High oxygen wafers from 100 mm diam Si crystals grown by the Czochralski process, but subjected to three different thermal histories in an experimental puller, were examined by Wright etching, transmission electron microscopy, and Fourier transform infrared spectroscopy after wafer heat-treatments at 775 ~ 1050 ~ 775 ~ + 1050 ~ and 1320 ~ + 775 ~ + 1050~ Only wafers near the seed end of each ingot were used, thus minimizing differences in parameters other than thermal history. The observations show that defect morphology and O precipitate number density (not total O precipitation) after the one-and two-step heat-treatments depend on thermal history in the puller. In particular, 775~ heated wafers which spent less than an hour in the puller below 1000~ show number densities down by more than a factor of 100 from those which spent longer. On the other hand, the observations indicate that effects of puller thermal history are erased with a short 1320~ anneal, or typical VLSI multistep pretreatments which enhance bulk oxygen precipitation. In addition, the results suggest two possible complications for simple models of oxygen precipitation. These are that 500~ annealing can be much more effective than 775~ annealing for nucleating oxygen precipitates, contrary to model predictions, and that critical sizes for precipitate dissolution at 1050~ can be much larger than predicted by classical calculations.) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 128.255.6.125 Downloaded on 2015-06-05 to IP ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 128.255.6.125 Downloaded on 2015-06-05 to IP