INTRODUCTION

Solution concavities and imperfect crystal outgrowths (Fig. 1) characterize the surface of sand grains freed as a result of prolonged weathering of quartzose sandstones of western Virginia. The parent sandstones range in age from early Cambrian to early Pennsylvanian and include such formations as the Clinch-Tuscarora sandstone of Silurian age and the Oriskany sandstone of Devonian age. Quartzose sandstones protect the surface of the Appalachian Plateau and form the ridges of the Valley and Ridge province as well as some of the ridges of the west slope of the Blue Ridge. A thick residual soil mantle was developed on these ridge-forming quartzose sandstones in Virginia during one or more long periods of weathering in Cretaceous and Tertiary time (Lowry, 1954).

End_Page 489------------------------------

The residual sandy mantle has been preserved only on or near relatively stationary divides which have suffered least from the ravages of erosion during late Tertiary and Quaternary time. Some sand grains, especially the coarsest ones at some localities, have well developed solution concavities yet relatively few well developed crystal outgrowths. Other grains may show equal development of solution concavities and outgrowths. The finer sands tend to be angular and have a relatively higher percentage of their surface characterized by outgrowths. Smooth shiny crystal faces of some of the outgrowths appear in marked contrast to the dull and minutely pitted or at best uneven surface of concavities as well as the surface of the grains as a whole.

CEMENTATION BY WELDING

The concavities and other less well shaped depressions of the quartz grains are the result of solution of silica at original grain contacts whereas the outgrowths are the result of the deposition of the dissolved silica in originally lower pressure interspaces. This lithifying process, which need involve no introduction of silica from outside sources, continued so long as permeability was sufficient to permit transfer of silica. The process results in a marked reduction in volume and is believed to be in direct response to pressure developed with deep burial or during deformation. The texture of the sandstone formed by this process is sutured or interlocked (Fig. 2), and in the case of the quartzose sandstones of Virginia studied so far, all or nearly all of the original porosity and ermeability has been destroyed. Hatch, Rastall, and Black (1938, pp. 83 and 103) refer to this mode of cementation of sand with no addition of cement from an outside source as "welding." Although Butts (1940, p. 230) notes that the quartz grains of the Clinch-Tuscarora sandstone of Virginia are commonly cemented and partly rebuilt into crystals by silica deposited from circulating water, he does not mention any change in grain shape as a result of solution. Perhaps introduction of silica from an outside source by circulating water, if that is what Butts really implied, may have been important locally, but certainly most of the cementation has been by welding with the silica of local derivation. Waldschmidt (1941) points out that sandstones of the Rocky Mountain region, ranging in age fro Pennsylvanian to early Cretaceous, developed their sutured texture as a result of solution of silica from quartz grains at contacts and contemporaneous precipitation of secondary quartz within voids. The sandstones studied by Waldschmidt are also like the Virginia sandstones studied in that both show no conclusive evidence of fracturing of quartz grains. However, the Rocky Mountain sandstones, unlike the Virginia sandstones studied so far, retain part of their original porosity.

CONTACT SOLUTION FACTORS

As shown by the accompanying sketch of a thin section of early Cambrian Antietam sandstone (Fig. 2) from about 5 miles southeast of Natural Bridge, Virginia, the contact between some quartz grains, such as between grains A

End_Page 490------------------------------

and B, is the result of both grains losing silica whereas the contact between others, such as between A and K, is the result of only one, in this case grain A, losing silica. Perhaps Kennedy's study (Kennedy, 1950) of the solubility of quartz plates in water at temperatures ranging to 610°C. and at pressures ranging to 1,750 bars may explain part of these relationships. He notes that certain groups of quartz plates, all of which had the same orientation, required as little as 4 hours to attain equilibrium with water at 360°C., whereas others required as much as 16 hours. Kennedy suggests that the difference in the rate of solution is the result of minute compositional differences. Frederickson and Cox (1954) also studied the solubility of quartz in pure water at elevated tem erature and pressure. They note that water differentially etches or corrodes quartz and they attribute this to a preferential removal of a postulated cement whose chemical composition is the same as the bulk of the crystal but whose physical properties are different.

Although slight differences in chemical composition or structure of quartz grains are undoubtedly important in explaining the nature of the intimate contact between some grains whose contact was originally tangential, another factor may be of equal or greater importance. Kennedy (1950, p. 635) notes that plates cut parallel to the basal plane (0001) of a quartz crystal appear to reach equilibrium with the solution in one-fourth to half the time required by a plate cut parallel to the rhomb face. This phenomenon may possibly help explain

Click to view image in GIF format. Fig. 1. [Grey Scale] Photographs of originally well rounded quartz sand grains showing outgrowths and solution concavities. Grain A is from early Cambrian Erwin quartzite near Wytheville, Virginia, and grain B is from Silurian sandstone near Roanoke, Virginia. Some outgrowths of grain A (approx. ×50) show shiny crystal faces. Solution concavities of grain B (approx. ×30) were developed where numerous smaller grains were in contact with this large grain. Compare with thin section of Figure 1.

End_Page 491------------------------------

some of the relationships between grains at the contacts noted in the thin section figured. However, the work done so far in this regard has not given conclusive results. Merely determining the attitude of the "C" axes of two adjacent quartz grains in a thin section of sandstone is not sufficient to determine which grain should have lost the greater amount of silica. It is also necessary to determine the approximate attitude of the original surface of contact between the two grains because in a tightly packed sand the grain-contact surfaces are variously oriented. Much work on variously oriented thin sections of well rounded and exceptionally well sorted sandstones might reveal the role played by grain orientation in regard to contact solution.

As shown in the sketch of the Cambrian quartzite (Fig. 2), such as between grains A and L, the contact is the result of the growth outward from two grains originally not in contact insofar as the plane of the thin section is concerned. As was first noted by Sorby (1880), the secondary quartz of the outgrowths has the same orientation as the quartz of the detrital grains. The quartz of the outgrowths in the thin section of Antietam sandstone is in places separated from the quartz of the original grain by a very thin film of iron oxide. The quartz

Fig. 2. Thin section of early Cambrian Antietam quartzose sandstone from about 5 miles southeast of Natural Bridge, Virginia. Sutured texture, result of welding. Original shape of several grains indicated by dashed line along which is developed thin film of iron oxide. Silica has been lost by both grains A and B at their contact; grain A has lost silica at contact with K but K has lost none there. Both grains A and L have gained silica by outgrowth at one point along their contact. Roughly parallel lines within each grain are lines of strain, evident under crossed nicols. Some are bands rather than lines because not all surfaces of strain are perpendicular to plane of thin section. Outgrowths are strained fully as much as rest of grain. ×60.

End_Page 492------------------------------

of the outgrowths in thin sections of Oriskany sandstone in northern Virginia is difficult to distinguish from that of the detrital grain, but the quartz of the outgrowths tends to be free of inclusions.

The writer is aware that certain sandstones, such as the Potsdam in New York, have been cemented mainly by the development of outgrowths whose silica was largely introduced (Dietrich, 1953). However, in the western Virginia sandstones examined in thin section, the loss of silica by solution is equal to or nearly equal to the gain by outgrowth. In the few cases where solution losses do not appear to be as great as the gain by outgrowth, the introduced silica need not have moved far.

In the Virginia thin sections examined the quartz grains are slightly to markedly strained. The strain bands of most grains of the Antietam sandstone sectioned include the "C" axis or are nearly parallel with it, but some bands do show marked deviations from "C." Because the quartz of the outgrowths appears to be strained fully as much as that of the original grains, most of the welding took place before the development of any appreciable amount of permanent strain. In spite of this apparent fact, there appears to be in the thin section a tendency for outgrowths to be developed best where the strain bands intersect the periphery of a grain to form a large angle and for solution losses to occur either where the strain bands intersect the periphery to form a small angle or where the str in bands are parallel with the edge of the grain. Thus Riecke's well known principle of greater solution of stressed areas may be the factor most important in determining what part of a particular quartz surface is to lose silica if other factors are equal.

Additional petrographic work might show that the straining of the quartz grains of some sandstones is the result of the same stresses as those which controlled solution prior to the development of permanent strain. If the orientation of a thin section be known, then it may be possible to say whether the force responsible for folding a particular sandstone bed also produced the solution and strain effects shown by the quartz grains. Thus the relative importance of deep burial and deformation in regard to welding a particular sandstone might be estimated. Much petrographic work must be done before the relationships between the grains of sutured quartzose sandstones are fully understood.

APPLICABILITY OF EXPERIMENTAL FINDINGS

Although the writer does not question that burial to depths appreciably greater than 20,000 feet is capable of welding certain quartzose sands and completely destroying their porosity and permeability, he seriously questions the applicability of the recent experimental findings of Maxwell and Verrall (1954) to the study of the loss of porosity by Rocky Mountain and Appalachian quartzose sandstones. Maxwell and Verrall subjected rounded quartz sands to elevated temperature and pressure in the presence of several different solutions including sea water in an attempt to evaluate the effect of deep burial on the porosity of the sands. They state that their experiments indicate that a pure,

End_Page 493------------------------------

well sorted quartz sand, saturated with sea water or fluid of similar composition and pH, probably would not retain appreciable porosity (FOOTNOTE 3) if buried to a depth of 25,000 feet or more at temperatures above 270°C. Maxwell and Verrall also note that the mechanism largely responsible for consolidating the quartz sand used in their experiments is fracturing of grains with resulting interpenetration of grains and rotation of fragments. In none of their experiments was the porosity completely destroyed, and it is doubtful if crushing alone could destroy porosity. However, they do note that solution and redeposition of silica did occur. In questioning the applicability of their results to Appalachian and Rocky Mountain quartzose sandstones, two points are raised. First, petrog aphic work by Waldschmidt (1941) on Rocky Mountain sandstones and the work done so far by the writer on Virginia sandstones have failed to show any fracturing of quartz grains even though the grains and the outgrowths of some of the sandstones are markedly strained. Taylor (1950) in her study of the Morrison and Cretaceous sandstones in two deep wells in Wyoming noted the crushing and yielding of micas, feldspars, and rock fragments, but reported no such effects for the quartz grains. The maximum depth of burial of the sandstones studied by her can not be determined accurately, but they probably were buried at one time a few thousand feet deeper than their present depth. Taylor points out that tangential contacts between sand grains decrease exponentially with increasing depth and are no present in the Frontier sandstone at a present depth of 7,271 feet. She also notes that the depth at which sutured contacts first develop is between a present depth of 4,535 feet and 6,832 feet and that they increase progressively to a maximum in the deepest sand. Herein lies the second point in regard to the applicability of the results obtained experimentally by Maxwell and Verrall (1954). The force which they applied resulted in stresses at grain contacts far greater than those which actually develop with deep burial. By the time that sands of the Rocky Mountains and Appalachians were buried to a depth where the confining pressure was equivalent to the force that was applied in the experiments, the sands already had undergone appreciable lithification. The surface or area of contact etween grains had been greatly increased so that the stresses at the contacts were greatly reduced. If the rounded sand grains that Maxwell and Verrall used in their experiments had been spherical, the stress developed at grain contacts would have been tremendous because of the very small area of contact between such grains. Also if the quartz sand used had been very coarse, the stresses would have been even greater, for the number of contacts between grains would have been fewer.

The importance of an increase in grain contact area in preventing fracturing with burial can be emphasized in a different way. The rounded quartz grains of the friable St. Peter sandstone of the Ottawa, Illinois, district show shallow but well developed solution concavities. These concavities were developed at grain contacts. Although this Ordovician sandstone certainly was not as deeply buried or as strongly deformed as some of the Virginia sandstones considered

FOOTNOTE 3. About 10 per cent, personal communication from John C. Maxwell.

End_Page 494------------------------------

here, one might ask why the solution at grain contacts did not go to completion once it had begun. Probably the best explanation is that the process failed because as the contact surfaces became larger, the stresses developed at grain contacts became insufficient to cause further solution.

Perhaps the experimental results obtained by Maxwell and Verrall (1954) may be applicable to Tertiary sands whose burial to great depth occurred relatively rapidly. Experiments similar to those conducted by Maxwell and Verrall would be useful in attempting to predict the depth at which porosity and permeability are lost by quartz sands which do not experience deep burial catastrophically. In such experiments the deforming force must be not produce rupturing of the quartz grains. At the start of an experiment when the contacts between rounded grains tend to be tangential, less force should be applied than later in the experiment when the contact areas are larger. Sufficient time must be allowed for solution rather than fracturing to occur at grain contacts. These experiments should be erformed on coarse as well as fine sands in order to evaluate the size factor. The sandstones produced in these experiments should have strengths comparable with those found in nature.

Some of the Virginia sandstones studied thus far did not develop their sutured textures as a result of really deep burial. The Lower Pennsylvanian quartzose sandstones of the Cumberland Plateau of southwestern Virginia probably were never buried much deeper than 50,000 feet. Although these sandstone beds are little deformed, they apparently were subjected to sufficient, compression to weld the quartz grains together. The rocks of the Cumberland Plateau comprise a thrust sheet with a horizontal movement ranging from about 2 to 9 miles (Miller and Brosge, 1954). Some of the Lower Pennsylvanian sandstones there have lost all or nearly all of their primary porosity whereas those of comparable age in the Rocky Mountains retain appreciable porosity.

The original shape of the quartz grains of some sandstones of western Virginia has been little if any modified. This is true of calcite-cemented quartz grains of part of the Oriskany sandstone and it is also true of iron-cemented sandstone of the Clinton group which overlies the Clinch-Tuscarora sandstone. These cements are essentially primary and were introduced prior to deep burial or strong deformation. The shape of the quartz grains of certain Clinton sandstones whose interstitial material was originally clayey has been only slightly modified as a result of contact solution. Although these sandstones, whose grain shapes have been modified little or only very slightly, do not retain an appreciable amount of primary porosity, they do point up one question. Were there in places condi ions which retarded alteration of the shape of quartz grains sufficiently so as to preserve porosity and permeability?

CONDITIONS FAVORING RETENTION OF POROSITY BY VIRGINIA SANDSTONES

The condition which probably would be the most favorable for retention of primary porosity and permeability by a sandstone is the migration of oil or gas prior to deep burial or strong deformation. Although a film of water coating

End_Page 495------------------------------

the grain of an oil-saturated sand or sandstone may permit transfer of silica from stressed points of contact to lower pressure pore spaces, such a film can not be as effective a vehicle as a much larger quantity of water. Also, if the sand is coarse, the amount of water present as films coating the grains of an oil-saturated sand is less than that in an equal volume of finer sand because the surface area of the coarse sand grains is less than that of a finer sand. In addition, the thickness of the film of water coating the grains of an oil-saturated sandstone must decrease somewhat in the advanced stages of welding when the surface area of the grains has become somewhat larger as the shape of the grains departs more and more from that of a sphere. Probably of even greater importance han a slight reduction in the thickness of the film of water is the great increase in the area of the surface of contact between grains in the advanced stages of welding, for it is from this surface that silica must be dissolved before further reduction in volume and porosity can occur.

Oil has not been found in commercial quantities in any sandstone in Virginia, and gas is known to occur only in Oriskany sandstone of Devonian age and in younger Paleozoic formations. It is only fair to point out that post-Cambrian sandstones are exposed in most of the major anticlines of western Virginia. However, in the Bergton-Crab Run anticline of Rockingham County, Virginia, where the Oriskany sandstone contains gas, the Silurian sandstones as well as the Cambrian have not been tested. Although Young and Harnsberger (1955) believe that the permeability of the Oriskany sandstone of the Bergton gas field is largely the result of fracturing, it may retain slight primary porosity. The character of the Cambrian sandstone is also not known in the Rose Hill oil field of Lee County, Virg nia, where the main reservoir rock is Trenton limestone. In addition to the oil- or gas-bearing structures here noted, the Lower Cambrian sandstones are buried in several major anticlinal structures of western Virginia. All of the Cambrian sandstones in these anticlines have undergone deep burial in addition to strong deformation. Although their chances of having retained primary porosity and permeability even in the most favorable structural and stratigraphic settings are not good, they have not been adequately tested. In Virginia buried and structurally high Cambrian sandstone is known to have been reached only in United Fuel Gas Company's unsuccessful test of the Burkes Garden dome in Tazewell County. Whether the entire sandstone section was penetrated is not known by the writer but b sement was not reached. Cuttings of the sandstone are fine- to medium-grained and characterized by a sutured texture.

If Cambrian sandstone in western Virginia ever becomes productive, the chances are that the sandstone will be found to be coarse-grained and well sorted and that oil or gas entered prior to deep burial. Warping not long after deposition of Silurian sandstone would have favored migration of hydrocarbons prior to deep burial. Although folding such as that which took place in eastern Pennsylvania in late Ordovician time did not take place northwest of the Blue Ridge in Virginia, gentle folding may have begun. As yet, we can not say that folding was

End_Page 496------------------------------

not active in the Piedmont of eastern Virginia in late Ordovician or early Silurian time, for the youngest known Paleozoic beds there apparently belong to the Upper Ordovician. In several places in western Virginia the Ordovician section appears to be either abnormally thick in synclinal positions or abnormally thin in anticlinal areas. Whether the result of contemporaneous longitudinal warping or proximity to source (Lowry, 1955), early migration is favored. In addition to abnormal thicknesses there is a well defined unconformity between Lower and Middle Ordovician limestones. In places the surface of unconformity has a relief of several hundred feet. Perhaps of much greater significance is the existence in at least one place of a gradational relationship between Lower and Middle Ord vician limestones in a synclinal area where as a well defined hiatus occurs in an adjacent anticline. Gentle folding in western Virginia in early as well as somewhat later Paleozoic time would not only have favored early migration and accumulation of hydrocarbons but also would have kept Cambrian and Silurian sandstones from being as deeply buried in upwarps as they were in synclines. Certainly it would be unfair in such a case to use the thickness of the Devonian and Carboniferous sections as measured in flank or synclinal settings as the correct figure for the comparable section once present in the crestal part of an anticline.

Most of the Virginia sandstones and residual sands examined were collected from the flanks of folds. From several standpoints it may be very unfair to judge the character of a sandstone once present in the crestal part or culmination of an anticline by that seen in the flanks. The breached character of many of the anticlines of the Valley and Ridge province might be in part the result of retention of some porosity and permeability in the highest part of the structure.

It is also unfair to judge the character of Lower Silurian sandstone everywhere by the Clinch sandstone encountered in the California Company's unsuccessful test of the Price Mountain anticline in Montgomery County, Virginia. Although the Clinch sandstone at present is at a depth of less than 7,000 feet, the rock is an impermeable vitreous quartzite. Several factors explain the character of the sandstone there. First, Price Mountain is a window in the Pulaski thrust sheet whose basal strata are Upper Cambrian dolomitic limestones. Although erosion undoubtedly stripped off the upper strata of the thrust block during thrusting, the chances are that all of the Ordovician and Silurian systems and much of the thick Devonian system were once present on top of the Mississippian and older str ta of the overridden block. Another importance factor which must be considered is that the Price Mountain anticline is a fold within the trough of a major syncline. This anticline could have developed long after any hydrocarbons had been formed and migrated updip to the adjacent major anticlines.

OTHER FACTORS BEARING ON POSSIBLE RETENTION OF POROSITY

Retention of primary porosity and permeability by sandstone may be governed in part by the behavior of a sand or sandstone bed during lateral compression. Certainly the sand grains in the crests and troughs of folds

End_Page 497------------------------------

undergo less intergranular movement than those in the steep flanks, for the attitude of the bed is less changed in the crests and troughs. Also at the crest of an anticline the sand grains in the upper part of a partly lithified sandstone should undergo less compression than the grains in the lower part of the same bed as the tension developed at the top should partly offset compression. Offsetting these factors somewhat is the smaller area of a vertical section of the bed at the crest than in the steep flanks. Still another factor to be considered in regard to possible retention of primary porosity and permeability is the thickness of the sandstone section. Although thickness can be neglected in connection with vertical force resulting from deep burial, it must be considered in the c se of horizontal compression. If a sandstone section is stronger than adjacent units, it has to transmit more than its share of force, especially in the later stages of folding. In such a case the sand grains of a thick sandstone section undergo less stress than those of a thinner section. If, on the other hand, a sandstone section is adjacent to a competent section of thick limestones or other competent rock, it may not have to transmit nearly as much force as the sandstone section whose adjacent beds are incompetent.

Except for the possible early introduction of oil or gas, the factor most favorable to the retention of primary porosity and permeability by a quartzose sand is the possession of high permeability initially. Grain size, shape, and degree of sorting largely determine the permeability of a well packed sand. Although an open-packed, uniform-size, angular or subangular sand may possess a greater porosity than an open-packed, uniform-size, rounded sand, the rounded sand may retain greater porosity when tightly packed. Also the average sand composed of rounded grains may be better sorted than the average sand composed of less abraded grains.

Of much greater importance than shape in regard to initial permeability is grain size. Other factors being equal, permeability varies roughly as the square of the diameter of the grains. Thus a coarse-grained, well sorted, well rounded, well packed, quartzose sand should possess greater permeability than a similar medium-grained one and certainly much greater permeability than a fine-grained sandstone, most of whose grains are less well rounded. Of the residual quartzose sands studied so far, some of the coarse, originally well rounded ones appear to have experienced the least change in shape. Although the grains of a coarse quartzose sand would lose silica at contacts more rapidly than a finer sand during the early phases of deformation or deep burial because the fewer points of cont ct experience greater stress, it may be that as the large grains are joined by contact solution, their contact surface increases in area sufficiently rapidly to offset, in part, the initial disadvantage. Also just prior to completion of welding, a coarse sand with its larger pores is more apt to retain slight primary permeability than a finer one with more numerous microscopic pores. Because of the several factors involved in the welding of quartz grains, an experimental approach to this phase of the problem may be better than a mathematical one.

End_Page 498------------------------------

As already suggested, the effect of grain size on reduction of primary porosity and permeability should be experimentally investigated. The use of solutions which would accelerate contact solution would facilitate the evaluation of this apparently very important factor.

If the possession of coarse grain size favors retention of primary permeability, then one other factor should be considered in regard to the possible presence of oil or gas in the still buried, untested, structurally high quartzose sandstones of western Virginia. An attempt should be made to find the coarsest and best sorted sands present in those areas farthest northwest where deformation was less severe. For example, if the possibilities of Silurian sandstones are being considered, one should study and compare the character of the Silurian sandstones west and northwest of the Massanutten syncline (Butts, 1940, p. 443), for this area appears to have lain near the mouth of a major stream bringing in sands from the area on the east. Also in such a northwest area the aggregate thickness of the Silurian sandstone might be somewhat greater and this, as previously noted, might also favor retention of primary permeability by the better sorted and coarser sandstone beds.

Until the origin of the permeability of the Oriskany sandstone in the Bergton gas field has been learned by studying cores rather than cuttings and until the older sandstones have been tested there as well as in the Rose Hill oil field, it certainly is not justified to rule out other parts of western Virginia as possible producers from early Paleozoic sandstones. Silurian sandstones in the Rose Hill district did produce a little oil (Miller and Fuller, 1954). Even should the older Paleozoic sandstones in these two fields be proved unproductive, we could not eliminate all of western Virginia and comparable areas elsewhere in the Appalachians until exposed sections of these rocks had been carefully studied. The thickness, lithologic character, stratigraphic relationships, and structural positions of these sandstone sections should be compared. Not only should an attempt be made to determine the environment of deposition of the parent sands, but also contemporaneous gentle folding should be studied.

The few wells in western Virginia which have reached the Silurian or Cambrian sandstones do not constitute an adequate testing of these beds. Additional testing of Lower Paleozoic quartzose sandstones in western Virginia as well as elsewhere in the Appalachians is warranted. The Lower Paleozoic sandstones would not be the only objectives, for fracture and perhaps solution permeability undoubtedly exist in places in some of the Cambrian and Ordovician dolomites or limestones. Locally, where exposed, some of the beds have a strong petroliferous odor.