Note that the claimed 20% energy overhead (reduced to 7% if ammonium-carbonate capture is used, IIRC) is for post-combustion carbon capture.  Pre-combustion capture can be far more efficient.

Also, CCS and CHP are hardly exclusive technologies; they can be used together.  Indeed, a historical business model is almost ideally suited to it:  the local gas works.  Instead of converting coal to producer gas, coal would be gasified (roughly 76% efficient), scrubbed, and the syngas steam-reformed to convert all the carbon monoxide to hydrogen:
CO + H2O -> CO2 + H2
The waste heat from the gasifier can power the required oxygen separation and provide the steam to the reformer.

This produces a hydrogen fuel stream and two side streams:  CO2 for sequestration, and H2S which can be dealt with in several ways (one possibility is to run it through molten copper to react Cu + H2S -> CuS + H2, then burn off the CuS in air:  CuS + O2 -> Cu + SO2.  React the SO2 with limestone to get gypsum and more CO2).  The hydrogen goes off to run remote CHP equipment, with minimal carbon emissions (some methane is produced in the gasifier and will go with the hydrogen unless it is removed and recycled).

Indeed, most CTL and IGCC plants produce both heat, and fuels/power.
The differences from the local gas works would be the scale, the reaction chemistry and the availability of low-ash coals.

This chart from the IPCC report posted by Rembrandt shows the distribution of energy costs.

Chart 1 - additional energy costs of carbon capture for different electricity plants, source: IPCC Special Report on Carbon Dioxide Capture and Storage.

Do you have links to support your claim of 7% - that seems lower than the compression costs alone - though I don't understand from the IPCC chart why the compression energy costs vary.

The ammonium-carbonate process doesn't require compression of any gases; the CO2 is handled as ammonium carbonate in aqueous solution, and can be pumped as a liquid.  It is liberated from the liquid at moderate temperatures, which can be done with low-pressure steam from the powerplant.

I did some digging and couldn't find the 7% figure again (it may have been superceded), but I did find this article which included this table:

The overhead for this process is under 9% per the table.

So what is the source of the ammonium? How much is needed and how is the ammonium carbonate disposed of?

Ammonium ion is formed by the dissociation of water in aqueous ammonia solutions:

NH3 + H2O -> NH4+ + OH-

Ammonium carbonate and bicarbonate are formed from ammonium hydroxide and carbonic acid:

NH4+ + CO32- + H+ -> (NH4)HCO3

2 NH4+ + CO32- -> (NH4)2CO3

Conversion of gaseous carbon dioxide to dissolved or precipitated carbonates reduces the volume and the required pumping work immensely.  Ammonium carbonate and bicarbonate can be dissociated into carbon dioxide and ammonia with heat at relatively low temperature; the references I found claimed that the heat input is around half that required for amine stripping, and the mass-loading of CO2 per volume of solution is much higher than for amines also, which requires less liquid to be pumped.