flipper, some possibilities:
an excess of carbonyldiimidazole and DEA was used in the first reaction, and the resulting residue would hence contain LSD along with any remainders of the other two reagents. this residue was then dissolved in 250mL of 2% tartaric acid, enough water to be sure to react with residual carbonyldiimidazole, which decomposes to CO2 + imidazole. Under certain physical conditions, pockets of carbonic acid may drop in pH enough to influence yields. i also found a couple of papers that implied that imidazole group may bind with heavy metals, a la EDTA. the fact that imidazole accounted for the bulk of the 19% impurity in the raw product indicates that virtually all of the catalyst was decomposed.
another possibility is that the activated charcoal in the first reaction contained trace amounts of heavy metals. a bit of both seems the most likely explanation for the large discrepancy, as the imidazole was still in solution when it was passed through the charcoal.
i notice that DMF is boiled off at 50-60degC under pressure in the first reaction. this seems an unusually high temperature to expose lsd to, so it may possibly explain the bizarre yield.
just a few educated guesses, please corrent me if i'm wrong