I'm a master student and I'm the phase of preparing my thesis. I've searched many papers related to my thesis but I'm a bit confused in what to focus. Do I need to take only a paper (or few papers) and to focus basically on that?
I think the best answer to this question has already been given as a comment:
This is a good answer to so many questions on this site (I have started to think that there should be an "Ask your advisor!" closure option), but it seems especially true here. From the OP's comment I can see that the master's thesis is being done in mathematics. The expectations and requirements for a master's thesis in mathematics are so highly variable across institutions and countries that I can think of almost nothing (other than "Ask your advisor!") which would be guaranteed to be universally applicable.
In mathematics programs in the US, it is especially unclear what work constitutes a master's thesis, especially in pure mathematics. Unlike the situation in many other countries, there are relatively few full-time master's students in American universities, even compared to the number of master's degrees awarded: in my experience, more master's degrees go to talented, ambitious undergraduates who get them alongside their bachelor's degrees, or to PhD students who have decided to drop out of the program and get a consolation degree. In my case I got a master's degree along with my bachelor's degree at the University of Chicago, and for such a top university you might be surprised to hear how minimal the requirements were: I had to complete all nine trimester courses that first year graduate (i.e., PhD) students take (I did so over two years) and pass a pro forma foreign language exam. [In particular I did not write a master's thesis.] When I went on to my PhD studies (at Harvard) I found that I was about as well prepared as most of the other students. I don't recall that my having a master's degree came up once during my five years in a PhD program...with the possible exception that some students would, a year or two into their program, fill out paperwork and pay a small fee to get a master's degree, whereas I already had one of those so chose to keep my money.
Having been heavily involved with the graduate (mostly PhD) program at the University of Georgia in recent years, I can say that here a master's thesis is whatever the advisor and student agree that it is, subject to the approval of two other committee members. Writing a thesis is one route; there is another route involving more coursework and some exams. Among master's thesis advisors I've talked to, the sense is that the student should take the thesis as an opportunity to engage with some piece of mathematics at a deeper level than they have done before, to the extent that they have mastered it and can show this mastery with an original (or at least, independent) exposition. This description seems rather at odds with the one given in another answer to this question:
This is probably a correct description of some academic fields, but not for mathematics. Very few mathematics undergraduates are reading "real" math papers at all. The task of reading, understanding and writing about even one "real" math paper may in fact be sufficient for a master's thesis. There are even certain papers out there for which rewriting them so as to contain the level of detail and completeness that would satisfy a master's thesis committee would be a real service to the mathematical community. I think that most PhD students in mathematics do not use every seminal work in [their] field: I didn't, for instance.
I am not saying that just any old thing will suffice for a master's thesis in mathematics: I am saying that the global requirements are very few, so it becomes more important to talk to your advisor.
I have so far supervised one master's thesis. My student carefully read and wrote about two papers concerning geometry of numbers and Legendre's Equation ax^2+by^2+cz^2 = 0. She then tried to extend the techniques of the second paper to diagonal quadratic equations in n \geq 4 variables. Much of this was easy, but the key was the existence of a "magic sublattice" defined in the three-dimensional case by the necessary congruence conditions for Legendre's Equation to have a solution. After much trouble (and some help from me), she was eventually able to prove that for more than three variables such a magic sublattice did not exist. She wrote up a thesis which contained expositions of the two papers she read plus a proof of this result (plus a few more small things, totalling a few pages). I remember that she was concerned that her thesis was rather short: 45 pages double-spaced. I thought her thesis was an unusually strong one and told her so. (If people are wondering, I think her result is not yet publishable -- it's along the lines of showing that a certain proof strategy cannot work, and such things are hard to publish in mathematics -- but that the work could be continued and made publishable. I honestly think that makes it an above average American master's thesis in mathematics.)
NO. Undergrads may get away with using one (or a few) papers to synthesize the essence into their own viewpoint, but at the post-grad level you should be reading for width and depth, which means using every seminal work in your field as well as exploring the newer papers on the topic. This is particularly important since you state that most of the published research in this area dates from 2007.
And even though you feel that you are not prepared to tackle original research in this area, consider at least having a section outlining possible future research.