Electrodes ranging from 25 nm to
500 nm of RuO2 were hydrated in pH 7 buffer for one week to
eliminate any effects of ageing (as before). Electrodes were then equilibrated
for 1 hour at pH 12, and looped from pH 12 to 2 with pH 12 between each step.
Here measurement times were reduced form 180s to 90s, since examination of the
previous data showed that electrodes had fully equilibrated within that time
and had begun to drift. It should be noted that this reduction in loop time
should reduce hysteresis values (compared to the previous data) as less shift
should be incurred at pH 2 15, which is advantageous
from an end user point of view.
Results presented in table 8
reveal that as RuO2 thickness increases there is in-general an
increase in sensitivity, accompanied by a reduction in hysteresis and drift.
The increase in sensitivity is approximately linear between 30min and 400 hours;
after which the rate of increase slows as sensitivity approaches Nernstian;
which agrees with the trend predicted by Bousse (figure 5).
Examination of electrodes by
scanning electron microscope (SEM) reveals a pronounced change in structure as
RuO2 material increases. The 30min RuO2 electrode has a
relatively flat surface with few features; whilst the 4 hours electrode
exhibits a much more rough and grainy structure. This rough grainy structure
would result in a high number of defects in the
crystal lattice, providing a high density of RuIII sits, allowing
the electrode to attain a stable near-Nernstian response.
Hysteresis decreased markedly
between 30min and 3hours, followed by a gradual decrease as thickness
increased. This too agrees with the predictions made above. As a higher density
of active sites will make the electrode less sensitive to changes in
sensitivity from changes in active-site density.
There was no observable trend
for drift-rate; a large decrease and then the values were all approximately the
same. It should be noted that the drift rate values calculated here represent
the drift in potential over the experimental period; which was only 30 minutes.
As a result, they probably do not provide an accurate estimation of the
electrode’s true longer-term drift rate.
Also presented is a calculation
for the precision “a pH senor would hypothetically exhibit if it were
manufactured with RuO2 electrodes of different thicknesses; precision
was calculated by dividing hysteresis by sensitivity. It can be seen that to
obtain a precision of 0.01 pH units a 500nm electrode would be required (if an
“ideal” reference electrode was used). However, it can also be seen that
electrodes with thicknesses between 50 and 300 nm would provide precision
values better than 0.05; which is acceptable for many pH applications.