John,

From a metallurgical standpoint this is not a good idea.

However, the lower temperature your product operates at, the longer the
gold/tin/lead intermetallics will survive, but it is still a matter of time
before you see changes in interfacial resistance, strength, etc.  The wedge
(2nd) bond will typically fail before the ball (1st), due to the differences
in cross-sectional gold thickness (volume) present in the bond.

Glob-top, and cure of the glob top, will also add additional stresses to the
system.

Ultrasonic aluminum wedge bonding to the bare copper, or to nickel plating
is a far more reliable metallurgy - albeit slower bonding.  If given a
choice, I would flip chip it.

Steve