Identification and Paleoclimatic Significance of Magnetite Nanoparticles in Soils
Total Page:16
File Type:pdf, Size:1020Kb
Identification and paleoclimatic significance of magnetite nanoparticles in soils Imad A. M. Ahmeda,1 and Barbara A. Maherb,1 aDepartment of Earth Sciences, University of Oxford, OX1 3AN Oxford, United Kingdom; and bCentre for Environmental Magnetism and Palaeomagnetism, Lancaster Environment Centre, University of Lancaster, LA1 4YQ Lancaster, United Kingdom Edited by Lisa Tauxe, University of California, San Diego, La Jolla, CA, and approved December 27, 2017 (received for review November 2, 2017) In the world-famous sediments of the Chinese Loess Plateau, fossil from the Chinese loess/paleosol sediments indicates that abiotic (ex- soils alternate with windblown dust layers to record monsoonal tracellular) precipitation of the paleosol ferrimagnets (Fig. 1 C and D) variations over the last ∼3 My. The less-weathered, weakly magnetic is dominant, rather than intracellular formation of magnetosomes, dust layers reflect drier, colder glaciations. The fossil soils (paleosols) of controlled size and shape, by magnetotactic bacteria (Fig. 1E). contain variable concentrations of nanoscale, strongly magnetic iron Both the mineralogy of the soil ferrimagnets and their pathways oxides, formed in situ during the wetter, warmer interglaciations. of formation have been hotly debated. Opposite hypotheses link Mineralogical identification of the magnetic soil oxides is essential soil magnetic enhancement not with redox-related formation of the + for deciphering these key paleoclimatic records. Formation of mag- Fe2 -bearing magnetite but with the formation from ferrihydrite of 2+ 3+ netite, a mixed Fe /Fe ferrimagnet, has been linked to soil redox an oxidized, maghemite-like phase (hydromaghemite), which itself oscillations, and thence to paleorainfall. An opposite hypothesis transforms to hematite upon ripening and aging (11–13). This latter states that magnetite can only form if the soil is water saturated 3+ 2+ pathway would thus be temperature dependent rather than mois- for significant periods in order for Fe to be reduced to Fe ,andsug- ture and redox dependent. Indeed, the transformation of ferrihy- gests instead the temperature-dependent formation of maghemite, 3+ drite to maghemite has been identified in vitro (14), albeit under an Fe -oxide, much of which ages subsequently into hematite, hydrothermal experimental conditions (150 °C for 120 d), which typically aluminum substituted. This latter, oxidizing pathway are environmentally unrealistic. More recently, it has been sug- would have been temperature, but not rainfall dependent. Here, gested that a magnetically ordered ferrihydrite, before its oxidation through structural fingerprinting and scanning transmission elec- to hematite, might contribute to soil magnetic enhancement (15). EARTH, ATMOSPHERIC, tron microscopy and electron energy loss spectroscopy analysis, AND PLANETARY SCIENCES Magnetite has an inverse spinel structure that contains tetrahedral we prove that magnetite is the dominant soil-formed ferrite. 2+ 3+ (Td) and octahedral [Oh] sites accommodating Fe and Fe Maghemite is present in lower concentrations, and shows no evidence 3 + 3 + 2 + cations with a spin arrangement of ½Fe ↓ ½Fe ↑Fe ↑ . of aluminum substitution, negating its proposed precursor role for the Td Oh Magnetite and maghemite are end members of a solid solution aluminum-substituted hematite prevalent in the paleosols. Magnetite γ dominance demonstrates that magnetite formation occurs in well- series. Although maghemite ( -Fe2O3) has the same composition as α drained, generally oxidizing soils, and that soil wetting/drying oscil- hematite ( -Fe2O3), it has the structure of a spinel (a cation-deficient 3+ SI lations drive the degree of soil magnetic enhancement. The magnetic spinel, lacking sufficient Fe ions to fill the available Fe sites; variations of the Chinese Loess Plateau paleosols thus record changes Appendix). Maghemite is thus ferrimagnetic, with very similar in monsoonal rainfall, over timescales of millions of years. magnetic properties to magnetite. It is metastable with respect to hematite (it inverts to hematite upon heating), but can be stabilized soil magnetite | Quaternary paleoclimate | monsoon rainfall | magnetic susceptibility | structural fingerprinting Significance he windblown sediments of the famous Chinese Loess Pla- In the famous Chinese Loess Plateau (CLP), weakly magnetic, Tteau (CLP), spanning hundreds of meters in thickness and windblown dust layers alternate with variably magnetic fossil >600,000 km2 in extent, potentially contain the longest, most soils, recording monsoonal variations through the last ∼3 My. detailed terrestrial records of East Asian monsoonal evolution. The soils contain strongly magnetic iron oxides, formed in situ, Interleaved layers of glacial-stage windblown dust (loess) and the mineralogy and paleoclimatic significance of which are + interglacial/interstadial-stage fossil soils (paleosols) span the late controversial. Reduction of iron to form Fe2 -bearing magne- Pliocene and Quaternary geological periods; i.e., the last ∼3My. tite has been linked to soil wetting/drying. Conversely, oxida- 3+ Compared with the less-weathered loess layers, the interglacial tion of iron to form Fe -bearing maghemite, which ages into paleosols contain varying but higher concentrations of nanoscale, Al-substituted hematite, has been linked to paleotemperature. strongly magnetic (ferrimagnetic) iron oxides, formed in situ during This study uses structural fingerprinting and electron energy soil development. These variations in ferrimagnetic concentration are loss spectroscopy to resolve this debate, proving that magne- readily apparent even from simple room-temperature measurements tite is the dominant soil-formed magnet. The magnetic varia- of magnetic susceptibility (Fig. 1A and SI Appendix,Fig.S1). A climatic tions of the CLP paleosols thus record changes in monsoonal cause for the varying soil nanomagnet concentrations is evidenced rainfall, providing a key time series for testing of general cir- by their strong correlation with the deep-sea oxygen isotope record, culation climate models. principally a record of continental ice volume (Fig. 1B). Author contributions: I.A.M.A. and B.A.M. designed research; I.A.M.A. and B.A.M. per- Modern loessic soils across the Chinese Loess Plateau, and formed research; I.A.M.A. contributed new reagents/analytic tools; I.A.M.A. and B.A.M. similar regions (e.g., the Russian steppe, the North American Great analyzed data; and I.A.M.A. and B.A.M. wrote the paper. Plains), display a direct, strongly significant correlation of their soil- The authors declare no conflict of interest. formed (pedogenic) magnetic concentrations with mean annual This article is a PNAS Direct Submission. – rainfall (1 4).Thiscorrelation,usedtoobtainquantifiedestimates Published under the PNAS license. of paleorainfall (5) for the Chinese monsoon region, has been 1 2+ 3+ To whom correspondence may be addressed. Email: [email protected] or linked causally to pedogenic formation of the mixed Fe /Fe iron [email protected]. oxide, magnetite, through redox changes in soil microsites following This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. rainfall events (6–9). Electron microscopy of magnetic concentrates 1073/pnas.1719186115/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1719186115 PNAS Latest Articles | 1of6 Downloaded by guest on September 25, 2021 0 2 probable projected crystallographic symmetries for a candidate A B C phase are then obtained, together with geometric information 6 (i.e., reciprocal lattice spacings and interfringe angles). The com- 200 bination of geometric information on the reciprocal lattice, plane symmetry, and elemental composition (from EDXA) within a crystal 8 100 nm 10 leads to unambiguous identification of the nanoferrite structure. 400 The soil ferrites were extracted from: a modern, magnetically 12 D enhanced soil (a cambisol, from Exmoor, United Kingdom; Fig. 1B and ref. 17); paleosol S1 (of last interglacial age, ∼125,000 y 3 14 600 before present [BP]) from the central region of the CLP 16 (Luochuan; Fig. 1A and SI Appendix,Fig.S2); and a paleosol (∼5 10 nm 18 My BP) from the Mio/Pliocene Red Clay sequence (at Lingtai), Age (10800 year) which underlies the Quaternary-age loess and paleosols. To ex- 22 E amine any possible oxidation effects related to postsampling lab- 26 oratory storage, we compared recently collected (2011) and “old” 1000 30 (collected in 1990) samples from the Luochuan S1 paleosol. For independent verification of our structural fingerprinting approach, 100 nm we additionally used coupled scanning transmission electron mi- 1200 croscopy and electron energy loss spectroscopy (STEM/EELS) to 90 200 400 -5.0 -4.0 -3.0 + quantify the Fe3 =ΣFe ratios and to identify the dominant ferrites in -8 -3 -1 18 “ ” χ (10mkg) δ Ο (‰) the new and long-stored Luochuan S1 paleosol samples. Fig. 1. (A) The magnetic susceptibility of the loess/paleosol sequence at Results Luochuan, central Chinese Loess Plateau (5); (B) the deep-sea oxygen isotope Magnetic concentrates were obtained from the soil and paleosol record from site 677 (10) (glacial stages numbered); (C–E) TEM micrographs samples; the extraction efficiency (SI Appendix,TableS1) quan- of nanoscale, low-temperature ferrimagnets formed in the environment, (C) tified by before- and after-extraction measurements of magnetic CLP paleosol, (D) United Kingdom modern cambisol (Exmoor), and (E)the susceptibility,