Tech-Talks Bregenz | Jan 30, 2015

Tech-Talks BREGENZ - Prof. Mehmet Arik, University of Ozyegin, Turkey

Thermal management is an ongoing challenge and a topic of continued research and innovation. Prof. Mehmet Arik from the University of Ozyegin, Istanbul, also known as the ‘Edison of the 21st Century’, discussed core thermal issues and new innovative solutions, trends in cost reduction, packaging and how we can learn from electronics to achieve better thermal management of the system as a whole.

LED professional: Could you tell us about your basic research and findings from your former activities in the USA?

Professor Arik: I worked at the GE Research Centre from 2000 to 2011. We started working on LEDs in the year of 2000 and my first assignment was on the thermal management and packaging of the chip and light engine levels of an LED lamp. I drove this research quite a lot and looked at problems at chip level, local hot spots and packaging to minimize the chip junction to board thermal resistance. As you know blue light to white light conversion is not 100% efficient, and we were looking to see if we have local hot spots in the LED chip as well as phosphor particles because of this. Later on, we also published a lot of papers on thermal management. My very first paper on thermal management of LEDs was published around 2002 and was one of the most cited papers on this topic! We also started looking at system level issues like how to place an LED light engine on top of a heat sink, what kind of materials we need to utilize and how to reduce the heat sink size by novel microfluidics cooling technologies like synthetic jets. If you want to produce 1500 lumens you needed to use 10-15 LEDs and if some of them are not attached well, some are getting more electrical current; some will heat up more than other components. The only way to abate this problem is using a liquid cooling system. We published a lot of papers on this research. For the last three years of my time at GE, I was supported by the Department of Energy (DOE) to develop the first LED active cooling luminaire to generate 1500 lumens and it was published by DOE as a success story. In 2008, we were able to get half the size, half the volume, half the number of LEDs and develop the whole system for a bright demonstration. Right now, I have started a research group (ARTgroup) at the University of Ozyegin, Istanbul, working on new and novel ideas like active cooling for LEDs, fluid mechanics, heat transfer, liquid cooling systems and looking at local phosphor particles cooling to name a few.

LED professional: Did you get the name ‘Edison of the 21st Century’ based on your research at GE?

Professor Arik: When I developed the LED lamp, GE selected my technology as one of the best technologies at the time. They made a press release on the day of the invention of the first Edison light bulb, so Jeffrey Immelt, CEO of GE, gave me this title! I may not be as smart as Edison but I was able to drive a technology from scratch to product and GE was very supportive of the journey. I will always be thankful to GE and the great team behind these inventions. I always tell my students: if you have an idea, go ahead and never turn back. Always try, work hard and make it happen.

LED professional: We have heard time and again that rather than exploring new ideas or new applications, a lot of research tends towards replacing the incandescent light bulb. What would you say to this trend?

Professor Arik: I attended DOE meetings in the USA, and I see it as a great research, application and implementation vision-driven meeting. It is not easy to change people and adopt radically different technologies! In the USA, people change homes very often and some houses are over a hundred years old. There is an existing infrastructure for retrofit lamps. New constructions may take into account new LED bulbs or designs but would you really spend thousands of dollars replacing old structures to install new and bright LED lamps? I think retrofit can win only if it is made cost effective or by better educating people of its clear advantages.

LED professional: It is a common consensus that LEDs are getting more and more efficient. This will continue and maybe plateau by the end of the decade. So do we really need thermal management if the efficiency keeps rising?

Professor Arik: Yes, LEDs are getting more efficient, but at the end of the day they are solid-state devices and they are temperature limited. We have learned our lessons from electronics when we went from vacuum tubes to bipolar and CMOS technologies, and today even though we have very advanced manufacturing techniques for transistors, we still have thermal problems getting more and more aggressive each day. LEDs will come to a certain efficiency level and we will then look at how to drive them harder, get the number of LEDs lower and lower the cost.

LED professional: So would you say that engineers would always go to the limit and try to find cost optimization etc. but that thermal management issues will always exist?

Professor Arik: Basically we want better and better in compact volumes! The comment that I made to Marc McClear, VP Sales of CREE, at the LpS 2014 opening ceremony is that electronics is a very old and a more mature industry. So we can look at electronics to see what they had done and learn from them. LEDs are a relatively young technology and there are some differences but a lot of similarities. We will still have thermal problems because we will never have 100% efficiency. The problems will be different and we will need cost effective solutions.

LED professional: Could you elaborate on some options for thermal management and the limits?

Professor Arik: These are semiconductor devices with temperature limitations and if operated over 100°C, lifetime is drastically reduced. You can operate them at 150°C or 200°C, but they will die faster because high temperatures are always difficult for many materials and engineering. I believe 100°C is the allowable temperature for electronics but higher temperatures will cause reliability issues. If we are able to, we will always use passive cooling in any design and we will tend to avoid active cooling e.g. synthetic jets, fans, piezo fans, thermo electrics, etc. If you don’t have enough surface area to remove heat by convection combined with radiation, you need a heat sink and we will almost always use a heat sink over 400 lumens tight space LED applications. The majority of current heat sink technologies are based on metals that are heavy and large. Therefore, we need to come up with an unconventional way of cooling, leading to innovations, like lightweight heat sinks. Some companies are working on plastics filled with graphite particles or some boronitride or metal particles. Ceramics may also be an option. We would like to reach a certain thermal performance goal, and it needs to be lighter weight with a compact size option. Thermal technologies are progressing very rapidly for electronics, and we can leverage that for LEDs. When I am asked about cooling solutions, I offer that electronic cooling equals to LED cooling. It is all connected and we need a solution that is highly reliable, low cost, easy to use, easy to maintain and available everywhere.

LED professional: One part is the light source and LED itself, but what about the quality of the driver and electronics at the moment?

Professor Arik: The weakest part in the electronics, to the best of my knowledge, is the capacitor and capacitors need to be reliable. You have to keep the capacitors at 80°C unless you use very expensive ones. At the end of the day, you are selling a product that is competing with a 30-cent incandescent light bulb or $2.00 (US) fluorescent bulbs, so you need to make sure you can bring cost down or provide something else that people will buy. A lot of times when I look at systems, they are not very well thermally managed as a whole! We do have some research programs where we would like to integrate LEDs with driver electronics and have a whole system level thermal management to reach ultimate lumen and lifetime goals.

LED professional: Do you have any comments on the power LED trends and thoughts on significant lifetime improvement options?

Professor Arik: I think it is great that we can talk about 50,000 hours; that is 8 hours a day for over 17 years! I don’t think we have to go further but we have to work on making it cheaper. If you consider the analogy, a computer chip costs less than 5% compared to the cost of the whole system. The cost should be reduced by advanced manufacturing techniques and we should be able to drive them harder, thereby reducing the number of LEDs needed in a system. About four years ago we were only driving them at 400 mA now we can drive over 1000 mA. So power LEDs are the way to go, and in the future we may only use one LED to get 500 - 700 lumens and it may be 2 x 2 mm2 or even smaller in size! It will be a thermal and optical challenge and that will also be a major advancement in LED technology.

LED professional: We had a presentation at LpS 2014 about LiFi and combining light with communications. What do you think of this trend?

Professor Arik: LiFi or visible light communications is an amazing technology and I believe that it will provide tremendous advantages in some areas! It also does not have the health hazards, compared to current electromagnetic waves and WiFi and it is very secure. I think it is the future and will be a part of the system. I would expect some products in a 3-year time frame. I could imagine smart controls, LiFi and maybe a technology using a bird’s eye camera to take into account human factors. Each person needs or likes a different kind and amount of light. Light and color affect our performance and could help or hinder our lives, giving us better or poorer quality of life. What if you could control the lighting with your smartphone? LiFi would be limited in the beginning, but if we can find a way to generalize that, and if we can use this for communication, then why not.

LED professional: Could you comment on communication standards and what might be the direction to go?

Professor Arik: I am not a controls expert but I will comment from an engineering point of view. There are a lot of control standards and methods in thermal management, reliability or lifetime and we already have a vast amount of knowledge in the electronics industry. We shouldn’t reinvent the wheel. We can do smart controls. I have seen some good examples at the LpS 2014 Exhibition. It is just fantastic that people are working in this field and sharing their knowledge with the community.

LED professional: You are also working in packaging trends. What can you tell us about these?

Professor Arik: I like Chip on Board (CoB) and we first proposed CoB in 2002 at GE. Package on board is fine but we just need an LED light engine and a PCB. In earlier days, people had a hard time handling 1mm2 chips; today you have affordable pick-and-place machines that can handle small chips in the thousands! If you look at conventional packages from well-known companies, the thermal resistance is 5-7 K/W. However, if you put the LED on the PCB directly, you can get rid of that. That means if you drive an LED at 2.0 - 2.5 W, in the first case you could have a 15°C temperature drop! That is a lot of sacrifice. It could increase the light output by 5%! As long as we can get the best chip, I would use it in CoB.

LED professional: On a more general level, after over a decade in an industrial research laboratory working very closely with lighting, medical, energy and aviation industries, you have now started teaching and performing research at a university. Do you still work with the industry and how is this synergy working out?

Professor Arik: It is working out great and is very fulfilling! For the last two years, I have been very busy looking to grow my network especially in Europe, for research funding and training students. Very few people knew about the fundamental of LEDs when I first moved to Turkey so I am teaching the basics to a lot of students and industrial partners through the EVATEG center at Ozyegin University. We just spent over 1.1 million dollars to put in the infrastructure at the university. The Turkish government, Turkish National Science Foundation (TUBITAK), European Union FP7 program, Istanbul Development agency (ISTKA) and a number of local companies mainly funded this. The lab is used for LEDs and electronics packaging. If I want to look at a 10-micron hot spot on a chip, which is similar to a 10-micron hot spot on a computer chip, I can use the same IR thermography. We don’t work with many European or US companies yet, but we are just opening up our doors towards Horizon 2020, US collaborations and Asian companies. So EVATEG is really a global center for anyone that wants to come and do groundbreaking research, or work on a practical problem. We are encouraging researchers, graduate students and support staff, trying to give them some exceptional incentives to solve problems.

LED professional: You are at the LpS 2014 for the first time. How has your experience of the Symposium been and what are your comments?

Professor Arik: I think it is very nice, well thought out and very professionally organized. The quality of the conference and presentations are very high and the speakers are well selected. The Expo is nice in terms of the companies and I am certain it will grow to bring more people from around the world and become even more impactful.

LED professional: Thank you very much for your comments and insights!

Professor Arik: Thank you! It has been a great experience to be at LpS 2014 in Bregenz.