System Bits: March 14

Neuromorphic computing
While for five decades, Moore’s law held up pretty well, today, transistors and other electronic components are so small they’re beginning to bump up against fundamental physical limits on their size, and because Moore’s law has reached its end, it’s going to take something different to meet the need for computing that is ever faster, cheaper and more efficient.

Stanford University professor of bioengineering and of electrical engineering Kwabena Boahen has a pretty good idea what that ‘something more’ is. It is brain-like — neuromorphic — computers that are vastly more efficient than the conventional digital computers we’ve grown accustomed to.

Professor Kwabena Boahen has written “A Neuromorph’s Prospectus” outlining how to build computers that directly mimic in silicon what the brain does in flesh and blood. (Source: Stanford University)

He describes in the latest issue of Computing in Science and Engineering, that this is not a vision for the future, it is now, and “we’ve gotten to the point where we need to do something different.” Boahen is also a member of Stanford Bio-X and the Stanford Neurosciences Institute.

While others have built brain-inspired computers, Boahen said, he and his collaborators have developed a five-point prospectus for how to build neuromorphic computers that directly mimic in silicon what the brain does in flesh and blood.

Time crystals
With possible application to future quantum computers, researchers at Harvard University, and separately, at the University of Maryland, among others, have deduced that time crystals — hypothetical structures that pulse without requiring any energy — are possible.

Think ticking clock that never needs winding: The pattern repeats in time in much the same way that the atoms of a crystal repeat in space.

Illumination with green light reveals a time crystal formed in a network of electron spins (red) within the defects of a diamond. [Source: Harvard University]

Further, the team of researchers led by physicists at the University of Maryland-based Joint Quantum Institute (JQI) have reported creating the world’s first time crystal using a chain of atomic ions. 

They explained that crystals such as ice or diamond are made of atoms arranged in a repeating pattern in space. These new time crystals have atoms follow a repeating pattern, but in time rather than space. The UMD-led team’s creation brings to life the exotic idea that it might be possible to create such time crystals that was proposed in 2012 by Nobel-prize winning MIT physicist Frank Wilczek.

Much like freezing destroys the symmetry of liquid water, a time crystal disturbs a regularity in time. This is somewhat surprising, says lead author and JQI/ UMD postdoctoral researcher Jiehang Zhang, since nature usually responds in sync to things that change in time but a time crystal doesn’t follow this expectation, instead responding with a slower frequency—like a bell struck once a second that rings every other second. The atomic ions in the Maryland experiment, which researchers manipulated using laser pulses, responded exactly half as fast as the sequence of pulses that drove them.

Zhang, Christopher Monroe, a UMD Distinguished University Professor of Physics and a JQI Fellow, and a group of experimentalists at UMD teamed up with a theory group at the University of California, Berkeley to create their time crystal. The Berkeley group, led by physicist Norman Yao, had previously proposed a way to create time crystals in the lab. For a chain of atomic ions, the challenge came down to finding the right sequence of laser pulses, along with assembling the sea of mirrors and lenses that ensured the lasers impinged on the ions in the right way.

Sensing sweat for health
Sweating it out on a treadmill, or racing to finish a half marathon, a runner might risk a potentially dangerous buildup of electrolytes in her blood, reminded UC Berkeley researchers who have created a network of sensors imbedded in a sweatband to monitor moment-by-moment changes in electrolytes and metabolites.

In theory a “sweat sensor” could monitor electrolyte levels in real time or track diabetes risk by measuring quick spikes in blood sugar levels. Such a device could find wide use, and make an impact in the marketplace.

Devised by Ali Javey can monitor moment-by-moment changes in electrolytes and metabolites, a potential boon to weekend athletes, diabetics and people exposed to heavy metal concentrations. [Source: UC Berkeley]

Current tests monitor these telltale signs only periodically, missing short-term fluctuations or suddenly spiking concentrations.

But in Berkeley’s Cory Hall, a lab that’s been converted into a high tech mini-fitness center, can now trace these metabolic changes second by second in a substance any good work out produces: sweat.

Ali Javey, a materials scientist and professor of electrical engineering and computer sciences, has combined innovative materials, sensor technology and integrated circuits to develop a wearable sweat sensor network that can measure rapid fluctuations in electrolytes and metabolites, and even the buildup of heavy metal concentrations in perspiration.

Prototype sweat sensors are printed on thin plastics and are embedded in headbands or wristbands to monitor concentration levels of these metabolic markers in real-time.

The lightweight sensor network tracks half a dozen chemical markers in sweat as volunteers toil on a bike in his lab. The sensors within the film are connected to a flexible electronic board with silicon Integrated Circuits. The circuit board converts the voltage and current measures of the sensors to a readout of electrolyte or metabolite concentration.

As they huff and puff, runners can monitor spikes or dips in their electrolytes, metabolites and skin temperature on a smart phone or other mobile device via Bluetooth. The readouts can also be transmitted wirelessly to other sites for more detailed analysis.

Javey is now refining the sensor fabrication process to make it more commercially practical for fitness training, athletics, health diagnostics and even large-scale population studies.