Why Centrifuges Separate Mixtures

Whole blood looks uniform in a test tube, but spin it in a centrifuge for ten to fifteen minutes at 1,000 to 2,000 times Earth's gravity and it splits cleanly into three visible layers: dense red blood cells at the bottom, a thin white buffy coat of white blood cells and platelets in the middle, and pale yellow plasma on top. Nothing was added, nothing was filtered. Spinning alone did the sorting, by exploiting a difference that was already there: density.

That single principle, that particles of different density move at different speeds through a fluid under an applied force, underlies an enormous range of technology. It's the same physics whether the machine is a tabletop unit in a hospital lab separating serum from clotted blood, an ultracentrifuge in a structural biology lab pelleting viral particles, or an industrial gas centrifuge cascade enriching uranium isotopes for nuclear fuel. Understanding how a centrifuge actually works, and why it isn't simply "a fast spinning machine that throws things apart," explains a surprising amount of modern laboratory medicine, molecular biology, and even nuclear engineering.

Sedimentation, sped up

Left alone under normal gravity, denser particles suspended in a liquid will eventually settle to the bottom on their own. This slow, familiar process is called sedimentation — the gradual downward movement of particles denser than the surrounding fluid, driven by gravity. Anyone who has left a glass of muddy water standing overnight has watched sedimentation happen: the silt collects at the bottom while clearer water remains above.

A centrifuge doesn't introduce a new physical principle. It accelerates one that already exists. By spinning a sample at high speed around a fixed axis, a centrifuge generates centripetal acceleration far stronger than Earth's gravitational acceleration of 9.8 meters per second squared (32.2 feet per second squared). The sample container, and everything inside it, is forced to travel in a circular path rather than a straight line, and the force required to keep it on that circular path scales with both the rotor's speed and its radius.

Because denser particles experience more net outward force per unit volume than less dense particles suspended in the same medium, they migrate toward the bottom of the tube — which in a spinning rotor means radially outward from the spin axis — faster than lighter components do. Modern laboratory ultracentrifuges can generate forces exceeding 1,000,000 times Earth's gravity, according to equipment specifications published by manufacturers such as Beckman Coulter, compressing a separation process that might take days or weeks under gravity alone into a matter of minutes or hours.

It's worth being precise about what "faster settling" actually buys a researcher. Sedimentation under gravity alone is often too slow to be useful, particularly for very small particles like proteins or viruses, because thermal jostling from surrounding water molecules (Brownian motion) keeps them suspended almost indefinitely. Colloidal particles, those small enough to stay dispersed rather than settle, simply don't sediment in any practical timeframe under 1 x g. Multiplying the effective gravitational pull by a factor of tens of thousands or more is what makes separating these particles feasible at all.

Relative Centrifugal Force: why "how fast" isn't the full story

Centrifuge performance is measured in relative centrifugal force (RCF) — the force experienced by a sample expressed as a multiple of Earth's standard gravity, written as "x g" — rather than simply rotations per minute (RPM). This distinction matters because the actual force generated depends on both rotational speed and the radius from the spin axis to the sample, not on rotational speed alone. The relationship is described by the formula RCF = 1.118 × 10⁻⁵ × r × N², where r is the radius in centimeters from the center of rotation to the sample, and N is the rotational speed in RPM.

This is why the same rotor spun at the same RPM produces different forces on samples positioned at different distances from the axis, and why a smaller rotor spinning faster can sometimes generate the same RCF as a larger rotor spinning more slowly. It's the reason clinical and research protocols specify RCF (x g) rather than RPM alone: RPM by itself doesn't fully describe the physical force a sample actually experiences during a run.

In practice, this shows up constantly in laboratory work. A widely used clinical protocol for preparing blood plasma calls for spinning a sample at 1,000 to 2,000 x g for ten to fifteen minutes; for platelet-depleted plasma, some protocols call for 2,000 x g for fifteen minutes. If a technician transfers that protocol from one centrifuge model to another without recalculating RPM for the new rotor's radius, using the same RPM number could under- or over-process the sample, because the two machines' rotors may have very different radii and therefore produce very different RCF at the "same" speed.

Urenco's TC-21 gas centrifuge, used in commercial uranium enrichment, spins its 20-centimeter-diameter rotor at speeds where the outer wall travels at roughly 770 meters per second, faster than the speed of sound in air.

What is relative centrifugal force and why does it matter more than RPM?

RCF measures the actual force a sample experiences as a multiple of Earth's gravity, and it depends on both rotational speed and the radius from the spin axis to the sample. Reporting only RPM omits the radius term, meaning the same RPM figure can produce very different forces on different centrifuge models, which is why lab protocols standardize on RCF (x g).

The playground analogy

Picture a merry-go-round with two riders holding onto the center pole at different distances: one close to the middle, one further out toward the edge. Both riders complete the same number of rotations per minute, but the rider further from the center feels a much stronger outward pull, because the actual force experienced depends on both spin speed and distance from the axis, not spin speed alone. Anyone who has ridden a playground merry-go-round and scooted toward the edge knows the feeling intensifies sharply the further out you go, even though the ride's rotational speed hasn't changed at all.

Particles in a spinning centrifuge tube experience exactly the same relationship. Those farther from the axis of rotation, meaning nearer the bottom of an angled or vertical tube in most rotor designs, feel more outward force at a given RPM. This is precisely why the RCF formula includes the radius term as a squared multiplier on speed but a linear multiplier on radius, and why simply reporting "10,000 RPM" without specifying the rotor's radius leaves out essential information about the actual separating force being applied to the sample.

A second, complementary analogy helps explain why different particles separate from each other rather than all moving together at the same rate. Imagine dropping a steel ball bearing and a ping-pong ball into a swimming pool at the same moment. Both experience the same gravitational pull, but the denser steel ball sinks quickly while the less dense ping-pong ball barely moves, or even floats. A centrifuge does the same thing to particles suspended in a tube, just with a far stronger "gravity" substitute and a fluid medium (often water, saline, or a sugar solution) standing in for the pool.

Why density, not size alone, determines separation

A common misconception is that centrifuges separate purely by particle size, the way a sieve does. What actually governs settling velocity in a centrifugal field, following principles related to Stokes' Law, the physics equation describing the drag force on a small sphere moving through a viscous fluid, is a combination of three factors: particle density relative to the surrounding medium, particle size, and the viscosity of that medium.

Two particles of identical size but different density will separate from each other in a centrifuge, because the denser one experiences more net force and moves faster toward the bottom. Two particles of identical density but different size will also separate at different rates, with the larger particle typically settling faster because it presents proportionally less surface drag relative to its mass, but in a density-gradient setup, both will eventually reach the same equilibrium position once they hit a layer matching their own density, regardless of their original size.

This distinction matters directly in blood separation. Red blood cells sediment fastest because they're the densest cellular component in whole blood, at roughly 1.09 to 1.10 grams per milliliter. White blood cells and platelets follow, forming the thin buffy coat, because they're slightly less dense. Blood plasma, the liquid component with a density close to 1.025 grams per milliliter, remains on top, floating above the pelleted cells. None of this depends primarily on the size of these components; a platelet is far smaller than a red blood cell, yet it's the density difference, not the size difference, that dictates their relative final positions in the tube.

Viscosity, the third variable, is why some separation protocols use sugar or salt solutions of graded concentration rather than plain water or saline as the medium. A denser, more viscous medium slows the settling of all particles proportionally, which paradoxically makes fine separations easier to control, because it stretches out the time window during which particles of similar but not identical density are still moving at meaningfully different rates.

Do centrifuges separate mixtures by particle size or by density?

Primarily by density, though particle size and medium viscosity also affect settling speed. Two particles of the same size but different density will separate from each other, while density-gradient centrifugation specifically exploits this relationship by letting particles migrate through a graded medium until each reaches a layer matching its own density, then stop moving entirely.

Two centrifuge strategies: pelleting versus density-gradient

Laboratories generally use one of two distinct approaches, depending on what needs to be isolated and how pure the result needs to be.

Method	How it works	Common application
Differential (pelleting) centrifugation	Spins sample until denser components form a solid pellet at the tube bottom; lighter material stays suspended in the supernatant liquid above	Separating blood cells from plasma; isolating cell organelles in successive stages of increasing speed
Density-gradient centrifugation	Sample is layered onto a pre-formed gradient of increasing density, often built from sucrose or cesium chloride solutions; particles migrate until they reach a layer matching their own density, then stop	Purifying viruses; separating DNA fragments by density; isolating specific cell organelles with high precision
Isopycnic (equilibrium) gradient	A variant of density-gradient centrifugation run long enough that every particle reaches its exact density-matched equilibrium point, regardless of starting position	Determining a molecule's buoyant density precisely, as in the Meselson–Stahl DNA replication experiment

Differential centrifugation is fast and simple but yields cruder separations; gradient methods take longer but resolve particles that differ by only small density fractions.

Differential centrifugation, sometimes called pelleting, is the workhorse technique. A researcher isolating cell organelles typically runs a series of spins at successively higher speeds: a low-speed spin first pellets whole cells and large debris, a moderate-speed spin next pellets mitochondria and larger organelles, and a high-speed spin finally pellets smaller structures like ribosomes. Each step discards the pellet or keeps the supernatant, depending on which fraction is needed, and moves the remaining material to the next, faster spin.

Density-gradient centrifugation trades speed of execution for resolving power. Rather than simply driving everything toward the bottom of the tube, this method suspends particles within, or layers them on top of, a pre-formed gradient where density increases gradually from top to bottom. As the rotor spins, each particle migrates through the gradient until it reaches a point where the surrounding medium's density matches its own, at which point the net force on it drops to essentially zero and it stops moving. Run long enough, this becomes an isopycnic (equal-density) separation, meaning the final position of a particle is determined entirely by its buoyant density, independent of its size, shape, or how long the run continues past equilibrium.

Why does spinning a mixture separate it by density instead of just mixing it more?

Centrifugal force acts on particles in proportion to their mass and their density relative to the surrounding fluid, pushing denser particles outward, toward the tube bottom in a swinging-bucket rotor, faster than less dense ones. Because this force scales with the density difference between a particle and its surrounding medium rather than acting uniformly on everything, particles of different density separate into distinct layers rather than becoming more thoroughly mixed together.

A field guide to centrifuge types

Not every centrifuge does the same job, and the differences in scale are dramatic. A clinical lab's benchtop unit and an industrial gas centrifuge cascade both rely on the same underlying physics, but they differ enormously in the forces they generate, the materials they process, and the precision of separation they achieve.

Centrifuge type	Typical RCF / speed	Sample form	Primary use case
Clinical/lab benchtop centrifuge	1,000–3,000 x g	Liquid (blood, urine, cell suspensions)	Separating serum or plasma from blood cells in diagnostic testing
High-speed refrigerated centrifuge	Up to ~30,000 x g	Liquid (cell lysates, organelle suspensions)	Pelleting bacteria, cell debris, or organelles for research
Preparative/analytical ultracentrifuge	Up to ~1,000,000 x g	Liquid (proteins, viruses, macromolecules)	Purifying viruses and proteins; measuring molecular weight and shape via sedimentation
Gas centrifuge (uranium enrichment)	~50,000–70,000 RPM, ~770 m/s peripheral speed (Urenco TC-21)	Gaseous uranium hexafluoride (UF6)	Separating uranium-235 from uranium-238 isotopes by mass difference for nuclear fuel

Force ranges and speeds are approximate and vary by manufacturer and rotor design; gas centrifuge figures are drawn from published Urenco specifications.

The clinical and research categories separate mixtures on the basis of density, exactly as described above. The gas centrifuge used in uranium enrichment works on a related but distinct principle: because uranium-235 and uranium-238 are chemically identical isotopes differing only in a few neutrons of mass, they cannot be separated chemically. Instead, uranium is converted into gaseous uranium hexafluoride and spun in a rotor at extremely high peripheral speed, where the very slight mass difference between the two isotopes causes the heavier uranium-238 gas to concentrate marginally more toward the rotor's outer wall while the lighter uranium-235 concentrates marginally more near the center. A single spin achieves only a tiny enrichment; commercial facilities cascade thousands of centrifuges in series to build up a usable concentration of uranium-235 for civilian reactor fuel. This is a matter of publicly documented nuclear engineering, monitored internationally: the International Atomic Energy Agency (IAEA) inspects enrichment facilities worldwide to verify uranium enrichment activity, whatever its stated purpose, complies with international safeguards.

Where centrifugation became foundational science

Theodor Svedberg developed the analytical ultracentrifuge in Uppsala, Sweden, during the 1920s, working with collaborators including J. Burton Nichols, specifically to study the size and shape of large molecules like proteins by observing their sedimentation behavior under intense centrifugal force. His first working ultracentrifuge, completed in 1924, generated forces up to roughly 5,000 times Earth's gravity, a striking figure at the time, though modest compared to the forces modern ultracentrifuges can produce. Svedberg used the device to determine, with unprecedented precision, the molecular weights of complex proteins such as hemoglobin, work that earned him the 1926 Nobel Prize in Chemistry for his studies of colloids and his invention of the ultracentrifuge. The sedimentation-rate unit still used today in biochemistry, the svedberg (S), is named in his honor; it's the same unit that gives ribosomal RNA subunits their familiar names, such as the 30S and 50S subunits of bacterial ribosomes.

Density-gradient ultracentrifugation later became central to one of molecular biology's landmark experiments. The 1958 Meselson–Stahl experiment used a cesium chloride density gradient, spun at extremely high speed for extended periods to reach isopycnic equilibrium, to directly track how DNA density changed across successive generations of replicating bacteria grown first in a heavy nitrogen isotope and then switched to a normal, lighter one. The resulting density patterns confirmed that DNA replicates semi-conservatively, exactly as predicted by James Watson and Francis Crick's 1953 double-helix model, with each new DNA molecule retaining one original strand and synthesizing one new strand. The experiment is frequently cited in molecular biology courses as one of the cleanest, most elegant confirmations of a major theoretical prediction in the field's history, and it would have been essentially impossible without the resolving power of density-gradient centrifugation.

Who developed the scientific use of ultracentrifugation?

Theodor Svedberg developed the analytical ultracentrifuge in the 1920s in Uppsala, Sweden, to study the size and shape of large molecules like proteins based on their sedimentation behavior under intense centrifugal force. His work earned him the 1926 Nobel Prize in Chemistry, and the sedimentation-rate unit "svedberg" is named after him.

How was centrifugation used to prove how DNA replicates?

The 1958 Meselson–Stahl experiment used cesium chloride density-gradient ultracentrifugation to track how DNA density changed across generations of replicating bacteria. The resulting density pattern directly confirmed the semi-conservative replication model proposed by Watson and Crick's 1953 double-helix structure, in which each new DNA molecule retains one original strand.

What people get wrong about centrifuges

A common mistake is assuming centrifuges work by "throwing out" lighter material, the way a spinning basin flings water droplets outward off its rim. In reality, all particles in the sample experience the same centrifugal acceleration at a given point in the rotor; it's the difference in density relative to the surrounding medium, not an active expulsion mechanism that somehow "selects" lighter material, that causes different particles to migrate at different rates and end up in different final positions.

Another mistake, common enough that lab manuals explicitly warn against it, is treating RPM as an adequate substitute for RCF when comparing protocols across different lab equipment. Because RCF depends on rotor radius as well as rotational speed, following an RPM value published in a paper or protocol that used a different centrifuge model can produce a meaningfully different actual force on the sample, potentially under-processing a separation that needed more force, or over-processing one that needed gentler handling to avoid damaging fragile particles like intact organelles or cell membranes.

A third misunderstanding conflates size-based filtration with density-based centrifugal separation. Filtration through a membrane or mesh sorts particles strictly by whether they're physically small enough to pass through pores of a given size; a centrifuge sorts particles by how fast they migrate through a fluid under an applied force, a rate governed jointly by density, size, and medium viscosity. It's entirely possible for a smaller, denser particle to settle faster than a larger, less dense one in a centrifuge, an outcome that would be impossible in simple pore-size filtration.

Finally, some assume that spinning a sample longer always improves separation. Past the point where particles reach a stable pellet or, in gradient work, their equilibrium density position, additional spin time adds no further separation and can in some cases degrade sample quality through prolonged mechanical stress or, for biological samples, extended exposure to non-physiological pressure and temperature conditions inside the rotor.

The takeaway

A centrifuge does not add anything new to a mixture; it amplifies an existing physical property, density difference, into a fast, controllable separation by generating centrifugal forces that can exceed Earth's gravity by a factor of a million or more in modern ultracentrifuges. That amplification, first harnessed rigorously by Theodor Svedberg to study protein structure in the 1920s and later used to confirm how DNA itself replicates in the Meselson–Stahl experiment, remains one of the most direct demonstrations in science of how spinning something fast enough turns a slow natural process, sedimentation, into a routine and precisely controllable laboratory tool. The same basic physics, scaled up enormously and applied to a gas rather than a liquid, also underlies uranium isotope enrichment, a reminder that a single principle, force proportional to mass and radius, can span from a hospital blood draw to industrial-scale nuclear fuel production.

---

Read next:

• How Crystals Grow From Solutions
• How Mass Spectrometry Identifies Molecules
• Why Blood Types Matter for Transfusions

FAQ

Q: What is relative centrifugal force (RCF) and why does it matter more than RPM?

RCF measures the actual force a sample experiences as a multiple of Earth's gravity, and it depends on both rotational speed and the radius from the spin axis to the sample. Reporting only RPM omits the radius term, meaning the same RPM can produce very different forces on different centrifuge models.

Q: Do centrifuges separate mixtures by particle size or by density?

Primarily by density, though size and medium viscosity also affect settling speed. Two particles of the same size but different density will separate; density-gradient centrifugation specifically exploits this by letting particles migrate until they reach a layer matching their own density, then stop.

Q: How does centrifugation separate blood into its components?

Spinning whole blood generates centrifugal force that moves the densest component, red blood cells, to the bottom fastest, followed by white blood cells and platelets forming a thin buffy coat, while the least dense component, plasma, remains on top.

Q: Who developed the scientific use of ultracentrifugation?

Theodor Svedberg developed the analytical ultracentrifuge in the 1920s to study the size and shape of large molecules like proteins based on their sedimentation behavior, earning the 1926 Nobel Prize in Chemistry for this work.

Q: How was centrifugation used to prove how DNA replicates?

The 1958 Meselson-Stahl experiment used cesium chloride density-gradient ultracentrifugation to track DNA density changes across generations of replicating bacteria, directly confirming the semi-conservative replication model proposed by Watson and Crick's 1953 double-helix structure.

Q: How does a gas centrifuge enrich uranium?

A gas centrifuge spins gaseous uranium hexafluoride at extremely high peripheral speed, exploiting the tiny mass difference between uranium-235 and uranium-238 isotopes, so the heavier isotope concentrates slightly more toward the rotor's outer wall. Because a single spin achieves only minimal enrichment, commercial facilities cascade thousands of centrifuges in series, and operations are monitored internationally by the IAEA.