FTIR.fun FTIR.fun Make FTIR Analysis Easier
RESULT PAGE

Epoxy resin

View the report above. Use the comments below if you need follow-up discussion.

Result No.: 20260520111818017698846 Owner: aadhithya Comments: 0
FTIR ANALYSIS REPORT

Spectrum Analysis Result

No.: 20260520111818017698846 Date: 2026-05-20 05:52:20 Reported by: FTIR.fun Contact: [email protected]

Download a professional, formatted PDF of this analysis.

Executive conclusion

Epoxy resin

Class / direction
Moderate confidence
#151435 Initial rank 1 Current rank 1 Library lead match 84.2%
Result
  1. Aggregate family-key resolution (llm_existing_family_choice): Library, KG, and LLM diverge materially: the library axis gives the strongest single-family assignment and specifically clusters its top-15 consensus to epoxy_resin at 0.743, though still below the 0.8 lock threshold; the KG axis is weaker and broader, only supporting a nonspecific oxygen_containing direction (0.397) from peaks/groups such as 1646 and 3420 plus mixed oxygen/metal/nitrogen clues, with no usable literature backing because online enrichment failed; the LLM axis is literature-backed for an amide/protein-like interpretation around 1646 and 3420 (amide_like, 0.58), but that evidence is functional-group level and does not identify a tighter canonical family from the allowed list. Given this mismatch, the best aggregate family is epoxy_resin because it is the most specific defensible canonical assignment available, it comes from the highest-confidence axis, and it still remains broadly compatible with the KG oxygen-containing signal, whereas the LLM amide-like reading appears narrower and potentially driven by a shared 1646 band that is not sufficient on its own to override the stronger library family call.
  2. Library view: Library retrieval keeps Sassafras oil (#151435) as the nearest match.
  3. Reference clues: KG derives the direction oxygen containing from group-level evidence. top graph recall is spectrum #17826. Online literature search did not yield a usable peak-backed or direction-matched source. Errors: google_scholar:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116) | springer:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116) | crossref:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116). semantic bridge links the current direction to known directions such as "acetic acid; aniline", "acetic acid; aniline". secondary component clues remain visible through nitrogen containing, main group element containing.
Primary risk

Independent RAG axis flags conflicting groups: missing clear amide ii support in provided peaks

Library comparison

Reference spectrum for the identified material type vs. uploaded sample

This comparison uses a reference spectrum chosen by the analyzer for the identified material type (e.g. siloxane). It may differ from the library similarity Top-1 shown in the verification workbench below, because the Top-1 is ranked purely by spectral similarity without material-type reasoning.

Library top match ready.
Figure 1. Overlay of the chosen library reference and the uploaded sample (teal curve).
Show Top-15 verification workbench (similarity-ranked) These 15 candidates are ranked by spectral similarity only, without material-type reasoning. The top match here may differ from the reference spectrum above.
Library spectrum loads when this workbench opens.
Library spectrum Interactive sample curve Move the pointer to show the vertical guide line.
Top 15 candidates

Reference library candidates

This list is a verification appendix. It helps you compare nearby candidates without repeating the report conclusion.

Rank Match % Compound Name Formula / SMILES Library preview Action
Reference candidates load when this verification workbench opens.
Peak analysis

Detected peaks and interpretation

Rows marked with ★ are the characteristic peaks selected by the LLM analyzer.

Index Characteristic Wavenumber Absorbance Sources One-line interpretation KG citation LLM confidence
1 ★ 3420 0.76 RAG / KG 3420 cm-1: possible source: proteinaceous / amide-containing organic material; water-bearing / hydroxyl-containing organic residue; KG hint: oxygen; nitrogen. - Low confidence
2 ★ 1646 0.38 RAG / KG 1646 cm-1: possible source: proteinaceous / amide-containing organic material; water-bearing / hydroxyl-containing organic residue; KG hint: amide; oxygen. - Low confidence
3 ★ 765 0.12 RAG / KG 765 cm-1: possible source: aromatic-containing organic material; KG hint: oxygen; nitrogen. - Low confidence
4 ★ 2071 0.04 RAG / KG 2071 cm-1: possible source: insufficiently resolved minor feature around 2071 cm-1; KG hint: oxygen; nitrogen. - Low confidence
Show full reasoning, evidence, and recommendations Expand to read the conclusion evidence, reasoning chain, full risk list, recommendations, and mixture notes.
Evidence

Conclusion evidence

Only evidence already present in the current result contract is shown here.

Library lead candidate LIB
Epoxy resin #151435 | match 84.2%
Material direction AGG
Epoxy resin Likely material direction: proteinaceous / amide-containing organic material (57% support).
Supporting peaks
1646 cm-1 3420 cm-1 765 cm-1
Supporting groups
amide_i o_h_/_n_h aldehyde amide carboxyl metal_hydroxyl metal_oxygen metalloid_oxygen
Reasoning

Reasoning chain

This section explains how the current conclusion was reached from the available evidence.

  1. Aggregate family-key resolution keeps the final direction at "Epoxy resin" based on the tri-axis family decision rules.
  2. Current policy keeps this result at material-direction level instead of a specific substance.
  3. Library retrieval keeps Sassafras oil (#151435) as the nearest match.
  4. KG derives the direction oxygen containing from group-level evidence. top graph recall is spectrum #17826. Online literature search did not yield a usable peak-backed or direction-matched source. Errors: google_scholar:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116) | springer:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116) | crossref:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116). semantic bridge links the current direction to known directions such as "acetic acid; aniline", "acetic acid; aniline". secondary component clues remain visible through nitrogen containing, main group element containing.
  5. The strongest locally supported signal is 1646 cm-1, which repeatedly maps to amide I / protein-related absorption in the retrieved literature: disordered protein/amide I around 1645-1657 (S1), protein amide I near 1650 with 1600-1700 cm-1 base region (S3), amide I/protein around 1641-1650 (S2, S4), and protein random coil near 1644 (S5). The broad 3420 cm-1 band is also locally associated with hydroxyl functional groups and aromatic protein-like material in one retrieved source (S6), but the sample itself flags this region as possible moisture background, so it is supportive only weakly. Overall this supports an amide-containing organic direction, likely proteinaceous at class level, but not a specific substance.
  6. Peak notes: 3420: LLM: proteinaceous / amide-containing organic material; water-bearing / hydroxyl-containing organic residue | KG: oxygen; nitrogen; 1646: LLM: proteinaceous / amide-containing organic material; water-bearing / hydroxyl-containing organic residue | KG: amide; oxygen; 765: LLM: aromatic-containing organic material | KG: oxygen; nitrogen; 2071: LLM: insufficiently resolved minor feature around 2071 cm-1 | KG: oxygen; nitrogen.
Risk

Uncertainty and blockers

These items limit how far the current conclusion can be trusted or refined.

  • Independent RAG axis flags conflicting groups: missing clear amide ii support in provided peaks
  • dual-axis confidence 0.397 is below dynamic threshold 0.770
  • LLM/KG agreement 0.060 is below 0.600
  • Library lead over the runner-up remains narrow (gap 0.028)
  • Broad O-H / moisture background is visible above 3400 cm-1
  • KG limit: online literature enrichment failed: google_scholar:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116)
  • online literature enrichment failed: google_scholar:multiple sample side groups align with the curated family direction oxygen containing semantic bridge links this direction to known direction "acetic acid; aniline". 765 cm-1 stretching vibration FTIR: Online literature cooldown state is not valid JSON: /home/bob/projects/ftirfun/runtime/online_literature_engine_cooldowns.json: Extra data: line 46 column 2 (char 2116)
Recommendation

Recommended next steps

These are the next verification actions suggested by the current reasoning output.

  • Reacquire or reprocess the spectrum before making a stronger identification claim because Broad O-H / moisture background is visible above 3400 cm-1.
  • Use discriminative bands around 1646, 3420, 765 to decide between the LLM-supported proteinaceous / amide-containing organic material direction and the KG-supported oxygen containing direction.
  • Compare the library lead against the nearest runner-up explicitly because the current lead gap is only 0.028.
  • If mixture remains plausible after the peak-level checks, reopen component peeling only after better library coverage or cleaner acquisition is available.
Mixture

Component and mixture notes

When present, mixture or component-peeling diagnostics are summarized here.

  • Two-component formulation analysis is paused because the current library does not have enough coverage for reliable formulation-level inference.
  • The old two-component code path is kept in the repository, but the runtime no longer uses it to avoid pushing the search toward unsupported composition claims.
  • Component peeling is paused in runtime because conservative low-match replay did not produce verified wins after excluding self-peeling.
  • The exploration was still useful: relaxing the gates restored peelability, but most new extractions were self-peeling or chemically implausible, which clarified the current boundary of the library and gating design.
  • The peeling implementation remains in the repository for later reactivation after stronger low-match evidence appears.
Appendix

Sample information and raw spectrum

The raw uploaded spectrum is retained here for audit and manual review.

Baseline correction method: Asymmetric Least Squares Smoothing

The wavelength range for analysis(cm-1): [(400, 4000)]

Raw spectrum without baseline correction or other processing:

Raw spectrum image
Discussion

Comments and follow-up evidence

Use this area to continue the interpretation, ask questions, or add extra verification evidence.

  • Page 1 of 1
Reply
RELATED TOPICS

Other analyses in this category

Submit Request Form