To translate CAS Registry Numbers, EC numbers and UN numbers accurately, a translator must separate chemical substance identity from dangerous-goods transport identity. Safety data sheets, regulatory dossiers, laboratory records, labels, customs documents and shipping papers can contain chemical names that need translation beside numerical identifiers that must remain exact. Changing a code can connect the target document to a different substance or transport entry even when the surrounding prose sounds correct.
This guide explains how to translate chemical and dangerous-goods documents without changing CAS RNs, EC numbers or UN numbers. It solves one specific search intent: what does each identifier actually identify, which labels and chemical names can be localized, which values must remain protected, and how can translators verify that a multilingual document still refers to the same substance and the same transport classification entry?
The working rule is preserve the identifier, translate the human-language description, and verify the relationship between them. CAS describes a CAS Registry Number as a unique identifier for a specific substance. ECHA uses EC numbers as numerical identifiers for substances in the EC Inventory and distinguishes them from ECHA list numbers. UN numbers belong to the dangerous-goods transport system and identify entries in the UN Model Regulations. This specialist branch connects to Master Art of Translation, the protected Vocabulary Learning Hub, and How English Works without replacing the existing SDS/GHS translation owner.
A one-minute orientation: substance identity and transport identity are different layers
A chemical substance can have a CAS RN and an EC number, while a dangerous-goods shipment involving that substance can also be associated with a UN number and proper shipping name. These codes may appear together because one document connects chemistry with transport. They are not interchangeable identifiers.
Before translating, ask what each field identifies. Is it the substance itself, a European regulatory inventory entry, or a transport entry defined by hazard and shipping conditions? The correct translation behavior follows from that classification.
1. Chemical names are language; identifiers are data
A chemical can have systematic names, common names, trade names, abbreviations and names in many languages. The identifier exists partly because names can vary. Translation therefore changes human-readable naming while preserving the stable code used to connect records.
Do not translate digits inside CAS, EC or UN identifiers. Do not localize hyphens or replace them with typographic dashes. Copy the code from an authoritative record and verify that the translated chemical name still refers to the same entry.
2. CAS RN identifies a specific substance record
CAS states that a CAS Registry Number provides a unique and unambiguous identifier for a specific substance within CAS REGISTRY. That makes it a powerful anchor when chemical names differ across languages, naming systems or commercial contexts.
A translator can render the chemical name into an approved target-language form while retaining the CAS RN. If the target name is uncertain, the identifier helps the reviewer check whether both languages point to the same substance.
3. CAS number formatting should be preserved
CAS RNs are commonly displayed as groups of digits separated by hyphens. The hyphens are part of the conventional representation. Do not change them to en dashes, spaces or target-language punctuation merely to match typography.
Word processors can automatically replace hyphens with typographic dashes. Disable smart punctuation in protected fields or store identifiers separately from running prose so formatting automation cannot alter them.
4. A check digit helps detect some CAS transcription errors
CAS numbers include a check digit that can help identify certain transcription mistakes. That is useful for QA, but a syntactically valid CAS RN can still be the wrong substance if it was copied from another row.
Use validation as one layer only. The stronger test is whether the authoritative substance record matches the name, formula and context described in the source document.
5. Common chemical names can hide identity ambiguity
Everyday chemical names can refer broadly to mixtures, families or substances whose exact form matters. Translation can make this ambiguity worse if a familiar target term is broader than the source. The CAS RN is therefore useful as an identity anchor, not as a substitute for semantic judgment.
If the source name and CAS RN appear inconsistent, do not choose the one that “sounds right.” Flag the discrepancy and verify it against the responsible chemical database or document owner.
6. Trade names should not replace chemical identity
A commercial product can have a brand or trade name alongside one or more chemical constituents. A trade name is not automatically the chemical name, and its brand identity may not translate at all. Preserve the distinction between product name and substance identifier.
When a product contains several ingredients, each ingredient may have its own CAS RN. Do not assign the product name one constituent’s CAS RN as though they were the same object.
7. Mixtures complicate the idea of a single substance code
Mixtures can contain multiple substances and may be regulated or described at the product level. Translators should not invent a single CAS RN for a mixture simply because the document contains several component identifiers.
Preserve the document structure: product identity, ingredient identity, concentration and hazard information are separate fields. If an identifier is blank, leave the absence visible rather than filling it with the nearest code.
8. EC numbers belong to the European chemicals identity framework
ECHA explains that an EC number is the numerical identifier for substances in the EC Inventory, which combines historical European substance lists. In European regulatory documentation, the EC number can therefore function as a core substance identifier alongside CAS information.
Translate the label “EC number” according to established regulatory terminology if necessary, but preserve the numerical code. Do not replace an EC number with a CAS RN simply because both identify the substance in different systems.
9. ECHA list numbers are not the same as official EC numbers
ECHA explicitly distinguishes EC numbers from ECHA list numbers. Its guidance notes that list numbers are technical identifiers and do not have the same legal status as official EC numbers. They can share a similar visual format, which makes careless translation dangerous.
If the source says “List number,” preserve that label rather than upgrading it to “EC number.” A translation should not make a technical identifier appear to have regulatory status the source did not claim.
10. Similar formatting does not make two identifier systems equivalent
CAS and EC identifiers both contain digits and hyphens, but their grouping patterns and issuing systems differ. Translators should not infer identity type from visual appearance alone when labels have been lost during extraction.
Use the surrounding field name, database schema and authoritative record. If a PDF extraction merges columns, inspect the original page image before assigning codes to labels.
11. UN numbers belong to dangerous-goods transport classification
The UN Model Regulations on the Transport of Dangerous Goods organize dangerous-goods entries by UN number and proper shipping name. The UN number is therefore a transport identity, not simply another chemical database number.
A single transport entry can cover substances or articles under defined conditions, and a chemical substance can have transport treatment that depends on concentration, physical form or hazard classification. Do not treat a CAS RN and UN number as one-to-one translations of each other.
12. Proper shipping name and UN number must remain paired
Dangerous-goods documentation combines a UN number with a proper shipping name and other transport information such as hazard class or packing group where applicable. The name is language; the UN number is the code. Their relationship is the important unit.
If the target-language proper shipping name is prescribed by the applicable transport regime, use the official form. Do not create a casual translation that changes the scope of the UN entry while leaving the number unchanged.
13. A UN number is not a CAS number with fewer digits
UN numbers are used to identify dangerous-goods entries for transport. CAS numbers identify specific substances in a chemical registry. One is rooted in transport regulation; the other in substance identity. Their functions overlap only because chemical transport needs both chemistry and logistics.
Translate explanatory text so the reader can see this distinction. Labels such as “CAS No.” and “UN No.” should remain unambiguous even when they appear in the same table.
14. UN numbers can cover articles as well as substances
The dangerous-goods system includes entries for articles and products such as batteries, equipment or other regulated items, not only pure chemical substances. This is another reason a UN number cannot be interpreted as a chemical-substance identifier.
Translate the proper shipping name and technical description according to the applicable rules, while leaving the UN number intact. The target document should preserve the same regulated transport entry.
15. Hazard class is a separate field from UN number
Dangerous-goods entries also have class or division information. Do not merge the class number into the UN number or translate it as ordinary prose. It is another structured regulatory field.
When the target language has official names for hazard classes, translate the class description while preserving the numerical class designation. This makes the record understandable to humans without altering regulatory structure.
16. Packing group carries degree-of-danger information
Some dangerous-goods entries use packing groups such as I, II or III. These Roman numerals are not headings to be renumbered by an editor. They carry transport meaning within the classification.
Translate “Packing Group” if necessary, but preserve the assigned group. A layout tool that converts Roman numerals to local list numbering can create a serious compliance error.
17. Hazard statements and transport entries belong to different systems
GHS hazard statements, pictograms and signal words support chemical hazard communication, while UN numbers and proper shipping names support dangerous-goods transport identification. The same product may contain both systems on one SDS or label.
This article therefore complements rather than replaces Top Ways to Translate Correctly | Safety Data Sheets, GHS Labels and Hazard Statements. The existing owner covers hazard communication broadly; this guide owns identifier integrity.
18. Chemical formulas are another protected representation
A molecular formula uses element symbols and subscripts to describe composition. It should not be translated into words inside a formula field. The descriptive chemical name may be localized while formula and identifiers remain exact.
If the source formula conflicts with the CAS RN or name, flag the inconsistency. Do not silently choose which representation is “probably” correct.
19. Stereochemistry and salt forms can change substance identity
Chemical names that differ only by stereochemical descriptor, hydration state, salt form or other qualifier can refer to different substances or regulatory records. A translator who drops a small qualifier can make the target name inconsistent with the identifier.
Preserve technical qualifiers and consult approved nomenclature. Small lexical details can carry large identity consequences in chemistry.
20. Concentration belongs to the formulation, not the identifier
An SDS may list the same chemical with a concentration range. The CAS RN remains the substance identifier while the percentage describes how much is present. Do not incorporate concentration into the identifier or translate a percent range into a different basis.
Keep concentration units, ranges and inequality signs exact unless the target regulatory format explicitly requires another representation. The relationship between identity and amount must survive together.
21. Synonyms should not be mistaken for separate substances automatically
A substance can have several names: systematic, common, historical, trade or translated names. The CAS RN helps connect them. A target terminology database should therefore record synonyms around one verified substance identity rather than treating every string as a new chemical.
At the same time, do not merge similar-looking names merely because they share a root word. Verify the identifier and structure before declaring them synonyms.
22. Regulatory names may differ from everyday laboratory language
Scientists may use a short common name while regulatory databases use a more formal substance name. Translation should respect the function of the document. An SDS or regulatory filing may need the official name; a teaching explanation may add a familiar synonym.
Do not replace an official substance name with a friendlier term if that weakens traceability. Readers can be given both, clearly labeled, while the identifier anchors the record.
23. Language versions of regulatory databases can help terminology, not replace verification
ECHA and other authorities publish multilingual terminology and substance information. These resources can help translators choose established target-language forms. The fact that a translated name exists does not remove the need to check that the accompanying identifier is the same substance record.
Copy the official target name and code together when possible. This reduces the risk of pairing a correct translation with a neighboring substance’s identifier.
24. Tables need protected identifier columns
Chemical inventories commonly use columns for product name, substance name, CAS RN, EC number, concentration, UN number and hazard class. Protect the code columns from translation. Expose only the fields that genuinely require language work.
After export, compare row counts and key fields. A row shift can produce a chemically plausible but completely wrong combination of name and identifier, which ordinary spell-checking will not detect.
25. Spreadsheet import settings can damage chemical identifiers
CAS RNs and EC numbers contain punctuation that can be reinterpreted by spreadsheet software, while UN numbers can be treated as ordinary integers. Store all identifier fields as text. Automatic formatting can remove leading zeros, replace hyphens or turn values into dates.
Define column types before opening large inventories. A translation workflow should never depend on whatever type a spreadsheet guesses from the first few rows.
26. CSV exports can preserve errors perfectly
A CSV file can carry damaged identifiers just as faithfully as correct ones. If software already converted a CAS RN or stripped punctuation, exporting the sheet does not restore the original. Keep a protected source file and compare identifiers after every round trip.
For high-volume chemical localization, use automated source-target equality checks for CAS, EC and UN fields. This is faster and more reliable than visual inspection alone.
27. OCR is risky in old SDS files and scanned labels
Scanned safety sheets can confuse 0 with O, 1 with I, 5 with S, or join punctuation to nearby digits. An OCR system can also split a hyphenated CAS RN across lines. A plausible-looking code may therefore be a transcription artifact.
Verify critical identifiers against CAS, ECHA, supplier data or another authoritative source. Do not “correct” a code merely because a guessed version returns a search result.
28. PDF extraction can merge chemical table columns
A multi-column SDS may visually separate CAS number, EC number and concentration, yet text extraction can concatenate them. Translators working from extracted text should inspect the original layout whenever field boundaries are unclear.
The rule is simple: never assign an unlabeled code to a substance based only on position in extracted text. Restore the table structure first, then translate.
29. Decimal commas and percentage signs belong to amounts, not identifiers
Target locales may use different decimal separators for concentration values. That formatting decision should not spill into CAS or EC numbers. Chemical identity fields and quantity fields follow different rules even when both contain digits and punctuation.
If a document converts 12.5% to 12,5% for locale display, the CAS RN beside it should remain exactly the same. Treat formatting rules as field-specific.
30. UN proper shipping names need official-language discipline
Dangerous-goods rules prescribe proper shipping names. A translator should use the form recognized by the applicable transport regime rather than creating a free paraphrase. The purpose is not literary elegance but consistent transport identification.
If the source language and destination language use official versions of the regulations, align the UN number with the corresponding authorized proper shipping name. Do not translate one word at a time and assume the result has regulatory status.
31. N.O.S. entries require careful preservation of scope
Dangerous-goods lists contain “not otherwise specified” entries. These names can be easy to over-simplify in translation. The qualifier indicates a regulatory category, not casual uncertainty.
Use the official target-language abbreviation or phrase required by the applicable rules. Preserve technical names or qualifiers that the shipping description requires, and keep the UN number paired with the same entry.
32. Technical names in transport descriptions can differ from marketing names
A commercial product may be shipped under a regulatory proper shipping name that looks nothing like the label brand. Do not replace the shipping name with the product name simply because it is more familiar to customers.
Translate each identity layer according to its function: brand for product communication, chemical name for substance identity, proper shipping name for transport compliance, and identifiers for exact lookup.
33. Physical form can affect the dangerous-goods entry
Solid, liquid, gas, solution and mixture qualifiers can change which transport entry applies. A translator who drops “solid” or “liquid” from a proper shipping name can weaken the relationship between wording and UN number.
Preserve state and form qualifiers exactly where the regulated entry requires them. Do not infer physical form from a chemical name alone.
34. Concentration thresholds can change transport classification
A substance may move between transport entries or regulatory treatments depending on concentration, formulation or other criteria. Translation must preserve those threshold statements and numerical conditions. This is another reason CAS identity and UN transport identity cannot be mechanically mapped one-to-one.
When the source describes a conditional classification, translate the condition rather than simplifying it into one unconditional label.
35. Hydrates and solvates can have distinct registry identities
A hydrate, anhydrous material or solvate can have different substance records. A casual target-language name that omits the qualifier may point readers toward the wrong CAS RN or regulatory entry.
Keep stoichiometric and form descriptors visible. If the project glossary lacks an approved translation, verify nomenclature rather than shortening the name for readability.
36. Isomers can be different substances even when formulas match
Structural or stereochemical isomers can share molecular formulas while having distinct identities and properties. Translation must therefore preserve locants, stereochemical descriptors and other naming detail that distinguishes them.
Do not use the molecular formula alone to “confirm” a CAS RN. Substance identity requires more than matching atom counts.
37. Polymer and UVCB identities may not behave like simple small molecules
Some regulated substances are polymers or substances of unknown or variable composition, complex reaction products or biological materials. Their identity descriptions can be more complex than a single molecular formula.
Translate the official substance description faithfully and retain its identifiers. Avoid “simplifying” a long regulatory name into a common chemical that only resembles part of the composition.
38. Supplier-specific product codes are a separate layer again
An SDS can include catalogue number, product code, CAS RN, EC number and UN number. Catalogue and product codes identify the supplier’s commercial item; CAS and EC identify substances; UN identifies dangerous-goods transport entries. Keep all labels distinct.
This layered structure mirrors other technical domains in the Translate series: one document can legitimately carry several identifiers because each system answers a different question.
39. Confidential composition must stay confidential in translation
Chemical documents can omit or mask certain identifiers when confidentiality rules permit. A translator should reproduce the authorized disclosure level. Do not search external databases and insert a hidden CAS RN merely because you think you have identified the ingredient.
Translation is not declassification. Preserve phrases such as “proprietary,” “confidential,” or “not disclosed” according to the source and legal brief.
40. Multilingual synonyms need a controlled substance glossary
Build a terminology resource containing source name, approved target name, CAS RN, EC number, formula and relevant notes. Use the identifier to connect synonyms, but let a qualified reviewer confirm that all names truly refer to the same substance.
This is where the Vocabulary Learning Hub idea becomes technically useful: meaning is learned through relationships. In chemical translation, those relationships should be anchored to verified substance identity.
41. Build a separate dangerous-goods lookup table
Do not overload the substance glossary with transport assumptions. Maintain a transport table containing UN number, proper shipping name, class, subsidiary risk, packing group and any project-specific mode or jurisdiction notes.
This keeps chemical identity and transport classification connected but not conflated. Reviewers can see when one substance appears under different shipping conditions.
42. Automated QA should check identifier equality and label pairing
Extract protected identifiers from source and target. Compare CAS, EC and UN fields exactly, then check that each remains attached to the same row and label. Pattern matching alone cannot detect a valid code attached to the wrong chemical.
Flag additions as well as deletions. A translator should not introduce a new identifier into the target unless the project explicitly authorizes data enrichment.
43. Worked example: one substance, three identifiers
Imagine a fictional chemical with a CAS RN and EC number that is transported under a particular UN entry. The target document translates the chemical name and proper shipping name but preserves all three codes. Each identifier continues to answer its original question.
The example works because the translator does not assume that the three numbers are competing synonyms. They are parallel keys into different identity and regulatory systems.
44. Worked example: mixture with several CAS RNs and one transport entry
A fictional cleaning product contains three hazardous ingredients, each with its own CAS RN, while the shipped product is assigned one UN entry based on the transport classification. A careless translation might copy the UN number into each ingredient row.
The correct target preserves the ingredient table separately from the transport section. Substance codes stay with ingredients; the UN number stays with the dangerous-goods shipping description.
45. Error clinic: ten failures that can survive fluent translation
One: replacing CAS hyphens with typographic dashes. Two: relabeling an ECHA list number as an EC number. Three: treating UN number as a chemical registry number. Four: pairing a correct CAS RN with the wrong row. Five: dropping a hydrate qualifier. Six: translating an official shipping name freely. Seven: changing packing-group Roman numerals. Eight: exposing confidential identifiers. Nine: importing code columns as dates. Ten: assuming one CAS RN always maps to one UN entry.
These mistakes show why chemical localization must protect structure as well as language. The target can be grammatically excellent and chemically misleading if identifier relationships drift.
46. A practical release checklist
CAS: exact number preserved and matched to substance. EC: correct regulatory identifier preserved. List number: not mislabeled as official EC number. UN: preserved and paired with correct proper shipping name. Names: approved chemical and regulatory terminology used. Form: salt, hydrate, isomer and physical-state qualifiers retained.
Tables: no row shifts. Amounts: concentration and units preserved. Confidentiality: source disclosure level maintained. Transport fields: class and packing group checked. QA: source-target code equality plus relationship verification completed.
47. Frequently asked questions
Should I translate a CAS RN? No. Translate the chemical name if needed, not the identifier. Is an EC number the same as a CAS number? No. They belong to different systems. Is an ECHA list number always an EC number? No; ECHA explicitly distinguishes technical list numbers from official EC numbers.
Is a UN number a substance identifier? It identifies a dangerous-goods transport entry, which can involve substances or articles under defined transport conditions. What is the best final test? Confirm that every translated name remains paired with the same substance or transport record as the source.
48. Connect this specialist guide to the wider eduKateSG translation architecture
This article owns the narrow intent of translating CAS, EC and UN identifiers without corrupting chemical or dangerous-goods identity. It does not replace the existing SDS/GHS translation guide or broader chemical-manufacturing material. It supplies the identifier layer those wider pages can link to when exact codes matter.
The deeper rule is transferable: names communicate meaning, while identifiers stabilize identity across names, languages and systems. Translate the name carefully, preserve the code exactly, and verify that they still point to the same thing.
Authoritative reference points
For substance identity, consult CAS REGISTRY. For EC and ECHA list numbers, use ECHA’s About EC and list numbers guidance and substance information. For dangerous-goods transport entries, consult the UNECE UN Recommendations on the Transport of Dangerous Goods Model Regulations or the applicable modal regulation. These systems serve different purposes and should remain distinct in translation.
49. Revision control matters because chemical identities and classifications are maintained over time
Chemical databases, supplier dossiers and dangerous-goods regulations are maintained through revisions. A translation should record which source edition or revision it follows. Do not silently replace historical identifiers or shipping information with newer data unless the project explicitly requests an updated edition.
Keep source date, revision number and target revision linked. When an authority changes a classification or a supplier updates an SDS, reviewers should be able to tell whether the target reflects the old source faithfully or has been deliberately brought forward.
50. Supplier updates require field-by-field comparison, not blind reuse
A new SDS version may retain the same product name while changing composition, concentration ranges, identifiers or transport classification. Translation memory can encourage reuse because much of the wording looks identical. Identity fields must be compared independently before old target text is propagated.
Build a change report that highlights CAS, EC, UN, concentration, hazard-class and revision fields. A single changed row can be more important than hundreds of unchanged sentences.
51. Emergency and transport documents need identifier visibility under pressure
Shipping papers and emergency information can be read in stressful conditions. Codes should remain legible, clearly labeled and separated from translated prose. Do not bury the UN number inside a long sentence or allow line wrapping to detach it from the proper shipping name.
Good multilingual layout supports fast verification: identifier first, official name clearly paired, then class and other regulated fields. Translation should reduce ambiguity without altering the structured record.
52. Multilingual labels need one source of truth for identity fields
A package may carry several language panels. Do not let each language team retype CAS, EC or UN values independently. Populate protected identifiers from one controlled data source and localize only the surrounding labels and instructions.
This architecture prevents cross-language drift. If the product record changes, one identifier update can flow to every language while terminology remains independently reviewed.
53. Change logs should distinguish linguistic edits from identity edits
When a reviewer changes a translated chemical name, that is a linguistic edit. When a CAS RN or UN number changes, that is a data or regulatory edit and deserves separate authorization. Use change tracking that makes the difference visible.
This prevents an editor from “fixing” a code as though it were punctuation. Protected-field changes should trigger a verification step and record who approved the new value.
54. Final audit: confirm identity, regulation and document purpose together
The final audit should not ask only whether every code survived. Confirm that each CAS RN still identifies the named substance, each EC or list number is labeled with the correct status, and each UN number remains paired with the correct transport entry for the document’s stated context.
Then review concentration, physical form, revision date and regulatory language. When exact token fidelity and correct regulatory relationships both survive, the target can be trusted as a translation of the same chemical record rather than a merely similar-looking document.