If you are writing a science research paper, you should avoid topics that you cannot prove through the scientific method. This means that vague, philosophical, and theological topics are all out -- you need to be able to show how a scientific process or topic was proved by stating a problem, creating a hypothesis and carrying out an experiment. If you cannot do this for a science topic you should avoid it for science research papers.
Science research papers should also avoid topics on which you have strong opinions. Even if your experiment and research are both completely unbiased, the fact that you have such strong opinions will influence people's perception of the results. What's more, it may influence them subconsciously, even if you take stringent steps to avoid this. It is best to research topics that you feel neutral about, and let other scientists research topics about which you feel strongly.
All research papers should avoid topics that are vague, whether they are science papers, political science papers, commerce papers or papers that deal with any other subject. The reason for this is because it makes research too difficult. If you are going to write a paper on war, you are going to need years and years and a word limit in the millions. Rather, you should approach a topic from an angle -- war and economics, war from a specific nation's viewpoint, or war and culture in relation to how one group changed after it fought in a war. These are much easier to research and write about in a clear, coherent manner that does the topic the justice it deserves.
You should also avoid topics that have been beaten to death. Even if you have a strong interest in an issue, if it has been researched and written about by thousands of other writers the odds are that yours will just fade away into white noise. A research paper's effectiveness depends on whether people read it or not; if you write about something that people have been reading about for eons they will avoid it, thus wasting your research and writing time.