Welcome to the world of corrosion! As you have found out, mixtures don't always have the same corrosivity as their pure components and this can be complicated by the presence of impurities. Having said this, I am not sure about thionyl chloride "being almost non-corrosive" to Alloy C-22. Schweitzer's Corrosion Resistance Tables rates Alloy C-276 (which is similar to Alloy C-22) as "good" with 10% thionyl chloride up to about 75 °C. But a "good" rating is < 20 mpy which I would not consider almost non-corrosive. You could try moving to a generally more resistant alloy such as tantalum (which is very expensive) or perhaps Alloy 400, which has good resistance to thionyl chloride at higher maximum temperatures. If a reactor is involved, a glass-lined steel might be a good choice. When dealing with mixtures such as these, it is always a good practice to run laboratory corrosion tests to confirm potential material of construction choices.