Efficiency

IBM T.J. Watson Research Center will develop a two-pronged approach to improve future datacenter efficiency.. New optical interconnect solutions can provide a path to energy-efficient datacenters at higher bandwidth levels, but they must also meet key metrics including power density, cost, latency, reliability, and signal integrity. IBM's team will use their expertise with vertical-cavity surface-emitting lasers (VCSELs) to develop VCSEL-based optical interconnect technology capable of meeting the necessary future demands. VCSEL-based interconnects offer an appealing combination of low power consumption, small size, high performance, low cost, and manufacturability. The team will work to increase the operating speed of VCSELs, detectors, and the associated circuits, while also developing packaging solutions to install optical interconnects on the integrated circuit. This integration will allow the system to eliminate the traditional driver and receiver electronics of most board-mounted optical modules, greatly reducing the cost and energy use of data transfer. The team will eventually use single-ended signaling to drive and receive signals from the modules directly, increasing the bandwidth of the system chips by at least two times and improving power efficiencies across the datacenter.

PV inverters convert DC power generated by modules into usable AC power. Ideal Power's initial 30kW 94lb PV inverter reduces the weight of comparable 30kW PV inverters by 90%--reducing the cost of materials, manufacturing, shipping, and installation. With ARPA-E support, new bi-directional silicon power switches will be developed, commercialized, and utilized in Ideal Power's next-generation PV inverter. With these components, Ideal Power will produce 100kW inverters that weight less than 100lb., reducing the weight of conventional 3,000lb. 100kW inverters by more than 95%. The new power switches will cut IPC's $/W manufacturing cost in half, as well as further reduce indirect shipping and installation costs.

iMetalx is scaling up an advanced electrochemical process to produce low-cost titanium from domestic ore. While titanium is a versatile and robust structural metal, its widespread adoption for consumer applications has been limited due to its high cost of production. iMetalx is developing an new electrochemical titanium production process that avoids the cyclical formation of undesired titanium ions, thus significantly increasing the electrical current efficiency. iMetalx will test different cell designs, reduce unwanted side reactions to increase energy efficiency, and minimize the heat loss that occurs when processing titanium. By developing a scalable and stable electrochemical cell, iMetalx could significantly reduce the costs and energy consumption associated with producing titanium.

INFINIUM is developing a technology to produce light metals such as aluminum and titanium using an electrochemical cell design that could reduce energy consumption associated with these processes by over 50%. The key component of this innovation lies within the anode assembly used to electrochemically refine these light metals from their ores. While traditional processes use costly graphite anodes that are reacted to produce CO2 during refining, INFINIUM's anode can use much cheaper fuels such as natural gas, and produce a high-purity oxygen by-product. Revenue from this by-product could significantly affect aluminum production economics. Traditional cell designs also waste a great deal of heat due to the necessity of keeping the reactor open to the air while contaminated CO2 rapidly exits the chamber. Since INFINIUM's anode keeps the oxygen or CO2 anode gas away from the main reactor chamber, the entire system may be far more effectively insulated.

Iowa State University (ISU) will develop new sensors that measure the amount of nitrogen in soils and plants multiple times per day throughout the growing season. Nitrogen fertilizer is the largest energy input to U.S. corn production. However, its use is inefficient due to a lack of low-cost, effective nitrogen sensors. Year-to-year variation in nitrogen mineralization, due to differences in soil water and temperature, creates tremendous uncertainty about the proper fertilizer input and can cause farmers to over-apply. As a result, nitrogen fertilizer is lost from croplands to the surrounding environment where it pollutes air and water resources. To address this problem, the team will develop a novel silicon microneedle in-plant nitrogen sensor and a microfluidic soil nitrogen sensor. The microscale needles can be inserted into multiple sites of the plant to provide frequent and accurate monitoring of nitrate uptake, and for the first time provide a view of plant nitrogen use as the plant and roots develop. The team will also develop an automated microfluidic sensor which will measure the amount of nitrate in soil by extracting very small amounts of solution from the soil. The microfludic technology on which soil sensors are based can be produced at low cost. The combination of these two sensors will allow for a deeper understanding of plant nitrogen use and how it correlates with nitrate levels in the soil. These new sensors will accelerate the effort to identify, select, and breed new crops with improved nitrogen use efficiency. And the project will help increase the energy efficiency of our agriculture systems while reducing input costs, greenhouse gas emissions, and nitrate pollution of aquatic ecosystems.

IR Dynamics will develop a low-cost nanomaterial technology to be incorporated into flexible window films that will improve thermal insulation and solar heat gain. The team's nanomaterial will incorporate two materials. First, low-cost nanosheets will increase thermal resistance. Second, a new type of nanomaterial will allow heat, in the form of infrared radiation (IR) from the sun, to pass through the window when it is cold outside, helping to warm the room in cold weather. When it is hot outside, the material will block the solar IR from passing through the window and warming the interior. This same material reflects thermal radiation and displays a tunable emissivity, contributing more to its insulation value and energy retention. The dynamic IR reflectivity and emissivity are passive by nature, requiring no electronics or power source to shift, and only rely on environmental temperature changes. IR Dynamics' technology creates a window film that automatically adjusts depending on outside temperatures and can have a substantive impact in performance on single-pane and older variants of double-pane windows.

Johns Hopkins University will develop and assess components of a self-powered system to convert methane (the main component in natural gas) into carbon fiber. Methane can be separated into carbon and hydrogen, or burned for energy. The team will develop processes to use methane both to power the system and serve as carbon feedstock in a four stage system. First, methane is decomposed into hydrogen and carbon, and combined into a carbon/metal aggregate. Second, the carbon/metal aggregate is melted, producing a liquid melt containing carbon dissolved within it. Third, the melt is solidified into a homogeneous ribbon. Fourth, carbon is extracted from the ribbon in the form of fiber or fiber precursor. Finally, the metal content of the ribbon is reclaimed and recycled back to the start of the process for further methane decomposition. The project will focus on resolving the materials science challenges of directing carbon crystal growth into fiber and/or fiber precursors (steps 3 and 4). The final goal is to produce fibers that have the strength and stiffness of traditionally produced carbon fiber while requiring a fraction of energy and cost to produce.

Kyma Technologies will develop a cost-effective technique to grow high-quality gallium nitride (GaN) seeds into GaN crystal boules, which are used as the starting material for a number of semiconductor devices. Currently, growing boules from GaN seeds is a slow, expensive, and inconsistent process, so it yields expensive electronic devices of varying quality. Kyma will select the highest quality GaN seeds and use a proprietary hydride vapor phase epitaxy growth process to rapidly grow the seeds into boules while preserving the seed's structural quality and improving its purity.

Lawrence Berkeley National Laboratory (LBNL) will develop an imaging-modeling toolbox to aid in the development of more efficient crops at field scales. The approach is based on a root phenotyping method called Tomographic Electrical Rhizosphere Imaging (TERI). TERI works by applying a small electrical signal to a plant, then measuring the impedance responses through the roots and correlating those responses to root and soil properties. Key target traits of the LBNL project include root mass, root surface area, rooting depth, root distribution in soil, and soil moisture content and texture. The TERI technology will be sensitive enough to distinguish between various plant varieties. The process is minimally invasive, and by doing repeated TERI measurements over the growing season, critical root architectural traits and their dynamic changes over time can be quantified for a range of soil conditions. From laboratory studies, LBNL and its partners will integrate hardware and software tools to develop a field deployable instrument based on the TERI technology. LBNL is partnered with the Noble Foundation to apply the TERI technology to wheat breeding and identify wheat varieties with improved root characteristics, and also link visible above-ground phenotypes with the desired root characteristics. The team will utilize the TERI technology to characterize plants in both controlled laboratory and field studies, and use the data generated to improve ecological models predicting plant performance in the environment.