Automated Separation of Micro-Grains in a Mixed Solid Sample

Areas such as industrial failure analysis, geology, geochronology and others that use solid-phase extraction require accurate measurement of specific micro-grains, which are small particles mixed into a larger matrix. Manual separation works well for grains larger than 100 μm, while grains under this size are best separated using a micromanipulator.

Manual grain separation

Manual separation of micro-grains can be done as follows:

  1. Tungsten probe, for grains smaller than 20 μm: an electrostatic interaction occurs as the tungsten probe nears the grain, causing the grain to adhere to the probe, and allowing it to be moved.
  2. Vacuum, for grains that are 10–100 μm in size: a micropipet with a diameter smaller than the grain is affixed to the motorized arm of a micromanipulator. The operator moves the pipet close to the grain, initiates a vacuum to provide suction and attaches the grain to the tip of the micropipet. Next, the micropipet tip is moved to the location at which the grain should be deposited, the vacuum is turned off and the grain is released.
  3. Micro-tweezers, for grains 20–500 μm in size: micro-tweezers are mounted onto one of the manipulator arms. The user positions the tweezers around the grain and closes them around the particle. After the grain is moved to the new location, the tweezers are opened and the grain is deposited onto the new surface. This approach is useful for moving grains from a sticky surface.

Automated grain separation

 Figure 1 – Axis Pro micromanipulator.

An automated system for grain separation combines an Axis Pro micromaniplator (Micro Support Co. Ltd., Shizuoka, Japan) with WinROOF software (Mitani Corp., Tokyo, Japan) (see Figure 1). The software provides image analysis and identifies grains based on shape, size and/or color.

Using the automated XY stage, the micromanipulator generates a large image of the area of interest by tiling individual images. The user defines the types of grains to be separated, and the system automatically identifies grains in the image area to be separated (see Figure 2).

Once they have been identified, an automatic vacuum micropipet transfers the grains, one at a time. One grain is typically separated every 20–30 seconds.

 Figure 2 – Display of WinROOF software illustrating tiled images and grains to be separated.

Grain identification

WinROOF software distinguishes the grains that should be separated from those that should remain with the bulk matrix based on the following criteria:

  • Color: determined using hue, lightness, saturation or RGB standards
  • Area and perimeter: area and perimeter are measured, without taking into account any voids or holes in the grain
  • Diameter: diameter of a circle that is approximately the size of the grain
  • Maximum length across the grain
  • Maximum length/breadth: ratio of the maximum length divided by the breadth
  • Circularity: degree to which the shape of the grain is approximated by a circle.

Conclusion

Recently, more than 6000 micro-grains were separated in an overnight run using the automated micro-grain separation system, with accuracy of approximately 90%. This combination of a software-driven micromanipulator with image-recognition algorithms provides a powerful tool for the separation of micro-grains in mixed solid matrices.

A video on the micro-support system for automated grain separation can be seen here: https://goo.gl/iAjooP

Steven M. Barnett is principal, Barnett Technical Services, 5050 Laguna Blvd., Ste. 112-620, Elk Grove, Calif. 95758, U.S.A.; e-mail: [email protected]; www.barnett-technical.com. Toshinori Maebayashi is general manager, Micro Support Co. Ltd., Shizuoka, Japan. Satoshi Takeshima is assistant manager, Mitani Corporation, Tokyo, Japan.

Related Products

Comments