Research Horizons

Evaluating Lithium Batteries

Lithium-ion batteries tend to be smaller and lighter than the nickel-hydrogen batteries commonly used in satellite power systems. They also offer a significantly lower rate of self-discharge—a phenomenon that affects all batteries, causing them to lose power over time. For these and other reasons, lithium-ion is expected to be the next dominant type of space vehicle battery. Still, not much is known about the long-term performance of these batteries. Real-time life tests would require 10 years or more, and no validated accelerated test methods have yet been designed.

To learn more about the performance of these batteries, a group of Aerospace researchers led by Albert Zimmerman, Distinguished Scientist in the Electronics Technology Center, has developed a variety of new diagnostic techniques. The group has also mapped out a meta-analysis approach that can digest industry-wide life-test data (from ground tests) for statistical analysis. The goal is to develop and validate a life model based on wear rates, degradation processes, and cell statistics.

"Lithium ion cells degrade steadily during life by a combination of capacity loss and resistance growth," Zimmerman said. "If we can predict the amount of capacity loss and the amount of resistance growth during cycling, we can determine cell lifetime."

Some current research is directed at assessing the stress factors that affect degradation. For example, Zimmerman's group devised a new method for measuring self-discharge. The method involves measuring the amount of charge required to periodically restore cell voltage to a fixed level, along with the rate of voltage decay between periodic voltage restorations. Using this technique, researchers determined that lithium-ion cells, like most types of cells, will develop capacity imbalance over time. This is a significant concern for power system designers. "If one cell in a battery becomes much lower in state of charge than the others, it will fail prematurely during discharge," Zimmerman explained. "If one cell becomes higher in state of charge than the others, it can be overcharged during recharge, and potentially ignite or explode." No mechanisms exist to prevent the imbalance, except for cell-balancing electronics, which add cost and complexity to the power system.

Zimmerman's research further suggested that external losses through insulation resistance could contribute to cell imbalance. "If one cell in a series has inadequate insulation resistance," he explained, "that cell connection point will provide a current leakage path that will bleed capacity from some of the cells in the battery, thus causing them to become imbalanced."

Other diagnostic methods based on impedance, residual capacity, entropy, and thermal measurements were developed to characterize the electrical and thermal performance of the electrodes. "A parasitic lithium-metal plating process can be significant in some electrodes at high charge rates or low temperatures and cause them to degrade faster than normally expected," Zimmerman said. Some materials are better, he noted, primarily because they offer improved lithium ion transport rates.

Complete validation of a lithium-battery life model would require that both real-time and accelerated tests have been completed to cell failure, but few such cases are available. Nonetheless, Zimmerman's group has developed a capacity-loss and resistance-growth model based on observed life-test behavior and cell-failure statistics. Validation efforts have shown that the model correctly predicts the observed cycle life within 10 percent.

The ultimate service life that can be expected from lithium-ion batteries is so far unknown, but Zimmerman predicts it will be about half of that possible from the best nickel-hydrogen batteries (which can last more than 20 years in some applications). That "may be adequate for many future space missions," he said.

Thermal Properties of Ball Bearings

In terrestrial applications, ball bearings and their attached rotational components are often cooled by convection, either by the atmosphere or by a flood of lubricant. But neither method of cooling exists in space. Bearings, for instance, cannot be flooded with lubricant because of the potential for spacecraft contamination, and the lack of air eliminates convection altogether. Thus, the dominant mode of cooling is conductance through the bearing itself. As such, temperature predictions for rotating mechanisms in space require knowledge of a bearing's thermal conductance; however, such information is generally not known, and little has been published on the subject.

Yoshimi Takeuchi at the testing apparatus devised to measure bearing thermal conductance and study the influence of variables such as speed, lubricants, axial load, bearing size, and temperature.

To address these concerns, an interdisciplinary group of researchers led by Yoshimi Takeuchi of the Mechanical Systems Department devised experiments to assess the thermal conductance of bearings in vacuum. The studies were designed to allow control of parameters such as axial load, thermal environment, and speed. The investigation identified variables of importance for bearings in dry, lubricated, static, and dynamic states.

A static bearing is one that remains nearly stationary. A pointing mechanism, for example, might require a finely tuned bearing that moves in extremely small increments, and may remain motionless for some periods of time. "Although the bearing is not generating heat in these applications, knowledge of bearing thermal conductance is still important because heat is being transferred between the housing and the sensor bed through the bearings," Takeuchi explained. At the other extreme are dynamic bearings—for example, the bearings in momentum wheels, which typically run between 6000 and 9000 rpm. "This may not seem like much compared with terrestrial applications," Takeuchi said, "but keep in mind that on Earth, the atmosphere cools the bearings. In space, there is no convection cooling, so heat generated by the bearings creates an upper limit to component speeds."

Takeuchi's team developed a testing apparatus in which the outer race of a single ball bearing remains stationary while the inner race rotates at speeds ranging from 0 to 20,000 rpm. The test bearing supports the balance of its outer fixture and a dead weight, creating a constant axial load. A heat lamp and a cooling channel provide temperature control, and pyrometers and thermocouples take measurements for calculating thermal conductance.

The tests yielded some useful information, Takeuchi said, and showed how operational conditions affect thermal conductance differently depending on whether a bearing is static, dynamic, lubricated, or dry.

A diagram of the test rig for measuring the thermal conductance of a bearing.

For example, the thermal conductance of a static dry bearing appeared insensitive to temperature, but increased to the 1/3 power of axial load. For a static and oil-lubricated bearing, thermal conductance did not change with axial load, but did change with temperature. For a dynamic and oil-lubricated bearing, thermal conductance increased linearly with axial load and linearly with temperature; the degree of temperature sensitivity depended on the axial load.

"Of all variables, lubrication and lubricant quantity could potentially dominate the thermal conductance properties of a bearing," Takeuchi said. "The presence of oil, for instance, could increase bearing conductance by an order of magnitude."

The research could give engineers a new tool in designing mechanisms. Typically, engineers use heritage information from similar bearings when designing a rotational component. Where no heritage information is available, they sometimes rely on a closed-form solution known as the Yovanovich model; however, this model is only applicable to a dry static bearing. "This may be a good approximation for some applications, such as a bearing for a pointing mechanism with solid lubrication," Takeuchi explained. "But our experiments show that with oil or grease lubrication or significant motion, these assumptions no longer hold, and predicted conductance values could be drastically different. Our research gives a thermal analyst an idea of what bearing thermal conductance values would more likely be."

Contamination Detection and Resistance

Hardware cleanliness is a major issue for spacecraft. For example, polymeric materials, such as conformal coatings, thermal blankets, or the epoxy used in composite structures, contain molecular species that can outgas and collect on sensitive spacecraft hardware, such as optics or solar cells. Identifying and controlling such contaminants can help ensure that space instruments will meet their required service life. In general, however, the space industry suffers from a lack of advanced laboratory tools to detect and examine contamination and techniques to address contamination on the ground or on orbit.

A look inside the contamination chamber, where surfaced-enhanced Raman scattering is used to characterize the molecular vibrations of contaminant films.

In response to this problem, Aerospace has been conducting research geared toward detecting ultrathin contaminant films and developing space materials that resist contamination. "The goal is to replace standard spacecraft materials with smarter, multifunction materials that not only serve the originally intended purpose, but also impart the ability to reduce contamination exposure and thus protect hardware," explained Randy Villahermosa, lab manager for the Contamination Control Section of the Materials Processing and Evaluation department. Villahermosa's group used surfaced-enhanced Raman scattering (SERS) to characterize the molecular vibrations of ultrathin contaminant films. In Raman scattering, a laser pulse directed at a sample is deflected at a different wavelength based on the vibrational frequency of the sample's constituent elements. This data can be used to identify and characterize contaminants, both in the lab and out in the field. "In essence, a surface that is microrough, with feature sizes on the order of nanometers, will act like an amplifier of the Raman-scattered light," said Villahermosa. "With SERS, we can boost a standard Raman signal by a factor of 1,000,000 or more."

Recent work has involved the detection and characterization of submonolayer films on SERS-active surfaces. The analysis so far has also yielded some curious insights. "For the most part, the vibrational characteristics of the contaminants looked the same whether they were in bulk solution or cast as an ultrathin film," Villahermosa explained. "In essence, the contaminant doesn't really care if the surface is there or not—which is good from the standpoint of understanding how to treat surface effects in our contamination modeling analyses." Still, this finding was somewhat unexpected. "Surface-bound contaminants have been shown in numerous studies to act differently than their bulk counterparts, which makes this result interesting and something worth exploring further," he said.

Raman scattering has been used successfully in other industries, such as semiconductor manufacturing, but the use of SERS to address spacecraft anomalies is rare, if not unique to Aerospace. But based on recent advances, Villahermosa expects to see the technique used much more widely. "Just recently, we had very good success analyzing samples containing hard-particle contamination," he said. "These particles are believed to play a role in reducing the life of certain spacecraft mechanisms." Working with the contractors, Aerospace analyzed the samples via scanning electron microscope and Fourier-transform infrared—the usual techniques for samples of this type. "But it was Raman that gave us the definitive identification," Villahermosa said.

Outside view of the contamination chamber. The use of surfaced-enhanced Raman scattering to address spacecraft anomalies is rare, but highly useful.

Aerospace has been synthesizing nanofibers of polyaniline, a conducting polymer, using a variety of techniques that are capable of forming fiber diameters from 100 to 500 nm (about 100 times thinner than a human hair). "Bruce Weiller and his team have developed a new class of highly sensitive chemical sensors using these nanofibers," Villahermosa said. "Weiller's work spawned other research efforts, including one led by Alan Hopkins, who is developing new spacecraft materials that take advantage of the electrical conductivity properties of the nanofibers." In the case of contamination detection, researchers are trying to exploit the chemical mechanisms that give rise to conductivity in the nanofibers so as to make them respond to otherwise inert analytes. "Many outgassed contaminants are fairly benign from a chemistry standpoint, so we need a new way to detect and measure their presence," Villahermosa said. One important goal is to create a sensor that can not only detect certain contaminants, but filter and identify them based on chemical class or structure.

Looking at the bigger picture, Villahermosa one day expects to incorporate contamination sensors directly into spacecraft materials and structures. Unlike other on-orbit contamination-control technologies, which typically involve separate hardware, this approach would have little affect mass or installation. Other projects are also in the works to mitigate contamination on orbit, including materials that will absorb contamination before it ever reaches a detector or other sensitive surface.

"I'd like to think that someday, we'll launch spacecraft with multifunction materials that will provide thermal, radiation, and contamination resistance and protection all in one package," Villahermosa said. "Moreover, the materials will be smart because they will sense when the space environment is becoming dangerous and respond accordingly."

Characterization of Defects in Advanced Solar Cells

Modern solar cells are far more complicated than their early counterparts, containing many more materials, interconnects, and metallization layers. Defects introduced in the chemical-vapor deposition process are believed to contribute to failure mechanisms in multijunction photovoltaic cells, but no study has attempted to correlate defect centers to cell degradation or to develop a set of failure modes for calculating mean time between failure.

To address this need, Aerospace has developed a set of nondestructive, noncontact techniques to inspect multilayered photovoltaic semiconductors for crystalline defects. The first technique, optical-beam-induced current (OBIC), can be used to identify millimeter-sized areas of high defect concentration in solar cells. The second technique, microwave-detected photo-induced-current transient spectroscopy (MD-PICTS), can then be used to map the concentration of defects in those areas to 1-micron resolution as well as determine their cross section and activation energy.

As explained by Brad Reed, engineering specialist in the Electrical and Electronic Systems department, the OBIC system counts the number of 5- to 50-micron features and correlates them to a semiconductor defect frequency per unit area to define the scope of the problem. The system performs a laser scan of a region of interest, mapping both electrically defective features and electrically functional features. Researchers have theorized that some of these electrically functional features, which appear to operate nominally at the beginning of life, may change to become localized shunt sites, or electrical defects, as the cell ages.

John Nocerino with the OBIC system, developed to help understand the nature of latent defects in solar cells.

MD-PICTS is part of a family of spectroscopic techniques that use transient measurements at various temperatures to determine the relative concentration and activation energy of deep-level defects in semiconductor materials, explained Maribeth Mason of the Microelectronics Technology department. While most of these techniques use capacitance transients to find this information, MD-PICTS uses photoconductivity transients. "Because carriers can be locally excited with a laser beam, this allows measurement of the spatial distribution of defects as well," Mason said. "The photocurrents can be detected from changes in the quality of a microwave resonant cavity near the sample, making MD-PICTS a contactless and nondestructive defect characterization method."

The greatest challenge, said Mason, has been to design a microwave bridge sensitive enough to detect the small change in photoconductivity induced by laser illumination of the solar cell. The 20-gigahertz bridge is still being optimized, although the MD-PICTS equipment has not yet been fully assembled. "We are in the process of constructing an improved prototype system" in conjunction with specialized software for modeling microwave transmission through small apertures, Mason said.

The researchers have been using the OBIC and MD-PICTS techniques to test solar cells from major manufacturers. Analysis of the data will indicate failure modes and frequencies and support development of the first statistical model to predict solar-array failure. The data will also help refine Aerospace models of solar-cell semiconductors and arrays, particularly in regard to how temperature, applied electrical bias, optical illumination, and defect density affect solar-cell response.

To Fall 2005 Table of Contents