An osmometer counts particles. It does not weigh them, size them or identify them, and that single fact governs everything a reading licenses you to conclude about a peptide solution. Osmotic pressure is the minimum pressure that must be applied to a solution to prevent pure solvent flowing across an ideal semipermeable membrane into it, and osmolality is the effective quantity of solute per kilogram of solvent inferred from a colligative measurement on an assumption of ideality [1]. Put a synthetic peptide into a buffer and the instrument sees the peptide, every counterion that came with it, the buffer acid and its conjugate base, and any excipient, all as undifferentiated contributions to one number.
Freezing point depression and vapour pressure do not detect the same solutes
A freezing-point osmometer supercools the sample several degrees below the freezing point of the solvent and then induces crystallisation; as crystals form the heat of fusion is released and the temperature rises to the solution's freezing point, and the depression relative to pure solvent is converted to osmolality using the cryoscopic constant for water, 1.86 K mol−1 kg−1 [1]. The measurement is fast, under 0.1 h, takes 200–250 µL, and on many instruments carries a practical ceiling near 3 Osm/kg; it also reports only at the freezing point of that particular solution, so the temperature is not yours to choose [1].
A vapour-pressure osmometer measures the water vapour pressure above the sample and converts it to water activity, using sensors such as a resistive electrolytic hygrometer whose resistance changes as it absorbs or loses water vapour with the humidity of the sample cell [1]. It spans a large osmotic pressure range, 25–2800 bar, needs 0.5–3 mL, and can be run at a temperature you select — at the cost of a longer measurement, 0.5–6 h, and a sensor whose chemical compatibility with the sample has to be established rather than assumed [1].
The two techniques part company over anything that can leave the liquid. Volatile solutes are not measured by vapour-pressure osmometry, so the difference between a freezing-point reading and a vapour-pressure reading on the same sample is attributable to volatile content; in serum, the difference was used to estimate ethanol, and osmolality showed a linear relationship with isopropanol or methanol added in vitro [3]. That is the clean case. The messy case is a species only partly able to escape: in bicarbonate-buffered culture medium at pCO2 up to 250 mmHg and pH 6.7–7.5, freezing-point and vapour-pressure osmometers returned similar results, because the great majority of total CO2 in solution is present as bicarbonate, and further depletion of bicarbonate is inhibited both by the pH rise that accompanies it and by the need to maintain charge neutrality; degassing therefore lowered pCO2 sharply with little associated decrease in osmolality [4]. Volatility is not a label you can read off a component list. What matters is the fraction of the osmotically active population that can actually partition out.
With no volatile component at all the two methods can still disagree. Across 11 polymers and surfactants measured at several concentrations each, osmolality rose with concentration by both methods and showed no correlation with polymer molecular weight, but the two methods differed significantly overall, with freezing-point depression tending to return values similar to or higher than vapour pressure and the two methods giving different osmolality values for the polyvinylpyrrolidones, the polyethylene glycols and sodium carboxymethylcellulose [5]. At low osmolality the choice was not critical; at higher excipient concentrations, or with thermosensitive excipients, the vapour-pressure method was judged the better fit [5].
Osmolality is not osmolarity, and the gap opens where you work
Osmolality is referenced to a kilogram of solvent water; osmolarity is referenced to a litre of whole solution [2]. In a dilute buffer the two are interchangeable to within the precision of any instrument you own. They stop being interchangeable once the solutes occupy an appreciable share of the volume, because a litre of concentrated solution holds appreciably less than a kilogram of water. The mass-of-water basis is the one to report, precisely because the volume basis becomes misleading when solutions of differing water content are compared [2]; osmolality is also the quantity with the more nearly linear relationship to osmotic pressure [1].
The consequence is visible in any calculated value. In high-concentration monoclonal antibody formulations, osmolality measured by freezing-point depression runs much higher than the theoretical value computed from composition, and for two antibodies formulated at 150 mg/mL the measured osmolality increased approximately linearly with antibody concentration [6]. Those were intact antibodies rather than peptides, so do not carry the magnitudes across a molecular class; carry the arithmetic. A theoretical osmolarity computed from formula molarity presumes a fixed solution volume and complete dissociation into independent particles, and both presumptions decay as concentration rises.
What you are mostly measuring is the counterion and the buffer
Solid-phase synthesis and purification lean on trifluoroacetic acid as cleavage agent and ion-pairing reagent, so peptides are obtained as TFA salts [7]. The amount of counterion that arrives with the peptide depends on the sequence, and on its positive charges in particular: in a systematic study of exchange from TFA to chloride by freeze-drying from HCl, with residual TFA quantified by FT-IR, 19F-NMR and HPLC with evaporative light-scattering detection, all validated to ICH guidelines and with 10 mM HCl identified as the optimal exchange condition, the measured counterion contents were used to interpret the effect of sequence differences, especially positive charges, on counterion amount and likely localisation [7]. Counterions are measurable rather than assumable: ion chromatography, capillary electrophoresis and isotachophoresis have all been applied to acetates, trifluoroacetates and chlorides in synthetic peptides, with ion chromatography giving the best results and isotachophoresis judged least useful because of difficulty with chloride [8].
Run the arithmetic before you interpret a reading. A 3,000 g/mol peptide at 1 mg/mL is 0.33 mM. If it carries three protonated basic sites it brings roughly 1 mM of counterion with it. Set that against a 10 mM buffer contributing acid and conjugate base, and 150 mM of a fully dissociating sodium salt contributing two ions each, and the peptide plus its counterions account for a low single-digit percentage of the particle count. The osmolality of a typical working solution is a property of the buffer, not of the peptide, and this is why excipient identity matters more than excipient mass, which is what the absence of any molecular-weight correlation in that excipient comparison already showed. Grams on the label are not particles in solution.
Ideal against measured: the osmotic coefficient
Osmolality is an empirical quantity determined on the assumption of ideality and various colligative relationships [1]. The osmotic coefficient is the correction that assumption needs: the ratio of the colligative effect actually observed to the effect the same formal concentration would produce if every particle behaved independently. For the reference electrolyte, traceable three-parameter extended Hückel equations describe the activity coefficient of sodium chloride and the osmotic coefficient of water in aqueous NaCl from 273.15 to 373.15 K, applying within experimental error to all the thermodynamic data available for those solutions up to the molality of the saturated solution [9]. That is what a well-characterised solute looks like; it is not a licence to assume your peptide's coefficient tracks it.
The coefficient falls when solutes attract one another, which is what makes it informative. Osmotic coefficients of a series of amino acids computed with the AMBER ff99SB-ILDN force field were compared against available experimental data: with TIP4P-Ew water, interactions between aliphatic residues agreed well with experiment, while interactions of the polar residues serine and threonine were excessively attractive, and for every amino acid tested the osmotic coefficients were lower with the TIP3P water model [10]. Coefficients for charged amino acids depended strongly on the parameters assigned to the salt ions, and for four two-residue peptides agreement with experiment improved when partial charges were rederived for each peptide [10]. The whole comparison rests on one link, and it is the link you should carry to the bench: where the modelled solutes attracted one another too strongly, the osmotic coefficients came out low against experiment. The second thing to carry is that the counterion is part of the system being measured rather than a spectator.
Vapour-pressure osmometry is sensitive enough to resolve interaction, not merely to count. Isoosmolal preferential interaction coefficients determined by the technique for native bovine serum albumin against six small solutes were all negative, indicating preferential exclusion of those solutes from the protein at dialysis equilibrium, with glycine betaine the most excluded and glycerol the least [11]. That was a protein rather than a peptide, and the coefficients were determined over a wide range of small-solute concentrations, with the dialysis coefficient proportional to bulk solute molality at least up to 1 molal — but it establishes what the measurement can reach when the arithmetic is set up to isolate an interaction term.
Self-association shows up as particles that go missing
Colligative counting has one consequence that makes an osmometer useful for something it was not built for. If n monomers assemble into one n-mer, the particle count falls by a factor of n for that population, and the solution reads below what its formula concentration predicts. Osmometry therefore becomes an indirect reporter of association state.
Peptides give it plenty to report. Semaglutide, a lipidated peptide, aggregates slowly in aqueous solution into spherical micelles at sufficiently high concentration, alongside a small population of needle-shaped fibrils, while at lower concentration dimer and trimer structures form; the micelles, once formed, were stable toward further ageing, as shown by cryogenic transmission electron microscopy, small-angle X-ray scattering and molecular dynamics [12]. The equilibrium also moves with the salt you chose. Glucagon trimerisation constants at pH 10.6 in 0.76 M K2HPO4, calculated from circular dichroism between 5 and 50 °C, together with a comparison of fluoride and chloride against phosphate at 25 °C, were consistent with binding of approximately one molecule of salt to the trimer and a systematic decrease in the number of water molecules bound to the trimer relative to the monomer across the series K2HPO4, KF, KCl [13]. Temperature moves it too: vapour-pressure osmometry was used to parameterise the thermosensitivity of interactions between peptide amphiphiles in aqueous solution, and the measurement showed the peptides becoming more hydrophobic as temperature rose, quantified through the Flory–Huggins χ parameter [14]. That experiment is available to vapour-pressure osmometry because the technique lets you set the temperature, and closed to freezing-point depression, which reports only at the freezing point [1].
Be strict about what a deficit buys you. It does not give you the association number, the size distribution, or the shape of anything. It does not separate association from ordinary thermodynamic non-ideality, since both depress the measured coefficient. And it is only as good as the inventory you compared against: an osmolality that reads low against a calculation is meaningless if the counterion stoichiometry in that calculation was assumed rather than measured [7, 8]. Treat the number as a flag that sends you to a sizing or scattering method, not as the answer.
Tonicity is a property of a pairing, not of the tube
Osmotic pressure is defined against an ideal semipermeable membrane [1]. A colligative osmometer has no membrane in the measurement at all: it counts every particle, whether or not a real barrier would reject it. Direct membrane osmometry makes the difference concrete. Solute pass-over across the membrane raises solute concentration on the solvent side, lowers the chemical potential difference, and depresses the observed osmotic pressure; measured pass-over ranged from none detectable for Eosin Y to over 7% for a 1500 g/mol polyethylene glycol, with the polymer figure attributed to low-molecular-weight oligomers in the sample rather than to the polymer itself [1]. The osmolality you record is a property of the solution alone. Anything described as effective osmotic activity is a property of the solution and a named barrier together, and no benchtop osmometer reports it.
Where the reading actually goes wrong
Start with what the instrument contributes, because it is usually the smallest term. A freezing-point micro-osmometer evaluated against linearity studies and split-sample correlation with vapour-pressure osmometry, freezing-point osmometry and an external quality assurance programme showed a within-run coefficient of variation of 0.59% and a between-day coefficient of variation of 0.58%, accuracy across 0–2000 mmol/kg, and a 20 µL sample requirement [15]. Sub-1% imprecision means that a several-percent discrepancy between measured and calculated osmolality is chemistry, not noise — provided the calibration is current across the range you are working in, which the linearity data define rather than extend.
Freezing-point depression fails in specific, recognisable ways. It cannot be used when freezing triggers a second phase transition, such as gas release, precipitation of a solid, or liquid–liquid phase separation of a solute like a surfactant [1]. It suffers from pre-freezing, in which particles nucleate crystallisation and prevent the supercooling the method depends on, and it cannot handle solutes that inhibit water crystal formation; both problems worsen as viscosity rises [1]. The obvious remedy is not clean: dilution changes the answer non-proportionally, because non-idealities differ from substance to substance, so an osmolality measured on a diluted aliquot cannot simply be scaled back [1].
Vapour-pressure osmometry fails differently. Volatile molecules can interact with the sensor and disturb the vapour pressure measurement, and in the reported case of a small volatile amine, sensor interaction prevented a stable and reliable reading, leaving the value plausibly higher than the true one [1]. Most vapour-pressure instruments are not isopiestic measurements and their sensors are often of low accuracy at low osmotic pressure, given that osmotic pressure depends on the natural logarithm of water activity [1]. Sample geometry deserves a thought as well: the smaller sample and larger surface area of a vapour-pressure osmometer would be expected to give more extensive degassing than a freezing-point instrument, an expectation that made no difference in the bicarbonate system but is the right argument to apply whenever a buffer holds a species that can leave [4].
Two further practicalities. In that culture-medium study, osmolality did not change in samples frozen at −20 °C for up to one year [4], which is an encouraging precedent for salt-buffered aqueous systems and not a guarantee for a peptide whose association state may not survive a freeze-thaw. And the methods do not converge: even with care, vapour-pressure, freezing-point and direct membrane osmometry give different osmotic pressure values for concentrated aqueous solutions, with the first two consistently reading higher than the third across small-molecule amines in water, small molecules in carbonated water and polymers in water [1]. An osmolality without a stated method, temperature and concentration is not a comparable number.
Reading it beside the other concentration-dependent properties
Osmolality, viscosity and net charge are three instruments pointed at the same crowded solution, and they answer different questions about it. Osmolality counts; viscosity reports on how the molecules obstruct one another; net charge sets whether they repel or approach at all. The three are coupled in practice — viscosity and osmotic pressure both generally rise with concentration, and rising viscosity is what degrades the freezing-point measurement and promotes concentration gradients in the membrane method [1]. The isoelectric point guide in this library covers the charge state that governs whether self-association is favoured at your working pH, and the viscosity guide covers what that association does to flow. When you record an osmolality, record with it the method, the measurement temperature, the calibration range, the concentration, and the full counterion and excipient inventory the calculation was compared against. Without those, the number describes a tube nobody can reconstruct.