X hits on this document





7 / 13

W C Maskell

EMF deviations previously reported (Fouletier 198213, Mogab 1973, Pizzini 1973). Where this occurs it can often be alleviated by raising the cell temperature. which results in desorption of much of the strongly bound CO.

2.7. Degradation ofperformance The performance of oxygen sensors deteriorates with time. This may involve either the degradation of the ceramic or of the electrodes or both. The former aspect has been discussed previously (Steele et al 1981, Maskell and Steele 1986). Electrodes may degrade for a number of reasons.

(i) At high operating temperatures (e.g. > 800 OC with Pt) a porous metallic electrode tends to sinter with a consequent decrease of surface area and concomitant reduction of the length of the three-phase boundary. These changes reduce the exchange current density on the electrode and its catalytic activity, and response may become sluggish.

(ii) When operating in atmospheres containing aggressive components, e.g. H2S, S02: Pb, dust etc, the electrodes may take up these contaminants and become poisoned. In particular, as was discussed in 0 2.5, HIS or Pb compounds may reduce the catalytic activity of the electrode: poisoning by the former is partially reversible while poisoning by the latter is irreversible. This contamination can result in a sensor that records an oxygen content closer to the actual oxygen level rather than the equilibrium value (Fleming 1977).

  • 3.

    Amperometric sensors

  • 3.

    I . General theory

The discussion in this section is taken principally from the work of Dietz (1982).

A schematic diagram of the limiting-current sensor is shown in figure 7. The device is used to pump gas X2 from the cathode to the anode, the processes occurring at the electrodes being


a n o d e X 2 + 2 V j ; + 4 e + 2 X f .


This equation assumes that X is divalent for simplicity but (34) may readily be transformed for alternative valency of the active species.

A porous barrier is fixed in front of the cathode to restrict transport of XZto the electrode. If a sufficient voltage is applied between the anode and the cathode then the partial pressure of X Zat the cathode is reduced to a value close to zero. This is the limiting condition and the current flowing, i,,,, is controlled by the rate of diffusion of X Zthrough the porous barrier according to Fick’s first law,

ill, =nFD(Q/L)cx,.


n is the number of electrons transferred per molecule of X Z

(n=4 for 0 2 in equation (34)) and D is the diffusion coefficient; Q is the sum of the cross sections of the pores of effective length

L ; C X , is the concentration of X Z in the sample gas. Equation (35) shows that the limiting current is proportional to the gas

Applied voltage


Figure 7. Schematic diagram of the limiting-current sensor. (Diagram courtesy North-Holland Physics Publishing, Amsterdam.)


concentration (compare equation (4) for the potentiometric device),

One of the problems involved in constructing practical devices is the very high diffusion coefficients of gases; e.g. Do, (in N2)is 16 and 150 mmZs-’ at 20 OC and 700 O C respectively. These values are four to five orders of magnitude higher than for

  • 0

    2 in aqueous solutions. Consequently, diffusion barriers must

restrict gases very severely, particularly since pump currents through solid electrolytes are restricted by the ionic conductivity of the ceramic. Dietz (1982) described various forms of diffusion barrier which were successfully employed for O2sensing.

The diffusion mechanisms of importance are bulk and Knudsen diffusion and the predominant type is dependent upon the pore size of the diffusion barrier. Bulk diffusion occurs where the pore size is larger than the mean free path (typically 10pm at atmospheric pressure): Knudsen diffusion predominates for

very small pores (e.g.

  • -

    10 nm). In the intermediate



types of diffusion contribute to molecular transport.

Dietz (1982) has considered the effect of temperature and total gas pressure, p, on i,,, for both types of diffusion. For bulk diffusion the limiting current is proportional to both the mole fraction, x,, of the active component X2 and to To’ and is independent of e.For Knudsen diffusion i,,, is proportional to both Teo5and topx,. It should be possible. by using a barrier with pores in the transition region between bulk and Knudsen behaviour, to obtain a characteristic showing little dependence upon temperature.

Characteristics of an amperometric oxygen sensor are shown in figure 8 (Dietz 1982) for 02-N2 mixtures. The limiting current plateau is clearly evident and i,,, is proportional to -yo,.



- - 4:

c 5 0 . t






0.5 Applied voltage ( V )


Figure 8. Characteristics of a limiting-current sensor where the diffusion harrier was a laser-drilled hole, 100,um in diameter. filled with porous ceramic. Percentage of 0 2 in N2: A, 1; B, 2; C, 3; D, 4;E, 5; F, 10. (Dietz 1982). (Diagram courtesy North-Holland Physics Publishing, Amsterdam).

Document info
Document views29
Page views29
Page last viewedThu Dec 08 06:09:19 UTC 2016