Subtract 99.0 from 99.9 and the answer is 0.9. That arithmetic is correct and almost entirely useless, because the informative quantity is not the headline figure but the remainder. What does 99 percent purity actually mean in a gram of material? It means 10 mg of something that is not the named compound. At 99.9 percent it means 1 mg. The remainder has not shrunk by 0.9 percent; it has shrunk by a factor of ten, and the question of what that remainder consists of matters more than either number.
Two batches can carry the same purity figure and behave differently. Two laboratories can measure the same bottle and print different figures without either being wrong. Both facts follow from the same source: a purity value is a summary of a particular measurement, not a property of the powder.
What 99 Percent Purity Actually Means, and What It Does Not
A percentage purity is a ratio, and a ratio needs a numerator and a denominator that someone chose. The three commonest choices give three different answers.
- Area normalisation — the main peak area divided by the total integrated area of a chromatogram. It assumes every component responds identically to the detector, and it cannot see anything the detector is blind to.
- Assay against a standard — the response of the sample compared with a characterised reference of known content. This gives mass fraction, but inherits the uncertainty of the standard.
- Mass balance — one hundred percent minus every impurity class that was separately measured: water, residual solvents, inorganic residue, related organic substances. It is the most laborious and the most honest about its own gaps.
An area-normalised 99.5 percent and a mass-balance 99.5 percent are not the same statement. The first says the main peak dominated one chromatogram. The second says four categories of non-analyte were quantified and summed.
Why Area Percent and Mass Percent Disagree
Detector response is compound-specific. A 2025 paper in Analytical Chemistry made the case bluntly for peptide impurity profiling: impurities arising from insertion, truncation, deamidation, isomerisation and oxidation do not have equal responses to the parent molecule, so relative response factors above and below one both occur. Applying a default relative response factor of one, the paper notes, leads to over-estimation where the true factor is above one and under-estimation where it is below.
The magnitude of that error is measurable. A 2026 study using liquid chromatography with charged aerosol detection, a detector chosen precisely because it responds to mass rather than to chromophore, reported a relative response factor relative standard deviation of 13 percent across nineteen natural toxins from six structural classes, against an average relative response factor of 4.8 percent within a class. Even a deliberately near-universal detector is not universal.
Ultraviolet detection is worse in a specific and predictable way. A molecule with no chromophore in the monitored wavelength range produces no peak at all. Its contribution to the denominator is zero, which inflates the main peak’s share of the total. The impurity did not vanish; the instrument never saw it.
What Is in the Remaining Fraction
The ICH Q3A guideline on impurities in new drug substances sorts the remainder into three categories. Organic impurities, arising during manufacture or storage, cover starting materials, by-products, intermediates, degradation products, and reagents, ligands and catalysts. Inorganic impurities cover reagents, ligands and catalysts again, heavy metals or other residual metals, inorganic salts, and other materials such as filter aids and charcoal. Residual solvents form the third category.
That list is the reason a single chromatographic number cannot close the question. A reversed-phase gradient with ultraviolet detection addresses part of the first category and none of the other two. Water is invisible to it. Palladium is invisible to it. Residual methanol, lacking a useful chromophore and eluting unretained, is invisible to it.
Published certification work shows how large the unseen fraction can be. A 2025 study assigning a certified purity to L-thyroxine used mass balance alongside quantitative nuclear magnetic resonance, quantifying impurities by Karl Fischer coulometric titration, ion chromatography, gas chromatography–mass spectrometry and inductively coupled plasma mass spectrometry. Ten structurally related organic impurities were separated and five quantified. The single largest impurity was not any of them: it was water, at 3.563 percent. The certified value came out at 94.90 percent with an expanded uncertainty of 0.34 percent.
Why Two Figures Can Be Indistinguishable
Certified values carry an interval. The NIST certificate for a D-glucose standard states 99.6 ± 0.3 percent as mass fraction, the interval covering the measurand with approximately 95 percent confidence. That interval runs from 99.3 to 99.9. A second material certified at 99.8 ± 0.3 percent would overlap it almost completely, and no comparison between the two headline numbers would be meaningful.
Most routine purity figures arrive with no interval at all. Two such figures from different laboratories, produced by different methods on different batches, cannot be subtracted and treated as a difference. The gap between 99 and 99.9 percent survives that objection, because a tenfold difference in remainder exceeds any plausible method spread. The gap between 99.5 and 99.7 often does not.
The same glucose certificate illustrates how narrow the certified statement really is. Moisture is reported as less than 0.06 percent and non-volatile inorganic content, measured as ash residue, as less than 0.02 percent. Those figures exist because mass fraction alone would not have told anyone whether the remainder was water or metal oxide.
Order of Magnitude in the Impurity Budget
Regulatory thresholds put the difference into a scale. ICH Q3A sets thresholds that scale with the maximum daily dose of a substance; in the lower-dose band, an impurity above 0.05 percent is reported, one above 0.10 percent is identified, and one above 0.15 percent is qualified, with the latter two also capped in absolute daily intake where that cap is the more restrictive of the two. The guideline also defines an unidentified impurity as one whose structure has not been established and which is defined solely by qualitative analytical properties such as chromatographic retention time.
Set those against the two remainders. A 1 percent remainder can accommodate twenty separate components each sitting exactly at the reporting threshold, or six components each at the qualification threshold. A 0.1 percent remainder accommodates two at the reporting threshold and none at all at the qualification threshold. Moving from 99 to 99.9 percent does not merely reduce a quantity; it moves the entire remainder out of the range where individual components require structural identification.
| Question | At 99.0 percent | At 99.9 percent |
|---|---|---|
| Non-analyte mass in one gram | 10 mg | 1 mg |
| Components possible at 0.05 percent each | Up to twenty | Up to two |
| Components possible at 0.15 percent each | Up to six | None |
| Likelihood an individual impurity needs identification | High | Low |
| Effect of one undetected non-chromophoric impurity | Modest relative error | Can dominate the remainder |
The last row is the one that bites. At 99.9 percent, a single impurity that the assay cannot see is not a rounding error in the remainder — it can be most of it.
What Mass Balance and Orthogonal Methods Add
The metrological answer to method dependence is to avoid relying on one method. A 2025 certification of quantitative NMR internal standards at a national metrology institute compared mass balance, quantitative NMR, freezing-point depression and coulometry, deliberately assigning certified values by routes independent of NMR so that the traceability chain would not rest on the technique the materials were meant to support. Delivered purities exceeded 999.8 milligrams per gram with uncertainties between 0.6 and 3 milligrams per gram at a coverage factor of two.
A 2025 purity reference material of gamma-linolenic acid methyl ester followed the same logic on a lower plateau, determining purity by mass balance and validating it by quantitative NMR, with a certified value of 98.9 percent and an expanded uncertainty of 0.4 percent. The uncertainty combined contributions from characterisation, homogeneity and stability — three sources, none of which appears on an area-percent chromatogram.
Response-factor bias has attracted the same treatment from the other direction. A 2025 paper on extractables and leachables analysis noted that variability in response factors complicates quantitation, that an uncertainty factor only partially corrects it, and that a multi-detector approach improves characterisation without fully resolving the bias. The honest reading of a purity figure therefore includes the question of which detector produced it.
Frequently asked questions
Is 99.9 percent purity ten times purer than 99 percent?
Not in purity, but in the remainder, yes. One gram at 99.0 percent contains 10 mg of material that is not the named compound; at 99.9 percent it contains 1 mg. The headline figures differ by 0.9 percentage points, while the quantity of everything else differs by a factor of ten. The remainder is the informative quantity.
What is the difference between area percent and mass percent?
Area percent is the main peak’s share of total integrated detector response on one chromatogram. Mass percent is the fraction of sample mass that is the named compound. They coincide only if every component responds identically to the detector and every component is detected. Neither assumption normally holds, so the two figures differ by an unpredictable amount.
Why do impurities have different response factors in HPLC?
Because ultraviolet response depends on the chromophore and its molar absorptivity at the monitored wavelength, which varies with structure. Published work on peptide impurities reports relative response factors both above and below one for insertion, truncation, deamidation, oxidation and isomerisation products. Assuming a factor of one over-estimates some impurities and under-estimates others.
Does an HPLC purity figure include water and residual solvents?
Usually not. Water gives no ultraviolet absorbance and elutes unretained, and many residual solvents behave the same way. Determining them requires separate techniques: Karl Fischer titration for water, headspace gas chromatography for solvents, and elemental analysis for inorganic residue. A certification of L-thyroxine found water to be its single largest impurity at 3.563 percent.
How is purity assigned by mass balance?
By measuring each impurity class separately and subtracting the total from one hundred percent. The classes are water, residual solvents, non-volatile inorganic residue and structurally related organic compounds, each determined by a technique suited to it. The approach is laborious but explicit about its own coverage, since anything not measured is visibly missing from the subtraction.
Why does a certified purity value carry an uncertainty?
Because the certified value combines contributions from characterisation, batch homogeneity and stability, and each contributes spread. A NIST glucose standard is certified at 99.6 plus or minus 0.3 percent mass fraction, an interval covering the measurand with roughly 95 percent confidence. Two certified values whose intervals overlap cannot be ranked against each other.
How many impurities can sit below a reporting threshold?
Arithmetically, a 1 percent remainder accommodates twenty components at 0.05 percent each, while a 0.1 percent remainder accommodates two. This is why the step from 99 to 99.9 percent changes the character of a material rather than just its grade: the remainder moves out of the range where individual components normally require structural identification.
References
- ICH Q3A(R2) — impurities in new drug substances, classification and thresholds
- US Food and Drug Administration — Q3A guidance, independent reading of the same classification and threshold tables
- Analytical Chemistry, 2025 — relative response factors in the assessment of impurities in peptide therapeutics
- ACS Measurement Science Au, 2026 — charged aerosol detection and response-factor uniformity across toxin classes
- Molecules, 2025 — L-thyroxine certified reference material by mass balance and quantitative NMR
- NIST Standard Reference Material 917d — certificate of analysis with certified purity, moisture and ash residue
- Analytical and Bioanalytical Chemistry, 2025 — certification of qNMR internal standards by orthogonal primary methods
- Food Science and Nutrition, 2025 — purity certified reference material of gamma-linolenic acid methyl ester
- Analytical and Bioanalytical Chemistry, 2025 — uncertainty factors and relative response factors in extractables and leachables studies
Research use only. Nonsensia supplies analytical reference standards for laboratory and research applications. This article is published for scientific and educational purposes. It is not medical advice, it does not describe any use in humans, and nothing in it should be read as a recommendation to administer any substance to a person or animal.
Filed under: Purity & Quality Control
This article is part of our guide to What Purity Actually Means: Impurities, Certificates and Specifications.