Tải bản đầy đủ - 0trang
II. Relative Stability of Minerals; Weathering Sequences and Indexes
M. L. JACKSON AND Q. DONALD SHERMAN
out for minerals of heavy specific gravity, coarse-grained minerals, colloidal minerals, and combinations of these categories.
a. Weathering Sequence of Heavy Ninerals. Of the minerals of heavy
specific gravity, the less stable were found to occur in decreasing amounts
in rocks of increasing age (Pettijohn, 1941). Thus olivine was not found
in rocks older than Pleistocene; augite is rare or unknown in rocks of
pre-Cambrian age but becomes increasingly common in younger rocks.
Pettijohn’s weathering sequence (most stable mineral first) follows :
( -3) anatase, (-2) muscovite, (-1) rutile, (1) zircon, (2) tourmaline,
(3) monazite, (4) garnet, (5) biotite, (6) apatite, (7) ilmenite, (8)
magnetite, (9) staurolite, (10) kyanite, (11) epidote, (12) hornblende,
(13) andalusite, (14)topaz, (15) sphene, (16) zoisite, (17) augite, (18)
sillimanite, (1919)hypersthene, (20) diopside, (21)actinolite, and (22)
olivine. The first three minerals were assigned negative numbers to
indicate their tendency to formation rather than disappearance during
long periods of burial.
Longstaff and Graham (1951) found hornblende to be more stable
than olivine. This discrepancy from Pettijohn’s sequence may be a result of the different properties of the many hornblendes found in nature.
A grouping of heavy minerals in four categories of stability is suggested
by Weyl (1952):
Extremely unstable : Olivine, hornblende, augite.
Slightly stable : Garnet, epidote.
Stable : Staurolite, kyanite, sillimanite, andalusite.
Very stable : Tourmaline, zircon, rutile, titanite, magnetite.
Determination of total zirconium was reported by Marshall and Haseman (1943) to be a satisfactory method for estimation of zircon, the
most resistant of the heavy minerals. As is indicated by Pettijohn’s sequence given above, zircon should be the most appropriate reference
mineral (rather than alumina, commonly employed) from which to calculate the chemical weathering losses of materials. Calculations of this
type based on zirconium are made for a number of soil profiles by Haseman and Marshall (1945).
Buckhannan and Ham (1942) found that epidote, augite, hornblende,
and kyanite had disappeared from soils on old (Permian) materials and
had largely disappeared from old alluvium but were common i n the soils
on Tertiary and Pleistocene materials in Oklahoma. This study illustrates the way in which the relative stability of heavy minerals can be
employed to establish the age of the rock formation from which soils axe
derived. Carroll (1944) concluded that heavy minerals (specific gravity
of 2.9 and over) in the sand fraction of soils give a useful means of identifying parent rock formations for a number of west Australian soih.
CETEMICAL WEATHERING OF MINERALS IN SOILS
b. Weathering Seqwence of Coarse-Grained Minerals. To the extent
that the susceptibility to weathering of a mineral overshadows specific
surface, a weathering sequence can be defined strictly in terms of mineral
species without reference to their particle size. As an extreme case,
gypsum is moderately soluble and is thus easily removed by leaching;
consequently, no attention need be paid to the crystal size in rating its
susceptibility to weathering.
Goldich (1938) presented a branched stability series of coarse-grained
minerals (most stable last) :
This sequence is similar to the Bowen (1922) sequence for temperature
of crystallization (olivine, 1890" C. grading to quartz a t 570" C.), a fact
that has been interpreted as reflecting the increasing instability of minerals with the increasing departure of the temperature from that of the
environment of formation.
Marel (1947, 1948a, 1949) has made extensive observations on the
relative stability of magmatic minerals, when weathering conditions and
particle size are kept equal, and has summarized his observations (personal communication) in the following weathering sequence (most resistant last) : basic volcanic glass, olivine, hypersthene, biotite, augite,
amphibole, anorthite, epidote, bytownite, andesine, oligodase, muscovite,
garnet, orthoclase, microcline, albite, allanite, zircon, staurolite, rutile,
tourmaline, and quartz. He summarizes : " The most resistant of the minerals accumulate, so that . . . the weathering state of the soil, e.g. the
power of a soil to supply nutrient elements to the plant" can be characterized, as will be further considered in Section V. Marel (1947)
found that rhyolitic tuff contained 42 per cent of amphibole in the heavy
fraction, the slightly weathered material contained 32 per cent, the moderately weathered contained 14 per cent, the strongly weathered 5 per
cent and the very strongly weathered revealed 0 per cent of amphiboles.
I n this same investigation, the resistant mineral zircon increased from
6 per cent in the heavy fraction of rhyolitic tuff to 40 per cent in the
very strongly weathered sample.
Quartz is listed as the most stable mineral in the two sequences just
given because coarse-grained quartz is extremely resistant to weathering.
M. L. JACKSON AND Q. DONALD SEERMAN
Humbert and Marshall (1943) have shown how quartz grains can be
weathered by iron-oxide crystal growth in cracks. Engelhardt (1937)
similarly noted sharp-angled (fractured) quartz grains in the finer fraction and rounded grains in the coarse. Quartz particles of clay size have
a lower rela.tive stability, as will be brought out in the weathering sequence for clay-size particles.
A sequence of orthoclase weathering in acid and alkaline environments is given by s,A. Tyler of the University of Wisconsin in his lectures :
The broken arrow was inserted by the writers to represent the accumulation of gibbsite crusts on weathered rocks (Alexander et at., 1942).
A weathering sequence of both light and heavy minerals, referring
largely to coarser-size particles of the silt, has been worked out by
Graham (1940,1950, and personal communication) on the basis of reactions of the minerals with H-clay as follows :
Least stable : Olivine, apatite, anorthite, and bytownite.
Moderately stable : Biotite, augite, hornblende, garnet, epidote, labradorite, and andesine.
Stable : Staurolite, orthoclase, microcline, albite, and oligoclase.
Most stable : Quartz, muscovite, zircon, tourmaline, rutile, ilmenite,
anatase, kyanite, titanite (sphene), and magnetite.
Vanderford (1942) studied the changes in quartz and calcium content of the silt in seven locations extending from northeastern Iowa to
central Mississippi, and found that the quartz content increased from
about 60 to about 80 per cent as the temperature and rainfall increased
from the northerly to the southerly location. The calcium content decreased from about 80 to about 60 meq. per 100 g. in the same traverse.
Quartz and calcium in the silt thus appear to be a moderately sensitive
measure of degree of weathering within the temperate zone. The difference can be explained by the weathering away of plagioclase in accordance with the relative stability sequences given above.
Springer (1949) showed that the ratio of K :Ca in the silt was 0.3
greater in the A horizon than in the underlying loess, which indicated
CHEMICAL WEATHERINXI OF MINERALS I N SOILS
relatively more weathering of calcium than of potassium minerals in the
A horizon. He also concluded that the quartz: Ca ratio is a sensitive
measure of the degree of weathering of loessial materials. Younger and
thicker loess contains more feldspars and other easily decomposed minerals than thinner and older loess occurring along the Mississippi River,
according to Wascher et al. (1948).
c. Weathering Bequelzce of Clay-Size Mineral Particles. With increasing fineness of mineral particles, the mineral stability sequence is
somewhat different from that of coarse-grained mineral particles because
specific surface becomes great enough to hasten weathering of minerals
which are more stable in coarser sizes. By clay-size mineral particles is
meant mainly mineral particles less than 211. in diameter, but the sequence
applies to some extent to the size range of 2-5p (fine silt-size particles)
also. The mineral weathering sequence of Jackson et al. (1948, 1952),
consisting of thirteen stages for fine particles (most stable stage given
last), follows : (1)gypsum (also halite, sodium nitrate, ammonium chloride, etc.) ; (2) calcite (also dolomite, aragonite, apatite, etc.) ; (3) oLivine-hornblende (also pyroxenes, diopside, etc.) j (4) biotite (also
glauconite, magnesium chlorite, antigorite, nontronite, etc.) j ( 5 ) albite
(also anorthite, stilbite, microcline, orthoclase, etc.) ; ( 6 ) quartz (also
cristobalite, etc.) ; ( 7 ) muscovite (also 10-A. zones of sericite, illite, etc.) ;
(8) interstratified 2 : 1 layer silicates and vermiculite (including partially
expanded hydrous micas, randomly interstratified 2 : 1 layer silicates with
no basal spacings, and regularly interstratified 2 : 1layer silicates) ; (9)
montmorillonite (also beidellite, saponite, etc.) ; (10) kaolinite (also
halloysite, etc.) ; (11) gibbsite (also boehmite, allophane, etc.) ; (12)
hematite (also goethite, limonite, etc.) ; and (13) anatase (also zircon,
rutile, ilmenite, leucoxene, corundum, etc. ) .
Included in the above list are slight changes from the original, including the inclusion of a.patite at stage 2 ; the substitution (Jackson
et al., 1952) of interstratification and vermiculite terminology for “mica
intermediates’’ in stage 8 ; the assignment of allophane to stage 11
(Tamura et al., 1953) ; and the inclusion of zircon (Marshall and Haseman, 1943 ; Marel, 1947) and leucoxene (Tyler and Marsden, 1938 ; Carroll and Woof, 1951) in stage 13. Vermiculite in quantities sufficient to
give a distinct 14-A. spacing was given (Hseung and Jackson, 1952) the
weight 8.5 in the weathering mean calculation, Section 11, 2c.
Kelley t$ al. (1939a) reported chlorite in the colloid of a California
soil, and later Jeffries and Yearick (1949) reported chlorite in the clay
fraction of several Pennsylvania soils. Stephen (1952a) reported chlorite i n soils of the Malvern Hills of England and attributed its origin to
weathering of basic rocks (high in hornblende, biotite, etc.) The weath-
Y. L. JACKSON AND Q. DONALD SHERMAN
ering sequence suggested by Stephen (1952b) was from these ferromagnesian minerals to chlorite, and then to vermiculite, in accordance with
the weathering sequence given in the previous paragraph (Jackson e t al.,
1948, 1952). However, occurrence of a pedogenic chlorite formed in a
weathering sequence following mica-vermiculite has been proposed by
Jeffries et al. (1953). This indicates tlie possible occurrence of pedogenic chlorite in several positions in the 2 : 1layer silicate stages (stages
4, 7 through 9) according to its composition. Caillere e t al. (1947) reported formation of a 14-A. stable diffraction spacing as a result of treatment of a montmorin with MgC12. Thus one might expect the occurrence
of a n aluminous chlorite (from mica layers), in a regime of high magnesium such as with biotite weathering, in the same weathering stage
as montmorillonite. At the same time ferromagnesian chlorite of rocks,
listed in the previous paragraph, has stage 4 stability. Chlorites of intermediate composition would be expected to ha,ve stabilities intermediate
between stages 4 and 9. L. D. Whittig and the senior author have found
chlorite interstra,tified with biotite in weathered coarse silt of a Wisconsin
2 . Weathering Indexes
Weathering indexes refer to the numbers which express the degree of
chemical weathering of a mineral or soil material. The indexes condense
the data of mineralogical analysis and assist in its interpreta.tion, both
in terms of weathering processes and in terms of probable fertilizer needs
of soils. Several types of indexes are summarized here but undoubtedly
many more have been employed.
a. Molar Ratios. Ra.tios of moles of different constituents of rocks,
minerals, and soils were employed to condense elemental analyses of
these materials (Harrassowitz, 1926 ; Jenny, 1931 ; Marbut, 1935). The
use of such ratios was an important way of evaluating the relative rates
of loss of tlie different elements during the course of weathering, as
clearly summarized by Jenny (1941). These ratios were emphasized to
a greater extent before methods of mineralogical analysis of colloids and
the weathering sequence of colloidal mineral species were developed.
The chief molar ratios that have been employed are as follows :
=silica :alumina ratio, sa
=silica, : ferric oxide ratio, sf
CHEMICAL WEBTHERING OF MINERALS IN SOILS
K 20+ Naz0 +Ca0
=silica :sesquioxide ratio, ss
=bases :alumina ratio, ha
=alkali :alumina ratio, ha,
=alkaline earth : alumina ratio, bu3
=alumina : ferric ratio, af
= calcic : magnesic ratio,
= potassic :sodic ratio, p s
Alumina has often been employed as the basis for calculation of relative weathering losses of other constituents ; hence, the above ratios to
alumina serve as measures of depletion of the other constituents. Jenny
(1931) employed the ratio of KzO : Na20 as a measure of degree of weathering, since sodium is lost by weathering a t a higher rate than potassium.
The “shifting value” is this ratio for the original rock subtracted from
the ratio for the weathered product (Robinson, 1949). Also, Jenny
(1931) proposed the beta value which is the ratio
bal of leached horizon
hal of the parent material
Reiche (1943, 1950) formulated a more complex and inclusive measure of the degree of weathering of rocks based on molar ratios. H e
plotted “weathering potential” defined as
x moles (CaO + NaaO + MgO + KaO - HaO)
moles (combined SiOp +TiO, + A1,0, + FepO, + CraO, + CaO + Na,O + MgO + K,O)
against the “product index” defined as
100 x moles (combined SiOz)
moles (combined SiOz
+ TiOz + R203)
M. L. JACKSON AND 0. DONALD SHERMAN
The consistency in over-all results of the molecular ratios to weathering sequence of colloidal minerals is a t once apparent from the shift in
elemental composition of the minerals as several stages of the sequence
is traversed ; this general relationship with respect to the silica : sesquioxide ratio was pointed out at the time of the presentation of the weathering sequence for cla,y-size minerals (Jackson et al., 1948). The fact
that the silica : sesquioxide ratio has little significance among the 2 : 1
layer silicates (corresponding to stages 6-9 in the sequence) has been
pointed out by Marshall (1935) and further by Hseung and Jackson
Robinson and Richardson (1932) and Robinson (1949) expressed the
degree of weathering by the proportion of the total alumina of the soil
which was in the clay fraction. They also suggested an index similarly
calculated for iron oxide. Layer silicates in clay under this system of
evaluation are given the role of the final end product of weathering (that
is, a sediment consisting of layer silicate clay and quartz silt would have
a weathering index of 100 per cent). Contrary to this role, it is clear
from many reports in the literature that layer silicates of clay size continue to weather progressively ; thus, micas weather to partially expanding and expanding layer silicates, to kaolin, and to free oxides.
6 . Weathering Stages. The term “weathering stage” as applied to
the weathering sequence grew out of the common expression “stage of
weathering” of rocks. Polynov (1937) defined four stages of weathering
and characterized each as a measure of the extent to which weathering
has proceeded, as follows :
coarse detrital material
chlorides and sulfates
(removed to another site)
carbonate (deposited elsewhere)
siallitic (allophanic) products
The first stage is recognized as physical weathering and sorting, whereas
the other three stages concern chemical weathering. These broad stages
of chemical weathering of rocks to form parent materials and soils are
universally recognized. The extent to which the aluminosilicates and the
aluminous compounds of his stages I11 and IV are merely residues as
opposed to products of the synthesis mechanism and their mineral compositions is not established by Polynov. Polynov did make a contribution
in emphasizing a continuity between the various stages of weathering.
CHEMICAL WEATHERING O F M I N E R A L S IN SOILS
It is not yet known the extent to which longer time may offer a counterpart to greater intensity of weathering in arriving a t a given stage.
Jackson et al. (1948) defined weathering stage in terms of specific
minerals associated with a given degree of weathering. It was observed
that ‘ ‘three to five minerals . . . are usually present in the colloid of any
one soil horizon . . . in the form of a distribution curve, dominated (40
to 60 per cent) by one o r two minerals with other minerals of the sequence decreasing in amounts with remoteness in the sequence. ” When
the percentage of minerals present at each stage was plotted against the
weathering stage, distribution curves were obtained for many soils ranging along the weathering stage axis. Additional examples of such distribution curves, representative of a Ferruginous Humic Latosol, are
shown in Fig 1. Kelley and Dore (1938) had previously noted that one
FIG.1. Mineral distribution curves and weathering mean ( m ) of the Naiwa soil
of the Ferruginous Humic Latosols (from Tamura, 1951).
colloidal mineral generally predominates in any one soil horizon but that
several other minerals were generally present in lesser amounts, thus in
a way anticipating the occurrence of weathering distribution curves.
Dominance of one mineral in a given tropical soil has also been noted.
Y. L. JACKSON AND 0. DONALD SHERYAN
c. Weatherirtg Mean. The results of chemical weathering are measured both by the residual minerals (residue mechanism) and the solids
formed from solution (synthesis mechanism). In soils unsaturated with
bases, the stage of weathering can be measured by the clay-size mi’nerals
alone, whereas in supersaturated soils the calcite and gypsum must also
be included in the calculation in addition to clay-size minerals (Hseung
and Jackson, 1952). A mean stage of chemical weathering, termed
“weathering mean” ( m ) , is calculated in which p is the percentage of
a mineral in a soil, and s is the weathering stage of that mineral. The
summation (2) is the simple addition of the various p X s values of all
the minerals to be considered on a given soil. The weathering mean m
has the dimensions of weathering stage (1-13) and is considerably more
sensitive to weathering changes than is the silica :sesquioxide ratio
(Hseung and Jackson, 1952; Jackson e t d.,1948). It is sensitive to
changes in layer silicate weathering stage wherein, as pointed out by
Marshall (1935), the silica : sesquioxide ratio is not. The weathering
mean is used as an index of weathering in different soils and in different
horizons of the same soil profile and as a measure of the geochemical
weathering of parent materials. Where the silt fraction has been weathered appreciably (especially in Latosols), the weathering mean can also
be applied effectively to that fraction.
AFFECTINQ THE REACTION
The reaction rates of chemical weathering are controlled by various
intensity and capacity factors operating as a function of time, and different combinations of intensities, capacities, and times of weathering
may produce a given degree or stage of weathering. The several factors
which affect the reaction rates of chemical weathering may be identified
also to a considerable extent with the five factors usually considered to
govern soil formation, namely, climate, biotic forces, relief, parent material, and time. But the evidence from research on weathering of soil
minerals in relation to the development of soil groups points to a vastly
different rate of action and difference in relative importance of the different factors in affecting the two phenomena, weathering and soil formation. To the extent that the two phenomena run along parallel
(particularly in the tropics), one is able to consider the two processes
CHEMICAL WEATHERING OF MINERALS IN SOILS
simultaneously ; however, when sharp divergences occur (particularly in
temperate regions), it is necessary to consider them separately.
For convenience, the factors which affect the rate of chemical weathering reactions are to be considered in more specific categories than the
five usually listed as controlling soil formation. F o r example, the climatic factor is considered in terms of temperature, separately, and of
rainfall. The effect of leaching is considered as a single intensity factor
whether it is controlled by amount of rainfall, distribution of rainfall,
rate of evaporation, relief, or internal drainage. The nature and extent
of leaching is all-important in the determination of chemical weathering
processes. Oxidation and reduction are considered specifically, whether
arising from relief, texture of the material, valence of the ions in the
minerals, or other factors.
1. Methods of Measurement of t h e Factors Affecting Rate of Chemical
The methods of discovery and measurement of the factors affecting
the rate of chemical weathering reactions are to some extent similar to
the methods of discovery of the factors affecting soil development,
mmely, geographic correlation, catenary correlation, particle-size function, and depth function. These four methods operate in a consistent
pattern, in as much as exposure to weathering factors of various kinds
varies in cliff erent geographic and catenary locations, with different specific surfaces of the material, and in different degrees of proximity to
the earth (soil) surface.
a. Geographic Correlation. Marbut (1951, p. 17) points out that the
primary tool for determination of the effect of different soil-forming factors controlling “the conversion of soil (parent) material into soil . . .
(is) geographic correlation.” For example, the effect of temperature or
rainfall is noted by comparison of maps of these factors to maps of soils.
Muckenhirn et al. (1949) similarly emphasize isolation of individual factors of soil formation by comparison of soils in different localities having
identical sets of factors of formation except for one factor under examination. It was proposed (Jackson et al., 1948) that the effects of intensity and capacity factors controlling chemical weathering reactions
can be assessed in a similar way by geographic correlation, and this idea
was supported by the consistent indications of the mineral weathering
sequence given by geographic correlation, by the particle-size function,
and by soil depth function. They state : “The mineralogical composition
of soil colloids follows the weathering sequence geographically, in accordance with the geographic distribution of climate, together with time of