The electrical/electronics content in the average vehicle will increase from 22% of total value now to 35% in the next 10 to 15 years, projects Mike Gauthier, director, corporate technology, Siemens VDO Automotive. And a lot of that will come from an increase in sensors. "Sensors in the vehicle will increase at up to 25% annually," he predicts. That's just for vehicles with traditional internal combustion engine-based powertrains; add a hybrid electric system (Gauthier: "By 2015 most vehicles will be offered in a hybrid version.") and the number could more than double. But most of the sensors used in the future will probably not be the comparatively bulky magnetic or proximity-type sensors found in cars today, but Micro-Electro-Mechanical Systems (MEMS) that can be 0.001 the size. MEMS are made up of mechanical microstructures, sensors, actuators and microelectronics integrated together in the sealed environment of a silicon chip. Though they essentially mimic the workings of their larger predecessors, because the scale of the components is so tiny the response time to inputs is almost immediate. This makes MEMS good candidates for tasks that require reactions measured in milliseconds like air bag deployment. Their sealed configuration helps them withstand the temperature and vibration extremes of the engine compartment (some are already being employed in areas like engine knock detection). Because they are so small, MEMS can be put virtually anywhere in a vehicle without much concern for packaging problems. For example, some sensors used for stability control systems need to be placed as close to the central mass point of a vehicle as possible, which has often been a problem with larger sensors. According to Gauthier, using MEMS in this application effectively solves the problem by integrating the yaw, pitch and roll sensors into one tiny package.
The latest generation of heated and cooled seats in vehicles operate using Peltier devices that convert electricity into temperature differentials without resorting to compressors or liquid coolants. Thermotunneling diodes do just the opposite; they convert temperature differences between two surfaces into electricity; the greater the temperature variance the more electricity generated. Unlike Peltier devices which are only about 7 or 8% efficient, Gauthier says thermotunneling diodes can potentially recover up to 70% of the lost heat generated by a vehicle's engine. To put that into perspective, Gauthier calculates that for every 100 kW of wheel-driving output, another 100 kW of waste heat goes through the exhaust and 60 kW more is dissipated by the radiator. Coat the radiator and exhaust system with thermotunneling diodes and you can more than double an engine's efficiency. In fact, you can eliminate the pumping, friction and incomplete fuel combustion loss of an engine altogether by replacing it with a small furnace coated with diodes that generate the power needed to run an electric powertrain. Gas mileage would soar, emissions would plummet. Sound too good to be true? It is. At least at the technology's current maturity level.
The problem with thermotunneling diodes is that in order for electrons to undergo the quantum physics effect of thermo-tunneling that generates electricity, the two surfaces involved must be consistently held about one nanometer apart without touching. A company based in Gibraltar called Power Chips plc has developed a method of electrodepositing layers of metals (silver and titanium are the cheapest ones used) on a substrate and then thermally shocking it so that it snaps at the interface of the two materials. This produces surfaces that are microscopically rough, but fit together perfectly and can be adjusted with piezo devices to achieve the necessary one-nanometer gap. Maintaining that gap cost-effectively in a mass production environment and then later in a moving vehicle is the biggest hurdle now. But if that can be overcome, Gauthier says it means, "Bye, bye internal combustion engine."
Sometimes a technology can be less powerful or effective than the one it replaces but still succeed because it is really, really cheap. That is basically the case with organic semiconductors. Gauthier estimates that the manufacturing costs for these semiconductors, which use plastic as a substrate, could be as low as 0.01 that incurred by silicon-based semiconductors. One big reason is facility cost. Currently a silicon wafer fabrication plant costs around $3.5 billion because of the expensive photolithography and etching equipment needed to create the minute pathways that carry electrons through each chip, equipment sitting in large, assiduously monitored and maintained clean rooms. Organic semiconductor production would do away with all of that and substitute a process that resembles the continuous-feed printing of a newspaper. As it is currently envisioned, rolls of plastic substrate would be fed into presses and then printed with semiconductor "ink." The resulting flexible chips would not be as powerful as their silicon counterparts because the electron pathways would have to be bigger and consequently fewer, but for low-demand computing tasks they could prove attractive low-cost alternatives. "If you are talking about something like an automatic air conditioning system you don't exactly need a Pentium 4 to control it," quips Gauthier, "You can get by with a Commodore 64 processor running at 2 gigahertz." And with the number of microprocessors in vehicles growing rapidly, plastic Commodore 64-level processors may be just what automakers need to help keep electronics costs down.