THE ORIGIN OF HIGH-ALUMINA CLAY MINERALS

--A REVIEW
by
W , D. KELLER
University of Missouri at Columbia

ABSTRACT
High-alumina minerals refer, in this paper, to Al-rich, normal sedimentary phases,
including gibbsite, boehmite, diaspore, possibly "proto-diaspore" and "proto-alunite",
"Al-chlorite", and corundum, in association with kaolin minerals. They may be derived
from any common Al-containing rock. Processes of origin include direct bauxitization
of non-clay silicate minerals and rocks, and the desilication of any of the common clay
minerals, particularly of the kaolin group. Apparently aluminous gels were formed
within certain marshy basins, and/or were transported into basins, giving rise to
concretionary masses of high-alumina minerals. Concretionary deposits were iormed
by dissolution of A1 and reprecipitation.
Alumina is probably more soluble, inherently, than is interpreted from geologic
occurrences. Even after A1 is dissolved under appropriate pH, or complexing reactions,
its solubility is obscured by Al-immobilization through fixation within 3-layer clay
minerals, or absorbed by other phases. A 14~ phase, probably aluminous, is not un-
common in certain soils, and in flint and boehmite-diaspore-containing clays. Subtraction
of silica by plants, is accompanied by concentration of alumina.

INTRODUCTION

THE t e r m " h i g h - a l u m i n a c l a y s " h a s b e e n a p p l i e d to v a r i o u s clays (rocks)

t h a t c o n t a i n A120 3 in excess of c e r t a i n m i n i m u m a m o u n t s - - t h e m i n i m u m
c o m m o n l y h a s b e e n set b y t h e use t o which t h e h i g h - a l u m i n a c l a y was
p u t , a s for refractories, a b r a s i v e s , a l u m i n u m m e t a l e x t r a c t i o n , or o t h e r
purposes. P r o b a b l y t h e lowest a l u m i n a c o n t e n t a p p l i c a b l e to high a l u m i n a
h a s been a b o u t 35 p e r c e n t A120 a (Wilson a n d Treasher, 1938; Allen, 1946;
W i l l i a m s , 1960) or d o w n to 20 p e r cent AlsO 3 for h i g h - a l u m i n a kaolinitic
c l a y (Mark, 1962), for t h e p u r p o s e s of this p a p e r t h e low l i m i t of high-
a l u m i n a clays will be p l a c e d a r b i t r a r i l y a t 40 p e r cent, A1203 on t h e r a w
c l a y basis, i.e., j u s t a b o v e t h e t h e o r e t i c a l a l u m i n a c o n t e n t of k a o l i n i t e
(39.5 p e r cent A1,O3). This r e q u i r e s t h a t some a l u m i n u m oxide m i n e r a l
(usually h y d r a t e d ) be a s s o c i a t e d w i t h t h e c l a y m i n e r a l s c o m p r i s i n g t h e
c l a y rocks. T h e c l a y m i n e r a l s in such h i g h - a l u m i n a clays c o m m o n l y are
in t h e k a o l i n group, a n d t h e c l a y rocks (mixtures) ensuing m a y be t h o s e
classified as n o d u l a r clays, b u r l e y clays, b a u x i t i c kaolins, t e r r a rosa, a n d
129
130 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
perhaps lithomarge, bole, and others. Clays having the highest alumina
contents include the bauxites, diaspore and boehmite clays, and laterites
of certain writers.
Interestingly, the clays highest ill alumina content are composed of
hydrated aluminum oxides, and paradoxically not of clay minerals, in the
way they have been defined tentatively by CIPEA as "hydrated silicates
with layer or chain l a t t i c e s . . . " (Mackenzie, 1959).
Hydrated aluminum oxide minerals present in high-alumina clays
include alpha-bayerite, AI(OH)a; boehmite, AIO(OH); diaspore, HAl09.;
and gibbsite, AI(OH)3; and possibly "cliachite, a mineraloid, AllOa.xH~O"
(although high-resolution diffraction may show this to be comprised of
other crystalline minerals). Relatively unexplored as high-alumina
minerals, or their associates, are double hydroxides of A1 and Mg (Turner
and Brydon, 1962), and poorly ordered---possibly "proto" varieties of--
diaspore and alunite.
Corundum, although less commonly observed in normal sedimentary
bauxite, was so reported by Beneslavsky in 1958 (published in 1962) when
he quoted Terentieva's find of corundum in bauxite in 1949. Alpha alumina
(corundum) was observed by Dr. R. C. Cross* (Aluminum Company of
America) in the hard, dark, vitreous cores of bauxite pisolites ("Birds-eye
ore") in the Arkansas residual deposits. We have observed, weak but
recognizable, four major X-ray diffraction peaks for corundum from such
pisolite cores collected from Arkansas bauxite. Although Gordon, Tracey,
and Ellis (1958) do not mention corundum in the Arkansas bauxite
mineralogy, they report (1959, p.86) an analysis of "brown to black
pisolites" yielding "71 per cent Al~Oa" and "16.6 per cent ignition loss",
which probably indicates the presence of corundum. Since the ideal formula
for gibbsite calls for 65.4 per cent AlcOa and 34.6 per cent H~O, and for
boehmite 85 per cent A120 a and 15 per cent water, neither of which or-
dinarily occurs ideally pure, it is probable that some corundum contributed
to the 71 per cent AlzO a and relatively low ignition loss ill the brown to
black pisolites.
The processes of origin of high-alumina minerals, based on first-hand
observations and studies in particular districts, have been well documented
in a number of papers (Mead, 1915; Harrison, 1933; Harder, 1952; Bolger
and Weitz, 1952; Gordon and Tracey, 1952; and with Ellis, 1958; Bardossy,
1959; and Bates, 1962; and others). The purpose of this discussion is, not
to try to improve on those excellent papers, but to find from them, plus
some personal observations, common denominators of processes of origin
for high-alumina minerals. Proceeding from the principle that a mineral
is a response of appropriate geologic materals to energy impressed upon
them--from which it follows that the mineral formed tends to represent
the lowest-energy state available to the constituent elements--we shall

* Personal communication, 1963.

 THE ORIGIN OF HIGH-ALUMINA CLAY MINERALS--A REVIEW 131
inquire what materials, and what energies (by way of processes) are
productive of high-alumina minerals.

PARENT MATERIALS OF HIGH-ALUMINA MINERALS

Fox (1963) and Harder (1952) described bauxite deposits that had been
formed on essentially every type of common rock and rock-forming
mineral that contained A1. Thus, the parent material for the origin of high-
alumina clay is not critical, except that it must contain A1. Parent minerals
most favorable to high-alumina argillation were observed to be the felds-
pathoids and alkali varieties of dark minerals, i.e., subsilicic minerals, but
under sufficiently intense conditions of leaching, intermediate and sialic
(even quartz-containing) igneous rocks, clay residues from dissolving
limestones, and metamorphic rocks of most kinds also produced large
bodies of high-alumina clays. The formation of a high-alumina deposit,
especially where metastable high-alumina minerals occur, is a step in a
slow-reaction process; subsilicic minerals react more quickly, but others
eventually are desilicated under intense leaching. Likewise, solutions
carrying dissolved A1 ions yield fillings of high-alumina minerals in veins,
and aluminous gels or colloids also give rise to high-alumina minerals.
High-alumina minerals may arise by concentration of A1 compounds
during the removal of non-aluminous associated elements, by exceeding
the solubility of AI in aqueous systems, by sorption of A1, or ion-exchange
for A1, and b y flocculation or gelation of an M-containing colloid. In
ordinary geologic environments, the dissolution and precipitation of alu-
mina may be modified by the pH, the presence of alkali metal and alkaline

T A B L E 1 . - - P R O C E S S OF HIGH-ALUMINA ARGILLATION

Derivation from non-clay silicate minerals by weathering,

Derivation from clay minerals by weathering,
Dissolution and precipitation of A1 in relation to pH,
Flocculation and gelation of A1 compounds,
Precipitation with dissolved silica,
Sorption by clays (e.g., Al-chlorite), Al-bonding, and
" anti-gibbsitization",
Authigenesis,
Replacements,
Structure of parent material,
Organisms and organic residues.

earth cations, dissolved silica, the presence of sorbtive clay and other
minerals, macro- and micro-floras and their residues, temperature, and
pressure. Individual processes effective in high-alumina argillation are
listed in Table 1.
132 TWELFTH ~2~IATIONALCONFERENCE ON CLAYS AND CLAY MINERALS

P H A S E R E L A T I O N S IN T H E S Y S T E M A1,O3-H,O

The system Al,Os-H20 was studied in 1943 by Laubengayer and Weisz.

They found reactions to proceed sluggishly, but that the transition, diaspore
to corundum plus water, and reversal, could be made from both directions.
They suggested that boehmite is never a stable phase in tile Al~Os-HzO
system.
In 1951, Ervin and Osborn confirmed the results of Laubengayer and
Weisz and studied the system at higher temperatures and pressures. They
found the sequence of stability field with rising temperature to be gibbsite
to boehmite to corundum, and the "pressure and temperature minima for
the stable existence of diaspore" to be 2000 lb/in 2, and 275~ They noted
that geologic field relations do not support their conclusions regarding
minimum T and P for diaspore, and stated, in connection with experi-
mental difficulties, that, "evidence for equilibrium is only indirect, and it
is least satisfactory along the boehmite-diaspore transition curve".
Kennedy (1959) investigated further the A12Os-H20 system and con-
firmed much of the work of Ervin and Osborn but modified the previous
interpretation of the boehmite-diaspore relationship. Kennedy pointed out
that, "the gibbsite-boehmite transition and the boehmite-diaspore transi-
tion are bottl exceedingly sluggish. Further, in many low temperature runs
nucleation of diaspore is n e c e s s a r y . . . " "A good deal of evidence, internal
to the phase diagrams, suggests that the results . . . do not represent
equilibrium determinations." With regard to the boehmite-diaspore fields,
"observed slopes of the boundaries separating the phases suggests rather
strongly that a rate process rather than an equilibrium boundary is
involved, as Laugengayer and Weisz suggested".
Kennedy further, "concluded that all boehmite forms metastably. The
approximate slope of the boehmite-diaspore boundary, computed from
thermodynamic data, indicates that diaspore in bauxite and clay deposits
may have formed at atmospheric pressures and temperatures, a familiar
conclusion from field observations." To this point, he wrote, " I t thus
appears that within the limitations of the thermodynamic and experi-
mental data, boehmite is metastable at all temperatures and pressures and
probably always forms metastably in natural despoits. Diaspore is then the
stable phase in equilibrium with gibbsite."
From Kennedy's low H20-pressure diagram we would expect gibbsite
to be formed at the earth's surface and this to be dehydrated to diaspore
either directly, or via the metastable boehmite. Gibbsite, diaspore, and
boehmite go to corundum when heated sufficiently, ordinarily above
1300~ in industrial application. Bratton and Brindley (1962) showed
(as has also been observed by manufacturers of high-alumina refractories)
that the path of conversion is modified by structure of the clay. Diaspore
goes to corundum in a single step, whereas boehmite goes through gamma-
 THE ORIGIN OF HIGH-ALuMINA CLAY MINERALS---A REVIEW 133
alumina and gibbsite may go through either of two avenues of transition
before arriving at corundum. That heating far above room temperature to
convert gibbsite to corundum is not necessary from thermodynamic con-
siderations was calculated by Carpenter,* who found that the thermo-
dynamic conditions for transition of gibbsite to corundum-plus-3H,O to
be 1.2 per cent relative humidity at 25~ It is probable that such dehydra-
tion of gibbsite, or possibly syneresis of an alumina gel under low relative
humidity, gave rise to the corundum in the Arkansas bauxite.
Because high-alumina clays in large deposits are regularly associated with
kaolins or other clay minerals, the geological systems in which high-
alumina clays are formed usually contain also SiO~ along with A120 s and
H,O. Information on this system (Si02-A12Os-H,O) is scanty. Kennedy
(1961) reported preliminary studies on it at high temperatures and pres-
s u r e s - t o o high to have yielded much information pertinent to the problem
before us. At ordinary earth-surface temperature and pressure, reaction-
yields of minerals in the system SiO~-AI,Os-H20 are exceedingly low
(de Kimpe, Gastuche, and Brindley, 1961). Because chemical stability
relationships are incompletely known we shall turn to geological observa-
tions, which are next described.

DERIVATION FROM NON-CLAY SILICATE MINERALS

BY W E A T H E R I N G

Among early reports on high-alumina argiUation, Mead's work on the

Arkansas bauxite in 1915 is outstanding. At a time when favored inter-
pretations of origin of bauxite included action of hot springs, replacement
by rising thermal waters, bacterial action, and others (see Fox, 1932,
pp.72-77) Mead's conclusion that subaerial weathering was the process by
which bauxite was formed represented a distinct advance in thinking.
Mead described the formation of bauxite as a process in gradual desilication
of the nepheline syenite (Arkansas) in which kaolinite occurred as an inter-
mediate stage between syenite and bauxite.
Harrison (1933), however, in careful, detailed, chemical and mineralogical
work on Guiana bauxite and "pot clays" (largely kaolinite), concluded
that much gibbsite ("primary laterite") was formed directly (not preceded
by kaolinite) from basic and intermediate igneous and metamorphic rocks.
He stated that, "the mineral of first and direct formation from the plagioclase
is gibbsite" (p.37 of Harrison, his italics), and that resilication of the primary
laterite resulted in lateritic earths or argillaceous laterite. Obviously
this is a reverse mechanism of that proposed by Mead for the Arkansas
deposits.
* Personal communication, Alden B. Carpenter, May, 1963, from a study in prepara-
tion for the 13th conf.
134 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
Gordon, Tracey, and Ellis (1958), who studied the Arkansas deposits in
detail, concluded that "Bauxitization of the nepheline syenite resulted in
the formation of granitic-textured bauxite that contains microcrystalline
gibbsite as the main product. The place where this process occurred has
been termed the zone of leaching" (Type I deposits; p.146). "Kaolin was
produced chiefly b y weathering of the nepheliue syenite beneath the
bauxite, following a rise in the water table that accompanied the deposition
of the Eocene sediments." "Some of the silica released by this weathering
process went into resilication of the bauxite." Thus Gordon et al. con-
cluded that the nepheline syenite could alter to either gibbsite or kaolinite,
depending upon the nature of the environment, and that either gibbsite
or kaolinite could precede the other in genetic sequence.
Allen (1952), after extensive petrographic studies of bauxites and dias-
pore clays in the U.S., cited evidence for, and favored the process that,
"The desilication of feldspars and parent rocks is usually a two-stage pro-
cess in which clay minerals such as kaolinite, halloysite, and montronite are
formed first and then silica is removed from them to form gibbsite, boeh-
mite, and diaspore." Some occurrences in Georgia showed relationships,
however, that to Allen (1952, p.678), "suggest that some 'chimney rock'
specimens were formed by resilication of gibbsite".
Andesite in Hawaii is altered under present-day weathering mainly to
halloysite or Al-gel, and thence to gibbsite, according to Bates (1962).
He stated: "Halloysite forms from plagioclase usually by alteration of first
the core and subsequently the rim of the feldspar laths. No kaolinite was
observed, its absence being attributed to the absence of fine-grained mica
as an intermediate weathering product. Gibbsite is produced by (1) removal
of silica from halloysite, (2) dehydration of Al-gel, and (3) precipitation
from solution. Although it is possible that the mineral (gibbsite) may form
directly from feldspar, halloysite is the common crystalline intermediate
on both the megascopic and microscopic scales."
The mechanism that Bates describes for the gibbsitization of andesite
(feldspar in andesite) does not conform to that of Gordon et al. for nepheline
syenite, and of Harrison for basic to intermediate rocks. One wonders if the
different parent rocks, or different environments of weathering, account for
the different mechanisms; or even further, was there a submicroscopic film of
intermediary halloysite present in the rocks of Harrison, and Gordon et al.,
but undetected by them? I have examined in thin section, under high
magnification, the direct gibbsitization of a diabase from the "Hope
Quarry" of Harrison (collected by E. C. Harder, sent to C. S. Ross, and
given to me), and do not observe halloysite between the feldspar and the
gibbsite, but one can always ask, "Would a transition layer have been seen
if the resolution of the microscope had been better?" Since both Bates, and
Gordon et al., used both field and microscopic evidence for each of their
conclusions, it is reasonable to believe that the results of each investigation,
although antithetical, were reported correctly for the particular systems
 THE ORIGIN OF HIGH-ALUMINACLAY MINERALS--A REVIEW 135
with which each was concerned. We shall now inquire into possible ways
by which high-alumina argillation operates.

M E C H A N I S M S OF D I R E C T ( P R I M A R Y G I B B S I T E )
BAUXITIZATION

Gordon, Tracey, and Ellis (1958) considered in detail the effects of the
following mechanisms in t h e weathering of nepheline syenite to bauxite:
leaching, concretion, formation of boehmite and magnetic iron oxide,
migration of A13+ in solution, and formation of kaolin bodies. B~rdossy
(1959, Tomus VI, pp.21-53) emphasized the effect of pH on the desilication
of clay minerals to high-alumina minerals. Keller (1958) considered the
possible effect of pH and metal cations on the surface of hydrolyzing silicates
towards separating silica and alumina (direct bauxitization) during dissolu-
tion, or alternatively towards effecting their combination as clay mineral
aluminum silicates.
Harrison, Gordon et al., and Keller concur that direct bauxitization can

IO

a.

:5

-- 4
i
3
SIO2
2 - -

o I I I, I i
o J 2 3 4 9 s 7 a e to
pH
AI203 data from Corrans
5102 data from Krauskopf

fIGURE 1 . ~ S o l u b i l i t y of A I , O s a n d a m o r p h o u s SiO 2 i n w a t e r ,
136 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
result from the leaching of primary silicates. Harrison's conclusion has
been quoted (p. 133), and Gordon et al. state, "The first part of the bauxitiza-
tion process resulted in the formation of bauxite that preserved to some
degree the original texture of the nepheline syenite . . . . The principal end
product was the microcrystalline form of gibbsite, derived chiefly from
alteration of feldspar." Keller (1958) and B~rdossy (1959) independently
related the separation of silica and alumina by differential solubilities
influenced by pH. These relationships are explained as follows.
In Fig. 1 is shown the relation between pH and the solubilities of amorphous
silica (Krauskof, 1959) and AI~O3 (Correns, 1949) in distilled water. These
data may be used as guides to relative solubilities of SiO, and AI~O8 in
ground water, although the absolute solubility of SiO, is decreased signifi-
cantly by associated alumina in solution (Okamoto, Okura and Goto, 1957),
and may be greatly enhanced by the presence of certain organic matter
(Lovering, 1959). Above pH 10, both SiO 2 and AI~Oa are relatively highly
soluble, whereas between pH 5 and 9, AI~O3 is relatively insoluble although
Si02 still remains dissolved in low concentration.
The pH of the solution of gibbsite in H~O (at equilibrium) controls the
distribution of A1a+ and AlOe-. At pH 5.1, the activities of A1a+ and A10~---
are approximately equal (at thermodynamic equilibrium), whereas at pH 4,
the activities ratio of A13+ to A10~-- is approximately 104: 1, and at
pH 10.5, it is approximately 1 : 1032, respectively (personal communica-
tion, A. B. Carpenter).
Referring again to Fig. l, when hydrolysis of an aluminum silicate
occurs at p H l l, both Si02 and Al~O3 are relatively soluble and mobile,
and can be removed, even to long distances if the pH remains at 11. Alu-
minum is present predominantly as A102-, as has been noted. As the
solution becomes more acid, at pH 8 for example, aluminum precondi-
tioned as: A10~-- will be precipitated, probably as gibbsite, or possibly as
metastable boehmite (?), A10(OH), or diaspore, HAlOs, and the silica
may remain in solution if sufficiently dilute. In this way, primary bauxite
might be formed directly on the surface of the hydrolyzing silicate without
intervening kaolinization, or precipitated elsewhere so that the pH was
lowered.
Question may be raised if minerals yield pH 10 or above upon hydrolysis.
Nepheline, diopside, and actinolite yield abrasion pH's of 11 (Stevens and
Carron, 1948; Keller, 1958) showing that the surfaces of certain hydrolyzing
minerals are highly alkaline. In a solubility experiment (Keller, Balgord
and Ree~man, 1963), nepheline and diopside were pulverized under
distilled 'water exposed to the atmosphere by grinding 10 g of mineral
in 100 ml water for one hour, at which time the pH's of the slurries were
8.8 and 9.0, respectively, Water taken from an intermittent seep in jointed
bare nepheline syneite at Bauxite, Arkansas, about 3 hours after a summer
shower yielded pH 7.9, measured in the field (Keller, 1960). Therefore,
the pH range and solubilities shown in Fig. 1 are geologically realistic. In
 THE ORIGIN OF HIGH-ALUMINACLAY MINEI~ALS--A REVIEW 137
the experiments by Keller, Balgord, and Reesman (1963), previously cited,
nepheline yielded 39 ppm silica in solution (diopside 50 ppm), and 8.3 ppm
alumina. In contrast, labradorite and microcline both at pH 8.0 yielded
only 8.2 and 15.6 ppm SiO~ and 1.25 and 3.25 ppm A12Os, respectively.
These data lend support to a possibility of alteration consistent with
the conclusion by Harrison that "under tropical conditions basic and
intermediate r o c k s . . , leave an earthy residuum of aluminum trihydratet
9 . ." Under the tropical conditions of Harrison, it can be assumed that rain
fell almost every day, keeping the rocks continuously wet and flushing
away the hydrolysate so that concentrations of dissolved silica never
reached the saturation point. On the contrary, if saturation of the solution
had occurred for a considerable time, either because of (1) lack of rain, (2)
evaporation, or (3) ionic build-up during water logging, pIobably a com-
bination of silica and alumina would have formed kaolin minerals, Fligh
rainfall provides the important factor of always dilute solution of ground
water.
The mechanism based on control by high pH and low concentrations of
dissolVed alumina and silica, is proposed as one way hypothetically, by
which direct bauxitization of non-clay silicates may occur. Its operation
may not have been unequivocally proved, but it describes logically a way
by which the "cryptocrystalline or so-called amorphous form of gibbsite"
in Arkansas (Gordon et al., 1958, p,83) may be formed9
Kaolinization in association with bauxite9 Kaolin minerals (meant to
include endellite, hallysite, b-axis disordered kaolinite or "fire clay mineral",
and well-ordered kaolinite) may form by partial desilication of the primary
silicate rock or mineral (Bates, 1962) or resilication of gibbsite (Gordon
et al., 1958). Gordon et al. (1958, p. 149) state, "It seems likely that b o t h . . .
migration of kaolin in solution and resilication of bauxite . . . may have
taken place simultaneously, as Mohr has suggested that silica and kaolin
behave alike under the same ground-water conditions". Harrison (1933,
p.10) concluded that "under tropical conditions, acidic rocks, such as
aplites, pegmatites, or granites and granitic gneisses, do not undergo primary
laterization, but gradually change into pipe or pot-clays, and more or less
quartziferous and impure kaolins". Harrison, like Mead, concluded that
kaolins could be desilicated to bauxite: "Under tropical conditions, lateritic
earths, and even pot clays, may undergo desilication, with the formation
of concretionary and superficial masses of bauxite."
To account for kaolinization (not direct bauxitization) of aluminum
silicates, Keller (1958) has interpreted the control by pH of solubilities of
AltOs and SiO3 as resulting in the combination of AI~Os and SiO2 below
probably pH 6 to form kaolin minerals. In the laboratory, De Kimpe,
Gastuche, and Brindley (1961) synthesized kaolinite at room temperature
at pH 4.5 and lower. Patterson and Hosterman (1962) found the pH index
of all 11 flint clay (kaolinite) samples (Table 5, p.F56) cited from Kentucky
to range from 6.0 to 4.3. I have found the pH of water standing in pools
138 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
in deposits o1 flint fire clay (kaolinitic) in Kentucky, Ohio, Pennsylvania,
Maryland, and Missouri, and in kaolins of Alabama and Georgia, to be
consistently below 5. In occurrences where pyrite is or was present, acidity
would be conditioned by sulphuric acid formed from pyrite, but p H 4.0
to 4.5 was found regularly also where no pyrite was observed, as in some
Missouri flint clay pits. Pools of water in dolomite, on the other hand,
occurring a hundred feet from flint clay (in Missouri) yield p H 6.8 to 7,
obviously a function of reaction with the dolomite. A p H of 4.5, or less, in
association with kaolinite is so c o m m o n that it must be truly significant.
From free-energy calculations* in the system gibbsite-silica-water,
gibbsite exists (25~ and I atmosphere) in thermodynamic equilibrium
with kaolinite at pH 5.1, the activity of Si (from amorphous silica) at
10--a'~, and activity of A1 at 10-9.45 (where Ap+ and AlOe- are equal in
activity). This is a point of minimum activity for total AP + and A10 ~ .
When the activity of Si is increased to 10-3.0 at the same pH (5.1), and
activity of aluminum is only slightly less, (AP+) -1~ (A1Oz--)-1~
kaolinite is in equilibrium with amorphous silica. Therefore, a variation
of only about ten-fold in the activity of silica at pFt 5.1, ideally should
reverse the gibbsite-kaolinite reaction.
The kaolinite stability field in contact with gibbsite widens as pH be-
comes either lower or higher thalz 5.1 and simultaneously the activity
of silica remains almost constant but t h a t of aluminum increases. The
kaolinite-gibbsite stability .fields widen most prominently from pH 5.1 up
to pH 9.8, above which the gibbsite field widens at the expense of silica
even though the activity of Si may increase 10-fold.
Factors other .than pH that condition the reaction of gibbsite-kaolinite
include adequate concentrations (activities) of Al,03 and SiO 2 in solution
related to rainfall and efficiency of ground water leaching, various com-
plexing inorganic substances, complexing and protective organic com-
pounds, effects of alkali metal ions to promote solubility of silica, and
alkaline earth ions as flocculants of silica, and others.
Secondary, and other primary (?), effects by dissolution, reconstitution, and
precipitation in a low-pH range. The preceding discussion has emphasized
direct bauxitization in an environment of high pH wherein Also * is rela-
tively soluble. This viewpoint does not in the least preclude effective dis-
solution of alumina at low pH, 4.5 and less, and the development of endel-
lite, and possibly allophane and alumina gel, in this range of acidity.
Indeed, such acid environments have almost certainly been very important
in the dissolution, reconstitution, and precipitation of high-alumina
minerals.
Essentially every investigator of high-alumina deposits has reported
second-stage overgrowths, cavity fillings, and veins of gibbsite, boehmite,
diaspore and clay minerals in the deposits. I believe that most of this
transfer of alumina has been done at low pH for the following reasons. In
9 Personal communication, A. B. Carpenter, October, 1963.
 THE ORIGIN OF HIGH-ALUMINACLAY MINERALS--A REVIEW 139
the Arkansas bauxite region, a pyritic-marcasitic lignite overlies or is
interbedded with, the bauxite. Sulphuric acid Solutions yielding pH 1
(Hydrion paper test) are generated as oxidizing water percolates through
the coal and into the bauxite. This acidity, and pH up to 4, renders
alumina quite soluble. Water in the rills draining the bauxite under lignite
is commonly at pH 3.5 (Hydrion paper test). Mining engineers in the area
state that drainage water is exceedingly acid during a heavy rain following
a dry spell. The formation of endellite under present-day alteration as clear
massive porcelainous and white granular varieties by the action of sul-
phuric acid (from oxidizing iron sulphate) on a red, granular argillacious
progenitor has been observed in Kentucky (paper presented by title at
this Conference; Keller, McGrain, and Saum). The moist endellite at this
occurrence yields pH 3.0 to 3.5, and altering parent clay, pH 1.$. Our
observations support and extend further those of Ross and Kerr (1934)
who wrote that "hatloysite... forms through weathering and by the action
of cool sulphate-beating solutions on aluminous materials". I suspect that
the sulphate radical, in addition to the acidity, is significant in argillation
within the kaolin group. Sulphur and silicon apparently interact in clay
minerals to a certain extent (Lesar, Krinbell, Keller, and Bradley, 1946).
Veins of white and porcelainous endellite are common in parts of the
bauxite in Arkansas. The development of endellite observed in Kentucky
suggests strongly that the vein-type endellite within Arkansas gibbsite
was formed by sulphide-coal-derived acid solutions reacting with pre-
existing clay and/or bauxite.
One is tempted to query if, in addition to endellite, gibbsite may be
formed directly (without intermediate solution stage) from primary silicate
rocks exposed to pH 4.5 and less. Conceivably it might be, but probably
it is not, because highly acid weathering solutions produce silica-rich and
alumina-poor (also iron-poor) podsols rather than lateritic soils from
silicate rocks. I have seen good field confirmation of the loss of Fe,O 3 and
alumina, and enrichment of silica in the present-day formation of siliceous
refractory clay deposits (report in preparation with J. F. Westcott). There-
fore, I doubt that direct bauxitization occurred in a highly acid environ-
ment, but certainly high-alumina minerals have been observed to originate
in abundance secondarily in an acid environment:

D E R I V A T I O N OF H I G H - A L U M I N A M I N E R A L S F R O M
CLAY M I N E R A L S BY W E A T H E R I N G

The development of bauxitic-lateritic clays and compounds in the A-

horizon of the soil by the desilication of common clay minerals (all types)
is so well known that it need not be dwelt upon. Mohr and van Baren (1954)
clearly show that leaching (as opposed to water-logging) in which silica
and cations other than A1 and Fe are removed results in bauxitic-lateritic
soil minerals.
 140 TWELFTH ~ATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
In Hawaii, beneath the subsoil in the zone of fresh rock, the andesite
weathers to halloysite which, in turn, is desilicated to gibbsite. Bates (1962)
has shown this mechanism to proceed on both macroscopic and microscopic
scales. Bates (p.315) further stated: "An amorphous transition state,
probably ranging in composition from allophane to Al-gel exists as part of
the change from hallosite to gibbsite . . . . The nature and relative abundance
of intermediate anc~ end products of weathering depend primarily upon
the rainfall, angle of slope, and texture of the rock."
Another type of high-alumina argillation is exemplified by the Creta-
ceous (?) Hungarian bauxites, the Pennsylvanian-age nodular diaspore
clay in Pennsylvania, and the Missouri burley and diaspore (rarely boeh-
mite) clays of Pennsylvanian age. The Hungarian and Missouri deposits
were formed in low-lying, marshy to karst-type topographies into which a
clayey sediment had been transported. B~rdossy (1959, Part IV, p.21)
wrote of the Hungarian bauxffes, "The original substance of Hungarian
bauxites was deposited under a subtropical to tropical climate in periodic-
ally inundated depression of the ancient landscape. This original substance,
of a clayey nature, has altered in the karstic depressions of the carbonatic
bedrock into bauxite in situ. It is not yet fully clear as to where the
original material of bauxite was derived from. So much seems to be certain,
however, that it was transported to its present site by stow, quiet streams
(fluviatile), in the form of muddy and colloidal suspensions . . . . The pH of
the colloid solutions should have been presumably rather acid (pH, 4-6) as
this is the only way of accounting for the joint transportation of A1, Fe,
and Ti on a greater scale . . . . The suspensions gradually coagulated and
settled . . . probably facilitated by the carbonatic character of the bed-
rock . . . . In all probability the predominant mineral was allophane or
kaolinite of the fire clay type.
" T h u s the kaolinitic (and fire clay-like) nature of the original primary
substance indicates a subaerial or freshwater sedimentation, under intense
leaching in an acid or neutral medium and an intensely oxidizing environ-
ment. These features also indicate a hot, moist climate. (Italics are B~rdossy's).
" I n our opinion the gradual leaching of silica was begun by solutions
turning alkaline (pH, 7-10) filtrating through the loose unconsolidated
s e d i m e n t . . , in the course of the dissolution of finely dispersed calcium
carbonate in water, free of C02, Ca(OH)~ is formed resulting in an intense
alkalinity pH, 9-10, of ground water . . . . The resulting highly basic solutions
decomposed the fire-clay mineral and the still amorphous aluminum silic-
ates (allophane). The alkalinity of the solvent simultaneously made possible
the leaching of the silica-gel and its washing out across the bedrock. At the
same time aluminum was instantly precipitated on decomposition . . . . "
The mineralogy of Hungarian bauxites varies with deposits. Gibbsite
occurs in maximum abundance at 75 per cent in the Kincses Jozef fields,
boehmite up to 85 per cent in 4 other areas, and diaspore as much as 70
per cent of the deposit at Nezsa. B~rdossy proposed that upon destruction
 THE ORIGIN OF HIGH-ALUMINA CLAY MINERALS--A REVIEW 141
of the kaolinite lattice and addition of water, gibbsite or boehmite result.
If gibbsite forms, it may dehydrate to boehmite. If the pH decreases to the
"neutral state", bayerite, boehmite, or amorphous trihydrate may form.
In regard to diaspore formation, however, B~rdossy (Part IV, p.32)
believes that, " T h e formation of diaspore by the bauxitization of kaolinite
is indeed hard to imagine, in the first place due to reasons of lattice
g e o m e t r y . . , the anionic lattice of kaolinite is cubic close packing, that of
diaspore hexagonal close packing. It is well-known that the stablest part
of the lattice is the anionic grid, whose rearrangement consumes an ex-
ceedingly great amount of energy." In the Transdanubian Mountains, and
the Nagyxzal and Nezsa deposits, "the primary occurrence of diaspore is
unlikely". But, "the situation at the Nagyharsany locality is somewhat
different. The very small iron content and the greater percentage of chlor-
ite" (in some Missouri diaspore deposits chlorite also occurs) "of this
bauxite is similar to the ones described by Schfiller . . . . " "It is, in our
opinion, possible that the none too great diaspore (10 per cent average)
content of the Nagyharsany bauxite should be a primary formation." The
preceding quotations were carefully selected from B~rdossy's discussion
of the paragenesis of the Hungarian bauxites to present the most salient
features of his ideas of their origin; the entire, comprehensive report of
approximately 135 pages is recommended.
Description of the Hungarian bauxite deposits indicate many similarities
with parts of the Missouri diaspore deposits. The outstanding mineralogic
difference between them is that no gibbsite has been found in any Misc~ouri
high-alumina deposit, and boehmite is infrequent in occurrence. A com-
parison of theories of origin for the two regions will be made after the
Missouri deposits are described. Considerable agreement in theory may be
anticipated, but I find difficulty in accepting B~rdossy's proposal that water
dissolving CaCOa will be CO,-free, enabling it to yield pH 9-I0 (from
Ca(OH) e).
The origin of high-alumina minerals, dominantly diaspore and minor
boehmite, in Missouri diaspore clay deposits has been discussed by Allen,
1935, 1955; Burst, 1952; Keller, 1952; Keller, Westcott, and Bledsoe, 1954;
McQueen, 1943; and others. All of these writers concur in the first phase
of the origin of the clays, namely, that clay colloids were carried by
probably slow-moving streams into low-lying marshes, karstic depressions,
and sink holes dissolved into limestones and dolostones of Mississippiau
and Ordovician ages. Allen (1955) and McQueen (1943) interpreted that
ground water, moving predominantly downward, leached and desilicated
the clay minerals to yield diaspore and boehmite. Keller earlier subscribed
to the same hypothesis, but modified it in part when the Bucker clay de-
posit (Keller, 1952; Keller, Westcott, and Bledsoe, 1955)was opened and
exhibited unequivocal evidence that the high-alumina clay not only had
been formed, but was subsequently subjected to weathering and oxidation,
all of this occurring during Pennsylvanian time prior to the deposition of
142 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
the Excello Shale (Cabaniss Group) and Blackjack Creek Limestone of the
Fort Scott Limestone of Pennsylvanian age (Searight, 1955). Keller,
Westcott, and Bledsoe (1955) interpreted the removal of relatively large
amounts of silica and metal cations, by dialysis from the clay colloids by
means of the Donnan effect and flushing off the top of the low-lying marshy
basins during slow sedimentational filling of the basins. Reasons for signifi-
cant removal at the top that augmented downward leaching include, (1)
no evidence of silica migration or deposition into the underlying shale,
sandstone, or limestone (including dolostone) which enclose the clay
deposits, (2) the high impermeability of clay colloids and the flint clay,
(3) non-systematic distribution of high-alumina minerals in the flint clay,
(4) no evidence of downward solution channels and associated high-alumina
enrichment in the large diaspore deposits, (5) relatively horizontal layers
of high-alumina minerals at the top of deposits and accompanying con-
centration of silica and fluxing metals in lower parts of deposit as would
have occurred if fresh-water leaching started at the top and proceeded
downward, and (6) completion of high-alumina argillation in pre-Fort
Scott time. Desilication was aided extensively, it is thought, by the action
of plants growing in and around the marshes.
The source of the clay colloids in the Missouri deposits was the clay
fractions of weathered limestones of Mississippian and Ordovician ages
surrounding the clay deposits (Robbins and Keller, 1952). These colloids
may have ranged in alumina composition from illite to aluminous gels,
modifying accordingly the extent of desilication necessary to yield diaspore
and boehmite.
The formation of Missouri diaspore probably represents a rate change
towards diaspore rather than achievement of equilibrium. X-ray diffracto-
grams of certain diaspore-bearing clays exhibit variations in intensities
believed to be dependent upon degree of ordering of the crystals of diaspore
(not sample-orientation effects), and some of these may be interpreted as
being produced by "proto-diaspore". Likewise, in at least one K-rich (low
fusion, "alkali" clay) deposit, diffractograms suggestive of alunite have
been obtained from the high-alumina clay. Interplanar spacings suggest
alunite, although spacings and intensities deviate from those yielded by
first-quality reference alunite. Possibly a "proto-alunite" was being evolved.
The origin of the Mercer fire clay (Pennsylvanian in age) occurring in
Pennsylvania was interpreted in 1952 by Bolger and Weitz to be as
follows: "the source material of the Mercer was a colloidal mixture of
hydrous aluminum silicates and aluminum hydroxides. These colloidal gels
were then concentrated and crystallized to form kaolinite, diaspore, and
boehmite. After lithification, shrinkage cracks appeared, into which a
silica-rich kaolinitic material was introduced, resulting in the replacement
of diaspore by kaolinite. The silication of diaspore released free alumina,
which recrystallized in voids as a later stage of diaspore." Bolger and Weitz
do not elaborate on the source or composition of the silica-rich kaolinitic
 THE ORIGIN OF HIGH-ALUMINA CLAY M I N E R A L S - - A REVIEW 143

material. They do not explain the chemical drive by which incoming silica
is caused simultaneously to combine with diaspore to form kaolinite and
release alumina into solution.
Chlorite has been referred to in previous quotations from B~rdossy (1959)
on the Hungarian bauxites, and from Patterson and Hosterman (1962, pp.9,
54, 57) who noted probably chlorite in the flint clay in Kentucky. Eleven
years ago, Keller (1952, p.125) reported a "chloritic clay mineral", and
Hahn (1954) confirmed and studied this 14A clay phase, presumably
chlorite, in varying abundance in a Missouri diaspore clay deposit (Oches-
key) that was notably rich in boehmite. More recently, we have tried to
concentrate and isolate the 14/~ phase for detailed study, but repeated
efforts to separate it from admixed kaolinite and boehmite by grinding,
sedimentation in water, passing through an isodynamic magnetic separator,
and suspension in heavy liquids have enriched the fraction only to a
content of about 20 per cent (estimated conservatively) of the 14/~ phase
(Fig. 2). A chemical analysis of the best-fractionated sample yielded the
results listed in Table 2.

9O
eo CHO03?
?o~- dJa ? Kao
/ Kao 4.73 7.13
6o{- 3.56 I l

40 613 Ch

It II ill
65 55 45 55 25 ZO 15 Io 5
Hahn C6-10
FIGURE 2 . - - X - r a y diffractogram o5 the clay concentrate from the Ocheskey
pit showing characteristic peaks of chlorite (?), 14.1A, not destroyed by
heating to 525~ kaolinite, 7.13A, which is destroyed by heating to 525~
boehmite, 6.13/~; diaspore and 003 of chlorite (?), 4.73/~. The 060 spacing
record of the trace is not definitive with respect to a di- or tri-octahedral
character of chlorite.

Because the fractionated C-10 sample is not a pure single phase, a n d

the mixture with kaolinite, boehmite, TiO2, and iron oxides is complex,
it is not possible to estimate quantitatively the composition of pure 14A-
phase and calculate its structural formula. Nevertheless, the paucity of
MgO, 0.11 per cent, demands that the chlorite (?), if it is a chlorite, must
be aluminous. From Foster's (1962) study of chlorites, those highest in
alumina run approximately the ratio 0.11 MgO : 0.11 A1303 : 0.12 SiO 2: 0.06
H~O. In gram-per cent, utilizing all of the MgO for chlorite, 0.11 per
cent MgO would yield about 0.4 per cent chlorite in the C-10 sample,
144 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS

TABLE 2.--CHEMICAL ANALYSIS OF FRACTIONATED

C-10 SAMPLE, OCItESKEY DEPOSIT
SiOz 25.09
A120s 56.71
Fe20 a 0.70
FeO 0.20
CaO 0.07
MgO 0.11
NaaO 0.52
K20 1.04
TiO I 1.48
Loss Ignition 13.99
99.91

Analysis by Bruce Williams Laboratory, Joplin, Mo.

Cost of analysis defrayed by NSF GP-1057

which is obviously far, far below that estimated from the diffractogram--
indeed, the X-ray would not "see and record" less than one per cent
chlorite. A clay containing 20 per cent of "normal" Mg-containing chlorite
highest in A1 of those found by Foster should contain 4 to 6 per cent, not
0.11 per cent, MgO, which is actually present. Therefore, the 14A phase in
the Missouri clay must be high in alumina, presumably an Al-chlorite.
The origin for it is postulated to be precipitation-fixation by a 3-layer
clay, most probably a weathered, expanding illite ("montmorillonite"), of
migrant, dissolved alumina in the manner by which chlorite-like phases
were synthesized by Slaughter and Milne (1960).
Although properties of the Al-chlorite (?) cannOt be determined from
the mixture, it was observed that it did not expand with ethylene glycol,
the 14A reflection was not diminished appreciably by digesting the sample
in 6N HC1 at about 90~ for 12 hours, and heating the clay to 525~ over-
night did not modify the 14.A, reflection. In these respects it is similar
to a mineral Patterson and Hosterman (1962, p.F64) found in the Olive
Hill, Kentucky, clay bed, which they called "probably chlorite . . . this
mineral has a fundamental vasal spacing of 14A, and it does not swell with
ethylene glycol or c o l l a p s e ~ h e n heated to 300~ . . . it was neither des-
troyed nor enhanced by heat treatment at 659~ and was not destroyed
by 3N hydrochloric acid treatment . . . . Possibly the clay mineral here
reported to be chlorite is a highly aluminous variety."
Aluminous gels have been referred to repeatedly in the origin of high-
alumina deposits in contiguous U.S., Hawaii, Hungary and the U.S.S.R.
Beneslavsky (1962) emphasized strongly the role of Al-gel, or "complex A1,
Si, Ti, Fe-gel", believing that gels go to corundum by complete dehydra-
tion, or to boehmite by ageing. Boehmite, in turn, may be aged further to
diaspore or hydrated to hydrargillite. It appears more logical, however,
t h a t hydrargillite (gibbsite) should precede boehmite and diaspore in the
 THE ORIGIN OF HIGH-ALUMINA CLAY MINERALSuA REVIEW 145
sequences of ageing and alteration. Documentation is scanty, but it would
appear that Al-gels are compatible with an acid environment.

D I S S O L U T I O N , P R E C I P I T A T I O N , S O R P T I O N , AND
F I X A T I O N OF A13+ AND A L U M I N A

That A1 is dissolved readily at pH 11 or higher, and below pH 4 is

certain. Less clear, however, is how alumina can be carried into geologic
situations at presumably intermediate pH Values, such as vein and other
fillings for which there is indisputable evidence, as shown by Allen (1954,
1952), B~rdossy (1959), and Beneslavsky (1962). Gordon et al. writes of
the migration of A13+, Keller described a diaspore clay cast of fossil wood
in a Missouri diaspore pit (1938)--necessitating movement of diaspore.
Perhaps an assumption that the pH at which alumina is dissolved and
carried in intermediate in range is in error.
Repeatedly observed evidence, coupled with the known migrations of
alumina in soils and the dissolved alumina found in pulverized silicate
mineral slurries (Keller et al., 1963), indicates that alumina is notably more
soluble and mobile than its behavior classically interpreted would indicate.
Perhaps its "absence" (or extremely low dilution) in most geologic waters
is more apparent than real, and due not to inherent low solubility, but to
its property of acting as a bonding agent in A1 silicates (Jackson, 1963), i.e.,
being picked up, immobilized, and/or fixed by almost omnipresent clay
minerals, so that it is removed from mobile solution. Because it is trivalent
and small in size, the A13+ ion has a high energy of ion exchange. There-
fore, it may be sorbed during cation exchange and held tightly by any of
the clay minerals, and may be fixed by the 3-layer types to develop
vermiculitic minerals, corrensite, or chlorite. This sort of behavior in soil
clays has been called "anti-gibbsitization" by Jackson (1962-63).

C O N C E N T R A T I O N OF A L U M I N A BY
ORGANIC PROCESSES

Plants and their residues may operate to concentrate alumina in several

ways. Lovering (1959) in a thought-stimulating paper points out that
"Some species of Equisetum (horsetail), sedge, reeds, and bamboo contain
about 10 per cent SiO 2 (dry weight) and grasses are notorious high silica
forage p l a n t s . . , the leaves of corn contain at 2.5 per cent SiO~. . . . Many
tropical lumber trees and also some of the temperate-zone hardwoods are
silica accumulators." By extracting silica from the soil and/or rocks,
alumina may simultaneously be concentrated therein, assuming, of course,
that the silica is blown or washed away when the plant residue falls or
decomposes. That plant opal, opalized residues of squat, pear-shaped,
 146 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
rod-like, and irregular forms up to 0.3 mm in length obtained from grasses
and other plants is transported to marine sediments was observed by
Kanno and Arimura (1958). Beavers and Stephen (1958) found up to 5.5
per cent plant opal in certain Illinois soil, and regard it as a potential
index mineral for paleosols.
Lovering (1959, p.781) pointedly stated, "A forest of silica-accumulator
plants averaging 2.5 per cent silica and 16 tons dry weight new growth
each year would extract about 2000 tons of silica per acre in 5000 years,
equivalent to tile silica in 1 acre-foot of basalt. Comparison of lateritic soil
with parent rock indicates that silic-accumulator jungle could convert
basalt into lateritic soil rapidly--geologically speaking."
Alumina-accumulator plants which concentrate alumina removed from the
soil may require acid soil; Lovering cites two species of aluminum accumu-
lators in which the cell sap acidity is pH 3.6 to 5.2. One species of plant,
the silky oak Orites excelsa of Australia (not a true oak), actually secretes
aluminum succinate in cavities in the wood.
By increasing the solubility of silica in water, plant residues may en-
hance its removal and thereby concentrate the remaining alumina within
an aluminum silicate being desilicated. Lovering (1959, p.789) cited an
example, "extract from the Equisetum contained about 320 ppm silica--
more than twice the amount in a saturated solution of amorphous silica
in water". Indications are that geological chemists have hardly begun to
recognize the reactions that can be carried on by plants and their residues.
Complexing effects, and surface energies generated at interfaces by micro-
flora, m a y yield products considerably beyond those of simple, obvious,
inorganic reactions.
Summarizing, plants and plant residues exert diverse effects on alumina
in geologic systems as follows:
1. Alumina which has been put into solution at pH 10 or above by
other means may be acidified by plant acids and thereby precipitated.
2. Acids at pH 4 or less, derived from plants or their residues, may bring
alumina into solution.
3. Organic complexes with silica increase the solubility and removal
of silica, thereby concentrating less-soluble associated alumina.
4. Alumina-accumulating plants concentrate alumina in their tissues.
5. Humic acid solutions may mobilize alumina and iron (colloids?).

COMMON D E N O M I N A T O R ( S ) I N T H E O R I G I N OF
HIGH-ALUMINA MINERALS

The stated objective of this paper was to search for a common denomina-
tor(s) of, or factor(s) common to, the processes of origin of high-alumina
minerals in clay deposits. Those devoid of any qualification are few indeed;
 THE ORIGIN OF HIGH-ALuMINACLAY MINERALS---A REVIEW 147
it is better to add the qualifications. Some of the denominators found most
common follow.
1. Aluminum must occur in the parent material.
2. Relatively fresh water in (a) great abundance, and (b) continuous
supply must be available if the high-alumina argillation occurs by direct
desilication of an aluminum silicate parent. An abundant and continuous
supply of fresh water is necessary that the ambient ground water solutions
remain so dilute as to hold A1 and Si ions and compounds in solution. A
rain-forest climate meets these conditions. The continuous supply of fresh
water accommodates the required efficient removal of the solute (com-
monly a hydrolysate). The removal may be either downward through
permeable rock, or by flushing off the top of a basin lined with impervious
clay on its bottom.
3. Higher temperature accelerates chemical reaction, i.e., desilication
and/or resilication, and tropical warmth acts accordingly. Climatic tem-
perature of 77~ or above promotes destruction by the microflora of
residues from the macroflora. Under certain conditions, the presence of
plant humic residues causes precipitation of silica and mobilization of
alumina, whereas destruction of plant residues would permit silica mobihty
and consequently, in situ concentration of alumina.
4. Desilication may be enhanced by:
(a) pH above 4 and below 10 during dissolution
(b) Na and K ions in solution,
(c) growth of silica-accumulator plants in the parent material,
(d) action of certain organic complexing with silica to increase silica
solubility. Simultaneously it must be recognized that certain
other humic compounds cause silica to be precipitated, and alu-
minum and iron became mobilized.
5. Alumina may be precipitated-concentrated by
(a) slight acidification (COs dissolved, plant or mineral acids) of
Al-containing solutions originally at pH 10 or above,
(b) evaporation from Al-solutions,
(c) sorption, combination, cation-exchange, or fixation by clay
minerals (or other minerals), silica gels, or silica in solution,
(d) syneresis of Al-gel,
(e) recrystallization of microcrystalline A1 minerals,
(f) alumina-accumulator plants.

U N S O L V E D P R O B L E M S IN T H E O R I G I N OF
HIGH-ALUMINA MINERALS

Unsolved problems of the origin of high-alumina minerals include the

following:
148 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
What geochemical and geologic conditions control and differentiate the
formation of gibbsite vs. boehmite vs. diaspore in bauxite deposits? The
question of what is the environment of diaspore formation brings into
sharper focus the problem of all three.
In the Missouri deposits absolutely no gibbsite has been found, and
boehmite is a minor occurrence; why? Bridge's (1952) suggestion that long
geologic ageing (back to the Paleozoic) is vital for abundant diaspore forma-
tion is invalidated by the occurrence of diaspore in younger deposits.
Keller et al. (1954) suggested independently that gibbsite forms in the
zone of leaching (above ground water table) but diaspore subaqueously,
and found later that de Lapparent (1936, p.378), long before, had concluded
that gibbsite formation took place above the water table, boehmite within
the water table. If these concepts are true, the occurrence of all three
hydrates in deposits such as those in Hungary pose serious problems of
water-table movement. Eyles et al. (1952, p.79) noted that, "since diaspore
is a denser mineral (3.33) than boehmite (3.06) pressure is the essential
factor controlling its development on a large scale, and that rise in tem-
perature alone would not be effective" and (p.80), " t h a t the formation
of monohydrates in large amounts results from a change of geological
environment superimposed on gibbstitic laterites after the formation of
the l a t t e r . . . " (italics by Eyles et al.).
The very intimate association of highly varied ratios of the three
minerals, however, throws doubt on high pressure as a necessary factor to
form bauxite. Is it possible that each of the conclusions is correct in part,
as follows?:
Diaspore requires a higher energy level for its formation than does
gibbsite or boehmite, which can be suppled by either:
1. inheritance of structural energy as an octahedrally coordinated sheet
from kaolinite,
2. deep burial or tectonic movement,
3. high temperature from deep burial, intrusion, or frictional movement.
Under what specific geologic conditions is alumina sufficiently soluble
that it moves through joints and then is deposited as vein fillings?
What conditions promote specifically and critically the sequence from
feldspar to alloysite to gibbsite, as Bates observed, and conversely, feldspar
to gibbsite, to kaolinite, as Gordon et al. observed? Not only is the sequence
reversed, but halloysite is in one sequence and kaolinite in the other.
Is a tropical or subtropical temperature necessary for bauxitization to
operate? Is the minimum temperature (whatever it is) determined by
reaction rate of dissolution; or by the balance between accumulation of
organic matter from a macroflora off-set destruction of it by a microflora?
How extensive is alumina-gel present in weathering environments, in
stream transportation, and in high-alumina mineral diagenesis, and is it
necessary for the develoment of pisolitic structure in bauxite?
 THE ORIGIN OF HIGH-ALUMINA CLAY MINERALs--A REVIEW 149

W h a t are t h e specific organic c o m p l e x e s t h a t increase t h e solubility of

silica in w a t e r a n d a l t e r n a t i v e l y , w h a t mobilizing c o m p o u n d s of A1 a n d Fe?
A t a t i m e w h e n space, isotope, a n d n u c l e a r geology a p p e a r v e r y g l a m o r -
ous, t h e g e o c h e m i s t r y of th e t h i r d m o s t a b u n d a n t e l e m e n t on t h e e a r t h ' s
c r u s t q u i e t l y still i n v i t e s research.

REFERENCES
Allen, V. T. (1935) Mineral composition and origin of Missouri flint and diaspore clays:
Appendix IV, 58th An. Rpt., Mo. Geol. Surv. and Water Resources, pp.5-24.
Allen, V. T. (1945) Effect of migration of clay minerals and hydrous aluminum oxides
on the complexity of clay, Jour. Amer. Ceram. Sot., v.28, pp.268-275.
Allen, V. T. (1946) Sedimentary and volcanic processes in the formation of high alumina
clay: Ec. Geol., v.41, pp.124-137.
Alien, V. T. (1952) Petrographic relations in some typical bauxite and diaspore deposits:
Bull. Geol. Soc. Amer., v.63, pp.649~88.
Allen, V. T. (1955) Relation of porosity and permeability to the origin of diaspore clay:
Clays and Clay Mins., Nat'l. Acad. Sci.-Nat'l. Res. Council Pub. 395, pp.389~I01.
B~rdossy, Gy (1959) The geochemistry of Hungarian bauxites: A cta Geologica (Hungarian
Academy, Budapest 62), parts I and II, Tomus V, pp.103-155, 255-285, parts III
and IV, Tomus VI, pp.l-53.
Bates, T. F. (1962) Halloysite and gibbsite formation in Hawaii, Clays and Clay Min.
9th Conf. Pergamon Press, pp.315-327.
Beavers, A. H., and Stephen, I. (1958) Some features of the distribution of plant-opal
in Illinois soils: Soil Sci., v.68, pp.l-5.
Beneslasvky, S. J. (1962) Chemical and mineralogical composition of bauxites and some
problems concerning the genesis of their minerals: Acta Geologica (Hungary),
Tomus VI, Fascieuli I, pp.54-64. (In English).
Bolger, R. C., and Wertz, J. H. (1952) Mineralogy and origin of the Mercer fire clay
of north-central Pennsylvania: Prob. of Clay and Laterite Genesis, Syrup. of AIME
at St. Louis, Mo., New York, pp.81-93.
Bratton, R. J., and Brindley, G. W. (1962) Structure-controlled reactions in kaolinite-
diaspore-boehmite clays: Jour. Am. Ceram. Soc., v.45, pp.513-516.
Bridge, Josiah (1952) Discussion, in Problems of Clay and Laterite Genesis: Symposium
AIME, pp.212-214.
Burst, ]. F. (1952) New clay mineral evidence concerning the diagenesis of some
Missouri fire clays: in Prob. Clay and Laterite Genesis, AIME Symposium
pp.135-143.
Correns, D. W., 1949, Einfuhrung in die Mineralogie, Springer-Verlag, Berlin.
De Kimpe, C., Gastuche, M. C., and Brindley, G. W. (1961) Ionic coordination in
alumino-silicic gels in relation to clay mineral formation: Am, Min. v.46,
pp. 1370-1381.
Ervin, G., Jr. and Osborn, F. F. (1951) The system Al,Os-H20: Jour. Geol., v.59,
pp.381-394.
Eyles, V. A., et al. (1952) The composition and origin of the antrim laterites and bauxites:
Mere. Geol. Sum. No. Ireland, Belfast, 90 pp. With mineral determinations by
X-ray by Bannister, F. A., Brindley, G. W., and Goodyear, J.
Foster, M. D. (1962) Interpretation of the composition and a classification of the
chlorites: U.S. Geol. Surv. Prof. Pap. 414-A.
Fox, C. S. (1932) Bauxite and Aluminous Laterite: Crosby Lockwood & Son, Pub.,
London, 212 pp.
Gedeon, T. G. (1956) Bayerite in Hungarian bauxite: Acta Geologica (Budapest,
Hungary), Tom. IV, Fasc. I, pp.95-105.
150 TWELFTH NATIONAL CONFERENCE ON CLAYS AND CLAY MINERALS
Gordon, M., Jr., and Tracey, J. I., Jr. (1952) Origin of the Arkansas bauxite deposits:
Prob. of Clay and Laterite Genesis, Symp. AIME at St. Louis, New York, pp.
12-34.
Gordon, M., Jr., Tracey, J. I., Jr., and Ellis, M. W. (1958) Geology of the Arkansas
bauxite region, U.S. Geol. Surv. Prof. Pap. 299.
Hahn, G. W. (1954) Clay mineral relationships in the Ocheskey fire clay pit: unpublished
M.A. thesis, Univ. of Mo. Library, Columbia, Mo., 90 pp.
Harder, E. C. (1952) Examples of bauxite deposits illustrating variations in origin:
Prob. of Clay and Laterite Genesis, Symp. AIME at St. Louis, Mo., New York,
pp.35-64.
Harrison, J. B. (1933) The katamorphism of igneous rocks under humid tropical condi-
tions: Imperial Bureau of Soil Science, Rothamsted Experimental Station, Harpen-
den, England.
Jackson, M. L. (1962-63) Interlayering of expansible layer silicates by chemical weather-
ing: Clays and Clay Min., 11th Conf. Pergamon Press, New York, pp.29-46.
Jackson, M. L. (1963) Aluminum bonding in soils: a unifying principle in soil science:
Soil Sci. Soc. Amer. Proceed., v.27, pp.l-10.
Kanno, I., and Arimura, S. (1958) Plant opal in Japanese soils: Soil and Plant Food,
v.4, pp.62-67.
Keller, W. D. (1938) Diaspore clay cast of fossil wood in a Missouri diaspore pit: Amer.
Min., v.23, pp.461-463.
Keller, W. D. (1952) Observations on the origin of Missouri high-alumina clays:
Problems of Clay and Laterite Genesis, Symposium AIME, pp.llS-134.
Keller, W. D. (1958) Argillation and direct bauxitization in terms of concentrations of
hydrogen and metal cations at surfaces of hydrolyzing silicates: Bull. Amer. Assn.
Petrol. Geol., v.42, pp.233-245.
Keller, W. D. (1960) Acidity-alkalinity of surface drainage waters as related to under-
lying silicate rocks: Jour. Sed. Pet., v.30, pp.582-584.
Keller, W. D., Balgord, Wm. D., and Reesman, A. L. (1962) Dissolved products of
artificially pulverized silicate minerals and rocks, Part I: ]our. Sed. Petrol., v.33,
pp. 191-204.
Keller, W. D., McGrain, P., and Saum, N. M. (1964) Observations on the origin of
endellite in Kentucky and their extensions to "Indianaite": by title, this Conference.
Keller, W. D., Westcott, J. F., and Bledsoe, A. O. (1954) The origin of Missouri fire
clays: Clays and Clay Min., Natl. Acad. Sci.-Natl. Res. Coun. Pub. 327, pp.7-46.
Kennedy, G. C. (1959) Phase relations in the system Al~O3-H20 at high temperatures
and pressures: Am. Jour. Sci., v.257, pp.563-573.
Krauskopf, Konrad (1959) The geochemistry of silica in sedimentary environments, in
Silica in Sediments: (Soc. Econ. Paleo. and Mineral.), Spec. Pub. No. 7, pp.4-20.
de Lapparent, J. (1936) Raisons geologiques de la formation des trois hydroxides
d'aluminum naturels: Cong. Int. des Mines, de la Metallargie et de la Geol. applique,
VIIe Session. Section de Geol. applique, Tom. I, pp.375-379.
Laubengayer, A. W., and Weisz, R. S. (1943) A hydrothermal study of equilibria in the
system alumina-water: Am. Chem. Soc. Jour., v.65, pp.247-250.
Lesar, A. R., Krinbill, C. A., Jr., Keller, W. D. and Bradley, R. S. (1946) Effect of
compounds of sulfur on reheat volume change of fire clay and high-alumina
refractories: ]our. Amer. Ceram. Soe., v.29, pp.70-75.
Lovering, T. S. (1959) Significance of accumulator plants in rock weathering: Bull. Geol.
Soc. Amer., v.70, pp.-781-800.
Mackenzie, R. C. (1959) The classification and nomenclature of clay minerals, Clay
Min. Bull., v.4, pp.52-66.
McQueen, H. S. (1943) Fire clay districts of east central Missouri: Mo. Geol. Surv, and
Water Resour., 2nd Series, v.23.
Mark, H. R. (1963) High-alumina clay in the United States: U.S. Geol. Surv. M i n .
Invest. Resour., Map MR-37.
 THE ORIGIN OF HIGH-ALUMINA CLAY M I N E R A L s - - A REVIEW 151
Mead, W. J. (1915) Occurrence and origin of the bauxite deposits of Arkansas. Ec.
Geol., v.10, pp.28-54.
Mohr, E. C. 5-, and Van Baren (1954) Tropical Soils: Interscience, New York.
Okamoto, G., Okura, T., and Goto, K. (1957) Properties of silica in water: Geochim. el
Cosmochim. Acta, v.12, pp.123-132.
Patterson, S. H., and Hosterman, J. W. (1962) Geology and refractory clay deposits of
the Haldeman and Wrigley quadrangles, Kentucky: U.S. Geol. Surv. Bull. 1122-F,
113 pp.
Robbins, Carl, and Keller, W. D. (1952) Clay and other non-carbonate minerals in some
limestones: Jour. Set. Pet., v.22, pp.146-152.
Ross, C. S., and Kerr, P. F. (1934) Halloysite and allophane: U.S. Geol. Surv. Prof.
Paper 185-G, 13 pp.
Sasvari, K. (1956) The crystal structure of ~-bayerite, AI(OH)8: Acts Geologica
(Budapest, Hungary), Tom. IV, Fasc. I, pp.123-129.
Seafight, W. V. (1955) Guide Book to Field Trip Sec. Ann. Mtg. Assn. Mo. Geol., Mo.,
Geol. Sur. and Water Res. Rpt. Invest. No. 20, p.33.
Slaughter, M., and Milne, I. H. (1960} The formation of chlorite-like structures from
montmorillonite: Clays and Clay Mins., 7th Conf. Pergamon Press, pp.114-124.
Stevens, R. E,, and Carton, M. K. (1948) Simple field test for distinguishing minerals
by abrasion pH, Amer. Mineral, v.33, pp.31-50.
Terentieva, K. F. (1949) Corundum in bauxite: Trudy V.I.M.S. Nov. Ser. 2. In
Beneslavsky, 1962.
Turner, R. C., and Brydon, J. E. (1962) Formation of double hydroxides and the
titration of clays: Science, v.136, No. 3521, pp.1052-1054,
Williams, E. G. (1960) Relationship between the stratigraphy and petrography of
Pottsville sandstones and the occurrence of high-~lumina Mercer clay: Ec. Geol.,
v.55, p.1291-1302.
Wilson, H., and Treasher, R. C. (1938) Preliminary report of some of the refractory
clays of western Oregon: Ore. Dept. Geol. and Min. Ind. Bull. 6, 96 pp.