Tải bản đầy đủ - 0trang
CHAPTER 7. PHYSICAL PROPERTIES OF ALLOPHANE SOILS
T. MAEDA ET AL.
could have been used. The reader should interpret “allophane” in this less
The concept of allophane is changing as more becomes known about the
material. Imogolite is now distinguished from allophane, but is included in our
use of the term “allophane soils.” Imogolite, with its thread-shaped particles, has
long-range order in one direction. It has a characteristic form in electron
micrographs and has an identifiable X-ray diffraction pattern. Allophane, as the
term is now used, has only short-range crystalline order, but does have an
identifiable spherule form in electron micrographs. There are “amorphous materials” in soils which do not have this spherule form. It is likely that some of
these materials are of pyroclastic origin.
Possibly the term “andept” or “andosol” would have been a better choice to
describe these soils. However, for the nonspecialist, the term “allophane soil”
will probably best bring to mind the soils which we will be describing here. While
these soils are dominated by allophanic properties, other minerals are often
present in addition to allophane and imogolite. There are soils which contain
small amounts of imogolite and/or allophane which would not be considered
allophane soils here because they do not possess the physical properties associated with allophane.
Allophane soils are widely distributed. They occur frequently in the Caribbean and Andean lands, as well as in the Pacific areas of Indonesia, Japan, New
Zealand, and the United States. More studies on physical properties have been
carried out in Japan than in any other country. One of the purposes of this
review is to make the results of the published Japanese studies more readily
available to readers of the English language. The references cited are mostly in
Japanese, but usually have summaries in English. The more recent articles often
have legends for figures and tables in English. Some of the papers are written in
The justification for a review on physical properties of allophane soils is that
they have distinctive properties which distinguish them from other soils. Soils
can be divided into three groups on the basis of physical properties. In the first
group, void characteristics determine physical properties, and void volume
changes little with changes in water content. These soils, with sands as the
example, can be treated as rigid, porous media. In the second group, the nature
and extent of surfaces determines physical properties. Volume changes accompany water content changes; these changes are reversible even though they show
hysteresis. Physical-chemical descriptions of behavior are often more useful than
mechanical descriptions. Swelling clays are examples of soils in this group. In the
allophane soils, of the third group, void characteristics rather than surface area
determine physical properties. There are volume changes accompanying water
content changes, but the effects are largely irreversible. The matrix changes on
drying, and the dried soil can be considered a different material.
PHYSICAL PROPERTIES OF ALLOPHANE SOILS
Much of this review will, therefore, be in the form of comparing physical
properties of allophane soils with the properties measured for soils with crystalline clay minerals. It is assumed that most readers will be familiar with the latter.
B. GENERAL NATURE OF PHYSICAL PROPERTIES OF ALLOPHANE
The measured values of physical properties of allophane soils, in summary,
show: they have low natural bulk density, high 15-bar water content, and high
natural water content; that medium to low amounts of water are available to
plants; they have high liquid limit and low plasticity index; they are difficult t o
disperse; and that there are irreversible changes in all these properties on drying.
Several good general summary descriptions of physical properties of allophane soils have been published. Swindale (1964) lists the following features:
. . .deep
soil profiles, usually with distinct depositional stratification, and normally
friable in the upper part; topsoils as thick as one metre, and dark brown to black in
color, containing humic compounds which are comparatively resistant to microbial
decomposition; prominent yellowish brown to reddish brown subsoil colors with a
smeary feel when the soil is wet; very light and porous profiles with a low bulk density
and high water-holding capacity; rather weak structural aggregation, with easily destroyed porous peds lacking in cutans, and lack of horizontal differentiation in the
subsoil except for the occurrence of duripans in some soils;. .. Smeary consistencies
are marked only in soils in very humid or per-humid climates. The soils which form in
per-humid climates tend to dry irreversibly when they are allowed to dry out in road
cuts or banks. This feature of irreversible drying is a useful classification criteria,
although the soils in the field never become dry enough to exhibit the property to any
Fieldes and Claridge (1975) have summarized the early studies by Fieldes and
his co-workers on New Zealand allophane soils:
. . . allophane in its early stages of formation could be visualized as gel-like fragments of
random aluminosilicate held together by cross-linking at a relatively small number of
sites. The fragments have an open internal structure, which originally, in the hydrogel
state, enclosed much water. Until the water is removed, rearrangements of materials
with more ordered structure cannot occur, and the moist clay has a weak “waxy”
consistence. When the water is removed by drying, the structure collapses and further
cross-linking takes place so that the process cannot be reversed; and the resultant
material has considerable mechanical strength. Thus, it is not possible to reconstitute the
moist hydrogel structure by rewetting, although the more compact xerogel is still open
enough to have a strong affinity for water..
Allophane soils generally have a friable surface soil and massive structure in the
subsoil, which however has a relatively high permeability. The friable structure
of the surface soil is partly due t o effects of drying. Often allophane soils have
several layers with very different physical properties which affect water movement and water available for plant use.
T. MAEDA ET AL.
Many properties, such as volume or water retention which decrease on drying,
show an irreversible decrease beyond 10- to 15-bar suction.
Physical properties of allophane soils do not show the dependence upon
exchangeable cation which is prominent in soils with crystalline minerals. For
example, Kubota (1971) showed that there was no difference in glycerol
adsorbed on allophane with different exchangeable cations Mg, Ca, Sr, and Ba,
and only about 4% lower adsorption with K as compared with Li. For bentonite
there is a 20% difference for the divalent ions and nearly a 100%difference for
the monovalent ions. Water vapor adsorption showed a similar pattern.
I I. Index Properties
A. GRAIN SIZE DISTRIBUTION
The grain size distribution, also called particle size distribution or mechanical
analysis, of a soil is the most widely used index property for physical properties
of soils. Much effort has been spent in soil science on grain size measurements.
For soils with crystalline clay minerals, especially in glaciated areas, and for clay
contents less than 30%, one can predict many soil properties from the grain size
distribution (Warkentin, 1972). However, for allophane soils the grain size is not
an adequate index property.
The index properties used for allophane soils include water retention (Flach,
1964; Colmet-Daage et al., 1967) and plasticity (Warkentin, 1972). Packard
(1957) used surface area as a measure of clay content, as did Birrell (1966).
Flach (1964) recommends using the 15-bar water retention to estimate clay
The difficulty in obtaining dispersion, against both chemical and physical
forces, and the uncertainty of what is the unit particle of an allophane soil, are
the reasons for the limited usefulness of grain size in predicting physical
properties of allophane soils. Chemical deflocculation is the problem with wet
allophane subsoils, while in surface soils which have been dried the problem is
cementing to form larger particles. These cementing bonds can be broken to
different degrees. There is a considerable literature on the problems of dispersion
of allophane soils (Gautheyrou et al., 1976). It is not possible from most of the
studies to separate the effects of chemical deflocculation from physical dispersion. Therefore, the general term “dispersion” is used here.
The difficulty of dispersing allophane soils has been noted by many people.
Davies (1933) was one of the first to study the problem and recommended using
0.002 N HCl for dispersion. Kanno (1961) used 0.002 N HCl as a dispersant for
Japanese allophane soils. Optimum pH for good dispersion of Kanto loam, both
surface and subsoil, is in the pH range 2.5-3.5 (Tada and Yamazaki, 1963).
PHYSICAL PROPERTIES OF ALLOPHANE SOILS
AUophane soils flocculate in sodium silicate solutions, but polyphosphates can
sometimes be used as dispersants. Sodium pyrophosphate was a better dispersant
than sodium metaphosphate for Indonesian soils (van Schuylenborgh, 1953).
Low pH or calgon were used for Japanese soils (Kobo and Oba, 1964); the
former works better for subsoils and the latter for surface soils with organic
matter. Sherman et al. (1964) emphasized that the usual dispersing agents could
not be used for allophane soils. Ultrasonic vibration was found to increase
dispersion (Kobo and Oba, 1964), and is now generally used for allophane soils.
Oba and Kobo (1965) found that ultrasonic dispersion released clay size grains
from aggregates. A number of papers describe the use of this method (Gautheyrou et al., 1976). Espinoza et al. (1975) found that ultrasonic treatment still
gave much lower values for clay content than did the estimate from the 15-bar
water content, i.e., clay = FBP X 2.5.
Ahmad and Prashad (1970) have taken a different approach t o dispersion of
allophane soils. They found good dispersion by reversing the charge with
zirconium to get a positively charged particle.
Drying the sample decreases the measured clay content. This can be attributed
to the cementing on drying. The phenomenon has been described by many
workers, e.g., Sherman (1957), Birrell (1966), Wesley (1973). The magnitude of
the effect varies with the particular allophane soil. Kubota (1972) measured the
approximate soil suction at which irreversible bonding of clay into sand-size
grains occurs on drying. The clay and silt contents began to decrease when the
pF exceeded 3.5; fine sand-size grains were formed. Increases in coarse sand-size
grains were not measured until about pF 5, at which time the clay and silt-size
grains were at a minimum. The fine sand grains were then being bonded to coarse
sand-size grains. No further changes in grain size distribution occurred at suctions above pF 5.5.
Particles of pumice break down on stirring, and sand-size particles settle more
slowly than expected because of internal pores (Youngberg and Dyrness, 1964).
They also have a long wetting time because of entrapped air.
Kobo (1 964) has summarized the Japanese studies on dispersion of allophane
soils, and Colmet-Daage et al. (1972) report on an extensive series of tests of
dispersion of allophane soils of the Antilles and Latin America. Their results are
summarized as follows. There is no one best method which can be recommended. Surface and subsoils react differently, as do allophane soils containing
different components such as gibbsite or halloysite. Both flocculation and
incomplete physical dispersion occur. Undried samples always disperse more
completely than air-dried or oven-dried samples, the difference being much
larger for subsoils than for surface soils. Surface soil samples generally disperse
better at high pH of 10 or 11 with ammonium or sodium hydroxide (the Kanto
loam is an exception), while subsoils generally disperse better at pH 3 with HCl.
Subsoils generally flocculate at high pH. Sodium pyrophosphate is an effective
T. MAEDA ET AL.
dispersant for surface soils, but metaphosphate is not. AUophane soils containing
gibbsite are difficult to disperse in acid or basic suspensions. With small amounts
of gibbsite, dispersion appears to be better at pH 3; with larger amounts the best
dispersion is obtained at high pH. Soils with halloysite disperse at high pH, but
not at low pH. Ultrasonic vibration is recommended for dispersion. Birrell and
Fieldes (1952) had also noted that the presence of gibbsite makes allophane soils
more difficult to disperse.
Dispersion of surface soils at low pH might be attempted when solubilization
or organic matter at high pH would interfere with subsequent measurements.
The use of pyrophosphates might also interfere with other measurements on
separated soil fractions.
Control of pH is critical for dispersion at low pH, but not at high pH.
The results given by Kubota (1972) are representative of the effect of different
treatments on hydrandepts. On a B horizon sample the clay contents measured
were as follows: 1Okc-300W sonic dispersion, 56%; standard shaking on moist
sample, 31%; on air-dry sample, 5%; and on an oven-dry sample, 1%.
Baba (1971) was able to disperse allophane soils in an alkaline medium only
when ultrasonic treatment was used. Under these conditions sodium silicate was
a more effective dispersant than sodium polymetaphosphate.
While it is generally preferable to work with undried samples, this is not always
possible. Undried samples may not be desirable if the soil contains predominantly sand and gravel because of the difficulty in obtaining a representative
subsample of a wet soil.
Grain size analysis, therefore, has a limited usefulness in characterizing allophane soils. The measurement should be done on field-moist samples (e.g.,
Schalscha et al., 1965) and the details of the method should be given. Since
different allophanes react differently to dispersion treatments, some experimentation is necessary to obtain maximum dispersion (Colmet-Daage et al., 1972).
The method most generally used for dispersion is ultrasonic vibration and low
The plasticity is one of the physical properties which distinguish allophane
from crystalline materials. The name “allophane” comes from the striking
change on drying of allophane clays. Glassy when wet, the allophanes become
earthy on drying (Grim, 1953). The wet material is plastic, and the dry earthy
material is nonplastic.
Many papers have documented the plasticity limits, or Atterberg limits, and
their change on drying (Birrell, 1951; Gradwell and Birrell, 1954; van Schuylenborgh, 1953; Yamazaki and Takenaka, 1965; Wesley, 1973; Warkentin and
PHYSICAL PROPERTIES OF ALLOPHANE SOILS
Maeda, 1974; Kodani ef al., 1976; and others). Wet allophane soils have a high
liquid limit, but also a high plastic limit, and hence a low range of water content
over which they are plastic. As the samples are gradually dried, the liquid limit
decreases more rapidly than the plastic limit. Highly allophanic soils become
nonplastic before they reach the air-dry water content. The nonplastic state is
where the plastic limit cannot be measured, or where its measured value equals
or exceeds the liquid limit.
Since the samples are completely rewetted during the determination of plasticity limits, the decreases in liquid and plastic limits on drying indicate an
irreversible decrease in hydration of the allophane surfaces. The nature of these
irreversible changes is discussed in Section 111, A.
The values of plasticity limits are shown on a Cassagrande plot in Fig. 1.
Crystalline clays have plasticity values which fall near the “A” line. The allophane samples fall far from the line, with the most allophanic samples having the
highest liquid limit and the lowest plasticity index. This has suggested the use of
plasticity values in classification of allophane soils (van Schuylenborgh, 1953;
Gradwell and Birrell, 1954; Warkentin, 1972; Warkentin and Maeda, 1974). The
measurement of plasticity is r e a d y made, while other measures of allophane are
difficult to make. The intensity of allophanic characteristics would be highest
for samples with high liquid limit and low plasticity index, and lowest for values
approaching the A line. Samples which remain plastic on air-drying or oven-drying would have low allophanic characteristics.
The difficulty in using plasticity values for classification is that the measured
values depend upon degree of previous drying and upon content of organic
matter. Since the drying history of surface soil samples is usually not known,
air-dry samples may have to be used. The method might be more suitable for
FIG. 1. Cassagrande plasticity chart with values for Japanese allophane soils (from
Yamazaki and Takenaka, 1965). 0, Fresh soil; 0 , air-dried soil.
T. MAEDA ET AL.
Organic matter contributes t o the water held at the liquid limit (Kodani et aZ.,
1976), an effect which decreases irreversibly on drying. Maeda et al. (1976) have
shown the separate contributions of organic matter and mineral matter to the
liquid limit of allophane soils (Table I). The liquid limit is increased from 1.5 t o
3% for each 1% organic matter in the samples they used. Their results indicate
that not all the organic matter in a soil sample contributed to the liquid limit.
Bonfils and Moinereau (1971) found that both liquid limit (L.L.) and plastic
limit (P.L.) were strongly related to organic matter content (O.M.); the equations were: L.L. = 2.7 O.M. + 41 and P.L. = 2.7 O.M. + 34. The plasticity index
was not correlated with organic content.
The activity values (ratio of plasticity index to clay content) measured for
allophanes are variable. Northey (1966) gives values of 1.2 t o 1.5, Wesley (1973)
gives value below 0.6. The meaning of these values is uncertain because the
measurement of clay content is difficult.
The measurement of plasticity is more difficult for allophane samples than for
crystalline clays, and the precision is lower. This results from the low range of
water content over which the samples are plastic. The slope of the liquid limit
determination (water content versus log number of blows) is also more variable
for allophane clays than for crystalline clays. The one-point liquid limit method
(Sowers, 1965) cannot be recommended for allophanes. The degree of remolding
and working of the soil with water affects the liquid limit, especially of subsoils
(Ikegami and Tachiiri, 1966).
Measurements of liquid limit by Soma and Maeda (1974) on soils which were
gradually dried show a sharp break at a specific water content where irreversible
changes occur (Fig. 2). Drying does not produce irreversible change for the soil
shown until the water content falls below 100%. The shrinkage limit also occurs
at this water content.
Effect of Organic Matter on Liquid Limit of Allophane Soil'
at liquid limit
Soil content (%)
'From Maeda etul. (1976).
Decrease in liquid limit
Whole Organic matter Due to organic
soil (%) removed (%)
PHYSICAL PROPERTIES OF ALLOPHANE SOILS
60 80 IiO IhO I40 I60
Initial Water Content
FIG. 2. Decrease of liquid limit and volume of allophane soil o n gradual drying (from
Soma and Maeda, 1974).
Freeze-drying produces less particle bonding than air-drying, and results in
higher measured plasticity (Warkentin and Maeda, 1974). Ultrasonic vibration
increases the measured liquid limit (Soma and Maeda, 1974) because it causes
There is little change in liquid limit with different exchangeable cations. For
some samples the Na-soil has a slightly higher liquid limit than the Ca-soil, for
others it is reversed (Yazawa, 1976). The flow index (slope of the water content
versus number of blows curve in the determination of liquid limit) is also not
consistently different for Na or Ca allophane soils (Yazawa, 1976).
C. SURFACE AREA AND HEAT O F WETTING
Measurements of surface area for allophane soils have been reviewed recently
by Wada and Harward (1974). This literature will not be reviewed in detail here.
The surface area values are high, in the range of 300-600 mZg-’, but some of
this surface is not accessible to large molecules. This is not equivalent t o the
internal and external surfaces of swelling clays, but is due to small size of voids
and small necks leading to voids. The heats of adsorption indicate that physical
adsorption rather than chemical adsorption is dominant (Fieldes and Claridge,
Total surface area is often calculated from the amount of ethylene glycol
adsorbed, and external surface is measured from nitrogen adsorption; internal
surface area is estimated from the difference. Aomine and Egashira (1970) found
ratios of total to internal surface from 2.3 t o 3.0, while Fieldes and Claridge
(1975) report ratios of 1.8 to 2.6 when the nitrogen surface was measured after
heating to 600°C. Egashira and Aomine (1974) found that total surface area
measured on samples vacuum-dried over P2 O5 was higher than for oven-dried
T.MAEDA ET AL.
Aomine and Egashira (1970) measured the heat of immersion of allophane
soils in comparison with soils containing crystalline minerals. They found that
for equal surface areas, allophane soils had heats of immersion about twice as
large as montmorillonite soils. The ratio of heat of immersion to surface area
for allophane and from 0.023 to 0.029 for
ranged from 0.047 to 0.056 cal
montmorillonite soils. The exchangeable cation had only a small effect on heat
of immersion of allophane soils. For montmorillonite soils, hydration of exchangeable cations is a more important part of total heat of immersion. Water
molecules are bonded more strongly on allophane surfaces than on montmorillonite surfaces. When the heat of immersion was plotted against initial
water content of the allophane soils, the oven-dry samples fell below the smooth
curve joining the vacuum-dried and moist samples. This indicated to the authors
that oven drying had altered the nature of the allophane surface.
Maeda er al. (1976) measured heat of wetting values for an allophane soil
ranging from 7.3 cal g-' at 4% organic matter to 11.2 cal g-' at 26% organic
D. MINERAL DENSITY
Some measured values of mineral density, or specific gravity, of wet allophanes
are low, in the range of 1.8-1.9 g cm-3 (Fieldes and Claridge, 1975). These
values had been accepted in earlier studies. However, other measurements show
values of 2.7 or higher. Forsythe etal. (1964) quote values of 2.7-2.9 g ~ m - ~ .
Wada and Wada (1975) measured values of 2.72 to 2.78. They took special
precautions to remove entrapped air from the samples. If the unit particle of
allophane is accepted to be a hollow spherule of 50 A outside diameter and
about 30 A inside diameter, water movement into and out of this sphere would
be difficult and could account for the low mineral density sometimes measured.
Bonfils and Moinereau (1971) measured values of 2.32 to 2.70, the lower
values being for horizons with large amounts of organic matter-about 25%.
E. THERMAL CONDUCTIVITY
The thermal conductivity of a soil depends upon the conductivities of the
components-mineral, organic, water, and air. Because the path of heat transfer
from one component to another cannot be easily specified, the conductivity of a
soil cannot be readily calculated from the conductivities of the Components.
Models exist for some heat flow paths, but in general the preceding statement is
true. However, soil thermal conductivities vary in a predictable way with
properties of the components. On this basis allophane soils would be expected t o
PHYSICAL PROPERTIES OF ALLOPHANE SOILS
have thermal conductivity and thermal diffusivity values which are lower than
the corresponding values for soils with crystalline clay minerals. Diffusivity is the
ratio of conductivity to heat capacity and is the constant which relates temperature changes in the soil to the temperature gradient.
The conductivity of glass is lower than that of clay minerals or quartz
(Cochran et al., 1967). Allophane soils have a lower bulk density than soils with
crystalline minerals; this should result in lower thermal conductivity. The higher
water content would give a higher heat capacity and hence lower thermal
Some of the results obtained are summarized in Table 11.
Yakuwa (1 943) made extensive measurements on soil temperature and thermal
properties of different soils in Japan, including allophane soils.
Higashi (1951) measured a value of 0.32 cal g-' O C - ' for specific heat of dry
allophane soil. Thermal diffusivities were calculated from amplitude ratios and
phase differences of temperature waves; the phase differences gave more consistent results. While the dense packing seems to give higher values of diffusivity,
the scatter was such that no conclusions could be drawn on effect of bulk
density (Table 11). The maximum in the diffusivity curve occurred at 50% water
on a weight basis or about 0.25 on a volumetric basis. Thermal conductivity,
calculated from specific heat and diffusivity, was higher at high bulk density.
~ ; dense packing
The loose packing had a bulk density of around 0.5 g ~ m - the
varied from 1.0 at 0% water to 0.7 g cm-3 at 50% water. Higashi (1952) also
measured thermal properties of frozen allophane soils. The thermal diffusivity is
similar for frozen and unfrozen soils below a water content of about 30%, but at
higher water contents the diffusivity of frozen soils increases very quickly.
Cochran et al. (1967), using a line heat source probe, found that thermal
conductivity of a pumice material was very low, only 'slightly higher than for a
peat soil (Table 11). They used this to explain low night temperature and the
high incidence of frost in pumice soil areas of Oregon. The maximum in the
diffusivity-water content curve occurred at a low volumetric water content of
0.04 for the C horizons of the soil.
Maeda (1968) measured temperature gradients in soils near Sapporo, Japan. He
calculated thermal diffusivities from amplitude ratios and phase differences. The
diffusivity of fine pumice was higher than for an alluvial soil, and the allophane
soil was the lowest (Table 11).
Kasubuchi (1975a) found that the specific heat of allophane clay (0.229 cal
g-' "C-') was higher than bentonite (0.209), kaolin (0.201), or quartz (0.170).
Kasubuchi (1 975b) measured thermal conductivity of different soils using a line
heat source probe. The values again show conductivity and diffusivity values for
allophane which are 20 to 35% of those for soils with crystalline minerals. The
higher heat capacity and lower heat conductivity of allophane soils results in
slower temperature changes.
Measured Thermal Properties of Allophane Soils
(g cm--' )
(cal ~ r n - ~ ~ c - ' )(mcal cm-' sec-' "C'
Cochran et al. (1967)
Cochran ef a!. (1967)
Cochran et nl. (1967)
Cochran et al. (1967)
Cochran et al. (1967)
Maeda (196 8)
Kasubuchi (1 975b)
Miyazawa and Konno (1976)
Miyazawa and Konno (1976)