Key takeaways
- Patents and papers can address the same mechanism while providing different kinds of evidence about it.
- Compare the material or design, baseline, outcome and operating conditions before deciding whether two sources support the same claim.
- A source-linked comparison should distinguish supported findings, genuine conflicts and questions the available text cannot resolve.
An engineering researcher may find the same proposed solution in patent records and scientific papers. That overlap is useful, but it leaves a harder question: do the documents support the effect the project needs, under the conditions in which it must work?
The walkthrough below follows a team evaluating solid-state battery electrolytes. It moves from relevant patent families to papers, then compares the approaches and the evidence behind them.
Define the claim precisely
Frame the claim around four elements: the material or design, the baseline, the measured outcome and the operating conditions.
For solid-state battery electrolytes, start by identifying the material approach your team is evaluating and the performance claim it needs to check. A list of promising electrolyte types can help you choose what to investigate; a decision about a particular approach needs evidence about its composition, configuration and reported performance.
Write down the intended application and the criteria that would make an approach useful. Keep the reported test conditions beside each result, and mark conditions the source does not provide. This gives the patent and paper searches a shared question to answer.
The distinction matters when you read the results. A patent describing an electrolyte formulation and a paper reporting a property of a related material may address the same research area without supporting the same claim.
Build search terms around the approach and the outcome. Start with solid-state battery electrolytes, then use the material names and technical terms in relevant records to refine the investigation. Keep the search terms and any date limits with the results.
Search patents and papers for different parts of the question
Use patent records to find implementations: how the material is arranged, how components interact and which design variations are described. Read the description and examples as well as the claims. A claimed arrangement identifies technical subject matter; its inclusion in a claim does not itself demonstrate a measured effect.
The walkthrough starts with relevant patent families and short technical summaries. For a useful first pass, retain the publication number, title, assignee, earliest priority date, electrolyte type and technical approach. Open the linked publication to examine the relevant disclosure.
Then add papers on the same topic in the same conversation. Follow DOI links to the original studies before relying on a finding. Keep the patent and paper lists distinct so that each observation remains traceable to its source type.
Use two successive requests, adjusting the scope to your project:
Find relevant patent families on solid-state battery electrolyte materials. Include publication numbers, titles, assignees, earliest priority dates, electrolyte types and short technical summaries.
Now add scientific papers on the same topic. Include the reported findings, source links or DOIs, available test conditions and details that still need checking. Keep the paper evidence separate from the patent disclosures.
If you group the retrieved patents by assignee, describe the result as a view of that search. The walkthrough makes this distinction explicit: a relevance-based set of records is not a global filing count. Use it to identify an approach or organization to investigate further.
Compare evidence under matching conditions
Begin with what you can actually inspect. An abstract may identify a material and summarize an outcome while omitting the test method, configuration or comparison conditions. Open the relevant original text before drawing a conclusion that depends on those details.
For each important claim, compare the following fields:
| Question | What to extract |
|---|---|
| Is it the same material or design? | Electrolyte composition, material arrangement and configuration, where reported |
| Is it the same baseline? | Reference material or cell configuration and what changed in the comparison |
| Is it the same outcome? | Reported property or cell result, with its units and measurement method |
| Are the conditions comparable? | Test conditions, cell configuration and other limits stated in the source |
| What kind of evidence is available? | Proposed mechanism, described example, measured result or interpretation |
Ask the AI assistant to compare the approaches against your team's criteria, then inspect the source passages behind the comparison. For each row, retain the patent publication number, paper DOI or available identifier, and the location of the relevant passage.
Classify the relationship between sources carefully:
- Support under comparable conditions: comparable observations are consistent with the same technical claim under sufficiently similar conditions.
- Difference under nonmatching conditions: the sources use different baselines, configurations or operating conditions, so the difference does not yet establish a contradiction.
- Unresolved conflict: comparable observations differ and the available information does not explain why.
- Evidence gap: a necessary condition or result is absent from the inspected text.
Also check whether the sources are independent. A paper cited by a patent may be the basis for the patent’s technical statement. Treating that pair as two independent confirmations would overstate the evidence.
Write a conclusion that directs the next investigation
Bring the two sets of evidence together with three questions: where do the retrieved patents and papers overlap, where do they differ, and what still needs evidence? Answer for the documents you inspected. In the comparison, separate proposed formulations, reported results and unresolved questions.
Record source identifiers and relevant passages, and note any source text you could not access. Keep the search question, terms and date with the comparison.
Then name the next task: obtain the full methods, inspect a patent example, search for a missing operating condition or design a comparison under the team’s own conditions. This makes the cross-check useful to a technical review rather than leaving the team with two disconnected reading lists.
Once you know the sections you need, you can put the full research brief into one request. The walkthrough uses successive questions to make the reasoning visible; the same scope can also be stated at the outset.
Use Patsnap search in your AI research workflow
Patsnap's Patent & Literature Search MCP brings patent and paper retrieval into an MCP-compatible AI client. The Patsnap Eureka walkthrough shows the research sequence in an AI conversation: find patent families, add papers, compare approaches and inspect the evidence behind the answer.
The search interface selects one source type per call, so a request covering both requires separate patent and paper searches. The fetch tool provides patent descriptions, claims and drawings, while paper retrieval provides basic abstract metadata. Use the original publication for methods and other details that are not present in the retrieved record.
Connect the search tool in your client, then start with one question your team is already trying to answer.
Explore Patsnap Patent & Literature Search
