Several preparation techniques have been used widely for the microscopic examination of coal. Thin sections are generally preferred by investigators in the United States since their use makes possible the more certain identification of the various macerals and phyterals. However, specimens are sometimes prepared by polishing and by maceration. Polished section techniques have been described by Winter (1923), Stach (1928), Duparque (1933), and Roos (1937). The method involves the production of a plane, highly polished surface on the coal which may be modified by relief polishing or etching. The surface is then studied with a reflecting microscope. McCartney (1949) has suggested a refinement of the method for use with the electron microscope. The maceration technique was introduced by Schuize (1855) and has been discussed by von Gumbel (1883), White and Thiessen (1914, pp. 216-218), Schopf (1938), and others. It involves oxidizing the humic portions of coal and leaching them with alkaline solutions so as to leave the less soluble, translucent portions for microscopic examination.
Thin Section Technique
The method of preparing a representative sequence of thin sections from a column sample of coal was developed at the Bureau of Mines by Thiessen, Sprunk, and O'Donnell (1938) for use in the quantitative microscopic determination of petrographic components. The technique used in this study was the same except for minor variations. Other petrographic work has demonstrated that optimum analytical results are obtained when the coal is sampled in such a manner as to preserve the stratification of the coal. The ideal sample is an unbroken column of coal, about 12 inches square in cross-section, cut perpendicularly to the bedding plane and including all the coal from the top to the bottom of the bed. Since petrographic studies are usually made in conjunction with other tests, a channel sample may be taken immediately adjacent to the column sample. In the laboratory, a subcolumn approximately 3 inches wide and 3 inches deep is cut with a 12 x 0.0625 inch resinoid-bonded Crystolen cut-off wheel (standard designation C46-P-8B) operating dry at about 12,000 surface feet per minute. The subcolumn is then mounted in plaster of paris and cut into two parts normal to the bedding. One part of the subcolumn is reserved for polishing and the other for thin sectioning.
Mention should be made of the necessity of eliminating the coal dust resulting from dry cutting. A special saw was designed for the purpose. The cut-off wheel was mounted on a 5/8-inch belt-driven mandrel so as to project through a slotted table top. The lower part of the wheel was enclosed and connected by flexible tubing to the intake end of a motor-driven blower and the blower in turn connected to a 30 x 30 x 16 inch galvanized steel tank equipped with a series of baffles and fiberglass filter. A plexiglass shield one-half inch thick provided protection from the exposed part of the wheel.
Preliminary polishing of one mounted subcolumn is done on a 3 x 3 foot piece of plate glass using successively finer carborundum powder sludges. The final polish is achieved by stroking the column with a fine-grained, yellow Belgian hone and then buffing it with a Selvyt cloth and a paste of Lakeside polishing compound No. 27. After a thorough drying, the polished surface is painted with a 5 percent solution of water-soluble polyvinyl alcohol (Dupont Elvanol 51-05) to prevent surface oxidation. The polished columns thus serve as permanent sample specimens and can also be used for macroscopic studies.
Thin sections are prepared from the remaining part of the subcolumn. Beginning at the top and continuing to the bottom, parallel lines spaced about 0.8 inch apart are marked on the coal with a red wax pencil and small numbered blocks are cut with an 8 x 0.0312 inch cut-off wheel (standard designation C80-P-5B) using the lines as a guide. Each block is trimmed to a length of about 1 inch.
One face of each block is then ground on rotating laps with successively finer abrasives, polished on a yellow Belgian hone, and buffed on a mounted Selvyt cloth with Lakeside polishing compound No. 27. After thorough drying in an oven and proper orientation, the block is cemented to a frosted standard glass slide with Lakeside No. 70 cement. Frosting of the slides has proved to be especially beneficial in securing good sections as suggested by Gibbs and Evans (1950, p. 2). The excess coal is next sawed off with the 8-inch cut-off wheel leaving about one-fourth inch of coal on the slide.
The mounted blocks are then ground to a thickness of about 10 microns. This is done by using successively finer carborundum sludges on rotating laps until the section just begins to transmit light. This is a somewhat critical point and can be determined only by experience. Too close grinding on the lap will ruin the section.
After trimming the superfluous cement from around the edges of the section with a razor blade, the section is transferred to the yellow Belgian hone. The hone is mounted in a wooden block sloping away from the worker and a stream of water is played on it. The section is honed with forward strokes and slight pressure until it transmits light uniformly and shows an orange to red color. The section is then transferred to a light box and gently rubbed with a cork dipped in Lakeside polishing compound No. 27 to produce an even section. The section is finally labeled and painted with polyvinyl alcohol.
In previous work, each section had been ground to completion by a single operator. It was found that this procedure does not lend itself well to the production of a large number of sections. Consequently, the sections were ground in column lots with one operator doing the coarse grinding on the laps and a second the work on the hone and light box. However, in the interval between the two stages of preparation, the section frequently dried out, oxidized, and contracted so that a useless section resulted. On the suggestion of B. C. Parks of the Bureau of Mines (personal communication) the slides were immersed in glycerine except when being improved.
In 1930, Thiessen set up a method of microscopic analysis and a type classification of coal that has been generally followed as standard practice by the coal petrography laboratory of the Bureau of Mines. It has subsequently undergone such changes that Parks and O'Donnell (1948) have described the modified procedure. In the discussion which follows, this procedure is described, modified, and evaluated.
The sequence of thin sections, prepared in the manner described earlier, is essentially a disconnected, transparent, ribbon-like sample of coal about 10 microns thick and 1 inch wide, representing the entire coal bed. The ribbon sample of sections is not, however, a complete representation because of some loss from sawing and polishing. A recovery of 90 to 95 percent of the height of the original column is considered not unusual by the Bureau of Mines.
The ribbon transect method of statistically evaluating the relative amounts of anthraxylon, opaque attritus, translucent attritus, and fusain in a coal bed is nothing more than refined visual estimate employing the principle of the Rosiwal analysis (Head and others, 1932). It is based on the assumption that the sum of the areas of each of the components in a random section of uniform rock is proportional to the volume of that constituent in the rock. In practice, actual areas are seldom measured, but rather a linear traverse is made. Coal is inherently heterogeneous so that the assumption of uniformity is not valid. Nevertheless, it is probably valid to assume that each increment of the coal parallel to the bedding plane will exhibit homogeneity over a limited area. Thin sections cut normal to the bedding should therefore represent statistically the coal in the area and Rosiwal analysis may be used if its limitations are appreciated.
Any type of microscope equipped with a mechanical stage is satisfactory for the measurements. Although the Bureau of Mines prefers the binocular type, a petrographic microscope was used in this work since it facilitated identification of mineral components of the coal. A grid micrometer disc on which is centered a 10 mm square field divided into 100 1.0 x 1.0 mm constituent squares is inserted in the ocular. Each square is further subdivided into four subsquares 0.5 x 0.5 mm on a side. The values represent the real dimensions of the squares scribed on the disc. Since only one central vertical tier of squares is used, each transect field is 10 mm long by 1.0 mm wide and consists of 10 major 1.0 x 1.0 mm constituent squares and 20 1.0 x 0.5 mm subsquares. Because it is pertinent to later discussion, mention is made that the Bureau of Mines employs a Whipple disc, which is a 7 mm square field divided into 100 0.7 mm squares with the central one subdivided into 25 0.14 mm squares. Usage is largely a matter of convenience.
Under microscopic magnification, the dimensions change; therefore it is necessary to calibrate the disc with a stage micrometer for different powers of magnification as shown in Table 4.
Table 4--Calibrated values of grid micrometer with petrographic microscope B&L LM5919
It has been the practice of the Bureau of Mines to measure opaque attritus and fusain at a magnification of 60X and anthraxylon at 150X. Translucent attritus is determined by difference. Since it seems doubtful that any unique advantage is gained by changing magnifications, all components have been determined at 100X in this study.
Measurements are made by turning the mechanical stage of the microscope so as to move the central vertical transect field of the grid micrometer along a line extending from the bottom to the top of the section. Two such traverses are made for a better statistical average; one near the middle of the right half of the section and the other near the middle of the left half. The traverse is not made continuously; the thin section is moved a distance equal to the length of the transect field and when the number of constituent squares and subsquares occupied by each of the coal components has been estimated and tabulated, the section is again moved the length of the transect field. These field moves are continued until the section is crossed. A horizontal shift of the stage brings the section into position for the second traverse. The total number of transect fields is tabulated and checked against the total distance traversed as determined with the mechanical stage vernier by multiplying the number of transect fields by the calibrated length of each transect field. A sample data sheet and an illustration of the use of the grid micrometer are shown in Figure 3.
Figure 3--Sample data sheet and illustration of microscopic measurement of coal components.
After all the thin sections representing the column sample have been measured in this manner, the data for the components of each slide are converted to percentages using the relation:
The percentage distribution of anthraxylon, translucent attritus, opaque attritus, and fusain in each thin section is tabulated graphically by bar diagram as illustrated in Plate 1. On the basis of this distribution, the coal is classified according to type and a profile of the coal results. Due to loss in cutting and grinding, the tabulated distances and lengths on the bar diagrams are in error by the amount of loss.
Although the Bureau of Mines has published the results of numerous petrographic studies of coal, actual determinative procedures have always been omitted from the reports. Considerable uncertainty had existed regarding the validity of the method until Parks and O'Donnell (1948) fully described procedures and evaluated the influence of such factors as microscopic magnification, thin section coverage, and errors arising from the personal element. However, several important considerations were overlooked in the paper by Parks and O'Donnell. Personal communication with the authors and reference to a discussion of the paper by Schopf (1948) have contributed to the following examination of critical limits.
Subsize thresholds--Parks and O'Donnell (1948, p. 536) state:
No particular difficulty is experienced in recognizing attritus in thin sections under the microscope. The heterogeneous mixture of ingredients of different shape, structure, translucence, and color occur in layers that are easy to distinguish from the other banded components.
They further say:
Anthraxylon can also be easily recognized in a thin section when seen with transmitted light under the microscope. It is present in prominent orange bands, sometimes shaded toward brown to red, and usually shows well-preserved cellular structures of a woody tissue seen in cross-sectional or longitudinal view.
These statements are substantially true when the anthraxylon bands are wide and the attritus extremely fine. However, difficulty is experienced in deciding whether certain fairly fine translucent components of attritus shall be classified with the atrital or the anthraxylous material. In other words, at what point does the translucent component cease being preserved cellular tissues of stems, branches, twigs, etc., and become part of a mixture of finely divided plant debris. Realizing the question was largely one of size, Schopf discovered that according to standard practice of the Bureau of Mines laboratory, anthraxylon is not identified in any particles or strands thinner than 0.014 mm. This subsize threshold was chosen empirically because the subsquares of the calibrated Whipple disc were determined to be about 0.014 mm at a magnification of approximately 150X--the magnification used to measure anthraxylon.
It seems entirely possible that the use of this arbitrary limit may constitute a source of error which has been overlooked. It was earlier stated that all determinations for components other than anthraxylon were made at a magnification of 60X. At this magnification, the subsquares are no longer 0.014 mm but 0.037 mm. Hence, in traversing attrital material, it would be possible to miss strands of anthraxylon 0.014 mm wide.
Schopf also gives the subsize threshold for the microscopic determination of fusain as 0.037 mm--the size of a subsquare at a magnification of 60X. All smaller opaque material is assigned to opaque attritus.
In view of the large accumulation of data using these limiting values, it seems necessary for comparison purposes that they be tentatively accepted as part of the definition of anthraxylon and fusain in quantitative work. However, since they are only visual estimates, certain liberties may be taken for the sake of convenience. Anthraxylon is here defined as any translucent strand larger than one-fourth of a subsquare at a magnification of 100X. A subsquare at this magnification is 0.067 mm high; therefore one-fourth of a subsquare is 0.017 mm. as compared with the Bureau of Mines value of 0.014 mm. Similarly, the subsize threshold for fusain is given as one-half of a subsquare at a magnification of 100X. This is 0.033 mm as compared with the Bureau of Mines value of 0.037 mm.
The problem of establishing limiting subsize thresholds for certain of the petrographic components is not unique to the Bureau of Mines and should be subjected to closer scrutiny since it stems from some very fundamental considerations. Cady (1942, pp. 343-346), in discussing a parallel situation involving the Slopes classification, pointed out that uncertainty had developed concerning the application of the term "vitrain" to vitrain-like material of small dimension which may make up a considerable portion of a clarain band. He also concluded that the distinction was one of size and that limiting values were necessary. In an attempt to resolve the situation, he suggested that all the thin vitrain-like bands composing clarain be called "micro-vitrain" and set the lower limiting value for vitrain at 2 mm with a tolerance of 1 mm. Justification for this is based on the fact that, in the natural breakage of coal, the thicker vitrain bands tend to break away from the rest of the coal and concentrate in the small screen sizes whereas the finer bands tend to remain intimately associated with the clarain which concentrates in the larger sizes. The limiting value is thus a function of the physical properties of vitrain and clarain since vitrain is characteristically friable and clarain is not. This approach is not necessarily the complete answer but is at least suggestive that limits should be established on the basis of physical or chemical behavior.
Color, thickness, and opacity--Another important point which has not received sufficient consideration is the question of color and light transmission in connection with quantitative analytical work. In the description of each of the petrographic components, reference was made to its color or opacity. Essentially then, color comparison is an important basis for identification and can be appreciated only by direct examination since most photographs of thin sections are not reproduced in color. However, accurate comparison precludes that all sections be ground to the same thickness since light transmission is partially a function of the thickness of the section. This factor is of relatively minor importance in the identification of anthraxylon and translucent attritus but becomes extremely important for opaque attritus and fusain where identification is based largely on opacity. These components cannot, however, be regarded as totally opaque but only as possessing varying degrees of opacity since all of them probably can be made to transmit light if cut thin enough. There is, then, a very real problem for the petrographer who, for example, attempts to classify a dark-brown attrital material as opaque or translucent attritus when it is almost impossible to achieve uniformity in thickness of the section. He is beset by the same problem when he tries to classify opaque attritus and fusain since there is ample evidence that they, too, are gradational. The question involves not only the establishment of criteria for opacity but also a reconsideration of the fundamental basis of the classification. It is now apparent that opacity is an attribute which may be acquired by all kinds of plant materials to a varying degree and is dependent upon the activity and duration of the process which produced it. Study of numerous thin sections indicates that much so-called opaque attritus is actually crushed fusinized material and that cellular fusain maintains its open cell structure only because the spaces have been filled with mineral matter at an early stage in the process of coalification. This problem will be considered further in the discussion of the petrography of Kansas coals.
Kansas Geological Survey, Geology
Placed on web November 2005; originally published May 1953.
Comments to email@example.com
The URL for this page is http://www.kgs.ku.edu/Publications/Bulletins/102_1/04_prep.html