Past variations in lithium concentration and isotope ratio of seawater recorded in marine carbonates have shown great potential in tracing carbon cycles during critical climate transitions, especially coupled with the radiogenic strontium isotope system. However, acquiring reliable seawater Li isotope compositions from marine carbonates can be technically challenging because non‑carbonate phases co-existed in ancient marine carbonates can easily contaminate carbonate Li signal during inappropriate sample dissolution. Each contaminant phase, including Mn oxides, clays, and silicate minerals, can have different Li isotope compositions than that preserved in the carbonate lattice. More importantly, a significant portion of Li is adsorbed on the Mn oxides and clay minerals. This fraction is considered exchangeable and can be liberated using weak acid without destruction of mineral structure. Cleaning procedures and partial digestion have been applied to remove the exchangeable Li fraction in contaminative phases and selectively dissolve the carbonates. However, previous methods cannot be directly applied to ancient marine carbonates of various carbonate purity, which lack systematic testing on the threshold of sample purity to maintain the efficiency of the precleaning and partial digestion. Here we report an efficient sequential leaching protocol to extract primary seawater Li and Sr isotope compositions from bulk marine carbonates. The procedure consists of pre-cleaning steps using 1 M ammonium acetate and 1 M ammonium carbonate to completely remove the exchangeable fraction, followed by partial digestion using 0.3 M acetic acid to selectively target carbonate-bounded Li. We validate this method by mixing carbonate and contaminant endmembers in different weight percentages. We found this protocol works effectively for carbonates of >80% purity. Additionally, we demonstrate filtration is necessary when separating leachates from the solids to quantitatively avoid contaminations from non‑carbonate phases. We also suggest NIST-SRM-1d to be a cross-lab carbonate standard for Li isotope analyses and report its long-term external precision (two years). To sum, we propose an effective community standard procedure to reconstruct δ⁷Li_sw records using ancient bulk carbonate samples.

锂同位素,碳酸盐岩,化学淋滤,锶同位素