Long-chain alkenones were analyzed to estimate paleo-SSTs. The analytical method is based on Yamamoto et al. (2000) with slight modifications. A 1-g sample of freeze-dried sediment was ultrasonic extracted with dichloromethane:methanol (6:4) for 6 min then concentrated and passed through a short bed of Na2SO4 to remove water, yielding a solvent extraction.
An aliquot of the lipid extract was separated into four fractions by column chromatography (eluants are as follows: F1 = 3 mL n-hexane, F2 = 3 mL n-hexane:toluene [3:1], F3 = 4 mL toluene, and F4 = 3 mL toluene:methanol [3:1]). The column was 5.5 mm in diameter x 45 mm long and was filled with 5% deactivated SiO2. C36 n-alkane was added as an internal standard into fraction F3 (alkenone and alkenoates). Fraction F3 was dissolved in 100 µL of n-hexane, and a 1-µL aliquot was then analyzed by gas chromatography at Hokkaido University, on a Hewlett-Packard Model 6890 gas chromatograph equipped with a 6-m x 0.25-µm (0.25 µm film thickness) CP-Sil 5CB (Chrompack) capillary column. For the analysis of the F3 fraction, the oven temperature was programmed from 70° to 310°C at 20°C/min ramp rate and then held at 310° for 40 min. Identification of alkenone compounds was based on comparison of mass spectra and retention time with those in the literature (e.g., de Leeuw et al., 1980; Volkman et al., 1980). The unsaturation index of alkenone was calculated as
where [C37:2] and [C37:3] are the concentration of C37 in each compound (Prahl and Wakeham, 1987). The calculation of paleotemperature was conducted according to the equation
based on an experimental result for cultured Emiliania huxleyi (Prahl et al., 1988).
Organic carbon (Org-C), nitrogen (N), and sulfur (S) contents in acid-treated samples were determined for all 75 samples. Approximately 0.2-0.3 g of powdered sample was treated with 1-M HCl for ~12 hr, then was washed two times with 50 mL deionized filtered water (DIFW) to remove HCl and sea salt. These samples were then dried at 70°C in a glass centrifuge tube for ~24 hr. The samples were analyzed using a LECO-CNS2000 elemental analyzer at the Department of Earth and Planetary Science, Tokyo University. Approximately 0.1 g of sample was weighed in a ceramic crucible and oxidized at 1350°C. The evolved C and S were measured in an infrared absorption analyzer, and N was measured in a thermal conductivity analyzer. The standard deviations for six duplicate analyses are C = 0.251, N = 0.064, and S = 0.081.
The concentrations of 10 major element oxides (SiO2, TiO2, Al2O3, Fe2O3, MnO, CaO, Na2O, K2O, and P2O5) were determined for all 75 samples by X-ray fluorescence (XRF) analysis of fused glass beads using a Philips-PW1480 at the Department of Earth and Planetary Science, University of Tokyo. Powdered samples were dried at 110°C for >4 hr and were then ignited at 1000°C for 6 hr. Loss on ignition (LOI) was calculated from the weight loss. Approximately 0.4 g of ignited sample was mixed with ~4 g of Li2B4O7 flux with an exact 1:10 mixing ratio and was fused in a platinum crucible to make a glass bead. The accuracy of these analyses are ±0.7% for SiO2, ±1.2% for TiO2, ±1.3% for Al2O3, ±2.0% for Fe2O3, ±3.8% for MnO, ±1.5% for MgO, ±2.9% for CaO, ±3.9% for Na2O, ±3.9% for K2O, and ±4.6% for P2O5 (Yoshida and Takahashi, 1997).