Literature Review

All of the above studies were performed using relative pure systems. Their data are useful as they remove possible complications resulting from the interaction of different minerals and varying particle size but are not directly applicable to most soils and sediments. Soils and sediments are complex assemblages of minerals, organic material, living organisms, water, and gases. Size, shape, and chemical composition of soil minerals are highly variable soil properties and are among the principle controls of the chemical, physical, and biological behavior of soils. For that reason, observation of soil mineral morphology is very important in the study of soil behavior. Also, information obtained from mineral morphology may reveal the history of the soil in terms of new minerals formed and modifications of existing phases.

The application of SEM in the study of clay dispersibility can directly study of the effects of the dispersion process on the samples by examining the nondispersible fraction of the clay to determine the reason for the lack of dispersion. In general, the nondispersed clay may result from two processes. In the first process, soil aggregation is so strong that dispersion could not disaggregate the soil. In the second process, the soil was disaggregated but dispersion of the clay could not be maintained. The dispersibility of soils can also be examined using the presence and morphology of soil minerals to reveal the chemical character of the soil and aspects of the genesis of the soil.

The morphology of soil minerals results from interactions between several competing factors eg. physical and chemical conditions during formation, weathering history and solubility (Rai and Kittrick, 1989). The chemical composition of the solution along with the temperature and pressure of formation will determine the initial size and shape of an authigenic mineral. Many minerals persist in chemical systems far different from those required for formation because of their low solubility or slow rate of dissolution. Rutile is an example of a mineral which can persist for very long periods even when soil solution is undersaturated with respect to the mineral. Sometimes, however, evidence can be seen for the weathering of rutile and the formation of anatase, the form of TiO2 which is most stable at the surface. The presence of this evidence can be considered to be the result of extreme weathering over a long period of time. Once a mineral is formed, it will be weathered or undergo crystal growth depending upon the solution status of the sediments in which it is found. Some minerals may be relatively insensitive to the solution chemistry such as the aforementioned rutile but other minerals are very sensitive to soil chemistry. The differences in sensitivity to changes in soil chemistry by unrelated minerals have been used by soil genesis researchers for many years to evaluate the factors affecting soil formation. Such studies have resulted in several tabulations of the persistence of minerals in soils and sediments (e.g. Loughnan, 1969; Brewer, 1976). Mineral morphology can also reveal the sedimentary history of the minerals in a sample and chemical etching can help in examining the weathering intensity.

One of the earliest applications of the use of SEM in the study of minerals was the characterization of quartz sand grain morphologies (Krinsley and Doornkamp, 1973). The surface morphology of quartz has been intensively studied to determine the depositional history of the sediment and the degree of etching or secondary silica precipitation to determine the chemical history or relative age of a soil. Krinsley and Doornkamp (1973) concluded that the quartz surface features could be divided into four categories: 1) conchoidal fractures, 2) flat cleavage plates and their expression at grain margins, 3) upturned plates on cleavage or crystal faces, and 4) the alteration of the features above. By proper use of these features, they felt that the history of the quartz could be determined. The important environments according to their classification system are original source (residual rock type), diagenetic, glacial, littoral, glacial plus littoral, aeolian, and high energy chemical. The source effects are the relicts in grain morphology resulting from the conditions during quartz formation in rocks. Combinations of surface features resulting from sedimentary transport allow the detection of grains which have been in glacial, littoral, glacial plus littoral, and aeolian environments. The diagenetic and high energy chemical environments classification of Krinsley represent the physical features a quartz grain obtains from conditions in which there is a high concentration or a low concentration of solution silica.

The use of these features to distinguish sedimentary depositional environment is very controversial because grains may have relict features from a past depositional sequence or have been moved for such a short distance that depositional microfeatures may not have become expressed on most grains. There is also a problem in that the exact definition of an upturned plate is not clear. The lack of division between mechanical surface features formed during transport and chemical features derived after deposition has resulted in some confusion in the use of the Krinsley and Doornkamp (1973) classification system.

The environmental features useful in interpreting the geomorphic history and genesis of soils formed in stratified sediments in the Willamette Valley, Oregon were successfully explained in terms of the mechanisms proposed by Krinsley and Doornkamp (1973). Glassman and Kling (1980) found that differences in sediment provenance could be determined and the results used to establish lithologic discontinuities in the soils studied. The morphology of quartz grains was an aid to the recognition of soil parent material discontinuities in soils of the Edwards Plateau of Texas (Rabenhorst and Wilding, 1986). The degree of expression of the alteration of quartz surfaces useful in dating Quaternary soils by Douglas and Platt (1977). The amount of chemical surface alteration was determined to be proportional to the water holding capacity of the soil and to the age of the parent material. Residual soils were found to have the most severely altered quartz surfaces in agreement with Krinsley and Doornkamp (1973). Quartz surfaces were more altered by precipitation in the lower B and upper C horizons than in the A horizons in young soils and the differences decreased with soil age (Douglas and Platt, 1977). Quartz surface morphology was dependent on soil texture; more alteration occurred in soils containing more fine silt and clay. They hypothesized that the differences were due to the A horizons drying out faster than the B horizons or finer textured soils.

Later work on grain shape and environment by Mazzullo and colleagues (Mazzullo et al., 1986; Mazzullo and Magenheimer, 1987; Haines and Mazzullo, 1988; Pye and Mazzullo, 1994) cast doubt on the conclusions of Krinsley and Doornkamp (1973). Fourier analysis of 2-dimensional grain profiles showed that the shape of sedimentary quartz grains is controlled by source rather than method of deposition (Mazzullo and Magenheimer (1987). Mazzullo (personal communication, 1995) is quite critical of the methods of Krinsley and Doornkamp but examination of his research shows that his main criticism is that many people who try to use quartz morphology do not examine a sufficient number of grains. The experience of the first author with aberrant grains in these and other investigations support Mazzullo's concern about sample size. Some of the reasons that Krinsley and Doornkamp (1973) consider grain shape to be important while Mazzullo discounts this feature is that Krinsley and Doornkamp examined the total surface features of a grain to judge the roundness while Mazzullo used only the roundness of the grains as measured by shape profiles from binary visual images. Depending upon variability in grain surface features, several hundred grains may need to be observed to properly use the method of Krinsley and Doornkamp (1973).

The effects of weathering on a mineral surface may help in the identification of a mineral. Feldspar has very distinctive structurally controlled weathering patterns (Berner and Holdren, 1979). Many K feldspars are perthitic with microscopic intergrowths of Na feldspar. Likewise, the members of the plagioclase series intermediate between albite and anorthite are often demixed into Na rich (albite) and Ca rich (anorthite) domains with regions separated by crystallo-graphic boundaries and sizes recognizable by light microscopy. Weathering of these feldspars is almost always incongruent with one type of the feldspar weathered away almost completely leaving the other feldspar appearing pitted but with an EDS pattern characteristic of the end member composition. Incongruent weathering also is common for many other minerals (Nahon and Colin, 1982).

Authigenic phyllosilicate morphology has been the source of intense study by the oil industry because of its effects on oil reservoir properties and potential as a geothermometer. The most complete study of the SEM observation of authigenic clay is that of Wilson and Pittman (1977) which has been used as a standard reference for almost all later work on the subject. They found as a result of study of many samples using SEM, EDS, and X-ray diffraction that each major authigenic clay group exhibited a limited number of independent morphologies (pseudohexagonal plates for kaolinite, and laths for illite, for example). Other common minerals have a three dimensional morphology which does not give much aid in their identification. Minerals such as quartz and some feldspars potentially have mineral morphologies which are distinctive enough to allow for their identification but their shape is altered by weathering, transport, or other factors. In the latter cases, there may be other clues which will aid in their identification. How the mineral is broken up by weathering or transport can be used to aid in mineral identification. Some minerals such as feldspars and many carbonates have good cleavage in one or more directions. Cleavage produces flat surfaces at set angles which can help in mineral identification. Other minerals such as quartz have extremely poor or no cleavage. Lack of cleavage is a distinctive aid to identification as it can lead to a the pattern for breakage such as the conchoidal fracture of quartz.

When using EDS spectra to identify minerals, surface coatings and weathering can be misleading. Iron oxide coatings or secondary silica coatings will cause significant differences in EDS patterns from underlying minerals and should be taken in consideration when viewing grains with surfaces which do not appear fresh. It is also important to remember that many minerals do not weather congruently. That is, the EDS spectra obtained may be very much different from the ideal spectrum due to preferential leaching of one or more elements from the mineral surface. Feldspars (Berner and Holdren, 1979) and pyroxenes (Nahon and Colin, 1982) are two common minerals which weather incongruently. Each of these factors was important as the samples had undergone no pretreatment other than particle size fractionation using pure water.