trends and challenges in thermal management field
george meyer, coo and cto, celsia technologies

the field of electronics thermal management is in its middle ages and one would think that things would be getting a little boring. they are not! looking back at the history of thermal management of electronics one can find reference material dating to mid-century, 1900’s. seems odd saying mid-century when discussing electronics thermal management! the history of thermal management followed a pretty typical technology development. first engineers discovered heat was an issue, did some research and had something machined to fit. sometimes it worked, sometimes it failed. traditionally, brown board testing was done, which means a new design was run and brown spots on the board were looked for. seriously, this was an early development process. once the demand for these cooling things developed so did companies to serve the market. most of these were the same machine shops that made the parts for the first prototypes. these companies, mostly western companies enjoyed nice growth during the 80s and 90s.  the next big bang was due to the dot com bust or at least set off by the bust. once the crash occurred, surviving oem companies set about saving money anyway they could to stay afloat. this included sourcing all of their hardware with the lowest possible suppliers, e.g. asian sources.

this was a severe blow to western suppliers and even those that had asian facilities as the lowest bids came from low overhead asian based companies. the effect of this was that any real research and development in the area of thermal management ground to a halt. low margins equal little investment in r&d. the good news is that this transition is behind us, the wounds have pretty much healed and there are serious challenges/opportunities in the thermal field again.

some of the challenges that we face today have more to do with energy usage/costs than component or system cost and several studies have shown that higher performing/higher cost thermal solutions actually have a rather short term payback in the reduction in electrical consumption. of course there is still a large part of the market where cost is the primary driver and the designs are priced out per pound.

let’s take a look at some of the challenges in the thermal market today. we will start by taking a look at what darpa is funding. darpa currently has several programs in the area of thermal management. these are the tgp, mace and nti programs.

nanothermal interfaces (nti)

the primary goal of this program is the development and demonstration of ideas based on novel materials and structures that can provide significant reductions in the thermal resistance of the interface layer (often called the tim) between the backside of an electronic device and the next layer of the package, which might be a spreader or a heatsink.”

thermal ground plane (tgp)

proposed research should investigate innovative uses of 2-phase cooling, as in common heat pipes, where the benefits include very high thermal conduction and extreme reliability in a light-weight, thin, 2-d package that is also engineered to match the thermal expansion of semiconductor substrates.

microtechnologies for air-cooled exchangers (mace)

the primary goal of this program is the development and demonstration of air-cooled exchangers that offer significant reductions in thermal resistance (from case to air) and significant reductions in the total electrical power used to force the air through the system.

now keep in mind that these programs are aimed at the requirements for the military and may or may not mirror the needs in the commercial world but it does shed some light on the challenges.

interface materials continue to be a challenge. this is due to several causes, one is the bulk conductivity of materials that meet the requirements of the market and the second is the interface of these materials with the semiconductor and heat sink materials. we have had improvements in both areas but not often both combined. that is, it is easy to improve the bulk conductivity of interface materials or the wetting of the materials but it is difficult to do both. the advancement of nanomaterials holds promise but too early to tell.

the two other areas being addressed by darpa, two phase spreading and advances on the heat exchanger side, may or may not have any impact on commercial applications in the foreseeable future. this is because the commercial world has already done a pretty good job of addressing these two areas and darpa is looking to squeeze more performance for very high end applications.

one area ripe for an improvement is in the area of air movers. over the last 20 years there has been steady improvement in fan design giving us long lasting, high performance air movers but the issue of noise is a huge problem. cooling electronics using air quickly reaches a limit due to noise constraints. there are several new products being presented recently claiming to make improvements, the most interesting to me are the miniature blowers being developed by bergquist.

liquid cooling

liquid cooling is often treated as the black sheep of cooling even though it offers the best performance next to direct immersion or spray cooling. liquid cooling is simply a way to move heat efficiently and if it was not effective your automobiles would be using some other type of cooling system. the drawbacks for liquid cooling are primarily in two areas. first and foremost is the risk of a leak. the best liquid cooling systems use water as a working fluid which we all know is not a good thing to have in an electrical system. from observations, the leak risk is more a human factor than a technical factor. liquid cooling systems can be designed to be leak free. tubing, fittings, etc, are available that meet the requirements for reliability. the problems occur when costs cutting is done or when user induced damage occurs. the second concern is over cost. a liquid cooling system may cost several times that of a passive system but the cost/benefit analysis needs to be done at the system performance level and not the component level. we will continue to see the adoption of liquid cooling. one key example of this is ibms move back to using liquid cooling. for a good look at a liquid cooling system that meets todays requirements take a look at the coolit web site.

one last note:  about every ten years developments come along that threaten the thermal management business; first there was cmos and most recently the word on the street is that low power arm chips will take a bite out of the thermal market. i am not suggesting that it is not possible, all i know is that every time there is a reduction in power the demand for system performance quickly heats things up again. how long do you think 4g will satisfy the consumers hunger for bandwidth?

how do today’s engineers keep up to date on advances in this field? there are several forums for doing this. the traditional ones are events such as semi-therm, meptec – the heat is on, imaps-atw, i-therm, interpack and coolingzone and new forums such as linkedin/thermal management group and other thermal engineering groups on linkedin. for engineers in this field attending these events should be a must, not only for technical knowledge but for networking. generally speaking, someone at these events has encountered a problem you are facing and knowing who you can contact to ask questions is very valuable, not only to your company but to you and your career.