LETTERS
PUBLISHED ONLINE: 15 DECEMBER 2013 | DOI: 10.1038/NGEO2028
Copper enrichment in arc magmas controlled by
overriding plate thickness
Massimo Chiaradia
Porphyry copper systems supply about 75% of the world’s
copper1 . They form above subduction zones and are preferen-
tially associated with calc-alkaline magmas1,2 . Such magmas
result from continuous iron depletion during differentiation, in
contrast to tholeiitic magmas that show initial iron enrichment
during differentiation3 . The formation of calc-alkaline magmas
is favoured by high water content and oxygen fugacity4–7 .
These characteristics, as well as magmatic metal contents,
are thought to be imparted in the mantle source by fluids
of the subducted slab8–10 . Yet this process does not explain
why porphyry copper systems preferentially occur in thicker
arcs1,2 . Here I present a statistical assessment of more than
40,000 published geochemical analyses of magmatic rocks
from 23 Quaternary-aged volcanic arcs worldwide. I find that
magmas of thicker arcs are systematically more calc-alkaline
and more depleted in copper than magmas of thinner arcs. This
implies that the missing copper in the former accumulates as
copper sulphides within or at the base of thicker arcs. Such
copper accumulations are an essential step in forming porphyry
systems11–13 . These results suggest that the thickness of the
overriding plate provides a more important control on magma
differentiation than the composition of the mantle source, and
can explain the preferential association of porphyry copper
systems with calc-alkaline magmas and thicker arcs.
Copper contents of primary magmas are controlled by the
oxidation state of their source, which determines the stability of
residual sulphide phases, the main host of Cu (refs 8,11): higher
oxidation states in the source result in higher Cu contents of the
primary magma8 . Despite the commonly held view that arc magmas
are more hydrous and oxidized at the mantle source than mid-ocean
ridge basalts9,10 (MORB) and therefore should also be richer in
chalcophile metals (Cu, Au; ref. 8), recent studies14,15 suggest
that the oxidation state of the subarc mantle is not significantly
different from that of the mantle under ocean ridges and that
Cu contents of the primitive magmas generated in these two
contexts are similar11 . A previous study11 has also shown that arc
magmas sooner or later undergo Cu depletion during evolution and
attributed this to varying oxygen fugacity (fO2 ) conditions during
magmatic differentiation. The authors inferred that the Cu lost
by evolving arc magmas is sequestered by sulphide phases in the
lower crust. The above arguments suggest that the processes of Cu
enrichment/depletion in arc magmas, which may ultimately play
an important role in the Cu deficit of the continental crust and in
the formation of porphyry copper deposits11 , are controlled by the
upper plate. However, how this occurs is not clear.
Here I investigate the behaviour of Cu during magmatic
evolution (MgO from 10 to ∼0 wt%) in 23 Quaternary (–Miocene)
arcs (Supplementary Tables 1–2 and Dataset) with the aim
Department of Earth Sciences, University of Geneva, Rue des Maraîchers 13, Geneva 1205, Switzerland. *e-mail: Massimo.Chiaradia@unige.ch
NATURE GEOSCIENCE | VOL 7 | JANUARY 2014 | www.nature.com/naturegeoscience
© 2014 Macmillan Publishers Limited. All rights reserved
to identify upper plate factors determining magmatic Cu
enrichment/depletion. Owing to its large-scale approach, implying
comparisons among entire arc segments, this investigation is
necessarily about first-order factors controlling the behaviour of
copper in arc magmas.
The behaviour of Cu during magmatic evolution is illustrated
using plots of Cu and total iron expressed as Fe2 O3 (Fe2 O3tot ) versus
MgO contents (Fig. 1 and Supplementary Fig. 1). To reduce the bias
induced by outliers and to extract information on general trends,
data of each arc were statistically treated by calculating median
values of Cu, Fe2 O3tot and MgO for subpopulations corresponding
to intervals >∼0.5 wt% MgO (Supplementary Table 2).
Trends of Cu and Fe2 O3tot versus MgO contents of bulk rocks
differ depending on arc thickness (Fig. 1). Arcs developed on
crust >30 km thick have median Cu values decreasing at a low
rate between 10 and ∼5 wt% MgO (∼3.5 parts per million (ppm)
Cu/wt% MgO) and at a higher rate (∼9 ppm Cu/wt% MgO) below
∼5 wt% MgO. In these arcs also Fe2 O3tot decreases at a lower rate
(∼0.2 wt% Fe2 O3tot /wt% MgO) between 10 and 5 wt% MgO and at
a higher rate (1.3 wt% Fe2 O3tot /wt.% MgO) below ∼5 wt% MgO.
Starting from initial values between 50 and 100 ppm at 10 wt%
MgO (similar to Cu in thicker arcs and in MORB; ref. 11) median
Cu contents in arcs developed on crust <20 km thick steadily
increase at arc-dependent rates to maximum values of ∼200 ppm
at ∼4–6 wt% MgO (Fig. 1c). Below ∼4–6 wt% MgO, Cu contents
decrease sharply. Also in these arcs the Cu versus MgO trend
is mimicked by Fe2 O3tot : median Fe2 O3tot first increases at arc-
dependent rates from ∼10 wt% at 10 wt% MgO to a maximum
median value of ∼14 wt% at ∼4–6 wt% MgO, then sharply
decreases to ∼1 wt% at ∼0 wt% MgO. Arcs with crustal thickness
between 20 and 30 km display intermediate trends (Fig. 1e,f). It
follows that averages of median Cu and Fe2 O3tot contents of magmas
at ∼4–6 wt% MgO for the 23 arcs display significant linear negative
correlations with arc crustal thicknesses (r = 0.76, p < 0.0001 for
Cu and r = 0.71, p < 0.0003 for Fe2 O3tot , excluding the anomalously
thick Central Andes: Fig. 2a,b). Averages of median Cu and Fe2 O3tot
values at 4–6 wt% MgO are also significantly correlated for the 23
arcs (Fig. 2c). It is noteworthy that the Fe2 O3tot content at 4–6 wt%
MgO is equivalent to the tholeiitic index of ref. 4.
The global scale correlations among arc crustal thickness, Cu and
Fe2 O3tot have important implications for the understanding of first-
order factors controlling the development of calc-alkaline versus
tholeiitic trends and Cu enrichment/depletion in arc magmas.
Figure 2b shows that the thicker the arc crust the more calc-alkaline
(less Fe2 O3tot -enriched at MgO ∼4–6 wt%) are magmas (see also
refs 16,17). Magmatic H2 O contents and fO2 play a prominent role
in the development of tholeiitic versus calc-alkaline trends4–7,18 :
high water contents and fO2 lead to calc-alkaline trends through
43
LETTERS NATURE GEOSCIENCE DOI: 10.1038/NGEO2028

a Arcs on crust > 30 km thick c Arcs on crust < 20 km thick e Arcs on crust > 20 and < 30 km thick
400 400 400
(n = 6,518) (n = 3,868) (n = 3,011)
350 350 350
300 300 300

Cu (ppm)
250 250 250
Cu (ppm)

Cu (ppm)
200 200 200
150 150 150
100 100 100
50 50 50
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
MgO (wt%) MgO (wt%) MgO (wt%)

b 18 d 18 f 18
16 (n = 21,460) 16 (n = 10,752) 16 (n = 10,519)
14 14 14

Fe2O3 tot (wt%)

Fe2O3 tot (wt%)

Fe2O3 tot (wt%)

12 12 12
10 10 10
8 8 8
6 6 6
4 4 4
2 2 2
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
MgO (wt%) MgO (wt%) MgO (wt%)

Figure 1 | Plots of Cu and Fe2 O3tot versus MgO for magmatic rocks of 23 Quaternary arcs. a–f, Seven arcs are >30 km thick (a,b), 9 are <20 km thick (c,d)
and 7 are between 20 and 30 km thick (e,f). Black dots are individual bulk rock analyses (n indicates the total number of individual analyses in each plot).
Blue symbols (different shapes denote different arc thicknesses) are median values of Cu, Fe2 O3tot and MgO for subpopulations >0.5 wt% MgO
calculated for each individual arc (see Supplementary Information for plots of each individual arc and details of calculations). Thick translucent yellow lines
are liquid lines of descent obtained with MELTS (refs 21,22) under the conditions defined in Supplementary Table 3. Thick green dashed lines are liquid
lines of descent from ref. 4.

suppression of plagioclase fractionation and promotion of early
magnetite and amphibole fractionation4,6 . The inverse correlation
between tholeiitic character (Fe2 O3tot enrichment at ∼4–6 wt%
MgO) and arc thickness (Fig. 2b) suggests that magmas evolving
in thicker arcs are richer in H2 O and more oxidized than those
evolving in thinner arcs. This correlation cannot be explained by
source effects alone: in such cases the strong control of upper
plate thickness on the development of the tholeiitic/calc-alkaline
trends would not be expected. As far as the water content effect
on the calc-alkaline trend is concerned, this correlation is better
accounted for by the fact that magmatic evolution occurs on average
at deeper crustal levels in thick arcs compared with thin arcs17,19 . As
a result, water contents of magmas evolving at higher depth increase
rapidly to high values20 , even at relatively high MgO in the residual
melt17 , because water is incompatible and more soluble at higher
pressure. A simple fractional crystallization model using MELTS,
a software package for modelling phase equilibria in magmatic
systems (refs 21,22), reproduces the observed Fe2 O3tot –MgO trends
by varying H2 O contents (Fig. 1 and Supplementary Table 3).
Additionally, it is likely that fO2 correlates positively with water
contents during deep crustal evolution17 enhancing the effect of
crustal thickness on the development of the calc-alkaline trend.
Copper saturation in magmas is intimately bound to the onset of
magnetite crystallization23 , which induces sulphide saturation and
consequent Cu sequestration into either a fluid24 or, more likely,
a sulphide melt25,26 . The positive correlation between averages of
median Fe2 O3tot and Cu contents at ∼4–6 wt% MgO for the 23 arcs
(Fig. 2c) supports an Earth-scale role for magnetite controlling Cu
saturation in arc magmas. As the onset of magnetite crystallization is
controlled by crustal thickness (Fig. 2b), median Cu contents of arc
magmas at MgO ∼4–6 wt% also correlate significantly with crustal
thickness for the Quaternary arcs (Fig. 2a). Arc magmas evolving
in thinner crust are systematically enriched in Cu at intermediate
MgO contents because of retarded magnetite crystallization, which
also produces a tholeiitic trend therein (Fig. 1c,d). In contrast, arc
magmas evolving in thicker crust start to lose Cu early because of
early magnetite crystallization, which also produces a calc-alkaline
trend therein (Fig. 1a,b).
Although the magmatic rocks treated here are not associated
with mineralization, the previous discussion may explain why
specific large-scale environmental factors (that is, thick crust) and
magmatic features (that is, calc-alkaline rather than tholeiitic mag-
mas) represent a favourable general framework for the formation
of porphyry copper systems27 . The differential Cu evolution in thin
(tholeiitic) and thick (calc-alkaline) arcs implies that magmatic
products in thin arcs carry on average ∼12 Mt Cu Myr−1 km−1
arc, whereas those in thick arcs carry on average only 4 Mt Cu
Myr−1 km−1 arc (Supplementary Table 4 and Fig. 2). The ∼12 Mt
Cu Myr−1 km−1 arc carried by magmas in thin arcs correspond to
the theoretic average global Cu flux through arcs (Supplementary
Table 5), suggesting that in thin arcs Cu is quantitatively carried
by magmas. In contrast, following the same reasoning, there is
a deficit of ∼8 Mt Cu Myr−1 km−1 arc in magmas of thick arcs,
which implies loss of magmatic Cu and its sequestration within
the thick crust as copper sulphides. A ∼4–6-Myr-long magmatic
evolution in a thick crust arc could thus separate ∼30–50 Mt
of sulphide-bound Cu km−1 arc, that is, typical amounts of Cu
contained in supergiant porphyry-type copper deposits2 . It has been
suggested that sequestration of Cu by magmatic sulphides in the
lower crust or at the crust–mantle boundary during a protracted
period could play a critical role in the formation of porphyry
copper systems by creating Cu pre-concentrations in specific arc
segments that may be prone to mobilization by later magmas11–13 .
The results presented here suggest that larger amounts of Cu can be
concentrated within or at the base of the crust of thicker arc seg-
ments and may provide a consistent explanation to why porphyry
copper systems form preferentially (although not exclusively) in
thick arcs in association with calc-alkaline magmas1,2 . Nonetheless,
as shown by the heterogeneous spatial and temporal distribution
of porphyry-type deposits also within thick arcs12 , only a special

44 NATURE GEOSCIENCE | VOL 7 | JANUARY 2014 | www.nature.com/naturegeoscience

© 2014 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE DOI: 10.1038/NGEO2028 LETTERS
a 200
r = 0.76, P < 0.0001 200
160

Averages of median Cu (ppm)

160

Cu (ppm)
120

for MgO 4¬6 wt%

80
120 CA
40
0
80 5 15 25 35 45 55 65
Crust thickness (km)

40

0
5 15 25 35 45
Crust thickness (km)

b 14 14
r = 0.71, P < 0.0003 13

Fe2O3 tot (wt%)

13 12
Averages of median Fe2O3 tot

11
12 10
for MgO 4¬6 (wt%)

9 CA
11 8
7
10 6
5 15 25 35 45 55 65
9 Crust thickness (km)
8

6
5 15 25 35 45
Crust thickness (km)

c 14
Averages of median Fe2O3 tot

12
for Mg 4¬6 (wt%)

10

r = 0.78, P < 0.0001

6
0 40 80 120 160 200
Averages of median Cu (ppm)
for MgO 4¬6 wt%

Figure 2 | Correlations between crustal thickness, Cu and Fe2 O3tot in 23 Quaternary arc magmas. a–c, Individual points represent the averages of the
median Cu and Fe2 O3tot values comprised between ∼4 and 6 wt% MgO (see Methods and Supplementary Table 2) with associated 1σ uncertainties and
average crustal thickness values with associated 1σ uncertainties (Supplementary Table 1). Crustal thicknesses and uncertainties are from ref. 29 (see
Supplementary Table 1 for more details). See Supplementary Information for further information on how averages of Cu and Fe2 O3tot contents were
calculated. CA, Central Andes.

combination of regional to local scale conditions28 will ultimately
allow the successful transfer of the deep Cu accumulations to
shallower levels and the formation of large porphyry copper systems
within the generally favourable environment of thick arcs.
Methods
The data used here (Supplementary Tables 1–2; Figs 1 and 2; Supplementary
Fig. 1 and Dataset) and discussed in the text are from the Georoc database
(http://georoc.mpch-mainz.gwdg.de/georoc/) and from unpublished data of the
author (88 samples of Ecuador). They concern volcanic rocks to avoid the potential
problem of anomalous Cu concentrations of cumulitic plutonic rocks. Furthermore,
volcanic rocks are by large the most represented ones in the database for the 23
arcs. Nonetheless, plutonic rocks, with a few exceptions of anomalously high Cu
concentrations probably due to cumulate nature, have been also checked and found
to display the same general Cu behaviour as volcanic rocks. All samples that in the
Georoc database were described as affected by any kind and degree of alteration were
discarded. More than 40,000 individual bulk rock analyses of volcanic rocks from
23 Quaternary arcs (extending back to the Miocene in a few cases) have been treated
and plotted (Supplementary Tables 1–2; Figs 1 and 2; Supplementary Fig. 1).
To reduce the bias induced by outliers and to extract information on general
trends data, each arc was statistically treated by calculating the median values of
Cu, Fe2 O3tot and MgO for subpopulations corresponding to intervals >∼0.5 wt%
MgO (see Supplementary Tables 1 and 2 and Dataset). Median values of Cu and
Fe2 O3tot comprised between ∼4 and ∼6 wt% MgO were averaged for each one of
the 23 arcs (Supplementary Tables 1 and 2) and corresponding uncertainties were
calculated for the plots of Fig. 2.
Received 6 June 2013; accepted 7 November 2013;
published online 15 December 2013
References
1. Sillitoe, R. H. Porphyry copper systems. Econ. Geol. 105, 3–41 (2010).
2. Cooke, D., Hollings, P. & Walshe, J. L. Giant porphyry deposits: Characteristics,
distribution, and tectonic controls. Econ. Geol. 100, 801–818 (2005).

NATURE GEOSCIENCE | VOL 7 | JANUARY 2014 | www.nature.com/naturegeoscience 45

© 2014 Macmillan Publishers Limited. All rights reserved
LETTERS
3. Miyashiro, A. Volcanic rock series in island arcs and active continental margins.
Am. J. Sci. 274, 321–355 (1974).
4. Zimmer, M. M. et al. The role of water in generating the calc-alkaline trend:
New volatile data for Aleutian magmas and a new Tholeiitic Index. J. Petrol. 51,
2411–2444 (2010).
5. Berndt, J., Koepke, J. & Holtz, F. An experimental investigation of the influence
of water and oxygen fugacity on differentiation of MORB at 200 MPa. J. Petrol.
46, 135–167 (2005).
6. Sisson, T. W. & Grove, T. L. Experimental investigations of the role of H2 O
in calc-alkaline differentiation and subduction zone magmatism. Contrib.
Mineral. Petrol. 113, 143–166 (1993).
7. Osborn, E. F. Role of oxygen partial pressure in the crystallization and
differentiation of basaltic magma. Am. J. Sci. 257, 609–647 (1959).
8. Mungall, J. E. Roasting the mantle: Slab melting and the genesis of major Au
and Au-rich Cu deposits. Geology 30, 915–918 (2002).
9. Gill, J. B. Orogenic Andesites and Plate Tectonics (Springer, 1981).
10. Kelley, K. A. & Cottrell, E. Water and the oxidation state of subduction zone
magmas. Science 325, 605–607 (2009).
11. Lee, C-T.A. et al. Copper systematics in arc magmas and implications for
crust–mantle differentiation. Science 336, 64–68 (2012).
12. Sillitoe, R. H. Copper provinces. SEG Spec. Pub. 16, 1–18 (2012).
13. Shafiei, B., Haschke, M. & Shahabpour, J. Recycling of orogenic arc
crust triggers porphyry Cu mineralization in Kerman Cenozoic arc rocks,
southeastern Iran. Miner. Deposita 44, 265–283 (2009).
14. Lee, C-T. A. et al. The redox state of arc mantle using Zn/Fe systematics. Nature
468, 681–685 (2010).
15. Dauphas, N. et al. Iron isotopes may reveal the redox conditions of
mantle melting from Archean to present. Earth Planet. Sci. Lett. 288,
255–267 (2009).
16. Coulon, C. & Thorpe, R. S. Role of continental crust in petrogenesis of orogenic
volcanic associations. Tectonophysics 77, 79–93 (1981).
17. Lee, C-T. A., Lee, T. C. & Wu, C-T. Modeling the compositional evolution
of recharging, evacuating, and fractionating (REFC) magma chambers:
Implications for differentiation of arc magmas. Geochim. Cosmochim. Acta
http://dx.doi.org/10.1016/j.gca.2013.08.009 (2013).
18. Kelley, K. A., Cottrell, E. & Brounce, M. N. The relationship between fO2 and
calc-alkaline affinity of arc magmas. Mineral. Mag. 77, 1446 (2013).
19. Chaussard, E. & Amelung, F. Precursory inflation of shallow magma
reservoirs at west Sunda volcanoes detected by InSAR. Geophys. Res. Lett. 39,
L21311 (2012).
46
© 2014 Macmillan Publishers Limited. All rights reserved
NATURE GEOSCIENCE DOI: 10.1038/NGEO2028
20. Annen, C., Blundy, J. D & Sparks, R. S. J. The genesis of intermediate and silicic
magmas in deep crustal hot zones. J. Petrol. 47, 505–539 (2006).
21. Ghiorso, M. S. & Sack, R. O. Chemical mass transfer in magmatic processes. IV.
A revised and internally consistent thermodynamic model for the interpolation
and extrapolation of liquid–solid equilibria in magmatic systems at elevated
temperatures and pressures. Contrib. Mineral. Petrol. 119, 197–212 (1995).
22. Asimow, P. D. & Ghiorso, M. S. Algorithmic modifications extending MELTS
to calculate subsolidus phase relations. Am. Mineral. 83, 1127–1131 (1998).
23. Jenner, F. E., O’Neill, H. St. C., Arculus, R. J. & Mavrogenes, J. A. The magnetite
crisis in the evolution of arc-related magmas and the initial concentration of
Au, Ag and Cu. J. Petrol. 51, 2445–2464 (2010).
24. Sun, W., Arculus, R. J., Kamenetski, V. S. & Binns, R. A. Release of gold-bearing
fluids in convergent margin magmas prompted by magnetite crystallization.
Nature 431, 975–978 (2004).
25. Nadeau, O., Williams-Jones, A. E. & Stix, J. Sulphide magma as a source of
metals in arc-related magmatic hydrothermal ore fluids. Nature Geosci. 3,
501–505 (2010).
26. Chiaradia, M., Ulianov, A., Kouzmanov, K. & Beate, B. Why large porphyry Cu
deposits like high Sr/Y magmas? Sci. Rep. 2, 685 (2012).
27. Clark, A. H. Are outsize porphyry copper deposits either anatomically or
environmentally distinctive? SEG Spec. Pub. 2, 213–284 (1995).
28. Tosdal, R. M. & Richards, J. P. Magmatic and structural controls on the
development of porphyry Cu±Mo±Au deposits. Rev. Econom. Geol. 14,
157–181 (2001).
29. Zellmer, G. Some first-order observations on magma transfer from mantle
wedge to upper crust at volcanic arcs. Geol. Soc. Lond. Spec. Publ. 304,
15–31 (2008).
Acknowledgements
A review by C-T. A. Lee (Rice University) contributed to significantly improve an earlier
version of this work. This study was financially supported by the Swiss National Science
Foundation (Project no. 200020_137663).
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints.
Competing financial interests
The author declares no competing financial interests.
NATURE GEOSCIENCE | VOL 7 | JANUARY 2014 | www.nature.com/naturegeoscience