Tired of the laborious cutting of ultrathin corn-root specimens at the rate of 4 to 5 per hour for x-ray study, horticulturalists at Penn State (University Park, PA) tapped the university’s Applied Research Lab for help. In response, Benjamin Hall developed a laser ablation tomography (LAT) technique and formed Lasers for Innovative Solutions (L4iS; www.l4is.com; State College, PA) with graduate student Brian Reinhardt to commercialize the laser-sectioning technique, which speeds the analysis of specimens for biology and materials science. For corn-root sectioning, the method was several orders of magnitude faster than hand-sectioning.
Simultaneous ablation and imaging
Using a 355 nm Q-switched ultraviolet laser source, a sample material is placed on a servo-driven stage and moved into a laser “sheet” formed by coupling the laser to a galvanometer that scans the beam rapidly along a line. The laser sheet effectively ablates the sample material, revealing a cross section that is then imaged with a microscope’s CCD camera with a theoretical resolving power of 800 nm per pixel.
The surface of the specimen is ablated and subsequently imaged, and the process repeats. This allows scientists to image thousands of slices of the specimen, enabling a digital reconstruction of the specimen to be made. The slice data is then available for three-dimensional rendering and color-coding of structural elements within the sample (see figure).
For corn-root sectioning, a 5-mm-long line is created by sweeping the laser at a rate of 1 m/s and the stage is incremented, producing a slice thickness of approximately 0.5–1 μm. In this application, the laser has a pulse-repetition rate between 25 and 40 kHz with a pulse energy of approximately 150 μJ and a spot size of about 20 μm at its focus. A total of 1200 images spaced 3.3 μm apart typically consumes only 40 s of imaging time, with 15 min. of data-processing time.
For an apple-tree graft union, five passes of the laser sheet were required to slice through the material, with 3500 scans spaced 10 μm apart taking 5 hrs. to acquire. Mapping the trajectory of water vessels in the root stock is important to understanding water and nutrient pathways.
Imaged at a resolution of 5184 × 3456 pixels per slice, 1500 slices of a yellow-jacket stinger at 5 seconds of ablation per slice (and using a 160 μJ fluence level at a 28 kHz repetition rate) produced 27 billion voxels of data. To handle that amount of data, the use of a powerful video card, GPU processing, and significant subsampling are required to create digital models. Acquiring the data in full resolution allows for focusing in on particular features in high resolution using Visualization Sciences Group’s (Burlington, MA) Avizo Fire software.
And finally, a section of Marcellus shale consumed 14 s per slice compared to just 500 ms per slice for woody plant material.
“We’re excited to see where this new method takes us,” says L4iS president Benjamin Hall. “Through the use of various spectroscopic techniques, we are currently developing tools to allow deeper insight into the structure, composition, and chemistry of specimens.”
