Why are they building eLISA and what implications would it have? I understand that the next step after LIGO is to plan and build eLISA, I understand that out in space there are a lot less interferences compared to Earth which makes it a good way to detect things we couldn't initially. 
Based on the ideas of gravitational waves, would it be possible to detect dark matter and energy? What else could eLISA tell us about the universe?
 A: Space is, as you say, good for removing a lot of the background noise that spoils LIGO's data — like seismic noise, disturbances from traffic and logging activity, people shooting at the beam tubes, etc.  But another important reason to go into space is so that you can basically make a much larger version of LIGO.  LIGO's arms are 4km long; eLISA's arms will be 1,000,000km long.  The main effect this has is to lower the frequencies that you're looking at.  So instead of LIGO's ~100Hz, eLISA will be looking at ~0.01Hz.
Now this change in frequency also changes the sources that you'll be looking at.  LIGO's GW150914 detection was a pair of black holes, each with mass around 30 times the mass of our sun, and its peak frequency was just over 100Hz.  But if you have a heavier pair of black holes, the peak frequency will be proportionally lower.  For example, if you have a pair of black holes that are each about 300,000 times the mass of our sun, the peak frequency will be just over 0.01Hz.  So eLISA will be looking for this type of event.  It will also be looking for mergers where one of the black holes has such a large mass, but the other is small (these are called "extreme mass-ratio inspirals").
This paper gives a pretty readable overview of all the different expected sources.  These events will typically be invisible to LIGO, because they will happen at frequencies too low for LIGO to see.  More importantly, they will also typically be invisible to all our other ways of looking at the universe.  So if we ever want to understand these phenomena, we need something like eLISA.
Unfortunately, we don't really expect to see any cosmological effects in eLISA, or evidence of dark matter, dark energy, cosmic strings, etc.  But you never know if you don't look.  So if there is new physics that we don't yet understand, eLISA might be able to tell us.
A: Mike's answer is good but there Is more. You can actually see it in the papaer he referred you to, it is an excellent paper. It not only describes the possible sources, but it points to the new physics that it might see.  eLisa, also called NGO, will be sensitive to gravitational waves from the early universe (and after), down to 10 to the minus 18 sec after the Big Bang. It will see gravitational waves from merging massive and supermassive black holes going back quite a bit also in time, and so much more sensitively that it will see order of magnitude More details. It will see extra low frequencies which means extra large wavelengths, and so gravitational waves from much bigger cosmological and astrophysical objects. 
For instance, it will be sensitive to gravitational waves emitted by cosmological cosmic strings, if they exist. That would be new physics. It will also be sensitive to early universe phase transitions including things like fluctuations for leftover inflation with small bubbles forming,and other major changes in the cosmological history - see the paper. If it finds  any of those can also find evidence of higher dimensions in spacetime, and string theory, a huge physics impact if anything is found. 
As for black holes, it will probe if tHey behave as general relativity predicts. It'll test the no hair theorem, seeing if mass and spin is all they have, by measuring various multipole moments. Thus, it'll determine if general relativity in the ultra strong regime needs modification or not, or if there are other exotic kinds of matter around that could give them hair (for instant, some scalar or other fields). It'll determine if they can instead be boson stars (Higgs like particles collapsed like neutron stars). It'll be sensitive to significant new physics. 
It'll look for other kinds of symmetry breaking like what split the electroweak force into electromagnetic and weak. In the LHC the energies to be reached will be on the order of about 15 TEVs. The grav radiation seen from early times in the universe will include events in the 1000 TEV range.
The sensitivity of eLisa is hugely more than that of Ligo. Mainly because of the much longer interferometer arm lengths, 1 million kilometers vs Ligo's 5 Kms or so. It will see black holes spiraling in to merge for much longer than Ligo. Ligo saw the last 1/4 second of the black hole mergers, at relatively low SNR. eLisa will see supermassive black holes all over the universe spiraling in and then merging up to months before they merge, thus averaging a lot of data. It will see the actual mergers and ring downs much more sensitively than Ligo, so it will be able to see any minor deviation from the predictions of general relativity, in the Strong gravity domain. That will include a lot more detail in the merger phase, and possibly higher order effects describing the dynamics of the merger, and then the ring down.  It will determine a lower limit to the mass of the graviton by seeing any dispersion in the speed of the gravitational waves as function of frequency, down to orders or magnitude better than what we know now. It can eliminate various enahncementS or alternative gravity theories that way. It will see gravitational waves from binary neutron stars
It will detect so many supermassive black holes that it will be able to form a history of galactic formation, seeded by those black holes in their centers. It might, but I didn't see any specifics, thus determine some more details one the characteristics of dark matter that is also known to help in Galaxy formation. 
I'd didn't see anything in the paper on dark matter or energy. But I did not read it all carefully. 
I knew some of this, but the paper opened my eyes to the extraordinary new realm of physics these gravitational observatories open up.
Not bad for just 3 satellites.
