Negative Results Policy

Papers are not rejected due to lower efficiency or negative trends. Proper controls and analysis are the key acceptance criteria.

The problem

Photovoltaic literature is shaped by what gets published, and what gets published is dominated by improvement. The consequence is a field in which the same unsuccessful approaches are attempted repeatedly in different laboratories, in which degradation and failure are systematically under-reported relative to their practical importance, and in which reproducibility is difficult to assess because unsuccessful reproduction attempts rarely appear in print.

This is not a failure of individual researchers. It is a structural feature of a publishing system that rewards records. Solar Cell Insights is an attempt to provide a venue where the incentive runs differently.

What we mean

Negative result	A well-designed experiment in which the tested hypothesis was not supported — a modification that did not improve performance, an additive that did not passivate, an architecture that did not deliver the predicted gain.
Null result	A study in which no statistically or physically meaningful effect was observed, where the measurement was sensitive enough for the absence to be informative.
Non-optimised result	A device or process studied for the mechanism it reveals rather than the performance it reaches. The efficiency is not competitive and was never intended to be.
Failed reproduction	A careful attempt to reproduce a published result that did not succeed, with the discrepancy analysed rather than merely reported.

What we require

A negative result is only useful if the reader can trust that the absence of an effect is real rather than an artefact of the experiment. Manuscripts in this category must therefore include:

• A clearly stated hypothesis or expectation, and why it was reasonable
• Controls sufficient to distinguish a genuine absence of effect from an inadequate experiment
• Full description of materials, fabrication conditions, environment and measurement protocol
• Sample statistics — how many devices, batch-to-batch variation, and the spread of the data, not only champion or mean values
• An honest assessment of sensitivity: what size of effect the experiment could have detected
• Interpretation that stays within what the data support, including alternative explanations the authors cannot rule out

What we will not publish

We do not publish work that failed because it was poorly executed. An experiment without controls, without adequate methodological detail, or with too few samples to support any conclusion is not a negative result — it is an inconclusive one. The distinction is the whole basis of this policy, and reviewers are instructed to apply it.

How these manuscripts are reviewed

Reviewers are explicitly instructed that a modest or absent effect is not a reason for rejection, and are asked instead to judge whether the experiment was capable of detecting the effect in question and whether the conclusions are proportionate to the evidence. The review form used by the journal contains a specific question on this point.

How they are labelled and cited

Articles reporting negative, null or non-optimised results are published as full articles with a DOI, indexed in the same way as any other article, and carry a "Negative or Null Result" label in their metadata and on the article page. The label exists so that the work is discoverable by researchers looking for exactly this kind of study — not to mark it as a lesser category.

Examples of work we are looking for

• A passivation strategy that works in one perovskite composition and not in another, with the interfacial chemistry analysed
• A systematic study of batch-to-batch variability in a widely used fabrication route
• An attempt to reproduce a published stability result that did not succeed, with the differences in protocol identified
• A characterisation method that gives misleading values under commonly used measurement conditions
• A promising additive class that produced no measurable improvement across a well-controlled series