The literature

Search tips: finding and reading nanoscience papers

Searching the nanoscience literature well is mostly a matter of knowing its vocabulary and its habits. Reading it well is a matter of knowing where a paper keeps its evidence. This page covers both, in that order.

Searching: three handles on a paper

Almost every search tool lets you look for a paper in three ways: by author, by words in the title or abstract, and by journal. Each has its own traps.

Authors. Surnames repeat, initials vary, and transliterated names can be spelled several ways. When a search by name returns too much, add a second author you know works with the first, or restrict by journal or year. When it returns too little, try the surname with only the first initial and drop any accents. Author searches are most useful once you already know who works on a subject; they are a poor way to discover a field.

Title and abstract words. The nanoscience vocabulary is unstable. The same object may be called a nanocrystal, a quantum dot or a colloidal semiconductor particle; a nanotube may be single-walled, SWNT or SWCNT; a technique may appear as scanning tunnelling microscopy, scanning tunneling microscopy or STM. Search for the variants deliberately, using OR between them, and use truncation where the tool allows it so that one stem catches plurals and related forms. Put multi-word phrases in quotation marks when you mean the phrase and not the separate words.

Journals. Restricting a search to a journal is a quick way to see what one community is saying about a subject. It also helps when a phrase is common in one field and means something else in another. Journal abbreviations vary between databases, so check how a given tool expects the name to be written.

Using a controlled vocabulary

Biomedical databases add a layer that general search engines lack. The PubMed user guide explains that the system tries to map what you type onto its own indexing vocabulary, MeSH, through a process it calls Automatic Term Mapping, and that quotation marks can be used to search for an exact phrase instead. It also describes searching by author, journal and date, and combining terms with AND, OR and NOT. For the biological edge of nanoscience, from nanoparticle toxicity to peptide materials, starting from the controlled vocabulary and then narrowing with free-text words usually gives cleaner results than free text alone.

Reading: the abstract first, but not only the abstract

An abstract, as the Wikipedia article on abstracts puts it, is a brief summary placed at the start of a paper to help a reader see its purpose quickly; some journals use structured abstracts with headings such as background, methods, results and conclusions. Read it to decide whether the paper deserves more time. Then read it again at the end, because an abstract is also the authors' own advertisement for the work and is often the most confident sentence-by-sentence statement in the paper.

Figures and microscopy images

In nanoscience the figures carry most of the argument. Go to them early. For a microscopy image, check the scale bar and ask how the region shown was chosen: a single beautiful image of an ordered array says little about how common that order is across the sample. Look for statistics such as size distributions drawn from many particles, and for complementary measurements that confirm what an image suggests. For spectra and electrical measurements, look at the axes, the error bars and the number of samples or devices measured. A device result reported from one exceptional device is a demonstration, not a characterisation.

Methods and supplementary information

Short-format journals push much of the detail into supplementary files: synthesis recipes, additional images, control experiments and data from samples that did not work as well. For anyone trying to judge or reproduce a result, those files are essential reading. A method section that cannot be followed, or a supplement that contains only the best examples, is a signal to read the main claims with more caution.

Peer review and what it guarantees

Journal papers have passed peer review, which the Wikipedia article on peer review describes as evaluation of work by others with similar expertise, used in academia to decide whether a paper is suitable for publication. Review catches many errors and improves most papers, but it is performed by a few busy people on a fixed deadline. It does not reproduce experiments, and it cannot see data that were not submitted. Treat a published paper as a claim that has survived one round of scrutiny, not as settled fact. Later papers that confirm, extend or contradict it are the real test.

Preprints and published versions

Many physics and materials papers exist in two forms: a preprint posted before or during review, and the final journal version. The two can differ in their figures, their conclusions or even their authors. When you cite or rely on a result, note which version you read, and if only a preprint exists, say so. The page on keeping up with new papers describes preprint services, and the journals page explains which titles carry which kinds of work.

A short checklist

• Search for vocabulary variants, not just the term you know best.
• Read the figures before the discussion.
• Check scale bars, sample numbers and error bars.
• Open the supplementary files for anything you intend to rely on.
• Look for later papers that test the same claim.
• Record which version of a paper you read.

None of this requires specialist training. It requires only the habit of asking, of each figure and each claim, what evidence stands behind it and how that evidence was chosen. Readers who build that habit find that the nanoscience literature, for all its jargon, becomes one of the more rewarding bodies of scientific writing to follow.