16.4.1 Rock relationships
To gain insight into the formation of terrestrial rocks other than basalts requires understanding the chemical relationships between the major rock classes. Figure 16.3 illustrates these rela- tionships, for the major igneous rocks only; recall from Chapter 8 that metamorphic and sedimentary rocks begin their existence as igneous rocks. The primary distinguishing feature among the igneous rocks is their silica (or SiO2) content. Felsic rocks (richer in SiO2) are light colored, relatively low density; mafic rocks are poor in SiO2, have proportionately more iron and mag- nesium, and are therefore higher density and darker color. There is a progression of rocks from felsic to mafic. This progres- sion occurs along two tracks, corresponding to whether the rock erupted from a volcano, and hence isextrusive, or instead, the melt stopped rising within the relatively low-density continental crust (hence was no longer buoyant), and cooled to solidification as anintrusiverock.
The major extrusive/intrusive equivalent rock types, in pro- gression of chemical composition from mafic to felsic, are basalt/gabbro, andesite/diorite, rhyolite/granite. The more com- mon felsic rock class is the intrusive granites; the more com- mon extrusive mafics are the basalts. In some of what follows, we rather loosely describe oceanic crust as basaltic, continen- tal crust as granitic, and crust produced at continental margins above descending slabs as andesitic. Mantle rock is even more mafic than basalt; an example of this kind of rock is called peridotite.
16.4.2 Seismic waves and composition
As a brief aside, the seismic P-wave velocity (Chapter 11) changes smoothly with composition; granites have aP-wave velocity of 6 km/s, gabbro and basalts 7 km/s, and peridotite 8 km/s. The seismic-wave explorations described in Chapter 11 for Earth show that the oceanic crust is basaltic (along with some gabbro) with no granites, and the upper mantle is consistent with an ultramafic composition like that of peridotite. Interestingly, the thick continental crust is not entirely granitic; theP-wave velocities suggest that some gabbro is present in the lowermost continental crust. In this dual nature of continental crust lies further clues to continent formation.
16.4.3 Role of water in partial melting
As Table 16.1 shows, a granite is made from a basalt by reducing the iron and magnesium contents while increasing the sodium and potassium abundances. Andesites are, compositionally, an
extrusive
pumice
hyolite dacite
scoria tuff and breccia
andesite basalt
granite granodiorite diorite gabbro peridotite dunite intrusive
c i f a m c
i s l e f
glassy
finer
crystalline coarser
Figure 16.3Relationship between selected major igneous rocks. Rock types are positioned according to the amount of silica they contain, and whether they are plutonic (intrusive) or volcanic (extrusive). Plutonic rocks do not breach the surface during eruptions, but cool and solidify buried in the crust. Because these rocks cool slowly, they are composed of large crystals. Volcanic rocks, which erupt to the surface, are cooled suddenly and have glassy textures, or very fine crystals.
intermediate step. The challenge lies in identifying the geologic environment in which large amounts of such chemical alterations can occur. Simply cycling the basalt back into the mantle is not sufficient to produce andesites and granites; the melting of dry basalt occurs at a fairly high temperature, thus fairly deep, in such a fashion that the liquid retains a composition quite similar to that of the original rock. (Recall that, because basalt is the solidified product of partial melting of mantle rock, it will melt under different conditions than the original mantle rock.)
What is required to make more felsic rocks is the addition of a material that will alter the melting relationship of the basalt, and water is an excellent candidate. Figure 16.2 shows that, in the presence of sufficient amounts of water, the melting point of the basalt drops dramatically. Further, the melting is partial, with larger ions again partitioning preferentially into the melt. The resulting liquid has an andesitic composition, partway between basalt and granite.
The physical environment in which water plays a role is sup- plied by plate tectonics (Figure 16.4). The basalt at the mid- ocean ridges solidifies as it approaches the surface. Because the ridges are underwater, the basalts react with the water, which becomes incorporated in the crystal structure of the rocks, hydratingthe minerals. The process can be very efficient because the midocean ridges are filled with cracks, through which water circulates in an intricate network of hydrothermal systems. Thus, much like the cells in our body receive nourishment through an intricate network of capillaries, oceanic basalts enjoy extensive and intimate contact with water.
As oceanic crust moves away from mid-ocean ridges, it cools and thus becomes denser than the rock beneath. Sitting on less dense rock is an unstable situation, and the crust eventually founders and sinks into the mantle below. This sinking, or sub- duction, can take place at the boundary of a continent or purely within an ocean basin. Basalt that was not fully hydrated at the
ridges now has a second chance as it is warmed up during its descent below the continental edge. However, temperatures con- tinue to climb as the subductingslabdescends, and eventually become too high for water to remain stably bound in the basaltic rock. The water becomes unbound from the minerals and, being buoyant, rises upward through the boundaries between mineral grains.
The release of water has two profound effects on the subduct- ing slab and the adjoining mantle. First, the slab rock becomes denser, aiding subduction. Second, the water rises above the slab into the wedge of mantle above; as it contacts basaltic and mantle grains, the melting point of the grains plummets, and partial melting of the slab and mantle begins at temperatures and pressures much lower than that which originally formed the basalts. The partial-melt products are less dense than the mantle and hence rise, though more slowly than the water because the molten rock is closer in density to the mantle itself. The partial melt, andesitic in composition, erupts onto the surface in the form of volcanic lavas; some may come to rest near the surface as igneous intrusions of diorites.
Andesitic volcanism, orarc volcanism, is common along the margins of active subduction zones, for example, along the Andes, Japan, and the northwestern United States. Some of these margins are actually disconnected from the adjoining con- tinent (as with Japan). Aside from an enrichment in sodium and potassium, the andesitic or dioritic rock is enriched in the large ions uranium and thorium. Because these elements, along with potassium, have relatively abundant, long-lived radioactive iso- topes, the effect of the low-temperature melting of the mantle at plate margins is to concentrate the heat-producing elements in the andesitic crust.
The reader may notice an interesting correspondence between the cycling of water into the mantle and that of carbon, described in Chapter 14. Indeed, in both cases, plate tectonic subduction
(a) Post-Archean growth of new continental crust
(b) Growth of new continental crust during the Archean
felsic volcano
sea level
magma chamber
first cycle sediments mid-ocean ridge trench
old oceanic crust
slab dehydrates
granite batholiths sedimentary
cover andesite
volcano
magma mantle wedge
zone of melting basalt magma
sea level
magma separation partial melting in rising column
mantle
mantle
mantle mantle
mature continental crust
sea level
km 25
50
75
km 25
50 upper crust
oceanic crust
tonalite (sodium granite) intrusion
zone of mantle melting?
zone of shallow dehydration and partial melting
lower crust of residual basic granulites following melting
and extraction of granites plume head of basalt
underplated beneath crust?
young hot oceanic crust
Figure 16.4View of the formation of continental crust: (a) today and (b) in the Archean. The size of the continent is truncated in the top panel;
only the edge and a portion of the interior continental shield are shown. At the continental margin, formation of andesite is fairly well understood.
In the interior, it is suggested that the formation of granites results from melting of more mafic rocks in the lower continental crust, with the granites rising to the top as batholiths. In the bottom panel, it is speculated that Archean plate tectonics involve rapid recycling of ocean crust and small continental minishields, which are the products of partial melting of subducting slabs – made possible by the higher temperatures of the Archean mantle. However, because hydration of the basalt (that is, chemical combination with water) is not complete, and Archean temperatures were quite high even at shallow depths, the melt product is not andesite but unusual granite-like “granitoid” rocks rich in sodium, including the so-called (tonalites). From Taylor and McLennan (1995).
provides a mechanism for recycling volatiles trapped chemi- cally in oceanic sediments (carbonates) or basaltic crust (water) back onto the surface. In the case of carbon, the recycling is key to sustaining a warming atmospheric greenhouse; in the case of water, its release from rock provides the key step in low-temperature partial melting of basalts and mantle to form andesites and diorites.
16.4.4 The puzzle of granite formation
The upper crust of the continents is not made predominantly of andesites or diorites. The granite of which it is composed is even further away in composition from basalt than are the andesites,
and hence must represent an additional cycle of differentia- tion. However, no obvious simple process exists by which such differentiation might occur. Although it generally is acknowl- edged that arc (subduction zone) volcanism – the production of andesites – is a principal means by which mantle material is converted to a buoyant state and accreted onto continents, the eventual conversion of that material to a truly granitic composi- tion remains something of a puzzle.
The currently favored picture for granite formation, shown in Figure 16.4, is that partial melting within the continental crust itself produces felsic rocks, such as granites and the gra- nodiorites (intermediate between granites and diorites), which rise to the surface in the form of large intrusive masses called
batholiths, leaving behind a residue in the lower continental crust. However, samples of the lower continental crust are not consistent with being simply a residue of such melting. These xenolithscontain trace elements, such as the rare-earths, whose composition is altered by the formation of granite. The abun- dance pattern seen in these elements in granites versus xenoliths requires that granite formation be more complicated than simple melting and differentiation of the lower continental crust.
One way to explain the pattern is to invoke basaltic mag- mas beneath the continental crust that rise and plate (or essen- tially stick to) the bottom of the continental crust. These hot plumes might trigger episodes of deep continental melting and consequent differentiation, while contaminating the lowermost continental crust so as to produce the observed composition of xenoliths. The apparent presence of gabbro in lower continen- tal rock, based on seismic data, is consistent with this idea.
Such a model must be tentative at least, because xenoliths may or may not be representative of lower continental rock. More widespread samples of metamorphic rocks called granulites, which appear to have formed under high pressure and temper- ature, may have originated in the lower crust as well; however, there is no consensus on whether these are truly rocks from the lower continental crust.
The difficulty in understanding formation of granitic rock is in large measure a result of the very complex nature of the con- tinents. Unlike the ocean floor, with its simple geology that is erased on 100-million-year timescales, the continents are cumu- lates of geologic processes stretching over billions of years.
To understand how this process began requires examining the nature of rocks from the Archean time of Earth’s history.