On the disc , the upward normal selects only the third component, which equals . Hence the surface integral is the area of the unit disk:
The two orthogonal directions and are preserved by this symmetric matrix. Direct multiplication, using the hyperbolic function identities , gives the eigenvalue–eigenvector pairs
Each eigenvector can be multiplied by any nonzero scalar. For these give the two one-dimensional eigenspaces. When , the matrix is the identity: the only eigenvalue is , with multiplicity two, and every nonzero vector is an eigenvector.
The tangent half-angle substitution gives and . Dividing and then taking the reciprocal square of the cosine gives
These formulas apply where is defined and ; at the original tangent and secant function are undefined because .
Use integration by substitution with , so . Then
Differentiating the answer by the chain rule gives , confirming the antiderivative. The real constant of integration is arbitrary.
Let . Its values at useful points are
Since a polynomial is a continuous function, the intermediate value theorem gives one polynomial root in each of the disjoint intervals , and . A nonzero cubic polynomial has at most three distinct polynomial roots. Hence it has
These three distinct polynomial roots already exhaust its degree of a polynomial, so each has multiplicity of a root one.
Both segments start at , so their direction vectors are and . Their dot product is , and their norms are and . Thus
The angle between the rays lies in , so its sine is nonnegative:
The negative cosine correctly identifies an obtuse angle. One also obtains the sine from , using the cross product.
For , the product rule and chain rule give
The positive exponential factor makes increasing for and decreasing for . Its only interior stationary point is therefore a maximum at . Checking it together with both endpoints,
Consequently
Checking the endpoints is necessary for extrema on a closed interval.
Both sums are geometric series. For first term and common ratio with , the sum is . Taking the initial term to be gives
and
Both have absolute convergence because the series of absolute values has ratio .
Using the Pythagorean trigonometric identity and the double-angle formula,
The required value therefore occurs when . Within , the only possibilities are and . Hence
Equivalently equality requires .

There are unlisted articles, also show them or only show them.