This site has now written some version of the same sentence more than a dozen times. D- and L-amino acids are isobaric, so mass spectrometry is blind to chirality, and a standard certificate of analysis has no line that answers the question. It has appeared in the SS-31 deep-dive, the FOXO4-DRI deep-dive, the Dermorphin deep-dive, and in every methods piece that touched sequence confirmation.
What none of those posts did was say what the missing test actually is. That is this post. Chiral purity is a real, established, routinely performed analysis with its own reagents, its own failure modes, and its own literature. It is simply a different experiment from the one that produces a purity percentage.
Why the mass spectrum cannot see it
A D-residue and its L-counterpart have identical elemental composition, identical molecular formula, and identical exact mass. Okyem and Sweedler, reviewing the field in Mass Spectrometry Reviews (2026;45(2):218-230), group these with isoaspartate under a useful label: zero-dalton modifications. Mass spectrometry's whole advantage in peptide characterization is that most modifications carry a characteristic mass shift. These carry none.
One important nuance is usually lost. A peptide with a single epimerized residue is not the enantiomer of the parent — it is a diastereomer, because the rest of the chain retains its original configuration. Diastereomers do have different physical properties, which means an ordinary achiral reversed-phase column can in principle resolve an epimer as a separate peak. Sometimes it does. But the separation is not designed for, the retention difference is often small, and co-elution is common. An epimer that co-elutes is counted inside the main peak and reported as product.
So the correct statement is narrower than "chromatography cannot see it." A routine purity method is not a chirality method, and when it happens to separate an epimer, that is luck rather than design.
The standard route: hydrolyse, label, separate
The classical approach works on the free amino acids rather than the intact peptide. The chain is hydrolysed down to its constituent residues, and each residue's configuration is then determined.
Free amino acids released this way are enantiomers of one another, so an achiral column is genuinely useless. Two routes get around that. The first converts them into diastereomers with a chiral labelling reagent — the standard being Marfey's reagent, 1-fluoro-2,4-dinitrophenyl-5-L-alanine amide. Bhushan and Brückner's review in Journal of Chromatography B (2011;879(29):3148-3161) covers the reagent and the family of structural variants built around it, including the approaches known as the advanced Marfey's method. After derivatisation the D- and L-forms are diastereomers and separate on ordinary HPLC.
The second route separates the enantiomers directly on a chiral stationary phase, skipping derivatisation. A method from Lilly Research Laboratories published by Strege and colleagues in Journal of Chromatography B (2023;1219:123638) does exactly this, coupling chiral chromatography to tandem mass spectrometry. The authors describe validation feasibility for all nineteen chiral proteogenic amino acids — nineteen rather than twenty, because glycine has no stereocentre — and demonstrate recovery of D-substitutions at every residue of a model octapeptide across the 0.1 to 1.0 percent range, applied to four model peptides of eight to fourteen residues.
That range is worth noticing. Chiral purity is reported as a percentage of D-form at a named residue, not as a pass/fail on the molecule.
The test partly destroys what it measures
Here is the complication that makes this analysis harder than it sounds. Acid hydrolysis itself racemizes amino acids. The conditions that break the peptide bonds also scramble some fraction of the stereocentres, so a naive measurement reports pre-existing D-content plus D-content that the sample preparation just created.
The scale is not trivial. Amelung and Brodowski, in Analytical Chemistry (2002;74(13):3239-3246), quantified this for environmental samples under standard hydrolysis conditions and found that between zero percent for D-glutamic acid and roughly 85 percent for D-alloisoleucine of the detected D-amino acids had been generated during processing, varying by a factor of two to ten between sample types. That work is on soil and litter rather than pharmaceutical peptides, so the specific numbers do not transfer. The mechanism does.
The fix is elegant. Hydrolyse in deuterated acid. Any residue that racemizes during hydrolysis picks up deuterium at its alpha-carbon and is therefore mass-labelled, while a D-residue that was already there is not. Miyamoto and colleagues demonstrated this in Chemistry & Biodiversity (2010;7(6):1644-1650) using simple model tripeptides, and the Lilly method above builds the deuterated-acid correction directly into its sample preparation.
The same paper reports the mirror-image problem, which matters more for this shelf than the usual framing suggests: hydrolysis also converted D-phenylalanine to L. Racemization runs both directions. For a compound whose D-residue is the point of the molecule, careless sample preparation can erase the very feature the test was meant to confirm.
Methods that never hydrolyse anything
A second family of techniques works on the intact peptide, where the epimer is a diastereomer and conformational differences are available to exploit.
Ion mobility separates ions by shape rather than mass, so peptides identical in m/z can differ in drift time and collision cross-section. Abdulbagi and colleagues survey the approach in Critical Reviews in Analytical Chemistry (2025;55(2):306-315). The honest limitation is resolution: Du and colleagues, in Analytica Chimica Acta (2025;1355:344000), note that most current instruments have limited resolving power for this and develop a workaround forming non-covalent complexes with cucurbituril to amplify the structural difference, demonstrated on methionine-enkephalin epimers.
The freshest entry is more directly relevant to synthesis. Tyson and colleagues, in the Journal of the American Society for Mass Spectrometry (2026;37(5):1258-1267), open by calling epimerization control of histidine during solid-phase peptide synthesis a persistent challenge in the pharmaceutical industry. Their method cationizes the peptide with silver and reads stereochemistry-dependent fragmentation, detecting epimeric content as low as five percent in intact peptides across tripeptide and heptapeptide sets.
One older technique deserves a mention for the neatness of its logic. The D-amino-acid discovery workflow published by Livnat and colleagues in Analytical Chemistry (2016;88(23):11868-11876) uses resistance to aminopeptidase digestion as its first-pass screen, before any hydrolysis or labelling. A D-residue near the N-terminus blocks the enzyme. That is precisely the protease-resistance mechanism described in this site's peptidase and clearance explainer — the design goal, run backwards and used as an assay.
Which compounds on this shelf the question applies to
The catalog splits into three groups, and they need different things.
Group one: chirality is the identity claim. SS-31 carries D-arginine at position 1. FOXO4-DRI is all-D by construction. Dermorphin is defined by a naturally occurring D-alanine at position 2. Adipotide carries a D-amino-acid payload. Ipamorelin, Hexarelin, GHRP-6, Melanotan 2 and PT-141 all contain D-residues by design. For every one of these, a matching intact mass confirms composition while leaving the defining structural feature unaddressed.
Group two: all-L targets with epimerization-prone positions. Histidine is the residue the 2026 work above singles out, and it appears in ipamorelin, hexarelin, GHRP-6 and Gonadorelin. Aspartate is the other classic risk, through the aspartimide pathway covered in the synthesis-routes piece, which yields racemization as well as beta-peptide — relevant to Epithalon and BPC-157.
Group three: residues with no stereocentre at all. The alpha-aminoisobutyric acid in semaglutide and tirzepatide carries two identical methyl groups on its alpha-carbon, so it is achiral and cannot epimerize. Its protease resistance is geometric, not stereochemical.
A caveat cuts across all three. The Lilly method covers the nineteen chiral proteogenic amino acids. The non-proteinogenic residues that define several of these molecules — 2-naphthylalanine, dimethyltyrosine, the D-2-methyl-tryptophan in hexarelin — are not standard chiral analytes, exactly as they are not standard analytes for amino acid analysis.
What this means for a document
Three transfers.
Chiral purity is a separate test with its own named method, and its absence means it was not determined. It is not implied by a purity percentage, not implied by a mass match, and not implied by peptide mapping.
Chirality is a stability question, not only a synthesis question. The Lilly authors note that D-isomers can enter as impurities in the amino acid starting materials, form during synthesis, and in some cases form during product shelf life. That places chirality alongside deamidation and oxidation in the degradation-pathways piece, except that unlike those it produces no mass change to detect.
Regulatory expectations exist, and they do not attach to research compounds. FDA's 1992 policy statement on the development of new stereoisomeric drugs asks that stereoisomeric composition be known and quantified, and Badgujar and colleagues survey the requirements and methods in Chirality (2024;36(3), doi 10.1002/chir.23652) — while also observing that comparatively few reports on peptide enantiomeric purity exist at all. That framework describes a drug development pathway. It confers nothing on an unapproved research compound, and it is useful here only as a description of what a complete answer would look like.
Questions researchers ask
If chirality is this important, why is it not on a standard COA? Because the routine three-line package answers identity, purity and mass, and each of those is a different question. Chiral analysis requires a separate sample preparation, a separate column or reagent, and separate standards. It is omitted for the same reason endotoxin or copper content is omitted: nobody ran that instrument.
Does an unusual peak on the chromatogram indicate an epimer? It might. An epimer is a diastereomer and can resolve on an achiral column. But so can a deletion sequence, an oxidised variant, a des-acetyl species and a scavenger adduct, most of which also differ in mass. The peak is a prompt to ask what the mass is, not an answer.
Is any of this measurable after the fact? Yes, by a laboratory that runs the method, on a sample that no longer exists afterwards for the hydrolysis-based routes. The intact-peptide ion mobility and energy-resolved approaches are less destructive, but they are characterization techniques rather than release tests, and they are still being developed — the 2025 and 2026 papers cited above are all method-development work, not settled routine.
For related methods coverage see /quality/, and for compound profiles see /library/.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.