Yen-Lin Chen edited section_Wet_SEM_From_the__.tex  over 8 years ago

Commit id: 369e062b034902e3103b7ca1437eb3def00923cb

deletions | additions      

       

The main problem of the polyimide protection membranes is that the membranes will swell outwards like a balloon because of the pressure differences. The swelling might distort the specimen, making some parts of it deeper and some shallower. The image will appear brighter in the center and darker at the edge. A tiny leakage opening with radius $\approx 10 \mu m$ will be designed on the membrane to allow limited exchange of particles and reduce the strain in the membrane, as shown in Fig.5 b. The thickness of the membranes can also be reduced from $145 nm$ to $50 nm$. The tiny hole and the thinner membranes further enable imaging with SE signals. With this set up, it was claimed that the spatial resolution could achieve the best resolution of conventional SEM down to $10 nm$ with heavy atomic stains and $100 nm$ in stainless samples, shown in Fig.5 d.   The charging problem will no longer be solved by the ionized gaseous molecules. In wet SEM, the charging effects can be avoided by microfluidic flow inside the sample cell, as shown in Fig.5 c. The flow will remove the excess charges in the fluid and slowly direct the charges out of the sample. On average, since the flow rate is low, the sample carries an small number of negative charges all the time. There is a trade-off between the rate of liquid flow and the temporal and spatial resolution. The flow tends to drift the specimens of interest and the specimens might be displaced within the dwelling time due to Brownian motion and flow drive. As more viscous liquids are introduced to prevent the random motion and flow drive, the flow rate will be relatively low; the charges pile up fast. Another way to solve this issue in to make the liquid layer as thin as possible. The primary electrons are either backscattered or transmitted, leaving few charges in the sample.