Tải bản đầy đủ - 0 (trang)
IV. Physical Characteristics of Allophane Soils

IV. Physical Characteristics of Allophane Soils

Tải bản đầy đủ - 0trang











Water Conieni, %

FIG. 3. Shrinkage curves for two allophane soils. Reproduced from Soil Science Society

of America Proceedings, Volume 38, Page 375, 1974 by permission of the Soil Science

Society of America. X , Field moisture; @, air dry; 0,oven dry.

the shrinkage curve changes and the high value of residual shrinkage (between

the shrinkage limit and zero water content) indicates a random arrangement of

units. Measured shrinkage is isotropic, again indicating random arrangement. For

crystalline minerals, the shrinkage limit is lowest for high swelling soils. The

shrinkage limit is, therefore, diagnostic for allophane soils (Warkentin and

Maeda, 1974).

Despite the large amount of shrinkage on drying, one would predict little

visible cracking for allophane soils in the field because the shrinkage is taken up

in small spaces between clusters (Section 111, B). Cohesion of allaphane is low

and it decreases if the samples are dried (Soma and Maeda, 1974). AUophane

soils do not form dry clods with dimensions of tenths of a meter.

The effect of remolding a soil depends upon its degree of consolidation

(Croney and Coleman, 1954). Remolding an overconsolidated soil exposes new

surfaces for water retention. AUophane soils are underconsolidated. Remolding

breaks some of the fabric bonds, decreases the soil suction, and increases the

amount of shrinkage (Takenaka, 1965).


This section will deal with soil water characteristics measured on allophane soil

samples in the laboratory; the mechanisms of water retention will be discussed.

The field water regime is described in Section IV, D.



It has been difficult to relate soil suction measured in the laboratory to the

water regime of allophane soils in the field. Only two variables-water content

and suction-are usually measured; while the complete description of the waterretention curve of an allophane soil requires the specification of four variableswater content, suction, sample volume or percent water saturation, and initial

degree on drying of the sample. Only for coarse-grained soils, where suction and

water content determine the system, have water-retention curves been useful in

predicting the field water regime. Measurements in the laboratory on finegrained soils, especially swelling clays, have not been useful in predicting soil

water behavior in the field. The water-retention curve for swelling crystalline

clay soils requires specification of three variables-water content, suction, and

volume. If the structure is disturbed on sampling this can be considered a fourth

variable. Initial bulk density is assumed to be the same as that in the field, even

though this is not always assured.

Another difficulty in applying laboratory measurements to the field is the

unreliability of water-retention measurements during wetting of fine-grained

soils. The measurements are almost always made only on the drying part of the


Therefore, the main use for water-retention measurements for allophane soils

is in characterizing soil surfaces or void-size properties. In addition, ColmetDaage and his colleagues (e.g., Colmet-Daage et al., 1967) have used p F values

for wet and dry allophane soils as an index property (see Section 11, A). Hughes

and Foster (1970) have suggested that the water content at a specific suction can

be used to rank the degree of disorder, or the amount of extractable materials, in

halloysite and allophane. Galindo-Griffith (1974) showed that the 15-bar percentage was not correlated with surface area, but was correlated with reactive

alumina and with the point of zero charge. Many physical properties of allophane soils are interrelated, as has been shown in a number of papers (e.g.,

Warkentin and Maeda, 1974).

Water retention by allophane soils is determined by the size distribution of

voids, not by the amount of surface area (e.g., Fujiwara and Baba, 1973). This

fact differentiates allophane soils from soils with swelling clay minerals,

The amount of water held by wet allophane samples is lllgh (Table 111); the

very high water contents at 15-bar suction are striking. This results from the

large volume of small voids. Drying the samples decreases water retention at any

suction value. These measurements have been made by a number of people.

Msono et ul. (1953) made a detailed study of water-retention characteristics of a

number of Japanese allophane soils. Colmet-Daage and his co-workers (1967,

1970) have published measurements for a wide range of allophane soils from the

Caribbean, Central America, and South America. These reports provide very

valuable source material on properties of allophane soils, A complete list of

papers is available in the bibliography by Gautheyrou et al. (1976).




Measured Water Content at Different Suctions for Allophane Soils

Water content at suction







0.001 bar 0.3 or 0.5 bar 15 bar








Misono ef QI. (1953)

Misono et QI. (1953)







Maeda and Warkentin


Maeda and Warkentin





















0-1 10 cm, weakly allophanic



















Hydrandep t






50-300 cm, strongly allophanic



Flach (1964)

Colmet-Daage er QI.


Colmet-Daage et al.


Colmet-Daage et aL


Colmet-Daage et al.


Colmet-Daage et al.


Colmet-Daage er at.


Colmet-Daage er al.


Colmet-Daage et QI.


The measurements by Colmet-Daage and Cucalon (1965) illustrate the large

decreases on drying in water content at p F 4.2 and 2.8. The amount of water

held between these two suction values, an estimate of plant-available water, also

decreases on drying. In extreme cases for certain horizons this decrease was from

45% to 3% and from 68% to 5% by weight.

Soils with a high content of allophane have an S-shaped water retention curve,

similar t o the shape for a coarse-grained soil. The approximately linear portion

on the water-content-log-suction plot is between 0.01 and 1-bar suction (Maeda

and Warkentin, 1975). For swelling crystalline clays this linear portion extends

from less than 0.01 bar to around 100 bar.



Forsythe (1972) states that the loss in water retention on drying is greater at

lower suctions than at high suctions. Maeda and Warkentin (1975) confirmed

this for samples with moderate allophanic properties; for highly allophanic soils

the effect appeared to be reversed. Forsythe (1972) found that volumetric water

content increased on air-drying; Maeda and Warkentin (1975) found a lower

volumetric water content and lower percent saturation of samples after drying.

Parfitt and Scotter (1972) report on an allophane soil from Papua, New

Guinea, with a low 15-bar percentage of 36% and a 0.1-bar percentage of 120%.

This results in a high content of plant-available water.

Colmet-Daage et al. (1967) checked the influence of organic matter on water

retention by allophanic soils. They realized that treatment with peroxide could

aid in dispersion, so that the measured difference could not be attributed solely

to organic matter. They found that treatment with hydrogen peroxide to remove

organic matter had no influence on water retention at pF 2.5. At p F 4.2 some of

the treated samples had lower and some had higher water retention than the

control samples. They concluded that organic matter had only a small influence

on water retention for the soils which they studied. However, in humic dlophane soils, the organic matter content is important in water retention. Swindale

(1964) reports that for some Hawaiian soils, percent field water content (w)

increases with percent organic matter (O.M.): w = 4.2 + 34.7 O.M. Takenaka

(1973) and Maeda et al. (1976) pointed out that water retention by allophane

soils high in organic matter content was especially decreased on drying.

A number of authors have divided the water-retention curve into three or more

classes of water (e.g., Misono et aL, 1953). The breaks in the water-retention

curves indicate some validity in this approach, but there is no evidence that

forces of water retention are different and can be separated in this way. The

retention appears to be due to voids of different sizes except at high suction

values, larger than 100 bar, where surface adsorption is involved. The waterretention curve at suctions below 100 bar can then be used to obtain a

qualitative determination of void size distribution (Misono et al., 1953; Maeda

and Warkentin, 1975).

Masujima (1962) found that different exchangeable cations did not change the

water retained between 10 and 100 bar.


Rate of water transmission through allophane soils is high, due t o the low bulk

density and the granular structure of surface horizons (Table IV). At the same

void ratio, allophane soils have a higher saturated hydraulic conductivity than

soils with montmorillonite (Maeda and Warkentin, 1975).

Drying increases the saturated hydraulic conductivity at constant bulk density.

The increase is at least two orders of magnitude for highly allophanic soils but


25 1


Saturated Hydraulic Conductivity of Allophane Soils









Costa Rica


Kanto loam





(g ~ m1 - ~



Saturated K

(cm sec-' )


6 X lo4

Kubota (1972)


1 x 10-5

Maeda and Warkentin (1975)

Maeda and Warkentin (1975)

2 x 10-~

Colmet-Daage el al. (1967)

3 x 10-~

Forsythe (1975)


Tabuchi (1963)

Tabuchi (1963)

Yamanaka (1964)




2 x 10"

becomes smaller when the content of allophane decreases (Maeda and Warkentin,

1975). This indicates that dried samples have a larger proportion of large voids,

i.e., that the small voids are lost preferentially on drying (Misono et al., 1953).

The common observation that dried allophane soils resemble sands in their

physical properties is due to formation of aggregates with small internal porosity

and hence large interaggregate void volume.

Tada (1965) described non-Darcy flow of water in fresh allophane soil. The

conductivity was constant to hydraulic gradients of about 15, then increased

about 20 times t o gradients of 50, after which the conductivity again decreased.

Dried allophane soil did not show this effect.

Iwata (1 963, 1966) published a thorough study of water movement in unsaturated soils t o clarify water redistribution and field capacity concepts. He found

that allophane soils had a much higher unsaturated hydraulic conductivity than

soil with crystalline clay minerals when compared at the same suction. For this

reason the field capacity occurred at a higher suction in allophane soils. Iwata

(1963) measured unsaturated hydraulic conductivity values of 0.1 cm day-' at

0.1 bar and 1 cm day-' at 0.05 bar for an allophane soil. These values were 3 to

4 times higher than measurements on an alluvial soil. Maeda and Warkentin

(1975) measured values of 0.01 cm day-' at 1 bar and

cm day at 7 bar.

The unsaturated conductivity of allophane soils was higher than that of crystalline clay soils at suctions below 2 bar.

The advance of the wet front during infiltration also increases from wet to dry

allophane samples (Maeda and Warkentin, 1975). This is opposite t o the effect

for crystalline clay minerals. The diffusivity was four orders of magnitude larger



for the dry sample. The change in soil water diffusivity with change in bulk

density was about the same as for soils with crystalline clay minerals.

El-Swaify and Swindale (1968) found that relatively high levels of salinity and

sodium in irrigation water could be used on allophane soils and still maintain

adequate permeability. The sodium caused only slight changes in structure of

the soil. El-Swaify (1973) found anion effects on hydraulic conductivity for

allophane soils.


High percolation rates in the surface and subsoil present a problem in managing rice paddies in Japan. A percolation rate of 3-4 cm day-' is desired. Often

bentonite is used to decrease percolation rates because soil compaction by

rolling is not sufficient. Another difficulty is the variability of percolation rates

within a field; percolation does not follow a normal distribution and a few high

values can result in ratios of the mode to the mean of 0.05 to 0.3 (Ishikawa et

d.,1963). This makes effective bentonite dressings difficult to achieve. Birrell

(1952) comments that the natural variation in water content and compaction

characteristics of allophane soils in the field has made engineering investigations


The variability of physical properties of allophane soils in the field has received

considerable attention. A number of Japanese research workers cooperated in a

study organized by the Japan Society for Irrigation, Drainage, and Reclamation

Engineering on field variability of physical properties (e.g., Kuroda, 1971,

Tokunaga and Sato, 1975).

Forsythe (1975) has reviewed measured infiltration rates for allophane soils in

Central America. The values are generally high, initial infiltration rates of 20-70

cm hour-' and 2-hour rates of 5-20 cm hour-' are reported. He points out that

these high rates make the soils unsuitable for furrow or flood irrigation.

Nakano et al. (1970), in a series of three papers, reported laboratory measurements of infitration and evaporation from columns of layered volcanic ash soils

of different grain size, and field measurements of water movement and evaporation. Internal soil drainage can often be increased by mixing the layers to

overcome boundary effects (Kon, 1967).


Relatively little information has been published on field studies of water

available to plants in allophane soils. Many numbers for available water are based

on water held between 15-bar and 0.5- or 0.3-bar suction. There is evidence that



15 bar is too high for the upper limit; that 5- to 8-bar suction is observed in the

field (e.g., Misono and Terasawa, 1957; Masujima and Mori, 1962). It is also to

be expected that the lower limit may be closer to 0.1 bar for allophane soils with

relatively high permeability. Kira et al. (1963) found the field capacity t o be at

0.08 to 0.1 bar, and commented on the difficulty of estimating field water

regime from water-retention curves. Chichester er al. (1969) found the field

capacity occurred at suctions of less than 0.05 bar for a pumice soil in Oregon.

Youngberg and Dyrness (1964) found the value was below 0.1 bar.

Shiina and Takenaka (1961) used the water content after 24 to 48 hours

drainage as the upper limit. They found that crop growth decreased markedly

when the suction exceeded 1.5 bars in the root zone. They also found considerable water movement up from wet subsoil layers, enough to supply 50% of

evapotranspiration in their experiment. Masujima and Kon (1963) also found

that water movement up from the subsoil contributed to available water.

Shiina (1963) rejected the use of 0.5- and 15-bar suctions to define the

available water. He said this must be based upon growth period and weather

conditions as well as soil suction. Plants use water held at suctions as low as 0.03

bar, and initial wilting can occur at 2 bar. Masujima and Mori (1962) noted that

water was not equally available over the range of available water, and that

availability depended upon plant and soil factors.

Masujima and Mori (1 962) studied physical properties and plant growth;

increasing noncapillary porosity was associated with decreasing water availability. Optimum porosity for growth of beans was 30% at 0.5 bar and 20% at

0.03 bar. They suggested that land improvement could be achieved by mixing in

different particle sizes to achieve an optimum void size distribution.

Forsythe et al. (1964) contains a good summary review of measured physical

properties such as bulk density and available water content. They rate available

water in many allophane soils as average to low, although some allophane soils

have a high available water content. On a volume basis the available water in

allophane soils is not markedly different from soils with crystalline minerals

(Swindale, 1964). On a weight basis, the numbers are very high because of the

low bulk density of allophane soils.

Bonfils and Moinereau (1971) report relatively high values of available water

(15 to 21%), with the available water increasing with increase in organic matter


V. Soil Engineering

The term “cohesive volcanic ash soils” is often used in the soil engineering

literature. Cohesion indicates clay properties. The term allophane soils is used in

this review in the same sense, and will, therefore, be used in this section as well.



Many of the engineering properties of allophane soils in Japan have been

studied on “Kanto loam,” an allophane soil found on the Kanto plain north of

Tokyo. The subsoil, especially, of the Kanto loam shows typical allophane



Soils are compacted when large earth-moving equipment is employed in land

reclamation. From the viewpoint of engineering, it is important to know the

compaction characteristics of soils. The compaction curve for a soil is a plot of

water content versus the bulk density whch can be achieved by a specified

compactive effort at that water content. This curve shows a maximum bulk

density at a water content called the “optimum water content” for Compaction.

Compaction produces a moderate increase in density and a large increase in

strength of allophane soils. However, on resaturation the strength is again

decreased (Northey, 1966).

AUophane soils have a remarkably high natural water content compared with

soils containing crystalline clay minerals, and moreover water-holding characteristics change during drying. These differences are reflected in differences in

compaction characteristics. Allophane soils have low maximum bulk densities, in

the range 0.8-1.3 g cmd3and relatively high optimum water contents (Table v>.

The optimum water content is much below the natural water content.

The compaction curve of a soil with crystalline clay minerals is the same

regardless of initial water content at the beginning of the compaction test;

however, allophane soils show different curves depending upon the initial water

content and the amount of remolding produced during testing (Birrell, 1951).

An undried allophane soil does not show a distinct maximum in bulk density,


Compaction Characteristics of Two Typical Allophane Soils


Kanto loam




Java andosol




Maximum dry density


(g cmW3

Optimum water content

(w, %)














Kuno and Mogami (1949)

Kuno and Mogami (1949)

Kuno and Mogami (1949)

Wesley (1973)

Wesley (1 973)

Wesley (1973)



and hence no identifiable optimum water content. Since the natural water

content exceeds the optimum water content, this curve can be measured only by

gradually drying samples for compaction. The bulk density increases only

gradually as the water content is decreased. Once’the soils are dried, they show

typical compaction curves. This behavior is illustrated in Fig. 4, which is typical

of the results obtained. These characteristics are described by Kuno and Mogami

(1949), Birrell (1951)’ Tada (1965), Tokunaga (1965), Northey (1966), Frost

(1967)’ Takenaka (1973), Wesley (1973), Adachi and Takenaka (1973)’ and

others .

The compaction curve obtained by first drying and then rewetting a sample,

therefore, forms a loop with only the drying portion showing a maximum

density (Takenaka, 1973). This is due t o the reduction in the amount of water

retained at any suction value. This loop forms when the soil is dried below a

critical water content, which Tada (1965) found to be at a suction of 15 bar for

the Kanto loam subsoil. This is the highest suction which the subsoil would be

subjected to under natural soil conditions with drying only by plant roots.

Since it is often difficult to find the optimum water content and the maximum

dry density on the compaction curve of fresh allophane soil, the criterion of

maximum dry density cannot be applied for embankment work. The degree of

saturation is used in Japan in place of the maximum dry density to follow the

degree of compaction of allophane soil (Kuno and Yabe, 1960, 1962).

The permeability of compacted soils is an important engineering consideration.

When allophane soils are compacted, with a decrease in water content, the dry

density increases slightly, but the permeability also increases. According to

Tada’s (1965) experiments, coefficient of permeability reaches a minimum near

the natural water content, and tends to gradually increase on lowering of water

content. Tokunaga’s (1969) experiments studied these phenomena in relation to

Zero air voids











Water Content




FIG. 4. Compaction curve for allophane soil on gradual drying (from Northey, 1966). A,

Dried from natural water content (both samples, 7592B and 7592C); B, dried to 53% and

rewet (7592B); C, dried to 70% and rewet (7592C).




Permeability of Compacted Allophane Soils‘

Water content (%)

Permeability (cm sec-’ )



7 x lod




Bulk density (g ~ m - ~ )





‘From Tokunaga (1969).

consistency and structure of the soil. Microscopic observations revealed that in

the range of low water content the soil has a granular structure with large voids.

Therefore, it has a high permeability coefficient in spite of the higher bulk

density. As the water content increases near the plastic limit, aggregated blocks

are developed causing hindrance to water passage and lowering permeability.

Further increase of water content up to around the natural water content causes

the aggregated blocks to flow into a pastelike structure. In this range of water

contents the coefficient of permeability is a minimum, and the dry density is

low because of high water content (Table VI).

Allophane soils rich in humus show a different behavior. The organic matter

forms stable humus which combines with soil particles forming an aggregated

structure. Permeability increases slightly at high water content for compacted



Allophane soils can be stable in the undisturbed state, often occurring in

relatively steep banks (Northey, 1966; Wesley, 1973). They are also relatively

resistant to erosion, although landslides and water erosion occur on steep slopes.

The strength of allophane soils disturbed by excavation and embankment is

remarkably lower than that of the undisturbed soils. Once disturbed, allophane

soils are too weak t o ensure traffcability for construction equipment (Highway

Research Board, 1973).

The Kanto loam has the bearing capacity to support buildings of four or five

stories in spite of its high natural water content. The unconfined compressive

strength of Japanese allophane soils is in the range of 1.O-2.3 kg cm-’ (Highway

Research Board, 1973). This is more than five times the strength of alluvial

nonvolcanic soils at the same water content. It is also recognized that the bearing

capacity, estimated from the N-value obtained from the standard penetration

test, is higher than that of the nonvolcanic clays. For allophane soils in the



undisturbed condition, the experimental equation between the N-value of the

standard penetration test and the bearing capacity (4a) is given as follows:

4a = (2 to 2.5)N

For alluvial clay soils the equation is:

qa = (1 to 1.3)N

Birrell (1951) found small f,riction angles, 0-8", and low shear strength,

0.2-0.4 kg cm-', from triaxial test results. Pope and Anderson (1960) measured

values as low as 1" for friction angle and 0.5 kg cm-' for the cohesion parameter.

Wesley (1973) found undrained shear strength values of 1.0 to J.2 kg cm-2, and

in situ vane shear results of 0.7 to 1.0 kg cm-2. The unconfined compressive

strength of allophane soils decreases with increasing organic matter content

(Yamanouchi and Yasuhara, 1972).

Most authors comment on the large variability over short distances in engineering properties of allophane soils (Pope and Anderson, 1960; Northey, 1966;

Wesley, 1973). This high heterogeneity in the field makes it difficult t o use

measured results in design of earth structures.

Another feature of allophane soils is that strength does not increase with

depth, i.e., with increasing overburden pressure. Since it is not possible to

estimate strength from the natural water content, no practical indices exist to

estimate the strength.

The strain at failure is only 2-3% for allophane soil, against 2-6% for other

clay soils. When allophane soils are disturbed, the compressive strain increases

(Gradwell and Birrell, 1954; Takenaka, 1965). Gradwell and Birrell (1954) and

Komamura and Takenaka (1973) show the Mohr circles for shearing resistance as

a function of stress for undisturbed and remolded allophane soils.

Allophane soils have moderate measured sensitivity, the ratio of undisturbed

to remolded shear strength. Birrell (1951) gives values of 6-12, Wesley (1973)

measured values between 1 and 3 . Wells and Furkert (1972) showed that water

in undisturbed allophanes was held in hydrogen-bonded clusters. Remolding

breaks up these clusters, and the water becomes distributed as single-linked


Mophane soils have high water contents and a well-developed soil structure.

The lowering of the strength on remolding is due to change in water-holding

characteristics and structure peculiar t o allophane soils. On disturbance, water

which was held firmly in the voids is released and free to flow. The measured soil

suction has decreased; the decrease is largest at low suction values. In consequence, the soils soften and lose strength. The soils least resistant t o this

softening are buried soils with high organic matter, especially from old volcanic

ash layers (Takenaka, 1966; Komamura and Takenaka, 1973; Takenaka and

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

IV. Physical Characteristics of Allophane Soils

Tải bản đầy đủ ngay(0 tr)