Naked in an LED Factory

30 May 2013

Interview with

Barry Dennington, COO of Plessey Semiconductors

Technology developed by Professor Colin Humphreys at Cambridge University has been liscensed to a company called PlesseyNaked in an LED Factory Semiconductors, who have set up what is currently the UK's only LED factory. Naked Scientist Dominic Ford hopped on a train to find out how the technology is being commercialised.....

Dominic - To see Plessey's facilities for myself, I'm heading to Plymouth to meet some of the people behind the venture. Hi, Dominic Ford from the Naked Scientists.  I'm here to see Barry Dennington.

Receptionist - Hi, that's right Dominic. Come with me.

Dominic - So, here at Plessey, you're bringing Colin Humphrey's technology into industry.

Barry - That's correct. So, Colin's technology is the growth of gallium nitride on silicon and we are taking that technology and we're scaling up to be able to build more than one wafer in a single reactor run, going into high volume production, so we're leveraging the wafer size, the improved yield to get very cost competitive products out to the world market.

Dominic - So, I think this is the first factory making LEDs on silicon in the UK. I guess we think a lot about electronic products being made in the far east. Can the UK's manufacturing industry compete with that?

Barry - Absolutely. The first thing is that we have a unique technology which is not available in Asia right now. So, that's a breakthrough that's disruptive for the LED market. Furthermore, it's the machinery that creates the volume and we have the machinery here and the capacity to increase that machinery so that we can increase the volume.

Dominic - So, can I see some of the products you're making here?

Barry - Yes, of course. We have a demonstration table over here which shows you some of the applications that we have.

Dominic - So, you've got a wafer here, a couple of centimetres across and you got this - what looks like a yellow plastic dome over the top. Is it that dome which is converting the white light to blue light?

Barry - So, the dome is a plastic dome which is impregnated with phosphor and it's the phosphor impregnated in the plastic dome that converts the blue light to the white light. The phosphor dome, sits over a small printed circuit board and on that printer circuit board, there are 8 LEDs mounted. Elsewhere in the bulb, we have a lot of attention paid to heat management. So, there's a rather large heat sink with lots of fins to increase the surface area and finally, we have the driver circuitry. The LED is a 3.3-volt component. We take 240 volts at 50 hertz in and the driver is a voltage converter to make sure that we're getting the right amount of current and the right voltage to the LED to allow it to illuminate.

Dominic - So, we've got these fins here to dissipate the heat. How efficient is this light bulb?

Barry - This is a 60-watt equivalent light bulb. That doesn't mean to say 60 watts go in to it. In fact, only 12 watts go in to the light bulb. So, around 8 watts of heat is dissipated from the light bulb.

Dominic - So, this bulb is producing an equivalent amount in light to what an inefficient, 60 watt incandescent bulb would produce.

Barry - That's right. So, this is producing an equivalent amount of light to a 60-watt incandescent bulb with 1/5 of the energy going in.

Dominic - So, can we go and take a look at the production line where you're producing these bulbs?

Barry - Of course. The people who work in the MFA, this is the manufacturing line we make the magic LEDs in, they're the biggest source of contamination that there is. And so, in order to keep contaminants to a minimum, we have to make sure that all of our external clothes are covered by these antistatic coveralls.

Dominic - So, what is the problem that dust can cause?

Barry - So, the dust can cause open circuits or short circuits on the LED and simply cause LEDs not to function. So, there is a specific order to gowning up.  The first is the face mask.

Dominic - Okay, so I will pop this over my head and I've got a sort of veil in front of my face now.

Barry - Yes, perfect. Next then is the overall.

Dominic - Quite some suit this isn't it?

Barry - It is, yes. It's a little bit like a clean room onsie. So, the suit itself is impregnated with very fine wires and that stops static.

Dominic - I'm certainly feeling quite well-dressed now.

Barry - And in a moment, you'll be feeling very warm as well.

Dominic - It's quite incredible lengths you go to to keep the cleanroom clean.

Barry - The cleaner the cleanroom, the higher the yield on the product and the more cost competitive we can be. We now have to enter the cleanroom through our air shower. The air shower is there just to make sure that any loose debris of particulates that might be on the suit are blown off before you get into the cleanroom. We do a little pirouette as we go through and now, we can enter the cleanroom.

Dominic - So, I guess my picture of a cleanroom is they're white and clinical. It really is very white and clinical in here, isn't it?

Barry - Very much so, many times cleaner than any surgical operating theatre.

Dominic - So here, we've got a glass box with in fact some gloves coming out of it. What are we looking at here, Barry?

Barry -  This is the chamber inside which the gallium nitride is grown on the silicon.

Dominic - So, this is the process where you're performing the technique that was pioneered in Cambridge.

Barry - Yes and we've industrialised it and commercialised it.

Dominic - So, I can see you've got what looks like a laser in there. What's happening in this chamber?

Barry - The system delivers all of the gases and chemicals that are required to grow the gallium nitride. And in fact, the process uses vast quantities of hydrogen and ammonia gas as well as metal organic sources to bring the gallium, the indium and the aluminium into the reaction at the appropriate times.

Dominic - I get the impression you can't say too much about exactly what that recipie is though.

Barry - I could, but I have to kill you afterwards.

Dominic - Probably best not then.

Barry - Probably best not. This is the machine that we use to print the patterns onto the wafer.

Dominic - So, this is basically printing the electronic circuits onto those silicon wafers.

Barry - You can think of it as a very large camera and it images the pattern which is contained on a quartz master, images it onto the surface of the wafer and enables us to make the patterns on the wafer as small as about 0.6 microns in size. But it's very important to put those patterns in the right place. So, the placement accuracy of the patterns is even finer, to within about 1/10 of a micron. So, once the pattern is printed onto the wafer, we then go through an etch process to selectively remove some of the GaN layers in order that we can make contact to the anode and contact to the cathodes.

Dominic - So that you can put the wires into the final product.

Barry - Yes. So, we have a number of etch technologies available to us.  This is our GaN etcher. This contains all the chemicals and the gases that you need to selectively etch the Gallium Nitride itself. Once the gallium nitride is etched, we then need to make electrical contact to the anode and to the cathode, and we use various metal schemes in order to make those contacts. We have a number of different technologies for depositing those metals. We have a technology we call evaporation and that's literally where the metal is heated up with an ion beam and the metal evaporates into a cloud and that cloud then condenses onto the surface of the wafer. It's little bit like seeing snow form, small crystals of frozen water come together to make a film of snow, very similar to the way we deposit metals onto the GaN on silicon wafers.

So, in this bay, this is our evaporator tool. The evaporator we use to deposit metals onto the surface of the wafer, mostly precious metals such as gold and silver.

Dominic - So, we've got what looks like a very strong metal door in front of us.  I guess I find it incredible given what lengths you're taking to keep everything clean in controlled conditions, how many LEDs you're managing to get through this plant. How often, for example, would you be opening this door and putting new wafers in?

Barry - So, in this particular tool, a run would take in the region of 2 hours and 9 wafers would be deposited in a single run. So, 9 wafers with 15,000 LEDs per wafer is 135,000 LEDs every 2 hours through this particular tool.

Dominic - So, in terms of going to a supermarket and buying a light bulb, when do you think I might see a Plessey LED based light bulb in the supermarket?

Barry - I think possibly first quarter next year and of course, you can buy LED lamps in the supermarket today, but they are very expensive. This technology will enable the price breakthrough to be achieved, reach their strategic inflection point where people will naturally choose an LED over and above any other kind of lamp.

Add a comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.