As you might expect from its name, a hyperspectral camera does more than reproduce the appearance of objects. The name connotes spectroscopy: For each of its pixels, a hyperspectral camera yields a continuous spectrum over the same, wide waveband.

Given that molecules’ distinctive fingerprints lie in the IR, hyperspectral imaging is especially useful when the camera’s waveband encompasses both the IR and the visible. Indeed, IR-to-visible hyperspectral imagers have found a host of applications, including remote sensing of vegetation, prospecting for oil, and surveillance of criminal suspects.

Most hyperspectral cameras owe their spectroscopic ability to a diffraction grating, which spreads the light from a narrow slit-shaped aperture over a sensor. If the slit is oriented in, say, the x direction, then sweeping the aperture over a scene by means of a movable mirror builds the image in the y direction.

The narrow slit and long focal length yield fine spectral and spatial resolution, but at the expense of throughput (because the aperture is small), camera size (because of multiple optical components), and mechanical complexity (because the optics move). In 2009 Andy Lambrechts and his colleagues at IMEC in Leuven, Belgium, set out to design a cheaper, more compact camera. Last week at SPIE Photonics West, the camera was publicly unveiled for the first time.

Chips and filters

Besides its standard optics, the IMEC camera (shown here) consists of two main components, a CMOS sensor and a Fabry–Pérot or dichroic filter. The camera’s CMOS sensor is the commercially available 2048 × 2048-pixel CMV4000 made by CMOSIS of Antwerp, Belgium. It operates in the same way as other CMOS sensors: Each of its silicon-based pixels converts photon energy into electronic charge, which is then read out by the pixel’s transistor-based electronics.

The Fabry–Pérot filter was developed by Lambrechts’s team and is what makes the IMEC hyperspectral camera special. In cross section the filter resembles a staircase of tiny steps. Interference between the top surface of each step and the bottom surface of the staircase ensures that the step transmits only one spectral band to the sensor below.

The camera unveiled at Photonics West has 100 spectral bands that range from 560 nm (green) to 1000 nm (near-IR), but different and wider wavebands are possible. The upper limit of a CMOS sensor’s waveband is limited to 1125 nm, which is the size of silicon’s bandgap. The lower limit depends on the choice of material and how it is modified. Lambrechts’ team is currently working on a new sensor that reaches 400 nm (violet).

The filter, which is made as a wafer, is fixed directly above the CMOS sensor to create a compact hyperspectral sensor that has no moving parts and which makes simultaneous use of all the light that falls on it. The spectral resolution of the IMEC camera is lower than that of a standard hyperspectral camera, but, thanks to its wide aperture, it has the offsetting advantage of high throughput and therefore fast operation.

At the IMEC booth in the Photonics West exhibition hall, Lambrechts’s colleague Bart Masschelein demonstrated the camera. He had it scan, in less than a second, over a collection of green objects: real leaves, life-like tissue leaves, and pieces of plastic. Masschelein’s computer interface quickly displayed the images and their corresponding spectra, which clearly distinguished the different materials. Masschelein’s software could also classify the objects based on their spectra—that is, knowing what the spectrum of a real leaf looks like, it could find the real leaves among the other objects.

Lambrechts hopes that the IMEC camera, being cheaper and more compact than standard hyperspectral imagers, will find applications beyond the traditional uses. But IMEC is a research center, not a manufacturer. If its cameras are put to work, say, inspecting vegetables on a conveyor belt or watching for poison gas on a battlefield, they will be built by IMEC’s industrial partners.

In 2005 Vanderbilt University’s Anita Mahadevan-Jansen and her colleagues made a remarkable discovery: Pulsed IR light triggers a tightly localized response in mammalian peripheral nerves. What’s more, the pulse energy needed is far below the threshold for damaging tissue.

Mahadevan-Jansen’s team had found, to quote from the abstract of the discovery paper in Optics Letters, “a simple yet novel approach to contact-free in vivo neural activation that has major implications for clinical neurosurgery, basic neurophysiology, and neuroscience.” Followup research supports that optimism. Infrared neural stimulation (INS) evokes responses not just in nerves that control a mammal’s limbs, as was the case in the original paper, but also auditory and cardiac nerves.

The mechanism behind INS is thermal. In a series of experiments published five years ago in the Biophysical Journal, the Vanderbilt team and their collaborators from the University of Texas at Austin, investigated effects that could plausibly underline INS, including photochemistry, pressure, and electromagnetism. Their conclusion: The moderate, transient tissue heating engendered by the IR pulse activates transmembrane ion channels and begets action potentials—that is, triggers the nerves to fire.

Unlike electrical stimulation, which requires electrodes to be inserted to tissue, INS can be delivered without contact. Given the compact size of solid-state light sources, INS-based cochlear implants, pace makers, and other medical devices are conceivable—at least in principle.

From PNS to CNS

At this year’s SPIE Photonics West, one of Mahadevan-Jansen’s Vanderbilt colleagues, Anna Wang Roe, reported on her team’s efforts to advance INS beyond the peripheral nervous system and into the central nervous system. The goal is challenging. As Roe notes in a recent paper, peripheral neurons tend to be organized in parallel bundles like wires and cables in utility pipes. In contrast, nerves in the brain are organized in complex hierarchical three-dimensional architectures. What’s more, cerebral neurons interact with astrocytes and other brain cells that may affect the neurons’ response to INS.

The focus of Roe’s talk was on her team’s use of both functional magnetic resonance imaging (fMRI) and optical imaging to map the cerebral cortex of macaque and squirrel monkeys. The cerebral cortex, which forms the outer part of the mammalian brain, is important for several high-level functions including attention, memory, and language. Research has shown that specific functions are controlled by clusters of nerves that occupy distinct domains about 200 μm across.

The fMRI scans can pinpoint which regions of the brain become active in response to a stimulus, such as a blinking light, a burst of noise, or even—in the case of human subjects—a verbal suggestion. The technique works because brain cells, when called into action, require energy, which they obtain, in part, from a sudden inrush of oxygenated blood. Oxygenated hemoglobin, being less magnetic than deoxygenated hemoglobin, shows up in fMRI. Roe’s lab is one of the few in he world that has an MRI scanner (shown here) whose bore is of a suitable size and orientation for monkey studies.

Oxygenated hemoglobin is also redder and brighter than deoxygenated hemoglobin. The sudden inrush of blood to a region of the brain is also visible—literally—if the region under investigation is exposed and close to the surface. The difference in red brightness is not difficult to detect. Roe’s lab uses a red laser and a commercial CCD camera.

It’s more difficult to expose an area of the monkey’s cerebral cortex to visualize those changes in oxygenation. To meet that goal, Roe and her collaborators surgically implanted an optical window in the monkey’s skull, a procedure involving replacement of a section of the skull and native dura (the protective membrane surrounding the brain) with a circular chamber and a biocompatible transparent, artificial dura. Besides providing a way to view the cerebral cortex, the window also allows an INS signal to be delivered.

At this stage of their research, Roe and her team have succeeded in associating some brain functions, like attention, with cortical domains. They have observed the telltale reddening of the domains when the monkeys are called on to perform a task. And they have confirmed, as her colleague Mahadevan-Jansen has done for the peripheral nervous system, that INS does not cause visible, lasting damage to tissue.

What remains to be demonstrated is whether INS can potentially be used to achieve therapeutic effects in the central nervous system. Enhancing or restoring lost function is one possibility. At the end of her talk, Roe speculated that some symptoms of autism and attention deficit hyperactivity disorder might one day be treatable with INS.

Microfluidics is a technique for controlling, processing, and analyzing small volumes of liquid in submillimeter devices. The small scale befits biology especially. Not only are cells and microorganisms themselves small, but they are often available only in small, precious quantities.

Monitoring a microfluidic device is usually done through an optical microscope, which is typically a hundred times as big as the device itself. What’s more, because conventional microscopes rely on lenses, high spatial resolution is obtained at the expense of a small field of view. Paradoxically, big microscopes yield small images.

At this year’s SPIE Photonics West, graduate student Serhan Isikman from Aydogan Ozcan’s biophotonics lab at UCLA described a lensless tomographic imaging system whose 4- by 6-mm CMOS sensor array is not much bigger than a typical microfluidic device. Like a hospital CT scanner, the UCLA imager takes multiple two-dimensional images from which 3D images are reconstructed. Isikman and Ozcan call the technology optofluidic tomography.

The figure shows the basic principle. Each 2D image consists of a so-called in-line hologram. Spherical wavefronts emanate from a coherent light source and either pass right through the objects in the microfluidic device or are scattered and absorbed by them. What the CMOS sensor records is the interference pattern formed by the unaffected and affected wavefronts. The 2D image is reconstructed directly from the hologram (that is, the interference pattern) without the need to send a second reconstruction beam through the hologram, as is the case in standard “off-line” holography.

The 2D images are created in quick succession by a bank of LEDs arranged in a circular arc above the microfluidic device. A second round of reconstruction, called filtered back projection, is used to create a 3D image from the 2D images.

Because the UCLA system lacks magnification, its spatial resolution is ordinarily limited by the pixel size of the sensor, 2.2 μm. But finer resolution can be obtained thanks to a technique called pixel superresolution (PSR). Assuming that an object doesn’t change size or shape as it passes over a patch of pixels, you can use the signals recorded in the pixels to interpolate the object’s shape. If the object is stationary, rocking the light sources with electromagnetic actuation enables PSR to be applied.

Using PSR, Isikman, Ozcan, and their colleagues have obtained a resolution of less than 1 μm. That’s comparable to the resolution of an optical microscope, but it’s achieved over a bigger field of view (about 100 times as wide as an optical microscope at ×40 magnification). Despite its significantly enlarged field of view, the lensless tomographic microscope can fit in a small volume of 96 × 89 × 40 mm, and weighs only about 110 grams.

In his talk, Isikman reported results of imaging two microorganisms, the nematode worm and model invertebrate Caenorhabditis elegans and the eggs of the dwarf tapeworm Hymenolepis nana.

The lensfree tomography and computational microscopy work of Ozcan’s bio-photonics lab has yielded several patent applications over the last few years. They’re licensed by Holomic LLC, a startup based in Los Angeles.

I’m spending the first half of this week at the Biophysical Society’s annual meeting, which is being held in Baltimore. The meeting is large. More than 6500 biophysicists from around the world are attending. Rather than try to provide an overview, I’ll focus instead on a few talks that I find surprising and interesting. If I liked them, maybe you will, too.

In yesterday’s New and Notable symposium, I learned about an innovative optical technique for tracking the three-dimensional paths taken by biomolecules as they carry out their evolution-ordained tasks. In general, you can’t use an optical microscope to locate individual protein molecules because the wavelength of visible light (380–750 nm) is a hundred times larger than the molecules themselves.

However, if you know how your microscope blurs the light from a point source—that is, if you know your microscope’s point spread function (PSF)—you can circumvent that diffraction limit. The trick is to insert a fluorescent tag into the molecules you want to track. Most microscopes have axially symmetric PSFs. If yours does too, you’ll know that when you observe glowing disks of fluorescence hundreds of nanometers in diameter, the protein isn’t just anywhere in the disk; it’s dead center.

Superlocalization, as this principle is known, can track molecules as they move from side to side, but it has a tougher time following their up and down motion, even within the microscope’s shallow focal plane. That’s because the PSF’s z component resembles a vertically oriented sausage. You can’t tell whether a tag is in the middle or at one of the ends of the sausage.

In his talk yesterday, Stanford University’s W. E. Moerner explained how his group is able to apply superlocalization in 3D. Following a recipe devised in 2000 by Rafael Piestun, Yoav Schechner, and Joseph Shamir, Moerner and his group create focused beams of light whose PSF in the z direction resembles not one but two sausages.

What’s more, the sausages wind around each other in helical pattern. Consequently, a horizontal slice through the PSF consists of two spots whose rotation angle in the xy plane corresponds to depth in the z direction. The upshot, as Moerner and his team demonstrated in 2009, is that they can pinpoint the location of a fluorescent tag with 20-nm precision throughout their microscope’s 2-μm focal depth.

In the 2009 proof-of-principle experiment, fluorescent molecules were scattered about and fixed in a polymer substrate. In his talk yesterday describing recent 2010 work in collaboration with Stanford’s Patrick O. Brown, Moerner showed the results of tracking the progress of messenger RNA molecules as they emerge from transcription and participate in translation.

For me, the most amazing aspect of the technique is not the valuable biological information that it yields, but the way one generates the double-helix PSF: by simply inserting a suitably patterned mask in an otherwise standard microscopy system. Of course, you do need to know how to pattern the mask. See the 2008 paper by Sri Rama Prasanna Pavani and Rafael Piestun for details.

Charles Day

For some products, even fairly new ones, you can predict whether the latest versions will be bigger or smaller than their predecessors. Disk and flash drives are getting smaller (or staying the same size while storing more data), TV screens are getting wider, and laptops are getting slimmer.

But if you’d asked me before today to predict the size trend of adaptive optics systems, I wouldn’t have said they’re shrinking.

A perfect parabolic mirror will bring an object into perfect, diffraction-limited focus, provided the wavefronts radiating from the object are parallel and flat. Starlight arriving at a ground-based telescope doesn’t have parallel, flat wavefronts. Fluctuating, uneven refraction caused by atmospheric turbulence bends the wavefronts out of shape. The focus is imperfect.

In 1953 Horace Babcock of the Mount Wilson and Palomar Observatories in California proposed a way—adaptive optics—to compensate for atmospheric blurring. If you could measure the distortions as they happen, and if you could deform the telescope mirror quickly and arbitrarily, you could restore the wavefronts to their flat, parallel state.

Babcock’s idea was ahead of its time. Earth’s atmosphere fluctuates on a time scale of 10 to 100 milliseconds. Measurement and control systems from the 1950s through the 1980s couldn’t keep up. But since the early 1990s, adaptive optics systems have been installed at several observatories, including the two Keck telescopes on Mauna Kea in Hawaii and the Very Large Telescope (VLT) on Cerro Paranal in Chile. The three huge telescopes under development now—the European Extremely Large Telescope (E-ELT), the Thirty Meter Telescope, and the Giant Magellan Telescope—would not be worth building without adaptive optics.

The E-ELT’s primary mirror is 42 meters wide. That’s five times bigger than the primaries in each of the four VLT telescopes. Because I had just read a news story in Nature about the E-ELT, the association of adaptive optics with mirrors the size of swimming pools was freshly established in my mind.

Michael Helmbrecht’s talk at SPIE Photonics West soon wiped out my mistaken prejudice. Helmbrecht is the CEO and owner of Iris AO Inc, a company based in Berkeley, California, that makes miniaturized adaptive optics systems.

The photo shows Iris AO’s PTT111-X deformable mirror. The product fact sheet lists its impressive features, among them its aperture: 3.5 mm. One hundred eleven tiny MEMS (microelectromechanical systems) actuators deform the mirror’s 37 segments.

NASA and the Pentagon are Iris AO’s current customers. After the talk, I asked Helmbrecht what the first commercial applications for his products would be. High-end microscopes for biology was one area. Lasers for semiconductor fab plants was another. As for retinal imaging, an area where bulkier adaptive optics systems are already in use, he said the price of his devices would first have to drop significantly.

Disk and flash drives, big-screen TVs, and laptops have all gotten cheaper. I expect MEMS-based deformable mirrors will, too.

Charles Day

When pharmacologists develop anticancer drugs, they need a way of seeing whether the drugs really do shrink and kill tumors. Magnetic resonance imaging can do the job without harming human patients or lab animals, but it’s expensive, especially if you need high spatial resolution. Biopsies are cheaper, but they risk interfering with the growth—or, one hopes, the shrinkage—of the tumors under investigation.

Yesterday at SPIE Photonics West, I learned about an optical method for tracking tumors: diffusive optical tomography. In DOT beams of near-IR light are sent through tissue at various angles and detected when they emerge, much depleted by both absorption and scattering. Wavelengths are chosen to reveal differences in concentration of various molecules, including hemoglobin.

Forming a three-dimensional image from those diffuse signals is doubly challenging. First, detecting the signals requires a sensitive instrument. You can’t arbitrarily increase the incident intensity to boost the output signal lest tissue be damaged. Second, untangling the paths taken by the light as it makes its way through a scattering medium is a formidable mathematical problem. Those and other challenges have been met. Diffuse optical tomography is now a cheap, fast, and effective imaging modality for use in the lab and the hospital.

In one talk I heard yesterday, Molly Flexman of Columbia University in New York City described how her group is using DOT to monitor the efficacy of bevacizumab, a drug designed to kill tumors by cutting off their ability to grow blood vessels and therefore their ability to obtain nutrients.

Bevacizumab is controversial. Marketed by its maker Genentech under the name Avastin, the drug appears to shrink some tumors, but not others. That known, variable performance suits Flexman because it gives her the opportunity to test whether DOT can indeed monitor the efficacy of any anticancer drug. Moreover, whether or not the drug targets a tumor’s blood vessels, DOT can readily detect them thanks to its sensitivity to hemoglobin.

Flexman and her colleagues looked at two types of cancer: Ewing’s sarcoma, for which bevacizumab has some beneficial effect, and neuroblastoma, for which bevacizumab is less effective. It turned out that DOT could reveal the difference in how the two cancers responded to the drug.

For her study, Flexman used lab mice. In the introduction to her talk, she explained why she and her colleagues chose to focus on Ewing’s sarcoma and neuroblastoma. Both cancers afflict children. Because children are still growing, therapies that attack cells’ ability to divide, such as ionizing radiation, have worse side effects for children than for adults.

Charles Day

“Snapshots of cooperative atomic motions in the optical suppression of charge density waves” might not seem like an exciting title. But the paper, which appeared in last Thursday’s Nature was my favorite of last week.

Charge density waves are periodic distortions in crystals that consist of stacked chains or planes of atoms. Above a critical temperature, the valence electrons are free to move about the crystal. Below that temperature, the atoms shift positions slightly, moving closer to some neighbors than to others. The shift has the effect of marooning the electrons in puddles around the atoms, like fish trapped in rockpools at low tide. The crystal becomes an insulator and acquires an additional periodicity.

Rudolf Peierls worked out the basic physics of charge density waves in a 1934 paper. One-dimensional crystals, Peierls pointed out, are intrinsically unstable and inevitably buckle. Charge density waves are manifestations of the Peierls distortion in bulk materials. Neutron scattering experiments confirmed the waves’ existence in the 1950s.

Charge density waves, then are well established. What got me excited about the new paper are the words “snapshot” and “cooperative atomic motions” in its title. The nine-author team from universities in Canada, Germany, and Slovenia has succeeded in observing a Peierls distortion on the few-hundred-femtosecond timescale over which it occurs.

To pull off that coup de recherche, the team took a thin crystal of one of the best-studied charge-density-wave materials, 1T-TaS₂, and chilled it below the temperature at which charge density waves appear. Next, the team zapped the crystal—first with a brief pulse of light and then, after a precisely controlled and adjustable delay, with a brief pulse of electrons.

As an electron pulse passed through the crystal, it diffracted off the regular array of atoms to form a pattern that embodied the crystal’s instantaneous structure. From those patterns collected at different delays, the team could track how charge density waves form.

The light pulse gave the valence electrons enough energy to flow freely. Responding to the electrons’ liberation, the more sluggish atoms started to shift back to their undistorted positions. However, even as the atoms were shifting, the electrons began to cool and sank back into confinement around the atoms. The atoms, too, regained their original, Peierls-distorted configuration.

In the schematic, which depicts the process, the atoms are red and the electron distribution is purple: light for low density; dark for high density. The whole process took less than 4 picoseconds to play out.

The paper’s implications go beyond charge density waves. As the authors point out at the very beginning of the abstract:

Macroscopic quantum phenomena such as high-temperature superconductivity, colossal magnetoresistance, ferrimagnetism and ferromagnetism arise from a delicate balance of different interactions among electrons, phonons and spins on the nanoscale. The study of the interplay among these various degrees of freedom in strongly coupled electron–lattice systems is thus crucial to their understanding and for optimizing their properties.

Femtosecond electron diffraction, as the technique is called, should yield a host of interesting results in the future. It could even prove decisive in solving one of the hardest problems in physics: explaining high-T_c superconductivity.

Charles Day

Some new techniques and ideas are so interesting and potentially important that their inventors don’t wait to prove they’re useful before writing a paper. Instead, after describing the ground-breaking work, the inventors outline possible applications that they and others might one day realize.

The delay between invention and application can be brief. In April 2003 I wrote a news story about the creation of precisely timed bursts of extreme UV light that last a few hundred attoseconds (10⁻¹⁸ s). Within a year, one of the inventors, Ferenc Krausz, had used the technique to probe the motion of electrons inside neon atoms on timescales shorter than the time it takes the electrons to orbit the atoms.

Krausz was applying his own technique. This morning I encountered a freshly published paper in Nature Nanotechnology, which, if taken at face value, suggests that its authors, Glasgow University’s Malcolm Kadodwala and his collaborators, implemented someone else’s idea within six months.

The idea was published in the 23 April issue of Physical Review Letters. Harvard University’s Yiqiao Tang and Adam Cohen used James Clerk Maxwell’s famous equations to identify electromagnetic waveforms whose degree of chirality, or handedness, is stronger than that of left- or right-handed circularly polarized light. Such light, Tang and Cohen proposed, could serve as a sensitive probe of amyloid fibrils, viruses, and other biomolecular aggregates whose constituents are themselves chiral.

Although I can’t be sure, I’m guessing that Kadodwala or one of his colleagues read Tang and Cohen’s paper and sprang into action. In the introduction of the Nature Nanotechnology paper, Kadodwala writes:

Recently, it has been postulated that under certain circumstances superchiral electromagnetic fields could be produced that display greater chiral asymmetry than circularly polarized plane light waves. We have realized that such superchiral electromagnetic fields are generated in the near fields of planar chiral metamaterials (PCMs), which can greatly enhance the sensitivity of a chiroptical measurement, enabling the detection and characterization of just a few picograms of a chiral material.

Kadodwala and his coauthors go on to describe fabricating a PCM, whose periodic features are electrically conducting and of the same few-hundred-nanometer scale as UV and visible wavelengths.

Illuminating the PCM excites plasmons, thereby generating short-range superchiral light. When Kadodwala’s team immersed the PCM in a solution of chiral molecules, they detected strong, telltale resonances whenever the chirality of the PCM matched that of the molecules—just as Tang and Cohen had predicted.

It’s possible that Kadodwala really did find out about Tang and Cohen’s idea by reading their paper. One of the marvels of modern science is how conveniently papers are disseminated.

On the other hand, it’s just as conceivable—and perhaps more comforting—that members of the two teams met one day at a conference and decided over beer, coffee, or other social lubricant to work together.

Charles Day