The only thing I think you could do is to change the sacrificial layer
I would suggest PMGI with the following process explained in this paper
Polydimethylglutarimide (PMGI) as a sacrificial material for SU-8
surface-micromachining:

    http://iopscience.iop.org/0960-1317/18/7/075011

On Wed, Jan 4, 2012 at 10:40 PM, Shiv GD  wrote:

> Hello Everyone,
>
> I am working on a Microchannel device, and facing some problem with AZ 4620
> and SU-8 Layer. The device is fabricated on a glass substrate and I am
> using AZ 4620 as sacrificial layer ( I need ~20um height). I patterned the
> AZ layer using standard lithography process ( Spincoat, Bake on Hotplate
> for 2mins @ 110'C, Expose and Develop using AZ 400K. ) and I want to keep
> it until the last step. I than, spincoat SU-8 2025 (on top of the AZ 4620
> layer), bake, expose, post bake. But when I try to develop the SU-8
> structure ( using Microchem SU-8 developer ), I observe that the SU-8
> structure develops alright but the AZ sacrificial layer is gone. As a
> result the SU-8 structure collapse, cracks and I can't continue to the next
> step.
>
> What do you think I can do to resolve the situation ? Please suggest.
>
> Your help is much appreciated.
>
> Thank you.
>
> Shib