@s.harp Interesting question. What if $f_n$ is the identity except on union of balls of radius $1/2$ a distance $1,2,\dots,n$ away from the origin? (Do a $90^\circ$ rotation on half the ball, and then "unrotate" to join it to the identity by the boundary of the big ball.) So, pointwise, $f_n$ converges to the identity, but clearly that won't be uniform on compacta.
The derivative of $f_n$ is a rotation at every point, so it's an infinitesimal isometry. This might need a little repairing, but ...