If the A260/A230 or A260/A280 values of your sample decrease while optimizing the washing stage, impurities likely weren’t fully washed away.
Your method will likely involve several washes. The first is the most challenging and most important to optimize. There are many proteins in the sample at this point. These proteins can cause the beads to aggregate and trap impurities. The magnetic particles must be fully resuspended to release and wash away these impurities. Good mixing is crucial.
Try increasing the shaker speeds to improve bead suspension during the first wash. Volumes should be low at this stage, so the risk of splashing is low. If you do need to mix by pipetting, consider using a particle mover robot or a 96-channel pipette on the platform or your throughput will drastically decrease.
If faster mixing doesn’t improve bead suspension, try washing the sample with water or TE buffer and mixing vigorously. This elutes nucleic acids from the beads and breaks aggregates. Rebind the nucleic acids by adding binding buffer at the exact ratios used in the binding step.
If the beads are fully suspended during the wash step and you still see low purity, there might be a problem with liquid handling on the robot. If the robot does not fully take up supernatant during the wash steps, impurities can persist. You can adjust liquid transfer volumes. You can also add another wash to dilute the lingering contaminants. Use a polar, non-viscous solution like 80–95% ethanol or the solution used for the recommended final wash.