RUSSIAN JOURNAL OF EARTH SCIENCES VOL. 10, ES6006, doi:10.2205/2008ES000304, 2008

Correlation Between the Content of Metallic Iron and Fe-Hydroxides, Magnetite, and Titanomagnetite

2008ES000304-fig05
Figure 5
[10]  As noted above, tight positive correlation between the total iron content as determined by chemical analyses, the intensity of paramagnetic magnetization Mp, and the amount of ferromagnetic Fe-hydroxides (goethite) was found in sediments close to the K/T boundary in all sections studied; this observation was accounted for by prevailing role of Fe-hydroxides in the intensity of paramagnetic magnetization Mp [Pechersky, 2008]. This Mp -goethite correlation is well illustrated in Figure 5, where the data for five sections close to the K/T boundary and the Khalats sedimentary section (age 17-3 My) are summarized. When the values of Mp and goethite are compared with concentrations of magnetite, titanomagnetite and metallic iron, two main groups of data points are found (Figure 5): (a) the first group reveals clear positive correlation between accumulation of magnetite, titanomagnetite and metallic iron, in contrast to what is observed in the boundary layer (Figure 2); (b) the second zero-correlation group, where no magnetic minerals is detected but, naturally, Mp is measurable. The coefficients of linear correlation are computed both for each pair of observed values and their logarithms

2008ES000304-fig06
Figure 6
[11]  First of all, I would like to pay attention to the steady increase of correlation coefficients when the values are replaced by their logarithms; this observation points to the decisive role of lognormal distribution of the analyzed magnetic characteristics. This is well illustrated by Figure 6, where the rows of measured characteristics and their logarithms are presented in Figure 6a and 6b, respectively. The rows for the values differ greatly, even the general trends showing the similarity just for some intervals (Figure 6a); whereas, the rows for the logarithms are similar in all cases, as well as the general trends (Figure 6b). The strongest correlation is found between the logarithms of paramagnetic magnetization, i.e., the concentration of paramagnetic Fe-hydroxides, and the goethite content. The weaker, but still noticeable, correlation exists between the logarithms of paramagnetic magnetization and titanomagnetite content, whereas the weakest correlation between the logarithms of paramagnetic magnetization and magnetite content. Rather high and practically the same positive correlations are between the logarithms of metallic iron, paramagnetic magnetization, and the goethite content, while the correlation between the logarithms of metallic iron and magnetite + titanomagnetite is much weaker.

2008ES000304-fig07
Figure 7
2008ES000304-fig08
Figure 8
2008ES000304-fig09
Figure 9
2008ES000304-fig10
Figure 10
2008ES000304-fig11
Figure 11
[12]  Let us see if this pattern is observed on the regional level on the base of data on the studied sections (Figures 7, 8, 9, and 10). For comparison, the results are presented on the Khalats section (Figure 11), where the age of sediments is noticeably younger than the K-T boundary. A common feature for all sections is that all points form two groups, similarly to the case of the boundary layer (Figure 2): Group A where metallic iron, Fe-hydroxides, and titanomagnetite are present, and Group B, or zero-group, without metallic iron but with Fe-hydroxides and/or titanomagnetite. Emphasize that the range of values is similar in both groups. In contrast to the boundary layer, where no correlation between the components of Group A is found, perceptible positive correlation is present between these components (Figures 7, 8, 9, 10, and 11), which is described for each object separately.

[13]  For the Gams section (Figure 7), weak correlation between MT+TM and Mp and MT+TM and iron is observed, while no correlation is found between Mp and iron as well as between goethite and iron.

[14]  For the Klyuchi and Teplovka sections (Figure 8), no correlation is observed, except for weak correlation between MT+TM and Mp.

[15]  For the Koshak (Figure 9), Tetritskaro (Figure 10), and Khalats (Figure 11) sections, all components of the Group A reveal relatively weak correlation.

[16]  Therefore, the common feature is expressed by the presence of two groups A and B on the plots; note that the values on the abscissa axis have similar ranges, and the values in the Group B are often higher than in the Group A. In the latter, the characteristics are variably correlated. Hence, the degree of the connection between concentrations of Fe-hydroxides, titanomagnetite and iron varies from highly positive correlation to the lack of it. More high correlation is found in the carbonaceous Koshak and Tetritskaro sections, whereas much slight or no correlation is found in the sections with high terrigenous input. Among all correlations between separate characteristics, the weakest one is between goethite and Fe, and the strongest one is between MT+TM and Mp, except for correlation between goethite and Mp (Table 1). The latter pair, in difference to the others, does not have the "zero'' Group B (Figure 5), which is mainly due to the fact that Mp strongly depends upon Fe-hydroxides that are nearly omnipresent in all studied sediments [Pechersky, 2008]. It is likely that the connection between MT+TM and Mp is predictable and reflects the largely terrigenous provenance of magnetite and titanomagnetite in sediments, although both minerals, in particular titanomagnetite as indicated by its composition, had been originally produced by basaltic volcanism. It is worth recalling that a rather large "zero'' Group B is found for all pairs, when one component of a pair is absent, while the other one is abundant. Such a two-ways connection excludes the "secondary'' origin of the correlation between iron and titanomagnetite, magnetite, and Fe-hydroxides. It is possible to assume that the degree of oxidation of iron to hydroxides is approximately equal and thus to account for positive correlation between goethite and iron; in other words, the higher is the iron content, the higher is the concentration of its oxidized part. The concentrations of iron and goethite, however, differ by three orders of magnitude, and, therefore, the original concentration of iron had to be as high as several percent. A reverse dependence has to exist too, when iron particles approach the fully oxidized state. This is in sharp contrast to the observed pattern, when the Group B is distinct and comprises a large number of samples where metallic iron is absent altogether. Moreover, large Groups B and A coexist within similar range of TM+MT in the plot of MT+TM versus Fe, which cannot be attributed to secondary oxidation of iron for the correlation between titanomagnetite and iron. Really, titanomagnetite is definitely unrelated to oxidation of metallic iron, as the particles of metallic iron, nickel, and their alloys do not contain large amounts of other elements, titanium in particular.

[17]  So the correlations (Figures 5, 6, 7, 8, 9, 10, and 11) are of primary, either terrestrial or extra-terrestrial, origin. Different degree of correlation between components is to be expected for very different conditions of accumulation of Fe-hydroxides, magnetite + titanomagnetite and metallic iron. Indeed, the most interesting features are high correlations between metallic iron and Fe-hydroxides and between metallic iron and titanomagnetite. As the secondary origin of the observed correlations is ruled out, the common way accumulation for the above listed minerals becomes evident. It would not be surprising if the particles of metallic iron are of terrestrial origin. In principle, authigenic metallic iron may be formed in sediments. A review of publications, however, revealed that this is exceptionally rare, while, according to our data, metallic iron is omnipresent in sediments, albeit in tiny amount. Hence metallic iron can be of terrigenous origin only, and its source is most likely to be magmatic rocks most likely of basaltic composition. However, we fail here again, as the grains of metallic iron in basalts are exceptionally rare. Therefore, metallic iron has en mass to be of extraterrestrial origin, which is in full accord with numerous publications. Although some grains may be of terrestrial origin, their share is negligible on the global scale. If so, the strange and enigmatic correlation between accumulation of the particles of very different provenance and origin, i.e., Fe-hydroxides, magnetite, titanomagnetite, and metallic iron, remains unexplained.

[18]  It is worth mentioning that, irrespective of the origin of metallic iron, its accumulation may result from erosion of older rocks and subsequent redeposition. This is similar to formation of placers, similarly to the recent (Quaternary) gold placers in Northeast Asia that were formed by erosion of gold-bearing rocks of mainly Cretaceous age. However, such "far-reaching'' erosion is not necessary. Several millenniums or even less will suffice; it is important that a positive correlation must appear between heavy iron minerals that are extracted from rocks and accumulate in sediments under the study. So it can be concluded that the observed variable positive correlations between concentrations of Fe-hydroxides, magnetite, titanomagnetite, and metallic iron stems from changeable role of redeposition in accumulation of these minerals in sediments.


RJES

Citation: Pechersky, D. M. (2008), Metallic iron in sediments at the Mesozoic-Cenozoic (K/T) boundary, Russ. J. Earth Sci., 10, ES6006, doi:10.2205/2008ES000304.

Copyright 2008 by the Russian Journal of Earth Sciences

Powered by TeXWeb (Win32, v.2.0).