Mv. Space I/es. Vol. 12, No.5, pp. (5)141—(5)149, 1992 0273—1177/92$15.

tt)
Printed inGreat Britain. All rights reserved. Copyright Cl 1991 COSPAR

EVALUATION OF LIGHT EMITFING DIODE

CHARACTERISTICS FOR A SPACE-BASED
PLANT IRRADIATION SOURCE
D. J. Barta, T. W. Tibbitts, R. J. Bula and R. C. Morrow*
Wisconsin Centerfor Space Automation and Robotics, and the Department of
Horticulture, University of Wisconsin-Madison, Madison, WI 53706, U.S.A.

ABSTRACT
Light emitting diodes (LEDs) are a promising irradiation source for plant growth in space. Improved semiconductor
technology has yielded LED devices fabricated with gallium aluminum arsenide (GaA1As) chips which have a high
efficiency for converting electrical energy to photosynthetically active radiation. Specific GaAIAs LEDs are available
that emit radiation with a peak wavelength near the spectral peak of maximum quantum action for photosynthesis.
The electrical conversion efficiency of installed systems (jJmol ~-1 of photosynthetic photons per watt) ofhigh output
LEDs can be within 10% of that for high pressure sodium lamps. Output of individual LEDs were found to vaiy by
as much as 55 % from the averageof the lot. LED ratings, in mcd (luminous intensity per solid angle), were found
to be proportional to total photon output only for devices with the same dispersion angle and spectral peak.
Increasing current through the LED increased output but also increased temperature with a consequent decrease in
2 ~-1 has been produced on
electrical conversion
surfaces using arrays efficiency. A photosynthetic
with LEDs mounted photon
7.6mm apart, flux as high
operating as 900 of
at a current ~.tmolrn- device-i and at an installed
50 mA
density of approximately 17,200 lamps m—2 of irradiated area. Advantages ofLEDs over other electric light sources
for use in space systems include long life, minimal mass and volume and being a solid state device.

INTRODUCTION
Supplying adequate photosynthetic irradiance to plants in space-based research or bioregenerative life support
facilities offers many challenges /1/. Direct sunlight may be unsuitable because of problems associated with its
duration, intensity or safe collection /2/. Electric light sources can be used as an alternative to solar radiation to
provide photosynthetic irradiance. A candidate electric light source must have high electrical efficiency, an optimal
spectral output for photosynthesis and photomorphogenesis, small mass and volume and an excellent record of safety
and reliability.
No currently available commercial lamp meets all of these criteria. High intensity discharge lamps have the highest
electrical efficiency and greatest expected life of currently available commercial lamps /3/. However, they produce a
large amount of long wave radiation /4/ and pose a safety risk because of their high operating temperature and
explosion hazard. Specially mounted fluorescent lamps have met safety requirements for STS flight experiments;
however, because of their relatively low electrical efficiency and bulky size, high photosynthetic irradiation levels can
not be achieved /5/.

Recent advances in semiconductor technology have led to the development of red light emitting diodes (LEDs) with
intensities and electrical efficiencies competitive with commonly used, commercially available lamps /6/. These new
devices have a peak wavelength around 660 nm, corresponding with the maximum photosynthetic quantum action of
plants 17/. Being solid state devices, these lamps have superior safety and reliability characteristics. A plant lighting
system based on LEDs has been developed /8/. Several crops including lettuce, spinach, wheat and potatoes have
been successfully grown under LED irradiation /9,10,11/. This paper describes the characteristics of these recently
introduced LEDs and how various LED characteristics impact the design and construction of LED irradiation systems
for use in plant growing facilities.

DESCRIPTION OF LED DEVICES

Red LEDs constructed with gallium aluminum arsenide chips (GaAlAs or A1GaAs) have undergone significant
technological developments. Changes in chip (synonymous with die) structure have led to more efficient production
and emission of photons. The new chip designs have a double-heterostructure and transmitting substrate (DH-TS).
They are also referred to as double-power, double-heterostructure devices (DDH). In commercial literature, LEDs
with a DH-TS chip have been termed ultra-bright, super-bright, hi-super bright and sun-brightLEDs.

A typical LED lamp is illustrated in Figure 1. It is composed of an LED chip mounted on a spherical or parabolic
reflector, encapsulated in a transparent epoxy casing. The reflector at the base of the chip and a lens at the tip of the
casing direct emitted radiation at a specific angle ofdispersion (also termed viewing angle). Angle ofdispersion will
affect LED focus, as illustrated in Table 1, but will not affect total radiant output
* The authors wish to acknowledge the contributions of J. Vignali, Associate Research Specialist.

(5)141
(5)142 D. J. Barta et at.

TABLE 1 Effect of Angle of Dispersion to On-axis An le of

Intensity of Typical GaAlAs LEDsY. Dispersion ~

Angle of dispersion On-axis intensityz

Transparent
Casing 1,
1
~‘
1
(degrees) (mcd) GaAIAs II
DH-TSChip CC:.’
8 1000 ‘CC!
17 500 Reflective
24 250 Mount
30 200
45 100 Wire__—~-
60 60

‘ Adapted from reference /12/.

z Centered under and perpendicular to the LED. Fig. 1. Significant components of an LED.

LED devices are powered by DC current. The radiant output is proportional to the forward current provided to the
chip (Figure 2). After a critical forward voltage is reached, the device begins to draw current rapidly. The power
supply’s output should be current-limited to prevent LED damage from accidental power overload.

2.0

“a 1.0~

0 25 50 75 100 125 150

Forward current (mA)

Fig. 2. Relative photon output of a GaA1As LED at different levels of forward current. The maximum

recommended forward current without supplemental heat-sinking is 50 mA.

CHARACTERISTICS OF GaA1As LEDS

Spectral Outout

Most of the available DH-TS GaAlAs LEDs have a narrow spectral output with a peak wavelength near 660 nm
(Figure 3). This peak wavelength corresponds to the maximum photosynthetic quantum action of plants 17/ and a
maximum absorption region of chlorophyll /13/. The half-power bandwidth is about 20 nm. Radiation in the
ultraviolet, blue, green, yellow, orange and infrared regions of the spectrum is absent (Table 2). GaA1As LEDs
produce significantly less non-photosynthetic irradiance than other commonly used electric lamps (Table 3).

The peak wavelength of different GaAIAs LEDs from several manufacturers was found to vaiy from 650 to 670 nm.
The peak wavelength also varied among individual devices obtained from a single manufacturer. Variations in peak
wavelength among devices were correlated with differences in lamp intensity. Individual lamps with higher radiant
output tended to have a higher peak wavelength. The peak wavelength can also be shifted by the specific
composition of the chip, device temperature (Figure 4) and electrical load on the device (Table 4). The effect of
electrical load on the peak wavelength is probably a temperature effect. Increased power from increased forward
currents will result in increased lamp temperatures (Table 4). Differences in peak wavelength among devices within
the same lot may be due to variations in composition of the wafer used in their manufacture or differences in cooling
rates among devices. Stricter manufacturing tolerances or improved quality control practices could minimize these
observed differences.
 Light Emitting Diode Characteristics for Plant Irradiation (5)143

1.2

~1.0 —
‘a Photosynthetic
quantum action

I
~0.6I

~0.4
()
I-
~0.2 I
LED output
0.0 . • I • I~

300 400 500 600 700 800

Wavelength (nm)

Fig. 3. Emission of a GaA1As LED plotted against the photosynthetic quantum action ofplants 17/.

TABLE 2 Spectral Characteristics of Radiation Sources.

Percent ofphoton output (%)y

300- 400- 500- 600- 700-

Lamp 400 nm 500 nm 600 nm 700 nrn 1100 nm

DH-TSGaA1A5LEDz 0.0 0.0 0.1 98.9 1.0

High Pressure Sodium 0.3 4.0 23.2 31.8 40.7
Metal Halide 6.5 14.5 38.2 14.7 26.0
Cool-white Fluorescent 2.4 18.7 49.2 23.5 6.2

yPercent ofphoton output in selected wavebands over the spectral region of 300-1100 nm.
Measurements were made using a LICOR LI-1800 spectroradiometer.
Double-heterostructure, transmitting-substrate light emitting diode.

TABLE 3 Short and Longwave Emissions from Different Radiation Sources.w

2)
Irradiance(Wm-
Photosynthetic
photon flux Non-
(p.mol rn-2 s-i) Photosynthetic photosynthetic Total
Lamp (0.4 - 0.7 pm) (0.4 - 0.7 jim) (.3-.4 & .7-100 jim) (.3-100 jim)

DH-TS GaA1As LED5X 400 72 51 123

High Pressure SodiumY 400 78 143 219
Metal Halide~ 400 87 97 184
Cool-white Fluorescentz 400 88 127 215
w Measurements were taken under lamps and fixtures without barriers. Fixtures were hung in a room with white
enamel walls. Air temperature was maintained at 20’C. PPF was measured using a LICOR LI-19OSA quantum
sensor. Photosynthetic irradiance was determined using a LICOR LI-1800 spectroradiometer. Total irradiance
was determined using a total hemispherical radiometer(Radiation Energy Balance Systems, model THRDS3).
Double-heterostructure, transmitting-substrate light emitting diodes. An array of 144 LEDs powered at 50 mA
per device, 14.2 watts (W) total power, was used.
Y Single 400W lamps in fixture.
z Four 110W lamps in fixture.
(5)144 D. J. Barta ‘a al.

Although most commercially available red photon emitting GaAlAs LEDs have a peak wavelength at 660 nm,
GaAlAs devices with other peak wavelengths can be fabricated. By adjusting the amount of Al and Ga in the active
layers of the chip, GaAlAs devices with a peak wavelength ranging from 630 to 940 nm can be obtained. The
spectral outputs of two such devices are presented in Figure 5. One company currently manufactures devices of any
desirable peak wavelength within this range /14/. This is significant to photobiologists because it is now possible to
have high-output narrow-band devices with peak emission at 660 and 730 nm, closely matching the peak absorptance
ofthe red and far-red forms of phytochrome. LEDs with peak wavlengths between 630 and 660 nmwould also be
useful for photosynthetic lighting. Bandwidth increases with peak wavelength. Devices with a peak wavelength at
940 nm would be expected to have a half-power bandwidth of about 40 nm.

668

666 .0—

I1i1’

10 20 30 40 50 60
Temperature (°C)

Fig. 4. Relationship between peak wavelength and device temperature. A single LED was mounted to a
temperature-controlled heat sink. Peak wavelength at different device temperatures was determined using a
LICOR LI-1800 computerized spectroradiometer.

TABLE 4 Peak Wavelength and LED Temperature at Different

Forward Current Levels.

Forward Peak Temperature

Current Wavelength of Device
(mA) (nm)Y (‘C)z

10 658 27.1
30 659 28.9
50 661 30.7
90 665 34.5
150 672 41.1

~Measuredusing a LICOR LI-1800 spectroradiometer.

‘Measured at base of LEDs. Air temperature 25’C.

Highly efficient LEDs with peak wavelengths shorter than 630 nm are not currently available. Gallium phosphide
(GaP) and gallium arsenide phosphide (GaAsP) LEDs are available with peak wavelengths between 550 and 640 nm,
but have considerably less output compared with GaA1As LEDs (Figure 5). Silicon carbide (SiC) LEDs are available
with output between 400-500 nm, but achievable intensities are only about one one-thousandth of that available from
GaAlAs LEDs (Figure 5). High output LEDs having a peak wavelength in the blue region of the spectrum would
greatly enhance the development of LED-based plant irradiation systems, because some higher plants have a
photomorphogenic requirement for blue light /15,16/.
Photosynthetic Photon Oumut
Brightness of most LEDs is rated in terms of their perpendicular or on-axis intensity at a given forward current and is
expressed as millicandellas (mcd). Device dependent factors such as angle of dispersion have a dramatic effect on
mcd intensity without changing total radiant output or device efficiency (Table 1). Since the mcd is a measure of
luminous intensity (lumens per solid angle) and not total radiant output, it is not useful for describing output of a
device to be used for irradiation of plants. In addition, because the mcd is a measure of luminous intensity
 Light Emitting Diode Characteristics for Plant irradiation (5)145

(sensitivity of the human eye), it is not an appropriate unit of photosynthetic response /17/. These considerations
make comparisons of different lamp types, based on manufacturer’s information, extremely difficult.

10
GaAlAs

.1 GaP

~ .01

SiC

.0001 II’IIIIII”I”

400 450 500 550 600 650 700 750 800 850 900
Wavelength (nm)

Fig. 5. Emission spectra of different types ofLEDs plotted on a log scale.

The photosynthetic photon output (photons 400-700 nm) ofGaA1As LEDs varied greatly. LEDs were compared at a
forward current of 50 mA [approximately 0.1 watt (W), depending on the resistance of the specific LED]. To
overcome problems associated with differences in angle of dispersion between devices, we devised a simple
reflective cylinder that fits over a quantum sensor to collectmost of the downward radiation emitted by the LED lamp
(Figure 6). A small amount of emitted radiation escapes from the back of the LED and is not included in these
measurements. This device has been used to screen individual LEDs for differences in photosynthetic photon output
and efficiency. By using a current-controlled power supply, set at a standardized forward current, the output of
thousands of individual LEDs have been determined for selection of those with exceptionally high output and
efficiency.

~1 ~-LED
Aluminum _____________________________
Cylmder 11 ~
Reflective
Inner
~ fl“—‘

Wall _________

Quantum ________ ~ ~s~s~s

Sensor ~ __________

Fig. 6. Aluminum cylinder with reflective inner wall mounted overa quantum sensor to collect photons emitted
by an LED.

We have observed significantdifferences in photosynthetic photon output among individual LED5 from the same lot
and among lots from the same company (Table 5). A frequency distribution of photosynthetic photon output,
summarized over seven lots, is given in Figure 7. The output ofindividual LEDs within a specific lot varied as much
as 56 % of the average of the specific lot. A similar variability in photosynthetic photon output was found among
individual LEDs from different manufacturers.

The effect of lamp aging on photosynthetic photon output is not well understood. Factors such as on/off cycle, level
offorward current, operating temperature, chip type, treatment to prevent reaction of the chip with the encapsulating
material (passivation) and presence and type of transparent encapsulating material all may affect aging characteristics
JASR 12:5-K
(5)146 D.J. Barta et al.

of LEDs. Manufacturers have not tested devices under conditions likely to occur in plant growth facilities. Some
industry representatives indicate the life expectancy of DH-TS GaAlAs LEDs to be as long as 100,000 hours. Some
studies have actually shown increases in LED efficiency after 1000 hours of use /12/. We have observed 10% a
degradation in PPF underlamps after 3000 hours of continuous use at the maximum recommended forward current
of 50 mA. Degradation may be less, or negligible, if lower forward currents are used. It is obvious that more
research on this important consideration is needed before conclusions about the aging effects on LED output can be
made.

TABLE 5 Differences in Photosynthetic Photon Output of Individual

Devices from 7 Lots of the Same Type ofLED Obtained from a Single
Manufacturer.
Photosynthetic Photon Output
(% of mean of all lots)z

Lot N Mean Mm Max

A 330 86 52 125
B 200 92 44 130
C 144 94 55 126
D 352 100 77 128
E 80 105 81 129
F 362 114 81 153
G 90 115 79 153

All lots 1558 100 44 153

z Photons between 400-700 nm. Data are expressed as a percentage of

the mean photosynthetic photon output of all devices from all lots.
Photons emitted from the LEDs were collected by a reflective cylinder
that fit over a LICOR LI-19OSA quantum sensor, as pictured in
Figure 6. LEDs were powered at a forward current of 50 mA.

25

____
0.4 0.5 0.6 0.7 0.8 0.9
I I
1
~

1.1
~

1.2 1.3 1.4 1.5 1.6

Photon output relative to the mean

Fig. 7. Frequency distribution of LED photosynthetic photon output. 1760 devices were screened at a
common forward current of 50 mA using the setup pictured in Figure 6.

Efficiency

The electrical conversion efficiency of a highly efficient LED plant irradiation unit is given in Table 6. Installed
1 W-i were achieved when devices were operated at 10 mA forward
electrical efficienciesdecreased
current. Efficiency as high asas0.911 jimol s- current applied to the LED unit was increased. At the maximum
the forward
recommended forward current of 50 mA, electrical conversion efficiency was 18 % lower than at 10 mA. The
installed electrical conversion efficiency of the LED unit was superior to published information for fluorescent
systems and close to that for a high pressure sodium lamp system (Table 7). It is possible that further improvements
 Light Emitting Diode Characteristics for Plant Irradiation (5)147

in LED technology will lead to installed electrical conversion efficiencies of LED units that are equal to or greater than
those for high pressure sodium lamps.

TABLE6 Photon Flux and Electrical Efficiency of an LED Lighting System at 5 Levels of
Electrical Load.Y

LED Forward Power per Photon Electrical

Current Unit Area Flux Efficiency
2) (jimol m-2 s-i)’ (jimol s-’ W-i)
(mA) (W rn-

10 215 196 0.911

20 447 394 0.882
30 694 583 0.839
40 952 756 0.794
50 1226 913 0.745

‘ An array of 144 high efficiency LEDs were mounted on a 10.8 cm by 10.8 cm reflective
board at a density of 17,200 devices m-2. Four reflective walls were fashioned around the
array to collect scattered photons. The photon flux was determined using a LICOR
LI-19OSA quantum sensor 5 cm beneath the LED array.
Photosynthetically active radiation (400-700 nm)

TABLE 7 Electrical Efficiency ofSelected Photosynthetic Radiation Sources Installed

in Plant Growth Facilities.

Electrical Efficiencyw
Photosynthetic Radiation Source (jimol s-1 W1)

High Pressure Sodium’ 1.00 - 1.52

DH-TS GaA1As LEDs~ 0.20 - 0.91
Cool White Fluorescentz 0.13 - 0.75
w Efficiency dependent on specific lamps selected, type of luminaires, reflective char-
acteristics ofplant growth chamber, electrical load, distance from lamps and lamp age.
‘Data for typical growth rooms, derived from personal experience of authors and Dr.
B. Bugbee /18/.
y LEDs selected for very high photon output. Measurements performed in a small plant
growth unit.
z Data for typical growth rooms and cabinets, derived from personal experience of
authors and reference /19/.

A plant irradiation system that provides 400 pruol m-2 s-1 and has an installed electrical conversion efficiency of0.88
jimol s-1 W-i would consume —10.9 kW of electrical energy to irradiate the area required for life support of one
person. This calculation assumes 24 m-2 of irradiated growing area is required for each person for life support
/20,21/. For a six-person crew on a space base, the power requirement would be —65 kW of electrical power. For
comparison, preliminary designs provide only 75 kW of total power for all purposes on Space Station Freedom /22/.
Consequently, conversion efficiency of electrical energy to photosynthetically active radiation is a critical factor in the
selection of a lighting system for use in space-based plant growth facilities. It is important that comparative
efficiencies of different lighting systems be based on measurements made in an installed facility, because losses
related to growth room geometry and reflector quality ofthe walls and the lamp mounting can significantly affect PPF
at a defined distance from the light source.

An important advantage of the LED irradiation system is that the peak spectral output coincides closely with the
waveband of maximum photosynthetic quantum action of plants (Figure 3). Thus, a greater photosynthetic
utilization efficiency can be expected from photons emitted from LEDs than from lamps with spectral output spread
over all photosynthetically active wavebands /23/. At equal levels of photosynthetic photon flux, net assimilation
rates of plants under LED irradiation systems would be expected to be greater than under other irradiation systems.

Increases in the effectiveness of LEDs may be possible by use of pulsed irradiance. GaA1As LEDs have an
extremely rapid rise and fall time (—80 ns) which would enable the devices to be pulsed on and off at frequencies as
high as 2 MFIz /12/. On/offduty cycles that would provide a cooling period between current pulses, would allow the
devices to be powered at instantaneous forward currents significantly greater than with continuous operation. At
short duty cycles (10%) the instantaneous PPF level during pulses can be up to 6 times greater than the level
maintained during continuous operation. Some studies suggest that the efficiency of photosynthesis could be
increased by the use of rapidly pulsed PPF, either through an off-cycle enhancement of photosynthetic rate or by
minimizing photon losses resulting from the difference in time constants between the light and dark reactions of
photosynthesis /24/. Careful experiments are required because plants tend to integrate PPF, particularly at low
(5)148 D. J. Barta eta!.

intensities and at pulse frequencies less than —2000 Hz /25/. Research has been initiated at NASA’s Ames Research
Center to evaluate the potential value of pulsing an LED irradiation source to increase efficiency through increased
plant activity and decreased electrical requirements.

LED PLANT IRRADIATION SYSTEMS

LED irradiation systems can be constructed by mounting individual LED lamps on boards and connecting devices in
senes until a desired series voltage is reached. A GaA1As DH-TS device typically operates at —2.0 VDC. Twelve
devices in series requires a series voltage of —24 VDC. We have observed that the forward voltage of GaA1As
DH-TS LEDs at a specific current level is quite uniform among individual devices within the same lot and among
different lots of the same device. Several sets of devices in series could be connected in parallel and powered by a
single current-controlled DC power supply. However, because the effect of aging on driving voltage has not been
defined, it may be prudent to install current controlling circuits on each series of devices. If the space facility has a
2 would be required.
DC power
Ballasts, as source,
requiredonly simple current-controlling
for fluorescent and high intensitycircuits having
discharge a mass
systems, of not
would 1 tobe2 necessary.
kg rn-

The effect offorward current (mA) on the PPF 5 cm below an array, constructed with highly-efficient selected LEDs,
is presented in Table 6. Levels as high as 900 j.Lmol rn-2 s-1 have been achieved at a density of approximately 17,200
devices rn-2 and at a forward current of 50 mA. Photon level is directly proportional to LED lamp density. An array
constructed at an LED lamp density of 8,600 lamps m-2 yielded half as much output. PPF at greater distances will
depend on the reflective characteristics of the particular growth chamber in which the unit is used. Because the
output from an LED is highly focused, PPF levels will not decrease as rapidly with increasing distance from the lamp
as compared to other lamp systems. In addition, irradiance across the plant canopy can be quite uniform because of
the many individual devices that make up the array.

LED plant irradiation systems will take up very little head space in a plant growth chamber. This is very significant
to space-based plant growth facilities where available growing height and volume will be limited. LED arrays built
with commercially available lamps are only about 1.5 cm deep. LED chips, without the transparent casing, could be
directly mounted to reflective circuit boards that could be used as the growth chamber ceiling. The entire ceiling
would essentially be an emitting surface and the lighting system would have minimal volume and mass.
When red photon emitting LEDs are used as an irradiance source for some plant species, a quantity of blue photons is
required (—5 to 10 % of the total photons) /10,15/. The electrical conversion efficiency of commercially available
blue photon emitting SiC LEDs, however, is too low at this time to be useful in LED-based plant irradiation systems.
Efficient LEDs that emit in the blue region of the spectrum may become available in the future. To provide blue
irradiance when required by specific crops, we have supplemented LED irradiation systems with blue-phosphor
fluorescent lamps.

ADVANTAGES OF LED IRRADIATION SYSTEMS FOR SPACE APPLICATIONS

LEDs are a promising alternative irradiation source for plant growth facilities in space because of their low mass and
volume. They have a near optimum spectral output for photosynthesis. Devices are available with electrical
conversion efficiencies approaching the most efficient lamp sources used for plant irradiation. Because they are
solid-state devices, they offer safety and reliability characteristics superior to presently used lamps. Research is
needed to enhance device efficiency, optimize emission geometry and optics for plant lighting and define factors that
affect device aging and useful life. Improvements in LEDs with photon output in the blue region of the spectrum
would enhance the development ofan LED plant irradiation system for applications in a space environment.

REFERENCES

1. R.W. Langhans and D.R. Dreesen, Challenges to plant growing in space, HorrScience 23, #2, 286-293 (1988).

2. R.L. Olson, M.W. Oleson and T.J. Slavin, CELSS for advanced manned mission, HortScience 23, # 2,
275-286 (1988).
3. J.E. Kaufman and J.F. Christensen (eds), JES Lighting Handbook. Reference Volume, illuminating Engineering
Society of North America, New York, 1984.
4. D.L. Bubenheim, B. Bugbee and F.B. Salisbury, Radiation in controlled environments: influence of lamp type
and filter material, J. Amer. Soc. Hort. Sci. 113, # 3, 468-474 (1988).
5. Annonymous, Plant growth unit. Comment edition. Ames Research Center, NASA, Moffett Field, California
(April 1983).
6. L.W. Cook, M.D. Camras, S.L. Rudaz and F.M. Steranka, High efficiency 650 nm aluminum gallium arsenide
light emitting diodes, in: mt. Symp. GaAs and Related Compounds, Heraklion, Greece, 1987. Inst. Phys.
Conf. Ser. No. 91, 1988, p. 777-780.
7. K.J. McCree, The action spectrum, absorptance and quantum yield of photosynthesis in crop plants, Agr.
Meteorol. 9, 191-216 (1972).

8. R.W. Ignatius, T.S. Martin, R.J. Bula, R.C. Morrow and T.W. Tibbitts, Method and apparatus for irradiation
 Light Emitting Diode Characteristics for Plant Irradiation (5)149

of plants using optoelectronic devices, U.S. Patent Anplication 07t283.245 (1988).

9. D.J. Barta, R.J. Bula and T.W. Tibbitts, Wheat growth under a light emitting diode irradiance source, ASGSB
Bull. 4, # 1, 50 (1990).

10. R,J. Bula, R.C. Morrow, T.W. Tibbitts, D.J. Barta, R.W. Ignatius and T.S. Martin, Light emitting diodes as a
radiation source for plants, HortScience, in press (1991).

11. A. Hoehn and M.W. Luttges, Tuning the efficiencies of plant lighting systems: from primary energy to edible
biomass - the LED as an alternative light source, in: Life Sciences Research in Space, Fourth European
Symposium, 28 May - 1 June 1990. Trieste, Italy, ESA-SP-307, European Space Agency (1990).

12. F.M. Steranka, D.C. DeFevere, M.D. Camras, C. Tu, D.K. McElfresh, S.L. Rudaz, L.W. Cook and W.L.
Snyder, Red A1GaAs light-emitting diodes, Hewlett-Packard 1., August, 84-88 (1988).

13. L.P. Vernon and G.R. Seely (cdi), The chiorophylls, Academic Press, New York 1966.
14. Annonymous, QDI-DDH series, application specific LED’s (640 nm - 940 nm), Quantum Devices, Inc., P.O.
Box 100, Barneveld, Wisconsin, 53507, U.S.A. (1990).
15. M.E. Hoenecke, R.J. Bula and T.W. Tibbitts, Lettuce seedling response to red light emitting diodes
supplemented with varying levels of blue photons, ASGSB Bull. 3, # 1, 59 (1989).
16. W. Schmidt, Bluelight physiology, BioScience 34, # 11, 698-704 (1984).

17. M.G. Holmes, W.H. Klein and J.C. Sager, Photons, flux and some light on philology, HortScience, 20, # 1,
29-31 (1985).

18. B. Bugbee, private communication (1990).

19. T.W. Tibbitts, J.C. McFarlane, D.T. Krizek, W.L. Berry, P.A. Hammer, R.W. Langhans, R.A Larson and
D.P. Ormrod, Radiation environment of growth chambers, J. Am. Soc. Hort. Sd. 101, 164-170 (1976).

20. F.B. Salisbury and B.G. Bugbee, Wheat fanning in a lunar base, in: Lunar Bases and Space Activities of the
21st Century, ed. W.W. Mendell, Lunar and Planetary Institute, Houston, 1985, p 635-645.

21. R.M. Wheeler and T.W. Tibbitts, Utilization of potatoes for life support systems in space: Ill. Productivity at
successive harvest dates under 12-h and 24-h photoperiods, Amer. Potato J. 64, 311-320 (1987).

22. Annonymous, Space Station Freedom. Opportunities for commercial users and providers, Pre-workshop,
Denver, Colorado, NASA, Office of Commercial Programs (October 25, 1988).

23. J.C. Sager, J.L. Edwards and W.H. Klein, Light energy utilization efficiency for photosynthesis, Transactions
of the ASAE 25, # 6, 1737-1746 (1982).
24. R. Emerson and W. A. Arnold, A separation of the reactions in photosynthesis by means of intermittent light. I.
Gen. Physiol. 15, 391-420 (1932).

25. J.C. Sager and W.G Giger, Re-evaluation of published data on the relative photosynthetic efficiency of
intermittent and continuous light, Agric. Meteorol. 22, 289-302 (1980).