The analysis of oxygen isotopes from diatom silica (δ18OSi) in sediment cores has obtained importance for palaeoclimate reconstruction especially where carbonate proxies are either rare or not available. Compared to the widely accepted relation of oxygen isotopes of carbonate origin to climate-relevant parameters, challenges still occur using biogenic silica. These questions arise at sample preparation as well as and the analysis itself, but are especially related to the removal of loosely bound oxygen of the hydrous layer. It is the common view that diatoms consist of an isotopically homogenous inner Si-O-Si layer and a less dense, hydrous layer forming Si-OH bonds, which has to be removed from the sample prior to analysis. Three methods have been accepted so far to perform this step: Controlled Isotopic Exchange (CIE) followed by fluorination, Stepwise Fluorination (SWF) and inductive High-Temperature carbon reduction (iHTR). The former method of vacuum dehydration (VD) proved to be unable to remove all exchangeable oxygen.Here, a new, remotely-operated laser-fluorination based mass spectrometry unit is used for the analysis. The silica is reacted with a CO2 laser in a BrF5 atmosphere and oxygen is then transferred to and analysed in a mass spectrometer (PDZ Europa 20-20). As CIE is both time-consuming and work-intense and SWF is impractical for this setup mainly due to the high pressure increase during dehydration, a new, efficient and fast method should be developed to remove the hydrous layer using the laser-fluorinationprocess.Two approaches were tested to remove the Si-OH layer and the impact on δ18OSi was assessed by performing tests on internal standard materials of marine and lacustrine biogenic silica and of quartz. For VBM, a minimum of 1.5 mg of pure sample is melted to a bead with a defocused laser to eliminate the hydrous outer layer and to reduce the surface. After the bead has formed it is transferred into the reaction chamber completely reacted with a focused laser under BrF5 atmosphere and subsequently analysed on-line.The second method, HFD is an improvement of the outdated VD technique. The sample is heated to 1050°C in a He flow transporting away any removed exchangeable oxygen immediately and thus, not allowing it to re-react with the sample. Various tests have been performed considering pre-heating at 200°C, He flow adjustments and the time of the sample exposed to 1050°C.VBM has difficulties to fully remove the hydrous layer, which results in comparatively lower δ18O values. The final set-up was not found yet. The HFD generated similar data than SWF in other laboratories with a high reproducibility and accuracy (standard deviation <0.2 ). Best results could be achieved by pre-heating the sample at 200°C and later on expose it for 15 minutes under a Helium counter flow at 1050°C. Experience with both techniques will be discussed and the reliability of the data compared to other methods.