assuming you can reach a working temperature differential why not continue deeper with the well? From that threshold depth onwards you can extract heat along a line segment that extends to the ultimate depth (which is pressumably dictated by engineering limitations)

caveat: thinking like a physicist, not an engineer :-)