As for perpendicular axis of maximal moment it is well studied. In the actual case of buckling of I-beam I didn't find the plot of tangent stresses in a section when the tangent load Q is perpendicular to a member. Though the section may be regarded as a simple combinaton of 3 rectangles the load balance between flanges and the member is unknown. We may assume that all the stresses are parallel to the Q then a strong gradient of the stresses should be both in thin member and in its joins with flanges.