I'm a little confused about satisfaction of formulas (in particular of sentences).
Consider a concrete example. Let $L=(\cdot, ^{-1},e)$ be the language of groups and consider its model $M$ with universe $C_3=\{1,g,g^2\}$ and standard interpretations. Let $\phi$ be the formula in the empty set of variables $1\cdot g=g$. Satisfaction is defined inductively. But what case are we dealing with? $g$ is not a term, is it? So $1\cdot g=g$ is not an atomic formula $t_1=t_2$, is it? Also, it's not T or F, nor is it the conjunction/disjunction of atomic formulas... (I didn't mention all cases, but the formula is also not of one of the types I didn't mention, as far as I can see).