THE TEXTURAL EVOLHTION OF A E’ROGKESSWELY MLOMITIZED LIMESTONE 400 I01: COLD—x Thesis for fine Degree of M. S. MECHIGAN STATE UNIVERSITY Jeanne L. Hewitt 197 5 ’" l\\\\\\\\\\\\\\\\\\\\\\\\ M \\\\\\\\\\\\\\\\\\\\\\l “' 3 1293 10406 6893 h.-- - LIEEARY Michigan State University This is to certify that the thesis entitled Textural Evolution of a Fine-Grained Lim stone presented y Jeanne L. Hewitt Department of Geology has been accepted towards fulfillment of the requirements for 4.5.,— degree in MEX— ; Major professor % Date—M ' 7L 0-7639 ABSTRACT THE TEXTURAL EVOLUTION OF A PROGRESSIVELY DOLOMITIZED LIMESTONE By Jeanne L. Hewitt Many reactions that produce textural changes—-such as replacement and grain growth--are initiated at grain boun- daries. Textural change can be studied quantitatively using a statistical technique that measures surface area of grains of individual phases in a unit volume of rock. The Platteville Formation in Wisconsin was chosen to obtain samples exhibiting all stages in the evolution of a progressively dolomitized limestone. Surface area data and modal analyses show that dolomite nucleates randomly throughout the rock volume and grows as well-formed rhombohedra replacing the micrttic matrix until the rock is approximately 70% dolomite. This produces a regular decrease in total rock surface area as lime mud is replaced by dolomite grains. Further dolomitization results in the replacement of the remaining matrix calcite and pro— duces a non-equilibrium texture characterized by high—angle boundaries between adjacent dolomite rhombs. In the final stage of dolomitization, the grains have developed smooth, rounded boundaries with no distinct rhombohedral outline, uniformly stable size and show no microscopically observable porosity. Studies in solid-state reactions involving silicate rocks have shown the importance of surface energy in producing equilibrium textures; it may be that this is also an important kinetic variable in the textural evolution of carbonate rocks. THE TEXTURAL EVOLUTION OF A PROGRESSIVELY DOLOMITIZED LIMESTONE By Jeanne L. Hewitt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE College of Natural Science 1975 Dedicated to all the people who helped me, but especially Tom Vogel, Gary Byerly and most of all my husband, Larry Dukerich. ACKNOWLEDGMENTS I gratefully acknowledge the assistance of Mr. West of the United States Geologic Survey for the patience and guidance he gave in the field. I would like to thank my advisor, Tom Vogel, who helped me when I needed it and yelled at me when I needed that too. I especially thank Gary Byerly, Lee and Jean Younker, Steve Tilmann and Vivian Bust for the long talks about science and life in general. My committee members, Bob Anstey, Bill Cambray and Duncan Sibley deserve special thanks for all the times I changed the date of the thesis defence on them. Duncan I thahk for the many long arguments about the words printed in this thesis. And most of all I wish to thank my husband, Larry Dukerich and my parents for accepting me and understanding me while I worked on this. y vi LIST OF FIGURES ..... TABLE OF CONTENTS INTRODUCTIONS ...... .... ... .. . ..... .. ........ .. PREVIOUS WORK AND GENERAL OBSERVATIONS ..... . . .. . DATA .......... . . ,. .................... . . .. . DISCUSSION ......... ... . . .................... .. ..... REFERENCES ......... . .......... . ................... vii viii 15 22 LIST OF FIGURES Figure Page 1. Hand sample approximately 5 inches long showing mottled areas of dolomitization. Area A is dolomite; Area B is still pre— dominantly calcite. A layer containing crinoid stems and fossil debris is also shown at the margins of two dolomitic patches............................. .............. 6 2. Photomicrograph showing rounded grain boundaries of Platteville dolomite...... .......... l9 3. Total Rock Surface Area vs. Percent Dolomite with approximate regression curve drawn in ........ ll 4. Dolomite-Dolomite Surface Area vs. Percent Dolomite with approximate regression line.. ....... 12 5. Calcite-Dolomite Surface Area vs. Percent Dolomite with approximate regression line.. ....... l3 6. Calcite-Calcite Surface Area vs. Percent Calcite with approximate regression line ....... ... 14 viii INTRODUCTION In the past many studies have been done pertaining to rock compositions and the record of chemical reactions found in sedi- mentary rocks. Little attempt has been made toward quantifying and correlating concomitant textural change with progressive chem- ical reactions. These changes in textural parameters may poten- tially record details of the chemical processes as well as yield different kinds of information pertaining to reaction pathways. In this paper, a textural parameter—-surface area-—is examined in the role it plays during progressive dolomitization of a fine- grained limestone. Most reactions such as nucleation and grain growth begin at the surfaces of grains of the individual phases (Byerly and Vogel, 1973). By observing the responses associated with grain surfaces--as measured by surface area changes in calcite, calcite- dolomite contacts and dolomite, it is possible to monitor the reactionpathways of the dolomitization process. By considering the rock as a three-dimensional network of grain boundaries the changes in the number and types of grain boundaries is a variable that may be as important as chemical variables. If these changes in texture are compared with the amount of dolomite in the rock, 1 the overall textural response of a fine-grained limestone under- going dolomitization can be examined. The changes in chemical composition can be monitored by standard petrographic techniques such as point counts and the changes in surface area can be easily quantified by means of a technique developed for metallurgists. Such a method could help explain the histories of two rocks with similar chemical composition but different textures in terms of processes which overall chemistry of the system cannot reflect (Ehrlich, Vogel and others, 1972). Data for this study are obtained from a progressively dolomitized lithology in the Platteville Formation in southwestern Wisconsin and are expressed as total surface area of the rock and surface area of phase-to-phase contacts. PREVIOUS WORK AND GENERAL OBSERVATIONS While there have been no previous quantitative textural studies of the response to dolomitization, there have been a number of important observations. Many other workers, among them Bergenback and Terriere, 1953, Parker, 1956, and Edie, 1958, have noted that where dolomitization is incomplete, and constitutes less than 50% of the rock volume, it selectively replaces fine-grained lime mud rather than well-sorted and more permeable carbonate sands or clastics. HighAmagnesium fossils are also selectively dolomitized before other fossils and fossil fragments. It appears that the more highly reactive surface area of the small calcite particles (or fragments already containing a considerable amount of magnesium) are more easily dolomitized than the larger fossils and fossil fragments which are relatively resistant to replacement(Murray, 1960). Where replacement has begun, the original grain size of the calcite in the matrix is unchanged during the dolomitization process. The original tex- tures are usually preserved in the Platteville dolomites. Once nucleation has begun, dolomite grows as euhedral rhombs scattered throughout the lime matrix with no apparent orientation. Other workers have reported that the dolomite rhombs seem to "float" in the matrix and grow to become several 3 orders of magnitude larger than the neighboring particles in lime mud (summarized by Murray, 1960). Hohlt (1948) published data that demonstrated that dolomite has no preferred crys- tallographic orientation even though it has grown in limeStone with strongly preferred orientation and concluded that dolomite is unaffected by orientation of the limestone it is replacing. As the dolomite content increases, the number though not necessarily the size of the dolomite rhombs increases. Other researchers have noted that there appears to be a remarkably homo- geneous grain size in dolomitized rocks, with most grains falling between five and twenty microns in size (Murray,1960). Had the first-formed rhombohedra grown at the same rapid rate as later rhombs, the result would be a true porphroblastic texture with a few large grains, a range of intermediate-sized grains, and some small late rhombs filling in the remaining space. This does not happen and the resulting texture is composed of equal-sized dolomite crystals in random orientations. The lime— stone matrix still left at this point is "walled off" in the interstices between the euhedral dolomite rhombs, leaving only small patches showing the original texture. This texture can be extremely porous and resembles the sucrosic textures associated with typical dolomites. Porosity at this point increases dram- atically as shown by a number of studies summarized by Murray, 1960, if the remaining interstitial calcite is leached away. The Platteville Formation is middle Ordovician and re- presents a spectrum of lithologies from limestone to dolostone. It is Overlain by the Decorah Formation (shaly carbonates) and underlain by the St. Peter's Formation (sandstone). Dolomi— tization increases progressively as the Wisconsin Arch is ap- proached from the west (Badiozamani, 1973). The limestones are fine—grained, unrecrystallized, often contain well-preserved fossils and are generally low in quartz and other impurities. In this formation dolomitization occurred after lithification by the replacement of limestone when magnesium-bearing fluids passed through the rocks (Deininger, 1963). Consistent with this model is the observation that dolomitization begins in areas having higher initial porosity—~along joints, fractures and par- ticularly bedding planes (Deininger, 1963). Because this process did not continue and completely dolomitize all the rock in some areas, patchy dolomitized zones can be found coexisting on the hand sample level with zones of unaltered limestone. Others have also called attention to the unusual mottling pattern in this formation (Hohlt, 1948) and suggested that this may predate joint formation. The borders between these patches are quite distinct and occur abruptly over a few millimeters (See Fig. 1). In other areas, higher on the Arch, dolomite has replaced all the limestone in an interlocking fabric of roughly equal—sized dolomite crystals which have enlarged to occupy all available space. Because it contains a wide range of structures and compo- Hand sample approximately 5 inchesilong showing mottled areas of dolomitization. Area A is dolomite; Area B is still pre- dominantly calcite. A layer containing crinoid stems and fossil debris is also shown at the margins of two dolomite patches. sitions, the Platteville Formation was selected in order to study the textural relationships as measured by the surface area of the whole rock and phase-to—phase contacts during the dolomi- tization process. The Platteville Formation extends across southwestern Wisconsin and over the Wisconsin Arch. Numerous outcrops and quarry sites make this an easy formation to sample extensively. A suite of rocks was selected from thirteen sites in an attempt to represent the complete spectrum of dolomitization. While no attempt was made to precisely locate sampling sites either geo- graphically or stratigraphically, care was taken to sample frmm equivalent units (i.e. Magnolia member) in both the undolomié tized and dolomitized regions of this formation. It was hoped that this would represent equivalent primary lithologies that had been subjected to different dolomitizing environments. A few samples were taken from the Galena Formation, a hydrothermally- altered dolomite for the purposes of comparison with another dolomitic texture. Thin sections were prepareci representing the range of dolomitization observed within all samples from all thirteen sites. Because of the range of variation present, small areas of minimal variance representing equilibrium textures and compositions were selected for petrographic analysis. These homogeneous structural areas (Ehrlich, 1964) were con- sidered necessary since the amount of information obtained from a series of observations is inversely proportional to the variance, and a sampling plan that produces minimal variance is optimal. These homogeneous structural areas were used in the analysis of surface area, partitioned components and volumetric percentages. Surface area determinations were made using the line transect method, as modified from Kendall and Moran (1963), by counting all grain boundaries between both like and unlike phases along a known distance. The formula for converting the number of grain boundary intersections to phase rock surface area per unit volume is: Surface Area = 2 N L Number of boundaries crossed 2 II where: I." ll Length of traverse When the total rock surface area is to be determined the formula becomes: Surface Area = 4 N The individual phase—to-phase boundaries (e.g. calcite-calcite, calcite—dolomite, and dolomite-dolomite) can be tabulated during data collection and individual phase-to-phase surface area can be determined by the same method. Modal percent composition was determined by point counts, tabulating at least 700 points for each homogeneous structural area. DATA The data is presented as surface area of either total rock or phase-to-phase contacts. Prior to dolomitization, most samples are mudstones to wackestones and have total 3 3 rock surface areas ranging from 331 mm2/mm to 444 mmz/mm with a mean value of 372 mmzlmm3 (See Fig. 3). Calcite- calcite surface area is equal to the total rock surface area in the undolomitized units (See Fig. 6). Initially the measurements were made excluding large fossil fragments from the low-dolomite end of the spectrum, since they represented a special type of surface area configuration. Because it appeared that large stable fragments were relatively resis- tant to the dolomitization process (and also tended to lose their distinct shapes), they were considered as the rock be- came more dolomitic. The range of values from 331 mmZ/mm3 to 444 mmz/mm3 thus represents a maximum surface area range for the fine-grained micrite. Some decrease in that would be expected as fossil content of the rock increased, but probably would not affect the range of values greatly. In the early stages of dolomitization (i.e. rocks with less than approximatley 40% dolomite) dolomite rhombs are scattered throughout the limestone matrix. In these samples, 9 10 the total rock surface area (Fig. 3) decreases because large dolomite rhombs with low surface area to unit volume ratios are replacing fine-grained limestone. The plot of total surface area versus percent dolomite (Fig. 3) shows that this trend is relatively constant until approximately 60% to 70% of the rock is dolomitized. Calcite-calcite surface area steadily decreases over this range as more and more calcite grains are being replaced by dolomite rhombs (Fig. 6). Calcite-dolomite surface area (Fig. 5) increases steadily and reaches a maximum at approximately 50% to 70%. 173; 358- 401. 443. 315. WT. I44- IOI. 583 ‘Er‘ 'C’F 11 TOTAL ROCK SURFACE AREA vs PERCENT DOLOMITE _ I 40.0 . . . 90.0 lob. Total Rock Surface Area vs. Percent Dolomite with approximate regression curve drawn in. \4-4 Ie.o IO.8 I. — ...- .— 12 DOLOMITE-DOLOMITE SURFACE ‘ .. AREA vs PERCENT. DOLOMITE‘ Fig. 4. Dolomite—Dolomite Surface Area vs. Percent Dolomite with approximate regression line. 13 A CAiCiTE-DOLOMITE :- ‘FACE AREA vs * PERCENT DOLOMITE 5-87 7-33 8-80 lO-3 F7 l 4.40 153 Fig. 5. Calcite-Dolomite Surface Area vs. Percent Dolomite with approximate regression lines. "Inf? nnn3 ‘7- | . l‘ .. (1‘1 m n) 14 CALCITE-CALCTTE SURFACE AREA vs PERCENT CALCITE. . 6. Calcite-Calcite Surface Area vs. Percent Calcite with approximate regression lines. 3 A___. ’D r DISCUSSION As the dolomite grains nucleate and grow, several forces affect their growth and development. Some of these, like chemical potential and diffusionirates oftthe ionic species involved, determine if dolomitization will take place at all. Assuming all the conditions required for dolomiti- zation can be met, then the energy associated with grain boundaries can play a large role in the dolomitization process. Parameters like surface area of the growing grains and distribution of the phases within the rock may be determined by such factors as surface energy, original grain size and the space-filling effects associated with crystal growth (DeVore, 1959). Previous workers have documented that dolomite nucleates randomly throughout the rock volume and selec- tively replaces fine-grained mud components first. None have considered the rapidly increasing surface area of the dolomite rhomb and the effect of neighborhood on its growth and devel- opment. Does the neighborhood that the growing grain "sees" have an effect and at what stage in its growth is this most critical? These points will be considered during the sequence of dolomitization. Once a number of dolomite nuclei have developed, there are two ways in which the net amount of dolomite can be 15 16 increased in the rock: either 1) enlarging the existing dolomite grains in a kind of porphroblastic growth or 2) in- creasing the number of nucleation sites and forming a large number of small grains. The data shown in Fig. 4 supports the observation also noted by others that few dolomite grains are growing so close as to impinge upon one another. This is expected if the dolomitization does take place randomly throughout the mud matrix. If this were not true, there would be a much greater increase in the dolomiteidolomite surface area over this range than the 0 mmZ/mm3 to 7 mm2/mm3 increase shown. By random chance, however, a few grains are likely to grow into contact and this is probably responsible for the slight increase. The implication here is very strong: dolomite grains do not provide host sites for nucleation and growth of other dolomite grains and that dolomite-dolomite contacts may inhibit further growth of dolomite. If doloa mitization is to continue at this stage, it is easier to re— place a calcitic mud. These dolomite grains resemble por- phroblasts in that they often become several orders of mag- nitude larger than the surrounding calcite grains and con- tain inclusions and zones of trace metal enrichment, both typical characteristics of metamorphic porphroblasts. Calcite-dolomite surface area (Fig. 5) increases steadily and reaches a maximum between 50% and 70% dolomite. In this range enough dolomite has been formed so that a maximum number of calcite-dolomite contacts have formed. 17 Space-filling models imply that this point of maximum surface area between two phases would be skewed toward the larger— sized phase of two unlike phases. If there is a large dis- parity in the size of one phase relative to the other as there is in the fine-grained calcite and large-grained dolomite, a large number of grains could fit around a dolomite rhomb. For this reason, the point of maximum interpenetration of calcite and dolomite as measured by surface area occurs when more than 50% of the rock is dolomitized. When roughly 60% of the rock has been replaced by dolomite, dolomite grains have grown to occupy most of the free matrix and begin to impinge upon one another. Dolomite— dolomite surface area begins to increase dramatically (Fig. 4) since small increases in the total amount of dolomite now produce large increases in the contact area between adjacent dolomite grains. The remaining matrix phase is contained in the interstices between the rhombs of dolomite. The calcite- calcite surface area (Fig 6) drops rapidly to a very low value as more and more limestone is "walled off" by dol- omite rhombs. Some of this decrease in calcite—calcite surface area is due to replacement by dolomite, but recrys- tallization and grain growth which have been occurring in the limestone are also contributing causes. From observations made while the data was being collected it was noticed that fossils, which can be considered as large, single crystals of calcite, were most stable with respect to dolomitization. When more than 60% of the rock has been dolomitized, it can 18 no longer be determined whether the last remaining calcite is recrystallized matrix limestone or a fossil fragment that has lost most of its original detail. These data do not indicate which mechanism is most important in decreasing calcite-calcite surface area or if the recrystallization is caused as a result of or response to dolomitizing conditions. Total rock surface area (Fig. 3) now assumes a value char- acteristic of a dolostone, with some scatter associated with the last remaining calcite fragments. During the final stages of dolomitization the dolomite crystals enlarge their boundaries to fill in the remaining voids or replace the last traces of limestone. The surrounding grains enlarge their boundaries by growth normal to the rhom- bohedral crystal faces until the void is filled. The grains appear to "bulge" their surfaces into the last remaining space and thus lose their rhombohedral shape. The final result is a rock that is 100% dolomite with tightly-packed, rounded and smooth grains that show little of their original rhombohedral character (See Fig. 2), and at the last stage of the process the total rock surface area is equal to the dolomite‘dolomite surface area. Other writers have commented upon dolomites with smooth grain boundaries and essentially no pore space but none have speculated upon the conditions necessary to effect this. What conditions would produce this type of dolomite texture? Pressure solution can produce carbonate rocks l9 Photomicrograph showing rOunded grain boundaries of Platteville dolomite. 20 having smooth rounded boundaries. This may be unlikely for several reasons. There is excellent preservation in many hand samples showing bedding, burrowing and relict fossils. There are no highly sutured boundaries characteristic of rapid rates of pressure solution. These rocks were collected from the flanks of the Wisconsin ARch and thus subject to little load due to burial. Probably some pressure solution has taken place but it is unlikely to have produced such a wide— spread, uniform texture. Adjustments at grain boundaries can produce equilibrium textures such as this. High angle boun- daries, such as a point of a rhombohedron, would be more readily destroyed than the face of a rhomb. Whether these adjustments occur by pressure solution or by migration of grain boundaries, the same equilibrium texture would result. If equilibrium textures are produced by grain boundary adjustments, the nature of the surfaces of neighboring grains (surface energy) would be an important controlling force. Surface energy arises because the environment around mole- cules and atoms on the surface of a grain is different from the environment around molecules and atoms in the center of the grain. While the physical separation between grains may be quite small, the forces associated with surface energy cause strain effects which are felt for tens of atomic dis- tances inside the surfaces of both grains, perhaps up to 50 microns, (Westbrook, 1967). Surface energy is influenced by composition and structure of the solid surface, orientation of 21 the crystal lattice, and the type and amount of impurities concentrated along the grain boundaries. While the surface energy is only a small part of the total energy of the system it may be important in recrystallization and can help to pro- duce a texture of uniformly—sized grains with smooth boundaries. The effects of this force on solid-state silicate reactions have been investigated (Kretz, 1969; Flinn, 1969) but as yet there is little data on how important this factor may be in the formation of carbonate rocks. In summary, the forces affecting dolomitization of a fine—grained limestone may be loosely grouped into two classes: thermodynamic, of which the effect can be seen by the bulk responses of the system to conditions affecting dolomitization, and kinetic, which is seen by the textural variability within the system. While the bulk responses of a rock have been extensively studied both in the laboratory and field, quan- titative textural responses have received less attention. In this study, both surface area of individual grains and their changing neighborhoods are monitored throughout the dolomi- tization process. While the problems of explaining and under- standing dolomites are far from solved, we have shown how a textural parameter can be used to explain and quantify some of the processes involved. LIST OF REFERENCES Badiozamani, Khosrow. 1973. The Dorag Dolomitization Model-- Application to the Middle Ordovician of Wisconsin. Jour. Sed. Pet., v. 43(4): 965-984. Bergenback,R.E. and Terriere, R.T. 1953. Petrography and Petrology of Scurry Reef, Scurry, County, Texas. Amer. Assn. Petrol. Geol. Bu11., v. 37(5): 1014-1029. Byerly, Gary R. and Vogel, Thomas A. 1973. Grain Boundary Processes and Development of Metamorphic Plagioclase. Lithis., v. 6: 183-202. Deininger, R.W. 1964. Limestone-Dolomite Transition in the ' Ordovician Platteville Formation in Wisconsin. Jour. Sed. Pet.,v. 34: 281-288. ' DeVore,G.W. 1959. Role of Minimum Interfacial Free Energy in Determining the Macroscopic Features of Mineral Assemblages. Jour. Geol.,v. 67: 211-226. Edie, Ralph. 1958. Mississippian Sedimentation and Oil Fields in Southeastern Saskatchewan. Amer. Assn. Petrol. Ehrlich, Robert. 1964. The Role of the Homogeneous Unit in Sampling Plans for Sediments. Jour. Sed. Pet.,v34(2): 437-439. Ehrlich, Robert et al. 1972. Textural Variation in Petro- genetic Analyses. Geol. Soc. of Amer. Bull,v. 83: 665-676. Flinn, D. 1969. Grain Contacts in Crystalline Rocks. Lithos., Hohlt, Richard B. 1948. The Nature and Origin of Limestone Porosity. Col. School of Mines Quart.,v. 43(4): '1-510 .‘ Kendall, M.G. and Moran,P.A.P. 1963. "Geometrical Probability. Charles Griffin, London, 125 p. Kretz, R. 1966. Interpretation of the Shape of Mineral Grains ' in Metamorphic Rocks. Jour. Petrol.,v. 7{: 68-94. 22 23 Murray, R.C. '1960. The Origin of Porosity in Carbonate Rocks. Jour. Sed. Petrol.,v. 30(1): 59-84. Westbrook, J.H. 1967. Impurity effects at Grain Boundaries in Ceramics. in Stewart, G.H., editor. Science of Ceramics 3. Academic Press, New York: 263-284.