Do you have a reference for the paper? The term 'satellite' is used to refer to any peak(s) additional to the main photoemission signal(s), and can arise from one of several phenomena. I seem to recall that the Fe 2p spectrum of ferrocene is a clean doublet (i.e., no satellites), but don't have any data at hand to confirm this.

check this reference http://pubs.rsc.org/-/content/articlehtml/2016/ta/c6ta00313c

Efficient approach to iron/nitrogen co-doped graphene materials as efficient electrochemical catalysts for the oxygen reduction reaction

Ferrocene is just a precursor that the authors use to prepare iron oxyhydroxide films. It is the photoemission spectra of these that are reported in the paper, not of ferrocene. The iron in the oxyhydroxide has a formal oxidation state of III. The satellites are 'shake-up' peaks, as described here: http://www.xpsfitting.com/search/label/Shake%20Up%20Structure

At what binding energies do the two ferrocene peaks and the two additional peaks appear?

can you plz share the actual Figure, in which you should mention the concern point, so that you get clear explanation.

Ferrocene has iron in +2 oxidation state. In Fe 2p XPS spectra, Fe2+ is expected to show a doublet due to Fe2p1/2 and Fe 2p3/2, and the satellite peak for Fe2+ may occur at around 5-6 eV difference from Fe 2p3/2 peak. In the given question, if the Fe 2p3/2 peak occurs at 708 eV then, the peak at 721 eV should be for Fe 2p1/2.