Non-destructive particle analyses from different surface sediment samples of Lake Towuti, Indonesia

At the University of Cologne, Germany, a subsample was taken from one aliquot and used to produce smear slides for identification of sedimentary components using transmitted light microscopy. On selected samples, sponge spicules and diatom frustules were additionally investigated using a Zeiss Gemini Sigma 300VP scanning electron microscope (SEM; Carl Zeiss AG, Oberkochen, Germany). Furthermore, some magnetic mineral grains were identified with energy dispersive X-ray spectroscopy (EDX) of the Sigma SEM system.
Based on smear slide analyses, a set of 40 samples that contain sponge spicules, diatoms and/or tephra particles were selected for automated, non-destructive particle image analyses using a dynamic imaging system (Benchtop B3 Series VS FlowCAM®; Fluid Imaging Technologies, Inc., Scarborough, ME, USA) to quantify the abundance of these particles. Aliquots of wet bulk samples were treated with hydrogen peroxide (H2O2; 30%) for seven days at room temperature to remove organic matter (OM) and disaggregate the siliceous biogenic particles, and were subsequently sieved with 25 and 80 µm meshes. The pre-treated sample fractions were diluted with deionized water (samples <25 µm) or polyvinyl pyrrolidone (PVP, 2 %; samples 25 to 80 µm and >80 µm). Particle recording in the <25 µm and 25 to 80 µm fractions was carried out using a 100 µm flowcell, a 10x objective lens with a collimator, and a 1 ml syringe-pump (flow rate 0.3 ml/min), whereas the >80 µm fraction was recorded using a 300 µm flowcell, a 4x objective lens without collimator, and a 5 ml syringe-pump (flow rate 0.6 ml/min). Data were acquired using the software VisualSpreadsheet (Fluid Imaging Technologies, Inc., Scarborough, ME, USA) until 10.000 images were recorded or 30 ml of the sample was investigated. An automated catalogue based on training sets developed for sponge spicules, diatoms and tephra particles was compiled to differentiate and group components with comparable characteristics in the measured sample fractions.

本数据集的子样本取自德国科隆大学（University of Cologne）的一份等分试样，将其制作为涂片玻片后，通过透射光显微镜开展沉积组分鉴定。针对部分选定样品，研究人员额外使用蔡司Gemini Sigma 300VP扫描电子显微镜（SEM）对海绵骨针与硅藻壳瓣进行观测；此外还借助该Sigma SEM系统的能量色散X射线能谱（EDX），完成了部分磁性矿物颗粒的鉴定工作。基于涂片玻片的分析结果，研究人员筛选出40份含有海绵骨针、硅藻及/或火山碎屑颗粒的样品，采用动态成像系统（台式B3系列VS FlowCAM®; Fluid Imaging Technologies, Inc., Scarborough, ME, USA）开展自动化、非破坏性颗粒图像分析，以定量上述颗粒的丰度。具体实验流程如下：将湿态块状样品的等分试样置于室温下，用30%过氧化氢（H₂O₂）处理7天，以去除有机质（OM）并解离硅质生源颗粒；随后分别使用25 μm与80 μm孔径的筛网进行筛分。将预处理后的样品组分分别用去离子水（粒径<25 μm的样品）或2%聚乙烯吡咯烷酮（PVP）溶液（粒径25~80 μm及>80 μm的样品）进行稀释。对于<25 μm与25~80 μm粒径组分的颗粒成像，采用100 μm流通池、带准直器的10倍物镜以及1 ml注射泵（流速0.3 ml/min）完成数据采集；而>80 μm粒径组分则使用300 μm流通池、不带准直器的4倍物镜以及5 ml注射泵（流速0.6 ml/min）进行成像。数据采集通过"VisualSpreadsheet"软件（Fluid Imaging Technologies, Inc., Scarborough, ME, USA）完成，直至采集到10000张图像或处理完30 ml样品。此外，研究人员基于针对海绵骨针、硅藻及火山碎屑颗粒构建的训练集，编制了自动化分类目录，用于对测量样品组分中特征相似的颗粒进行区分与归类。