K-Mn oxides of hollandite group minerals such as cryptomelane (K-1-2(Mn3+, Mn4+)(8)O(16)nH(2)O) are often precipitated authigenically in weathering profiles. The presence of structural K allows these minerals to be dated by the K-Ar and 40Ar/39Ar methods, making it possible to study the progression of oxidation fronts during weathering processes. Within the context of a recent 40Ar/39Ar study of cryptomelane from the Azul Mn deposit in the Carajas region (Amazonia, Brazil), Vasconcelos et al. (1994) defined three age clusters (65-69, 51-56, and 40-43 Ma) and proposed that they correspond to the episodic precipitation of the three generations of Mn oxide that have been identified in the deposit (Beauvais et al., 1987). We performed a laser probe 40Ar/39Ar and Rb-87/Sr-87 study on new samples from the same Mn deposit. Our 40Ar/39Ar data confirm that cryptomelane is a suitable mineral for 40Ar/39Ar dating, although in some cases we clearly identify the existence of 39Ar recoil effects. Although the corresponding age spectra are generally strongly disturbed, our results also confirm that the earliest cryptomelane generation is of Upper Cretaceous-Paleocene age. We obtained good plateau ages from veins and concretions of the second cryptomelane generation. Some of these results allow definition of a well-constrained age cluster at 46.7-48.1 Ma not observed by Vasconcelos et al. (1994). A petrographic study confirms that none of the samples analyzed in the present study contained material associated with the third generation of cryptomelane. We propose that these new results support the idea of a more or less continuous crystallization of K-Mn oxides, mainly constrained by local factors, rather than the model advanced by Vasconcelos et al. (1994), which suggests that each cryptomelane generation corresponds to distinct weathering events related to global climatic changes. Sr-87/Sr-86 data show large variations, clearly inherited from the 2.1 Ga parent rock of the Mn protore. The Rb/Sr results demonstrate that minimum fractionation occurs during cryptomelane crystallization, except for the latest generation, which is depleted in Sr. This precludes use of the Rb/Sr radiochronometer for dating secondary Mn oxides in laterites. K-Mn oxides of hollandite group minerals such as cryptomelane (K-1-2(Mn3+, Mn4+)(8)O(16)nH(2)O) are often precipitated authigenically in weathering profiles. The presence of structural K allows these minerals to be dated by the K-Ar and 40Ar/39Ar methods, making it possible to study the progression of oxidation fronts during weathering processes. Within the context of a recent 40Ar/39Ar study of cryptomelane from the Azul Mn deposit in the Carajas region (Amazonia, Brazil), Vasconcelos et al. (1994) defined three age clusters (65-69, 51-56, and 40-43 Ma) and proposed that they correspond to the episodic precipitation of the three generations of Mn oxide that have been identified in the deposit (Beauvais et al., 1987). We performed a laser probe 40Ar/39Ar and Rb-87/Sr-87 study on new samples from the same Mn deposit. Our 40Ar/39Ar data confirm that cryptomelane is a suitable mineral for 40Ar/39Ar dating, although in some cases we clearly identify the existence of 39Ar recoil effects. Although the corresponding age spectra are generally strongly disturbed, our results also confirm that the earliest cryptomelane generation is of Upper Cretaceous-Paleocene age. We obtained good plateau ages from veins and concretions of the second cryptomelane generation. Some of these results allow definition of a well-constrained age cluster at 46.7-48.1 Ma not observed by Vasconcelos et al. (1994). A petrographic study confirms that none of the samples analyzed in the present study contained material associated with the third generation of cryptomelane. We propose that these new results support the idea of a more or less continuous crystallization of K-Mn oxides, mainly constrained by local factors, rather than the model advanced by Vasconcelos et al. (1994), which suggests that each cryptomelane generation corresponds to distinct weathering events related to global climatic changes. Sr-87/Sr-86 data show large variations, clearly inherited from the 2.1 Ga parent rock of the Mn protore. The Rb/Sr results demonstrate that minimum fractionation occurs during cryptomelane crystallization, except for the latest generation, which is depleted in Sr. This precludes use of the Rb/Sr radiochronometer for dating secondary Mn oxides in laterites. K-Mn oxides of hollandite group minerals such as cryptomelane (K-1-2(Mn3+, Mn4+)(8)O(16)nH(2)O) are often precipitated authigenically in weathering profiles. The presence of structural K allows these minerals to be dated by the K-Ar and 40Ar/39Ar methods, making it possible to study the progression of oxidation fronts during weathering processes. Within the context of a recent 40Ar/39Ar study of cryptomelane from the Azul Mn deposit in the Carajas region (Amazonia, Brazil), Vasconcelos et al. (1994) defined three age clusters (65-69, 51-56, and 40-43 Ma) and proposed that they correspond to the episodic precipitation of the three generations of Mn oxide that have been identified in the deposit (Beauvais et al., 1987). We performed a laser probe 40Ar/39Ar and Rb-87/Sr-87 study on new samples from the same Mn deposit. Our 40Ar/39Ar data confirm that cryptomelane is a suitable mineral for 40Ar/39Ar dating, although in some cases we clearly identify the existence of 39Ar recoil effects. Although the corresponding age spectra are generally strongly disturbed, our results also confirm that the earliest cryptomelane generation is of Upper Cretaceous-Paleocene age. We obtained good plateau ages from veins and concretions of the second cryptomelane generation. Some of these results allow definition of a well-constrained age cluster at 46.7-48.1 Ma not observed by Vasconcelos et al. (1994). A petrographic study confirms that none of the samples analyzed in the present study contained material associated with the third generation of cryptomelane. We propose that these new results support the idea of a more or less continuous crystallization of K-Mn oxides, mainly constrained by local factors, rather than the model advanced by Vasconcelos et al. (1994), which suggests that each cryptomelane generation corresponds to distinct weathering events related to global climatic changes. Sr-87/Sr-86 data show large variations, clearly inherited from the 2.1 Ga parent rock of the Mn protore. The Rb/Sr results demonstrate that minimum fractionation occurs during cryptomelane crystallization, except for the latest generation, which is depleted in Sr. This precludes use of the Rb/Sr radiochronometer for dating secondary Mn oxides in laterites.
