how low can you go?
There's an old saying in optical fabrication: If you can't measure it, you can't make it. In work that has made the ultrafast community sit up and take notice, researchers have not only produced attosecond (as) x-ray pulses, they've also developed a method to reliably measure them. Ferenc Krausz and collaborators from the Vienna University of Technology (TU Wien; Vienna, Austria), the Steacie Institute for Molecular Sciences (SIMS; Ottawa, Canada), and the University of Bielefeld (Bielefeld, Germany) used a high-order harmonic generation by a few-cycle driver pulse to produce single 90 eV soft-x-ray pulses of 650 ± 150 as in duration; then they measured the pulses using a visible/soft-x-ray cross-correlation scheme.
Using a laser system supplying 7 fs driver pulses of 700 µJ energy at 750 nm and 1 kHz repetition rate, the group focused ~1015 W/cm2 of optical power onto a 3-mm-long, 200-mbar neon target, generating an attosecond pulse, collinear beam; filtering removed all but the highest order (n > 55) harmonics.
The group directed the two beams to the krypton target by a monolithic laser/soft-x-ray cross correlator consisting of two concentric mirrors--an inner 3-mm diameter part with molybdenum/silicon multilayer coating and an outer 10-mm diameter part for the laser radiation. By recording the modulation of the width of the freed electron kinetic energy distribution, the group deduced the attosecond x-ray pulse duration.
The contour plot shows the modulation of the kinetic energy distribution of photoelectrons arising from a gas of krypton atoms exposed simultaneously to the two collinear beams. As the x-ray pulse is scanned through the laser pulse, the photoelectron spectrum broadens and contracts once over half the laser field oscillation cycle (~ 1.25 fs), displaying thereby the oscillating laser electric field. At the center of the laser pulse a sudden decrease in the oscillation period due to an abrupt blue shift of the laser carrier frequency is observed.
To measure the pulse duration, the group used the x-ray pulse to photoionize atomic krypton gas in the presence of the driver pulse. The x-ray photons trigger the ejection of krypton electrons with varying angular distributions of momenta modulated by the oscillating laser field. Momentum transferred parallel to the polarization of the driver pulse downshifts the kinetic energy of the electrons traveling orthogonal to the polarization and broadens their spectrum. The spectral broadening is subject to modulation as the x-ray pulse is scanned through the laser pulse. Because the laser pulse carried a strong and sudden frequency shift at its peak, from this cross correlation the group could not only determine the duration of the sub-fs x-ray pulse but also rule out the existence of substantial satellite pulses preceding or following the main x-ray pulse. Cross-correlation techniques allowed the group to evaluate the variation in kinetic energy between the absorption of the x-ray pulse and the peak of the driver pulse, thus obtaining the pulse duration.
"Because the generated sub-fs x-rays are too weak to be used as a pump and as a probe pulse simultaneously, our 'two-color' illumination scheme offers the only feasible route to attosecond time-resolved spectroscopy at present," explains Krausz. Future work will aim to optimize the setup to further compress the pulse duration of the x-ray pulses. The team's generation and characterization method offers sub-fs resolution, which effectively paves the way for experimental attosecond science.
visualization aids bio cleanup
Bioremediation offers the potential to improve both organic and metal contamination in a cost-effective manner. To perform it correctly, however, engineers need to obtain spatially correlative biogeochemical data of the microbial-mineral interface within naturally occurring geologic matrices. Because minerals are very heterogeneous, it is crucial to know which contaminants are associated with different mineral phases and what microbes are attracted to specific minerals on a micrometer scale.
In this image of basalt mineral, created with a 48 x 48 array of 4 µm spots, red is assigned for olivine, blue for plagioclase, green for augite, and yellow for ilmenite.
Using a 355-nm pulsed neodymium-doped yttrium aluminum garnet (Nd:YAG) laser system with a 10 Hz repetition rate, researchers at the Department of Energy's Idaho National Engineering and Environmental Laboratory (INEEL; Idaho Falls, ID) developed an imaging internal laser desorption Fourier transform mass spectrometer (I2LD-FTMS) to chemically map mineral crystals and bacteria growth on basalt with micrometer precision.
The group's design manipulates the laser beam instead of using a translation stage to shift the sample, which allows the user to scan the sample layer by layer while still correlating the information spatially. The beam passes into an optical assembly consisting of cradled prisms connected to leadscrews and stepper motors that allow for horizontal and vertical motion, which creates a virtual source. "The key is that the motion of the laser beam as it exits the virtual source is indexed to the center of the focusing element such that the beam always passes through the center of the lens," explains Jill Scott of the research group.
By scanning the 19-mm disk area of INEEL basalt with a 2-µm sample spot size before and after it was inoculated with bacteria found in neighboring soil, the team is able to evaluate the surface of the mineral (see figure). The tool was developed to determine and test a bioremediation scheme for uranium contaminants in the vadose zone underneath the INEEL, but it can also be used for various imaging studies. --Phillip Espinasse
a new tune for lasers
Long in the research phase, quantum-dot (QD) lasers are beginning to approach commercialization. New Mexico start-up Zia Laser (Albuquerque, NM) is hoping to take advantage of developments in the field to provide widely tunable sources for telecommunications starting later this year.
Zia's device is a distributed feedback laser. Molecular beam epitaxy is used to grow the QDs through the Stranski-Krastanov method, a self-assembly process. Initial layers are grown lattice-matched to the substrate, which is gallium arsenide (GaAs) for the 1300-nm emitters and indium phosphide (InP) for devices in the 1550 window. Engineers then deposit monolayers of indium arsenide (InAs), with a much larger lattice, which leads to surface tension. Tension created by the lattice mismatch forces the InAs to form quantum dots, which measure 20- to 40-nm wide and 7- to 10-nm tall. Only the quantum dots, which make up 10% to 20% of the quantum-well (QW) plane, are active.
Unlike QW lasers, which have a continual energy spectrum, QD structures have an energy gap between the lowest state that lases and the next state (see oemagazine, January 2002, page 18). It takes significant thermal energy to get to the next level, says Victor Klimov of Los Alamos National Laboratory (Los Alamos, NM), which is exploring colloidal chemical synthesis to produce the dots. Because of this, QD lasers emit a narrower spectral band, which translates into a higher differential gain.
So-called quantum dots are actually more like pyramids 70 to 90 Å high and 200 Å along the base.
"The smaller the dot size, the wider the energy gap and the larger the energy of the quanta, which is emitted by the dot," Klimov says. "It's a very simple way to tune." In reality, the devices come out with dots of different sizes in the shape of pyramids or discs (see figures). Zia would use its chips in an external cavity laser and would tune the output by tilting a grating to provide the desired wavelength.
"What you have now is single-band tunable laser products, and what we want to do is cover all wavelengths at once," says chief technology officer Luke Lester. The company's laser can tune from 1480 nm to 1620 nm with output of 10 mW. Zia is also developing lasers for the 1300-nm window for use in metropolitan area networks, but they would not be tunable, because metro networks are not currently using much dense wavelength-division multiplexing.
The company says its laser has a threshold current of 10 A/cm2. The gain spectrum of a QD laser is symmetrical, producing a device with less chirp and feedback sensitivity than a QW laser. "A quantum-dot laser looks more like having a gas laser in a semiconductor," Lester says. The characteristics make the design simpler, removing the need for such add-ons as external modulators, electrical isolators, and thermoelectric coolers.
He believes that in three to five years, quantum-dot lasers could widely replace quantum-well designs, just as those lasers replaced earlier double heterostructure designs. "The goal is to have a device that's tunable in two bands or more at $1,000," compared with $5,000 for current devices, Lester says. Right now, though, Zia is still in the process of proving that its laser can last 5000 hours at 85°C.
Other groups in Japan, Russia, and Germany are also working on applying quantum dots to telecom applications. Fujitsu (Tokyo, Japan), for instance, is developing 40-Gb/s optical amplifiers using the technology. Klimov says it seems likely QD lasers could replace QW devices, providing that they overcome some hurdles. QD lasers have lower output right now than QW lasers, and stacking more QD structures on top of one another to increase power leads to problems.
"There are several steps in technology that need to be done to show they are better than quantum-well lasers," Klimov says, although he concedes that from a physics standpoint they clearly are superior. "They may eventually take over. The question is: when?" --Neil Savage
solar cells shine in Europe
Photovoltaic technology is making headlines in Europe as organizations explore new technologies and form joint ventures. The Fraunhofer Institute of Solar Energy Systems (ISE; Freiburg, Germany), for example, has developed monolithic tandem solar cells based on gallium indium phosphide (GaInP) and gallium indium arsenide (GaInAs) that have set new world records for efficiency. The driving force behind this work is researcher Andreas Bett. "For the first time in Europe, ISE has developed novel solar cells that yielded 28% efficiency for terrestrial and 24.5% for space applications," he says. "This is a single process, so it is less costly compared with two separately made cells mechanically stacked. The different materials provide different absorption characteristics that yield a more efficient sunlight conversion. We have added a Fresnel lens to concentrate the sunlight. With such a device, ISE has recently achieved a record efficiency of 32%."
Meanwhile, I.M. Dharmadasa's group at Sheffield Hallam University (Sheffield, UK) is also focused on improving efficiencies of solar cells and reducing costs (see figure). "These objectives can be achieved by the reduction of expensive semiconductors and the use of lower-cost process techniques," says Dharmadasa.
Sheffield's cells are only 2-µm thick compared with the 200 µm used for crystalline silicon cells. An electrochemical fabrication method also lowers cost. "In eight years of research, we have shown that these electro-deposited materials produced under the right conditions are as good, if not superior, to those made by the expensive growth methods," Dharmadasa says.
The II-VIs, such as cadmium sulfide (CdS) and zinc selenide (ZnSe), are being studied as window materials, and exotics, such as cadmium telluride (CdTe) or copper indium selenide (CuInSe2) for example, are being studied as absorbers. "We are most excited by our latest material, copper indium gallium selenide (CuInGaSe2)," Dharmadasa continues. "Our first structure is based on glass/FTO/n-ZnSe/p-CuInGaSe2/Au. It has shown 15% efficiency. To our knowledge, this is the best so far for one-step electro-deposited semiconductors." Moreover, it is not far from the 18.8% measurement for cells made by costly vacuum growth by the U.S. National Renewable Energy Laboratory (NREL; Golden, CO). "We are presently directing our research to see if we can match such efficiencies with our low-cost electro-deposited materials," he notes.
heading to market
On the commercial front, energy companies Electrabel (Brussels, Belgium) and TotalFinaElf (Courbevoie, France) have joined forces with independent electronics research center IMEC (Leuven, Belgium) and IMEC spinoff Soltech (Heverlee, Belgium) to form Photovoltech. The new company will produce the photovoltaic cells and modules that form the basic components for photovoltaic systems. Photovoltech will begin construction of Belgium's first solar cell and modules factory in Tienen. The company expects to leverage IMEC production processes to manufacture higher-efficiency polycrystalline silicon solar cells at lower cost than by today's technology.
I. M. Dharmadasa of Sheffield Hallam University holds a solar panel in the lab. He is standing with teammates Anura Samantilleke, Nandu Chaure, and Tian Fang.
Sheffield Hallam University
According to Reed Electronics Research (RER; Surrey, UK), the market for semiconductor solar cells in Europe will grow from $154 million in 2000 to reach $228 million by 2005. "This, like other opto device markets, suffered one of the first reversals of fortune in its history in 2001," says Andrew Fletcher, RER publisher. "The reasons for this were complex but had much to do with a lack of confidence in the telecom sector for which the higher-value solar-cell market segment has so much to contribute. There has also been a general decline in manufacturing worldwide, which shaved a few more points off the growth of opto and other devices in this period. Continuing downward unit price pressure is also affecting the market."
In terms of wafer area, solar cells are still one of the biggest consumers of wafers, the bulk of them germanium (Ge). Depending on the particular solar cell design, n- or p-type wafers are used. Single and dual GaAs-on-Ge cells use n-type Ge, the triple-junction and newer four-junction cells use p-type. Today's efficiencies average 25% for production cells with improved cells of up to 27% with four-junction cells capable of exceeding 30%.
"Apart from the material quality defined through its conductivity type, dislocation density and wafer finish are the most important customer criteria," says Ignace de Ruijter, business line manager for substrates at Umicore (previously UM Electro-Optic Materials; Olen, Belgium).
Still, interest in solar cells in Europe has never been stronger--not only in silicon and germanium, which provide the foundation for the bulk of the world's PV market, but also for a range of advanced semiconductors such as chalcogenides and multielement III-Vs. The field is evidently becoming ever more interesting via continual efficiency improvements and efforts to reduce costs. Companies are using a broader range of semiconductors and working with more efficient fabrication processes to produce high-efficiency devices with less material.
DNA analysis techniques are increasingly moving to center stage in bioresearch laboratories, which offers great benefits for drug discovery and disease profiling applications. To be truly useful for drug discovery and customization, however, the analysis process must be faster. The Nippon Laser & Electronics Lab (NLE; Nagoya, Japan) recently developed a quicker way of reading DNA chips that is used for rapid analysis of biological samples. By illuminating all the genes on the chip simultaneously, researchers can extract data from DNA chips at speeds five times faster than conventional laser-based methods.
DNA chips are highly miniaturized devices containing an orderly, matrix-like arrangement of synthesized oligonucleotides or pure DNA fragments. To detect genes, DNA chips are first prepared by affixing error-free fragments of DNA that accurately reflect the exact genes of interest to the chip surface. Researchers mix the biological material to be tested with fluorescent labels and place it on the chip. The genes in those samples attach themselves to the matching DNA fragments on the chip. When the chip is illuminated with laser light, the labels on the genes attached to the DNA fragments fluoresce. Identification of the fluorescence patterns enables researchers to establish the structures of many different genes.
In the laser illumination method, each DNA fragment is illuminated, and the fluorescent labels are read in turn. Serial readout can be a time-consuming process; this prompted NLE to develop its new readout method. According to company president Katsumi Yoneda, the company feels that the faster reading method should help shorten development times for new drugs, among other applications.
Instead of lasers, NLE researchers use white light produced by a metal halide lamp. The DNA chip is illuminated edge-on so that the light travels through the glass as though it were a waveguide. When this happens, the surface of the glass evanescent waves at 200 to 300 nm. This weak light triggers fluorescence in labels on the genes attached to the DNA fragments on the chip; a CCD camera captures this signal for analysis.
With this method, it takes about two minutes to read a standard 26 x 76 millimeter chip, which is more than five times faster than the laser systems most commonly used in research laboratories today, says NLE spokesman Kureho Takeuchi. In addition, because the method does not use a laser, it can be set up in much less space, and maintenance is simpler. The NLE method offers 5-µm resolution, which is comparable to that of the laser systems. Further, NLE says its method can divide the DNA fragments into six areas by changing the white light source slightly and adding various filters to the CCD detector camera.
NLE plans to market its method with appropriate hardware and software to universities and research institutes. The company aims to sell 50 units per year at a forecast price of about $110,000 (¥14.5 million).
According to Compugen Ltd. (Princeton, NJ), a computational genomics company, the market for DNA chips alone is expected to jump from $40 million in 1998 to more than $400 million by 2003, which by extrapolation indicated a burgeoning market for DNA-chip reading devices.
On the other hand, not everyone is enamored with higher DNA-chip reading speeds. "Chips are already very fast," says Fulton T. Crews, professor of pharmacology at the University of North Carolina (Chapel Hill, NC). "The issue is how to understand the masses of information the chips produce." --Charles Whipple