Interface, explanation

The classical architecture of an expert system comprises a knowledge base, an inference engine, and some kind of user interface. Most expert systems also include an explanation subsystem and a knowledge acquisition subsystem. This architecture is given in Figure 9-34 and described in more detail below. 

Corrosion Control. Sihca in water exposed to various metals leads to the formation of a surface less susceptible to corrosion. A likely explanation is the formation of metahosihcate complexes at the metal—water interface after an initial dismption of the metal oxide layer and formation of an active site. This modified surface is expected to be more resistant to subsequent corrosive action via lowered surface activity or reduced diffusion. 

Combination of Eq. 7 or Eq. 8 with the Young-Dupre equation, Eq. 3, suggests that the mechanical work of separation (and perhaps also the mechanical adhesive interface strength) should be proportional to (I -fcos6l) in any series of tests where other factors are kept constant, and in which the contact angle is finite. This has indeed often been found to be the case, as documented in an extensive review by Mittal [31], from which a few results are shown in Fig. 5. Other important studies have also shown a direct relationship between practical and thermodynamic adhesion, but a discussion of these will be deferred until later. It would appear that a useful criterion for maximizing practical adhesion would be the maximization of the thermodynamic work of adhesion, but this turns out to be a serious over-simplification. There are numerous instances in which practical adhesion is found not to correlate with the work of adhesion at ail, and sometimes to correlate inversely with it. There are various explanations for such discrepancies, as discussed below. 

In some cases, e.g., the Hg/NaF q interface, Q is charge dependent but concentration independent. Then it is said that there is no specific ionic adsorption. In order to interpret the charge dependence of Q a standard explanation consists in assuming that Q is related to the existence of a solvent monolayer in contact with the wall [16]. From a theoretical point of view this monolayer is postulated as a subsystem coupled with the metal and the solution via electrostatic and non-electrostatic interactions. The specific shape of Q versus a results from the competition between these interactions and the interactions between solvent molecules in the mono-layer. This description of the electrical double layer has been revisited by... 

For the explanation of macroscopic phenomena, the thickness of the phase boundary (interface) often plays no important role. As an example, we describe the movement of a phase boundary in two dimensions or the movement of a step edge on a crystal surface. We start with a Ginzburg-Landau equation [69]... 

A mechanism such as that given above provides explanations for the known effects of many process variables ". The reductive dissolution and undermining processes require access of the acid to the metal surface, hence the benefits obtained by the deliberate introduction of cracks in the oxide by cold-working prior to pickling. Also the increase in pickling rate with agitation or strip velocity can be explained in terms of the avoidance of acid depletion at the oxide-solution interface. 

Usually, the molecular strands are coiled in the glassy polymer. They become stretched when a crack arrives and starts to build up the deformation zone. Presumably, strain softened polymer molecules from the bulk material are drawn into the deformation zone. This microscopic surface drawing mechanism may be considered to be analogous to that observed in lateral craze growth or in necking of thermoplastics. Chan, Donald and Kramer [87] observed by transmission electron microscopy how polymer chains were drawn into the fibrils at the craze-matrix-interface in PS films [92]. One explanation, the hypothesis of devitrification by Gent and Thomas [89] was set forth as early as 1972. 

Another reaction mechanism, which is conveniently mentioned under this heading, is due to Hill [479] who suggested that ions (atoms or molecules) frorh the product may move through the dislocation network of the reactant and activate potential nuclei, particularly in the vicinity of the reaction interface. Thus a reaction zone, within which potential nucleusforming sites are activated, is developed in front of an advancing interface. With appropriate assumptions, this reaction model provides an alternative explanation of the exponential rate law, eqn. (8), which in Sect. 3.2 was discussed with reference to chain reactions. 

Fig. 5. Various dispositions of reaction interface which result in obedience to the zero-order kinetic equation. Product is shown shaded for explanation see text.

There have been several reviews of literature reports of compensation behaviour [36,521,522]. The observations made are relevant in the present context since kinetic characteristics of surface processes may be applicable also to changes proceeding at a solid—solid interface (i.e., two surfaces). Some of the explanations proposed for compensation behaviour (discussed in greater detail, with citations, in ref. 36) are that... 

Fig. 11. Schematic representation of dehydration followed by recrystallization of the product phase behind the advancing interface. This model provides an explanation of Smith—Topley behaviour (see text).

The above explanations, (i)—(iv), of S—T behaviour suggest that changes at the reaction interface may include the following factors which... 

In reviewing reported values of E for calcite decompositions, Beruto and Searcy [121] find that most are close to the dissociation enthalpy. They suggest, as a possible explanation, that if product gas removal is not rapid and complete, readsorption of C02 on CaO may establish dissociation equilibria within the pores and channels of the layer of residual phase. The rate of gas diffusion across this barrier is modified accordingly and is not characteristic of the dissociation step at the interface. 

The catalytic activity of doped nickel oxide on the solid state decomposition of CsN3 decreased [714] in the sequence NiO(l% Li) > NiO > NiO(l% Cr) > uncatalyzed reaction. While these results are in qualitative accordance with the assumption that the additive provided electron traps, further observations, showing that ZnO (an rc-type semi-conductor) inhibited the reaction and that CdO (also an rc-type semi-conductor) catalyzed the reaction, were not consistent with this explanation. It was noted, however, that both NiO and CdO could be reduced by the product caesium metal, whereas ZnO is not, and that the reaction with NiO yielded caesium oxide, which is identified as the active catalyst. Detailed kinetic data for these rate processes are not available but the pattern of behaviour described clearly demonstrates that the interface reactions were more complicated than had been anticipated. 

C, E curves have been obtained for Zn(0001) andZn(lOTO) at various crci with different additions of tu.630,634-636 The data for Zn(0001) at Cju = const have been used to obtain C"1, Ql plots. Nonlinear plots have resulted, with the value of the reciprocal slope remarkably dependent on ctu- At c-ru = 0 1 M, the reciprocal slope of the PZ plot is 1.1, increasing with decreasing c-ru Such an effect has been related to the weak specific adsorption of OH" on Zn. This explanation has been critically discussed by Vorotyntsev,74 who has assumed that the effect635,636 is connected with the variation in the compact layer composition of the Z11/H2O + TU interface as cjv varies. 

The reason for the exponential increase in the electron transfer rate with increasing electrode potential at the ZnO/electrolyte interface must be further explored. A possible explanation is provided in a recent study on water photoelectrolysis which describes the mechanism of water oxidation to molecular oxygen as one of strong molecular interaction with nonisoenergetic electron transfer subject to irreversible thermodynamics.48 Under such conditions, the rate of electron transfer will depend on the thermodynamic force in the semiconductor/electrolyte interface to... 

---