Constant angle surfaces in product spaces (Q414522)

Angles between linear subspaces \( U \) and \( V \) of a Euclidean vector space \( W \) were originally introduced by \textit{C. Jordan} [Bull. S. M. F. III, 103--174 (1875; JFM 07.0457.01)]. Modern variants are due to, e.g., \textit{C. Vitner} [Cas. Pěst. Mat. 87, 415--422 (1962; Zbl 0124.13301)] and \textit{W. Degen} [Rev. Roum. Math. Pures Appl. 21, 503--509 (1976; Zbl 0336.50004)]. They amount, again modified, to the following. Take the projection \( P: U \to V \) along the orthogonal complement of \( V \), form \( S := P^{\displaystyle\ast}P \) (\( P^{\ast}: V \to U \) the adjoint of \( P \)) which is a symmetric endomorphism of \( U \) with spectrum in \( [0,1] \), and define the characteristic angles \( \theta_{i} \) between \( U \) and \( V \) via the eigenvalues \( \lambda_{i} \) of \( S \) by \( \cos\,\theta_{i} = \sqrt{\lambda_{i}} \). The present interesting paper deals with surfaces \( M^{2} \) immersed in the Riemannian product of two space forms \( M_{1} = M_{1}(c_{1}) \) and \( M_{2} = M_{2}(c_{2}) \) of dimension \( 2 \) whose tangent planes enclose constant angles with the tangent planes of the components (always assuming that the components are not both flat). The definition of angles adapted to this four-dimensional situation is somewhat different, but one can check its equivalence. In particular, following \textit{K. Yano} and \textit{M. Kon} [Structures on manifolds. Singapore: World Scientific. Distr. by John Wiley \& Sons Ltd., Chichester (1984; Zbl 0557.53001)], a \( (1,1) \)-tensor field \( F \) on \( M_{1} \times M_{2} \) is used such that \( (I+F)/2 \), resp., \( (I-F)/2 \), projects vectors to the first, resp., second, component. With these means, the authors locally classify all constant angle surfaces, first observing that each copy has constant Gaussian curvature. Besides trivial cases, their result presents special surfaces generated by curves in both factors and a general type of surface which is based on a coupled nonlinear system of PDEs. After due reformulation, this system of PDEs turns out to be a Bäcklund transform of the sine-(resp., sinh-)Gordon equation. Additionally, there are intermediate types where the original system is a deviation of the general case. Finally, attention is paid to the totally geodesic surfaces in \( M_{1} \times M_{2} \) which are shown to be constant angle surfaces.