1. Cs@C n K@C n Rb@C n Group I Group I Metallofullerenes Donate 1 electron to cage M + @C 2n − Distribution similar to empty cages C 60 and C 70 dominant n = 60 n = 70

2. Ca@C n Ba@C n Sr@C n Group II metals Donate 2 electron to cage M 2+ @C 2n 2− M@C 50 exhibits dominance as well as M@C 60 M@C 70 no longer dominant n = 50 60

3. Sc@C 2n Sc@C n Y@C n Group III metals Donate 3 electron to cage M 3+ @C 2n 3− Now, M@C 44 dominates M@C 50 is also abundant. M@C 60 no longer dominates n = 44 n = 50 n = 60

5. Clear correlation to charge transfer and growth More charge transferred, the distribution of fullerenes shifts to very small metallofullerenes Thus, when more charge is transferred they “grow slower”. As the number of transferred electrons are increased, the growth distribution is shifts to smaller fullerenes. This can explain why the larger metallofullerenes appear to only form as a small fraction of, for example, empty cage C 60. Under conditions where empty cage C 60 dominates of C 2n distribution, the M@C 2n distributions exhibits mostly small fullerenes…….which will likely “react away” in the solid state, in solution, or air. Charge transfer appears to be extremely important in determining metallofullerenes formation. It is likely the most important growth factor

6. C 84 + carbon vapor C 84 Metallofullerenes “grow slower” than empty cages in carbon vapor? + carbon vapor C 84 Compare Gd@C 80 to C 84 from our Nature Communications paper

7. Cs@C 2n Rb@C 2n Ca@C 2n Ba@C 2n Sc@C 2n Y@C 2n Cs@C 60 Cs@C 70 Ca@C 50 K@C 2n K@C 60 K@C 70 Rb@C 70 Rb@C 60 Ba@C 50 Ca@C 60 Sc@C 44 Ca@C 70 Ba@C 70 Ba@C 60 Sr@C 2n Sr@C 60 Sr@C 50 Sr@C 70 Sc@C 50 Sc@C 60 Y@C 50 Y@C 44 Y@C 50 Group I metals Donate 1 electron to cage M + @C 2n − Distribution similar to empty cages (ie, C 60 and C 70 dominate) Group II metals Donate 2 electron to cage M 2+ @C 2n 2− M@C50 exhibit much greater abundance, although M@C 60 just a bit more dominate. M@C 70 forms in lower abundance Group III metals Donate 3 electron to cage M 3+ @C 2n 3− Now, M@C 44 dominates over other endo cages. M@C 50 is also highly abundant. M@C 60 no longer dominates Comparison of Group I, II, III metallofullerene growth distributions

8. Metallofullerenes “grow slower” than empty cages in carbon vapor? La 2 @C 80 + carbon vapor m/z 1,8001,7001,6001,5001,4001,3001,2001,1001,000 La 2 @C 82 La 2 @C 80 C 84 + carbon vapor C 84 Compare growth of La 2 @C 80 to C 84 from our Nature Communications paper

9. C 84 + carbon vapor C 84 Metallofullerenes “grow slower” than empty cages in carbon vapor? + carbon vapor C 84 Compare Gd@C 80 to C 84 from our Nature Communications paper

10. Does more charge transfer from metal to cage render metallofullerenes less reactive? That could explain striking difference in growth of metallofullerenes vs empty cages Thus, as one moves from Group I to Group III metals, M@C 2n should become less reactive due to more transfer to cage. Thus, Group III would exhibit a greater distribution of smaller fullerenres than Group II, and Group I A good test of this is to look at the Lanthanides It is known there are two groups of lanthanides, those that transfer 3 electrons to cage  M 3+ @C 2n 3-  La, Ce, Pr, Nd, Pm, Gd, Tb, Dy, Ho, Er, Lu And those that only transfer 2 electrons to cage  M 2+ @C 2n 2-  Sm, Eu, Tm, Yb

11. La@C 2n Ce@C 2n Pr@C 2n Sm@C 2n Eu@C 2n Tb@C 2n Dy@C 2n Ho@C 2n Er@C 2n Nd@C 2n Yb@C 2n Lu@C 2n Gd@C 2n Comparison of Lanthanides La@C 44 La@C 50 La@C 60 Tb@C 44 Tb@C 50 Tb@C 60 Sm@C 44 Sm@C 50 Sm@C 70 Sm@C 60 M 3+ @M@C 2n 3− Lanthanides M 2+ @M@C 2n 2− Lanthanides

12. Striking difference for Sm, Eu, Tm, Yb  These are M 2+ @M@C 2n 2−  M@C 60 > M@C 50 >M@C 70  M@C 44 is weak  Matches Group II metal M@C 2n distributions (for example, Sr@C 2n below), which can, of course, only donate 2 electrons La@C 2n Sm@C 2n Comparison of Lanthanides La@C 44 La@C 50 La@C 60 Sm@C 44 Sm@C 50 Sm@C 70 Sm@C 60 M 3+ @M@C 2n 3− Lanthanides M 2+ @M@C 2n 2− Lanthanides La, Ce, Nd, Gd, Tb, Dy, Ho, Er, Lu all exhibit the same distributions  M@C 44 > M@C 50 >M@C 60  M@C 60 relatively weak for these M 3+ @M@C 2n 3− Lanthanides  Matches the growth distributions of the Group III metals….for example, Y@C 2n below Sr@C 2n Sr@C 60 Sr@C 50 Sr@C 70 Y@C 2n Y@C 50 Y@C 44 Y@C 50

13. Clear correlation to charge transfer and growth More charge transferred, the distribution of fullerenes shifts to very small metallofullerenes Thus, when more charge is transferred they “grow slower”. As the number of transferred electrons are increased, the growth distribution is shifts to smaller fullerenes. This can explain why the larger metallofullerenes appear to only form as a small fraction of, for example, empty cage C 60. Under conditions where empty cage C 60 dominates of C 2n distribution, the M@C 2n distributions exhibits mostly small fullerenes…….which will likely “react away” in the solid state, in solution, or air. Charge transfer appears to be extremely important in determining metallofullerenes formation. It is likely the most important growth factor