On the possibility that electronically excited products may be formed in the reaction H+2+H2→H+3+H