Silicon Nanoporous Membrane

Ali Khourshaei Shargh

University of Rochester, Department of Mechanical Engineering

1. Introduction

Silicon Nanoporous membranes have captured a comprehensive interest around the world due to its great potential in molecular seperation applications. Thus, several research laboratories conduct research on synthesis of Silicon nanomembranes.

However, commercial nanomembranes are over 1000 times thicker than molecules they are designed to separate, leading to poor size cut-off properties, filtrate loss within the membranes, and low transport rates. This limitations obviated by a new class of nanoporous membrane synthesized by McGrath et. al in 2007 where the membrane is almost as thick as the molecules being separated, and it can filter nanoparticles at rates an order of magnitude faster than existing nanoporous memebranes.

Figure 1: This diagram illustrates a 6'' wafer with more than 450 membrane chips. Each chip is a 5.4 mm square with smaller free-standing membrane window.

Briefly, McGrath et. al grow a 500nm thick layer of SiO2 on both sides of a silicon wafer. On the back side of the wafer, SiO2 is patterned. The front oxide layer is then removed, and a three layer film stack (20nm SiO2, 15nm amorphous Si, and 20nm SiO2) is deposited on the front surface. To form Si Nanoporous membrane, the substrate is exposed to 715-770oC for 30s, which results a Nanocrystalline film. The patterned wafer back side is then exposed to a highly selective silicon etchant, which removes the silicon wafer along (111) crystal planes until it reaches the first SiO2 layer of the front side film stack. Finally, exposing the three layer membrane to buffered oxide etchant removes the protective oxide layers, leaving only Si Nanoporous membrane.

In this project, Silicon nanoporous membranes synthesized by McGrath et. al were explored using SEM, TEM, and AFM microscopy. Moreover, X-ray spectroscopy with TEM was done to measures the elemental composition of the sample. Additionally, ImageJ software was utilized to calculate the prosity of a section of the nanomembrane.

2. Scanning Electron Microscopy

Scanning electron microscopy (SEM) is a type of electro microscope that is utilized for imaging specimens using a focoused electron beam. The image is formed as a result of the interaction between the specimen and the beam.

For the sake of imaging, secondary electron (SE) imaging mode with In-lens detector was used in this project. Shortly, secondary electrons are emitted from very close to the specimen surface through an inelastic scattering event. Due to the simplicity of the composition of the sample, details on topography of the nanomembrane was a priority, and that is the reason for using In-lens detector here. All images were collected on a Zeiss Auriga CrossBeam SEM-FIB.

Figure 2: Porous nanomembrane under 1.8kv accelerating voltage, and working distance of 1.6mm with In-lens detector: Low magnification (left) High magnification ( right)

Figure 2(left) shows one of the window on the Silicon wafer. It is obvious that a portion of the nanomembrane has been destroyed due to the fact that the sample was built a couple of years ago. As the magnification increases, i.e. Figure 2 (right), pores on the wafer become visible through dark point. It can be noticed that in some regions, dark point merge together because of the time and made big unique dark regions. Based on the author's experience, In-lens detector is the best detector to obtain appropriate image of the nanoporous membrane. Since the sample was too thin, the image was so sensitive to the accelerating voltage in a way that even a small change in the chosen voltage ended up with blur image.

3. Atomic Force Microscopy

Atomic Force Microscopy was the second technique used to look at the sample. For imaging process, AFM of the Institute of Optics at the University of Rochester was utilized. In this method, a cantilever with a sharp tip is thoroughly scanned a sample to produce the image. The most important advantage of this method comparing with the former one is the high resolution and drawing a 3D image of the sample. However, due to the fact that the sample was too thin, taking an image with AFM was so challenging and time consuming.

Figure 3: 2D images of nanoporous membrane sample: (left) Beam on the window, and (right) Beam far from the window

Figure 3 (left) shows the snapshot when the beam was on the window. In the image, dark points are the pores, and one can say that there is a relative appropriate anology between this image and image in Figure 2 (right). Due to the unstable nature of the window, a setpoint of 8nA with a rate of 0.2HZ was used. 3D distribution of the window has been shown in Figure 4 which represent the topography of the surface realistically. On the other hand, Figure 3 (right) shows the snapshot that the beam was far from the window.

Figure 4: 3D images of nanoporous membrane sample with beam on the window

4. Transmission Electron Microscopy

To better understand the microstructure of the Silicon nanoporous membrane, Transmission Electron microscopy owned by the Institue of Optics at the University of Rochester  was used. Here, the beam of electrons from the electron gun is focused into the beam using the condenser lens. It then collides with the specimen and parts of it will be transmitted. This transmitted portion is focused by the objective lens into the image. The image then passed down the column through the intermediate and projector lenses, magnifying all the way, and turns into the final image that can be observed.

A snapshot of the sample containing several pores has been illustrated in figure 5 (left). It is clear that the image has been taken underfocoused for the sake of  better distinction of the pores' boundary. In figure 5 (right), magnification increased significantly. Therefore, new phenomena including Lattice structure, and grain boundary appeared.

Figure 5: TEM bright field image of Silicon nanporous membrane with: (left) 88kx, and (right) 800kx magniication

4.1. Dark field V.S. Bright field

In the dark field imaging mode, the transmitted beam is excluded from the image formation process. Thus, components who fail to diffract electron beams effectively including pores, look darker than the other components. This explanation can be concluded by a comparison between Figure 6 (left), and (right). Thus, pores look as dark point in Figure 6 (right).

Figure 6: TEM image of sample with 39kx magnification with: (left) Bright field, and (right) Dark field

4.2. X-ray Spectroscopy

To gain information on the elemental composite of the sample, an Energy Dispersive X-ray spectroscopy was done using Transmission Electron microscopy owned by the Institue of Optics at the University of Rochester. In this way, a high-energy beam of particles is focused into the sample. This action, may excite an electron in an inner shell, ejecting it from the shell while creating an electron hole which will be filled by an electron from the outer shell. The difference in energy between the higher energy shell and the lower one may be released in the form of an X-ray which will be measured by an spectrometer. Figure 7 shows the EDX spectrum of the sample. In compliance with the expectation, a high peak of Silicon can be observed in the spectrum. However, a small amount of Cupper is also visible which the is probably related to the sample holder.


Figure 7: EDX spectrum of the sample

5. ImageJ

In this part, porosity of a chosen section of the nanomembrane was calculated based on a rough approxiimation using ImageJ software which is is a public domain, Java-based image processing program developed at the National Institutes of Health. Briefly, the area, the number of the pores, and the are of them was calculated based on the image in Figure 8 (left). It was found that the area of the chosen section is 1.461 um^2 and it contains 102 pores with total area of 0.013 um^2. Thus, the total porosity will be 0.88%.

Figure 8: TEM image of Silicon nanoporous membrane with: (left) 39kx, and (right) 800kx magnification

6. Conclusions

In this project, microstructure of Silicon nanoporous membrane, synthesized by McGrath et. al in 2007,  was investigated using 6 different techniques including SEM with In-lens detector, TEM in bright field, dark field and X-ray spectroscopy, AFM, and ImageJ. Moreover, porosity of a chosen section of the sample was roughly approximated using ImageJ.


I would like to thank Brian McIntyre for advising me through this project, and my TA, Caleb Whittier, for his helps in the SEM labs. I should also thank my Ph.D. advisor, Prof. Abdolrahim, for all her support and useful suggestions in the project, and Prof. McGrath for sharing Silicon nanoporous membrane samples with me.


Please enter any comments, criticisms, questions, etc. below.

Your name:

Email address:


Striemer, C. C., Gaborski, T. R., McGrath, J. L., & Fauchet, P. M. (2007). Charge-and size-based separation of macromolecules using ultrathin silicon membranes. Nature, 445(7129), 749.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature methods, 9(7), 671.