Coupled U-Pb and trace-element analyses of accessory phases in crustal xenoliths from the Late Devonian Udachnaya kimberlite (Siberian craton, Russia) are used to constrain Moho temperature and crustal heat production at the time of kimberlite eruption. Rutile and apatite in lower-crustal garnet granulites record U-Pb dates that extend from 1.8 Ga to 360 Ma (timing of kimberlite eruption). This contrasts with upper-crustal tonalites and amphibolites that contain solely Paleoproterozoic apatite. Depth profiling of rutile from the lower-crustal xenoliths show that U-Pb dates increase gradually from rim to core over µm-scale distances, with slower-diffusing elements (e.g., Al) increasing in concentration across similar length-scales. The U-Pb and trace element gradients in rutile are incompatible with partial Pb loss during slow cooling, but are consistent with neocrystallization and re-heating of the lower crust for <1 Myr prior to eruption. Because Paleoproterozoic rutile and apatite dates are preserved, we infer that long-term ambient lower-crustal temperatures before this thermal perturbation were cooler than the Pb closure temperature of rutile and probably apatite (<400 °C). The lower-crustal temperature bounds from these data are consistent with pressure-temperature arrays of Udachnaya peridotite xenoliths that suggest relatively cool geothermal gradients, signifying that the mantle xenoliths accurately capture the thermal state of the lithosphere prior to eruption. Combined, the xenolith data imply low crustal heat production for the Siberian craton (~0.3 µW/m³). Nevertheless, such values produce surface heat flow values of 20–40 mW/m², higher than measured around Udachnaya (average 19 mW/m²), suggesting that the surface heat flow measurements are inaccurate.

Electron probe microanalyses

Quantitative elemental analyses of major mineral phases were done using a Cameca SX-100 electron probe micro analyzer (EPMA) housed at the University of California, Santa Barbara (UCSB). Measurements were made using the following beam conditions: an accelerating voltage of 20 kV, a beam current of 200 nA and a defocused beam of 2–5 µm diameter. A series of natural and synthetic standards were analyzed for calibration purposes.

Laser ablation split-stream analyses

The xenoliths were processed using standard mineral separation techniques to recover whole grains (hand crushing, magnetic separation, and heavy liquid separation). Coupled U-Pb and trace element analyses on zircon, rutile, and apatite were done by laser ablation split stream inductively coupled plasma mass spectrometry (LASS) at UC Santa Barbara (Kylander-Clark et al., 2013; Kylander-Clark, 2017). Analyses were conducted on polished grain interiors for conventional LASS analyses and on the unpolished exterior of whole grains for depth profiling. Analytical protocols are detailed in Supplementary Information. The quoted uncertainties for all laser ablation data are 2σ and incorporate analytical uncertainty as well as additional uncertainties associated with reproducibility of secondary matrix-matched reference materials (RMs) (Horstwood et al., 2012). Reported uncertainties for trace-element data are 2σ and only include analytical uncertainties. All U-Pb data are shown in Tera-Wasserburg concordia plots (Vermeesch, 2018) and are uncorrected for common-Pb. The quoted U-Pb dates for apatite and rutile are common-Pb corrected using a common-Pb intercept defined by a linear regression through the U-Pb data or assuming a Stacey and Kramers (1975) initial ²⁰⁷Pb/²⁰⁶Pb composition. The quoted dates for zircon are ²⁰⁷Pb/²⁰⁶Pb dates that are within 5% concordance of their respective ²⁰⁶Pb/²³⁸U date.

The sampling resolution of laser ablation spot analyses described above (50-µm-diameter spots for rutile) is too coarse to resolve age gradients. For this reason, depth profiling was done using continuous laser pulsing at a low frequency (e.g., Cottle et al., 2009b; Smye and Stockli, 2014; Holder et al., 2019; Apen et al., 2020; Garber et al., 2020). Details regarding depth profiling are also reported in the Supplementary Information. Final pit depths were measured on the SEM and are on average ~13 µm. The depth profile data are reported at 1 s intervals and the reported U-Pb uncertainties take into account the reproducibility of multiple reference rutiles (~6% on ²³⁸U/²⁰⁶Pb and ~4% on ²⁰⁷Pb/²⁰⁶Pb). In addition to U-Pb ratios, we also report Pb concentration profiles, which we divided into radiogenic Pb (Pb*) and non-radiogenic (common) Pb (Pb_c) (see also Holder et al., 2019; Garber et al., 2020). The total concentrations of ^207,206Pb at each depth interval were calculated using the measured ²³⁸U/²⁰⁶Pb ratio, ²⁰⁷Pb/²⁰⁶Pb ratio, and U concentration and assuming an invariant ²³⁸U/²³⁵U ratio of 137.818 (Heiss et al., 2012). Concentrations of ^207,206Pb* were determined using common-Pb corrected dates and measured U concentrations. From this, we calculated the proportion of Pb_c by subtracting ^207,206Pb* from the total ^207,206Pb concentrations. The final ^207,206Pb* profiles were further corrected for radiogenic Pb ingrowth following kimberlite eruption at 360 Ma.

Folder contains a README.txt file and data to reproduce figures from the associated manuscript.

See the preprint article for more details.

Files can be opened with Excel.