2. Geologic setting

The spherulitic and cone-in-cone concretions described here occur on Île-aux-Grues/Île-aux-Oies (Crane Island/Goose Island), a mid-stream island in the St Lawrence Estuary, 80 km northeast of Quebec City, Canada (Fig. 1), where the white Canadian geese stop over twice a year on their migration to and from the Arctic. The island is made up of a succession of Cambro-Ordovician turbidites and black shales of undetermined stratigraphic age. It consists of a lower unit of thin-bedded (less than 10 cm thick) spill-over turbidites (fine-grained sandstone and siltstone alternating with beds of black shale). Beds of this unit are in fault contact with higher stratigraphic units and are well exposed on the broad tidal flats that surround most of the island (Fig. 1, inset). A succession of red and green slumped siltstones follows the spill-over unit. Overlying the slump unit is another spill-over turbidite succession of fine sandstone and siltstone beds that alternate with thin beds of black shale. A number of channels occur throughout this unit. Channel-fill deposits at the top of the second spill-over turbidite represent the highest stratigraphic unit on the island. The carbonate concretions occur in the lower spill-over unit. They are usually arranged in layers following distinct stratigraphic horizons within the black shales. No correlation has been found between a particular type of concretion and the lithology of the host rocks.

Fig. 1. Eastern Canada geography with Île-aux-Grues/Île-aux-Oies in the St Lawrence Estuary about 80 km northeast of Quebec City, Canada, shown as inset. Stars – sample locations; stippled area – tidal flats.

4. Morphology and petrography of concretions

Among the three groups of carbonate concretions, the following distinctive types are recognized: (1) end-member micritic concretions; (2) transition of micritic concretions into spherulitic exterior shell (Fig. 2a); (3) micritic concretions with a cone-in-cone rind; (4) end-member spherulitic concretions (Fig. 2b); (5) spherulitic concretions with cone-in-cone rims; (6) end-member cone-in-cone concretions (Fig. 2c); (7) spherulitic lenses; (8) bedded cone-in-cone (Fig. 3a, b).

Fig. 2. (a) Transition of micritic concretion centre into exterior shell of spherules. (b) End-member spherulitic concretion. Two black shale seams occur in the upper half forming notches on the periphery of the concretion. Left half of thin-section is not stained. Numerous small cone-in-cone structures form rims on the spherules that appear light owing to their calcitic nature and absence of quartzine. (c) End-member cone-in-cone concretion. Left half of thin-section stained with Alizarin Red-S revealing the ferrous nature of the calcite. Note the uninterrupted zig-zag patterned seam of black shale that resulted from displacement by the growing cone-in-cone calcite structures. (d) End-member cone-in-cone concretion. Left half of thin-section: carbonate dissolved by HCl showing quartzine fibres (white). Scale in centimetres.

Fig. 3. (a) Two layers of cone-in-cone structure with a wavy boundary. The growth direction is inward. Thickness of the sample is c. 5 cm. (b) ‘Beef’ calcite cone-in-cone structure. Note cross-lamination in overlying fine sandstone bed. Scale in centimetres.

The concretions range in size from 7 to 40 cm in diameter. Thin beds of ‘beef’ cone-in-cone structures are usually laterally extensive (Fig. 3b). Spherulitic lenses reach a thickness of up to 20 cm and a length not exceeding 1 m.

Exposed surfaces of concretions or lenses reflect their internal structures. The surfaces of micritic concretions and those with a spherulitic rim are usually smooth. Spherulitic concretions are rather warty looking. This is the result of numerous spherules protruding irregularly a few millimetres beyond the surface into the host sediment (Fig. 4). Concretions with an outer shell of cone-in-cone structure invariably exhibit characteristic conic scales. Furthermore, on both spherulitic and cone-in-cone concretions there are bedding-plane parallel notches that correspond to either a change from spherulitic to a cone-in-cone structure or a thin layer of silt or shale (Fig. 2b). Concretions composed entirely of cone-in-cone structure commonly are shaped like a jelly doughnut. The two polar regions may be depressed and underlain and overlain by black shale (Fig. 2d). On exposed vertical faces of bedded cone-in-cone structures, the characteristic ‘teepee’ forms are well exhibited as a result of differential weathering.

Fig. 4. 3D view of end-member spherulitic concretion showing a warty surface reflecting the make-up of the concretion by myriads of spherules. Scale: 30 cm.

Seams and stringers of organic-matter-rich black shale and silt are common in the concretions. These seams are more or less straight and little affected by concretionary growth. Internal sedimentary structures of the host sediment, which include parallel lamination and rare small-scale ripple cross-lamination, are well preserved (Fig. 3b). In a few spherulitic concretions silt stringers inhibit the growth of spherules (Fig. 5a).

Fig. 5. (a) Silt stringers inhibiting the growth of spherules. Microscopic view under crossed nicols displaying the extinction cross in the radiating carbonate fibres of the spherules. Field of view is c. 1 cm wide. (b) Close-up of near-perfect spherical spherule with some compromise boundaries under crossed nicols showing extinction cross. Spherule is 2 mm in horizontal diameter.

Under the microscope the spherules are composed of an intimate intergrowth of fine fibres of calcite and quartzine. The presence of fine silica fibres was not identified in XRD, but recognized in the larger-area SEM image mosaics and in the EDS analyses (Figs 6, 7, 8).

Fig. 6. Large-area image mosaic of the carbonate concretion showing the plane-polarized light image mosaic and the upper half from which a large-area Atlas 5 SEM (BSE) image mosaic was acquired.

Fig. 7. Higher-magnification SEM-BSE images and EDS element maps of the part of the concretion from which the carbonate was removed showing the distribution of the quartz fibres and the deformed shale bands.

Fig. 8. Higher-magnification SEM-BSE images of the concretion. (a) Structure of the quartz fibres. (b) EDS point analyses of location 1 in (a). (c) Deformed part of a shale band consisting almost entirely of chlorite. (d) EDS point analyses of location 2 in (c). (e) Area in the shale that shows framboidal pyrite and xenotime. (f–i) EDS point analyses of locations 3 to 6 in (e).

The SEM image mosaic allowed for the investigation of an entire half of a large concretion. The internal structure could be investigated at higher resolution and the chemistry of the components could be determined. The PPL image mosaic of the entire thin-section shows the various components of the carbonate concretion (Fig. 6; Atlas 5 BBV dataset https://www.petapixelproject.com/mosaics/RH/CC-BBV/Carbonate-Concretion/index.html). The carbonate appears pink in the stained area of the thin-section. The deformed shale bands appear as a brown to dark brown colour. The area of the thin-section that had undergone acid treatment was imaged with the SEM (Fig. 6). The area that was occupied with carbonate before that treatment now appears as black empty voids within the network of quartz fibres and deformed shale bands (Fig. 6; Atlas 5 BBV dataset https://www.petapixelproject.com/mosaics/RH/CC-BBV/Carbonate-Concretion/index.html). These areas appear white in PPL since the light is transmitted through the underlying thin-section glass and remaining glue that holds the actual thin-section (see Atlas 5 BBV dataset https://www.petapixelproject.com/mosaics/RH/CC-BBV/Carbonate-Concretion/index.html).

The EDS element maps impressively show the distribution of the quartzine fibres and the deformed shale bands (Fig. 7; see also EDS maps in Atlas BBV dataset https://www.petapixelproject.com/mosaics/RH/CC-BBV/Carbonate-Concretion/index.html). A close investigation of the quartzine fibres reveals that they are formed of various shapes that look like casts of carbonate crystals and former cracks in the carbonate that were filled with quartzine, which were left behind after the carbonate was removed by etching with acid (Fig. 8a). The shale bands are composed of various types of phyllosilicates, chlorite, pyrite, quartz grains, titanium oxides, sodium feldspar, and in some cases xenotime (Fig. 8; see also linked EDS maps in the Atlas 5 BBV dataset https://www.petapixelproject.com/mosaics/RH/CC-BBV/Carbonate-Concretion/index.html). The carbonate to silica ratio in the spherules is c. 60:40; at the transition to the cone-in-cone structure, silica disappears.

Spherules in these carbonate concretions range in size from 0.5 mm to 12 mm in diameter. Spherules of larger size usually exhibit a perfect spherical shape (Fig. 5b), whereas medium- to small-sized spherules up to 3 mm in diameter are commonly tightly packed and may exhibit a shoebox shape as well as other irregular shapes (Fig. 9). The nuclei of the spherules are commonly very small and their identity cannot be deciphered. From the nuclei, numerous fine fibres of calcite and quartzine 2 to 6 μm in diameter radiate outwards. Under crossed nicols a characteristic black extinction cross is displayed by the spherules (Fig. 5). Within a given concretion, spherules in the same horizontal plane are usually of equal size. When adjoining spherules come into contact they share a compromise boundary (Fig. 9). Spherules with compromise boundaries depart in shape from the ideal spherical form. Spaces surrounding three or more spherules may be filled with patches of black shale that have been displaced into this space by the growth of the neighbouring spherules. In spherulitic concretions the total volume of black shale or silt is <20 %. At the contact between spherules and black shale a transition from spherule to cone-in-cone structure is commonly observed (Fig. 10). Where cone-in-cone structure follows the periphery of a spherule, the high concentration of quartzine fibres in the inner parts of the spherule is reduced and gives way to clear ferroan calcite (Fig. 11). Calcite in these cone-in-cone structures frequently exhibits distinctively curved twin lamellae and cleavages characteristic of fascicular optic and radiaxial mosaic (Kendall, Reference Kendall, Schneidermann and Harris1985; Fig. 12a). In some cases, the cone-in-cone structure appears shaped like a bottle brush (Fig. 12b). The foregoing observations document the intimate relationship between the spherulitic and cone-in-cone structures.

Fig. 9. Tightly packed spherules with compromise boundaries.

Fig. 10. Photomicrograph of cone-in-cone structure rimming spherules. The high concentration of quartzine fibres in the inner parts of the spherules is reduced and gives way to clear ferroan calcite in the rims of the spherules. Hand specimen 79072802-1. Horizontal width c. 36 mm.

Fig. 11. Close-up of cone-in-cone structure in the periphery of two spherules.

Fig. 12. (a) Calcite in cone-in-cone structure exhibiting distinctively curved twin lamellae and cleavages characteristic of fascicular optic and radiaxial mosaic (c.f. Kendall, Reference Kendall, Schneidermann and Harris1985). (b) Cone-in-cone structure shaped like a bottle brush.

In the present study, all of the examined cone-in-cone structures are made up of fine fibres of ferroan calcite. The smallest basic construction units in a cone-in-cone structure are slender conical tufts of fine fibres of calcite, which form various fan shapes with apical angles ranging from 15^o to 70^o (Figs 10, 11). The length of the fibrous tufts ranges from 0.5 to 15 mm. When three-dimensionally reconstructed, these slender conical tufts should be conical bundles of fine calcite fibres. The characteristic clay rings on cone-in-cone structures are the result of growth of fibrous tufts. As correctly interpreted by Franks (Reference Franks1969), the clay patches were pushed aside by the slender conical tufts until the clay patches butt against the sheaths of an earlier cone.

5. Carbon and oxygen stable isotopes

The carbon isotope values for the various types of concretions range from −28 ‰ to −7 ‰ (VPDB) (Fig. 13a); the oxygen values show a narrow spread between c. −17 ‰ and −9 ‰ (VPDB) with one runaway value of −29 ‰ (VPDB) (Fig. 13b).

Fig. 13. (a) Carbon and (b) oxygen isotope traverses across five different types of spherulitic and cone-in-cone concretions. XA – spherules grading downwards into cone-in-cone structure; XB – spherulitic concretion; XC – cone-in-cone structure on rim of spherulitic concretion; XD – micritic concretion with spherule rim; XE – cone-in-cone concretion.

6. Interpretation: mechanism of formation of spherules and cone-in-cone structures

Since the classic monograph of Bernauer (Reference Bernauer1929), who described experiments in which spherulitic growth was induced, a number of examples of successful experimental precipitation of fibrous spherules in the laboratory have been reported (Morse et al. Reference Morse, Warren and Donnay1932; Brooks et al. Reference Brooks, Clarke and Thurston1950; Heinisch, Reference Heinisch1970; Schwartz et al. Reference Selles-Martinez1971; McCauley & Roy, Reference McCauley and Roy1974). As initially suggested by Bernauer, most authors attribute the precipitation of spherules to two factors: the presence of impurities and one form or another of a gel as the medium of crystallization.

Based on modern concepts of crystal growth, Keith & Padden (Reference Keith and Padden1963) developed a phenomenological theory of spherulite crystallization from a viscous magma, emphasizing the role of impurities in the spherulite-forming melt. Once impurities are added to a melt, the rate of advance of a crystallization front no longer depends upon the diffusion of latent heat away from the growth site alone, but rather upon an interplay between the transport of heat and the diffusion of impurities. The growing crystal rejects the impurities preferentially, and the concentration of the impurity in the liquid adjacent to the crystal builds up to higher than the average value in the melt. If the melt is not disturbed, an impurity-rich layer of ‘thickness’ δ = D/G forms in the liquid at the interface with the growing crystal, where D is the diffusion coefficient for the impurity in the melt, and G the growth velocity of the crystal. When the ambient temperature of the impurity-rich layer is smaller than the equilibrium liquidus temperature, small projections can be formed at the interface without additional latent heat having to be transported away. The surface of the growing crystal becomes broken up into numerous hexagonal cells. In this cellular crystallization, impurities are collected in the grooves between the hexagonal cells.

In a spherulite-forming melt, D values are reduced by several orders of magnitude, as evidenced by high melt viscosity. Together with very low G values, small δ values result (in most cases <10⁻⁴ cm or <1 μm). Since the equilibrium liquidus temperature gradient in the impurity-rich layer is steep, the surface of a hypothetical growing sphere is rendered extremely unstable, and cellular crystallization is to be expected. In the spherule-forming system, cellulation advances to an extreme and fibres grow rapidly on the sites of the hexagonal cells while impurities fill the grooves. In other words, the properties of spherulite-forming melts are such that the surface of a hypothetical growing spherule is forced to develop long fibrous projections separated one from another by uncrystallized melt. Owing to the lower temperature gradient in the impurity-rich layer, the uncrystallized melt will solidify slowly through secondary crystallization to fill in the overall structure. In this theory, Keith & Padden (Reference Keith and Padden1963) showed that the growth of fibrous crystals in spherulite-forming melts is attributable to cellulation and fibrillation on a fine scale, and brought about by a diffusive segregation of impurities in the manner described above.

Thus far, no consideration has been given to the curvature of the spherical surface and divergence of the cellular fibres. For the hypothetical growing sphere to complete its development into a spherulitic habit, the fibrous crystals should possess the ability to branch non-crystallographically at a small angle. The former authors stated that it is a fundamental property of cellular crystallization in an impure melt that any perturbation of the surface profile on a growing crystal is stable and persistent only if its size is commensurate with δ. Small disordered regions or minute singularities of the growing crystal have an opportunity to act as sources of persistent new growths only if δ is very small. Impure melts having high viscosities meet this condition. As δ is decreased from a relatively large value in such a melt, there is at first little likelihood of non-crystallographic branches being developed. As δ decreases, there is an increasing probability that one of the larger singularities at the shoulder of a growing fibre is of sufficient size that, with further growth, it becomes a persistent surface feature. If this new growth has a crystallographic orientation departing slightly from that of the parent fibres, it gives rise ultimately to a new fibre, which diverges from the original. Thus, one initial fibre has split into two fibres, each about δ in width and each growing along the same preferred crystal axis but misaligned slightly with respect to the other. As δ decreases further, the probability of such an occurrence increases correspondingly and, when δ is very small, almost any island of surface disorder becomes a possible source of non-crystallographic branching. As far as branching is concerned, the essential difference between crystallization in simple liquids and in spherulite-forming melts is in the magnitude of δ. Only in spherulite-forming melts do sufficiently small diffusion coefficients exist to ensure values of δ that are small enough for branching to become predominantly non-crystallographic. Keith & Padden (Reference Keith and Padden1963) observed that in polymers where δ was relatively large, spherules with coarse fibres and open textures were formed. Although the phenomenological theory of spherulite crystallization was developed for melts, it should be applicable to precipitation from pore solutions in clastic sedimentary environments as it depends critically only on the numerical value of the δ parameter and the presence of two chemical species, where one represents an impurity that temporarily poisons the growing crystal surface.

The intimate association of quartzine and calcite in carbonate concretions from Île-aux-Grues/Île-aux-Oies at first sight seems hard to explain in terms of ordinary precipitation processes of these minerals. However, applying the phenomenological theory of Keith & Padden (Reference Keith and Padden1963), a convincing solution to the problem emerges.

Most spherules precipitated experimentally in the laboratory crystallized in a gelatinous medium in the presence of impurities. Silica gel is the most common medium used. The occurrence of silica gel has not been reported from clastic or biogenic siliceous pelagic deep-sea sediments (e.g. Hesse, Reference Hesse, McIlreath and Morrow1990a,b), not even in hydrothermally sourced ponds of mid-ocean ridges like the Atlantic II Deep in the Red Sea. However, experimental spherulite precipitation has also been achieved from alkaline aqueous solutions (e.g. Johnston et al. Reference Johnston, Merwin and Williamson1916). Alkaline conditions are established during early diagenesis in the lower sulfate-reduction zone (Hesse et al. Reference Hesse, Shah and Islam2004). If sufficient biogenic or detrital calcium carbonate has been dissolved in the oxidation zone, and the nitrate-reduction and upper sulfate-reduction zones where acidic conditions prevail, then fibrous calcium carbonate precipitation in the presence of impurities may start in the lower sulfate-reduction zone or the carbonate-reduction (methane generation) zone. The presence of ferroan calcite in the studied concretions speaks for the beginning of the precipitation in the carbonate-reduction zone, because in the sulfate-reduction zone Fe²⁺ is removed by sulfide and not available for incorporation into carbonate. The carbon isotope results (Fig. 13a) support an early diagenetic precipitation at the top of the methane-generating zone at a level where the effect of methane fractionation is still subordinate. (Note: C. Fong considers that precursor concretions formed at shallow subsurface depth, and their recrystallization leading to spherulitic and cone-in-cone structures occurred at depths in the order of 4 km after tectonic stacking.) No systematic variations between spherules (XB) and cone-in-cone structures (samples XC and XE) are recognizable. The oxygen isotope values would give temperatures too high even if a δ¹⁸O of Early Palaeozoic seawater of <−7 ‰ is assumed (Fig. 13b; cf. Hesse et al. Reference Hesse, Shah and Islam2004). However, low δ¹⁸O values may result from advection of pore waters from a greater subsurface depth (Hesse & Schacht, Reference Hesse, Schacht, Hünecke and Mulder2011).

Carbonate precipitation may occur alternating with silica precipitation, because carbonate crystallization lowers the pH (García-Ruiz et al. Reference García-Ruiz, Melero-Garcia and Hyde2009). Under alkaline conditions dissolved silica is present, depending on pH, either as orthosilicic acid (H₄SiO₄) or in its dissociated forms H₃SiO₄⁻¹ (for pH >9) or H₂SiO₄² (for pH >13). Dissociation raises the silica solubility significantly (Hesse, Reference Hesse, McIlreath and Morrow1990a). Silica precipitation from supersaturated solutions involves polymerization in which monomeric acid first forms oligomers (dimers, tetramers, etc.) through the development of siloxane (Si–O–Si) bonds and then high-molecular-weight polymers (with molecular weights up to 10<2>000) which have colloidal dimensions (greater than 5 nm) and remain in suspension as sols unless precipitated as gels by metal hydroxides. The most commonly used hydroxide for silica precipitation in industrial applications is Mg(OH)₂, also thought to be instrumental in the nucleation of opal-CT (Kastner et al. Reference Kastner, Keene and Gieskes1977). Presently known environments of inorganic silica precipitation are the playa lakes of the Coorong Lagoon, Australia, which have a pH of ~10, Lake Bogoria in the Kenya Rift Valley (Renaut et al. Reference Renaut, Tiercelin, Owen, Frostic, Renaut, Reid and Tiercelin1986; Renaut & Owen, Reference Renaut and Owen1988) and others. Here the silica probably goes through a gel stage. Precipitation takes place when the pH drops owing to decaying plant matter during evaporation. Conditions of the shallow lacustrine environment of the Coorong Lagoon would not have been applicable to the continental slope setting of Laurentia in Cambrian or Ordovician times.

However, as pointed out, pH-lowering through carbonate precipitation would have triggered silica precipitation. The presence of dissolved silica supplied by the dissolution of unstable biogenic opal-A (in Palaeozoic time provided by radiolarians and sponge spicules) during shallow burial leads to local supersaturation of silicic acid next to the growing carbonate crystals, initiating the adsorption and precipitation of silica on the positively charged carbonate crystal surfaces (Fig. 8a). Silica precipitation poisons carbonate crystal growth and causes bifurcation of the carbonate crystals. The chemical coupling of carbonate and silica precipitation enables continuous crystal splitting at non-crystallographic angles, thus producing the radiating spherulitic fibres. If the silica supersaturation in the pore water is low enough, direct precipitation of the silica as stable chalcedony or quartzine is possible instead of the metastable precursor opal-CT phase.

In the concretions described, the cement (represented by the crystallized spherules) to solid ratio is as high as 80:20. The type of sediment at the depth of concretion growth is a soupy mud with high water content (Müller, Reference Müller, Larsen and Chilingar1967). The presence of calcite nuclei dictates that calcite is the initial mineral precipitated, while silica, although in excessively high concentration, constitutes the impurity attributed to the crystallization of fibrous calcite. The silica that filled interfibre grooves underwent secondary crystallization and formed a fibrous network. In well-formed and fully developed spherules, these silica fibres dominate the entire network, consequently obscuring the manner of calcite branching of the calcite fibres. Could the silica have been introduced by replacement as in the case of silica replaced cone-in-cone structures reported by Lugli et al. (Reference Lugli, Reimold, Koeberl, Koeberl and Henkel2005)? This is next to impossible because of the fine fibrous intergrowth of ferroan calcite and silica in the spherules. The rimming cone-in-cone structure should then also have been replaced by silica.

Observation under the petrographic microscope reveals that transition from spherule to cone-in-cone structure is brought about by an increase in the angle of branching of the fibres. Since the transition is accompanied by the disappearance of silica fibres from the structure as confirmed by the SEM/EDS analyses, it appears that exhaustion of the silica reservoir is the cause of the change from spherulitic to cone-in-cone structure.

According to Lugli et al. (Reference Lugli, Reimold, Koeberl, Koeberl and Henkel2005), hypotheses concerning the formation of calcareous cone-in-cone structure belong to two groups: those assuming early displacive formation of concretions in soft, unconsolidated sediment, and those that emphasize late fracturing of concretions with or without excess pore fluid. In detail, the following hypotheses have been proposed (see also Mozley, Reference Mozley, Middleton, Church, Coniglio, Hardie and Longstaffe2003):

1. (1) Recrystallization of fibrous aragonite to calcite forming open cone fractures intruded by argillaceous matter (Tarr, Reference Tarr and Twenhofel1922; Gilman & Metzger, Reference Gilman and Metzger1967).

2. (2) Differential pressure solution on calcite fibres along conical shear planes induced by the weight of overlying strata (Tarr, Reference Usdowski1932).

3. (3) Settling and volume shrinkage during the slow dewatering of highly saturated and loosely packed subaerially exposed sediments (Shaub, Reference Shead1937).

4. (4) Precipitation of fibrous calcite along fractures or in concretionary masses. The fibres would then be deformed by ‘tractional forces’ during deformation of the host strata or during concretionary growth of the calcite component (Bonte et al. Reference Bonte, Denaeyer and Goguel1947).

5. (5) Crystallization of calcite fibres under gravity-induced differential stress, during gradual compaction of the sediment. The angles of cone apices would be determined by the plasticity of the enclosing sediment: large apical angles would indicate relatively low plasticity (Woodland, Reference Woodland1964).

6. (6) Early displacive growth of cone-shaped plumose aggregates of fibrous calcite (Franks, Reference Franks1969).

7. (7) Diagenetic water-escape structures promoting the ‘reordering of phyllitic elements within a sediment undergoing diagenetic compaction and which exhibits differences in porosity and competency in its lithological composition; cone-in-cone can be compared to a schistosity acquired in the realm of hydroplastic deformation’ (Becq-Giraudon, Reference Becq-Giraudon1990).

8. (8) Brittle fracturing of crystalline calcite aggregates, grown by a crack-seal mechanism in overpressured environments, and induced by a decrease in pore pressure of the plastic host sediment (Selles-Martinez, Reference Shaub1994).

9. (9) Shallow burial (<1500 m) precipitation from modified marine pore fluids (Hendry, Reference Hendry2002).

10. (10) Early diagenetic growth (depth <40 m), probably microbially mediated, in sandstone units beneath flooding surfaces and sequence boundaries (McBride et al. Reference McBride, Picard and Milliken2003).