- Top of page
- Why care about the chemistry of biomineralized tissues?
- Recent views of biomineralization
- Trace element chemistry of biominerals and long-term diagenetic process
Biomineralized tissues are chemically altered after death, and this diagenetic alteration can obscure original biological chemical features or provide new chemical information about the depositional environment. To use the chemistry of fossil biominerals to reconstruct biological, environmental or taphonomic information, a solid appreciation of biomineralization, mineral diagenesis and biomineral–water interaction is needed. Here, I summarize the key recent developments in the fields of biomineralization and post-mortem trace element exchange that have significant implications for our understanding of the diagenetic behaviour of biominerals and the ways in which biomineral chemistry can be used in palaeontological and taphonomic research.
Biomineralized tissues represent the overwhelming majority of body fossils and as such provide the bulk of fossil tissue available for geochemical analysis. Interpreting the geochemistry of fossil biominerals is extremely challenging, however, requiring an understanding of the biochemistry of the living animal, the nature of the biomineral when fresh, its interaction with the immediate burial environment and its behaviour in the longer-term geological setting. This list of prerequisites is rather daunting, but lack of understanding has seldom prevented geochemists from pursuing interesting ideas, or even slowed them down appreciably; consequently, there is an extremely large literature treating the measurement, interpretation and application of the geochemical composition of fossil biominerals. Almost all of this literature depends to some extent on tacit or explicit assumptions regarding the post-mortem behaviour of the biomineral in question; thus, chemical taphonomy is a critical component of all geochemical palaeontology.
In the following article, I will review a few aspects of the geochemistry of fossil biominerals. The aim is to highlight specific recent developments that show promise for future exploration and suggest some areas where more work is required or where application of a geochemical perspective may yield interesting new taphonomic insights.
Trace element chemistry of biominerals and long-term diagenetic process
- Top of page
- Why care about the chemistry of biomineralized tissues?
- Recent views of biomineralization
- Trace element chemistry of biominerals and long-term diagenetic process
As reviewed above, recovery of biological signals from fossil biominerals is complicated by the relatively reactive and unstable nature of the original biomineral. Exacerbating this problem, the more unstable or reactive the biomineral, the less likely it is to preserve an intact signal through diagenesis. One of the most reactive and unstable biominerals is bone mineral (as opposed to enamel that has relatively large crystals, high initial order and low organic content). This is in many ways disappointing as there are obvious vested interests in recovering biological signals from fossil vertebrate remains. However, the relative ease with which bone is diagenetically altered opens new possibilities. If elements are added to bone rapidly after burial, then they could serve to provide a geochemical fingerprint of the environment of burial. This might be useful for reconstruction of environmental conditions, tracking post-mortem movement of bones or radiometric dating. However, all these applications depend on knowing the rate of uptake of trace elements into bone during diagenesis.
Much of the work involving trace element uptake into bone has revolved around the rare earth elements (REE). This group of 15 elements with similar properties is geochemically useful as they are only present in very low concentrations in vivo, are readily adsorbed onto bone mineral surfaces and, most importantly, provide a range of palaeoenvironmental proxies (Williams 1988; Trueman and Tuross 2002; Tütken et al. 2011). Neodymium is a member of the rare earth series, with five stable isotopes and a range of radioactive isotopes. The ratio of 143Nd to 144Nd isotopes is a convenient tracer for ocean water masses as the residence time of neodymium in the ocean is significantly lower than the mixing time. The isotopic composition of neodymium within ocean waters varies spatially with the age and composition of land masses and hydrothermal or volcanic inputs. Consequently, neodymium isotopes are important tracers of past ocean circulation (Martin and Haley 2000; Thomas 2004). Neodymium is present in very low concentrations in biominerals in vivo, but is scavenged into bioapatite post-mortem (Martin and Haley 2000). Consequently, fossil fish debris has become an important host material for neodymium isotope palaeoceanography. The validity of any palaeo-proxy based on diagenetic incorporation of trace elements rests on the assumption that element uptake into bioapatite occurs during early diagenesis when the tooth or bone is in contact with bottom ocean waters or that any subsequent uptake occurs from a trace element source with the same isotopic composition as the fluid in the initial depositional environment.
Uptake of REE into bone has been studied intensively both in marine and terrestrial environments (recently summarized by Trueman and Tuross 2002; Trueman 2007; Kocsis et al. 2010; Herwartz et al. 2011; Tütken et al. 2011). Three main mechanisms have been proposed for the introduction of metal cations into bone, and one or more of each may be involved in the incorporation of REE into bone under differing environmental conditions: (1) adsorption of metal cations onto crystal surfaces, (2) direct substitution of ions into the apatite crystal lattice and (3) growth of discrete metal phosphate phases disseminated within the bone.
The last of these mechanisms has been observed experimentally (Valsami-Jones et al. 1998). Elemental mapping of fossil bone, however, reveals zoned but finely disseminated distributions of trace metals, not the highly concentrated and localized distribution that would be expected were trace metals concentrated in authigenic metal phosphates (Williams 1988; Koenig et al. 2009). It seems unlikely, therefore, that trace metals are held as discrete metal-phosphate phases, but rather are contained within the apatite lattice as adsorbed or substituted ions. However, Molleson et al. (1998) reported a hitherto unique example of extensive replacement of bone mineral by the lead phosphate mineral pyromorphite in human remains within a lead coffin.
Elements with relatively low solubility and high affinities for metal oxide phases may be concentrated as disseminated aggregations. Thorium, in particular, is frequently strongly associated with authigenic Fe–Mn oxide phases rather than apatite. This was clearly demonstrated by Grün et al. (2010) who showed that thorium within c. 35 ka old bones was associated with an iron-oxide phase. However, Grün et al. (2010) further argued that REE were also likely to be held in iron-oxide phases. This is highly unlikely as thorium (and other incompatible elements) have a very low affinity for apatite, whereas more compatible elements such as strontium, barium, uranium (in oxidizing conditions) and especially the REE are strongly partitioned into apatite phases within natural waters, to the extent that bone apatite is an effective scavenging material for the remediation of contaminated land (Valsami-Jones et al. 1998). Bone apatite therefore concentrates REE from surrounding waters and acts as an effective host reflecting early diagenetic pore water elemental compositions. However, as discussed below, these may be modified to differing extents after initial uptake.
Due to large ranges in partition and adsorption coefficients between water and bioapatite, trace elements are fractionated during adsorptive uptake. Fractionation of trace elements between bone mineral and pore water is readily demonstrated by spatially resolved analyses of element distribution in archaeological and fossil bones (Williams 1988; Trueman et al. 2008; Koenig et al. 2009; Trueman et al. 2011). Trace elements are supplied to buried bone from the external margins and subsequently diffuse inwards through water-saturated pores (Millard and Hedges 1996). This directional addition leads to the formation of concentration gradients; the steepness of the gradient is a function of the ratio of adsorption and diffusion coefficients (D/R ratio) and time (Millard and Hedges 1996). Adsorption will continue until equilibrium is reached with the surrounding pore waters, at which point concentration profiles will be flat (Millard and Hedges 1996; Pike et al. 2002). Elements with relatively low adsorption coefficients but high concentrations in pore waters (e.g. Sr) will reach equilibrium relatively rapidly. By contrast, elements with high adsorption coefficients but low abundances in pore waters (such as the REE) rarely reach equilibrium with pore waters. The rate at which bone approaches equilibrium with pore waters for any given trace element is therefore dependent on the adsorption coefficient and the element concentration and mobility in the burial environment. Uranium is a particularly interesting element in this context, as its distribution within archaeological bones is rather varied. Bone mineral has a similar affinity for uranium as for the REE; concentration profiles are commonly developed in archaeological bones and form the basis for improved U-series age assessments (Millard and Hedges 1996; Pike et al. 2002; Eggins et al. 2003). In palaeontological bones, concentration profiles for uranium are frequently flat, implying equilibrium, while those for the REE often remain relatively steep (Trueman et al. 2011). This may reflect increased solubility and mobility of uranyl ions in the presence of carbonate, but after the pioneering work of Millard and co-workers, surprisingly little experimental work has been published regarding the uptake and transport of uranium into bone post-mortem.
It is clear from archaeological age samples, from the preservation of REE patterns and isotopic compositions consistent with early depositional environments, and from measured concentration gradients, that initial uptake of REE occurs rapidly post-mortem, likely within 103–106 years post-mortem depending on the size and nature of the bone and the hydrology of the burial environment (Trueman and Tuross 2002; Trueman et al. 2006). If trace elements are indeed incorporated into bone rapidly post-mortem and remain stable over longer timescales, then radiometric dating of bone should be possible. One potential radiometric pairing is the REE lutetium and its decay product hafnium (Barfod et al. 2003). If lutetium is incorporated into bone relatively rapidly after death (e.g. within 106 years), then age assessments from Lu–Hf dating should be consistent with known depositional ages. Due to the relatively long half-life of 176Lu, and the current limits of analytical precision, only bones of tertiary age and older are available for Lu–Hf dating. Recent independent studies on Lu–Hf systematics of fossil bone arrived at similar, and troubling, conclusions (Kocsis et al. 2010; Herwartz et al. 2011, 2013). In almost all investigated cases, calculated Lu–Hf ages were significantly younger than the known depositional age, pointing to continued uptake of REE after initial diagenetic recrystallization of bone and/or loss of radiogenic hafnium (Fig. 2). Simple modelling suggested that relatively low rates of continued uptake of lutetium would be sufficient to produce the measured isochron ages, resulting in up to 50 per cent of the total REE content of the fossil bone being added after initial recrystallization. A similar argument was proposed by Tütken et al. (2011), who showed that neodymium concentrations increased rapidly in the outer portions of fossil bones on timescales significantly less than 106 years, but that neodymium concentrations in the central cortex increased steadily over timescales of 106–108 years. These results are troubling for the assumptions underpinning the use of fossil bones and, to a lesser extent, teeth for palaeoenvironmental reconstruction centred on REE-based tracers such as neodymium isotopes. However, it should be noted that most investigations to date were conducted on relatively large bones likely to experience more protracted exchange with pore waters. Few systematic, radiometric-isotope-based investigations of the rate of uptake of REE in small teeth or bones have yet been published. While these studies are undeniably complicated to perform, it is anticipated that they will be especially revealing about the nature of long-term diagenesis in different bioapatite materials under differing depositional settings.
Figure 2. Testing long-term diagenetic uptake of rare earth elements (REE) into bone; Lu–Hf isochrons in bones and teeth from four Mesozoic localities. Heavy black line represents the isochron expected if bones remained a closed system after initial diagenetic recrystallization. In all cases, the apparent isochron ‘age’ recovered from measured bones is too young, indicating continued uptake of REE after initial diagenetic recrystallization. After Kocsis et al. (2010).
Download figure to PowerPoint
Some of the first attempts to use direct radiometric dating of fossil bone to test ages in isolated remains of disputed origin almost inevitably were centred on dinosaur remains recovered from Paleocene sediments (Fassett et al. 2011). Using laser-ablation-based U–Pb dating techniques, this study suggested a Paleocene age for a single disarticulated bone. Given the implications of such a finding, presented data must be analytically and statistically rigorous. The data presented by Fassett et al. (2011) have been challenged by several authors (Koenig et al. 2012), specifically the former's exclusion of a large proportion of measured U–Pb ages based on a posteriori criteria. Fassett et al. (2011) argued that bone is compositionally heterogeneous, and the spatially resolved nature of the ablation-based analyses allowed them to analyse unaltered portions of the bone. This may be possible, but the recovered age is itself used in part to determine whether a particular portion of bone is suitable for dating, and this circular reasoning means that it is difficult to see how a purely objective age assessment can be accomplished. In their Lu–Hf studies, both Kocsis et al. (2010) and Herwartz et al. (2011) recovered a range of isotopic values from a single bone or assemblage that could be filtered a posteriori to produce discrete subsets of data points yielding several apparent isochron ages. Some of this variation may reflect phases of disturbance of the isotopic system, in which case recovery of distinct groups of ages, marking episodes of diagenetic alteration, may be possible. Currently, direct dating of fossil bone is considered a success if a known age is recovered; consequently, the method is not sufficiently reliable to apply to unknown age remains. Until a systematic understanding of the duration of elemental uptake into fossil bone emerges, direct dating remains an exciting possibility and a potentially fruitful (but challenging) field of study. However, recovery of an age known to be ‘incorrect’ via direct radiometric dating definitely implies diagenetic disturbance and is thus useful both for validating palaeo-proxy records and for investigating the nature of diagenesis in biominerals.
Most studies investigating the chemistry of fossil biominerals attempt to extract either biological or environmental information. The diagenetic chemistry of biominerals can also be used to study post-mortem movement, assemblage accumulation history and to quantify preservation bias. Taphonomic applications of biomineral chemistry rely on the logic that if biominerals incorporate elements from their initial burial environments, then these early diagenetic signals should serve as fingerprints to group fossils from a common origin, distinguish allochthonous fossils within a mixed assemblage, or to compare spatial and temporal averaging between deposits (Fig. 3). Clearly, this logic relies on several fundamental assumptions; that the duration of early uptake is fast relative to the timescales of post-mortem movement that the study is interested in; that subsequent diagenetic change is relatively minor; and that variation between locations is greater than variation between fossils from a common location (Trueman 2007). Geochemical taphonomy has only been applied systematically to bone assemblages, presumably because of the relatively rapid uptake of a range of elements into bone post-mortem. A large number of studies have demonstrated the use of the REE composition of fossil bones to either identify burial environments (Williams 1988; Trueman 1999; Kemp and Trueman 2003; Metzger et al. 2004) or to track reworking through stratigraphic sections (Trueman and Benton 1997; Staron et al. 2001; Trueman et al. 2005, 2006; MacFadden et al. 2007). REE taphonomy has been reviewed recently (Trueman 2007). The subsequent recognition of continued uptake of REE in bone after initial recrystallization suggests that the resolution available to REE taphonomic studies will, in general, be less in older rocks.
Figure 3. Rare earth elements (REE) taphonomy. Bone fragments from lakeshore soils and a small (<1 m maximum breadth) palaeochannel in Pleistocene sediments of the Olorgesailie Formation, Kenya. Shale-normalized REE ratio values of 38 bones taken from the channel fill and 134 bones from 7 contemporaneous, laterally adjacent soil sites spread over an area of c. 100 m2. REE analysis effectively distinguishes between these depositional microenvironments and demonstrates relatively high variance in channel-hosted bones compared with adjacent soils. After Trueman et al. (2006).
Download figure to PowerPoint
The isotopic composition of elements such as strontium and neodymium can also be used to track reworking within assemblages. Bones and teeth clearly incorporate most diagenetic strontium and neodymium from the pore waters of the media in which they are embedded. As these pore waters themselves gain strontium and neodymium from the weathered bedrock, and as strontium and neodymium isotopes are not fractionated significantly during fluid–rock interactions, bones and teeth will inherit an isotopic composition similar to their enclosing sediments. In an elegant study, Tütken et al. (2011) demonstrated the close correspondence in isotopic composition of neodymium in bone and in the enclosing sediment. Differences in neodymium isotopic compositions between bones and enclosing sediments may indicate reworking. However, as the neodymium isotopic composition of bedrock (particularly sedimentary bedrock) may be relatively homogenous over large areas, these techniques are best suited to tests where high spatial or temporal variance in neodymium isotopic composition of depositional pore waters is expected. In addition, the sensitivity of such tests will be influenced by the relative timing of reworking compared with the extent of late diagenetic uptake. Tütken et al. (2011) suggested comparing neodymium isotope compositions between the external and internal sections of bones, and the enclosing sediment. This is a very promising approach, which may provide a subtle test for postburial reworking and movement of bones between depositional settings, and warrants further exploration.
Methodological and technical considerations
Most of the taphonomic studies reviewed by Trueman (2007) used bulk dissolution techniques to extract the REE compositions. Spatially resolved analysis of REE compositions in fossil bones by laser-ablation-ICP-MS is now a simple, widely available technique that has been instrumental in quantifying mechanisms of REE uptake. Spatial heterogeneity in trace element signals within fossils presents a problem for any form of comparative analysis: should the investigator attempt to explore the full spatial variation or compare averaged data? If comparing averages, should they average spatially varied compositions analytically by sampling mixed parts of a biomineral or arithmetically after obtaining spatially resolved data? These questions are far from unique to the field of fossil bone geochemistry. All biominerals contain elemental compositions that are spatially heterogeneous at some scale. In modern fish otoliths (aragonitic biomineralized tissues in the teleost ear), spatial variations in stable isotopes and trace element compositions across otoliths have been used to track migrations between fresh and saltwater in diadromous fish (Walther and Limburg 2012) and to identify connectivity and movement in marine fish populations (Sturrock et al. 2012; Trueman et al. 2012). Otolith geochemists have used two approaches to deal with spatial heterogeneity, either taking averages of otolith compositions across fixed portions of the otolith and comparing averages between individuals from different locations (Longmore et al. 2011) or fitting statistical models to the elemental transects and comparing model parameters between locations and populations (Sandin et al. 2005). The distribution of trace elements within bones is dominantly controlled by diffusion/adsorption processes and is usually approximated effectively by exponential decay or error function models (Millard and Hedges 1996; Kohn 2008; Trueman et al. 2011). The spatial distribution of trace elements within biominerals can therefore be described by fitting functions to measured concentration gradients and using the fitted model terms as comparative taphonomic characters. This approach has been used to compare the relative rates of recrystallization of bones and to demonstrate that the distribution of elements within fossil bones is not predicted by measurements of crystallinity indices and does not itself predict the presence of organic remains (Trueman et al. 2008, 2011).
Avenues for future work
To conclude this review, some suggestions for further work in the broad field of taphonomic geochemistry are provided:
- Quantifying crystal disorder in modern biominerals. Differences in mineral stability between biominerals and their inorganic equivalents provide the impetus for diagenetic change. Understanding and quantifying these differences is relatively simple with modern spectroscopic, XRD and imaging techniques. We are only beginning to recognize the mineralogical complexity in modern biominerals and quantifying this will significantly aid our mechanistic understanding of the post-mortem behaviour of biominerals and provide a better framework to understand nonequilibrium ‘vital effects’.
- Characterizing diagenetic change in terms of changes in crystal order. Linked to point (1) above: a better appreciation of the crystallographic and mineralogical heterogeneity of modern biomineralized tissues should lead to a more systematic description of the post-mortem behaviour of biominerals. Different mechanisms associated with diagenetic recrystallization have widely different implications for the preservation of intact geochemical information. Spectroscopic techniques that distinguish between long-range and short-range order will help to identify when recrystallization is a hindrance or a blessing for the recovery of palaeoenvironmental or ecological information.
- Quantifying uptake rates of elements post-mortem. Measuring the exact rate of diagenetic element uptake is possible using radiometric dating tools. This approach provides an absolute test for the preservation of early diagenetic chemical signals in biomineralized tissues and could identify tissue types and depositional settings suitable for direct dating of fossil assemblages and their associated sediments. This is an analytically challenging field, but the potential benefits are extremely high.
- Developing new taphonomic characters. The distribution of trace elements within bones describes the interaction between burial hydrology and bone recrystallization and is thus directly associated with preservation potential. Within a single depositional assemblage, the absolute value and range in concentration gradients expressed for a single element between bones could be used a taphonomic character linked to the rate of recrystallization of the bone, and thus directly linked to preservation potential. To my knowledge, this approach has not yet been used as a taphonomic characteristic of an assemblage or as a test of the effect of differing burial environments on the rate of fossilization (and thus preservational bias).
- Improving sampling design and statistical treatment of geochemical studies.
Drawing from points (3) and (4) above, field sampling should be conducted to select tissues with matched rates of element uptake and diagenetic recrystallization. Sampling design and statistical modelling should be improved reflecting advances in statistical theory and the availability of packages dealing with complex models (e.g. nonlinear mixed models and maximum-likelihood and Bayesian approaches to classification).
I would like to thank P. J. Orr, University College, Dublin, Ireland, for inviting me to submit this review and participate in the 2012 Annual Meeting of the Palaeontological Association. I am indebted to S. Weiner and L. Addadi, Weizmann Institute of Science, Israel, for introducing me to biomineralization and to M. R. Palmer, University of Southampton, UK, for geochemical discussions. As is often the case in reviews, much of the work described is based on studies performed by past and current postdocs and PhD students, especially L. Kocsis, A. Sturrock and K. M. MacKenzie. T. Tütken and an anonymous reviewer provided helpful and insightful comments. This manuscript is a contribution to the NERC-funded project ‘The Evolution of Modern Marine Ecosystems’.