We do not use particularly citrate buffer, you can use EDTA buffer etc., some even use urea. The aim is to break down the molecular cross links formed by formalin fixation by heating or enzymes.. their is a very good source for ihc where I learnt details; ihcworld.com and I love the technical book there: http://ihcworld.com/_books/Miller_handout.pdf
You need an antigen retrieval step in your IHC because formalin fixation of the tissue cross-links with antigens. You can apply heat (i.e. hot citrate buffer or EDTA) or use an enzymatic digestion (i.e. proteinase K or trypsin). I agree with Kivilcim in the ihcworld.com suggestion.
There are many antigen retrieval techniques you can use and which one depends on the antigen you are looking at. I use Citrate pH6, EDTA pH 8 and Tris-EDTA pH9 in a microwave. The higher the pH the more vigorouse the retrieval. I treat the choice of buffer as part of the optimisation as the higher pH can lead to background staining.
The retrival step breaks the formalin cross links that are formed during fixation so that the antigen is unmasked. Make sure your sections are baked and or on extra adhesive slides otherwise they can come off during this process.
I had the chance to improve my experience in a segnifficant european laboratory. Every time a new immunohistochemistry was set up, the technician in cooperation with the MD evaluate the propper technique for antigen retrieval and antibody dilution. This means there were tested protease digestion using at least two type of enzimes, heat retrieval in pressure cooker, steam oven, microwaves using all of three types of unmasking solutions: at pH 6, 8 and 9. The key point is the pH of the unmasking solution and surprisingly, one company ( the one I like their unmasking products very much) provides cytrate based sollutions at pH 6 and 8.
In conclusion, we are using cytrate unmasking sollutions because they are working.