… splitting our ‘Collectors’ field, which currently contains all the collectors separated by a semicolon, into five or six fields (e.g. ‘Collector 1’, ‘Collector 2’, etc.) might be the only option. Is that right?
Yes, that is the only option. Each field must be represented as a distinct column in the data set, so you cannot use a separator like a semicolon to indicate this separation in the WorkBench.
I would also like to mention that our ultimate goal is to export to Darwin Core Archive, ideally with all the collector information aggregated into one field separated by a vertical bar (|) rather than a semicolon.
You can configure the separator for exports by modifying the table aggregation format. This is entirely possible and is not restricted by the format used during data import!