Scanning Electron Microscopy of the Human Cornea

Anant Mathur

University of Rochester

OPT407: SEM Practicum
Spring 2005
Final Project

  1. Background
  2. Proposal
  3. Anticipated Time and Techniques
Sample Preparation
  1. Fixing
  2. Critical Point Drying
  3. Epoxy and Embedding
  4. Microtome
  5. Sputter Coating
Imaging Techniques
  1. Light Microscope Imaging
  2. Secondary Imaging
  3. Transmission Electron Microscopy
  4. Colorized Images
Conclusions and Remarks
  1. Conclusions
  2. Remarks
  3. References


1. Background

The cornea is the eye's outermost layer. It is the clear, dome-shaped surface that covers the front of the eye. Although the cornea is clear and seems to lack substance, it is actually a highly organized group of cells and proteins. Unlike most tissues in the body, the cornea contains no blood vessels to nourish or protect it against infection. Instead, the cornea receives its nourishment from the tears and aqueous humor that fills the chamber behind it. The cornea must remain transparent to refract light properly, and the presence of even the tiniest blood vessels can interfere with this process. To see well, all layers of the cornea must be free of any cloudy or opaque areas. The corneal tissue is arranged in five basic layers, each having an important function. The five layers are: epithelium, Boman's membrane, stroma, Descemet's membrane, and endothelium. Figure 1 (Fig. 1), an image acquired from, shows the anatomy of the eye.

Fig. 1. Anatomy of the human eye (


2. Proposal

The cornea sample that is imaged for this project is one that has recently failed for an unknown cause. Cornea’s can fail for multiple reasons including congenital abnormalities, trauma with severe injury and scarring of the cornea, trauma and scarring from infected ulcers of the cornea, hereditary degeneration, and corneal edema. Normally, failed corneas are examined using a light microscope to observe for any apparent reasons of failure. I would first like to image this cornea to observe whether all its layers and structures are in place. I would also like to observe the cornea for any causes of failure such as infiltration, including bacteria build-up or keratocytes. This way, I will be able to identify what may have caused the cornea to fail. Figure 2 (Fig. 2) below is a pictograph of the human cornea at 160X. The image has been taken from A Textbook of Histology written by William Bloom and Don Fawcettto to show an accurate figure of the cross section of the human cornea. The goal of this project will be to image a similar cross section, as well as take detailed images of the five layers.

Fig. 2. Cross section of human cornea at 160X


3. Anticipated Time and Techniques

I. Anticipated Microscopic Techniques:

Secondary electron imaging, critical point drying, coating, microtoming, light electron microscopy, and transmission electron microscopy.

II. Anticipated Non-Microscopic Techniques:

Image colorization using Paint Shop Pro 9

III. Estimated Sample Preparation Time:

I anticipate preparing this cornea sample as I would prepare any other biological sample. This process will include splicing the sample, properly drying the sample, microtoming, and then coating the sample. I will also need to create a fixture for the sample to be able to image its cross section. This will be accomplished by embedding the cornea in epoxy after the critical point drying process. I anticipate the total sample preparation process to take 5-10 hours.

IV. Estimated Microscopic Analysis Time:

10-15 hours

V. Estimated Image Processing Time:

5 hours


1. Fixing

To preserve all important biological structure and keep it intact, it was first import to fix the cornea sample. This was completed by soaking the sample in 2.5% Gluteraldehyde for approximately 30 minutes. Normally, 10-15 minutes is a sufficient amount of time, but because my sample was slightly larger than an average sample, it was soaked for a longer period of time. After soaking in the gluteraldehyde, the sample was sequentially soaked in 100% ethanol in order to replace the moisture in the sample with 100% ethanol. Since ethanol has less surface tension than water, during critical point drying, it can be released from the structure without causing any damage.

2. Critical Point Drying

The big hurdle to overcome in biological samples is the presence of water in living tissue. Though important for the biological sample, water is bad for vacuum systems, and therefore needs to be removed from the sample before going into the SEM vacuum chamber. Simply air drying a sample will not allow for the sample to remain properly intact and can result in the destruction of important structural detail. Therefore, it was important to perform critical point drying on the sample to gradually replace the water in the sample with another solvent that could be conveniently removed. The water in the sample is first replaced by a miscible fluid, commonly a graded ramp of alcohol. This miscible fluid was then replaced by a transitional fluid, in my case, liquid CO2, in a high pressure bomb. This was then removed by transitioning the temperature and pressure of the bomb past the critical point in the CO2 phase diagram. Finally, as pressure is released, the gas was vented, leaving the sample perfectly intact and dry of water, solvent, and CO2.

3. Epoxy and Embedding

The dried cornea sample was next embedded in epoxy and heated in an oven for 24 hours. The purpose of embedding is to ease the process of microtoming a clean edge on the sample, and keeping the small sample intact.

4. Microtome

To obtain a clean edge and a proper cross-section of the cornea sample, the microtome was used to gradually cut the epoxy the cornea was embedded in until the diamond cutting edge of the microtome reached the cornea sample. The microtome then slowly cut a clean edge on the cornea which could be cross-sectionally imaged. During this process, cross-sectioned samples of the cornea were picked up on TEM copper grids to use in the TEM.

5. Sputter Coating

Another hurdle in biological samples is that they serve as insulators to electron beams used for irradiation in SEM. This therefore results in a localized charge on the surface of the sample. To avoid this situation, the cornea sample was next sputter coated with gold for approximately 60 seconds. This provided a thin conductive layer on top of the samples to repel electrons and avoid localized charge.


Identifying biological samples is a difficult task in electron microscopy. A well versed knowledge is required to interpret obtained images. Proper imaging techniques and parameter settings can ease interpretation. In order to image the cornea properly, three imaging techniques were used - light microscopy, secondary imaging in scanning electron microscopy, and transmission electron microscopy.

1. Light Microscope Imaging

Light microscope images are useful prior to electron microscopy. It prepares the user to anticipate what to look for when imaging in the SEM or TEM. Light microscopy also allows for colorized imaging, and most importantly, a quick image of a sample can be collected without the need of a vacuum system. The image below (Fig. 3) is cross-section of the cornea imaged under a light microscope. Though the quality of the image isn't high nor can much be interpreted from it, it does show signs of layers in the cornea and helped identify that the cornea was microtomed properly. It also helped locate the position of the microtomed edge, which thus helped finding the cornea in the SEM quicker. The two layers of the cornea distinguishable in this image are the endothelium and the stroma.

Fig. 3. Light microscope image of cornea sample. Identified layers include the endothelium and stroma.

2. Secondary Imaging

To image the cornea in the Scanning Electron Microscope, secondary imaging and mix modes was used. For the SEM images below, all working distances were set between 14-22mm. The aperture size was set at 1 in order to keep charging of the sample minimum. The accelerating voltage was also kept low, 3-5kV in order to keep the energy of the electrons low, also in order to avoid charging. Magnification of the images varied from 100x to 3000x. At these magnifications, astigmatism generally did not play a large factor, and therefore did not need to be adjusted much. All images were acquired using a slow scan setting of 4. Best results were achieved by mixing the in-lens SE detector and the in-chamber SE detector. The mixing of their images allowed for better quality images and also allowed for adjustment of the contrast and brightness of the images.

Figure 4 (Fig. 4) below is a image of the cornea at 100x. This image nicely shows the image embedded in the epoxy. Because the cornea sample is so small, it needed to be embedded in epoxy in order to properly keep it in place. After embedding the cornea in the sample, a clean edge needed to be cut, and therefore was microtomed. Figure 5 (Fig. 5) is a cross-section image of the cornea at 420X. Visible in these figure is the endothelium, and stroma layers of the cornea, as well as a corneal nerve, which will be elaborated on further ahead. Both of these images were imaged at an accelerating voltage of 5kV, as higher accelerating voltages resulted in charging. A smaller working distance was used for Figure 5 (Fig. 5) because it was placed in the SEM sample chamber in a different orientation than how the sample was placed in figure 4 (Fig. 4). This therefore changed the distance the sample was from the last lens.

Fig. 4. Cornea embedded in epoxy

Fig. 5. Cross-section of cornea at 420X

The first layer of the cornea is the epithelium. However, during the sample preparation process, it is likely that the epithelium was stripped off. The epithelium of the cornea is extremely sensitive and contains numerous free nerve endings. Though the epithelium of the cornea has been stripped in this sample, one of these nerve endings is identified in Figure 6 (Fig. 6). The corneal epithelium rests upon the Bowman's membrane, which consists of randomly arranged collagen fibrils. Though faintly visible, some of these randomly arranged collagen fibrils are identified in Figure 7 (Fig. 7). To adjust contrast and brightness to increase the amount of information that could be interpreted from the images, the mix imaging mode was used for both of these images. This allowed the image to be adjusted for quality to help bring out the necessary detail. Once again, to keep charging of the sample minimum, a small aperture size was used as well as a low accelerating voltage.

Fig. 6. Corneal nerve

Fig. 7. Collagen fibrils

As can be seen from the referenced pictograph in figure 2 (Fig. 2), the stroma of the cornea consists of several long slender fibroblasts. These fibroblasts were imaged in the sample cornea, and can be seen in figure 8 and 9 below (Fig. 8 and 9). These two images are oriented 90 degrees opposite to one another and therefore is the reason why their fibroblasts aren't oriented in the same direction.

Fig. 8. Stroma layer at 3000X

Fig. 9. Fibroblasts in stroma layer

The endothelium is the last layer of the cornea. It consists of a single layer of squamous cells that lines the posterior surface of the cornea, and the Descemets membrane is the basement membrane of the endothelium. Figure 10 (Fig. 10) distinctly shows the endothelium layer of the cornea. Though not clearly visible, Figure 11 (Fig. 11) shows some cells of the endothelium layers. Imaging this section of the cornea was challenging because there was a lot of drifting of the sample. Techniques such as using a smaller aperture and a small accelerating voltage were tried to correct for this drifting, but with no success. Therefore, the sample was removed from the chamber and quickly sputter coated once again for approximately 20 seconds. This attempt did solve the problem with drifting and allowed for proper imaging.

Fig. 10.Endothelium layer

Fig. 11. Endothelium cells

3. Transmission Electron Microscopy

The resolution of a transmission electron microscope (TEM) is higher than a SEM. It is therefore very useful for imaging biological samples. The samples used for the TEM did not have to be embedded or coated. Instead, cross-sectioned slices of the cornea were collected on copper TEM grids as the embedded cornea was being microtomed in preparation for SEM imaging. Images were taken at magnifications of 30X to 125X. The contrast of the images were enhanced by using apertures of various sizes. However, it appeared that during the handling of the sample from the microtome to the copper TEM grids, the samples may have been damaged, and therefore did not provide any relevant images. Time constraints for this project did not permit any further imaging using the TEM.

4. Colorized Images

Colorizing SEM images is quickly becoming a useful technique in electron microscopy. Besides making images more appealing and artistic looking, colorizing images also allows accentuation of contrast and allows one to bring out details that can be missed in a black and white image. The images below were colorized using Paint Shop Pro 9. The images chosen to be colorized were based on images that needed fine details to be better emphasized. Therefore, figures 6 (Fig. 6), 7 (Fig. 7), 10 (Fig. 10), and 11 (Fig. 11) were chosen to accentuate the corneal nerve, collagen fibroblasts, endothelium layer, and endothelium cells. Figure 12 (Fig. 12) was colorized to accentuate the corneal nerve. In the image below, it has been colorized with a pink hue with the rest of the cornea colorized with a purple hue. Figure 13 (Fig. 13) was colorized to bring out the collagen fibrils, which are colored with a red hue and the rest of the cornea colorized with a green hue. Figure 14 (Fig. 14) was colorized to highlight the endothelium layer, indicated with a purple hue. Figure 15 (Fig. 15) was colorized to bring out the endothelium cells in the endothelium layer, which were only faintly visible in figure 11 (Fig. 11). The cells are colorized with a green hue in the endothelium layer, colorized with a purple hue.

Fig. 12. Corneal nerve colorized

Fig. 13. Collagen fibrils colorized

Fig. 14. Endothelium layer colorized

Fig. 15. Endothelium cells colorized



1. Conclusions

I. Techniques Employed:

Fixing, critical point drying, embedding, microtome, coating, light microscopy, secondary imaging in-lens, mix mode, transmission electron microscopy, image colorization.

II. Results:

Though this was a failed cornea, prior to image acquisitions, I anticipated being able to image all five layers of the cornea. Unfortunately, during image acquisition, only 2 layers were clearly distinguished, with the epithelium layer of the cornea likely stripped off during sample preparation. Biological samples are often difficult to interpret without a sound understanding of the sample. Better understanding of the pathology of the cornea would have improved image acquisition and interpretation. Regardless, the cornea sample used in this project allowed for very interesting images and an increased understanding of the cornea as well as imaging techniques of the SEM, TEM, and light microscope. Charging did not play a major issue in imaging, and proper parameter settings such as working distance, aperture size, and accelerating voltage, as well selecting the proper imaging modes and careful sample preparation, allowed for the acquisition of successful images. Originally estimated to take 10-15 hours to acquire images, a total of approximately 17-18 hours was dedicated solely to acquire SEM images.


Thank you to Brian McIntyre for his help and guidance, Ms. Janet Wagner for providing me the cornea sample and guidance, and Dr. James Aquavella for his advice and guidance.


Bloom, William, Don Fawcett. A Textbook of Histology. Philadelphia: W.B. Saunders Company, 1975.

© Anant Mathur, May 2, 2005

--> value="">

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

Your name:

Email address: