Now that I'm in my office, I can look at the Buderi book. Indeed it has some interesting stuff: from page 117 (near the bottom),
"From that point, it did not take long to identify the Brisith receiver's silicon crystal detectors as the main factor behind its superior showing. The Radlab radar employed a grounded grid triode. The finding stunned the Americans, since they ahd started with crystals but abandoned them earlier in the year after tests showed vacuum tubes to be superior. Only later, Ramsey related, did it become apparent that the crystal used for the American tests had partially burned out. The British, by contrast, had gone with crystals from the start. By spring 1941, Oliphant's lab had designed a better capsule that lessened susceptibility to shock and vibration, and rendered crystals superior to anything in the United States; these had gone into mass production at British Thomson-Houston. Around the same time the duplexing, or TR, problem was solved by the ingenious application of a reflex klystron, itself an adaptation of the Varian brothers' original creation. The so-called soft Sutton tube was filled with a low-pressure gas that was rapidly ionized by the transmitter pulse, providing a short circuit that protected the crystal from burnout. Once the pulse ended, the gas recovered, allowing received signals passage to the detector."
There's a nice section at pages 314-320 covering the discovery of the silicon P-N junction.
Cheers, Tom