Blood does not freeze at 32°F (0°C). The freezing point of human blood sits at roughly 26.8°F to 28.4°F (−2.9°C to −2.2°C), because dissolved salts, proteins, and glucose depress the freezing point below that of pure water. Whole blood begins forming ice crystals near 28°F, while the plasma component freezes around 27°F. This gap of roughly 4°F below water's freezing point explains why a person can survive exposure to subfreezing air for a time. It also explains why donated blood units are never stored at 32°F.
The mechanism: why dissolved particles lower blood's freezing point
Freezing is not a single event. It starts when water molecules lock into a crystal lattice, and that lattice can only form if the water is pure enough to assemble cleanly. Every dissolved particle in blood gets in the way.
Plasma is about 92 percent water by weight. The remaining 8 percent is a crowded mix: sodium at roughly 140 mmol/L, chloride, potassium, bicarbonate, glucose near 90 mg/dL, and 6 to 8 g/dL of albumin and globulins. Each of these particles occupies space and interrupts the hydrogen bonding that ice needs.
The physics is colligative, meaning it depends on particle count rather than particle identity. A single sodium ion depresses the freezing point about as much as one glucose molecule. When you add up all the solutes, blood's osmolality lands around 275 to 295 mOsm/kg.
That osmolality produces a freezing point depression of about 0.52°C per osmole. Doing the arithmetic gives 2.2°C to 2.9°C of depression, which matches the measured range. In plain terms: blood freezes lower than water because it is salty, and the saltiness is precisely measurable.
There is a second layer. Ice formation needs a nucleation site, a tiny surface where the first crystal can start. Without one, water can stay liquid well below its freezing point, a state called supercooling. Blood contains abundant nucleation sites: cell membranes, protein aggregates, and damaged tissue. That is why blood in a living body does not supercool much. It crystallizes close to its true freezing point.
The quantified picture: freezing points of blood and its components
Different parts of blood freeze at different temperatures, and the spread matters for medicine. The table below reconciles the numbers so no threshold is left ambiguous.
| Fluid or component | Freezing point | What drives the value |
|---|---|---|
| Pure water | 32.0°F (0°C) | No solutes; reference point |
| Whole blood (fresh) | 28.0°F to 28.4°F (−2.2°C to −2.0°C) | Combined solute load of plasma plus cells |
| Blood plasma | 26.8°F to 27.3°F (−2.9°C to −2.6°C) | Highest solute concentration of the fractions |
| Blood serum | 27.0°F to 27.5°F (−2.8°C to −2.5°C) | Plasma minus clotting factors |
| Red blood cell suspension | 28.5°F to 29.0°F (−1.9°C to −1.7°C) | Cells plus a smaller volume of plasma |
| Packed red cells (glycerolized) | −110°F to −120°F (−79°C to −84°C) | Glycerol cryoprotectant added before freezing |
The reconciliation sentence: whole blood at about 28°F, plasma at about 27°F, and water at 32°F form a consistent ladder, with plasma lowest because it carries the full dissolved load that red cells dilute in a suspension. If you see a source quote a single figure such as 27.5°F, that is a plasma value being applied to whole blood.
One comparison the top search results usually skip: blood's freezing point depression of roughly 2.5°C is about 2.5 times smaller than seawater's, which drops to about 28.6°F (−1.9°C to −2.1°C) at typical salinity. Blood is less salty than the ocean. That is a useful intuition check, because people often assume blood and seawater are similar in solute load.
In the living body: frostbite, hypothermia, and tissue freezing
The freezing point of a fluid is not the same as the temperature at which tissue is damaged. Living tissue freezes at a slightly higher effective temperature than pure plasma would suggest, because ice forms in the extracellular space first and pulls water out of cells by osmosis.
Frostbite injury begins with ice crystal formation in the skin and subcutaneous tissue. The clinical threshold commonly cited is tissue temperature below about 28°F (−2.2°C) for sustained periods. At that point, extracellular ice forms, cellular dehydration follows, and microvascular damage starts.
There is a non-obvious detail here: the first ice to form in a frostbitten finger is not inside the cells. It is in the interstitial fluid between them. The crystals grow outward, concentrate the remaining extracellular solutes, and suck water out of adjacent cells until the cells shrink and their membranes rupture. This is why rapid rewarming, not slow warming, is the standard treatment. Slow thawing lets the crystals grow and re-form during the refreeze cycles.
Hypothermia is a separate problem. Core temperature below 95°F (35°C) defines hypothermia, and severe hypothermia below 82°F (28°C) carries high mortality. But the blood itself does not freeze at these core temperatures. A person with a core temperature of 75°F (24°C) has dangerously cold blood that is still liquid. The freezing point of blood in the body is a floor that core hypothermia almost never reaches, because the heart stops before that.
In plain terms: hypothermia kills by disabling the heart and brain, not by freezing the blood. Frostbite freezes tissue locally at the surface, where temperatures can plunge far below core values.
Timeline: how long blood and tissue take to freeze
Freezing is a process, and the time course matters for first aid, for lab storage, and for understanding survival stories. The rows below describe a typical adult in cold exposure and a typical blood sample in a freezer.
| Time from exposure | What is happening |
|---|---|
| 0 to 5 minutes | Skin temperature drops toward ambient; vasoconstriction shunts blood to the core |
| 5 to 15 minutes | Superficial tissue approaches freezing; numbness and pallor appear |
| 15 to 30 minutes | Extracellular ice forms in exposed skin at tissue temperatures near 28°F or lower |
| 30 to 60 minutes | Crystal growth and cellular dehydration; frostbite becomes established |
| 1 to 3 hours | Deep tissue involvement possible; microvascular thrombosis begins |
| 2 to 4 hours | Core hypothermia can progress to severe range if exposure continues |
| 8 to 24 hours | A blood sample in a freezer at 0°F (−18°C) is fully frozen; ice crystal damage to cells is complete |
| Days to weeks | Red cell hemolysis from freeze-thaw becomes extensive without cryoprotectant |
The rows deliberately separate the living-body process from the lab process. A person and a blood bag freeze on different clocks, which is one reason clinical guidance for frostbite and laboratory guidance for blood banking look so different.
Where the number changes: additives, anticoagulants, and disease states
The freezing point of a blood sample is not fixed. It shifts with additives and with the donor's condition, and these variations are the part most explainers leave out.
Anticoagulants change the solute load. A unit of whole blood collected into a citrate-phosphate-dextrose bag has a slightly lower freezing point than native blood, by a fraction of a degree, because the citrate salts add particles. Glycerol, used to cryoprotect red cells for long-term storage, drops the freezing point dramatically, to about −110°F, which is why glycerolized units are frozen at −65°F or colder.
Disease states shift the value too. Severe hyperglycemia raises glucose and therefore lowers the freezing point slightly. Hypernatremia and renal failure with high urea do the same. These shifts are small, usually less than 1°F, but they are real and measurable in a laboratory osmometer.
There is a professional caveat worth stating plainly. The freezing point of blood is a thermodynamic property of its solutes. It is not the same as the temperature at which a blood sample can be safely stored or the temperature at which a transfusion becomes dangerous. Those are clinical decisions governed by separate standards. The 2016 AABB standards, for example, specify storage temperatures for red cells at 33.8°F to 42.8°F (1°C to 6°C), which is deliberately above the freezing point to prevent hemolysis.
In plain terms: the freezing point tells you when ice appears, not when blood is clinically usable.
Common misconceptions and edge cases
Several beliefs about frozen blood are wrong in ways that matter. The bulleted list below covers the corrections.
- Misconception: blood freezes at 32°F like water. Correction: the solute load depresses it to about 27°F to 28°F, so blood stays liquid at ordinary refrigerator temperatures near 39°F.
- Misconception: a frozen blood bag is ruined. Correction: red cells frozen with glycerol survive for 10 years or more at −65°F or colder, but red cells frozen without cryoprotectant hemolyze extensively.
- Misconception: a person with frostbite has frozen blood throughout the body. Correction: frostbite is a local tissue injury. Core blood remains liquid until the heart fails from hypothermia, not from freezing.
- Misconception: adding salt to ice makes blood freeze faster. Correction: salt lowers the freezing point of the surrounding water, which slows ice formation and is how salt-ice baths work as cooling agents, not as freezing accelerators.
- Edge case: a person with very high blood glucose or urea has a measurably lower freezing point, but the shift is under 1°F and does not change survival thresholds.
The most consequential of these is the second bullet. Blood banking relies precisely on the gap between 28°F and 32°F. Storing red cells at 39°F keeps them above the freezing threshold, which prevents the ice crystal injury that would destroy the units.
What to actually do: practical actions that follow from the numbers
The freezing point is not just trivia. It drives real decisions in emergency response, laboratory handling, and cold-weather safety.
For cold exposure, treat any tissue that has been below 28°F for more than 15 minutes as frostbite until proven otherwise. Remove wet clothing, protect the area from refreezing, and rewarm in water at 98.6°F to 102.2°F (37°C to 39°C) for 15 to 30 minutes. Never rub the area with snow or use dry heat, because both increase crystal damage and burn risk.
For blood handling, keep refrigerated units above 33.8°F and transport them in validated containers. Never place a blood bag directly against a freezer wall or dry ice, because local temperatures can drop below the freezing point and cause hemolysis.
For household thermometers and weather planning, remember that a reading of 28°F means the air is cold enough to start freezing exposed tissue, even though the blood in your core stays liquid. Dress in layers, cover extremities, and limit exposure below 28°F to under 30 minutes where possible.
For anyone monitoring a medical freezer, set the alarm at 30°F, not 32°F. That margin acknowledges the freezing point of blood and gives staff time to respond before ice forms.
Limitations of the numbers and what remains uncertain
The measured freezing point of blood is well established, but several adjacent questions are not fully settled. The exact freezing point varies between individuals by a fraction of a degree depending on hydration, protein levels, and glucose. The 2.2°C to 2.9°C range is the accepted envelope, not a single universal constant.
The tissue-freezing threshold is more hypothesis-heavy. The commonly cited 28°F figure comes from clinical observation and animal models, not from a precise controlled trial in humans. Frostbite severity depends on wind chill, contact with cold metal, and duration, so a single temperature cannot capture the risk.
The supercooling behavior of blood in the body is also incompletely characterized. Some evidence suggests small volumes of blood in narrow vessels can remain liquid below the nominal freezing point for limited periods, which is a plausible explanation for some survival cases. That is hypothesis, not established fact.
What is certain is the direction and rough magnitude. Blood freezes below water, plasma freezes lowest of the fractions, and the body's cold injuries arise from local tissue freezing and systemic hypothermia rather than from blood turning to ice in the vessels. Those three statements are the operational core, and they hold across every edge case above.