I am guessing with your screen name, you might have a clue about this pano stuff? Is 14mm wide enough on a d700? I am not even sure what software works best for post?
Panopeeper is far more experienced than I, I believe, however here's my own experience.
Firstly, as others have said, lens distortion is irrelevant when creating VRs. Most people use fisheye lenses. The images are remapped by the software into the equirectangular form (2x1, 360 x 180 deg equivalent) which is the basis of the viewed image.
I had been using a D200 with the Nikkor 10.5 fisheye for about a year and a half before I bought a D700 and, amongst others, the 14-24 2.8 - which is in itself a wonderful lens. It can be used to do VRs - but so could a 200mm lens if you wanted to spend a month creating one from the hundreds of shots that this focal length would require. As some people have done, with amazing high-resolution results.
However I would say that the 14-24 lens is completely unsuitable for the purpose. I've used it; it requires that you shoot 2 rows of six shots (assuming you're not doing HDR, which would require exposure bracketing too) plus however many shots required by your chosen method for patching the nadir - which many people don't find necessary. This approach requires a LOT of PP time in stitching - at least it does for me using PTGui Pro, which I find to be excellent, if a little difficult to learn initially. Others swear by AutoPano Pro. Both will do a good job once you know how to use them I believe. Also the lens is huge and heavy, requiring a very substantial pano head. It's also very easy to nudge the zoom ring and accidentally shoot at >14mm, possibly scr3wing up the whole job.
Give that the final use for most VRs is a tiny .mov or flash file on a website, it's fairly pointless to take this approach. I have had the lens hood shaved from my 10.5 FE which enables me to use it on the FF body, which gives a circular image. In theory this enables you to shoot 3-round, plus zenith, plus nadir(s), however I usually shoot 4-round as the greater overlaps make for more accurate, faster stitching. The resulting tif created by this approach is about 70Mb (I can't recall the actual pixel values as I'm on my laptop.)
In fact I now usually revert to shooting with my D200/10.5 FE. For the time taken to shoot an additional 2 shots the difference is negligible. In this format the lens fills the entire frame (at an effective equivalent f/l of about 15mm). So the resulting stitched equi.tif inclusive of the patched nadir is about 200Mb. Once reduced to a 1Mb .swf file these two approaches are effectively indistinguishable. I usually shoot HDR so the s/n ratio and DR difference between the D200 and D700 has an insignificant impact on the final output on a website.
I now tend to use the D700/14-24 combination to shoot "stills" asociated with the VR job, so less lens changing involved. FWIW I believe that most people tend to like the Sigma 8mm fisheye for VR shooting. I started using the Nikkor 10.5 assuming that the full-frame fisheye would be more generally useful even when "defished": it isn't, particularly now that I have the D700/14-24 combination.
Anyway, I hope that this helps. Others may have better and/or more useful advice.