Wake lows are known to cause extensive wind damage. The "eye" was simply a very very strong wake low and gave the appearance of a hurricane's eye on radar. Just as they call it in the article it is an "eye-like" structure. A great look alike but not actually an eye and eye wall.
Here is a response from Clark Evans, who has done extensive work on this case and was quoted in the UCAR StaffNotes:
"The eye-like feature and intense vortex associated with the 8 May 2009 derecho was
not associated with the wake low. Rather, this vortex is a convectively-generated, northern book-end vortex along the squall line associated with the derecho. There is a separate wake low feature removed approximately 100-150 km to the west of the eye-like feature both in surface observations and model output. Recall, wake lows are typically not associated with active convection and are instead found in the stratiform rain regions of MCSes. Most have only weak surface reflections. This event is no different in that regard with respect to its wake low.
What makes this event unique, however, is the intense vortex on the northern end of the squall line. Diagnostics of this feature do suggest a warm-core vortex structure. You are right in that it has an 'eye-like' structure, as do a number of other vortices of both tropical and non-tropical origin. How it formed, though, is still an open research question -- as is how the intense convectively-generated vortex developed. It probably wasn't an "inland hurricane," though it does appear to share some characteristics with other surface-based convectively-generated vortices as well as tropical cyclones that reintensified while inland, including Erin in 2007 over Oklahoma.
We were limited to about 300-350 words in the UCAR StaffNotes write-up that started this thread, so we couldn't tell the full story. (Plus, those articles are intended for a general audience, meaning we likely could not have been that specific even if we had wanted to be.) We'll be presenting some preliminary results at the Severe Local Storms conference in Denver, CO in mid-October and hope to publish final results sometime in early 2011. In the interim, our abstracts for those talks -- along with the abstracts for five other talks on the 8 May 2009 derecho to be given at SLS -- can be found at
http://ams.confex.com/ams/25SLS/techprogram/session_25095.htm.
Best,
Clark Evans"
If you have the opportunity to go to the SLS, I highly recommend that section on The 8 May 2009 Derecho. I know most everyone who is involved in that session and have seen some of the results already through seminars and collaboration through operationally-based work done at my office.