Should the impact factor of the journal it is published in be important, or is it based on its citations? I think individual qualities of papers should be used to judge the same.
Although I have a rather limited experience with publishing, but more with reading journal articles, I have observed that good quality journal papers share the following characterstics:
(1) The paper is well written from a scientific language point of view (average Fog Number = 18),
(2) The idea of the paper is well articulated and clear,
(3) The different sections of the paper are logically coherent and flow naturally,
(4) Relevant concepts are defined,
The above points may form the basis for judging the quality of a paper.
Paper either should have contributed something new, or analysed the problem from a new perspective bringing out hidden dimensions of the problem and be extendable.
I agree that a good paper should contribute something to science (e.g. in terms of a new topic, angle or method, a good overview of literature of some field(s) or a combination of these) and it should be easily understandable for scholars in the same field and maybe also in some other fields. It should reach a relatively wide audience: e.g. a conference paper would not easily achieve that, so it should be published in a relatively popular journal (book chapters are usually not read so much). The number of citations indicates the impact of the paper to some extent, although in narrow fields, reaching a certain number of citations is more difficult than in wide fields.