Thermodynamic Constraints

23 Cementation and dissolution are governed mainly by the saturation state. In general, temperature, pressure, and composition of the groundwater(s) determine the saturation state(s) and, hence, precipitation and dissolution, either directly or indirectly in one or several ways.

1. Pore fluids in sedimentary basins, if buried and not changed significantly through water-rock interaction or mixing, become supersaturated with respect to carbonates and sulfates with depth, even if the concentrations of the cations (calcium, magnesium, iron, etc.) and carbonate ions remain essentially unchanged. Conversely, ascending formation/ groundwaters that cool tend to become undersaturated with respect to carbonates and sulfates, with the potential to generate significant secondary porosity in carbonates and/or clastics that contain carbonate cements (e.g., Giles and deBoer, 1989). The reverse is true for silicates and most other mineral groups. This is because of the temperature-dependent retrograde and prograde solubility of carbonates and sulfates versus silicates and other mineral groups, respectively.
2. The concentrations of the cations and carbonate ions in solution generally increase because of pressure solution and/or a variety of diagenetic reactions involving non-carbonate minerals, such as silicates and sulfates. A common result is precipitation of carbonates because of the common-ion effect (e.g., Hanor, 1987).
3. An increase in acidity can take place in many ways, including meteoric water penetration, mixing of waters, carbon dioxide generation from maturation of organic matter, generation of organic acids, and formation of acids from clay mineral reactions (e.g., Giles and Marshall, 1986); it leads to carbonate dissolution.
4. Mixing of groundwaters may lead to pronounced undersaturation or supersaturation with respect to carbonate minerals, depending on temperature, pressure, and fluid composition, particularly pCO₂ of the mixed water. Mixing of shallow groundwaters at relatively low temperatures commonly results in under-saturation and dissolution, which is well known from karst terrains (e.g., Bonacci, 1987; James and Choquette, 1987; Smart et al., 1988) and from caves in coastal seawater-freshwater mixing zones (e.g., Machel and Mountjoy, 1990). However, mixing of fresh groundwater with brine in deep burial settings commonly leads to supersaturation and carbonate precipitation (Morse et al., 1997).
5. Pressure decreases, with or withouta significant temperature decrease, almost invariably leads to carbonate supersaturation and precipitation (Morse et al., 1997), either because of reverse solubility, CO₂ -degassing, or both.

24 Under the assumption that minerals and solutions attain local thermodynamic equilibrium along the flow path, the above processes can be quantified in computer simulations to predict the precipitation and dissolution of diagenetic minerals along the flow path in basin strata as groundwaters migrate along temperature and pressure gradients. Using this approach, rock porosity evolution can be calculated from relative amounts of mineral precipitation and dissolution (e.g., Lee, 1997; Morse et al., 1997). Permeability changes, however, can merely be inferred, as pore throat sizes and tortuosity cannot be predicted from thermodynamics.

Display large image of Figure 3

Diagenetic Potential and Kinetic Factors

25 Both carbonates and clastics have different diagenetic potentials during burial, i.e., in any diagenetic setting certain minerals or types of grains exist that dissolve, form, or recrystallize preferentially over others. Such phenomena are governed by textural and/or kinetic factors. One of the best known examples is that small grains dissolve more rapidly than large grains during recrystallization and/or Ostwald ripening (e.g., Folk, 1965; Morse and Casey, 1988). Another example of microstructural control is that aragonite with complex microstructures can dissolve more rapidly than thermodynamically less stable magnesian calcite (Walter, 1985). Kinetic factors can override thermodynamic constraints during precipitation, particularly in near-surface and shallow diagenetic settings. Examples include the well-known kinetic inhibition of dolomite formation from seawater, and the formation of ‘ protodolomite' rather than dolomite from evaporated groundwater or seawater (e.g., Machel and Mountjoy, 1986; Usdowski, 1994). Such kinetic problems are ignored in the remainder of this paper, however, under the assumption that most diagenetic settings have had enough time to adjust to a thermodynamic equilibrium mineral assemblage. This certainly is true for most intermediate- and deep-burial settings, where sudden incursions of "alien" formation fluids are relatively rare, and where flow rates are generally low, except in fracture systems.

26 Another important factor, especially in intermediate- and deep-burial settings, is that hydrocarbons commonly arrest diagenesis, either because they coat potentially reactive grain surfaces in oil-wet systems, or because the relative permeability to water is very low or zero in the case of high oil saturation in water-wet systems (e.g., Laudon, 1996). In both cases, ionic exchange through the aqueous phase (dissolution, precipitation), and hence diagenesis, is inhibited, except for local and volumetrically insignificant mineral redistribution by diffusion. However, some hydrocarbons may react with the rocks, as in the case of thermochemical sulfate reduction (discussed below).

Depositional Controls

27 In carbonates, much more so than in clastics, depositional environment and facies, especially cycle stacking patterns, play an important role in determining the porosity and permeability distribution and their subsequent evolution through burial diagenesis. A case in point is the Knox Group in Tennessee, which contains transgressive and regressice carbonate cycles (Montañez, 1997). The transgressive cycles, which originally were fairly porous and permeable, now consist of limestone and <10 to >50 vol% "late-diagenetic" dolomite having an average porosity of 8.2% and an average permeability of 70.2 md, which are interpreted to have formed during intermediate to deep burial. Petrographic data also show that these rocks were affected by extensive dissolution in intermediate- to deep-burial settings. However, the facies within regressive cycles are almost completely replaced by tight fine-crystalline dolomite that formed syn-depositionally from evaporated seawater. The transgressive carbonate cycles behaved like aquifers during later burial diagenesis, whereas the regressive cycles became aquitards almost syn-depositionally. Other examples of depositional control on diagenesis are reef and reef margin carbonates, which typically are much more porous and permeable than back-reef and deep fore-reef carbonates (e.g., Schneidermann and Harris, 1985; Schroeder and Purser, 1986). This commonly leads to a pronounced differentiation into "proto-aquifers" and "protoaquitards" at the time of deposition.

Marine Diagenesis

28 Most carbonate rocks originate as marine sediments having primary porosities around 40-45%. Hence, the first pore water to affect the carbonate sediments is seawater, and the diagenetic setting is near-surface marine, where the predominant hydrologic drives are wind/wave action and tidal pumping. The main effect is partial to extensive filling of primary pores with internal sediments and marine carbonate cements (James and Choquette, 1990), which leads to significant porosity reduction. In general, highly porous and permeable facies types affected by vigorous waves and tides tend to become more cemented than other facies types in hydrogeologically less vigorous settings. Recrystallization of metastable aragonite and magnesian calcites also occurs, but this generally has little effect on porosity/permeability development.

Coastal Mixing Zone Diagenesis

29 In coastal areas where seawater mixes with fresh water, the predominant diagenetic process affecting carbonate sediments and rocks is dissolution (Plummer, 1975). Groundwater fluxes in coastal mixing zones (also called "zones of dispersion") typically are quite high, and the effects are enhanced porosity and permeability, including caves, which are well known from many such settings. Other well known processes in coastal mixing zones are dolomite cementation and replacive dolomitization. However, the effects of these processes are volumetrically insignificant (less than 1 vol-%) in almost all cases (Machel and Mountjoy, 1990). Similarly, some coastal mixing zones form aragonite or calcite cements (Csoma and Goldstein, 2004), but mixing zone cementation is volumetrically negligible in most settings.

Meteoric Diagenesis

30 Meteoric waters in near-surface settings are almost always undersaturated with respect to carbonates because rain water is essentially devoid of earth alkali metals yet slightly acidic because of dissolved atmospheric CO₂. Where the ground has a significant soil cover, pCO₂ in the downward percolating rain water of the vadose zone can easily increase by two orders of magnitude because of extensive plant and microbial activity in the soil zone. This increase results in extensive dissolution in the upper few metres of burial, and the effects are increased porosity and permeability and/or physical weakness of the rocks. For example, in the Caribbean islands, many of which have a cover of Cenozoic carbonates that have not been buried by more than a few metres, the rocks are frail and crumbly because of a lack of near-surface cementation and/or subsequent meteoric dissolution (James and Choquette, 1984).

Hypersaline Diagenesis

31 The only volumetrically important and widespread diagenetic effect of ground-water flow on carbonates in near-surface evaporitic settings is reflux dolomitization. Typically, fresh groundwater and/or seawater that are evaporated to or beyond gypsum saturation flow seaward within the uppermost few metres of the carbonate sediments via density-drive. These fluids form finely crystalline and geochemically distinct dolomite, either as a replacement or as cement. The resulting dolostones typically are laterally extensive, yet thin, aquitards that may form effective seals for hydrocarbon trapping (e.g., Machel and Mountjoy, 1986; Moore and Wilde, 1986).

Physical (Mechanical) Compaction

33 With increasing burial, physical (mechanical) compaction leads to loss of water and porosity. Carbonate sediments can compact to about one-half of their original thickness under as little as 100 m of overburden by physical compaction alone (Choquette and James, 1987, 1990). Thereby, sedimentary particles and structures are modified and rearranged until a self-supporting framework is achieved at an average porosity of about 40%. Further burial will lead to grain deformation in the form of ductile (squeezing) or brittle (breakage) failure, followed by chemical compaction. Common textures generated by mechanical compaction include thinning of laminae between, and draping over, concretions; flattening of burrows, fenestrae, gas-escape structures, desiccation cracks, and skeletal or detrital grains; rotated grains; spalling of coated grains; swirling structures; telescoping (conversion of grain-poor to grain-supported textures); and planar to curviplanar grain contacts. In contrast to most other diagenetic processes and effects, the extent of physical compaction and concurrent porosity/permeability loss strongly depends on depositional setting and near-surface diagenesis. Specifically, shallow-water carbonates generally become strongly cemented and solidified in the near-surface marine setting, which inhibits later physical compaction. However, cementation is typically poor in deep-marine carbonates, such as chalks, which permit much more extensive physical compaction during burial (e.g., Choquette and James, 1987, 1990; Maliva and Dickson, 1992).

Chemical Compaction

34 Chemical compaction results from pressure solution that is caused by load or tectonic stress at grain contacts, because stress increases mineral solubility. In addition, clay minerals embedded in the carbonate matrix may enhance or reduce the carbonate solubility (depending on the types of clays and their composition). Other factors that influence the inception and course of pressure solution include the burial depth, pore water composition, mineralogy, and the presence or absence of organic matter (including oil and gas). Depending on the interaction of these factors, pressure solution in carbonates commonly begins perhaps within 200 to 300 m of burial, with the first microstylolites forming at minimum depths of ca. 500 m of burial (Lind, 1993; Fabricius, 2000). Where dolostones are associated with limestones, the latter tend to be more susceptible to pressure solution.

35 Pressure solution textures are diverse and include stylolites (one or several generations, typically with different amplitudes), other types of sutured seams, isolated or fitting nodules, nonsutured seams (Choquette and James, 1987, 1990), and pseudo-bedding that can resemble layering caused by sedimentary processes (Bathurst 1987; Munnecke et al., 2001; Fig. 4). Stylolites can be used to determine the number of burial episodes (number of generations) and the amount of compaction (calculations involve the number and amplitudes of stylolites: Bodou, 1976; Choquette and James, 1987, 1990; Ricken, 1987).

36 The type of interpenetration of adjacent grains during pressure solution depends on which minerals are juxtaposed. Where two grains are the same mineral, the pressure solution contact is sutured, as both grains dissolve to the same extent. However, the contacts are smooth where the mineralogy differs, as one mineral effectively grows at the expense of the other. For example, dolomite (mineral A) commonly grows into calcite (mineral B), and quartz commonly grows into calcite and/or dolomite, as illustrated in Figure 5. Such textures have been known for a long time, and the most common explanation given is that the solubility of mineral B is higher than that of mineral A at the contact point or surface. An alternative and more convincing explanation has been provided relatively recently by Fletcher and Merino (2001). These authors redefined and thereby expanded the principle of "force of crystallization" as "interfacial normal stress" (or "induced stress"). This stress is generated by growth driven by supersaturation of mineral A between crystals of minerals A and B, and the magnitude of this stress is dependent on the degree of supersaturation, temperature, as well as what the other mineral is. For example, at a temperature of 100 °C and a super-saturation of OMEGA = 2, quartz growing at the expense of calcite (quartz replacing calcite) generates a negligible stress, whereas calcite growing at the expense of quartz generates a huge stress of about 100 MPa (Fletcher and Merino, 2001). This effectively means that quartz penetrates calcite like a knife cutting into butter, as shown in Figure 5, where the pore water is supersaturated with respect to quartz yet saturated or undersaturated with respect to calcite. Conversely, calcite tends to deform or fracture adjacent quartz grains, if the pore water is supersaturated with respect to calcite yet saturated or undersaturated with respect to quartz. Both these scenarios are possible during pressure solution, depending on fluid composition. Figure 6 shows this and several other scenarios. Mineral pairs in the top left quadrant will show replacement textures, like "a knife cutting into butter", whereas the inverted mineral pairs in the bottom right quadrant will show deformation and/or brittle failure of the "weaker" mineral.

37 Chemical compaction begins during the latter phases of physical compaction and continues into the metamorphic realm (Choquette and James, 1987, 1990). Chemical compaction is most active in rocks that are being buried (as opposed to being stationary at some depth), and that have some porosity left to facilitate grain interpenetration and ionic movement from the points of dissolution into the pore fluids. The liberated ions move, via advection or diffusion, to areas of lower stress, either just adjacent to the area of dissolution or elsewhere in the pore system, where the dissolved material may reprecipitate, forming burial-diagenetic cements. Chemical compaction is known to result in thickness reductions of at least 25-35% in addition to the thickness losses caused by physical compaction. In the Smackover oolites, which were used to construct the porosity/depth curves in Figure 3, physical and chemical compaction combined account for a loss of 33% of the original 45% porosity (Heydari, 1997).

38 Chemical compaction generally is detrimental to reservoir rock characteristics and/or petroleum production. The major effects are porosity and permeability reduction as a result of the overall thickness reduction and reprecipitation of the material dissolved at grain contacts.

Display large image of Figure 4

Cementation

39 Cementation is common in intermediate- and deep-burial settings because of elevated temperatures, fluid mixing, and chemical compaction (see above). Cementation of several volume percent, which is common, requires large fluid fluxes, however, because the solubility of most carbonates is rather low (e.g., K = 10 ^-8.52 at 25°C for calcite, decreasing with increasing temperature, Robie et al., 1979). Nevertheless, numerous case studies have demonstrated extensive intermediate- and deep-burial cementation (and concurrent porosity/permeability losses) in carbonate aquifers, not only with carbonate minerals but also with sulfates, sulfides, and clay minerals (e.g., Scholle and Halley, 1985; Walls and Burrowes, 1985; Woronik and Land, 1985; several articles in Montañez et al., 1997). By contrast, some studies have shown deep burial cementation to be relatively minor (e.g., Prezbindowski, 1985).

40 Burial cements in carbonate rocks are mainly calcite, dolomite, and/or anhydrite, and they have a number of characteristics: (a) crystals cross-cut stylolites, or microstylolites and pressure solution seams terminate at cement crystals; (b) crystals occur in fractures or spalled ooid cortices; (c) crystals fill compacted pores; (d) crystals enclose compacted grains; (e) crystals contain hydrocarbons in fluid inclusions; (f) crystals partially replace earlier burial cements; (g) crystals have relatively large size (diameters from several tens of micrometres up to several centimetres); (h) crystals are ‘ blocky' and form equant mosaics, increase in mean diameter toward pore centres, or are poikilotopic (poikilitic); (i) crystals are not recrystallized; (j) carbonate crystals show well-defined cathodoluminescence zonation that ranges from bright towards quenched (dull), in some cases repeatedly, or they are dull-luminescent throughout, commonly caused by ferroan composition (divalent iron >500 ppm); (k) carbonate crystals have oxygen isotope ratios that are depleted relative to marine cements of the same stratigraphic age as the host rocks; and (l) crystals contain two-phase fluid inclusions with elevated temperatures of homogenization (>50°C) and large freezing point depressions (indicating high salinities). Characteristics (a) to (f) above are distinctive and indicate a burial origin (Figs. 7 and 8), but are relatively rare in some rocks. The other characteristics are not unique to burial diagenetic cements. Therefore, any interpretation of burial cementation should combine as many lines of evidence as possible.

Display large image of Figure 5

Display large image of Figure 6

Dissolution

41 Dissolution is common in intermediate-and deep-burial settings (albeit volumetrically it is not as important as cementation, or else a general porosity decrease with depth would not occur - see Figure 3), especially near the oil window where decarboxylation and certain other mineral reactions generate acidity. Secondary porosity is well known from carbonates (Choquette and Pray, 1970) and clastics (Schmidt and MacDonald, 1979). Significant dissolution may or may not involve large fluid fluxes. If acidity is generated within or near the rock with secondary porosity, large-scale pore water flow is not necessary.

42 A hitherto controversial process to generate secondary porosity in deep burial settings is thermochemical sulfate reduction (TSR). Model calculations suggest that partial dissolution of carbonate rocks may occur at least initially during TSR (Hutcheon, 1992; Nicholson, 1992), but most published case studies of TSR do not show a significant increase in porosity (e.g., Machel et al., 1995a). Dissolution via TSR, however minor, usually is an in situ process because TSR commonly takes place in hydrologically closed or nearly closed systems (e.g., Machel 2001).

Recrystallization

43 The matrix, allochems, and cements formed at shallow depths become thermodynamically metastable or unstable with increasing burial because of increasing temperature, pressure, and changing composition of the groundwaters. Barring kinetic inhibition, recrystallization and replacement of metastable and unstable minerals will occur. However, porosity and permeability are generally little affected by recrystallization.

Display large image of Figure 7

Display large image of Figure 8

44 The most common textural manifestations of recrystallization of carbonate cements are the enlargement of crystals (Ostwald ripening) and partial elimination of primary crystal or grain boundaries (Figs. 9A, B). Cathodoluminescence images of recrystallized carbonates commonly are ‘ blotchy'. Recrystallization also results in changes in crystal structure, isotopic composition, and paleomagnetic properties (e.g., Machel, 1997).

45 For practical applications, the absence, presence and/or degree of recrystallization is/are important for genetic interpretation. Consider, for example, a dolostone. If recrystallization changes or resets the original properties of the cystals, such as texture, structure, composition, and/or paleomagnetic properties, beyond what they were in the pristine (unrecrystallized) rock, the dolostone is said to be "significantly recrystallized". The reset properties are no longer representative of the fluid and environment during dolomitization but they reflect the recrystallization event. However, some measurable properties may not be reset and they will represent the dolomitization event. If recrystallization resets only one of 10 properties (e.g., ⁸⁷ Sr/⁸⁶ Sr-ratios) beyond the range in the pristine rock, the dolostone is significantly recrystallized with respect to Sr-isotopes, yet insignificantly recrystallized with respect to the other nine properties. In either case, values of those properties that are identical to the pristine reference values can be used for genetic interpretations of dolomitization.

46 The concept of "significant recrystallization" is of great use in genetic interpretations. In particular, for a reliable interpretation of the chemistry of the diagenetic fluids, it is sufficient to recognize that a crystal is insignificantly recrystallized (Machel, 1997).

Replacement

47 Next to dolomite (discussed separately below), the most common replacive mineral in deep carbonate aquifers, is anhydrite. For example, extensive anhdrite replacement of dolostones, with or without concomitant anhydrite cementation, is well known from the Devonian strata of western Canada (e.g., Walls and Burrowes, 1985). Replacive anhydrite commonly is massive but may also resemble primary nodular or bedded sulfates, which is best recognized where reef rock is replaced (Fig. 9C). At depths of less than 600m (up to 1000m depending on temperature and fluid composition), gypsum forms rather than anhydrite because the gypsum-anhydrite stability boundary is crossed somewhere in this interval. Furthermore, replacive gypsum generally dewaters and recrystallizes to anhydrite, generating distinctive microtextures (Fig. 9D).

Display large image of Figure 9

48 Dolomite and silica (chert) are other common replacements of limestone. The resulting textures can be predicted from thermodynamic considerations to some degree. As Fletcher and Merino (2001) have shown, the induced stress by quartz growing replacively toward and within calcite is very small (Fig. 6); hence, quartz tends to form euhedral crystals in limestone. The same is true for dolomite in limestone, at least at relatively low temperatures. However, where gypsum grows in limestone, the induced stress at the grain boundaries is very large; thus, calcite tends to deform or fracture, unless there is enough porosity to accommodate the sulfate.

49 Replacement of carbonates by calcium sulfate can have a pronounced effect on porosity and permeability. Regardless of the actual quantities of anhydrite formed relative to dolomite replaced or cemented, dolostones commonly are porous, whereas anhydrite is tight. Hence, any anhydrite cement or replacement will reduce the porosity significantly. Permeability, however, is not necessarily affected to a proportional degree. Anhydrite generally has a low nucleation rate and thus a tendency to grow in a few patches of large crystal aggregates, rather than forming pervasively throughout the rock. As a result, more than about 60% of the total rock volume has to be filled or replaced by anhydrite cement before the permeability is reduced to any significant degree (Lucia et al., 2004). At lesser amounts of anhydrite in the rock, the formation fluids (water and/or petroleum) tend to flow around the large patches of anhydrite.

Epimetamorphic Fluids and Tectonic Injection

50 Diagenesis in the deepest parts of the diagenetic realm is a poorly understood phenomenon because it grades into the epimetamorphic realm. Some authors (Land, 1997; Schroyen and Muchez, 1998) advocate that material can be transferred from the crystalline basement into the overlying deep burial diagenetic setting. Prograde metamorphism and devolatilization reactions liberate water and CO₂ , which are added to the sedimentary basin above. In the case of the Western Canada Sedimentary Basin, there is petrographic, isotopic, and circumstantial evidence for injection of Laramide metamorphic fluids into the Devonian carbonate aquifers in the deepest part of the basin (Machel and Cavell, 1999). The fluid fluxes appear to be low, however, as shown by the relatively small amounts of cement formed, commonly white sparry calcite (Machel et al., 2001). The effects of this type of fluid flow on reservoir properties are not yet understood.

Fractures

51 Where fractures are restricted in lateral or horizontal extent within one of the diagenetic burial zones (Fig. 2), diagenesis within the fractures is generally similar or identical to that within the adjacent wall rocks. Where fractures cross from one diagenetic zone into another, however, relatively cold groundwaters may descend or relatively hot groundwaters may ascend fairly rapidly. Hence, the fractures in any diagenetic zone may contain hydrofrigid (and underpressured) or hydrothermal (and overpressured) fluids relative to those in the surrounding wall rocks (Machel, 1999; Machel and Lonnee, 2001). This situation raises numerous possibilities for mineral reactions, especially when the fractures transect different rock types (e.g., Pedersen and Bjørlykke, 1994; Stoute and Harris, 1995). A comprehensive discussion of these possibilities is beyond the scope of this paper. The following examples characterize the most common and perhaps most important aspects of formation water flow though fractures.

52 Rapid descent of cool meteoric waters to depths of several kilometres through deep-reaching faults/fractures is common in mountainous recharge areas. In Western Canada, for example, such hydrologic systems are common in the Rocky Mountain Front Ranges, where meteoric waters convectively descend and return to the surface in sulfurous hydrothermal springs, e.g., at Miette, Banff, and Radium. The diagenetic effects in the descending parts of these flow systems are virtually unknown, except that the waters appear to equilibrate isotopically rather fast in at least some locations (Nesbitt and Muehlenbachs, 1997). The major effect in the ascending parts of these flow systems is massive cementation within the fractures, which commonly reach centimetre and locally decimetre dimensions, and are filled with sparry calcite or dolomite, quartz, and to a minor degree with other minerals, including sulfates and fluorite (Nesbitt and Muehlenbachs, 1997).

53 Similarly, from a hydrogeological point of view, rapid convection of meteoric waters is known from intermontane rift basins. In one of the best studied examples, Tóth and Otto (1994) demonstrated that meteoric waters move preferentially through fractures from recharge areas in the mountains that fringe the Rhine Graben toward the enclosed valley, where they discharge as hydrothermal and petroliferous brines. Radiocarbon data from waters of producing wells range from 3,165 to 31,000 years, indicating that convection is very rapid. The diagenesis within the fractures has not been investigated, however.

54 Rapid ascent of hydrothermal fluids derived from shallow-metamorphic and deep-burial diagenetic settings is well known from many sedimentary basins. Perhaps the best known effect of such "hydrothermal groundwater flow" is Mississippi-Valley-Type Pb-Zn-(saddle) dolomite mineralization that may be restricted to fractures, or that may partially replace adjacent wall rocks (e.g., Anderson and Macqueen, 1982; Braithwaite and Rizzi, 1997; Gregg et al., 2001).

55 Fibrous minerals in fractures indicate a special mode of fracture opening and filling, commonly referred to as "crack-seal" (Ramsey, 1980). In this case, the pore fluids precipitate minerals as fast as the fractures are opening, which may happen in multiple episodes, such that the actual space between the facture walls never exceeded a few nanometres or micrometres (Machel, 1985). An example of fibrous anhydrite is shown in Figure 7C. Ramsey (1980), Machel (1985) and others have interpreted the fibrous crystals to be passively precipitated in veins that opened under externally applied stress. An alternative interpretation has been provided by Wiltschko and Morse (2001), who advocated that such fibrous mineral veins originate from the vein mineral(s) actively pushing the wall rock apart.

56 Fractures commonly serve as migration pathways for petroliferous fluids. In such cases, the fluids may form minerals with petroliferous fluid inclusions and/or solid bitumen in the fractures, which can be used to determine hydrocarbon migration pathways and related aspects, such as timing of migration and fluid composition (e.g., Levine et al., 1991). In deeply buried and tightly sealed reservoirs, fractures are not only used by hydrocarbons as migration pathways but they may be generated by the pressure from in-situ hydrocarbon generation, either oil or gas from kerogen, or gas from oil. Especially well documented examples of this phenomenon occur in Devonian reefs in the deepest part of the Alberta Basin, Canada, where in-situ cracking of oil to gas has led to intense hairline fracturing that appears as pervasive "shattering" of the rocks (Marquez and Mountjoy, 1996; Fig.10).

Bacterial and Thermochemical Sulfate Reduction

57 Dissolved sulfate and hydrocarbons are thermodynamically unstable together in virtually all diagenetic environments. Hence, redox-reactions occur, whereby sulfate is reduced by hydrocarbons either bacterially (bacterial sulfate reduction = BSR) or inorganically (thermochemical sulfate reduction = TSR). Their geological and economically significant products are similar and can be formed in or across several burial zones. Based on empirical evidence, BSR and TSR occur in two mutually exclusive thermal regimes (Fig. 11): low-temperature and high-temperature diagenetic environments, respectively (Machel et al., 1995a; Machel, 2001).

58 BSR is common in diagenetic settings from zero to about 80°C. Above this temperature range, almost all sulfate-reducing microbes cease to metabolize, although they may persist in a dormant state. Those few types of hyper-thermophilic microbes that can form H₂S at temperatures up to 110°C appear to be very rare and they do not normally occur and/or metabolize in geologic settings that are otherwise conducive to BSR.

59 TSR appears to be common in geologic settings with minimum temperatures in the range of 100 to 140°C, while in some settings minimum temperatures of up to 180°C are necessary. TSR does not have a sharply defined minimum temperature because the onset and rate of TSR are governed by several factors that vary from place to place. These factors include the composition of the available organic reactants, kinetic inhibitors and/or catalysts, anhydrite dissolution rates, wettability, as well as migration and diffusion rates of the major reactants toward one another (Machel, 1998).

Display large image of Figure 10

Display large image of Figure 11

60 The most important redox-reactions involving BSR and TSR can be represented by a common set of mass and charge balance reactions (Table 1). All reactions representing inorganic processes are reversible but are shown only in that direction to which they proceed under the specified conditions. Each microbial reaction involves numerous enzyme-catalyzed steps that are omitted for clarity. Reactions 1 and 2 are specific to BSR and TSR, respectively, whereas most redox-reactions involving hydrocarbons and inorganic sulfur compounds are valid for both BSR and TSR. For a more comprehensive discussion of the geological, experimental and theoretical evidence for these reactions, see Machel (1987) and Machel et al. (1995a).

61 The main organic reactants for BSR are organic acids, methane, and other products of aerobic or fermentative biodegradation. The main organic reactants for TSR are branched and nalkanes, followed by cyclic and monoaromatic species, in the gasoline range. Sulfate is derived almost invariably from the dissolution of primary or secondary gypsum and/or anhydrite at or near the redox-reaction site(s).

62 The products of BSR and TSR are similar, but their relative amounts vary widely and are determined by a number of locally variable factors, including availability of reactants, formation water chemistry, and wettability (Machel, 1997, 2001; Machel et al., 1995a). The primary inorganic reaction products in both thermal regimes are H₂S (HS ^- ) and HCO₃ ^- (CO₂ ). Carbonates, particularly calcite and dolomite, commonly form if alkali-earth metals are present in the formation waters. Other carbonates, i.e., ankerite, siderite, witherite, strontianite, may form if the respective metal cations are available. Iron sulfides, galena, and sphalerite form as by-products of hydrogen sulfide generation, if the respective transition or base metals are present or transported to the BSR/TSR reaction site. Elemental sulfur may accumulate as a volumetrically significant net reaction product, generally when the system runs out of reactive hydrocarbons. Water may form as a by-product and can cause a local dilution of the formation waters at or near the reaction site. There are case studies of TSR, however, where no dilution of the formation water has occurred, indicating that the amount of water released during TSR was negligible. Porosity may be generated during TSR, but most case studies show that TSR does not normally increase porosity significantly.

63 TSR is likely to take place in fairly narrow reaction zones, where the irreducible water saturation in the hydrocarbon-containing pores is low. In the Nisku Formation of Alberta, Canada, this zone is about 10 - 20 m thick (Machel et al., 1995b). However, where the irreducible water saturation is high, TSR may take place throughout the entire hydrocarbon-containing pore volume. Solid bitumen may form as a byproduct of both BSR and TSR. Finally, the net reaction may be exothermic or (more commonly) endothermic, depending on reaction stoichiometry (Simpson et al., 1996; Simpson, 1999).

64 The mere presence of any of the above reaction products and by-products does not permit a distinction between BSR and TSR. However, a number of petrographic relationships and geochemical criteria can be used to discriminate between these two processes (Machel et al., 1995a). Specifically, most solid products of BSR and TSR, although similar in gross composition, can be distinguished from one another petrographically. Most BSR products form early, whereas most TSR products form late, in the paragenesis, and have distinctive crystal sizes, shapes, and/or reflectivity. However, there also are cases where BSR products formed relatively late diagenetically, such as in uplifted reservoirs after hydrocarbon migration. Gas chromatography, δ¹³ C-, δ³⁴ S-analyses, and/or a combination thereof, offer the best distinguishing geochemical criteria (Machel et al., 1995a).

Table 1. Reaction scheme for bacterial and thermochemical sulfate reduction (BSR and TSR, respectively). Reactions 1 and 2 appear to be pre-requisites for BSR and TST, respectively. Reactions 3 to 5 exemplify all important redox steps for both BSR and TSR. The net reaction represents reactions 3, 4, and 5 if all elemental sulfur generated in reaction 4 is used up in reaction 5. See text for further explanation.

Display large image of Table 1

Dolomitization

65 One of the most important effects of groundwater flow on carbonate aquifers and potential reservoir rocks is pervasive replacive dolomitization, which commonly but not necessarily results in a concomitant increase in porosity and permeability. Mass balance calculations indicate that large fluid fluxes are necessary for extensive, pervasive dolomitization, because of the generally low Mg-content of natural waters. From 6,500 to 10,000 m³ of brackish to fresh groundwaters (the former containing about 10% seawater) would be required to dolomitize one cubic metre of limestone with an initial porosity of 40%, or 650 m³ of seawater, the most common Mg-rich water in nature, would be required (Land, 1985). Porosity generally increases during dolomitization because generally about two moles of calcite are replaced by one mole of dolomite, with a concomitant decrease in the molar volume by about 13%. Other reaction stoichiometries, however, may result in little porosity change or in porosity loss during dolomitization (e.g., Machel and Mountjoy, 1986). Also, continued flux of fluids supersaturated with respect to dolomite may lead to cementation of pore spaces with dolomite after complete matrix replacement, a process termed "overdolomitization "(Lucia 2000). In extreme cases, the resulting dolostones are essentially non-porous and impermeable.

66 Massive dolostones with enhanced porosity and permeability relative to their limestone precursors are common all over the world. Popular dolomitization models (Fig. 12) typically depict limestones embedded in some type of local, intermediate, or regional groundwater flow system that is invoked to pump the necessary amounts of Mg through the rocks (various articles in Purser et al., 1994; Mountjoy et al., 1997; Machel, 2004). Thereby, the distribution, texture, and geochemical composition of the dolostones are fitted to perceived or possible models, in an attempt to obtain a viable genetic interpretation. Such models span all diagenetic settings from near-surface to the metamorphic realm, as well as almost any type of hydrologic flow system and geochemical composition.

Microbial Action

67 Most microbes live in near-surface or shallow subsurface diagenetic settings where temperatures and pressures are relatively low and nutrients are abundant. Temperature and pressure increase with increasing depth and nutrients become scarce, which leads to a general decrease in the rates of metabolic activity and eventually to death of all microbes. The various types of microbial activity and products in petroliferous subsurface settings have recently been summarized by Machel and Foght (2000), in an attempt to place absolute depth limits on the geologically significant groups of microbes.

Display large image of Figure 12

68 Four groups of microbes are known to be geologically significant in petroliferous subsurface settings (Machel and Foght, 2000), namely aerobic respiratory bacteria, and three anaerobic groups that commonly live in consortia (symbiotic communities): fermentative, sulfate-reducing, and methanogenic bacteria. These microbes form a number of economically important products and by-products in subsurface settings. The aerobic microbes, generally called biodegrading bacteria, form mainly napthenic crude oils and tar in the form of tar sand deposits (Tables 2 and 3). The role of fermentative microbes is mainly in the partial breakdown of organic molecules that then serve as nutrients for the sulfate reducers and the methanogens. Sulfate-reducing microbes form mainly H₂ S, metal sulfides, and elemental sulfur, wheras methanogenic bacteria form mainly methane. In all cases, carbonate cements with distinctive isotopic compositions may be formed as by-products. In addition, nanobacteria may be important in clay mineral diagenesis in buried sandstones. Various types of microbes from the above groups can be used for microbially enhanced recovery of oil. Ultramicrobacteria constitute a special class, as they are injected in a dormant state and then resuscitated in situ to form biobarriers.

69 In the subsurface, temperature appears to be the principal factor limiting microbial metabolism (and life), followed by the availability of suitable nutrients. All four groups of bacteria generally become inactive by a subsurface depth of about 2000 m, except for rare cases where aerobic biodegradation or sulfate reduction can occur as deep as about 3000 m (Machel and Foght, 2000). Sulfate-reducing bacteria appear to be able to tolerate the highest temperatures, up to at least 110°C. However, they normally are subject to the same depth limitations as the aerobic biodegraders (Fig.13), whose "waste" products they may use as nutrients. Similarly, most methanogenic bacteria appear to become dormant or are dead at depths greater than about 2000 m (Fig. 14). This, however, does not preclude that sulfate-reducing and/or methanogenic microbes may metabolize in rare and unusual cases at greater depths. Furthermore, the lower limit of the biosphere probably is marked by other types of bacteria or archeae at much greater depths.

Porosity and Permeability in Limestone-Dolostone Sequences

70 Porosity and permeability development transgresses all diagenetic zones and is a direct result rather than a process of diagenesis. For petroleum engineering, this development is the most important aspect, especially when comparing limestone with dolostone sequences.

71 It has long been claimed that most dolostones are more porous and more permeable than limestones (Van Tuyl, 1914; Blatt et al., 1972), a circumstance of obvious importance for the petroleum industry. A related aspect is that dolostones generally form aquifers and preferential migration pathways for hydrocarbons, or both. However, Schmoker and Halley (1982) and Halley and Schmoker (1983) demonstrated that many dolostones have porosities equal to or less than those of adjacent limestones, based upon porosity/depth profiles of Cenozoic carbonates in southern Florida, which have not been buried too deeply (less than 1 km). Budd (2001) showed that the same is true for the permeability in these particular rocks, i.e., many dolostones have permeabilities equal to or less than those of the adjacent limestones. On a larger scale, Schmoker et al. (1985) compared thousands of limestones and dolostones from across the U.S.A. and found that dolostone reservoirs commonly have lower matrix porosities and permeabilities, yet higher fracture porosities and permeabilities, than limestones. Amthor et al. (1994), however, in a study of 31 wells from Devonian reservoirs in Alberta that span a depth range of several thousand metres, found that there is a distinctive dependence on depth. If considered irrespective of depth, limestones and dolomitic limestones are more porous than dolostones, whereas at burial depths of greater than 2000 m dolostones are significantly more porous and permeable than limestones. There also are notable examples of very young, near-surface dolostones that are tight and apparently devoid of porosity, as in the Plio-Pleistocene carbonates of Bonaire (Lucia and Major, 1994).

Table 2. Aerobic biodegradation effects on gross properties of petroleum (modified from Connan, 1984).

Display large image of Table 2

Table 3. Changes in molecular properties with increasing level of aerobic biodegradation (compiled from Volkman et al., 1983; Connan, 1984; Hunt, 1996).

Display large image of Table 3

72 The major processes involved in porosity-permeability (poroperm) generation are summarized in Figure 3. Six specific processes are especially important in the case of limestone-dolostone sequences: (a) mole-per-mole replacement; (b) dissolution of unreplaced calcite (solution undersaturated for calcite after all Mg in excess of dolomite saturation is exhausted); (c) dissolution of dolomite (without externally controlled acidification); (d) acidification of the pore waters (via decarboxylation, clay mineral diagenesis, etc.); (e) fluid mixing (Mischungskorrosion); and, (f) thermochemical sulfate reduction, which may generate porosity under certain circumstances (Machel, 2001, 2004). The wide scatter and lack of systematic relationships between the porosity of limestones and dolostones in Florida and Alberta probably reflects locally and regionally heterogeneous interplays among the various processes that generate, preserve, or destroy porosity (Fig. 3).

73 Dolomitization almost invariably involves the reorganization of permeability pathways. Commonly permeability increases along with porosity, and vice versa. This is documented through studies such as the one on the Late Devonian Grosmont Formation in eastern Alberta, which hosts a giant heavy-oil reservoir (Luo et al., 1994; Luo and Machel, 1995; Machel and Huebscher, 2000). A comparison of porosity and permeability data from 237 core plugs reveals an overall positive correlation, despite considerable scatter. This correlation is also expressed in the displacement pressures from mercury injection capillary measurements, which permit the identification of four major and two minor dolomite reservoir rock types. Using a similar approach (thin section and SEM petrography combined with helium porosimetry and mercury injection capillary measurements), Woody et al. (1996) also documented positive and statistically significant correlations between porosity and permeability for planar dolomites in the Cambrian-Ordovician Bonneterre Formation of Missouri, U.S.A., which is host to one of the world's largest MVT-sulfide deposits.

Display large image of Figure 13

Display large image of Figure 14

74 Other authors have disputed that there is a systematic correlation between porosity and permeability in dolostones, or that these two petrophysical parameters are enhanced in dolostones relative to limestones. Halley and Schmoker (1983), in the absence of reliable or sufficient permeability data, attempted to assess the permeability of carbonate rocks from porosity data. They found that carbonate aquifers and carbonate aquicludes cannot be distinguished on the basis of porosity. Lucia (2002, 2004) claimed that "....there is no relationship between porosity and permeability in dolostones...... and dolomite crystal sizeand the precursor fabric are key elements in predicting permeability", and "Dolomitization of grain-dominated limestones usually does not change porosity-permeability relationships. Instead, the precursor fabric controls pore-size distribution." While this may be so in some, perhaps many cases, the Grosmont and the Bonneterre examples clearly show that there is a relationship between porosity and permeability in at least some major and economically important dolostone sequences.