2.2 The challenge for innovation
For a Pirani sensor, the basic task is to provide a reliable measurement of pressure in a vacuum system as it varies from atmospheric pressure down to a value at least as low as 1 Pa. This statement can be further qualified by saying that unless its performance or cost is a fantastic improvement on the existing type, the micromachined sensor must be compatible with existing interface electronics, such that only minor modifications to its design are needed. This implies an electrically resistive sensor whose resistance varies with pressure in a similar way to the conventional hot-wire sensor. Vacuum systems are often operated continuously for weeks, even months at a time. Therefore, the sensors monitoring the vacuum must be built to last for many thousands of hours. Given that during the pump-down phases, at least, there will be a significant pressure of gas around the sensor, it must therefore be very resistant to chemical attack. The sensors must also be resistant to vibration (common in vacuum systems) and mechanical shock during transportation and installation.
From this short list, it is already clear that the specification imposes a number of mechanical, electrical and chemical constraints on the material properties of the sensor. By the time you've finished this next bit, you may well wonder whether it's worth all the trouble – and there's no doubt it is a lot of trouble – when you can do the same thing with a bit of wire, some insulation and some turned metal components. Well, you have to bear in mind that the processes I'm about to describe are being applied simultaneously to hundreds of similar devices on each wafer, and wafers are usually processed in batches of at least twenty-five, so the reward is that we make thousands, or tens of thousands, of sensors at once.
Now it's time to see how those constraints listed earlier are dealt with in the design and manufacture of a microengineered Pirani sensor.