I have always wondered why there is an opening in the solder mask whenever a microstrip line is laid out on a PCB.
For reference consider the following board:
There is clearly a large opening in the board around the microstrip lines and stubs.
Does it have something to do with avoiding to change the characteristic impedance of the line, or does the solder mask somehow effect the propagation in the microstrip line?
I have searched but could not find a clear enough answer.
It appears that there is no soldermask on that board at all, and the opening you are referring is in the copper. When it comes to RF or other high-frequency signals, surrounding objects play a large role in the impedance of the line. If you have copper surrounding the microstrip trace(s), the impedance will be affected. Besides this, signals travelling on nearby copper can be coupled in to the high speed line, and this crosstalk could be disastrous for your high-speed signal integrity.
The reason the board doesn't have any soldermask at all is likely because of the reasons already mentioned - soldermask may change its properties over time, in different environments, etc. Eliminating it altogether ensures a more consistent impedance by doing away with that extra variable.
Because at RF, solder mask behaves differently than air; especially, it has higher losses.
Also, yes, even if it was lossless, the materials below and above the microstrip affect the wave impedance. However, one would be able to counter that by changing geometry of the line.
Does it have something to do with avoiding to change the characteristic impedance of the line, or does the solder mask somehow affect the propagation in the Microstrip line.
Yes - both your statements mean the same thing, in the end.
It may also have a lot to do with Rev 1 vs Rev 2. Your example BFU730 (2010) vs Rev 2 BFU730F (2012), which I'd guess as 6 actual internal revision.
Until it has been prototyped and tested, engineers will err on the side of caution to combat issues the other answers have described. As they identify, simulate and correct issues the design will change.
From AN11007 Single stage 5-6 GHz WLAN LNA with BFU730F - Rev 2:
In this revision, you can clearly see ground planes cover the majority of PCB. For a 5-6GHz signal, I'd expect solid ground planes shielding the circuit and wide traces interconnecting components.
They have refined their understanding of the best way to lay out this circuit. The large openings around the microstrip lines and stubs has disappeared in this revision.
