Upper Lowenheim-Skolem Theorem
Fix a language L. Take a L-structure M of infinite cardinality and a cardinal κ≥∣M∣+∣L∣. Exists an elementary superstructure N of M of cardinality κ.
Take a language L and L-structure M satisfying the theorem requirements.
By1 a construction essentially the same as the Henkin construction, exists a model N⋆ of Diagel(M) and of size κ. Letting N denote the L-reduct of N⋆, by a lemma of diagrams there exists an elementary embedding j:M→N. Hence, the image j(M) is an elementary substructure of N of the desired size.
I think