E
Earth’s Atmosphere
Daniele L. Pinti
GEOTOP, Research Center on the Dynamics of Earth System,
Université du Québec à Montréal, Montréal, QC, Canada
Definition
The Earth’s atmosphere is the outer layer of our planet. This
layer contains gases, commonly referred to as air.
Overview
The atmosphere is retained by the Earth’s gravity, although
some of it is lost to outer space through thermal escape,
known as Jeans escape. By volume, dry air contains
78.08% N2, 20.94% O2, 0.933% Ar, and 409 ppmv CO2
(CO2 data from the Mauna Loa Observatory, April 2017 at
https://www.esrl.noaa.gov/gmd/ccgg/trends/monthly.html).
The noble gases He, Ne, Kr, and Xe, together with CH4 and
H2O (as water vapor), occur in trace concentrations (see
Table 1). The total mass of the atmosphere is approximately
5.15
1018 kg, mainly concentrated in the first layer, called
the troposphere, which extends from the topographic sur-
face of the planet to 12 km above it.
The atmosphere is divided into five layers, from the sur-
face to outer space: the troposphere (0–12 km), the strato-
sphere (12–50 km), the mesosphere (50–80 km), the
thermosphere (80–700 km), and the exosphere
(700–10,000 km). The exosphere merges with the solar
wind. The current chemical composition of the atmosphere
is the result of several extraterrestrial volatile sources – wet
asteroids and comets – deeply modified by the activity of life.
Due to these latter processes, the Earth’s atmosphere can be
classified as habitable. The most apparent effect of biological
activity on Earth is the progressive accumulation of molecular
# Springer International Publishing AG, part of Springer Nature 2018
W. M. White (ed.), Encyclopedia of Geochemistry,
https://doi.org/10.1007/978-3-319-39312-4
oxygen (O2) since the end of the Archean and the formation of
the ozone (O3) layer.
How the atmosphere formed has fascinated philosophers,
scholars, and scientists for centuries, since ancient Greece.
The advent of cosmochemistry and isotope geochemistry in
the twentieth century provided valuable contributions to
understanding how the Solar system formed and how rocky
planets acquired their atmospheres and oceans. However,
recent discoveries made by the Rosetta mission on comet
67P/Churyumov–Gerasimenko might revolutionize our theo-
ries of the Earth’s formation and the evolution of its differen-
tiated geochemical reservoirs.
History of the Origins
One of the first treatises on the physics of the atmosphere is
Meteorologica by Aristotle (384–322 BC), the title of which
comes from the ancient Greek mete
oron (metéōron, “high in
the sky”). In this work, Aristotle attempted to describe every-
thing of a physical nature in the sky. One of the philosophical
thoughts of Aristotle was that earthquakes resulted from
winds within the Earth caused by the Earth’s own heat, as
well as heat from the Sun. Volcanoes, he thought, marked the
points at which these winds finally escaped from inside the
Earth into the atmosphere (Aristotle and Lee 1952). Prior to
Aristotle, Anaximander of Miletus (611–549) speculated
that the Earth began as fluid, some of which dried to become
solid earth and some of which evaporated to become the
atmosphere.
This vision, in which the atmosphere was derived from the
degassing of the interior of the Earth, lasted for centuries. In
one of the first treatises of geology, Mundus subterraneus,
quo universae denique naturae divitiae, the German Jesuit
and scholar Athanasius Kircher (1602 – 1680) illustrated the
interior of the Earth with a central core from which conduits
extend to the surface, with intermediate cavities
(pyrophylacies) filled with fluids and gases. The conduits
384 Earth’s Atmosphere

Earth’s Atmosphere, Table 1 Twenty-first century elemental compo-

sition of dry air by volume (Data modified after Gatley et al. 2008)
Constituent (chemical symbol) Volume %
Nitrogen (N2) 78.0818
Oxygen (O2) 20.9435
Argona (Ar) 0.9332
Carbon dioxideb (COb) 0.0385
Subtotals 99.9970
Neon (Ne) 0.001818
Helium (He) 0.000524
Methane (CH4) 0.00015
Krypton (Kr) 0.000114
Hydrogen (H2) 0.00005
Nitrous oxide (N2O) 0.00003
Xenon (Xe) 0.0000087
Subtotals 0.0026947
Eight trace gases, including CO, O3, SO2, NO2, NH3, 0.0003053
and H2O
Total 100.0000
Notes: Because Gatley et al. (2008) assumed that increases in CO2 result Earth’s Atmosphere, Fig. 1 Kircher’s vision of the Mundus Sub-
from combustion and respiration processes, and the increase is perfectly terraneus and its interior, in which conduits extend to the surface, with
correlated with a decrease in O2, then the volume % of O2 in the 2016 dry intermediate cavities (pyrophylacies) filled with fluids and gases (From
air should be 20.9417 Kircher 1678)
a
The amount of CO2 reported was measured in 2008. The present-day
value (April 28th 2017) is 409.48 ppmv (Mauna Loa Observatory at
https://www.esrl.noaa.gov/gmd/ccgg/trends/monthly.html)
b
that the terrestrial mantle contains at least five times the
NASA continues to use 0.9340 as the volume % for argon, as calculated
amounts of hydroxides as the global ocean mass in high-
at the beginning of the twentieth century
pressure mineral phases, such as ringwoodite (Pearson et al.
2014), so any hypothesized source of water on Earth must be
were thought to reach the surface to form volcanoes from able to explain the occurrence of a total water amount of
which the internal fire escaped to the atmosphere (Fig. 1) ~8–9
1021 kg.
(Kircher 1678). Rubey (1951) noticed that the amount of volatiles “in
In 1833, the French physicist André-Marie Ampère con- excess” of those possibly formed by rock weathering was
cluded that the Earth’s atmosphere resulted from the reduction similar to the combined amounts found in gases escaping
of oxygenated bodies by metals, leading to the release of from volcanoes, fumaroles, hot springs, and in gases trapped
nitrogen originally combined with oxygen, also explaining in igneous rocks. He concluded that during the crystallization
the origin of oxygen in the atmosphere. of melts, water and carbon dioxide were preferentially
Mid-twentieth century, American geologist William partitioned in the residual melt and then released during the
W. Rubey (1951) wrote a seminal paper entitled Geological last stages of magma differentiation. Intrusive rocks might
History of Seawater. An attempt to state the problem, in which have been the main sources of these volatiles, escaping to the
he attempted to resolve the origin of the atmosphere and the atmosphere through hot springs. Again, the hypothesis of a
oceans. volatile source from the interior of the Earth prevailed.
He noticed that the most common volatiles – H2O, CO2,
Cl, N, and S – are too abundant in the atmosphere, hydro-
sphere, biosphere, and in ancient sediments to be produced by Modern Theories of Earth’s Atmosphere
the weathering of the continental masses, which he assumed Formation
to be constant. For example, he estimated that only one
hundredth of the 1.37
1021 kg of H2O present in the The hypothesis that the volatiles forming the atmosphere and
hydrosphere and atmosphere could be derived from silicate hydrosphere were degassed from the Earth’s interior became
rock weathering. Continents growth throughout the Hadean even stronger with the advent of noble gas geochemistry. In
and Archean eons, reaching the present volume somewhere at 1969, American geochemist Harmon Craig and colleagues
the end of the Archean or in the Proterozoic, (e.g., found 3He anomalies in seawater samples taken at the prox-
Hawkesworth and Kemp 2006). The supply of volatiles imity of the East Pacific Rise (Clarke et al. 1969). In the same
from rock weathering would therefore be even lower than year, anomalous helium isotopic compositions (3He/4He
that estimated by Rubey. Furthermore, it is now recognized ratios) were found in the volcanic gases of Kamkatcha
Earth’s Atmosphere 385

(Mamyrin et al. 1969). Mamyrin and his colleagues argued clearly indicating the occurrence of solar 3He in the Earth’s
that the high 3He/4He ratios found was due to 3He production interior. These values are lower than the expected solar ratio
from decay of tritium released during Russian tests. Later of 108 Ra (Geiss et al. 1995), because of the ubiquitous
tritium measurements in the area dismissed this hypothesis. atmospheric contamination of terrestrial rocks.
3
He is a primordial isotope (of solar origin) trapped in the There is still debate as to whether 20Ne on Earth is also from
mantle during Earth’s accretion from the protosolar nebula the PSN (Sarda et al. 1988). 20Ne excesses over 22Ne compared
(PSN hereafter). It is extremely rare in the atmosphere and the to the ratio in the present atmosphere ((20Ne/22Ne)air = 9.80;
hydrosphere. The presence of 3He in seawater was a clear Ozima and Podosek 1983) were measured in deep CO2 well
indication that volatiles are still degassing from the mantle, gases (20Ne/22Ne up to 11.85  0.04; Ballentine and Holland
mainly at the mid-ocean ridges (presently, the 3He flux at 2008) and in magmatic rocks from the lower, non-degassed
MOR is 4  1 atoms cm2s1; Craig et al. 1975) and in mantle (20Ne/22Ne up to 13  0.7; Matsumoto et al. 1997). E
minor amount in back-arc basins (0.6–1.6 atoms cm2s1; More recently, Yokochi and Marty (2004) gave a more precise
Lan et al. 2010). This degassing is the residual of a cata- value of 13.0  0.2 for the Kola plume, while Mukhopadhyay
strophic degassing event which took place within (2012) gave a closer value of 12.88  0.06. 20Ne excesses
50–100 Ma of the Earth’s formation (Marty 1989; Kunz could either be accounted for by protosolar Ne
et al. 1998). This hypothesis was reached after the discovery (20Ne/22Ne = 13.8; Benkert et al. 1993; = 13.4; Heber et al.
of the extinct radionuclides, particularly the 129I-129Xe system 2007) trapped during the accretion of planetesimals or solar
(e.g., Butler et al. 1963). 129Xe is produced by 129I, a short- wind-implanted (SWI)-Ne (20Ne/22Ne = 12.5; Wieler 2002;
lived nuclide with a half-life of 16 Ma. 129Xe, normalized to Raquin et al. 2008). This latter Ne component is mainly
the primordial 130Xe isotope, is less abundant in the atmo- recorded in gas-rich meteorites and may have been produced
sphere (129Xe/130Xe = 6.496; Ozima and Podosek 1983) than on the meteorite surface by irradiation by the solar wind.
in the mantle basalts (129Xe/130Xe  8.2; Moreira et al. 1998). Models of solar system dynamics suggest that the Earth
This means that atmospheric 129Xe was outgassed from the formed in a region of the nascent Solar system too energetic to
mantle when the parent element was still present, and the two permit the trapping of volatile elements in the planetary
reservoirs were then essentially isolated from each other. embryos (Fig. 2). The accreting material was essentially dry
Residual 129I in the mantle would have continued to produce (e.g., Wanke and Gold 1981). In his model, Javoy (1997)
129
Xe, creating the isotopic difference between the two reser- assumed that the Earth was made almost entirely of enstatite
voirs. This model is consistent with that of the evolution of chondrites, the driest objects in the solar system, consisting of
40
Ar/36Ar in the atmosphere (e.g., Sarda et al. 1985). Like
3
He, 36Ar is a primordial isotope. 40Ar, present in negligible
amounts at the time of terrestrial accretion, has been produced
by the decay of 40K, present in the continental crust through-
out the Earth’s history. The 40Ar/36Ar ratio in the atmosphere
evolved from nearly zero (2
104) to the present-day value
of 298.6 by addition of radiogenic 40Ar. All models which
take continental crust growth into account predict that more
than 80% of the 36Ar was degassed very early in the Earth’s
history, while 40Ar slowly built up in the atmosphere
(Hamano and Ozima 1978; Turner 1989).
If the idea that the atmosphere was derived from volatiles
added during Earth’s accretion and then degassed from its
interior is basically correct; the source of these volatiles is still
not completely clear. The Earth’s atmosphere, also made up of
gases such as oxygen that were added more recently by the
activity of life, does not have a solar composition, except for
3
He (and possibly 20Ne; see below). The primary atmospheres
enveloping the rocky planets were stripped by powerful stel-
lar winds during the first phases of young Sun activity, known
as the T-Tauri stage, and only survived in the giant planets of
the Solar system. Earth retained a small amount of hydrogen Earth’s Atmosphere, Fig. 2 Water content (as a mass fraction of the
planetary body) in different objects of the solar system, compared to
and helium (3He). Indeed, primitive magmatic rocks show
3 the Earth’s water content, including the mass of the global oceans and
He/4He ratios of up to 60 times (Stuart et al. 2003) the five times this mass in the terrestrial mantle (8–9
1021 kg) (Redrawn
atmospheric ratio (Ra = 1.382
106; Sano et al. 2013), from Dauphas and Morbidelli (2014))
386
Earth’s Atmosphere,
Fig. 3 Major volatile (HCN) and
noble gas abundances in different
objects of the solar system (Mars,
CI-CM type chondrites, and
comets) compared to abundances
in the Earth’s atmosphere (labeled
“Earth”). Abundances are reported
as the logarithm of the ratio of the
abundances, normalized to the
solar composition. Comet
abundances are derived from
noble gas-trapping experiments in
amorphous ice (calculated at 55 K)
(Owen and Bar-Nun 1995). Note
the split in the X-axis for clarity
(Data are from Dauphas and
Morbidelli (2014))
only 0.01% water (Fig. 2). Drake and Righter (2002)
suggested an unknown primitive meteoritic material as the
building blocks of the Earth, something between ordinary and
enstatite chondrites, and almost entirely dry. In conclusion,
when the Earth arrived at the waning stages of accretion, it
was basically depleted in any sort of volatile.
This ascertainment led to the late veneer theory (Chou
1978). Basically, wet bodies, akin to carbonaceous chon-
drites, brought the main volatiles and water to a dry proto-
Earth at the end of accretion and after core formation (Wanke
and Gold 1981). The late veneer theory also elegantly
resolved another conundrum: the presence of platinum
group elements (PGE) in the mantle. Indeed, it is expected
that these highly siderophile elements (HSE) were all drained
into the core during the differentiation of the Earth. In order to
explain the mantle excesses of PGE, Morgan et al. (1981)
postulated the need for a chondritic veneer (C1 type). The
mass of chondritic material added to a dry proto-Earth was
estimated to range between 0.12% and 0.74% of the whole
mantle (Morgan et al. 1981; Dauphas and Marty 2002; Wang
and Becker 2013). We will see in the last section that the late
veneer did not contribute substantially to the volatile inven-
tory of the hydrosphere-atmosphere, which was instead con-
tributed by chondritic planetary bodies during the Earth’s
building stages.
Asteroids or Comets: The Long-Standing
Controversy
The development of cosmochemistry and isotope geochem-
istry in the second half of the twentieth century provided a
Earth’s Atmosphere
great deal of help in refining the model of Earth’s atmosphere
and hydrosphere formation.
By looking at the total atmospheric inventory of major
gaseous species and noble gases, a complex formation sce-
nario is revealed. In Fig. 3, the total inventories of 1H, 14N,
12
C, 20Ne, 36Ar, 84Kr, and 130Xe in the present-day Earth
atmosphere, on Mars’ surface, in volatile-rich carbonaceous
chondrites (CI-CM class meteorites), and in comets are nor-
malized to their solar abundances (data from Dauphas and
Morbidelli 2014). Because noble gases are chemically inert,
their abundances and isotopic ratios should provide solid
constraints on the cosmic reservoirs from which the volatiles
originated.
The major volatile elements (H, C, N) in the Earth’s
atmosphere are one to two orders of magnitude lower than
in the chondrites and comets but show a similar pattern.
However, the cometary pattern shows a slight enrichment in
N with respect to C compared with the present-day Earth’s
atmosphere (i.e., the C/N ratio in comets is lower than on
Earth; Fig. 3). Primordial noble gas isotopes are also depleted
in the Earth’s atmosphere compared with the other reservoirs,
with mass-dependent depletion; Ne being the most depleted
and Xe the least depleted compared to the solar abundance.
The patterns of Mars and Earth are identical and very close to
that of chondrites, except for Xe, which is depleted compared
to that in chondrites. This is known as the missing Xe prob-
lem. Comets are more depleted in Ne and more enriched in Kr
than Earth, but they show Xe depletion, like Mars and Earth.
It is important to note that the cometary abundances reported
in Fig. 3 (and in Dauphas and Morbidelli 2014) are those
derived from noble gas-trapping experiments in amorphous
ice (calculated at 55 K) (Owen and Bar-Nun 1995). The recent
Earth’s Atmosphere
Earth’s Atmosphere, Fig. 4 Isotopic composition of major volatiles
(HCN) in comets, CI-CM chondrites, and solar compared to those in the
Earth (Redrawn from Dauphas and Morbidelli (2014))
Rosetta mission on comet 67P/Churyumov–Gerasimenko
only allowed 36Ar to be measured (Marty et al. 2016).
The similarity of the noble gas abundance patterns
between the Earth and chondrites suggests that chondrites
could be the main external carrier of volatiles (including
water) on Earth. This is also supported by similar isotopic
ratio patterns of major volatiles (R = 2H/1H, 13C/12C,
15 14
N/ N), as reported in Fig. 4 in delta notation
(di(‰) = [Rreservoir/Rstd]-1)
1000, where the reference
standard (std) is V-SMOW for H, PDB for C, and air for N).
The isotopic composition of H, C, and N of Earth and of
chondrites (data compiled from Dauphas and Morbidelli
2014) are comparable (Fig. 4), while clearly different from
that of comets (except for C). Mars is not reported because its
atmospheric isotopic composition is strongly fractionated
toward higher values by massive loss of the hydrosphere-
atmosphere system (Atreya et al. 2013).
These simple observations suggest that volatile-rich chon-
drites, such as carbonaceous chondrites, are the main carrier
of water and of the atmophile elements on Earth. However,
the depletion of Xe (the so-called missing Xe problem) calls
for a possible contribution from comets. There are numerous
hypotheses to explain the missing Xe, such as hidden reser-
voirs on Earth (in clays (Bernatowicz et al. 1984); in high-
pressure phases in the mantle (Sanloup et al. 2005; Dewaele
et al. 2016); in the core (Lee and Steinle-Neumann 2006)) or
the progressive escape of Xe ions from the early Earth atmo-
sphere, possibly through the interaction with UV light from
the young Sun (Pujol et al. 2011). However, the contribution
387
of cometary xenon cannot be excluded. Based on missing Xe,
Dauphas (2003) calculated a maximum cometary mass
accreted to the Earth of 3
1018 kg (i.e., 0.00005% of the
Earth’s mass). Comets could only have contributed 2–3% the
total water budget on Earth by mass (8–9
1021kg).
There are two other convincing pieces of evidence that major
volatiles and some of the noble gases were brought to Earth by
chondritic bodies rather than by comets: the isotopic signature
of the oceanic water on Earth and the very recent analysis of
volatiles form the coma of the 67P/Churyumov–Gerasimenko
comet (Marty et al. 2016). E
The water molecule has two elements, H and O. Hydrogen
has two stable isotopes (1H and 2H, or D for deuterium), and
oxygen has three stable isotopes (16O, 17O, and 18O). The
major hydrogen reservoir in the universe is atomic or molec-
ular hydrogen. The isotopic composition of water in several
objects of our Solar system, such as comets, chondrites, or
planets, shows a systematic enrichment in deuterium, which
can be determined by measuring the 2H/1H or D/H ratio (e.g.,
Robert 2001a) (Fig. 5).
Models of the PSN evolution indicate that once they had
entered the nebula, interstellar ice grains vaporized, and the
D/H ratio of the resulting deuterium-rich water vapor was
lowered through isotopic exchange with molecular hydrogen.
Water vapor successively condensed into microscopic icy
grains, with decreasing D/H ratios, the closer they were to
the Sun. This isotopic gradient reflects the fact that the closer
to the Sun, the higher the temperature and the faster the
isotopic exchange between water and molecular hydrogen.
The D/H heliocentric gradient predicts that the D/H ratio for
water condensed at Earth’s distance (1 AU) from the Sun
should be ~80
106 or dD = 490‰ depleted compared
the global ocean (V-SMOW = 155.67
106 or dD = 0‰ by
definition) (Robert 2001a; Fig. 5). The water isotopic com-
position of the bulk Earth is only slightly depleted compared
to the global ocean (149  3
106 or dD = 43‰; Lécuyer
et al. 1998) and is very different from the PSN value of
21  5
106. Giant planets (which have primary atmo-
spheres of solar composition) have D/H ratios ranging from
21  8 to 65  2
106 (dD from 865 to 580‰), while
long period comets from the Oort cloud have D/H ratios much
higher than bulk Earth (average D/H of 296  25
106 or
dD = +900‰; Robert 2001a, b; Hartogh et al. 2011).
Conversely, carbonaceous chondrites show an average
D/H of 149  6
106 (dD = 43‰; data compiled by
Pinti 2005), which encompasses the bulk Earth value (Fig. 5).
Antarctic micrometeorites (which have a CM-CR carbona-
ceous chondrite composition; Kurat et al. 1994) show an
average value of 154  16
106 or dD = 11‰ (Engrand
et al. 1999, data compiled by Pinti 2005; Fig. 5).
The comparison of D/H ratios of water supports the
hypothesis that chondrites are the main carrier of volatiles
on Earth. However, defenders of the theory of cometary
388 Earth’s Atmosphere

Earth’s Atmosphere,
Fig. 5 Variation in the D/H ratio
in the solar system. Proto-Earth is
the expected initial D/H ratio in
the region where the Earth formed,
based on isotopic exchange
models between molecular H and
water (Robert 2001a). The bulk
Earth value range is from Lécuyer
et al. (1998). Gaseous planets,
protosolar nebula (PSN) and Oort
cloud comet D/H ratios are
compiled by Robert (2001b) and
Hartogh et al. (2011). Chondrite
and Antarctic Micrometeorite
(AMM) data are from a literature
compilation in Pinti (2005). D/H
for the
67P/Churyumov–Gerasimenko
comet is from Altwegg
et al. (2015)

contribution to water and volatiles on Earth contested that
comets formed within the Solar system, in the Kuiper belt (the
short-period Jupiter-family comets), are the best candidate to
have contributed to the Earth’s wetness. Most of the data
available are from Oort cloud comets that might have con-
densed water from the outer Solar system, where isotopic
exchange is slow and deuterium is enriched in the water
phase. The measurement of a “terrestrial” D/H ratio of
161  24
106 or dD = +34‰ in the Jupiter family
comet, 103P/Hartley 2, by the Herschel telescope refueled
the debate (Hartogh et al. 2011). However, very recent D/H
data obtained by in situ coma measurement of the Jupiter
family 67P/Churyumov–Gerasimenko comet by the
ROSINA mass spectrometer onboard the ROSETTA orbiter
seems to have settled the controversy regarding the origin of
water. Indeed 67P/Churyumov–Gerasimenko is a short-
period comet, presumably originating from the Kuiper belt
and yet has the highest D/H ratio among comets. Altwegg
et al. (2015) measured a D/H of 530  70
106 or
dD = +2400‰ in the 67P/Churyumov–Gerasimenko comet
(i.e., three times higher than in the bulk Earth; Fig. 5).
More interestingly, the first measurements of noble gas
36
Ar, together with major volatiles H, C, and N in the coma
of 67P/Churyumov–Gerasimenko comet, allowed to calcu-
late the contribution of chondritic versus cometary material to
the Earth volatile inventory (Fig. 6; Marty et al. 2016).
Figure 6 reports the 36Ar/H2O (a), 36Ar/N (b), and 36Ar/C
(c, d) versus D/H mixing diagrams between cometary and
chondritic (asteroidal) end-members (after Marty et al. 2016).
Bulk Earth and Earth surface (hydrosphere + atmosphere)
volatile compositions are also reported. From the mixing
curves, the cometary % of H, C, and N are estimated to vary
from less than 1% for H2O and C to less than 10% for N.
On the other hand, 10 to 90% of 36Ar, and by extension the
other heavy noble gases, Kr and Xe, could be of cometary
origin. To reconcile these findings with a general model where
carbonaceous chondrite material brought the main volatiles to
Earth, Marty et al. (2016) suggested that 10% of cometary
material could have been added during the Late Heavy
Bombardment (LHB), bringing argon, but not substantially
affecting the original amounts of major volatiles, brought
much earlier by asteroids of chondritic composition.
Timing and Processes of the Earth’s Atmosphere
Formation
The major volatiles composing the early atmosphere and
hydrosphere reservoirs on Earth were brought by asteroids
of chondritic composition very early during the accretion of
the Earth. There are several lines of evidence suggesting this
scenario. First, models of Solar system formation indicate that
only a few wet asteroids of carbonaceous chondrite compo-
sition, formed originally in the outer asteroid belt, were able
to deliver water in the inner Solar system during the waning
phases of the Earth’s accretion (~70% of the Earth’s mass
accreted) (Morbidelli et al. 2000). Planetary formation
models are also able to simulate the migration of wet plane-
tary embryos from distances of up to 5 AU into the Earth’s
orbit, delivering water (Raymond et al. 2006). Volatiles added
to the planet were partitioned preferentially into melt during
the magma ocean episode (complete melt of the planet at the
end of its accretion) or during the melt of its surface by the
successive impacts. Volatiles, including water, were thus
incorporated into the interior of the planet. This process
would have been brief and happened very early in the history
of the Earth. A model of catastrophic mantle degassing and
secondary atmosphere formation based on 129I-129Xe
Earth’s Atmosphere
Earth’s Atmosphere, Fig. 6 36Ar/H2O (a), 36Ar/N (b), and 36Ar/C
(c, d) versus D/H mixing diagrams, between cometary and chondritic
(asteroidal) end-members. Bulk Earth composition, labeled differently,
is from various literature sources (see Marty et al. (2016) for details on
bibliographic references). Solid curves represent hyperbolic mixing lines
indicated that mantle degassing was likely produced within
the first 50 Ma after the Solar system formation (e.g., Kunz
et al. 1998; Avice and Marty 2014), which is also the mini-
mum age of the formation of the Earth (30–60 Ma) based on
82
Hf-182 W systematics (e.g., Kleine et al. 2002; Yin et al.
2002). Consequently, the delivery of volatiles must precede
this episode of massive degassing. The impact of a Mars-sized
object colliding with the Earth 50–100 Ma after the Solar
system formation (Canup and Asphag 2001) could have
been the trigger of the catastrophic volatile degassing, includ-
ing water, from the terrestrial interior (Sleep et al. 2001; Pinti
2005).
The amount of carbonaceous chondrite material added to
the mass of the Earth would have had to have been approxi-
mately 4% of the Earth’s mass (Albarède et al. 2013) (Fig. 7)
to explain the entire volatile inventory of the Earth (H, C, Ne,
Ar, Kr, and halogens). N and Xe are more depleted on Earth
389
between cometary and chondritic end-members, the curvature factor
calculated using known elemental and isotopic compositions of the
end-members. Dotted lines and associated numbers represent the per-
centages of volatiles of cometary origin in the mixture (From Marty et al.
(2016))
than other volatiles and would require a much smaller amount
of chondritic material to be added, equivalent to that of the
late veneer (Fig. 7). To reconcile these diverging views, Marty
(2012) suggested that the missing N is now partitioned into
the core. The missing Xe could be related to late cometary
contributions (Dauphas 2003) or to Xe-specific atmospheric
loss through geological periods of time (Pujol et al. 2011).
Mass balance results reported in Fig. 7 also place strong
constraints on the timing of volatile delivery to Earth.
Excesses of PGE in the mantle require a smaller contribution
of chondritic (or cometary) material, probably less than 0.5%,
while volatiles require 4% of accreted mass. To reconcile the
volatile and PGE mass balances, we need to assume that
volatiles were mainly brought prior to core differentiation.
In this case, the accompanying PGE would have been totally
segregated into the core, leaving the mantle depleted in these
HSEs. The timing of core formation is highly variable and
390
Earth’s Atmosphere, Fig. 7 Comparison of the CI-normalized con-
centrations of terrestrial volatile elements (amounts in the mantle +
surface), with the amounts predicted by the conventional late veneer
assumption (pale brown) and by the addition of 4% CI chondrites (blue)
(Redrawn from Albarède et al. (2013))
isotopic model dependent. It varies from 3060 Ma after
Solar system formation based on 182Hf-182W systematics and
possibly terminating within 80–200 Ma after Solar system
formation, based on coupled Hf-W and Pb-Pb systems
(Nimmo and Kleine 2015). Successive delivery of PGE by
0.5–1% of chondritic material and a smaller percentage of
cometary material during a late veneer episode or even during
the Late Heavy Bombardment could explain the current PGE
excesses in the mantle (Dauphas and Marty 2002) and a late
delivery of cometary Ar (and possibly Xe) without consider-
ably affecting the inventory of major volatiles (Marty and
Meibom 2007; Marty et al. 2016).
Summary
Resolving the origin of the Earth’s atmophile elements has
been an intriguing scientific problem for centuries. Develop-
ment of geochemistry as an Earth science and planetary
discipline, particularly isotope geochemistry (noble gases,
radiogenic isotopes, and extinct radionuclides) and cosmo-
chemistry, has been an exceptional tool for deciphering
between different theories. If the scientific community gener-
ally accepts the fact that water and volatiles were brought to
Earth by planetary bodies assimilated to CI-CM class carbo-
naceous chondrites, numerous questions remain unresolved.
Among them we can mention, the mechanisms and timing of
Earth’s ingassing and successive degassing, and the relative
contribution to Earth’s volatile budget of the late veneer
during the LHB compared to the early delivery from wet
asteroids. If new volatile data from the ROSETTA mission
seem to support the hypothesis of a chondritic origin of the
Earth’s Atmosphere
main volatiles on Earth (the D/H ratio dismisses comets as the
water carriers to Earth), a new line of questioning is also
opened on the origin of the trace gases, particularly the heavier
noble gases, Ar, Kr, and Xe. This points out of the importance
of in situ measurements or home return samples in future space
missions to asteroids (Osirix-Rex, Hayabusa 2) and comets, so
as to obtain precise and pristine isotopic data to constrain a
univocal mechanism and timing of atmophile arrival to our
planet to form a stable atmosphere-hydrosphere system.
Cross-References
▶ Argon Isotopes
▶ Atmophile Elements
▶ Atmospheric Evolution
▶ Carbon
▶ Formation and Evolution of the Earth
▶ Helium Isotopes
▶ Hydrogen
▶ Krypton
▶ Neon Isotopes
▶ Nitrogen
▶ Noble Gases
▶ Water
▶ Xenon
References
Albarède F, Ballhaus C, Blichert-Toft J, Lee C-T, Marty B, Moynier F,
Yin Q-Z (2013) Asteroidal impacts and the origin of terrestrial and
lunar volatiles. Icarus 222:44–52
Altwegg K et al (2015) 67P/Churyumov-Gerasimenko, a Jupiter family
comet with a high D/H ratio. Science 347
Aristotle, Lee HDP (1952) Meteorologica. Harvard University Press,
Cambridge, MA
Atreya SK et al (2013) Primordial argon isotope fractionation in the
atmosphere of Mars measured by the SAM instrument on curiosity
and implications for atmospheric loss. Geophys Res Lett
40:5605–5609
Avice G, Marty B (2014) The iodine–plutonium–xenon age of the
moon–earth system revisited. Phil Trans R Soc A 372:20130260
Ballentine CJ, Holland G (2008) What CO2 well gases tell us about the
origin of noble gases in the mantle and their relationship to the
atmosphere. Philos Transact A Math Phys Eng Sci 366:4183–4203
Benkert J-P, Baur H, Signer P, Wieler R (1993) He, ne, and Ar from the
solar wind and solar energetic particles in lunar ilmenites and pyrox-
enes. J Geophys Res Planets 98:13147–13162
Bernatowicz TJ, Podosek FA, Honda M, Kramer FE (1984) The atmo-
spheric inventory of xenon and noble gases in shales: the plastic bag
experiment. J Geophys Res Solid Earth 89:4597–4611
Butler WA, Jeffery PM, Reynolds JH, Wasserburg GJ (1963) Isotopic
variations in terrestrial xenon. J Geophys Res 68:3283–3291
Canup RM, Asphaug E (2001) Origin of the moon in a giant impact near
the end of the Earth’s formation. Nature 412:708–712
Chou CL (1978) Fractionation of Siderophile elements in the Earth’s
upper mantle. Proceedings of 9th lunar and planetary Science Con-
ference, vol 1, pp 219–230
Earth’s Atmosphere
Clarke WB, Beg MA, Craig H (1969) Excess 3He in the sea: evidence for
terrestrial primordial helium. Earth Planet Sci Lett 6:213–220
Craig H, Clarke WB, Beg MA (1975) Excess 3He in deep water on the
East Pacific rise. Earth Planet Sci Lett 26:125–132
Dauphas N (2003) The dual origin of the terrestrial atmosphere. Icarus
165:326–339
Dauphas N, Marty B (2002) Inference on the nature and the mass of
Earth’s late veneer from noble metals and gases. J Geophys Res
Planets 107:5129. https://doi.org/10.1029/2001JE001617
Dauphas N, Morbidelli A (2014) 6.1 - Geochemical and Planetary
Dynamical Views on the Origin of Earth’s Atmosphere and Oceans.
In: Holland HD, Turekian KK (eds) Treatise on Geochemistry,
2nd edn. Elsevier, Oxford, pp 1–35
Dewaele A, Worth N, Pickard CJ, Needs RJ, Pascarelli S, Mathon O,
Mezouar M, Irifune T (2016) Synthesis and stability of xenon oxides
Xe2O5 and Xe3O2 under pressure. Nat Chem 8:784–790
Drake M, Righter K (2002) Determining the composition of the Earth.
Nature 416:39–44
Engrand C, DeLoule E, Robert F, Maurette M, Kurat G (1999) Extrater-
restrial water in micrometeorites and cosmic spherules from Antarc-
tica: an ion microprobe study. Meteorit Planet Sci 34:773–786
Gatley DP, Herrmann S, Kretzschmar H-J (2008) A twenty-first century
molar mass for dry air. HVAC&R Research 14:655–662
Geiss J, Gloeckler G, Von Steiger R (1995) Origin of the solar wind from
composition data. Space Sci Rev 72:49–60
Hamano Y, Ozima M (1978) Earth-atmosphere evolution model based
on Ar isotopic data. In: Alexander Jr EC, Ozima M (eds.), Terrestrial
rare gases. Japan Societies Press, Hakone, Kanagawa, Japan,
pp 155–171
Hartogh P, Lis DC, Bockelee-Morvan D, de Val-Borro M, Biver N,
Kuppers M, Emprechtinger M, Bergin EA, Crovisier J, Rengel M,
Moreno R, Szutowicz S, Blake GA (2011) Ocean-like water in the
Jupiter-family comet 103P/Hartley 2. Nature 478:218–220
Hawkesworth CJ, Kemp AIS (2006) The differentiation and rates of
generation of the continental crust. Chem Geol 226:134–143
Heber VS, Wiens RC, Reisenfeld DB, Allton JH, Baur H, Burnett DS,
Olinger CT, Wiechert U, Wieler R (2007) The genesis solar wind
concentrator target: mass fractionation characterised by neon iso-
topes. In: von Steiger R, Gloeckler G, Mason GM (eds) The compo-
sition of matter. Springer, New York, pp 309–316
Javoy M (1997) The major volatile elements of the earth: their origin,
behavior, and fate. Geophys Res Lett 24:177–180
Kircher A (1678) Athanasii Kircheri Mundus Subterraneus Athanasii
Kircheri Mundus Subterraneus. Apud Joannem Janssonium à
Waesberge & filios, Amsterdam
Kleine T, Munker C, Mezger K, Palme H (2002) Rapid accretion and
early core formation on asteroids and the terrestrial planets from
Hf-W chronometry. Nature 418:952–955
Kunz J, Staudacher T, Allègre CJ (1998) Plutonium-fission xenon found
in Earth’s mantle. Science 280:877–880
Kurat G, Koeberl C, Presper T, Brandstätter F, Maurette M (1994)
Petrology and geochemistry of Antarctic micrometeorites. Geochim
Cosmochim Acta 58:3879–3904
Lan TF, Sano Y, Yang TF, Takahata N, Shirai K, Pinti DL (2010) Eval-
uating earth degassing in subduction zones by measuring helium
fluxes from the ocean floor. Earth Planet Sci Lett 298:317–322
Lécuyer C, Gillet P, Robert F (1998) The hydrogen isotope composition
of seawater and the global water cycle. Chem Geol 145:249–261
Lee KKM, Steinle-Neumann G (2006) High-pressure alloying of iron
and xenon: “Missing” Xe in the Earth’s core? Journal of Geophysical
Research: Solid Earth 111:B02202
Mamyrin BA, Tolstikhin IN, Anufriev GS, Kamenskyi IL (1969) Anom-
alous isotopic composition of helium in volcanic gases. Dokl Akad
Nauk SSSR 1197. (in Russian)
Marty B (1989) Neon and xenon isotopes in MORB – implications for
the earth-atmosphere evolution. Earth Planet Sci Lett 94:45–56
391
Marty B (2012) The origins and concentrations of water, carbon, nitro-
gen and noble gases on earth. Earth Planet Sci Lett 313–314:56–66
Marty B, Meibom A (2007) Noble gas signature of the late heavy
bombardment in the Earth’s atmosphere. eEarth 2:43–49
Marty B, Avice G, Sano Y, Altwegg K, Balsiger H, Hässig M,
Morbidelli A, Mousis O, Rubin M (2016) Origins of volatile ele-
ments (H, C, N, noble gases) on earth and Mars in light of recent
results from the ROSETTA cometary mission. Earth Planet Sci Lett
441:91–102
Matsumoto T, Honda M, McDougall I, Yatsevich I, O’Reilly SY
(1997) Plume-like neon in a metasomatic apatite from the Australian
lithospheric mantle. Nature 388:162–164
Morbidelli A, Chambers J, Lunine JI, Petit JM, Robert F, Valsecchi GB,
Cyr KE (2000) Source regions and timescales for the delivery of E
water to the earth. Meteorit Planet Sci 35:1309–1320
Moreira M, Kunz J, Allègre C (1998) Rare gas systematics in popping
rocks: isotopic and elemental compositions in the upper mantle.
Science 279:1178–1180
Morgan JW, Wandless GA, Petrie RK, Irving AJ (1981) Composition of
the earth’s upper mantle – I. Siderophile trace elements in ultramafic
nodules. Tectonophysics 75:47–67
Mukhopadhyay S (2012) Early differentiation and volatile accretion
recorded in deep-mantle neon and xenon. Nature 486:101–104
Nimmo F, Kleine T (2015) Early differentiation and core formation:
processes and timescales. In: Badro J, Walter M (eds) The early
earth: accretion and differentiation, Geophysical Monograph series,
vol 212. American Geophysical Union, Washington, DC, pp 83–102
Owen T, Bar-Nun A (1995) Comets, impacts, and atmospheres. Icarus
116:215–226
Ozima M, Podosek FA (1983) Noble gas geochemistry. Cambridge
University Press, Cambridge
Pearson DG, Brenker FE, Nestola F, McNeill J, Nasdala L, Hutchison
MT, Matveev S, Mather K, Silversmit G, Schmitz S, Vekemans B,
Vincze L (2014) Hydrous mantle transition zone indicated by
ringwoodite included within diamond. Nature 507:221–224
Pinti DL (2005) The origin and evolution of the oceans. In: Gargaud M,
Barbier B, Martin H, Reisse J (eds) Lectures in astrobiology.
Springer, Berlin, pp 83–112
Pujol M, Marty B, Burgess R (2011) Chondritic-like xenon trapped in
Archean rocks: a possible signature of the ancient atmosphere. Earth
Planet Sci Lett 308:298–306
Raquin A, Moreira MA, Guillon F (2008) He, ne and Ar systematics in
single vesicles: mantle isotopic ratios and origin of the air component
in basaltic glasses. Earth Planet Sci Lett 274:142–150
Raymond SN, Quinn T, Lunine JI (2006) High-resolution simulations of
the final assembly of Earth-like planets I: terrestrial accretion and
dynamics. Icarus 183:265–282
Robert F (2001a) The origin of water on Earth. Science 293:
1056–1058
Robert F (2001b) L’origine de l’eau dans le Système Solaire telle que est
enregistrée par son rapport isotopique D/H. In: Gargaud M,
Despois D, Parisot J-P (ed) L’environnement de la Terre primitive.
Presses Universitaires de Bordeaux, Bordeaux, France, pp 79–89
Rubey WW (1951) Geologic history of seawater: an attempt to state the
problem. GSA Bull 62:1111–1147
Sanloup C, Schmidt BC, Perez EMC, Jambon A, Gregoryanz E,
Mezouar M (2005) Retention of xenon in quartz and Earth’s missing
xenon. Science 310:1174–1177
Sano Y, Marty B, Burnard P (2013) Noble gases in the atmosphere. In:
Burnard P (ed) The Noble gases as geochemical tracers. Springer,
Berlin, pp 17–31
Sarda P, Staudacher T, Allègre CJ (1985) 40Ar/36Ar in MORB glasses:
constraints on atmosphere and mantle evolution. Earth Planet Sci
Lett 72:357–375
Sarda P, Staudacher T, Allegre CJ (1988) Neon isotopes in submarine
basalts. Earth Planet Sci Lett 91:73–88
392
Sleep NH, Zahnle K, Neuhoff PS (2001) Initiation of clement surface
conditions on the earliest Earth. Proc Nat Acad Sci USA 98:
3666–3672
Stuart FM, Lass-Evans S, Godfrey Fitton J, Ellam RM (2003) High
3He/4He ratios in picritic basalts from Baffin Island and the role of
a mixed reservoir in mantle plumes. Nature 424:57–59
Turner G (1989) The outgassing history of the Earth’s atmosphere.
J Geol Soc Lond 146:147–154
Wang Z, Becker H (2013) Ratios of S, se and Te in the silicate earth
require a volatile-rich late veneer. Nature 499:328–331
Wanke H, Gold T (1981) Constitution of terrestrial planets [and discus-
sion]. Philos Trans R Soc A Math Phys Eng Sci 303:287–302
Wieler R (2002) Cosmic-ray-produced Noble gases in meteorites. Rev
Mineral Geochem 47:125–170
Yin Q, Jacobsen SB, Yamashita K, Blichert-Toft J, Telouk P, Albarède F
(2002) A short timescale for terrestrial planet formation from Hf-W
chronometry of meteorites. Nature 418:949–952
Yokochi R, Marty B (2004) A determination of the neon isotopic com-
position of the deep mantle. Earth Planet Sci Lett 225:77–88
Earth’s Continental Crust
Roberta L. Rudnick
Department of Earth Science, University of California, Santa
Barbara, CA, USA
Department of Geology, University of Maryland, College
Park, MD, USA
Definition
The continental crust is typically defined as that portion of the
outer rocky layer of the Earth that extends vertically from the
surface (subaerial or submarine) to the Mohorovicic discon-
tinuity (or Moho) and laterally to the slope break on conti-
nental shelves (Cogley 1984).
Introduction
The continental crust has an average thickness of around 35 km
(Hacker et al. 2015; Huang et al. 2013), considerably thicker
than oceanic crust, which averages 6.5 km in thickness (White
and Klein 2014). The lower density and greater thickness of the
continental crust compared to oceanic crust causes it to ride
higher on the mantle; consequently, a large proportion (70% by
area) is exposed above sea level. Data for the physical proper-
ties of the continental crust are given in Table 1.
The continental crust is arguably the best-studied and
understood region of the solid Earth, given that we live on
the continental crust, that our species evolved on the conti-
nents, and that nearly all of the resources required to sustain
our civilization derive from it. It thus has an importance
disproportionate to its mass, which is only ~0.5% of the
silicate portion of the Earth. Despite this proximity, however,
Earth’s Continental Crust
there are still first-order questions about the continental crust
for which no consensus answers exist. What is the average
composition of the continental crust, and how might that
composition changed over time? When and how did the
continental crust form? How has the volume of continental
crust changed over Earth history? What was the nature of the
earliest crust on Earth, and how did that crust, which is no
longer preserved, compare to the continental crust of today?
Each of these major questions are addressed, in turn.
Composition and Structure
While the composition of the upper continental crust is
derived by study of rocks exposed at Earth’s surface, the
structure and composition of the crust at depth have generally
been inferred from geophysical data (specifically, seismic and
heat flow data, e.g., Christensen and Mooney 1995;
McLennan and Taylor 1996; Rudnick and Fountain 1995;
Wedepohl 1995). Seismic refraction studies generally divide
the crust into layers, varying in number from two to six (see
compilations in Pasyanos et al. 2014; Rudnick and Fountain
1995). In general, geochemical models for the crust compo-
sition divide the crust into either two or three layers, compris-
ing the upper and lower (e.g., Hacker et al. 2011; Taylor and
McLennan 1985; Wedepohl 1995) or upper, middle, and
lower continental crust (Gao et al. 1998; Hacker et al. 2015;
Rudnick and Fountain 1995). In two-layer crust models the
upper crust is generally considered the heat-producing layer,
which does not extend below the upper 1/3 of the crust; the
one exception is the study of Wedepohl (1995), which splits
the crust into roughly equal proportions of upper and lower
crust. In reality, it’s unlikely that the crust is truly layered, and
these hypothetical layers are used as a matter of convenience
because many seismic refraction profiles report layers, and
also to allow discussion of different portions of the crust.
The upper crust is the most accessible portion of the
continental crust, and hence, its composition is the best
constrained. Estimates of the composition of the upper conti-
nental crust are based on widescale surface sampling (e.g.,
Eade and Fahrig 1971, 1973; Gao et al. 1998; Shaw et al.
1967, 1976; Togashi et al. 2000), and analyses of fine-grained
sediments or sedimentary rocks such as shales (Haskin and
Haskin 1966; McLennan and Taylor 1980; Nance and Taylor
1976, 1977), loess (Chauvel et al. 2014; Park et al. 2012;
Peucker-Ehrenbrink and Jahn 2001; Sauzéat et al. 2015;
Taylor et al. 1983), and river sediments (Kamber et al.
2005), which provide natural averages for relatively insoluble
trace elements in the crust that is exposed to weathering and
erosion (see summary of methods in Rudnick and Gao 2003).
In addition, the fine-grained matrices of glacial diamictites
have recently been utilized to determine the major and trace
element composition of the upper continental crust (Chen
Earth’s Continental Crust
Earth’s Continental Density Thickness
Crust, Table 1 Physical
Units g/cm3 km
parameters of different
portions of the continental Sed 2.25 1.5
crust UC 2.76 11.6
MC 2.88 11.4
LC 3.05 10
Total 34.5
Average 2.87
Sed is uppermost sedimentary layer, UC upper crust, MC middle crust, LC lower crust
Data from Huang et al. (2013)
et al. 2016; Gaschnig et al. 2016), following on from the
original approach of Goldschmidt (1933). Table 2 provides
examples of these estimates of the upper continental crust.
All estimates of upper crust composition are felsic
(SiO2 = 66 to 70 wt.%, 1.6 to 2.5 wt.% MgO) with significant
heat-producing element (HPE: K, Th, and U) concentrations
(average heat production varies from 1.3 to 1.8 mW m3). The
compositional estimate derived from river sediments and
glacial diamictites shows the influence of chemical
weathering (i.e., depletion of relatively soluble elements
such as Mg, Ca, Na, Sr, and Mo), as both sample the upper-
most crust (upper 1–2 m) that has experienced significant
chemical modification due to weathering (Gaschnig et al.
2016; Kamber et al. 2005; Li et al. 2016). This strongly
weathered signature probably does not extend through the
entire upper continental crust. Excluding the elements thus
influenced, all major element estimates fall within a factor of
1.7 of each other, and many trace element estimates fall within
a factor of two. For three elements listed in the table (e.g., Br,
Rh, Te), there is only one estimate of average upper crust
composition, and our knowledge of the upper crustal abun-
dances of these elements should be considered preliminary, at
best. For some of the trace elements that show more than a
factor of two difference between estimates, the variability
reflects an outlier composition. For example, the Shaw et al.
(1967, 1976, 1986) estimates for C, Sc, Cr, Cu, and Bi are all
considerably lower than other estimates for these elements
and may reflect the limitations of their study, as they describe:
“few sampling areas, inadequate sampling within areas, lack
of stratigraphic control” (Shaw et al. 1976). When these out-
liers are eliminated, all of the remaining estimates for these
elements fall within a factor of two of each other. Other
outliers are likely related to analytical issues (e.g., the La
estimate of Eade and Fahrig 1973; the Nb and Ta estimates
of Shaw et al. 1976; and the Sn estimate of Taylor and
McLennan 1985), as discussed in Rudnick and Gao (2003).
Similarly, the Ru value in Rudnick and Gao (2003), which
derives from Peucker-Ehrenbrink and Jahn (2001), may be an
analytical artifact (see discussion in Park et al. 2012). Elimi-
nation of these compromised values brings the overall vari-
ability of different estimates generally to within a factor of
two. The exceptions include Be, B, S, Cl, Ni, V, As, Se, Mo,
393
Mass Total Mass Total Thickness Elevation

1021 kg % % m
0.7 3.4 4.3
6.7 32.5 33.6
6.9 33.5 33.0
6.3 30.6 29.0
20.6 100 100
800
E
Sb, I, Ba, some of the heavy rare earth elements (HREE), Re,
Hg, Tl, the platinum group elements (PGEs), and the elements
mentioned above for which there is only a single estimate.
Estimates for the composition of the middle continental
crust (that part of the crust between the upper and lower crust,
generally defined seismically, see Huang et al. 2013; Rudnick
and Fountain 1995) are based on the geochemistry of
amphibolite-facies rocks (Gao et al. 1998; Huang et al.
2013; Shaw et al. 1994; Weaver and Tarney 1984) or granu-
lites, corrected for loss of K, Rb, Cs, Th, and U that is
associated with high-grade metamorphism (Rudnick and
Fountain 1995). The values published in Rudnick and Gao
(2003) are mostly averages of previous studies, with some
exceptions noted by them. The estimates range from interme-
diate to felsic compositions (SiO2 = 62 to 69 wt.%, Table 3),
with magnesium numbers (Mg#’s = 100*Mg/(Mg + FeT),
where “T” denotes all Fe as ferrous iron) that overlap with but
range to higher values than observed in upper crustal esti-
mates. Relative to the upper crust, there are very few trace
element estimates yet made for the middle crust, and of those
available, there is much greater variability among the esti-
mates. Heat production ranges from 0.75 to 1.36 mW m3 and
is generally below that of upper crust estimates. The middle
crust is currently the least well-documented portion of the
crust from a geochemical perspective.
In contrast to the middle crust, there have been a large
number of investigations into the composition of the lower
continental crust, though some of these lump middle and
lower crust together (e.g., Hacker et al. 2011, 2015; Taylor
and McLennan 1985), making direct comparison of compo-
sitions potentially misleading. Table 4 presents a subset of
lower crustal compositions from the literature (see also
Table 7 of Rudnick and Gao 2003). Many investigations use
measured seismic velocities of the deep crust (i.e., that crust
below the upper crust) and ultrasonic wave speeds of different
types of deep crustal rocks determined in the laboratory to
infer rock type, and combine these with geochemical data for
granulite-facies rocks to derive lower crustal composition
(e.g., Rudnick and Fountain 1995; Christensen and Mooney
1995; Wedepohl 1995; see review of methods by Rudnick and
Gao 2003 and more recently, Huang et al. 2013). An alterna-
tive approach is to calculate equilibrium mineralogy and wave
 394

Earth’s Continental Crust, Table 2 Selected compositional estimates for average upper continental crust, along with minimum and maximum values for each element
1 2 3 4 5 6 7 8 9 10 11 Max/min Max/min
Gaschnig
Shaw Fahrig and et al.c
et al. Eade (1968) Hu (2016)
(1967, and Eade Gao Taylor and Togashi Rudnick Kamber and and Chen
1976, and Fahrig Condie et al.a Wedepohlb McLennan et al. and Gao et al. Gao et al. Outliers
Units 1986) (1973) (1993) (1998) (1995) (1985, 1995) (2000) (2003) (2005) (2008) (2016) Othersd All removed
SiO2 wt.% 66.8 66.2 67.0 67.97 66.8 65.89 67.53 66.62 68.5 70.4 1.07
TiO2 wt.% 0.54 0.54 0.56 0.67 0.54 0.50 0.62 0.64 1.38 0.7 2.76
Al2O3 wt.% 15.05 16.11 15.14 14.17 15.05 15.17 14.67 15.4 17.6 14.6 1.24
FeOTe wt.% 4.09 4.41 4.76 5.33 4.09 4.49 4.85 5.04 7.1 5.5 1.74
MnO wt.% 0.07 0.08 0.10 0.07 0.07 0.11 0.10 0.14 0.1 2.00
MgOf wt.% 2.30 2.23 2.45 2.62 2.30 2.20 2.53 2.48 1.6 2.0 1.65 1.65
CaOf wt.% 4.24 3.44 3.64 3.44 4.24 4.19 3.90 3.59 1.1 1.4 3.78 1.23
Na2Of wt.% 3.56 3.95 3.55 2.86 3.56 3.89 2.72 3.27 0.8 1.6 4.88 1.45
K2O wt.% 3.19 2.91 2.76 2.68 3.19 3.39 2.42 2.80 1.6 3.6 2.31
P2O5 wt.% 0.15 0.16 0.12 0.16 0.15 0.20 0.12 0.15 0.18 0.2 1.67
Mg# 50.1 47.4 47.9 46.7 50.1 46.6 48.2 46.7 28.5 39.3
Li mg/g 22 20 [22] 20 21 28.2 41 31.3 30.5 2.05
Be mg/g 1.3 1.95 3.1 3 2.1 1.92 1.9 2.34 2.38
B mg/g 9.2 28 17 15 17 47 5.11
Cg mg/g 290 3290 3240 11.34 1.02
N mg/g 83 83 150 1.81
F mg/g 500 561 611 557 1.22
S mg/g 600 309 953 621 3.08
Cl mg/g 100 142 640 370 631 6.40
Scg mg/g 7 12 13.4 15 [7] 13.6h 16 14 16.5 14 12.1 2.36 1.38
V mg/g 53 59 86 98 [53] 107h 110 97 120 106 74 2.26
Cr mg/g 35 76 112 80 [35] 85h 84 92 65 73 53 3.20 2.11
Cog,i mg/g 12 18 17 [12] 17h 15 17.3 22.4 15 11.8 1.90
Nii mg/g 19 19 60 38 [19] 44h 38 47 31.6 34 26.3 3.16 3.16
Cug,i mg/g 14 26 32 [14] 25 25 28 32.4 27 20.2 2.31 1.58
Zni mg/g 52 60 70 [52] 71 74 67 73 75 73 1.44 1.44
Ga mg/g 14 18 [14] 17 17.5 19.1 18.6 18.2 1.36
Ge mg/g 1.34 1.4 1.6 1.4 1.3 1.66 1.28
As mg/g 4.4 2 1.5 6.8 4.8 5.7 4.53
Se mg/g 0.15 0.083 0.05 0.09 3.00
Br mg/g 1.6 1.6
Rb mg/g 110 85 83 82 110 112 85 84 80 94 110 1.40
Srf mg/g 316 380 289 266 [316] 350 225 320 142 91.5 4.15 1.69
Earth’s Continental Crust
Y mg/g 21 21 24 17.4 [21] 22 26 21 32 33.2 1.91
Zr mg/g 237 190 160 188 [237] 190 135 193 199 220 1.76
Nbg mg/g 26 9.8 12 [26] 12h 9 12 15 11.6 12.7 2.89 1.67
Mof mg/g 0.78 1.4 1.5 1.1 0.6 0.3 5.00 2.50
Ruj ng/g 0.34 0.079 0.030 11.33 2.63
Rh ng/g 0.018
Pd ng/g 1.46 0.5 0.52 0.80 0.53 2.92
Earth’s Continental Crust

Ag ng/g 55 55 50 53 1.10
Cd ng/g 75 79 102 98 90 60 110 1.83
In ng/g 61 50 56 66 62 1.32
Snk mg/g 1.73 2.5 5.5 2.1 2.8 2.2 2.5 3.18 1.62
Sb ng/g 300 310 200 610 400 750 418 3.75
Te ng/g 27
I ng/g 1.4 1.4 218 156
Cs mg/g 3.55 5.8 4.6h 5.5 4.9 5.4 4.9 4.3 1.63
Ba mg/g 1070 730 633 678 668 550 458 624 396 731 2.70
Lal mg/g 32.3 71 28.4 34.8 [32.3] 30 21.7 31 32.5 37 3.27 1.68
Ce mg/g 65.6 57.5 66.4 [65.7] 64 46.4 63 71.1 73 1.57
Pr mg/g 6.3 7.1 5.54 7.1 8.5 8.54 1.54
Nd mg/g 25.9 25.6 30.4 26 20.8 27 32.9 33.3 1.60
Sm mg/g 4.61 4.59 5.09 4.7 4.5 4.28 4.7 6.88 6.84 1.61
Eu mg/g 0.937 1.05 1.21 0.95 0.88 1.07 1.0 1.57 1.37 1.78
Gd mg/g 4.21 2.8 3.8 3.65 4.0 6.36 6.05 2.27
Tb mg/g 0.481 0.66 0.82 [0.5] 0.64 0.70 0.7 0.99 0.98 2.06
Dy mg/g 2.9 [2.9] 3.5 3.94 3.9 5.89 5.57 2.03
Ho mg/g 0.62 [0.62] 0.8 0.72 0.83 1.22 1.13 1.97
Er mg/g 2.3 2.39 2.3 3.37 2.3 3.27 1.47
Tm mg/g 0.33 0.39 0.30 0.51 0.37 0.51 1.70
Yb mg/g 1.47 1.91 2.26 [1.5] 2.2 2.57 2.0 3.25 2.34 3.07 2.21
Lu mg/g 0.233 0.32 0.35 [0.27] 0.32 0.4 0.31 0.49 0.36 0.48 2.10
Hf mg/g 5.8 4.3 5.12 [5.8] 5.8 4.1 5.3 5.3 6.50 1.59
Tag mg/g 5.7 0.79 0.74 1.5 1.0h 0.72 0.88 1.12 0.92 0.94 7.92 2.08
W mg/g 0.91 1.4 2 1.9 1.6 1.4 1.10 2.20
Re ng/g 0.4 0.198 0.25 0.176 2.27
Os ng/g 0.05 0.031 0.059 1.90
Ir ng/g 0.02 [0.02] 0.022 0.036 0.022 1.80
Pt ng/g 0.5 0.8 0.599 1.60
Au ng/g 1.81 1.24 [1.8] 1.5 1.46
Hg ng/g 96 12.3 56 50 7.80
Tl mg/g 0.52 1.55 0.75 0.75 0.9 0.42 0.55 0.69 3.69
Pbi mg/g 17 18 17 18 17 17h 16.9 17 20.4 15.8 1.29 1.14
395

(continued)
E
 396

Earth’s Continental Crust, Table 2 (continued)

1 2 3 4 5 6 7 8 9 10 11 Max/min Max/min
Gaschnig
Shaw Fahrig and et al.c
et al. Eade (1968) Hu (2016)
(1967, and Eade Gao Taylor and Togashi Rudnick Kamber and and Chen
1976, and Fahrig Condie et al.a Wedepohlb McLennan et al. and Gao et al. Gao et al. Outliers
Units 1986) (1973) (1993) (1998) (1995) (1985, 1995) (2000) (2003) (2005) (2008) (2016) Othersd All removed
Big ng/g 35 230 123 130 160 230 240 6.86 1.95
Th mg/g 10.3 10.8 8.6 8.95 [10.3] 10.7 8.3 10.5 11.1 11.3 1.36
U mg/g 2.45 1.5 2.2 1.55 [2.5] 2.8 2.32 2.7 3.27 2.6 2.66 2.18
Heat mW/ 1.65 1.41 1.43 1.27 1.66 1.79 1.41 1.7 1.80 1.81
production m3
Major elements recast to 100% anhydrous
Mg# = molar 100*Mg/(Mg + FeT)
a
Calculated on a sedimentary rock-free carbonate basis
b
Wedepohl’s upper crust is largely derived from the Canadian Shield composites of Shaw et al. (1967, 1976). Values taken directly from Shaw et al. are shown in brackets
c
Preferred values from Table 4 of Gaschnig et al. (2016), except for Be, Ga, In, Sn, and Tl data, which come from preferred values of Table 5 in the same reference
d
Values taken from studies of individual elements: Li value from Sauzeat et al. (2015); N value from Johnson and Goldblatt (2015); platinum group elements and Re from Park et al. (2012)
e
Total Fe as FeO
f
Values of Kamber et al. (2005) and Gaschnig et al. (2016) reflect depletion due to chemical weathering of surficial layer
g
Values of Shaw et al. are outliers
h
Value from McLennan (2001)
i
Values of Kamber et al. (2005) elevated due to anthropogenic input
j
Value in Rudnick and Gao is analytical artifact
k
Value in Taylor and McLennan is analytical artifact
l
Value in Eade and Fahrig is an analytical artifact; see text for discussion of each of these outliers
Earth’s Continental Crust
Earth’s Continental Crust 397

Earth’s Continental Crust, Table 3 Selected compositional estimates for average middle continental crust, along with the ratio of minimum and
maximum values for each element
1 2 3 4 5 6 Max/min
Weaver and Tarney Shaw et al. Rudnick and Gao et al. Rudnick and Gao
Units (1984) (1994)* Fountain (1995) (1998) (2003) Othersa All
SiO2 wt.% 68.1 69.4 62.4 64.6 63.5 1.11
TiO2 wt.% 0.31 0.33 0.72 0.67 0.69 2.35
Al2O3 wt.% 16.33 16.21 15.96 14.08 15.0 1.16
FeOTb wt.% 3.27 2.72 6.59 5.45 6.02 2.43
MnO wt.% 0.04 0.03 0.10 0.11 0.10 3.51
MgO wt.% 1.43 1.27 3.50 3.67 3.59 2.88
CaO wt.% 3.27 2.96 5.25 5.24 5.25 1.78 E
Na2O wt.% 5.00 3.55 3.30 3.48 3.39 1.52
K2O wt.% 2.14 3.36 2.07 2.52 2.30 1.82 1.84
P2O5 wt.% 0.14 0.15 0.10 0.19 0.15 1.88
Mg# 43.8 45.5 48.6 54.5 51.5
Li mg/g 20.5 7 16 12 2.93
Be mg/g 2.29 2.29
B mg/g 3.2 17 17 5.31
C mg/g 3520
N mg/g
F mg/g 524 524
S mg/g 20 20
Cl mg/g 182 182
Sc mg/g 5.4 22 15 19 4.07
V mg/g 46 118 95 107 2.57
Cr mg/g 32 43 83 69 76 2.59
Co mg/g 30 25 18 22 1.67
Ni mg/g 20 18 33 34 33.5 1.89
Cu mg/g 8 20 32 26 4.00
Zn mg/g 50 70 69 69.5 1.40
Ga mg/g 17 18 17.5 1.06
Ge mg/g 1.13 1.13
As mg/g 3.1 3.1
Se ng/g 0.064 0.064
Br mg/g
Rb mg/g 74 92 62 67 65 1.48
Sr mg/g 580 465 281 283 282 2.06
Y mg/g 9 16 22 17.0 20 2.44
Zr mg/g 193 129 125 173 149 1.54
Nb mg/g 6 8.7 8 11 10 1.83
Mo mg/g 0.3 0.60 0.60 2.00
Ru ng/g
Rh ng/g
Pd ng/g 0.76 0.76
Ag ng/g 48 48
Cd ng/g 61 61
In ng/g
Sn mg/g 1.30 1.30
Sb ng/g 280 280
Te ng/g
I ng/g
Cs mg/g 0.98 2.4 1.96 2.2 2.45
Ba mg/g 713 1376 402 661 532 3.42
La mg/g 36 22.9 17 30.8 24 2.12
(continued)
398 Earth’s Continental Crust

Earth’s Continental Crust, Table 3 (continued)

1 2 3 4 5 6 Max/min
Weaver and Tarney Shaw et al. Rudnick and Gao et al. Rudnick and Gao
Units (1984) (1994)* Fountain (1995) (1998) (2003) Othersa All
Ce mg/g 69 42.1 45 60.3 53 1.64
Pr mg/g 5.8 5.8
Nd mg/g 30 18.3 24 26.2 25 1.64
Sm mg/g 4.4 2.8 4.4 4.74 4.6 1.69
Eu mg/g 1.09 0.78 1.5 1.20 1.4 1.92
Gd mg/g 2.11 4.0 4.0 1.90
Tb mg/g 0.41 0.28 0.58 0.76 0.7 2.71
Dy mg/g 1.54 3.8 3.8 2.47
Ho mg/g 0.82 0.82
Er mg/g 2.3 2.3
Tm mg/g 0.14 0.32 2.29
Yb mg/g 0.76 0.63 2.3 2.17 2.2 3.65
Lu mg/g 0.1 0.12 0.41 0.32 0.4 4.10
Hf mg/g 3.8 3.3 4.0 4.79 4.4 1.45
Ta mg/g 1.8 0.6 0.55 0.6 3.27
W mg/g 0.60 0.60
Re ng/g
Os ng/g
Ir ng/g
Pt ng/g 0.85 0.85
Au ng/g 0.66 0.66
Hg ng/g 7.9 7.9
Tl mg/g 0.27 0.27
Pb mg/g 22 9.0 15.3 15 15.2 2.44
Bi ng/g 0.17 0.17
Th mg/g 8.4 6.4 6.1 6.84 6.5 4.86 1.73
U mg/g 2.2 0.9 1.6 1.02 1.3 0.97 2.44
Heat mW/ 1.36 0.97 1.03 0.97 1.00 0.75
production m3
Major elements recast to 100% anhydrous
Mg# = molar 100*Mg/(Mg + FeT)
a
Values taken from studies of individual elements: K, Th, and U from Huang et al. (2013)
b
Total Fe as FeO
*
Taken from the Wawa average of Table 3 in Shaw et al. with Fe2O3 converted to FeO and major elements re-normalized to 100%

speeds for granulite-facies rocks as a function of pressure and
temperature to infer lithology and composition (Behn and
Kelemen 2003; Hacker et al. 2011, 2015).
In nearly every seismic survey, the velocity of the conti-
nental crust increases with depth (see summaries in
Christensen and Mooney 1995; Pasyanos et al. 2014;
Rudnick and Fountain 1995). Temperature and pressure also
increase with depth, and while P-wave velocities decrease
with increasing temperature by ~  4
104 km s1  C1,
they increase with pressure (2
104 km s1 MPa1; see
references in Rudnick and Fountain 1995 for the temperature
and pressure derivatives). Nevertheless, the in situ velocity of
a given rock will either stay approximately the same (in areas
of low heat flow) or will decrease (in areas of high heat flow)
with depth in the crust (see Fig. 1 of Rudnick and Fountain
1995). Therefore, the increasing seismic velocity with depth
reflects either metamorphic recrystallization, causing an
sffiffiffiffiffiffiffiffiffiffiffiffi
4
kþ m
increase in the bulk and shear moduli ðV p ¼ 3
r ,
where K is the bulk modulus, m is the shear modulus, and r
is the density) and/or increasingly mafic compositions with
depth, as seismic velocity increases with decreasing SiO2 for
rocks of a given metamorphic grade (Fig. 1). Despite this
correlation between composition and seismic velocities, the
latter may vary by over 10% for a given bulk composition
depending on mineralogy (Fig. 1). Moreover, metapelitic
rocks (metamorphosed mudstones and shales) show a huge
range in seismic velocities, depending upon the original pro-
portion of clay to quartz in the sample and the amount of melt
Earth’s Continental Crust, Table 4 Selected compositional estimates for average lower continental crust, along with ratios of the minimum and maximum for each element
Max/
1 2 3 4 5 6 7 8 9 10 min
Weaver and Rudnick and Gao Taylor and Rudnick Hacker
Tarney Fountain Wedepohl et al.a McLennanb and Gao et al.b Hacker et al. Hacker et al.
Units (1984) (1995) (1995) (1998) (1985), 1995 (2003) (2011) (2015) felsic (2015) mafic Othersc All
SiO2 wt.% 62.9 53.4 59.0 59.8 54.3 53.4 64.4 61.9 48.6 1.33
Earth’s Continental Crust

TiO2 wt.% 0.5 0.82 0.85 1.04 0.97d 0.82 0.7 0.78 1.4 2.00
Al2O3 wt.% 16.0 16.9 15.8 14.0 16.1 16.9 14.8 16.1 18.1 1.29
FeOTe wt.% 5.4 8.57 7.47 9.30 10.6 8.57 6.1 6.52 10.44 1.73
MnO wt.% 0.08 0.10 0.12 0.16 0.22 0.10 0.1 0.11 0.18 2.19
MgO wt.% 3.5 7.24 5.32 4.46 6.28 7.24 2.5 3.14 6.87 2.90
CaO wt.% 5.8 9.59 6.92 6.20 8.48 9.59 3.4 5.77 10.11 2.97
Na2O wt.% 4.5 2.65 2.91 3.00 2.79 2.65 2.8 3.92 2.85 1.48
K2O wt.% 1.0 0.61 1.61 1.75 0.64d 0.61 1.5 1.54f 1.22f 1.82 2.98
P2O5 wt.% 0.19 0.10 0.21 0.10 0.1 0.21 0.23 2.30
Mg# 53.4 60.1 55.9 46.1 51.4 60.1 42.2 46.2 54.0
Li mg/g 6 13 13 11 13 7 8 2.17
Be mg/g 1.7 1.1 1.0 1.4 1.70
B mg/g 5 7.6 8.3 2 4.15
C mg/g 588 2900 4.93
N mg/g 34 34 17 2.00
F mg/g 429 703 570 1.64
S mg/g 408 231 345 1.77
Cl mg/g 278 216 250 322 1.49
Sc mg/g 31 25 26 35d 31 21.8 1.42
V mg/g 196 149 185 271d 196 84 2.33
Cr mg/g 88 215 228 123 219d 215 51 4.47
Co mg/g 38 38 36 33d 38 23.5 1.62
Ni mg/g 58 88 99 64 156d 88 31 3.19
Cu mg/g 26 37 50 90 26 14 6.43
Zn mg/g 78 79 102 83 78 71 1.44
Ga mg/g 13 17 19 18 13 19 1.48
Ge mg/g 1.4 1.24 1.6 1.3 1.29
As mg/g 1.3 1.6 0.8 0.2 8.11
Se ng/g 170 170 50 200 4.00
Br mg/g 0.28 0.3 1.07
Rb mg/g 11 11 41 56 12d 11 41 14 17 5.05
Sr mg/g 569 348 352 308 230 348 206 320 289 1.71
Y mg/g 7 16 27 18 19 16 25 20.6 21 1.70
Zr mg/g 202 68 165 162 70 68 156 127 72 2.43
399

(continued)
E
 400

Earth’s Continental Crust, Table 4 (continued)

Max/
1 2 3 4 5 6 7 8 9 10 min
Weaver and Rudnick and Gao Taylor and Rudnick Hacker
Tarney Fountain Wedepohl et al.a McLennanb and Gao et al.b Hacker et al. Hacker et al.
Units (1984) (1995) (1995) (1998) (1985), 1995 (2003) (2011) (2015) felsic (2015) mafic Othersc All
Nb mg/g 5 5.0 11 10 6.7d 5 10 7.1 4.9 2.31
Mo mg/g 0.6 0.54 0.8 0.6 1.47
Ru ng/g 0.75
Rh ng/g
Pd ng/g 2.78 1 2.8 2.80
Ag ng/g 80 51 90 65 1.77
Cd ng/g 101 97 98 100 1.04
In ng/g 52 50 50 1.04
Sn mg/g 2.1 1.34 1.5 1.7 1.57
Sb ng/g 30 90 200 100 6.67
Te ng/g
I ng/g 140 140 11 12.73
Cs mg/g 0.3 0.8 2.6 0.47d 0.3 0.2f 8.60
Ba mg/g 757 259 568 509 150 259 390 419 204 3.79
La mg/g 22 8 27 29 11 8 19 20 8 3.60
Ce mg/g 44 20 53 53 23 20 53 46 19 2.79
Pr mg/g 7.4 2.8 2.4 8 5.1 2.5 3.37
Nd mg/g 19 11 28 25 13 11 26 22 13 2.55
Sm mg/g 3.3 2.8 6.0 4.65 3.17 2.8 5.4 4.1 2.9 2.14
Eu mg/g 1.18 1.1 1.6 1.39 1.17 1.1 1.6 1.1 1.1 1.45
Gd mg/g 3.1 5.4 3.13 3.1 4.6 3.7 3.5 1.74
Tb mg/g 0.43 0.48 0.81 0.86 0.59 0.48 0.8 0.66 0.52 1.80
Dy mg/g 3.1 4.7 3.6 3.1 3.4 3.1 3.6 1.52
Ho mg/g 0.68 0.99 0.77 0.68 0.68 0.61 0.75 1.62
Er mg/g 1.9 2.2 1.9 2 1.6 2 1.38
Earth’s Continental Crust
Tm mg/g 0.19 0.32 0.24 1.32
Yb mg/g 1.2 1.5 2.5 2.29 2.2 1.5 2.5 1.5 1.8 1.67
Lu mg/g 0.18 0.25 0.43 0.38 0.29 0.25 0.5 0.24 0.3 2.08
Hf mg/g 3.6 1.9 4.0 4.2 2.1 1.9 3.3 3.5 1.7 2.48
Ta mg/g 0.6 0.8 0.6 0.7d 0.6 0.4 0.42 0.34 2.47
W mg/g 0.6 0.51 0.6d 0.6 1.17
Earth’s Continental Crust

Re ng/g 0.4 0.18 2.22

Os ng/g 0.05 0.049 1.02
Ir ng/g 0.13 0.05 2.60
Pt ng/g 2.87 2.7 1.06
Au ng/g 1.58 3.4 1.6 2.16
Hg ng/g 21 6.3 14 3.33
Tl mg/g 0.26 0.38 0.23 0.32 1.65
Pb mg/g 13 4 12.5 13 5.0d 4 14 10.2 5.3 3.50
Bi ng/g 37 380 38 200 10.27
Th mg/g 0.42 1.2 6.6 5.23 2.0d 1.2 5.6 1.26 0.75 4.86 8.80
U mg/g 0.05 0.2 0.93 0.86 0.53d 0.2 0.7 0.29 0.25 0.97 4.85
Heat mW/m3 0.12 0.19 0.85 0.75 0.34 0.19 0.71 0.29 0.22
production
Major elements recast to 100% anhydrous
Mg# = molar 100*Mg/(Mg + FeT)
a
Model CEC3
b
Combined middle and lower crust
c
Values taken from studies of individual elements: K, Th, and U from Huang et al. (2013); Li from Teng et al. (2008); N from Johnson and Goldblatt (2015); Cl and I from Muramatsu and Wedepohl
(1998, mixing 62% felsic granulite with 38% mafic granulite; nb that lower crust represents approximately 50% of total crust)
d
Value from McLennan (2001)
e
Total Fe as FeO
f
Corrected value from misprint in published paper, via personal communication from the author
401

E
402
Earth’s Continental Crust, Fig. 1 (a) Correlation between SiO2 con-
tent and laboratory measured room-temperature ultrasonic velocities of
amphibolite- and granulite-facies rocks of different bulk composition:
extraction they experienced during metamorphism. The elas-
tic properties of quartz, the second-most common crustal
mineral, are also quite unusual with respect to most crustal
minerals, causing further uncertainty (Abers and Hacker
2016). Thus, constraining the proportion of metasedimentary
rocks in the lower crust from seismic velocity is difficult
(Rudnick and Fountain 1995).
This ambiguity in using seismic velocities to infer rock type
forms the cornerstone of one of the current debates regarding
the composition of the deep crust. Whereas most studies using
seismic velocities to infer deep crustal composition have con-
cluded that the crust becomes more mafic with depth
(Christensen and Mooney 1995; Huang et al. 2013; Rudnick
and Fountain 1995; Wedepohl 1995), an alternative hypothesis
(driven by the relamination scenario, which is described in the
next section) is that the lower crust is composed of relatively
felsic rock types, including metapelites (Hacker et al. 2011,
2015). Heat flow data currently provide only broad constraints
on the composition of the deep crust: heat production must
decrease with increasing depth (see Jaupart and Mareschal
2014, for a recent review), and a felsic lower crust produces a
surface heat flow that is at the upper bounds of those permitted
by heat flow data (Hacker et al. 2011).
As can be seen from Table 4, the different compositional
models for average lower crust vary from mafic to intermediate
(48.6 to 64.4 wt.% SiO2, MgO = 2.5 to 7.2 wt.%, and Mg# = 46
to 60). Interestingly, the two studies that consider the lower crust
as the lower ⅔ to ½ of the crust fall at different ends of the
compositional spectrum: the Taylor and McLennan (1985)
lower crust is the most mafic (54 wt.% SiO2, 6.3 wt.% MgO),
whereas the Hacker et al. (2011) lower crustal model is the most
felsic (64 wt.% SiO2, 2.5 wt.% MgO). Heat production varies by
a factor of seven among all models listed in Table 4, with the
least radioactive crust being the relatively felsic composition of
Earth’s Continental Crust
mafic = 45 to 52 wt.% SiO2, intermediate = 52 to 63 wt.% SiO2, felsic
>63 wt.% SiO2. (b) Histograms of the same data shown in panel
a (Modified from Huang et al. 2013)
Weaver and Tarney (1984), who used the extremely HPE-
depleted Scourian granulites to represent the lower crust.
Composition of the Bulk Continental Crust and
Implications for Its Formation
The bulk continental crust is estimated, by all who have under-
taken such an exercise, to be “intermediate” in composition (see
Rudnick and Gao 2003 for the most recent review on the
methods used to estimate crust composition and Hacker et al.
(2015) and Huang et al. (2013) for updated perspectives on its
average composition). If cast in terms of an igneous rock type,
the bulk crust would be called “andesitic,” sensu lato, in that all
estimates have between 57 and 65 wt.% SiO2 (Table 5). This
intermediate composition distinguishes the continental crust
from all other planetary crusts recognized in our solar system,
which are dominated by basalt (Fegley 2014; McCoy and
Nittler 2014; McSween and McLennan 2014). The significance
of an “andesitic” continental crust on Earth is twofold: (1) it
points to more protracted processes of crust formation on Earth
compared to other planets, as andesite, unlike basalt, is not a
primary melt of peridotitic mantle, and (2) the majority of
andesites on Earth are generated in subduction zones (indeed,
the type locality for andesite is the Andes), suggesting a link
between generation of the continental crust and subduction, one
of the defining processes of plate tectonics on Earth (e.g., Taylor
1977).
Despite the overall andesitic bulk composition of the con-
tinental crust, andesite (or its intrusive equivalent, diorite) is
not a particularly common rock type within the crust, and the
average andesitic composition reflects the fact that the conti-
nental crust is a mixture of a wide array of rock types,
particularly mafic igneous rocks such as basalts and gabbros
Earth’s Continental Crust, Table 5 Selected compositional estimates for average bulk continental crust, along with ratios of maximum/minimum values for each element
Max/
1 2 3 4 5 6 7 8 9 10 11 12 13 min
Hacker Hacker
Weaver Christensen Rudnick Taylor and Rudnick et al.b et al.b
and Shaw and and Gao McLennan and Gao Hacker (2015) (2015)
Taylor Tarney et al. Mooney Fountain Wedepohl et al.a (1985, (2003, et al. most most
Earth’s Continental Crust

Units (1964) (1984) (1986) (1995) (1995) (1995) (1998) 1995) 2014) (2011) mafic felsic Othersc
SiO2 wt.% 60.4 63.9 64.5 62.4 60.1 62.8 64.2 57.1 60.6 65.2 56.0 65.4 1.17
TiO2 wt.% 1.0 0.6 0.7 0.9 0.7 0.7 0.8 0.9 0.72 0.7 1.1 0.7 1.83
Al2O3 wt.% 15.6 16.3 15.1 14.9 16.1 15.4 14.1 15.9 15.9 15.0 16.7 15.7 1.18
FeOTd wt.% 7.3 5.0 5.7 6.9 6.7 5.7 6.8 9.1 6.7 5.8 8.7 5.3 1.83
MnO wt.% 0.12 0.08 0.09 0.10 0.11 0.10 0.12 0.18 0.10 0.10 0.17 0.1 2.22
MgO wt.% 3.9 2.8 3.2 3.1 4.5 3.8 3.5 5.3 4.7 2.5 5.2 2.4 2.17
CaO wt.% 5.8 4.8 4.8 5.8 6.5 5.6 4.9 7.4 6.4 3.4 7.1 4.3 2.17
Na2O wt.% 3.2 4.2 3.4 3.6 3.3 3.3 3.1 3.1 3.1 3.0 3.2 3.7 1.42
K2O wt.% 2.5 2.1 2.4 2.1 1.9 2.7 2.3 1.3 1.8 1.9 1.75 2.21 2.69
P2O5 wt.% 0.24 0.19 0.14 0.20 0.20 0.18 0.1 0.1 0.2 0.2 2.41
Mg# 48.7 50.5 50.1 44.8 54.3 54.3 48.3 50.9 55.3 43.5 51.5 44.9
Li mg/g 20 11 18 17 13 17 13 1.82
Be mg/g 2.8 2.4 1.7 1.5 1.9 1.87
B mg/g 10 9.3 11 18 10 11 1.92
C mg/g 200 1990 3120 15.6
N mg/g 20 60 56 120 3.00
F mg/g 625 525 602 553 1.19
S mg/g 260 697 283 404 2.68
Cl mg/g 130 472 179 244 448 3.63
Sc mg/g 22 13 22 16 19 30 21.9 19 2.31
V mg/g 135 96 131 98 128 230 138 89 2.58
Cr mg/g 100 56 90 119 126 92 185 135 65 3.30
Co mg/g 25 26 25 24 24 29 26.6 21.3 1.36
Ni mg/g 75 35 54 51 56 46 105 59 31 3.39
Cu mg/g 55 26 24 25 38 75 27 19 3.95
Zn mg/g 70 71 73 65 81 80 72 69 1.25
Ga mg/g 15 16 15 18 18 16 18 1.21
Ge mg/g 1.5 1.4 1.25 1.6 1.3 1.28
As mg/g 1.8 1.7 3.1 1.0 2.5 3.07
Se ng/g 50 120 130 50 130 2.60
Br mg/g 2.5 1.0 0.88 2.83
Rb mg/g 90 61 76 58 78 69 37e 49 55 39 41 2.31
Sr mg/g 375 503 317 325 333 285 260 320 246 276 385 2.04
(continued)
403

E
 404

Earth’s Continental Crust, Table 5 (continued)

Max/
1 2 3 4 5 6 7 8 9 10 11 12 13 min
Hacker Hacker
Weaver Christensen Rudnick Taylor and Rudnick et al.b et al.b
and Shaw and and Gao McLennan and Gao Hacker (2015) (2015)
Taylor Tarney et al. Mooney Fountain Wedepohl et al.a (1985, (2003, et al. most most
Units (1964) (1984) (1986) (1995) (1995) (1995) (1998) 1995) 2014) (2011) mafic felsic Othersc
Y mg/g 33 14 26 20 24 17.5 20 19 24 22 18 2.36
Zr mg/g 165 210 203 123 203 175 100 132 169 120 148 2.10
Nb mg/g 20 13 20 8f 19 11 8f 8 11 7.4 8.1 2.70
Mo mg/g 1.5 1.1 0.65 1.0 0.8 2.31
Ru ng/g 0.1 0.57 5.70
Rh ng/g 0.06
Pd ng/g 0.4 1.74 1 1.5 4.36
Ag ng/g 70 70 52 80 56 1.54
Cd ng/g 200 100 80 100 80 2.50
In ng/g 100 50 50 50 2.00
Sn mg/g 2.0 2.3 1.5 2.5 1.7 1.69
Sb ng/g 200 300 200 200 200 1.50
Te ng/g 5
I ng/g 500 800 700 119 1.60
Cs mg/g 3.0 2.6 3.4 2.8 1.5e 2 2 1.70
Ba mg/g 425 707 764 390 584 614 250 456 473 329 534 3.06
La mg/g 30 28 18 30 31.6 16 20 23 17 23 1.98
Ce mg/g 60 57 42 60 60.0 33 43 56 36 49 1.82
Pr mg/g 8.2 6.7 3.9 4.9 8 4.5 5.5 2.10
Nd mg/g 28 23 20 27 27.4 16 20 26 19 23 1.75
Sm mg/g 6 4.1 3.9 5.3 4.84 3.5 3.9 5.1 3.9 4.1 1.71
Eu mg/g 1.2 1.09 1.2 1.3 1.27 1.1 1.1 1.4 1.1 1.1 1.28
Gd mg/g 5.4 4.0 3.3 3.7 4.4 3.8 3.6 1.64
Tb mg/g 0.9 0.53 0.56 0.65 0.82 0.60 0.6 0.8 0.66 0.59 1.70
Dy mg/g 3 3.8 3.7 3.6 3.5 3.9 3.0 1.30
Earth’s Continental Crust
 Ho mg/g 1.2 0.80 0.78 0.77 0.73 0.84 0.61 1.97
Er mg/g 2.8 2.1 2.2 2.1 2.1 2.3 1.7 1.65
Tm mg/g 0.48 0.24 0.30 0.32 0.28 2.00
Yb mg/g 3.0 1.5 2.0 2.0 2.2 2.2 1.9 2.3 2.1 1.6 1.96
Lu mg/g 0.50 0.23 0.33 0.35 0.35 0.30 0.30 0.4 0.33 0.24 2.17
Hf mg/g 3 4.7 5 3.7 4.9 4.71 3.0 3.7 4.0 3.1 4 1.67
Ta mg/g 2 4 0.7f 1.1 0.6 0.8e 0.7 0.6 0.54 0.52 7.69
Earth’s Continental Crust

W mg/g 1.5 1.0 0.7 1.0 1 0.7 2.18

Re ng/g 0.4 0.4 0.19 2.13
Os ng/g 0.05 0.05 0.041 1.21
Ir ng/g 0.05 0.10 0.037 2.67
Pt ng/g 0.4 1.81 0.5 4.53
Au ng/g 4 2.5 1.21 3.0 1.3 3.31
Hg ng/g 80 40 9 30 2.9 8.89
Tl mg/g 0.45 0.52 0.39 0.36 0.50 1.44
Pb mg/g 12.5 15 20 12.6 14.8 15 8.0 11 15 9.5 11.7 2.50
Bi mg/g 0.17 0.085 0.27 0.06 0.18 4.43
Th mg/g 9.6 5.7 9 5.6 8.5 7.1 4.2 5.6 7.3 4.29 4.20 2.29
U mg/g 2.7 1.3 1.8 1.4 1.7 1.2 1.1 1.3 1.4 1.13 1.09 2.48
Heat mW/ 1.6 0.93 1.32 0.93 1.28 1.01 0.7 0.89 1.05 0.85 0.88
production m3
Major elements recast to 100% anhydrous
a
Model CEC3
b
This paper presents several compositional models based on differing assumptions regarding matching composition to seismic velocities and heat flow constraints; shown here are the end-member
compositions
c
Values taken from studies of individual elements: N concentration from Johnson and Goldblatt (2015); Cl and I concentrations from Muramatsu and Wedephohl (1998); W concentration from König
et al. (2011); Hg concentration from Canil et al. (2015)
d
Total Fe as FeO
e
Updated by McLennan (2001)
f
Updated by Barth et al. (2000)
405

E
406
and felsic igneous rocks such as rhyolites and granites
(Hacker et al. 2015; Rudnick and Fountain 1995). Indeed,
many have argued that andesites themselves result from
mixing of basalt and rhyolite (e.g., Eichelberger 1975; Kent
et al. 2010; Reubi and Blundy 2009), though crystal fraction-
ation is also an important process in generating evolved
igneous rocks (Keller et al. 2015; Lee and Bachmann 2014;
Yanagi and Yamashita 1994), and there is ongoing debate
regarding the relative importance of partial melting versus
crystal fractionation in generating evolved igneous rock
types. Assuming that crust grows by mass transfer from the
mantle via melt intrusion, and because the primary melt of
mantle peridotite is basalt (or more Mg-rich lavas such as
picrites or komatiites), processes in addition to mantle melting
must have modified the crust. This mismatch between crustal
building-blocks (mantle-derived basalts) and the crustal edi-
fice (andesitic crust) has been termed the “andesite paradox”
(Kelemen et al. 2003a) or the “crust composition paradox”
(Rollinson 2008).
Processes that have been suggested to solve the crust
composition paradox include weathering, slab melting, and
preferential recycling of mafic to ultramafic rocks from the
continental lithosphere into the mantle (see reviews of Hacker
et al. 2015; and Rudnick 1995). Each of these processes is
addressed, in turn.
The weathering hypothesis, whereby soluble elements
such as Mg are lost from the continents, are taken up in altered
oceanic crust, and then preferentially recycled back into the
mantle when that crust subducts (Albarède 1998; Anderson
1982; Lee et al. 2008; Liu and Rudnick 2011), serves to
deplete Mg (and possibly other soluble elements) from basal-
tic crust. The fact that most rocks of the continental crust have
oxygen (Simon and Lecuyer 2005) and lithium (Liu and
Rudnick 2011; Teng et al. 2004) isotopic compositions that
are offset from those of primary mantle-derived magmas in
the direction expected for a weathered regolith is prima facie
evidence that weathering played a role in generating the
continental crust that we see today. Estimates for the mass
of continental crust lost to chemical weathering vary from
>20% (Lee et al. 2008) to between 15% and 65% (Liu and
Rudnick 2011).
Whereas the thermal structure of the majority of modern
subduction zones is cool enough that melting within the slab
is primarily confined to the water-laden sedimentary layer 
uppermost volcanic rocks (van Keken et al. 2011), this may
not have been the case in the ancient past (although the
question of when subduction initiated is very much an ongo-
ing debate, as discussed below). In particular, higher heat flow
expected in the Archean due to greater heat production may
have led to dehydration melting of subducted, hydrated basal-
tic crust, generating Na-rich granitic rocks, the so-called
tonalite-trondhjemite-granodiorites (TTGs, Martin 1986)
that may then interact with mantle wedge peridotite before
Earth’s Continental Crust
they cross the Moho (Martin and Moyen 2002). In this sce-
nario, the igneous flux across the Moho is not basaltic, but
rather tonalitic.
Removal of mafic lower continental crust may happen
through a variety of processes that involve density sorting
(see Hacker et al. 2011; Lee 2014). One popular hypothesis is
that mafic to ultramafic lithologies that may be present in the
deep crust may be denser than underlying mantle (e.g., if the
crust is thickened so that eclogite forms) and can be lost from
the base of the crust through density foundering (Arndt and
Goldstein 1989; Herzberg et al. 1983; Jull and Kelemen 2001;
Kay and Kay 1991), which is often referred to erroneously as
“delamination.” A more recent hypothesis suggests that sub-
duction of compositionally stratified continental or differen-
tiated arc crust leads to density segregation as the buoyant
upper layers underplate the overlying crust, while the denser,
deeper sections founder into the convecting mantle (the
“relamination” hypothesis of Hacker et al. 2011; Kelemen
and Behn 2016). Estimates for the cumulative mass of conti-
nental crust lost to lower crustal recycling vary from 0.3 to
4 times the present mass of the continental crust (Table 6).
Tectonic Setting of Crust Formation
Creation of continental crust is fundamentally an igneous pro-
cess, as the crust grows by mass addition of melts from the
mantle. Today, mantle melting occurs mainly in three distinct
plate tectonic settings: rifts, where melts are generated due to
adiabatic melting; convergent margins, where hydrous fluids/
melts from the subducting slab flux the overlying mantle wedge;
and intraplate settings, which are often attributed to mantle
plumes. Debate centers on which tectonic environments have
been most important in generating the crust that we see today.
Many workers have cited the similarity between the trace
element composition of the continental crust and convergent
margin magmas, such as the significant depletions in Nb and
Ta relative to La (Fig. 2 the elements are plotted on this figure
in order of bulk partition coefficients during adiabatic mantle
melting) and enrichments in Pb relative to neighboring ele-
ments, as evidence that most continental crust formed above
subduction zones (e.g., Barth et al. 2000; Hawkesworth and
Kemp 2006; Kelemen 1995; Rudnick 1995). Such anomalies
in multielement plots are iconic features of convergent margin
magmas and are attributed to retention of Nb and Ta in the
slab (possibly due to the presence of a Ti-bearing phase such
as rutile) and enrichment of more soluble elements such as Pb
and Sr in slab fluids and/or melts (Elliott 2003; Kelemen et al.
2003b). Using the magnitude of Nb depletion relative to La,
mass balance between different end-member compositions
have been used to suggest that 80–95% of the present conti-
nental crust formed in subduction zones (Barth et al. 2000;
Hawkesworth and Kemp 2006).
Earth’s Continental Crust 407

Earth’s Continental Crust, Table 6 Estimates of the amount of lower continental crust recycled into the mantle
% of juvenile crust lost to Proportion of present crustal masses lost to
References foundering foundering Region Basis
Gao et al. (1998) 50–71% 1.0–2.4 Dabie-Sulu, Trace element mass
China balance
Ducea (2002) 50–60% 1.0–1.5 Sierra Petrological mass
Nevada, CA balance
Plank (2005)a 25–60% 0.3–1.5 Global Th/La ratio mass
balance
Lee et al. (2007) 60–75% 1.5–3 Sierra Petrological mass
Nevada, CA balance
Clift et al. (2009)b 30–43% 0.4–0.8 Global Compilation of regional E
studies
Jagoutz and Schmidt 52–74% 1.1–2.9 Kohistan arc Petrological mass
(2013) balance
Tang et al. (2015) 65–80% 2–4 Global Eu/Eua mass balance
a
Minimum estimate
b
For orogenic plateaus only and assuming crust formation for the past 2.5 to 4.5 Ga, density of foundered crust = 3200 kg/m3, and mass of present-
day crust is 2.06
1022 kg (Huang et al. 2013)

Earth’s Continental Crust,

Fig. 2 Primitive mantle-
normalized trace element
abundances of (a) different
compositional models of the bulk
continental crust (Wedepohl 1995;
Gao et al. 1998; Taylor and
McLennan 1985; Rudnick and
Gao 2003; Hacker et al. 2015
showing the mafic and felsic end-
members). (b) Average primitive
arc lavas from oceanic and
continental settings (From
Kelemen et al. 2003b). Vertical
gray bars highlight, from left to
right: Nb and Ta (depleted in
continental crust and arc lavas), Pb
(enriched in both), Sr (enriched in
arcs but not bulk crust), and Ti
(depleted in bulk crust and
variable in arc lavas). Mantle-
normalizing values are from
McDonough and Sun (1995)

By contrast, Albarède (1998) and Stein and Hofmann
(1994), citing the apparent episodicity in crustal growth
(discussed below), suggest that much of the continental
crust formed in intraplate settings by “superplume” or “man-
tle overturn” events, respectively. They explained the
similarities between the trace element composition of the
continental crust and convergent margin magmas as being
due to later reprocessing of original plume-generated crust
in subduction zones. A more recent hypothesis regarding the
role of plumes in generating episodic continental crustal
408
growth was proposed by Arndt and Davaille (2013) whereby
large mantle plumes displaced material from the upper mantle
and accelerated subduction to generate new crust.
It is worth noting that estimates of the trace element com-
position of the bulk continental crust, while similar to one
another and to convergent margin magmas, also show differ-
ences with the latter. For example, subduction zone magmas
are typically enriched in Sr relative to Nd (also attributed to
slab fluids and/or melts), yet the bulk continental crust does
not show such an enrichment (Fig. 2). This lack of Sr enrich-
ment, and perhaps even Sr depletion, in the continental crust
may be due, in part, to chemical weathering, as Sr is highly
soluble and is preferentially lost from the continents (e.g.,
Gaschnig et al. 2016; Goldstein 1989; Rudnick 1995). Stron-
tium depletion may also be generated by removal of
plagioclase-bearing residues or cumulates from the base of
the crust, as plagioclase incorporates Sr relative to
Nd. Indeed, Tang et al. (2015) presented evidence that the
bulk continental crust possesses a negative Eu anomaly
(Eu/Eu*) (“Eu/Eu* = EuN/(SmN*GdN)1/2, where “N” sig-
nifies concentrations normalized to chondrite”), which they
attributed to this process. In addition, Plank (2005) noted that
the high Th/La ratio seen in arc lavas is inherited from
subducted terrigenous sediments and is not generated through
subduction. She suggested that the continental crust attained a
high Th/La due to intracrustal melting, followed by removal
of the low Th/La residue. So, whereas the overall pattern of
trace elements within the continental crust is similar to that of
convergent margin magmas, there are also notable differences
that must relate to the signature of crustal recycling super-
imposed on the original igneous composition, just as the
major element composition of the bulk crust requires return
of mafic materials to the mantle.
Finally, it is also worth noting that arc-like trace element
signatures (light rare earth element (LREE) enrichment, Nb
and Ta depletion, Sr and Pb enrichment) can also be generated
in intraplate settings by melting of thickened mafic crust. An
example of such magmas has been documented in Iceland
(Willbold et al. 2009) and may be an analogy for crust for-
mation during the Archean. Indeed, if most of the continental
crust formed in the Archean (see below), modern plate tec-
tonic settings may bear little relevance to the environment in
which much of the continental crust formed, as discussed in
the next section.
Temporal Evolution of the Continental Crust
The continental crust today is the product of billions of years
of evolution involving mantle melting, igneous differentia-
tion (e.g., crystal fractionation, crustal melting, and crustal
assimilation), fluid and melt transfer in subduction zones,
weathering, metamorphism, and crustal recycling, which
Earth’s Continental Crust
may occur via both subduction and foundering, as described
above. A number of lines of evidence point to a fundamental
change in the composition of the upper continental crust at the
Archean-Proterozoic (A-P) boundary (i.e., 2500 Ma) (Fig. 3),
which may signal a shift in the processes by which continental
crust formed. Evidence for this change is seen in the compo-
sitions of fine-grained terrigenous sedimentary rocks
(Fig. 3a, b), in the composition of Archean chemical sedi-
ments, and the composition of magmatic rocks (Fig. 3c, d),
and surface exposures of igneous rocks, as discussed below.
Taylor and McLennan (1985) summarized the evidence for
crustal compositional change contained within the insoluble
trace element concentrations in fine-grained terrigenous sed-
imentary rocks such as shales, mudstones, and their metamor-
phosed equivalents. This includes a decrease in first-row
transition elements (e.g., Sc, Cr, Co, Ni) and an increase in
incompatible element concentrations (e.g., Th, La) as well as
Eu anomalies of shales at the A-P boundary. Figure 3a shows
how Th/Sc increases in fine-grained terrigenous sedimentary
rocks across the A-P boundary, which is due to both an
increase in Th and a decrease in Sc concentrations. Th is a
highly incompatible trace element while Sc is a compatible
trace element, so this change reflects a change from relatively
mafic to felsic crust across the A-P boundary. Later studies of
shales and glacial diamictites have confirmed these composi-
tional changes (e.g., Condie 1993; Gaschnig et al. 2016;
McLennan and Taylor 1991; Tang et al. 2016).
The composition of chemical sedimentary rocks (banded
iron formations, carbonates, and cherts) provide insights into
the chemical composition of the ancient oceans, which, like
the modern ocean, reflect a mixture of inputs from continental
weathering (rivers) and hydrothermal vents, and outputs
through exchange with oceanic crust. Veizer and Compston
(1976) noted that the 87Sr/86Sr ratio of Archean carbonates are
similar to values expected for the Archean mantle and
suggested that this low ratio arose from weathering of an
Archean continental crust that was dominated by greenstone
belt basalts. Later studies attributed the low 87Sr/86Sr in
Archean carbonates to the influence of hydrothermal vents.
However, Kamber (2010) suggested that the observed Sr and
Eu concentrations and Nd isotopes observed in Archean
chemical sediments do not match those expected from mixing
between hydrothermal vents and rivers draining evolved con-
tinents. He instead suggested that these rather enigmatic com-
positions reflect the weathering of widespread mafic to
ultramafic volcanic surfaces on subaerial oceanic plateaus.
Studies of surface exposures of igneous rocks also suggest
a compositional change of the upper continental crust at the
A-P boundary. Using geologic maps and measured strati-
graphic sections, Condie (1993) observed that greenstone
belts show an increase in the proportion of felsic volcanic
rocks and graywackes, whereas komatiites drop to almost
zero at the A-P boundary. He also observed a systematic
Earth’s Continental Crust
Earth’s Continental Crust, Fig. 3 Illustration of various lines of
evidence for a compositional change in the composition of the continen-
tal crust across the Archean-Proterozoic (A-P) boundary. (a) Average
Th/Sc ratio in fine-grained sedimentary rocks (Modified from Taylor and
McLennan 1985). Because both Th and Sc are insoluble elements, the
values reflect the average composition of the crust exposed to weathering
and erosion for a given time period. Boxes represent time interval under
consideration and 95% confidence interval of mean. (b) Changing aver-
age MgO concentration of upper continental crust as a function of time,
change during the Archean, with Late Archean upper crust
having lower Mg, Cr, Ni, Co, and Eu/Eu* and higher K, Rb,
Ba, Th, U, and HREE compared to that of the Early Archean.
These findings are consistent with a shift from more mafic to
more felsic crust within the Archean and across the A-P
boundary. Comparing shales to surface exposures of igneous
rocks in the Archean, he found that the shales, on average,
lack HREE depletion and are more enriched in Fe, V, and Sc
compared to the surface exposures and hypothesized that the
source of the shales, which had since eroded away, was
dominated by basalt and komatiite with little HREE-depleted
TTGs. This finding is supported by an inferred drop in aver-
age upper crustal MgO content from ~15 wt.% in the Early
409
inferred from Ni/Co and Cr/Zn ratios in fine-grained sedimentary rocks.
Also shown is inferred change in crustal mass based on addition of TTGs
to original mafic crust (From Tang et al. (2016). (c) Changing Rb/Sr ratio
of juvenile crust and correlated SiO2 as a function of time based on data
for over 13,000 igneous rocks (Modified from Dhuime et al. 2015). (d)
Changes in the geochemistry of felsic igneous rocks as a function of
time, based on data for over 70,000 igneous rocks (Modified from Keller
and Schoene 2012)
Archean, to ~4 wt.% in the Late Archean based on transition
metal ratios in fine-grained sedimentary rocks (Tang et al.
2016, Fig. 3b). It remains an open question whether this
change in MgO reflects a vertical or temporal change in
upper crust composition.
Dhuime et al. (2015) used initial Sr isotope ratios and Nd
model ages of igneous rocks to calculate the Rb/Sr ratio of
juvenile crust at the time of mantle extraction. Because Rb/Sr
ratio increases systematically with SiO2 concentration, they
used these results to estimate the average SiO2 concentration
of newly generated continental crust over time (Fig. 3c). They
found that both Rb/Sr and SiO2 increased dramatically around
3.0 Ga, which they attributed to the start of plate tectonics.
410
Keller and Schoene (2012) used statistical sampling of
over 70,000 dated igneous rocks compiled from geochemical
databases and other sources to examine secular trends in the
compositional data. They found that continental basalt com-
positions record a gradual decrease in mantle melt fraction but
with an abrupt change occurring at the A-P boundary. Like-
wise, felsic igneous rocks also show a rather abrupt change in
composition at the A-P boundary from high Na/K, La/Yb,
Eu/Eu* (typical of TTGs) to lower ratios (Fig. 3d). Keller and
Schoene attributed these changes to decreasing mantle tem-
perature and changing continental crustal thickness at the A-P
boundary and inferred that the Archean crust was originally
thick and mafic. Melting in its deeper sections produced
TTGs followed by recycling of the residues to leave a rela-
tively thin and felsic crust in Archean regions today. Keller
and Schoene’s interpretations suggest that the deep continen-
tal crust, as well as the upper continental crust, was more
mafic in the Archean than in the post-Archean.
Most recently, Lee et al. (2016) suggested that the first
significant peak seen in detrital zircon dates, at ~2.8 Ga (see
next section), may reflect the transition from a mafic crust
where the magmas are not generally saturated in zircon, to a
felsic crust that grew abundant zircon. Because production of
copious amounts of felsic magmas requires water (e.g., Arndt
2013; Campbell and Taylor 1985), they also suggested that
this transition may mark the evolution from stagnant lid
Earth’s Continental Crust, Fig. 4 A selection of continental crustal
growth or preservation curves, compiled by Cawood et al. (2013). The
curves of Goodwin (1996) and Hurley and Rand (1969) are based on the
present-day surface Rb-Sr and K-Ar dates and represent minima, as they
do not take into consideration isotopic resetting due to metamorphism.
The curve of Condie and Aster (2010) is based on a combination of U-Pb
dates of zircons from both orogenic granites for which Nd whole rock
data exist and U-Pb dates of detrital zircons for which Hf isotope data
Earth’s Continental Crust
tectonics to plate tectonics, which provides a ready mecha-
nism by which to carry water to depth.
Crustal Growth
One of the most persistent debates concerning the continental
crust is the question of how the volume of the crust may have
changed through time. Did continents form early and remain
more or less in steady state as new additions were balanced by
subtraction due to recycling (e.g., Armstrong 1981, 1991)? Or
has the volume of continental crust increased over Earth
history (e.g., Dhuime et al. 2012; McCulloch and Bennett
1994; Taylor and McLennan 1985)? If the latter, what is the
shape of the growth curve? Fig. 4 shows a recent compilation
of crustal growth curves by Cawood et al. (2013). While there
is far from a consensus on this topic, the burgeoning infor-
mation being gleaned from zircons, particularly detrital zir-
cons harvested from terrigenous sedimentary rocks, points to
significant generation of felsic continental crust in the mid- to
Late Archean.
Zircons are particularly powerful recorders of geologic
history because their crystallization age can be determined
precisely via U-Pb dating (see reviews by Scherer et al. 2007;
Schoene 2014), and they contain relatively high concentra-
tions of Hf. Decay of 176Lu to 176Hf (half-life = ~37.5 Ga,
exist, as well as map distributions of dated rocks. The Allègre and
Rousseau (1984) curve is based on Sm-Nd model ages in shales, while
the curves of Belousova et al. (2010) and Dhuime et al. (2012) are
different takes on Lu-Hf model ages of detrital zircons. The remaining
curves (Armstrong 1981; Taylor and McLennan 1985) are based on
concepts about how the crust may have grown (Modified from Cawood
et al. 2013)
Earth’s Continental Crust
Scherer et al. 2001; Söderlund et al. 2004) and the relatively
low Lu/Hf ratio of zircon means that mantle derivation ages
(or model ages) can be determined for each zircon, provided
that the Lu/Hf ratio of the pre-existing crustal reservoir and
the Hf isotope mantle evolution curve can be approximated
(see Kemp and Hawkesworth 2014; Payne et al. 2016;
Vervoort and Kemp 2016 for recent reviews of the methods
and potential pitfalls involved). In particular, with the wide-
spread application of in situ analytical methods such as sec-
ondary ion mass spectrometry (SIMS) and laser ablation
inductively coupled multi-collector plasma mass spectrome-
try (LA-MC-ICP-MS), both U-Pb dates and Hf isotopes can
be determined for the same zone within a given zircon
(zircons frequently contain different domains grown at differ-
ent times, Fig. 5), or even for the same volume (e.g.,
employing laser ablation split-stream methods, Fisher et al.
2014; Kylander-Clark et al. 2013; Yuan et al. 2008) to ensure
a direct match between U-Pb date and Hf isotope model age.
Finally, oxygen isotopes of the zircons can be used to identify
zircons derived, in part or in whole, from pre-existing crust,
which will generally have non-mantle-like d18O values. In
this way, zircons that crystallized from magmas that contain
components of older continental crust can be identified, in
order to infer the proportions of juvenile to pre-existing crust
and define crustal growth curves (Kemp et al. 2006). The
results from analyses of thousands of zircons suggest that
most (~60–75%) of the continental crust formed by the end
of the Archean, possibly by processes other than subduction
zone magmatism. Assuming that it is difficult for the zircon
record to be completely obliterated by crustal recycling
because weathering and erosion disperse detrital zircons, the
results also suggest that only a small fraction of present-day
continental crust (20–25%) was present at the start of the
Archean at about 3.8 Ga. In these discussions, it is important
to note, as pointed out by Lee et al. (2016) that the U-Pb
Earth’s Continental Crust, Fig. 5 Example of a complexly zoned
zircon crystal from a granite from Cameroon. Left: cathodoluminescence
image, right: 207Pb/206Pb age map; center panel shows the position of the
individual laser spot analyses that were used to create the map on the
411
zircon record is mostly a record of felsic crust, as zircon is not
saturated in mafic melts. In addition, detrital zircons (and their
associated sedimentary rocks) can only be generated if crust is
exposed above sea level. It is conceivable that evolved crust
submerged below the oceans could have been efficiently
recycled in the Hadean without leaving a trace. However,
based on the evidence for dominantly mafic crust in the
Archean cited above, this scenario is unlikely.
A striking feature of the age distribution of continental
crust is the well-defined peaks in U-Pb zircon dates (Fig. 6).
These peaks appear in zircons from both terrigenous sedi- E
ments and igneous rocks and are found on all continents.
Given the very large number of near-concordant U-Pb dates
for zircons from the continental crust (several hundred thou-
sand), the peaks appear to be robust and are unlikely to be an
artifact of incomplete sampling. Yet the significance of these
peaks is debated. One school of thought interprets the peaks
as episodes of enhanced crustal growth, possibly driven by
superplume events or slab avalanches in the mantle (Arndt
and Davaille 2013; Condie 1998; Parman 2015). This inter-
pretation is supported by peaks at similar time intervals in Os
model ages in mantle peridotites and platinum-group alloy
grains from ophiolites, which may date continental crust
extraction events (Pearson et al. 2007). An alternative inter-
pretation suggests that crustal preservation may be uneven
(Gurnis and Davies 1986) and that the U-Pb peaks may
represent zircon growth and preservation during superconti-
nent formation (Condie and Aster 2010; Hawkesworth et al.
2010, 2013; Kemp et al. 2006). This interpretation is
supported by the apparent absence of corresponding peaks
in Lu-Hf model ages from the same zircons (Voice et al.
2011), including those that have been filtered using oxygen
isotopes to identify only zircons from juvenile crustal addi-
tions (Dhuime et al. 2012; Kemp et al. 2006). If the Lu-Hf
model age distributions prove robust (see caveats in Payne
right. The analyses reveal a ~2.8 Ga igneous core with oscillatory zoning
that was partially resorbed and overgrown in the latest Archean, earliest
Paleoproterozoic. The outer rim grew around 2.1 Ga (Modified from
Cottle and Stearns 2017)
412
Earth’s Continental Crust, Fig. 6 U-Pb dates of zircons from detrital
terrigenous sedimentary rocks and granites. (a) Detrital zircons that are
less than 5% discordant, from Voice et al. (2011); (b) detrital zircons
<10% discordant, from Condie et al. (2009); (c) detrital zircons from
global modern river sediments, from Campbell and Allen (2008); (d)
zircons for orogenic granites that are <5% discordant, from Condie et al.
(2009). Vertical gray bars delineate the statistically significant peaks and
proposed times of supercontinent formation (labels at top) (Modified
from Voice et al. 2011)
et al. 2016), they indicate that crustal growth was a far more
continuous process than the U-Pb zircon age peaks suggest.
Earth’s Earliest Crust
The oldest known indigenous minerals found on Earth are
Hadean detrital zircons from Mt. Narryer and the Jack Hills,
Western Australia (Compston and Pidgeon 1986; Froude et al.
1983), which date back to 4.4 Ga (Wilde et al. 2001). Since
their discovery, these tiny grains have been the subject of
intense scientific scrutiny in order to divine the nature of the
Hadean Earth, for which we have no rock record (see review
by Harrison 2009). Key features suggest that these zircons
crystallized from hydrous, relatively low-temperature granitic
melts: encapsulation of mineral inclusions of typical granitic
composition (e.g., quartz, muscovite, K-feldspar, hornblende,
Hopkins et al. 2008; Maas et al. 1992), fractionated oxygen
Earth’s Continental Crust
isotopes, indicating incorporation of materials that experi-
enced low temperature alteration in the presence of liquid
water (Mojzsis et al. 2001; Wilde et al. 2001), and low
Ti-in-zircon temperatures indicative of relatively evolved
melt compositions (Fu et al. 2008; Watson and Harrison
2005). These results have been interpreted to suggest that
the Hadean Earth was temperate, with liquid water, and pos-
sibly also widespread continents created by plate tectonics
(Harrison 2009; Mojzsis et al. 2001; Wilde et al. 2001). Most
recently, isotopically light carbon isotopes measured in
graphite contained within a 4.1 Ga zircon suggests the possi-
bility of a biosphere at this time (Bell et al. 2015), and remnant
magnetism has even been detected in the Hadean zircons,
leading to the suggestion that the Hadean Earth possessed a
magnetic field (Tarduno et al. 2015), though debate exists
regarding the veracity of this signal (cf. Cottrell et al. 2016;
Weiss et al. 2015).
Origins alternative to a wet granite source for the zircons
have also been proposed. For example, the zircons may have
crystallized during differentiation of impact melts, as zircons
from the Sudbury impact melt sheet also record low Ti-in-
zircon temperatures (e.g., Darling et al. 2009; Kenny et al.
2016). Or the Hadean zircons may have crystallized during
the final stages of differentiation of basaltic melts (e.g.,
Rollinson 2008). Various lines of argument have been mus-
tered to counter these arguments. Perhaps the most compel-
ling evidence lies in the presence of muscovite inclusions,
which, after quartz, are the second most abundant inclusion
observed (Hopkins et al. 2008). Muscovite does not crystal-
lize during basalt differentiation and has not been recognized
in impact melts. Moreover, the trace element compositions of
the zircons are distinct from those formed in late-stage differ-
entiates of basalt (Grimes et al. 2007).
Although the weight of evidence supports crystallization
of the Jack Hills zircons from hydrous granitic melts, the
degree to which these rare crystals indicate the presence of
widespread continental crust and the operation of plate tec-
tonics is open to debate. Harrison (2009) made the case that
the inferred low geotherm determined from thermobarometry
of the mineral inclusions (Hopkins et al. 2008) could only
reasonably exist on a hotter Hadean Earth if there were
subduction zones, which are characterized by exceedingly
low temperatures in the modern Earth and, by analogy, the
Hadean Earth. However, others have suggested that it is
unlikely that widespread Hadean continental crust could be
obliterated with nary a trace. For example, the rarity of detrital
zircons having Hadean U-Pb ages is hard to explain if such
widespread felsic crust existed (Belousova et al. 2010; Voice
et al. 2011). Moreover, Lu-Hf model ages for the most ancient
zircons typically fail to exhibit evidence for significant vol-
umes of Hadean continental crust (Kemp et al. 2010, 2015;
Vervoort et al. 1996). In addition, the evidence reviewed
above for a dominantly mafic crust in the Archean does not
Earth’s Continental Crust
support the presence of widespread granitic crust in the
Archean or Hadean. Finally, one of Armstrong’s (1991) stron-
gest arguments for very early differentiation of Earth’s crust is
the comparison with other rocky planets in our solar system,
all of which show evidence of early crust formation (e.g.,
Taylor and McLennan 2009). It is indeed quite likely that such
differentiation occurred on Earth but that the earliest crust was
dominated by basalts and komatiites, consistent with the
higher mantle temperatures at that time, and with the basaltic
compositions of other ancient planetary crusts (Taylor and
McLennan 2009).
In addition to the Hadean zircons, evidence for the nature of
the earliest crust comes from some of the oldest rocks on Earth:
the 4.0 Ga Acasta gneisses in the Great Slave Craton, Canada,
the 3.8-3.9 Ga rocks of southwest Greenland, and unusual,
~3.8 Ga ortho-amphibole-bearing amphibolites and associated
TTGs of the Nuvvuagittuq Greenstone Belt, Canada.
The oldest dated rocks are tonalitic gneisses – the Acasta
gneisses from the Slave Craton, Northwest Territories, Can-
ada, which have U-Pb zircon ages dating back to 4.03 Ga
(Bowring and Williams 1999). Several lines of evidence
suggest that the original tonalitic magmas derived from pre-
existing mafic to ultramafic crust. Whole rock 142Nd isotope
data for the Acasta gneisses that deviate from that of the bulk
Earth have been interpreted to reflect their derivation from a
~4.3 Ga precursor mafic crust (Roth et al. 2014 and references
therein). Likewise, Hf isotope compositions of zircons from a
4.02 Ga Acasta gneiss indicate its derivation from mafic to
ultramafic crust, with no evidence for earlier granitic conti-
nental crust (Reimink et al. 2016).
The Itsaq gneiss complex and Isua supracrustal belt in
southwestern Greenland were recognized early on as
containing some of Earth’s most ancient rocks, dating to at
least 3.7 Ga (Moorbath et al. 1972, 1973). They form the
largest continuous track of early Archean rocks yet recog-
nized. More recent work shows these rocks to have formed as
far back as ~3.9 Ga (Nutman et al. 1996, 1997). The Itsaq
gneisses are highly deformed amphibolite facies rocks com-
prised of ~80% gray gneisses dating to >3.8 Ga, 10% supra-
crustal, mafic, and ultramafic rocks formed at 3.87 to 3.6 Ga
and 10% somewhat younger (3.65 Ga) granites (Nutman
et al. 1997). By contrast, the Isua supracrustal belt is com-
prised of less deformed amphibolite-facies chemical (BIF,
cherts, and carbonates) and clastic sedimentary rocks
interbedded with bimodal (mafic-felsic) volcanic rocks that
form a least two distinct stratigraphic units dating to 3.7 Ga
and 3.8 Ga (Nutman et al. 1997) and contain what may be
the oldest microbial mats (Nutman et al. 2016). The SW
Greenland rocks record both positive (Boyet et al. 2003;
Caro et al. 2003; Harper and Jacobsen 1992) and negative
142
Nd isotopic anomalies (Rizo et al. 2012), indicating their
derivation from rocks that originally differentiated in the
Hadean.
413
The most recently recognized ancient crust occurs in the
Nuvvuagittuq Greenstone Belt, in the Minto Block of the
northern Superior Province, which crops out on the eastern
shore of Hudson Bay in northern Quebec. The lithological
assembly comprises chemical sedimentary rocks (quartz-
magnetite rocks, ferruginous quartz-pyroxene rocks),
quartz-biotite schists containing clasts of quartzite and mafic
rocks (Cates and Mojzsis 2007), ultramafic sills, as well as
unusual cummingtonite-bearing amphibolites (O’Neil et al.
2008). The oldest dated rocks in the assemblage are felsic
schists, interpreted to be volcanic in origin with U-Pb zircon E
dates of 3.77 (Augland and David 2015) to 3.82 Ga (David
et al. 2009). The amphibolites generally have lower
142
Nd/144Nd than bulk Earth and correlate with Sm/Nd ratios
to yield an isochron of 4280 Ma (O’Neil et al. 2008), which
has been interpreted, with some controversy (e.g., Roth et al.
2013), to represent the age of the belt. Irrespective of this
debate, multiple lines of evidence suggest the derivation of
the Nuvvuagittuq igneous rocks from mantle that experienced
melt depletion early in the Hadean (Roth et al. 2013; O’Neil
et al. 2016). Most recently, coupled 142Nd and 143Nd isotopic
studies of much younger granitoids in the region point to their
derivation from extensive Hadean mafic crust (O’Neil and
Carlson 2017).
Summary
Earth is the only planet in our solar system that has a compo-
sitionally highly evolved crust – the continental crust – that
contains a significant fraction of Earth’s most strongly incom-
patible elements, including the heat-producing elements K,
Th, and U. All estimates agree that the present-day continental
crust has an average “andesite” composition and is vertically
stratified. The upper continental crust is evolved (66–70 wt.%
SiO2), and its composition is the best constrained of any
portion of the crust, with average estimates for major element
concentrations falling within 70% of one another and most
trace element concentration estimates falling within a factor
of two. The average compositions of the middle and lower
crust are much less constrained and the subject of debate
(mafic or intermediate?), but surface heat flow observations
require they be depleted in heat-producing elements. An
average andesitic crust composition cannot be generated by
single-stage melting of mantle peridotite. Multiple processes
likely helped to transform original basaltic crust to the crust
present today, including weathering and erosion, melting of
subducted or foundered basaltic rocks, and density founder-
ing and/or “relamination.”
A number of lines of evidence point to a fundamental
change in the bulk continental crust composition in the Late
Archean from dominantly mafic to felsic by 2.5 Ga. Such a
transformation may signal the onset of widespread plate
414
tectonics as a ready means to generate voluminous granitic
magmas, though hydrous, K-rich granites existed in the
Hadean, as inferred from Hadean-aged detrital zircons. The
timing of growth of the continental crust is a topic of major
debate. Early crust formation must have occurred in response
to a cooling Earth, but it is likely that this crust was domi-
nantly basaltic, with minor TTGs, but thick enough to rise
above sea level locally. Growth of volumetrically significant
felsic continental crust likely occurred in the Late Archean.
Prominent peaks in U-Pb zircon dates are observed on all
continents, but the interpretation of these peaks is uncertain:
they may represent periods of enhanced crust formation or
preservation.
Acknowledgments Peter van Keken, Gray Bebout, Ralf Halama, and
Sujoy Mukhopadhyay all provided help in tracking down compositional
estimates for the continental crust or other data. John Cottle kindly
provided the beautiful zircon images in Fig. 5. Mike Free, Allison
Greaney, and Ming Tang provided comments that helped me to clarify
different parts of the original manuscript. I appreciate review comments
from Kent Condie, Bruno Dhuime, Brad Hacker, Chris Hawkesworth,
and Scott McLennan. I am grateful for support of my research from the
National Science Foundation over the past two decades. Finally, I thank
Bill White for being exceedingly patient while the writing of this entry
was delayed due to my move across a continent.
References
Abers GA, Hacker BR (2016) A MATLAB toolbox and Excel workbook
for calculating the densities, seismic wave speeds, and major element
composition of minerals and rocks at pressure and temperature.
Geochem Geophys Geosyst. https://doi.org/10.1002/2015GC006171
Albarède F (1998) The growth of continental crust. Tectonophysics
296:1–14
Allègre CJ, Rousseau D (1984) The growth of the continent through
geological time studied by Nd isotope analysis of shales. Earth Planet
Sci Lett 67:19–34. https://doi.org/10.1016/0012-821X(84)90035-9
Anderson AT (1982) Parental basalts in subduction zones: implications
for continental evolution. J Geophys Res 87:7047–7060
Armstrong RL (1981) Radiogenic isotopes: the case for crustal recycling
on a near-steady-state no-continental-growth Earth. Phil Trnas R Soc
Lond A 301:443–472
Armstrong RL (1991) The persistent myth of crustal growth. Aust
J Earth Sci 38:613–630
Arndt NT (2013) The formation and evolution of the continental crust.
Geochem Perspect 2(3):405–533. https://doi.org/10.7185/
geochempersp.2.3
Arndt N, Davaille A (2013) Episodic Earth evolution. Tectonophysics
609:661–674. https://doi.org/10.1016/j.tecto.2013.07.002
Arndt NT, Goldstein SL (1989) An open boundary between lower
continental crust and mantle: its role in crust formation and crustal
recycling. Tectonophys 161:201–212
Augland LE, David J (2015) Protocrustal evolution of the Nuvvuagittuq
Supracrustal Belt as determined by high precision zircon Lu-Hf and
U-Pb isotope data. Earth Planet Sci Lett 428:162–171. https://doi.
org/10.1016/j.epsl.2015.07.039
Barth M, McDonough WF, Rudnick RL (2000) Tracking the budget of
Nb and Ta in the continental crust. Chem Geol 165:197–213
Behn MD, Kelemen PB (2003) The relationship between seismic
P-wave velocity and the composition of anhydrous igneous and
Earth’s Continental Crust
meta-igneous rocks. Geochem Geophys Geosys 4:1041. https://doi.
org/10.1029/2002GC000393
Bell EA, Boehnke P, Harrison TM, Mao WL (2015) Potentially
biogenic carbon preserved in a 4.1 billion-year-old zircon. Proc
Natl Acad Sci U S A 112(47):14518–14521. https://doi.org/
10.1073/pnas.1517557112
Belousova EA, Kostitsyn YA, Griffin WL, Begg GC, O’Reilly SY,
Pearson NJ (2010) The growth of the continental crust: constraints
from zircon Hfisotope data. Lithos 119:457–466. https://doi.org/
10.1016/j.lithos.2010.07.024
Bowring SA, Williams IS (1999) Priscoan (4.00-4.03 Ga) orthogneisses
from northwestern Canada. Contrib Mineral Petrol 134(1):3–16
Boyet M, Blichert-Toft J, Rosing MT, Storey M, Telouk P, Albarede F
(2003) 142Nd evidence for early Earth diferentiation. Earth Planet
Sci Lett 214:427–442. https://doi.org/10.1016/S0012-821X(03)
00423-0
Campbell IH, Allen CM (2008) Formation of supercontinents linked to
increases in atmospheric oxygen. Nat Geosci 1(8):554–558. https://
doi.org/10.1038/ngeo259
Campbell IH, Taylor SR (1985) No water, no granites – no oceans, no
continents. Geophys Res Lett 10:1061–1064
Canil D, Crockford PW, Rossin R, Telmer K (2015) Mercury in some arc
crustal rocks and mantle peridotites and relevance to the moderately
volatile element budget of the Earth. Chem Geol 396:134–142.
https://doi.org/10.1016/j.chemgeo.2014.12.029
Caro G, Bourdon B, Birck JL, Moorbath S (2003) Sm-146-Nd-142
evidence from Isua metamorphosed sediments for early differentia-
tion of the Earth’s mantle. Nature 423(6938):428–432. https://doi.
org/10.1038/nature01668
Cates NL, Mojzsis SJ (2007) Pre-3750 Ma supracrustal rocks from the
Nuvvuagittuq supracrustal belt, Northern Quebec. Earth Planet Sci
Lett 255(1–2):9–21. https://doi.org/10.1016/j.epsl.2006.11.034
Cawood PA, Hawkesworth CJ, Dhuime B (2013) The continental record
and the generation of continental crust. Geol Soc Am Bull
125(1–2):14–32. https://doi.org/10.1130/b30722.1
Chauvel C, Garcon M, Bureau S, Besnault A, Jahn BM, Ding ZL
(2014) Constraints from loess on the Hf-Nd isotopic composition
of the upper continental crust. Earth Planet Sci Lett 388:48–58.
https://doi.org/10.1016/j.epsl.2013.11.045
Chen K, Walker RJ, Rudnick RL, Gao S, Gaschnig RM, Puchtel IS,
Tang M, Hu Z-C (2016) Platinum-group element abundances and
Re-Os isotopic systematics of the upper continental crust through
time: evidence from glacial diamictites. Geochim Cosmochim Acta
191:1–16. https://doi.org/10.1016/j.gca.2016.07.004
Christensen NI, Mooney WD (1995) Seismic velocity structure and
composition of the continental crust: a global view. J Geophys Res
100(B7):9761–9788
Clift PD, Vannucchi P, Morgan JP (2009) Crustal redistribution, crust-
mantle recycling and Phanerozoic evolution of the continental crust.
Earth-Sci Rev 97(1–4):80–104. https://doi.org/10.1016/j.
earscirev.2009.10.003
Cogley JG (1984) Continental margins and the extent and number of the
continents. Rev Geophys Space Phys 22:101–122
Compston W, Pidgeon RT (1986) Jack Hills, evidence of more very old
detrital zircons in western. Aust Nat 321(6072):766–769. https://doi.
org/10.1038/321766a0
Condie KC (1993) Chemical composition and evolution of the upper
continental crust: contrasting results form surface samples and
shales. Chem Geol 104:1–37
Condie KC (1998) Episodic continental growth and supercontinents: a
mantle avalanche connection? Earth Planet Sci Lett
163(1–4):97–108. https://doi.org/10.1016/s0012-821x(98)00178-2
Condie KC, Aster RC (2010) Episodic zircon age spectra of orogenic
granitoids: the supercontinent connection and continental growth.
Precambrian Res 180(3–4):227–236. https://doi.org/10.1016/j.
precamres.2010.03.008
Earth’s Continental Crust
Condie KC, Belousova E, Griffin WL, Sircombe KN (2009) Granitoid
events in space and time: constraints from igneous and detrital zircon
age spectra. Gondwana Res 15(3–4):228–242. https://doi.org/
10.1016/j.gr.2008.06.001
Cottle JM, Stearns MA (2017) Application of single shot laser ablation
split stream to accessory phase petrochronology. In: Macrostructural
geochronology; lattice to atom-scale records of planetary evolution.
American Geophysical Union Monograph, Washington, DC
Cottrell RD, Tarduno JA, Bono RK, Dare MS, Mitra G (2016) The
inverse microconglomerate test: further evidence for the preservation
of Hadean magnetizations in metasediments of the Jack Hills, West-
ern Australia. Geophys Res Lett 43(9):4215–4220. https://doi.org/
10.1002/2016gl068150
Darling J, Storey C, Hawkesworth C (2009) Impact melt sheet zircons
and their implications for the Hadean crust. Geology
37(10):927–930. https://doi.org/10.1130/g30251a.1
David J, Godin L, Stevenson R, O’Neil J, Francis D (2009) U-Pb ages
(3.8-2.7 Ga) and Nd isotope data from the newly identified Eoarchean
Nuvvuagittuq supracrustal belt, superior Craton, Canada. Geol Soc
Am Bull 121(1–2):150–163. https://doi.org/10.1130/b26369.1
Dhuime B, Hawkesworth CJ, Cawood PA, Storey CD (2012) A change
in the geodynamics of continental growth 3 billion years ago. Science
335(6074):1334–1336. https://doi.org/10.1126/science.1216066
Dhuime B, Wuestefeld A, Hawkesworth CJ (2015) Emergence of mod-
ern continental crust about 3 billion years ago. Nat Geosci
8(7):552–555. https://doi.org/10.1038/ngeo2466
Ducea MN (2002) Constraints on the bulk composition and root foun-
dering rates of continental arcs: a California arc perspective. J
Geophys Res-Solid Earth 107(B11)
Eade KE, Fahrig WF (1971) Chemical evolutionary trends of continental
plates – a preliminary study of the Canadian Sheild. Canadian Geo-
logical Survey, Ottawa
Eade KE, Fahrig WF (1973) Regional, lithological, and temporal varia-
tion in the abundances of some trace elements in the Canadian Shield.
Canadian Geological Survey, Ottawa
Eichelberger JC (1975) Origin of andesite and dacite – evidence of
mixing of magmas at Glass Mountain in California and at other
circum-Pacific volcanoes. Geol Soc Am Bull 86(10):1381–1391.
https://doi.org/10.1130/0016-7606(1975)86<1381:ooaade>2.0.co;2
Elliott T (2003) Tracers of the slab. In: Eiler J (ed) Inside the subduction
factory, Geophysical monograph, vol 138. American Geophysical
Union, Washington, DC, pp 23–45
Fahrig WF, Eade KE (1968) The chemical evolution of the Canadian
Shield. Geochim Cosmochim Acta 5:1247–1252
Fegley B (2014) Venus. In: Davis AM (ed) Planets, asteroids, comets and
the solar system, Treatise on geochemistry, vol 2. Elsevier, Oxford,
pp 127–148
Fisher CM, Vervoort JD, DuFrane SA (2014) Accurate Hf isotope
determinations of complex zircons using the “laser ablation split
stream” method. Geochem Geophys Geosyst 15(1):121–139.
https://doi.org/10.1002/2013gc004962
Froude DO, Ireland TR, Kinny PD, Williams IS, Compston W, Williams
IR, Myers JS (1983) Ion micropre identification of 4,11-4,200
Myr-old terrestiral zircons. Nature 304(5927):616–618. https://doi.
org/10.1038/304616a0
Fu B, Page FZ, Cavosie AJ, Fournelle J, Kita NT, Lackey JS, Wilde SA,
Valley JW (2008) Ti-in-zircon thermometry: applications and limita-
tions. Contrib Mineral Petrol 156(2):197–215. https://doi.org/
10.1007/s00410-008-0281-5
Gao S, Luo T-C, Zhang B-R, Zhang H-F, Han Y-W, Hu Y-K, Zhao Z-D
(1998) Chemical composition of the continental crust as revealed by
studies in East China. Geochim Cosmochim Acta 62:1959–1975
Gaschnig RM, Rudnick RL, McDonough WF, Kaufman AJ, Valley JW,
Hu Z-C, Gao S, Beck ML (2016) Compositional evolution of the
upper continental crust through time, as constrained by ancient
glacial diamictites. Geochim Cosmochim Acta 186:316–343
415
Goldschmidt VM (1933) Grundlagen der quantitativen Geochemie.
Fortschr Mienral Kirst Petrog 17:112
Goldstein SL (1989) Decoupled evolution of Nd and Sr isotopes in the
continental crust and the mantle. Nature 336:733–738
Goodwin AM (1996) Principles of Precambrian geology. Academic,
London
Grimes CB, John BE, Kelemen PB, Mazdab FK, Wooden JL, Cheadle
MJ, Hanghoj K, Schwartz JJ (2007) Trace element chemistry of
zircons from oceanic crust: a method for distinguishing detrital
zircon provenance. Geology 35(7):643–646. https://doi.org/
10.1130/g23603a.1
Gurnis M, Davies GF (1986) Apparent episodic crustal growth arising
from a smoothly evolving mantle. Geology 14:396–399
Hacker BR, Kelemen PB, Behn MD (2011) Differentiation of the con- E
tinental crust by relamination. Earth Planet Sci Lett
307(3–4):501–516. https://doi.org/10.1016/j.epsl.2011.05.024
Hacker BR, Kelemen PB, Behn MD (2015) Continental lower crust. In:
Jeanloz R, Freeman KH (eds) Annu Rev Earth Planet Sci 43:167–205
Harper CL, Jacobsen SB (1992) Evidence from coupled 147Sm-143Nd
and 146Sm-142Nd systematics for very early (4.5 Gyr) differentiation
of the Earth’s mantle. Nature 360:728–732
Harrison TM (2009) The Hadean crust: evidence from >4 Ga zircons.
Annu Rev Earth Planet Sci 37:479–505. https://doi.org/10.1146/
annurev.earth.031208.100151
Haskin MA, Haskin LA (1966) Rare earths in European shales: a
redetermination. Science 154:507–509
Hawkesworth CJ, Kemp AIS (2006) Evolution of the continental crust.
Nature 443(7113):811–817. https://doi.org/10.1038/nature05191
Hawkesworth CJ, Dhuime B, Pietranik AB, Cawood PA, Kemp AIS,
Storey CD (2010) The generation and evolution of the continental
crust. J Geol Soc 167(2):229–248. https://doi.org/10.1144/0016-
76492009-072
Hawkesworth C, Cawood P, Dhuime B (2013) Continental growth and
the crustal record. Tectonophysics 609:651–660. https://doi.org/
10.1016/j.tecto.2013.08.013
Herzberg CT, Fyfe WS, Carr MJ (1983) Density constraints on the
formation of the continental Moho and crust. Contrib Mineral Petrol
84:1–5
Hopkins M, Harrison TM, Manning CE (2008) Low heat flow inferred
from >4 Gyr zircons suggests Hadean plate boundary interactions.
Nature 456(7221):493–496. https://doi.org/10.1038/nature07465
Hu ZC, Gao S (2008) Upper crustal abundances of trace elements:
revision and update. Chemi Geol 253(3–4):205–221. https://doi.
org/10.1016/j.chemgeo.2008.05.010
Huang Y, Chubakov V, Mantovani F, Rudnick RL, McDonough WF
(2013) A reference Earth model for the heat-producing elements and
associated geoneutrino flux. Geochem Geophys Geosyst
14(6):2003–2029. https://doi.org/10.1002/ggge.20129
Hurley PM, Rand JR (1969) Pre-drift continental nuclei. Science
164:1229–1242
Jagoutz O, Schmidt MW (2013) The composition of the foundered
complement to the continental crust and a re-evaluation of fluxes in
arcs. Earth Planet Sci Lett 371:177–190. https://doi.org/10.1016/j.
epsl.2013.03.051
Jaupart C, Mareschal J-C (2014) Constraints on crustal heat production
from heat flow data. In: Rudnick RL (ed) The crust, Treatise on
geochemistry, vol 4. Elsevier, Oxford, pp 53–72
Johnson B, Goldblatt C (2015) The nitrogen budget of Earth. Earth-Sci
Rev 148:150–173. https://doi.org/10.1016/j.earscirev.2015.05.006
Jull M, Kelemen PB (2001) On the conditions for lower crustal convec-
tive instability. J Geophys Res 106(4):6423–6446
Kamber BS (2010) Archean mafic-ultramafic volcanic landmasses and
their effect on ocean atmosphere chemistry. Chem Geol
274(1–2):19–28. https://doi.org/10.1016/j.chemgeo.2010.03.009
Kamber BS, Greig A, Collerson KD (2005) A new estimate for the
composition of weathered young upper continental crust from
416
alluvial sediments, Queensland, Australia. Geochim Cosmochim
Acta 69(4):1041–1058. https://doi.org/10.1016/j.gca.2004.08.020
Kay RW, Kay SM (1991) Creation and destruction of lower continental
crust. Geol Rundsch 80:259–278
Kelemen PB (1995) Genesis of high Mg# andesites and the continental
crust. Contrib Mineral Petrol 120:1–19
Kelemen PB, Behn MD (2016) Formation of lower continental crust by
relamination of buoyant arc lavas and plutons. Nat Geosci
9:197–205. https://doi.org/10.1038/NGEO2662
Kelemen PB et al (2003a) Along-strike variation in lavas of the Aleutian
island arc: implications for the genesis of high Mg# andesite and the
continental crust. In: Eiler J (ed) Inside the subduction factory, AGU
monograph, vol 138. Washington, DC, pp 293–311
Kelemen PB, Hanghoj K, Greene A (2003b) One view of the
geochemistry of subduction-related magmatic arcs, with an
emphasis on primitive andesite and lower crust. In: Rudnick
RL (ed) The crust, Treatise on geochemistry, vol 3. Elsevier,
Amsterdam, pp 593–659
Keller CB, Schoene B (2012) Statistical geochemistry reveals disruption
in secular lithospheric evolution about 2.5 Gyr ago. Nature
485(7399):490–493. https://doi.org/10.1038/nature11024
Keller CB, Schoene B, Barboni M, Samperton KM, Husson JM
(2015) Volcanic-plutonic parity and the differentiation of the conti-
nental crust. Nature 523(7560):301–307. https://doi.org/10.1038/
nature14584
Kemp AIS, Hawkesworth CJ (2014) Growth and differentiation of
the continental crust from isotope studies of accessory minerals.
In: Rudnick RL (ed) The crust, vol 4. Elsevier, Oxford,
pp 379–421
Kemp AIS, Hawkesworth CJ, Paterson BA, Kinny PD (2006) Episodic
growth of the Gondwana supercontinent from hafnium and oxygen
isotopes in zircon. Nature 439(7076):580–583
Kemp AIS, Wilde SA, Hawkesworth CJ, Coath CD, Nemchin A,
Pidgeon RT, Vervoort JD, DuFrane SA (2010) Hadean crustal evo-
lution revisited: new constraints from Pb-Hf isotope systematics of
the Jack Hills zircons. Earth Planet Sci Lett 296(1–2):45–56. https://
doi.org/10.1016/j.epsl.2010.04.043
Kemp AIS, Hickman AH, Kirkland CL, Vervoort JD (2015) Hf isotopes
in detrital and inherited zircons of the Pilbara Craton provide no
evidence for Hadean continents. Precambrian Res 261:112–126.
https://doi.org/10.1016/j.precamres.2015.02.011
Kenny GG, Whitehouse MJ, Kamber BS (2016) Differentiated impact
melt sheets may be a potential source of Hadean detrital zircon.
Geology 44(6):435–438. https://doi.org/10.1130/g37898.1
Kent AJR, Darr C, Koleszar AM, Salisbury MJ, Cooper KM (2010) Pref-
erential eruption of andesitic magmas through recharge filtering. Nat
Geosci 3(9):631–636. https://doi.org/10.1038/ngeo924
König S, Munker C, Hohl S, Paulick H, Barth AR, Lagos M, Pfander J,
Buchl A (2011) The Earth's tungsten budget during mantle melting
and crust formation. Geochim Cosmochim Acta 75(8):2119–2136.
https://doi.org/10.1016/j.gca.2011.01.031
Kylander-Clark ARC, Hacker BR, Cottle JM (2013) Laser-ablation split-
stream ICP petrochronology. Chem Geol 345:99–112. https://doi.
org/10.1016/j.chemgeo.2013.02.019
Lee CTA (2014) Physics and chemistry of deep continental crust recycling.
In: Rudnick RL (ed) The crust, vol 4. Elsevier, Oxford, pp 423–456
Lee C-TA, Bachmann O (2014) How important is the role of crystal
fractionation in making intermediate magmas? Insights from Zr and
P systematics. Earth Planet Sci Lett 393:266–274. https://doi.org/
10.1016/j.epsl.2014.02.044
Lee CTA, Morton DM, Kistler RW, Baird AK (2007) Petrology and
tectonics of Phanerozoic continent formation: From island arcs to
accretion and continental arc magmatism. Earth Planet Sci Lett 263
(3–4):370–387
Lee CTA, Morton DM, Little MG, Kistler R, Horodyskyj UN, Leeman
WP, Agranier A (2008) Regulating continent growth and
Earth’s Continental Crust
composition by chemical weathering. Proc Natl Acad Sci U S A
105(13):4981–4986
Lee CTA, Yeung LY, McKenzie NR, Yokoyama Y, Ozaki K, Lenardic
A (2016) Two-step rise of atmospheric oxygen linked to the growth
of continents. Nat Geosci 9(6):417–424. https://doi.org/10.1038/
ngeo2707
Li S, Gaschnig RM, Rudnick RL (2016) Insights into chemical
weathering of the upper continental crust from the geochemistry of
ancient glacial diamictites. Geochim Cosmochim Acta 176:96–117.
https://doi.org/10.1016/j.gea.2015.12.012
Liu X-M, Rudnick RL (2011) Constraints on continental crustal mass
loss via chemical weathering using lithium and its isotopes. Proc Natl
Acad Sci U S A 108(52):20873–20880. https://doi.org/10.1073/
pnas.1115671108
Maas R, Kinny PD, Williams IS, Froude DO, Compston W (1992) The
Earth’s oldest known crust – a geochronological and geochemical
study of the 3900-4200 Ma old detrital zircons from Mt. Narryer
and Jack Hills, Western Australia. Geochimica Et Cosmochimica
Acta 56(3):1281–1300. https://doi.org/10.1016/0016-7037(92)
90062-n
Martin H (1986) Effect of steeper Archean geothermal gradient on
geochemistry of subduction-zone magmas. Geology 14:753–756
Martin H, Moyen JF (2002) Secular changes in tonalite-trondhjemite-
granodiorite composition as markers of the progressive cooling of
Earth. Geology 30(4):319–322
McCoy TJ, Nittler LR (2014) Mercury. In: Davis AM (ed) Meteorites,
comets and planets, Treatise on geochemistry, vol 2. Elsevier,
Oxford, pp 119–126
McCulloch MT, Bennett VC (1994) Progressive growth of the Earth’s
continental crust and depleted mantle: geochemical constraints.
Geochim Cosmochim Acta 58:4717–4738
McDonough WF, Sun S-S (1995) Composition of the earth. Chem Geol
120:223–253
McLennan SM (2001) Relationships between the trace element compo-
sition of sedimentary rocks and upper continental crust. Geochem
Geophys Geosyst 2:art. no.-2000GC000109
McLennan SM, Taylor SR (1980) Th and U in sedimentary rocks: crustal
evolution and sedimentary recycling. Nature 285:621–624
McLennan SM, Taylor SR (1991) Sedimentary-rocks and crustal
evolution – tectonic setting and secular trends. J Geol 99(1):1–21
McLennan SM, Taylor SR (1996) Heat flow and the chemical composi-
tion of continental crust. J Geol 104:396–377
McSween HY Jr, McLennan SM (2014) Mars. In: Davis AM
(ed) Meteorites, comets and planets, Treatise on geochemistry,
vol 2. Elsevier, Oxford, pp 251–300
Mojzsis SJ, Harrison TM, Pidgeon RT (2001) Oxygen-isotope evidence
from ancient zircons for liquid water at the Earth’s surface 4,300 Myr
ago. Nature 409(6817):178–181. https://doi.org/10.1038/35051557
Moorbath S, Gale NH, Pankhurst RJ, McGregor VR, Onions RK
(1972) Further rubidium-strontium age determinations on very
early Precambrian rocks of Godthaab district, West Greenland. Nat
Phys Sci 240(100):78–82
Moorbath S, Onions RK, Pankhurst RJ (1973) Early Archaean age for
Isua iron formation, West Greenland. Nature 245(5421):138–139.
https://doi.org/10.1038/245138a0
Muramatsu Y, Wedepohl KH (1998) The distribution of iodine in the
earth’s crust. Chem Geol 147(3–4):201–216. https://doi.org/10.1016/
s0009-2541(98)00013-8
Nance WB, Taylor SR (1976) Rare-Earth element patterns and crustal
evolution. 1. Australian post-Archean sedimentary-rocks. Geochim
Cosmochim Acta 40(12):1539–1551. https://doi.org/10.1016/0016-
7037(76)90093-4
Nance WB, Taylor SR (1977) Rare-Earth element patterns and crustal
evolution. 2. Archean sedimentary-rocks from Kalgoorlie, Australia.
Geochim Cosmochim Acta 41(2):225–231. https://doi.org/10.1016/
0016-7037(77)90229-0
Earth’s Continental Crust
Nutman AP, McGregor VR, Friend CRL, Bennett VC, Kinny PD
(1996) The Itsaq Gneiss Complex of southern west Greenland; the
world’s most extensive record of early crustal evolution (3900-3600
Ma). Precambrian Res 78(1–3):1–39. https://doi.org/10.1016/0301-
9268(95)00066-6
Nutman AP, Bennett VC, Friend CRL, Rosing MT (1997) ~ 3710 and
>3790 Ma volcanic sequences in the Isua (Greenland) supracrustal
belt; structural and Nd isotope implications. Chem Geol
141(3–4):271–287. https://doi.org/10.1016/s0009-2541(97)00084-3
Nutman AP, Bennett VC, Van Kranendonk MJ, Chivas AR (2016) Rapid
emergence of life shown by discovery of 3,700-million-year-old
microbial structures. Nature. https://doi.org/10.1038/nature19355
O’Neil J, Carlson RW (2017) Building Archean cratons from Hadean
mafic crust. Science 355:1199–1202
O’Neil J, Carlson RW, Francis D, Stevenson RK (2008) Neodymium-
142 evidence for hadean mafic crust. Science 321(5897):1828–1831.
https://doi.org/10.1126/science.1161925
O’Neil J, Rizo H, Boyet M, Carlson RW, Rosing MT (2016) Geochem-
istry and Nd isotopic characteristics of Earth’s Hadean mantle and
primitive crust. Earth Planet Sci Lett 442:194–205. https://doi.org/
10.1016/j.epsl.2016.02.055
Park JW, Hu ZC, Gao S, Campbell IH, Gong HJ (2012) Platinum group
element abundances in the upper continental crust revisited – new
constraints from analyses of Chinese loess. Geochim Cosmochim
Acta 93:63–76. https://doi.org/10.1016/j.gca.2012.06.026
Parman SW (2015) Time-lapse zirconography: imaging punctuated con-
tinental evolution. Geochem Perspect Lett 1:43–52. https://doi.org/
10.7185/geochemlet.1505
Pasyanos ME, Masters TG, Laske G, Ma ZT (2014) LITHO1.0: an
updated crust and lithospheric model of the earth. J Geophys
ResSolid Earth 119(3):2153–2173. https://doi.org/10.1002/
2013jb010626
Payne JL, McInerney DJ, Barovich KM, Kirkland CL, Pearson NJ, Hand
M (2016) Strengths and limitations of zircon Lu-Hf and O isotopes in
modelling crustal growth. Lithos 248:175–192. https://doi.org/
10.1016/j.lithos.2015.12.015
Pearson DG, Parman SW, Nowell GM (2007) A link between large
mantle melting events and continent growth seen in osmium isotopes.
Nature 449(7159):202–205. https://doi.org/10.1038/nature06122
Peucker-Ehrenbrink B, Jahn B-M (2001) Rhenium-osmium isotope
systematics and platinum group element concentations: loess and
the upper continental crust. Geochem Geophys Geosyst
2:2001GC000172
Plank T (2005) Constraints from Th/La on sediment recycling at sub-
duction zones and the evolution of the continents. J Petrol
46:921–944
Reimink JR, Davies JHFL, Chacko T, Stern RA, Heaman LM, Sarkar C,
Schaltegger U, Creaser RA, Pearson DG (2016) No evicence for
Hadean continental crust within Earth’s oldest evovled rock unit.
Nat Geosci. https://doi.org/10.1038/ngeo2786
Reubi O, Blundy J (2009) A dearth of intermediate melts at subduction
zone volcanoes and the petrogenesis of arc andesites. Nature
461(7268):1269–1273. https://doi.org/10.1038/nature08510
Rizo H, Boyet M, Blichert-Toft J, O’Neil J, Rosing MT, Paquette JL
(2012) The elusive Hadean enriched reservoir revealed by Nd-142
deficits in Isua Archaean rocks. Nature 491(7422):96–U109. https://
doi.org/10.1038/nature11565
Rollinson H (2008) Secular evolution of the continental crust: implica-
tions for crust evolution models. Geochem Geophys Geosyst
9. https://doi.org/10.1029/2008gc002262
Roth ASG, Bourdon B, Mojzsis SJ, Touboul M, Sprung P, Guitreau M,
Blichert-Toft J (2013) Inherited Nd-142 anomalies in Eoarchean
protoliths. Earth Planet Sci Lett 361:50–57. https://doi.org/10.1016/
j.epsl.2012.11.023
Roth ASG, Bourdon B, Mojzsis SJ, Rudge JF, Guitreau M, Blichert-Toft
J (2014) Combined Sm-147,Sm-146-Nd-143,Nd-142 constraints on
417
the longevity and residence time of early terrestrial crust. Geochem
Geophys Geosyst 15(6):2329–2345. https://doi.org/10.1002/
2014gc005313
Rudnick RL (1995) Making continental crust. Nature 378:571–578
Rudnick RL, Fountain DM (1995) Nature and composition of the con-
tinental crust: a lower crustal perspective. Rev Geophys 33(3):
267–309
Rudnick RL, Gao S (2003) The composition of the continental
crust. In: Rudnick RL (ed) The crust, vol 3. Elsevier, Amster-
dam, pp 1–64
Sauzéat L, Rudnick RL, Chauvel C, Garçon M, Tang M (2015) New
perspectives on the Li isotopic composition of the upper continental
crust and its weathering signature. Earth Planet Sci Lett
428:181–192. https://doi.org/10.1016/j.epsl.2015.07.032 E
Scherer E, Munker C, Mezger K (2001) Calibration of the lutetium-
hafnium clock. Science 293(5530):683–687. https://doi.org/10.1126/
science.1061372
Scherer EE, Whitehouse MJ, Munker C (2007) Zircon as a monitor of
crustal growth. Elements 3(1):19–24. https://doi.org/10.2113/
gselements.3.1.19
Schoene B (2014) U-Pb geochronology. In: Rudnick RL (ed) The crust,
vol 4. Elsevier, Oxford, pp 341–378
Shaw DM, Reilly GA, Muysson JR, Pattenden GE, Campbell FE
(1967) An estimate of the chemical composition of the Canadian
Precambrian shield. Can J Earth Sci 4:829–853
Shaw DM, Dostal J, Keays RR (1976) Additional estimates of continen-
tal surface Precambrian shield composition in Canada. Geochim
Cosmochim Acta 40:73–83
Shaw DM, Cramer JJ, Higgins MD, Truscott MG (1986) Composition of
the Canadian Precambrian shield and the continental crust of the
earth. In: Dawson JB, Carswell DA, Hall J, Wedepohl KH (eds)
The nature of the lower continental crust, vol 24. Geol Soc London,
London, pp 257–282
Shaw DM, Dickin AP, Li H, McNutt RH, Schwarcz HP, Truscott MG
(1994) Crustal geochemistry in the Wawa-Foleyet region, Ontario.
Can J Earth Sci 31(7):1104–1121
Simon L, Lecuyer C (2005) Continental recycling: the oxygen isotope
point of view. Geochem Geophys Geosyst 6:Q08004. https://doi.org/
10.1029/2005GC000958
Söderlund U, Patchett JP, Vervoort JD, Isachsen CE (2004) The Lu-176
decay constant determined by Lu-Hf and U-Pb isotope systematics of
Precambrian mafic intrusions. Earth Planet Sci Lett
219(3–4):311–324. https://doi.org/10.1016/s0012-821x(04)00012-3
Stein M, Hofmann AW (1994) Mantle plumes and episodic crustal
growth. Nature 372:63–68
Tang M, Rudnick RL, McDonough WF, Gaschnig RM, Huang
Y (2015) Europium anomalies constrain the mass of recycled
lower continental crust. Geology 43(8):703–706. https://doi.org/
10.1130/g36641.1
Tang M, Chen K, Rudnick RL (2016) Archean upper crust transition
from mafic to felsic marks the onset of plate tectonics. Science
351(6271):372–375. https://doi.org/10.1126/science.aad5513
Tarduno JA, Cottrell RD, Davis WJ, Nimmo F, Bono RK
(2015) A Hadean to Paleoarchean geodynamo recorded by single
zircon crystals. Science 349(6247):521–524. https://doi.org/10.1126/
science.aaa9114
Taylor SR (1964) Abundance of chemical elements in the continental
crust – a new table. Geochim Cosmochim Acta 28:1273–1285
Taylor SR (1977) Island arc models and the composition of the conti-
nental crust. In: Talwani M, Pitmann WC III (eds) Island arcs, deep
sea trenches and back-arc basins, vol 1. American Geophysical
Union, Washington, DC, pp 325–336
Taylor SR, McLennan S (1985) The continenal crust: its composition
and evolution. Blackwell, Oxford
Taylor SR, McLennan SM (1995) The geochemical evolution of the
continental crust. Rev Geophys 33:241–265
418
Taylor SR, McLennan SM (2009) Planetary crusts their composition,
origin and evolution. Cambridge University Press, Cambridge
Taylor SR, McLennan SM, McCulloch MT (1983) Geochemistry of
loess, continental crustal composition and crustal model ages.
Geochim Cosmochim Acta 47:1897–1905
Teng FZ, McDonough WF, Rudnick RL, Dalpe C, Tomascak PB,
Chappell BW, Gao S (2004) Lithium isotopic composition and con-
centration of the upper continental crust. Geochim Cosmochim Acta
68(20):4167–4178
Teng FZ, Rudnick RL, McDonough WF, Gao S, Tomascak PB, Liu YS
(2008) Lithium isotopic composition and concentration of the deep
continental crust. Chem Geol 255(1–2):47–59. https://doi.org/
10.1016/j.chemgeo.2008.06.009
Togashi S, Imai N, Okuyama-Kusunose Y, Tanaka T, Okai T, Koma T,
Murata Y (2000) Young upper crustal chemical composition of the
orogenic Japan arc. Geochem Geophys Geosyst 1. https://doi.org/
10.1029/2000GC000083
van Keken PE, Hacker BR, Syracuse EM, Abers GA (2011) Subduction
factory: 4. Depth-dependent flux of H2O from subducting slabs
worldwide. J Geophys ResSolid Earth 116. https://doi.org/10.1029/
2010jb007922
Veizer J, Compston W (1976) SR-87-SR-86 in Precambrian carbonates
as an index of crustal evolution. Geochim Cosmochim Acta
40(8):905–914. https://doi.org/10.1016/0016-7037(76)90139-3
Vervoort JD, Kemp AIS (2016) Clarifying the zircon Hf isotope record
of crust-mantle evolution. Chem Geol 425:65–75. https://doi.org/
10.1016/j.chemgeo.2016.01.023
Vervoort JD, Patchett PJ, Gehrels GE, Nutman AP (1996) Constraints on
early Earth differentiation from hafnium and neodymium isotopes.
Nature 379(6566):624–627. https://doi.org/10.1038/379624a0
Voice PJ, Kowalewski M, Eriksson KA (2011) Quantifying the timing
and rate of crustal evolution: global compilation of radiometrically
dated detrital zircon grains. J Geol 119(2):109–126. https://doi.org/
10.1086/658295
Watson EB, Harrison TM (2005) Zircon thermometer reveals minimum
melting conditions on earliest Earth. Science 308(5723):841–844.
https://doi.org/10.1126/science.1110873
Weaver BL, Tarney J (1984) Empirical approach to estimating the
composition of the continental crust. Nature 310:575–577
Wedepohl H (1995) The composition of the continental crust. Geochim
Cosmochim Acta 59:1217–1239
Weiss BP, Maloof AC, Tailby N, Ramezani J, Fu RR, Hanus V, Trail D,
Watson EB, Harrison TM, Bowring SA, Kirschvink JL, Swanson-
Hysell NL, Coe RS (2015) Pervasive remagnetization of detrital
zircon host rocks in the Jack Hills, Western Australia and implica-
tions for records of the early geodynamo. Earth Planet Sci Lett
430:115–128. https://doi.org/10.1016/j.epsl.2015.07.067
White WM, Klein EM (2014) Composition of the oceanic crust. In:
Holland HD, Turekian KK (eds) Treatise on geochemistry, vol 4,
2nd edn. Elsevier, Oxford, pp 457–495
Wilde SA, Valley JW, Peck WH, Graham CM (2001) Evidence from
detrital zircons for the existence of continental crust and oceans on
the earth 4.4 Gyr ago. Nature 409(6817):175–178
Willbold M, Hegner E, Stracke A, Rocholl A (2009) Continental geo-
chemical signatures in dacites from Iceland and implications for
models of early Archaean crust formation. Earth Planet Sci Lett
279(1–2):44–52. https://doi.org/10.1016/j.epsl.2008.12.029
Yanagi T, Yamashita K (1994) Genesis of continental-crust under island-
arc conditions. Lithos 33(1–3):209–223. https://doi.org/10.1016/
0024-4937(94)90061-2
Yuan HL, Gao S, Dai MN, Zong CL, Gunther D, Fontaine GH,
Liu XM, Diwu C (2008) Simultaneous determinations of
U-Pb age, Hf isotopes and trace element compositions of zircon
by excimer laser-ablation quadrupole and multiple-collector
ICP-MS. Chem Geol 247(1–2):100–118. https://doi.org/10.1016/j.
chemgeo.2007.10.003
Earth’s Core
Earth’s Core
William F. McDonough
Department of Geology, University of Maryland, College
Park, MD, USA
Definition
The composition of the Earth’s core is essentially 85% Fe, 5%
Ni, and ~10% other, by weight. The details are more compli-
cated, with the “other” being a light element component,
which results in the core having on average a lower mean
atomic number than that of iron. Birch (1952, 1964) recog-
nized a 10% difference in the density of the outer core relative
to an iron-nickel alloy at core pressure and temperature con-
ditions. Anderson and Isaak (2002) recommend that the core
density deficit is between 3% and 7%. Both the liquid outer
core (~95% by mass) and solid inner core (~5%) contain a
fraction of a light element (nominally in a ~2/1 proportion,
respectively).
Properties
For more than a century the Earth has been described
as having a metallic core surrounded by a silicate shell.
A brief history of this topic is given in McDonough
(2014). Emil Wiechert in 1897 first subdivided the Earth’s
interior into these two main components and later, his
graduate student Beno Gutenberg, using seismological
data, established the core-mantle boundary (CMB) at
2900 km depth, which is within uncertainty the accepted
depth for the boundary. First-order physical aspects of the
core are listed in Table 1 and are based on fundamental
constraints derived from seismology, geodetics of the
planets, mineral physics, and experimental studies, with
limited additional constraints from geochemistry and cos-
mochemistry. Seismology reveals that the CMB is more
than half an Earth’s radius. The outer core (OC) is liquid
(absence of a shear-wave) and the inner core (IC) is solid
(given observation of PKJKP waves (i.e., P-waves in
mantle and OC that converts to an S-wave in IC then
back to P-wave as it leaves the IC). Geodesy documents
the coefficient of the planet’s moment of inertia (i.e., a
dense core surrounds a less dense silicate shell) and
libration confirms the liquid state of the outer core. Ther-
modynamic analysis and experimental studies document
the conditions of melting and crystallization of Fe-Ni
alloys at core conditions and set the upper temperature
limits at the inner-outer core boundary. Meteorites con-
firm that Fe and Ni are chemical partners and their ratios
Earth’s Core 419

Earth’s Core, Table 1 Physical properties of the Earth’s core

Mass Value () Units Ref (% Planet)
Earth 5. 97218(60)
1024 kg Chambat et al. (2010) 100%
Inner core 9.68
1022 kg Yoder (1995) 1.6%
Outer core 1.84
1024 kg Yoder (1995) 30.7%
Core 1.93
1024 kg Yoder (1995) 32.3%
Mantle 4.015
1024 kg Chambat et al. (2010), Yoder (1995), 67.7%
Huang et al. (2013)
Crust (ocean + continental) 2.73(48)
1022 kg Huang et al. (2013) 0.46%
Inner core to core (%) 5.0%
Radius
Inner-outer core boundary 1220  10 km Masters and Shearer (1995) 19% E
Core-mantle boundary 3483  5 km Masters and Shearer (1995) 55%
Mean radius of the Earth 6371.23 (0.01) km Chambat et al. (2010) 100%
Volume
Inner core 7.61
109 km3 0.7%
Outer core 1.69
1011 km3 16%
Bulk core 1.77
1011 km3 16%
Silicate Earth 9.14
1011 km3 84%
Earth 1.083
1012 km3 100%
Density
Outermost inner core (solid) 12,830 kg/m3 Masters and Gubbins (2003)
Innermost outer core (liquid) 12,010 kg/m3 Masters and Gubbins (2003)
Inner-outer core D density 820 (180) kg/m3 Masters and Gubbins (2003)
Average outer core density 11,160 (60) kg/m3 Masters and Gubbins (2003)
Average inner core density 13,070 (260) kg = m3 Masters and Gubbins (2003)
CMB (Core-mantle boundary)
Peak-to-peak topography <3 km Sze and van der Hilst (2003)
CMB ellipticity ~0.2 km Sze and van der Hilst (2003)
Thermal and pressure data
Top of outer core 4000  500 K Tsuchiya et al. (2016), Fischer (2016) 137 GPa
Top of inner core 5700  550 K Tsuchiya et al. (2016), Fischer (2016) 330 GPa
Outer core adiabatic gradient 0.55 (0.1) K/km Tsuchiya et al. (2016)
Heat flow across CMB 5–15 TW Tsuchiya et al. (2016)
Moment of inertia constants
Equatorial moment of inertia (I) 0.3299765 Ma2 Yoder (1995)
Mean moment of inertia (I) 0.330690 (9) MR20 Chambat et al. (2010)

in primitive and iron meteorites inform us of their initial Formation

proportions in the Earth.
The maximum in the Earth’s gravitational acceleration
occurs at just above the CMB, where being close to the
high density core has its greatest effect on the gravity field.
Consequently, “g” (i.e., acceleration of gravity) increases
in roughly linear fashion from center to CMB (i.e., approx-
imately a uniform density sphere) and then remains
between 9.8 and 10.5 m/s2 from the surface to the CMB.
The Earth’s dynamic figure for the gravitational field is
known to six significant figures and in combination with
seismological constraints indicates that the excess elliptic-
ity of the CMB is slight (~0.2 km) and that the CMB
surface topography is smooth at long-wavelength being
on the order of 1–3 km of peak-to-peak topographic ampli-
tude (Table 1).
The core’s origin dates back to the formation of the planet,
with planet accretion occurring via accumulation and agglom-
eration of increasing size of particles and later planetesimals.
Our solar system owes it origin to a supernova (or multiple
supernovae) where the mass and momenta of the supernova
combined with a fragment of an interstellar gas-dust cloud to
rotate, collapse, condense, and centrally concentrate mass,
internally controlled by a balance of gravitational and cen-
tripetal forces. Because of its mass the Sun, and to a lesser
extent Jupiter, dominated accretion and condensation of the
inner rocky planets, which formed relatively early (a few
106 years) and grew by the accretion of material flowing
into and less so out of their respective feeding zone (first
few AU surrounding the Sun).
420
Growth by accretion of a planet does not require differ-
entiation of the body into a core and mantle, although
differentiation likely occurred in stages during growth and
included accretion of predifferentiated planetismals. For
the Earth, it is likely that as accretion progressed the
proto-Earth at some point reach melting temperatures for
metallic liquids (lower than that for silicate liquids) and
because of the large density contrasts (i.e., silicates being
~3000 and iron metal being ~7900 kg/m3 at surface condi-
tions) gravitational separation of a metallic core quickly
ensued. The mean time of formation of the bulk of the
Earth’s core is, however, an open question with best esti-
mates being between 30 and 60 Ma (million years) after the
formation of the solar system (an age established by the
oldest dated materials in the solar system, Hf-W isotope
systematics of iron meteorites, and the age of Moon for-
mation; more about this topic later). Note, timing of core
formation can be considered as the mean age (or mean
lifetime), which is a scaling time reflecting when much of
the mass of metallic atoms were separated. The initial mass
associated with core formation was likely significant and
rapidly separated. Subsequent additions to the core might
have followed some exponentially decreasing mass versus
time function.
A second constraint on the time of core formation comes
from understanding the origin of the Moon and the dating of
its oldest rocks. This constraint is based on the fact that the
Moon possesses a bulk composition that approximates that
of the Earth’s mantle (i.e., they share the same oxygen
isotopic composition to six significant figures), with a bit
more iron in the Moon (Young et al. 2016). The Moon’s
origin is often ascribed to being due to a giant impact
(collision) between a nearly completely formed Earth and a
Mars sized impactor (Mars is ~10% by mass the size of the
Earth, whereas the Moon’s mass is 7.346
1022 kg, which is
1.2% of the mass of the Earth) (Canup 2012; Cuk and
Stewart 2012). Alternative scenarios for the moon-forming
event(s) allow for a series of lesser, but still substantial,
impacts that also satisfy the observational constraints
(Cuk et al. 2016). The material ejected during this event
coalesced into a corotating accretion disk around the Earth,
possessed only a small fraction of metal from the Earth’s
core or the impactor, and subsequently accreted to form the
Moon. The general features of this model (Canup 2012;
Cuk and Stewart 2012; Cuk et al. 2016) are consistent with
hydrodynamic models of accretion that seek to reconcile the
angular momentum of the Earth-Moon system and isotopic
constraints.
Major questions about the composition of the Earth’s core,
its age, the age of the inner core, present conditions, and
dynamics of its formation remain unresolved. Debate focuses
on the following questions for which we have limited obser-
vational constraints:
Earth’s Core
Formation Conditions and Composition
– What was the integrated composition of material feeding
the Earth (i.e., chondrites)?
– What is the temperature and pressure evolution of planet
growth and core formations?
– What was the integrated fO2, fH2, and fS2 (gas fugacity) of
core formation?
– How did the redox state of core formation evolve (early to
late evolution)?
– Was segregation and sinking of metal from silicate an
equilibrium process throughout the mantle or only for a
portion thereof?
– Was core separation accomplished in the presence a
magma ocean or not?
– Did the process of giant impact collision coalesce and
emulsify the cores of the Earth and its impactor and
re-establish metal-silicate equilibrium?
– What is the duration and timing of core formation (e.g.,
182
Hf-182W. isotope system)?
– How much S, C, N, and H (i.e., volatile elements) are in
core?
– How much O, Si, and Mg (i.e., major elements) are in
core?
– Are there radioactive elements in the core and if so, how
much?
Inner-Outer Core
– When did the inner core start to form?
– What is the crystal structure of the inner core?
– How much and what is (are) the light element(s) in the
inner core?
– How much and what is (are) the light element(s) in the
outer core?
– Does enrichment of buoyant, light-element-rich liquids,
produced at the inner-outer core boundary, drive compo-
sitional convection and the geodynamo?
– To what degree does crystallization of the inner core drive
compositional convection?
Many unknowns remain and detailed constraints on the core’s
composition come almost exclusively from the chemical and
isotopic composition of the silicate Earth and meteorites.
Building Compositional Models of the Planet
and Core
Our view of planetary compositional modeling is explicitly
informed by the compositions of chondrites, the most primi-
tive rocks of the solar system. Chondritic meteorites, which
make up the bulk of the meteoritic material (~95% of the
falls), fell to the Earth during historical times and thus shape
our view of the building blocks of the planets (Engel and
Earth’s Core
McDonough 2016). CI1 carbonaceous chondrites have the
most primitive composition of the various chondrites and
closely match the spectroscopically derived composition of
the solar photosphere for all but H, C, N, O, and the noble
gases (McDonough 2016). [Note, that the chondrite type is CI
and the metamorphic grade is 1, hence CI1.] The element
abundance curve for the solar photosphere, for at least the
dozen most abundant elements, is the same as that for other
stars in the galaxy and nearby galaxies. Thus, the principles
developed below are applicable for exoplanet building too.
As building blocks, chondrites provide the basis for com-
positional models that can plausibly describe what is in the
Earth’s core. Some proposed models of the Earth’s composi-
tion fall outside of the observed range for chondrites, with
enrichments in refractory elements and stronger depletions in
volatile element abundances than seen in chondrites (Fig. 1)
(McDonough 2014). Element classification (i.e., volatile ver-
sus refractory) is defined in terms of half-mass condensation
temperatures of an element during cooling of a nebular accre-
tion disk and precipitation into common mineral phases.
Given the plausible range of volatile and refractory element
abundances in different groups of chondrites that accreted to
form the Earth, it is not possible to uniquely define the
planet’s composition. Moreover, while chondrites represent
a compositional guide, the Earth is unlikely to match the
composition of any specific, single chondrite. The standard
rules for planet building are as follows:
– Use orbital and seismic constraints, if available, to define
the physical state
– Use chemical and isotopic data from chondritic (i.e., prim-
itive) meteorites and the solar photosphere
– Assume chondritic proportions for the refractory elements
and model their absolute abundances based on chemical
trends observed for mantle samples (basalts and
peridotites),
Earth’s Core, Fig. 1 Chondrite normalize compositions for carbona-
ceous chondrites (CM, CO, CV subtypes), ordinary chondrites (H, L,
and LL subgroups), enstatite chondrites (EH and EL subgroups) and the
Earth. The figure is double normalized with all compositional models
plotted as having the same Si content, which corrects for differing
421
– Model the absolute abundances of Mg, Fe, Si, and O; these
are the major elements, which are nonrefractory elements,
and constitute >90% by mass of the rocky planets
– Recognize a chemical gradient in solar system (e.g., rocky
planets versus the gas giants)
– Model the relative and absolute abundances of the non-
refractory elements in the planet
Model core compositions are estimated by subtracting the
composition of the silicate Earth from the composition of the
bulk planet. The relative distribution of lithophile elements E
(those that tend to bond with oxygen and are excluded from
the core) versus siderophile elements (those that tend to bond
with iron and are dominantly stored in the core) in the crust
and mantle define the composition of the silicate Earth. For
refractory elements, including lithophile and siderophile,
their relative proportions in the planet are the same as in
chondrites (i.e., the lithophiles have constant ratios at 10%
or better, constant ratios of siderophiles are at 15% or better,
and ratios of lithophile to siderophile have uncertainties at
circa 25%, with the greater difference for the latter being
due to variations in the proportion of metal to silicate in
chondrites). Therefore, if we can predict the abundance of
one refractory element in the Earth, we can predict the plan-
etary inventory of all 37 refractory elements, which include:
Be, Al, Ca, Sc, Ti, V, Sr, Y, Zr, Nb, Ba, REE, Hf, Ta, Th, and
U (these are the lithophiles); and Mo, Ru, Rh, W, Re, Os, Ir,
and Pt (these are the siderophiles, with V showing both
behavior). How and why elements are distributed between
the core (mostly siderophile) and the surrounding silicate
shell (lithophile) depends on the conditions of pressure, tem-
perature, and redox state during core formation. The condi-
tions of core formation are critical to establishing its final
composition.
Among the dozen most common elements in the solar
system, the rocky planets are variably depleted in the most
amounts of highly volatile element contents (i.e., H, C, N, O). Normal-
izing reference frame assumes a CI1 chondrites composition. Data
for the chondrites are from Wasson and Kallemeyn (1988) and
McDonough (2014)
422 Earth’s Core

Earth’s Core, Fig. 2 Atomic

proportions of O, Fe, Si, and Mg in
chondrites and the Earth. Data
sources listed in Fig. 1. The
O context is 50% in the
chondrites and the total atomic
fraction for these four elements in
all of these materials adds up to
>90%

volatile constituents (i.e., the gases: H, He, C, N, O, Ne, Ar).

The remaining common elements are O, Fe, Si, and Mg,
which constitute between 90% and 95% the mass of the
Earth, Mars, and Venus roughly in the mass proportion
30:30:20:20 (or in the approximate atomic proportions of
50:15:15:15), respectively (Fig. 2). The high density and
small radius of Mercury indicates that it likely has a much
higher proportion of iron than the other rocky planets. Com-
positional models for the Earth, other terrestrial planets, and
exoplanets with bulk density in the range 3000–6000 kg/m3
will be dictated by mixtures of metal and silicate. The silicates
are predominately a bimineralic mix of olivine and pyroxene
at pressures equivalent to the Earth’s upper mantle. At the
seismically define base of the Transition Zone (i.e., 660 km
depth, ~22 GPa), olivine disproportionates into bridgmanite
(Mg-Fe silicate perovskite) and ferro-periclase, whereas
the pyroxene component (which in the upper mantle has
gradually dissolved into a majoritic garnet) converts to
bridgmanite and Ca perovskite over a broader range of pres-
sures (i.e, ~19–24 GPa).
A big unknown in predicting the bulk composition of a
planet is establishing the proportion of olivine to pyroxene
accreted by the body, which is indicated by its Mg/Si ratio.
There is no fixed Mg/Si ratio for chondrites and thus there is
no chondritic constraint (Fig. 3). Olivine has a 2:1 molar
Mg/Si content, whereas pyroxene’s molar ratio is 1:1. Astro-
mineralogical observations of accretion disks reveal radial
variations in the proportion of olivine to pyroxene, some
with inner disk regions being richer in olivine relative to the
disk wide composition (van Boekel et al. 2004), whereas
other have more olivine in the outer part of the circumstellar
disk; differences in disk mineralogy may relate to type of star
(e.g., T Tauri vs. Herbig Ae/Be stars) (Bouwman et al. 2010).
Likewise, one can also envisage disk evolution in time with
temperature decreasing as the disk cools leading to decreasing
Mg/Si (less olivine ppt with time). Moreover, crystallinity
scales radially, with inner disk regions (a few AU) having
Earth’s Core, Fig. 3 The compositional space for the chondrites and
the Earth in a tetrahedral volume plotted from the apices of Si, Mg, Fe,
and O, with the data plotted onto the surface of the Si-Fe-Mg plane. Data
sources listed in Fig. 1. The Earth’s composition falls outside of the field
of the known chondrites. This model composition for the Earth is
acceptable given observations of the spatial variations in rates of olivine
to pyroxene crystallization in accretion disks (See text for further details)
higher abundances of large grains and generally higher crys-
tallinity as compared to outer disk regions, suggesting grain
growth occurs more rapidly in the inner disk regions
(D’Alessio et al. 2005; Sargent et al. 2009). Thus, planetary
modelers must acknowledge that there is no fixed Mg/Si value
for a planet.
After O, Fe, Si, and Mg, the next most abundant elements
in chondrites and rocky planets are Al, Ca, and Ni and
together these seven elements, plus S, make up >98% by
mass of these solid bodies. Both Ca and Al are refractory
elements, and their planetary weight ratio is 1.1 (atomic ratio
is 0.72), equal to that of chondrites, but their absolute abun-
dances in the Earth are model dependent; these elements are
exclusively lithophile and thus concentrate in only the silicate
portion of the Earth. Nickel and sulfur are both nonrefractory
elements and are partitioned into both the core and mantle,
with Ni being dominantly siderophile (>90% of the Earth’s
Earth’s Core
Earth’s Core, Fig. 4 A plot of the oxidized iron to Si content of
chondrites versus the metallic and sulfur-bearing content to Si content
of chondrites. This diagram is also referred to as a Urey-Craig diagram.
The Earth plots intermediate to the highly reduced enstatite chondrites
and the less reduced ordinary chondrites. The CM and CI chondrites
have nearly (or all) iron as oxidized, with no metallic iron
Ni is hosted in the core) and sulfur is the definition of what is a
chalcophile element (>95% of the Earth’s sulfur is in the
core) (McDonough 2014).
The ratio of metals to oxides is a basic feature used to
classify chondrites. The chondrites, carbonaceous (highly
oxidized), ordinary (an intermediate redox state), and
enstatite (highly reduced state) can be divided into subgroups
of high, low, and low-low iron contents. Therefore, the
amount of Fe relative to O, Si, and Mg in chondrites and
planets can vary considerably. The redox state of a planet can
be readily established by determining the Mg # (atomic
Mg/(Mg + Fe)) of mantle silicates in combination with a
value for the mass proportion of metallic to silicate shells.
The Earth with 90% of its iron in metallic state is a reduced
body with a redox state intermediate between ordinary and
enstatite chondrites (Fig. 4).
Importantly, there are a number of critical element ratios
for the nonrefractory elements that are relatively constant in
chondrites and, if these element pairs have similar condensa-
tion temperatures, then these fixed ratios can be used to place
compositional constraints on the bulk planet. The fixed chon-
dritic ratios of Fe/Ni (17.4  0.5) and Ni/Co (20.5  0.6)
constrains the planetary and core abundances of Ni and Co,
given well established values for the mantle. To this end, a
major advance was made when Thibault and Walter (1995)
and Li and Agee (1996) identified the pressure dependence
423
partitioning of Ni relative to Co and found that at mid-mantle
conditions the relative partitioning of these two elements
approached 1, with both the core and mantle having chon-
dritic Ni/Co values. This finding was fundamental as the
silicate Earth and core are assumed to have a mass balance
in these ratios.
The initial growth of Earth and core-mantle segregation is
often consider in terms of occurring under conditions of
homogeneous or heterogeneous accretion, with the former
model assuming that the accreting materials had a fixed,
constant composition throughout planetary growth, while E
latter model models envisages significant changes in the
composition of material that accretes to the Earth as it grew
(see also discussion in McDonough (2014)). These different
models seek to explain the siderophile element composition
of the mantle. For the most part models of homogeneous
accretion require a continuous and restricted set of physical
and chemical conditions to achieve the present composition
of the core and mantle, whereas heterogeneous accretion
models allow for a range of probable conditions envisaged
in an evolving proto-planetary disk, including temporal var-
iations in redox conditions and the relative proportions of
volatile to refractory materials.
Given an average condition for temperature, pressure,
and oxygen fugacity of core formation we can construct
models for the core that fit the existing constraints and
match compositional models of the silicate Earth and chon-
drites. The abundance, absolute and relative, of various
redox sensitive siderophile elements (e.g., W, V, Cr, Nb,
Mn, Mo, and others) in the silicate Earth further define the
relative silicate-metal partitioning of elements during core
formation (Badro et al. 2015; Badro et al. 2014; Corgne et al.
2008; Fischer et al. 2015; Rubie et al. 2015). At this stage
there remains one further consideration before estimating
the composition of the core, and that is identifying the
chemical identity of the element(s) accounting for its density
deficit.
Finally, an important constraint on core formation came
from the observation that the primitive mantle abundances of
the highly siderophile elements (i.e., those with metal/silicate
partition coefficients of 104 and greater and include the plat-
inum group metals (Ru, Rh, Pd, Os, Ir, Pt), Re and Au) are in
chondritic proportions and are enriched in the mantle at con-
centration levels well beyond that expected for their known
metal/silicate partition coefficients (Kimura et al. 1974). Fol-
lowing a range of experimental studies it has been demon-
strated that that the relative metal/silicate partitioning
behavior of the highly siderophile elements are 104 and
greater and that the relative differences between different
member of this element group would result in grossly non-
chondritic ratios for the Re-Os-Pt isotope system, which is not
observed (Brenan et al. 2016; Day et al. 2016). Thus, the
mantle signature requires the addition of a late veneer of
424
added chondritic material, with a mass contribution equiva-
lent to about 0.5% to <1% the mass of the mantle (Day
et al. 2016).
Core Composition and Its “Light Element(s)”
Table 2 presents a model for the composition of the Earth, its
core, and the silicate Earth. The core model was constructed
as the difference based on estimates of the bulk silicate Earth
and the bulk planet compositions. The core composition
proposed here has a mixture of O, Si, and S that accounts
for the density deficit (3–7%). The relative proportions of
O and Si in the core were established by metal/silicate
partitioning (Badro et al. 2015; Fischer et al. 2015; Rubie
et al. 2011). Assuming the core’s mean atomic number to be
23 (Birch 1964), the absolute proportions of O and Si is then
fixed. By way of comparison, McDonough (2014) proposed
two end-member models, with one model containing Si and
S and no O as the light elements and the other model contains
O and S and no Si. Other compositional models are presented
for comparison in Table 1 in Hirose et al. (2013), where they
have sorted the different models in terms of simple single light
element versus multiply, mixed light element models. One
should view all of these models as acceptable, given the
available constraints.
Let me start by saying that I do not know what is (are) the
light element(s) in the core. The issue has dogged our science
for half of a century and will continue to do so for some time.
Words of caution are needed before discussing further the
composition of the core. My bias as a geochemist states that
observations on the composition of the silicate Earth and the
range of compositions in chondrites defines the recipe that
must be matched by experimental studies to define the P-T-
fO2-Xi (i.e., pressure, temperature, oxygen fugacity, and
compositional) space for metal-silicate equilibrium during
core formation. Models describing the dynamics of core
formation and its final composition are probability positions
and need to be recognized as such. As a second cautionary
note, the periodic table is not a one-stop shopping table, if it
is argued that a minor or trace element is in the core, then
other elements in the same column are likely there too in
some proportional relationship (e.g., Cr: Mo and W, Mn: Re,
V: Nb and Ta). This principle is fundamental chemistry
reflecting the importance of outer orbital electrons in con-
trolling the partitioning behavior of elements. Thirdly, all
partition coefficients (D-values) go to 1 as temperature goes
to infinity and thus metal-silicate fractionation becomes
muted at higher temperature conditions. Finally, appeals to
adding specific elements into the core are also understand-
ably done so to explain the power driving the geodynamo;
resolving the energy equation for the geodynamo is intracta-
ble and will be so for some time.
Earth’s Core
The identity of the light element(s) in the core is elusive.
Many suggestions have been published over the decades
(e.g., H, C, N, O, Mg, Si, S). In general, any element with a
lower atomic number than 26Fe is a likely candidate. Assum-
ing Birch’s law
 0:5
K
Vf ¼ ; V p ¼ Mr þ B þ AðT  T 0 Þðr  r Þ (1)
r
where Vf is hydrodynamic or bulk sound velocity, VP is
compressional wave velocity, k is bulk modulus, r is density,
M and B are room temperature coefficients of Birch’s law, and
A and r* account for its temperature dependence (see also
Sakamaki et al. (2016))], one can predict the mean atomic
weight of the mantle or core. Recently, Helffrich (2015)
reported a similar relationship between sound speed and
molecular weight using a hard sphere model for liquids. In
addition, Poirier (1994), Anderson and Isaak (2002) and
Badro et al. (2014) present simple graphical models to deter-
mine the amount (in wt%) of light element mixture in the core
that scales as a function of core temperature. The model
presented in Table 2 uses these principles to constrain the
proportion of O and Si in the core. The S content of the core is
established on cosmochemical grounds (Dreibus and Palme
1996) and constraints from the modeling of the bulk silicate
Earth (McDonough and Sun 1995).
Chemical and physical observations are useful for devel-
oping plausible models for the nature and amount of a light
element(s) in the core. Lithophile elements like O, Si, and Mg,
in addition to Fe, are the most abundant elements in the Earth
and can be readily defended as being dissolved in a core
forming metallic liquid. Likewise, the gases of H-C-N-O are
abundant in the solar system and would have been abundant
in the nebular disk. The rocky planets might have had a dense
surrounding atmosphere, particularly early in their evolution,
which increases the solubility of these gases in a magma
ocean and the potential for partitioning these elements into a
core forming metallic melt. Thus, models invoking H, C,
and/or N as constituents of the core cannot be dismissed. On
the other hand, however, the Earth, and potentially the rest of
the terrestrial planets in the solar system, is volatile depleted,
as documented by its Rb-Sr, K-Ar, K-U, and He-Ne-Ar iso-
topic systems and total radiogenic heat production. Therefore,
the Earth’s volatile depletion curve establishes the limit of 2%
(by mass) of sulfur in the core. Overall, chemistry offers
guidelines, but a few constraints in eliminating competing
core compositional models. Physical observations provide
the best limits on acceptable core models.
The small addition of the lightest elements (e.g., H and C)
is appealing because they offer considerable leverage in low-
ering the mean atomic weight of the core. It is noteworthy that
the core composition reported in Table 2 posits significant
amounts of C and H in the core given the Earth’s volatility
Earth’s Core 425

Earth’s Core, Table 2 Composition of the Earth, silicate Earth, and core in weight percent and atomic proportions
wt% Earth Silicate Earth Core Atomic prop. Earth Silicate Earth Core
Fe 31.9 6.26 85.5 Fe 0.148 0.024 0.750
O 30.4 44 2 O 0.493 0.581 0.061
Si 15.3 21 4 Si 0.142 0.158 0.070
Mg 15.4 22.8 0 Mg 0.164 0.198 0
Ni 1.83 0.196 5.10 Ni 0.008 0.001 0.043
Co 0.090 0.0105 0.249 Co 0.0004 0.00004 0.002
Ca 1.71 2.53 0 Ca 0.011 0.013 0
Al 1.59 2.35 0 Al 0.015 0.018 0
S 0.650 0.025 1.9 S 0.005 0.0002 0.029
Cr 0.43 0.263 0.75 Cr 0.002 0.001 0.007 E
Na 0.18 0.27 0 Na 0.002 0.002 0
P 0.072 0.009 0.02 P 0.001 0.0001 0.0003
Mn 0.08 0.105 0.03 Mn 0.0004 0.0004 0.0003
C 0.073 0.012 0.2 C 0.002 0.0002 0.008
H 0.026 0.01 0.06 H 0.007 0.002 0.029
Total 99.7 99.8 99.8 Total 1.000 1.000 1.000
Absolute and relative proportions of O and Si in the core were established by metal/silicate partitioning and fixing the core’s mean atomic number to
22.9 (c.f. Birch 1964)

curve (McDonough 2014). However, their appeal is dimin-
ished if their addition to the core also comes with conse-
quences for the mantle (e.g., limited metal-silicate
fractionation for H and C during core formation). Composi-
tional models for the mantle have proposed a range of H and
C abundances (Dasgupta and Hirschmann 2010; Halliday
2013; Marty 2012) and variable estimates for the contents of
H and C in the core and some model compositions proposed
for the Earth’s core suggest abundant carbon (Wood 1993) or
hydrogen (Williams and Hemley 2001) for consideration.
Considerable attention has been paid to models with O, Si,
or both as likely light element candidates (Badro et al. 2015;
Fischer et al. 2015; Li and Fei 2014; Ringwood 1984). Based
on the differences in the silicon isotopic composition of
chondrites and the Earth, Fitoussi and Bourdon (2012) and
Fitoussi et al. (2016) proposed that the Earth core contains
some 4–7 wt% Si, with the range taking into account metal-
silicate isotopic fractionation at presumed core separation
temperatures. Initially, the presence of silicon and oxygen in
the core was considered an either or condition, whereas today
many have found a range of acceptable conditions of core
formation that result in iron alloys with mixed amounts of
silicon and oxygen. More recent models suggest that Mg may
have once been present as a light element in the core (Badro
et al. 2016; O’Rourke and Stevenson 2016). This suggestion
is permitted given that Mg is one of the four most abundant
elements in the Earth. Moreover, these authors note that as the
core cools down it exsolves MgO (or Mg silicate) and this
reaction is exothermic and thus provides a power source for
driving the geodynamo. Alternatively, Hirose et al. (2017)
have suggested that exsolution of SiO2 from the outer core is
another potential power source for the geodynamo. The
present-day situation we face is that there are many competing
models to explain the core composition and potentially also
energy sources for driving the core’s dynamo.
Power to Drive the Geodynamo and Radioactive
Elements in the Core
The geodynamo is produced in the outer core and demands
power for its operation. Several papers have considered the
presence of radioactive, heat producing elements (i.e.,
HPE = K, Th, and U) in the Earth’s core, as they provide
power to drive the core’s dynamo. There is no compelling
experimental or geochemical constraint that requires HPE in
the core. Moreover, doing so often requires incorporation of
other lithophile elements for which there is no supporting
evidence for their depletion in the silicate Earth.
The energy budget of the core does not require HPE, as
alternative energy sources (e.g., oxide exsolution, see above)
may also provide the needed power. In addition, unknowns
in understanding the generation and maintenance of a
geodynamo are great, which includes being able to describe
accurately the core’s energy equation.
As is the case with the discussion of the light element in the
core, so too for the presence of radioactive elements in the
core: there are big unknowns. Thus, one should not confuse
the conditions of what is plausible and what is needed.
Potassium: Incorporation of K in the core has been con-
sidered by numerous workers (Corgne et al. 2007; Li and Fei
2014; Watanabe et al. 2014) and generally found that a few to
a couple tens of m g/g of K, or less, can be accommodated in
the core. Using ab initio calculations Lee et al. (2004)
426
suggested that the metallization of potassium at pressure, due
to s-to d-orbital transitions, as a mechanism for alloying
potassium with Fe. Both Cs and Rb, larger alkali metals,
would undergo similar orbital transitions, but at lower pres-
sures. There is no evidence, however, that Rb is anomalously
depleted in the mantle with respect to the planetary volatility
curve, contrary to an expectation from the pressure-dependent
metallization process.
Uranium and thorium: Wheeler et al. (2006) and
Malavergne et al. (2007) show that under core segregation
conditions U remains strongly lithophile and is for the
most part excluded from the core. Similarly, thorium has not
been found to partition into a core forming metal phase; it
is understood to be wholly lithophile. An additional constraint
on the abundance and relative distribution of Th and U in
the bulk silicate Earth comes from Pb isotope studies. The
kPb ratio (i.e., the time integrated Th/U ratio derived from the
208 
Pb/206 Pb ratio) of the silicate Earth is comparable to that
of chondrites at 25% (Andersen et al. 2015; Castillo 2016;
Elliott et al. 1999; Galer and O’Nions 1985; Kumari et al.
2016; Paul et al. 2003), consistent with negligible amounts of
Th and U in the Earth’s core. Studies on the partitioning of
thorium between metal and silicate demonstrate that it is
wholly lithophile in its behavior.
The Age of Core-Mantle Separation
Age of core formation is constrained by the 182Hf-182W
isotope system (t1/2 of 9 Ma) and known to be on the order
of 30–150 million years after the age of solar system forma-
tion (Kleine et al. 2002; Kleine et al. 2009; Yin et al. 2002).
The solar system formation age (t0 = 4568 Ga) is established
mostly from the absolute system, U-Pb isotopes, and the
relative system 26Al/26Mg isotopes (t1/2 of 0.7 Ma), as applied
to the earliest condensates, which are high temperature,
refractory materials (i.e., CAIs, calcium alumina inclusions)
found as inclusions in chondritic meteorites (Brennecka et al.
2015; Connelly et al. 2012; MacPherson et al. 2012).
Together these isotopic systems indicate that the bulk of
Earth’s accretion occurred some 106–107 years after t0 and
core formation followed as accretion progressed, including
some modification during and following the Moon-forming
impact event.
The current core formation paradigm posits segregation of
sinking metal diapirs that chemically interact with the imme-
diately surrounding silicate (liquid or solid) during or follow-
ing a magma ocean(s) environment of the early Earth. The
degree of metal-silicate equilibrium occurring during descent
of these diapirs is unknown. However, as metal aggregates
ponds and then begin to sink the potential for chemical
exchange becomes less efficient as the length scale increases.
Chemical consequences of a large Moon-forming impact event
Earth’s Core
introduces other unknowns, for which we do not know whether
or not the impactor’s core merged, emulsified in the mantle, or
otherwise superimposed additional complications on the core’s
composition (Nakajima and Stevenson 2015).
Hafnium and tungsten readily and effectively separate
between the metallic (core) and silicate Earth (planet minus
the core), with Hf hosted nearly exclusively in the silicate
Earth and W almost wholly (90%) in the core. Due to the solar
system initiating supernova, the solar system (and hence the
Earth) inherited a fraction of live 182Hf (Lugaro et al. 2014)
that decayed into 182W and consequently the age of core
formation reflects the competing processes of metal segrega-
tion versus 182Hf decay.
Chronologically, what has been established recently is that
small bodies, as represented by groups of iron meteorites,
accreted and differentiated quickly (i.e., 0–5 Ma after t0)
(Kruijer et al. 2014). During the waning stages of small-
body accretion (t0 + 3–5 million years), and when radiogenic
heating from the 26Al/26Mg isotope system was running out
of energy, the formation of chondritic bodies occurs, which
are the most primitive and undifferentiated planetesimals in
the solar system. Hence, the preservation of chondritic parent
bodies represents a waning stage of accretion of small bodies.
It is likely that Earth’s formation occurred throughout this
sequence and beyond and consequently its inventory of 182W
was shared between the core and mantle, revealing that the
amount of 182W in the silicate Earth establishes a lower limit
to core formation age, while the timing of Moon formation
sets the upper limit. Our present estimate is the silicate Earth
has a m182W isotopic composition of zero (it is the reference
point for the isotope system and units are expressed as devi-
ations in parts in 106), chondrites are approximately 200
and the Earth’s core is slightly more negative (~220)
(Walker 2016).
Inner Core and Its Age of Crystallization
The inner core is <2% the mass of the Earth and 5% the mass
of the core. It is slightly smaller than the Earth’s moon and is
at about the temperature of the Sun’s surface. The ICB (inner-
outer core boundary) surface provides a constraint for esti-
mating the temperature at the center of the Earth. An upper
temperature limit is obtained from the melting temperature of
Fe at relevant temperatures and pressures (~6200 K, 330 GPa)
and an assumed conductive thermal gradient (Fischer 2016).
Experimental challenges associated with determining the
melting curve of iron and the upper temperature limit of the
ICB are formidable; there is a range of 1500 K at 330 GPa for
the Fe-alloy melting curves (Fischer 2016). Nonetheless,
given the potential light element candidates and their influ-
ence on melting point depression, an ICB temperature of
5700  550 K covers the range of estimates (Fischer 2016).
Earth’s Core
The inner core has about half the density deficit relative to
that of the outer core (Anderson and Isaak 2002; Jephcoat and
Olson 1987; Masters and Gubbins 2003); proportionally, one
might intuit that this translates to approximately half the
amount of light element. The phase diagram dictates that the
inner core crystallization temperature will be lower than that
for pure iron due to freezing point depression. Thus, an
estimate of the temperature of the ICB depends on one’s
estimate for the light element composition of the outer core,
which is a concatenation of unknowns.
Compositional estimates for the inner core are not forth-
coming, as the unknowns are great, and come with consider-
able nonuniqueness (Antonangeli et al. 2010; Badro et al.
2007; Sakamaki et al. 2016). We know that its solid, likely
in an hcp (hexagonal close packed) structure, there is
820  180 kg/m3 density increase across the ICB (Masters
and Gubbins 2003), it has a density deficit, and it is aniso-
tropic (polar vs equatorial wave speed). The inner core may be
undergoing thermal convection (Cottaar and Buffett 2012),
which may explain its seismic anisotropy.
Solidification of the inner core produces latent heat and
results in a buoyant, residual liquid enriched in light elements.
The Coriolis force from the Earth’s rotation, combined with a
heated, compositionally buoyant fluid, gives this ascending
fluid the power to generate a magnetic field. In addition, heat
extracted from the core by the mantle, as the planet cools,
gives rise to the core’s thermal convection. In combination
with rotational forces, the thermal and compositional effects
are the most probable energy sources driving the geodynamo.
When did inner core crystallization begin? This question
hinges greatly on understanding the thermal history of the
core. There are many unknowns that contribute to this ques-
tion and most are active areas of research accompanied by
considerable debate, these include: the thermal and electrical
conductivity of core materials (where there is an ongoing
debate over the factor of 2 difference in the assumed values
for these parameters), the initial and present day core temper-
ature, heat flow across the CMB (see Table 1), and whether
there are (and how much) radioactive elements in the core.
Debates on these topics will continue for some time as
technical and computational challenges are great, including
understanding the energy budget needed to drive the
geodynamo.
For about a decade there were models that proposed early
inner core crystallization, as early as 4.5 Ga. Models of early
inner core growth were nearly exclusively based on anoma-
lous 186Os-188Os and 187Os-188Os isotopic compositions in
basalts whose source region experience core additions; the
origins of these anomalies came from decays in the Re-Os and
Pt-Os isotope systems and parent daughter fractionation due
to inner core crystallization (Walker 2016). More recently,
however, corroborating 182W isotopic evidence of this core-
mantle exchange process occurring was not found in these
427
basalts and thus the model of early inner core growth based on
isotopic evidence has been abandoned (Walker 2016).
Geophysical estimates for the age of the inner core are
based on a full understanding of the core’s energy equation
and usually includes a trade-off between internal heat produc-
tion and heat dissipation across the core-mantle boundary.
Labrosse et al. (2001) examined the question assuming a
role for heat generation from potassium, modeling up to
200 mg/g K in the core, and showed that core crystallization
might have started anytime from the Archean to nearly the
present depending on how much potassium is in the core. E
Later, addressing this issue following the recently revised
thermal and electrical conductivity values for the core,
Gomi et al. (2013) and Labrosse (2015) concluded that the
amount of potassium in the core plays less of a role in
establishing the initiation of inner-core crystallization, but
that the core cooling rate is significant. From these analyses
a young age (circa ~05 Ga) for initiation of inner core is
favored. Alternatively, estimates of an inner core having
begun crystallization from 1.5 to 0.5 billion years ago are
common (Labrosse et al. 2001). Overall, there are few con-
straints regarding the physical state and evolution of the core
and inner core system, leaving us with open questions about
when the inner core began to crystallize, the heat flux across
the core-mantle boundary, and the power budget for the
geodynamo.
Summary
The Earth’s core comprises the innermost 3483 km of the
planet. Both the inner solid core and outer liquid core consist
of about 85% iron and 5% nickel with 10% other elements
including lighter elements O, Si, and/or S as well as most of
the inventory of highly siderophile elements such as the
platinum group elements and gold. A metallic liquid core
had likely formed within the first 30–60 million years of
solar system history. The solid inner core has grown by
crystallization of the outer core, a process that may have
begun as recently as 0.5–1.5 billion years ago.
Cross-References
▶ Chondrites
▶ Cosmic Elemental Abundances
▶ Formation and Evolution of the Earth
▶ Fugacity
▶ Giant Impact Hypothesis
▶ Heat-Producing Elements (HPEs)
▶ Iron
▶ Iron Meteorites
428
▶ Lithophile Elements
▶ Nickel
▶ Partitioning and Partition Coefficients
▶ Siderophile Elements
▶ Sulfur
References
Andersen MB, Elliott T, Freymuth H, Sims KWW, Niu Y, Kelley
KA (2015) The terrestrial uranium isotope cycle. Nature 517:
356–359
Anderson OL, Isaak DG (2002) Another look at the core density deficit
of Earth’s outer core. Phys Earth Planet Inter 131:19–27
Antonangeli D, Siebert J, Badro J, Farber DL, Fiquet G, Morard G,
Ryerson FJ (2010) Composition of the Earth’s inner core from high-
pressure sound velocity measurements in Fe-Ni-Si alloys. Earth
Planet Sci Lett 295:292–296
Badro J, Brodholt JP, Piet H, Siebert J, Ryerson FJ (2015) Core forma-
tion and core composition from coupled geochemical and geophys-
ical constraints. Proc Natl Acad Sci USA 112:12310–12314
Badro J, Cote AS, Brodholt JP (2014) A seismologically consistent
compositional model of Earth’s core. Proc Natl Acad Sci USA
111:7542–7545
Badro J, Fiquet G, Guyot F, Gregoryanz E, Occelli F, Antonangeli D,
d’Astuto M (2007) Effect of light elements on the sound velocities in
solid iron: Implications for the composition of Earth’s core. Earth
Planet Sci Lett 254:233–238
Badro J, Siebert J, Nimmo F (2016) An early geodynamo driven by
exsolution of mantle components from Earth’s core. Nature
536:326–328
Birch F (1952) Elasticity and Constitution of the Earth’s Interior.
J Geophys Res 57:227–286
Birch F (1964) Density and Composition of Mantle and Core. J Geophys
Res 69:4377–4388
Bouwman J, Lawson WA, Juhasz A, Dominik C, Feigelson ED,
Henning T, Tielens AGGM, Waters LBFM (2010) The protoplane-
tary disk around the M4 Star Recx 5: Witnessing the influence of
planet formation? Astrophys J Lett 723:L243–L247
Brenan JM, Bennett NR, Zajacz Z (2016) Experimental results on
fractionation of the highly siderophile elements (HSE) at variable
pressures and temperatures during planetary and magmatic differen-
tiation. In: Harvey J, Day JMD (eds) Highly siderophile and strongly
chalcophile elements in high-temperature geochemistry and cosmo-
chemistry, vol 81. Mineralogical Society of America, Washington,
DC, pp 1–88
Brennecka GA, Budde G, Kleine T (2015) Uranium isotopic composi-
tion and absolute ages of Allende chondrules. Meteorit Planet Sci
50:1995–2002
Canup RM (2012) Forming a Moon with an Earth-like composition via a
Giant Impact. Science 338:1052–1055
Castillo PR (2016) A proposed new approach and unified solution to old
Pb paradoxes. Lithos 252:32–40
Chambat F, Ricard Y, Valette B (2010) Flattening of the Earth: Further
from hydrostaticity than previously estimated. Geophys J Int
183:727–732
Connelly JN, Bizzarro M, Krot AN, Nordlund A, Wielandt D,
Ivanova MA (2012) The absolute chronology and thermal processing
of solids in the solar protoplanetary disk. Science 338:651–655
Corgne A, Keshav S, Fei YW, McDonough WF (2007) How much
potassium is in the Earth’s core? New insights from partitioning
experiments. Earth Planet Sci Lett 256:567–576
Corgne A, Keshav S, Wood BJ, McDonough WF, Fei YW (2008) Metal-
silicate partitioning and constraints on core composition and oxygen
Earth’s Core
fugacity during Earth’s accretion. Geochim Cosmochim Acta
72:574–589
Cottaar S, Buffett B (2012) Convection in the Earth’s inner core. Phys
Earth Planet Inter 198:67–78
Cuk M, Stewart ST (2012) Making the Moon from a fast-spinning earth:
a giant impact followed by resonant despinning. Science
338:1047–1052
Cuk M, Hamilton DP, Lock SJ, Stewart ST (2016) Tidal evolution of the
Moon from a high-obliquity, high-angular-momentum Earth. Nature
539:402–406
DAlessio P, Merin B, Calvet N, Hartmann L, Montesinos B (2005)
WWW database of models of accretion disks irradiated by the central
star. Rev Mex Astron Astrofis 41:61–67
Dasgupta R, Hirschmann MM (2010) The deep carbon cycle and melting
in Earth’s interior. Earth Planet Sci Lett 298:1–13
Day JMD, Brandon AD, Walker RJ (2016) Highly siderophile elements
in Earth, Mars, the Moon, and Asteroids. In: Harvey J, Day JMD
(eds) Highly Siderophile and Strongly Chalcophile Elements in
High-Temperature Geochemistry and Cosmochemistry, vol 81. Min-
eralogical Society of America, pp 161–238
Dreibus G, Palme H (1996) Cosmochemical constraints on the sulfur
content in the Earth’s core. Geochim Cosmochim Acta 60:
1125–1130
Elliott T, Zindler A, Bourdon B (1999) Exploring the kappa conundrum:
the role of recycling in the lead isotope evolution of the mantle. Earth
Planet Sci Lett 169:129–145
Engel K, McDonough WF (2016) Geochemical models of the Earth and
the crustal geoneutrino flux. In: Ludhova L (ed) Geo-neutrino. Open
Academic Press
Fischer RA (2016) Melting of Fe alloys and the thermal structure of the
core. In: Terasaki H, Fischer RA (eds) Deep Earth: physics and
chemistry of the lower mantle and core. John Wiley & Sons, Inc.,
Washington, DC, pp 3–12
Fischer RA, Nakajima Y, Campbell AJ, Frost DJ, Harries D,
Langenhorst F, Miyajima N, Pollok K, Rubie DC (2015) High pres-
sure metal-silicate partitioning of Ni, Co, V, Cr, Si, and O. Geochim
Cosmochim Acta 167:177–194
Fitoussi C, Bourdon B (2012) Silicon isotope evidence against an
enstatite chondrite Earth. Science 335:1477–1480
Fitoussi C, Bourdon B, Wang XY (2016) The building blocks of Earth
and Mars: a close genetic link. Earth Planet Sci Lett 434:151–160
Galer SJG, O’Nions RK (1985) Residence time of thorium, uranium and
lead in the mantle with implications for mantle convection. Nature
316:778–782
Gomi H, Ohta K, Hirose K, Labrosse S, Caracas R, Verstraete MJ,
Hernlund JW (2013) The high conductivity of iron and thermal
evolution of the Earth’s core. Phys Earth Planet Inter 224:
88–103
Halliday AN (2013) The origins of volatiles in the terrestrial planets.
Geochim Cosmochim Acta 105:146–171
Helffrich G (2015) The hard sphere view of the outer core. Earth Planets
Space 67:73
Hirose K, Labrosse S, Hernlund J (2013) Composition and State of the
Core. Annu Rev Earth Planet Sci 41:657–691
Hirose K, Morard G, Sinmyo R, Umemoto K, Hernlund JW, Helffrich G,
Labrosse S (2017) Crystallization of silicon dioxide and composi-
tional evolution of the Earth’s core. Nature 543:99–102
Huang Y, Chubakov V, Mantovani F, Rudnick RL, McDonough WF
(2013) A reference Earth model for the heat-producing elements
and associated geoneutrino flux. Geochem Geophys Geosyst
14:2003–2029
Jephcoat A, Olson P (1987) Is the inner core of the Earth pure iron.
Nature 325:332–335
Kimura K, Lewis RS, Anders E (1974) Distribution of gold and rhenium
between nickeliron and silicate melts. Geochim Cosmochim Acta
38:683,781
Earth’s Core
Kleine T, Munker C, Mezger K, Palme H (2002) Rapid accretion and
early core formation on asteroids and the terrestrial planets from
Hf-W chronometry. Nature 418:952–955
Kleine T, Touboul M, Bourdon B, Nimmo F, Mezger K, Palme H,
Jacobsen SB, Yin QZ, Halliday AN (2009) Hf-W chronology of the
accretion and early evolution of asteroids and terrestrial planets.
Geochim Cosmochim Acta 73:5150–5188
Kruijer TS, Touboul M, Fischer-Godde M, Bermingham KR, Walker RJ,
Kleine T (2014) Protracted core formation and rapid accretion of
protoplanets. Science 344:1150–1154
Kumari S, Paul D, Stracke A (2016) Open system models of isotopic
evolution in Earth’s silicate reservoirs: implications for crustal
growth and mantle heterogeneity. Geochim Cosmochim Acta
195:142–157
Labrosse S (2015) Thermal evolution of the core with a high thermal
conductivity. Phys Earth Planet Inter 247:36–55
Labrosse S, Poirier JP, Le Mouel JL (2001) The age of the inner core.
Earth Planet Sci Lett 190:111–123
Lee KKM, Steinle-Neumann G, Jeanloz R (2004) Ab-initio high-
pressure alloying of iron and potassium: implications for the Earth’s
core. Geophys Res Lett 31, L11603, https://doi.org/10.1029/
2004GL019839
Li J, Agee CB (1996) Geochemistry of mantle-core differentiation at
high pressure. Nature 381:686–689
Li J, Fei Y (2014) Experimental constraints on core composition. In:
Carlson RW (ed) Treatise on geochemistry: the mantle. Elsevier,
New York, pp 527–557
Lugaro M, Heger A, Osrin D, Goriely S, Zuber K, Karakas AI,
Gibson BK, Doherty CL, Lattanzio JC, Ott U (2014) Stellar origin
of the Hf-182 cosmochronometer and the presolar history of solar
system matter. Science 345:650–653
MacPherson GJ, Kita NT, Ushikubo T, Bullock ES, Davis AM
(2012) Wellresolved variations in the formation ages for Ca-Al-
rich inclusions in the early solar system. Earth Planet Sci Lett 331:
43–54
Malavergne V, Tarrida M, Combes R, Bureau H, Jones J, Schwandt C
(2007) New high-pressure and high-temperature metal/silicate
partitioning of U and Pb: implications for the cores of the Earth and
Mars. Geochim Cosmochim Acta 71:2637–2655
Marty B (2012) The origins and concentrations of water, carbon,
nitrogen and noble gases on Earth. Earth Planet Sci Lett 313:
56–66
Masters G, Gubbins D (2003) On the resolution of density within the
Earth. Phys Earth Planet Inter 140:159–167
Masters TG, Shearer PM (1995) Seismic models of the Earth: elastic and
anelastic. In: Ahrens TJ (ed) Global Earth Physics. American Geo-
physical Union, Washington, DC, pp 88–103
McDonough WF (2014) Compositional models for the Earth’s core. In:
Carlson RW (ed) Treatise on geochemistry: the mantle, 2nd edn.
Elsevier, New York, pp 559–576
McDonough WF (2016) The composition of the lower mantle and core.
In: Terasaki H, Fischer R (eds) Deep Earth: physics and chemistry of
the lower mantle and core. AGU-Wiley, Washington, DC,
pp 145–159
McDonough WF, Sun SS (1995) The Composition of the Earth. Chem
Geol 120:223–253
Nakajima M, Stevenson DJ (2015) Melting and mixing states of the
Earth’s mantle after the Moon-forming impact. Earth Planet Sci Lett
427:286–295
O’Rourke JG, Stevenson DJ (2016) Powering Earth’s dynamo with
magnesium precipitation from the core. Nature 529:387–389
Paul D, White WM, Turcotte DL (2003) Constraints on the Th-232/U-
238 ratio (kappa) of the continental crust. Geochem Geophys
Geosyst 4
Poirier JP (1994) Light-elements in the Earth’s outer core: a critical-
review. Phys Earth Planet Inter 85:319–337
429
Ringwood AE (1984) The Bakerian lecture, 1983: the Earth’s core – its
composition, formation and bearing upon the origin of the Earth.
P Roy Soc Lond a Mat 395:1–46
Rubie DC, Frost DJ, Mann U, Asahara Y, Nimmo F, Tsuno K, Kegler P,
Holzheid A, Palme H (2011) Heterogeneous accretion, composition
and core-mantle differentiation of the Earth. Earth Planet Sci Lett
301:31–42
Rubie DC, Jacobson SA, Morbidelli A, O’Brien DP, Young ED,
de Vries J, Nimmo F, Palme H, Frost DJ (2015) Accretion and
differentiation of the terrestrial planets with implications for the
compositions of early-formed Solar System bodies and accretion of
water. Icarus 248:89–108
Sakamaki T, Ohtani E, Fukui H, Kamada S, Takahashi S, Sakairi T,
Takahata A, Sakai T, Tsutsui S, Ishikawa D, Shiraishi R, Seto Y, E
Tsuchiya T, Baron AQR (2016) Constraints on Earth’s inner core
composition inferred from measurements of the sound velocity of
hcp-iron in extreme conditions. Sci Adv Vol. 2, no. 2, e1500802.
https://doi.org/10.1126/sciadv.1500802
Sargent BA, Forrest WJ, Tayrien C, McClure MK, Li A, Basu AR,
Manoj P, Watson DM, Bohac CJ, Furlan E, Kim KH, Green JD,
Sloan GC (2009) Silica in protoplanetary disks. Astrophys
J 690:1193–1207
Sze EKM, van der Hilst RD (2003) Core mantle boundary topography
from short period PcP, PKP, and PKKP data. Phys Earth Planet Inter
135:27–46
Thibault Y, Walter MJ (1995) The influence of pressure and tem-
perature on the metal-silicate partition-coefficients of nickel and
cobalt in a model-C1 chondrite and implications for metal seg-
regation in a deep magma ocean. Geochim Cosmochim Acta 59:
991–1002
Tsuchiya T, Kawai K, Wang X, Ichikawa H, Dekura H (2016) Temper-
ature of the Lower Mantle and Core based on ab initio mineral
physics data. In: Terasaki H, Fischer R (eds) Deep Earth: physics
and chemistry of the lower mantle and core. American Geophysical
Union, Washington, DC, pp 13–30
van Boekel R, Min M, Leinert C, Waters LB, Richichi A, Chesneau O,
Dominik C, Jaffe W, Dutrey A, Graser U, Henning T, de Jong J,
Kohler R, de Koter A, Lopez B, Malbet F, Morel S, Paresce F,
Perrin G, Preibisch T, Przygodda F, Scholler M, Wittkowski M (2004)
The building blocks of planets within the “terrestrial” region of proto-
planetary disks. Nature 432:479–482
Walker RJ (2016) Siderophile elements in tracing planetary formation
and evolution. Geochem Perspect 5:1–142
Wasson JT, Kallemeyn GW (1988) Compositions of chondrites. Philos
Trans R Soc A Math Phys Eng Sci 325:535–544
Watanabe K, Ohtani E, Kamada S, Sakamaki T, Miyahara M, Ito Y
(2014) The abundance of potassium in the Earth’s core. Phys Earth
Planet Inter 237:65–72
Wheeler KT, Walker D, Fei YW, Minarik WG, McDonough WF
(2006) Experimental partitioning of uranium between liquid iron
sulfide and liquid silicate: Implications for radioactivity in the Earth’s
core. Geochim Cosmochim Acta 70:1537–1547
Williams Q, Hemley RJ (2001) Hydrogen in the deep Earth. Annu Rev
Earth Planet Sci 29:365–418
Wood BJ (1993) Carbon in the core Earth planet. Sci Lett 117:
593–607
Yin QZ, Jacobsen SB, Yamashita K, Blichert-Toft J, Telouk P,
Albarede F (2002) A short timescale for terrestrial planet
formation from Hf-W chronometry of meteorites. Nature 418:
949–952
Yoder CF (1995) Astrometric and geodetic properties of the Earth and
the solar system. In: Ahrens TJ (ed) Global Earth Physics. American
Geophysical Union, Washington, DC, pp 1–31
Young ED, Kohl IE, Warren PH, Rubie DC, Jacobson SA, Morbidelli A
(2016) Oxygen isotopic evidence for vigorous mixing during the
Moon-forming giant impact. Science 351:493–496
430
Earth’s Oceanic Crust
Michael Perfit
Department of Geological Sciences, University of Florida,
Gainesville, FL, USA
Definition
Oceanic crust is the outermost solid layer of the lithospheric
tectonic plates under the oceans that covers much of the
Earth’s surface. It has a distinctive basaltic composition char-
acterized by rocks that have relatively low concentrations of
potassium and other highly incompatible trace elements
(those typically excluded from minerals that comprise the
mantle). Not including a sedimentary cover of variable thick-
ness and composition, the oceanic crust consists of three
layers: (1) a relatively thin uppermost volcanic layer of basal-
tic lavas known as mid-ocean ridge basalts (MORB) erupted
on the seafloor; (2) a thicker layer of more coarsely crystal-
line, intrusive basaltic dikes that feed the volcanic layer; and
(3) a thick lowermost layer of intrusive coarse-grained
gabbros that mostly represent minerals that have settled out
of slowly cooled melt bodies and lenses in a crystal mush
zone below ridge crests.
Introduction
Nearly three-quarters of Earth’s surface lies beneath the
oceans and marginal seas where most of the oceanic crust
has been formed by magmatic processes at mid-ocean ridges
(MOR) and near-ridge seamounts. However, there are some
regions where significant volumes of crust have formed
locally such as at intraplate hotspot volcanoes, back-arc
basins, and oceanic plateaus that will not be discussed here.
Oceanic crust formed at MOR is primarily basaltic in compo-
sition and thin (~3–10 km thick) compared to continental
crust that has an average thickness of 35–40 km and a roughly
andesitic composition (Taylor and McLennan 1985; Rudnick
1995). In-situ sampling of the oceanic crust has largely been
restricted to surface lava flows by dredging, drilling, remotely
operated vehicles and human-occupied submersibles. There-
fore, in order to estimate the bulk composition of the entire
oceanic crust, these surficial samples have been augmented by
dredging steep faults, several deep drill holes, investigations
of ophiolite complexes (fragments of oceanic and back-arc
crust and upper mantle that have been thrust upon to conti-
nents), and comparisons of the seismic structure of the oce-
anic crust with laboratory determinations of seismic wave
velocities in known rock types. Integration of observations
and analysis of oceanic rocks on the seafloor and exposed
Earth’s Oceanic Crust
above sea level, combined with geophysical information,
provides us with a reasonable view of Earth’s oceanic crustal
composition. Here the focus is on the composition of rocks
that comprise the crust formed at MOR with the understand-
ing that there is a range of compositions associated with other
oceanic magmatic environments (e.g., intraplate hotspots,
island arcs).
Seismic investigations of the seafloor have determined that
the thickness of oceanic crust averages about 6–7 km at fast-
and intermediate-spreading rate ridges, but typically is much
thinner at slow-spreading MOR where the crust exhibits
greater variation in thickness and is remarkably complex
compared to crust formed at magmatically robust fast-
spreading centers (Cannat 1996; Dunn and Forsyth 2007).
In general, seismic layers 2a and 2b form the top 2 km of the
oceanic crust and are comprised of extrusive basaltic lava
flows and intrusive dikes that serve as conduits that supply
the magma erupted on the seafloor. Compared to basalts
erupted in other tectonic environments, mid-ocean ridge
basalts (MORB) have rather limited compositions. However,
they are far from being a homogeneous group, reflecting their
diverse sources and melting conditions, as well as the mag-
matic processes that cause them to differentiate into different
rock types (Melson et al. 1976; Langmuir et al. 1992; Perfit
and Chadwick 1998; Arevalo and McDonough 2010; O’Neill
and Jenner 2012; Gale et al. 2013). The lower and thicker part
of the crust (~5 km thick) consists of a wide variety of coarse-
grained plutonic rocks that range in composition from ultra-
mafic cumulate rocks to a wide variety of cumulate and
isotropic gabbros and even to silica-rich plagiogranites
(Coogan 2014). Although detailed surficial sampling and
mapping has been completed at a large number of MOR and
some sampling has been carried out along steep escarpments
and rifts in the crust (tectonic windows), only very limited
direct sampling of the deep oceanic crust has been accom-
plished. Consequently, there are significant uncertainties
about the detailed internal structure and overall composition
of the crust.
Composition of the Upper Oceanic Crust: Major
Elements
The chemical characteristics of MORB provide critical infor-
mation about their origin (source) and petrologic evolution
(Langmuir et al. 1992; Rubin and Sinton 2007; Rubin et al.
2009). Deciphering the magmatic history of MORBMid-
ocean ridge basalts (MORB) suites can be relatively straight-
forward and provides information about the mantle without
the complexities that arise, for example, when magmas transit
through thick, old continental crust or where subducted sed-
iments contaminate mantle sources beneath island arcs. Slight
differences in oceanic mantle sources and the extents and
Earth’s Oceanic Crust 431

depths of melting of these sources, however, are known to
generate “parental melts” with compositions that differ within
individual ridge segments and relative to spreading rates (e.g.,
Klein and Langmuir 1987). Variations in the amounts of FeO,
Na2O (Fig. 1), and K2O (Fig. 2) at specific MgO contents
reflect the effects of source composition and/or extents and
depths of mantle melting.
Although as a global rock type MOR lavas are thought to
be relatively homogeneous, significant differences in major
element compositions also occur on both large (ridges with
different spreading rates) and small (adjacent ridge seg-
ments) spatial scales as well as over short-time intervals
(decades) (Langmuir et al. 1986; Reynolds et al. 1992; Perfit
et al. 1994; Perfit and Chadwick 1998; Smith et al. 2001;
Rubin and Sinton 2007; Rubin et al. 2009; Goss et al. 2010).
Both the most primitive (high MgO wt%) and evolved (low
MgO, high SiO2) tend to erupt at discontinuities in the linear
ridge segments such as transform faults and overlapping
spreading centers where magma mixing and recharge are
limited.
To eliminate some of the complexities of different sources
and amounts and depths of melting, Fig. 3 shows only the
compositional variation of lavas (glasses) from a limited
region of the northern East Pacific Rise. As observed in
many other ridges, MORB can range in composition from
rare picrites with high MgO contents to ferrobasalts and FeTi
(iron-titanium) basalts containing lower MgO concentrations
and high concentrations of FeO and TiO2. The pronounced E
iron-enrichment trend with decreasing MgO contents (Fig. 3
bottom) is primarily related to fractional crystallization as
magmas cool. It is this trend together with their low alkali
contents that classifies MORB as belonging to what is known
as the subalkaline “tholeiitic magmatic suite.” Hence MORB
are sometimes also known as ocean floor tholeiites. More

Earth’s Oceanic Crust,

Fig. 1 Major element diagram
showing the variation of MgO
wt% versus FeO (total) wt% (top)
and Na2O wt% (bottom) in
quenched glass from lavas from
three mid-ocean ridges that have
different spreading rates. Basalts
from more slowly spread ridges
have relatively higher Na2O and
lower FeO. Volcanic rock names
are shown adjacent to their typical
range of compositions in top
diagram. The general elemental
trends with decreasing MgO are
primarily due to fractional
crystallization for typical MORB
suites. The variations in FeO and
Na2O at any given MgO reflect
different extents and integrated
depths of melting (Klein and
Langmuir 1987; Langmuir et al.
1992; Gale et al. 2013; White and
Klein, 2015)
432
Earth’s Oceanic Crust, Fig. 2 Variations in K2O wt% versus MgO wt
% in MORB from different MOR. Differences in K2O in lavas with high
MgO (e.g., parental N-MORB vs. E-MORB) arise from differences in
the K2O in their mantle source and/or extents of melting of the source.
Potassium is an incompatible element in MOR magmatic systems and as
such also exhibits large increases in concentration with decreasing MgO
as a consequence of fractional crystallization (typical fractional
rarely, and almost exclusively, found on fast- and
intermediate-spreading rate ridges, are even more differenti-
ated, silica-enriched lavas that include low-potassium MOR
basaltic andesites, andesites (icelandites), dacites, and
rhyodacites. They have low MgO (<4 wt%) concentrations
and SiO2 contents that range from 52 wt% to more than 70
wt% (Fig. 3 top).
Composition of the Upper Oceanic Crust: Trace
Elements
MORB are similar to tholeiitic basalts erupted on oceanic
islands like Iceland, the Galápagos, and the Hawaiian Islands
in that they are composed of about 50 wt% silica (SiO2) and
lesser amounts of the oxides of magnesium, iron, calcium,
and aluminum. However, MORB differ from ocean island
basalts (OIBOcean island basalts (OIB)) and oceanic arc
basalts in their distinct trace element and isotopic chemical
characteristics (Sun et al. 1979; Sun and McDonough 1989;
White and Klein 2014). In comparison to other magmatic/
tectonic environments, MORB contain very low abundances
of K2O and trace elements such as cesium, rubidium, barium,
(large ion lithophile elements), uranium, thorium, niobium,
the LREE, and volatile elements. These elements are all
considered to be highly incompatible in typical mantle min-
eral assemblages, and their low concentrations indicate
MORB mantle was depleted in these elements by previous
melting events. The trace element abundances in MOR lavas
Earth’s Oceanic Crust
crystallization path shown as green curve). Dacite lava compositions
reflect additional enrichments from crustal assimilation coupled with
fractional crystallization (AFC) (Data from Wanless et al. 2010). Some
variability may also be due to mixing between N- and E-MORB melts
(purple dashed line) or between highly evolved, low-MgO magmas and
evolved basalts (red dashed line) (Modified from Karson et al. 2015)
relative to an accepted value of “primitive mantle” show the
differences in MORB types (Fig. 4). So-called normal
mid-ocean ridge basalts, or N-MORB, are those with very
low concentrations of highly incompatible trace elements and
consequently are characterize as “depleted” relative to basalts
from continents, island arcs, or ocean islands. On normalized
elemental abundance diagrams, N-MORB exhibit character-
istic smooth concave-down patterns reflecting their incom-
patible element depleted nature. MORB containing slightly
greater concentrations of highly incompatible elements,
including K2O, are termed “enriched” or E-MORB (Figs. 3
and 4). In some localities, samples have been recovered that
are even more depleted than N-MORB and are consequently
classified as D-MORB (Gale et al. 2013).
Differences in the absolute abundances of elements are
commonly related to the amount of crystal fractionation that
magmas experience during cooling. In contrast, the relative
abundances of elements, reflected in the slope of the elemen-
tal patterns, are generally considered to reflect source compo-
sition and/or extents of melting of different mantle sources
(or mixing between different MORB types). Relatively
enriched MORB are volumetrically minor on most normal
ridge segments but can comprise a significant proportion of
the crust around regions influenced by plume magmatism
such as the Galapagos Islands, the Azores, and Iceland (e.g.,
White and Schilling 1978; Schilling 1985; White et al., 1993).
Isotopic investigations have conclusively shown that
values of the radiogenic isotopes of strontium, neodymium,
hafnium, and lead in N-MORB are consistent with their
Earth’s Oceanic Crust
Earth’s Oceanic Crust, Fig. 3 Major element variation diagrams of
MORB glasses from 9 to 10 N on the East Pacific Rise (EPR) as an
example of the range of compositions along a fast-spread MOR (Perfit
et al. 2012). The plot shows the compositions in ~1600 lavas erupted
between 9 170 N and 10 N (red crosses) compared to those from an
Overlapping Spreading Center (OSC) at 9 N (black asterisks). The
general compositional trends in the basalt data (MgO > ~6 wt%, SiO2
< 52 wt%) are primarily due to the effects of low to moderate pressure
fractional crystallization. Highly evolved andesites and dacites, exem-
plified by those erupted at around 9 N (Wanless et al. 2010; asterisks),
are a result of extreme fractional crystallization coupled with crustal
assimilation (AFC) along a propagating limb of the OSC, as well as a
component of magma mixing between high-SiO2 magmas and relatively
evolved basaltic liquids. Also shown are the composition of lavas from
9 500 N that erupted in 1991 and again in 2005–2006 showing the
relatively restricted compositional variations that characterize individual
MOR eruptions. There are, however, distinct differences in the compo-
sitions of lavas erupted in 1991 (green diamonds) versus 2005–2006
(blue circles) that reflect magma cooling and mixing during the ~15 years
between eruptions at the same locality (Goss et al. 2010)
incompatible element-depleted characteristics and indicate
this depletion occurred via one or more episodes of partial
melting of upper mantle sources beginning more than 2 billion
years ago (Hart 1971; Zindler and Hart 1986; Hofmann 1988;
White and Klein 2014). E-MORB, on the other hand, have
433
obtained their geochemical characteristics from mantle that
has been enriched in incompatible trace elements relatively
recently in Earth history (Schilling 1985; Hofmann 1997;
Stracke 2012).
The geochemical characteristics of MORB have been used
to decipher the composition of their mantle source(s), how
much mantle melting occurred, and at what temperatures and
pressures (Langmuir et al. 1986, 1992; Putirka et al. 2011;
Stracke 2012; Gale et al. 2013). It has been clearly shown that
different MOR segments have distinctive chemical character-
istics that can be related to differences in regional mantle E
sources and melting (Schilling et al. 1983; Graham et al.
1992; Michael 1995). On a much smaller scale, detailed
geochemical investigations of lavas have documented the
dominant effects of fractional crystallization and magma
mixing on MORB geochemistry (e.g., Rhodes et al. 1979;
Perfit and Fornari 1983; Perfit et al. 1983; Langmuir et al.
1992; Frey et al. 1993; Perfit and Chadwick 1998; Wanless
et al. 2010; Rubin 2014). High melt supply at fast-spreading
ridges leads to greater degrees of differentiation and homog-
enization of parental magmas in shallow magma chambers or
lenses. At slower-spreading ridges where magma supply is
low, episodic magma accumulation in deeper, poorly
connected, and more insulated melt reservoirs produces less
differentiated and more heterogeneous lavas (Rubin and
Sinton 2007; Rubin et al. 2009).
Recent examinations of trace element systematics in
MORB and mineralogic reactions in gabbros and xenoliths
enclosed in some lavas have indicated that other magmatic
processes such as reactive porous flow, in situ crystalliza-
tion, and frequently replenished, open system magma cham-
bers likely occur in the magma chambers or crystal-melt
mush zones beneath ridge axes (Langmuir 1989; Ridley
et al. 2006; Kvassnes and Grove 2008; Lissenberg and
Dick 2008; Lissenberg et al. 2013; Coogan and O’Hara
2015). The greatest extents of chemical diversity occur at
MOR discontinuities such as environments proximal to
transform faults, propagating rift tips, and overlapping
spreading centers where highly differentiated FeTi basalts
are common and iron-rich andesites and dacites are often
found (Perfit et al. 1983; Sinton et al. 1983; Rubin and
Sinton 2007; Wanless et al. 2010). This is largely due to
extensive fractional crystallization coupled with some
crustal assimilation (AFC- assimilation fractional crystalli-
zation) in crustal magma chambers as a consequence of the
relatively cooler thermal regimes and magmatic processes
associated with ridge propagation and/or proximity to trans-
form faults.
High-resolution sampling has also documented the inher-
ent short-term and spatial variability of MOR eruptions (Perfit
et al. 1996; Smith et al. 2001; Stakes et al. 2006; Bergmanis
et al. 2007; Rubin and Sinton 2007; Goss et al. 2010). Sig-
nificant geochemical differences in closely spaced dikes from
434
Earth’s Oceanic Crust, Fig. 4 Incompatible trace element abundances
are commonly used to chemically characterize different types of MORB
and to provide information about magmatic processes and sources.
Concentrations of incompatible elements (on X-axis) plotted on this
primitive mantle-normalized diagram (Values from Sun and
McDonough 1989) show the variations in trace element “patterns” for
representative E-MORB (enriched), N-MORB (normal), and D-MORB
(depleted). Transitional varieties are sometimes referred to as T-MORB.
The degrees of enrichment (e.g., increased La/Sm) in MORB show good
correlations with K2O contents (see Fig. 2) and K/Ti ratios. E-MORB
sheeted dike complexes that presumably fed overlying lava
flows are also consistent with the short-term temporal and
spatial variability in ridge magmatic systems (Stewart et al.
2002, 2003; Pollock et al. 2009).
MOR Dikes
Dikes that comprise seismic layer 2b in the oceanic crust have
compositions and mineralogies similar, although not identi-
cal, to the overlying MORB lavas they supply. Basaltic dikes
have been recovered together with lavas at a few tectonic
windows and from a few deep-drilling sites. Where unaltered,
dikes have major element compositions that, for the most part,
overlap the composition of spatially associated lavas (Stewart
et al. 2002, 2003; Pollock et al. 2009; Sano et al. 2011). Dikes
tend to have coarser textures than lavas, as a consequence of
their intrusive nature where cooling rates can be much slower
than magma extruded on the seafloor. Ubiquitous patchy,
hydrothermal metamorphism of both the lowermost lavas in
layer 2a and the sheeted dike complex complicates direct
comparisons between the geochemistry of lavas and dikes in
some instances.
Earth’s Oceanic Crust
have strong enrichments in elements that are the most incompatible
during mantle melting (elements on far left) relative to the middle to
heavy rare earth elements and may be the result of melting enriched
mantle source or smaller extents of melting. Highly differentiated
N-MORB (low-MgO) FeTi and dacite lavas exhibit overall enrichments
of incompatible elements but also some significant elemental depletions
(e.g., Sr, P, Eu, Ti) due to their compatibility in minerals crystallizing out
of more evolved magmas. All samples shown are from the 8 –10
N segment of the EPREast Pacific Rise (EPR) including the D-MORB
from the Siqueiros Transform (From Karson et al. 2015)
Gabbroic Layer
The middle to lower portions of the oceanic crust (seismic
layer 3) are composed of a complex and varied assemblage of
coarse-grained, gabbroic, and ultramafic plutonic rocks.
These materials primarily formed by slow cooling and crys-
tallization of melts stored in mushy magma chambers or melt
sills (mixtures of crystals and melt) located at depths of
~1–4 km below the seafloor. Unlike crystal-poor volcanic
lavas on the seafloor, and the dikes that feed them, these
plutonic rocks are comprised of crystalline minerals that
reflect the physical and chemical conditions under which
they formed. Mafic and ultramafic plutonic rocks constitute
the bulk of the oceanic crust (>60% by volume). Most of the
gabbroic rocks have mineralogies dominated by plagioclase,
olivine, pyroxenes, and minor spinel (Fig. 5). In more evolved
rock types, hornblende and Fe-Ti oxides are more prevalent
and olivine and spinel less so.
Various magmatic processes that occur in crustal magma
chambers such as fractional crystallization coupled with crys-
tal settling and accumulation form cumulate plutonic rock
types, commonly exhibiting igneous layering, such as dunites
rich in accumulated olivine, anorthosites rich in plagioclase,
Earth’s Oceanic Crust 435

Earth’s Oceanic Crust,

Fig. 5 Major element variations
in gabbros, from various ocean
floor localities compared to lavas
and dikes from Hess Deep and
Pito Deep rifts in the southern
Pacific. Overall, gabbros exhibit a
much wider compositional range
than lavas and dikes due to
accumulation of crystals during
their formation. MgO-rich olivine
gabbros and troctolites have trends
that project toward magnesian
olivine (Ol) compositions (green E
field) whereas high Al2O3, low
MgO gabbros project toward
calcic plagioclase (Plag)
compositions (yellow field). Some
more evolved upper crustal
gabbros (e.g., IODP Hole
894 from Hess Deep and IODP
Hole 1256 from the EPR) have
compositions similar to lavas and
dikes from Hess Deep (red
outlined field). These more
evolved gabbros, including
ferrogabbros, follow the same
differentiation trends observed in
the lavas and dikes (blue arrow)
due to low-pressure fractional
crystallization (Modified from
Karson et al. 2015)

and pyroxenites rich in pyroxenes. Various combinations of
these mineral phases form a range of rock types including
troctolite, olivine-gabbro, gabbronorite, and ferrogabbros,
collectively referred to as gabbroic rocks. Rare silica- and
plagioclase-rich intrusive bodies, commonly found in the
upper parts of layer 3, are collectively known as
plagiogranites.
On the seafloor, plutonic rocks are only exposed on slow-
and ultraslow-spread ridges where major faulting has
removed the overlying upper crustal units exposing core
complexes or where deeply rifted crustal sections provide
tectonic escarpments. Along segments of the melt-poor,
ultraslow-spread Southwest Indian and Gakkel ridges,
serpentinized peridotites (altered mantle rocks) are exposed
directly on the seafloor with minor amounts of gabbro (Dick
et al. 2003; Cannat et al. 2006). Only the uppermost portions
of the plutonic section, away from tectonic windows have
been penetrated and sampled by crustal drilling (Coogan et al.
2002; Coogan 2014; Karson et al. 2015). Despite their abun-
dance and importance in constructing oceanic crust, the
paucity of exposures and difficulties in recovering gabbros
in situ on the seafloor limit our understanding of their origins
and overall compositions.
Gabbroic rocks recovered from the oceans display a much
greater compositional variation than their spatially related
dikes and lavas (Fig. 5). Whereas the major element trends
in the lava and dike compositions primarily reflect the differ-
entiation and evolution of primitive liquids via crystal frac-
tionation and magma mixing, the spread in gabbro data
mainly reflects crystal accumulation within MORMid-ocean
ridges (MOR) magma chambers and crystal mush zones as
well as the addition of melts that permeate and react with the
cumulate minerals. For example, elemental trends toward
high Al2O3 and low FeO and MgO contents in plutonic
rocks reflect the accumulation of calcic plagioclase in many
of the troctolites, gabbros, and anorthositic gabbros whereas
more mafic samples with high MgO contents primarily show
the effects of magnesian olivine accumulation (Fig. 5).
Gabbroic rocks recovered from the seafloor exhibit a large
range of trace element abundances, particularly the highly
436 Earth’s Oceanic Crust

Earth’s Oceanic Crust,

Fig. 6 Normalized trace element
concentration diagram showing
the relative abundances of
incompatible trace elements in
gabbros (lower plot) from the
Atlantis Massif (IODP Site
U1309) relative to spatially
associated MORB from the MAR
(upper diagram). The primitive
mantle-normalized values are
mean values for different types of
gabbroic rocks corresponding to
the range of compositions shown
in Fig. 5. Excluding the highly
differentiated oxide gabbro and
leucocratic dike, the gabbroic
rocks are unlike MORB,
characterized by variable but
overall low abundances of
incompatible trace elements and
anomalously high concentrations
of Sr and Eu relative to the REE
indicative of their cumulate
nature. Incompatible trace element
pattern for a typical MORB (red)
from the Mid-Atlantic Ridge is
shown relative to the gabbros
(lower plot) highlighting the
differences between intrusive
rocks and erupted basalts (From
Godard et al. 2009)

incompatible elements that correspond to their varied major
element compositions (Fig. 6). The trace element abundances
partly reflect the concentrations in parental melts but are also
largely dependent on the proportion of cumulus phases
(including accessory phases) and intercumulus melts.
Chondrite-normalized REE patterns vary from very LREE-
depleted troctolites to slightly LREE-enriched ferrogabbros
and plagiogranites (e.g., Dick et al. 2000; Coogan et al. 2002;
Godard et al. 2009; Gillis et al. 2013; Coogan 2014). Relative
to primitive mantle concentrations, some cumulate gabbros
have low incompatible element and REE abundances with
characteristic positive Sr- and Eu-anomalies – indicative of
plagioclase accumulation (Fig. 6). The most evolved ferro-
gabbros exhibit elevated incompatible element contents and
negative Sr- and Eu-anomalies indicative of plagioclase
removal from evolving magmas, consistent with their rela-
tively low Al2O3 contents. Relatively few gabbros have trace
element patterns that mimic those in dikes and lavas. Because
the compositions of gabbroic rocks reflect the variable pro-
portions of accumulated phases and different amounts of
interstitial melt, they cannot be directly related to mantle-
derived melts. However, detailed mineralogic studies suggest
many gabbros have experienced complex crystal-melt reac-
tions within crystal mush zones in the lower crust (Natland
and Dick 2001; Gao et al. 2007; Suhr et al. 2008; Lissenberg
et al. 2013; Coogan 2014).
Bulk Composition of the Oceanic Crust
Because oceanic crust has been incompletely sampled and is
heterogeneous on a small scale, it is difficult to determine its
bulk chemical composition with confidence. However some
estimates have recently been made of the compositions of the
volcanic and plutonic layers (Arevalo and McDonough 2010;
Gale et al. 2013; Coogan 2014; White and Klein 2014) that,
together with their relative volumes, allow some constraints
to be placed on the bulk unaltered composition of the oceanic
crust. However, one needs to be aware that it is uncertain what
proportion of the oceanic crust has been chemically modified
by the pervasive hydrothermal alteration associated with
submarine volcanism focused at the ridge crest and
low-temperature submarine weathering that occurs as crust
moves away from the magmatic portion of ridges (Ridley
et al. 1994; Alt 1995; Alt 2004; Staudigel 2013). Basically,
the interaction of seawater with oceanic igneous rocks in
low-temperature (<100  C), off-axis hydrothermal recharge
environments results in increased oxidation and alkali fixa-
tion, an increase in volatiles and some large ion lithophile
elements (e.g., Rb, K, Cs), and a depletion of Ca, Sr, S that is
sometimes associated with depletions in Mg, Si, Co, and Ni
(Alt 1995). In deeper regions of the crust where temperatures
are higher (>150  C) and fluid flow may be more restricted,
there are significant gains in Na, Mg, H2O, and CO2 and
Earth’s Oceanic Crust
losses in S, Ca, Si, alkalis, and Au. Where temperatures are
even higher (>350  C), which typically occurs in focused or
diffuse reaction zones in the dikes and upper gabbros beneath
hydrothermal vents and stockwork zones, significant deple-
tion of liable trace metals (e.g., Cu, Zn, Co, Fe) occurs along
with the stripping of S and alkalis (Alt 1995). A study of an in
situ section of crust drilled at IODP hole 1256D showed
significant depletions (> ~8% relative to fresh MORB) in
Au, As, Se, S, Cu, Zn, and Pb in the lower portions of the
dike layer and uppermost gabbroic samples. However, the
same elements are commonly enriched in altered rocks at
higher levels in the crust beneath zones of high-temperature
venting (Patten et al. 2016). In sections of the upper crust
dominated by hydrothermal fluid-rock interactions, major
element enrichments and depletions are quite variable and
difficult to generalize (Ridley et al. 1994). Overall, the extent
of alteration and elemental exchange in the oceanic crust can
be highly variable depending on the temperature of the circu-
lating seawater or hydrothermal fluids and the water to rock
ratio. Although alteration of the upper portions of the crust
clearly plays an important role in determining the bulk com-
position of the oceanic crust, the overall effects and volume of
crust affected are uncertain.
Summary
The composition of the oceanic crust is roughly basaltic;
however, an accurate estimate of the bulk composition is
difficult to make for many of the reasons discussed above
(also see Rubin et al. 2009). Reasonable estimates of the
composition of the bulk, unaltered oceanic crust might be
made by using two methods. First, one can take a mass-
balance approach by determining the compositions and volu-
metric proportions of the various rocks that comprise the crust
but the current state of our knowledge of the lower crust
precludes any precision.
Alternatively, one can estimate the composition of pri-
mary, mantle-derived magmas that crystallized and differen-
tiated to form the different crustal lithologies observed.
Current estimates of primary MOR mantle melts are still
somewhat speculative and debatable (Kinzler and Grove
1993). High-MgO lavas that have compositions comparable
to some high-pressure melting experiments of likely mantle
source rocks include rare picritic lavas and magnesian glasses
that have been recovered from a few ocean floor locales
(Elthon 1979; van Heerden and le Roex 1988; Natland
1989; Perfit et al. 1996; Wendt et al. 1999). MgO contents
for those rocks range from ~10 wt% to over 15 wt% and could
have been primary or near-primary mantle melts. But are they
representative of the composition of the oceanic crust as a
whole? The range of their observed compositions likely
reflects variations in source composition, extents of melting,
437
types of melting and minor effects of crustal differentiation.
Consequently, the estimates are not very precise and the bulk
composition of the oceanic crust remains controversial and a
focus of current investigations.
Cross-References
▶ Fractional Crystallization and Assimilation
▶ Hydrothermal Vents
▶ Incompatible Elements E
▶ Lanthanide Rare Earths
▶ Large-Ion Lithophile Elements
▶ Lithophile Elements
▶ Mantle Geochemistry
▶ Mid-Ocean Ridge Basalts (MORB)
▶ Mineralogy
▶ Volcanism
References
Alt JC (1995) Subseafloor processes in mid-ocean ridge hydrothermal
systems. In: Humphris SE, Zierenberg RA, Mullineaux LS, Thomp-
son RE (eds) Seafloor hydrothermal systems: physical, chemical,
biological, and geological interactions, Geophysical monograph,
vol 91. American Geophysical Union, Washington, DC, pp 85–114
Alt JC (2004) Alteration of the upper oceanic crust: mineralogy, chem-
istry and processes. In: Davis EE, Elderfield H (eds) Hydrogeology
of the oceanic lithosphere. Cambridge University Press, Cambridge,
pp 495–533
Arevalo R, McDonough WF (2010) Chemical variations and regional
diversity observed in MORB. Chem Geol 271(1):70–85
Bergmanis EC, Sinton JM, Rubin KH (2007) Eruptive history and
magma reservoir dynamics on the East Pacific rise at 17 300 N.
Geochem Geophys Geosyst 8(12). https://doi.org/10.1029/
2007GC001742
Cannat M (1996) How thick is the magmatic crust at slow spreading
oceanic ridges? J Geophys Res 101:2847–2857
Cannat M, Sauter D, Mendel V, Ruellan E, Okino K, Escartin J, Baala M
(2006) Modes of seafloor generation at a melt-poor ultraslow-spread-
ing ridge. Geology 34:605–608
Coogan LA (2014) The lower oceanic crust. In: Holland H, Turekian K
(eds) Treatise on geochemistry, Reference module in earth systems
and environmental sciences, Vol. 4: the crust, 2nd edn. Elsevier,
Oxford, pp 497–541. https://doi.org/10.1016/B978-0-08-095975-
7.00316-8
Coogan LA, O’Hara MJ (2015) MORB differentiation: in situ crystalli-
zation in replenished-tapped magma chambers. Geochim
Cosmochim Acta 158:147–161
Coogan LA, Gillis KM, MacLeod CJ, Thompson GM, Hékinian R
(2002) Petrology and geochemistry of the lower oceanic crust formed
at the East Pacific rise and exposed at Hess deep: a synthesis and new
results. Geochem Geophys Geosyst. https://doi.org/10.1029/2001/
GC000230
Dick HJB, Natland JH, Alt JC et al (2000) A long in situ section of the
lower oceanic crust: results of ODP Leg 176 drilling at the Southwest
Indian ridge. Earth Planet Sci Lett 179:31–51
Dick HJB, Lin J, Schouten H (2003) An ultraslow-spreading class of
ocean ridge. Nature 426:405–412
438
Dunn RA, Forsyth DW (2007) Crust and lithospheric structure – seismic
structure of mid-ocean ridges. Treatise Geophys 1:419–443
Elthon D (1979) High magnesia liquids as the parental magma for ocean
floor basalts. Nature 278:514–518
Frey FA, Walker N, Stakes D, Hart SR, Nielsen R (1993) Geochemical
characteristics of basaltic glasses from the AMAR and FAMOUS
axial valleys, Mid-Atlantic Ridge (36 –37 N): petrogenetic impli-
cations. Earth Planet Sci Lett 115(1):117–136
Gale A, Dalton CA, Langmuir CH, Su Y, Schilling J-G (2013) The mean
composition of ocean ridge basalts. Geochem Geophys Geosyst
14:489–518. https://doi.org/10.1029/2012GC004334
Gao Y, Hoefs J, Hellebrand E, Von Der Handt A, Snow JE (2007) Trace
element zoning in pyroxenes from ODP hole 735B gabbros: diffu-
sive exchange or synkinematic crystal fractionation? Contrib Min-
eral Petrol 152:429–442. https://doi.org/10.1007/s00410-006-
0158-4
Gillis KM, Snow JE, Klaus A et al (2013) Primitive layered gabbros from
fast-spreading lower oceanic crust. Nature. https://doi.org/10.1038/
nature12778
Godard M, Awaji S, Hansen H-E et al (2009) Geochemistry of a long in-
situ section of intrusive slow-spread lithosphere: results from IODP
site U1309 (Atlantis Massif, 30 N Mid-Atlantic Ridge). Earth Planet
Sci Lett 279:110–122
Goss AR, Perfit MR, Ridley WI et al (2010) Geochemistry of lavas from
the 2005–2006 eruption at the East Pacific Rise, 9 460 N–9 560 N:
implications for ridge crest plumbing and decadal changes in magma
chamber compositions. Geochem Geophys Geosyst 11:Q05T09.
https://doi.org/10.1029/2009GC00297
Graham DW, Jenkins WJ, Schilling JG, Thompson G, Kurz MD,
Humphris SE (1992) Helium isotope geochemistry of mid-ocean
ridge basalts from the South Atlantic. Earth Planet Sci Lett
110(1):133–147
Hart SR (1971) The geochemistry of basaltic rocks. In: Carnegie Insti-
tution of Washington yearbook, vol 70. Carnegie Institution, Wash-
ington, DC, pp 353–355
Hofmann AW (1988) Chemical differentiation of the Earth: the relation-
ship between mantle, continental crust and oceanic crust. Earth
Planet Sci Lett 90:297–314
Hofmann AW (1997) Mantle geochemistry: the message from oceanic
volcanism. Nature 385(6613):219–229
Karson JA, Kelley DS, Fornari DJ, Perfit MR, Shank TM (2015) Dis-
covering the deep: a photographic atlas of the seafloor and ocean
crust. Cambridge University Press, Cambridge, UK. 430 p
Kinzler R, Grove T (1993) Corrections and further discussion of the
primary magmas of mid-ocean ridge basalts, 1 and 2. J Geophys Res
98:22339–22347
Klein EM, Langmuir CH (1987) Global correlations of ocean ridge
basalt chemistry with axial depth and crustal thickness. J Geophys
Res 92:8089–8115
Kvassnes AJ, Grove TL (2008) How partial melts of mafic lower crust
affect ascending magmas at oceanic ridges. Contrib Mineral Petrol
156(1):49–71
Langmuir CH (1989) Geochemical consequences of in situ crystalliza-
tion. Nature 340(6230):199–205
Langmuir CH, Bender JF, Batiza R (1986) Petrologic and tectonic
segmentation of the East Pacific Rise, 5 300 –14 300 N. Nature
322:422–426
Langmuir CH, Klein EM, Plank T (1992) Petrological systematics of
mid-ocean ridge basalts: constraints on melt generation beneath
ocean ridges. In: Phipps-Morgan J, Blackman DK, Sinton J (eds)
Mantle flow and melt generation at Mid-Ocean ridges, Geophysical
monograph series, vol 71. American Geophysical Union, Washing-
ton, DC, pp 183–280
Lissenberg CJ, Dick HJB (2008) Melt–rock reaction in the lower oceanic
crust and its implications for the genesis of mid-ocean ridge basalt.
Earth Planet Sci Lett 271(1):311–325
Earth’s Oceanic Crust
Lissenberg CJ, MacLeod CJ, Howard KA, Godard M (2013) Pervasive
reactive melt migration through fast-spreading lower oceanic crust
(Hess Deep, equatorial Pacific Ocean). Earth Planet Sci Lett
361:436–447
Melson WG, Vallier TL, Wright TL, Byerley G, Nelen J (1976) Chem-
ical diversity of abyssal volcanic glass erupted along Pacific, Atlantic
and Indian Ocean Sea-floor spreading centers. In: Sutton GH,
Manghnani MH, Moberly R (eds) The geophysics of the Pacific
Ocean basin and its margins. American Geophysical Union, Wash-
ington, DC, pp 351–367
Michael PJ (1995) Regionally distinctive sources of depleted
MORB: evidence from trace elements and H2O. Earth Planet Sci
Lett 131(3):301–320
Natland JH (1989) Partial melting of lithologically heterogeneous man-
tle: inferences from crystallization histories of magnesian abyssal
tholeiites from the Siqueiros fracture zone. In: Saunders AD, Norry
MJ (eds) Magmatism in the ocean basins, Geological Society of
London special publication, vol 42. Geological Society of London,
London, pp 41–70
Natland JH, Dick HJB (2001) Formation of the lower ocean crust and the
crystallization of gabbroic cumulates at a very slowly spreading
ridge. J Volcanol Geotherm Res 110(3–4):191–233
O’Neill HSC, Jenner FE (2012) The global pattern of trace-element
distributions in ocean floor basalts. Nature 491(7426):698–704
Patten CGC, Pitcairn IK, Teagle DAH, Harris M (2016) Mobility of Au
and related elements during the hydrothermal alteration of the oce-
anic crust: implications for the sources of metals in VMS deposits.
Mineral Deposita 51:179–200
Perfit MR, Chadwick WW Jr (1998) Magmatism at mid-ocean ridges:
constraints from volcanological and geochemical investigations. In:
Buck WR, Delaney PT, Karson JA, Lagabrielle Y (eds) Faulting and
magmatism at mid-ocean ridges. American Geophysical Union,
Washington, DC, pp 59–115
Perfit MR, Fornari DJ (1983) Geochemical studies of abyssal lavas
recovered by DSRV ALVIN from the eastern Galapagos rift – Inca
transform – Ecuador rift: II. Phase chemistry and crystallization
history. J Geophys Res 88:10530–10550
Perfit MR, Fornari DJ, Malahoff A, Embley RW (1983) Geochemical
studies of abyssal lavas recovered by DSRVALVIN from the eastern
Galapagos rift – Inca transform – Ecuador rift: III. J Geophys Res
88:10551–10572
Perfit MR, Fornari DJ, Smith MC, Bender JF, Langmuir CH, Haymon
RM (1994) Small-scale spatial and temporal variations in mid-ocean
ridge crest magmatic processes. Geology 22:375–379
Perfit MR, Fornari DJ, Ridley WI, Kirk PA, Casey J, Kastens KA,
Edwards M, Shuster R, Paradis S (1996) Recent volcanism in the
Siqueiros transform fault: picritic basalts and implications for MORB
magma genesis. Earth Planet Sci Lett 141:91–108
Perfit MR, Wanless VD, Ridley WI, Klein E, Smith MC, Hinds JS,
Kutza S, Fornari DJ (2012) EPR ISS lava geochemistry – 2
decades of comprehensive sampling and 2 eruptions. Oceanography
25(1):89–93
Pollock MA, Klein EM, Karson JA, Coleman DS (2009) Compositions
of dikes and lavas from the Pito deep rift: implications for accretion at
superfast spreading centers. J Geophys Res 114(B3). https://doi.org/
10.1029/2007JB005436
Putirka K, Ryerson FJ, Perfit M, Ridley WI (2011) Mineralogy and
composition of the oceanic mantle. J Petrol 52(2):279–313
Reynolds JR, Langmuir CH, Bender JF, Kastens KA, Ryan WB (1992)
Spatial and temporal variability in the geochemistry of basalts from
the East Pacific rise. Nature 359(6395):493–499
Rhodes JM, Dungan MA, Blanchard DP, Long PE (1979) Magma
mixing at mid-ocean ridges: evidence from basalts drilled near 22
N on the Mid-Atlantic Ridge. Tectonophysics 55(1–2):35–61
Ridley WI, Perfit MR, Jonasson IR, Smith MF (1994) Chemical path-
ways of alteration in oceanic volcanics: a detailed study at the
Electron Probe Microanalysis (EPMA)
Galapagos fossil hydrothermal field. Geochim Cosmochim Acta
58:2477–2494
Ridley WI, Perfit MR, Fornari DJ, Smith MC (2006) Magmatic
processes in developing oceanic crust revealed in a cumulate
xenolith collected at the East Pacific Rise, 9 500 N. Geochem
Geophys Geosyst 7:Q12O04. https://doi.org/10.1029/2006GC00
1316
Rubin KH (2014) Mid-Ocean ridge magmatism and volcanism. In: Harff
J, Meschede M, Petersen S, Thiede J (eds) Encyclopedia of marine
geosciences. Springer, Dordrecht, pp 1–21. https://doi.org/10.1007/
978-94-007-6644-0_28-3
Rubin KH, Sinton JM (2007) Inferences on mid-ocean ridge thermal and
magmatic structure from MORB compositions. Earth Planet Sci Lett
260:257–276
Rubin KH, Sinton JM, Maclennan J, Hellebrand E (2009) Magmatic
filtering of mantle compositions at mid-ocean-ridge volcanoes. Nat
Geosci 2(5):321–328
Rudnick RL (1995) Making continental crust. Nature 378:571–577
Sano T, Sakuyama T, Ingle S, Rodriguez S, Yamasaki T (2011)
Petrological relationships among lavas, dikes, and gabbros from
Integrated Ocean drilling program hole 1256D: insight into the
magma plumbing system beneath the East Pacific Rise. Geochem
Geophys Geosyst 12(6):Q06013. https://doi.org/10.1029/
2011GC003548
Schilling J-G (1985) Upper mantle heterogeneities and dynamics. Nature
314:62–67
Schilling J-G, Zajac M, Evans R et al (1983) Petrologic and geochemical
variations along the Mid-Atlantic Ridge from 29 N to 73 N. Am J
Sci 283:510–586
Sinton JM, Wilson DS, Christie DM, Hey RN, Delaney JR (1983)
Petrologic consequences of rift propagation on oceanic spreading
ridges. Earth Planet Sci Lett 62:193–207
Smith MC, Perfit MR, Fornari DJ et al (2001) Magmatic processes and
segmentation at a fast spreading mid-ocean ridge: detailed investiga-
tion of an axial discontinuity on the East Pacific Rise crest at 9 370 N.
Geochem Geophys Geosyst 2:1–32. https://doi.org/10.1029/
2000GC000134
Stakes DS, Perfit MR, Tivey MA et al (2006) The cleft revealed:
geologic, magnetic and morphologic evidence for construction
of upper oceanic crust along the southern Juan de Fuca Ridge.
Geochem Geophys Geosyst 7:Q04003. https://doi.org/10.1029/
2005GC001038
Staudigel H (2013) Chemical fluxes from hydrothermal alteration of the
oceanic crust. In: Treatise of geochemistry, vol 4, Holland, pp
583–606
Stewart MS, Klein EM, Karson JA (2002) The geochemistry of dikes
and lavas from the north wall of the Hess Deep Rift: insights into the
four-dimensional character of crustal construction at fast-spreading
mid-ocean ridges. J Geophys Res 107(B9):2181. https://doi.org/
10.1029/2000JB000051
Stewart MA, Klein EM, Karson JA, Brophy JG (2003) Geochemical
relationships between dikes and lavas at the Hess Deep Rift: impli-
cations for magma eruptibility. J Geophys Res 108(B4):2184. https://
doi.org/10.1029/2001JB001622
Stracke A (2012) Earth’s heterogeneous mantle: a product of convection-
driven interaction between crust and mantle. Chem Geol 330–331:
274–299
Suhr G, Hellebrand E, Johnson K, Brunelli D (2008) Stacked gabbro
units and intervening mantle: a detailed look at a section of IODP Leg
305, Hole U1309D. Geochem Geophys Geosyst 9. https://doi.org/
10.1029/2008GC002012
Sun S-S, McDonough WF (1989) Chemical and isotopic systematics of
oceanic basalts: implications for mantle composition and processes.
In: Saunders AD, Norry MJ (eds) Magmatism in the ocean basins,
Geological Society of London special publication, vol 42. Geological
Society of London, London, pp 313–345
439
Sun S-S, Nesbitt RW, Sharaskin AY (1979) Geochemical character-
istics of mid-ocean ridge basalts. Earth Planet Sci Lett 44:
119–138
Taylor SR, McLennan SM (1985) The continental crust: its composition
and evolution. Blackwell, Oxford, UK, p 312
van Heerden LA, le Roex AP (1988) Petrogenesis of picrite and associ-
ated basalts from the southern Mid-Atlantic Ridge. Contrib Mineral
Petrol 100:47–60
Wanless VD, Perfit MR, Ridley WI, Klein EM (2010) Dacite petrogen-
esis on mid-ocean ridges: evidence for crustal melting and assimila-
tion. J Petrol 51:2377–2410
Wendt JI, Regelous M, Niu Y, Hékinian R, Collerson KD (1999) Geo-
chemistry of lavas from the Garrett transform fault: insights into
mantle heterogeneity beneath the eastern Pacific. Earth Planet Sci E
Lett 173:271–284
White WM, Klein EM (2014) Composition of the oceanic crust. In:
Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn.
Elsevier, Oxford, pp 457–496. https://doi.org/10.1016/B978-0-08-
095975-7.00315-6
White WM, Schilling J-G (1978) The nature and origin of geochemical
variation in Mid-Atlantic Ridge basalts from the Central North
Atlantic. Geochim Cosmochim Acta 42(10):1501–1516
White WM, McBirney AR, Duncan RA (1993) Petrology and geochem-
istry of the Galápagos Islands: portrait of a pathological mantle
plume. J Geophys Res 98(B11):19533–19563. https://doi.org/
10.1029/93JB02018
Zindler A, Hart S (1986) Chemical geodynamics. Annu Rev Earth Planet
Sci 14:493–571
Electron Probe Microanalysis (EPMA)
Stuart Kearns
School of Earth Sciences, University of Bristol, Bristol, UK
Definition
A microanalytical technique yielding quantitative major and
minor elemental analysis from solid materials using an elec-
tron source and X-ray spectrometers.
The technique of electron probe microanalysis (EPMA) was
developed in the 1950s, primarily the outcome of a PhD thesis
by Castaing (1951). In the following decade, it became
established as the preferred method to characterize the in situ
chemistry of very small volumes of mainly inorganic materials,
including minerals. The instrument, often referred to as the
“microprobe” or sometimes simply the “probe,” works by
firing a beam of electrons accelerated to high potential
(1–30 kV) and focused to a small convergent spot which strikes
a flat, usually polished sample. Atoms within the sample
undergo inner shell ionization resulting in characteristic X-ray
emission. These X-rays are detected and measured by X-ray
spectrometers, and their count rates are compared to known
standards to yield a primary intensity ratio – the k-ratio. As both
the incoming electron beam and outgoing X-rays are subject to
other interaction events, the k-ratio needs to be corrected for
matrix effects to yield a quantitative analysis.
440
LabFB 0.0 nm
551.1 nm
1102.3 nm
1653.4 nm
Labradorite 15kV
2204.5 nm
800 nm
-1600.0 nm -800.0 nm -0.0 nm 800.0 nm 1600.0 nm
Electron Probe Microanalysis (EPMA), Fig. 1 Monte Carlo simulation for labradorite plagioclase at 15 kVand 5 kVaccelerating voltage. Electron
interaction volume is shown (red lines display emitted backscattered electron trajectories). Generated using Casino (Hovington et al. 1997)
Electron Source
The electron gun is supplied as one of three options –a tungsten
filament, lanthanum hexaboride (LaB6), and thermal Schottky
Field Emission Gun (FEG). All three sources have been used
extensively in scanning electron microscopes (SEM). EPMA
has traditionally utilized the tungsten filament as it offers suffi-
cient longevity and good stability, critical for quantitative anal-
ysis. Technology advances now mean that both LaB6 and FEG
can also supply a very stable beam with the benefit of higher
beam brightness yielding a smaller spot size and therefore the
promise of higher resolution analysis, with the FEG offering the
highest. EPMA instruments include a scan generator and elec-
tron detectors such that standard SEM images can be acquired.
Analytical Resolution
The analytical resolution can be determined from Monte
Carlo simulation of the electron beam interaction and X-ray
emission. The volume of material excited is much greater than
that for either secondary electron (SE) or backscattered elec-
tron (BSE) yields. It is strongly dependent on both the beam
voltage and sample properties, importantly density (Fig. 1).
Typically silicate minerals will yield excitation volumes of
several cubic micrometers using a 20 kV electron beam. At
5 kV when the beam penetration is maybe only a few hundred
nanometers, submicron resolution is achievable. However, at
such low voltage, only low energy X-rays, whose critical
excitation potential is exceeded, can be detected.
X-ray Spectrometers
The microprobe usually consists of one Energy Dispersive
Spectrometer (EDS) and multiple wavelength dispersive
Electron Probe Microanalysis (EPMA)
LabFB 0.0 nm
551.1 nm
1102.3 nm
1653.4 nm
Labradorite 5kV
2204.5 nm
800 nm
-1600.0 nm -800.0 nm -0.0 nm 800.0 nm 1600.0 nm
spectrometers (WDS) around the electron column. The
EDS benefits from parallel detection of all X-rays. Contrast-
ingly each WDS has a limited wavelength range (depending
on its diffracting crystal d-spacing) but benefits from much
greater resolution and better peak to background ratio
(Fig. 2).
WDS operates on the basis of diffraction of X-ray photons
by a crystal of known d-spacing according to Bragg’s
Law – as the angle of incidence changes, it detects X-rays
of varying wavelengths in a serial fashion.
Today’s EDS consists of a silicon drift detector (SDD) with
the older Si (Li) detectors slowly being replaced. The SDD
offers much higher throughput such that it can operate at high
beam currents without excessive deadtime – a limitation of Si
(Li) detectors. As such, SDD is an ideal accompaniment to
WDS on a microprobe. Additionally, to the delight of labora-
tory managers, the SDD only requires Peltier cooling; the
constant necessity to replenish the liquid nitrogen dewers of
Si(Li) detectors is no longer needed.
A modern addition to the standard detectors is a soft X-ray
emission spectrometer (Terauchi et al. 2012) based on a diffrac-
tion grating which offers spectacular resolution improvement in
the <1 kV X-ray range. Other detectors usually associated with
the SEM, e.g., cathodoluminescence and electron backscatter
diffraction are occasionally added to microprobes.
Quantitative Analysis
The conversion of the emitted X-ray signals into a quantita-
tive analysis requires firstly measurement of a set of primary
standards containing known concentrations of each of the
elements of interest. The derived k-ratio from the unknown
is then subject to an iterative matrix correction algorithm to
take account of the effects of Rayleigh scattering, Compton
scattering, and photoelectric absorption. The relationship
Electron Probe Microanalysis (EPMA)
Electron Probe Microanalysis 1 EDS
(EPMA), Fig. 2 Comparison of
Energy dispersive (red) and
wavelength dispersive (blue) 0.8
spectra of LREEs from a monazite
Normalised intensity
recorded under identical operating
conditions 0.6
0.4
0.2
0
4 4.5
between the X-ray intensity and the elemental composition is
established using the phi-rho-z method – an integral as a
function of mass depth provided the depth distribution of
ionizations is known. Several models to predict this function
are available, e.g., the double Gaussian model, X-Phi (Merlet
1994). The calculation is performed online and yields a quan-
titative analysis with elemental concentrations expressed as
atomic or weight percents. We would expect an accuracy of
better than 2% on samples of known concentration.
As the wavelengths of X-rays increase so the uncertainty
becomes greater; absorption effects are larger for “soft” X-rays
and the correction factors increase. Consequently at low accel-
erating potentials, selection of primary standard materials as
compositionally similar to the unknowns as possible, and
measured under the same operating conditions, is required to
yield the best quantitative results (Bastin and Heijligers 2004).
Routinely in earth sciences, oxygen is assumed to have a fixed
stoichiometry and is rarely analyzed directly – EPMA cannot
readily distinguish between varying valence states.
The X-ray spectrum includes a continuum of
“bremsstrahlung” – literally a braking radiation which
forms a background of variable intensity. This must be mea-
sured and subtracted from the characteristic peak intensity.
The detection limits of EPMA may be defined as 3 standard
deviations of the background count rate. Typically, minimum
detection limits are in the range 100–500 ppm but in cases
where multiple WD spectrometers are assigned to a single
element and long count times and high beam currents are
employed, less than 10 ppm of some elements may be detect-
able (e.g., analysis of Ti in quartz, Wark and Watson 2006).
X-ray Mapping
If the beam is scanned across a sample surface, and spectrom-
eters record the X-ray intensity as the scan progresses, a map
of that X-ray, and therefore element distribution, can be
441
WDS
E
5 5.5 6 6.5
Energy (keV)
collected (Wuhrer and Moran 2014). The X-ray signal is
subject to the same matrix effects as with spot analyses, but
quantitative element maps can be produced. Using EDS, it is
now commonplace to acquire a full ED spectrum for each
pixel on an image permitting complex phase mapping. As
WDS signals are serially collected, a single element is
mapped on each spectrometer. As with quantitative analysis
the background intensity needs to be subtracted especially
when mapping multiphase samples. This can either be mea-
sured directly or estimated using a mean atomic number
calculation (Donovan and Tingle 1996).
Sample Preparation and Materials
It is imperative that samples are clean, flat, and polished and,
as most earth science materials are dielectric, are rendered
conductive by application of a coating of conductive material,
typically carbon or alternatively, gold, silver, or aluminum.
Many materials (especially those containing water or
hydroxyl) suffer from beam damage, often a result of phonon
excitation. For this reason, a range of optimal operating con-
ditions exist depending upon the task in question. For exam-
ple, rhyolitic glasses suffer from alkali migration during beam
exposure; this is mitigated by using a defocused beam and
low beam currents (2–5 nA) (Morgan and London 2005).
Time-dependent intensity calculation can be used to restore
the intensity back to that expected at t = 0 (e.g., halogen
analysis in apatites, Goldoff et al. 2012).
Conclusions
Since the 1950s, EPMA has developed into a mature tech-
nique and continues to play an important role in the chemical
characterization of minerals and glasses. Modern advances
offer the prospect of better light element analysis,
442
quantification of smaller volumes of material, and improved
X-ray mapping systems. EPMA occupies a unique analytical
niche and compliments many other techniques such as XRF,
XRD, SEM, SIMS, LA-ICPMS.
References
Bastin GF, Heijligers HJM (2004) Quantitative electron probe micro-
analysis of nonconducting specimens: science or art? Microsc Micro-
anal 10:733–738
Castaing R (1951) Application of electron probes to local chemical and
crystallographic analysis. PhD thesis, University of Paris. [English
translation by P Duwez and DB Wittry, California Institute of Tech-
nology, 1955]
Donovan JJ, Tingle TN (1996) An improved mean atomic number
background correction for quantitative microanalysis. Microsc
Microanal 2:1–7
Goldoff B, Webster JD, Harlov DE (2012) Characterization of
fluor-chlorapatites by electron probe microanalysis with a focus on
time-dependent intensity variation of halogens. Am Mineral
97:1103–1115
Hovington P, Drouin D, Gauvin R (1997) CASINO: a new era of Monte
Carlo code in C language for the electron beam interaction – Part I:
description of the program. Scanning 19:1–14
Merlet C (1994) An accurate computer program for quantitative electron
probe microanalysis. Mikrochim Acta 114/115:363–376
Morgan G, London D (2005) Effect of current density on the electron
microprobe analysis of alkali aluminosilicate glasses. Am Mineral
90:1131–1138
Terauchi M, Takahashi H, Handa N, Murano T, Koike M, Kawachi T,
Imazono T, Koeda M, Nagano T, Sasai H, Oue Y, Yonezawa Z,
Kuramoto S (2012) Ultrasoft-X-ray emission spectroscopy by using
a newly designed WDS spectrometer attached to a transmission
electron microscope. J Electron Microsc 61:1–8
Wark D, Watson EB (2006) TitaniQ; a titanium-in-quartz
geothermometer. Contrib Mineral Petrol 152:743–754
Wuhrer R, Moran K (2014) Quantitative X-ray mapping, scatter dia-
grams and the generation of correction maps to obtain more infor-
mation about your material. IOP Conf Ser Mater Sci Eng 55:012021
Electronegativity
Barry R. Bickmore and Matthew C. F. Wander
Department of Geological Sciences, Brigham Young
University, Provo, UT, USA
Definition
Electronegativity is a measure of the electron attracting power
of an atom, or group of atoms, in a molecule. Several electro-
negativity scales exist, based on different physical properties,
but all of them are highly correlated. Electronegativity is used
for a variety of purposes in chemistry, providing both an
intuitive concept with which to rationalize chemical trends
and values that are highly correlated with a number of chem-
ical and physical properties.
Electronegativity
Introduction
Electronegativity is one of the most ambiguously defined but
useful concepts commonly employed in chemistry. In 1806,
Humphry Davy posited that the chemical affinity between two
substances was related to the degree of difference in their
electrical properties, i.e., their ability to obtain and hold posi-
tive or negative charge. Avogadro expanded upon it and gave
similar suggestions in 1809, when he envisioned a relative
scale on which all simple substances could be ranked, calling
it the scale of “oxygenicity.” Over the next few years, Jacob
Berzelius constructed a similar ranking, calling it the scale of
“electronegativity.” As theories of atomic structure progressed,
the term came to mean the ability of an atom to attract and hold
electrons (Jensen 1996). Linus Pauling was the first to create a
quantitative scale of electronegativity, based on thermochemi-
cal data (Pauling 1932). Since then, a number of electronega-
tivity scales have been proposed, with precise definitions based
on various types of chemical and spectroscopic data, and the
concept has been extended to encompass groups of atoms or
even entire molecules. The fact that there is no unique defini-
tion of electronegativity has caused some to question the use-
fulness of the concept (Bergmann and Hinze 1996), but all
these scales are highly correlated, implying they all are based
on a fundamental atomic characteristic.
In this entry, we explore the origin of some of the
most commonly used electronegativity scales and their
applications.
Electronegativity Scales
Pauling (1932) created the first numerical electronegativity
scale based on the dissociation energies of single bonds. He
noted that there was considerable evidence for the proposition
that the energies of normal covalent, single bonds are addi-
tive. In other words, if there are two molecules A2 and B2,
internally held together by covalent single bonds, the disso-
ciation energy of the molecule AB (EAB) will be approxi-
mately equal to the mean of the dissociation energies of A2
and B2 (EAA and EBB). However, this is not the case when AB
is ionically bonded; instead, the dissociation energy of AB is
larger than the average of the other two by some amount DAB.
Pauling represented this using Eq. 1.
EAA þ EBB
EAB ¼ þ DAB (1)
2
Pauling constructed his electronegativity scale to follow the
definition in Eq. 2, where wA and wB represent the electroneg-
ativity values of elements A and B, respectively.
DAB ¼ ðwA  wB Þ2 (2)
Electronegativity
Thus, the difference in electronegativity values is related to
the polarity of the bond between the two atoms, with the more
electronegative atom attracting the bonding electrons to a
greater extent than the other. Pauling rationalized bond polar-
ity in terms of his valence-bond theory, in particular the
resonance between the states AB, AB+, and A+B. That is,
the relative electronegativity of the two atoms determines the
degree to which some resonance states are favored over the
others in the molecule. The calibration of the Pauling electro-
negativity scale is somewhat arbitrary, in that it is based on the
assignment of w = 4.0 for F (Malito 1995). The scale has
since been updated and expanded several times, based on
larger sets of thermochemical data (e.g., Allred 1961).
In the Pauling definition, electronegativity is conceived as
an intrinsic property of an atom, but its definition is tied to the
energies of bonds between atoms. Allred and Rochow (1958),
in contrast, developed an alternative scale in which electro-
negativity is more closely tied to a physical, though over-
simplified, picture of the atom itself. Here, electronegativity is
conceived as the electric force experienced by an electron on
the “surface” of an atom, proportional to the force of attrac-
tion between an atomic core (nucleus + core electrons) and a
valence electron in terms of a simple Coulomb’s Law calcu-
lation. Specifically, the Allred-Rochow electronegativity
(wAR) is proportional to the effective nuclear charge (Zeff),
obtained by applying Slater’s rules (Slater 1930), and
inversely proportional to the squared covalent radius (rcov),
as shown in Eq. 3, where e is the charge on an electron. The
values obtained are then scaled to be comparable to the 4-
point Pauling scale (Malito 1995).
Zeff e
wAR / (3)
r 2cov
This physical picture is oversimplified, in that rcov is a very
rough approximation of the distance of valence electrons from
the nucleus, and orbital shapes are not taken into account.
Furthermore, Allred-Rochow electronegativity values can be
problematic, due to the difficulty in obtaining accurate Zeff
and the fact that the covalent radii of transition metals lead to
lower electronegativity values than other scales.
Although the Allred-Rochow scale is based on an
oversimplified physical depiction of the atom, one of its advan-
tages is that it makes a very direct link between electronega-
tivity and its definition as the relative ability of an atom to
attract electrons. The Pauling scale, in contrast, is linked to the
definition more indirectly via bond dissociation energies.
Another scale, the Mulliken-Jaffe electronegativity scale is
similar to the Allred-Rochow scale, in that it has a direct
connection to the ability of an atom to attract electrons, but
the former has the advantage of being based on precisely
measurable quantities. Mulliken-Jaffe electronegativity values
(wMJ) are calculated by taking the arithmetic mean of an atom’s
443
first ionization energy (the energy required to remove the most
loosely bound electron from the neutral atom to make a posi-
tively charged ion) and its electron affinity (the energy released
when the neutral atom captures one extra electron to make a
negatively charged ion). This produces a value in energy units,
which is often subsequently scaled to be comparable to Pauling
electronegativity values (Malito 1995).
Several other electronegativity scales are commonly used,
but all of them are highly correlated with one another, espe-
cially among main group elements. Certainly this indicates
that all of the precise definitions used are based on highly E
related atomic and molecular characteristics, but the particular
definition employed determines how useful electronegativity
values will be for a given application.
Electronegativity and Chemical Context
If the purpose of an electronegativity scale is to provide insight
about trends in chemical reactivity, then viewing electronega-
tivity as an intrinsic property of a certain element, rather than a
property of an atom in a particular chemical environment, may
prove too inflexible for some purposes.
One obvious deficiency is that electronegativity values are
often presented as an unchanging property of a given element,
when in fact the relative attraction of an atom for electrons
should depend strongly on the oxidation state. Electronega-
tivity values for individual oxidation states of some elements
have been reported, and, at least for some definitions, there is
no a priori reason why they should not be. The main obstacle,
in these cases, is usually a paucity of available data for the less
common oxidation states (Malito 1995).
Even for a given oxidation state of an element, electroneg-
ativity should vary somewhat with chemical context, due to the
competition for electrons among all the atoms in a particular
neighborhood. For example, all else being equal, the carbon
atom in a trifluoromethyl group (R–CF3) must hold its elec-
trons more tightly, with respect to the rest of the molecule, than
that in a methyl group (R–CH3), because its association with
fluorine would decrease its polarizability. Organic chemists,
therefore, commonly assign “group electronegativity” values
based on several methods (Mullay 1987).
Finally, the assignment of electronegativity scales to the noble
gases has been particularly difficult. For those scales that attempt
it, widely varying answers are obtained. While these elements
are able to undergo chemical reactivity, their special natures
make them difficult to analyze with such a simple model.
Electronegativity Equalization
Some of these methods for calculating group electronegativity
rely on the principle of “electronegativity equalization,” which
444
has received strong support from density functional theory
(Mullay 1987). That is, as atoms come together and electron
density is redistributed toward the more electronegative ele-
ments, the electronegativity values of the atoms change with
their atomic charge, until all the atoms in a system have equal
electronegativity values. In Sanderson’s electronegativity
equalization method, for example, the final electronegativity
of all atoms is taken as the geometric mean of the atomic
electronegativity values, and atomic partial charges are
assumed to vary linearly with electronegativity (Sanderson
1983).
Sanderson was able to accurately calculate many bond
dissociation energies with this method, but it presents certain
problems for large molecules and condensed phases. If all
atoms in a molecule have the same electronegativity value,
for example, and partial charge on atoms of the same ele-
ment is linked only to their electronegativity, then it follows
that all atoms of the same element in a system must have the
same partial charge. This contradicts the argument presented
above about the behavior of methyl and trifluoromethyl
groups, at least if they exist in the same molecule. Therefore,
others have developed partial equilibration schemes to more
adequately deal with proximity effects (Mullay 1987). Sev-
eral of these schemes have been implemented to estimate
partial charges in molecular mechanics calculations
(Oliferenko et al. 2006).
Applications
The utility of the electronegativity concept is that it
condenses complex atomic characteristics in a way that
correlates with important chemical properties in an intui-
tively understandable manner. One can readily imagine
how a property like electronegativity would affect the
behavior of atoms in a molecule, so a model meant to
predict some chemical behavior based on correlation with
a quantity derived from electronegativity values often also
provides a simple, intuitive rationale for that behavior.
For instance, the distribution of charge around the atoms
in a compound would obviously affect the energies of
available orbitals, so it is no surprise that atomic charges
calculated from electronegativity equalization correlate
well with chemical shifts in X-ray photoelectron spectros-
copy data (Bergmann and Hinze 1996). Charge distribu-
tion obviously affects bond energies (Pauling 1932), so it
is no surprise that the ionic character of a bond is an
important predictor of equilibrium constants for acid dis-
sociation reactions (Li 2000; Bickmore et al. 2004). Sim-
ilar arguments can be made for the relationships between
electronegativity and stretching force constants, bond
lengths, and many other properties (Bergmann and
Hinze 1996).
Electronegativity
Summary and Conclusions
The definition of electronegativity, the power of an atom to
attract electrons, may be somewhat vague, but it has never-
theless proven to be an extremely useful concept in chemistry
and geochemistry. Certainly we have available more rigorous
theoretical constructs, like quantum mechanics, but given the
difficulty in applying rigorous theory in some contexts, sim-
pler, more intuitive constructs are warranted for creating
models of chemical behavior. Furthermore, intuitive concepts
like electronegativity are useful as a means to facilitate fluent
thinking about chemical problems (Brown 2003).
Although electronegativity is usually presented as an invari-
ant atomic property, it nevertheless depends strongly on the
chemical context. Oxidation state and the electronegativity of
surrounding atoms, for instance, both affect the attraction of a
particular atom for electrons. In fact, it has been shown that the
electronegativity values of atoms in a compound at least par-
tially equilibrate with one another, and several electronegativ-
ity equilibration schemes have proven useful for calculating
on-the-fly atomic charges for molecular simulations.
Cross-References
▶ Chemical Bonds
▶ Equilibrium Constant
References
Allred AL (1961) Electronegativity values from thermochemical data. J
Inorg Nucl Chem 17:215–221
Allred AL, Rochow EG (1958) A scale of electronetativity based on
electrostatic force. J Inorg Nucl Chem 5:264–268
Bergmann D, Hinze J (1996) Electronegativity and molecular properties.
Angew Chem Int Ed Engl 35:150–163
Bickmore BR, Tadanier CJ, Rosso KM, Monn WD, Eggett DL (2004)
Bond-valence methods for pKa prediction: critical reanalysis and a
new approach. Geochim Cosmochim Acta 68:2025–2042
Brown TL (2003) Making truth: metaphor in science. University of
Illinois Press, Urbana. 215 p
Jensen WB (1996) Electronegativity from Avogadro to Pauling. Part I.
Origins of the electronegativity concept. J Chem Educ 73:11–20
Li Y-H (2000) A compendium of geochemistry: from solar nebula to the
human brain. Princeton University Press, Princeton. 475 p
Malito J (1995) A plethora of electronegativity scales. Chim Oggi Chem
Today 13:57–64
Mullay J (1987) Estimation of atomic and group electronegativities.
Struct Bond 66:1–25
Oliferenko AA, Pisarev SA, Palyulin VA, Zefirov NS (2006) Atomic
charges via electronegativity equalization: generalizations and per-
spectives. Adv Quantum Chem 51:139–156
Pauling L (1932) The nature of the chemical bond IV. The energy of
single bonds and the relative electronegativity of atoms. J Am Chem
Soc 54:3570–3582
Sanderson RT (1983) Polar Covalence. Academic, New York. 240 p
Slater JC (1930) Atomic shielding constants. Phys Rev 36:57–64
Elements: Metalloids
Elements: Metalloids
David A. Dixon and Monica Vasiliu
Department of Chemistry, University of Alabama,
Tuscaloosa, AL, USA
Definition
The metalloids comprise a group of main group elements
found along the stair-step line that divides metals
(transition metals and other metals) and the nonmetals.
The metalloids usually include B (Group 13) in the first
row, Si (Group 14) in the second row, Ge (Group 14) and
As (Group 15) in the third row, Sb (Group 15) and Te
(Group 16) in the fourth row, and At (Group 17) in the
fifth row.
Overview
The metalloid are main group elements that are at the
border of solids that are metallic (itinerate electrons that
are shared freely among the atoms) in contrast to those
that are covalent with chemical bonds. Note that as one
proceeds down the Periodic Table, the metalloids move to
the right in the Table.
The metalloids exhibit mixed properties as they are solids
and may even look metallic but may not conduct electricity
and heat in the same way as a metal. For example, silicon
looks like a metal as it is shiny, but it is not malleable like
a metal, and is a poor conductor of heat and electricity. The
metalloids can exhibit different crystal structures. For exam-
ple, white tin is a metallic conductor with a tetragonal crystal
structure, whereas below 13  C, it is known as gray tin and
forms a powder with the diamond crystal structure and is a
poor conductor. A number of the metalloids are semiconduc-
tors as they have intermediate conductivity which is temper-
ature dependent. The ability of the metalloids to conduct
electricity and heat is far better than the nonmetals, for example,
diamond, which are insulators. As such, the metalloids play a
major role in the electronics industry, and modern society is
built around the properties of the metalloids and their
compounds.
These elements can be quite toxic and are associated with
ore formation and hydrothermal chemistry. Thus, compounds
of the metalloids play an important role in geochemistry as
their compounds make up much of the Earth’s surface, for
example, the critical role of silicon oxides in geochemical
systems. As another example, there is an important role for
boron in geochemical systems, and boron isotope fraction-
ation factors play an important role in understanding
445
paleoclimates as the equilibrium constant for the isotope
exchange reaction
10
BðOHÞ3 ðaqÞþ11 BðOHÞ 11 10 
4 ðaqÞ$ BðOHÞ3 ðaqÞþ BðOHÞ4 ðaqÞ
is the basis for a method commonly used to infer the pH of the
ancient oceans and, hence, the PCO2 of the ancient atmo-
sphere (Hemming and Hanson 1992; Spivack et al. 1993;
Sanyal et al. 1995; Rustad et al. 2010).
Silicon is the most abundant of the metalloids on Earth,
and it is the second most abundant element in the Earth’s crust E
after oxygen. Tellurium is the least abundant among the
nonradioactive metalloids, and it is one of the rarest stable
elements on Earth with an abundance similar to platinum.
Because the metalloid elements are in different columns in
the Periodic Table, the properties and reactivity of these
elements are quite different from each other due to having
different numbers of valence electrons. Boron with three
valence electrons forms trivalent molecules with an empty
2p orbital which often makes its compounds good Lewis
acids, electron pair acceptors. Boron also forms electron-
deficient bonds in compounds such as the boron hydrides.
Silicon with four valence electrons forms tetravalent com-
pounds but can expand its valency up to six, for example,
SiF62 which is important in the fluoridation of water (Jagtap
et al. 2012; Urbansky 2002). Germanium with four valence
electrons is also tetravalent and like silicon can expand its
valency up to six. Arsenic with five valence electrons forms
both trivalent and pentavalent compounds as does antimony.
The pentafluorides of As (AsF5) and Sb (SbF5) are very strong
Lewis acids (Christe et al. 2000; Dagani 2003), and SbF5 with
FSO3H is the basis of magic acid, a super Brönsted acid (Olah
et al. 2009). Tellurium can form divalent, tetravalent, and
hexavalent compounds. Little is known about the chemistry
of astatine, but as it is in the halogen column with seven
valence electrons, it would be expected to be able to form
mono-, tri-, penta-, and heptavalent compounds.
Conclusions
Metalloids have properties of both metals and nonmetals.
Metalloids are solids under standard conditions and can
appear to be metallic in appearance, but they are brittle and
generally form alloys with metals. They usually do not con-
duct heat or electricity that well in many forms. Some of the
metalloids (silicon and germanium) are semiconductors mak-
ing them useful in chips for the semiconductor industry. Other
metalloids are often used as dopants for semiconductor
manufacturing. Because of the variation in the number of
valence electrons, the chemistry of these elements can be
very different. Overall compounds containing the metalloids
are mostly nonmetallic in their properties.
446
References
Christe KO, Dixon DA, McLemore DK, Wilson WW, Sheehy J,
Boatz JA (2000) On a quantitative scale for Lewis acidity and
recent progress in polynitrogen chemistry. J Fluor Chem
101:151–153
Dagani R (2003) Feast for Fluorine Aficionados. Chem Eng News
Mar 3, pp. 44–49
Hemming NG, Hanson GN (1992) Boron isotopic composition and
concentration in modern marine carbonates. Geochim Cosmochim
Acta 56:537–543
Jagtap SM, Yenkie K, Labhsetwar N, Rayalu S (2012) Fluoride in drinking
water and defluoridation of water. Chem Rev 112:2454–2466
Olah GA, Prakash S, Molnar A, Sommer J (2009) Superacid chemistry,
2nd edn. Wiley, New York
Rustad JR, Bylaska EJ, Jackson VE, Dixon DA (2010) Calculation of
boron-isotope fractionation between B(OH)3(aq) and B(OH)4(aq).
Geochim Cosmochim Acta 74:2843–2850
Sanyal A, Hemming NG, Hanson GN, Broecker WS (1995) Evidence for
a higher pH in the glacial ocean from boron isotopes in Foraminifera.
Nature 373:234–236
Spivack AJ, You C-F, Smith HJ (1993) Foraminiferal boron isotope
ratios as a proxy for surface ocean pH over the past 21-myr. Nature
363:149–151
Urbansky ET (2002) Fate of fluorosilicate drinking water additives.
Chem Rev 102:2837–2854
Enthalpy
Carl O. Moses
Provost and Vice President for Academic Affairs, Grand
View University, Des Moines, IA, USA
Enthalpy (from the Greek root for “heat”), H, is a thermody-
namic function that provides a measure of the energy in a
system. Enthalpy is fundamentally important to geochemistry
owing to its relationship to heat capacity (see ▶ Geochemical
Thermodynamics), its role in defining and giving the temper-
ature dependence of free energy (see ▶ Free Energy), and its
relationship to the entropy of the system’s surroundings (see
▶ Entropy). The conditions under which H represents a sys-
tem’s energy are different from those under which internal
energy, U (see ▶ Geochemical Thermodynamics), represents
the energy of the same system. In fact, U is the basis for the
definition of enthalpy, H & 6; U + PV, so a brief back-
ground discussion of U is warranted.
The first law of thermodynamics (Eq. 1), in which q is heat
and w is work, is the basis for measuring energy changes.
DU ¼ q þ w (1)
Originally published in: Marshall CP, Fairbridge RW (eds) Encyclopedia
of Geochemistry, Springer Netherlands, 1999
Enthalpy
Measuring changes in U entails partitioning the work term
into P-V (pressure-volume) work (w =  P d V) and another
form of work (e.g., electrical work, we) that can be measured
to give the energy change for the process in question (Eq. 2).
ð
DU ¼ q þ we  P dV (2)
Because q cannot be measured, calorimeters are well insu-
lated so that processes are as adiabatic
Ð as possible (q ! 0). In
principle, one could measure P dV, but it is generally more
practical to make dV = 0 (i.e., a constant volume process) so
that Eq. 2 yields DU = we. Such measurements are called
constant volume calorimetry (see ▶ Calorimetry).
An alternative measurement of the energy in a system is
obtained if P is held constant. In the adiabatic limit, Eq. 2
gives, under constant pressure conditions,
DU ¼ we  PDV 
(3)
¼ we  PV f  PV i ;
where the subscripts f and i denote the final and initial states,
respectively, of the system. Because V is a state function, the
potentials PVf and PVi depend only on the states and are
independent of the means by which the system is transformed
from initial to final state. Similarly, DU = Uf  Ui, and
rearranging Eq. 3 gives (Uf + PVf) – (Ui + PVi) = we. Thus,
the constant pressure calorimetric measurement is we = D
(U + PV) or we = DH. Because the difference between con-
stant volume and constant pressure conditions is so funda-
mental, it is convenient to use U and H = U + PV to obtain
different measures of energy in the system. In general, the
enthalpy function is more relevant to geochemical processes
than internal energy because very few natural systems operate
at constant volume.
Differentiating the definition of enthalpy gives
dH ¼ dU þ P dV þ V dP: (4)
Equation 1 can be written dU = q  P dV, excluding all
forms of work besides P-V work, so Eq. 4 can be rewrit-
ten as dH = q + V dP, an alternative statement of the first
law. Another first law statement is obtained from Eq. 4
when the state function entropy, S, is introduced (see
▶ Entropy), and Eq. 1 written dU = T dS  P dV:
dH = T dS + V dP.
By convention, the enthalpy of elements in the standard
state (usually 298 K, 1 bar) is 0. For compounds or elements
in a state other than the standard state, the standard state
enthalpy of formation (from the standard state elements),
D0f, represents the energy of making the bonds of the com-
pound and, therefore, stored in those bonds. For example,
graphite, the standard state form of elemental carbon, has
Entropy
DH0f = 0, while for diamond, DH0f = 1.9 kJ/mol, indicating
that converting the planar trigonal bonds of a mole of graphite
into the tetrahedral bonds of diamond requires 1.9 kJ. Thus,
the diamond bonds are “stronger” than those of graphite by
that amount. Note that, although diamond is a polymorph of
elemental carbon, it is not the standard state form, so its DH0f
is nonzero.
Generally, a convenient notation includes a subscript on
DH0, where the superscript 0 denotes standard state, which
indicates the process (e.g., chemical reaction, phase transi-
tion) for which the enthalpy change is calculated:
X
DH 0process ¼ vi DH 0f , i ;
where the ni are the stoichiometric coefficients of the constit-
uents of the system in the final state (e.g., reaction products,
ni > 0) or the initial state (e.g., reactants, ni < 0). For exam-
ple, the evaporation of liquid water, H2O(l) = H2O(g),
involves an enthalpy change given by DH0evap = DH0f,
H2O(g)  DH0f,H 2O(l) = (241.8)(285.8) = 44 kJ/mol
at 298 K. Because DH0evap > 0, it is said to be an endother-
mic process, one in which the final state (water vapor) has
greater energy than the initial state. Equation 4 shows that, at
constant pressure, this energy must be in the form of
heat. Another example: the enthalpy change for the quartz-
fayalite-magnetite O2-buffer reaction, 3Fe2SiO4(-
fay) + O 2 (g) = 2Fe O
3 4(mag) + 3SiO 2(qtz) is given by
    Atkins PW (1986) Physical chemistry, 3rd edn. W. H. Freeman,
DH 0QFM ¼ 2DH 0f , mag þ 3DH 0f , qtz  3DH 0f , fay þ 0
¼ 529:9 kJ=mol
at 298 K. For this reaction, DH0QFM < 0, indicating an exo-
thermic process that releases energy to the surroundings
(as heat at constant pressure) and yielding a final state with
less energy than the initial state.
Like most properties of condensed phases (solids and liquids),
enthalpy has only a weak pressure dependence, but the pressure
dependence for gases can be quite strong. The differential expres-
sion of the isothermal pressure dependence of enthalpy, (@H/@P)T
can be inverted to give (@T/@P)H(@H/@T)P. At constant pres-
sure, Eq. 4 shows that dH = q or dH = dq. Because constant
pressure heat capacity is defined as CP = (@q/@T)P (see ▶ Geo-
chemical Thermodynamics), we have the isobaric (constant pres-
sure) temperature dependence of enthalpy: (@H/@T)P = CP. Thus
the isothermal pressure dependence of enthalpy becomes (@H/
@P)H CP. Finally, the isenthalpic pressure dependence of temper-
ature, (@T/@P)H, is known as the Joule-Thomson coefficient, mJT,
so we have (@H/@P)T =  mJT Cp.
Enthalpy is temperature dependent in both gases and con-
densed phases. If Cp is independent of T over the temperature
range of interest, i.e., Cp 6¼ f(T)), integration of (@H/@
T)P = Cp gives the simple relationship DH = CP DT or HT2 =
447
HT1 + CP DT. If, however, Cp = f(T), an analytical function
for the T dependence of Cp must be substituted and integrated.
Another important relationship can be derived for the
temperature dependence of enthalpy, this one for isochoric
(constant volume) temperature dependence, (@H/@T)V,
although we will not pursue the complete derivation
here. The isobaric thermal expansivity, a, is given by (1/V)
(@V/@T)P, and the isothermal compressibility, k, is given
by (1/V)(@V/@T)T. If we begin by writing a total differential
for H = f(T, P), dH = (@H/@P)T dP + (@H/@T)P dT, we can
derive (@H/@T)V = [1  (amJT/k)]Cp. E
Cross-References
▶ Activation Parameters: Energy, Enthalpy, Entropy, and
Volume
▶ Chelation
▶ Entropy
▶ Equilibrium
▶ Free Energy
▶ Geochemical Thermodynamics
▶ Kinetics of Geochemical Processes
References
Anderson GM, Crerar DA (1993) Thermodynamics in geochemistry.
The equilibrium model. Oxford University Press, New York. 588 pp
New York. 857 pp
Nordstrom DK, Munoz JL (1994) Geochemical thermodynamics, 2nd
edn. Blackwell, Cambridge. 493 pp
Robie RA, Hemingway BS, Fisher JR (1978) Thermodyna mic properties
of minerals and related substances at 298.15 K and 1 Bar (105Pascals)
pressure and at higher temperatures, US Geological Survey bulletin,
vol 1452. US Government Printing Office, Washington, DC, p 456
Entropy
Carl O. Moses
Provost and Vice President for Academic Affairs, Grand
View University, Des Moines, IA, USA
Introduction and Definition
The Greek word evtrope refers to a property that indicates
the direction of change. In thermodynamics, the entropy
function, S, provides the criteria for predicting the direction
Originally published in: Marshall CP, Fairbridge RW (eds) Encyclopedia
of Geochemistry, Springer Netherlands, 1999
448
of spontaneous natural processes and the equilibrium state of
a system. Entropy is a state function, which means that its
value is determined by the state of a system and is indepen-
dent of the process or path by which the system arrived at its
state.
Background and Framework
The need for a function such as entropy arises from the fact
that the zeroth and first laws of thermodynamics (see ▶ “Geo-
chemical Thermodynamics”) fail to explain the direction of
natural processes. The zeroth law introduces the concept of
thermal equilibrium and provides a basis for temperature
measurements, while the first law is a statement of the con-
servation of energy principle and the interconvertibility of
heat (q) and work (w). Heat exchange and work contribute
to changes in the internal energy (U) of a system. The first law
gives us a tool for meeting one of the principal objectives of
thermodynamics: keeping track of energy changes in a system
(Eq. 1).
DU ¼ q þ w (1)
A system is a region whose boundaries can be specified.
A system’s surroundings only include those entities with
which it actually interacts, and we define the sum of the
system and its surroundings as a thermodynamic universe
(not the same as the cosmologic Universe; thermodynamics
does not presume to apply on such vast scales). Accordingly,
a universe is itself an isolated system (one that does not
interact with its surroundings). We can quantify an extensive
property (one that depends on the amount of substance pre-
sent) of a universe by adding the contribution from the system
of interest and the contribution from the system’s surround-
ings. For example, S = S univ = S sys + S surr.
The Thermodynamic Concept of Entropy
As important as it is to quantify the conservation of energy
principle, the first law cannot explain why natural processes
proceed spontaneously in one direction only. For example,
why is heat spontaneously transferred from a warm object to a
colder object, and why do we never observe the cold object to
transfer some of its heat spontaneously to the warmer object?
One of the earliest statements of the second law, attributed to
Rudolph Clausius, holds that there exists no process in which
the sole result is the transfer of heat from a cooler to a hotter
body. Another early statement, attributed to William Thom-
son, Lord Kelvin, holds that there exists no process in which
the sole result is the complete conversion of heat absorbed by
a system into work. Clausius’ statement emphasizes the
Entropy
a P1 > P2
P1 P2
system surroundings
w
b T sys > T surr
T sys T surr
system surroundings
q
Entropy, Fig. 1 (a) In response to a pressure gradient, the system
transfers energy to its surroundings by expanding and, thereby, doing
work on the surroundings. The heavy line encloses a universe, and the
system is separated from its surroundings by a movable or deformable
partition. (b) In response to a temperature gradient, the system transfers
energy to its surroundings as heat. As in (a), the heavy line encloses a
universe, but here the system is separated from its surroundings by a
rigid, unmovable wall
unidirectional nature of natural spontaneous processes while
Kelvin’s emphasizes the idea that energy is degraded when it
is transferred as heat, so not all of the absorbed heat is
available for work. The interconvertibility of heat and work
is thus asymmetrical. Because energy is conserved. Kelvin’s
degradation of energy requires a quantity (a potential) into
which energy can be converted but not recovered, which is
Clausius’ unidirectional change.
Consider a mechanical analog that provides a basis for
analyzing the second law (Fig. 1a). The apparatus is enclosed
by rigid unmovable walls and is well insulated, so it is a
universe. Energy will be transferred from the system to its
surroundings as the system’s volume expands against the
lower pressure of the surroundings. The amount of work
done by the system on its surroundings is proportional to the
change in volume of the system: w/  DV. Pressure gives
the magnitude of the proportionality (higher pressure yields
greater work for a given volume change), hence w =  PDV.
Decompression may be carried out reversibly or irrevers-
ibly (Fig. 2). The irreversible process takes place by, for
example, a combination of isothermal and isobaric steps.
The changes in P, T, and V are finite, and derivatives relating
P, T, V, and U (e.g., dP/dT) are undefined. DV or DU can be
calculated from the P and T coordinates of any pair of inter-
mediate states along the path using an equation of state,
relying on the fact that V and U are state functions and DV =
Vr  Vi and DU = Uf  Ui, regardless of the sequence of
Entropy
P
i case of the pressure gradient, but now energy is transferred
f
T T
Entropy, Fig. 2 Decompression of the system in Fig. 1a, from
initial state i to final state f, in P-T coordinates. There is one reversible
path (the diagonal arrow); an irreversible path, comprising one set of the
infinite number of irreversible steps between i and f, is also shown
steps (f and i denote final and initial states, respectively).
During an isothermal decompression step, the system must
absorb heat from the surroundings (qsys > 0), while during an
isobaric cooling step, the system must transfer heat to the
surroundings (qsys < 0). For irreversible processes,
w = DU – q depends on the net heat exchange between
system and surroundings, so the work depends on the char-
acter of the path rather than the change in state of the system.
Because we can know DVj and Pj for each j step of the
irreversible process, we can calculate wj for each step, and
the total work is given by wirrev =  S Pj DVj.
If we make the isothermal and isobaric steps increasingly
smaller, DP ! dP, DT ! dT (accordingly, DV ! dV and
DU ! dU, the behavior of the system approaches reversibil-
ity. The reversible process is continuous, and derivatives like
dP/dT are defined. As a consequence of the infinitesimal size
of the steps, q ! 0. The reversible behavior of the system is
singular, because there can be only one path in which the steps
are infinitely small. Now, wrev = DU because no heat is
exchanged between system and surroundings, but w is no
less dependent on path than before because we have stipulated
a very specific path. As the steps become infinitesimal, the
summationÐ used for the total work becomes an integral, so
wrev =  fi P dV, and dwrev =  PdV. Because no heat is
exchanged, all of the energy of the system is available for work,
so wmax = wrev and for any arbitrary process, w  wrev.
Now consider a universe in which the system is separated
from its surroundings by a unmovable diathermal wall (one
that readily permits heat exchanges but does not permit work;
Fig. 1b). If heat is transferred in infinitesimal increments (dq)
by allowing the temperature of the system and its surround-
ings to differ only very slightly, the heat transfer can be made
449
reversible. This energy transfer is completely analogous to the
down the gradient as heat rather than work, entropy change is
the quantity analogous to volume change, and in place of
w/  DV, we have qsys sys
rev p DS . Because higher temperatures
yield greater heat transfer for a given entropy change,
T determines the magnitude of the proportionality between
rev =T dS .
the heat transfer and the entropy change, and dqsys sys
Whether applied to the system or to the surroundings, the
defining expression for entropy holds that entropy is the state
function whose differential is given by the reversible transfer E
of heat divided by temperature (Eq. 2).
dqrev
dS ¼ (2)
We might similarly say that volume is the state function
whose differential is given by work divided by pressure
[dV = (dw rev/P]. Both expressions emphasize the relation-
ships between energy changes as work or heat and
corresponding changes in volume (energy/pressure) or
entropy (energy/temperature). A given amount of work
(heat) creates a larger volume (entropy) change in a system
of lower pressure (temperature).
Because work and heat are both potentials, dw and dq
correspond to the potential gradients P dV and T dS, and PV
and TS are potentials. With entropy defined and work
restricted to volume changes, the first law may be written as
dU = T dS – P dV (each of U, S, and V is measured for the
system), a form that is sometimes called the first part of the
second law of thermodynamics.
The Clausius Inequality
Reversibility means that a transfer of energy between the
system and its surroundings can be reversed without changing
the state of the surroundings. At every stage of the process, the
system and its surroundings are in equilibrium. Reversibility
is an idealization, but one that can be approached very closely
in many systems often well within the limits of measurement
error; it is a pragmatically viable concept. In the reversible
heat transfer example above, the heat transferred from the
system to the surroundings was exactly balanced by the heat
gain in the surroundings. We can write dqsys rev ¼  dqrev or
surr
dqrev =T ¼  dqrev =T , so dS = dS . We conclude
sys surr sys surr
that dS univ = dS sys + dS surr = 0 for a reversible process.
What if we carry out the heat transfer irreversibly? In this
case, the process could not be reversed without causing a
definite change in the state of the surroundings. Because no
change results to the state of the surroundings in a reversible
process, the change in the state of the surroundings must be
larger for an irreversible process than for a reversible one.
450 Entropy

From the point of view of the system, therefore, dqsys irrev < d conserved, but Suniv is a very different sort of function because
qsys
rev (generally, dq sys
 dq sys
rev , just as we concluded for work it is conserved only under the special conditions of the equi-
in the previous section), which means that dqsys irrev =T < rev =
dqsys librium state; if there are internal gradients, entropy will be
sys
Tand dqirrev =T < dS . Generally, dq /T  dS ; the “<”
sys sys sys
generated!
condition applies to irreversible processes, while the “=”
condition appliesÞ to reversible processes. Þ Because S is a
state function, dS sys = 0, where the denotes a loop inte- Using the Second Law Entropy Criteria for
gral and implies a path from state A to state B to state A (i.e., a Direction and Equilibrium
cycle in which the Þfinal state is congruent with the initial
state). Accordingly, dqsys rev =T ¼ 0, but generally,
Only the sum of DS sys and DS surr must be 0, so it is possible
for dS sys 0. For a spontaneous process, any decrease in S sys
þ must be accompanied by an offsetting increase in S surr, and at
dqsys
 0: (3) equilibrium, where dS univ = 0, dS sys =  dS surr (including
T
the possibility that both dS sys and dS surr are zero). Because the
If we take a system from an initial state i to a final state f by an second law criteria for direction and equilibrium are
arbitrary process, which may be reversible or irreversible, and expressed only in terms of dS univ, it seems that we need to
then return it to i by a reversible process, we can write: know something about both dS surr and dS sys in order to use
those criteria.
þ ðf ði Because TS sys represents a potential quantity unavailable
dqsys dqsys dqsys
¼ þ rev
: for doing work, we define free energy as the net energy
T i T f T
available for doing work. With two measures of the total
energy in a system (internal energy, U, and enthalpy, H; see
The
Ð second term on the right hand side is equivalent to ▶ “Enthalpy”), there are two measures of free energy:
– fi dS sys = DS sys (switching the limits of integration
A U – TS sys and G H – TS sys (see ▶ “Free Energy”).
while changing the sign is legitimate only for a state
For our present purposes, we will concentrate on G, which is
function), so:
called Gibbs free energy. In differential form and under iso-
þ þf thermal conditions (dT = 0),
dqsys dqsys
¼  DSsys : (4)
T i T dG ¼ dH  T dSsys : (6)

By combining Eqs. 3 and 4,
ðf
dqsys
0  DSsys ;
i T
or
ðf
dqsys
DSsys  ; (5)
i T
which is the Clausius, or fundamental, inequality. In an iso-
lated system, dqsys = 0 always, and because an isolated sys-
tem is the same as a universe, the Clausius inequality leads to
DSuniv  0, which is the second part of the second law.
A spontaneous process in a universe must be irreversible (i.e.,
caused by an internal potential gradient), so DSuniv >0. For a
universe at equilibrium (no internal gradients), DS univ = 0.
The second part of the second law provides the criterion for
the direction of spontaneous processes (the direction that yields
an increase in the entropy of a universe) and the criterion for
equilibrium (the state in which the entropy of a universe
remains constant). There are no known processes that cause a
decrease in the entropy of a universe. Energy must be
Each quantity in Eq. 6 is a property of the system. Is there a
way to use free energy to determine the direction of sponta-
neous processes or equilibrium?
For an arbitrary (reversible or irreversible) isothermal pro-
cess, dS surr = dqsurr/T = dqsys/T. Summing the contribu-
tions to dS univ from the system and the surroundings and
referring to the second part of the second law,
dqsys
dSuniv ¼ dSsys   0: (7)
T
With Eq. 7 we can express dS univ in terms of quantities
measured for the system, but we need a proxy for dq sys
because we cannot measure q. Under constant pressure
conditions, dqsys = dH, so we can write dS sys – dH/T 
0, identifying dH/T with dqsurr and dS surr. With this
identity and the result of dividing Eq. 6 by T, we
have an expression that is congruent with dS univ =
dS surr + dS sys (Eq. 8; an analogous expression can be
established for A and U).
dG dH
¼ þ dS (8)
T T
Entropy 451

Equation 8, in which G, H, and S are all properties of the 0 = 0. The so-called absolute entro-
convenience, we use Ssys
system, is one of the most important equations in chemical pies of substances, Ssys 0
T , are given the symbol ST and have
thermodynamics, and it allows us to translate the second law units of energy/temperature/amount (e.g., J/K/mol).
criteria for direction and equilibrium into free energy terms: Values of S0T, based on calorimetric and heat capacity
dG/T  0. Spontaneous processes lead to minimizing measurements, are tabulated and can be used to calculate
G (dG < 0), and at equilibrium, dG = 0. standard state entropy changes, DS0R, for reactions. Entropy
changes in processes are calculated from:
X
Measuring Entropy DS0process ¼ vj DS0T , j ;

For entropy to be a useful function, we need a way to measure where the n j are the stoichiometric coefficients of the constit- E
it. From Eq. 2, a system heated from an initial temperature Ti to a uents of the system in the final state (e.g., reaction products,
final temperature Tf, must undergo an entropy change given by: nj > 0) or the initial state (e.g., reactants, nj < 0). The sub-
ð Tf script R may be substituted by a more informative subscript
dqsys denoting the specific process in question (e.g., melt or boil).
DSsys ¼ rev
:
Ti T Note that S0T is not the same as DSf 0.By convention, the DG0f
for an element in the standard state (usually 298 K, 1 bar) is
If we treat Ti as a reference temperature (Tref), we define a 0, as are DH0f and DS0f . Absolute entropy gives the entropy
scale of entropy measurements, and for any T, change for a substance from 0 K to T, so S0T > 0. Furthermore,
for a nonelement, S0T includes the entropy of formation, DS0f ,
ðT
dqsys of that substance plus the absolute entropies of its constituent
Ssys sys
T ¼ ST ref þ
rev
:
T ref T elements. For example, at 298 K, graphite, the standard state
form of elemental carbon, has DH0f = DG0f = DS0f = 0,
Given that constant pressure heat capacity Cp is defined as while for diamond, DH0f,dia = 1.9 kJ/mol and DG0f,dia = 2.9
(@q/@T)P (see ▶ “Heat Capacity”), we have: kJ/mol. The difference, 1 kJ/mol, is the potential – TDSf,dia
0
, so
DSf,dia =  3.36 J/mol/K. Note that DSf,dia = Sdia  Sgra
0 0 0 0
=
ðT (2.38  5.74) J/mol/K. Values of DSf are not usually tabu-
0
CP dT
Ssys
T ¼ Ssys
T ref þ : (9) lated. In calculating DS0R, this distinction is transparent. For
T ref T
the graphite = diamond transition,
If Cp 6¼ f(T), Eq. 9 gives Ssys
T ¼ SsysT þ Cp In(T/T ref). Over
temperature ranges for which Cp = f(T), we need at DS0R ¼ S0dia  S0gra ¼ ð2:38  5:74Þ J=mol=K
temperature-dependent expression for Cp that is integrated ¼ DS0f , dia  DS0f , gra ¼ 3:36  0 J=mol=K
in Eq. 9. If the temperature range spans a phase transition,
an entropy change DSsystrans ¼ DH trans =T trans must be added to It is important to recognize, however, that the distinction
Eq. 9. A rational choice for Tref is 0 K, so a complete expres- matters in calculations of DS0f for nonelements, which are in
sion for the entropy of a gas (i.e., T > Tboil) is: turn required with values DH0f of in calculations of DG0f .

ð T melt ð T boil
Csolid
P dT DH melt Cliquid dT
Ssys sys
T ¼ S0 þ þ þ P
T T melt T Clapeyron Equation
0
ðT T melt
DH vap Cgas
P dT
þ þ Entropy has practical utility beyond providing criteria for the
T boil T boil T
direction of spontaneous change and equilibria. If we remove
For a liquid (Tmelt < T < Tboil), the last two terms are not the isothermal constraint used in deriving Eq. 6 and substitute
needed, and the second integration is carried from Tmelt to T; dH = T dS + V dP, we have:
for a solid (T < Tmelt), the last four terms are not needed, and
the first integration is carried from 0 to T. In practice, Cp data dG ¼ V dP  S dT (10)
for the lowest temperatures (<
10 K) are extrapolated from a
function like aT 3, which is fitted to Cp data for temperatures (see ▶ “Free Energy” for more discussion of the derivation of
up to
50 K. Eq. 10). Then, if we define chemical potential, m, as
The third law of thermodynamics holds that all pure sub- the partial molar Gibbs free energy of a substance
stances have the same entropy at T = 0 K. Thus, for all pure j, m j = (dG/dn j) T,P,nk 6¼ j (the subscripts denote constant
substances, Ssys
0 , the reference entropy, is the same. For pressure, temperature, and composition; see ▶ “Free Energy”
452
and ▶ Geochemical thermodynamics), Eq. 10 can be rewrit-
0 0
ten in terms of mðP, T Þ : mj ¼ V j dP  Sj dT, where the over-
0
bar on V j and Sj denotes a molar quantity. At equilibrium, two
coexisting phases (A and B) must have the same m. Accordingly,
0 0 0 0
m A = mB, and V A dP  SA dT ¼ V B dP  SB dT .
After collecting like terms, we have:
dP DS0R
¼ (11)
dT DV 0R
Equation 12 is the Clapeyron equation, which is the basis for
deriving the equations for phase boundaries in P-T coordi-
nates (P = f(T)). Note that DS0R and DV0R are taken to be molar
quantities. No assumptions or approximations are involved in
the derivation of Eq. 12, so it is exact and applies to all phase
changes (see section “Clapeyron Equation”).
The Molecular Basis of Entropy
From the thermodynamic point of view, there is no need to
establish a molecular basis for any process; all that is neces-
sary is to account for the energy changes. From this perspec-
tive, the idea that entropy is the state function given by Eq. 2 is
sufficiently powerful and leads to many other useful relation-
ships. A somewhat deeper measure of understanding can,
however, be acquired by inquiring into the molecular basis
of entropy.
The fundamental proposition of statistical thermody-
namics is that a system contains a very large number of
molecules, the average properties of which determine the
behavior or state of the system. The goal, then, is to
calculate the average properties. The most basic property
of a molecular system is the distribution of molecules
among the system’s energy levels, which are quantized
according to the rules of quantum mechanics. The lowest
energy level is the ground state of a molecule; all other
states have higher energy. The configuration of a system
expresses how many molecules are in each state. There
are W different ways to organize N molecules into a given
configuration:
N!
W¼ ;
n0 !n1 !n2 ! . . .
where n j represents the number of molecules in state j, and
j = 0 represents the ground state. The most probable config-
uration of a system is the one with the largest W subject to the
constraints Snj = N and Snj ej = E (ej is the energy of state j)
so that the number of molecules N and the total energy E of
the system are fixed. For the most probable configuration,
W is denoted by W and each nj by nj . The Boltzmann
Entropy
distribution gives the ratio of each nj to N as a function of
the discrete (quantized) ej :
 
nj exp ej =kT
¼ ;
N q
where k is the Boltzmann constant (1.381
1023J/K), and
q is the molecular partition function, Sj exp(ej/kT). Deriving
and interpreting partition functions for different modes of
molecular motion and different systems is one of the principal
tasks of statistical thermodynamics.
For the sake of brevity, we will assume that macroscopic
thermodynamic systems are ergodic (properties averaged
over a very large number of identical systems are the same
as the properties of one such system averaged over a long
period of time) and bypass the development of canonical
ensembles, but the reader is cautioned that the ergodic
hypothesis and our assumption are untested. The total energy
term, E, used above is the difference between the internal
energy of a system at T and U0, the system’s internal energy at
T = 0, so the most probable configuration gives U  U 0 ¼
nj ej . By differentiation, we obtain an expression for changes
in U with contributions from changes in U0, changes in the
energy states of the system, and changes in the distribution of
molecules among the energy states (Eq. 12).
X X
dU ¼ dU 0 þ nj dej þ ej dnj (12)
We observe that reversible work changes internal energy
by altering the energy levels of a system but does not alter the
n*j distribution of molecules in the levels, so we identify
dw rev = dU0 + n*j dej. On the other hand, reversible heat
transfers change the distribution without altering the levels, so
rev = TdS
we identify dqsys = Sej dnj, or:
sys
 X
1
dS sys
¼ ej dnj : (13)
T
Under isothermal conditions, N is fixed, so Sdn = 0 on its
own. If molecules shift from low energy levels elow to high
energy levels ehigh, however, and dnlow < 0 and dnhigh > 0
(in fact, dnlow = dnhigh). Accordingly, elowdnlow < 0 and
ehighdnhigh > 0. Because of the energy difference, jelow dnlow j
< ehigh dnhigh so Sej dnj >0 and TdS sys > 0. If we define
entropy as being proportional to W ,
S ¼ k ln W  ; (14)
we can derive the established identity (Eq. 13) by another
route beginning with Eq. 14. That derivation is beyond the
scope of this article, but we conclude that it validates
the definition in Eq. 14. If we increase the temperature of
Epigenesis
the system, more energy levels become accessible, so the
number of ways to distribute molecules among those energy
levels and still satisfy the constraints increases. Accordingly,
W  and S also increase. Decreasing temperature yields lower
values of W  and S. In the limit of T ! 0 K, the only accessible
energy level is the ground state, so limT!0 W  = 1. It follows
that limT!0 S = 0, which is consistent with our earlier discus-
sion of the third law and absolute entropies.
Cross-References
▶ Activation Parameters: Energy, Enthalpy, Entropy, and
Volume
▶ Chelation
▶ Enthalpy
▶ Equilibrium
▶ Free Energy
▶ Fugacity
▶ Geochemical Thermodynamics
▶ Henry’s Law
▶ Kinetics of Geochemical Processes
▶ Solid Solution/Exsolution
References
Anderson GM, Crerar DA (1993) Thermodynamics in geochemistry.
The equilibrium model. Oxford University Press, New York. 588 pp
Atkins PW (1986) Physical chemistry, 3rd edn. W. H. Freeman, New
York. 857 pp
Denbigh K (1971) The Principles of chemical equilibrium, 3rd edn.
Cambridge University Press, Cambridge. 494 pp
McGlashan ML (1966) The use and misuse of the laws of thermody-
namics. J Chem Educ 43:226–232
McGlashan ML (1979) Chemical thermodynamics. Academic, New
York. 345 pp
Nordstrom DK, Munoz JL (1994) Geochemical thermodynamics, 2nd
edn. Blackwell Scientific Publications, Cambridge. 493 pp
Epigenesis
Jamie J. Wilkinson
Department of Earth Sciences, Natural History Museum,
London, UK
Department of Earth Science and Engineering, Imperial
College, London, UK
Definition
Epigenesis is a term derived from the Greek
(epi+genesis = after formation) that is primarily used to
describe a geological process involving the addition,
453
modification, and/or removal of minerals from a rock subse-
quent to its formation. It typically refers to a metasomatic
process that has, as well as transforming the mineral assem-
blage, modified the chemical composition and even textural
properties of the bulk rock via the addition or removal of
elements. The term may be applied in geomorphology to the
formation of glacial valleys, in soil science to soil formation,
and in sedimentology to the process of formation of a mineral,
texture, or structure subsequent to compaction – thus partly
synonymous with diagenesis. However, it is most commonly
used in geochemistry in relation to the formation of mineral E
deposits – in particular those formed from hydrothermal
solutions such as sediment-hosted mineral deposits (Brown
2014; Ramanaidou and Wells 2014; Wilkinson 2014) and
volcanogenic massive sulfide deposits (Franklin et al. 1981).
Epigenetic Mineral Deposits
The early subdivision of mineral deposits into those that
formed by syngenesis (process of formation similar to that
which formed the enclosing rock and, in modern usage, also
implying a temporal connection) and epigenesis (introduced
into a preexisting rock) recognized the fundamental differ-
ence in mineral deposit morphology that can result from these
end-member processes (Lindgren 1913). The morphology is
important for mineral exploration and mining because differ-
ent methods are used to recognize and mine ore from syn-
genetic deposits that are lenticular and conformable with
stratigraphic units (described as stratiform) and epigenetic
deposits that may have tabular, pipelike, or irregular geome-
tries that are contained within specific stratigraphic units
(stratabound) or crosscut stratigraphy. In sedimentary ore
systems, there is overlap in the usage of terms, with syn-
genesis often considered synonymous with early diagenesis
and epigenesis with late diagenesis (or epidiagenesis; partic-
ularly in the USA). Syndiagenesis has been used as a term by
economic geologists to separate the stages of syngenesis and
epigenesis (e.g., Bartholomé et al. 1972; Boast et al. 1981). In
carbonate sedimentology, syndiagenesis is broadly synony-
mous with eogenesis.
Epigenetic deposits may form in two general ways: either
by infiltration of fluids into intergranular porosity with pre-
cipitation of ore and gangue minerals in pore space and by
replacement of existing grains or by infiltration of fluids along
fracture permeability or brecciated zones with precipitation of
minerals to form veins or breccia cements. Gradations
between the two end-members frequently occur, where min-
eralization spreads into the wall rocks surrounding vein sys-
tems or into breccia clasts. Replacement can be either highly
selective, preserving primary rock components such as fossils
or primary rock-forming minerals and textures, or may be
texturally destructive so that primary features are obscured.
454
Controversy Regarding the Syngenetic or
Epigenetic Origin of Mineral Deposits
The distinction between syngenesis and epigenesis is not
straightforward and has led to extensive debate in the litera-
ture surrounding the processes responsible for the formation
of certain types of mineral deposits, such as sediment-hosted
Pb-Zn deposits (e.g., Leach et al. 2005). Syngenesis, as
applied to sediment-hosted or volcanogenic massive sulfide
deposits, implies that ore-forming processes occurred near the
water-sediment interface, effectively in connection with the
hydrosphere and biosphere. By contrast, epigenesis can have
occurred tens or hundreds of millions of years after rock
lithification, potentially at significant burial depths, and pos-
sibly at high temperatures. Clearly, the processes involved in
the transport and precipitation of ore-forming elements in
such contrasting regimes are very different. The relative
timing of ore formation also has major implications for min-
eral exploration: syngenetic deposits are likely to be found in
a specific stratigraphic interval in a particular terrane, whereas
epigenetic deposits could be found in any suitably receptive
rocks that are older than the hydrothermal system involved. In
a number of cases, both genetic models have been proposed
for the same ore deposits, with perhaps the best-known exam-
ples being the Proterozoic sediment-hosted Zn-Pb deposits of
the McArthur Basin in Australia (Large et al. 1998; Perkins
and Bell 1998) and the Carboniferous deposits of the Irish
Midlands orefield (Hitzman and Beaty 1996; Wilkinson et al.
2005). The principal arguments for the timing of mineraliza-
tion are based on geological relationships and textural
observations – whether the ore minerals form stratigra-
phically conformable sedimentary layers that can be
accounted for by deposition from suspension in water or if
they can be shown to transgress stratigraphic contacts and
replace preexisting minerals in the rock. Unfortunately, where
a mineral deposit is subject to post-formation diagenetic,
tectonic, or metamorphic overprinting, remobilization of min-
erals can obscure primary depositional textures and generate
secondary “epigenetic” characteristics. Thus, because of the
ambiguity often present in such observations, research efforts
increasingly utilize isotope geochemistry or geochronology to
distinguish between syngenesis and epigenesis (e.g.,
Hnatyshin et al. 2015).
Mineral Systems with Syngenetic and Epigenetic
Components
Other complications include the fact that syngenetic deposits
can also have epigenetic components in the form of vein,
breccia, and/or alteration “feeder” zones that underlie the
main deposit, such as in sedimentary-exhalative (SEDEX)
deposits (Leach et al. 2005) or volcanic-hosted massive
Epigenesis
sulfide deposits (Franklin et al. 1981). These reflect the
upward passage of buoyant hydrothermal fluids into the
ore-forming environment and so are an intrinsic part of the
overall hydrothermal system. The temporal relationship
between the metasomatic process and rock lithification
implicit in the term epigenesis is also subject to ambiguity
when the host rocks are carbonate rich. Such rocks lithify very
quickly, potentially at the seafloor (Bathurst 1970); as a result,
strictly epigenetic processes and textures can be generated,
although the processes are effectively operating within the
syngenetic domain (e.g., Wilkinson and Hitzman 2014).
Summary
Epigenesis has wide usage in the geoscience literature but is
most commonly applied in the description of mineral
deposits. Because it describes a genetic process that may not
be easily constrained from the rock record, and because it also
represents an “end-member” of a spectrum of linked pro-
cesses, its usage is quite commonly controversial. For more
details on distinctions between the usage of “epigenesis,”
“diagenesis,” and “syngenesis” in the context of mineral
deposits, see Misra (2000).
Cross-References
▶ Diagenesis
▶ Hydrothermal Solutions
▶ Ore Deposits
References
Bartholomé P, Evrard P, Katekesha F, Lopez-Ruiz J, Ngongo M (1972)
Diagenetic ore-forming processes at Kamoto, Katanga, Republic of
the Congo. In: Amstutz GC, Bernard AJ (eds) Ore in sediments.
Springer, New York, pp 21–41
Bathurst RGC (1970) Problems of lithification in carbonate muds. Proc
Geol Assoc 81:429–440
Boast AM, Coleman ML, Halls C (1981) Textural and stable isotopic
evidence for the genesis of the Tynagh base metal deposit, Ireland.
Econ Geol 76:27–55
Brown AC (2014) Low-temperature sediment-hosted copper deposits.
In: Holland HD, Turekian KK (eds) Treatise on geochemistry, Geo-
chemistry of mineral deposits, vol 13, 2nd edn. Elsevier, Amsterdam,
pp 251–272
Franklin JM, Sangster DM, Lydon JW (1981) Volcanic-associated mas-
sive sulfide deposits. In: Skinner BJ (ed) Economic geology 75th
anniversary volume. Lancaster Press, Pennsylvania, pp 485–626
Hitzman MW, Beaty DW (1996) The Irish Zn-Pb-(Ba) orefield. In: Sangster
DF (ed) Carbonate-hosted lead-zinc deposits, Special publication, vol 4.
Society of Economic Geologists, Littleton, pp 112–143
Hnatyshin D, Creaser RA, Wilkinson JJ, Gleeson SA (2015) Re-Os dating
of pyrite confirms an early diagenetic onset and extended duration of
mineralization in the Irish Zn-Pb orefield. Geology 43:143–146
Equilibrium
Large RR, Bull SW, Cooke DR, McGoldrick PJ (1998) A genetic model for
the HYC deposit, Australia: based on regional sedimentology, geochem-
istry, and sulphide–sediment relationships. Econ Geol 93:1345–1368
Leach DL, Sangster DF, Kelley KD, Large RR, Garven G, Allen C R,
Gutzmer J, Walters S (2005) Sediment-hosted lead-zinc deposits: a
global perspective. Econ Geol (100th Anniversary Volume,
1905–2005):561–607
Lindgren W (1913) Mineral deposits, vol 11, 1st edn. McGraw-Hill
Book Company, New York, pp 140–165
Misra K (2000) Understanding mineral deposits, vol 4. Springer, Neth-
erlands, pp 228–237
Perkins WG, Bell TH (1998) Stratiform replacement lead-zinc deposits:
a comparison between Mount Isa, Hilton, and McArthur River. Econ
Geol 93:1190–1212
Ramanaidou ER, Wells MA (2014) Sedimentary hosted iron ores. In:
Holland HD, Turekian KK (eds) Treatise on geochemistry, Geo-
chemistry of mineral deposits, vol 13, 2nd edn. Elsevier, Amsterdam,
pp 313–356
Wilkinson JJ (2014) Sediment-hosted zinc-lead mineralization: pro-
cesses and perspectives. In: Holland HD, Turekian KK (eds) Treatise
on geochemistry, Geochemistry of mineral deposits, vol 13, 2nd edn.
Elsevier, Amsterdam, pp 219–250
Wilkinson JJ, Hitzman MW (2014) The Irish Zn-Pb Orefield: the view
from 2014. In: Archibald SM, Piercey SJ (eds) Current perspectives in
zinc deposits. Irish Association of economic geology, Dublin, pp 59–73
Wilkinson JJ, Everett CE, Boyce AJ, Gleeson SA, Rye DM (2005)
Intracratonic crustal seawater circulation and the genesis of sub-sea-
floor Zn-Pb mineralization in the Irish orefield. Geology 33:805–808
Equilibrium
Heinz Gamsjäger
Leoben, Austria
Definition
Equilibrim is the condition toward which a system will
change in the absence of constraints. In a chemical system it
is the state with the minimum Gibbs Free Energy.
The state of equilibrium is a common feature of all
branches of physical science. It plays an important role
when either chemical, mechanical, or electromagnetic phe-
nomena are described. Three examples emphasizing chemical
applications must suffice. Analytical and preparative chemists
use equilibrium considerations to select the temperature and
concentration ranges best suited for their work. Chemical and
metallurgical processes are often designed on the basis of
equilibrium calculations. Moreover, equilibrium models for
the distribution of species in natural waters are frequently
applied by geochemists.
Heinz Gamsjager is deceased.
Originally published in: Marshall CP, Fairbridge RW (eds) Encyclopedia
of Geochemistry, Springer Netherlands, 1999
455
As originally demonstrated by Gibbs, the concept of equi-
librium is described in an entirely general way by classical
thermodynamics (Gibbs 1875). Let us consider the contem-
porary treatment also based on the second law and using
modern nomenclature given by McGlashan (1979):
If a spontaneous process occurs in an isolated system
(L) consisting of one or more phases a, b, ..., the entropy SS of
the system increases. If nothing is happening, the entropy remains
constant and the isolated system is said to be in a state of
equilibrium.
In an isolated system the quantities internal energy US, E
volume VS and the total amount of substance nS are kept
constant. Expressing the entropy variation by its time deriv-
ative @S/@t, the statement can be cast into a mathematical
relation where the inequality and equality signs refer to spon-
taneous process and equilibrium respectively:
 
@SS =@t US , V S , nS
P0 (1)
The definition @SS/@t = 0 includes metastable equilibria,
provided no change is detected during the period and with the
instruments of observation. Thus, a mixture of methane and
air at ambient temperatures is metastable unless it is lighted,
whereupon it burns and transforms into the even more stable
state of flue gas.
Other, sometimes more convenient, sets of independent
variables are obtained from Eq. 1 simply by applying the
rules of differentiation. By this method, the extensive quan-
tity U is successively transformed into the likewise exten-
sive quantities H (enthalpy), A (Helmholtz function), and
G (Gibbs function). Now, the conditions for chemical as well
as thermal, hydrostatic, diffusive, etc. equilibrium can be
derived from Eq. 1 or one of its alternative forms
(McGlashan 1979). It is easy to carry out and study chemical
reactions at constant thermodynamic temperature T and
pressure p; consequently, Eq. 2 is the most important from
the viewpoint of chemistry.
 
@GS =@t T , p, nS O0 (2)
It can be shown that the general homogeneous reaction:
aA þ bB Ð cC þ dD (3)
is in a state of equilibrium when the chemical potentials m
satisfy the following equation:
cmC þ dmD  amA  bmB ¼ 0 (4)
This result can easily be generalized to systems, composed
of more than one phase, in which two or more simultaneous
reactions occur.
456 Equilibrium Constant

Once a system has attained chemical equilibrium the con-
centrations of the chemical species involved no longer
change, on a macroscopic level it has chemically come to a
rest. On the molecular level, the reactions are still going on,
however, as can be demonstrated by spiking an equilibrated
system with an isotopic tracer.
where DrG(T) is the change in the standard molar Gibbs
energy due to the reaction. The value of K(T) depends on the
choice of the standard state, which must be specified. The
standard equilibrium constant is, like all standard thermody-
namic quantities, independent of pressure and composition
(Ewing et al. 1994; Cohen et al. 2008).

Cross-References Introduction

▶ Activation Parameters: Energy, Enthalpy, Entropy, and The following definitions and examples are based on recom-
Volume mendations by the Commission on Thermodynamics of the
▶ Chemical Bonds International Union of Pure and Applied Chemistry (Ewing
▶ Clapeyron’s Equation et al. 1994), cf. also the recommendations given in IUPAC’s
▶ Enthalpy Green Book (Cohen et al. 2008). Further details can be found
▶ Entropy in textbooks on Chemical Thermodynamics (e.g., McGlashan
▶ Equilibrium Constant 1979).
▶ Fluid–Rock Interaction According to IUPAC, all standard chemical potentials
▶ Free Energy and standard equilibrium constants are defined in terms of
▶ Fugacity an arbitrarily chosen standard pressure. Alternative defi-
▶ Geochemical Thermodynamics nitions with explicit pressure dependencies of these quan-
▶ Gibbs-Duhem Equation tities are possible for reactions in condensed phases (e.g.,
▶ Henry’s Law Wood and Battino 1990). This, however, requires differ-
▶ Kinetics of Geochemical Processes ent relationships for reactions in condensed and gaseous
▶ Metamorphic Reactions and Processes phases, because for the latter the standard thermodynamic
▶ Partitioning and Partition Coefficients quantities are usually defined for a standard pressure of
▶ Phase Equilibria 1 bar.
It should be emphasized that other choices can be made as
well. For instance, fixed-temperature and fixed-pressure stan-
References dard states have been proposed for aqueous solutions
(Nordstrom and Munoz 1986, p. 186).
Gibbs JW (1875) On the equilibrium of heterogeneous substances. In:
The collected works of J. Willard Gibbs, vol I: Thermodynamics.
Yale University Press, New Haven, p 56. Reprinted 1948
McGlashan ML (1979) Chemical thermodynamics. Academic, London, Equilibrium Condition
345 pp
Consider a chemical reaction written as
X
vB B ¼ 0, (1)
B
Equilibrium Constant
where each vB represents the stoichiometric coefficient of the
Erich Königsberger substance B in the balanced reaction; the sign is positive for
Chemical and Metallurgical Engineering and Chemistry, products and negative for reactants.
School of Engineering and Information Technology, The Gibbs energy, G(T,p,x), of the reacting system (1) is
Murdoch University, Murdoch, WA, Australia defined as
X
GðT, p, xÞ ¼ nB ðxÞmB ðT, p, xÞ,
Definition B

The standard equilibrium constant, K(T), for a chemical where nB(x) is the amount of substance B and mB(T,p,x)
reaction is a dimensionless quantity defined by denotes the chemical potential of substance B which depends
on temperature, pressure, and the extent of reaction, x. Chem-
K  ðT Þ ¼ expfDr G ðT Þ=RT g, ical equilibrium will be attained when G assumes a minimum
Equilibrium Constant 457

with respect to x at constants T and p. The extent of reaction is referred. Suitable reference values, known as standard ther-
defined by modynamic quantities, can be obtained from the standard
chemical potential by the usual mathematical operations
nB ðxÞ ¼ nB ð0Þ þ vB x, (2) (cf., e.g., McGlashan 1979).
The standard chemical potential is defined in terms of a
where nB(0) denotes the initial amount of substance B. Hence, specified state of aggregation, a specified temperature, and an
the condition of equilibrium for reaction (1) is given by arbitrarily chosen but specified standard pressure. For a solute
substance in a solution, an arbitrarily chosen but specified
@G X @nB X @m standard molality is required too. Consequently, all standard
¼ mB þ nB B ¼ 0: (3)
@x B
@x B
@x thermodynamic quantities, including standard equilibrium
constants, are independent of pressure and composition, but E
According to the Gibbs-Duhem equation, the second sum they remain functions of temperature.
in Eq. 3 is identically equal to zero. Consequently, by the use The recommended value of the standard pressure p is
of Eq. 2, 10 Pa (1 bar). If the value of p changes from p
5 
1 to p2 , then
the change in the standard chemical potential is given by one
X
vB mB ðT, p, xeq Þ ¼ 0, (4) of the following three equations:
B
1. For a gaseous substance B, whether pure or mixed,
where xeq denotes the extent of reaction at chemical       
m   
B g, T, p2  mB g, T, p1 ¼ RT ln p2 =p1 :
equilibrium.
2. For a substance B in a condensed phase (cd), whether pure
or mixed, or for a solvent substance B in a solution
Standard Equilibrium Constant
ð p

 
 
 
 2

The standard equilibrium constant K(T) for reaction (1) is a mB cd, T, p2  mB cd, T, p1 ¼ V B ðcd, T, pÞ dp,
p
1
dimensionless quantity defined by
X where V B ðcd, T, pÞ is the molar volume of the pure
RT ln K  ðT Þ ¼  vB m 
B ðT Þ ¼ Dr G ðT Þ (5) substance B.
B
3. For a solute substance B in a solution
   
where m B ðT Þ denotes the standard chemical potential of m   
B solute, T, p2  mB solute, T, p1
substance B and DrG(T) is the standard Gibbs energy of ð p
2
reaction (1). ¼ V1
B ðsolute, T, pÞ dp,
Combination of Eqs. 4 and 5 yields p
1

X

RT ln K  ðT Þ ¼ vB mB ðT, p, xeq Þ  m where V 1
B ðsolute, T, pÞ is the partial molar volume of the
B ðT Þ : (6)
B solute B at infinite dilution.

Evaluation of the differences For example, if p increases from 105 Pa to 101,325 Pa

 (1 atm), then, for a gaseous substance, the standard chemical
mB ðT, p, xÞ  m
B ðT Þ
between the chemical potential of substance B and its stan-
dard chemical potential for the same temperature and state of
aggregation will be discussed below.
Standard Chemical Potential
Since a variation of temperature, pressure, or composition
changes the values of thermodynamic quantities, it is impor-
tant to define a reference value of each quantity to which such
changes, which can be measured or calculated easily, may be
potential increases by about 0.013 RT and the standard molar
entropy decreases by about 0.013 R. For condensed phases,
however, the same variation in the standard pressure will result
in a negligible difference in the standard thermodynamic quan-
tities compared with the current uncertainties in these data.
The recommended value of the standard molality m is
1 mol kg1.
In the following three sections, definitions are given for the
standard chemical potential of a substance B in the following
states of aggregation: pure or mixed gas phase, pure or mixed
condensed phase or solvent substance in a solution, and solute
substance in a solution. Using these definitions, methods
for the calculation and measurement of the differences
458 Equilibrium Constant

 ð p
mB ðT, p, xÞ  m
B ðT Þ , and hence of the standard equilib-
rium constant, can be derived. m 
B ðcd, T Þ ¼ mB ðcd, T, pÞ þ V B dp, (9)
p

where mB ðcd, T, pÞ and V B ðcd, T, pÞ are the chemical potential

Gaseous Systems and the molar volume of the pure substance B, respectively.
The activity coefficient fB of a substance B in a liquid or
The standard chemical potential of a gaseous substance B, solid mixture is a dimensionless quantity defined by
whether pure or mixed, is defined by the relation
mB ðcd, T, p, xÞ  m
B ðcd, T Þ
mB ðg, T, p, xÞ  m 
B ðg, T Þ ¼ RT lnðf B =p Þ, (7)
ð p
where x denotes the set of mole fractions xB, xC,    of B, C, ¼ RTlnðxB f B Þ  V B dp, (10)
p
   and fB is the fugacity of B defined by
ð p   where x denotes the set of mole fractions xB, xC,    of B, C,
f B ¼ ðxB pÞexp
VB 1
 dp , (8)    in the mixture. The (relative) activity of B in a mixture is a
0 RT p dimensionless quantity given by

where VB = VB(g, T, p, x) is the (partial) molar volume of aB ¼ x B f B : (11)

B. Consequently, the fugacity has the same dimension as
pressure. Since the integral in Eq. 8 vanishes for a perfect It follows from the definitions of aB and fB that
gas, the fugacity is sometimes used as a measure of the
deviation of the gas or gaseous mixture from perfection. For lim aB ¼ lim f B ¼ 1:
xB !1 xB !1
an expression of the fugacity in terms of the virial equation of
state see, e.g., Ewing et al. (1994) or McGlashan (1979). As
A mixture of substances B, C,    is called an ideal mixture
the symbol fB is also used for the activity coefficient of
when aB = xB, aC = xC,    or fB = 1, fC = 1,  .
substance B referenced to Raoult’s law, the fugacity is some-
The standard equilibrium constant for a reaction in a con-
times denoted as p~B (Cohen et al. 2008).
densed phase is given by
The standard equilibrium constant for a reaction in the gas
phase is derived by combining Eqs. 6 and 7: " ð  #
Y p X
vB V B
Y Y K  ðcd, T Þ ¼ vB
ðxB f B Þ exp  dp : (12)
K  ðg, T Þ ¼ ðf B =p Þ ðxB p=p Þ B : RT
vB v
¼ lim B p B
p!0
B B
The pressure integral in Eq. 12 is often neglected, espe-
The second form of this equation, which follows from cially when p p. Then, the quantity
Eq. 8, provides the most direct method of determining
K(g, T) experimentally and is useful when the partial Y
K xf ðcd, T, pÞ ¼ ðxB f B ÞvB
molar volumes VB(g, T, p, x) are not known. At least at low B
to moderate pressures, the extrapolation to zero pressure is
often not necessary. In practice the quantity is a good approximation to K(cd, T), whereas the quantity
Y Y
K p ðg, T, p, xÞ ¼ ð x B p Þ vB K x ðcd, T, p, xÞ ¼ ð x B Þ vB
B B

which has the dimension of ðpressureÞSB vB is often used. usually varies strongly with composition and is therefore a
However, Kp varies with pressure and composition and poor approximation to K(cd, T) because only few real mix-
should be clearly distinguished from the dimensionless stan- tures are even approximately ideal.
dard equilibrium constant K(g, T).

Solutions
Condensed Phases
A solution is a special description of a mixture for which it is
The standard chemical potential of a substance B in a con- convenient to distinguish between the solvent A, usually
densed phase, whether pure or mixed, is defined by the relation present in excess, and the solutes B, C,   .
Equilibrium Constant 459

The standard chemical potential m B of a solute substance B in The pressure integral in Eq. 14 is often neglected, espe-
a solution of B, C,    in a solvent substance A is defined by cially when p p. Then, the quantity

 1 Y
m
B ðsolute, T Þ ¼ fmB ðsolute, T, p, mÞ  RT ln ðmB =m Þg K a ðsolution, T, pÞ ¼ ðaA ÞvA ð a B Þ vB
ð p B

þ V1
B dp,
p varies with pressure but may still be a good approxima-
(13) tion to K. The solvent term, which is close to unity for
sufficiently dilute solutions, is also often omitted. Then,
where m denotes the set of molalities mB, mC,    of B, C,    the quantity
in the solvent A, m is the standard molality,
Y  vB E
V1 1
B ¼ V B ðsolute, T, pÞ is the partial molar volume of B at K mg ðsolution, T, p, mÞ ¼ mB gm, B =m
infinite dilution, and 1 denotes the limit at infinite dilution, B
i.e., BmB ! 0. In Eq. 13, both terms inside {  }1 diverge
as mB ! 0, but their difference remains finite. varies with composition but may still be a good approxima-
The activity coefficient gm,B of a solute B in a solution is a tion to K. The quantity
dimensionless quantity given by Y
K m ðsolution, T, p, mÞ ¼ ð m B Þ vB ,
mB ðsolute, T, p, mÞ  m
B ðsolute, T Þ B
ð p
 
¼ RT ln mB gm, B =m  V1
B dp: which has the dimension of ðmolalityÞSB vB , is sometimes used
p
in practice; however, it should be clearly distinguished from
the dimensionless quantity K (solution, T). The dependence
The (relative) activity aB of a solute B in a solution is a
of Km on solution composition may be severe, especially for
dimensionless quantity given by
aqueous electrolyte solutions.
aB ¼ mB gm, B =m : In many experimental studies on complex formation and
solubility, the solution composition has been kept essen-
It follows from the definitions of aB and gm,B that tially constant by adding an “inert” electrolyte (constant
ionic medium method). Also, some natural waters maintain
1 a rather constant composition, such as surface seawater.
ðaB m =mB Þ ¼ lim gm, B ¼ 1:
SB mB !0 For reactions involving minor species B at p = p, a
so-called stoichiometric equilibrium constant may be
A solution of solutes B, C,    in a solvent A is called an defined as
ideal-dilute solution when aB = mB/m, aC = mC/m,    or
gm,B = 1, gm,C = 1,   . Y
K st ðsolution, T, mC Þ ¼ lim ðmB ÞvB ,
The definition of the standard chemical potential of the SB mB !0
B
solvent A in a solution is identical with the definition (9) of
mC6¼B ¼ constant:
the standard chemical potential of any substance B in a
condensed phase. The activity aA of the solvent A is defined
The quantity Kst, which has the dimension of
by relations analogous to Eqs. 10 and 11.
ðmolalityÞSB vB, varies with solution composition and is only
Since the solvent A is treated differently from the
valid for a particular set of mC6¼B.
solutes B, C,   , the expression for a chemical reaction in a
For any solution at p = p, Eq. 14 can be written in the
solution may be written in the form
form
X
vA A þ vB B ¼ 0: Y
K  ðsolution, T Þ ¼ lim ðmB =m Þ B :
v
B
SB mB !0
B
The standard equilibrium constant then takes the form

K  ðsolution, T Þ " ð  #
Y p
vA V A þ SB vB V 1 This equation provides the most direct method of deter-
mining K (solution, T) experimentally and is useful when the
vA vB
¼ ðaA Þ ðaB Þ exp  B
dp :
RT
B p
activity aA for the solvent and the activity coefficients gm,B,
(14) gm,C,    for the solutes are not known.
460 Erbium

Effect of Temperature on the Equilibrium Cross-References

Constant
▶ Equilibrium
The standard molar enthalpy of substance B, H 
B , is given by ▶ Free Energy
the Gibbs-Helmholtz equation: ▶ Geochemical Thermodynamics

H   
B ¼ mB  TdmB =dT ¼ RT dðmB =RT Þ=dT.
2

(cf. McGlashan 1979). Remember that all standard quan- References

tities are functions of temperature only. Then, by the use of
Eq. 5, van’t Hoff’s Eq. 15 is obtained: Cohen ER, Cvitaš T, Frey JG, Holmström B, Kuchitsu K, Marquardt R,
Mills I, Pavese F, Quack M, Stohner J, Strauss HL, Takami M, Thor
X AJ (2008) Quantities, units and symbols in physical chemistry,
d ln K  =dT ¼ vB H  2 
B =RT ¼ Dr H =RT :
2
(15) IUPAC Green Book, 3rd edn, 2nd printing. IUPAC & RSC Publish-
B ing, Cambridge
Ewing MB, Lilley TH, Olofsson GM, Rätzsch MT, Somsen G (1994)
Integration of Eq. 15 yields Standard quantities in chemical thermodynamics. Pure Appl Chem
66:533–552
ð T2 McGlashan ML (1979) Chemical thermodynamics. Academic, London


ln K  ðT 2 Þ  ln K  ðT 1 Þ ¼ Dr H  ðT Þ=RT 2 dT: (16) Nordstrom DK, Munoz JL (1986) Geochemical thermodynamics. Black-
T1 well, Palo Alto
Wood SE, Battino R (1990) Thermodynamics of chemical systems.
Cambridge University Press, Cambridge
Thus, if the value of K is known for a temperature T1, its
value at another temperature T2 can be calculated, provided
that the standard enthalpy of reaction, DrH(T), is known over
the temperature range T1 to T2. Usually, the standard molar
heat capacities at constant pressure, C
p, B, have been measured
Erbium
and tabulated (e.g., as coefficients of a power series in T), and
DrH(T0) is known at some reference temperature T0 (which
Scott M. McLennan
may be outside the range T1 to T2). Since
Department of Geosciences, Stony Brook University, Stony
ðT X Brook, NY, USA
Dr H  ðT Þ  Dr H ðT 0 Þ ¼ vB C 
p, B ðT Þ dT,
T0 B

Eq. 16 can be rewritten as

Element Data
ln K  ðT 2 Þ  ln K  ðT 1 Þ ¼ Dr H  ðT 0 ÞðT 2  T 1 Þ=RT 1 T 2 Atomic Symbol: Er
ð T 2 (ð T X ) Atomic Number: 68

 
þ vB Cp, B ðT ÞdT 1=RT 2 dT: Atomic Weight: 167.259(3)
T1 T0 B Isotopes and Abundances: 162Er, 0.139(5)%; 164Er,
1.601(3)%; 166Er, 33.503(36)%; 167Er, 22.869(9)%;
Examples 168
Er, 26.978(18)%; and 170Er, 14.910(36)%
1 Atm Melting Point: 1529  C
In practice, many chemical reactions involve more than one 1 Atm Boiling Point: 2868  C
phase. Then, the expressions for the standard equilibrium Common Valences: +3
constant are appropriate combinations of those given above Ionic Radii: 89.0 pm (CN6), 100.4 pm (CN8), and
for reactions in a single phase (an illustrative example is 121.4 pm (CN12)
presented by Ewing et al. 1994). Pauling Electronegativity: 1.24
Standard equilibrium constants can be defined and also First Ionization Energy: 589.30 kJ mol1
determined experimentally for electron-transfer reactions, Chondritic (CI) Abundance: 0.166 ppm
proton-transfer reactions (e.g., standard acidity constant, K  A ), Silicate Earth Abundance: 0.370 ppm
water dissociation (standard ionization constant, K  w ), complex Crustal Abundance: 2.2 ppm
formation (standard stability constant), dissolution of solid elec- Seawater Abundance: 1.09 ppt
trolytes (standard solubility product, K 
s ), etc. (cf. McGlashan Core Abundance: ~0
1979).
Erbium
Properties
Erbium (Er), named after the Swedish village of Ytterby, is a
soft, malleable, dense (9.066 g cm3), bright silvery-white
metal. Its electronic configuration is [Xe]4f126s2. It is stable at
room temperature and readily dissolves in mineral acids. It is
a group 3 or IIIB inner transition element, with +3 being its
only valence state in geological environments, and is one of
the lanthanide rare earth elements (REE). In geochemical
terminology, it is grouped with heavy rare earth elements
(HREE; Gd-Lu). Erbium has six natural, stable isotopes,
listed above with their abundances.
More than 270 minerals contain lanthanides as essential
structural constituents, but those with Er as a major compo-
nent are restricted to Y and HREE varieties, such as xenotime
[Y(PO4)], bastnäsite [REE(CO3)F], allanite [(REE,Ca,
Y)2(Al,Fe2+,Fe3+)3(SiO4)3(OH)], and britholite [(REE,
Ca)5(SiO4,PO4)3(OH,F)]. In addition, Er may be enriched in
ion adsorption lateritic kaolinite-halloysite clay deposits that
are mined for REE.
Details of properties, mineralogy, history, and uses of Er
were compiled from Goonan (2011), Chakhmouradian and
Wall (2012), Zaimes et al. (2015), Gschneidner (2016), Ham-
mond (2016), and Meija et al. (2016a, b).
History and Use
Erbium was discovered in 1843 by Carl Gustaf Mosander, but
early naming is somewhat confused as the names erbium and
terbium were switched in 1864; it was first isolated as a metal
in 1934. Erbium oxide has a distinctive pink color and so is
used as a colorization additive to glass, cubic zirconia jewelry,
and glazes on ceramics. Due to the Er ion 2940 nm emission,
it is also used to produce Er:YAG lasers, commonly used for
certain types of cosmetic and dental surgeries. Erbium-doped
fiber amplifiers (EDFA) can be used to amplify modulated
laser beams and accordingly are used in fiber-optic systems.
Geochemical Behavior
The fundamental importance of Er in geochemistry is that it is
one of the small trivalent rare earth elements, among the most
useful trace elements in all areas of geochemistry and cosmo-
chemistry due to their coherent and systematic behavior as a
group, largely a consequence of the “lanthanide contraction.”
Erbium is typically a trace element in most rocks and
minerals. It is classified geochemically and cosmochemically
as a highly refractory lithophile element, with a 50%
Tcondensation of 1659 K at 104 bars (Lodders 2003). In most
igneous systems, Er is incompatible with bulk partition
461
coefficients, D < 1. In aqueous systems, Er normally has
very low fluid/rock partition coefficients (1). As a result,
Er contents of clastic sediments typically reflect their average
provenance.
In seawater, Er has very low concentrations (~ppt) and
residence times (~2420 year) with speciation dominated by
Er3+, ErCO3+, and Er(CO3)2. In other aqueous systems (e.g.,
magmatic, hydrothermal), Er can reach ppm levels due to
elevated temperatures, pH effects, and complexing with var-
ious ligands (e.g., F, Cl, OH).
Concentration and residence time data are from compila- E
tions in Taylor and McLennan (2009) and Nozaki (2001).
Biological Utilization and Toxicity
There are no documented biological uses for Er, but in China
REE have been added to fertilizers and to stock feed to
promote growth. REE (including Er3+) likely substitute for
Ca2+ in biological materials and processes. REE concentra-
tions (including Er) in human tissues and fluids are very low
(tens of ppb to <ppb levels, respectively) due to very low
concentrations in natural waters and low uptake rates through-
out the food chain (Bulman 2003). At low levels of ingestion,
there are no established toxic effects that pose a threat to
human health.
Cross-References
▶ Incompatible Elements
▶ Lanthanide Rare Earths
▶ Lithophile Elements
▶ Trace Elements
▶ Transition Elements
References
Bulman RA (2003) Metabolism and toxicity of the lanthanides. In: Sigel
A, Sigel H (eds) Metal ions in biological systems, vol 40, The
lanthanides and their interrelations and biosystems. Marcel Dekker,
Basel, pp 683–703
Chakhmouradian AR, Wall F (eds) (2012) Rare earth elements. Elements
8:333–376
Goonan TG (2011) Rare earth elements – end use and recyclability. US
Geol Surv Sci Invest Rpt 2011–5094. 15pp
Gschneidner KA Jr (2016) Physical properties of the rare earth metals.
In: Haynes WM (ed) CRC handbook of chemistry and physics, 96th
edn. CRC Press, Boca Raton, pp 4-115–4-120
Hammond CR (2016) The elements. In: Haynes WM (ed) CRC hand-
book of chemistry and physics, 96th edn. CRC Press, Boca Raton,
pp 4-1–4-42
Lodders K (2003) Solar system abundances and condensation tempera-
tures of the elements. Astrophys J 591:1220–1247
462
Meija J, Coplen TB, Berglund M, Brand WA, De Bièvre P, Gröning M,
Holden NE, Irrgeher J, Loss RD, Walczyk T, Prohaska T (2016a)
Atomic weights of the elements 2013. Pure Appl Chem 88:265–291.
(IUPAC Technical Report)
Meija J, Coplen TB, Berglund M, Brand WA, De Bièvre P, Gröning M,
Holden NE, Irrgeher J, Loss RD, Walczyk T, Prohaska T (2016b)
Isotopic compositions of the elements 2013. Pure Appl Chem
88:293–306. (IUPAC Technical Report)
Nozaki Y (2001) Rare earth elements and their isotopes in the ocean. In:
Steele JH et al (eds) Encyclopedia of ocean sciences. Academic,
London, pp 2354–2366
Taylor SR, McLennan SM (2009) Planetary crusts: their composition,
origin and evolution. Cambridge University Press, Cambridge.
378pp
Zaimes GG, Hubler BJ, Wang S, Khanna V (2015) Environmental life
cycle perspective on rare earth oxide production. ACS Sustain Chem
Eng 3:237–244
Europium
Matthias Willbold
School of Earth, Atmospheric and Environmental Sciences,
University of Manchester, Manchester, UK
Element Data
Atomic Symbol: Eu
Atomic Number: 63
Atomic Weight: 151.965 g/mol
Isotopes and Abundances: 151Eu 47.8%, 153Eu 52.2%
1 Atm Melting Point: 822  C
1 Atm Boiling Point: 1529  C
Common Valences: 3+, 2+
Ionic Radii: Eu 3+ 109 pm (sixfold), 121 pm (eightfold);
Eu 2+ 131 pm (sixfold), 139 pm (eightfold)
Pauling Electronegativity: 1.2
First Ionization Potential: 5.6704 eV
Chondritic (CI) Abundance: 0.056 mg/g
Silicate Earth Abundance: 0.166 mg/g
Crustal Abundance: 1.1 mg/g
Seawater Abundance: <~0.3 pg/g
Core Abundance: ~0
Properties
Europium is a silvery-white rare-earth element with a pale
yellow tint (Greenwood and Earshaw 1997). It is a ductile
metal and has the second lowest melting point and the lowest
density of all rare-earth elements (Haynes 2016). Europium
metal is commercially produced by isolation from rare-earth
element-bearing minerals such as bastnäsite and more rarely
Europium
monazite. Currently, the Bayan Obo iron ore deposit (China)
is one of the most important sources of bastnäsite worldwide
(Wu 2008).
Europium is the most reactive element of all rare-earth
elements, and it readily oxidizes in air and water to form
colorful Eu(III) compounds. Although predominantly occur-
ring in the trivalent state, it is the only rare-earth element that
occurs in the divalent state under reducing conditions. Diva-
lent Eu is highly reducing and has an ionic radius and coor-
dination number similar to that of Ca2+ (calcium), Sr2+
(strontium), and Ba2+ (barium). Many Eu compounds exhibit
phosphorescence properties.
History and Use
Europium was discovered in 1890 by Paul Émile Lecoq de
Boisbaudran. It was first isolated in 1901 by Eugène-Anatole
Demarçay who named it after Europe. Commercial use of Eu
is limited and centered on its phosphorescence characteristics.
Europium has been used as a dopant in laser crystals and as a
fluorescent material in cathode ray tube television sets and
fluorescent lamps.
Geochemical Behavior
Europium is a refractory, lithophile element (Goldschmidt
1930; Lodders 2003) and is concentrated in the silicate
fraction of terrestrial planets, i.e., the Earth’s mantle and
crust. It is classified as a trace element. The redox-sensitive
property of Eu makes it a useful geochemical tracer. In
oxidized magmatic systems, it exists – like most other rare-
earth elements – as a trivalent ion rendering it incompatible
in rock-forming minerals. Under reducing magmatic condi-
tions, Eu is compatible in feldspar because Eu 2+ can substi-
tute for divalent alkaline earth elements in its crystal lattice.
This geochemical characteristic can lead to relative enrich-
ment of Eu in feldspar-rich rocks. When compared to other
rare-earth elements of similar geochemical properties
(samarium and gadolinium), rocks with an overabundance
of Eu relative to a reference material (most commonly a
chondritic meteorite or an average mid-ocean ridge basalts
composition) are said to have a positive Eu anomaly, while
those with a deficiency are referred to as having a negative
Eu anomaly.
Being an important rock-forming mineral in crustal
lithologies, fractional crystallization of feldspar in mag-
matic systems has a significant effect on the distribution of
Eu within the crust of the Earth, the Moon, and other
planetary bodies. Feldspar-rich lithologies and geochemical
reservoirs, such as the lower continental crust on Earth or
the lunar highland terrains, exhibit large positive Eu
Europium
anomalies, while the Earth’s upper continental crust and the
lunar mare basalts comprise complementary negative Eu
anomalies (Rudnick and Gao 2014; Taylor and McLennan
2010). This makes Eu the most heterogeneously distributed
rare-earth element within the crust. The dichotomy of Eu
distribution in the Earth’s continental crust has recently
been used to trace recycled upper and lower continental
crust materials in enriched mantle reservoirs (Willbold and
Stracke 2010).
In chondritic meteorites, negative Eu anomalies have been
measured in some chondrules and interpreted as the result of
rare-earth element gas-solid fractionation at reducing condi-
tions within the solar nebula (Pack et al. 2004).
In aquatic environments, enrichment of Eu relative to other
rare-earth elements can only occur in reduced hydrothermal
systems (Klinkhammer et al. 1994; Olivarez and Owen 1991).
Experimental data suggest that in hydrothermal fluids Eu
exists in the divalent form, which exhibits higher solubility
than trivalent rare-earth elements (Bau 1991; Haas et al.
1995).
In seawater, Eu only occurs in trivalent form (Goldstein
and Hemming 2014; Piper and Bau 2013). Like other rare-
earth elements, Eu complexes have a low solubility in
seawater. Europium has an average residence time in the
oceans of up to 820 years (Alibo and Nozaki 1999; Li
1982).
Both naturally occurring isotopes of Eu can be regarded as
stable. 151Eu has recently been found to undergo a-decay to
147
Pm, albeit with an extremely long half-life of 4.6
1018
years (Casali et al. 2014). Natural mass-dependent variations
of Eu isotopes have so far only been reported for two calcium-
aluminum inclusions from a carbonaceous chondrite
(Moynier et al. 2006).
Biological Utilization and Toxicity
Europium is not known to have a biological role and is not
classified as a toxic or harmful element in the environment.
Owing to its susceptibility to oxidation, metal powder of Eu
can ignite in air at temperatures >150  C and may thus
present a fire and explosion hazard.
Summary
The unique geochemical property of Eu to change valence
state depending on ambient redox conditions can be used to
understand the magmatic history of igneous rocks and mete-
oritic components, investigate crustal evolution and mantle
recycling models, constrain the sources of sedimentary rocks
and water masses, as well as to study hydrothermal systems
and their deposits.
463
Cross-References
▶ Barium
▶ Calcium
▶ Chondrites
▶ Chondrules
▶ Earth’s Continental Crust
▶ Mantle Geochemistry
▶ Gadolinium
▶ Hydrothermal Alteration
▶ Lanthanide Rare Earths E
▶ Lithophile Elements
▶ Mid-Ocean Ridge Basalts (MORB)
▶ Oxidation-Reduction Reactions and Eh-pH (Pourbaix)
Diagrams
▶ Trace Elements
▶ Samarium
▶ Solubility
References
Alibo DS, Nozaki Y (1999) Rare earth elements in seawater: particle
association, shale-normalization, and Ce oxidation. Geochim
Cosmochim Acta 63:363–372
Bau M (1991) Rare-earth element mobility during hydrothermal and
metamorphic fluid-rock interaction and the significance of the oxi-
dation state of europium. Chem Geol 93:219–230
Casali N, Nagorny SS, Orio F, Pattavina L, Beeman JW, Bellini F,
Cardani L, Dafinei I, Domizio SD, Vacri MLD, Gironi L, Kosmyna
MB, Nazarenko BP, Nisi S, Pessina G, Piperno G, Pirro S, Rusconi C,
Shekhovtsov AN, Tomei C, Vignati M (2014) Discovery of the151 Eu
a-decay. J Phys G Nucl Part Phys 41(7):075101. https://doi.org/
10.1088/0954-3899/41/7/075101
Goldschmidt VM (1930) Geochemische Verteilungsgesetze und
kosmische Häufigkeiten der Elemente. Naturwissenschaften 18
(47–49):999–1013
Goldstein SL, Hemming SR (2014) Long-lived isotopic tracers in ocean-
ography, paleoceanography, and ice-sheet dynamics. In: Holland
HD, Turekian KK (eds) Treatise on geochemistry, vol 8, 2nd edn.
Elsevier, Amsterdam/Oxford/Waltham, pp 453–483
Greenwood NN, Earshaw A (1997) Chemistry of the elements.
Pergamon Press, Oxford
Haas JR, Shock EL, Sassani DC (1995) Rare earth elements in hydro-
thermal systems: estimates of standard partial molal thermodynamic
properties of aqueous complexes of the rare earth elements at
high pressures and temperatures. Geochim Cosmochim Acta
59:4329–4350
Haynes WM (2016) Handbook of chemistry and physics. CRC Press,
New York
Klinkhammer GP, Elderfield H, Edmond JM, Mitra A (1994) Geochem-
ical implications of rare earth element patterns in hydrothermal fluids
from mid-ocean ridges. Geochim Cosmochim Acta 58:5105–5113
Li Y-H (1982) A brief discussion on the mean oceanic residence time of
elements. Geochim Cosmochim Acta 46:2671–2675
Lodders K (2003) Solar system abundances and condensation tempera-
tures of the elements. Astrophys J 591:1220–1247
Moynier F, Bouvier A, Blichert-Toft J, Telouk P, Gasperini D, Albarède
F (2006) Europium isotopic variations in Allende CAIs and the
nature of mass-dependent fractionation in the solar nebula. Geochim
464
Cosmochim Acta 70(16):4287–4294. https://doi.org/10.1016/j.
gca.2006.06.1371
Olivarez AM, Owen RM (1991) The europium anomaly of seawater:
implications for fluvial versus hydrothermal REE inputs to the
oceans. Chem Geol 92:317–328
Pack A, Shelley JMG, Palme H (2004) Chondrules with peculiar REE
patterns: implications for solar nebular condensation at high C/O.
Science 303:997–1000
Palme H, O’Neill HSC (2014) Cosmochemical estimates of mantle com-
position. In: Holland HD, Turekian KK (eds) Treatise on geochemis-
try, vol 3, 2nd edn. Elsevier, Amsterdam/Oxford/Waltham, pp 1–39
Palme H, Lodders K, Jones A (2014) Solar system abundances of
the elements. In: Holland HD, Turekian KK (eds) Treatise on geochem-
istry, vol 2, 2nd edn. Elsevier, Amsterdam/Oxford/Waltham, pp 15–36
Piper DZ, Bau M (2013) Normalized rare earth elements in water,
sediments, and wine: identifying sources and environmental redox
conditions. Am J Anal Chem 04(10):69–83. https://doi.org/10.4236/
ajac.2013.410A1009
Rudnick RL, Gao S (2014) Composition of the continental crust. In:
Holland HD, Turekian KK (eds) Treatise on geochemistry, vol 4, 2nd
edn. Elsevier, Amsterdam/Oxford/Waltham, pp 1–51
Taylor SR, McLennan SM (2010) Planetary crusts – their composition,
origin and evolution. Cambridge University Press, Cambridge
Willbold M, Stracke A (2010) Formation of enriched mantle components
by recycling of upper and lower continental crust. Chem Geol
276:188–197
Wu C (2008) Bayan Obo Controversy: carbonatites versus Iron Oxide-
Cu-Au-(REE-U). Resour Geol 58(4):348–354. https://doi.org/
10.1111/j.1751-3928.2008.00069.x
Zheng X-Y, Yang J, Henderson GM (2015) A Robust procedure for high-
precision determination of rare earth element concentrations in sea-
water. Geostand Geoanal Res 39(3):277–292. https://doi.org/
10.1111/j.1751-908X.2014.00307.x
Evaporites
1.10 g/cc. As
John Warren
Petroleum Geoscience Prog., Chulalongkorn University,
Bangkok, Thailand
Synonyms
Bittern salts; Marine salts
Definition
A widely accepted definition of an evaporite is “A salt rock
Evaporites
undergoing evaporation, which has reached a state of super-
saturation with respect to a particular mineral salt or salts.
This simple definition encompasses a broad range of naturally
precipitated salts and includes alkali earth carbonates as well
as some halide salts (see also halide salts).
Brine Chemistry and Evolution
The mother brine that precipitates a suite of evaporite salts
can be either marine or nonmarine. Sometimes the source
of an ancient evaporite salt can be inferred from the chemical
proportions of the ions in the deposited salt sequence
(Figs. 1 and 2).
Marine Brine
When seawater evaporates, a predictable suite of primary
evaporite salts crystallizes from increasingly concentrated
hypersaline waters (Fig. 1). Today, the chemical makeup
and the proportion of the major ions in seawater are near
constant in all the world’s oceans and are dominated by Na
and Cl, with lesser amounts of SO4, Mg, Ca, K, CO3, and
HCO3 (Fig. 1a). Using the brine classification of Eugster and
Hardie (1978), modern seawater is a Na–(Mg)–Cl–(SO4)
solution, with a density of ~1.03 g/cc and a salinity around
35 ‰. By the time a modern seawater brine attains the bittern
stage (defined by the onset of the potassium and magnesium
salts), it has become a Mg–SO4–Cl brine (Fig. 1a). The first
mineral to precipitate in the seawater evaporation series is
CaCO3, usually as aragonite. This begins in mesohaline
waters when the brine reaches twice the concentration of
seawater (40–60 ‰) and achieves a density
the brine continues to concentrate and approaches four to five
times the concentration of seawater, that is 130–160 ‰, gyp-
sum precipitates from penesaline waters, with densities
around 1.13 g/cc. At 10–12 times the original seawater con-
centration (340–360 ‰) and densities around 1.22 g/cc,
halite drops out of supersaline marine waters.
After halite, the bittern salts (potassium or magnesium
sulfates/chlorides) precipitate from supersaline waters with
concentrations more than 70–90 times that of the original
seawater. Brine density by this stage is more than 1.30 g/cc,
while brine viscosity and feel approach that of olive oil.
Because magnesium and sulfate, along with chloride, are the
major ionic components of a modern seawater bittern, car-
originally precipitated from a saturated surface or near-surface
brine in hydrologies driven by solar evaporation” (Warren 2016).
How Evaporite Salts Form
Evaporites are the resultant mineral precipitates accumulating
in and around an increasingly saline residual brine
nallite (KMgCl3  6(H2O)) and epsomite (MgSO  7H2O) are
the dominant bittern precipitates (Fig. 1b). In contrast, sylvite
(KCl), not carnallite, was a significant primary bittern salt at
different times in the Phanerozoic when seawater contained
lesser amounts of dissolved Mg and SO4. The varying pro-
portion of magnesium relative to calcium and sulfate in Phan-
erozoic seawater was likely controlled by rates of rock–fluid
interactions. Proportions varied according to the rate of
Evaporites
Evaporites, Fig. 1 Chemical evolution of marine brine. (a) Changes in
ionic proportions due to sequential precipitation of aragonite, gypsum,
halite and bittern salts as modern seawater is evaporated to 100 times its
creation of newly formed oceanic crust at mid-ocean ridges or
varying extents of epeiric seaways and associated rates of
dolomitization (Holland et al. 1996; Lowenstein et al.
2014). Further back in time, in the early Archean, when the
atmosphere was hotter, lacked oxygen, and contained high
levels of CO2 and CH4, the typical marine evaporites were
aragonite, nahcolite (NaHCO3), and trona
465
original concentration (in part, after McCaffrey et al. 1987; Warren
2016). (b) Chemical divides and the evolution of three major brine
types in a framework of geological time (in part after Hardie 1990)
silica-rich Ca–Na–HCO3 waters. Pyritic shale and other
sulfide-rich sediment under today’s oxidizing atmosphere
contribute sulfate ions, whereas basic and ultrabasic rocks
tend to produce alkaline Mg–HCO3 mother waters. In all,
Eugster and Hardie (1978) distinguished five major water
types in what they termed “hydrologically closed” continental
evaporite basins (Fig. 2a):
(Na2CO3  NaHCO3  2H2O), along with halite, with little
or no primary gypsum. Given the vast age of these sediments,
the chemically aggressive Archean atmosphere, and the inher-
ent high reactivity of most mineral salts in the subsurface,
only silicified or baritized pseudomorphs of Archean evapo-
rite salts remain (Sugitani et al. 2003; Lowe and Tice 2004).
Nonmarine Brine
More diverse, less predictable suites of evaporite minerals
precipitate during evaporation of continental waters. In non-
marine settings, rivers and groundwaters source most of the
ions that ultimately deposit as evaporite salts. Unlike marine-
derived brines, evaporating continental waters do not draw on
the near-isochemical reservoir that is the ocean. So, in closed
nonmarine hydrologies, the composition of nonmarine brines
and the resulting evaporite suite depends on lithologies
leaching in the surrounding drainage basin (Eugster and
Hardie 1978; Eugster 1980). Flow through limestone aquifers
in hydrologically closed basins produces inflow waters rich in
Ca and HCO3, dolomite dissolution generates more Mg,
while igneous and metamorphic matrices tend to yield
I. Ca–Mg–Na–(K)–Cl
II. Na–(Ca)–SO4–Cl
III. Mg–Na–(Ca)–SO4–Cl
IV. Na–CO3–Cl
V. Na–CO3–SO4–Cl waters
As any one of these waters concentrates within a partic-
ular evaporite basin, it deposits a characteristic suite of
evaporite minerals (Fig. 2b). Early precipitates are the alka-
line earth carbonates: low-magnesian calcite, high-
magnesian calcite, aragonite, and dolomite. Mineralogy of
initial precipitates depends on Mg/Ca ratio in the parent
brine. Subsequent evaporation pathway (types I, II, and III)
of a continental brine is controlled by relative proportions of
calcium, magnesium, and bicarbonate ions in inflow waters.
That is, early carbonate precipitation sets up geochemical
divides that then control subsequent geochemical and min-
eralogical (Fig. 2a).
Archean seawater, which due to much higher levels of CO2
in the atmosphere, had a chemical makeup much like a type
466
Evaporites, Fig. 2 Nonmarine brine evolution pathways (After Eugster and Hardie 1978). (a) Hydrologic classification and brine evolution
pathways of concentrating nonmarine waters (L lake, V valley). (b) Listing of major evaporite minerals associated with the different brine types
I brine, tended to precipitate trona and nahcolite, along with
some halite but no gypsum (Lowe and Fisher-Worrell, 1999;
Sugitani et al. 2003). Thermally dependent type I pathways
explain widespread occurrence of nahcolite in saline lacus-
trine deposits of the Eocene in the USA and China. At that
time, atmospheric CO2 levels were significantly higher than
today. During the Eocene climatic optimum, atmospheric
CO2 ranged between 680 and 1260 ppm, compared to current
levels around 400 ppm (Jagniecki et al. 2015).
Genetic Mechanism and Salt Texture
For a water molecule to escape into the vapor phase in an
evaporitic setting, and so increase the salinity of the residual
brine lake, or pore water in a sabkha, the molecule must
(1) absorb heat energy, (2) be located near the liquid surface,
(3) be moving in the proper direction, (4) have sufficient
Evaporites
energy to overcome liquid-phase intermolecular forces, and
(5) pass through the surface tension interface.
Simple physics (rate and intensity) of molecular escape
during solar evaporation largely controls the hydrological
nature of the remaining brine. Brine bodies, either as
perennial surface waters or as shallow capillary pore
brines, are where various evaporite salts crystallize. Pre-
cipitative textures preserved in an accumulating salt bed
are the result of concentration levels in the brine and its
hydrological setting/stability. In combination, relative
brine density and specific heat capacity of adjacent brine
masses control density and thermal stratification in saline
brine bodies located at or near the earth’s surface. Subse-
quent burial and interaction with other regional shallow
and deep phreatic hydrologies control later textural evo-
lution (Warren 2016).
Specific heat, or heat capacity, is the amount of heat needed
to raise 1 g of a substance by 1  C. For a given amount of heat
Evaporites
input, a unit volume of hypersaline water will show a greater
increase in temperature than a less saline water. Given the same
degree of insolation, this means density-stratified water columns
tend to be heliothermic, with the lower, denser bottom brine
layer being warmer than the somewhat less saline, less dense,
upper water layer (Fig. 3). The condition of a warmer lower
water and a cooler upper water is known as heliothermality. The
level in a water column where a marked change in temperature
occurs in the water column is called the thermocline, and in a
density-stratified brine, mass corresponds to the halocline, some-
times also referred to as the chemocline. The combination of
high temperature, high salinity, and lowered oxygen in the lower
brine mass of a heliothermal brine system means only a special-
ized biota can survive in these waters, often with a specialized
bacterial population occurring in waters just above a saline
thermocline (Warren 2011).
Holomixis allows deposition of a consistent salt layer
across the whole basin floor, beneath both shallow and
deep brine columns. Density stratification allows evaporitic
salts to crystallize only in the upper water mass or at the
brine–brine interface, so bottom nucleation tends to occur on
the shallower lake floor, where it lies above the halocline.
That is, long-term (ectogenic) column stratification means
bottom nucleation of salts can only occur where the upper
salt-saturated brine mass intersects the sediment bottom,
with pelagic settling of salts occurring deeper out in the
depositional basin. Bottom growth of crystals cannot occur
on the deeper bottom beneath a density-stratified system as
there is no mechanism to drive ongoing supersaturation in
the lower water mass. For the same reason, constant
brine reflux driving sinking of a dense brine into sediments
below the flow of the evaporite basin can only proceed in
significant volumes if the overlying brine mass is
holomictic. Deposition of capillary salts (sabkha deposits)
occurs in sub-aerial settings, wherever the saline capillary
zone intersects the land surface.
When salts are accumulating beneath a holomictic brine
mass, the nature of bottom nucleates is controlled by the
stability of the overlying brine column (Fig. 3a). When the
overlying column is deep (>30–100 m), then, other than areas
on the deep bottom of local phreatic outflows, there is no
widespread hydrochemical mechanism to drive fluctuations
in bottom-brine chemistry. The resulting deep bottom pre-
cipitates tend to be monomineralogic crystal clusters, possi-
bly encased by re-transported material washed in from the
shallower surrounds (Fig. 3a). In contrast, when the overlying
brine column is shallow (<30 m and typically <5–10 m), then
the chemistry and stability of the brine varies on a shorter-
term (daily weekly) basis, so more layered bottom nucleates
can accumulate as layered to laminated salt beds. In addition,
all evaporite sediments can be mechanically reworked by
process sets that also typify siliciclastic and carbonate sedi-
ments (Fig. 3b).
467
World-Scale Depositional Associations
At the world scale, most ancient evaporites are marine and
were deposited in either of two large-scale settings that at
times merged into one another, namely, (1) platform evapo-
rites and (2) basinwide evaporites (Fig. 4).
Platform Evaporites
Platform evaporites are made up of stratiform beds, usually
<50 m thick and composed of stacked <1–5 m thick para-
sequences or evaporite cycles, with a variably present E
restricted marine carbonate at its cycle base (Table 2). Salts
were deposited as mixed evaporitic mudflat and saltern evap-
orites, sometimes with local accumulations of bitterns. Typi-
cally, platform salts were deposited in laterally extensive
(>50–100 km wide), hydrographically isolated, subsea level
marine-seepage seaways known as salterns or in evaporitic
mudflats made up of a mosaic of sabkhas and salinas (Warren
2016). These regions had no same-scale modern counterparts
and extended as widespread depositional sheets across large
portions of marine platform areas, which passed seaward
across a sub-aerial seepage barrier into open marine sedi-
ments. In marine-margin epeiric settings, such as the Jurassic
Arab/Hith and Permian Khuff cycles of the Middle East or the
Cretaceous Ferry Lake Anhydrite in the Gulf of Mexico, these
platform evaporites are intercalated with shoalwater marine-
influenced carbonate shelf/ramp sediments, which in turn
pass basinward across a sub-aerial sill into open marine car-
bonates. Landward, they pass into arid zone continental
siliciclastics or carbonate mudflats.
Platform evaporite deposition occurred in pericontinental or
intracratonic (epicontinental) settings, at times of low-amplitude
sea level changes, which typify greenhouse eustasy. Platform
evaporites cannot survive in the high-amplitude, high-
frequency sea level changes of icehouse eustasy. The 100 m +
amplitude oscillations of icehouse times mean sea level falls off
the shelf edge every 100,000 years, so any evaporite that had
formed on the platform is sub-aerially exposed and leached.
Fourth-order high-amplitude icehouse eustatic cycles also tend
to prevent laterally continuous carbonate sediment barriers
forming at the top of the shelf to slope break, and so icehouse
evaporite systems tend not to be hydrographically isolated
(drawdown) at the platform scale. Rather icehouse eustasy
favors nonmarine evaporites as the dominant style, along with
small ephemeral marine-margin salt bodies.
Platform evaporite successions may contain halite beds,
especially in intracratonic basinwide settings, but in peri-
platform settings are typically dominated by 5–40 m thick
Ca-sulfate beds intercalated with normal marine platform
carbonates. The lateral extent of these epeiric platform salts,
like the Khuff Anhydrite with its area of 1,206,668 km2,
constitutes some of the most aerially extensive evaporite
beds ever deposited.
468 Evaporites

Evaporites, Fig. 3 Textures are controlled by hydrology (After Warren nucleation is restricted to shallower areas, when holomictic, bottom
2016). (a) Inter-relationships between brine stratification, stability and nucleated growth possible over whole basin floor
depositional texture. (b) When brine column is stratified, bottom
Evaporites
Evaporites, Fig. 4 Schematic cross sections of ancient evaporite settings that have no modern counterparts. (a) Platform evaporites. (b) Basinwide
evaporites
Basinwide Evaporites
Basinwide evaporites are made up of thick evaporite units
>50–100 m thick made up of varying combinations of deep-
water and shallow water evaporites. They retain textural
evidence of different but synchronous depositional settings,
including mudflat, saltern, slope, and basin (Fig. 4). When
evaporite deposition took place, the whole basin was evapo-
ritic, holomictic, and typically saturated with the same min-
eral phase across vast areas of the basin floor (as when today
in the Dead Sea basin, but within its more limited lateral scale,
halite forms decimeter-thick chevron-dominated beds on the
saline-pan floor and coarse inclusion-poor crystal meshworks
on the deep basin floor). Basinwide successions are usually
dominated by thick massive salt beds (generally more than
100 m thick), which are made up of stacked thick halite beds,
but can also contain substantial volumes of thick-bedded
Ca-sulfate and evaporitic carbonate, as in the intracratonic
basinwide accumulations of the Delaware and Otto Fiord
Basins.
Some basinwide systems (mostly marine-fed
intracratonic) also entrain the larger accumulations of the
world’s potash salts, even though our only Quaternary exam-
ples of economic potash are continental systems (Warren
2010). Owing to inherent purity and thickness of the depos-
ited halite, many halite-dominant basinwide beds are also
remobilized via loading or tectonics into various halokinetic
geometries (Hudec and Jackson 2007).
Basinwide evaporite deposits require tectonically and
hydrologically isolated widespread subsea level depressions
that are not currently present on the world’s surface. They
469
were last active in the Late Miocene, in association with soft-
suture collision basins tied to the Alpine–Himalaya orogenic
belt and to Middle Miocene basins developed in the early rift
stages of the Red Sea. Basinwide systems will be active again
in the future at sites and times of appropriate plate–plate
interaction, when two continental plates are nearby and the
intervening floor at or near a plate-edge rift or suture is both
subsea level and hydrographically isolated.
Lacustrine (Nonmarine) Evaporites
Quaternary continental playa/lacustrine is constructed of
stratiform salt units, with the greater volume of saline sedi-
ment accumulating in lower, more saline portions of the
lacustrine landscape. Beds are usually dominated by nodular
gypsum and displacive halite, deposited in widespread evap-
oritic mudflats and saltpans, rather than as bedded bottom-
nucleated layers on the subaqueous floors of perennial brine
lakes. In ancient counterparts, the total saline lacustrine thick-
ness ranged from meters to hundreds of meters, with lateral
extents measured in tens to hundreds of kilometers. Lacus-
trine salt beds are separated vertically and usually surrounded
by deposits of lacustrine muds, alluvial fans, ephemeral
streams, sheet floods, eolian sands, and redbeds. As today,
ancient lacustrine salts accumulated in endorheic or highly
restricted discharge basins, with perennial saline water
masses tending to occur in the drainage sumps of steep-
sided drainage basins (Warren 2010, 2016). Saline lake basins
accumulating gypsum, or more saline salts like halite or
glauberite, typically have a shallow water table in peripheral
saline mudflat area and so are dominated by sabkha textures.
470
The lowermost part of the lacustrine depression or sump is
typified by ephemeral ponded brine pan deposits rather than
permanent saline waters with subaqueous textures lacking
evidence of drawdown and exposure.
Across the Quaternary, less-saline perennial saline lake
beds tend to occur during more humid periods in the same
continental lacustrine depressions where saline pan beds form
today (e.g., Lake Magadi, Great Salt Lake, Lake Urmia). On a
smaller scale, in some modern saline lake basins, parts of the
lake floor can be permanently located below sea level
(Northern Basin in the Dead Sea, Lake Assal). In some
modern saline sumps dominated by mudflats, a perennial
saline lake water mass can be located toward the edge of a
more central salt-flat zone, forming a “moat” facies (Salar de
Atacama, Salar de Uyuni, Lake Magadi, Lake Natron). These
permanent saline water “moat” regions are typically created
where fresher inflows come in contact with saltier beds of the
lake center, dissolve them, and so form water-filled peripheral
depressions.
As mentioned earlier, saline lacustrine evaporite mineral-
ogies depend on compositions of inflowing waters (Fig. 1);
suitable settings can accumulate thick sequences of non-
marine salts such as trona, glauberite, and thenardite, as well
as more typical sequences of halite, gypsum, and anhydrite.
High-water stage perennial saline lacustrine sediments tend to
be carbonate-rich or silica-rich (diatomaceous) laminites.
Ancient examples of large saline lacustrine deposits include
the Eocene Green River Formation of Wyoming and the
Permian Pingdiquan Formation of the Junggar Basin, China.
Evaporites deposited in a supra-sealevel lacustrine basin
(especially Neogene deposits) have numerous same-scale
quaternary analogs, unlike the more voluminous ancient
marine platform and basinwide evaporites.
Summary
Evaporites are sedimentary salts precipitated at or just below
the earth’s surface in arid hydrologies where the concentration
of the mother brine is driven by solar evaporation. Inflow
brines can be either nonmarine (continental) or marine
(seawater supplied). The largest modern evaporite settings
are continental endorheic sumps with base levels that may
be hundreds to thousands and meters above sea level. The
largest ancient evaporite basins were marine, with vast
expanses of bedded salt accumulating in seawater-seepage
fed, sub-sealevel basins. Ancient marine evaporite basins
were of two types, (a) platform evaporite systems and
(b) basinwide evaporite systems. Platform evaporite systems
typify epeiric and intracontinental seaways during times of
basin restriction in arid settings when the earth was in green-
house climate mode. Basinwide evaporite systems are tecton-
ically induced and accumulated large volumes of salt at times
Evaporites
of plate tectonic-scale continent–continent proximity in both
extensional (early rift) and compressional (suture basins)
plate-scale settings. Platform and basinwide settings require
hydrographical isolation of the evaporite basin, that is, no
long-term surface connection to waters of the open ocean,
but typically had an ongoing marine seepage (drawdown)
connection. Because of the continual seepage connection to
the large isochemical of the ocean, volumes of salt found in
ancient marine evaporite basins are much larger than any
continental lacustrine evaporite system (modern or ancient).
Because of the earth’s current icehouse mode, and a lack of
appropriate zones of continent–continent proximity, there are
no regions on the earth’s surface where large-scale platform or
basinwide hydrologies are currently active. The last time a
basinwide system was active was across the various
Messinian (Late Miocene) basins of the Mediterranean,
while the last time a widespread set of bedded platform
evaporite salts accumulated was during the Eocene in the
Middle East.
Cross-References
▶ Aqueous Solutions
▶ Carbonate Minerals and the CO2-Carbonic Acid System
▶ Halide Minerals
▶ Heat Capacity
▶ Ocean Biochemical Cycling and Trace Elements
▶ Sulfate Minerals
References
Eugster HP (1980) Geochemistry of evaporitic lacustrine deposits. Annu
Rev Earth Planet Sci 8:35–63
Eugster HP, Hardie LA (1978) Saline lakes. In: Lerman A (ed) Lakes;
chemistry, geology, physics. Springer, New York, pp 237–293
Hardie LA (1990) The roles of rifting and hydrothermal CaCl2 brines in
the origin of potash evaporites: an hypothesis. Am J Sci 290:43–106
Holland HD, Horita J, Seyfried W (1996) On the secular variations in the
composition of Phanerozoic marine potash evaporites. Geology
24:993–996
Hudec MR, Jackson MPA (2007) Terra infirma: understanding salt
tectonics. Earth Sci Rev 82:1–28
Jagniecki EA, Lowenstein TK, Jenkins DM, Demicco RV (2015) Eocene
atmospheric CO2 from the nahcolite proxy. Geology 43:1075–1078
Lowe DR, Fisher-Worrell G (1999) Sedimentology, mineralogy, and
implications of silicified evaporites in the Kromberg formation,
Barberton Greenstone Belt, South Africa. In: Lowe DR, Byerly
GR (eds) Geologic evolution of the Barberton Greenstone Belt,
South Africa, Special paper, vol 329. Geological Society of Amer-
ica, pp 167–188
Lowe DR, Tice MM (2004) Geologic evidence for Archean atmospheric
and climatic evolution: fluctuating levels of CO2, CH4, and O2 with
an overriding tectonic control. Geology 32:493–496
Lowenstein T, Kendall B, Anbar AD (2014) Chapter 8.21. The geologic
history of seawater, treatise on geochemistry, 2nd edn. Elsevier, pp
569–621
Experimental Mineralogy and Petrology
McCaffrey MA, Lazar B, Holland HD (1987) The evaporation path of
seawater and the coprecipitation of Br  and K+ with halite. J
Sediment Petrol 57:928–937
Sugitani K, Mimura K, Suzuki K, Nagamine K, Sugisaki R (2003)
Stratigraphy and sedimentary petrology of an Archean volcanic-
sedimentary succession at Mt. Goldsworthy in the Pilbara Block,
Western Australia: implications of evaporite (nahcolite) and barite
deposition. Precambrian Res 120:55–79
Warren JK (2010) Evaporites through time: tectonic, climatic and
eustatic controls in marine and nonmarine deposits. Earth Sci Rev
98:217–268
Warren JK (2011) Evaporitic source rocks: mesohaline responses to
cycles of “famine or feast” in layered brines, Doug Shearman memo-
rial volume, IAS special publication number, vol 43, Wiley-Black-
well, pp 315–392
Warren JK (2016) Evaporites: A Compendium. Springer, Berlin, 1854 p.
ISBN 978-3-319-13511-3
Experimental Mineralogy
and Petrology
Charles A. Geiger1 and Tatsuhiko Kawamoto2
1
Department of Chemistry and Physics of Materials,
Section Mineralogy, University of Salzburg,
Salzburg, Austria
2
Institute for Geothermal Sciences, Graduate School of
Science, Kyoto University, Kyoto, Japan
Definition
Experimental mineralogy and petrology is an Earth science
discipline that models or simulates mineralogical, petrologi-
cal, and geochemical processes within the laboratory in order
to understand the physical and chemical properties of min-
erals, rocks, melts, and fluids and their interactions. For
example, most of Earth’s interior and that of other planets
are not accessible to direct study. Thus, various types of high-
pressure and high-temperature experiments are carried out in
order to obtain an understanding of deep Earth and planetary
structure and behavior.
History and Introduction
The field of experimental mineralogy and petrology, which
constitutes a part of the geological or Earth sciences, can be
traced back for at least 200 years. The Scottish geologist Sir
James Hall (1761–1832) is considered to be the father of
experimental petrology. His laboratory experiments on the
melting and crystallization of basaltic magmas were decisive
in resolving the debate between the Plutonist and Neptunist
schools of geologists in the late eighteenth and early nine-
teenth centuries and in favor of the former. His results
471
illustrated the power of laboratory investigations for a scien-
tific field that was field orientated and observational. The
publication of N.L. Bowen’s well-known book The Evolution
of the Igneous Rocks in 1928, which was largely based on his
pioneering laboratory melting investigations on simple model
silicate systems at the Geophysical Laboratory of the Carne-
gie Institute of Washington, D.C., proved the necessity of
experimentation in understanding the diversity of igneous
rocks found in Earth. It was not until the last half of the
twentieth century, though, that experimental mineralogy and
petrology became an integral part of geology departments at E
universities.
The main, but not only, goal of experimental mineralogy
and petrology is to model or simulate certain conditions
within the laboratory in order to understand the physical and
chemical properties of minerals, rocks, melts, and fluids and
their interactions. As such, it is closely linked to the field of
geochemistry and also partly geophysics. Most of Earth’s
interior and that of other planets are not accessible to direct
study. Thus, it is necessary to simulate conditions or model
geochemical or geophysical processes in order to gain a more
complete understanding of these bodies. In addition, experi-
mental mineralogical studies can have technical applications
relating it, in part, to the fields of inorganic and physical
chemistry, solid-state physics, and materials science.
The physicist Percy Bridgman (1882–1962), a foremost
scientist in high-pressure research and a Noble prize recipient,
developed methods to reach conditions up to 2 GPa in 1905
while a student at Harvard University. Since then, studies
have increased greatly in variety and complexity with techni-
cal advances so great and wide-ranging that it is presently
possible to make experiments in the laboratory between pres-
sures of 0.01 Pa and 100 GPa on mm-size samples using
various large-volume presses (Fig. 1) and up to 500 GPa on
micrometer-size samples with a small hand-sized diamond
anvil cell (DAC, Fig. 2). Accessible temperatures range
from those of liquid helium (~3 K) for SQUID magnetic
susceptibility measurements in the DAC up to 6,000 K
using focused laser heating in the same device. Hence, it is
now technically possible to simulate the full range of
pressure-temperature (P-T) conditions encountered on and
within Earth, other rocky planets, and extraterrestrial bodies
as well the large gaseous planets to substantial depths.
Phase Equilibrium Experiments: The Mineralogy
and Petrology of the Earth
Traditionally, phase equilibrium studies have played the cen-
tral role in experimental mineralogy and petrology. They
normally consider the P-T-X stability relations, where
X represents the composition of the system under study,
namely, a mineral, mineral assemblage (i.e., rock), melt, or
472
fluid phase. In most experiments, a sample, either synthetic or
natural in nature, is held for a certain length of time at the
desired P-T condition in some device, quenched to ambient
conditions, and the experimental products are examined. Nor-
mally, the assumption is made that chemical equilibrium was
achieved at the P and T of the experiment and it is preserved in
the quenched state. The chemical and structural state of the
different phases from a series of experiments can then be
characterized and a P-T phase diagram, for example,
constructed. Hence, various solid-state mineral reactions,
melting relations, solid phase transitions, fluid-rock interac-
tions, or changes in chemical and physical properties can be
described. In the 1980s, in situ experiments using the DAC
and simple optical observation were made, which is possible
because of the transparency of diamond. Intense synchrotron
radiation can be used for in situ experiments using large-
volume multi-anvil type (Fig. 1) or toroidal-cell-type devices,
and the DAC (Fig. 2). X-ray diffraction, absorption, and
fluorescence measurements are possible.
Phase equilibrium studies are coupled to the field of chem-
ical thermodynamics. The stability of a given phase with
Experimental Mineralogy and Petrology, Fig. 1 A large volume
1,500-ton Kawai-type multi-anvil press used in connection with syn-
chrotron radiation at the SPring-8 facility, Japan, in order to study
various processes in situ at high P and T. (Kawamoto et al. 2012,
Proceedings of the National Academy of Sciences, U. S. A., 109, 18695)
Experimental Mineralogy and Petrology
respect to another phase or phase assemblage is a function
of the P-T-X conditions of the system. Thermodynamic anal-
ysis enables experimental results to be extrapolated outside of
the P-T conditions under study and also to extract thermody-
namic properties of the participating phases.
Some important areas involving phase equilibrium study
have involved or are:
i. The determination of phase diagrams showing P-T sta-
bility relations for key minerals and their reactions in Earth’s
crust and upper mantle. Here, cold-seal (for P < 0.5 GPa) and
solid-media piston-cylinder (for P < 5 GPa) devices are often
used. Various mineral geothermometers and geobarometers
have also been calibrated. They can be used, as with phase
diagrams, to determine paleo P-T conditions experienced by
buried rocks, which have been returned to Earth’s surface in a
metamorphic cycle, for example, to understand tectonic
events of the crust.
ii. In terms of igneous petrology, laboratory studies made
in the 1950s at relatively low pressures and temperatures
(P < 0.5 GPa and T < 750  C) showed that most natural
granitic bodies are the products of partial melting processes in
the continental crust and were not formed by fluid-related
metasomatic processes, as had been advanced by some field
geologists.
iii. To understand the mineralogy of the deeper mantle,
experiments using various multi-anvil devices and the DAC
over the past 50 years have been made. They show that
Earth’s transition zone, located between seismic discontinu-
ities at 400 and 670 km depth, consists largely of b-Mg2SiO4
(wadsleyite) or g-Mg2SiO4 (ringwoodite) and a composition-
ally complex garnet solid solution containing a majorite com-
ponent [Mg3(Mg,Si)Si3O12]. The 400 km discontinuity is
related to the relatively sharp phase transition of upper-mantle
olivine (Mg2SiO4) to b-Mg2SiO4. Experiments using the
DAC and multi-anvil devices have shown that below the
670 km discontinuity, in the lower mantle, MgSiO3
bridgmanite and magnesiowüstite [(Mg,Fe)O] are stable.
The former phase is considered the most abundant silicate in
Earth. At even higher P and T conditions, DAC experiments
led to the discovery of a new phase called post-bridgmanite,
which is thought be a main constituent of the D00 layer at the
mantle-core boundary.
iv. DAC experiments at extreme P and T are now being
made on the system Fe–O–S with or without H, for example,
in order to better understand the nature of Earth’s Fe-rich
outer and inner core.
Fluid, Melt, and Rheological Behavior at High
Pressure and Temperature
Research has been carried out to study the phase relations and
thermodynamic behavior of single or multicomponent fluids
Experimental Mineralogy and Petrology
Experimental Mineralogy and Petrology, Fig. 2 (left) Externally
heated Bassett-type hydrothermal diamond anvil cell (DAC) setup used
for high-P and high-T investigations in connection with synchrotron
radiation at the Soleil facility, France. The cell size is 75 mm in diameter.
(Kawamoto et al. 2014, Earth, Planets and Space, v. 66, 61). (right)
Schematic figure showing just the two diamond anvils. Between the
(e.g., H2O, H2O–CO2, H2O–CO2–CH4). Equations of state
have been developed that describe P-V(volume)-T relations to
extreme conditions. In addition, much work over the years has
been made to understand mineral or rock phase relations with
fluids at high P and T. Research has showed that water is
soluble in silicate melts as molecular H2O and OH. Starting
in the 1960s, the critical endpoints between SiO2 and aqueous
fluids were determined at 1 GPa, and it was determined that
those between H2O and various natural rocks occurred at
higher pressures. The compositions of basaltic and high-Mg
andesitic partial melts derived from mantle peridotite in the
presence of H2O were determined at 1 GPa in the 1990s. At
pressures greater than 3 GPa, partial melts of hydrous peridotite
are more mafic in composition than those obtained by anhy-
drous melting. Experiments also showed that the composition
of aqueous fluids changes with pressure. Those coexisting with
forsterite (Mg2SiO4) and enstatite (MgSiO3) are relatively sil-
ica rich from 1 to 2 GPa, but they become more mafic than
enstatite in composition at 3 GPa and are close to peridotite
composition from 3 and 5 GPa. The interpretation is that the
solvus between melt and aqueous fluid in the peridotite-H2O
system closes at pressures greater than 3 GPa.
More recently, the critical endpoints between various nat-
ural rock compositions and aqueous fluids have been studied
in situ with multi-anvil devices using synchrotron X-ray
radiography. The second critical endpoints are located at
2.5 GPa and 700  C for pelite-H2O, 2.9 GPa and 800  C for
high-Mg andesite-H2O, 3.4 GPa and 900  C for basalt-H2O,
and 3.8 GPa and 1,000  C for peridotite-H2O. These results
are significant because beyond the endpoints, there is no
phase separation between silicate melt and aqueous fluid.
Only a single fluid phase exists.
The solubility of minerals in aqueous fluids has also been
investigated in situ using the DAC via simple visual
473
diamond faces, a sample is enclosed in NaCl, for example, which serves
as the pressure-transmitting medium and thermal insulator. Here,
focused heating is done internally using an IR laser, and Raman and
synchrotron X-ray measurements can be made (Uts et al. 2013, Reviews
of Scientific Instruments, v. 84, 103904)
observation or using synchrotron X-ray fluorescence. The
results agree with those obtained using the larger devices.
Finally of note, the solubility of hydrogen in nominally anhy-
drous minerals (e.g., olivine, pyroxene, garnet, plagioclase,
wadsleyite, ringwoodite, bridgmanite) coexisting with free
aqueous fluids, as well as the hydrogen partitioning behavior
between minerals and silicate melts, has been determined
from high-P and high-T quenched experiments. It has been
argued through these investigations that Earth’s mantle
possibly contains several “oceans of water” relative to its
hydrosphere.
The physical processes governing melt generation, segre-
gation, and extraction, from an initially solid source, are
important in understanding magma genesis and transport.
Experimental studies have determined the amount of melt
that must be present and under which physical conditions,
before it can separate from a crystalline matrix. Mg-rich
silicate melts, which are difficult to quench from high P and
T, can be studied using the “floating-bead method” with laser
heating, with which it is possible to make forsterite
(Mg2SiO4) glass. Laboratory measurements have shown
that some silicate melts are more compressible than their
equivalent crystalline materials and that at some elevated
pressure, a crossover point occurs such that melts will have
negative buoyancy and crystals will float even under hydrous
conditions. Investigations have also focused on measuring
the macroscopic physical properties, such as density and
viscosity, of silicate- and oxide-based melts for both model
and natural composition systems. High-pressure compres-
sion experiments on some crystalline silicates have shown
that pressure-induced amorphization can be achieved with-
out heating or melting. Hence, the realm between the amor-
phous state of glass, having only short-range order, and
crystalline materials, having long-range translational
474
symmetry, may not be as widely separated as previously
believed.
Rheological studies have been made on the deformational
properties of Earth materials and in the presence or absence of a
fluid and differential stress. Using synchrotron X-rays, defor-
mation experiments in large-volume multi-anvil apparatus
have been carried out at controlled strain rates at high-P and
high-T conditions. High-pressure experiments to investigate
the conversion of olivine to b-Mg2SiO4 have been undertaken
to understand the physical properties of subducting lithosphere.
The metastable persistence of olivine beyond its equilibrium
pressure stability in relatively cool subducted slabs, followed
by a sudden transition to b-Mg2SiO4, has been proposed as a
mechanism for triggering deep-focus earthquakes.
Rock physics and fracture-mechanics investigations, often
undertaken by the engineering community, have now also
found a place within the Earth sciences. Deformation and
textural studies of polymineralogic rocks, highlighting their
anisotropic physical properties, are important for interpreting
seismic profiles of the lower crust and upper mantle. The
lattice preferred orientation of olivine is used as a hydrometer
and also to determine stress conditions in the upper mantle.
Fracture mechanics studies, concerning the strength and duc-
tility of crustal rocks, are undertaken to better understand, for
example, the cause of earthquakes generated in active fault
zones.
Low-Temperature Mineral Solubility and Kinetic
and Diffusion Studies
At low P-T conditions, with T less than roughly 300  C,
solubility measurements of minerals in H2O can be made.
The results can be used to extract thermodynamic data for
phases that are not stable at high temperatures or where the
kinetics of growth are slow, and, hence, thermodynamic equi-
librium is not reached. These experiments are designed to
understand, for example, near surface geological processes
and hydrosphere-lithosphere interactions. Studies have been
undertaken to model seawater-basalt interactions occurring at
ocean spreading centers and to understand the processes
causing submarine sulfide precipitation and ore-forming
reactions.
Phase equilibrium work is complemented by atomic diffu-
sion studies of minerals, glasses, and melts and kinetic studies
of reactions. Experiments investigating the rates of cation or
anion diffusion in minerals at different P and T conditions
have been done and interpreted using the theory of chemical
kinetics. In these studies, atomic diffusion coefficients have
been determined for important rock-forming crystals such as
olivine and garnet to better understand their compositional
zoning. It is possible to estimate the cooling histories of some
rocks and meteorites using such data.
Experimental Mineralogy and Petrology
The study of crystalline point defects in minerals is impor-
tant in kinetic-related experiments, as it is in investigations of
electrical and thermal conductivity behavior. Impedance
spectroscopy and EMF measurements using oxygen-sensitive
solid electrolytes are used in studies of point defects and
electrical conductivity.
Mineral Physics and Experimental Measurement
Techniques
Mineral physics is a scientific field related to experimental
mineralogy and petrology that largely focuses on the physical
properties of minerals often, but not only, at extreme P-T
conditions. The fields of geophysics and mineralogy have
merged, in part, creating the field of high-P mineral physics.
Investigations employ an ever-increasing variety of tech-
niques to study condensed matter properties. Diffraction mea-
surements often use synchrotron radiation. In addition,
infrared, Raman, Mössbauer, nuclear magnetic resonance
(NMR), X-ray absorption (i.e., EXAFS and XANES),
UV/VIS optical absorption, electron paramagnetic resonance,
and neutron spectroscopies are being used routinely.
Magic-angle spinning NMR work has been made on
glasses synthesized in the laboratory to determine their local
properties such as the type of polyhedral coordination units
that are present and their degree of polymerization. Some
NMR measurements can even be made directly at high T in
the stability field of the melt phase. Raman and IR spectros-
copy are used in DAC experiments at high-P and high-T
conditions. Inelastic synchrotron X-ray scattering measure-
ments are made to study the electronic state of various ele-
ments and their oxygen coordination behavior at high
pressures.
The determination of density and bulk modulus and its
pressure derivatives (equations of state) of mantle phases and
silicate melts is needed for the interpretation of seismic veloc-
ity profiles measured in Earth. They are used in models that
attempt to understand early Earth differentiation and Earth
structure with depth. The elastic constants at room tempera-
ture of most of the important rock-forming silicates have been
measured by ultrasonic and Brillouin methods. They, together
with diffraction measurements made in the DAC, allow the
density and bulk moduli of minerals to be determined. X-rays
at synchrotron facilities permit, further, both the P and
T effects on the elastic properties and sound wave velocities
to be measured. The results are important because density and
sound velocity properties of minerals can be compared with
geophysical seismic measurements through the whole Earth
to analyze mantle structure and its mineralogical composition
with depth.
Other measurement techniques that are being employed
include electron back-scattered diffraction to determine
Experimental Mineralogy and Petrology
single-crystal orientations. Computer tomography is used to
make nondestructive analyses of the 3D distributions of min-
erals, melts, and pore spaces in both natural and synthetic
rocks. Very minute sections of experimental charges retrieved
from laser-heated high-P and high-T DAC experiments can be
excised using focused ion beams for careful transmission
electron microscopy examination that indicate reaction and
melting features.
Crystallographers and mineral physicists have contributed
knowledge to other fields concerned with crystal chemistry
and chemical bonding behavior such as inorganic and phys-
ical chemistry and solid-state physics. An emphasis is being
placed on understanding the atomistic scale behavior of Earth
materials (e.g., minerals, melts, fluids) and their relationship
to macroscopic properties and under different, often extreme,
P-T-X environments.
Mineral Synthesis
The discipline of mineral synthesis is also an important, but
often overlooked, part of experimental mineralogy and petrol-
ogy. The growth or production of single crystals, mineral
assemblages, or glasses in the laboratory are invaluable for
other investigations. For example, compositionally simple
end-member and solid-solution minerals, which cannot be
found in nature, have been synthesized for calorimetric mea-
surements to determine their thermodynamic properties.
These thermodynamic data have been used to construct
geothermobarometers and calculate various P-T positions of
mineral reactions and complement those calibrated in phase
equilibrium experiments. Single crystals are grown for
crystal-chemical study using different spectroscopic and dif-
fraction techniques. Polycrystalline nano-diamonds are syn-
thesized directly from graphite at high P and T, and they are
harder than single crystal diamond. They can be used, for
example, to generate the highest pressures in large-volume
multi-anvil devices.
Present and Future Developments
Today, there is an extensive and impressive arsenal of exper-
imental techniques available for investigating the chemical
and physical properties of Earth and planetary materials.
Traditional methods used in experimental mineralogy and
petrology of just 30 years ago largely involved optical micro-
scope examination, X-ray powder diffraction measurements,
and chemical analysis using the electron microprobe. Techni-
cal developments have made it possible to make measure-
ments on continually smaller amounts of material and at
higher spatial resolutions. The development and construction
of large national and international research facilities has
475
increased. Dedicated third-generation synchrotron sources,
characterized by extremely intense radiation, and advanced
neutron spallation sources are permitting new condensed
matter experiments unimaginable 30 years ago. Novel exper-
iments and measurements on minerals and other materials are
being made, and investigations are of increasing complexity
and cost and sometimes involve teams of specialists and
researchers.
Applied research, though not discussed here, is becoming
more widespread, and collaborations between mineral
physicists and material scientists and condensed matter phys- E
icists are increasing. Experimentalists are needed to apply
their expertise to solve environmental problems in such
areas as radioactive waste disposal, pollution control and
cleanup and in finding ways to reduce atmospheric CO2 by
storage in underground reservoirs and through CO2 binding
by forming carbonate minerals. Here, the field of experimen-
tal mineralogy and petrology strongly overlaps that of
geochemistry.
Concomitant with these experimental developments,
there has been an increase in theoretical and computer
studies (computational or theoretical mineralogy). Molec-
ular dynamic simulations, empirically based static lattice
energy, lattice dynamic, and ab initio quantum mechanical
calculations are being applied in a variety of mineralogi-
cal, geochemical, and geophysical studies. Experimental-
ists can use these calculations to guide their avenues of
research and to interpret their results. Calculations can be
used to plan experiments and to predict certain behavior,
but ultimately it is in the laboratory or in nature where they
must be tested.
Summary
The field of experimental mineralogy and petrology has
become extremely diverse encompassing many different
approaches and areas of research. The principles of physics
and chemistry are of increasing importance in many areas of
the Earth sciences. Developments point to the breakdown of
traditional borders both within the Earth sciences and between
it and other scientific disciplines. The synergy generated by
closer collaboration between different scientific disciplines is
spawning new methods and types of experiments and is
providing an evermore quantitative and comprehensive
understanding of a variety of Earth processes.
Cross-References
▶ Fluid–Rock Interaction
▶ Formation and Evolution of the Earth
▶ Fractional Crystallization and Assimilation
476 Experimental Mineralogy and Petrology

▶ Geochemical Thermodynamics Kawamoto T, Kanzaki M, Mibe K, Matsukage KN, Ono S (2012)

▶ Geothermometry and Geobarometry
▶ Kinetics of Geochemical Processes
▶ Magmatic Process Modeling
▶ Partial Melting
▶ Partitioning and Partition Coefficients.
▶ Phase Equilibria
References
Henderson GS, Neuville DR, Downs RT (eds) (2014) Spectroscopic
methods in mineralogy and material sciences. Reviews in mineralogy
and geochemistry. Mineralogical Society of America and Geochem-
ical Society, vol 78. Minerological Society of America, Chantilly,
800 pp
Holloway JR, Wood BJ (1988) Simulating the earth: experimental geo-
chemistry. Unwin Hyman, Boston. 196 pp
Separation of supercritical slab-fluids to form aqueous fluid and
melt components in subduction zone magmatism. Proc Natl Acad
Sci USA 109:18695
Kawamoto T, Mibe K, Bureau H, Reguer S, Mocuta C, Kubsky S,
Thiaudière D, Ono S, Kogiso T (2014) Large ion lithophile elements
delivered by saline fluids to the sub-arc mantle. Earth Planets Space
66:61
Miletich R (ed) (2005) Mineral behaviour at extreme conditions,
European Mineralogical Union – notes in mineralogy, vol 7,
McLean. 488 pp
Oelkers EH, Schott J (eds) (2009) Thermodynamics and kinetics of
water-rock interaction, Reviews in Mineralogy and Geochemistry,
vol 70. Mineralogical Society of America and Geochemical Society,
Chantilly. 569 pp
Uts I, Glazyrin K, Lee KKM (2013) Effect of laser annealing of pressure
gradients in a diamond-anvil cell using common solid pressure
media. Rev Sci Instrum 84. 103904 pp