Cheap T5HO "Plant" bulbs and longevity

[naoki]

There has been a couple discussion about different T5 HO bulbs (Stephen's thread and Polyantha's thread). With T5, the bulb choices are still somewhat limited (unlike T8 or T12). There are some specialty bulbs, but they are expensive. This spring I found a cheap source of plant bulbs (the purplish, gro-lux type). It is from TopDogSellers in ebay (I think the name of the company is GobiiDirect). 2' 24W bulbs are $26 + $12 shipping for 4, and 4' 54W bulbs are $34 + $20 shipping for 4 bulbs. I think the shipping is a bit expensive, but if you order a lot of bulbs in one shot, you might save some. It is still cheaper than some of the plant bulbs you can get from Pet/aquarium Shop.

It is marked as Odyssea Plant bulb.

I wasn't sure about the quality, so I took some simple measurements. I used 2' 24W bulbs. The fixture is Hydrofarm FLT-24, but used only 1 bulb per fixture. Room temp was 66F. Measurement is at the middle, 1 foot from the bulb.

Brand new bulbs:

bulb  PPFD fc.1  fc.2 watt PPFD/fc.2 PPFD/watt
plant   81  230   475 33.3      0.17    2.43
6400K   58  230   472 29.3      0.12    1.98

The 6400K bulb was AgroBrite 6400K.

columns:
PPFD: photosynthetic photon flux density in micromoles/m^2/s, measured by old Li-Cor Quantum sensor (not calibrated recently).
fc.1: foot-candles measured by Gossen Ultra Pro (calibrated, but the spectral response curve deviates from CIE luminosity function).
fc.2: foot-candles measured by a cheap LX-1330B meter (not calibrated, but fairly new, the manual shows a nice spectral response curve, but not sure if it is the true spec.)
watt: this is measured by Kill-a-watt. Since I'm putting only 1 bulb in the 4-bulb fixture, this energy usage is quite a bit of overestimation. Indeed, if you put 4 bulbs, it uses 72.6 watt (not 30wattx4).

PPFD and PPFD/watt columns are the most relevant data for plants. So the plant bulbs give 40% more PPFD per bulb and 23% more PPFD for a given watt. PAR (PPFD) doesn't put weight for red light (which is a bit more efficient for photosynthesis than green light), and it counts the number of photons in 400-700nm range. 6400K contains quite a bit of green (compared to plant bulbs), so the advantage of the plant bulb here is an under estimation.

I have only measured 1 bulb of each type, so I can't say too much, but the cheap plant bulb seems to be pretty good.

Longevity of T5HO.

According to some web site, T5 and T5HO seem to have a bit better longevity (i.e. the output doesn't decline as fast as T8). A couple sites about the lumen maintenance curve can be found: here 1 and here 2. However, when I measured my older bulbs, the output decline seems to be pretty significant.

2 year old AgroBrite 6400K had only 57% output in PAR compared to the brand-new bulb (and 61% in fc).

About 5 months old Odyssea Plant had 70% in PAR (79% in fc).

About 3 months old Odyssea Plant had 87% in PAR and in fc.

The decline in output seems to be much more than I expected. I used to think T5HO should be replaced every two years, but now I think I have to replace them much more frequently. I wonder why there is such a big difference in the reported lumen maintenance vs my measurement. It's possible that since I put the fixture directly on top of acrylic enclosure and the bulbs have shorter life due to the heat. Any ideas?

Inverse square law.
The inverse square law states that the light intensity should decrease to 1/4 if you increase the distance twice. But since T5HO is not a point light source, the reduction of light with distance is slightly different.

The fixture with 4 bulbs of T5HO gives following fc at several distances:
1' 970fc
2' 336fc
4' 93fc
So at 4', it gives about 1/10 instead of 1/16 (=1/4 * 1/4).

Ray has a more detailed info.

 

[TyroneGenade]

The inverse square law will only hold for a point light source. T5 lamps tend to come with reflectors nowadays that refocus the light and nullify the central tenet of the inverse square law. Your data at 1, 2 and 4 feet confirms this.

Have you found spectra for these lamps? The plant lamp must have a lot of red in the spectra for such a large difference in PPFD. The more blue light the less PPFD as generating blue photons uses more energy than red photons. Many 6500 K lamps have large green and blue peaks and smaller red peaks hence the lower PPFD counts but then some plant lamps also have low PPFD measures because the lamps mostly produce blue and red. The high f.c. and PPDF suggests the plant lamp is mostly green and red light (green/yellow light is the what is measured by lumenometers).

The low longevity is surprising. Industrial data from GE, Osram etc... for their T5 lamps have them maintaing 80% of their output for 70% of the lamps functional lifespan.

Personally, I think LEDs are the way of the future.

 

[naoki]

Have you found spectra for these lamps? The plant lamp must have a lot of red in the spectra for such a large difference in PPFD. The more blue light the less PPFD as generating blue photons uses more energy than red photons. Many 6500 K lamps have large green and blue peaks and smaller red peaks hence the lower PPFD counts but then some plant lamps also have low PPFD measures because the lamps mostly produce blue and red. The high f.c. and PPDF suggests the plant lamp is mostly green and red light (green/yellow light is the what is measured by lumenometers).

Tyrone, I don't know the spectra (other than it is purplish one). I did make the DIY Spectrophotometer with DVD-R and a cheerio box. The optics side works well, but capturing the image with camera and taking the measurement is a bit sketchy at this moment (then I got bored with refining it).

Is it really true that blue photon from fluorescent light requires more energy to produce? This seems a bit more counterintuitive. Fluorescent light shoot UV, and when it hits different types of phosphors, they fluoresce in different color. So the total energy consumed by a bulb is from the UV generation. Then you are saying that blue phosphors have lower conversion efficiency (quantum yield?) than red phosphors.

LED is different, but with phosphor based white LED, bluer cool white produces more photon than redder warm white per given watt in theory, right? When I measured PPFD of cool white vs warm white LED, the difference was pretty small, though.

Yes, I'm very surprised/disapointed with the longevity. I'll probably follow the longer term decay when I put new bulbs next time. It is true that the reported longevity is for the continuous lighting, and turning on and off wear the bulbs quicker. But for us, we turn on and off once a day, so I thought it shouldn't cause too much wear.

 

[TyroneGenade]

The energy of a blyue (450 nm) photon is 4.4 x 10^-19 J. A red photon (720 nm) is 2.76 x 10^-19 J (http://staffweb.psdschools.org/rjensen/gchemlabs/photon_energy.htm). The power rating of the lamp (e.g. 24 W) is how much energy is consumed. A blue lamp would produce fewer photons than a red lamp if it was using the same energy. This is in accordance to the laws of thermodynamics (you can't get more energy out than you put in, and you get out less than you put in). I was quite shocked by this when I realized this was happening but it explained what I was seeing with PAR/W values from different lamps, and why some low PAR/W lamps grew plants better even though they had lower PAR/W. In retrospect it seems so simple.

 

[naoki]

Hmm, I don't think that's how florescence work, but I may be wrong. UV photons have the higher energy. Phosphors absorb some of the energy, and release the lower energy (longer wave length) photons. In other words, phosphor molecule has to absorb only small amount of energy to release blue light, but it needs to absorb lots of energy to release red light. So quantum yield for different phosphor doesn't seem to be just related to the energy of each photon, and it seems to be more complicated than your explanation.

 

[DavidCampen]

Hmm, I don't think that's how florescence work, but I may be wrong. UV photons have the higher energy. Phosphors absorb some of the energy, and release the lower energy (longer wave length) photons. In other words, phosphor molecule has to absorb only small amount of energy to release blue light, but it needs to absorb lots of energy to release red light. So quantum yield for different phosphor doesn't seem to be just related to the energy of each photon, and it seems to be more complicated than your explanation.

With phosphors, 1 absorbed photon results, at most, in 1 emitted photon (assuming 100 percent efficiency). If the energy of the photon being absorbed is greater than the energy of the photon being emitted then the excess energy is converted to heat.

The same is true of photosynthesis, it is also a quantized effect, a certain number of photons are required to fix a CO2 molecule. The photons need only have the energy of 650 nm (red) photons. If blue photons are supplied instead, the same number of photons are required as with red photons and the higher energy of the blue photons is converted to heat.

 

[naoki]

That's my understanding, too, David. So the question of which of the blue or red florescent bulbs creates more photons per given watt depends on the efficiency of phosphors, right? In other words if you make a bulb which emit mostly blue or mostly red, which one gives more photons per given watt of electricity. Quantum yield is defined here as number of photons emitted/number of photons absorbed. So the answer depends on the quantum yield of different types of phosphor (not the energy of each photons). Well it's a minor point, but since Tyrone brought it up, I was curious to know.

 

[DavidCampen]

That's my understanding, too, David. So the question of which of the blue or red florescent bulbs creates more photons per given watt depends on the efficiency of phosphors, right? In other words if you make a bulb which emit mostly blue or mostly red, which one gives more photons per given watt of electricity. Quantum yield is defined here as number of photons emitted/number of photons absorbed. So the answer depends on the quantum yield of different types of phosphor (not the energy of each photons). Well it's a minor point, but since Tyrone brought it up, I was curious to know.

Yes, so with phosphors as used in fluorescent lamps, if the quantum yield of a red and a blue emitting phosphor were equal (say 100%) then each phosphor will emit the same number of photons when used in the fluorescent lamp. You will not get more photons from a red emitting phosphor than you would from a blue emitting phosphor because the lamp is using the same ultraviolet photons to excite the phosphors. At 100% quantum efficiency, 100 ultraviolet photons emitted the mercury vapor discharge inside the lamp will produce 100 blue photons or 100 red photons.

 

[Bjorn]

Yes, so with phosphors as used in fluorescent lamps, if the quantum yield of a red and a blue emitting phosphor were equal (say 100%) then each phosphor will emit the same number of photons when used in the fluorescent lamp. You will not get more photons from a red emitting phosphor than you would from a blue emitting phosphor because the lamp is using the same ultraviolet photons to excite the phosphors. At 100% quantum efficiency, 100 ultraviolet photons emitted the mercury vapor discharge inside the lamp will produce 100 blue photons or 100 red photons.

Isnt it the same thing for white LEDs?

 

[DavidCampen]

Isnt it the same thing for white LEDs?

Yes, the common white emitting LED modules that are sold for household lighting have a combination of red, green and blue phosphors that are being excited by photons from a blue emitting LED. So energy is being wasted in that you are using a high energy blue (almost ultraviolet) photon to produce a lower energy green or red photon. There are some white emitting LED modules intended for decorative and theater lighting that do not use phosphors but instead combine light from individual red, green and blue LEDs but these are not commonly used.

Therefore, if the quantum efficiency of red, green, and blue LEDs were all the same then using red and blue LEDs for plant lighting would be more energy efficient than using white phosphor LEDs. There is one factor though that compensates for this and that is that blue emitting LEDs seem to be available with significantly higher quantum efficiencies compared to red emitting LEDs.

 

[TyroneGenade]

Are there any physicists in the forum?

My understanding of thermodynamics, is that if you are getting 1 J our you had to put more than 1 J in. So, if you get 1J of red photons out, you had to put more than 1 J in... Same for blue/violet. But if blue photons have more energy than red photons, and we assume the phosphors respond the same (all losing the same amount of E as heat per photon) then surely we would still have violet photons costing more input energy per red photon because you can only get out what you put in to start with. The laws of thermodynamics must be obeyed (or some physicist's head will explode). Photons are little particles/waves of energy (quanta). They must obey the same laws as everyone else.

FYI: the W output (as light)/W input efficiencies for fluorescents are more or less constant at, if memory serves, about 30% of the input is lost as heat and the rest is emitted as photons of various wavelengths and energies.

 

[aquacorps]

I had great results with run of the mill T-5 bulbs when I was growing.

 

[TyroneGenade]

(from http://www.controlledenvironments.org/Light1994Conf/1_1_Geiger/Geiger text.htm )

Normal, cool white fluorescents have large blue, green and yellow-orange peaks and as it turns out, for terrestrial plants*, they photosynthesize just fine at those wavelengths even though they actually prefer red light.

(From http://www.photobiology.info/Gorton.html ) gives action spectra for crop plants and you can see that terrestrial plants seem to use green/yellow light very well. There is some question as to whether shade plants are more biased in their use of light (shade is due to canopy cover that will screen out red and green light). If you look here: http://www.slippertalk.com/forum/showthread.php?t=32896&highlight=Fluora , Osram Fluora lamp stimulated photosynthesis (CO2 fixation) more effectively than the daylight lamp suggesting a red/blue bias. That being said, bombarding the plants with lots of yellow light (cool white) will get you the same outcome as a little plant-lamp lighting. I am not convinced, for terrestrial plants, that there is a major advantage.

If you look at http://www.waynesthisandthat.com/fluorescent.html, the GE Bath & Kitchen (1st row, top) was as good as the GE Plant & Aquarium (5th row):

The Gro-lux (third row) was rather unimpressive.

A lot of our thinking of what type of light plants need to grow is based more on myth than fact.

If someone has a link to an action spectra of a typical shade plant please let me know... I am very curious. If someone had data on orchids that would be even better!

*Aquatic plants have very different action spectra and yield photon fluxes and can't really be compared. Special aquarium plant lamps may actually be better than normal daylight/cool white lamps.

 

[naoki]

Yes, the common white emitting LED modules that are sold for household lighting have a combination of red, green and blue phosphors that are being excited by photons from a blue emitting LED. So energy is being wasted in that you are using a high energy blue (almost ultraviolet) photon to produce a lower energy green or red photon.

I agree that phosphor wastes energy, and I believe that improving phosphor is one of the areas where LED companies are working on. But one thing I don't quite know is why the emission spectra look so different between white fluorescent (with a few major peaks) and white (phosphor-based) LED (with smooth fluoresced emission + blue). I know quite a few major peaks in fluorescent lights are from Mercury (which isn't in LED), but they still look quite different.

 

[naoki]

Are there any physicists in the forum?

My understanding of thermodynamics, is that if you are getting 1 J our you had to put more than 1 J in. So, if you get 1J of red photons out, you had to put more than 1 J in... Same for blue/violet. But if blue photons have more energy than red photons, and we assume the phosphors respond the same (all losing the same amount of E as heat per photon) then surely we would still have violet photons costing more input energy per red photon because you can only get out what you put in to start with. The laws of thermodynamics must be obeyed (or some physicist's head will explode). Photons are little particles/waves of energy (quanta). They must obey the same laws as everyone else.

FYI: the W output (as light)/W input efficiencies for fluorescents are more or less constant at, if memory serves, about 30% of the input is lost as heat and the rest is emitted as photons of various wavelengths and energies.

It is following the thermodynamics. UV light has higher energy. When it hits phosphor, it emits lower energy light like David said. I don't think phosphors respond the same. Phosphors which release longer wave length has to absorb more energy to meet the thermodynamics. And different types of phosphor has different quantum yield (so this part is related to whether blue fluorescent bulb or red fluorescent bulb releases more photons per given watt).

You are right that in a ball park, certain amount of energy is released as heat etc. And this efficiency is quite different between systems (e.g fluorescent vs HPS vs LED). But different bulbs and ballasts contributes to some differences within a system.

 

[naoki]

(from http://www.controlledenvironments.org/Light1994Conf/1_1_Geiger/Geiger text.htm )

Thanks for this link! It is very interesting. Plants are accustomed to gradually increasing and decreasing light. So when we use artificial constant light, plants need to acclimate to the constant light condition. Acclimation capacity of plants are quite amazing.

If you look at http://www.waynesthisandthat.com/fluorescent.html, the GE Bath & Kitchen (1st row, top) was as good as the GE Plant & Aquarium (5th row):

This is interesting. I wonder what are the PAR and energy consumption of each bulb. A part of the response is likely to be due to photomorphogenesis; blueish light causes smaller, more compact plants. So above and below ground biomass would be better than the rosette size. Actual growth experiment is better than measuring PAR, spectra, or photosynthetic rate, but it takes longer time.

A lot of our thinking of what type of light plants need to grow is based more on myth than fact.

If someone has a link to an action spectra of a typical shade plant please let me know... I am very curious. If someone had data on orchids that would be even better!

For crop plants, there are lots of data about effects of light on plants (both photosynthesis and photomorphogenesis). If you look at the original paper which came up with the action spectrum for the crop plants, there are some amount of variation among different species. So you are right that it would be interesting to see the action spectra for shade plants. I looked into this at one point, but I didn't find it.

 

[TyroneGenade]

GE Plant and Aquarium: 1.16 microEinsteins/W (peaks at 435, 475, small 550, 635 & 665, very little > 700 nm)
GE Kitchen & Bath: 0.72 microEinsteins/W (peaks at 440, 490, 550, 590, 615 & 715 nm)
(values computed from spectra output in W/5nm/1000 lm)

Ok, Naoki, lets see if I understand you: red light costs more than blue light because more of the UV energy is lost as heat in exciting the phosphor to release a red photon? Looking at my lamp data, there is r = 0.598 (P = 0.000000314, n=62, Spearman) correlation between the fraction red relative to the total photons emitted (the more red, the more photons total). Conversely, there is a r=-0.443 (P=0.000342, n=62) for blue relative to total (the more blue, the fewer photons total). So the more blue, the less photons, the more red, the more photons. Transformation efficiencies (input power to photons) is pretty much the same per lamp, so the data suggests that blue photons come at an increased cost of other photons, i.e. it costs more energy to generate a blue photon. (I suppose, red phosphors could be more efficient, and blue phosphors less efficient so the cost isn't exactly the energy of the emitted photon.) PM me your email address and I will send my dataset to you. The explanation, as I understand it, doesn't fit the data. I would like some resolution on this. If you are right I am generating wrong hypotheses for my aquarium plant experiments...

 

[Ray]

Phosphors which release longer wave length has to absorb more energy to meet the thermodynamics.

Is that correct, Naoki?

If I think about an incandescent light bulb, lower energy input means lower energy, longer wavelength output. If the opposite were true, we'd see blue first, not heat and red. Why is emission of a phosphor different?

Light emission occurs when the excited electrons fall back into their shells in both cases, right?

 

[naoki]

Is that correct, Naoki?

If I think about an incandescent light bulb, lower energy input means lower energy, longer wavelength output. If the opposite were true, we'd see blue first, not heat and red. Why is emission of a phosphor different?

Light emission occurs when the excited electrons fall back into their shells in both cases, right?

Ray, it's an interesting point, and I don't know whether the principle of light emission is same for incandescent and phosphor. Is it?

Well, here is my understanding (I could be wrong). In the case of florescent bulbs, input energy for the phosphor doesn't differ (energy of UV), but different types of phosphors have different emission spectra. You analogy to incandescent light is like changing the wavelength of the excitation light (input energy) without chaing the type of phosophors. I'm not sure what happens in this case.

When you look at the photochemisty section of:
http://en.wikipedia.org/wiki/Fluorescence#Physical_principles
it seems that more energy has to be released as heat with phosphors which shift the light frequency more. I said "absorbed", but heat release is probably a better term?

Also http://en.wikipedia.org/wiki/Fluorescent_lamp#Principles_of_operation at the end of 2nd paragraph in Principles of Operation section says the same thing.

 

[naoki]

GE Plant and Aquarium: 1.16 microEinsteins/W (peaks at 435, 475, small 550, 635 & 665, very little > 700 nm)
GE Kitchen & Bath: 0.72 microEinsteins/W (peaks at 440, 490, 550, 590, 615 & 715 nm)
(values computed from spectra output in W/5nm/1000 lm)

Are these numbers (1.16 and 0.72) photon efficiency? Did you mean micro moles (=microE) / J (instead of W = J/s)? The unit of PAR is micro moles (=microE)/s and input electricity can be measured by W (=J/s). So if you divide PAR by W, the denominator becomes J (joules) instead of W.

But that is a big difference, and it makes me wonder how Kitchen & Bath can be so good.

Ok, Naoki, lets see if I understand you: red light costs more than blue light because more of the UV energy is lost as heat in exciting the phosphor to release a red photon?

That's not quite what I mean (sorry I wasn't clearer), Tyrone. I'm not sure red light or blue light costs more. My point is that the question cannot be answered by the fact blue light photon has more than red light photon nor by the fact that the phosphor which emits longer wavelength need to release more heat. My point is that it depends more on how different types of phosphor behaves (which I don't know how they do).

But your real data is better than the theoretical background. So I sent you my email in PM.