Life Sciences in Space Research 2 (2014) 4353

Contents lists available at ScienceDirect

Life Sciences in Space Research

www.elsevier.com/locate/lssr

Signicant reduction in energy for plant-growth lighting in space

using targeted LED lighting and spectral manipulation
L. Poulet a, , G.D. Massa b , R.C. Morrow c , C.M. Bourget c , R.M. Wheeler b , C.A. Mitchell a
a
b
c

Purdue University, West Lafayette, IN 47907, USA

NASA Kennedy Space Center, FL, USA
ORBITEC, Madison, WI, USA

a r t i c l e

i n f o

Article history:
Received 13 January 2014
Received in revised form 6 June 2014
Accepted 9 June 2014
Keywords:
Crop
ESM
Food production
Life-support systems
Light-emitting diodes
Lettuce

a b s t r a c t
Bioregenerative life-support systems involving photoautotrophic organisms will be necessary to sustain
long-duration crewed missions at distant space destinations. Since sucient sunlight will not always
be available for plant growth at many space destinations, ecient electric-lighting solutions are greatly
needed. The present study demonstrated that targeted plant lighting with light-emitting diodes (LEDs)
and optimizing spectral parameters for close-canopy overhead LED lighting allowed the model crop leaf
lettuce (Lactuca sativa L. cv. Waldmanns Green) to be grown using signicantly less electrical energy
than using traditional electric-lighting sources. Lettuce stands were grown hydroponically in a growth
chamber controlling temperature, relative humidity, and CO2 level. Several red:blue ratios were tested
for growth rate during the lag phase of lettuce growth. In addition, start of the exponential growth
phase was evaluated. Following establishment of a 95% red + 5% blue spectral balance giving the best
growth response, the energy eciency of a targeted lighting system was compared with that of two
total coverage (untargeted) LED lighting systems throughout a crop-production cycle, one using the
same proportion of red and blue LEDs and the other using white LEDs. At the end of each cropping
cycle, whole-plant fresh and dry mass and leaf area were measured and correlated with the amount
of electrical energy (kWh) consumed for crop lighting. Lettuce crops grown with targeted red + blue
LED lighting used 50% less energy per unit dry biomass accumulated, and the total coverage white LEDs
used 32% less energy per unit dry biomass accumulated than did the total coverage red + blue LEDs. An
energy-conversion eciency of less than 1 kWh/g dry biomass is possible using targeted close-canopy
LED lighting with spectral optimization. This project was supported by NASA grant NNX09AL99G.
2014 The Committee on Space Research (COSPAR). Published by Elsevier Ltd. All rights reserved.

1. Introduction
Planetary exploration and expansion of humanity into the solar
system to establish permanent settlements are grand challenges
of the 21st century (NASA, 2010, 2012). The goal of human exploration set by the Global Exploration Roadmap is a rst human
mission to Mars by 2040 (ISECG, 2013). Standard mission-to-Mars
scenarios envision a crew of 6 people and a total mission duration
of approximately 1000 days (Drysdale et al., 2003), requiring a total
mass of consumables (food, water, oxygen) that current heavy-lift
vehicles are unable to launch at once. Resupplying consumables

Corresponding author at: Eduard-Grunow Str. 24b, 28203 Bremen, Germany.

Tel.: +49 17 664 738 184.
E-mail addresses: luciepoulet@yahoo.fr (L. Poulet), gioia.massa@nasa.gov
(G.D. Massa), morrowr@orbitec.com (R.C. Morrow), bourgetm@orbitec.com
(C.M. Bourget), raymond.m.wheeler@nasa.gov (R.M. Wheeler), cmitchel@purdue.edu
(C.A. Mitchell).

with a cargo vessel like the Automated Transfer Vehicle (currently

used to resupply the International Space Station) would not be
cost effective, either, since the average cost to launch a kilogram
of mass into Low-Earth Orbit has been estimated to be $10,000
(Futron Corporation, 2002). In-situ Resource Utilization might be
used to recover water and oxygen from Lunar or Mars regolith,
but it does not directly enable generation of food (ISECG, 2013).
Thus, bioregenerative life-support systems coupled to controlledenvironment food-crop growth modules are needed for food production on the Moon or Mars (Massa et al., 2006; Wheeler et al.,
2001, 2003). Such systems could sustain a crew at distant space
destinations since seeds can be stored viably for a long time and/or
(re)generated in space for very long missions.
High energy radiation on both the Moon and Mars also is a
problem for both humans and higher plants that human explorers
will have to address, most likely by using some form of radiation shielding. In addition, any pressurized structures at or near
the planetary surface must withstand tremendous differences in

http://dx.doi.org/10.1016/j.lssr.2014.06.002
2214-5524/ 2014 The Committee on Space Research (COSPAR). Published by Elsevier Ltd. All rights reserved.

44

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

temperature between night and day (from 233 C to +123 C),

large pressure differentials, and the likelihood of frequent micrometeorite impacts. Thus, human habitats and crop-growth modules very likely will be sheltered or located underground (ACCESS
Mars, 2009). Moreover, sunlight will be reduced or not available
at all times due to local conditions such as long dust storms on
Mars, the periodically increased distance between Mars and the
Sun, or extended dark periods on most locations of the Moon
(Cockell and Andrady, 1999; Cockell, 2001; Horneck et al., 2003;
Ront et al., 2003; Salisbury, 1992; Wheeler, 2004), making electric lighting a more reliable option for growing food plants in space
(Massa et al., 2006). However, it has been estimated that 40 to
50 m2 of cropping area, in continuous use, would be necessary
to fully sustain each crew member (Mitchell et al., 1997), which
would require considerable energy for a traditional electric croplighting scenario (Drysdale, 2001). Ikeda et al. showed that 45% of
the total power needed for growing lettuce in a controlled environment under uorescent lights was consumed by lamps and ballasts
and 35% by air-conditioning to reject waste heat (Ikeda, 1991).
Several ground-based bioregenerative Life-Support-System studies have been conducted since the 1970s, all including electric or hybrid (electric + solar) lighting (Gitelson et al., 1989;
Masuda et al., 2005; Nitta, 2005; Tako et al., 2010; Lasseur et al.,
2010). Experiments using the Minitron II growth-chamber/cuvette
system to determine maximum growth response for hydroponic
lettuce using red-rich incandescent lighting had an associated
power cost per unit growth area ranging from 1 to 10 kW/m2
(power density, where the area term refers to crop-growth area)
and an energy consumption per unit dry biomass produced between 953 and 1680 kWh/g (Knight and Mitchell, 1988). NASAs
Biomass Production Chamber (BPC), which also was not designed
for optimal light delivery but as a closed plant-production system with HPS lamps (Wheeler, 1992), used 2.1 kW/m2 of electrical
power for lighting (Wheeler et al., 1996), which translated into
4.7 kWh/g for Waldmanns Green lettuce (Wheeler et al., 2008).
An engineering concept for an inatable Mars surface greenhouse
estimated that a greenhouse module of 90 m2 using HPS lamps
12 h/day at 1000 mol/m2 /s would require 2.47 kW/m2 (Hublitz et
al., 2004). Using an intracanopy light-emitting diode (LED) system
to grow cowpea crop stands in a controlled environment, Massa
et al. (2005) reduced power density to 0.83 kW/m2 , which corresponded to 1.02 kWh/g of dry plant biomass. More recently, Gomez
et al. (2013) showed that intracanopy LED lighting during highwire greenhouse tomato cultivation enabled the consumption of
4.3 times less energy for supplemental lighting than with overhead HPS lamps. In recent years, LED plant research has indeed
become more active in the horticulture greenhouse industry because of potential energy savings, as indicated by the study of
Dueck et al. (2012) on hybrid supplemental lighting utilizing overhead HPS lamps and interlighting with LEDs.
LEDs also are promising candidates for space life support as
their small size, mass, and ballast-free operation would contribute
positively to reducing the Equivalent System Mass (ESM) of a space
lighting system (Drysdale and Hanford, 1999) compared to traditional high-intensity discharge (HID) lighting systems. In addition,
their solid-state electronics ensure reliability and safety (Tibbitts et
al., 1991). LED lifetime is more than twice as long as for any other
type of light source for plant growth, which would spare maintenance time for astronauts on a given mission (Bourget, 2008) and
reduce the launch mass of needed spares.
Eciencies of the Philips Luxeon LEDs per se in 2012 reached
38% for red (630 nm) emitters and 50% for blue (455 nm) (Philips
Lumileds Lighting Co., 2012). High-intensity discharge lamps such
as HPS and Metal Halide (MH) have eciencies comparable to red
LEDs, but, because of their intensely hot lamp surfaces, must be
placed much farther away from plants than LEDs (Tibbitts et al.,

1991), thereby resulting in much higher operating power required

to get sucient photosynthetic photon ux (PPF) at leaf level. Another very important advantage of LEDs is that they emit pure
colors, which can be selected to match the absorption peaks of
plant pigments (Tibbitts et al., 1991) and thus improve spectral
eciency for optimal plant growth and development (Kim et al.,
2007). Over the past decade, studies have shown that red and blue
LEDs are an effective lighting source for plant growth (Yorio et al.,
2001). Even though blue light is photosynthetically less ecient
than red light (Dougher and Bugbee, 2001; McCree, 1971/1972), it
has important photomorphogenic effects on stem elongation, leaf
expansion (Dougher and Bugbee, 2001; Hoenecke et al., 1992), and
is important for water relations (Sharkey and Raschke, 1981).
Despite the energy-saving advantages of LEDs compared to traditional crop-lighting sources, when using an overhead lighting
system for rosette plants such as leaf lettuce, light in a xedspacing growth system still is wasted falling on empty spaces
between small plants before they grow. To avoid such losses, the
concept of targeted lighting was investigated in the present study,
switching on LEDs positioned only directly above individual plants.
Changing the space between plants as they grow (variable spacing) is another option for solving this problem (Both et al., 2009;
Field, 1988; Davis, 1985; Prince and Bartok, 1978), as was done
at Phytofarms of America using automatic spacing (Prince et al.,
1981), but that solution was designed for large-scale terrestrial
agriculture and required additional energy for daily plant-position
adjustments. For space applications, simplicity of operation and
ESM minimization are always preferred, and with effective targeted
lighting, xed spacing between growing seedlings should not be
critical.
The ultimate goal of the present study was to test the concept and demonstrate the energy eciency of targeted lighting to
grow Waldmanns Green leaf lettuce under optimizing red + blue
LED lighting conditions, compared to total coverage red + blue or
white LED lighting.
To achieve this objective, a three-part study was conducted for
hydroponic lettuce in a growth chamber using red and blue LEDs
for sole-source crop lighting. First, a temporal characterization was
carried out to determine kinetics of lag and exponential phases of
growth; secondly, a determination was made regarding optimizing conditions of red:blue ratio during both growth phases; and
nally, a comparison was conducted for eciency of targeted vs.
total coverage LED lighting. Previous studies found a power density
of 2.47 kW/m2 needed to grow plants using HID lighting (Hublitz
et al., 2004), and 2.1 kW/m2 was demonstrated for HPS lighting in
the NASA Biomass Production Chamber (Wheeler et al., 1996). It
was hypothesized that such gures may be lowered by at least an
order of magnitude using targeted LED lighting technologies under
optimizing spectral conditions, as tested herein.
2. Materials and methods
2.1. ORBITEC lighting system
Two identical custom LED lighting arrays were provided by
the Orbital Technologies Corporation (ORBITEC, Madison, WI, USA).
Both rectangular lighting arrays measuring 61 61 cm were arranged in four two-by-two 27.5 27.5 cm panels, each containing
36 red LEDs (max 630 nm) and 9 blue LEDs (max 455 nm). The
red peak is 19 nm Full Width at Half Maximum (FWHM), and
the blue peak is 21 nm FWHM (Fig. 1). The red LEDs are 29%
ecient and the blue 41% (manufacturer data). Fig. 2a shows a
detailed image of the emission surface of the array. The irradiance of red and blue, as well as photoperiod, are adjustable from
custom-control software, and light intensities were measured using a spectroradiometer (Apogee) and a quantum sensor (Apogee).

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

45

70 5% during the day and 80 5% during the night. Temperature was measured continuously at a central point in the growth
chamber, but additional temperature checks were performed periodically under the two lighting systems, which conrmed that the
reported temperature of the growth chamber equated to the temperature of air under the LED arrays suspended above the plants.
2.4. Hydroponic system

Fig. 1. Spectroradiometer scan of the ORBITEC red + blue LED array. (For interpretation of the references to color in this gure legend, the reader is referred to the
web version of this article.)

Color uniformity was ensured at all times by keeping the LED

panel at a threshold distance above the crop surface. Heat is
removed from the interior of the array via air-cooled heat exchangers and ductwork (Fig. 2b). Power and energy consumption
were measured with current sensors, which were calibrated during the initial system checkout and monitored via custom software.
Power use displayed on a control computer screen is that used by
LEDs only, including LED driver circuits. The power supply is connected to an AC power source with input voltages in the range of
100240 VAC, 50/60 Hz.
2.2. FlexFire LED lighting system
A third LED lighting array was built in-house using highintensity UltraBright LED strip lights from the FlexFire LEDs Company (Mountain House, CA). It was composed of four 77-LED,
54-cm-long strips, with two 7-LED, 5-cm-long-strips between
rows. The LEDs are cool-white phosphor-based, with a Correlated
Color Temperature (CCT) of 5914 K and a Color Rendering Index
(CRI) of 69.0. There were two main peaks, one blue (448452 nm)
and one green (520525 nm) (Fig. 3). A 100 to 0% dimmer was
added so that light output could be controlled as needed. Electrical
energy and power consumption were monitored with a Kill-A-Watt
meter (P3 International Corporation, New York, NY, USA), which
measures energy consumption with an accuracy of 0.2% (manufacturer data).
2.3. Growth chamber
The growth chamber in which the experiments were conducted
is an Environmental Growth Chambers (EGC, Chagrin Falls, OH)
walk-in chamber of 9.29 m2 oor area. Temperature, relative humidity, and CO2 concentration are controlled and monitored by
EGC ControlNet software (version 4.5). When LED light was continuous (from days 1 to 3 after sowing), temperature was maintained at 25 1 C and relative humidity at 80 5%. When
plants were subsequently maintained at a 16-h photoperiod on a
24-h light/dark cycle, temperature was 25 1 C during the day
and 20 1 C during the night, and relative humidity was kept at

A deep-batch-recirculating hydroponics system was used consisting of a tub that served as a root compartment (35 L) with
a 0.372 m2 polystyrene foam lid that served as a growing surface, both mounted above a large reservoir (70 L) and pump that
continuously recirculated nutrient solution between the reservoir
and root compartment (Frantz et al., 2000). Active root-zone aeration was practiced within the root compartment using a bubbling
wand and external aquarium pump. The nutrient solution used was
Hoaglands no. 1 (Hoagland and Arnon, 1950). During the rst 10
days of growth, quarter-strength nutrient solution was used, keeping solution conductivity at 0.6 dS/m. From day 11, concentration
of the nutrient solution was incrementally increased on a daily basis until it reached full strength with a conductivity of 2.4 dS/m
on day 18 after planting. The pH was maintained between 5.4 and
5.9. Conductivity and pH were monitored and adjusted as needed
every other day during the rst 10 days of crop development, and
then every day during the last 8 to 10 days of growth. This was
done to avoid water status and mineral nutrition becoming variables affecting plant growth. Conductivity was adjusted by adding
more nutrient solution concentrate if the EC value was too low or
by adding water if the EC value was too high; pH was adjusted by
titrating with 0.1 N H2 SO4 if too high or 0.1 N KOH if too low.
2.5. Plant material and experimental design
The test species used was Lactuca sativa L. cv. Waldmanns
Green leaf lettuce, provided by High Mowing Organic Seeds (Wolcott, VT 05680), which was grown hydroponically in the manner of Frantz et al. (2000). The growing surface was composed
of 2-cm-thick extruded polystyrene insulation FOAMULAR (Owens
Corning, Toledo, OH, USA) sheet in which holes were drilled to
insert closed-cell polyethylene foam plugs (Log Home Center Inc.,
Noblesville, IN, USA). Pregerminated seedlings were placed and
oriented root-down within a rolled polyester wick, which was inserted into a slitted polyethylene foam plug, and then inserted into
a pre-drilled hole in the polystyrene foam sheet.
Overhead LED lighting arrays were mounted above separate hydroponics systems in the growth chamber by a series of ropes
and pulleys so that height could be adjusted and the arrays leveled just above the growing surface on which lettuce plants were
grown. CO2 level was uncontrolled at ambient levels during the
rst 8 days of seedling development. From day 8, CO2 level was
set at 700 50 ppm and was increased by 100 ppm at the beginning of each photoperiod until it reached 1000 50 ppm.
Growth data collected immediately following harvest included, on
a per plant basis, number of leaves, hypocotyl length, leaf area
(Model LI-3000: LAMBDA Instruments Corporation, USA), shoot
fresh and dry masses, and root dry mass. Lettuce was dried for
55 7 hours in a forced-air oven (Model Blue M: Blue island, IL,
USA) at 70 5 C. Fresh and dry masses were obtained using an
electronic balance (Model 1219 MP: Sartorius GmbH, Gttingen,
Germany).
Preliminary experiments determined that yield was signicantly
enhanced by having a reective, enclosed growth compartment
(pink hydroponic growth surface together with reective white
plastic walls) to prevent light from escaping the growth compartment vs. using a at-black growth surface and no wall lm.

46

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

(a)

(b)
Fig. 2. ORBITEC LED array: (a) Bottom view of the combined four panels, each containing nine LED clusters with each cluster composed of four red LEDs and one blue LED,
and zoom on one panel showing the position of red and blue LEDs. (b) Top view showing ductwork for fan-driven heat rejection from the back of the LED array.

Fig. 3. Spectroradiometer scan of the home-made FlexFire White LED array. Source: FlexFireLEDs.

2.6. Experiment 1: establishing lettuce-growth-optimizing red:blue

ratio at low irradiance
Experiments were conducted to determine optimizing red:blue
ratio for establishment of viable seedlings during the lag phase of
growth (duration estimated from preliminary experiments). Sample number per treatment was 8 plants grown for 10 to 14 days
from seed to harvest. For germination, 30 seeds were sown on
double-layer Whatmann number 1 lter paper in a Petri dish wetted with 2 mL of tap water, imbibed for 6 h under dim uorescent
light (15 mol/m2 /s) at 22 C, followed by 24 h under dim 630-nm
red LED light (about 2 mol/m2 /s) at 20 0.5 C (8 h) and 25
0.5 C (16 h). After transplantation to hydroponics in the growth
chamber, seedlings were kept for 6 h in darkness, followed by
2 days under continuous treatment light at 100 5 mol/m2 /s.
Then the photoperiod was set to 16 h. The lighting schedule is
graphically described in Fig. 4. Since plants during lag phase did
not response to high light, the goal was to keep seedlings healthy
while saving energy, and thus low light intensities were used in

this experiment. PPF was increased by 25 mol/m2 /s daily until it

reached 225 10 mol/m2 /s. The LED panel was maintained 6 cm
above the growing surface throughout light treatments.
During the rst two days after germination, light was continuous. The different light treatments during these 2 days had the
following red:blue PPF ratios: 100:0, 95:5, or 90:10. After this period, each treatment had a red:blue ratio of 95:5 or 90:10, as
summarized in Fig. 5.
2.7. Experiment 2: growth curve
For the growth curve experiment, 120 plants were split evenly
between the two growing systems. The experiment lasted for
26 days from seed to nal harvest. During the rst 11 days in
the growth chamber, plants in both systems were maintained under the same lighting treatment, with a red:blue ratio of 95:5.
On day 12, one system was switched to 90:10, whereas the other
system was kept under 95:5. From day 5, four plants per system
were harvested every 2 days, and then harvested every day for

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

47

Fig. 4. Detailed lighting schedule for experiment 1 (red:blue ratio experiment). The x-axis indicates days of treatment, while the y-axis displays PPF in mol/m2 /s.

Fig. 5. Six different red:blue ratio treatments tested during the lag phase of lettuce growth. The x-axis indicates days of treatment, while the y-axis displays treatment
identication.

Fig. 6. Detailed lighting schedule for experiment 2 (growth curve experiment). The x-axis indicates days of treatment, while the y-axis displays PPF in mol/m2 /s.

48

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

Fig. 7. Detailed lighting schedule for experiment 3 (targeted vs total coverage experiment). The x-axis indicates days of treatment, while the y-axis displays PPF in mol/m2 /s.

the last 8 days. These regular harvests required plant-respacing

adjustments, so that plants would always be equally spaced over
the growing area. Plants were grown as described above with the
exception that, since plants were grown longer than in experiment 1 (red:blue ratio), PPF was increased by 25 mol/m2 /s daily
until it reached 350 10 mol/m2 /s to accommodate their higher
light-intensity needs. The lighting schedule is graphically described
in Fig. 6.
During the rst 11 days in the growth chamber, plants in both
systems were maintained under the same lighting treatment, with
a red:blue ratio of 95:5. On day 12, one system was switched to
90:10, whereas the other system was kept under 95:5.
2.8. Experiment 3: targeted red + blue vs. total coverage red + blue vs.
total coverage white
Under optimizing conditions determined from the growth-curve
experiment and the red/blue-ratio experiments described above,
the eciency of targeted red + blue LED lighting was compared
to that of a system in which all red and blue LEDs in the array were energized during the photoperiod of a crop-production
cycle as well as a lighting array with all white LEDs energized (total coverage). Targeted LED lighting refers to visual estimation
of overhead LED position relative to plant position and plant size.
At the beginning of each photoperiod, only LEDs that were located
directly above individual plants were manually energized, and irradiance was adjusted daily using the ORBITEC lighting-system controller.
This experiment was replicated three times for 21 days from
seed to harvest each time. Plants were grown as described for
experiments 1 and 2 with the following exceptions: after newly
transplanted seedlings were kept for 6 h in the dark, they were
then maintained for 1 day under continuous light of 75
5 mol/m2 /s and one day at 85 5 mol/m2 /s; PPF was increased
by 25 mol/m2 /s daily until it reached 350 10 mol/m2 /s. The
lighting schedule is graphically described in Fig. 7. This difference
in light intensity compared to experiment 1 comes from the fact
that plants were grown longer (lag plus exponential phases), thus
requiring higher light intensities post-lag phase. The red:blue ratio
used for both treatments was 95:5.
Sixteen plants were mounted on each 0.372 m2 growth board,
giving a xed plant density of 43 plants/m2 . Plants were assembled

Fig. 8. Diagram to scale representing plant spacing on growing lid for experiment 3.
Circles represent where a plant was placed. All dimensions are in cm.

in 4 clusters of 4 plants and spacing between them did not change

during the course of an experiment. Each plant in a cluster was
separated from the other by 10 cm, and clusters were separated
from each other by 15 cm, due to the lighting-array conguration,
which is separated in four identical panels (Fig. 8).
2.9. Statistical analysis
Statistical analysis was performed using SAS software (SAS 9.2).
To perform analysis of variance (ANOVA) within treatments, the
ANOVA procedure and Duncans multiple range test were used. For
experiment 3, data were pooled and treated as one sample with
a larger N number and the assumption was made for complete
randomization of treatments.

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

49

Table 1
Growth responses of lettuce (per seedling) subjected to different red:blue ratios of LED lighting for the rst 9 days of lag-phase seedling development following transplanting.
Treatments I and III had a constant red:blue ratio throughout the experiment. Treatments II, IV, V and VI were split into 2 parts: 2 days of continuous light at a given red:blue
ratio and the remaining 8 days at another red:blue ratio and 16 h photoperiod. N = 8 plants per treatment. Comparisons made within rows. Different letters within rows
indicate signicant treatment differences at P < 0.05. SLA is specic leaf area.
Shoot growth
parameters

95R:5B
(Treatment I)

100R:0B then
95R:5B
(Treatment II)

90R:10B then
95R:5B
(Treatment III)

90R:10B
(Treatment IV)

100R:0B then
90R:10B
(Treatment V)

95R:5B then
90R:10B
(Treatment VI)

Shoot fresh mass (mg)

Shoot dry mass (mg)
Hypocotyl Length (cm)
Leaf area (cm2 )
SLA (cm2 /g)
% Dry Mass

174 a
23 a
0.76 b, c
7.0 a
327 a
13.0 a

109 c
17 a
1.3 a
5.4 b, c
366 a
15.6 a

139 b
19 a
0.83 b
5.4 b, c
280 a
14.2 a

133 b
18 a
0.61 c
5.2 c, d
306 a
14.9 a

114 c
17 a
1.3 a
4.5 d
314 a
15.2 a

160 a
20 a
0.65 b, c
6.2 b
340 a
12.7 a

3. Results
3.1. Experiment 1: establishing lettuce-growth-optimizing red:blue
ratio at low irradiance
Preliminary experiments (results not presented), compared effects of 5, 10, 15, or 20% blue light at low irradiance on lettuceseedling development during the lag phase of seedling development. Those experiments indicated that seedlings grown under
low-irradiance red light (<20 mol/m2 /s) during the lag phase led
to stem breakage and seedling death. A consistent trend observed
from preliminary experiments was that lag-phase plants developed
shorter, stronger stems with increasing proportions of blue light,
but above a certain level of blue, leaf expansion was inhibited.
Results for the lag-phase red:blue-ratio experiment 1 are summarized in Table 1. The highest seedling fresh and dry masses, plus
leaf area, were achieved with treatment I (red:blue ratio of 95:5).
Treatment VI (red:blue ratio of 95:5 initially, then 90:10) gave similar results (not signicantly different) for fresh and dry masses.
Minimum hypocotyl length occurred for treatment III (red:blue ratio of 90:10). Treatments II and IV (continuous red light for the
rst two days) gave similar results: same hypocotyl length (significantly different from other treatments) and dry mass, and similar
fresh mass. Specic leaf area and percentage of shoot dry matter
were not signicantly different between treatments. Similar results
for fresh and dry mass, percentage of dry matter, and hypocotyl
length were found for plants that grew with the same light treatment during the rst two days (red:blue ratios of 100:0, 95:5 or
90:10), independently of the treatment they were under for the
rest of the experiment.
3.2. Experiment 2: growth curve
Regression analysis was used to determine transition time between lag and exponential growth phases for four important metrics of plant productivity, including leaf area, shoot fresh mass,
shoot dry mass, and root dry mass (Fig. 9a, b, c, d). Given the signicantly different results found for lag-phase testing of different
sole-source LED red:blue ratios (experiment 1), a difference also
was expected for exponential-phase growth metrics between the
two treatments tested (95:5 vs. 90:10), but treatments were not
statistically different for any metric. Thus, growth data were pooled
and only one regression curve is presented per variable.
For each growth parameter measured, two growth phases were
identied, including a lag phase, when growth was very slow, followed by an exponential phase when growth accelerated at an
increasing rate. The beginning of exponential growth corresponds
to intersection of the horizontal asymptote of the lag phase and
the oblique asymptote of the exponential phase. On average for
all variables, exponential growth began on day 15 for the environmental conditions of these experiments. Energy expended for

lighting per unit dry mass of edible lettuce parts during lag phase
was 3.22 kWh/g, but was only 0.40 kWh/g during exponential
growth (Table 2). Specic leaf area during lag phase averaged
331 cm2 /g versus 201 cm2 /g during exponential phase. Percentage
dry matter in the shoot (ratio between dry mass and fresh mass)
was 45% larger during exponential phase (6.84 %) than during lag
phase (4.67 %).
3.3. Experiment 3: targeted R:B vs. total coverage R:B vs. total coverage
white light for the entire production cycle
Three replicate experiments were performed and results are
summarized in Tables 3a and 3b. Hypocotyl length was signicantly shorter for lettuces grown under total coverage white LEDs.
Leaf and whole-plant dry masses and leaf area were statistically
larger for the total coverage red + blue treatment than for the
other two treatments. On average, total and edible biomass produced by this treatment also was higher. However, percent dry
matter in the shoot was highest for lettuces grown under the targeted red + blue treatment and statistically higher than that of
lettuce grown under the total coverage white treatment.
The total coverage red + blue LEDs used 2.5 times more absolute energy and almost twice as much energy per unit dry biomass
accumulated than did the targeted red + blue LEDs; the total coverage red + blue LEDs also used 2.2 times more absolute energy
and 1.5 times more energy per unit dry biomass than did the white
LEDs. Crop productivity per unit area was highest for the total coverage red + blue LEDs with 13 to 16 g/m2 more than for the two
other treatments, that is to say 32% greater than targeted red +
blue and 50% greater than total coverage white.
4. Discussion
4.1. Red:blue ratio experiments
Light spectrum (i.e., red:blue ratio) was addressed for lag phase
to obtain normal morphogenesis while minimizing light intensity,
which otherwise causes fatal hypocotyl elongation and insucient
leaf expansion. The results revealed that lettuce grew best under
a red:blue ratio of 95:5 giving the better fresh and dry mass and
leaf area. Our results agree with the ndings of Hoenecke et al.
(1992) and Dougher and Bugbee (2001) that stem length of lettuce
decreases with increasing percentage of blue light. The trend found
by Dougher and Bugbee (2001) that leaf area of lettuce decreases
with increasing blue light did not appear clearly in our results.
However, treatments IV and V (respectively 90R:10B and 100R:0B
followed by 90R:10B), with the largest percentages of blue light,
also had the smallest leaf area.
Shoot fresh mass and hypocotyl length were similar and not
statistically different between pairs of treatments that had in common their rst two days of growth but different red:blue ratios from days 2 to 9. Light quality during the rst 2 days of

6.84

growth thus seems critical for seedling establishment and future

plant developmental characteristics. Continuous red light during
the rst 2 days did not result in acceptable growth, especially
because hypocotyls became too long, conrming that sole-sourceLED-lighted lettuce needs a minimum amount of blue light early
on for healthy seedling development.

Specic leaf area

(cm2 /g)

331

201

Conversion
eciency1
(g/kWh)

0.35

2.78

4.2. Growth curve

This study allowed us to determine precisely when the lag
phase of seedling growth ends and when the exponential phase
starts, during which time lettuce seedlings are most responsive to
optimizing environments. There actually are more phases, but the
experiment was intentionally stopped before plateau and senescence phases because the aim was to characterize the lighting systems during the most active phase of growth, and phases beyond
exponential are not productive in terms of biomass accumulation.
Determining the inection point between the two phases had two
direct consequences on subsequent experiments:

74.44
0.40

0.36

5.96
3.22

2.85

incrementally beginning on the appropriate day for optimal

use by lettuce during its exponential growth phase.
Energy expenditure for each phase of growth could be determined in terms of kWh per gram dry mass of edible biomass
produced.
The exponential phase of crop growth started on day 15, and
the crop was terminated before an inection point occurred curving over to plateau phase. Total dry biomass accumulation during
this phase was 12.5 times more than during lag phase, which also
is reected in the percentage of dry matter in the shoot biomass
(6.84% for exponential phase versus 4.67% for lag phase).
Absolute energy consumption was least during lag phase
(17 kWh versus 26.8 kWh during exponential phase), which was
predictable. However, energy-biomass conversion eciency during exponential growth was 8-fold higher than that during the
lag phase (2.78 g/kWh versus 0.35 g/kWh). This stresses the importance of lag-phase light-use optimization to further increase
overall energy savings over an entire cropping cycle.
One objective of experiment 3 (targeted vs. total coverage lighting) was to minimize energy consumption during the lag phase of
growth, when plants are small, by focusing incident light on plant
material and not wasting it on empty space.

66.25
26.8

Column 6 is the inverse of column 5.

1

5.28
17

4.3. Targeted red + blue vs. total coverage red + blue vs. total coverage
white

Lag phase
(days 1 to 14)
Exponential phase
(days 15 to 25)

Edible biomass
accumulation (g)

Energy spent per

unit edible
biomass
(kWh/gedible )

Total biomass
accumulation (g)

Energy spent per

unit of biomass
produced
(kWh/gtotal )

Nutrient-solution strength and light level could be increased

Total energy
consumption
(kWh)

Table 2
Energy consumption and lettuce dry mass accumulation Experiment 2 (Growth Curve). Data pooled for both light treatments (red:blue ratio of 95:5 and 90:10). N = 120.

4.67

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

% Dry matter in
the shoot

50

The shortest hypocotyls occurred for the lighting treatment using white LEDs (Table 3a), which have a rich blue component
of approximately 40% (Fig. 3). This conrms previous ndings of
Dougher and Bugbee (2001), who found a decrease of 72% in lettuce stem length when increasing blue light from 0 to 2%, and a
further 13% decrease when increasing blue-light percentage from
2 to 6%. The blue light used in treatments using red + blue LEDs
accounted for 5% of total PPF. Similar trends were found for the
specic leaf area (SLA) index. It was highest for the targeted red +
blue treatment, though not signicantly different from other treatments, and lowest for the total coverage white LED treatment. This
also reinforces the ndings of Dougher and Bugbee (2001), which
showed that specic leaf area of lettuce decreased with increasing
blue-light fraction from 5 to 25%.
The targeted red + blue treatment was the most ecient of
the three treatments: 20% more biomass produced per kWh than
with total coverage white; energy consumption per unit dry mass

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

51

Fig. 9. Leaf area (a), root dry mass (b), shoot fresh mass (c), and shoot dry mass (d) of lettuce plants over 25 days of growth, average per plant. Circles represent data for
treatment 1 (90 R:10 B) and triangles represent data for treatment 2 (95 R:5 B). The line is the regression curve.

Table 3a
Growth responses of lettuces (per plant) subjected to three different lighting treatments (Targeted R + B, Total coverage R + B, and Total coverage White) Experiment 3,
cumulative growth after 21 days of treatment N = 16 in each treatment comparisons are within rows. Different letters within rows indicate signicant treatment
differences at P < 0.05.
Growth parameter

Targeted R + B

Total coverage R + B

Total coverage White

Hypocotyl length (cm)

Leaf area (cm2 )
Leaf dry mass (g)
Root dry mass (g)
Shoot/root ratio
Whole plant dry mass (g)
SLA (cm2 /g)
% Dry matter in the shoot

0.50 a
410.56 b
0.82 b
0.12 a
7.17 a, b
0.97 b
514.09 a
6.9 a

0.53 a
510.53 a
1.06 a
0.15 a
7.98 a
1.28 a
508.49 a
6.2 a, b

0.14 b
345.47 b
0.70 b
0.12 a
6.35 b
0.85 b
502.09 a
6.1 b

Table 3b
Relationship between total biomass of plants subjected to three different lighting treatments (Targeted R + B, Total coverage R + B, and Total coverage White) and their
respective energy and power consumptions N = 16 Experiment 3, cumulative growth after 21 days of treatment comparisons are within rows.
Growth parameter

Targeted R + B

Total coverage R + B

Total coverage White

Total edible dry biomass (g)

Total dry biomass (g)
Total energy consumption (kWh)
Energy per unit edible biomass1 (kWh/g)
Energy per unit total biomass2 (kWh/g)
Conversion eciency (g/kWh)
Power per unit area3 (W/m2 )
Dry biomass per unit area (g/m2 )

13.59
15.48
9.6
0.71
0.62
1.61
82.2
41.6

18.06
20.41
23.6
1.31
1.16
0.86
202
54.9

11.74
13.64
10.8
0.92
0.79
1.27
92.5
36.7

1
2
3

The edible biomass is the biomass of the leaves.

The total biomass is the biomass of the shoot and the root together.
Power averaged on lighted periods.

52

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

(a)

(b)
Fig. 10. ORBITEC LED lighting system providing close-canopy overhead lighting of a young hydroponic lettuce crop (a) with photons falling on empty spaces between plants
in a total coverage lighting scenario testing a 95:5 red:blue ratio, (b) with photons falling only on plants in a targeted lighting scenario testing a 90:10 red:blue ratio. (For
interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

of lettuce was reduced by almost half compared to the total coverage red + blue. Light falling between plants on a growing surface
is wasted because it is not used eciently by plants. This is particularly true when small plants are in the lag phase and thus widely
spaced (6 cm) apart from each other (Figs. 10a and 10b). The excess light of the total coverage R + B treatment, however, had a
positive effect on crop productivity, which was 21% higher than
for the other two treatments. The most likely explanation is reectance of PAR off of the pink polystyrene growth surface onto
the leaves of nearby plants, thereby enhancing their photosynthesis.
The energy expenditure per gram of dry biomass produced with
the total coverage white LEDs (0.79 kWh/g) was not far from that
of the targeted red + blue (0.62 kWh/g), and it was 32% lower than
for the total coverage red + blue LEDs. This eciency likely is because the white LED technology was 3 years newer than that of the
red and blue LEDs used in the present study. Another possibility is
that broad-band (white) light contains wavelengths promoting certain photomorphological characteristics affecting productivity in a
positive way that neither red nor blue LEDs alone can provide. An
interesting follow-on study would be to include a targeted whiteLED treatment of this same technology.
For the three LED lighting treatments tested, energy expenditure per unit dry biomass formed was about 80% lower than that
reported by Wheeler et al. (1996, 2008) for Waldmanns Green
lettuce in the BPC, which consumed 4.7 kWh/g dry biomass. The
targeted red & blue LEDs spent 39% less energy per unit dry
biomass accumulated and the white LEDs 23% less than the solesource intracanopy (vertical) red + blue LEDs reported by Massa et
al. (2005), which consumed 1.02 kWh/g dry cowpea biomass. The
total coverage red + blue LED treatment, however, used 14% more
energy per unit dry biomass accumulated than what Massa et al.
(2005) reported. This leads to two conclusions:

Targeted lighting enables signicant savings of energy expenditure.

Ongoing progress in LED technology should further reduce the

cost of lighting in coming years.
The three LED lighting treatments compared here (targeted red

+ blue, vs. total coverage red + blue or white) used at least 90%
less power per unit growing-surface area than did the HPS lamps
of the BPC, which consumed 2.1 kW/m2 (Wheeler et al., 1996).

5. Conclusions
Regardless of light treatment, the types of LEDs used in the
present study performed better in terms of energy savings than in
the previous lighting experiments cited, with power-use per unit
growing-surface area reduced by at least an order of magnitude. It
should be noted, however, that previous lighting studies were not
designed for energy-use optimization but rather to explore maximum limits of crop productivity.
The ndings of the present investigation reinforce that targeted lighting coupled to ecient LED technology plus optimization of red/blue ratio are key to reducing the overall energy cost
of growing food crops in controlled environments, including those
intended for life support in space.
Acknowledgements
The authors would like to thank Judith Santini for help with
statistical analysis of the data, as well as Cannon Cheng, Michael
Dzakovich, Celina Gomez, David Lotz, and Miranda Smith for their
help in experiment management.
References
ACCESS Mars, Al Husseini, A., Apeldoorn, J., Ashford, Jr. K., Bennell, K.M., de Carufel,
G., Haider, O., Hirmer, T., Jivnescu, I.E., Pegg, C.J., Shaghaghi Varzeghani, A., Shimmin, R., Suresh, R., Mar Vaquero Escribano, T., Vargas Muoz, M., Winetraub, Y.,
Gallardo, B., Fernndez-Dvila, A., Zavaleta, J., Laufer, R., 2009. The role of caves
and other subsurface habitats in the future exploration of Mars. IAC-09-A5.1.4.
Both, A.J., Albright, L.D., Langhans, R.W., 2009. Design of a demonstration greenhouse operation for commercial hydroponic lettuce production. ASAE paper
No. 994123, 12 pp.
Bourget, C.M., 2008. An introduction to light-emitting diodes. HortScience 43 (7).
Cockell, C.S., 2001. The Martian and extraterrestrial UV radiation environment.
Part II: Further considerations on materials and design criteria for articial
ecosystems. Acta Astronaut. 49, 631640.
Cockell, C.S., Andrady, A.L., 1999. The Martian and extraterrestrial UV radiation environment 1, Biological and closed-loop ecosystem considerations. Acta Astronaut. 44, 5362.
Davis, N., 1985. Controlled-environment agriculture: past, present, and future. Food
Technol. 39 (10), 124126, 134.
Dougher, T.A.O., Bugbee, B., 2001. Differences in the response of wheat, soybean and
lettuce to reduced blue radiation. Photochem. Photobiol. 73 (2), 199207.
Drysdale, A.E., 2001. Life support trade studies involving plants. SAE Technical,
2001-01-2362.
Drysdale, A.E., Hanford, A.J., 1999. Advanced life support research and technology
development metric-baseline, NASA, Lyndon B. Johnson Space Center, Houston,
TX. CTSD-ADV-384, JSC 39503.

L. Poulet et al. / Life Sciences in Space Research 2 (2014) 4353

Drysdale, A.E., Ewert, M.K., Hanford, A.J., 2003. Life support approaches for Mars
missions. Adv. Space Res. 31 (1), 5161.
Dueck, T.A., Janse, J., Eveleens, B.A., Kempkes, F.L.K., Marcelis, L.F.M., 2012. Growth of
tomatoes under hybrid LED and HPS lighting systems. Acta Hort. 952, 335342.
Field, R., 1988. The green machine: indoor farming, Discover, pp. 4652.
Frantz, J.M., Mitchell, C.A., Joly, R.J., 2000. Intracanopy lighting inuences radiation
capture, productivity, and leaf senescence in cowpea canopies. J. Am. Soc. Hort.
Sci. 125 (6), 694701.
Futron Corporation, 2002. Space transportation costs: trends in price per
pound to orbit 19902000. http://www.futron.com/upload/wysiwyg/Resources/
Whitepapers/Space_Transportation_Costs_Trends_0902.pdf. pp. 17.
Gitelson, J.I., Terskov, I.A., Kovrov, B.G., Lisoviski, G.M., Okladnikov, Yu.N., Sidko, F.Ya.,
Tubachev, I.N., Shilenko, M.P., Alekseev, S.S., Pankova, I.M., Tirranen, L.S., 1989.
Long-term experiments on mans stay in biological life-support system. Adv.
Space Res. 9 (8), 6571.
Gomez, C., Morrow, R.C., Bourget, C.M., Massa, G.D., Mitchell, C.A., 2013. Comparison
of intracanopy light-emitting diode towers and overhead high-pressure sodium
lamps for supplemental lighting of greenhouse-grown tomatoes. HortTechnology 23 (1), 9398.
Hoagland, D.R., Arnon, D.I., 1950. The water-culture method for growing plants
without soil. Circ. - Calif. Agric. Exp. Stn., 347 (rev.).
Hoenecke, M.E., Bula, R.J., Tibbitts, T.W., 1992. Importance of blue photon levels for
lettuce seedlings grown under red light-emitting diodes. HortScience 27 (5),
427430.
Horneck, G., Facius, R., Reitz, G., Rettberg, P., Baumstark-Khan, C., Gerzer, R., 2003.
Critical issues in connection with human missions to Mars: protection of and
from the Martian Environment. Adv. Space Res. 31, 8795.
Hublitz, I., Henninger, D.L., Drake, B.G., Eckart, P., 2004. Engineering concepts for
inatable Mars surface greenhouses. Adv. Space Res. 34, 15461551.
Human exploration destination systems roadmap, Technology Area 07, NASA
space technology roadmaps, April 2010. http://www.nasa-usa.de/pdf/
501327main_TA07-ID_rev7_NRC-wTASR.pdf.
Human health, life support and habitation systems, Technology Area 06,
NASA space technology roadmaps, April 2012. http://www.nasa-usa.de/pdf/
500436main_TA06-ID_rev6a_NRC_wTASR.pdf.
Ikeda, A., 1991. Environmental control and operation monitoring in plant factoring
using articial light. In: Proc. First International Workshop on Sensors in Horticulture.
International Space Exploration Coordination Group (ISECG), The global exploration
roadmap, 2013. http://www.globalspaceexploration.org.
Kim, H.H., Wheeler, R.M., Sager, J.C., Norikane, J., Yorio, N.C., 2007. Electric lighting
considerations for crop production in space. In: ISHS 2007. Acta Hort. 761.
Knight, S.L., Mitchell, C.A., 1988. Effects of CO2 and photosynthetic photon ux on
yield, gas exchange, and growth rate of Lactuca Sativa L. Waldmanns Green.
J. Exp. Bot. 39 (200), 317328.
Lasseur, C., Brunet, J., De Weever, H., Dixon, M., Dussap, G., Godia, F., Leys, N.,
Mergeay, M., Van Der Straeten, D., 2010. MELiSSA: the European project of
closed life support system. Gravit. Space Biol. 23 (2).
Massa, G.D., Mitchell, C.A., Jeffrey, G.D., Morrow, R.C., 2005. Development of a recongurable LED plant-growth lighting system for equivalent system mass reduction in an ALS. SAE International, 2005-01-2955.

53

Massa, G.D., Jeffrey, G.D., Emmerich, C., Morrow, R.C., Bourget, C.M., Mitchell, C.A.,
2006. Plant-growth lighting for space life support: a review. Gravit. Space
Biol. 19 (2).
Masuda, T., Arai, R., Komatsubara, O., Tako, Y., Harashima, E., Nitta, K., 2005. Development of a 1-week cycle menu for an advanced life support system (ALSS)
utilizing practical biomass production data from the closed ecology experiment
facilities (CEEF). Habitation 10, 8797.
McCree, 1971/1972. The action spectrum, absorptance and quantum yield of photosynthesis in crop plants. Agric. Meteorol. 9, 191216.
Mitchell, C.A., Dougher, T.A.O., Nielsen, S.S., Belury, M.A., Wheeler, R.M., 1997. Costs
of providing edible biomass for a balanced vegetarian diet in a controlled ecological life-support system. In: Suge, H. (Ed.), Plants in Space Biology. Tohoku
University Press, Sendai, Japan, pp. 245254.
Nitta, K., 2005. The Mini-Earth facility and present status of habitation experiment
program. Adv. Space Res. 35, 15311538.
Philips Lumileds Lighting, Co., 2012. LUXEON Rebel Color Portfolio. Datasheet
DS68:20120421. Philips Lumileds Lighting, Co., Amsterdam, The Netherlands.
Prince, R.P., Bartok Jr., J.W., 1978. Plant spacing for controlled environment plant
growth. Trans. ASAE 21 (2), 332336.
Prince, R.P., Bartok Jr., J.W., Protheroe Jr., D.W., 1981. Lettuce production in a controlled environment plant growth unit. Trans. ASAE 24 (3), 725730.
Ront, G., Brces, A., Lammer, H., Cockell, C.S., Molina-Cuberos, G.J., Patel, M.R., Selsis, F., 2003. Solar UV irradiation conditions on the surface of Mars. Photochem.
Photobiol. 77, 3440.
Salisbury, F.B., 1992. Some challenges in designing a Lunar, Martian, or microgravity
CELSS. Acta Astronaut. 27, 211217.
Sharkey, T.D., Raschke, K., 1981. Effect of light quality on stomatal opening in leaves
of Xanthium strumarium L. Plant Physiol. 68, 11701174.
Tako, Y., Arai, R., Tsuga, S., Komatsubara, O., Masuda, T., Nozoe, S., Nitta, K., 2010.
CEEF: closed ecology experiment facilities. Gravit. Space Biol. 23 (2), 1324.
Tibbitts, T.W., Bula, R.J., Morrow, R.C., Barta, D.J., 1991. Light-emitting diodes as a
radiation source for plants. HortScience 26 (2), 203205.
Wheeler, R.M., 1992. Gas-exchange measurements using a large, closed plant growth
chamber. HortScience 27, 777780.
Wheeler, R.M., 2004. Horticulture for Mars. Acta Horticult. 642, 201215.
Wheeler, R.M., Mackowiak, C.L., Stutte, G.W., Sager, J.C., Yorio, N.C., Ruffe, L.M., Fortson, R.E., Dreschel, T.W., Knott, W.M., Corey, K.A., 1996. NASAs biomass production chamber: a testbed for bioregenerative life support studies. Adv. Space
Res. 18 (4/5), 215224.
Wheeler, R.M., Stutte, G.W., Subbarao, G.V., Yorio, N.C., 2001. Plant growth and human life support for space travel. In: Pessarakli, M. (Ed.), Handbook of Plant and
Crop Physiology. 2nd edition, pp. 925941.
Wheeler, R.M., Sager, J.C., Prince, R.P., Knott, W.M., Mackowiak, C.L., Stutte, G.W.,
Yorio, N.C., Ruffe, L.M., Peterson, B.V., Goins, G.D., Hinkle, C.R., Berry, W.L.,
2003. Crop production for advanced life support systems Observations from
the Kennedy Space Center breadboard project. NASA Technical Memorandum
TM-2003-211184. NASA Kennedy Space Center, pp. 125.
Wheeler, R.M., Mackowiak, C.L., Stutte, G.W., Yorio, N.C., Ruffe, L.M., Sager, J.C.,
Prince, R.P., Knott, W.M., 2008. Crop productivities and radiation use eciencies for bioregenerative life support. Adv. Space Res. 41, 706713.
Yorio, N.C., Goins, G.D., Kagie, H.R., Wheeler, R.M., Sager, J.C., 2001. Improving
spinach, radish, and lettuce growth under red light-emitting diodes (LEDs) with
blue light supplementation. HortScience 36, 380383.