Sensors are on the rise: computer chips that measure air pressure, liquid flow velocities and a host of other quantities. One problem: how to protect these super-sensitive flakes of silicon against the rough environment they're supposed to measure.
They're so fragile, you'd better encase them completely: computer chips. But if your chip is a sensor, or measuring chip, a direct connection between the chip and the outside world may be unavoidable.
Sensors are applied more and more widely. They measure air pressures, flow velocities, chemical concentrations and other quantities, sometimes in rough physical of chemical circumstances.
Dr. Carmen Cotofana, a former electrotechnics PhD. student, worked out a new method to package sensors. Take your chip, engulf it in plastic, but leave open a small window, that's what it comes down to. ''People often ask: why didn't anyone think of this before?'' Cotofana reminisces. Last Monday, she defended her thesis, entitled 'Low-cost sensor packaging'.
''Normal chips are packaged using standard methods,'' Cotofana continues,'' the cheapest alternative is to cast them in plastic.'' Designing a mold for this is quite simple.
However, a similar standard packaging method doesn't exist for sensors. These often demand custom-made packaging, depending on the function and the environment the sensor is designed for. Often, in order to measure anything, there should be direct contact between the chip surface and the outer world.
The lack of a standard packaging method for sensors drives up the price, Cotofana declares, even to the extent that the packaging, not the chip itself, is the major hurdle to cheap mass production.
OWPP, Open Window Plastic Packaging, is the term Cotofana coined for the packaging procedure that she and her thesis adviser invented. The trick is in the nipple structure, a small rectangular part within the mold, touching the chip surface. When fluid plastic engulfs the chip, the nipple leaves a window between the chip and the outer world.
''It's a charming idea,'' Cotofana enthuses, ''but it's not at all simple to apply. You should be very careful not to damage the chip surface.'' The metal mold, for one, is much too hard to form the nipple structure.
Cotofana chose to make a nipple made out of silicones, a rubbery material more widely known from their application in breast augments, which are also supposed to be soft.
The silicone nipple structure is fitted in the mold, which subsequently can be used to package large amounts of sensors. ''Working with this kind of soft structure is quite new,'' Cotofana contends. In the past, silicones used to be nowhere near as easy to handle as they are now. ''You can do lots of things with it now. Silicones used to be known for not sticking to metals, but now there are types that stick very well.'' Along with silicones, the methods to handle them improved dramatically as well.
This may be the reason why nobody thought of OWPP before, Cotofana suspects. The time just wasn't ripe.
To prove the principle, Cotofana packaged a sensor of the type also used in digital video cameras. She left the light-sensitive part of the chip surface open, so that light could reach it. OWPP seemed to work, but her first chip malfunctioned when it was soldered to a print board. After an analysis of mechanical stresses in her package, the improved sensor worked without problems. OWPP turned out to work fine.
After she thought up silicone nipples, Cotofana invented a packaging method for sensors featuring ultra-sensitive structures. Miniature membranes or microscopic mechanical parts would even be damaged by the soft touch of a silicone nipple.
In this case, Cotofana thought, all you have to do is make a ring around the sensor surface when shaping the chip out of silicon. Using modern lithography techniques, that's an easy thing to do. The circular ring surrounds the sensitive structures like a dam surrounds a plot of land.
The silicone nipple is brought down upon the silicon ring, just like a lid covers a cake tin. The underlying space is left open, saving the sensitive structures when plastic is cast around the rest of the chip.
Now, Cotofana hopes that someone picks up the method. Several companies showed in interest in OWPP, she says, and fellow scientists asked about it repeatedly.
She herself has left the sensor circles for a while, working in a new job at Philips Semiconductors, in Nijmegen, though she remained faithful to her big interest: chip packaging.
''To be quite honest, Philips is lagging behind a bit in this area,'' the packaging expert concedes, ''a lot must be done here.''
Busy Cotofana had to write her thesis during evening hours, she reveals, not intending it as a complaint. ''It's hard work, here at Philips,'' Cotofana concludes. ''I like that.''