CR-invariance of the Steinness index (Q2663067)

Let \(\Omega \subset \mathbb{C}^n\), \(n\ge 2\), be a bounded \(C^k\), \(k\ge 1\), smooth domain. The Diederich-Fornæss exponent of a defining function, \(\rho\), for the domain, \(\Omega\), is given by the supremum of all \(\eta\in (0,1)\), such that \(-(-\rho)^{\eta}\) is strictly plurisubharmonic on \(\Omega\) (if there is no such exponent, we define it to be 0). And the Diederich-Fornæss index of the domain, \(\Omega\), is given by the supremum over defining functions, \(\rho\), of \(\Omega\) of all Diederich-Fornæss exponents. The Diederich-Fornæss index of the domain, \(\Omega\), will be denoted by \(DF(\Omega)\). The author also introduced, in [J. Geom. Anal. 29, No. 2, 1583--1607 (2019; Zbl 1437.32006)], the definition of the Steiness index of a domain, which is the infimum over defining functions, \(\rho\), of Steiness exponents associated to \(\rho\), which in turn are defined to be \(\inf \tilde{\eta}\) for \(\tilde{\eta}>1\), such that \(\rho^{\tilde{\eta}}\) is strictly plurisubharmonic on \(\Omega^c \cap U \) for some neighborhood \(U\) of \(\partial \Omega\). If there is no such exponent for a particular defining function, the exponent of that defining function is defined to be \(\infty\). The Steinness index of \(\Omega\) is defined by the infimum over \(C^k\) defining functions of Steiness exponents. The Steiness index of the domain, \(\Omega\), will be denoted by \(S(\Omega)\). The paper relates the CR-invariance of both the Diederich-Fornæss index and Steiness index of a domain to a 1-form introduced by \textit{J. P. D'Angelo} in [J. Differ. Geom., 14, 1, 59--66 (1979; Zbl 0411.32008); Lect. Notes Math. 1268, 103--110 (1987; Zbl 0647.32011)] and given by \(\alpha_{\eta}:=-\mathcal{L}_T\eta\), where \(\eta = \partial \rho - \bar{\partial}\rho\) \(T=L_n - \bar{L}_n\), for \(L_n\) the vector field dual to \(\partial\rho\), and \(\mathcal{L}_T\) denoting the Lie derivative in the direction of \(T\). Let \(\Omega_1\) and \(\Omega_2\) be bounded domains in \(\mathbb{C}^n\) with \(C^k\)-smooth, \(k \ge 3\), boundaries. \(\partial\Omega_1\) and \(\partial\Omega_2\) are assumed to be CR-equivalent (there exists a \(C^k\)-diffeomorphism, \(F\), such that \(F(T_p^{(1,0)} (\partial\Omega_1)) \subset T_{F(p)}^{(1,0)}(\partial\Omega_2)\)). A main result of the paper (Theorem 1.2) states that \(DF(\Omega_1) = DF(\Omega_2)\) and \(S(\Omega_1) = S(\Omega_2)\). This result is proved by relating both indices to the D'Angelo 1-forms associated to the domains (in Theorem 1.1) and showing the 1-form itself is CR-invariant.