noted previously, Hf content increases with decreasing

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

crustal melts & ES-12807

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

X ppy Qtz ppy

1200 Qzt ppy

0.6 600 L, A, &

0.5 500 X ppy

0.3 Qtz ppy 300

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

Downloaded from http://pubs.geoscienceworld.org/segweb/economicgeology/article-pdf/112/2/245/726452/245-273.pdf

0.9 Age < 41.5 Ma Age 46.0 - 43.5 Ma

0.6 Late, well mineralized