constant level of relative humidity (%RH). This operation is neces-
sary in order to minimize the contribution of the water vapour to
the signal of the sensors. Fig. 1 shows the used assembly which
was made up by a bottle with some distilled water inside closed by
a stopper with two holes. The synthetic dry air from the cylinder
was split into lines before the humidity regulator: one line enters
the bottle, while the other half bypasses the bottle. Both lines were
regulated by a needle valve. The two lines were rejoined after the
bottle to go into the electronic nose.
The response of the system after samples injection implies a
reduction in the sensors resistance and a subsequent recovery with
reference air. For the measurements, this process is divided into
four steps: before sample injection, during sample exposure, after
sample exposure (recovery) and a waiting time for complete signal
stabilization. These two last steps are similar in terms of sensor
response and are only used to optimize the operation time (waiting)
maintaining a minimum recovery interval (after sample exposure).
In our case, the time of each step was set as follows: before 60 s,
during 60 s, after 60 s and wait 420 s. Each sample analysis lasted
600 s and was analysed in triplicate.
Volatile compounds were directed to the sensor chamber by the
carrier gas at a 150 cm3/min flow-rate while the temperature of
the sensor chamber was kept at 55 ◦C. With the exception of the
“during” step, the gas valve was maintained in a switched position
and only carrier air was blown into the sensor chamber; this caused
the baseline of the sensor signal to return. Ambient air filtered with
activated silica and charcoal was used as reference gas during the
recovery phase of the measurement cycle. The previous conditions
ensured that the baseline reading has indeed been recovered before
performing the next analysis.