The 5-primary homotopy exponent of the exceptional Lie group \(E_8\) (Q816493)

The paper provides a very sharp upper bound for the exponent of the 5-primary component of the exceptional Lie group \(E_8\). Recall that the exponent is the highest power \(5^r\) such that there is an element in the homotopy \(\pi_*(E_8)\) which has such order. The Lie group \(E_8\) localized at the prime \(5\) is known to split as a product of two spaces \(X\times Y\) where the hard part consists in estimating the exponent of the factor \(X\) which has cohomology \(H^*(X;\mathbb Z_p)=\mathbb Z/5\mathbb Z[x_{12}]/(x_{12}^5)\bigotimes \Lambda(x_3, x_{11},x_{27},x_{35})\). The cohomology of this space projects to the algebra \(\mathbb Z/5\mathbb Z[x_{12}]/(x_{12}^5)\bigotimes \Lambda(x_3,x_{11})\). This is the cohomology of Harper's rank two mod-\(p\) finite \(H\)-spaces which have torsion an odd prime, for \(p=5\). The author constructs a fibration \(\Omega K_5 \to B(27,35) \to X\) in order to estimate the exponent of \(X\), where \(B(27,35)\) has cohomology \(\Lambda(x_{27},x_{35})\). This is a crucial step and is not easy to construct. Then he uses his own results about the exponent of Harper's spaces \(K_p\) and obtains: Theorem: \(\exp(E_8)\leq 5^{31}\). The result is at most one power of 5 from being optimal since it is known that there are elements of order \(5^{30}\) in the homotopy groups of \(E_8\).