Crystal Growth in Geology: Patterns on the Rocks

Full text

Provided for non-commercial research and educational use only.
Not for reproduction, distribution or commercial use.
This chapter was originally published in the book Handbook of Crystal Growth, Volume
II. The copy attached is provided by Elsevier for the author's benefit and for the
benefit of the author's institution, for non-commercial research, and educational use.
This includes without limitation use in instruction at your institution, distribution to
specific colleagues, and providing a copy to your institution's administrator.
All other uses, reproduction and distribution, including
without limitation commercial reprints, selling or
licensing copies or access, or posting on open
internet sites, your personal or institution’s website or
repository, are prohibited. For exceptions, permission
may be sought for such use through Elsevier’s
permissions site at:
http://www.elsevier.com/locate/permissionusematerial
From García-Ruiz JM, Otálora F. Crystal Growth in Geology: Patterns on the
Rocks. In: Nishinaga T, Rudolph P, editors. Handbook of Crystal Growth, Vol. II.
Elsevier; 2015. pp. 1–43.
ISBN: 9780444633033
Copyright © 2015, 1993 Elsevier B.V. All rights reserved.
Elsevier
Author's personal copy

1
Crystal Growth in Geology:
Patterns on the Rocks
Juan Manuel García-Ruiz, Fermín Otálora
INSTITUTO ANDALUZ DE CIENCIAS DE LA TIERRA, LABORATORIO DE ESTUDIOS
CRISTALOGRÁFICOS CSIC-UNIVERSIDAD DE GRANADA, GRANADA, SPAIN
CHAPTER OUTLINE
1.1 Introduction ..................................................................................................................................... 1
1.2 Geological Scenarios for Crystal Growth..................................................................................... 4
1.2.1 Magmatic Environments.......................................................................................................5
1.2.2 Metamorphic Environments................................................................................................. 8
1.2.3 Hydrothermal Environments ................................................................................................ 9
1.2.4 Sedimentary Environments ................................................................................................ 10
1.2.5 Biomineralization and Biologically Induced Crystallization............................................ 11
1.3 Deciphering Geological Information from Crystal Morphology.............................................. 12
1.4 Decoding Polycrystalline Textures from Nucleation and Growth........................................... 16
1.5 The Case of Giant Crystals........................................................................................................... 21
1.6 Decoding Disequilibrium Mineral Patterns ................................................................................ 24
1.6.1 Compositional Zoning......................................................................................................... 25
1.6.2 Liesegang Structures........................................................................................................... 28
1.6.3 Fractal Dendrites ................................................................................................................. 30
1.7 Early Earth Mineral Growth, Primitive Life Detection, and Origin of Life............................. 32
1.8 From Deep Earth to Outer Space................................................................................................ 36
Acknowledgments ............................................................................................................................... 39
References................................................................................................ ............................................. 39
1.1 Introduction
Rock-forming minerals are mostly crystalline. In most cases, rock minerals are micro-
metric in size and are irregularly embedded within other minerals, so they are not really
attractive to the naked eye. However, in a few cases, minerals display their geometrical
beauty as isolated single crystals of several centimeters or even larger in size. The
geometry of crystals (as those shown in Figure 1.1), which is so different than any other
Handbook of Crystal Growth. http://dx.doi.org/10.1016/B978-0-444-63303-3.00001-8 1
Copyright ©2015 Elsevier B.V. All rights reserved.
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

natural object, triggered the curiosity of rational thinkers such as Theophrastus, who
classified minerals by their physical properties [1]; Pliny the Elder, who wrote what was
for centuries the most extensive and rational approach to minerals [2]; Kepler, who
envisaged the internal order of ice crystals [3]; Biringuccio, who studied metals; Stenon,
who discovered the morphological properties of quartz crystals [4]; and Rome
´de L’Isle,
who demonstrated unambiguously the law of constancy of angles [5].
The science of crystals began historically with the study of minerals, and the same
goes for the study of crystals’ formation and growth. In his prodromus, Nicolas Steno
provided a rational analysis of crystal growth and concluded that crystals do not grow by
germination or vegetative growth, as thought at his time, but by accretion of growth
units. Until the late nineteenth century, most crystallographic studies were devoted to
the morphology and symmetry properties of single and twinned mineral specimens [6];
the main crystal growth challenge was the synthesis of natural crystals in the laboratory,
from common minerals to precious stones [7]. The progress of these studies during the
twentieth century powered the emergence of a huge industry related to crystal growth,
not only for gems but also for single crystals for new technologies [8a,b].
Most of the chapters in this book deal with the study of the formation of crystals in
the laboratory or industry. In those cases, it is possible to obtain reproducible and
comprehensible data about the composition of the mother phase under highly precise
and stable conditions of pressure, temperature, and compositional purity, thus providing
FIGURE 1.1 Minerals have played a main role in the history of crystals and crystal growth. The morphology of
crystals found in nature was the most astonishing feature to explain, the main challenge. The law of constancy of
angles rediscovered up to four times until his definitive discovery by Rome de L’Isle- and the concept of the
“molécule intégrant”of René Haüy set the basis for the development of morphological crystallography in XIX
century. Minerals were also the targets that triggered the development of the first crystallization technology to
grow synthetic crystals. Natural samples of minerals: (a) Galena, (b) Calcite, (c) Pyrite, (d) Quartz.
2 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

feedback about the growth conditions and products of the crystallization process. In this
chapter, we face a completely different problem. Here, the crystals to be studied already
formed during the long history of this planet, a history that dates back 4500 million years.
Except for the contemporary surface or near-surface geological scenarios, the compo-
sition of the mother phase or the crystallization conditions of those crystals is not known
and the temperature, pressure, and composition are not controlled. These parameters
change with time, sometimes even suddenly. A large number of reactions proceed
simultaneously in a complex, multicomponent fluid, including chemicals that modify
crystallization behavior in many nonlinear ways. In addition, these processes are usually
difficult (or even impossible) to observe, particularly those occurring at very high tem-
peratures and pressures, and the typical timeframe for geological processes spans a
range from a few seconds to a few million years.
In mineral growth studies, there is nothing but crystals. In fact, the main contribution
expected from crystal growth to this field is to decode the information contained in these
crystalline rock-forming minerals to reveal the physicochemical scenario from which
they grew—that is, to help uncover the geological history of this planet. This is not an
easy task because the relevant information derived from classical crystal growth studies
is rather scarce. We have adequate knowledge on crystallization to control the nucleation
flow and crystal growth rates at the laboratory and industrial scales. However, we still
have difficulties understanding in detail the variables that control crystal growth
morphology and crystal defects, which are the crystal properties most used to under-
stand mineral growth thanks to the pioneering work of Sunagawa [9]. Therefore, we need
to continue the analysis and interpretation of single crystal features—morphology, im-
purities, inclusions, surface topography, and isotopic signature—to better understand
their geological context. However, it is very important to note that minerals almost never
appear as single crystals in rocks. They only rarely appear as the beautiful well-faceted
single or twin crystals displayed in museums and sold in mineral shops. In most
cases, minerals appear on forming rocks as polycrystalline materials [10]. Unfortunately,
the information on polycrystalline crystallization was only interesting for metallurgists,
so it is scarce and restricted to alloys. Crystal growth studies will only be useful for earth
and planetary sciences if there is an extensive exploration of pattern formation, such as
how minerals self-assemble into peculiar structures and how they do so when crystal-
lizing massively from seas and lakes, from magma and vapor, and how they experience
phase transitions during their spatiotemporal trip triggered by the slow but continuous
dynamics of the planet.
All of these circumstances and drawbacks make it very difficult to understand crystal
nucleation and growth in geological media; however, at the same time, they make
crystals richer in behavior and products and open up the possibility of performing
studies at the frontier of science and technology. It is essential to push a new research
program in crystal growth beyond classic studies because novel theoretical and exper-
imental tools are necessary to understand what mineral crystals can tell us about the
past of this planet, as well as its current inner and outer dynamic; they will also allow us
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 3
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

to properly analyze the results of our incipient exploration of solar planets and extra-
terrestrial images. This chapter reviews the current status and the future trends of this
fascinating research program.
1.2 Geological Scenarios for Crystal Growth
Minerals form in a variety of different geological environments, such as solutions, melts,
or vapor, as well as by solid–solid transition. Therefore, it is convenient to analyze the
different geological scenarios in order to understand how they work in terms of crys-
tallization. Figure 1.2 shows the main processes in the Earth’s crust. The solid arrows
indicate processes of crystallization, whereas the dashed arrows indicate dissolution or
melting; the dotted arrows correspond to the transport of material. The two gray rect-
angles contain the fluids from which crystallization proceeds in the majority of
geological crystallization processes: magmas and solutions of different origin and
properties. For simplicity, the atmosphere is not included, although it plays a very
important role, along with life, in processes such as the precipitation of carbonates.
Igneous rocks are formed by the crystallization of magmas, either in the interior of the
crust (plutonic rocks) or at the surface (volcanic rocks). Crystallization produces frac-
tionation and liquids with increasing concentration of metals and other ions that do not
enter into the crystal lattice of the crystallized minerals. Finally, a high-temperature
solution remains, which gives rise to hydrothermal deposits and contributes to the
composition of surface waters through volcanoes and smokers. All types of rocks at the
Earth’s surface undergo weathering owing to the interaction with meteoric waters,
the atmosphere, and living organisms. This produces fragments of rocks (clasts) and
dissolved species in the surface waters that, after precipitation, can combine with the
FIGURE 1.2 Sketch of the relation between the main crystallization/dissolution processes in the Earth’s crust that
produce minerals from solutions or melts. Solid line arrows indicate processes of crystallization, dashed line
arrows indicate dissolution or melting and dotted line arrows correspond to transport of material. The two gray
rectangles contain the sources from which crystallization proceeds in virtually all geological crystallization
processes: magmas and solutions having different origin and properties. For simplicity, the effects of atmosphere
or life are not included, although they play an important role in processes like the precipitation of carbonates.
4 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

clasts into sediments. Compaction, crystallization, and recrystallization of minerals
during diagenesis transform these sediments into sedimentary rocks. The minerals of
any rock can become unstable with temperature and pressure changes. This can lead to
dissolution, recrystallization, and nucleation of new phases during metamorphism.
Sedimentary rocks, forming at the surface and later buried by new sediments, typically
suffer modification by changes of pressure, temperature, and redox potential.
1.2.1 Magmatic Environments
Crystallization from the melt, as in the case of Czhochralsky growth of semiconductors,
is characterized by high temperature, very condensed systems, small or no interaction
with solvents, and continuous growth of rough interfaces. However, this only applies to
the solidification of single phases from a melt of their components. Magmatic crystal-
lization normally implies that the growing crystals are in contact with a complex
multicomponent melt that acts as a solvent. Crystals in most magmatic solidification
scenarios grow by layer growth, not by continuous growth; therefore, faceted crystals are
far more common than rounded or dendritic forms. For these reasons, magmatic min-
eral growth can be considered as high-temperature, high-viscosity solution growth with
silicates and volatile components acting as solvents, rather than melt growth [11]. Due to
the high viscosity, mass transport is slow and growth kinetics are controlled by the
transport of latent crystallization heat rather than by mass transport, although mass
transport may become very important for several processes.
Magmas can be either generated from materials in the mantle, such as peridotitic
magmas, or from the melting of igneous or metamorphic rocks, typically in subduction
zones. Crystallization of magmas upon cooling deep into the crust is slow, starting with
crystals having the highest fusion temperature among all the possible minerals that can
be nucleated from the composition of the magma. This first crystallization produces a
melt that is enriched in the component not present in the first crystals produced, which
eventually gives rise to new minerals with a lower fusion temperature. As a consequence,
the composition of the melt evolves with time. This evolution, along with the movement
of the magma within the crust, differentiates the various types of igneous rocks and
magmas that contain the remaining components. The large number of degrees of
freedom characterizing a crystallizing magma can be summarized in the form of a phase
diagram containing the most relevant chemical species and divided into a series of
stability fields for different minerals. The crystallization of one of them—the most stable
one for the composition of the magma—moves the position of the magma within this
diagram, eventually entering the stability field of a different mineral that starts crystal-
lizing. This trajectory around the phase diagram is called the solidification path, which
leads to magma differentiation in the sense of the classic Bowen’s reaction series. The
minerals present in the resulting rock and the textural relations between them encode
the solidification path, which can be investigated looking at the phase relations in the
phase diagram. The main variables to be considered when dealing with mineral growth
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 5
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

in magmatic environments are the latent heat value for each of the crystallizing minerals,
the nucleation, growth thermodynamics and kinetics, the rate of convective flow, the
temperature of the melt, and the constraints (e.g., the thermal contrast between magma
and country rock) if the temperature is not constant.
Along with metamorphic media (see below), crystal growth in magmatic environ-
ments is difficult to study experimentally because of the high temperatures and pres-
sures involved, and it is almost inaccessible in in-situ studies, although some
experiments have been successfully performed [12,13]. Most of the experimental in-
vestigations in igneous petrology have studied magmas at equilibrium, with the aim of
obtaining thermodynamic data on mineral/melt and mineral/mineral partition co-
efficients to study the genesis of igneous rocks by analytical methods on geological
samples. Unsteady systems have also been studied to investigate the crystallization of
magmas upon cooling in closed (magmatic chambers) [14] and open (magmatic ocean)
[15] conditions. Experiments in which magma was cooled to reproduce crystallization
conditions [16] have been used to investigate crystal growth in magmatic environments,
particularly the dependency of crystal growth rates on temperature and convective mass
transport, the texture, crystal shape and composition as a function of cooling rate, and
the development of compositional zonation.
Upon the nucleation of crystals in the magma and after a highly transient stage,
crystallization tends toward equilibrium between heat production (latent heat release)
and heat loss. Because the main controlling factor is the fast release of latent heat during
the nucleation and the first stages of fast growth, nucleation usually is the rate-limiting
process, occurring as sharp pulses, producing thermal oscillations, and followed by
longer periods of crystal growth. The main parameters controlling crystal size distribu-
tion, especially close to the margins of cooling igneous bodies, are the initial thermal
conditions and the nucleation and growth kinetics of the main minerals [16]. Crystal size,
number, and shapes resulting from crystal growth kinetics have important consequences
on the fractionation of magmas and their rheological properties. Fractional crystalliza-
tion occurs when crystals can freely sediment within the magma and accumulate at the
bottom of the magmatic chamber, leaving the fluid depleted in the chemicals entering
their composition, which leads to a chemical evolution of magmas. This process is
responsible for the differentiation of the light minerals making up the Earth’s crust. On
the other hand, nonfractional crystallization implies differentiation within the melt
percolating the already-crystallized solid matrix; the crystal size distribution determines
the degree of differentiation in this regime due to the competition between percolation
and solidification rates. The crossover between fractional and nonfractional crystalli-
zation, and the chemical and mineralogical evolution of the magma in both cases, are
controlled by the crystal size and shape [15], which are the macroscopic output of the
nucleation and growth kinetics operating in the magmatic body.
The rheological properties of the lava that produces volcanic rocks are also controlled
by crystal nucleation and growth kinetics. The content and shape of the crystals deter-
mine the formation of basic lava crusts and their rheological properties [17]. In volcanic
6 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

rocks, and in general in magmatic rocks for which temperature and pressure can change
quickly during the growth of crystals, the crystallization conditions are usually encoded
in the crystal compositional zoning, morphology, and texture, either in the different
parts of large crystals (phenocrystal cores crystallized deeper than their rims), or in
different crystals of the same composition (microphenocrystals and microlites formed at
subaerial conditions) [14].
Many natural rocks contain clear evidence of disequilibrium, such as compositional
zoning, because they have formed from magmas crystallizing along nonequilibrium,
time-dependent pathways [12,18]. The main contribution of crystal growth to igneous
petrology is probably the interpretation of the features related to mineral growth during
the time-dependent evolution of these igneous systems. Changes in temperature,
pressure, and composition of the magma do modify crystal growth kinetics, so the
history of the rock is recorded in the disequilibrium geochemical features and the
petrographic texture of the rock. An example of this approach to decoding the genetic
information contained in igneous rocks is the application of growth rate studies into
zoning profiles in plagioclase crystals. This problem has been modeled and tested
against experimental data [19]. Recent models, including nonlinear multicomponent
diffusion equations, show that growth rate and crystal composition may vary stepwise,
even under linear cooling of the magma [20].
The nonequilibrium features of crystal growth are relevant for the fine-tuning of the
thermodynamic indicators used in petrology. Crystal geothermometers and geo-
barometers are extensively used in igneous and metamorphic petrology. They are based
on the temperature-dependent partition coefficient of a component or a chemical
species between coexisting phases, with at least one of them being a crystal, such as
clinopyroxene-melt, orthopyroxene-melt, or orthopyroxene-clinopyroxene. The process
of impurity trapping depends on the thermodynamics of lattice-strain accumulation
within the crystal lattice, as well as on the growth kinetics and the crystal growth
mechanism, which makes them sensitive to changes in both the overall growth rate [21]
and the relative growth rate of different crystal faces (i.e., to the morphology and
sectorial structure) [22]. These features can hinder the use of geothermometers or make
them less accurate, but they also open the possibility for advanced experimental
methods, such as using single-crystal thermometers in which the partition coefficient
between different growth sectors is used as a geothermometer or geobarometer, thus
avoiding any assumption about the thermodynamic equilibrium between two phases in
contact.
Extreme examples of crystal growth are found in pegmatites, which are very coarse-
grained intrusive igneous rocks made of interlocking crystals, from millimeter size up to
large crystals of tens of meters in length. The genesis of pegmatites is still controversial,
and there is not a model explaining all the features of these rocks. The most accepted
model is the formation of granitic pegmatites from residual melts derived from the
crystallization of granitic plutons. Shallow-level pegmatites cool much more rapidly than
previously believed [23]. This is consistent with the disequilibrium crystal growth
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 7
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

features observed in these rocks, including skeletal, radial, and graphic morphologies;
strong compositional zoning patterns; sharp changes in grain size; and anisotropic,
oriented fabrics such as comb structures. These features seem contradictory to the ex-
istence of large crystals, which in classical petrology textbooks are identified as having
slow-cooling rates. The exceptional textural properties of these rocks are most probably
due to the uncommon composition of the residual magmas from which they grow. After
magmatic differentiation, these magmas are enriched in rare, incompatible components,
and they have high concentrations of fluxes and volatiles. In these conditions, one can
expect anomalously high mass transport rates and solubility and anomalously low
crystallization temperatures, nucleation rates, and melt polymerization [24]. The joint
effect of these anomalies leads to wide metastability ranges and rapid depletion of the
high-mobility growth units, which can explain the development of large crystals due to
inhibited nucleation rates and enhanced growth rates.
1.2.2 Metamorphic Environments
Metamorphic rocks undergo changes in texture, mineralogy, and chemical composition
through partial or complete recrystallization of their minerals, either solid-state or
solution-mediated [25–27]. These rocks remain essentially solid during metamorphism,
but they can flow in a plastic-like manner due to differential pressure. The main factors
driving metamorphism are temperature, pressure, the presence of fluids, and the
composition of the original rock. Temperature drives the chemical changes that result
in the recrystallization of existing minerals, destabilization of previous ones (particu-
larly hydrated minerals), or the nucleation and growth of new phases. Many crystals
will grow larger than they were in the parent rock. Pressure, like temperature, increases
with depth and contributes to changes in a rock’s mineralogy and texture. Confining
pressure (in all directions) causes the spaces between mineral grains to close, pro-
ducing a more compact rock. Directed pressure, on the other hand, guides the shape
and orientation of the new metamorphic minerals, producing plate-like crystals
extended in the directions perpendicular to the pressure gradient. Fluids (mostly water
but also other volatile components, such as carbon dioxide) play an important role in
metamorphism, dissolving and transporting ions along the boundaries between crystals
or catalyzing some reactions during metamorphism. Clay minerals can contain up to
60% water in their crystal structure, which is mobilized when these minerals are
transformed during metamorphism. Apart from this loss or accumulation of volatiles,
most metamorphic rocks have the same overall composition as the parent rock from
whichtheywereformed.
Metamorphism creates new textures on the rock, and it is the realm of recrystalli-
zation and transformation of some of the original minerals into new ones that grow at
high temperatures and pressures. Metamorphic rocks can exhibit great variation in
crystal size; in general, the size of crystals increase as the grade of metamorphism in-
creases. During the recrystallization process, certain metamorphic minerals, including
8 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

garnet, staurolite, and andalusite, tend to develop a few large crystals [28,29]. In contrast,
minerals such as muscovite, biotite, and quartz typically form a large number of small
crystals. The mineral grain shape and orientation is also altered: changes in orientation
lead to foliation, whereas the most relevant effect of crystal size increases in some
minerals is the development of granoblastic textures. Foliation can develop by solid-
state plastic flow because of the intracrystalline movement of lattice defects within
each grain or due to the dissolution of crystals from areas of high stress, transport along
the intercrystalline surface, and deposition in low-stress areas [30,31]. Granoblastic
rocks, such as quartzite, have a massive or coarse granular appearance that is produced
by recrystallization during metamorphism [32,33] and exhibit no directional deforma-
tion. They are composed mainly of crystals that grow in equidimensional shapes in
restricted spaces. Aside from quartzite, the most common granoblastic rock is marble,
resulting from the metamorphism of limestone or dolostone leading to a complete
recrystallization of the original rock into a polygonal interlocking mosaic of calcite,
aragonite, and/or dolomite crystals [34]. Metamorphic rocks are classified by their
metamorphic grade, which is defined as the maximum temperature and pressure to
which the rock was subjected. However, metamorphism is a dynamic process, and a
metamorphosed rock may have a very complex history. This history influences the
successive stages of crystal growth or dissolution; therefore, it is encoded in the crys-
tallography, chemistry, and texture of the minerals that grow during metamorphism (i.e.,
in the properties of the crystals making up the rock). For example, by studying the
compositional zoning of the growth rims or the inclusions into crystals of garnet, it is
possible to reconstruct complex metamorphism sequences [35].
1.2.3 Hydrothermal Environments
After the solidification of magmas, high-temperature hypersaline solutions enriched in
elements incompatible with magmatic minerals remain close to the magmatic chamber
and flow through the pores or fractures, provoking leaching of rocks and the precipi-
tation of some of the most important ore deposits. The growth of hydrothermal minerals
occurs in relatively open spaces and with a continuous supply of ions from solution,
which produces large-faceted crystals that stand out among the most beautiful crystal
specimens in mineralogical collections; some of them show very interesting growth
structures, such as the quartz specimens from the Alpine veins [36]. Hydrothermal
systems comprise solutions of magmatic origin but also meteoric waters, which can
account for up to 95% of the waters transported during cooling of an intrusion in the
shallow crust [37]. Hot springs and volcanic fumaroles are the active, observable
equivalents to these systems.
Compositional zonation is very common in hydrothermal crystal growth due to the
open character of the system with continuous restoring of solution. This zonation
records information on the hydrothermal processes. An interesting example is the
zonation patterns of the andradite-rich garnets in a contact aureole in Drammen
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 9
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

granite, which record at least five intermittent growth periods related to changes in the
oxygen fugacity [38].
1.2.4 Sedimentary Environments
Crystallization from low-temperature aqueous solutions in sedimentary or diagenetic
environments has traditionally been a prolific field of collaboration between geology and
crystal growth, mainly because of the ease for laboratory experimentation and the
widespread distribution of sedimentary rocks. The main topics of this collaboration have
been carbonates, evaporates, and clay minerals. The studies have focused mostly on the
morphology of crystals in relation to growth conditions [39] and the use of morphology
as a geological indicator [40].
Three main scenarios are common in solution growth studies in geologic media: (1)
the crystallization of minerals from free superficial brines, (2) the precipitation of
minerals in the pore space in sediments, and (3) the crystallization and transformation
of minerals during diagenesis. The temperature as well as the pressure increase from
scenario (1) to scenario (3), while the volume of solution and the effect of living or-
ganisms decrease. These differences obviously influence the thermodynamics and the
kinetics of mineral growth, but also the degree of equilibration with the hosting
sediments and the morphology of the resulting crystals. The driving force for crys-
tallization is produced in surface waters by evaporation or by the mixing of solutions
of different compositions. These processes are usually fast, so nucleation and growth
rates tend to be high in comparison with diagenetic environments, where the super-
saturation develops because of the destabilization of minerals and slow mixing of
phreatic waters. The effect of living organisms is also mostly restricted to superficial
solutions, although investigations suggest a deep biosphere that is more active than
thought [41].
The range of temperatures and pressures involved makes the laboratory study of
these crystal growth processes possible; therefore, sedimentary environments are the
most-often addressed environment for experimental studies in mineral crystal growth.
Abundant literature is available on the thermodynamics, kinetics, and mechanisms of
crystal nucleation, growth, and dissolution, as well as the effect of mass transport and
the morphology of the grown crystals.
Nucleation and crystal growth play a fundamental role in the deposition of chemical
and biochemical sediments, as well as the compaction and cementation of sedimentary
rocks. Chemical and biochemical processes in the sedimentary environment lead to the
precipitation of minerals such as calcite, gypsum, halite, or apatite and to the formation
and deposition of calcareous or siliceous plant or animal parts, such as shells. The rocks
resulting from these precipitation processes are limestones, cherts, evaporates, or
phosphorites. During the burial diagenesis of any type of sediment, authigenic minerals
(e.g., quartz, feldspars, clay minerals, calcite, gypsum, hematite) precipitate, compacting
the rock and partially changing its chemical composition.
10 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

Carbonate rocks account for 20–25% of the sedimentary record. They are mainly
made of calcium carbonate (mostly calcite and aragonite) or calcium magnesium car-
bonate (dolomite) in many different forms, from chemically precipitated fine lime mud
to small grains (peloids) generally formed by biological activity to skeletal grains or
carbonate-coated grains such as ooids. The texture of these constituents depends on the
genetic processes involved in their precipitation. For instance, when ooids [42] are made
of randomly oriented aragonite crystals, a marine origin is indicated. On the other hand,
the presence of radially distributed aragonite crystals can indicate marine, lacustrine, or
hypersaline environments. Tangentially distributed aragonite crystals in an ooid are only
possible in high-temperature environments. If the crystals are made of calcite, ooids are
of nonmarine origin; a tangential distribution of the crystals indicates that they formed
in a caliche, while radial or random crystal orientations can form in lakes and radial
distributions are only possible in caves of fluvial deposits. “Sparry calcite” is the term
used for large crystals formed during diagenetic cementation. They usually have rich
textures: granular fillings of voids, fibrous overgrowths of grains, and even large
monocrystalline overgrowths on echinoderm fragments. Carbonate crystallization
leading to carbonate rocks in sedimentary environments occurs at close to ambient
temperatures (typically 0–40 C) from sea (salinity 3.0–4.5%) or continental waters.
Crystallization in porous media can be successfully mimicked by crystal growth in
gels [43,44]. Actually, gels such as silica gels, agarose, or the use of other substrate (e.g.,
clays, bentonites, sands, muds) that reduce convective mass transport can be used as
laboratory-analogous crystallization [45,46]. The technique has great potential, but the
complex spatiotemporal pattern of supersaturation, ratio of concentration of reactants,
and speciation of reactants, among other factors (see Section 1.6.2) makes analysis of the
experiment very demanding and difficult to extrapolate to the natural-analogous.
1.2.5 Biomineralization and Biologically Induced Crystallization
The impact of life on the nucleation and growth of minerals is qualitative and quanti-
tatively important. First, there is a direct impact on mineral formation because living
organisms have managed a way to produce minerals themselves, the so-called bio-
minerals. Everything—the nucleation, crystal phase, crystal morphology, texture, and
growth mechanisms of biominerals—was dramatically affected by biological molecules
and macromolecules, as it is today. Secondly, beyond biomineralization, there is an
indirect but important influence of life on the chemical history of the ocean and at-
mosphere. An evident example is the impact that photosynthetic bacteria had on the
oxygenation of the hydrosphere and then on the atmosphere of the planet, starting
2.5 billion years ago [47]. However, this chapter will not deal with the growth of bio-
minerals; instead, we refer the reader to Chapters 19 and 20 in Volume IB and Chapter 31
of Volume IIB in this book. We will only discuss the problem of life detection and the
morphogenesis of microstructures and macrostructures that have an ambiguous origin,
either biologically or geochemically induced.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 11
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

1.3 Deciphering Geological Information
from Crystal Morphology
The feature of the minerals most often used in geological studies aimed at decoding
genetic information during crystal growth is crystal morphology. The external
morphology of a crystal is the result of its growth history; fortunately, this history (the
evolution of the crystal morphology during crystal growth) is accessible through the
zonal and sectorial structure of the crystal. Changes in the composition, temperature, or
concentration of impurities in the fluid phase in contact with the crystal produce dis-
tortions in the crystal lattice, which result in greater incorporation of impurities or in-
clusions (either liquid or solid), changes in the composition of the crystalline phase, or
accumulation of stress or nucleation of crystal defects. All of these defective lattice
volumes, called the zonal structure, appear within the crystals at the positions where the
crystal surface was at the time of the change in the fluid phase; they allow the recon-
struction of the morphological history of the crystal during growth.
Faceted crystals are made of pyramidal volumes left behind on a single face
during the growth of the crystal. Each such pyramid is called a sector.Thedistri-
bution and properties of sectors is called the sectorial structure of the crystal. Stress
accumulates into the surfaces joining two nonequivalent sectors due to the aniso-
tropic growth and impurity accumulation kinetics, producing slight misfits at the
edges of the crystal (a common place for the accumulation of defects). The presence
of these defects may render may render the sectorial structure visible. Both the zonal
and the sectorial structure of crystals, when observable, make it possible to study the
growth history of crystals.
As in any type of crystal, mineral morphology is a function of the supersaturation at
which the crystal grew (Figure 1.3). This is due to the different growth mechanisms as a
function of supersaturation or supercooling. At very low supersaturation, crystals grow
by screw dislocations and develop slightly convex faces containing dislocation hillocks
on an essentially flat surface. Above a given supersaturation, the dominant growth
mechanism is two-dimensional nucleation on the crystal faces, which are flat with some
roughness where two-dimensional (2D) islands are growing. This growth mechanism
becomes unstable at higher supersaturation values due to the higher supply of growth
units at the edges and vertices of the crystal, which start growing faster, producing the
concave crystal faces typical of hopper crystals, but are mostly flat. Further increases of
supersaturation make continuous growth the dominant process and the surface be-
comes rounded, giving rise to the unstable growth of dendrites, spherulites, and bow-tie
morphologies. Even higher supersaturation values produce fractal aggregates and
coagulation textures.
Crystals grown at variable supersaturations can show their morphological history in
their sectorial and, more likely, zonal structures. The inset in Figure 1.3 shows the zonal
structure of a crystals that, after growing for some time at high supersaturation as
dendrite, underwent a change of growth conditions to a lower supersaturation; it
12 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

gradually changed morphology to a hopper crystal and then a polyhedral shape with flat
faces. After growing in this regime for some time, further decreases of supersaturation
made the fluid undersaturated with respect to the crystal, provoking dissolution of the
crystals with rounded edges. After this period, the supersaturation raised again to
moderate values, and the crystals restarted their growth until their final size.
Figure 1.3 shows a cubic crystal. However, in general, supersaturation at the crystal
surface is controlled by the specific surface energy, depending on the structure of the
crystal surface and, consequently, is anisotropic. The relative growth rate of different faces
in the crystal changes with supersaturation; this change leads to changes in the crystal
habit and to a supersaturation-dependent growth morphology. An extreme example of
this fact is ice: the faces parallel to the hexagonal axis grow very fast by continuous growth
under atmospheric conditions, while the faces perpendicular to this axis grow slowly layer
by layer, producing the well-known flat, dendritic snowflakes. In general, the observed
morphologies of minerals are growth morphologies, which deviate from the equilibrium
FIGURE 1.3 Growth rate and surface growth mechanism as a function of supersaturation. Vertical dashed lines
indicate the crossover between the spiral growth, 2D nucleation and Normal growth regimes. The typical
morphology of the crystal surface is sketched for spiral and 2D nucleation mechanisms. The overall morphology of
the dendritic crystals and the spherulitic aggregates is also shown at the supersaturation values where they
develop. The inset at bottom right shows consecutive morphologies of a crystal grown at decreasing (and later
increasing) supersaturation illustrated by the gray “c”-shaped curve. The crystal displays a morphological
sequence from dendritic crystal to hopper crystal and then a faceted morphology followed by a dissolution period
(rounded edges) when supersaturation falls below 0, and finally a new growth period at low supersaturation. The
consecutive morphologies can show in the zonal structure of the crystal.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 13
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

or structural morphology, resulting from the bonding energies of each face; they can be
computed using methods such as the Hartman and Perdok PBC (Periodic Bond Chain)
theory [48].
Differential changes of the growth rate as a function of supersaturation among
different faces of the same crystal can be studied through the zonal and sectorial
structure of the crystal to get information on the compositional, thermal, or flow history
of the environment. Figure 1.4 shows sketches of the growth morphology of a simple
cubic crystal with two families of faces: {100} faces (those vertical or horizontal) and
{101} faces (the diagonal ones). Zonal structure is shown as solid lines and sectorial
structure as dashed lines. The growth rate of the {100} faces is constant for all cases. The
three sketches on top show cases in which the {101} growth rate is also constant.
Figure 1.4(a) shows the morphological evolution when the {100} and {101} growth rates
are equal, leading to an octagonal shape of increasing size. In Figure 1.4(b), the growth
rate of the {101} faces is constant and 20% smaller than that of {100}. The slowest faces
{101} are more developed, but the morphology (the relative development of each face) is
preserved over time. The opposite situation, with the growth rate of {101} being 20%
larger than the growth rate of {100}, is illustrated in Figure 1.4(c). Again, the slowest
faces—those corresponding to the form {100} in this case—are more developed, but the
relative size of faces is preserved along the growth history.
The sketches on the bottom row show the cases in which supersaturation changes
with time. In all of them, the {100} growth rate is assumed to be constant (i.e.,
FIGURE 1.4 Morphological evolution of crystals growing at constant (top row) and variable (bottom row)
supersaturation. (a) Development for R
100
¼R
101
(R
hkl
being the growth rate of the (hkl) face). (b) 0.8R
100
¼R
101
.
(c) R
100
¼0.8R
101
. (d) R
100
¼constant, R
101
¼0.77R
100
–1.40R
100
(linear). (e) R
100
¼constant, R
101
¼1.40R
100
–0.80R
100
(linear). (f) R
100
¼constant, R
101
¼1.00R
100
–0.75R
100
–1.00R
100
.
14 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

independent of supersaturation). In Figure 1.4(d), the growth rate of {101} linearly in-
creases from 77% that of {100} (leading to diamond shaped crystals in the first stages) to
140% that of {100} (leading to square-shaped crystals in the last stages). The opposite
situation is shown in Figure 1.4(e), where the growth rate of the {101} faces linearly
decreases from 140% that of {100} at the beginning to 80% by the end of the growth
history. Finally, Figure 1.4(f) shows the morphological output of a more complex situ-
ation where the growth rate of the {101} faces first decreases smoothly from 100% of the
growth rate of {100} to 75% of this value by the fifth outline; it then increases smoothly
back to 100% the growth rate of the {100} face by the last stage.
As the number of crystals increases at higher nucleation rates, the interrelation be-
tween different crystals—that is, their relative orientation, size distribution, geometrical
interrelations, and grain boundaries—start to play an important role, restricting the
growth of crystals and their morphology. This collective morphological control by the
ensemble of crystals in the system can be the most determinant aspect of mineral growth
in cases such as the development of granoblastic textures (Figure 1.5(a)), made by an
interlocked, space-filling set of crystals of similar size. This texture is characteristic of
marbles and quartzites and develops by regrowth of a large density of crystals. When the
growth is only restricted in two spatial dimensions but is unrestricted in the third one,
competitive growth develops crystal aggregates; characteristic parallel, columnar tex-
tures of partially oriented crystals also develop as an emergent property (Figure 1.5(b)).
This kind of texture is typical of hydrothermal veins, geodes, and druses. Crystals ori-
ented with faster growth rates in the direction allowing unrestricted growth will overtake
their less favorably oriented neighbors, thus stopping their growth. Competitive growth
selects the crystals that will continue growing and the ones that will be buried behind the
advancing front. This collective growth as a front made of closely oriented crystals
FIGURE 1.5 Textures produced by competition for space. Granoblastic textures (a) are generated by three-
dimensional competition by growing crystals. Sub-parallel growth textures (b) are generated by one or two-
dimensional competition for space.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 15
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

synchronizes the zonal structure of the crystals contributing to the front and can lead to
banding parallel to the substrate.
In natural environments, other actors can play a major role in determining the crystal
growth morphology, notably the action of living organisms and the presence of impu-
rities increasing or reducing the relative growth rate in different directions. Life effects
are discussed elsewhere in this volume (see Section 1.2.5). The effect of impurities is
thoroughly explained in Ref. [49].
1.4 Decoding Polycrystalline Textures
from Nucleation and Growth
The texture of a rock is defined by the crystal size distribution—that is, the distribution
for each mineral of crystals having a given size. This distribution is determined by both
nucleation and growth kinetics. In natural crystallization, heterogeneous nucleation is
expected to be the dominant process by far because homogeneous nucleation requires
larger activation energy. This is mainly due to the large number of impurities, as well as
the widespread presence of particles of different natures and multiphase, rough surfaces.
Secondary nucleation can be important in superficial waters moving by turbulent flow,
such as in a river or at a coastline.
A simple model for the dynamics of nucleation and growth is the Johnson-Mehl-
Avrami-Kolmogorov model, which describes phase transformations by nucleation and
growth [50–52]. This theory was initially formulated by Kolmogorov to explain the so-
lidification of metals, but it is general enough to be applied to many different nucleation/
growth processes because it is a geometrical formalism without any energetic term and
only involves generic nucleation and growth rates I(t) and G(t). Assuming that crystals
grow spherically and that the growth rate is independent of crystal size, it is possible to
write the volume, at time t, of a crystal nucleated at time sas follows:
V¼4p
30
@Zt
s
Gdt1
A
3
The number of such crystals is V
sys
I(s)ds, where V
sys
is the volume available for nucle-
ation. Therefore, at time t, the total volume of crystals nucleated between sand sþdsis
dV ¼4p
3VsysIðsÞds0
@Zt
s
Gðt0Þdt01
A
3
Finally, by integrating over all times, it is possible to compute the total volume of crystals
in the system:
Vext ¼4p
3Vtot Zt
0
IðsÞ0
@Zt
s
Gðt0Þdt01
A
3
ds
16 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

The process is illustrated in Figure 1.6. We can see crystals nucleating at random posi-
tions with a given frequency I(t) (constant in Figure 1.6, top) and then growing at a given
rate G(t) (also constant in the illustration). The process is straightforward and well
described by the previous equations until t¼7. At this time, crystals start impinging
during growth on other crystals and with the boundaries of the system. The equations
then start to fail because these interactions are not included in the model. This is
indicated in the last equation by writing V
tot
(the total volume of the system) instead of
V
sys
(the volume available for nucleation, which is hard to compute after crystals start to
interact) and V
ext
(the fictitious crystal volume if interactions are ignored) instead of V
(the actual crystal volume that indicated in the last equation) by writing V
tot
(the total
volume of the system) instead of VV (the actual crystal volume that, again, is too difficult
to compute). In the Avrami formulation, this approximation is handled correcting the
equations using the mean field approximate relationship:
dV ¼1V
Vtot dVext
Integrating both sides of this relation and solving for V, we get
V
Vtot
¼1expVext
Vtot 
Inserted into the expression for V
ext
, that produces
4hV
Vtot
¼1exp0
B
@4p
3Zt
0
IðsÞ0
@Zt
s
Gðt0Þdt01
A
3
ds1
C
A
That is the general form of the Avrami equation. For constant nucleation and growth
rates, the equation reduces to
4¼1exp4
3IG3t3
The equation is commonly used for interpreting experimental results as
4¼1exp(kt
n
), which has the well-known sigmoidal form shown in Figure 1.6 (down).
The low transformation rates at the beginning and at the end are due to, respectively, the
limited number of nuclei existent at the beginning and the limited volume of
liquid feeding growth at the end. Using this equation, experimental data can be analyzed
by plotting log(log1/1 4) as a function of log(t), which produces a linear plot with
slope nand intercept log(k). An example of these curves from experimental data is
shown in Figure 1.7, where selected data on the transformation of aragonite into
calcite are shown using this type of plot (data from Ref. [53]). The curvature in the plot is
due to the nonlinear behavior of I(t) and G(t), which is assumed to be constant in the
equations.
The next step in developing a model relating the nucleation and growth rates
measurable in the laboratory with the actual texture of mineral grains in a rock is to
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 17
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

FIGURE 1.6 Time evolution of the crystal size distribution of crystals nucleating and growing at a constant rate
(top). Notice the overlapping of crystal “volumes”for t>7. Evolution of the crystallized volume fraction 4
(bottom).
18 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

develop a model for the number of crystals of a given length per unit length per unit
volume. This quantity nis called the population density:
nhdN
dL
where Nis the number of crystals per unit volume with size less than L. To obtain an
expression for n, we will use the Avrami formalism to define efficient rates of nucleation
and growth (I
eff
(t) and G
eff
(t)), related to the laboratory measured values without inter-
crystals interferences I(t) and G(t).
Ieff ¼IðtÞð14ðtÞÞ
The exponent 1/3 in the growth rate equation is due to G
eff
(t) being defined as a linear
growth rate because we are interested in the linear size Lof the crystals.
Given these values, we can compute the size at time tof a crystal nucleated at time s:
L¼Zt
s
Geff ðt0Þdt0
This equation introduces a functional dependency of son L, so we can write the time at
which a crystal having size snucleated s(L). In these terms, the number of crystals
nucleated between sand sþdsis, using the Avrami corrected nucleation rate,
dN ¼Ieff ðsÞds¼IðsðLÞÞð14ðsðLÞÞÞds
FIGURE 1.7 Plot of selected data on the aragonite/calcite transformation (From Ref. [53].) in log(log(1/(1 V)))
versus log t coordinates used to estimate the kinetic constants nand kfrom experimental data using the
approximate Avrami equation.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 19
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

and, from the definition of n,
nhdN
dL ¼IðsðLÞÞð14ðsðLÞÞÞ ds
dL
The minus sign in this equation comes from the inverse relationship between dsand dL:
higher dsmeans shorter L. Using the previous equation for L, this equation can be
rewritten as follows:
n¼IðsðLÞÞð14ðsðLÞÞÞ 1
Geff ðsðLÞÞ
This equation, along with the Avrami equation for 4, defines a fully general model for
the relationship between the experimental nucleation and growth kinetics and the
population density (i.e., the texture of the rock resulting from the crystallization of the
mineral grains). The inversion of this equation, along with values of the dependency of
the experimental rates on some variable, such as temperature, allows the decoding of the
evolution of this variable during the growth of mineral grains. By modeling the time-
dependent nucleation and growth rates in a functional form, we can use the previous
equation to compute population density curves. Assume, for example, that you can
model the nucleation and growth rates by Gaussians:
IðtÞ¼I0expAIttI
02
GðtÞ¼G0expAGttG
02
FIGURE 1.8 Computed distribution of the
number of crystals of a given length per
unit length per unit volume (population
density) as a function of crystal size. The
five curves correspond to increasing from
bottom to top.
20 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

with three adjustable parameters each. Using these functions, Figure 1.8 shows
the evolution of a typical population density for five different values of time (60, 70,
80, 90, and 100 years).
1.5 The Case of Giant Crystals
Beyond their beauty, the spectacular giant crystals of gypsum found in the Naica mine
(Chihuahua, Mexico) offer a typical example of how knowledge of crystal growth can be
used to decode geological information present in mineral crystals grown in geological
settings [54,55]. Crystal growth in the laboratory often assumes equilibrium, or at least
steady-state, but the time scales for laboratory studies are too short for close-to-
equilibrium studies. However, crystal growth in natural media proceeds over millions
of years, providing a singular and extremely valuable opportunity to study close-to-
equilibrium growth processes. This is the case of the giant crystals from Naica.
Among the many caves and cavities filled with gypsum crystals within the limestones
of the mountain of Naica, the best known are the Cave of Swords [56,57] and the Cave of
Crystals [54]. The walls of the former cave are completely covered with crystals of several
centimeters in length displaying two different morphologies (one elongated along the c-
axis and the other more equidimensional). By contrast, in the Cave of Crystals, there are
fewer crystals, but they are larger and more transparent; they also display the same
morphologies as in the Cave of Swords. In addition to these notable differences, the
number of crystals differs between the caves. In the upper cave, the crystals are smaller
but more abundant than in the Cave of Crystals, where there are fewer crystals but of
much bigger size. The existence of large crystals can be explained by nucleation. The
nucleation rate is an exponential function of supersaturation, so the probability of
having a nucleation event is very unlikely for small supersaturation values; however,
beyond a critical value, the rate increases so fast that small changes in supersaturation
can lead to orders-of-magnitude changes in the number of crystals produced. Therefore,
nucleation kinetics is an extremely sensitive probe for chemical conditions during crystal
nucleation and growth. The low number of crystals in the Cave of Crystals allows the
crystals to grow to a very large size, which is due to the fact that they nucleated at very
low supersaturation. In addition, supersaturation must be kept low enough, without
large fluctuations, during the entire crystal growth history to avoid triggering further
massive nucleation.
Garcı
´a-Ruiz et al. [54] have shown that the only geologically plausible mechanism
capable of producing this large and steady supply of Ca
2þ
and SO
4ions is a self-feeding
mechanism based on a solution-mediated anhydrite-gypsum phase transition occurring
in a slow and smooth cooling scenario. Calcium sulfate may precipitate as different
hydrates, with each of them being stable in a different temperature range. Among them,
the dihydrate is the less soluble phase at low temperatures (below 54.5) and the
anhydrous calcium sulfate is the stable phase above this temperature. The existence of a
crossover between the solubility of anhydrite and gypsum provides the “ion pump”
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 21
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

required for a steady and continuous supply of growth units. Figure 1.9 shows the sol-
ubility of gypsum and anhydrite plotted against temperature along with the difference in
solubility, which is the maximum amount of Ca
2þ
and SO
4ions transferred from
anhydrite to gypsum as a function of temperature. This amount is small enough to
ensure a slow and steady supply (10
3
to 10
3
mol/L in the 40–80 C range). This
extremely slow cooling rate leads to the close-to-equilibrium anhydrite dissolution and
gypsum growth at progressively lower levels within the phreatic layer.
For this mechanism to operate, there must be enough anhydrite in the area, and the
temperature at which the crystals grow must be close to the transition temperature. Both
conditions are met in the Naica mines. The temperature of both current groundwaters
and the solution in equilibrium with the crystals during their growth (obtained from fluid
inclusions trapped in the crystals [58]) show that current groundwaters have a mean
temperature of T ¼53.2 C, whereas fluid inclusions indicate that T ¼52.5 C during
crystal growth. These temperatures correspond to a very low undersaturation value for
anhydrite (Dm ¼0.04), which makes the self-feeding mechanism possible. Additional
FIGURE 1.9 Plot of the solubility of gypsum and anhydrite (left). Gypsum is the stable phase to the left of the
dotted vertical line and will grow from dissolving anhydrite. The amount of transferred Ca and sulfate is the
difference in solubility also plotted. The two vertical gray bars indicate the range of crystal growth temperatures
for the Cave of Swords (right top) and the Cave of Crystals (right bottom) in the Naica mine. These ranges were
obtained from the statistical analysis of fluid inclusion homogenization temperatures (insets) measured in samples
from these caves. The crystals in the Cave of Crystals grew very close to equilibrium from a very low driving force.
22 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

pieces of evidence supporting this self-feeding mechanism for the nucleation and
growth of the giant gypsum crystal [54] are the low salinity values of the mother waters,
the isotopic composition of these crystals, and the presence of celestite coating the walls
of the Cave of Crystals, as a result of the excess Sr that the structure of anhydrite can hold
with respect to that of gypsum.
The giant gypsum crystals also offer a unique opportunity to study the mechanisms
and kinetics of crystal growth at very low supersaturation values, as well as the
agreement between the morphology of crystals growing in this regime and the equi-
librium morphology. Obviously, the giant crystals of the Naica mines must have grown
very slowly at very low supersaturation, but what was their growth rate? The super-
saturation values of the current Naica waters are below the experimental error of
modern analytical techniques for measuring concentrations. Therefore, measurement
of growth kinetics in the laboratory using Naica’s waters must be performed with well-
designed growth cells in a short period of time, because significant shifts in super-
saturation could be produced if the solution evaporates. Experimental studies of
gypsum crystal growth kinetics show that the growth of gypsum crystals at low su-
persaturation proceeds mainly by two-dimensional nucleation on {010} faces and the
advance of steps, usually as macrosteps [59]. Even at lower supersaturation values,
crystal growth kinetics have been measured for the {010} face of gypsum in contact with
waters collected from the Naica mine using an advanced high-resolution white-beam
phase-shift interferometry microscope [60]. The growth rates found in this work for
temperatures in the range of 50–55 C are extremely slow, ranging from 1.6 10
6
to
2.1 10
5
nm/s. This very small growth rate can be enhanced by screw dislocations
and similar crystal defects, colloidal particles incorporated into the crystal surface, or
surface instabilities such as macrosteps or fluid inclusions, which are quite frequent on
the surface of gypsum crystals.
As previously mentioned, gypsum crystals in the Naica mines show two completely
different morphologies, both of which have been described by Foshag in the Cave of
Swords [56]: blocky crystals, which in some cases cluster to form parallel or radial ag-
gregates, and much more elongated crystals that grow from some groups of blocky
crystals. Neither of these morphologies agree with the classical equilibrium morphology
of gypsum [61,62]. The more developed faces in the giant crystals are the {1k0} family
instead of {010}, as would be predicted by theory. Several faces in this family (at least
2k6) appear with an overall orientation of {140} instead of the predicted {120}. In
addition, {111} is more developed than {011}. These contradictions gave rise to new
investigations on the equilibrium morphology of gypsum, leading to refinements of the
surface energy calculations that considered surface relaxation and semi-empirical po-
tentials, which were carefully selected for each interaction that predicts equilibrium
morphologies showing better agreement with the giant Naica crystals [63]. The origin of
the striated {1k0} forms was also explained using similar methods [64].
The second type of morphology, the colossal beams, is more challenging and is
not predicted by any of the existent theories. An explanation for this morphology has
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 23
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

been proposed [55]. It has been observed that the Naica beams are {100} contact
twins showing reentrant dihedral angles between {111} faces. These reentrant angles
are known to (1) operate as an additional source of steps under low supersaturation
in crystals containing few or no screw dislocations (although this is less effective with
an increasing number of these defects) [65]; (2) produce elongated morphology due
to the enhanced growth rate in the crystallographic direction of the reentrant angle;
and (3) produce crystals that are much larger than the coexisting single crystals and
show crystal faces that are uncommon in single crystals. All of these features are
exhibited by the twin beams in the Cave of Crystals, which makes them so different
from the blocky crystals precisely because of these characteristics. Consequently, the
elongated giant crystals in Naica are crystals that developed, by chance, {100} contact
twins with the reentrant angle between their {111} faces pointing approximately
upwards.
1.6 Decoding Disequilibrium Mineral Patterns
In geological terms, self-organized processes leading to pattern formation operate at all
scales, from the atomic scale, as in the case of the kinetic behavior of cation ordering in
crystal lattices such as Na feldspars with partially ordered Al, Si positions [66], to the
tectonic processes controlling the distribution of tectonic plates at the scale of the whole
planet [67]. There are two requirements for self-organization: the system is sufficiently
far from equilibrium and at least two active processes in the system are coupled [68].
Because several types of reaction-transport loops can operate in geochemical systems
and because these systems are most often out of equilibrium, geochemical self-
organization should be expected to be commonplace [69,70].
Crystallization is an inherently nonequilibrium process in which surface aggregation
processes and mass transport often operate simultaneously and with comparable rates.
Consequently, pattern formation during crystallization is a common phenomenon. In
natural systems, this is facilitated by the ubiquitous presence of noise in all the variables
involved. In the laboratory, it is usual practice to work with smooth containers having
constant and isotropic chemical and physical properties to hold ultrapure solutions of
well-defined composition at constant temperature and pressure. In a natural system
where minerals are nucleating and growing, this is not the case. Self-organization can be
thought of as the ability of a system to select one pattern from all possible noise-
generated patterns and amplify it by feedback loops into a well-ordered observable
structure [71].
Mineral pattern formation can be the result of the following: (1) intrinsic instabilities
during the growth of the crystal, as in the case of Mullins–Sekerka instability leading to
dendritic crystals or the compositional zonation of crystals; (2) self-organized processes
involving coupled mass transport and precipitation, as in the case of Liesegang struc-
tures; or (3) the mineralization of patterns produced by biological or convective mass
transport processes, as in the case of stromatolitic structures or fractal dendritic
24 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

structures. Examples of these processes are provided in this section. More information
on mineral pattern formation in geological media can be found in references [72–75].
1.6.1 Compositional Zoning
The chemical composition of minerals may change within the crystal volume in a
distinctive (patterned) way. This is very common in solid solutions and in the minor
element compositions of many minerals. The kinetics of the incorporation of impurities
and minor components to the surface of a growing crystal are anisotropic, as for the
regular molecules or ions making up the crystal. As a result, the composition of a crystal
may vary from sector to sector. This phenomenon called sectorial zonation, is well
known in minerals such as quartz, calcite, staurolite, topaz, and zircon [76–79];itisa
result of the differences in the structures of the surface, bond distribution, and specific
energy of the different faces of the crystal. In addition, the presence of structurally
nonequivalent growth steps, typically at adjacent flanks of polygonized spiral growth
hillocks, can result in differential incorporation of minor components into symmetry-
equivalent growth sectors or even in the same sector [80]. This feature, called
intrasectorial zoning, has been described for calcite [81,82]; it illustrates both the
fundamental need for crystal growth knowledge for the decoding of geological infor-
mation encoded in mineral features and the stringent requirements for these studies,
which can go up to the atomic details of the surface of growing crystals.
Compositional zoning along volumes parallel to the growing crystal surface is
commonly associated with, and more common than, sectorial zoning. This composi-
tional zoning occurs in most major classes of minerals in a wide range of geological
environments. When this zoning is quasi-cyclic, such as the well-known compositional
alternation in plagioclase crystals, it is called oscillatory zoning. This rhythmic pattern
of chemical composition or physical properties can be alternatively explained as the
result of (1) self-organized coupling between the interface kinetics and the diffusion of
chemical species in the melt or (2) changes in the chemical composition of the magma
from which crystals grow. Moreover, both types of effects can happen simultaneously;
in both cases, the information on the conditions during crystal growth is encoded
within the pattern [83,84]. Oscillatory zoning has been explained by changes in the
crystal growth regime due to changes in the concentration in the boundary layer
[85,86], changes in the crystal growth kinetics as a function of the undercooling [19],
the nonlinearities of the partitioning coefficient of anorthite concentration [87],and
the cross-terms in the component diffusion coefficients [20,88]. All these models
explain to some degree the oscillatory zoning of plagioclase and have their particular
drawbacks. More studies, particularly with the support of experimental data, are
required to understand this process and to be able to use oscillatory zoning as a tool for
decoding the genetic information contained in plagioclase crystals. Notable advances
are being achieved in this direction for the case of decompression-driven crystallization
of volcanic rocks. Both the morphology and the composition of plagioclase in
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 25
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

decompression experiments have been demonstrated to be a function of the decom-
pression rate [89,90].
Other minerals—at least 75 rock-forming and accessory minerals comprising most
major mineral classes—show oscillatory zoning [91]. The widespread presence of
oscillatory zoning in natural minerals contrasts with the very limited observation of this
phenomenon in the laboratory. This fact leads to the common interpretation of oscil-
latory zoning being due to systematic variations in the crystal growth environment
(extrinsic mechanism) instead of being a result of self-organized nonlinear behavior in
systems driven far from thermodynamic equilibrium (intrinsic mechanisms). Several
models for self-organized oscillatory zoning in natural systems have been proposed for
plagioclase [92]. The model most accepted by geologists is the interpretation of this
zoning in terms of extrinsic mechanisms based on relative reductions and increases of
the total pressure during convective movement of the magma, leading to higher and
lower Ca concentrations within the plagioclase crystal, respectively.
The best-known case of a clearly intrinsic mechanism producing oscillatory zoning is
the barite (BaSO
4
)–celestite (SrSO
4
) solid solution system [93–96]. These studies corre-
spond to the observation of spontaneous oscillatory growth of Ba and Sr sulfate-rich
terms of the solid solution under controlled experimental conditions; therefore, the
observation cannot extend to extrinsic mechanisms, as it qualifies as a self-organized
oscillatory mineral pattern [94]. This behavior is explained by the coupling of diffusive
mass transport of growth units to the surface of the crystal and the autocatalytic
continuous growth, with a rate dependent on the mineral surface composition [97]. The
incorporation of cations at kink sites is energetically favored at sites having the same
sulfate species as neighbors. Therefore, when the surface is composed of mostly BaSO
4
,
more Ba is incorporated into the crystal and the solution in contact with the crystal gets
enriched in Sr. This increasing concentration of Sr leads to a progressively larger
incorporation of this cation to the surface, which, after some time, will start favoring the
incorporation of Sr Replenishment of the two cations by diffusive mass transport keeps
this oscillatory growth going. This interplay between mass transport and surface kinetics
is summarized by these authors [97] in a balance equation for mass transport at the
liquid/crystal interface:
vmi
vt¼Di
v2mi
vx2þVvmi
vx
where m
i
and D
i
are the concentration and diffusion coefficients of the ith species in
solution. The crystal growth kinetics are
V¼VbþVc
Vb¼bbmBamSO4m0
Bam0
SO4Xþpb2
Vc¼bcmSr mSO4m0
Sr m0
SO41Xþpc2
where the band csubscripts stand for barite and celestite, respectively; Vis the growth
rate, bis a kinetic coefficient; m
0
is the equilibrium concentration; and pis the
26 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

probability for the attachment of a Ba-unit (Sr-unit) on a pure Sr (Ba) surface. Xis the
solid-phase barite mole fraction having dynamics modeled by
aLdX
dt ¼½Xþð1XÞa2½VbXðVbþVc=aÞ
where ais the ratio between the barite and celestite molar volumes and Lis the effective
thickness that characterizes the roughness of the interface between the crystal and the
aqueous solution.
This model accounts for different crystal growth regimes in the barite
(BaSO
4
)–celestite (SrSO
4
) solid solution system, including the oscillatory composi-
tional zoning observed experimentally [93,94,95,96]. It is therefore demonstrated to be
an intrinsic, self-organized behavior of the system. Separate regions of stable (SF),
unstable (UF), and multistable (MS) composition during crystal growth appear in the
phase space, as shown in Figure 1.10(a) and (b), corresponding to two different con-
centration values. Within the stable zone, compositional fluctuations do not exist or
are quickly damped to a constant value of the Ba/Sr ratio (Figure 1.10(c),pointcin the
phase space plot). Conditions within the unstable zone (Figure 1.10(d) and
Figure 1.10(e)) show sustained, self-organized compositional zoning of different
(a)
(b)
(c)
(d)
(e)
(f)
FIGURE 1.10 Stability phase diagram (left) in the (p
b,
L) space. SF ¼stable focus, UF ¼unstable focus, S2 ¼saddle,
SN ¼stable node, and MS ¼multistability. The upper phase diagram corresponds to bulk concentrations of 10 mM
(a) and 15 mM (b). The four plots c–f show the molar fraction of barite in the resulting crystal as a function of the
distance from the crystal nucleus. They illustrate the behavior at conditions labeled c–f in the phase diagram at
the bottom b (Adapted from Ref. [97].).
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 27
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

frequencies and amplitudes. Finally, two different Ba/Sr ratios may coexist in the
multistable zone (Figure 1.10(f)).
The fact that these types of self-organized oscillatory zoning processes are wide-
spread in nature but difficult to observe in the laboratory is commonly explained by
noise-induced effects [98], which are very common in nature but almost absent under
controlled laboratory conditions.
Beyond the single crystal scale, rhythmic patterns are also very common in geological
environments. Again, the periodic changes in composition or other properties may arise
from environmental variations or from self-organization.
1.6.2 Liesegang Structures
On a larger scale, compositional banding due to precipitated minerals is also common in
the form of concentric or parallel stripes reminiscent of Liesegang bandings in many
geological systems, such as agates, cherts, rhyolites, sandstones, and limestones [99,100].
Liesegang bandings form sequentially and can be explained in the context of Ostwald’s
supersaturation theory when one or more of the reacting species diffuses into a space
occupied by other species. Figure 1.11 shows a sketch of the coupled mass transport of a
reactant A into a space initially filled with a homogeneous concentration of reactant B.
The supply of A molecules from the left side induces the supersaturation with respect to
the precipitating phase until nucleation happens at a region where the concentration of
A(C
A
) reaches a critical value C
A. This region, where the concentration of reactants is
close to the equilibrium value C0
Abecause crystals are growing, typically has the shape of
a band parallel to the diffusion front because all the points having a given value of su-
persaturation are located in this band. The precipitated band is marked as a gray bar in
the plots. This nucleation event and the further growth of the precipitating phase de-
pletes the concentration of both A and B (plot at t
2
), preventing the nucleation of further
mineral near the precipitated band (see the trajectories in phase diagrams for the points
x
2
and x
3
near time t
2
). Continuous diffusion of A eventually increases the supersatu-
ration value at points distantly ahead of the previously deposited bands, where the
concentration stays close to equilibrium as long as precipitation continues. At these
points, supersaturation can grow past the critical value required for nucleation, and a
new band is formed (t
3
). The process repeats itself until the full region is homogenized by
the diffusive supply of reactants and their consumption by nucleation and growth
processes.
The bands produced by this phenomenon are regularly spaced in space and time. The
spacing law, first described by Jablczinsky in 1923 [101], states that the positions where
bands form follow a geometrical series. That is, the ratio between the positions of any
pair of consecutive bands in the pattern is constant:
P¼xnþ1
xn
28 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

The time periodicity of the pattern is controlled by the dynamics of diffusive mass
transport. The time at which the nth band forms is proportional to the square of the
distance at which the band forms:
K¼x2
n
tn
FIGURE 1.11 Sketch of the coupled diffusion and crystal growth processes producing Liesegang rings. To the left a
series of four consecutive time snapshots is shown as plots of the A and B concentration profiles (C
A
,C
B
), the
critical (nucleation) and equilibrium concentration values ðC
A;C0
AÞ. Three points are marked x
1
–x
3
. The time
evolution of A and B concentrations is shown in the three plots to the right. In these phase diagram sketches, the
equilibrium (solubility) and nucleation curves are shown. The instants illustrated in the left plots are marked by
dots in the phase diagrams.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 29
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

1.6.3 Fractal Dendrites
Self-organized kinetic instabilities at the surface of growing crystals, generally inter-
preted under the Mullins-Sekerka stability model, occur at high supersaturation (or
undercooling) on surfaces growing by continuous growth. In these conditions, the
competition between capillary forces and diffusion kinetics (as opposed to interface
kinetics) favors configurations in which the growing surface has as large a surface area as
possible, which allows, for example, a faster dissipation of latent heat [102]. The best-
known example of this kind of instability is the development of snowflakes, with their
distinctive dendritic development in the hexagonal direction and their unique shape
reflecting the relative changes in temperature and humidity as the crystal moves within
the cloud during its growth. In mineralogy, the most common occurrence of this phe-
nomenon is among metal ores, where dendritic growth is typically associated with
framboidal and colloform textures.
Single-crystal dendrites are thoroughly discussed in Chapter 16 of Volume IB of this
handbook [103]. Here, we will concentrate on a type of mineral precipitation commonly
called “dendrites,” although they form by entirely different processes. “Pyrolusite den-
drites” is a somewhat misleading name for the beautiful tree-like fractal patterns formed
by manganese and iron oxyhydroxide minerals. Because of their very common occur-
rence, knowledge of the genetic conditions of manganese and iron dendrites would be of
great practical interest in understanding geological environments. These fractal den-
drites are sometimes embedded in quartz crystals and agates, but the most common and
geologically significant are associated with cracks and sedimentary laminations or, in
general, with quasi two-dimensional spaces. They appear on or inside many different
types of rocks, such as limestone and sandstone, which suggests that the mineralogy of
the host rock is not a limiting factor for their formation.
There are two different mechanisms that may explain the formation of these struc-
tures. Chopard et al. proposed that pyrolusite dendrites can be explained by a diffusion-
limited aggregation (DLA) [104–106], a mechanism that may work for tridimensional iron
and manganese dendrites found in quartz and agates. Alternatively, Garcı
´a-Ruiz et al.
showed that manganese and iron dendrites are the mineral record of Saffman-Taylor
instability [107]. Both mechanisms can plausibly operate in geological environments,
and only a thorough study of the mineralogical and textural properties of the minerals
and the geological environment can shed light on the actual origin of the patterns. The
controversy illustrates very well the pathway of decoding genetic information in mineral
patterns and also is an example that morphology, by itself, does not contain genetic
information [108]. The values of the fractal dimensions measured for these mineral
dendrites are compatible with both proposed mechanisms.
The fractal dendrites analyzed by Garcı
´a-Ruiz et al. appear in the sedimentary lam-
inations of sandstones from the flysch of Tarifa (Spain). Their fractal dimension, as
measured by box-counting methods, was l.69, matching that for Laplacian growth pat-
terns. They appear to be associated with other deposits of the same composition and
30 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

aspect that have higher fractal dimensions. As conjectured by Van Damme [109] for
manganese dendrites, this is consistent with a genetic mechanism based either on DLA,
which implies the irreversible aggregation of diffusing units to a stationary growing
pattern, or on viscous fingering [110], which is produced when a fluid is injected into a
more viscous and non-Newtonian one. In any case, it must be assumed that the for-
mation of the dendrites occurs under highly irreversible conditions that are far from
equilibrium, in which the nucleation kinetics are faster than the pattern formation
process.
Infrared spectroscopic studies of these deposits [111] showed that they are formed by
nonsingular, variable mineral associations, including various species of the
Romanecheita and Hollandite groups, along with minor and optional iron oxides that,
when present, can also form dendritic tree-like patterns. All of these minerals have low
crystallinity, as shown by powder X-ray diffraction and high-resolution electron micro-
scopy. The very low crystallinity of the precipitates is a clear indication of very fast
growth kinetics and therefore high supersaturation, which is consistent with the fractal
geometry of the patterns. Under scanning electron microscopy, the material forming the
dendrite shows a colloidal aspect and appears to be coating the mineral grains forming
the matrix rock.
The colloidal texture of manganese oxides, the manganese concentration profile
around the dendrites, and the coexistence of well-differentiated Mn and Fe dendrites are
difficult to explain by a diffusion reaction mechanism, at least in the cases where these
patterns appear to be associated to sedimentary laminations (Figure 1.12). In these
cases, a genetic model based on the mineral record of fluid flow structures is more
plausible. Viscous fingering is a pattern created when one fluid pushes another of higher
viscosity in a confined quasi-2D space. In porous media for very low velocities, the flow
provokes a percolation pattern with a fractal dimension of 1.82, which decreases for
higher velocities as the flow patterns become reminiscent of DLA patterns with a
dimension of 1.70 [112]. These patterns fit very well with the geometrical properties of
manganese dendrites and related structures with higher fractal dimensions.
Consequently, the most probable genetic mechanism for the formation of manganese
dendrites is based on the mineralization of flow structures formed during the invasion of
sedimentary discontinuities by Mn- and Fe-bearing fluids rising through cracks. For this
process to operate, the host rock body to be injected must be at least partially cemented
and the injected fluid must displace another of higher viscosity. Both facts are possible
during diagenesis, when a colloidal suspension fills the interbedding laminations and
cavities of partially cemented sediments. Under these conditions, the parameters con-
trolling the great variety of dendritic and nondendritic patterns observed in field studies
are the width and thickness of the uncemented muddy lamina, the roughness of the
cemented surfaces, the pressure of the injection, and the viscosity of the pushed fluid.
Sediments containing Mn
4þ
and Fe
3þ
compounds are reduced upon burial below the
anoxic-oxic interphase and thus become enriched in Mn
2þ
and Fe
2þ
. Solutions con-
taining these cations can migrate upwards through the cracks of the sedimentary body
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 31
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

and invade the upper alkaline and oxygenated zones, pushing the colloidal suspension
that fills the sedimentary laminations and consequently forming the dendritic patterns.
Finally, preservation of the pattern can be due to the cementation of the host rocks,
plugging of the system by the formation of colloidal iron oxides, and the exhaustion of
the source of reducing chemicals.
1.7 Early Earth Mineral Growth, Primitive Life
Detection, and Origin of Life
Once living organisms appeared on our planet, most crystal growth phenomena
occurring in surface and subsurface environments were affected by life. If life appeared
near or within hydrothermal vents, not only the hydrosphere and atmosphere but also
the fluid transport and geochemistry of hydrothermal systems were deeply affected by
the influence of living organisms. This influence started more than 3 billion years ago,
and today it is ubiquitous and evident.
Thus, it is not strange that when the geological record is analyzed with the actualistic
view characteristic of geological studies, biomimetic structures are interpreted by
comparison with contemporary structures created by processes working today—that is,
processes on which life plays an important role. Mineral textures and structures that look
FIGURE 1.12 On the left, a mineral dendrite of iron and manganese oxi-hydroxides. The vertical length of the
rock is 45 cm. This type of dendrites with fractal dimension close to 1.68 can be generated either by a diffusion
limited aggregation mechanism (top right) or by a viscous fingering mechanism (bottom right). This is an example
that the morphology of an object—by itself—does not contain information about its genesis.
32 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

like biofilms [113], microfossils [114], or extant stromatolitic structures [115] were
straightforwardly interpreted as life remnants, as real fossils, because it was thought that
such complex structures could not be produced by mineral processes alone. This is no
longer true since the discovery of silica biomorphs [116,117], which are self-organized
mineral structures with hierarchic textures and shapes of continuous curvature, such
as those thought to be exclusive of life. They can be produced not only without the aid of
life but even without organic compounds. In fact, there are several physicochemical
routes to producing inorganic materials displaying shapes reminiscent of life, but silica-
carbonate biomorphs have an exceptional relevance for life detection. This is because of
the formidable mimicking of the structures proposed to be the earliest remnants of life,
but mostly because they self-assemble from a simple chemical cocktail that is
geochemically plausible and similar to the chemistry of the rocks in which these putative
microfossils are embedded.
The elements required to form silica biomorphs in the laboratory—silicon, water, and
alkaline earth metals, particularly calcium and barium—were all abundant in Archean
times. In addition, all the Archean putative microfossils reported to date were found in
cherty rocks (i.e., rocks that either formed from silica sols or were silicified by silica-rich
fluids). More astonishingly, in several cases the rocks containing the oldest putative
remnants of life were barium-rich cherts [114,118]. With such a geochemical framework,
silica biomorphs are a clear inorganic alternative explanation to some of the micro-
structures considered to be the oldest remnants of life on the planet. Further experi-
ments also showed that the carbonaceous composition of the kerogen decorating
Archean putative microfossils and claimed to be a proof of biogenesis [119] could also be
obtained when silica biomorphs are grown with the same chemical recipe used to create
inorganic hydrocarbons from siderite decomposition [120,121]. The Raman spectra of
both Archean putative microfossils and their laboratory inorganic counterpart are
characteristic of graphite-like and diamond-like carbon (see Figure 1.4 in Ref. [122]).
Figure 1.13 shows a selection of images illustrating the morphological variety of silica
biomorphs. Each of these structures are made by millions of nanocrystals of barium,
strontium, or calcium carbonate, which remain co-oriented to each other no matter the
curvature of the shape. They are obtained either in single-pot solution experiments or by
the counterdiffusion technique in gels, where these millions of nanocrystals self-
assemble to create the bioforms. How do these bizarre structures crystallize?
Everything starts with an initial single crystal of the orthorhombic carbonate (witherite,
strontianite, or aragonite), followed by the breaking of the symmetry of the crystal
structure. This occurs by fibrillation, by the splitting of the two ends of the single crystal
corresponding to their basal faces. The continuous splitting creates fractal cauliflower-
like structures that are characteristic of pH values in the range of 9–10. At higher pH
values, the fibrillation takes place by nucleation events rather than by splitting, thus
creating 2D lamellae that grow radially until the lamellae curl on their own in some
singular point of the growth front. Then, the radial growth stops and the curls propagate
along the rims of the structure. All the precise morphological diversity of the biomorphs
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 33
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

(e.g., globular, twisted ribbons, spirals, worms) can be explained by three parameters of a
pair of counterpropagating curls, namely their relative chirality, their propagation ve-
locity with respect to the radial growth velocity of the lamellae, and their relative radius
of curling.
All of the above phenomenology are driven by the role that polymeric silica plays as a
modifier and inhibitor of carbonate precipitation. What is singular in this case is the
chemical coupling between the growing carbonate crystals and the silica impurities.
They have reverse solubility with respect to pH. Thus, as shown in Figure 1.14, there is a
feedback that provokes the oscillatory behavior of the supersaturation of both phases, as
the very growth of the carbonate induces the precipitation of the impurity and, reversely,
the precipitation of the impurity provokes the precipitation of the carbonate [123].
Beyond the impact on life detection, the morphology of silica biomorphs, as in the
case of fractal dendrites, does not by itself contain genetic information; therefore, it
cannot be used as an unambiguous tool to decipher the formation mechanism. Another
example of this misuse of morphology is the case of nanobacteria, which are living or-
ganisms less than a micron in size that have been claimed to play important roles in
FIGURE 1.13 Complex self-assembled inorganic materials from chemical coupling: a selection of images illustrating
the morphological variety of silica/carbonate biomorphs.
34 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

several diseases. It has been demonstrated that abiotic nanostructures formed of
amorphous calcium carbonate meet existing criteria for identification as living nano-
bacteria [124,125]. Today, most paleontologists realize that morphology cannot be
claimed as a proof of biogenicity for Archean microstructures [78,79,89,90,126].
However, none of the putative Archean microfossils have been demonstrated to be
unequivocally silica biomorphs, either. A deeper investigation to search and decipher
putative, dubio, and nonfossil microstructures in Archean rocks should be carried out in
the future.
In parallel, the search for self-organized biomimetic structures in current alkaline
silica-rich environments must be performed. The precipitation of tubular structures of
calcium-silicate-hydrated phases have been found during the alkaline weathering of
granites (Figure 1.15). These tubules have been demonstrated to be reverse silica gardens
formed by the leaching of silica from plagioclase and quartz in the presence of an
alkaline, calcium-rich solution flow [127]. This finding is important because alkaline
silica-rich environments are natural locations investigating the complexity of and
pathways to prebiotic chemistry. Serpentinization reactions known to yield alkaline
solutions and be the main source of abiotic organic carbon will therefore be investigated
in the future [128].
FIGURE 1.14 Mineral complexity requires a feedback mechanism to trigger autocatalysis. In the case of self-
assembled silica biomorphs the mechanism is the reverse solubility of silica and carbonate versus pH.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 35
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

1.8 From Deep Earth to Outer Space
A new challenge for crystal growers is to unravel the structure of minerals in the deep
Earth. Until recently, information about the layered structure of the interior of our planet
was mainly provided by the analysis of the propagation of seismic waves. The devel-
opment of new technologies for high-pressure, high-temperature crystal growth opened
a new field of study of geological interest. The goal is to know how mineral phases
change under the pressure and temperatures found in the interior of the planet. The
discontinuities that are delimited by the change of the speed of the seismic waves (e.g., at
FIGURE 1.15 Top: tubular structures of
calcium silicate hydrate forming by the
weathering of granitic rocks by highly
alkaline water flow. Below: scheme of the
proposed origin as reverse silica garden.
From Ref. [127].
36 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

660 km, dividing the upper and lower mantle) can now be understood from the physical
properties of the phases that are stable below and above the pressure-temperature
conditions of these discontinuities. The new diamond anvils are simple but fascinating
machines, allowing one to reach pressure values characteristic of the lower mantle (tens
to hundreds of GP) at laser-induced high temperatures while recording in situ the phase
transformations by X-ray diffraction [129].
The field is ready for discoveries of surprises, such as the exotic NaCl
3
phase found
when NaCl is squeezed at 20 GP [130]. Experimental studies on the phase transitions
observed when the main minerals forming the upper mantle (namely pyroxene, olivine,
and garnets) undergo high pressure have already started to shed light on the structure of
the Earth. Thus, the silicate with perovskite structure (Mg,Fe)SiO
3
undergoes a trans-
formation to postperovskite at pressure-temperature conditions of 2900 km deep where
the mantle–core discontinuity is located [131,132]. More detailed experiments will be
performed in the near future. There is much to discover, and certainly laboratory studies
will try to reach the pressure values required to mimic the actual core below the deep
lower mantle.
The formation of the solar system can be described as a history of phase transitions
from gas, melt, and vapors to solid phases, which in most cases ends with a crystalline
mineral. Thus, in some way, understanding our solar system is a matter of crystallization.
Therefore, decoding growth information contained in the earliest crystals ever formed
can, in turn, be useful for revealing the actual history of the solar system they have
helped to create. The exposed rocks of our own planet contribute poorly to that research
program because there is no geological record older than 4.0 billion years, with the
exception of some zircon crystals recycled from the oldest extinct rocks [133,134]. Our
solar systems started 4.6 billion years ago with a solar nebulae from which chondrules,
meteorites, and interstellar dust particles started to form. These objects are the main
pieces of information, and their study in terms of crystal growth is one of the research
avenues for the future. In addition, the exciting robotic missions to the moon and Mars
are sending important analytical information, allowing for the first time the study of
extraterrestrial mineral growth processes that are vital for the future of planetary mis-
sions [135,136]. Some examples of what we call astrocrystallization are described here.
The nucleation and growth of silicate melts in space has been studied in the labo-
ratory, both on Earth and also under microgravity conditions. For instance, small silicate
spherules called chondrules are a major component of primitive meteorites; therefore,
they have been extensively studied in the past. Computer simulation and experimental
work performed under microgravity conditions provided new insights on the subject.
Crystallization experiments using a floating melt droplet by gas-jet levitation found that
heterogeneous nucleation by collisions of tiny cosmic dust particles likely explain the
formation of crystalline chondrules. Homogeneous nucleation is very difficult, and the
melted spheres are able to crystallize only at significantly high supercooling tempera-
tures (w1000 K) [137]. These studies have also successfully mimicked the textures of
chondrules. For instance, to form the rim characteristic of some chondrules, a
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 37
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

temperature difference of w100 K is required between the center and the surface of the
droplet. Such a large difference in temperature can be achieved in the case of rapid
cooling at a rate of w10
3
K/s [138].
The formation of nanometer-sized cosmic dust has been also reproduced in the
laboratory based on nanoscale properties, such as a decrease of melting point by 50%
[139] and an increase in the diffusion coefficient of more than nine orders of magnitude
[140], as well as the fusion of nanoparticles [141]. Indeed, various analogs of cosmic dust
have been reproduced in laboratories and successfully proposed in growth processes,
such as silicate nanoparticles having non-mass-dependent oxygen isotope fractionation
formed in a plasma field [142], composite particles of TiC-core and carbon-mantle
formed by decomposition of CO gas [143], fullerene and its existence around evolved
stars [144], and carbonaceous hollow particles [145].
Another application of crystal growth in astrocrystallization is the study of the for-
mation process of framboidal magnetite. These are particles made of three-
dimensionally aligned nanocrystals of magnetite found in fractures of the parent
asteroid and formed in microgravity as the result of shaking due to collision with other
solar bodies [146]. The uniformity of the size distribution and the similar morphology of
the magnetite nanoparticles in each of the colloidal crystals suggest that they were
formed through homogeneous nucleation from a highly supersaturated isolated solution
in a single nucleation event. The colloidal nanomagnetites arrange by repulsive force to
form a colloidal crystal in the solution. To overcome this force, the density of magnetite
particles in a solution must be sufficiently high in an isolated solution. The diameter of
each droplet is roughly 25% larger than that of the final product (several micrometers in
diameter). The increase in the volume fraction of magnetite can be achieved through the
evaporation of water. If the distance between particles and the Debye screening length of
each particle is equivalent, they can arrange and make a colloidal crystal. Further
evaporation reduces the interparticle spaces until the colloidal crystals are stabilized by
desiccation. The very last droplets of aqueous solution in an ancient asteroid are very
important because the mutual interaction of minerals and organics in water-
concentrated chemical species play critical roles in the final minerals in meteorites
and in the early evolution of organics.
One of the most pristine materials in our solar system that we can collect on earth are
interplanetary dust particles (IDPs). Their mineralogy has been investigated and
compared with the spectra of dust around other stars [147,148], showing that they have
similar mineral phases. It is still unclear when and from where all these minerals orig-
inate, if all of them formed in the solar nebula, or if some components have a presolar
origin. One of the more intriguing components of IDPs is glass with embedded metals
and sulfides, which may have a presolar origin. It seems that silicates and other minerals
in IDPs formed directly out of the gas phase, so that gas-to-solid condensation is the
fundamental mechanism of grain growth in nebula environments [149]. Beyond the
information provided by observational astronomy, it seems clear that future experi-
mental crystal growth studies will be key for revealing the formation mechanism of the
38 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

components of IDPs and the particles themselves, as well as for a better understanding
of the formation of the solar system.
Acknowledgments
We acknowledge support from ‘‘Factorı
´a de Cristalizacio
´n’’ (Consolider Ingenio 2010, Spanish
MINECO), the Junta de Andalucı
´a (project RNM5384), and the Spanish MICINN (projects CGL2010-
16882 and AYA2009-10655). We thank our colleagues M. Prieto and C. Ayora for useful discussions
and suggestions.
References
[1] Caley ER, Richards JFC. Theophrastus on stones. Colombia: Ohio State University; 1956.
[2] [Heales JF, Trans.] Pliny the elder, natural history. Harmondsworth: Penguin; 1991.
[3] Kepler J. Strena Seu de Nive Sexangula. Frankfurt: Godfrey Tampach; 1611 [Hardie C, Trans.]. The
Six-Cornered Snowflake, Oxford, Clarendon Press, 1966.
[4] Steno N, De Solido Intra Solidum Naturaliter Contento Dissertationis Prodromus, Florence, 1669,
[Winter JG, Trans.] The prodromus of Nicolous Steno’s Dissertation concerning a solid body
enclosed by process of nature within a solid, New York, Hafner, 1968.
[5] Burke JH. Origin of the science of crystals. Berkeley: University of California Press; 1966.
[6] Amoro
´s JL. La gran aventura del cristal. Editorial Universidad Complutense. Madrid 1978.
[7] Scheel HJ. Historical aspects of crystal growth technology. J Cryst Growth 2000;211:1.
[8] [a] Galazka Z. Hand book of crystal growth: bulk crystal growth. In: Rudolph P, editor. Growth
measures to achieve bulk single crystals of transparent semiconducting and conducting oxides
vol. 2; 2014.
[b] Wang J, Zhang G, Haohai Y, Wang Y, Chen C. Hand book of crystal growth: bulk crystal growth.
In: Rudolph P, editor. Czochralski and flux growth of crystals for lasers and nonlinear optics vol. 2;
2014.
[9] Sunagawa I. In: Hurle DTJ, editor. Handbook of Crystal Growth, vol. 2. Amsterdam: Elsevier; 1994.
[Chapter 1], pp. 2–49.
[10] Grigoriev DP. Ontogeny of minerals. Jerusalem (Israel): Israel Program for Scientific Translation;
1965.
[11] Sunagawa I. Dev Sedimentol 1994;51:19.
[12] Nakamura H, Tsukamoto K, Sunagawa I. J Cryst Growth 1990;99:1227.
[13] Tsukamoto K, Abe T, Sunagawa I. J Cryst Growth 1983;63:215.
[14] Brandeis G, Jaupart C. Contrib Mineral Petrol 1987;96:24.
[15] Solomatov SV, Stevenson DJ. J Geophys Res 1993;98(E3):5407.
[16] W Dunbar N, Jacobs GK, Naney MT. Contrib Mineral Petrol 1995;120(3–4):412.
[17] Hoover SR, Cashman KV, Manga M. J Volcanol Geotherm Res 2001;107:1.
[18] Muncill GE, Lasaga AC. Am Mineral 1987;72:299.
[19] Lasaga AC. Am J Sci 1982;282:1264.
[20] Gorokhova NV, Melnik OE. Fluid Dyn 2010;45:679.
[21] Lofgren GE, Huss GR, Wasserburg GJ. Am Mineral 2006;91:1596.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 39
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

[22] Hinsberg VJ, Schumacher JC. Contrib Mineral Petrol 2006;153:289.
[23] Chakoumakos B, Lumpkin G. Can Mineral 1990;28:287.
[24] Simmons WBS, Webber KL. Eur J Mineral 2008;20:421.
[25] Barker AJ. Introduction to metamorphic textures and microstructures. Kluwer Academic
Publishers; 1990. ISBN 10: 0216926858/0-216-92685-8.
[26] Bard JP. Microtextures of igneous and metamorphic rocks. Series: petrology and structural
geology, IX. Springer-Verlag; 1986. 264 p.
[27] Kretz Ralph. Metamorphic crystallization. Wiley; 1994. 530 p. ISBN: 978-0-471-94214-6.
[28] Kretz R. Lithos 1974;7:123.
[29] Thanh NX, Sajeev K, Itaya T, Windley BF. Lithos 2011;127:552.
[30] Imon R, Okudaira T, Kanagawa K. J Struct Geol 2004;26:793.
[31] Wassmann S, Sto
¨ckhert B, Sto
¨ckhert. J Struct Geol 2013;46:200.
[32] Augenstein C, Burg J-P. J Struct Geol 2011;33:244.
[33] Otani M, Wallis S. Geology 2006;34:561.
[34] Barnhoorn A, Bystricky M, Burlini L, Kunze K. J Struct Geol 2004;26:885.
[35] Zhou L-G, Xia Q-X, Zheng Y-F, Chen R-X. Lithos 2011;127:101.
[36] Jourdan A-L, Vennemann TW, Mullis J, Ramseyer K, Spiers CJ. Eur J Mineral 2009;21:219.
[37] Hedenquist JW, Lowenstern JB. Nature 1994;370:519.
[38] Jamtveit B, Wogelius RA, Fraser DG. Geology 1993;21:113.
[39] Sunagawa I. Crystals: growth, morphology and perfection. Cambridge University Press; 2005.
295 pp.
[40] Rodriguez-Clemente R. Estud Geol 1982;38:55.
[41] Edwards KJ, Becker K, Colwell F. Annu Rev Earth Planet Sci 2012;40:551.
[42] Richter DK. In: Peryt TM, editor. Coated grains. Berlin: Springer; 1983. p. 71.
[43] Garcı
´a-Ruiz JM. In: Rodrı
´guez R, Sunagawa I, editors. Crystal growth processes in sedimentary
enviroments. Estudios Geolo
´gicos, vol. 38; 1982. p. 209.
[44] Garcı
´a-Ruiz JM. Key Eng Mater 1991;58:87.
[45] Cody RD, Shanks HR. J Cryst Growth 1974;23:275.
[46] van Rosmalen GM, Marche
´e WGJ, Bennema P. J Cryst Growth 1976;35:169.
[47] Holland HD. Philos Trans R Soc Biol Sci 2006;361:903.
[48] Hartman P. In: Hartman P, editor. Crystal Growth: an Introduction. Amsterdam: North-Holland
Publishing Cy; 1973 [Chapter 14]. pp. 367–402.
[49] Sangwal K. Impurities and crystal growth. Pergamon; 1996. 169 p.
[50] Avrami M. J Chem Phys 1939;7:1103.
[51] Avrami M. J Chem Phys 1940;8:212.
[52] Avrami M. J Chem Phys 1941;9:177.
[53] Kunzler RH, Goodell HG. Am J Sci 1970;269:360.
[54] Garcı
´a-Ruiz JM, Villasuso R, Ayora C, Canals A, Ota
´lora F. Geology 2007;35:327.
[55] Ota
´lora F, Garcı
´a-Ruiz JM. Chem Soc Rev 2014;43:2013.
[56] Foshag W. Am Mineral 1927;12:252.
40 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

[57] Degoutin N. Soc Cient Antonio Alzate Rev 1912;32:35.
[58] Kru
¨ger Y, Garcı
´a-Ruiz JM, Canals A, Marti D, Frenz M, Van Driessche AES. Geology 2013;41:119–22.
[59] Van Driessche AES, Garcı
´a-Ruiz JM, Delgado-Lo
´pez JM, Sazaki G. Cryst Growth Des 2010;10:3909.
[60] Van Driessche AES, Garcı
´a-Ruiz JM, Tsukamoto K, Patin
˜oLopez LD, Satoh H. PNAS 2011;108:
15721.
[61] Simon B, Bienfait M. Acta Cryst 1965;19:750.
[62] Weijnen MPC, van Rosmalen GM, Bennema P, Rijpkema JJM. J Cryst Growth 1987;82:509.
[63] Massaro FR, Rubbo M, Aquilano D. Cryst Growth Des 2010;10:2870.
[64] Massaro FR, Rubbo M, Aquilano D. Cryst Growth Des 2011;11:1607.
[65] Kitamura M, Hosoya S, Sunagawa I. J Cryst Growth 1979;47:93.
[66] Salje E. J Phys Condens Matter 1993:4775.
[67] Anderson DL. Geology 2002;30:411.
[68] Nicolis G, Prigogine I. Self-organization in nonequilibrium systems: from dissipative structures to
order through fluctuations. New York: Wiley; 1977. 491 p.
[69] Hergarten S. Self-organized criticality in earth systems. Berlin Heildelberg: Springer-Verlag; 2002.
250 p.
[70] Ortoleva P, Merino E, Moore C, Chadam J. Am J Sci 1987;287:979.
[71] Ortoleva P, Chadam J, Merino E, Sen A. Am J Sci 1987;287:1008.
[72] Jamtveit B, Meakin P, editors. Growth, dissolution, and pattern formation in geosystems.
Dordrecht and Boston: Kluwer Academic Publishers; 1999. p. 409.
[73] Meakin P, Jamtveit B. Proc R Soc A 2010;466:659.
[74] Turcotte DL. In: Fractals and Chaos in geology and geophysics. 2nd ed. Cambridge: Cambridge
University Press; 1997. p. 416.
[75] Bak P. How nature works, the science of self-organized criticality. New York: Copernicus Press;
1996. 212 p.
[76] Hollister LS. Am Mineral 1970;55:742.
[77] Dowty E. Am Mineral 1976;61:460.
[78] Reeder RJ, Paquette J. Sediment Geol 1989;65:239.
[79] Vavra G. Contrib Mineral Petrol 1990;106:90.
[80] Reeder RJ, Rakovan J. In: Jamtveitand B, Meakin P, editors. Growth, dissolution and pattern-
formation in geosystems. Dordrecht: Kluwer; 1999. p. 143.
[81] Paquette J, Reeder RJ. Geology 1990;18:1244.
[82] Paquette J, Reeder RJ. Geochim Cosmochim Acta 1995;59:735.
[83] Cashman KV. Contrib Mineral Petrol 1992;109:431.
[84] Gorokhova NV, Melnik OE, Plechov PY, Shcherbakov VD. J Volcanol Geotherm Res 2013;263:172.
[85] Alle
`gre CJ, Provost A, Jaupart C. Nature 1981;294:223.
[86] Tsune A, Toramaru A. Am Mineral 2007;92:1071.
[87] L’Heureux I, Fowler AD. Geophys Res Lett 1996;23:17.
[88] Oishi Y. J Chem Phys 1965;43:1611.
[89] Brugger CR, E Hammer J. J Petrol 2010;51:1941.
[90] Mollard E, Martel C, Bourdier J-L. J Petrol 2012;53:1743.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 41
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

[91] Shore M, Fowler AD. Can Mineral 1996;341:111.
[92] L’Heureux I, Fowler AD. Can Mineral 1996;34:1137.
[93] Putnis A, Fernandez-Diaz L, Prieto M. Nature 1992;358:743.
[94] Prieto M, Putnis A, Fernandez-Diaz L. Geol Mag 1993;130:289.
[95] Prieto M,Fernandez-Gonza
´lez A, PutnisA, Fernandez-DiazL. Geochim Cosmochim Acta 1997;61:3383.
[96] Pina CM, Enders M, Putnis A. Chem Geol 2000;168:195.
[97] L’Heureux I, Katsev S. Chem Geol 2006;225:230.
[98] Horsthemke W, Lefever R. In: Noise-induced transitions: theory and applications in physics,
chemistry, and biology, vol. 15. Berlin Heidelberg: Springer Series in Synergetics; 1984. p. 318.
[99] Stern KH. Chem Rev 1954;54:79.
[100] Jamtveit B, Hammer Ø. Geochem Perspect 2012;1:341.
[101] Jablczinsky K. Bull Soc Chim Fr 1923;4:592.
[102] Langer JS. Rev Mod Phys 1980;52:1.
[103] Glicksman M. Handbook of Crystal Growth, vol. IB. Amsterdam: Elsevier; 2014. [chapter 03].
[104] Witten TA, Sander LM. Phys Rev B 1983;27:5686.
[105] Chopard B, Herrmann HJ, Vicsek T. Nature 1991;353:409.
[106] Witten TA, Sender LM. Phys Rev Lett 1981;47:1400.
[107] Garcia-Ruiz JM, Ota
´lora F, Sanchez-Navas A, Higes-Rolando FJ. In: Kruhl JH, editor. Fractals and
dynamic systems in geoscience. Berlin: Springer-Verlag; 1994. p. 307.
[108] Garcı
´a-Ruiz JM. In: Garcı
´a-Ruiz JM, Louis E, Meakin P, Sander L, editors. Growth patterns in
physical sciences and biology. Pergamon Press; 1993. p. 183.
[109] Van Damme H. Flow and interfacial instabilities in newtonian and colloidal fluids. In: Avnir D, editor.
The fractal approach to heterogeneous chemistry. Chichester: John Wiley & Sons; 1989. p. 199.
[110] Vicsek T. In: Fractal growth phenomena. Singapore: World Scientific; 1989. p. 495.
[111] Potter RM, Rossman GR. Am Mineral 1979;64:1199.
[112] Lenormand R, Daccord G. Flow patterns in porous media. In: Stanley HE, Ostrowsky N, editors.
Random fluctuations and pattern growth. Dordrecht: Kluwer Academic Publishers; 1988. p. 69.
[113] Noffke N, Christian D, Wacey D, Hazen RM. Astrobiology 2013;13:1.
[114] Schopf JW. Science 1993;260:640.
[115] Allwood AC, Walter MR, Kamber BS, Marshall CP, Burch IW. Nature 2006;441:714.
[116] Garcı
´a-Ruiz JM. Biosphere 1994;24:451.
[117] Garcı
´a-Ruiz JM. Geology 1998;26:843.
[118] Buick R, Dunlop JSR, Groves DI. Alcheringa 1981;5:161.
[119] Schopf JW, Kudryavtsev AB, Agresti DG, Wdowiak TU, Czaja AD. Nature 2002;416:73.
[120] McCollom TM. Geochim Cosmochim Acta 2003;67:311–7.
[121] Zuilen MA, van Lepland A. Arrhenius G. Nature 2002;418:627.
[122] Garcı
´a-Ruiz JM, Hyde ST, Carnerup AM, Christy AG, Van Kranendonk MJ, Welham NJ. Science
2003;302:1194–7.
[123] Garcia-Ruiz JM, Melero E, Hyde ST. Science 2009.
[124] Martel J, Young JD. Proc Natl Acad Sci USA 2008;105:5549.
42 HANDBOOK OF CRYSTAL GROWTH
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy

[125] Wu C-Y, Young L, Young D, Martel J, Young JD. Bions: a family of biomimetic mineralo-organic
complexes derived from biological fluids. PLoSONE 2013;8(9):e75501. http://dx.doi.org/10.1371/
journal.pone.0075501.
[126] Martel J, Young D, Peng H-H, Wu C-Y, Young JD. Annu Rev Earth Planet Sci 2012;40:167.
[127] Satoh H, Tsukamoto K, Garcı
´a-Ruiz JM. Formation of chemical gardens on granitic rock: a new
type of alteration for alkaline systems. Eur J Mineral 2014;26(3):415–26. http://dx.doi.org/10.1127/
0935-1221/2014/0026-2378.
[128] NcCollom TM, Seewald JS. Elements 2013;9:129.
[129] Duffy TS. Rep Prog Phys 2005;68:1811.
[130] Zhang W, Oganov AR, Goncharov AF, Zhu Q, Boulfelfel SE, Lyakhov AO, et al. Unexpected stable
stoichiometries of sodium chlorides. Science 2013;342(6165):1502–5. http://dx.doi.org/10.1126/
science.1244989.
[131] Organov AR, Ono S. Nature 2004;430:445.
[132] Hirose K, Brodholt J, Lay T, Yuen D, editors. An introduction to post-perovskite: the last mantle
phase transition. American Geophysical Union; 2007. http://dx.doi.org/10.1029/174GM02.
[133] Wilde SA, Valley JW, Peck WH, Graham CM. Nature 2001;409:175.
[134] Bell EA, Harrison TM, McCulloch MT, Young ED. Geochem Cosmochim Acta 2011;75:4816.
[135] Bish DL, Blake DF, Vaniman DT, Chipera SJ, Morris RV, Ming DW, et al. Science 2013;341:1238932.
[136] Vaniman DT, Bish DL, Ming DW, Bristow TF, Morris RV, Blake DF, et al. Science 2013;343:
1243480.
[137] Nagashima K, Tsukamoto K, Satoh H, Kobatake H, Dold P. J Cryst Growth 2006;293:193–7.
[138] Miura H, Yokoyama E, Nagashima K, Tsukamoto K, Srivastava A. J Appl Phys 2010;108:114912.
[139] Lee J, Tanaka T, Mori H. Nanotechnology 2009;20:475706.
[140] Mori H, Komatsu M, Takeda K, Fujita H. Philos Mag Lett 1991;63:173.
[141] Kimura Y. In: Mann I, Meyer-Vernet N, Czechowski A, editors. Nanodust in the solar system:
discoveries and interpretations. Springer-Verlag; 2012. pp. 31–46.
[142] Kimura Y, Nuth III JA, Chakraborty S, Thiemens MH. Meteorit Planet Sci 2007;42:1429.
[143] Kimura Y, Nuth III JA, Ferguson FT. Meteorit Planet Sci 2006;41(5):673.
[144] Kimura Y, Nuth III JA, Ferguson Frank T. Astrophys J Lett 2005;632:L159.
[145] Saito M, Kimura Y. Astrophys J Lett 2009;703:L147.
[146] Kimura Y, Sato T, Nakamura N, Nozawa J, Nakamura T, Tsukamoto K, et al. Nat Commun 2013;4:
2649.
[147] Malfait K, Waelkens C, Waters LBFM, Vandenbussche B, Huygen E. Graauw MS van. Astron.
Astrophys 1998;332:L25.
[148] Molster FJ, Waters LB, Tielens A, Koike C, Chibara H. Astron Astrophys 2002;382:241.
[149] Apal D, Lauretta DS, editors. Protoplanetary dust. Cambridge University Press; 2010. p. 396.
Chapter 1 • Crystal Growth in Geology: Patterns on the Rocks 43
Handbook of Crystal Growth, Vol. II, Second Edition, 2015, 1–43
Author's personal copy