MOST ULTRAVIOLET RADIATION THAT bombards the earth from the sun never reaches the planet's surface because of a thick, vacuous blanket of pale-blue toxic gas known as ozone. Were it not for this chemical shield, ultraviolet rays would strike with such intensity that most organisms would perish. As human activities and such natural events as volcanic eruptions alter the composition of the atmosphere, it would be wise to monitor the ozone layer and ultraviolet radiation to observe possible changes.

Thanks to many ground stations and several satellites, workers have learned much about the density and distribution of ozone in the atmosphere. Yet investigators lack a comparable network to observe the ultraviolet radiation that seeps through the ozone layer. Aside from instruments operated by the Smithsonian Institution and several other organizations, the only network of ultraviolet monitoring stations in the U.S. comprises fewer than two dozen Robertson-Berger meters.

These devices are designed to detect the wavelengths of ultraviolet radiation that cause erythema, or reddening of the skin, and eventual sunburn. Erythema develops most rapidly when the skin is exposed to ultraviolet radiation with a wavelength near 300 nanometers. This wavelength falls within the ultraviolet-B spectrum, which extends from 280 to 320 nanometers.

Since 1974 the average flux of ultraviolet B has been measured with a network of eight Robertson-Berger meters From 1974 to 1985 the average flux fell some .7 percent per year. Because stratospheric ozone over the network decreased about .3 percent per year from 1978 to 1985, an increase in ultraviolet B would have been expected.

The decrease in ultraviolet-B flux was probably related to the fact that all the Robertson-Berger meters were located in urban areas. An independent investigation had shown that meters in some rural areas received from 5 to 7 percent more ultraviolet-B flux than urban meters. Another study has revealed that since 1981 ultraviolet-B flux has increased in the remote regions of the Swiss Alps. Were the results of the 1974-1985 study skewed by urban air pollution, which often contains gases and particles that absorb or scatter ultraviolet rays?

Here is where the amateur scientist can make a valuable contribution. With a little effort, you can construct an ultraviolet-B radiometer to record daily the flux of radiation. Comparing your observations with those of others from different regions would provide important information about how air pollution affects ultraviolet B.

Before building a radiometer for this purpose, one needs to understand how ultraviolet-B radiation travels through the atmosphere. Some ultraviolet-B rays scatter off air molecules; the remainder penetrate directly through the atmosphere. The sum of scattered and direct ultraviolet radiation is called global radiation.

Global radiation is of high interest in studies of the deleterious effects of ultraviolet on both living systems and materials such as paints and plastics. Global measurements are also helpful in determining how clouds affect ultraviolet B. (Robertson-Berger meters measure global ultraviolet B.)

Measurements of direct ultraviolet radiation yield valuable information about the presence and effect of absorbing and scattering agents in the earth's atmosphere. Because of the unpredictable nature of clouds and the presence of such barriers as buildings and trees, measurements of direct radiation are preferred to global ones for comparing the effects of air pollution on the relative magnitude of ultraviolet B at two or more locations.

So should the amateur observer monitor global or direct ultraviolet B? I recently discussed this question with John E. Frederick of the department of geophysical sciences at the University of Chicago. Frederick has devised a computer model that predicts the levels of ultraviolet B at the earth's surface for a range of conditions. He suggests that the amateur first concentrate on measuring direct ultraviolet B, since it is less affected by the many variables that can impair global measurements. The instruments described below, therefore, are both designed to detect direct ultraviolet B.

An ultraviolet-B radiometer requires a detector and a means for selecting the wavelength to be detected. The detector signal is amplified and transmitted to a digital voltmeter, an analog chart recorder or a computerized data-acquisition system.

Wavelength can be selected either with a monochromator or an optical interference filter. A monochromator provides a convenient but expensive way to measure ultraviolet B across a wide range of discrete wavelengths. An optical interference filter offers a much cheaper and more compact method for selecting a reasonably narrow band of ultraviolet-B wavelengths. Interference filters also allow considerably more radiation to reach the detector.

Ultraviolet-B interference filters, however, transmit a slightly wider band of wavelengths than do monochromators Moreover, interference filters transmit low but detectable levels of radiation outside their specified bands, which can cause significant errors in measurement. An ultraviolet detector that eliminates a filter's secondary bands is said to be solar-blind.

Some of the various radiometers I have designed and assembled are solar-blind. One of them is relatively easy to build because its detector incorporates both an interference filter and amplifier. The detector is a DFA-3000, made by EG&G Judson (221 Commerce Drive, Montgomeryville, PA 18936). A single detector costs $125. A calibrated detector, which costs another $75, will enable you to make absolute measurements of ultraviolet B. But even with an uncalibrated detector you can monitor relative trends in ultraviolet B.

Although the DFA-3000 detector greatly simplifies the assembly of a radiometer, its ultraviolet filter transmits a low-amplitude band of red light, which is much nearer the detector's peak spectral response than the ultraviolet wavelengths it is intended to detect. When the sun is high in the sky, the red response is perhaps 10 percent of the detector's signal. (Later I will describe a simple method for correcting this effect.)

The only electronic components necessary to transform a DFA-3000 detector into a 300-nanometer radiometer, are a resistor, a potentiometer (variable resistor) and two nine-volt batteries [see illustration at left below]. The DFA3000 contains a silicon photodiode and an operational amplifier. The diode generates a small electric current when radiation strikes its active surface. The amplifier transforms the current into a voltage, which equals the product of the current and the resistor labeled R1 in the diagram.

Because the current produced by the photodiode can be less than 10 millionths of an ampere, R1 must provide a resistance of at least 10 million ohms (or 10 megohms) to provide an output of a volt or so. I have found that 30 -megohms (three 10-megohmiresistors connected in series) provides a satisfactory resistance for R1, but some readers may want to increase it because the ultraviolet-B flux is weaker north of my latitude (29 degrees 35 minutes north). EG&G Judson suggests that up to 200 megohms can be used. At very high resistances, however, considerable care must be exercised to preclude false signals caused by leakage currents between the detector's inverting input pin (9) and ground (1). A thin film of dust, moisture or oil may provide a path for an error-generating current. Eltec Instruments, Inc. (P.O. Box 9610, Daytona Beach, FL 32020), manufactures miniature resistors of from 10 to 100 megohms.

I installed the radiometer in a pocket-size plastic enclosure purchased from Radio Shack (part number 270291). If you are inexperienced at assembling electronic circuits, you will find it easier to use a larger enclosure. Keep in mind that the enclosure should be as light-tight as possible. Light leaking through the enclosure may cause a false signal, because the base of the detector may not be totally opaque.

Because I installed the detector in a small enclosure, it was necessary to incorporate a potentiometer that can be adjusted with a miniature screwdriver. I connected the resistors to the detector with wrapping wire, a thin wire wound around component pins or leads by a special tool. If you prefer, you can solder standard wires to the pins of the components. It is important to keep the connections between the input of the amplifier (pin 9), the photodiode (pin 10) and R1 short and direct. To provide electrical contact with the nine-volt batteries, use a pair of battery clips. Inspect the wiring carefully before connecting the batteries to the circuit.

The unit I assembled includes a pair of output leads equipped with pin K, jacks that receive the probes from a miniature digital Multimeter. You may prefer to install both the radiometer and a digital voltmeter module inside a larger enclosure.

Because the radiometer is designed to measure the direct radiation from the sun, a collimator is required to restrict the detector's field of view. Thin-walled brass tubing from a hobby shop works well. A tube with an outside diameter of one centimeter should slip over the detector. Wrap a layer of tape around the detector if it fits too loosely in the tube. Coat the inside of the tube K, with flat, black enamel. A tube around 90 millimeters long will provide a field of view of approximately four degrees when the tube is pushed all the way to the detector's base.

Before installing the collimator, you should clean the surface of the detector's filter since dust and oil absorb ultraviolet B. Remove fingerprints by swabbing the surface of the filter with ethyl alcohol and wiping away the residue with lens cleaning paper. Blow away dust with clean, compressed air.

The assembled radiometer is simple to operate. First, look down the collimator tube. If you see a reflection of the pupil of your eye, the detector is perfectly centered. If not, realign the tube. After the voltmeter is connected and the radiometer's power switch is toggled on, block the opening of the collimator tube and adjust the potentiometer under the output voltage is zero. (Repeat this procedure before each measurement session.) Then point the tube toward the sun and align the tube until its shadow disappears. The detector will now be aimed directly at the sun. Record the voltage and make another measurement. You will soon discover that even on a clear day the signal level fluctuates, sometimes considerably-especially around noontime and whenever the atmosphere is obscured by clouds, smoke or dust.

Your readings will include an error factor because the detector responds to the red light that leaks through its filter. You can eliminate the error simply by following each reading with a second one during which you block the ultraviolet rays by placing a filter over the entrance of the collimator tube. An ultraviolet filter intended for a camera works well, and so does a WG-345 clear glass filter.

If you have an uncalibrated detector, subtract the second reading (B) from the first (A) to get a voltage that will be correct with respect to measurements you make at other times.

If you have a calibrated detector, you can compute the absolute spectral irradiance at 300 nanometers in terms of K' watts per square meter. Ultraviolet-blocking filters typically reflect about 8 percent of the incident nonultraviolet radiation. The nonultraviolet radiation without the filter is therefore approximately equal to the B reading divided by 92 percent. Because the active area of the DFA-3000 is about 9.9 square millimeters, the detector signal must be multiplied by 101,000 to find the signal per square meter. The formula that results is

where Dr is the detector's calibrated responsivity and F is the filter's band pass. (The band pass is the number of nanometers between the two points where the filter's transmission falls to half the maximum.) The ideal filter would have a band pass of less than a nanometer. Real filters have a wider band pass.

My detector has a responsivity of .04 ampere per watt and a filter band pass of 10.4 nanometers. Readings at noon on a clear August day are typically 1.50 (A) and .116 volt (B). InserUng these values into the formula gives

or .011 watt per square meter per nanometer. Remember, this is direct ultraviolet. The diffuse contribution from radiation scattered by molecules in the atmosphere adds at least 30 percent to this value at my latitude.

You can write a computer program to solve the formula. An even better approach is to program a computerized spread sheet to solve the formula, record the time and date, graph the measurements and save them.

An ultraviolet-B radiometer based on a gallium phosphide diode can also be built. These detectors, unlike silicon photodiodes, do not respond to red light and thus provide true solar-blind operation. Gallium phosphide diodes are made by Hamamatsu Corporation (P.O. Box 6910, Bridgewater, NJ 08807). The G1961 ($27.45 plus shipping and tax) is housed in a TO-18 package and has an effective surface area of 1.0 square millimeter. The G1962 ($35) is housed in a larger TO-5 package and has a surface area of 5.2 square millimeters. For a limited time and for readers of this department, Hamamatsu will calibrate the G1962 at 300 nanometers for an additional $65.

The most expensive component of the solar-blind radiometer is the optical filter. High-quality filters are made by Barr Associates (P.O. Box 557, Westford, MA 01886). Barr makes filters only on a custom basis. Therefore, unless you are connected with an institution that can afford to place a custom order, you will need to go elsewhere.

MicroCoatings (One Lyberty Way, Westford, MA 01886) makes a 12.5-millimeter-diameter filter that transmits 300-nanometer radiation and has a band pass of 10 nanometers (catalogue number ML3-300). The price is $77, a very reasonable amount for an interference filter. Twardy Technology, Inc. (P.O. Box 2221, Darien, CT 06820), sells a 25-millimeter-diameter filter with the same specifications for $210.

Most important in building the solar blind radiometer is to install the detector and the filter in a light-tight housing. If you have access to a machine shop, you can make one. Or you can install a 12.5-millimeter filter and detector in a brass compression fitting or union coupling [see Figure 3]. The coupling and the required O rings are available from hardware and plumbing stores. A two-conductor phone plug is inserted into one of the union's caps and secured in place with a rubber O ring. The leads of the detector are inserted into a light-emitting diode socket soldered to the plug's terminals. You can, however, solder the detector directly to the terminals. The cathode lead should be soldered to the terminal that is common to the hp of the plug. In either case, some couplings will accept only detectors in miniature TO-18 packages.

The filter, protected by a pair of O rings, is installed in the second end cap. A conical cap works best but may be hard to find. If the filter and O rings do not leave sufficient space for the end cap's threads to engage those of the union, replace one of the O rings with a paper spacer. Screw the end cap down so that it stays in place but does not apply pressure to the filter. If necessary, cement the end cap in place with a drop of removable glue. Be sure the filter remains clean during the installation procedure.

Depending on the detector's dimensions, a conical end cap will give a field of view of around 10 degrees. You should therefore attach a collimator tube to the opening in the end cap to reduce the field of view to four degrees or less. Brass tubing can be soldered or cemented to the opening in the end cap. Coat the inside of the tube with flat, black paint.

I store filter-detector assemblies in an airtight plastic refrigerator container along with a package or two of silica gel desiccant. The desiccant helps to protect the filters from deterioration caused by long-term exposure to water vapor. Ask a pharmacist or a salesperson at an electronics or camera store to save silica gel packets for you.

Because the solar-blind radiometer includes a separate detector, filter and amplifier, it is more difficult to assemble than an instrument in which these components are combined in a single package. Nevertheless, its circuit is functionally identical to that of the first instrument, as can be seen by referring to the circuit diagram.

An important advantage of this radiometer is that it consumes very little current. Because it can be powered by a single nine-volt battery, it easily fits inside a compact housing such as Radio Shack's part number 270-257. The operational amplifier specified in the accompanying circuit diagram (TLC271CP) can be purchased from major electronics distributors or Texas Instruments (P.O. Box 225012, Dallas, TX 75265). Other operational amplifiers can also be employed if they have a very low input bias current. Check the manufacturer's data sheet, since they may require different connections.

The TLC271CP can be damaged by static electricity. Therefore, do not touch the pins of the operational amplifier while installing it. You can avoid this problem by soldering an eight-pin integrated-circuit socket to the circuit. Insert the operational amplifier into the socket after assembling the circuit but before applying power.

The solar-blind radiometer operates much like the previous instrument except that its output will not fall below zero, because only a single battery powers the circuit. Therefore, adjust the potentiometer until the output just reaches zero when the collimator is blocked. The formula for the DFA-3000 detector system will help you calculate the ultraviolet level. The filter manufacturer should provide the filter's peak transmission wavelength and its transmission in percentages. Literature available from Hamamatsu states the detector's area and a graph of typical responsivity versus wavelength.

Regularly measuring direct solar ultraviolet B with either a calibrated or an uncalibrated detector can yield significant data. Always try to take a measurement at solar noon. For more information about determining solar noon, consult any standard reference on astronomy or sundial construction.

Rarely does the peak ultraviolet-B reading occur precisely at solar noon. Instead the signal fluctuates constantly as the 300-nanometer radiation is attenuated and scattered by the atmosphere and its constituent gases. For this reason, allow at least five minutes to make a single measurement.

Virtually every day for two years I have measured the direct solar flux at four ultraviolet-B wavelengths and six additional wavelengths. I have found that direct solar irradiance at 300 nanometers is significantly attenuated by fog, haze, clouds and aircraft contrails. During the summer of 1988 I observed that ultraviolet-B radiation decreased when smoke from the Yellowstone fires drifted over Texas.

The passage of a cold front, which raises barometric pressure, is more often than not followed by a reduction in ultraviolet B, even when the atmosphere is exceptionally clear and dry. Satellite data and my daily ozone measurements confirm that this phenomenon is caused by an increase in ozone of as much as 15 percent and sometimes more. Low barometric pressure, at least where I live in central Texas, is typically accompanied by a decrease in ozone. For example, when Hurricane Gilbert passed nearby in September of 1988, my radiometer registered a significant increase in ultraviolet B when the sun was visible. Several weeks later a packet of data from Arlin Krueger of the Goddard Space Flight Center revealed a simultaneous, pronounced decrease in ozone.

My observations may not hold true for your location, which is reason enough for you to conduct your own measurements. For example, can you explain and measure an increase in ultraviolet B caused by snow or a decrease caused by grass fires or a rapidly approaching squall line? If you make measurements at two closely spaced ultraviolet-B wavelengths, you can compute the ozone in a column through the atmosphere. For the relevant formulas and a list of references, send me a stamped, self-addressed envelope in care of SCIENTIFIC AMERICAN.

Bibliography

BIOLOGICALLY EFFECTIVE ULTRAVIOLET RADIATION: SURFACE MEASUREMENTS IN THE UNITED STATES, 1974 TO 1985. Joseph Scotto et al in Science, Vol. 239, No. 4841, Part 1, pages 762-764; February 12, 1988.

THE BUDGET OF BIOLOGICALLY ACTIVE ULTRAVIOLET RADIATION IN THE EARTH-ATMOSPHERE SYSTEM. John E. Frederick and Dan Luhin in Journal of Geophysical Research, Vol. 93, No. D4, pages 3825-3832; April 20,1988.

INDICATION OF INCREASING SOLAR ULTRAVIOLET-B RADLATION FLUX IN ALPINE REGIONS. Mario Blumthaler and Walter Ambach in Science, Vol. 248, No. 4952, pages 206-208; April 13, 1990.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.