Examination of Snowflake Replicas

Dan Easterly
University of Rochester
The Institute of Optics



Snowflakes are difficult to examine outside of a frozen environment. In order to facilitate this process, I researched some methods of preserving snowflakes so that they could be examined in the hostile environment of a scanning electron microscope. The method I found involves using a 0.1% solution of Formvar (formal acetate) dissolved in Chloroform. The sample stages are placed outside in the snow, along with the Formvar solution, and allowed to reach ambient temperature. After everything is at temperature, the stages are exposed to a light snowfall. When some flakes have accumulated on the stages, several drops of the solution can be placed on the snowflakes, and the stages can be covered once again. After a short time, 10 or 20 minutes, the chloroform will have evaporated and leave behind the the Formvar and snow. At this point, the sample stages are ready to be brought inside. The Formvar will have made a replica of the snowflakes so that they can melt and evaporate. Now that replicas have been cast, the stages can be viewed under a light microscope in order to select the best samples. In order to view the snowflakes in the SEM, they need to be sputter coated with gold. Sputter coating consists of placing the stages on an anode in a vacuum jar. Current is applied to the cathode, and gold atoms are conformally deposited onto the replicas. Coating the samples with conductive materials prevents electrons from building up on the sample inside the SEM.


The primary focus of this project lies in being able to replicate snowflakes and utilize the different imaging modes of the SEM to view them. In addition, I have colorized several of the images and even made use of stage tilt in order to later create stereo pairs and anaglyphs. Stereo pairs are two offset images set side by side so that, either by crossing your eyes or defocusing them, the images will merge into a three dimensional image. One interesting trick I discovered while making pairs was to colorize the images differently so that the merged image will actually shift back and forth between the two colors.

Excepting my color-blindness, anaglyphs aren't too difficult to create either. The same images that were used to make stereo pairs are opened in PhotoShop and then colorized as follows: the image on the left is colored red and the image on the right is colored cyan. The two images are then overlapped with one of the layers being partially transparent so that both images can be viewed simultaneously. If the viewer has stereoscopic glasses, (left eye has red filter and right eye has cyan filter), then the resultant image will be viewed in three dimensions.

Another technique I utilized in trying to make images seem more interesting was to colorize the images. Given the complexity of the snowflake images, I didn't feel justified in adding in more than one shade of color per image.

Here is an interesting link that shows some colorful photographs of snowflakes.


I must admit that some of my earlier attempts at snowflake replication weren't nearly as successful. Here's an image using a mixture of backscattered and secondary electron detection followed by a colorized version.

In addition to imaging with in-chamber secondary electron detector, I also used in-lens SE detector.


Lastly, this is an image using the invert function on the SEM with post-processing colorization.



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

Your name:

Email address: