No subset of technology permeates the electronic industry more than power. More than any other characteristic, it defines the design engineer's job. And it's on the verge of some very big changes. Breakthroughs at several scales are about to alter the power landscape.
On the small scale, galvanic battery technology has struggled to get past the limitations of lithium cells. On a larger scale, we're still refining the lead-acid battery, which is inherently handicapped by the weight of the element that gives it its name.
Moving up the power scale, conventional modes of electrical generation need help on two fronts: It's becoming harder to get approval for energy plant siting, and simply, there are better things to do with fossil reserves than burn them up to light the cave.
A number of technologies at the high-risk/high-reward end of power engineering aren't quite ready for prime time. Nonetheless, they're being pursued not only by ivory-tower researchers, but also by feet-on-the-ground engineers out to turn them into next year's everyday reality.
FUEL CELLS AT LAST?
Conceived in the mid-19th century, fuel cells have been used in space systems—where cost is seldom an object—since the 1960s. Terrestrial development has been slow, though. They're tantalizing because fuel-cell energy conversion is twice as efficient as combustion.
Typically, a terrestrial fuel cell's oxygen supply is derived from the atmosphere. The hydrogen comes from a tank or is extracted from a hydrocarbon compound such as methanol, propane, butane, natural gas, or diesel fuel. Except for direct methanol fuel cells (DMFCs), this process requires an external device called a reformer. Reformers tend to be big and expensive. They also generate pollution, but considerably less than burning the same amount of fuel in an internal combustion engine.
Several types of fuel-cell systems exist today. Alkaline systems, which use a liquid electrolyte, have been integrated within space systems for decades. Though bulky, they're cheap to manufacture and operate. They also achieve about 70% efficiency, but they require pure oxygen and hydrogen to operate.
Proton exchange membrane, also called polymer-electrolyte-membrane (PEM) cells, use an electrolyte membrane between the cell's anode and cathode. Hydrogen flows in at the anode, where a platinum catalyst causes it to split into protons and electrons. The membrane allows only the positively charged protons to pass through to the cathode.
To reach the cathode, the electrons must travel through an external circuit, the load. When they reach the cathode, electrons and protons combine with oxygen to form water, which flows out of the cell. Many electrical vehicles now use larger PEM cells.
Direct methanol fuel cells (DMFCs), a smaller variety of PEM cells that extract hydrogen directly from methanol, are the technology of choice for powering personal electronics (Fig. 2). They typically run on methanol and promise up to five times the energy density of lithium-ion batteries. Also, they presently achieve 20% efficiency and 3000 hours of useful life before the cells must be replaced.
A single PEM cell produces 0.6 to 0.8 V, so cells are stacked and connected in series to obtain usable voltages. Efficiency is about 50%, but the required methanol pumps lower system efficiency. They're far happier running continuously, rather than start/stop. Estimated stack life lasts around 40,000 hours in continuous use, but, except in some DMFCs, only 4000 in vehicular use. So far, PEM cells are expensive and fussy about hydrogen purity and how much water they contain.
Big, mains-power-generating stationary fuel-cell systems may include solid-oxide fuel cells (SOFCs) or, alternatively, designs that use either phosphoric acid or molten carbonate electrolytes. SOFCs can reach 60% efficiency, but they require a 1000°C operating temperature (new SOFCs run at about 700°C). High temperatures become an advantage for large-scale power generation because that heat can, in turn, run steam generators to improve overall efficiency.
Phosphoric-acid electrolyte fuel cells, running on natural gas, are fairly inefficient. Generally, 200-kW ac power plants are 40% efficient, and large 11-MW units manage 45%. However, they operate at fairly low temperatures of 160°C to 220°C.
Fuel cells that utilize molten-carbonate electrolyte operate around 600°C to 650°C, so gas turbines could be used to extract additional energy from waste heat. Molten-carbonate cells promise efficiencies up to 60%. They also can run on coal-based fuels with reforming accomplished inside the stack. One downside is that the molten carbonate electrolyte is corrosive.
Earlier this year, researchers at the University of Illinois at Urbana-Champaign described a fuel cell that doesn't require a membrane between fuel and oxidant. Instead, it has a Y-shaped microfluidic channel in which two liquid streams containing fuel and oxidant merge and flow between catalyst-covered electrodes without mixing.
This is possible because fluids flowing through channels of microscale dimensions maintain a laminar flow, eliminating the need for a membrane. Protons and electrons simply diffuse through the liquid-liquid interface. These University of Illinois membrane-less fuel cells are compatible with alkaline chemistry, meaning they use OH ions instead of protons.
Meanwhile, researchers at Penn State came up with a process in which bacteria extract four times as much hydrogen out of biomass than could be generated by simple fermentation. They nudge the bacteria with about 0.25 V, which enables them to convert acetic acid, an otherwise "dead end" fermentation product, into carbon dioxide and hydrogen.
The need for a flat, 2D PEM tends to limit power density in DMFCs. Neah Power Systems in Bothell, Wash., is developing a DMFC that uses a 3D porous silicon matrix for the catalyst, with much more surface area than a membrane (Fig. 3). In a recent demonstration, a stack achieved a power density of greater than 80 mW/cm2 at room temperature.
In the long term, Neah will shoot for the same portable consumer market as the Japanese. But in the short term, the company aims at a deeper-pocketed market—the military. Neah is working with the U.S. Marine Corps on a methanol-powered refillable battery pack with the same form factor used in a great deal of today's battlefield communications equipment. The company says its technology can potentially decrease the portable-power-related weight soldiers must carry—over a 72-hour mission—by up to 65%.
SOLAR POWER BOOSTED BY 50%
Among all alternative technologies, solar is the closest to economic sustainability. In the U.S., this is largely due to the issuance of standards for grid-connect inverters (see "California Sets Green Power-Supply Standards," January 20, 2005, ED Online 9480). Over the past five years, solar's compound annual growth rate has pushed past 40%; last year the rate was over 50%.
Globally, three key markets form the nexus of activity. Japan has held the lead over the last five years or so, annually installing about 50,000 3- to 4-kW residential systems. The German market began to take off more recently, driven by a government incentive program. In 2004, Germany surpassed Japan as the world's largest consumer of solar cells. The North American market grows at a rate of more than 100% a year, starting from a smaller base, but with a larger potential than either Japan or Germany.
In January, in a report on grid-connected solar power in North America, I stated that the highest solar-cell efficiencies were around 18% for multilayer/multibandgap III-V cells in space applications and around 12% to 15% for the best commercial crystalline-silicon cells (see "California Sets Green Power-Supply Standards," again). So I was surprised to see a local news story about SunPower, a subsidiary of Cypress Semiconductor, which said that the company claimed to be building commercial crystal-silicon cells with efficiencies greater than 20% (Fig. 4).
Peter Ashenbrenner, SunPower's vice president of sales and marketing, said that SunPower's gain in efficiency primarily comes from moving all of the collection grid from the front of the cell to the back. Here, it doesn't block any light and can be made thicker, reducing ohmic losses.
A white paper on SunPower's site, www.sunpowercorp.com, says several other design features contribute to high efficiency, besides that gridless front surface that permits optimization of light trapping and passivation. Localized back contacts reduce contact recombination losses. And, a back-side metallization approach supplies internal rear-surface reflection and very low series-resistance loss. But there's even more to the story.
Although SunPower uses single-layer crystalline-silicon wafers, the back-contact approach requires minority carriers to diffuse through the entire wafer thickness to reach the collecting junctions, which requires a carrier lifetime of greater than 1 ms. This was only made possible via extraordinarily high-lifetime float-zone silicon starting material only recently made available by Danish company Topsil Semiconductor Materials A/S.
A further unique factor in SunPower's process technology development was the development of low-cost screen-printing technology to replace photolithography. SunPower also adapted diffusion furnaces, wet etching, and cleaning equipment.
Tests performed on cells by the U.S. National Renewable Energy Laboratory (NREL) substantiate the claim of per-cell conversion efficiencies greater than 20%. A further real-life series of tests at NASA's Dryden Flight Research Center at Edwards Air Force Base in California deployed multicell modules in the Mojave sun. The 32-cell module, made with cells produced on SunPower's Texas pilot line, demonstrated an aperture area efficiency of 18.2% and produced 95 W from a total irradiance of 1030 W/m2 applied to an array with an area of 5192 cm2.
The cells for that module were produced on a pilot line at Cypress' plant in Richardson, Texas. Today, SunPower is up and running on its first production line in a dedicated plant in the Philippines. By 2006, there will be four lines, each producing about 25 MW of product per year, equivalent to about 8 million cells on each line.
Green nuclear power? Okay, that's controversial. But the term "peak oil" is getting its share of press attention these days. Stripped of rhetoric, the peak-oil message is that regardless of how much oil is under our feet and regardless of what we can do to make sure we don't leave any behind, we must be pretty close to the point at which production increases are no longer possible. The corollary is that everything is about to get more expensive.
That much is hard to deny. As a result, researchers continue to explore new ways of generating nuclear power, hopefully without the consequences of another Sellafield or Chernobyl catastrophe. The best representative of that today is the South African 170-MW Pebble Bed Modular Reactor (PBMR).
The South African reactor is the work of a consortium comprising British Nuclear Fuels, the South African electrical power utility Eskom, and the state-owned Industrial Development Corp. of South Africa. This helium-cooled, graphite-moderated, high-temperature reactor uses coated uranium pebbles encased in graphite in the form of 60-mm (2.36-in.) diameter spheres slightly smaller than a tennis ball (Fig. 5). Each pebble weighs 210 g and contains 9 g of uranium.
This reactor is based on smaller German pebble-bed research reactors of the mid-1980s. Problems uncovered in those designs led to modifications in the South African design. The PBMR uses a steel pressure vessel instead of concrete, and control rods are inserted in the reflector instead of the core. It also uses more efficient turbine technology. China currently operates a relatively small-capacity (10-MW) research pebble-bed reactor based on this German technology.
The reactor uses helium as its coolant and to transfer energy to a closed cycle gas turbine and generator. It's contained in a vertical steel pressure vessel, 6 m (~20 ft) in diameter and 20 m (~66 ft) high, lined with graphite bricks that reflect neutrons generated by the nuclear reaction back into the core and transfer heat away from the core. Vertical holes in the graphite lining contain control elements.
The core where the reaction occurs measures 3.7 m (~12 ft) in diameter and 9.0 m (~30 ft) high. Fully loaded, it could contain 456,000 fuel spheres. In the middle of the core sits a graphite column, so the nuclear reaction takes place in an annular chamber filled with recirculating carbon-coated uranium balls. They can be recirculated several times before being discarded as spent fuel.
Helium enters the reactor at about 500°C and 80 atmospheres. It flows down through the pebble bed, is heated to about 900°C, and is sent to a three-stage turbine/generator.
The PBMR's supporters cite several safety advantages. First, negative feedback places an upper limit on fuel temperature. It rises to a designed "idle" temperature and stays there. Second, there's no physical process that can cause an induced radiation hazard outside the site boundary (400 m). Third, the silicon-carbide coatings surrounding the uranium fuel particles in each pebble essentially constitute a pressure vessel, which serves as a barrier against the release of fission products that would contaminate the primary circuit.
Fourth, the graphite used in the fuel sphere remains stable up to 2800°C, much hotter than the normal-operation maximum temperature (1200°C) or the 1600°C maximum temperature anticipated in the event of a depressurized loss of forced cooling. Essentially, the core can't melt down. Fifth, helium is much better than steam as a working fluid. It won't change phase, and it's chemically inert.
Sixth, the PMBR's power density is lower than the power density of other nuclear reactors. Along with the good thermal conductivity of the graphite and the thermal inertia of the core, this helps ensure that the core temperature won't exceed 1600°C, even with a loss of forced cooling. Seventh, the fuel particles have been shown to operate at 1600°C without releasing fission products. Eighth, the continuous fueling scheme ensures that there's never a large amount of excess reactivity within the core.
Finally, a PBMR of this size will generate about 19 tons of spent fuel pebbles per year, of which less than one ton is depleted uranium. The spent fuel is much easier to store than fuel rods from pressurized water reactors. That's because the silicon-carbide coating around the fuel particles will keep the radioactive decay products isolated for approximately a million years. This is longer than the activity of any of the radioactive products, including plutonium.
Even so, the U.S. group Three Mile Island Alert raises several objections to pebble-bed reactors. First, a containment building should be provided as the last line of defense for containing a radiological release, leaving the reactor or reactors open to a terrorist attack. Second, the graphite covering the uranium is covered by several other layers of materials including silicon carbide. However, its graphite could burn if defects in the fuel defeat the outer coverings. At Windscale (Sellafield) and Chernobyl, radioactivity was dispersed by burning graphite.
Third, while the volume of waste is low, the amount of high-level waste by weight is higher than in older nuclear reactors. Since the pebbles are less radioactive than conventional fuel assemblies, it takes more of them. This involves additional risky transportation. Fourth, the industry acknowledges it cannot produce 100% "defect-free" fuel, making accidents inevitable. Finally, accidents already have occurred in experimental plants. Pebbles were damaged, releasing radiation.
The South African plant is in judicial limbo. The consortium planned to begin operating a PBMR in 2007 and construct up to 10 more in the future. In 2002, two Japanese firms joined the construction project. Nuclear Fuel Industries Ltd. announced plans for a factory to fabricate the PBMR fuel pebbles. And, Mitsubishi Heavy Industries Ltd. announced that it will develop the helium-powered turbine generators.
In 2003, Eskom announced that it was beginning the next phase of PBMR implementation—the development, construction, and commissioning of a demonstration unit, subject to obtaining a nuclear license and an environmental impact assessment record of decision.
In 2004, the South African government approved the PBMR demonstration project and decided to invest in it. Also in 2004, PBMR Ltd. informed the U.S. Nuclear Regulatory Commission that it would submit an application for design certification in 2007.
But in January, the Cape High Court ruled that Eskom should stop the development of pebble-bed nuclear reactors, overturning approval by the department of environmental affairs. That's where the demonstration project stands today. Stay tuned.