This is an old article that was published. Please, before sinking money into LEDs, consider the arguments made here. So, without further ado, what the LED grow light manufacturers don't want you to read!
Light emitting diodes (LEDs) are an exciting newer type of lighting for indoor horticulture with a number of new manufacturers coming out with LED grow lights. The LED lighting manufacturers make some pretty impressive claims about the performance of their products. Does the hype live up to their claims? We'll examine some of the claims about LED grow lights from an engineering and photo biology perspective and try to separate fact from fiction so that the consumer can make an informed choice about whether or not to take the plunge into LED grow lighting.
A major claim about LED grow lights is that they'll last for 50,000 or 100,000 hours (one manufacturer claims 11.41 years...!). Can a grow light really last being on for 10 years or so? Well, it depends. For one thing, a few of the manufacturers are giving numbers regardless of the actual LEDs used in their lamps. Most Chinese made LEDs, for example, actually have a rather poor track record as to their longevity (and the accuracy of the data sheets!).
There are two numbers that we need to look at for longevity of LEDs: the Lumen Maintenance level (L70) and mortality rate (B10). The Lumen Maintenance level is expressed in hours and is the point where LEDs have degraded to the point where they give off 70% of the light from when the LEDs were new. This is the lifetime rating of the LEDs. At this point there is also typically a 10% mortality rate so if you have an array of 100 LEDs you can expect 10 of the to burn out at a 90% confidence level. The harder that you drive a LED the shorter its life will be (higher temperature also shorten the life a the LED). This is a hidden cost that one must consider when making a long term capital investment in a lighting system that costs $5 per watt. Can you replace the LED that's mounted to a heat sink yourself? If not, there could be some down time and costs to get your light fixed. It's likely that by the time that the LEDs will start burning out the warranty will have expired. Wisely, at least one LED grow light manufacturer designed their light to make it easy to replace LEDs.
In a cost analysis of LED grow lighting it is also important to take into account the temperature of the LEDs. For example, the electrical efficiency of a high power red led is around 20% but this is for a junction temperature of 25 degrees Celsius. Run the LED with a junction temperature of 40 degrees Celsius, which is typical with a good heat sink, and the red LED drops down to about 17% efficiency, a 15% drop in light levels. In comparison, a high pressure sodium light is around 25-26% efficient with a magnetic ballast and closer to around 30% efficient with a digital ballast (the digital ballast is able to ionize the gases in a HPS bulb more efficiently. Also in a digital ballast the bulb will last longer due to the soft start feature found in digital ballasts).
It is possible to get red LEDs that are closer to 25% efficient but you can expect to pay a premium for them. LED manufacturers bin (test and sort based on characteristics) their parts. Unfortunately it's also possible to get red LEDs that are closer to 15% efficient if you get the wrong LEDs so it's important that one understands the bin code if ordering LEDs.
What about thermal efficiency? A claim made by a few manufacturers and writers is that LED grow lights put out no heat. This is very wrong; there is no way to get around the laws of thermodynamics. If you have a LED grow light with 100 watts input, at least 80 watts of heat will remain in the lamp housing which is why all higher power LED arrays on the market are either air cooled or water cooled and why the LEDs will be mounted on an aluminum heat sink. Nearly all of the 20 watts of light energy in this example also ends up as heat in accordance with the 1st Law of Thermodynamics (heat is the lowest form of energy and there's a conservation of matter and energy according to the 1st law, very little of the light is actually converted to matter). Even low power (5mm and 10mm LEDs) arrays generate heat proportional to their power levels. Try wrapping a low power array in a towel (OK, don't really try this!) and see how hot it gets. Trying running a high power LED at its rated current without a heat sink and see how many seconds it takes before the LED burns up (a 15 watt LED I accidentally smoke tested without a heat sink started burning in about two seconds at full power. I was talking with an engineer working for the LED manufacturer and he was surprised it look that long to burn! The 15 watt LED is about half the size of a dime).
If you have a series of closed rooms the same size and put a high pressure sodium lighting system with 1000 watts power input in one room, 1000 watts input of LEDs in the next room, a 1000 watt input air conditioning unit with no output exhaust in the next room and a 1000 watt electric heater in the last room, all of the rooms will be at the same temperature. There's nothing magical that makes LED grow lights cooler than other types of lighting watt for watt in an enclosed grow chamber. High power LED grow lights are, however, very efficient in dumping the heat off the lamp housing because of the aluminum heat sink and the active cooling but I can keep my 250 watt HPS bulb cooler by blowing air right on the bulb with a computer fan. If the 250 watt bulb were also efficiently thermally coupled someway to a proper size aluminum heat sink then the bulb would also stay fairly cool with active cooling but the total system heat output would be the same. The only way to lower the total system heat output is to use less power.
There's a high power LED grow light on the market that claims to be 90 watts. Is it? Nope! They do have 90 one watt LEDs in their lamp but a one watt LED isn't one watt. The way the manufacturers that actually make the LEDs rate any LED that runs at 350mA (mA stands for milliamp- an amp is a measure of current) is a one watt LED, 700mA is a three watt LED and 1000-1500mA is a five watt LED (10 and 15 watt LEDs usually have 4 LEDs on the same die in series). A typical one watt red LED will have a voltage drop of 2.25 volts- 2.25 volts times 350mA of current equals .79 watts of power. A typical blue LED will have a voltage drop of 3.4 volts- 3.4 volts times 350mA equals 1.19 watts. The 90 watt LED grow light has 82 red and 8 blue LEDs. This is closer to 75 watts rather than 90 watts. Unfortunately, a few grow light manufacturers use this trick to make their lights seem like they have more power than they do. One LED grow light manufacturer, homegrownlights.com and the maker of the Procyon 100, gives a true rating of their light (one hundred watts input to the LEDs and 125 watts input to the whole lamp).
It gets even worse with three and five watt LEDs. For example, one 5 watt red LED that I have gives a voltage drop of 2.5 volts at 1000mA. So that five watt red LED is actually a 2.5 watt LED while a blue five watt LED is closer to 3.5 watts. A three watt red LED is is more like 1.75 watts and a blue three watt LED is closer to 2.4 watts. It is important that the consumer understands this as more manufacturers start putting these higher power LEDs in their lamps.
Most LED grow light manufacturers measure the power rating of their lamps by power input. With one 48 watt LED grow light I've seen on Ebay, the seller actually claimed that the 48 watt rating was because the plug-in power supply said 24 volts at 2 amps. Unfortunately, this doesn't tells us how much power that the lamp is actually using and may suggest that they're using a constant voltage power supply instead of a more proper constant current power supply.
It would be much easier on the consumer if all grow light manufacturers would just give the true power input to the LEDs just like makers of metal halide and high pressure sodium lighting systems rate the power of their products by the power input to the bulb.
Bottom line from an engineering perspective, watt for watt LEDs generate as much heat as any other type of lighting and they currently have a lower electrical efficiency (and lower luminous efficacy!) than high intensity discharge (HID) lighting such as high pressure sodium. LEDs will last longer than HID bulbs but there's many more of them to burn out and since LEDs are wired in series in arrays it's possible for the entire lamp to go out if one LED burns out. The good news is that if you're technically inclined, most LED lamps are fairly easy to fix. In addition, 50 watts of LEDs, for example, might not actually be a true 50 watts of LEDs.
Well, we made it past the quick engineering perspective of LED grow lighting. Now, how do they do with plants? To start off with, you need to understand two different charts. One is the absorption spectra of chlorophyll A and B and the other is the net photosynthesis chart. You can look these charts up online and make sure that they're for land plants and not algae!
The absorption spectra of chlorophyll A and B shows very narrow absorption spikes centered around 660nm and 430nm for chlorophyll A while it's 640nm and 450nm for chlorophyll B. One can look at this chart and say we need LEDs to fall on these numbers and indeed a few LED lighting manufacturers actually use this graph to try to back their claim that LED grow lights are 20 or 30 times better than HPS (some really claim this), after all, the chart shows that there is very little absorption at 589nm which is where HPS has its spectral peak. There is just one major problem. The chlorophyll absorption charts are for isolated chlorophyll molecules suspended in a solvent and does not reflect total photosynthetic activity. In fact, different solvents can give slightly different numbers. If you're dealing with a LED lighting manufacturer or dealer hyping their product based on this chart then walk away. They are completely wrong, they likely don't realize it and no matter how much you argue with them you aren't going to change their minds. I speak from personal experience multiple times!
What we want to look at are the charts that show net photosynthesis and may also be referred to as the action spectra. I know I've already said this but it needs to be repeated, don't get algae charts mixed up with the land plant charts, they are different. Also, these charts tend to show relative results and not absolute results. They are normalized so that the lighting spectrum that is most efficient in photosynthesis is at 100%. This does not mean that 100% of that lighting spectrum is used in photosynthesis. It is only 100% compared to other lighting spectrum. This gets people confused. Also, there is no one size fits all chart, different plants will have a different optimal lighting spectrum, for example, purple basil and sweet basil are quite likely different since the purple basil is going to absorb more yellow light.
The action spectra charts show us something different than the chlorophyll absorption charts and explains why high pressure sodium lighting is so efficient. Most charts will show a high level of absorption in the 620-660nm range but at 589nm, the spectral peak of high pressure sodium lights, you're still getting 80-85% relative absorption and even with green you're looking at perhaps 50% absorption depending on the chart (yes, plants can use green light and any absorbed green photon is available for photosynthesis. I've grown lettuce under green high power LEDs and my results show about half the efficiency compared to red LEDs in terms of dry biomass. Plants aren't green because the absorb no green light, they're green because they absorb less green light. Green light will also stimulate auxins, an important class of growth hormones, which are needed in the biosynthesis of ethylene but that's another article!).
As far as lighting spectrum, it could be the case that the narrow LED lighting spectrum is not taking the full advantage of photosynthesis compared to a broader light source. Try looking up Emerson effect and look at the research.
Let's also look at photon flux density. A 100 watt LED light can put out lighting levels of roughly 100 umol/meter^2/sec. umol is pronounced micro mole and if you remember back to your high school chemistry classes the mole is Avogadro's number of 6*10^23. This makes one umol 6*10^17 photons and this unit is used quite often in photo biology. How far will that 100 umol/meter^2/sec of light get you? Well, to put it into perspective, full sunlight is 2000 umol/meter^2/sec (2000 umol per square meter per second), the photo saturation point for many food crops is around 1000 umol/meter^2/sec and most food crops thrive at 500 umol/meter^2/sec especially in flowering. The answer for the 100 watt LED light at an intensity of 500 umol/meter^2/sec is roughly two square feet. The 100 watt LED lamp can definitely grow in a larger area but the rate of photosynthesis will proportionally go down. I'll put my flame suit on now, some people are not going to be happy with this paragraph!
Well wait a second! What about these manufacturers claims about LEDs being 2, 4, 10, 20 (heck, pick your number, they're all over the place!) times more efficient than high pressure sodium lights? Here's the big reality check, if you get nothing else out of this article just remember this: there is no peer reviewed research paper backing these claims up. NASA has done extensive research, universities and private contractors working under Small Business Innovative Research contracts have done extensive research, yet not a single paper, not one, backs the claims that LED grow light manufacturers make about their lights and my 18 months in playing with LED grow lights also does not back these claims up. If you are a large grower about to drop thousands of dollars on LEDs please, I urge you, ask the LED grow light manufacturer to refer you to a peer reviewed paper that demonstrates the validity of their claims before making the purchase. If the manufacturer can not give you a link to a peer reviewed research paper then you need to ask why this is. LEDs have been well studied, where's the independent peer reviewed evidence showing how much better they are?
Check out 
