Levodopa Reference Standard — CAS 59-92-7, (2S), free amino acid, 1000 mg

1. What this material is

This page describes levodopa supplied as an analytical reference material: a weighed quantity of a single identified chemical substance, intended to serve as the point of comparison against which another sample is measured. A reference material has one narrow job. When a laboratory reports that a batch conforms to specification, that a plant extract contains a given percentage of an analyte, or that an active-ingredient consignment is what its label says it is, that report is only as good as the material the instrument was calibrated against. Everything downstream inherits the identity, the purity and the stability of one vial.

Levodopa is an unusually demanding molecule to hold in that role, and the reason is that it has two independent failure modes that the routine identity test cannot detect. The first is stereochemical. The molecule has one stereogenic atom, its enantiomer is a real and separately registered substance, and mass spectrometry returns the same answer for both. The second is chemical. The molecule carries a catechol — a benzene ring bearing two adjacent hydroxyl groups — and catechols oxidise. The registry record says so in capital letters, quoting a reference work from 1976 [1]. A vial that has taken up moisture and air is progressively no longer the substance on the label, and again the mass spectrum of what remains will keep looking correct for as long as any parent molecule survives in it.

Sections 5 to 8 set both problems out in detail, because they are what distinguishes this substance from an ordinary reagent purchase. They are also why this page runs long. For the general basis on which this shop supplies reference materials, the surrounding reference standards category collects the rest of the catalogue.

One further thing belongs at the outset rather than buried in a later section. Levodopa is the active substance of medicines available on prescription in the European Union and in the United States. That status attaches to those finished products and to the companies holding their marketing authorisations. It does not attach to the material on this page, which is supplied as a laboratory reagent under the terms in section 15. An authorised medicine and an analytical reference material are two different legal articles even when the molecule is identical, and the difference determines who may supply the article, to whom, with what documentation, and for what purpose. Section 12 sets out the regulatory position jurisdiction by jurisdiction, as a measurement against named registers rather than as a summary from memory.

2. Identity and registry codes

Every identifier below is quoted from a public registry, with the registry named. Where a registry carries more than one value for the same field, both are given rather than a choice being made silently.

The classification code is not one code, and the register does not use the one you expect

The classification code deserves a paragraph of its own, because it is a reliable source of confusion. The World Health Organization system assigns N04BA01 to levodopa as a single substance, N04BA02 to the combination of levodopa with a decarboxylase inhibitor, and N04BA03 to the combination of levodopa with a decarboxylase inhibitor and a catechol-O-methyltransferase inhibitor [13]. Three codes, one molecule, and what separates them is not the substance but what is packaged alongside it.

Measured against the Polish national register, the distribution across the 49 records matching this substance is N04BA02 in 25 records, N04BA03 in 22, N04BA01 in exactly one, and the bare stem N04BA in one further record [10]. The single-substance code is therefore the rarest of the four in that register, which is the opposite of what a reader would predict from a page about the substance itself. Section 12 identifies which record carries it, and why the bare-stem record is a different chemical entity altogether.

Machine-readable descriptors

Two rows in that table deserve a second look, and they point in opposite directions.

The two SMILES strings are not interchangeable. The first carries no stereochemical information whatsoever; the second encodes the single centre through the [C@@H] token. If a supplier record, a database import or a laboratory information system stores the first, the stereochemistry is simply gone and nothing downstream will flag its absence. This is not a hypothetical failure mode: a stereochemistry-free record for exactly this connectivity exists in the registry in its own right [3], and any process that rounds a structure down to its connectivity will land on it.

A computed partition coefficient of −2.7 and a polar surface area of 104 Å² describe a strongly hydrophilic molecule, and they are worth holding in mind when reading any solubility statement about this compound. A substance with a partition coefficient that negative, four hydrogen-bond donors and a zwitterionic ground state does not dissolve in chloroform. The registry's own experimental entry agrees: practically insol in ethanol, benzene, chloroform, ethyl acetate [1]. Descriptions to the contrary circulate widely in commercial listings for this compound; they are not registry values, and the descriptor table above is by itself enough to reject them without needing a second opinion.

3. Five ways to buy the wrong 197

Before any question of purity arises there is a question of identity, and for this molecule the identity question has more traps in it than usual. Each of the five below has been checked directly against the registry, and each is a distinct substance with its own record, its own InChIKey and in most cases its own mass.

Trap 1: the enantiomer and the racemate, which weigh exactly the same

This is the trap that governs the rest of the page, and section 5 is devoted to it. In summary: three registry records carry this connectivity, and the numeric fields that a laboratory ordinarily uses to confirm identity are identical across all three.

All three: formula C₉H₁₁NO₄, molecular mass 197.19 g·mol⁻¹, monoisotopic mass 197.06880783 Da, formal charge 0, one covalently bonded unit. The only field in which they differ is the second block of the InChIKey and, for the flat record, the stereocentre counter.

Trap 2: the hydrochloride, which is a real salt and weighs 18.5% more

Levodopa as supplied here is a free amino acid: one covalently bonded unit, formal charge zero, no counter-ion and no water of crystallisation [1]. In the solid and in water it exists as a zwitterion, which is a charge arrangement within the same molecule rather than a salt with something else. The registry describes the record as a tautomer of the levodopa zwitterion and as the conjugate acid of the levodopa anion; neither statement introduces a second component.

A hydrochloride nonetheless exists, and the numbers are worth setting out because the arithmetic is unforgiving:

The ratio is 233.65 / 197.19 = 1.1849. One thousand milligrams of the hydrochloride contains 844.0 mg of the free amino acid; one thousand milligrams of the free amino acid corresponds to 1184.9 mg of the salt. Weigh out the salt and calculate as though it were the base and every result in the run is 18.5 per cent too high — reproducibly, in a sequence that passes every system-suitability check. This is the same class of error described at length on the sildenafil card, where a citrate salt moves the arithmetic by 40.5 per cent, and it is the reason a form declaration belongs on lot documentation rather than in catalogue prose. For a substance supplied deliberately as a hydrochloride, compare procaine hydrochloride, whose registry record likewise shows two covalently bonded units.

Trap 3: the methyl ester, which the registry itself relates back to this record

Melevodopa is the methyl ester of levodopa, CID 23497, formula C₁₀H₁₃NO₄, molecular mass 211.21 g·mol⁻¹, monoisotopic mass 211.08445790 Da, one covalently bonded unit and one defined stereocentre [6]. The registry record for levodopa itself describes the compound as the active moiety of melevodopa, which means a search for one will reasonably surface the other. Fourteen mass units separate them, so this trap at least is one that a mass spectrometer can resolve — but only if someone thinks to look. There is also a hydrochloride of the ester, at 247.67 g·mol⁻¹, and a sodium salt of the parent acid at 220.18 g·mol⁻¹, both separately registered.

Trap 4: the phosphate prodrug, which shares the stereo block of the InChIKey

Foslevodopa is a phosphorylated derivative, CID 127766, formula C₉H₁₂NO₇P, molecular mass 277.17 g·mol⁻¹, monoisotopic mass 277.03513872 Da [7]. The mass difference from the parent is large and easy to see. What is not easy to see is the key: its InChIKey is YNDMEEULGSTYJT-LURJTMIESA-N. The first block differs, as it must, because the connectivity differs — but the second block, LURJTMIESA, is character-for-character the same as levodopa's, because the stereochemical description of the single centre is the same. Any tooling that matches on the stereo block, or that displays a truncated key, will treat these two substances as related more closely than they are. The mass ratio 277.17 / 197.19 = 1.4056 means that treating one as the other in a gravimetric calculation is a 40.6 per cent error.

Trap 5: the hydrate that belongs to the other component

Levodopa itself has no registered hydrate. That is a measured statement rather than an assumption: the name levodopa monohydrate returns HTTP 404 from the registry name endpoint, on the same instrument on which levodopa hydrochloride returns a valid CID. But the compounds levodopa is habitually paired with do have hydrates, and the arithmetic there is live. Carbidopa monohydrate is CID 38101, formula C₁₀H₁₆N₂O₅, molecular mass 244.24 g·mol⁻¹ against 226.23 for the anhydrous form — two covalently bonded units, the water counted as the second [8]. Section 12 shows the Polish register recording that distinction inconsistently across its own entries, sometimes naming the hydrate and sometimes not, for what is otherwise the same pairing.

4. Structure and stereochemistry

Levodopa is an aromatic amino acid: a propanoic acid backbone carrying an amino group at C-2 and a 3,4-dihydroxyphenyl group at C-3. The dihydroxyphenyl group is a catechol, and it is responsible for everything in section 7. The C-2 carbon is the single stereogenic atom, and it is responsible for everything in sections 5 and 6.

The registry record states the stereochemistry without ambiguity. The four counters below are quoted verbatim from the record rather than paraphrased, because they are the fields on which the whole configuration argument rests:

One defined centre, none undefined, and no stereogenic double bonds. The substance is a single, fully specified stereoisomer in the (2S) configuration, which in the older amino-acid convention is written L. The last two rows dispose of any salt or hydrate reading of the formula: a single covalently bonded unit at formal charge zero is a free amino acid by definition, and section 3 sets out what the salts look like when they are present.

Three independent statements of the same fact, and why that matters

Configuration is one of the fields most often lost in transcription, so it is worth noting that the record asserts it three times over, in three different notations, and that the three agree:

• the counter: defined atom stereocentre count 1, undefined 0;
• the InChI stereo layer, /t6-/m0/s1, which encodes one fully specified centre and no ambiguity;
• the InChIKey second block, LURJTMIESA, which is a hash of that stereo layer.

Contrast the racemic record [3], where the same three fields read: undefined atom stereocentre count 1, no /t layer in the InChI at all, and the second block UHFFFAOYSA — the string that the algorithm emits when there is no stereochemistry to hash. That is a negative control on the same instrument: the fields do change when the configuration is absent, so their content in the levodopa record is informative rather than boilerplate. And the mirror-image record [2] reads: defined atom stereocentre count 1, undefined 0, InChI layer /t6-/m1/s1, second block ZCFIWIBFSA. One character in the InChI layer, m0 against m1, is the entire difference between the substance on this page and a different one.

The optical rotation, and what it is worth

The record carries one chiroptical measurement: [α] −13.1° at 13 °C on the sodium D line, at a concentration of 5.12 in 1 N hydrochloric acid, attributed to the Merck Index [1]. The negative sign is the origin of the older name (−)-dopa and of the levo- prefix in the international name. The same entry carries the ultraviolet maxima measured in 0.001 N hydrochloric acid: 220.5 nm with log ε 3.79, and 280 nm with log ε 3.42.

Two observations about that rotation figure, one encouraging and one not.

The encouraging one: a polarimeter is ordinary equipment, and unlike the instruments in section 6 it is likely to be in the building already. A rotation measurement on this substance is one of the few configuration checks that a routine laboratory can actually perform. Compare tadalafil, where the public record carries no specific rotation at all despite the molecule having two stereocentres — a laboratory measuring a rotation there has a number with nothing to compare it against.

The discouraging one: the measurement conditions are specific and old, and the value is not additive in the way a novice might assume. The temperature is stated as 13 °C, which is not a temperature anyone reproduces casually; the solvent is 1 N hydrochloric acid, not water, because the compound is only sparingly soluble in water and because protonation state affects the rotation; and the concentration is quoted as 5.12 in the older convention. Reproducing that value requires reproducing those conditions. More importantly, a rotation measurement detects an enantiomeric imbalance, and at low levels of contamination the sensitivity is poor: a 1 per cent excess of the wrong enantiomer shifts the observed rotation by 2 per cent of its magnitude, which for a magnitude of 13.1° is around a quarter of a degree. Polarimetry answers the question is this grossly racemic reliably. It does not answer the question is the (R)-enantiomer below the pharmacopoeial limit, and section 6 lists the methods that do.

Where levodopa sits among chiral reference materials

Stereochemical burden varies enormously across a catalogue, and the contrast is instructive because it shows that the amount of work a certificate has to do is not the same for every product. Forskolin sits at one extreme with eight defined atom stereocentres and none undefined — a molecule whose stereochemical description is a paragraph rather than a letter. CDP-choline carries four defined centres. Noopept carries one defined centre, exactly as levodopa does.

At the other extreme sit records with a defined count of zero and an undefined count of one, which is the registry's way of saying this record is the racemate or the unresolved material: modafinil and oxiracetam both read 0 defined and 1 undefined, the same pattern as DL-DOPA. And then there are the molecules with nothing to get wrong at all: aniracetam, apigenin and paracetamol all read zero defined and zero undefined. For those, a certificate that omits configuration has omitted nothing, and an enantiomeric purity figure would be a fabricated number rather than a conservative one.

Levodopa is not the hardest case in that group by arithmetic — one centre gives two isomers, not eight. It is arguably the hardest case in practice, for three reasons that compound one another: its enantiomer is a named pharmacopoeial impurity rather than an academic curiosity [22]; the enantiomer is also a commercially available substance with its own registry record and its own suppliers [2]; and the molecule is simultaneously unstable in a way that consumes it into non-chiral degradation products (section 7). Two centres on a stable, unavailable isomer is a smaller problem than one centre on an unstable, purchasable one.

5. The blind spot: one atom, three records, one mass

This is the central analytical fact about levodopa as a reference material, and it deserves stating in the bluntest available terms.

Nor does fragmentation help. Published product-ion spectra are a function of connectivity, and connectivity is precisely what these three substances share. The deposited transitions from the registry record, recorded on an LTQ Orbitrap XL and, for one negative-mode deposit, on an Agilent 6530 Q-TOF, both with electrospray ionisation, are:

Every stereoisomer produces this series. The loss of ammonia (17.027) from the protonated molecule to give 181.050, the loss of formic acid (46.005) to give the 152.071 ion, and the combined loss of both (63.032) to give the 135.044 ion, are reactions of the amino-acid side chain and the catechol ring; neither reaction knows anything about the arrangement of substituents at C-2. A laboratory reporting levodopa confirmed by LC-MS/MS, retention time and two transitions within tolerance has confirmed connectivity and nothing more.

The consequence is worth spelling out because it is counter-intuitive. If the reference standard used to fix that retention time and those transitions was itself contaminated with the wrong enantiomer, the tolerance window closes around the wrong answer, and the resulting report is internally consistent, fully documented, signed, and wrong. Nothing in the chromatogram flags it. The system-suitability criteria all pass, because system suitability tests the chromatography, not the identity of the material that defined it.

What else fails to see the difference

Mass spectrometry is the technique most often named, but it is not alone. Every one of the following returns effectively the same answer for L-DOPA and D-DOPA:

That last row is the one that catches laboratories out most often, because it looks as though it should not be true. A validated, fully qualified, stability-indicating reversed-phase method with excellent resolution of every process impurity will still show one peak for a fifty-fifty mixture of the two enantiomers, at exactly the expected retention time, with a peak-purity check that passes. The chromatography is not failing. It is answering a different question from the one being asked of it.

The problem is fifty-three years old, and it is also current

Two references bracket this. In 1972 a method was published for the quantitative determination of D-DOPA present in levodopa samples [21] — that is, the contamination question was recognised and addressed within a few years of the substance entering large-scale production. In 2025 a method was published for the detection of D-DOPA in levodopa tablets by derivatisation coupled with LC-MS/MS, whose authors state plainly that detection of the chiral impurity is difficult due to the similarity of polarity [23].

Fifty-three years separate those two papers and the problem in them is the same problem. That is not a sign of a neglected field; it is a sign of a difficulty that is intrinsic rather than technological. Enantiomers are not hard to separate because instruments are insufficiently sensitive. They are hard to separate because in an achiral environment they genuinely have the same properties, and the only route to a difference is to build an asymmetric environment on purpose.

6. Methods that resolve configuration, and methods that cannot

Four families of approach are documented as capable of distinguishing levodopa from D-DOPA. Each has a published method, and each carries the same prerequisite: a reference material of confirmed configuration. None of them is self-calibrating.

Capillary electrophoresis with a chiral selector

The method that most directly addresses the pharmacopoeial question is a capillary electrophoresis assay developed and validated for the simultaneous determination of the impurities of levodopa listed in the European Pharmacopoeia, including the (R)-enantiomer [22]. The reported conditions are specific enough to reproduce: a fused-silica capillary, sulfated β-cyclodextrin at 6 mg·mL⁻¹ as the chiral selector, 0.1 M phosphate buffer at pH 2.0, 20 kV, 18 °C, with L-phenylalanine as internal standard. The method was validated over the range 0.1 to 1.0 per cent at a levodopa concentration of 2 mg·mL⁻¹, and it was applied to several samples of levodopa including the chemical reference substance of the European Pharmacopoeia.

Two details from that paper deserve emphasis for anyone planning to use the method. First, the authors examined the influence of different batches of sulfated β-cyclodextrin, which is an acknowledgement that the selector is not a defined single substance and that migration behaviour is selector-lot dependent. Second, the critical pair reported is levodopa against L-tyrosine — that is, the hardest separation in the system is not the enantiomer pair but a structurally related impurity. A method tuned only for the enantiomer may lose the tyrosine resolution, and the reverse is also true.

Chiral derivatisation followed by ordinary chromatography

The most recent published approach inverts the problem. Rather than building a chiral environment in the separation, it converts the enantiomers into diastereomers, which differ in ordinary physical properties and can then be separated on a conventional column. A 2025 method derivatises with a chlorinated chiral labelling reagent and runs LC-MS/MS on a standard C18 column, reporting a resolution of R_s = 4.15 between D-DOPA and L-DOPA, a derivatisation yield of 97 per cent, a mass-spectrometric response for the labelled D form 2.78 times that of the L form, and a limit of detection of 7.88 µg·L⁻¹ [23]. Three further impurities are determined in the same run.

This is the general escape route from section 5, and it is worth naming as a principle: when a technique cannot separate two things, make them differ in a dimension the technique can see. It is sound analytical engineering. It is also a standing admission that the original distinction was invisible — and it introduces its own dependency, because a derivatisation reagent that is itself not enantiomerically pure will manufacture the very impurity it is being used to measure.

Polarimetry

Discussed in section 4. Cheap, available, and adequate only for gross questions. The reference value against which any measurement must be compared is [α] −13.1° under the conditions stated there [1], and those conditions have to be reproduced for the comparison to mean anything.

Chromatography on a chiral stationary phase

The generic route, and the one a laboratory is most likely to attempt first. It works, and the published review literature on chromatographic determination of levodopa maps the options — HPLC with ultraviolet detection, HPLC with electrochemical detection and LC-MS/MS — across the range of sample types [31]. What the review makes clear is that method choice for this compound is governed as much by the matrix and by the analyte's instability as by the separation itself, which is a theme that returns in section 7.

The method that a purity figure alone does not tell you

A single number described as purity is ambiguous for this substance in a way it is not for an achiral one, and the ambiguity is not pedantic.

These two numbers are independent. A material can be 99.9 per cent chemically pure and simultaneously carry 1 per cent of the wrong enantiomer, because the chromatographic method that produced the first figure counted the second substance as levodopa. It co-eluted. It has the same retention time by physical necessity, not by coincidence. Any purity claim for this compound that does not say which of the two quantities it refers to is, strictly speaking, uninterpretable — and figures of the form 99.5 per cent or 99.52 per cent circulate widely in commercial listings for this substance without that distinction being drawn. They are supplier statements, not registry values, and the arithmetic above shows why the missing word matters more here than on an achiral product.

7. The catechol problem: when the standard changes colour

Everything above concerns which molecule is in the vial. This section concerns whether it is still there.

The mechanism, and the number attached to it

Catechols oxidise to ortho-quinones. For this molecule the product is dopaquinone, and the reaction does not stop there: the quinone is electrophilic, the molecule carries a nucleophilic amine on the same carbon skeleton, and the resulting intramolecular chemistry runs on into coloured polymeric material. That is why the visible symptom is darkening rather than simple discoloration.

A 2024 computational study put a number on the rate-limiting step. Using an empirical valence bond simulation, the authors report a free-energy barrier in water of 30.93 ± 1.12 kcal·mol⁻¹ against an experimental value of 27.55 kcal·mol⁻¹, and identify the reactive step as an intramolecular Michael addition concerted with proton transfer from the amino group [27]. They further note that the dopaquinone product enters a redox cycle producing hydrogen peroxide continuously. The agreement between calculation and experiment matters less than the fact that both exist: the darkening of this compound is a characterised chemical reaction with a measured activation barrier, not a vague quality complaint.

The redox chemistry of the catechol group has been examined directly in the context of tyrosine oxidation, where levodopa acts as the reducing partner [26]. That work explains why this particular fragment of the molecule is the reactive one, and by extension why storage under an inert atmosphere is a requirement rather than an abundance of caution. For a contrast in the same catalogue, methylene blue is a substance whose entire analytical usefulness rests on a reversible redox couple that is deliberately exploited; levodopa is a substance whose redox chemistry is entirely unwanted and runs in one direction only.

How much material is at stake: a measured figure

The most useful published number on the scale of the loss comes from work on a botanical raw material. Analysing 56 accessions of four varieties of a plant source by high-performance thin-layer chromatography with densitometry, the authors report levodopa contents of 0.58 to 6.42 per cent of dry weight, and — the figure that matters here — an average of 52.11 per cent degradation of levodopa on repeated extraction over 24 hours [25]. The degradation product appears as a second chromatographic band at R_f 0.41 ± 0.02 against 0.34 ± 0.02 for levodopa itself, and the authors describe the products in plain terms as damaging quinones and reactive oxygen species.

Light is a second, separately characterised pathway

Oxidation in the dark and photodegradation are not the same process, and both are documented. A study of levodopa microparticles reports photodegradation kinetics together with a developed and validated stability-indicating liquid chromatographic method [24]. Two things follow. First, the instruction to protect the material from light is backed by a kinetic study rather than by convention. Second — and this is the more useful point for a laboratory — a stability-indicating method exists and has been validated, meaning a method in which the degradation products are chromatographically resolved from the parent rather than hidden beneath it. An assay that is not stability-indicating will report a darkening sample as pure for as long as the degradation products happen to co-elute.

What this means for a vial on a shelf

The practical response is not exotic: an inert headspace, a tightly closed container, amber glass or foil, a dry storage location, dispensing in one operation rather than many, and preparation of solutions immediately before use. Section 13 sets this out as a table. The reason for spending a section on it is that none of those precautions look necessary if the only characterisation on file is a mass spectrum — and the mass spectrum will continue to look correct for exactly as long as the material continues to degrade.

8. The compendial method has a co-elution

Sections 5 to 7 concern the standard itself. This section concerns what happens when the standard is used, and it is the strongest single argument in the published literature for holding a set of reference materials rather than one.

A 2026 study set out to build a robust impurity profile for levodopa and carbidopa tablets using formal experimental design. In the course of verifying the compendial procedure, the authors report finding that a process-related impurity — (1R,3S)-6,7-dihydroxy-1-methyl-1,2,3,4-tetrahydroisoquinoline-1,3-dicarboxylic acid, designated IP2 in the paper — co-elutes with levodopa, which the authors state compromises selective quantification and peak-purity assessment. They further report that small univariate adjustments improved the levodopa and IP2 separation but created new critical overlaps among carbidopa, catechol and levodopa related compound C [32].

The same conclusion arrives from a second direction. A rapid liquid chromatography and tandem mass spectrometry method has been published for the simultaneous determination of levodopa, carbidopa, entacapone and six related compounds in film-coated tablets [29]. Six named related substances, in one run, is a design that presupposes six authenticated materials to establish retention and transitions. A laboratory attempting that method with a single levodopa standard has calibrated one of seven axes.

And the reason any of this is worth the trouble is documented too. A comparison of seven generic levodopa and benserazide products against an originator product examined colour, powder appearance, disintegration, release, tablet mass, content, identity and the amount of impurities, against shelf-life specifications. The reported outcome, quoted from the abstract, is that each of the seven generic products had one or two parameters outside the specifications [28]. That is seven for seven. Whatever else it demonstrates, it demonstrates that analytical control of this substance is not a formality, and that the parameters that failed included both colour and impurity content — the two things sections 7 and 8 are respectively about.

Where the material itself comes from, and why that changes the impurity profile

One further consideration belongs here, because it determines which impurities a given batch is likely to carry at all. Levodopa is manufactured by at least two fundamentally different routes.

The first is asymmetric hydrogenation, and it is the route that made the compound historically important far beyond its own field. The industrial process developed at Monsanto used a rhodium complex of the diphosphine ligand DIPAMP to hydrogenate a prochiral enamide precursor enantioselectively, and it was the first commercialised asymmetric catalytic process of any kind [18]. The work was recognised with a share of the 2001 Nobel Prize in Chemistry, and the corresponding Nobel lecture is a compact primary account of how the process was arrived at [19]. For a page about a chiral reference standard this is more than an anecdote: it is the clearest possible illustration that the same molecular formula does not mean the same material, since the entire industrial problem being solved was how to obtain one enantiomer rather than the statistical mixture.

The second is biocatalysis, principally tyrosinase-mediated hydroxylation of tyrosine and related enzymatic routes, reviewed comprehensively in 2015 [20]. Enzymatic production delivers the single enantiomer by construction rather than by catalyst design.

The two routes carry different impurity profiles. A material from asymmetric hydrogenation may carry residual metal, ligand-derived species and the unreduced precursor; a material from biocatalysis may carry substrate, over-oxidation products of the catechol and biological residues. Neither profile is worse than the other, but they are not interchangeable, and an impurity method developed against one may be blind to the characteristic impurities of the other. A reference material whose synthetic origin is stated is more useful than one whose origin is not — which is a reasonable thing to ask a supplier and an unreasonable thing for a catalogue page to assert without documentation.

9. Physicochemical data, with attribution per value

The registry record carries a populated experimental-properties section, and it also carries internal disagreements that a careful reader should see rather than have resolved for them. Every value below is given with the source named in the record. Where sources conflict, the conflict is reported rather than averaged.

The melting point is three numbers, and the spread is the information

Read the three melting-point rows together: 295 °C, 284–286 °C and 285 °C, from three sources within one record. That is an eleven-degree spread on a quantity that for a well-behaved crystalline solid is usually reproducible to better than a degree.

The spread is not sloppiness. It is what a melting point looks like for a compound that decomposes rather than melts cleanly. When the observed transition is a decomposition, the temperature at which it is recorded depends on the heating rate, on the atmosphere, on the sample size and on where in the process the observer decides the event has occurred. Different laboratories using different protocols will legitimately report different numbers, and averaging them produces a figure that is no more true than any of its inputs.

The solubility entries agree, and the agreement is worth stating explicitly

Three of the water-solubility rows are the same number in three unit systems: 5000 mg·L⁻¹, 5.0 mg·mL⁻¹ and 5 mM all describe a saturation concentration of about five milligrams per millilitre at ambient temperature. Multiplying 5 mM by the molecular mass of 197.19 gives 0.986 mg·mL⁻¹, which does not match — so the millimolar entry is either loosely rounded or refers to a different condition, and it should not be relied on as an independent confirmation. Two of the three are firm.

Five milligrams per millilitre is modest but perfectly workable for a chromatographic standard, and it is far from the picogram-scale problem that some reference materials present. The important part of the solubility entry is the other half of it: practically insoluble in ethanol, benzene, chloroform and ethyl acetate, and readily soluble in dilute hydrochloric and formic acids. That pattern is exactly what the descriptors in section 2 predict for a zwitterionic amino acid, and it dictates how a stock solution has to be made: an acidified aqueous medium, not an organic one.

There is a hidden cost to that, and it connects back to section 7. Dilute acid is the medium in which levodopa dissolves and also, being aqueous, the medium in which the oxidation described in section 7 proceeds. The acid helps — catechol oxidation is markedly slower at low pH — but a stock solution is nonetheless a decaying system rather than a stable one, and it should be prepared fresh and used promptly rather than stored. The validated stability-indicating chromatographic method reported alongside the photodegradation kinetics [24] and a separately validated reversed-phase method with ultraviolet detection [30] are the appropriate tools for checking whether a given solution has held up.

The two partition coefficients disagree, and one of them is unusable

The record carries logP = 0.05 from one source and log K_ow = −2.39 from another [1]. Those are not close: they differ by 2.44 log units, a factor of about 275 in partition ratio. The computed XLogP3 value, −2.7, sits close to the second and far from the first.

Two of the three agree, and they agree with the chemistry. A zwitterionic amino acid bearing a catechol, with four hydrogen-bond donors and a polar surface area of 104 Å², is a strongly hydrophilic substance and should have a distinctly negative partition coefficient. A value of 0.05 would describe a molecule with roughly equal affinity for octanol and water, which is not consistent with any other property in the record. This page reports both figures because both are in the source, and states plainly which one the rest of the data supports.

One number this page will not report

10. Spectra: what exists, what does not

Deposited spectroscopic coverage for levodopa is broad in mass spectrometry and unexpectedly narrow everywhere else. The pattern of the gaps is not random, and for this particular molecule it is close to the worst possible pattern.

Four observations follow, in ascending order of how much they matter for this specific substance.

First, the carbon spectrum is a placeholder. The ¹³C subsection contains exactly two information fields: a database compilation copyright line, and a thumbnail image URL. It contains no chemical shifts. This is a measured statement with a positive control on the same instrument: the ¹H subsection of the same record contains nine fields including instrument, frequency, solvent and pH, and the HSQC subsection contains eight including a list of cross-peak coordinates. The extraction sees data when data is present. For the carbon spectrum, it is not present. Anyone needing assigned carbon shifts for this molecule has to go to a method paper, not to the registry.

Second, both available NMR spectra were recorded in water at pH 7.00 — that is, on the zwitterion. That is a legitimate and chemically sensible choice, and for a metabolomics database it is the right one. It is nonetheless a single condition. There is no spectrum in dimethyl sulfoxide, none in deuterium oxide with added acid, and therefore no reference data for the protonated form that a laboratory will actually have in the vial after dissolving the material in dilute acid per section 9. Chemical shifts of a catechol and of an amino-acid backbone both move with pH; a comparison against the deposited spectrum requires reproducing its pH.

Third, the vibrational spectra come from three different suppliers and three different samples — a Hoffmann-La Roche sample for the FTIR, a Spectrum Chemical lot for the ATR-IR, a Spectrochem sample for the Raman. This is the opposite of the more common single-lot situation, and here it is a genuine strength: three independent materials giving consistent vibrational spectra is a stronger statement about the substance than three techniques applied to one bottle. It is worth noting explicitly, because it is the one part of the spectral record that is better than it looks.

Three further absences, stated as absences

A page that names these gaps is more useful than a page that fills them with plausible-looking numbers. Where a value is absent from the public record, this card says so, and section 14 restates the boundary in the form of a table.

11. Hazard classification

The registry carries an aggregated hazard classification for levodopa, and it is important to read what that aggregation actually is before relying on it.

Signal word: Danger. The precautionary statements listed in the same record are P203, P260, P261, P264, P264+P265, P270, P271, P273, P280, P301+P317, P302+P352, P304+P340, P305+P351+P338, P318, P319, P321, P330, P332+P317, P337+P317, P362+P364, P391, P403+P233, P405 and P501.

Two further regulatory facts belong here rather than in section 12, because they are properties of the substance rather than of any product. Levodopa is a registered substance under the European chemicals regulation, with an active registration dossier last updated in March 2020 [9]. And in the Australian assessment programme it has been through a tier I evaluation for human health and the environment under the name 3-hydroxy-L-tyrosine. Neither of these confers or removes any classification; they establish that the substance is known to the chemicals regulators as a chemical, alongside being known to the medicines regulators as an active substance.

12. Regulatory status

Each statement in this section is a measurement against a named document, and each negative statement was made with a positive control on the same document, so that a zero can be distinguished from a broken search.

Anti-doping

Levodopa does not appear on the World Anti-Doping Code Prohibited List for 2026 [11]. Searching the document returns zero occurrences for levodopa, zero for L-DOPA, zero for dopa as a standalone token, zero for carbidopa, zero for benserazide and zero for entacapone.

Those zeros are meaningful because of the controls. The same search of the same document returns testosterone 32 times, salbutamol 7, modafinil 4, meldonium twice and bromantan twice — five positive controls, all firing. More usefully, there is a paired control of the same shape as the target: selegiline returns two occurrences, listed among the stimulants. Selegiline is an antiparkinsonian agent, which is to say a substance from the same therapeutic shelf as levodopa. A therapeutically neighbouring compound is on the List and levodopa is not; the instrument is therefore pointed in the right direction and is discriminating, not merely empty. The same document and the same counting method underpin the anti-doping statements elsewhere in this catalogue, so a zero here and a non-zero result on another card are comparable measurements rather than differently worded assurances.

Controlled substances

Levodopa does not appear in the Polish schedules of narcotic drugs and psychotropic substances. The document searched is the consolidated text of the Regulation of the Minister of Health of 17 August 2018 on the schedules of psychotropic substances, narcotic drugs and new psychoactive substances (Journal of Laws 2024, item 1139), together with the two amendments in force, Journal of Laws 2025 item 598 and Journal of Laws 2026 item 934. Searching them returns zero occurrences for the Polish spelling lewodopa, zero for levodopa, zero for L-DOPA, zero for karbidopa and zero for the standalone stem dopa. Positive controls in the same document, on the same pass: the morphine stem returns 58 occurrences, fentanyl 38, the amphetamine stem 27 and the cocaine stem twice. The instrument fires on four separate substance classes in the document being searched.

There is no United States controlled-substance schedule for levodopa. The registry record carries a single entry in the forensic-laboratory reporting system, classified there as Other Substances and added in November 1999. That system is an inventory of what forensic laboratories encounter; it is not a schedule, and appearing in it does not place a substance under control.

Medicinal product status in Poland

This is the section where the widely repeated summary turns out to be wrong in three separate ways, and the corrections all run in the same direction: the register is more heterogeneous than the summary suggests.

The measurement was made by retrieving the entire national register — 22 871 records, retrieved in full with the record count matching the register's own declared total — and filtering on the substance name across three fields, rather than by looking up trade names known in advance. That distinction matters: a list of names known in advance can never prove a general statement about a register, because the names it does not contain are invisible to it. Positive controls on the same filter and the same dataset: ibuprofen returns 229 records, paracetamol 293, metformin 231. Negative controls on names that should not be present in the Polish register: Sinemet returns zero, Isicom zero.

First correction: there is exactly one single-component record, and it exists. The summary most often repeated about this register is that every levodopa product in it is a combination. Filtering the active-substance field for entries containing no separator returns one record, and it is Inbrija: active substance Levodopum, common name Levodopum, strength 33 mg, pharmaceutical form powder for inhalation in hard capsules, holder Acorda Therapeutics Ireland Limited, authorised through the centralised procedure, register identifier 100424346 [10]. It is also the only record in the set carrying the single-substance classification code. The general statement is therefore false, and it is false by exactly one record — which is the most easily missed way for a general statement to be false.

Second correction: three codes, not one. The distribution given in section 2 — N04BA02 in 25 records, N04BA03 in 22, N04BA01 in one and the bare stem N04BA in one — means that the triple combinations are not merely a minority variant but very nearly half the register entry for this substance.

Third correction, and the most interesting one: the register does not name the forms consistently, and the inconsistency runs inside single trade names. Read the active-substance column of the table above, then read these counts, taken across all 49 records.

Fifteen records pair levodopa with benserazide, and they split nine to six between declaring that second component as the free base and declaring it as the hydrochloride. Eleven records pair levodopa with carbidopa outside the triple combinations, and four of the eleven name the monohydrate while the other seven do not. Component order reverses in one pattern only: five records place carbidopa first (Bascar twice, Parkador three times). Latin and vernacular spellings of the same second component both appear — Carbidopum throughout the active-substance field, Carbidopa once in the common-name field.

The sharpest version of this is not the aggregate but what happens within one trade name. Under Madopar 250 mg, four records declare Levodopum + Benserazidum and a fifth declares Levodopum + Benserazidi hydrochloridum. Under Madopar 125 mg, it is one record each way. Under Corbilta, six records list the three components in one order and the seventh lists them in the reverse order. Same product name, same holder, same register — different declarations of what the material is.

None of this makes the register unreliable about the medicines it describes; a marketing authorisation is defined by its dossier, not by the word order in a search field. It makes the register unreliable as a machine-readable source of substance identity, which is a different claim and the one that matters to anyone building a substance list. A parser that treats the active-substance field as canonical will conclude that two forms of benserazide are in use where one is, will miss a hydrate in seven cases out of eleven, and — as the next paragraph shows — will import a prodrug as its parent compound.

For completeness on the remaining register fields across the 49 records: authorisation procedure is centralised in 23 records, decentralised in 10, national in 9, mutual recognition in 3 and parallel import in 4 (procedure code IR — the four Madopar 250 mg parallel-import permits 228/22 and 289/22 held by InPharm and 134/23 and 171/23 held by Delfarma). Validity is recorded as indefinite in 12 records, blank in 23, and carries a specific future date in the remainder, the earliest of which falls in 2027. This page does not reproduce the indications, contraindications or administration details of any of those products; the authorised product information is the only proper source for that, and none of it is summarised here.

United States

Levodopa is a prescription-only substance in the United States according to the availability field of the registry record, with a maximum development phase recorded as approved and a first approval recorded in 1970 [1]. Routes of administration recorded are oral and topical. The substance appears in the national listing of approved products with therapeutic-equivalence evaluations. The first approved fixed combination containing it was authorised in May 1975, and a single-substance inhalation product exists in that market as well.

European Union and pharmacopoeial position

Beyond the national register above, several of the products in the table hold centralised or mutual-recognition authorisations, as the procedure counts indicate. On the chemicals side, the substance holds an active registration dossier [9].

13. Handling, storage and documentation

The guidance below follows from the classification in section 11 and the stability findings in section 7. It concerns handling of a laboratory reagent by trained personnel and nothing else.

14. What we certify and what we do not

This section exists because the difference between a supplier's statement and a certified value is the difference a reference material is bought for. We would rather state the boundary plainly than let a page imply more than it can support.

15. Terms of supply

This material is supplied as a laboratory reagent and analytical reference material, for in-vitro laboratory use by qualified personnel in an appropriately equipped facility. It is not a medicinal product, not a dietary supplement, not a food, feed or cosmetic ingredient, and it is not supplied for administration to humans or animals in any form or by any route. This shop is not a pharmacy and does not dispense medicines.

By ordering, the purchaser confirms that they are a professional user acquiring the material for laboratory purposes; that they will handle it in accordance with section 13 and their own institutional risk assessment, including a specific assessment of the reproductive hazard set out in section 11; that they will not administer it to humans or animals, nor supply it to any person who intends to; and that they are responsible for compliance with all laws applicable at the destination, which for a substance that is an active pharmaceutical ingredient in the destination jurisdiction may include requirements that do not apply to ordinary reagents.

Nothing on this page is medical advice, nor an offer of a medicinal product, nor guidance on the use of any medicine. Statements in section 12 about authorised medicines describe those medicines and their register entries; they describe neither this article nor any use of it. Where the literature cited on this page reports findings from published research, those findings are reported as bibliographic facts about that literature and are not attributed to the material supplied here.

16. Questions and answers

Is this the same substance as the active ingredient in prescription levodopa medicines?

It is the same chemical substance. It is not the same article in law. An authorised medicine is a finished product manufactured under pharmaceutical quality rules, released against a marketing authorisation, and accompanied by approved product information. This is a reagent supplied for laboratory work, released against no such framework. The molecule being identical is exactly what makes the distinction worth stating rather than assuming, and section 12 counts the authorisations that exist without claiming any relationship between them and this article.

Why does this page spend so long on a single stereocentre?

Because it is the one property of this molecule that the usual identity test cannot see, and because the wrong answer is commercially available. Levodopa has one stereogenic atom. Its mirror image, D-DOPA, has its own registry record, its own suppliers and its own catalogue numbers. Both weigh 197.19 g·mol⁻¹ and both have a monoisotopic mass of 197.06880783 Da. The European Pharmacopoeia lists the (R)-enantiomer among the impurities of levodopa that have to be controlled [22], which means the distinction is a regulatory requirement and not an academic nicety. Sections 4 to 6 set out the consequences.

Can I confirm the identity of this material by LC-MS/MS?

You can confirm its connectivity, which is most of what identity usually means and is not enough here. Accurate mass, isotope pattern and product-ion spectrum will all match — and they would match equally well for D-DOPA and for the racemate, because all three share the same formula, the same exact mass to eight decimals and the same fragmentation chemistry. If the configuration matters to your work, it needs one of the methods in section 6. If it does not, mass spectrometry is entirely adequate and this answer is longer than your problem.

What does a purity figure mean for this compound?

It depends which purity is meant, and for this substance the two are independent. Chemical purity is the proportion of the material that is levodopa rather than some structurally different compound; it is what a reversed-phase chromatogram measures. Enantiomeric purity is the proportion of the levodopa present that is the (2S) form; it is what a chiral method measures. A material can be 99.9 per cent chemically pure and simultaneously carry one per cent of the wrong enantiomer, because the wrong enantiomer co-elutes with the right one by physical necessity. Figures around 99.5 per cent circulate in commercial listings for this substance without saying which quantity they refer to; they are supplier statements, not registry values, and without the qualifier they are uninterpretable.

My powder has darkened. Is it still usable as a standard?

No. The registry record states that in the presence of moisture this substance is rapidly oxidised by atmospheric oxygen and darkens [1]. A colour change means oxidation has occurred, and the material is now a mixture of levodopa, dopaquinone and polymerisation products in unknown proportion. Its titre is lower than certified by an unknown amount. The reference appearance is colourless to white crystals or crystalline powder [1], and any grey, pink, brown or greenish cast is a departure from it. A visual check before every use is a genuine analytical control for this compound rather than housekeeping.

Will a mass spectrum tell me whether the material has degraded?

No, and this is the most misleading thing about the compound. The parent ion remains present for as long as any parent molecule survives. A mass spectrum reports what is there, not what proportion of the sample it represents. To detect degradation you need a method validated as stability-indicating, meaning one in which the degradation products are chromatographically resolved from the parent rather than hidden under it. Such a method has been published and validated for this substance alongside its photodegradation kinetics [24].

Is this the hydrochloride?

No. The material described here is the free amino acid: formula C₉H₁₁NO₄, molecular mass 197.19 g·mol⁻¹, one covalently bonded unit and formal charge zero [1]. A hydrochloride exists as a separate registered substance at 233.65 g·mol⁻¹ with two covalently bonded units [4], and the two differ by 18.5 per cent in any gravimetric calculation. Section 3 sets out the arithmetic. Note also that in the solid and in water levodopa exists as a zwitterion — that is an internal charge arrangement within one molecule, not a salt, and it does not change the mass.

Is it a hydrate?

No, and that is a measured statement rather than an assumption. The registry name endpoint returns HTTP 404 for levodopa monohydrate, on the same instrument on which levodopa hydrochloride returns a valid record. Levodopa has no registered hydrate. Its usual pairing partner does: carbidopa monohydrate is a separate registered substance at 244.24 g·mol⁻¹ against 226.23 for the anhydrous form [8], and section 12 shows the Polish register recording that distinction in only four of the eleven relevant entries.

What is the difference between L-DOPA, D-DOPA and DL-DOPA?

They are three separate registered substances with the same molecular formula, the same molecular mass and the same monoisotopic mass. L-DOPA is the (2S) form and is the substance on this page [1]; D-DOPA is the (2R) form, its mirror image [2]; DL-DOPA is the record with no configuration assigned, corresponding to the racemate [3]. They are distinguished in the registry only by the second block of the InChIKey and, for the racemic record, by a stereocentre counter reading zero defined and one undefined. If your data system stores only the first fourteen characters of the InChIKey, it cannot distinguish them at all.

Which solvent should I use to prepare a stock solution?

Dilute acid. The substance is readily soluble in dilute hydrochloric and formic acids and practically insoluble in ethanol, benzene, chloroform and ethyl acetate [1]; aqueous saturation is about 5 mg·mL⁻¹ at ambient temperature. Descriptions of this compound as soluble in ethanol, chloroform or ether circulate in commercial listings; they are the opposite of the registry entry, and the computed descriptors in section 2 — a partition coefficient of −2.7, four hydrogen-bond donors, a polar surface area of 104 Å² — independently rule them out. Prepare the solution fresh and use it promptly: a published figure of 52 per cent loss over 24 hours in aqueous extraction [25] indicates the order of magnitude at stake.

Why do you quote three melting points instead of one?

Because the record carries three, from three sources, spanning eleven degrees: 295 °C, 284–286 °C and 285 °C [1]. That spread is not sloppiness; it is the signature of a compound that decomposes rather than melting cleanly, where the observed temperature depends on heating rate, atmosphere and sample size. Averaging three such values produces a number no more true than any of its inputs. A narrow two-degree range quoted for this substance implies a precision the physical quantity does not possess, and values in the high 270s circulate commercially despite sitting below the lowest of the three registry entries.

Can I use a polarimeter to check the configuration?

Partly. The reference value is [α] −13.1° at 13 °C on the D line, at a concentration of 5.12 in 1 N hydrochloric acid [1], and reproducing it means reproducing those conditions — the temperature in particular is not one anyone hits casually. Polarimetry reliably answers is this grossly racemic. It does not reliably answer is the (R)-enantiomer below a pharmacopoeial limit, because at low contamination the signal change is small: a one per cent enantiomeric excess of the wrong form shifts a 13.1° rotation by roughly a quarter of a degree. Section 6 lists the methods that resolve that question properly.

Is levodopa a controlled substance, or prohibited in sport?

Neither, on the documents measured. It does not appear in the Polish schedules of narcotic drugs and psychotropic substances, against positive controls in the same document returning 58 occurrences for the morphine stem, 38 for fentanyl, 27 for the amphetamine stem and two for cocaine. It does not appear on the 2026 Prohibited List [11], against positive controls returning testosterone 32 times, salbutamol seven, modafinil four, and — the control that matters most, because it has the same shape as the target — selegiline twice. Selegiline is an antiparkinsonian agent that is on the List; levodopa is not. It also carries no United States controlled-substance schedule.

Why does the Polish register show 49 records but only one carrying the single-substance code?

Because almost every authorised product containing levodopa also contains a decarboxylase inhibitor, and often a third component as well, and the classification system assigns those combinations their own codes: N04BA02 for the two-component combination and N04BA03 for the three-component one [13]. Across the 49 records the split is 25, 22, one and one [10]. The single record carrying N04BA01 is Inbrija, an inhalation powder whose active-substance field reads simply Levodopum. The one remaining record carries the bare stem N04BA and is discussed in the next answer.

What is foslevodopa, and why does it appear on a levodopa page?

It is a phosphorylated derivative registered as its own substance: formula C₉H₁₂NO₇P, molecular mass 277.17 g·mol⁻¹, monoisotopic mass 277.03513872 Da [7]. It appears here for two reasons. First, its InChIKey shares the second block with levodopa, LURJTMIESA, because the stereochemistry of the single centre is the same — so tooling that matches on stereo blocks or displays truncated keys will over-associate them. Second, one record in the Polish register carries the common name Foslevodopum + Foscarbidopum and the active-substance name Levodopum 170 mg/ml + Carbidopum 9 mg/ml in the same entry, which means a parser reading only the second field will import the prodrug as its parent. The mass difference between them is 40.6 per cent.

Do you supply a certificate of analysis, and is this a pharmacopoeial reference standard?

Lot documentation accompanies each unit, and it is the correct place for lot-specific statements: batch identifier, appearance, the method used and the date of analysis. This page deliberately does not assert a purity figure for a lot it cannot see. On the second question the answer is a plain no: this material is not supplied as a pharmacopoeial reference standard. A European Pharmacopoeia chemical reference substance for levodopa exists and is obtainable from the issuing body [22]; this is a different article and the two are not interchangeable for compendial testing. Section 14 states the full boundary.

Why does this page cost more attention than the reagent costs money?

Because for this molecule the failure modes are cheap to trigger and expensive to detect. A vial left open in a humid room degrades measurably; the degradation is invisible to a mass spectrum; the enantiomeric contamination is invisible to a conventional chromatogram; and a published verification of the compendial procedure found a process impurity co-eluting with levodopa itself, such that adjusting the method to fix that overlap created new overlaps among carbidopa, catechol and a designated related compound [32]. None of those problems announces itself. All of them are avoidable by knowing about them in advance, which is the only thing a page like this can usefully offer.

References

Registry records for the substance and its close relatives

1. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 6047, Levodopa." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/6047.
2. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 92222, D-DOPA." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/92222.
3. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 836, DL-DOPA." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/836.
4. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 12002502, Levodopa Hydrochloride." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/12002502.
5. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 22040, L-Dopa Hydrochloride (no stereochemistry recorded)." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/22040.
6. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 23497, Melevodopa." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/23497.
7. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 127766, Foslevodopa." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/127766.
8. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 38101, Carbidopa Monohydrate." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/38101.

Regulatory registers and inventories

9. European Chemicals Agency. 2026. "Levodopa — substance information and Classification and Labelling Inventory entry, EC 200-445-2." ECHA. https://chem.echa.europa.eu/.
10. Urząd Rejestracji Produktów Leczniczych, Wyrobów Medycznych i Produktów Biobójczych. 2026. "Rejestr Produktów Leczniczych — public register, complete retrieval of 22 871 records, filtered on the substance name levodopum." https://rejestry.ezdrowie.gov.pl/rpl/search/public.
11. World Anti-Doping Agency. 2026. "World Anti-Doping Code International Standard: Prohibited List 2026," in effect 1 January 2026. WADA. https://www.wada-ama.org/en/prohibited-list.
12. Office of Environmental Health Hazard Assessment, California Environmental Protection Agency. 2026. "Proposition 65 List — levodopa, developmental toxicity." OEHHA. https://oehha.ca.gov/proposition-65/proposition-65-list.
13. WHO Collaborating Centre for Drug Statistics Methodology. 2026. "ATC/DDD Index — N04BA01 levodopa, N04BA02 levodopa and decarboxylase inhibitor, N04BA03 levodopa, decarboxylase inhibitor and COMT inhibitor." https://atcddd.fhi.no/atc_ddd_index/.

Deposited structural and spectral data

14. Cambridge Crystallographic Data Centre. 1972 (deposition metadata year). "CCDC 291630: Experimental Crystal Structure Determination." CCDC. https://doi.org/10.5517/cc9sgfy.
15. Wishart, D. S., et al. 2026. "Human Metabolome Database entry HMDB0000181, L-DOPA — deposited ¹H NMR, ¹H-¹³C HSQC and MS-MS spectra." HMDB. https://hmdb.ca/metabolites/HMDB0000181.
16. Takahashi, H., S. Kanaya, and N. Ogasawara. 2026. "MassBank Europe records MSBNK-NAIST-KNA00173 to KNA00685, L-DOPA; LTQ Orbitrap XL, LC-ESI-ITFT, positive and negative ionisation." Graduate School of Information Science, Nara Institute of Science and Technology, via MassBank Europe. https://massbank.eu/MassBank/.

Crystal structure and manufacturing routes

17. Mostad, A., T. Ottersen, and Chr. Rømming. 1970. "X-Ray Crystal Structure Determination of 3,4-Dihydroxyphenylalanine (L-DOPA)." Acta Chemica Scandinavica 24: 1864–1865. https://doi.org/10.3891/acta.chem.scand.24-1864.
18. Knowles, W. S. 1983. "Asymmetric Hydrogenation." Accounts of Chemical Research 16 (3): 106–112. https://doi.org/10.1021/ar00087a006.
19. Knowles, W. S. 2002. "Asymmetric Hydrogenations (Nobel Lecture)." Angewandte Chemie International Edition 41 (12): 1998–2007. https://doi.org/10.1002/1521-3773(20020617)41:12<1998::AID-ANIE1998>3.0.CO;2-8.
20. Min, K., K. Park, D. H. Park, and Y. J. Yoo. 2015. "Overview on the Biotechnological Production of l-DOPA." Applied Microbiology and Biotechnology 99 (2): 575–584. https://doi.org/10.1007/s00253-014-6215-4. (Published online 2014; print issue 2015, the year cited here.)

Stereochemistry and the enantiomeric impurity

21. Coppi, G., A. Vidi, and G. Bonardi. 1972. "Quantitative Determination of D-Dopa Present in Levodopa Samples." Journal of Pharmaceutical Sciences 61 (9): 1460–1461. https://doi.org/10.1002/jps.2600610924.
22. Wongwan, S., M. Hammitzsch-Wiedemann, and G. K. E. Scriba. 2009. "Determination of Related Substances of Levodopa Including the R-Enantiomer by CE." Electrophoresis 30 (22): 3891–3896. https://doi.org/10.1002/elps.200900060.
23. Ding, L., C. Ye, J. Zhu, C. Sun, et al. 2025. "Detection of D-Dopa in Levodopa Tablets by Derivatization Coupled with LC-MS/MS." Journal of Chromatography A 1760: 466278. https://doi.org/10.1016/j.chroma.2025.466278.

Oxidation, photodegradation and stability

24. Pereira, R. L., C. S. Paim, A. B. Barth, R. P. Raffin, S. S. Guterres, and E. E. Schapoval. 2012. "Levodopa Microparticles for Pulmonary Delivery: Photodegradation Kinetics and LC Stability-Indicating Method." Pharmazie 67 (7): 605–610. https://pubmed.ncbi.nlm.nih.gov/22888517/. (No DOI is assigned to this record in either the bibliographic index or the citation registry; cited by PubMed identifier.)
25. Pulikkalpura, H., R. Kurup, P. J. Mathew, and S. Baby. 2015. "Levodopa in Mucuna pruriens and Its Degradation." Scientific Reports 5: 11078. https://doi.org/10.1038/srep11078.
26. Neyra Recky, J. R., M. P. Serrano, M. L. Dantola, and C. Lorente. 2021. "Oxidation of Tyrosine: Antioxidant Mechanism of l-DOPA Disclosed." Free Radical Biology and Medicine 165: 360–367. https://doi.org/10.1016/j.freeradbiomed.2021.01.037.
27. Prah, A., and J. Mavri. 2024. "L-DOPA Autoxidation: An Empirical Valence Bond Simulation of the Reactive Step." The Journal of Physical Chemistry B 128 (35): 8355–8361. https://doi.org/10.1021/acs.jpcb.4c03002.

Chromatographic methods, impurity profiling and product quality

28. Gasser, U. E., A. Fischer, J. P. Timmermans, and I. Arnet. 2013. "Pharmaceutical Quality of Seven Generic Levodopa/Benserazide Products Compared with Original Madopar / Prolopa." BMC Pharmacology and Toxicology 14: 24. https://doi.org/10.1186/2050-6511-14-24.
29. Burmaoglu, R. E., and S. Saglik Aslan. 2020. "A Rapid Liquid Chromatography/Tandem Mass Spectrometry Method for Simultaneous Determination of Levodopa, Carbidopa, Entacapone and Their Six Related Compounds in Film-Coated Tablets." Rapid Communications in Mass Spectrometry 34 (12): e8782. https://doi.org/10.1002/rcm.8782.
30. Tesoro, C., R. Ciriello, F. Lelario, A. Di Capua, R. Pascale, et al. 2022. "Development and Validation of a Reversed-Phase HPLC Method with UV Detection for the Determination of L-Dopa in Vicia faba L. Broad Beans." Molecules 27 (21): 7468. https://doi.org/10.3390/molecules27217468.
31. Jiang, R., J. Yang, S. Mei, and Z. Zhao. 2022. "Determination of Levodopa by Chromatography-Based Methods in Biological Samples: A Review." Analytical Sciences 38 (8): 1009–1017. https://doi.org/10.1007/s44211-022-00132-4.
32. Korkmaz, D. A., P. Gencer, and B. Arabaci. 2026. "Integrating Box-Behnken and Plackett-Burman Designs for the Robust Impurity Profiling of Levodopa and Carbidopa Tablets." Journal of Pharmaceutical and Biomedical Analysis 281: 117671. https://doi.org/10.1016/j.jpba.2026.117671.