In the field of solid state physics, the electron affinity is defined differently than in chemistry and atomic physics. For a semiconductor-vacuum interface (that is, the surface of a semiconductor), electron affinity, typically denoted by E_EA or χ, is defined as the energy obtained by moving an electron from the vacuum just outside the semiconductor to the bottom of the conduction band just inside the semiconductor:^[6]

${\displaystyle E_{\rm {EA}}\equiv E_{\rm {vac}}-E_{\rm {C}}}$

In an intrinsic semiconductor at absolute zero, this concept is functionally analogous to the chemistry definition of electron affinity, since an added electron will spontaneously go to the bottom of the conduction band. At nonzero temperature, and for other materials (metals, semimetals, heavily doped semiconductors), the analogy does not hold since an added electron will instead go to the Fermi level on average. In any case, the value of the electron affinity of a solid substance is very different from the chemistry and atomic physics electron affinity value for an atom of the same substance in gas phase. For example, a silicon crystal surface has electron affinity 4.05 eV, whereas an isolated silicon atom has electron affinity 1.39 eV.
The electron affinity of a surface is closely related to, but distinct from, its work function. The work function is the thermodynamic work that can be obtained by reversibly, isothermally moving an electron from the vacuum to the material; this thermodynamic electron goes to the Fermi level on average, not the conduction band edge: ${\displaystyle W=E_{\rm {vac}}-E_{\rm {F}}}$. While the work function of a semiconductor can be changed by doping, the electron affinity ideally does not change with doping and so it is closer to being a material constant. However, like work function the electron affinity does depend on the surface termination (crystal face, surface chemistry, etc.) and is strictly a surface property.
In semiconductor physics, the primary use of the electron affinity is not actually in the analysis of semiconductor–vacuum surfaces, but rather in heuristic electron affinity rules for estimating the band bending that occurs at the interface of two materials, in particular metal–semiconductor junctions and semiconductor heterojunctions.
In certain circumstances, the electron affinity may become negative.^[7] Often negative electron affinity is desired to obtain efficient cathodes that can supply electrons to the vacuum with little energy loss. The observed electron yield as a function of various parameters such as bias voltage or illumination conditions can be used to describe these structures with band diagrams in which the electron affinity is one parameter. For one illustration of the apparent effect of surface termination on electron emission, see Figure 3 in Marchywka Effect.