Was NMR/ESR spin flip-flop line ever measured?

Let $a$,$b$ be two unlike spins in a uniform magnetic field and RF excitation, and let $B_0$ be the magnetic field for which some transition energies of the spins coincide $E_a^{i,j}(B_0)=E_b^{k,l}(B_0)$. This is the classic condition for spin flip-flop due to dipolar interaction. Now we move away to $B=B_0+delta$, and for the flip-flop to occur, an RF excitation of frequency $f/h=E_a^{i,j}(B)-E_b^{k,l}(B)$ is required.
In an ESR/NMR B-f scan it should appear as an additional line.
It all looks like something that should be covered by theory in the 40s, and experimentally proven in the 50s, yet I see no sign of that.
For like spins it is indeed covered by double Larmor line literature.
What buzzword am I missing in my literature review?