when chaining models, ignored impurities must be added to output

[lars-petter-hauge]

for instance for solubility, there's only a few impurities that are incorporated. We must ensure that others are included in the output so that mass balance is not off

[kvashchuka]

related to #759

[lars-petter-hauge]

Had a chat with Sviat on this, we agree it's a good idea to do. fixed here #778