Recently in Microscopy Category

Whether rosy or rich, the color of human skin comes from a pigment called melanin, which occurs in two forms: Pheomelanin acts as a photosensitizer for UV radiation, while eumelanin has a protective role against light and seems to be overabundant in melanoma, a dangerous form of skin cancer. Melanoma is typically diagnosed by removing a lesion and examining it microscopically, layer by layer. Several noninvasive imaging techniques have been explored to detect the cancer, but they lack the molecular specificity needed to distinguish the two melanins. Enter researchers from Duke University who address that shortcoming with a nonlinear optical pump–probe technique. They excite the molecules with an ultrafast laser pulse and, after a varying time delay, measure the absorption of a second pulse—the probe. The contrast between the two melanins arises from their different time-delay absorption profiles. The white-light image on the left shows a melanoma lesion on human skin grafted onto a live mouse. The false-color image on the right, taken at a depth of 45 µm with the new technique, shows eumelanin in red, pheomelanin in green, and multiphoton fluorescence in blue. Previously the group used this method to image the tiny blood vessels in a mouse's ear. Such architectural views of skin coupled with the new functional mapping could provide the basis for a useful noninvasive diagnostic and screening method, particularly when removing tissue is problematic—as with melanoma in the eye. (T. E. Matthews et al., Biomed. Opt. Exp. 2, 1576, 2011.)—Stephen G. Benka

In 1805 Thomas Young derived the relationship between the static forces of a liquid droplet at rest on a solid substrate. As shown in the schematic, the droplet’s contact angle, θ, can be determined by balancing the horizontal solid–liquid and solid–vapor forces and the horizontal component of the liquid–vapor force, determined by the surface tension, γ/v (see Physics Today, February 2007, page 84). However, for more than two centuries, theories yielded unrealistic singularities for stress and strain at the three-phase contact line. That’s because Young’s equation does not account for the vertical, out-of-plane force pulling on the solid substrate, which naturally should be balanced by the substrate’s elastic response. Now, researchers at Yale University and at consumer products manufacturer Unilever have experimentally and theoretically resolved the out-of-plane contributions. Using a confocal fluorescence microscope, the researchers, led by Yale’s Eric Dufresne, laced a 20-micron-thick film of silicone gel with fluorescent beads and measured the deformation due to a water droplet. At equilibrium, a one-micron-high ridge, illustrated in the inset, formed in the gel at the contact line. When the researchers factored the gel’s surface tension and thickness into a linear elastic model, they arrived at a nonsingular theoretical solution for stress that closely fit their experimental data. Their model, however, underestimates the deformations in the solid-liquid contact plane, which they believe are caused by pinning or viscous drag. (E. Jerison et al., Phys. Rev. Lett., in press.)—Jermey N. A. Matthews

When imaging, monitoring, or stimulating samples in a scattering medium, even the most powerful optical microscopes and probes are hindered by the diffusion limit, the length scale beyond which the incident light uncontrollably scatters. To overcome that limit, some techniques focus the wavefront as it propagates through the sample; others iteratively shape it to amplify the target signal. Lihong Wang and colleagues at Washington University in St. Louis have developed a new approach that combines time reversal with ultrasound, whose waves scatter weakly in biological tissue, to focus light to a controllable position. (For an introduction to multiwave imaging, see the article by Mathias Fink and Mickael Tanter in Physics Today, February 2010, page 28.) In the team's setup, laser light, frequency-shifted by two acousto-optic modulators in series, entered the sample medium, where the diffused light was further modulated by an ultrasonic wave tuned to the frequency shift. The interaction between the light and the ultrasound produced a virtual point source within the sample. From a holographic record of the modulated diffused light, the researchers generated a time-reversed trajectory that produced optical focusing at the virtual source location. The new technique, known as time-reversed ultrasonically encoded (TRUE) optical focusing, generated a noticeably higher contrast for objects inside a slab of synthetic biological tissue than was attained by conventional ultrasound-modulated optical tomography, which cannot focus light. (X. Xu, H. Liu, L. V. Wang, Nat. Photonics, in press, doi: 10.1038/nphoton.2010.306M.)—Jermey N. A. Matthews

Messenger RNA (mRNA) is the shuttle that carries genetic information across a cell’s nuclear membrane and into the cytoplasm, where the information is translated into a protein sequence. However, the movement of mRNA, which is about 25 nm in diameter, through the nuclear pore complex, roughly 120 nm in diameter, has been difficult to resolve visually since both mRNA and the NPC are well below the 200-nm diffraction limit for optical microscopes. Now, Robert Singer at Yeshiva University’s Albert Einstein College of Medicine in New York and David Grünwald at the Delft University of Technology in the Netherlands have developed a new nanometer-resolution imaging technique that they used to track mRNA’s passage. Emission signals from mRNA and the NPC—labeled with spectrally different fluorescent probes and shown in the image as green and red, respectively—were chromatically separated and tracked by two synchronized high-speed CCD cameras. The researchers achieved 26-nm spatial resolution by resolving the misalignment between the mRNA and the NPC in the images collected before and during tracking by both cameras—a technique they’ve dubbed super-registration microscopy. In their experiments, they also observed that mRNA spent about 5-20 ms crossing the NPC, a fraction of the time it spends at the pore’s entrance and exit—as if, say the researchers, the mRNA was being double screened for quality. That information may support research into understanding how defective mRNA is prevented from escaping the nucleus. (D. Grünwald, R. H. Singer, Nature, in press, doi:10.1038/nature09438.)—Jermey N. A. Matthews

Related link:

Robert Singer explains super-registration microscopy

Genetic information is transcribed from DNA to RNA and translated from RNA to make proteins. Because each step entails a modest number of molecules, gene expression, as the DNA-to-protein conversion is termed, is inevitably noisy: Identical genes in identical cells don't yield identical numbers of proteins. But how noisy? Sunney Xie of Harvard University and his collaborators have used single-molecule fluorescence microscopy and microfluidics to find out. They started by modifying the DNA of Escherichia coli to create 1018 different strains of the single-celled bacterium. In each strain, the code for a yellow fluorescent protein (YFP) was inserted after the gene for a different protein. To see the rate at which one gene is expressed in one cell of one strain, you’d illuminate the cell with a laser and measure the YFP emission through a microscope. To gather gene-expression statistics for a sample of cells from all 1018 strains, the Harvard team sent streams of cells through channels cut in a microfluidic chip and imaged them. The figure shows sample images for three proteins, YjiE, AtpD, and Adk. Ninety-six strains could be processed at once at a total throughput of 160 cells per second. The team found that the least abundant proteins appear at 10⁻¹ molecules per cell; the most abundant, at 10⁴ per cell. Gene expression is indeed noisy, but with a twist. As you’d expect, the least abundant proteins have the largest cell-to-cell fluctuations. But for proteins whose mean abundance is 10 per cell or higher, the expression noise saturates, presumably because the various molecules that mediate gene expression inside a cell are in limited supply. (Y. Taniguchi et al., Science 329, 533, 2010.)—Charles Day

In recent years, notions of the ultrafast, the ultraintense, and the ultrasmall have been recurring themes in physics as those envelopes have been relentlessly pushed to reveal new phenomena. Caltech’s Brett Barwick, David Flannigan, and Ahmed Zewail have combined all three notions into a new technique they dub photon-induced near-field electron microscopy. PINEM exploits the fact that free–free interactions of electrons and photons are greatly enhanced when a third body, like a nanostructure, is present and when the electrons are more energetic than the photons. The physicists illuminated a carbon nanotube with an intense femtosecond laser pulse that generated an evanescent plasmonic field at the CNT’s surface. Simultaneously, a similar-duration pulse of 200-keV electrons from an electron microscope traversed the sample. During the few-hundred-attosecond interaction time, some of those electrons absorbed energy quanta from the 2.4-eV photon field. By selecting only those electrons that gained energy, the researchers could image the evanescent surface field with the spatial resolution of electron microscopy. That field extends about 50 nm into the vacuum from the dark surface of the roughly 150-nm-diameter CNT. As shown in the images, Zewail and colleagues also monitored the temporal decay of the surface field by varying the delay times between the exciting laser pulse and the probing electron pulse, from zero (top) to 400 fs (bottom) and beyond. With tunable and temporally controlled light pulses, PINEM enables visualization of dynamical optical responses of various nanostructures. (B. Barwick, D. J. Flannigan, A. H. Zewail, Nature 462, 902, 2009.) —Stephen Benka

MRI excels at revealing subtle features in soft tissue. Hydrogen nuclei are detected through the electromotive force induced in a nearby coil when their spins flip from an RF pulse. Typically, one coil transmits the pulse and another detects the induced signal. That configuration has been used in clinical settings for decades with imagers built from 1.5 T magnets. In recent years, imagers have been developed with greater field strength to boost sensitivity. But as the field increases, so does the resonance frequency required to excite nuclei. The corresponding wavelength in tissue in a 7-T magnet is about 12 cm, on par with the size of resonator coils that encircle a human head. The result: interference and standing-wave RF patterns. Those inhomogeneities in the RF field are disastrous because they perturb the image contrast between different types of tissue. A group led by Klaas Pruessmann at ETH Zürich has now solved the problem by removing the RF coils entirely and using the conductive lining in a 7-T MRI cavity as a waveguide with an antenna placed at one end. A patient inside the waveguide is exposed to a homogeneous traveling RF wave launched from the same antenna that subsequently detects the spin signals. The in vivo images of a human leg demonstrate that traveling-wave MRI (left) can excite spins more uniformly than can inductive MRI (right). The researchers speculate that, with the tight-fitting induction coils gone, patients may suffer less from claustrophobia and engineers may enjoy more design freedom. (D. O. Brunner, N. De Zanche, J. Frölich, J. Paska, K. P. Pruessmann, Nature 457, 994, 2009.) — R. Mark Wilson

Ultimately, the resolution of an electron microscope is limited by the electron's de Broglie wavelength. For the 300-keV electrons typical in scanning transmission electron microscopy, that limit is about 2 pm, or 1/25th of the radius of hydrogen's 1s orbital. But STEM images are formed by focusing a billion or so electrons per second onto a sample. The spherical aberration of the electromagnetic lenses and the finite size of the electron source cause the electrons to lose phase coherence, lowering the resolution to about 100 pm, or about twice the distance between atoms in many crystals. Now, a team from Lawrence Berkeley National Laboratory in California has halved the STEM resolution limit to 50 pm. The boost in performance comes from two novel components: an electron source that emits copious electrons from a region just 25 pm across and a hexapole corrector that can compensate for phase aberrations up to fifth order. Using their new microscope, the LBNL researchers looked at a piece of germanium foil. According to x-ray crystallography, Ge atoms are arranged in rows of dumbbell pairs aligned end-to-end. Ordinarily, the dumbbells are too small to be resolved with STEM. But, as the accompanying figure shows, the LBNL microscope could resolve the 47-pm separation between two paired atoms. The resolution is so fine that the thermal jiggling of the atoms during the room-temperature measurement acts as an additional source of blur. (R. Erni, M. D. Rossell, C. Kisielowski, U. Dahmen, Phys. Rev. Lett., in press.) — Charles Day

Inorganic crystal aggregates known as biomorphs earn their name by virtue of a remarkable resemblance to the fossils of primitive organisms. But although the structures can be varied and complex—leaflike sheets, wormy ropes, and helical filaments, among others—biomorphs are exceedingly simple to make. They self assemble when an alkaline earth halide such as barium chloride is mixed with a silica-rich solution under high pH conditions at ambient pressure and temperature. As carbon dioxide from the air dissolves into solution, barium carbonate and silica precipitate out and produce the complex structures. A long-standing question is how. Juan Manuel García-Ruiz, his postdoc Emilio Melero-García (both at the University of Granada), and Stephen Hyde (Australian National University) now propose a mechanism for the morphogenesis. As the carbonate crystallizes, it lowers the pH of the local environment and triggers the precipitation of silica. The silica precipitation, in turn, raises the local pH, which prompts another round of carbonate formation. The sensitivity of silica and carbonate species to opposite trends in pH fluctuation creates a chemical feedback that eventually produces rodlike, carbonate particles, each coated with silica. Freed from the hexagonal symmetry restrictions imposed by carbonate growth, the silica-coated nanoparticles form clusters that can adopt various shapes. For reasons unexplained by their mechanism, the clusters align themselves on the micron scale and grow as two-dimensional sheets. The edges of those sheets can then curl like a scroll to create the sort of curved and twisted filaments captured here by optical microscopy. (J. M. García-Ruiz, E. Melero-García, S. T. Hyde, Science 323, 362, 2009.) — R. Mark Wilson

To map molecules in cells and tissue, researchers prefer biomedical imaging techniques that rely solely on the intrinsic responses of chemical bonds to optical stimulation. Although fluorescence microscopy and other chemical tagging methods yield high-resolution images, they also introduce foreign species or synthetic derivatives that can alter the dynamics of intracellular processes. Spontaneous Raman scattering, which uses a single laser beam to excite the vibrational and rotational modes in chemical bonds, requires no chemical labels but generates a weak signal that gets muddled by Rayleigh scattering. A more sensitive technique known as coherent anti-Stokes Raman scattering uses multiple laser beams to generate coherent optical signals that enhance resonant frequencies in the sample; that method, however, also produces nonresonant background noise. Recently a team led by Harvard University chemist Sunney Xie demonstrated a new technique based on stimulated Raman scattering that tunes the difference between the frequencies of two laser beams to match a desired molecule's resonant frequency, thus amplifying the Raman signal. The measurable intensities of the transmitted beams change only when a match occurs; nonresonant signals are not picked up. The images show the top view (a) and the depth profile (b) of an acne medication (blue) that penetrated a mouse's skin, thus demonstrating the potential of the new technique to monitor drug delivery. (C. W. Freudiger et al., Science 322, 1857, 2008.)
— Jermey N. A. Matthews

Catalysts are ubiquitous in today's chemical industry, but there remains much to be learned about the specific mechanisms by which many of them work. Though such knowledge could lead to improved or new catalysts, obtaining atomic-scale information about in situ chemical changes in a hot environment at atmospheric pressure has presented a difficult challenge. A Dutch team led by Frank de Groot and Bert Weckhuysen of Utrecht University has recently demonstrated the potential of a new approach to imaging catalysts at work: scanning transmission x-ray microscopy. As a catalyst and reactants interact, the valence states and chemical bonding of the participating atoms evolve. STXM detects those changes by looking at the absorption of x rays by the atoms' inner electron shells. The researchers demonstrated the technique by looking at the iron-based catalyst for the Fischer–Tropsch process, in which hydrogen and carbon monoxide are converted to hydrocarbon chains. Soft x rays used in STXM are strongly attenuated in matter, so the research team used a nanoreactor of thickness 50 μm; the reactor was connected to gas lines and mounted on an adapter that was scanned in 35-nm steps through the focus of a monochromatic x-ray beam. In that way, two-dimensional absorption maps at various x-ray energies could be recorded. The researchers paid particular attention to energies near the absorption edges of carbon, oxygen, and iron. Analyzing the maps they obtained, the researchers could extract the carbon hybridization states and determine the extent to which the iron atoms, which started off in iron oxide, had been reduced, formed other oxides, or reacted with the silicon dioxide substrate or with carbon. The figure maps the distribution of the inferred iron compounds, each represented by a different color. With better optics and detection techniques, the team hopes to improve on its current 40-nm resolution and perhaps provide time-resolved and 3D imaging of complex chemical reactions in situ. (E. de Smit et al., Nature 456, 222, 2008.) — Richard J. Fitzgerald

In recent years, scientists have learned to make light-sensitive molecular probes and incorporate them into biological tissue. When stimulated by the correct wavelength, the probes then make available for study targeted dynamic processes of living systems. In a typical experiment, static optical elements such as lenses are used to focus light on a region of tissue, where all photoactivity—wanted or unwanted—is then observed. A new, active element for optical imaging is the liquid-crystal spatial light modulator, used to tailor light's distribution in such applications as optical tweezers and adaptive optics. Valentina Emiliani of the University of Paris V and her colleagues are now using LC-SLMs to create holographic illumination to selectively activate biological probes in specific locations. One such probe is glutamate, a major neurotransmitter, caged within a photoactive molecule. Shown here are data recorded from cells within a slice of mouse brain tissue perfused with caged glutamate and illuminated in the green regions with a 6-micron-diameter spot (left) and, in the same location, with a holographic shape that coincides with just the dendrite of interest (right). Plotted below each image are glutamate-stimulated currents resulting from brief light pulses. The responses show a large reduction in unwanted signal when the LC-SLM is used. The researchers also simultaneously activated several precisely positioned spots on the same neuron, demonstrating highly controlled stimulation of different neural inputs. (C. Lutz et al., Nat. Meth., published online 10 August 2008, doi:10.1038/nmeth.1241.) — Stephen G. Benka

In scanning microscopy, images are put together by sweeping a single narrow beam back and forth over a sample. If you had a wide array of multiple beams, one scan would suffice. And if the sample moved over the array, you wouldn't need to scan at all. That idea is behind a new optofluidic imaging scheme developed for biological applications by Caltech's Changhuei Yang and his colleagues. At the heart of the scheme is an off-the-shelf sensor whose CMOS pixels are read out individually. Contrast is achieved when a sample, under constant illumination, momentarily shadows the pixels as it passes over them in a microfluidic channel. At 10 × 10 µm², the pixel size is too big to resolve the parts of amoebae and other tiny organisms. To boost the resolution, Yang masks the pixels with a commensurate array of 1-µm-diameter holes. Although 97% of the sensor is masked, 100% of a sample is imaged because the sample's path over the lines of holes is canted at a slight angle. Thanks to the angle, an organism or cell is scanned not only along but also across its whole body. The Caltech team built and demonstrated two types of imager; they differ in how they stabilize the samples' orientation during a scan. In one type, suitable for imaging tiny worms and other elongated samples, gravity pulls the samples. Confinement suffices to prevent the samples from tumbling. In the other type, suitable for squatter, more rounded samples, pressure pushes the samples. Tumbling is forestalled by a strong DC electric field, which polarizes and aligns the samples. Both imagers are barely bigger than a US quarter and, as the accompanying images show, provide resolution comparable to that of a conventional optical microscope. (X. Cui et al., Proc. Natl. Acad. Sci. USA, in press.) — Charles Day