For , real nonzero , and a finite positive logarithmic time scale , the exponential-potential scalar-dominated power-law solution requires the displayed strict sign condition. Thus positive permits , while negative permits . The boundary cannot occur with nonzero in this ansatz. Setting gives a constant potential and a separate branch, including constant-field de Sitter spacetime for .
Write for the reduced Planck mass. For a nontrivial power law with , , , and . Matching its time dependence to the kinetic and Friedmann equation terms requires . The scalar field equation of motion and the Friedmann equation then reduce to
Solving gives the exponential-potential scalar-dominated power-law solution
Choose and the positive square root for . The existence conditions for exponential-potential scaling matter: if , a finite real requires , rather than an inclusive bound on signed . At , nonzero is incompatible with this logarithmic ansatz. If , the corresponding scaling solution instead requires . The sign of changes the field's rolling direction, not . For the potential is constant and this finite- ansatz does not apply; with positive , a constant field gives de Sitter spacetime expansion.