The Phospholipid Bilayer
Drop a spoonful of phospholipid into water. Within seconds it organizes itself into a sheet. That sheet is exactly two molecules thick. It seals at its edges into a closed compartment. No enzyme directs this. No energy from ATP is spent. The molecules simply obey the physical chemistry of their own shape. In doing so they build the single most important structure in cell biology: the plasma membrane. Every living cell is bounded by a phospholipid bilayer. This holds true from the simplest bacterium to a human neuron a meter long. This thin film is only about 4-5 nanometers thick. Yet it does something remarkable: it separates two aqueous compartments. It does this while itself being built from molecules that "like" water at one end and "hate" it at the other. Understanding how such a structure forms is essential. So is understanding why it self-seals. So is understanding why it selectively resists the passage of most solutes. So is understanding why its two faces are chemically distinct. These questions form the foundation for everything else in membrane biology — from ion channels to vesicle trafficking to signal transduction. This article addresses three tightly linked questions. First, the mechanism: how does the amphipathic shape of a phospholipid drive spontaneous, thermodynamically favorable self-assembly? Why does it form into a bilayer sheet rather than some other structure? Second, the consequence: why does this bilayer form an effective permeability barrier to ions and polar molecules? Why does it still let small nonpolar and uncharged polar molecules pass? Third, an often under-appreciated feature: why are the two leaflets of a real cell membrane compositionally different? This asymmetry is generated, maintained, and functionally exploited by the cell. The story here sets up almost everything that follows in membrane biology. Once you understand why the bilayer forms, and what it does and does not let through, you are ready for more. You are ready for Simple Diffusion Across Membranes. You are ready for the idea that a "fluid" bilayer studded with proteins constitutes a mosaic (the Fluid-Mosaic Model of Membranes). And you are ready for how cells actively tune the physical properties of this sheet through Membrane Fluidity and Its Regulation.
- Phospholipids are amphipathic: a hydrophilic head and two hydrophobic fatty acid tails joined via a glycerol backbone.
- In water, this shape drives spontaneous self-assembly into a bilayer, minimizing hydrophobic exposure and maximizing water's entropy.
- Bilayers are thermodynamically favored to close into seamless vesicles, eliminating exposed hydrophobic edges — this is why membranes self-seal.
- The hydrophobic core (roughly 3 nm) blocks ions and polar solutes but lets O₂, CO₂, and small uncharged molecules diffuse through.
- The two leaflets differ in lipid and protein composition — an asymmetry actively generated by flippases, floppases, and scramblases.
- Classic experiments (Gorter–Grendel, Langmuir monolayers, Bangham's liposomes, freeze-fracture EM) established both the bilayer structure and its permeability rules.
A phospholipid bilayer is not built by the cell so much as it builds itself — its sheet-like, self-sealing, selectively permeable form is the direct physical consequence of amphipathic molecules minimizing contact between their hydrophobic tails and surrounding water.
The core idea
A phospholipid is amphipathic — it has a hydrophilic ("water-loving") phosphate-containing head group and two hydrophobic ("water-fearing") fatty acid tails, connected through a three-carbon glycerol backbone. Placed in water, this molecule faces a physical problem: the head is happy to contact water, but the tails are not. Water molecules around an exposed hydrocarbon tail must arrange themselves into an ordered cage-like shell (to maintain hydrogen bonding among themselves while excluding the nonpolar solute), which lowers the entropy of the surrounding water. Because a system tends toward the arrangement of highest total entropy, the cell (or the test tube) does better, thermodynamically, when hydrophobic surfaces are hidden away from water rather than exposed to it. This is the hydrophobic effect, and it is the single physical principle that explains bilayer formation, protein folding cores, and much else in cell biology. Phospholipids solve the problem by clustering so that their tails point inward, away from water, and their heads point outward, toward water — either as a micelle (a ball, favored by single-tailed lipids with a large head-to-tail volume ratio) or, for lipids with two roughly cylindrical tails like phosphoglycerides and sphingomyelin, as a flat, two-molecule-thick bilayer, with head groups facing two aqueous surfaces and a hydrophobic core sandwiched between them. The bilayer is not held together by covalent bonds between lipids; it is a noncovalent, self-assembling structure stabilized by the collective hydrophobic effect together with van der Waals attractions between tails and electrostatic/hydrogen-bond interactions among head groups. This is why membranes can be dissolved by detergents, why lipids diffuse laterally within the plane of the membrane, and why isolated phospholipids will spontaneously re-form bilayers if mechanically disrupted. A flat bilayer sheet still has a problem: at its edge, hydrophobic tails would be exposed to water. The thermodynamically preferred solution is for the sheet to curve around and close on itself, eliminating all free edges — producing a sealed, continuous compartment. This is why phospholipids in water spontaneously form closed vesicles (liposomes) rather than infinite flat sheets, and it is the deep physical reason every cell membrane is a closed surface with an inside and an outside, never a sheet with a raw edge.
Mechanism
Step 1 — Molecular geometry sets the assembly. The overall shape of a lipid molecule, not just its chemistry, determines what it assembles into. Single-chain lipids (like many detergents and free fatty acids) are cone-shaped — a bulky head atop one thin tail — and pack into small spherical micelles (diameter often 3-10 nm, aggregation numbers of tens of molecules). Double-chain glycerophospholipids and sphingolipids are closer to cylindrical — the cross-sectional area of the two tails roughly matches that of the head group — so they pack most efficiently side by side into flat sheets: bilayers. This geometric packing rule (formalized as the "critical packing parameter," P = v/(a₀·lc), where v is tail volume, a₀ head-group area, lc tail length) predicts micelle formation when P < ⅓, bilayers when P is near 1, and inverted structures when P > 1.
Step 2 — Nucleation and growth. When phospholipids are introduced into water above a threshold concentration (the critical micelle/bilayer concentration, typically extremely low, ~10⁻¹⁰ M for long-chain phospholipids because their tails are so hydrophobic), individual monomers rapidly aggregate. Small bilayer fragments (discs) form first, exposing a hydrophobic rim. Because that exposed edge is energetically costly, further lipid addition preferentially occurs at the rim, growing the disc — until, past a certain size, closing the disc into a sealed vesicle (eliminating the edge entirely) becomes energetically cheaper than continuing to grow a flat sheet with an ever-longer perimeter. This self-sealing tendency is why damaged membranes reseal spontaneously and why phospholipid mixtures added to water form vesicles (liposomes) without any biological machinery.
Step 3 — The permeability barrier. The finished bilayer presents a roughly 3 nm-thick hydrophobic core (the fatty acid tails) flanked by two thin polar head-group zones, for a total thickness around 4-5 nm. Any molecule crossing the membrane by simple diffusion must transiently leave the aqueous phase and dissolve into this hydrophobic core — an energetically unfavorable step for anything polar or charged. Permeability, therefore, correlates strongly with a solute's oil/water partition coefficient and, for charged species, with the near-impossibility of stripping away a hydration shell:
| Solute class | Example | Relative permeability |
|---|---|---|
| Small nonpolar gases | O₂, CO₂, N₂ | Very high — cross almost freely |
| Small uncharged polar | H₂O, ethanol, urea, glycerol | Moderate — water crosses surprisingly fast (partly via bilayer, partly via aquaporins in real membranes) |
| Large uncharged polar | glucose, sucrose | Very low |
| Ions | Na⁺, K⁺, Ca²⁺, Cl⁻ | Extremely low (rates ~10⁹-fold slower than water) |
This is why a pure bilayer is an excellent barrier to ions and sugars but a poor barrier to gases — a fact essential for Simple Diffusion Across Membranes, and the reason real cell membranes need dedicated transport proteins (channels, carriers, pumps) to move ions and nutrients at physiologically useful rates. The bilayer supplies the barrier; the proteins supply the controlled leaks.
Step 4 — Building and maintaining leaflet asymmetry. In a living cell, the two leaflets of the bilayer are compositionally distinct — this is membrane asymmetry, and it is not a passive accident but an actively maintained state. In the human erythrocyte plasma membrane, for example, phosphatidylcholine (PC) and sphingomyelin are concentrated in the outer (exoplasmic) leaflet, while phosphatidylserine (PS) and phosphatidylethanolamine (PE) are concentrated in the inner (cytosolic) leaflet. Because spontaneous "flip-flop" of a phospholipid from one leaflet to the other requires its polar head group to pass through the hydrophobic core — an enormously unfavorable event — flip-flop is intrinsically very slow (half-times of hours to days for an unassisted phospholipid). Cells therefore generate and maintain asymmetry using three classes of ATP-dependent or ATP-independent lipid translocases:
- Flippases actively move specific phospholipids (notably PS and PE) from the outer to the inner leaflet, using ATP hydrolysis.
- Floppases move lipids from the inner to the outer leaflet, also ATP-dependent.
- Scramblases move lipids bidirectionally, nonselectively, without ATP — activated (for example) by rising cytosolic Ca²⁺.
This machinery matters physiologically: in a healthy cell PS stays hidden in the inner leaflet. When a cell undergoes apoptosis, scramblase activation and flippase inhibition let PS appear on the cell surface, where it acts as an "eat me" signal recognized by macrophage receptors — a direct functional payoff of leaflet asymmetry.
The evidence
Gorter and Grendel (1925) — the first evidence for a bilayer. Dutch scientists Evert Gorter and F. Grendel extracted the total lipid from a known number of human and animal erythrocytes (red blood cells, chosen because they have a plasma membrane and no internal membranes, simplifying the extraction). They spread this lipid as a monomolecular film on a Langmuir trough — a water tray with a movable barrier that compresses a floating lipid film until it packs into a single continuous monolayer, whose area can then be measured precisely. They compared the area of this monolayer to the calculated surface area of the intact red blood cells. They found the lipid monolayer area was roughly twice the calculated cell surface area. Their conclusion: the plasma membrane must consist of a bilayer — two molecules thick — so that unfolded onto a single layer, the lipid covers double the cell's surface. (Their original calculations contained compensating errors — they underestimated cell surface area and used an imperfect lipid extraction — but by remarkable good fortune the errors partly canceled, and the 2:1 ratio held up as later work corrected the individual numbers.) This was the first direct experimental argument for a bilayer structure of the cell membrane, decades before it could be directly visualized.
Freeze-fracture electron microscopy — visualizing the bilayer directly. In the 1950s-60s, the development of freeze-fracture EM allowed membranes to be frozen and physically split along the weak plane between the two lipid leaflets. This technique revealed a smooth fracture face studded with particles (later identified as embedded membrane proteins) and directly confirmed that the membrane splits into two distinct leaflets — visual, structural proof of the bilayer architecture that Gorter and Grendel had inferred indirectly. Standard transmission EM of stained, sectioned membranes (osmium tetroxide staining reacting with polar head groups) further showed the classic "railroad track" appearance: two dark parallel lines (the stained polar heads) separated by a light gap (the unstained hydrophobic core) — a pattern seen in essentially every cell membrane examined, supporting the bilayer as a near-universal structure.
Bangham (1965) — liposomes prove spontaneous self-assembly. Alec Bangham and R. W. Horne showed that when phospholipids are simply dried down and then dispersed in water with mild agitation, they spontaneously form closed, multilamellar vesicles ("liposomes") visible by electron microscopy as concentric bilayer shells enclosing an aqueous interior. Crucially, this happened with purified lipid alone — no protein, no cellular machinery, no added energy beyond simple mechanical mixing. This was the direct experimental demonstration that bilayer formation, closure into sealed vesicles, and enclosure of an aqueous compartment are intrinsic, spontaneous properties of phospholipid molecules in water, exactly as predicted by the hydrophobic-effect argument. Liposomes made this way went on to become both a research tool for measuring membrane permeability to different solutes, and (much later) a practical drug-delivery vehicle — the lipid nanoparticles used in mRNA vaccines are direct intellectual descendants of Bangham's liposomes.
Measuring permeability directly — the basis of the permeability table. Using synthetic liposomes of defined, protein-free lipid composition, researchers (from the 1960s onward) could load them with radiolabeled or fluorescent solutes, or with pH- or ion-sensitive dyes, and directly measure the rate at which different molecules leaked out or diffused in across a pure bilayer with no protein transporters present. These experiments established the empirical permeability ranking (gases > small uncharged polar molecules ≫ large polar molecules > ions) summarized above, and they showed that ions cross a pure lipid bilayer at rates many orders of magnitude slower than water — direct proof that the bilayer itself, independent of any protein, constitutes the primary permeability barrier of the cell, while real biological membranes need channels and transporters (the subject of later articles) to achieve physiologically necessary ion fluxes.
Worked examples
Example 1 — Reconstructing the Gorter–Grendel logic with modern numbers. Suppose you extract the total phospholipid from 1 × 10¹⁰ human erythrocytes and spread it as a compressed monolayer on a Langmuir trough, measuring a total monolayer area of 9,000 cm². Each erythrocyte is a biconcave disc with an estimated total surface area (both faces plus rim) of about 1.4 × 10⁻⁶ cm² (140 μm²). Calculate the ratio of monolayer area to total cell surface area and interpret it. Total cell surface area = 1 × 10¹⁰ cells × 1.4 × 10⁻⁶ cm²/cell = 1.4 × 10⁴ cm² = 14,000 cm². Ratio = monolayer area / cell surface area = 9,000 / 14,000 ≈ 0.64. At first glance this ratio (0.64) looks like it contradicts the expected 2:1 result — but notice the direction: here the monolayer area is smaller than the cell surface area, the opposite of Gorter and Grendel's finding. This should immediately flag a problem: either the lipid extraction was incomplete (some lipid was lost, so the true monolayer area is underestimated), or the assumed cell surface area is too large. Historically, this is exactly the kind of check that matters: Gorter and Grendel's original raw ratio was actually about 2.0, but only because their lipid extraction underestimated total lipid and their surface-area estimate (using a sphere approximation) underestimated true area — two errors that happened to cancel. The pedagogical lesson: a ratio consistent with a bilayer (≈2) is the diagnostic signature to look for, and any value far from 2 in either direction should prompt you to scrutinize the extraction efficiency and the geometric model of the cell before concluding the bilayer hypothesis is wrong.
Example 2 — Predicting relative permeability from partition coefficients. A liposome permeability assay measures the permeability coefficient P (in cm/s) for four solutes crossing a pure phosphatidylcholine bilayer:
| Solute | Oil/water partition coefficient (relative) | Measured P (cm/s) |
|---|---|---|
| Tryptophan (charged/polar, large) | very low | ~10⁻¹⁴ |
| Glucose | low | ~10⁻¹⁰ |
| Glycerol | moderate | ~10⁻⁷ |
| Water | high (small, uncharged) | ~10⁻³ |
A student is asked: if a drug candidate is a small, uncharged, moderately lipid-soluble molecule (similar partition behavior to glycerol, molecular weight ~150 Da), estimate whether it will cross a lipid bilayer fast enough to be orally absorbed passively, without a transporter, on the timescale of intestinal transit (minutes to a few hours). Reasoning: with P ≈ 10⁻⁷ cm/s (comparable to glycerol) and a typical enterocyte membrane thickness/geometry giving a diffusive flux J = P × ΔC, a molecule with this permeability coefficient crosses a cell membrane with a characteristic time (t ≈ 1/P, dimensionally, scaled by relevant surface-to-volume considerations) on the order of minutes to tens of minutes for reasonable concentration gradients — fast enough for meaningful passive absorption over an intestinal transit time of several hours. By contrast, glucose (P ≈ 10⁻¹⁰ cm/s) would take far too long by passive bilayer diffusion alone to be absorbed at physiologically necessary rates — correctly predicting that glucose absorption in the gut depends on dedicated transporters (SGLT1, GLUT2), not on diffusion through the lipid bilayer. This is precisely why membrane permeability coefficients, measured on pure liposomes, are used in pharmacology (Lipinski's Rule of Five reasoning) to predict which drugs can be delivered orally versus which require injection or transporter-mediated uptake.
Common misconceptions & traps
| Misconception | Why it's wrong |
|---|---|
| "The bilayer is held together by strong covalent bonds, like a polymer." | It is a noncovalent assembly, stabilized by the hydrophobic effect and van der Waals forces — which is exactly why it is fluid, self-healing, and dissolvable by detergents. |
| "Cells build the bilayer using enzymes, the way they build proteins." | Enzymes synthesize individual phospholipid molecules, but the bilayer sheet assembles spontaneously and physically once enough amphipathic lipid is present — self-assembly, not template-directed construction. |
| "A membrane is equally permeable to all small molecules." | Permeability depends on polarity and charge, not just size: O₂ (nonpolar) crosses easily; Na⁺ (small but charged, hydrated) crosses roughly a billion-fold slower. |
| "Both leaflets of a membrane have identical lipid composition." | Real membranes are asymmetric — e.g., PS and PE are concentrated on the cytosolic face of the human erythrocyte membrane, actively maintained by flippases/floppases/scramblases. |
| "Phospholipids flip between leaflets easily to maintain equilibrium." | Unassisted flip-flop is extremely slow (hours to days) because it forces a polar head group through the hydrophobic core; asymmetry persists because of this kinetic barrier plus active enzymes. |
| "Water cannot cross a lipid bilayer because it's polar." | Water is small enough and lacks net charge, so it actually crosses pure bilayers fairly quickly by simple diffusion — real membranes further speed this with aquaporin channels. |
The scientist behind it
Evert Gorter was a Dutch pediatrician and physiologist working in Leiden in the 1920s, whose interest in blood chemistry led him to a deceptively simple question: how much lipid does a red blood cell membrane actually contain, and how is it arranged? Working with his assistant F. Grendel, he adapted a technique developed a decade earlier by Irving Langmuir for studying oil films on water — the Langmuir trough — to precisely measure the area a fixed quantity of lipid occupies as a single molecular layer. Applying it to erythrocyte lipid extracts was a clever repurposing of a physical-chemistry tool for a cell-biological question, at a time (well before electron microscopy existed) when no one could simply look at a cell membrane and see its molecular structure. Gorter and Grendel's 1925 paper, "On Bimolecular Layers of Lipoids on the Chromocytes of the Blood," is now recognized as the founding experimental argument for the bilayer model of the cell membrane, even though the tools to see the structure directly would not exist for another three decades.
Origins · 1 minEvert Gorter and how this was discovered
A physician-scientist who used a physical-chemistry technique (the Langmuir trough) to make the first quantitative, indirect case that the cell membrane is a lipid bilayer — years before any microscope could resolve the structure.
Where this matters today
The physics of bilayer self-assembly, permeability, and asymmetry is not an abstract lesson — it underpins entire branches of modern medicine and biotechnology. Liposomal and lipid-nanoparticle drug delivery directly exploits Bangham's discovery: doxorubicin (a chemotherapy drug) is packaged in liposomes (Doxil) to reduce cardiotoxicity, and the mRNA COVID-19 vaccines depend on lipid nanoparticles whose self-assembly around mRNA follows exactly the same amphipathic-lipid physics described here. Anesthesiology rests on the century-old Meyer-Overton correlation, which showed that an anesthetic's potency correlates with its oil/water partition coefficient — direct evidence that general anesthetics act by dissolving into and perturbing the lipid bilayer (and/or associated membrane proteins), a correlation only intelligible once you understand bilayer permeability physics. Diagnostics of apoptosis use fluorescently labeled Annexin V, a protein that binds phosphatidylserine — clinically and in research this detects apoptotic cells precisely because PS, normally hidden on the inner leaflet by flippase activity, becomes exposed on the outer leaflet when scramblases activate during programmed cell death, a direct application of leaflet-asymmetry biology. Antifungal and antiparasitic drugs such as amphotericin B work by binding ergosterol in fungal membranes and forming pores that destroy the bilayer's barrier function selectively, sparing host cells whose membranes contain cholesterol instead. And in agricultural and food science, understanding phospholipid phase behavior underlies everything from the stability of milk emulsions to the design of biodegradable liposomal pesticide carriers. Every one of these applications is, at bottom, an engineering exploitation of the same simple physical fact: amphipathic lipids in water spontaneously build a selectively permeable, self-sealing, asymmetric sheet.
Ask the tutor: How do flippases, floppases, and scramblases actually recognize and move specific phospholipid head groups across the hydrophobic core?
Why don't phospholipids just dissolve completely in water like sugar does?
Sugar is fully hydrophilic, so individual sugar molecules can be surrounded by water and dissolve as separate molecules. A phospholipid is amphipathic: its head is hydrophilic, but its two fatty acid tails are strongly hydrophobic. Full dissolution as isolated monomers would force water to form ordered, entropy-lowering shells around every exposed tail — energetically costly. Instead, phospholipids minimize this cost by aggregating so tails cluster together, away from water, while heads remain in contact with water. This aggregation, not true molecular dissolution, is what produces micelles, bilayers, and vesicles, and it happens at extremely low concentrations because the hydrophobic effect is a strong driving force.
Is the bilayer held together by chemical bonds between lipid molecules?
No. The phospholipid bilayer is a noncovalent assembly. Individual lipid molecules are not bonded to their neighbors; they are held in place collectively by the hydrophobic effect (the entropic penalty of exposing tails to water), supplemented by weak van der Waals attractions between adjacent hydrocarbon tails and hydrogen bonding/electrostatic interactions among head groups. Because there are no covalent bonds between lipids, individual molecules can diffuse laterally within a leaflet, the membrane can flex and re-seal after mechanical disruption, and detergents (which insert into the bilayer and disrupt these noncovalent interactions) can dissolve the membrane entirely — none of which would be possible if lipids were covalently cross-linked.
If ions can barely cross the bilayer, how do cells manage ion gradients like the resting membrane potential?
Precisely because the pure lipid bilayer is such an excellent barrier to ions, cells can build and maintain steep, biologically useful ion gradients across it (for example, the roughly 10-fold higher K⁺ concentration inside versus outside a typical animal cell). If the bilayer were leaky to ions, these gradients would collapse within seconds. Controlled movement of ions across this otherwise near-impermeable barrier is instead handled by dedicated integral membrane proteins — ion channels (which open briefly to let ions flow down their gradient) and active pumps (like the Na⁺/K⁺-ATPase, which uses ATP to move ions against their gradient). The bilayer's low intrinsic ion permeability is what makes these proteins functionally meaningful.
What is membrane asymmetry, and why does it matter for a dying cell?
Membrane asymmetry means the two leaflets of a bilayer have different lipid compositions. In a healthy human red blood cell or most animal cells, phosphatidylserine (PS) and phosphatidylethanolamine are concentrated in the inner (cytosolic) leaflet, while phosphatidylcholine and sphingomyelin dominate the outer leaflet — maintained by ATP-dependent flippases pulling PS inward. During apoptosis (programmed cell death), rising intracellular calcium activates scramblases, which shuffle lipids nonselectively between leaflets, while flippase activity is shut down. The net effect is that PS becomes exposed on the cell's outer surface. Macrophages recognize surface PS as an "eat me" signal and phagocytose the dying cell, making membrane asymmetry (and its controlled loss) a key checkpoint in immune clearance of apoptotic cells.
How thick is the bilayer, and does its thickness matter biologically?
A typical biological lipid bilayer is about 4-5 nanometers thick overall, with a hydrophobic core of roughly 3 nm sandwiched between two thin polar head-group zones. This thickness is set by the length of the fatty acid tails and matters biologically because membrane proteins (such as ion channels) must span a hydrophobic stretch of matching length to sit stably in the membrane; mismatches between protein hydrophobic length and bilayer thickness can distort protein structure and function. Bilayer thickness also varies somewhat between organelles (the endoplasmic reticulum tends to be thinner than the plasma membrane), which cells use as one cue for sorting proteins to the correct membrane compartment.
Can a phospholipid bilayer form without any protein at all?
Yes, and this is one of the most important experimental facts in membrane biology. Alec Bangham showed in 1965 that purified phospholipid, simply dispersed in water with gentle agitation, spontaneously forms closed bilayer vesicles (liposomes) with no protein, no enzymes, and no external energy input beyond the mechanical mixing. This demonstrates that the bilayer's basic architecture — sheet formation, self-sealing into a closed compartment, and the resulting permeability barrier — is a direct, spontaneous consequence of lipid chemistry and physics alone. Proteins are essential for a real cell membrane's functions (transport, signaling, adhesion), but the bilayer's basic existence does not require them.
Why do general anesthetics correlate with lipid solubility rather than with a specific chemical structure?
The Meyer-Overton correlation, described around 1900, found that the potency of diverse general anesthetic gases correlates closely with their oil/water partition coefficient, regardless of otherwise unrelated chemical structures. This makes sense in light of bilayer physics: anesthetic potency tracks how readily a molecule partitions into and accumulates within the hydrophobic core of the lipid bilayer (and membrane-embedded proteins), rather than depending on binding a single specific receptor the way most drugs do. While modern anesthesiology recognizes that specific ion-channel and receptor interactions also matter, the century-old correlation remains a striking, still partly unresolved, piece of evidence for how deeply general anesthetics' action is tied to basic membrane lipid physics.