I’m Bruce Bugbee. I’m a professor of crop
physiology at Utah State University, and this is a seminar about calculating potential yield
of any agricultural commodity. It was first developed for fields like this and for outdoor
agriculture. But over the years we have modified this model, which is a series of calculations,
to get potential yield in indoor agriculture. And as we get better and better with indoor
agriculture, it becomes increasingly important to understand what the potential is. In this
talk, in the next 20-25 minutes, I’m going to just derive the potential yield from what
we call first principles. But first I want to acknowledge the funding agencies for this,
and NASA is a key funding agency. They’ve funded the research in my lab for over 30
years, and we’ve been deeply indebted to NASA for allowing us to keep refining these models.
Recently, we also got funded by the USDA. And this is a project, much more applied project,
to look at the use of supplemental lighting in controlled environments. This project is
called the LAMP project, “Lighting Approaches to Maximize Profits.” So this is a very
basic research project. This is a very applied research project. And back-and-forth, we zero
in on the reality of what we can get for food production in controlled environments. So
let’s now show you the units of this calculation. One gram per mole of photons. This is an output
from the system. One gram of dry biomass or one gram of yield per mole of photosynthetic
photons. So its output divided by input, which is the efficiency of the system. Now just
quickly I can show you how we derive this… and let’s make sure we’re on black… one
gram, let’s put it right over here, a gram per meter squared of ground area per day divided
by a mole of photons per meter squared of ground area per day. And you can see, the
meter squares, the days cancel out, and we get grams per mole. This mole per meter squared
right here is called the daily light integral. DLI. It’s a common measurement, moles per
meter squared per day. This is just biomass. This is tricky calculating the exact area
because we get guard row effects. But these are the input parameters to this. Now, where
do these come from and how can we calculate yield? I want to start by examining what to
any plant biologist is the most important equation in the world. Now, when I show this
to physicist colleagues, they’re pretty sure the most important equation in the world is
E=mc². What do you think this is? The most important equation in the world, particularly
to any life scientist? Here it is. Equations are inputs, and they go to outputs. Now this
is an every single high school biology textbook. It’s in every textbook in college. We have
20 courses on college campuses that emphasize this equation. Of course it helps to know
what this is. That is carbohydrate, CH₂O. And this equation, by now you’ve figured it
out, this is photosynthesis. So now we look at this. This takes a low… let’s get a different
color here… This takes a low-energy gas, CO₂, we use it to put out fires. And water.
Two really simple molecules. Very low energy. And it makes a high-energy solid food and,
well, O₂ is rocket fuel. So we get two very high energy outputs from two very low energy
inputs. We often we don’t think about that a lot. We make students memorize the stoichiometry,
but the energetics of this is what’s really important. Now, remember the first law of
thermodynamics says energy in equals energy out. Everything has to follow the first law.
How do we get two really low input molecules to go to two high input molecules? The answer
is in more advanced books, light. And particularly photons. And particularly 10 photons. This
is close to the theoretical minimum number of photons to push this equation across. So
how are we going to derive grams per mole of photons from this 10 photons? Every book
has this in it. The advanced books have the photons. This has always disturbed me that
it’s written like this because it doesn’t capture for our students the magnificence
of this equation and how you get these high energy outputs. So I would like to propose
that we revise the way we write this equation. Here’s the earth… BOOM! Here is our low
energy inputs, and here is our high energy outputs. And it’s many steps to get up there.
And now how do you get uphill from low energy to high? We bring in those 10 photons… KABOOM!
There’re a lot of energy in those photons. Now the energy balances because the 10 photons
drive it uphill. When we calculate the efficiency of this, and efficiency is… I’m abbreviated
I see… I’ll write it out… Efficiency of any system is just the output over the input.
So the output is carbohydrate, and the input is 10 photons. When we do this ratio, we get
30% efficient in theory. Here it is. Theoretical maximum 30%. That doesn’t sound very good
for a theoretical maximum, and engineers look at this and say, “Oh my goodness! 30%? That’s
all your important equation can do?” And it’s bothered me for a while. So then I thought,
wait a minute! I look at the biochemistry, and this is not one big step. This is many
many little steps. This goes to this, and then this goes to this. When we start to add
up all those steps, if you assume there’s 23 steps, and that’s about right, and each
step is 95% efficient, now look at the math in this. If we take point .95²³, that’s
30% efficient. So we should start thinking about photosynthesis as 23-automated steps,
each one’s 95% efficient, and out the end we get carbohydrates and food energy. This
is where this whole analysis comes from is this equation balancing. Now, let’s look at
these 10 photons. They… That’s the theoretical minimum. How close can we get to that when
we measure this in a plant? Here’s a graph of photosynthetic photon flux density down
here. Increasing light 0 is dark, go all the way up to 1,000. And this is the quantum yield.
So this ratio, moles of carbon fixed per mole of photons absorbed, is called quantum yield.
Remember I said 10? Ten is called the quantum requirement to this. So 1 over 10 is 0.10.
So same thing just different units. Now look at this. We can’t quite get to ten, but we
get close up here. The type of plant really doesn’t matter as long as it’s a nice green
leaf. Tomato. Spinach. Different kind of tomato. In this case, Grand Rapids leaf is quite a
light yellow leaf, and it had a little bit lower quantum yield. Now when I say the leaf
doesn’t matter, purple. Purple leaf lettuce that is full of anthocyanins, and anthocyanins
are a protective pigment in leaves. They absorb solar energy, and they don’t fix carbon in
photosynthesis. So when we have purple leaf cultivars, their quantum yield is fundamentally
less than nice green leaf cultivars. We only get to these very low very high quantum yields
and very low light. So if we’re gonna be more real about this, let’s assume that we’re growing
the plants at 400 µmols. It’s more typical. And we go up here. And we go across here.
And we’re at a quantum yield of 0.08 and quantum requirement of 12. So let’s take this now
and refine our model of grams per mole of photons. There’s 400 µmols. We know that
carbohydrate is 30 grams per mole. This is pretty simple. Carbohydrate is 12. Hydrogen
adds 2. Oxygen adds 16. That’s 30 grams per mole. Twelve moles of photons gives us 30
grams of biomass, and the efficiency of this system then, output over input, is 30 grams
per 12… 2.5. This is significantly higher than what we talked about a minute ago is
1 gram per mole of photons. What’s the difference in this? The key here is that this is one
way. When we run this in the light, the first stable product out of photosynthesis is not
just carbohydrate, it’s sucrose. If we’re only running at one direction we get sucrose,
and even sugarcane is not pure sucrose. We have proteins in plants, lipids, lignin, cellulose,
all kinds of things. How do those things get made if we’re running this only to get sucrose?
They get made at night, and they get made continuously by running this equation backwards.
This direction, and then we get ATP energy. And that ATP is used to synthesize all the
other things in plants. When we run it backwards we call it respiration. So respiration takes
about 40% of this, after we get here, about 40% is run backwards to make everything else
in plants. This is, we can measure this rigorously by canopy gas exchange. This is a picture
of some of our chambers. They’re little transparent chambers. We measure the gases coming in and
out. We keep track of all the CO₂ in this system. And we can keep track of the exact
number of photons from minute-to-minute. The exact rate of sucrose production, and then
respiration. Now respiration runs in the light and the dark all the time. So it’s running
back and forth. So we never get 2.5 grams of biomass. That would only be sucrose. How
efficient is respiration? 60%… I’m off the edge of the screen… to go back… 60%. Respiration
is net photosynthesis divided by gross photosynthesis. We also called that daily carbon gain divided
by gross photosynthesis. That ratio’s about 60. Sixty. How do we get down to that number?
Jonathan France was a PhD student with me many years ago, and he spent his whole time
studying this. Couldn’t we make respiration more efficient? Couldn’t we get this higher
than 60%? And we eventually did. We got it to like 65%. We couldn’t improve it much.
Cold nights. Long days. Continuous light. We were close to fixed at 60%. So now we take
2.5 going this way times 60% equals, now we’re down to, 1.5 grams per mole. And we call this
number Photon Conversion Efficacy or PCE. So, in the first slide I said we can get 1
gram per mole of photons. This is 1.5. But this number assumes every photon gets absorbed
by the plants. None of them bounce off. None of them hit the ground. 100% of the photons
get absorbed. And that never happens. Even in a perfect canopy, a few percent bounce
off and go back up to the ceiling, some hit the walls, and even worse some hit the ground
without hitting a leaf and those are completely gone. To go from this perfect system where
we just have photosynthesis and respiration to a real system, we need a little more sophisticated
model. And that sophisticated model is called the Energy Cascade model that I’ve worked
on with other people for many years, and we have several papers about it. Here’s the model.
The first… Here’s our two parameters right here that we’re talking about. And in this
typical case right here I’m using 0.06 because if it’s higher light levels, we get a little
lower quantum yield. But the very first step in this model is fraction of absorbed photons.
And we’re assuming 80%. This number is more like 50% in typical field production, but
with high planting density and maybe systems to move the plants apart, we capture more
photons: 80% times quantum yield times carbon use efficiency gives us biomass per mole of
incident photons. And finally the fourth one is called harvest index. That’s the ratio
of usable or edible product to total plant biomass. In the case of lettuce, this number
is 80-85%. We don’t eat the roots. We don’t eat the stem base. Lettuce is almost completely
edible. But in typical crop plants we’re at 50%. Wheat and rice. That’d be tremendous
if we could get those to 50%. They’re more like 40% typically. So we take this. We get
moles of carbon fixed per mole of photons, and now we multiply by 30 grams per mole,
and we’re down to 0.4 grams per mole. Well below our 1 gram per mole that I advertised
on that first slide. So now let’s look at potential. What do we have to do to get this
up to 1? Here’s a set of numbers. If we improve each one of these just slightly. We get that
80% absorbed to 90%, slight improvement. Quantum yield, we get it from 0.06 to 0.08. Low light
and, it’s one thing to get 0.08, but keep it there over the whole lifecycle. When the
plants are old try to still have this high quantum yield. Improve carbon use efficiency
to 0.65. Go to a very high harvest index crop. Multiply these together. Now we’re at 1 gram
per mole. There’s the magic 1 gram per mole. This is the potential yield. And in some systems
where we very carefully measure this, we can get 1 gram per mole. It is not an average,
it’s a potential. And you’ve got to hit it every single day of the lifecycle. Very hard
to do, but it’s still a potential yield. Most indoor farms that we have analyzed are down
around 0.4 grams per mole. Now mostly because they’re not absorbing all the light up here.
Many of the photons don’t fall on leaves. By the way this quantum yield only is achievable
with very elevated CO₂. Like 1,200 parts per million. It’s 400 outside. That causes
photorespiration. We can’t have that. To get to this number you have to have high CO₂.
It’d be much less than this without high CO₂. But we can enriched with CO₂ in indoor
agriculture. This approach came from this article right here, “Exploring the Limits
of Crop productivity,” that I wrote with my colleague Frank Salisbury about, well 1988,
that’s over 30 years ago. This particular article helped launch my career as a young
scientist. Showing the math, and then showing what we could get to in potential yield. And
of course we did this with wheat and keep track of all those parameters, and that was
the basis of the paper. So now let’s look, after we calculate potential yield, let’s
do an economic analysis of indoor ag. We know the cost of electricity. If we know we could
get 1 gram per mole, what’s the economics of this? This is, what I’m going to show you
on this slide now is the basis from this paper. It was a paper, an article we published in
“Nature.” “LEDs for photons, physiology, and food.” This is just last year, 2018.
Assuming 10 cents a kilowatt-hour, which is a sort of a national average, cost of photons
as a percent of the market price. So we took the market price as high as we could get.
Gourmet, premium quality lettuce. And look at this, 1% and 5%. No wonder people are rushing
to do indoor agriculture for leafy microgreens and for lettuce. But, most of the facilities
now where we have 5%, they’re at more like 30% of the cost of photons as a percent of
market price. And that’s because the systems are not yet optimized. They can run with 30%
energy, but they could get down to 5%. Now let’s look at tomatoes in this system. Tomatoes
have a much lower harvest index. You throw away the leaves and stems and roots. You only
eat the tomatoes. The tomatoes don’t have as much water as lettuce. Lettuce is 95% water,
and we only need photosynthesis to grow the 5% that’s dry weight. But tomatoes are much
harder. When we go to Tomatoes we get 18%. It’s harder to grow tomatoes than these. Still
possible, but boy we’re gonna have to have optimal conditions to ever think about growing
tomatoes. Now let’s look at other vegetables. Let’s take broccoli. Lots of broccoli we can’t
sell: stems, leaves, roots, and broccoli is more dry weight and less water. And it doesn’t
command the premium price that lettuce and tomatoes do. General vegetables, the cost
of the photons is equal to the market value of the product. So unless we can find a way
to get consumers to pay a lot more for general vegetables, including things like strawberries,
it’s going to be very very difficult. And by this analysis, at 10 cents a kilowatt hour
it’s not going to be cost-effective. The electric cost is going to be equal to the market value,
and you still have labor. You have depreciation. You have lots of other costs besides the photon
costs. Even worse… rice or wheat. The value of the photons is a hundred times greater
than the value of the rice and wheat. We are never going to grow our staple crops without
sunlight. We need free sunlight to grow these staple crops. One way to look at this—the
rapidly increasing cost of photons—is the market price. These are high market price.
These are low market price. That’s what drives this cost of photons up. And then secondly,
the harvest index. These have a high harvest index and it decreases down to 50%. There
are tomato lines that are 70%. So we could potentially do a little bit better with tomatoes.
The point here is tomatoes are much much harder to grow though than lettuce in indoor ag.
Now let’s do an analysis of what we could do, and this is my final set of slides in
this talk, what we could do with solar panels and LEDs so that we didn’t have to use fossil
fuels to get all that electricity. Let’s analyze this. What we’re going to compare, this is
sunlight, if we do total energy in sunlight it’s 30 mega joules per meter squared per
day of total short wave energy and 60 moles of photosynthetic photons per day. Now this
is what drives the photosynthesis. And the first step here is this is what drives solar
panels. So let’s do this side first. Sixty moles comes through a greenhouse, 70% transmission.
This is a good greenhouse. It’s a design goal for a greenhouse, and obviously we get with
this we get 42 moles, 70% of this, transmitted to the plant canopy. Now let’s analyze this
side. First step, photovoltaic panels 20% efficient. So that’s 30, goes to 6 moles per
meter square per day. That is captured as electricity from the sun per meter squared
of solar panels. Now we take the electricity. Put it through LEDs. Our very best LEDs can
get to 3 µmols per Joule of photons. µmols of photons per Joule. This is 10 to the, M
is 10 to the 6, this is 10 to the minus 6, so this is 3 times 6… 18 moles of photons
delivered to the plant canopy with solar takes. And it’s 42 here. This is a bigger number
than this. It takes more than 2 times the area of solar panels to make the equivalent
of what we could get in a greenhouse with nothing just a glass greenhouse. So I made
a diagram to emphasize that point. This is solar panels in a field, 2.3 acres of solar
panels feeding electricity to light up one area. So we’re not there for direct use of
solar energy through LEDs. But you say solar panels are getting better. LEDs are getting
better. Now let’s calculate the potential of this. What if we peer into the future and
look at what solar panels could be and what LEDs could be. Can we get this ratio better?
Here’s the same graph potential. This kept us the same, 70%, we might get this higher
with special coverings. But let’s keep this side the same. Now we go to 30% solar panels.
Now these, the stuff we fly on satellites, they’re expensive. They’re 30%. So it’s possible
to get these, and at some point in the future we’ll have consumer panels that are 30% efficient.
And the reason they’re so efficient is they can use energy out here in shortwave energy
by layering cells that photosynthesis can’t use. Photosynthesis uses this part of the
curve and solar panels, by layering them, can use these low-energy near-infrared photons.
So they get to 30%. Now we’ve captured nine mega joules of electricity. Now the theoretical
maximum if we had a 100% efficient LEDs, which we’re really not going to get there but we
might get to 90-95%. We can get close, 9.7 times 4.7… 42. So potentially we can have
a system where that’s 1:1. The one acre of solar panels feeds one acre of plants and
they don’t have a greenhouse. Now this is the cost of building a greenhouse versus all
these high efficiency solar panels and LEDs. This doesn’t show economics it just shows
energy flow in this system. So one final slide, this is from a recent paper that is now in
review with my colleagues Paul Kusuma and Morgan Pattison. It’s called “From physics
to fixtures to food: Potential efficacy of LEDs.” This is peering into the future.
These numbers come from the chip manufacturers of LEDs. Not the ones that make the fixtures.
Just the ones that make the chip. How good can they get those chips? Here’s the colors
we use. We use blues, we use reds, we use far-reds, and we use whites and cool white
is the most efficient of the whites. Look at these efficiencies now. This is watts of
photosynthetic energy out per watt of electricity in… Wow! 88% for blue LEDs. That is really
impressive. Whites are phosphor converted blue. So they take blue with a phosphor, and
we get white. So there’s some loss in that, and so whites are 80%. And the reds are less,
and the far-reds are less. Now this is energy. But remember energy doesn’t drive photosynthesis,
the number of photons do. So this column is the µmols of photons per Joule. And we review
this number. This is µmols per second over joules per second. And the seconds cancel.
So it is just µmols per Joule, and a Joule per second is a watt. So it simplifies to
µmols per Joule. Look at these numbers 3.3 up to 4. These are impressive numbers. We
don’t have fixtures this good yet, but on the horizon we could see where they could
get this good for these colors of light. Now I got a final box here that I haven’t shown
you yet. What’s in that box? If there’s efficiency shown this way. Efficacy. What about the cost
of these? Which do you think is the most cost-effective the easiest to make? Most lighting, almost
all lighting, is for humans. And so humans need white LEDs. So the white LEDs are much
much cheaper just because of economy of scale of this. So now let’s look at the price…
Wow… 1x for whites. And we go to 10x for reds. And 30x for these. These drive the cost
of fixtures that drive the cost of what we’re going to do in indoor agriculture. So this
is peering into the future and looking at what we can produce. All of our analysis indicates
that leafy greens can be very cost-effective once we get the systems optimized. Tomatoes
would be much harder, and things like strawberries and broccoli are extremely difficult to ever
produce without free sunlight. Thanks for listening. I look forward to another video