Ivl ve~l tlO ~le$
Inventiones math. 56, 231-238 (1980)
mathematicae ~ by Springer-Verlag 1980
,-Regularity of Exponen...
26 downloads
453 Views
342KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Ivl ve~l tlO ~le$
Inventiones math. 56, 231-238 (1980)
mathematicae ~ by Springer-Verlag 1980
,-Regularity of Exponential Lie Groups J. Boidol UniversitSt Bielefeld, Fakult/it fiir Mathematik, UniversitMsstrasse 1, D-4800 Bielefeld I, Federal Republic of Germany
Let ~r be a (*-semisimple) Banach-*-algebra and let C*(,~) be its C*-hull. Let Prim C*(.~t') be the primitive ideal space of C*(,4) and P r i m , , ~ the space of kernels of topological irreducible *-representations of ~_J in Hilbert space, both equipped with the Jacobson topology. Then we have a canonical mapping
7/: Prim C * ( , ~ ) ~ Prim,,~' ..r
J c~'
which is continuous and surjective. We want to call a (,-semisimple) Banach-*algebra ~J *-regular, if 7j is even a homeomorphism. Similarly we call a locally compact group G *-regular if its group algebra L~(G) is ,-regular. In [2] the problem was posed to determine the class [7j] of all *-regular locally compact groups. The main results were (A) I f G is *-regular, then it is amenable. (B) All groups G with polynomial growth are *-regular. (C) All semidirect products G =-A ~ N with abelian separable A and N such that IV/A is a To-space are *-regular. Furthermore it was shown that all connected simply connected solvable (real) Lie groups G with dirn G < 4, G 4=G4, 9(0) = exp 94.9 (0) in the notation of [1, p. 180] are *-regular. Then in [3] it was shown following a suggestion of D. Poguntke that G=G4,9(O ) is not *-regular. Thus there is exactly one connected simply connected solvable Lie group G with dim G _k
m I (X) = m1(X ) + (ad X) J mo(X ) = (ad X) .i re(J). Since ad X is nilpotent m o d u l o p there exists k o >_ k such that for all j>= k o
(adX)Jm(f)~_p,
hence m l ( X ) = [ X , mdX)]c_[p,p ].
We conclude that
(m1(X)) = m(./)~' _ [p, p] c_ k e r n f which contradicts our assumption. N o w let X e g ( f ) such that a d X is not nilpotent m o d u l o p. Let i o be the largest index such that ad X operates with nontrivial eigenvalue e on [io+l/tio. We get that tr ad X - tr (ad X)1% = tr (ad X)I%+ 1 - tr (ad X)I% 4=0, since Reck4=0 also, by the fact that .q is exponential. Thus with h=[io the proposition is proved. Proposition 4. The set {m(f)~'~'[feg *} is a finite set oJ ideals (~f g.
Proof. re(f) is an ideal of 9 and m ( f y a characteristic ideal of re(f), hence the re(f) ~' are ideals of 9 for all f e g * . We prove that even the set M TM
= {a~l[g, 9 ] - a