Electrochemical osjgen gus sensors
0.5 0, / C O r o t i o
Figure 4. Calculated sensor EMF against bulk 021CO ratio assuming rapid mass transfer to the sensing electrode. A. slow surface reaction
( k and k~ small); B, intermediate situation; C, rapid surface reaction ( k and k D large): D, equilibrium curve. (Anderson and Graves 1981). (Reprinted by permission of the publisher,The Electrochemical Society. Inc.)
0,/ C O rotio
Figure 5. Calculated sensor EMF against bulk 02/CO ratio assuming dissimilar adsorption rates of 0 2 and CO. A, rapid mass transfer: B, intermediate situation: C. slow mass transfer (Anderson and Graves 1981). (Reprinted by permission of the publisher,The Electrochemical Society, Inc.)
the mass transfer coefficients of the reacting gases differ: the latter vary as where M is the molecular weight of the gas. Thus, mass transfer coefficients for O2 and H 2 differ by a factor
of four whereas for O2and C O the ratio is 1.07.
2.4.3. Experimental results. The EMF of a zirconia sensor with an air reference was monitored against the ratio of combustible gas to oxygen for four oxidisable gases by Takeuchi et a1 (1978). Their results, shown in figure 6, were in close accord with the numerical analysis of Anderson and Graves, showing a linear relationship between & (the /1 value at which the potential step was observed) and the diffusion coefficient of the combustible gas.
Interestingly, Anderson and Graves (198 1) found the potential step for 0 2 / C H 4 mixtures to occur at an anomalously low dc value when using a Pt electrode indicating the difficulty of fully catalysing the oxidation of CH,,
In practice exhaust systems are not as far from equilibrium as the one in the above laboratory measurements and major deviations of LC from unity would not. in general, be anticipated.
0 1 3 4 5 * x
Figure 6. Experimentallydetermined EMF of a zirconia sensor for
various gas mixtures. A, H2-02-h'2;
CO-02-N2; D, C~H:o-02-N2. iis the actual oxygen to reductant ratio compared with the stoichiometricoxygen to reductant ratio (Takeuchi et a1 1978: this paper was originally presented at the 1978 Fall Meeting of The Electrochemical Society, Inc. held in Pittsburgh,
2.5. Selectii>it). The response of the sensor to reactive gases is dependent upon the choice of electrode material. In many applications, such as combustion control, the required oxygen measurement is usually that pertaining to thermodynamic equilibrium: in such cases Pt is normally the preferred choice of electrode material because of its high catalytic activity. However, there are other situations where the actual O2 content of a reactive gas mixture is required. Sandler (1971) found that in C H 4 / 0 2 mixtures Ag showed negligible catalytic activity and responded only to the total O2content of the mixture. Haaland (1980) confirmed this result and investigated a number of electrode materials in various reactive gas mixtures. He found that Ag deposited on Pt was non.catalytic while Au was only slightly catalytic.
Pt poisoned 500-550 "C. introduction
with lead was particularly non-catalytic above
possible problem with this electrode is the
would be expected to sweeping the oxygen Pb-PbO couple is in Steele 1986).
result in slow response particularly when pressure through the value at which the equilibrium (see figure 6 of Maskell and
Solid electrolyte sensors are normally considered to be highly selective because the ceramic may be chosen with a single mobile anion, e.g. A"- and thus only the gas A, or gases that
can take part in a redox equilibrium with A e.g. sA+yB+zC
are expected to influence the sensor EMF. In most cases this is true but there are exceptions. For example, 0 ; may be sensed using a fluoride electrolyte (Ramanarayanan et a1 1979, Siebert and Fouletier 1983, Kumata er a1 1984). This is probably the result of oxygen ions being inserted into fluoride vacancies
~f + 2e +0 ; .
2.6. Blocking The foregoing discussion has assumed electrode reversibility. However, if a species is very strongly adsorbed onto the electrode surface it can block sites for adsorption so that the electrode is no longer fully reversible towards the gas being sensed. It appears that this situation can arise with C O on Pt at relatively low temperatures ( < 500 "C). Referring back to figure 4 and equations (28) and (29) it is clear that if most of the adsorption sites are blocked by CO then the 02/02reaction cannot contribute to the mixed potential. Further, the Oco term (0, in equation 30) dominates so that y and A V (equation 32) assume very negative values compared with those expected for a reversible system. This may account for the large anomalous