A single silicon atom has a covalent radius of about 117 picometers, or roughly 0.117 nanometers. That is a little over one ten-billionth of a meter. The number sounds impossibly small, but it turns out to be one of the most consequential measurements in modern technology and science. Silicon’s atomic dimensions sit at the heart of how we build computer chips, define fundamental units of measurement, and push the boundaries of nanotechnology.
What “Size” Actually Means for an Atom
Atoms do not have hard edges the way a billiard ball does. An atom is a nucleus surrounded by a cloud of electrons, and that cloud does not end sharply. It fades out. So when someone asks how big a silicon atom is, the answer depends on which definition of “size” you use.
The most common measure for atoms bonded to their neighbors is the covalent radius, which describes half the distance between two bonded atoms of the same element. For silicon, that is about 117 picometers. A picometer is a trillionth of a meter, so you would need to line up roughly eight or nine silicon atoms to span a single nanometer.
A second measure, the van der Waals radius, captures how close two unbonded silicon atoms can sit before their electron clouds start pushing each other away. That radius is larger, around 210 picometers. And in a crystal, where silicon atoms arrange themselves in a repeating diamond-cubic lattice, the key number is the lattice parameter: the length of one side of the repeating cube. For silicon that value is about 543 picometers, with eight atoms packed into each cube. These three numbers, covalent radius, van der Waals radius, and lattice parameter, describe different aspects of silicon’s size, and all three matter in different contexts.
How Scientists Measure Silicon’s Size with Absurd Precision
You might think that once you know an atom’s radius to a few decimal places, the job is done. But some applications demand far more precision than “about 543 picometers.” Scientists at the National Institute of Standards and Technology (NIST) and Germany’s Physikalisch-Technische Bundesanstalt (PTB) have measured the difference in lattice spacing between different silicon samples with a relative uncertainty of just one part in a hundred million.
1PubMed Central. Precision Comparison of the Lattice Parameters of Silicon MonocrystalsWhy bother with that kind of precision? Because those measurements fed directly into one of the biggest changes in measurement science in over a century: the redefinition of the kilogram. The old kilogram was a single platinum-iridium cylinder locked in a vault near Paris, and it was slowly changing mass just from surface contamination. The replacement approach involved counting atoms in a nearly perfect sphere of isotopically enriched silicon-28. By knowing the lattice spacing and the volume of the sphere, researchers could calculate exactly how many atoms the sphere contained and tie the kilogram to a fixed value of Avogadro’s number and Planck’s constant.
An international collaboration measured the Avogadro constant using this method and arrived at a value of 6.02214082 × 10²³ per mole, with a relative uncertainty of about three parts in a hundred million.2Metrologia. Counting the atoms in a 28Si crystal for a new kilogram definition That result became one of the key inputs for the 2019 redefinition of the kilogram. Silicon spheres continue to serve as practical tools for realizing the definitions of both the kilogram and the mole.3Comptes Rendus Physique. Silicon spheres for the future realization of the kilogram and the mole In short, the size of a silicon atom is now baked into the very system we use to weigh things.
Why Chip Manufacturers Care About Individual Atoms
The semiconductor industry has spent decades shrinking the features on silicon chips, and the shrinking has reached a point where individual atoms start to matter. In the early 2000s, transistor gate lengths sat at roughly 50 nanometers. Projections at the time anticipated that fabrication advances would push gate lengths below 10 nanometers within the following decade.4Science. Silicon device scaling to the sub-10-nm regime That prediction came true: today’s leading-edge chips feature gate lengths of just a few nanometers, which translates to only a few dozen silicon atoms across.
At these scales, every atom’s position counts. Researchers have demonstrated that in atom-scale silicon devices, changing a tunnel junction gap by as few as nine rows of silicon surface dimers (each row about 0.77 nanometers wide) changes the junction’s electrical resistance by more than four orders of magnitude. Moving from about seven dimer rows of separation to twelve shifts a device from smooth, linear electrical conductance into a regime where quantum mechanical tunneling dominates and individual electrons can be counted one at a time.5Nature / Communications Physics. Atomic-scale control of tunneling in donor-based devices That is a dramatic electrical change driven by shifting a gap by a distance comparable to the width of a single-digit number of silicon atoms.
This sensitivity is exactly why atomic-scale precision in silicon fabrication is not just an academic exercise. A misplaced atom or an uncontrolled impurity can make or break a transistor at modern scales.
Dopants and the Strain They Cause
Pure silicon is a mediocre conductor. To make it useful for electronics, manufacturers introduce tiny amounts of other elements, called dopants, into the crystal. Boron, phosphorus, and arsenic are common choices. Each dopant atom replaces a silicon atom in the lattice, and because dopant atoms are a different size, they distort the lattice around them.
The distortion is measurable. When boron ions are implanted into silicon and then locked in place with a laser anneal, the resulting lattice strain near the surface can reach about −5.8 × 10⁻³, meaning the lattice is slightly compressed because boron atoms are smaller than silicon atoms. That strain decreases rapidly beyond a depth of about 0.2 micrometers, tracking the boron concentration profile.6Journal of Applied Physics. X-ray study of lattice strain in boron implanted laser annealed silicon Engineers deliberately use this strain to improve transistor performance: compressing or stretching the silicon lattice in controlled ways changes how easily electrons move through it, boosting the speed of the chip.
The Ultra-Thin Oxide Layer
One reason silicon dominates electronics is that it forms a high-quality oxide, silicon dioxide (SiO₂), when exposed to oxygen. This oxide acts as an insulator in transistors. But at the boundary between the silicon crystal and its oxide, there is a transition zone only a few atoms thick where the material is neither pure silicon nor pure oxide. That transition layer has been studied closely because its electronic properties determine how well a transistor can switch on and off.
Measurements of the SiO₂/Si interface show that the electronic boundary, the point where the oxide’s insulating energy barrier is fully established, sits about 0.51 to 0.61 nanometers away from the structural boundary, depending on which crystal face the oxide sits on.7Applied Surface Science. Chemical and electronic structure of SiO2/Si interfacial transition layer That is a distance of roughly four to five silicon atoms. In modern transistors where the entire oxide layer might be only a nanometer or two thick, that transition zone makes up a significant fraction of the insulator. Understanding it at the atomic level is not optional; it is central to making reliable devices.
What Happens When Silicon Gets Truly Tiny
Bulk silicon has an indirect bandgap of about 1.1 electron volts, which is why it absorbs infrared light well but does not glow visibly. Shrink a piece of silicon down to a nanocrystal just a few nanometers across, and the rules change. Quantum confinement kicks in: the electrons and holes are trapped in such a small space that their allowed energies shift, widening the effective bandgap and pushing it into the visible range.
Experiments on surface-oxidized silicon nanocrystals showed a bandgap shift of about 0.9 electron volts to higher energy, enough to produce visible-wavelength luminescence. Despite this dramatic shift, the nanocrystals still behaved fundamentally as indirect-gap materials with low oscillator strength, meaning they glowed but not brightly.8PubMed. Quantum confinement in size-selected, surface-oxidized silicon nanocrystals
More recent computational work has explored what the energy levels inside these nanocrystals actually look like. In a silicon nanocrystal just 1.8 nanometers across, the energy states are clearly discrete at low temperatures, more like a molecule than a chunk of solid. But at room temperature, thermal broadening smears those discrete levels into something resembling continuous bands, though small sub-gaps persist. Increasing the diameter to about 2.9 nanometers packs in enough states that the bands become denser and more closely resemble those of bulk silicon.9Scientific Reports. Quasi-band structure of quantum-confined nanocrystals The crossover between molecular-like and solid-like behavior happens across a size range of just one nanometer, roughly eight to ten atoms across. Silicon’s atomic size sets the scale at which its most basic electronic identity, whether it behaves like a molecule or a solid, transitions.
How Isotopes Subtly Shift Silicon’s Size
Not all silicon atoms are exactly the same. Natural silicon is a mixture of three stable isotopes: silicon-28 (about 92% of atoms), silicon-29 (about 5%), and silicon-30 (about 3%). These isotopes have the same number of protons and electrons, so their chemistry is identical, but they differ in mass because of extra neutrons. That mass difference subtly changes how strongly the atoms vibrate in the crystal lattice, which in turn changes the average spacing between atoms.
First-principles calculations predicted that the lattice parameter difference between isotopically pure silicon-28 and silicon-30 crystals would be largest at low temperatures and decrease as the crystal warms.10Metrologia. The Influence of Temperature, Isotope Composition and Impurities on the Lattice Parameter of Si But when experimentalists actually measured this difference across a range from about 5 K to 700 K, they found something unexpected. The maximum lattice parameter difference between silicon-30 and silicon-28 did not occur at the lowest temperature. Instead, the peak relative difference of about 57 parts per million showed up at around 75 K.11Physical Review Letters. Anomalous isotopic effect on the lattice parameter of silicon The explanation involves certain vibrational modes in the crystal that have negative Grüneisen parameters, meaning they cause the lattice to contract when they are excited rather than expand. At around 75 K, these modes contribute enough to produce the anomalous peak.
This may sound like an obscure detail, but it matters for the silicon-sphere metrology program mentioned earlier. If you are counting atoms to define a kilogram, you need to know the lattice spacing to extreme precision, and that spacing depends on both the isotopic makeup of the crystal and its temperature. The anomalous isotope effect was a surprise that had to be accounted for.
Silicene and the Flatland Version of Silicon
Carbon famously forms graphene, a single-atom-thick sheet of hexagonally arranged atoms. Silicon can form a similar structure called silicene, but it does not lie perfectly flat. Because silicon atoms are larger and bond somewhat differently than carbon atoms, each six-membered ring in silicene buckles: alternating atoms sit slightly above and below the average plane. This buckling is driven by a symmetry-breaking distortion and differentiates silicene from graphene in important ways.12Accounts of Chemical Research. Structures and chemical properties of silicene: unlike graphene
The buckling is a direct consequence of silicon’s atomic size. Silicon-silicon bonds are longer than carbon-carbon bonds, which makes the flat hexagonal arrangement unstable. The atoms relieve the strain by puckering out of plane. This structural difference gives silicene a stronger spin-orbit coupling than graphene, which makes it interesting for applications involving spin-based electronics. Silicene has been grown on silver and other metal substrates in the lab, though producing freestanding sheets remains a challenge.
Silicon’s Atomic Size in Biology
Given how central silicon is to technology, it is striking how minor a role it plays in biology compared to its neighbor carbon. Carbon forms the backbone of virtually all biological molecules. Silicon, despite sitting directly below carbon in the periodic table, almost never does. One reason is size: silicon atoms are about 50% larger than carbon atoms in terms of covalent radius, which makes silicon-silicon chains bulkier and less stable in water than carbon-carbon chains. Silicon-oxygen bonds are far more energetically favorable than silicon-silicon bonds in aqueous environments, so biological silicon chemistry tends to involve silicic acid, a simple molecule of silicon bonded to oxygen and hydrogen.
Some organisms have evolved dedicated transport systems for silicic acid. Diatoms, the single-celled algae that build intricate glass shells, use specialized membrane proteins called silicon transporters (SITs) to pull dissolved silicic acid from seawater. These transporters work as sodium-coupled symporters with a binding affinity for silicic acid of about 20 micromolar and a one-to-one protein-to-substrate ratio.13PubMed Central. Direct evidence of the molecular basis for biological silicon transport Four conserved amino acid motifs in the protein appear to be critical for this function, confirming a hypothesis that had been around for years.
So while silicon’s atomic size largely excludes it from the complex chain-based chemistry of life, organisms that need silicon for structural purposes have found molecular solutions to grab it from their environment in its dissolved, oxidized form. The difference between carbon-based life and the near absence of silicon-based biology traces back, at least in part, to those extra 40-odd picometers of atomic radius.