Should I cite a paper for its literature review?
There is more controversy on this topic than I expected, so I've done some further digging. The results of this surprised me!
Institution, Program, and Journal requirements
The rules of your neighborhood may differ, so you'll want to follow whatever more specific rules are given to you in your program. For example, Columbia College demands secondary source citation:
Indebtedness
You must cite any text you read that helped you think about your paper even if you do not reference it directly in the text of your paper.
However, the OWL APA style reference gives differing advice:
Work Discussed in a Secondary Source
NOTE: Give the secondary source in the references list; in the text, name the original work, and give a citation for the secondary source. For example, if Seidenberg and McClelland's work is cited in Coltheart et al. and you did not read the original work [emphasis mine], list the Coltheart et al. reference in the References. In the text, use the following citation...
So if you merely were directed to the original work, but then attained the original work and read it, there is not necessarily any need to cite the secondary source - but the guide doesn't prohibit it.
However, the IEEE has a stance as strong as Columbia, but in the opposite direction! As noted by York University in their IEEE style guide (page 5 of the pdf):
Should I use secondary references?
A secondary reference is given when you are referring to a source which you have not read yourself, but have read about in another source, for example referring to Jones’ work that you have read about in Smith. You should avoid using secondary references and locate the original source and reference that.
Murdoch University has an even stronger interpretation:
• IEEE style does not allow for the use of secondary source.
• Locate the original source of information which is cited in a work which you have read.
• If an original source cannot be located, it should not be cited.
Jeeze, no wonder there are so many different answers!
So it would seem there cannot be a one-size fits all answer - it depends. What is required in one field and style is forbidden in another, and optional in a third (and apparently even then people differ).
Of course you should cite the paper.
The content of the paper should be focussed on your topic and its field. But, this does not apply to the citations within the content. There is no such rule stating that the papers you cite should be within the scope of your topic.
Anything you refer in any domain deserves to be cited as long as you utilise its information within your publication (be it paper or thesis).
As witnessed by the varying answers, there are varying opinions on questions such as this. My opinion is that one should be honest about one's methodology and sources, including intermediate sources. In particular, acknowledgement of the sort of sources that have been helpful in finding more-primary sources. One sort of extreme case is acknowledgement that one has used Wikipedia in various way, e.g., to get started looking for things in a possibly somewhat-unfamiliar direction. No, wikipedia is not a primary, and maybe not even secondary, source, but to acknowledge its use serves several positive purposes. More modestly, but similarly, to acknowledge the helpfulness of survey/review/expository parts of papers is to acknowledge the utility of such things to the scholarly community. This is in contrast to the general editorial and stylistic pressure to not give context, not give too many external pointers, etc. It seems to me highly desirable to encourage/legitimize a broader scholarly writing style in STEM (science, tech, engineering, math, the latter my field) in contrast to a narrow technician's writing style that is nevertheless often adequate for "publishability". In summary, honesty both in sources and in methodology.