If I have understood correctly the return current for AC voltages will follow the red signal track but on the ground plane. For DC voltages the return current will spread out across the ground plane. Could adding the track shown in white on the image on the ground plane contain the DC return currents?
Also would the main EMC problem be the track acting as an antenna? What else could it be?
Thank you for any help you can give. Hopefully this makes sense
You could do that, but it's not generally done because the higher the frequency, the less of an issue EMC is. Frequencies that cause EMC issues are generally high enough that the return path follows the red line. Unless you have some extreme EMC requirements at frequencies below about 10MHz, or sensitive magnetic instruments nearby, it's not usually necessary.
There is no sharp cutoff for DC and switching signals as to how closely the return currents follow the signal; at DC the return path will indeed spread out and for switching signals, as the edge rate gets faster, the return path will follow the signal track more closely.
The return path will take the path requiring the least energy in my experience so even at DC (the plane has some resistance) the return currents will still be somewhat higher near the signal itself. This is due to the effective impedance of the plane (at high frequencies the plane is not a single lumped item but a distributed element).
In your case, isolating the return path in this manner will not necessarily prevent the return current from spreading out because the (isolated) ground you show must at some point return to the ground of the power source which will also be connected to the rest of the ground you show.
I went into controlling return currents in some detail in this answer.
The (simplified) power path is Power source positive -> IC -> power source negative at DC with perhaps some of the current flowing from an output to an input and then back to power source negative which depends entirely on the specifics of the circuit (and will have other return paths for a switching signal).
Any unbalanced track (which includes all single ended signals) has the potential to act as an antenna; just how much is radiated is very implementation dependent.
Your question is slightly unclear, but I will answer what I can.
If you want to prevent your signal trace from acting as an antenna, you could use the trick of parallel lines of exactly equal length and thickness, but you would have to use differential signalling instead of the ground-referenced one. You can see how that is done on computer mainboards for example.
The above image is from: http://www.viacircuits.com/differential-pair---matched-length.html
I was taught, by a guy from Bell Labs, that coax cables are not really effective until about 100,000 Hz. And even then, some of the energy will still venture outside the coax ---- because the energy uses ALL paths back home, and a bit of energy 80db down yet outside the coax ---is still allowed by mother nature.
