An inside look: LED bulbs

Foreword

Recently, a stir has been raised around the LED lamps, which should replace the ordinary Ilyich lamps. And as the chief nanotechnologist of Russia told , such lamps will soon go on sale in Moscow and St. Petersburg. Of course, everything was arranged with pathos: the first to appreciate the novelty of Vladimir Putin. I managed to get one of the first light bulbs from OptoGaN; besides, I had another Russian-made light bulb in my hands (SvetaLED or SvetaLED), though beaten with life, but working, and Chinese NoName, which can be easily buy on ebay or dealextreme.com.

When I get at least some valuable and interesting thing (from eyeshadow to a processor or CD , I just want to take it apart and look inside, see how it all works and works. Apparently, this is what distinguishes scientists from the townsfolk. Agree what a normal person will disassemble a light bulb for 1000 rubles, but what to do - the party said: it is necessary!

Part theoretical

Why do you think everyone is so concerned about replacing incandescent bulbs , which have become a symbol of a whole era, with gas-discharge and LED ?
')
Of course, firstly, it is energy efficiency and energy saving. Unfortunately, the tungsten coil emits more “thermal” photons (i.e., light from a long wave of more than 700-800 nm) than it gives light in the visible range (300-700 nm). It's hard to argue with this - the chart below will tell everything for itself. Taking into account the fact that the power consumption of gas-discharge and LED lamps is several times lower than that of incandescent lamps with the same illumination, which is measured in lux . Thus, we find that it is really beneficial for the end user. Another thing is industrial facilities (not to be confused with offices): lighting, albeit an important part, but still the main energy costs are associated precisely with the work of machines and industrial installations. Therefore, all produced gigawatts go to the rolling of pipes, electric furnaces, etc. That is, the real savings within the entire state is not so great.

Secondly, the service life of the lamps, which came to replace the “Ilyich light bulbs”, is several times higher. For an LED lamp, the service life is practically unlimited if the heat sink is properly arranged.

Thirdly, this is innovation / modernization / nanotechnology (underline the necessary). Personally, I do not see anything innovative either in mercury or in LED lamps. Yes, this is a high-tech production, but the idea itself is just a logical application in practice of knowledge about semiconductors, which is 50-60 years old, and materials known for about two decades.

Since the article is devoted to LED lamps, I will focus on their device in more detail. It has long been known that the conductivity of an illuminated semiconductor is higher than the conductivity of an unlit ( Wiki ). In some unknowable way, light causes electrons to run through a material with less resistance. A photon, if its energy is greater than the band gap of a semiconductor (E _g ), is able to knock an electron out of the so-called valence band and throw it into the conduction band.

The layout of the zones in the semiconductor. E _g is the forbidden band, E _F is the Fermi energy, the numbers indicate the distribution of electrons over the states at T> 0 ( source )

Let's complicate the task. Take two semiconductors with different types of conductivity n and p and connect together. If in the case of one semiconductor we just observed an increase in the current flowing through the semiconductor, now we see that this diode (namely, the pn junction, which appears at the border of semiconductors with different conductivity types) is differently called a mini-constant source current, and the magnitude of the current will depend on the illumination. If you turn off the light, the effect disappears. By the way, the principle of operation of solar batteries is based on this.

At the junction of semiconductors of p and n type arising after irradiation with light, the charges are separated and "leave" each to its electrode ( source )

Now back to the LEDs. It turns out that the opposite can also be done: connect a p-type semiconductor to a positive on a battery, and an n-type to a minus, and ... And nothing will happen, there will be no radiation in the visible part of the spectrum, since the most common semiconductor materials (for example , silicon and germanium) - are opaque in the visible region of the spectrum. The reason is that Si or Ge are not direct-gap semiconductors . But there is a large class of materials that have semiconductor properties and at the same time are transparent. The brightest representatives are GaAs (gallium arsenide), GaN (gallium nitride).

So, to get the LED, we just need to make a pn junction from a transparent semiconductor. On this, I think, I’ll stop, because, the further you go, the more difficult and confusing the behavior of the LEDs becomes.

Let me just say a few words about modern technologies for the production of LEDs. The so-called active layer is a very thin 10-15 nm thick alternating layers of p- and n-type semiconductors, which consist of elements such as In, Ga and Al. Such layers are epitaxially grown using the MOCVD method (metal-oxide chemical vapor deposition or chemical vapor deposition).

Schematic representation of the LED device

There is another problem that prevents 100% conversion (conversion of 1 electron into 1 photon) of electricity, and it is that even such thin layers of semiconductors absorb light to a certain extent. Even not strongly absorbed, just light “wanders” inside the crystal due to the effect of total internal reflection at the crystal / air interface: the path length increases until the light exits the crystal and, ultimately, such a wandering photon can be absorbed. One of the solutions is the use of structured substrates. For example, in the modern LED industry, the molded sapphire substrate method is widely used. Such a microstructuring leads to an increase in the light output efficiency of the entire diode ( more ).

For interested readers I can offer to get acquainted with the physics underlying the work of LEDs. In addition to this interesting work performed within the walls of the native Moscow State University, Svetlana and Optogan have a wonderful galaxy of research teams in St. Petersburg itself. For example, PhysTech . And you can read this article .

Methodical part

All measurements of the spectra of the lamps were made within 30 minutes (i.e., the background signal was changing slightly) in a darkened room using an Ocean Optics QE65000 spectrometer. Here you can read about the device spectrometer. In addition to 10 dependencies on each type of lamp was measured dark spectrum, which is then subtracted from the spectra of light bulbs. All 10 dependencies for each sample were summed and averaged. Additionally, each final spectrum was normalized to 100%.

Ocean Optics Spectrometer is a great tool in the right hands.

Part practical

So let's get started. We have six light bulbs in stock: 3 for complete analysis and 3 more for comparison (so to speak, control samples):
1. Bulb of Ilyich
2. Illich M lamp (i.e. a gas-discharge lamp that follows the usual Illich lamp)
3. Spiral of Ilyich (ordinary discharge lamp)
4. LED-lamp from "Optogan"
5. LED-lamp from Sveta LED
6. LED-lamp from China NoName

All light bulbs in the collection. We can start!

Spectra

There is nothing supernatural here we have not seen. The illich Ilyich lamp shamelessly lets all the electricity into the heat and the color of it either yellow or orange. All mercury lamps have a striped spectrum, which in the human eye, as the simultaneous inclusion of 3 pixels (RGB) on the screen (blue lines - ~ 420 nm, green - ~ 550 nm, orange and red - everything above 600 nm), is converted to white.

Spectrum of three comparison bulbs (for comparison, under the scale is a part of the spectrum that is perceived by the human eye)

But the range of LED lamps is very different. There are two components: actually, blue from the diode itself, and the second, spread across the spectrum, from a phosphor or, in Russian, a fluorescent dye, which is applied to the LEDs themselves and poured over with a protective layer of polymer. The ratio between the blue color of the diode and the emission (emission) band of the phosphor determines the color temperature of the lamp. We can see that Optogana has the warmest light, and China has the coldest light. It is advantageous to use 1 phosphor to control the color temperature, thus, the thickness of the phosphor layer, coupled with the power of the LED, determines the color temperature. It is worth noting that, apparently, the same phosphor is used in bulbs from China and from Svetlana, while Optogan uses its own (a significant difference in the maximum emission band of the phosphor).

Comparison of the spectra of LED lamps and the traditional lamp of Ilyich (for comparison, under the scale is presented the part of the spectrum that is perceived by the human eye)

We got the light bulb from Svetlana in a broken form, and we shot the spectrum without frosted glass. However, let me demonstrate a similar situation on the example of a lamp from China, since there were two of them. The normalized spectra differ little from each other, and a small increase in the intensity can be attributed to the fact that longer wavelength radiation is better scattered on frosted glass.

Comparison of Chinese-made lamps with and without a glass bulb (for comparison, under the scale is presented part of the spectrum that is perceived by the human eye)

If someone is interested, then here presented a rather detailed modeling of the characteristics of LEDs.

Price, materials and specifications

Three girls under the window broke late in the evening ... Right to left: Optogan, SvetaLED and NoName China

Chinese NoName

A light bulb from China was ordered through dealextreme.com and delivered to Russia within 2 months (you understand, Russian Post). Its cost is about $ 14 or about 420 rubles, including delivery. Color temperature 5000-6000K, which corresponds to the white cold light. The dimensions coincide with the usual light bulb Ilyich. Material "flask" - frosted glass. In my opinion, an ideal replacement for an ordinary incandescent lamp, if the color temperature was 1000-2000K lower than indicated.

"Optogan"

The light bulb was presented to mere mortals at a special presentation . Design by Artemy Lebedev, the noble materials of the case - polycarbonate and aluminum with proprietary symbols of "Optogan". The color temperature is 3050 K. A very soft and pleasant lamp, but the price bites - 995 rubles per piece. Who needs it for such money ?!

By the way, Optogan still has quality problems: endurance tests fail. He screwed / twisted a couple of times and got the following result:

Flimsy mount. Ladies light, what else to say!

SvetaLED

LED llamas of this company have not yet appeared on the Russian market, but they say that the price will be about 450-500 rubles. However, it came into my hands, packed in a stylish box (apparently, some kind of pilot batch), at which the temperature is 3500-4500K (it is like indicating that the equator is 35,000 km to 45,000 km long). The radiator is hidden under an aluminum hood (a trifle, but it's nice, as if you are holding an ordinary Ilyich light bulb, just a little “reworked”), and around an aluminum disk with mounted LED modules everything is covered with KT-8 thermopaste. They say that "Svetlana" somehow refers to the military, who apparently live by the principle of Jamie Heineman : "Doubt - lubricate!". For example, the Chinese lamp thermal grease applied only under the LED modules themselves.

Those who mercilessly beat the light bulbs Sveta LED and NoName from China say that the glass is quite fragile, and in terms of quality (a purely subjective assessment) is inferior to incandescent bulbs.

So the light bulb was picking ...

On the chip of the “OptoGaN” lamp
Tag #RusNT need to put!
And we will shine #RusNT
Both in September and February
(c) AP

A small photo report (the video camera for some reason refused to work) about how we disassembled the light bulbs:

It is necessary to approach the glamorous experiment glamorous! (Although all color matches are fictional)

The most important weapon is a hammer, how can it be without it ?!

Honestly, I tried, but the polycarbonate did not succumb. Everything was destroyed, the table, linoleum, aluminum radiator, but not polycarbonate, which was subsequently removed with a screwdriver. But the light, even in a semi-disassembled state, continued to burn.

Then it took a very long time to pick the driver, which is filled with some kind of polymer. As a result, both the driver and Optogan's pride — a monolithic LED chip — were pulled to the surface.

Driver

Below are all 3 drivers together. Rate the complexity of each of them ...

Top down: Optogan, SvetaLED and China

Let's start from the bottom. To be honest, I liked the Chinese driver: powerful capacitors, coils, a little transforming electronics (diode bridge, etc.). Everything is made very compact, because of what the lamp itself has a rather modest size. Also, a big plus is that all the leads are long, i.e. you can really "repair" the lamp! Or use the driver after the lamp life for some other purpose. Of course, most ordinary users don’t care, but still it can be attributed to potential advantages. The substrate itself with LED chips is mounted on 2 miniature bolts (after all, the Chinese ...), so that, in the literal sense, you can treat the lamp as a designer.

Chinese NoName LED Driver

The wiring is really very long ...

The lamp produced by OptoGaN has a very complex driver with solid capacitors and, as experts have convinced me, with a pulsed power supply (although all LED lamps must have such a power supply). At the same time, the driver itself, along with the light-emitting module, is a “chip” of the company and its main pride. It is rumored that the company will conduct R & D in the field of minimizing this driver and, possibly, in the near future will reduce the size of its giant light bulb to an acceptable size.

Pride of "OptoGaNa" - driver and light-emitting module - next to the main file - base

SvetaLED. Call it a driver language does not turn. Even China has some kind of "buns" that improve the consumer properties of the lamp (for example, protect against blinking), but there is nothing at all, except for a diode bridge, a fuse, a huge capacitor (10 μF, 450 V - a lot or a little It should be said that the energy stored in the capacitor is enough for the light bulb to be 1.5 minutes after the power is turned off) and, apparently, the load switch. Everything is so simple and primitive that at first I was a little surprised. The true brainchild of the gloomy Russian genius ...

Also the pride of the gloomy Russian genius ...

It is possible that the simplicity of performance is the trump card of the Sveta LED. Flickering with a frequency of 50 Hertz is most likely the average eye is unlikely to see, and there is no place for them to take there, since the powerful capacitor smoothes everything, and the phosphor cannot even so quickly highlight the energy pumped into it (no one canceled). From here the low cost of the lamp should flow ... hmm, but somewhere there is a catch, since the lamp is planned to be released at a price close to the Chinese equivalent, taking into account a one-time delivery to Russia!

NB It is important to remember that among other things, important and device-dependent driver parameters are: the ripple factor, which can adversely affect a person’s mental activity, and the background electromagnetic radiation, which inevitably arises from the use of various “rectifying” circuits. But that's another story ...

LEDs

So we got to the most tasty morsel of our research. There are many publications on the Internet ( one , two , three ), which compares the spectra of lamps from different manufacturers, their consumer characteristics (design, service life, etc.), but now we’ll go down a bit to get closer to the light-emitting elements lamps. I’ll just make a reservation that all 3 lamps of approximately the same power of 5-6 W (if you look closely at the technical characteristics of the Optogan lamp, we will find an image of this chip designed for 5 W, while the stated power of the lamp is 11 W) about the same light-emitting area. In total, we have a luminous flux per watt ( lumen n per watt): China - 70-90, Optogan - 65, Svetlana - 75. I think it is important if dear readers want to compare lamps with each other!

To be honest, to the Chinese LED, namely to the chip itself, I was filled with sympathy. The beauty of its internal structure is simply amazing. I was lucky: while I was tearing off all the layers from this LED, I accidentally damaged a large diode-chip, as a result of which a microstructured sapphire substrate was exposed:

Optical micrographs of the Chinese chip top view: the golden stripes on the chip - current-carrying contacts.

The layered structure of the light-emitting chip at maximum magnification on an optical microscope. The dark area corresponds to the sapphire substrate. Arrows indicate individual layers or groups of layers.

By the way, the chip itself is isolated from the outside world by at least 3 layers, but it seems to me that there are nevertheless 4 of them. The first is a polymer with a phosphor that turns part of the radiation in the blue region of the spectrum into yellow-orange. The second is a small layer of soft polymer, then a convex shell (a la lens) made of hard polymer, and two more layers of soft and hard polymers.

I would like to note that, compared to other lamps, the Chinese are as simple as possible. Only 4 wiring connect a large chip to the outside world (the rest of the lamps have much more), only 1 light-emitting chip per diode, which is already directly mounted on the board, correctly wired current-carrying contacts on the chip itself, allowing electric current to flow evenly across the surface ( which “Optogan” has a similar effect). I did not manage to find any obvious, significant shortcomings.

SEM images of a structured sapphire substrate

The layered structure shows that we are on the right path (a consequence of the chip creation method - MOCVD), but it is unlikely that we will be able to see the individual layers of the active region ...

Chip and contacts that feed it

«». , , – . . , «» «» 1 , , , «». , , «» , , . «» , , , , - , .

: , - , , ( ).

SEM-

. ( ~10 ) :

«»

, . , :

:

, , . , «» , «» :

SEM- ( )

, – ( )

,

– «LED». , , – – ! – ( ). , , «» , … , ! - . , , . , , … …

, , … – …

: . , . , «».

: -

: , «». , . , «» 1 , .

SEM- ( ). , (4 x3 ).

…

déjà vu?! .

, , «LED», : , , , , «» (, RU COM ). , , 1 , , - .

Scandals, intrigues, investigations…

- , . , , :
1. - «» «»:

2. «» «» . «» 5 5 , 2 «» 45 .. – «». déjà vu ?! ?!

, ,

«» (, «» ) ?! , . , : , , . , , , .

«» , , «» «» . , , «» , ( ) ( , - ?!). , 1000 . , – ( , , ..). , , v.1.0 v.1.1. , , – , , , 1 160 500 ( 50-80 ) «» «».

, «» , (), , 1 (, – ), , , , , «» .. .

«», , , , - . , , «» 1 (.. 12,5%), – . , , , .

. , ! , , …