Tải bản đầy đủ - 0 (trang)


Tải bản đầy đủ - 0trang



stituents. Mitchell et al., reviewed developments to 1964 in studies of amorphous inorganic material in the clay fraction and indicated the lines of

future studies.

Scientists are increasingly aware of the importance of amorphous phases

in soils and sediments. Amorphous clay materials are often described

by the term active because of their large specific surface area and high

chemical reactivity. When present in a substantial amount, they have a

very marked effect on both physical and chemical properties of soils. This

is particularly true for many soils derived from volcanic ashes, but their importance has also been suggested for other soils even though present in

a relatively small amount.

The objective of this article is to review recent progress in studies of

the amorphous clay constituents of soils and to outline our present knowledge on their nature, properties, and genesis. Effort was made to collect

analytical data on natural clay systems, although they often have been inconclusive, and to indicate similarities and differences between different

kinds of amorphous clay materials. Many of the data obtained for synthetic

amorphous hydrous oxides and aluminosilicates have not been included,

and their use as models is imperative in future studies. We will try to avoid

repetition of material presented in the excellent review by Mitchell et al.

( 1964). Rather, emphasis will be placed on the advances in research during the last ten years.


Definition and Scope

Brindley (1969) defined the amorphous state as one which lacks order

and is distinguished from the crystalline state which has order. Crystalline

is a relative term. By convention, X-ray diffraction has been used as the

criterion for crystallinity. Materials which are sufficiently well organized

to yield X-ray diffraction patterns are said to be crystalline whereas those

which do not are called amorphous. However, some materials are “amorphous” to X-rays but give diffraction phenomena when examined with an

electron beam. Consequently, Brindley ( 1969) recommended a more satisfactory distinction between them as matter having long-range order (crystalline state) and short-range order (noncrystalline state). Long-range

order has a unit cell repetition, generally in three dimensions, but sometimes in two, and possibly only in one, dimension. Some short-range, nonrepetitive order will normally be present in the so-called amorphous substances. Liquids and amorphous solids such as glass possess a degree of

short-range order due to coordination tendencies of atoms associated with

systems of electrostatic ionic packings and/or directed covalence bonds.


21 3

This article mainly deals with the clay materials which are, and have

been thought to be, amorphous. The term amorphous clay materials is retained as a conventional and inclusive term. This includes materials having

only short-range order and those having long-range order in only one dimension. For perspective and continuity of discussion of relationships to

properties and genesis, it is desirable to include poorly crystallized materials such as imogolite and certain forms of halloysite.

Chemically, the principal forms of the amorphous clay materials in soils

are oxides and hydroxides of iron, aluminum, and silicon, and the silicates

of aluminum and iron, all in various combinations with water. These materials commonly occur as particles smaller than 2 pm equivalent diameter

as weathering products of primary minerals, and it is these which are to

be considered here. Other amorphous components such as unweathered

volcanic glass and biogenic opal (phytoliths) are not included in the present discussion. Both of these are X-ray amorphous, but are present mostly

as particles larger than 2 pm in diameter. Most soil organic constituents

are also present in an amorphous state. They are extremely important in

physical and chemical reactions and properties of soils. However, limitations of space and our expertise preclude them from inclusion here. The

interaction between the amorphous inorganic and organic constituents will

be discussed briefly.

Where appropriate, comments will be given on the nomenclature of each

amorphous clay material. The AIPEA Nomenclature Committee has recommended that specific names be not given to poorly defined clay materials, such as irregular interstratified systems or imperfect structures, or

to amorphous constituents (Brindley and Pedro, 1972).



Nature of Materials



Not all the materials called opal in the past are X-ray amorphous. Jones

and Segnit (1971 ) classified opal into three well-defined structural groups:

Opal C (well-ordered a-cristobalite) , Opal Ct (disordered a-cristobalite

or a-tridymite) and Opal A (highly disordered, nearly amorphous). Most

precious opals give X-ray powder patterns with only a broad diffuse band

around 4.1 A. Electron microscopy shows that the play of colors, or “fire,”

in them is caused by diffraction effects from arrays of uniformly sized silica

spheres of 1700-3500 A in diameter (Jones et al., 1964). These spheres

are secondary aggregates of primary silica particles up to 400 A across,

often arranged in concentric shells. The perfection of shape and arrange-



ment of the secondary aggregates is greatest in precious opal and least

in varieties of natural opaline silica such as siliceous duricrust, wood opal,

and silica deposited in plants. In the latter, the secondary aggregates are

less than 1500 A in diameter and irregular in shape (Darragh et al., 1966).

Laminar opaline silica particles were found in the 0.2-20 pm fractions,

and most abundantly in the 0.4-2 pm fractions, of volcanic ash soils by

Shoji and Masui (1969a,b, 1971). They distinguished four morphological

types; circular, elliptical, rectangular, and rhombic. The thickness of the

particles varies from one-twentieth to one-fifteenth of their diameter and

they show fine-grained uneven surfaces. Weathered and alkali-treated

opaline silica particles appear to be very porous suggesting that they are

actually composed of extremely fine silica spheres. No data for chemical

composition of the isolated opaline silica particles are available. They are

soluble in hot 0.5 N NaOH. The Si0,/A1,08 ratios of the 0.5 N NaOH

soluble fraction of soil clays which contain many of the particles were in

the range from 2.4 to 21.7. Refractive indices and specific gravities of

opaline silica particles are in the range from 1.42 to 1.43 and from 2.1

to 2.3, respectively.

Shoji and Masui ( 1969a) indicated that opaline silica has two absorption

maxima on the infrared spectra at 1100 and 800 cm-l which are assigned

to the Si-0 vibrations. The difference spectrum of a 0.5 N NaOH soluble

fraction of a volcanic ash soil clay in which opaline silica is present in

a fair amount, showed absorption maxima at 1200 (shoulder), 1075, 935

(shoulder) and 790 cm-* (Tokashiki and Wada, 1972b). These frequencies are the same as those found for amorphous silica gels synthesized in

the laboratory (Mitchell et al., 1964; Leonard et al., 1964).





The name “limonite” has often been used for amorphous, hydrated iron

oxides with composition Fe,O,.nH,O (Brown, 1955) or with a molecular

formula such as 2Fe20,.3H,0. Rooksby (1961) points out that most of

the limonite with the latter composition are very finely divided goethite and

they hold moisture over that for a monohydrate, probably by adsorption


Very little information about the composition, structure, and morphology of discrete amorphous oxides and hydrous oxides of aluminum and

iron present in soils is available. Many of these materials exist as coatings

of gel-hull5 on other particles (Jones and Uehara, 1973) rather than discrete units. Under such conditions it is difficult to separate the amorphous

and crystalline components without changing the nature and properties.

De Villiers (1969) reported the occurrence of an amorphous alumina of

a boehmite character in certain tropical soils. The evidence provided is



that the amounts of gibbsite in these soil clays, as measured by differential

thermal analysis, were lower than those found using the selective dissolution procedure of Hashimoto and Jackson (1960) and that the weight

losses between 110” and 330°C fell short of those expected on the basis

of dissolution treatments. However, there is an implied but unwarranted

assumption in this that no crystalline constituents other than gibbsite are

dissolved in the hot 0.5 N NaOH.

The occurrence of ferric hydroxide gel in Akaka soils in Hawaii was

reported by Matsusaka and Sherman (1961). The magnetic attraction of

the soil increased from 2.2 mg per gram of soil when naturally moist, to

3.7 mg per gram of soil when air-dried, and to 9.2 mg per gram of soil

when heated to 600°C. On the basis of X-ray and differential thermal analyses, they suggested an alteration sequence of amorphous ferric hydroxide -+ cryptocrystalline lepidocrocite + maghemite system. Sherman et aE.

(1964) also observed gibbsite in the dried Akaka soils, but not in the

undried soils. Microscopic examination indicated the occurrence of gels

in channels and pores in the soils. They then suggested segregation and

crystallization of gibbsite from the amorphous aluminum-iron gels. No

mention was made, however, of the effect of moisture in detection of crystallinity by X-ray analysis in these highly hydrated materials.

The most common forms of amorphous aluminum hydroxy materials in

soils are probably those which exist as aluminum-hydroxy interlayer “islands” in expansible layer silicates. Their nature and occurrence has been

reviewed by Rich (1968). The aluminum-hydroxy interlayers appear in

a gibbsite-like monolayer structure, and disperse randomly as “islands”

in the interlayer space. The hydroxy/aluminum ratio of the interlayer material present in soils is a question that is largely unresolved, but synthetic

studies (Turner, 1965; Hsu, 1968) have shown that the hydroxy-aluminum

species with the OH/AI ratio of 2.5 to 2.7 are favorably retained by

montmorillonites. The extent of interlayering is small compared with the

amount required for formation of a complete gibbsite-like monolayer,

which is about 16 meq of aluminum per gram of montmorillonite. Brydon

and Kodama (1966) reported that amounts up to 8 meq of aluminum

per gram enters entirely into the interlayer space and that beyond this some

hydroxide is present external to the montmorillonite interlayer region.

That there are two kinds of OH-groups in the hydroxy-aluminum interlayers was indicated by Brydon and Kodama (1966) and Weismiller et al.

(1967) using infrared spectroscopy. The latter authors assigned the 3695

cm-I pleochroic band to the inner hydroxy groups of a gibbsite-like ring

[Al(OH),(H,O) ,Is6+ and the 3570 cm-1 nonpleohroic band, to the outer

OH-groups. A deuteration study by Ahrlichs (1968) showed that flushing

with D,O at room temperature removed water, and heating at 100°C in

D,O vapor exchanges interlayer OH groups but not the clay lattice OH



groups. These observations were made on interlayer materials prepared in

the laboratory and the results of similar studies on natural counterparts are


Hydroxy iron interlayers exist in the phyllosilicates although they apparently are less common than hydroxy aluminum. In 1940, Bower and Truog

reported that positively charged ferric dihydroxy ions were held by exchange positions; under certain conditions the clays adsorbed three times

the exchange equivalent of ferric iron. Synthetic hydroxy iron interlayers

in clay result in properties similar to those with hydroxy aluminum (Clark,

1964; Coleman and Thomas, 1964). Consequently, routine identification

procedures based on properties of systems would not always result in recognition of the presence of iron hydroxy interlayers. The little evidence for

naturally occurring interlayers composed largely of hydroxy-iron groups

led Rich (1968) to suggest that occlusions of Fe(OH), in positively

charged hydroxy-aluminum interlayers appear to be more likely. This

indicates that iron oxides and hydrous oxides are more stable than hydroxyFe3+interlayers. Singleton and Harward ( 1971 ) have presented evidence

for iron-hydroxy interlayers in two soil clays of Western Oregon.

The occurrence of iron-hydroxy interlayers probably relate to moisture

status, pH, and other factors which affect oxidation-reduction. Jackson

(1962) suggested that iron might be involved in interlayering of clays since

wetness and gleying seemed to clean up intergrade minerals. I t was implied

that under anaerobic conditions the ferric dihydroxy ion would be reduced

to ferrous, thus losing its capacity to be retained. Lynn and Whittig (1966)

obtained evidence which suggested chlorite formations may be correlated

with high ferrous ion concentrations in undrained tideland sediments.

Polymeric hydroxy-aluminum may also be present in association with

allophane. The negative charge of allophane is presumed to arise from substitution of aluminum for silicon in a tetrahedral silica framework. In order

to account for its characteristic pH-dependent negative charge and phosphate adsorption, blocking of exchange sites by hydroxy-aluminum was inferred by de Villiers and Jackson (1967), Cloos et al. (1968; 1969), and

de Villiers (1971 ). No specific information about the composition and structure of polymeric hydroxy-aluminum which may be present in allophane

has, however, been available. An aluminosilica gel prepared by addition of

monomeric silica to hydroxy-aluminum solution gave absorption on the

infrared spectra in the region from 860 to 930 cm-l and an endotherm with

maximum extending up to 200°C on the differentia1 thermal analysis curve

(Wada and Kubo, 1972, 1973). These features have not been seen for

either allophane or imogolite separated from soils.

Available evidence indicates that hydrous oxides of aluminum and iron

as well as amorphous silicates are the most important materials involved



in the interaction between clays and organic matter in soils (Greenland,

1971). The presence of aluminum and iron hydroxides in combination

with organic matter has been indicated by extracting soils with reagents

such as sodium pyrophosphate, Tamm’s acid oxalate, and citrate-dithtonite

solution, which will be described in Section IV. Again, there is very little

information about the composition and structure of these hydroxy compounds bound with organic matter.





The term allophane has been used with different meanings by different

investigators. Ross and Kerr described allophane as an amorphous member

of the kaolin group (Keller, 1964). This has a specific meaning and implies

some chemical combination involving aluminum and silicon. Apparently

Ross and Kerr also suggested that the term be used for all amorphous

clay materials regardless of their composition (Grim, 1953). Fieldes

(1966) and Furkert and Fieldes (1968) also proposed the use of the name

allophane as any clay size material characterized by structural randomness.

This concept likely results in discouraging recognition of important differences in chemical structure and surface properties of various amorphous

clay constituents in soils.

The problem was reviewed at the United States-Japan seminar on amorphous clay materials in 1969. A tentative definition adopted by this group

will be used in this article. This defines allophanes as members of a series

of naturally occurring hydrous aluminosilicates of widely varying chemical

composition and which are characterized by short-range order, by the

presence of Si-0-A1 bonds, and by a differential thermal analysis curve displaying a low-temperature endotherm and a high-temperature exotherm

with no intermediate endotherm (van Olphen, 1971) . Alumina and silica

would not be the end members of the series since allophanes are aluminosilicates containing AI-0-Si bonds.

The name hisingerite has been used in two ways. Brown (1955) defined

it as an iron analog of allophane with composition of Fe,O,-ZSiO,.nH,O.

However, no data have been available on hisingerite in this sense. On the

other hand, Gruner( 1935) described hisingerite as a mineral species which

gives broad diffraction lines which coincide with those of nontronite, but

which may be an “amorphous” variant. As pointed out by MacEwan

( 1961) this expression “amorphous” should perhaps be interpreted as

“very finely crystalline.”

Whelan and Goldich (1961) and Lindqvist and Jannson (1962) evaluated selected hisingerite samples and found three broad diffraction lines

at 4.3 to 4.6, 2.57 to 2.63, and 1.54 to 1.60 A. The molecular ratios


21 8



FeO) varied from 1.4 to 3.2 and the ratios


FeO) varied from 1.0 to 2.1. The infrared spectra

exhibited a broad, featureless Si-0 absorption band with a maximum at

1000 cm-I. The diffraction data suggested development of a layer lattice

but restricted crystal growth in the c axis. Whelan and Goldich (1961)

considered the material as principally poorly crystallized iron saponite

whereas Lindqvist and Jansson (1962) considered it to be mica mineral

with extensive substitution of Fe for Si and with interlayer hydronium ions.

Both investigators, however, stressed that hisingerite definitely warrants

further study.

Some of the more recent advances in characterization of amorphous

and poorly crystallized materials involve imogolite. Imogolite is a hydrous

aluminum silicate having a fine threadlike nature and unique diffraction

characteristics; it was first described by Yoshinaga and Aomine (1962b).

The name imogolite as a new mineral species has been approved by the

AIPEA Nomenclature Committee (Brindley and Pedro, 1970). Imogolite

exhibits long-range order in one dimension and is included herein because

of its frequent association with allophane and the similarities in chemical

and physical properties.


1 . Chemical Composition

Data from chemical analysis of allophane and imogolite derived from

weathering of volcanic ash and pumice were provided by Yoshinaga and

Aomine (1962a,b), Yoshinaga (1966, 1968), and Miyauchi and Aomine

(1966b). For these analyses, the <0.2 pm fractions were collected and

pretreated with dithionite-citrate (Mehra and Jackson, 1960) and

2%Na,C03 solution (Jackson, 1956). From these data, Wada and Yoshinaga (1969) showed that the SiO,/AI,O, ratio of the clays in which allophane predominates is in a range from 1.3 to 2.0, while that of the clay

in which imogolite predominates is in a fairly narrow range from 1.05 to

1.15. The H,O(+)/ALO, ratio is mostly in a range from 2.5 to 3.0 without any definite difference between the two groups of clays. Higher

SiO,/Al,O, ratios for imogolite, 1.5 (Russell et al., 1969) and 1.29 to 1.32

without any chemical pretreatment (Tazaki, 1971) , as well as lower values

near 1 for allophane also have been reported (Russell et al., 1969; Lai

and Swindale, 1969). Iimura (1969) gave Si0,/A1,03 ratios of 0.89 to

1.43 for five allophanic clays which had been dispersed at pH 4 without

any chemical pretreatment.

Egawa (1964) and Udagawa et al. (1969) used X-ray fluorescence

spectroscopy to obtain data on the coordination number of aluminum in

allophanic clay separated from a Kanuma pumice bed. Their samples were

different in Si02/A120, ratios, namely, 2.37 and 1.67, but alike in the



coordination status of aluminum; about 50-60% of the aluminum was estimated to be in 4-fold coordination at room temperature and at 110OC.

The aluminum K , line profile for samples treated with dithionite-citrate

and 2% Na,CO, suggest that a considerable part of the aluminum in allophane (Si0,/AI,03 ratio close to 2.0) is in 4-fold coordination while

nearly all of the aluminum in imogolite is in 6-fold coordination (T. Henmi

and K. Wada, unpublished).

Yoshinaga (1 968) confirmed that a small but significant amount of iron

(0.3-0.9% as Fe,O,) remains in allophane even after repetition of ten

dithionite-citrate treatments. Kitagawa ( 1973) interpreted electron resonance spectra of two allophanic clays as indicating the substitution of aluminum by iron similar to that in mica and montmorillonite.

Gotz and Masson (1970, 1971) developed a chemical procedure for

differentiating silicate anions possessing low degrees of polymerization. The

procedure is based on conversion of the anion to a trimethylsilylether, and

subsequent identification and quantitative determination of the volatile

ether by gas chromatography. Application of this technique to imogolite

gave a high yield of volatile products of which 95% was the orthosilicate

ether and 5 % the pyrosilicate ether. This furnished evidence in favor of

the presence of isolated orthosilicate groups (Cradwick et al., 1972).

2. Morphology

As a result of improvements in instrumentation and techniques of preparation, it has become possible to observe clay particles with a resolution

of about 5 A in the electron microscope. High resolution electron microscopy shows that the imogolite threads which previously appeared to be

of micron length with a diameter of 100-300 A actually consist of a finer

filiform unit with separations in the order of 18-22 A (Yoshinaga et al.,

1968; Russell et al., 1969). More recent studies on the samples dispersed

on a microgrid or cut in thin section (Wada et al., 1970) suggest that

this filiform unit is a tube with the inner and outside diameters of about

10-20 A (Fig, 1 top). This tubular unit is very useful for identification

and detection of imogolite even when present in a very small amount, as

illustrated in Fig. 2. In this particular sample, imogolite has not been identified by other analytical techniques.

Eswaran (1972) studied volcanic ash soils containing imogolite with the

scanning electron microscope and found threads or ribbons forming peculiar globules. The much thicker threads up to 30,000 A in diameter led

him to suggest that pretreatment for transmission electron microscopy

breaks up these threads of imogolite along planes of weakness. However,

neither these thicker threads nor globules have been found in the same

volcanic ash soil by scanning electron microscopy. Planar nets or films



FIG. 1. (Top): Electron micrograph of imogolite formed in saprolite of basalt

obtained by N. Yoshinaga. This sample was studied by Patterson (1964) and Wada

et al. (1972). Scalemarker: 1000 A. (Bottom): Scanning electron micrograph of

imogolite in a glassy volcanic ash soil “Irnogo” (Lab. No. 905) obtained by courtesy

of H. Yotsumoto of Japan Electron Optics Laboratory Company Ltd. Scalemarker:

10 pm.


22 1

consisting of twisting threads with diameter about 1000 A have been observed with unweathered glass shards [Fig. 1 (bottom)]. Similar objects

were previously observed in the carbon replica of the gel film which consists

exclusively of imogolite separated from pumice beds (Wada and Matsubara, 1968; Yoshinaga et al., 1968; Yoshinaga and Yamaguchi, 1970b).

Kitagawa (1971 ) obtained high resolution electron micrographs of several allophanic clays separated from weathered pumice beds in Japan. He

interpreted them as indicating that allophane has also a structural unit,

which is probably a hollow sphere of about 55 A in diameter. The presence

of similar objects described two-dimensionally as ringlets had been observed as an admixture of imogolite in the gel films from weathered pumice

beds (Yoshinaga, 1968; Wada and Yoshinaga, 1969). Figure 2 shows an

electron micrograph of allophane with a Si0,/A120, ratio 2.0 formed in

weathered volcanic ash. It seems to consist of “hollow” spherical particles

as described by Kitagawa (1971) though their outside diameter varies from

30 to 55 A. It is worth noting here that a prolonged exposure of both

allophane and imogolite to the electron beam likely results in loss of much

FIG.2. Electron micrograph of allophane formed in weathered volcanic ash obtained by T. Henmi. This sample (VA) was studied by Aoniine and Wada (1962)

and Wada and Tokashiki (1972). Scalemarker: 1000 A.



FIG.3. Electron micrograph of halloysite formed in weathered volcanic ash obtained by courtesy of H. Yotsumoto and S.Aida of Japan Electron Optics Laboratory Company Ltd. This sample (VH) was studied by Aomine and Wada (1962) and

Wada and Tokashiki (1972).

Tài liệu bạn tìm kiếm đã sẵn sàng tải về


Tải bản đầy đủ ngay(0 tr)