I've added a mq131 ozone sensor onto the detector.
It uses a heating element to move air through it and a reactive compound sensor to actually measure stuff. This type of sensor isn't very specific about the type of chemical it's reacting to. It'll likely react to several different gases to a lesser extent as well as ozone.
Originally it was being powered from the arduino board. The ozone readings seemed to drift and the other sensor readings seemed to be being affected in some way.
Most likely the heating element is using a relatively high current and this drain is changing the other sensor readings. I've added another power supply just for the heating element. This also has the added benefit of meaning that the battery for the main board will last longer and keep the other sensors working even if the ozone power has failed.