Review Warns Wine Oxygen Readings Cannot Be Safely Compared Across Wines or Devices

A scientific review published on Wednesday, November 4, warns that the dissolved oxygen numbers routinely reported in wine are less direct than they appear and may not be safely compared across different wines or measuring devices.

The review, published by IVES Technical Reviews and written by Ignacio Nevares and Maria del Alamo-Sanza of the University of Valladolid’s UVaMOX Group in Spain, argues that oxygen meters do not directly measure oxygen in milligrams per liter. Instead, the sensor detects oxygen partial pressure, known as pO2, or oxygen flux, and the instrument then converts that signal into mg/L using a solubility model. In wine, the researchers said, both the liquid itself and the type of sensor can affect that result.

That distinction matters because dissolved oxygen is closely watched in wineries as producers try to manage oxidation, shelf life, and sulfur dioxide additions. If a reading in mg/L is treated as universally comparable when it depends on the wine’s composition and the model used by the device, a winery could misread oxygen exposure or make sulfur dioxide adjustments on a weak basis. The issue could also matter more broadly in beverage production wherever oxygen control is used as a process tool.

The authors said a typical measurement appears straightforward: a technician inserts a probe, waits for a stable display and records a figure such as 0.42mg/L. But, they wrote, that screen value is the end of a conversion process, not a direct weighing of oxygen in a liter of wine. In a Clark electrode, the current depends on oxygen moving through a membrane. In an optical sensor, oxygen affects the luminescence of an indicator. In both cases, the primary response reflects oxygen availability at the sensor interface, not a direct mass concentration.

According to the review, the problem is not that modern instruments are necessarily inaccurate. The issue is that the meaning of the number on the screen depends on the solubility relationship used to turn the sensor signal into mg/L. An instrument can be precise and repeatable, the authors said, while still giving only an approximate concentration if the model behind the conversion does not match the wine being tested.

The researchers said wine should not be treated like water for this purpose. Many oxygen meters are calibrated in air-saturated water or rely on water-based solubility relationships. Those references can be useful for checking the sensor, they wrote, but they do not justify assuming that the relationship between pO2 and mg/L is the same in wine. Wine is a hydroalcoholic and reactive matrix in which ethanol, sugars, and extract can change oxygen solubility. Temperature and pressure also affect both solubility and transport. Headspace conditions and carbon dioxide add more variables.

As a result, the review said, two wines can show the same pO2 while containing different amounts of physically dissolved oxygen. A generic conversion to mg/L may hide that difference instead of clarifying it.

The authors also drew a line between two separate sources of distortion. One is physicochemical: the composition of the wine changes oxygen solubility and therefore the relationship between pO2 and mg/L. The other is instrumental: the wine matrix may interfere with the sensor response itself. In optical systems, that may involve the luminophore and its support. In polarographic systems, it may involve membrane diffusion.

The review cited earlier work by del Alamo-Sanza and colleagues from 2014, which found that these deviations were already visible when readings were expressed in hectopascals and that the effect varied by instrument. In tests on eight real wines, direct errors of 11% to 19% were reduced to 0% to 5% after device-specific compensation for ethanol and sugar, and to less than 0.5% with one optical system. The authors stressed, however, that this correction addressed the sensor response, not the true oxygen solubility of the wine.

The paper also revisits little-known research from the Soviet Union between 1970 and 1979. The authors said language barriers and limited access kept that work from becoming widely cited, but they argue that its findings remain relevant. Those studies reported that rising temperature reduced oxygen solubility while increasing the polarographic response, that membrane sensors responded to pO2 rather than directly to oxygen mass, and that extract and other wine components affected solubility enough to justify composition-specific equations.

The review says newer optical sensors have improved stability and reduced maintenance needs. They also avoid consuming oxygen during measurement and allow readings with less disturbance. But the authors argue that better sensor design does not remove the underlying model problem. A more stable reading, they wrote, is not automatically a more universal one.

For wineries, the researchers do not recommend abandoning mg/L reporting. They said those values remain useful when one wine is being followed under fixed conditions or when relative changes are being tracked within a stable protocol. They advise more caution when values are compared across different wines, devices, or laboratories, when they are used in stoichiometric calculations or sulfur dioxide requirement estimates, or when a single reading is treated as if it reflects equilibrium conditions.

Instead, the authors propose that wineries and laboratories record more information alongside the mg/L figure whenever instruments allow it. That includes the native signal, such as pO2, hectopascals, or saturation; the actual temperature and absolute pressure; the calibration and the model used to convert pO2 into mg/L; and details of the matrix, including alcohol, sugars, and extract. They also recommend noting the measurement context, including where and when the reading was taken, how long the wine was allowed to stabilize, whether there was agitation, the size of the headspace, and whether the wine was taking up or consuming oxygen at the time.

If an instrument reports only mg/L, the review says the measurement can still be useful for internal process control as long as the same device and protocol are kept constant. The authors warn against treating that result as universally comparable beyond those conditions.

They also argue that a single oxygen reading may say less than the behavior of the wine over time. In process studies, they wrote, the evolution of pO2, including its slope, depletion time, or area under the curve, can offer more useful information than one isolated concentration value.

The review is tied to a 2026 paper by the same researchers in Food Chemistry that reassessed dissolved oxygen measurement in wine and the limitations of mass-based reporting. Together, the papers challenge a common assumption in cellar work: that a dissolved oxygen number in mg/L means the same thing from one wine, instrument, or lab to another. For producers, that could lead to a more careful use of oxygen data in routine control, bottling decisions, and antioxidant management.