We've had a little debate at work about whether a paper that presents good research, but that is at the same time very poorly written (either sloppy presentation, or not written in the native language of the authors), would nevertheless get a lot of citations. Are there good examples of such papers, specifically ones that present seminal work in some field, but that are very difficult to read because of the language or are riddled with errors?
I think Heidegger's Being and Time might fit this criteria. Though the philosopher's work was important to the field, he was criticized for the use of convoluted, imprecise language that managed to confuse other philosophers (Bertrand Russell being one such critic). In fact, the philosopher Paul Edwards published a collection of articles under the title Heidegger's Confusions in order to
rescue [Heidegger’s] valuable ideas from the logical confusions in which they are embedded and from the willfully obscure and perverse language in which they are frequently expressed.
With regard to more recent (and research-oriented) writing: talk of the poor grammatical quality of academic journals has been around a while. Take a look at this quote from the Chronicle article "Bad Writing and Bad Thinking"
Many people—publishers of scholarly work, editors at higher-education publications, agents looking for academic authors capable of writing trade books—who think about the general quality of scholarly prose would admit that we're in a sorry state, and most would say there isn't much to do about it.
So, although an example other than Heidegger's writing doesn't come to mind, it doesn't seem you'll have to look very far.