Clumped isotope analysis of lacustrine endogenic carbonates and implications for paleo-temperature reconstruction: A case study from Dali Lake

Abstract

Carbonate clumped isotope (Δ47) is a new and reliable geothermometer. Endogenic carbonates in lake sediments are good archives to reconstruct lake water temperature using clumped isotope thermometry. However, applications of carbonate clumped isotope thermometry to lacustrine carbonates are still scarce because the existed organic and/or sulfur contaminants in such samples interfere with clumped isotope analysis and cause notable temperature biases. Therefore, exploring an effective way to remove contamination is a prerequisite to widely applying carbonate clumped isotope thermometry to lake sediments. By pretreating fine-grained endogenic carbonates from Dali lake sediments with hydrogen peroxide of different concentrations for different lengths of time, we conducted a series of conditional experiments to seek the optimal pretreatment condition for clumped isotope analysis. Δ47, Δ48 offset and 49 parameter were obtained from clumped isotope measurements to assess the effect of contamination removal. Results showed that untreated samples (“0%-0”) had a very high 49 parameter. After the treatment with hydrogen peroxide, the 49 parameter was significantly reduced while the Δ48 offset evidently increased. The Δ47 , Δ48 offset and 49 parameter of samples pretreated under different conditions showed significant variability, even though the measured Δ47 temperatures changed within a relatively narrow range (i.e., from 17.0±1.3 to 22.8±1.6°C). Among various treated samples, the sample “3%-8” yielded a smaller Δ48 offset and 49 parameter, and lower Δ47 temperature of 17.0±1.3°C. Therefore, “reacting with 3% hydrogen peroxide for 8 h” was suggested to be the optimal condition for removing contaminants from lacustrine endogenic carbonates prior to clumped isotope analysis. At the time when the pretreatment condition was too intense (i.e., H2O2 concentration >3% and/or reaction time > 8 h), secondary contaminants might have been generated in the closed reaction system, whose ultimately transformed state may interfere with masses 47, 48. Using the optimal pretreatment procedure, we obtained reasonable Δ47 temperature changes during the last deglaciation and found a temperature decrease of about 6°C in Northern China during the Younger Dryas period. Our study demonstrates a great potential of applying carbonate clumped isotope thermometry to lacustrine carbonates for paleo-temperature and paleo-elevation reconstructions in the future.