Silicification of originally non-siliceous sediments affects a wide variety of rock-types and materials and ranges from minor to pervasive. Partial and minor chertification occur mostly in Phanerozoic carbonates, carbonate-bearing sandstones, evaporites, and fossil wood. The source of the silica is predominantly biogenic. In petrified wood the silicification mechanism is a permeation or void-filling process, not a replacement. In this example, the sequence of silica-phase transformation is the same as that in deep-sea siliceous sediments. In many silicified rocks, particularly in certain carbonates, the transformation sequence is different from that in radiolarites or diatomites. The chemical environment and conditions of early diagenetic chert formation in shallow water carbonates are delineated by the general mixing model of Knauth (1979), but remain unknown for most other types. An exception are the flint nodules and bands of the English Chalk. A detailed geochemical study of the paramoudra flint structures by Clayton (1986) provided remarkable insight into the replacement process. Seven different recurring silica fabrics have been recognized in chert-replaced carbonates including equigranular (microcrystalline quartz or microquartz and megaquartz) and fibrous types (chalcedony, quartzine or length-slow chalcedony, lutecite, zebraic chalcedony and microflamboyant quartz). Among the latter, quartzine and microflamboyant quartz are common in, but by no means restricted to chert-replaced evaporites, for which Milliken (1979) recognized a sequence of seven quartz-fabrics. As a single criterion, only anhydrite inclusions in megaquartz, quartzine or microflamboyant quartz provide unequivocal evidence for an evaporite precursor. The relative timing between silicification and well-established diagenetic carbonate reactions shows that virtually all theoretically possible sequences occur. Chertification of carbonate host sediment thus may take place during early, intermediate or late diagenesis, and even during anchimetamorphism. Pervasive to complete silicification has been described in lacustrine, pedogenic and hydrothermal volcanogenic rocks and occurs on the scale of individual beds, members or entire formations. It may affect some of the aforementioned rock types too, for example carbonates. The source of the silica is predominantly inorganic. Chert formation in these environments includes direct chemical silica precipitation from solution through a gel stage. Magadi-type cherts result from the conversion of the hydrous sodium silicate magadiite into microcrystalline quartz which has occurred in East African rift valley lakes such as Lake Magadi in Late Pleistocene time. They are closely related to directly precipitated inorganic cherts which have been observed in alkaline environments of playa lakes. Silcretes originate from weathering and soil-forming processes under climatic and environmental conditions conducive to the formation of duricrusts and laterites. In more humid climates, however, non-weathering-profile silcretes occur which have been distinguished from weathering-profile silcretes. Four to five different silcrete fabric types have been established, including the (1) grain-supported (or ‘quartzitic’), (2) floating (terrazzo), (3) matrix (Albertinia and opaline), and (4) conglomeratic types. In volcanic edifices, silicification related to hydrothermal activity occurs (1) along the ascent-routes of the fluids (vents) within the volcanic complexes, (2) in isolated ponds and depressions on the sea-floor where the fluids discharge, mostly in the median rift valley of mid-ocean ridges, and (3) in geothermal areas on land associated with spreading lineaments, volcanic islands arcs, transform faults or continental hot spots. Attention has focussed on the associated base-metal concentrations and less on the silicification and accompanying iron enrichment processes. The Precambrian Banded Iron Formations may be indirectly or directly related to hydrothermal-volcanic processes. The origin of these cherty iron formations certainly requires more than one genetic model and at present remains a major unresolved problem despite massive efforts to tackle it.
Read full abstract