November 1973 Popular Electronics
Table of Contents
Wax nostalgic about and learn from the history of early electronics. See articles
published October 1954 - April 1985. All copyrights are hereby acknowledged.
As with most technologies, solar cells have come a long way in the last half century. Fabrication processes and efficiencies have improved significantly, motivated highly in the last twenty years or so by the global push to replace fossil fuels with other forms of power*. This article from a 1973 issue of Popular Electronics magazine is a snapshot of state of the art solar cells at the time. In the 1970s, there were no large scale solar cell arrays that were a critical part of an electric power grid. Ditto for wind turbines. One of the most significant uses of solar cells then was for powering satellites that operated near enough to the sun to generate useable energy (out to about Mars' orbit). Due to the relatively low output capacity, nuclear power supplies provided electricity for higher demand nearby loads and for deep space probes. A radioisotope thermoelectric generator; i.e., nuclear, powered the lunar rovers for Apollo astronauts. Yep, we left plutonium 238 on the moon.
* I reject the term "renewable" power since regardless of the source, no power source is renewable. The wind passing over a turbine blade has yielded that portion of its energy that resulted in motivating the blade, hence, not renewable. Solar rays impinging on a photovoltaic cell yields that portion of its energy that effected the conversion of light to electrical current, hence it is not renewable. Water falling over a dam yields that portion of its potential and kinetic energy that drives the generator, hence it is not renewable. "Renewable" is purely a feel-good marketing word.
Silicon Solar Cells - What makes them work; where they are used
By Joseph H. Wujek
The photoelectric effect that takes place in certain materials has long been known. However, it was not exploited in semiconductors until the mid-1950's when Bell Laboratories produced a solar cell, a device that converts the sun's energy into electrical energy. Then, as semiconductor technology expanded, improvements in solar-cell performance followed.
Although solar cells employ diverse materials, in this article we will discuss only the silicon device. Because of its high efficiency, stability, and reliability, silicon has become the most important and widely used material in solar cell technology.
Physical Principles. Before expanding in detail the characteristics of the silicon solar cell, let us review the physics involved in its operation to get a better understanding of the device's characteristics. Figure 1 shows a pn junction that is similar in principle to a simple semiconductor diode. While the events occurring in the junction are like those of an ordinary diode, several important differences in construction should be noted.
If useful application of the device as a solar cell is to be made, the surface area must be as large as possible to permit it to "see" the maximum amount of sunlight.
However, recognizing that silicon is an easily damaged brittle material, a surface area of roughly 2 cm on each side (0.8"x 0.8") was chosen. A means for drawing off the electrical energy generated by the cell was required; so, leads were attached to each side of the junction, with care taken to minimize the area used for connections on the "sun side" to avoid obscuring the sunlight from the active portion of the device. Then, to reduce the cell's resistivity, thin metallic "fingers" or grids are often used on the active surface, while the other side may be entirely metallized.
Fig. 1 - Cross section of silicon solar cell shows its similarity to simple pn junction.
Fig. 2 - Voltage-current characteristic for a silicon solar cell at 20°C with noon sun at right angles
Fig. 3 - The spectral response of a silicon solar cell in terms of peak short-circuit current for a constant light intensity (flux).
Fig. 4 - Relative spectral output of the sun viewed at surface of the earth.
The plot to the left is a 2015 measurement of solar flux as a function of wavelength at the Earth's surface (Wikipedia). I re-scaled and superimposed the original to fit over the 1973 version above. The new plot is more detailed, but the general curve is a very close match.
Fig. 5 - Plot of change in the short-circuit current as the angle of incidence is altered.
Depending upon the spectrum (wavelength characteristic) of the excitation light and the physical properties of the cell, electron-hole pairs are generated and a voltage appears across the terminals. If a resistance is placed across the terminals and the voltage and current for various loads are measured, the voltage-current (V-I) characteristic for the cell can be determined. A typical V-I plot for a commercial 2-x-2-cm silicon cell is shown in Fig. 2. (Note: the silicon solar cell has a negative temperature coefficient of 2 mV/°C; its output decreases by approximately 2 mV for each 1°C rise in temperature.)
To predict the performance of a solar cell, we must rely on spectral response measurements which state the cell's output as a function of the wavelength of the incident light for a given brightness. Energy is inversely related to wavelength (the higher the energy, the shorter the wavelength). The spectral response of the cell, as well as the spectral output information of the sun, enable us to predict the cell's performance.
In Fig. 3 is given the spectral response of a silicon solar cell. The wavelength is stated in angstroms (Å), equal to 10-10 meter, or 10-4 micron. The visible portion of the spectrum extends from 4000 Å (ultraviolet) to 7000 Å (infrared). The curve plot is given in terms of percent of peak (normalized) short-circuit current for a constant light intensity (flux). The curve's shape remains the same over a broad range of flux levels, although the short-circuit current increases with increasing flux. To find the total output of the cell, the function must be integrated, yielding the area under the curve.
The curve is statistical in nature. It is a measure of the number of electrons that are excited into the conduction band of the p-n junction. At the right, no output is obtained until the energy is increased to about 11,000 Å - which corresponds to the 1.1 eV of energy required to "trigger" the mechanism in silicon. The presence of excess electrons lowers this threshold somewhat so that some output appears at energies slightly below 11,000 Å.
As the energy is increased, more electrons are dislodged because an electron given additional energy can also collide with other electrons and transfer some of its energy, yielding additional electrons. If the energy is increased beyond a certain limit, the output decreases. This is due to the bluish-purple color of the silicon-dioxide passivation layer on the top of the cell; this layer filters the light in this energy region so that the junction does not "see" as much light. A further increase in energy would ultimately destroy or at least degrade the cell with X-rays.
The information given in Fig. 3 is useless unless something of the nature of the spectrum associated with the sun is known if the cell is to be excited. The relative spectral output of the sun, as viewed at the earth's surface, is shown in Fig. 4. The dips are due to atmospheric absorption of certain wavelengths.
The flux from the sun observed at the earth's surface depends upon the geographic location, time of year, and time of day, as well as local atmospheric and elevation conditions. For most of the U.S., the peak flux is 80-90 mW/sq cm, measured for clear-sky conditions at solar noon. For standardization, most workers in the field use the so-called air-mass zero (A.M.-0) flux, 140 mW/sq cm, based on measurements performed in high-altitude balloons and solar simulators. The peak earth-surface flux at noon for the above conditions corresponds to an air mass of 1 (A.M.-1). For times other than noon, the air mass is greater than 1, with an attendant decrease in flux. This effective increase in air mass, as well as the change in the incident angle of the sun throughout the day, cause variations in the output of a stationary terrestrial solar array.
The output current of a solar cell decreases as the angle of incidence of the sun changes from perpendicular to the cell. This roll-off is shown in Fig. 5. The "best" angle for mounting an array of solar cells will depend on geographic location, time of year, and time of day. A sun-tracking mechanism can be used to rotate the cells and at the same time vary the array's tilt to keep the sunlight perpendicular to the plane of the cells. Earth satellites use this scheme, although it is more common for them to have cells that "look" in all directions simultaneously. In Fig. 6 is shown a satellite of the latter category; so long as it is not oriented in an end-on attitude or is eclipsed behind the earth, power will be generated by the array.
For the most part, cells supplied by U.S. manufacturers are either rectangu1ar (1 X 2 cm) or square (2 X 2 cm), although other sizes and shapes are also manufactured. For example, Fig. 7 shows cells made by IRC. They have more than 95 percent of their surface area active. (Electrical contacts occupy a small fraction of the area.)
Most silicon solar cells have an efficiency of 8 to 11 percent. (At 90 mW /sq cm of sunlight, a 2- X -2-cm cell will have a 26-36-m W output.) The peak output also depends on the load.
Fig. 6 - Intelsat 4 communication satellite showing the solar cells that are mounted around circumference. By contrast, the Sky Lab satellite which is in orbit uses flat panels.
Fig. 7 - IRC silicon cells occupy 95 percent of the surface; contacts make up the rest
Since the individual solar cell has an output of about 0.5 V, cells must be interconnected in series and parallel configurations to raise the voltage and current, respectively, to the load's demands. To prevent one series string from driving current into another string, isolation diodes are used. These diodes are particularly important when sunlight is reduced or cut off from a string. If the diodes were not present, the shadowed string would appear as an ordinary diode across the unshadowed cells, reducing the voltage and perhaps destroying the cells by drawing excessive current. Diodes can also be used to protect individual cells in an array.
Types & Applications
Centralab Semiconductor manufactures a standard array that provides up to 6 watt-hours a day. The panels can be had at nominal outputs of 4.7, 7, 14, 16.5, or 28 volts at a nominal 1-watt output. The array's area is 5 1/4 X 6 3/4 in., and weight is about 2 lb. Panels such as these are used to power weather recorders, snow-depth recorders, and pipe-line monitors.
Centralab, along with Heliotek and International Rectifier Corp., furnish most of the silicon solar cells manufactured in the U.S. Although some panels are available as standard items, most of the arrays made by these companies are to customer specifications. Individual cells are also supplied for user-designed arrays.
An example of what can be done in designing an array is the FRUSA (Flexible Roled-Up Solar Array) developed by Hughes Aircraft Co. The system (Fig. 8) consists of two 16- X -55 1/2-ft. solar panels on which are mounted a total of 34,500 2- X -2- cm cells. The mounting medium is du Pont's Kapton® film, and a 0.006-in. glass cover protects the cells. Fabricated in this manner, the arrays are flexible and can be rolled up in window-shade fashion on a 10-in. diameter cylinder. The 1500-watt, 28-volt system was designed for satellite applications.
The FRUSA system represents an improvement of 300 percent/lb ratio over space power systems developed in the past. Hughes engineers are hopeful that the FR USA concept will find application in powering electrically propelled spacecraft in interplanetary exploration.
A proposal for building solar-powered stations has been advanced by E.L. Ralph of Heliotek. He points out the need for developing new sources of power to alleviate the fossil fuel shortages that are bound to occur. These stations would consist of great quantities of solar concentrators to reflect the sunlight to the solar cell over a broad range of solar angles. Ralph believes that such a system could be built at a cost of between $2000 and $10,000 per kilowatt for small power stations and as little as $1000 for large stations. The prospects of such installations is appealing since solar energy costs nothing and maintenance would be only a small effort when compared to present-day fossil-fuel stations.
Less exotic, but no less important perhaps, are other applications of solar cells. With the growing emphasis on preservation of our natural resources, the migratory habits of wildlife are of considerable interest. Schemes for implanting tiny transmitters in birds and sea otters have been discussed; they would be powered by small solar arrays attached to the subject. Information gathered in these studies is of fundamental importance in preserving the well-being of the species.
Fig. 8 - Panel of solar cells for a Hughes FRUSA (Flexible Rolled-up Solar Array).
With travel by automobile gaining in popularity in the U.S., the need to provide emergency roadside call boxes also grows. In those areas that are remote from power lines, solar cell powered call boxes could provide the means for furnishing this needed service. Even in those areas where power lines are accessible, solar power might prove economical. The development over the past several years of low-powered complementary symmetry MOS circuits enables designers to produce circuits that perform complex functions at power levels in the microwatt region. With high-density packaging made possible by large-scale integration (LSI) technology, these circuits, when combined with small batteries and solar-cell chargers, could provide new products that up to now were not feasible. Portable instruments, consumer products, and marine navigational aids are among the major product areas where this new technology is likely to appear.
To the interested amateur, solar cells are available from most of the major mail-order electronics supply houses. A variety of interesting circuits and gadgets can be built from these components, using the information contained in this article and manufacturer data sheets for the cells.
Beyond furnishing power to recharge batteries, the cells can be used as sunlight sensors to power relays or drive transistors. Used with small batteries, recharge capability is provided for activities in remote locations for backpacking campers, on small boats, and the like.
The solar energy, radiated to the earth's surface, is of the order of 1016 kW-hr/yr. This enormous source of energy is too plentiful to be overlooked. Solar cells will undoubtedly continue to provide a means for tapping this vital source.
Posted November 14, 2019