Hell published his idea in 1994 together with Jan Wichmann, a lab intern who helped with the numerical simulation. But as a postdoctoral researcher, Hell lacked the means to apply for grants and implement the idea, as well as cover the costs of patenting it. So, he turned to the Fraunhofer Society, a German scientific organization that, at the time, maintained an agency supporting freelance inventors. Fraunhofer provided Hell a stipend covering about 75% of the cost of applying for intellectual property protection in the US and the European Union. (The remaining 25%, about $30,000, he paid out of his own pocket.) In exchange, the Society retained the right to license the intellectual property itself.
What he needed next was a lab in which to actually build his system, and the money to fund the work. The Max Planck Institute provided the former, offering Hell a position in 1997, and even kicked in some of the latter. But it wasn't until the German government ponied up some additional funding in 1998 that he had the resources necessary to begin construction. Finally, in 1999, he had a working system.
Hell's first paper, published that year in Optical Letters, improved resolution from 150 nm to 106 nm—“clearly beyond the Abbe limit,” he wrote, and sufficient to resolve two adjacent nanocrystals that could not be distinguished using standard confocal microscopy. The follow-up report, in PNAS in 2000 (both Science and Nature passed on the manuscript), improved the resolution still further to about 97 nm, and demonstrated the technique's applicability to the life sciences by imaging both yeast and E. coli. The overall improvement amounted to six-fold in the axial (z) direction, and two-fold laterally; subsequent improvements bumped the lateral resolution to six-fold as well. Today, the maximal resolution of STED microscopy is around 50 nm, on the order of the size of a ribosome.The Road to Commercialization
Leica's decision to license STED did not come out of the blue. The company had a history with Hell, licensing an earlier super-resolution technique he had developed, called 4Pi.
“We knew that his ideas had the potential to be converted into products,” says Tanjef Szellas, who heads Leica's confocal team and previously was STED product manager.
Still, the road to STED commercialization wasn't easy. A photograph of Hell's original design, published in a 2006 report of the Max Planck Society entitled Wissen Hoch 12, bears little resemblance to a traditional microscope. Arrayed on an open optical bench measuring perhaps 1 m × 3 m are a series of mirrors, lenses, cables, and other assorted black boxes with silver fittings. A laser beam bounces from mirror to mirror, tracing a zigzag path across the bench. Conspicuously absent are the usual iconic microscope shape, oculars, objectives, and a stage.
The challenge for Leica's engineers was to shrink and fold those components, like optical origami, into a box measuring 25 cm × 25 cm × 15 cm, which could be plugged into the company's existing scanning confocal systems.
The core of the STED development team included just three or four people, says Hilmar Gugel, optics product design manager at Leica and the former STED project manager, though between ancillary players in software, electronics, and engineering, “probably 15 to 20 were involved all together.” Included in that group were several Hell lab ex-pats, including Gugel and physicists Marcus Dyba and Arnold Giske. (Two other Hell lab alumni have also held positions at Leica, helping both with the 4Pi system and a new Hell-developed super-resolution system called ground state depletion and individual molecule return (GSDIM).) “The transfer of the knowledge was facilitated significantly by that,” Szellas says.
According to Szellas, one of the biggest engineering challenges was a direct consequence of the system's heightened resolution. Because standard microscopes are diffraction-limited, their components can be built using tolerances in the 100-nm range. “Any jitter of, say, 50 nm in the laser beam wouldn't be that easy to see with conventional resolution,” he says. But on the other hand, “if you have a movement of the two beams in the STED system, you would see that immediately.”
Hell's team handles this problem the old-fashioned way: by finely adjusting mirrors. But this approach wouldn't work for a commercial system in a closed box, so Leica's engineers installed an autoalignment function. Every two hours or so (depending on environmental variables), users press a button that moves a reference object into the optical path (without removing the specimen under investigation). This object is then imaged using both the excitation and STED lasers, the system adjusting its mirrors until the two superimpose correctly.