Note that the zero in logarithm was specified as "positive zero". It had to
be investigated, since ln(0) does not exist. How was the "positive" part
deduced? For any x,
arctan(x) < π/2,
so the expression in
logarithm is always positive and goes to zero, hence positive zero.
No trick for the resulting indeterminate ratio can be seen, so it is up to the
standard procedure: Use the l'Hospital rule.