The interstellar medium (ISM) is the matter and radiation found in the space between star systems in a galaxy. This matter includes gas in ionic, atomic, and molecular forms, as well as dust and cosmic rays. It fills the space between stars and connects smoothly with the surrounding intergalactic medium. The energy in this space, in the form of electromagnetic radiation, is called the interstellar radiation field. Even though the number of atoms in the ISM is much lower than in the best laboratory vacuums, the distance between collisions is short compared to the size of interstellar space. This means the ISM behaves like a gas (specifically a plasma, which is a type of gas with charged particles) rather than a group of particles that do not interact.
The ISM has different phases based on whether matter is ionic, atomic, or molecular, and the temperature and density of the matter. It is mostly made of hydrogen, with helium and small amounts of carbon, oxygen, and nitrogen. The pressures from the heat of these phases are roughly balanced. Magnetic fields and chaotic movements also create pressure in the ISM and are usually more important than heat-related pressure. In the ISM, matter is mostly in molecular form, with up to 10 molecules per cubic meter (1 trillion molecules per cubic meter). In hot, spread-out regions, gas is highly charged, and the density may be as low as 100 ions per cubic meter. For comparison, air at sea level has about 10 quadrillion molecules per cubic meter, and a high-vacuum lab has about 10 molecules per cubic meter. In our galaxy, by mass, 99% of the ISM is gas, and 1% is dust.
Of the gas in the ISM, 91% of the atoms are hydrogen, 8.9% are helium, and 0.1% are atoms of heavier elements, called "metals" in astronomy. By mass, this is 70% hydrogen, 28% helium, and 1.5% heavier elements. Hydrogen and helium were mostly created in the early universe, while the heavier elements were produced by stars through nuclear reactions and released into the ISM during their life cycles.
The ISM is important in astrophysics because it connects processes on a star scale and a galaxy scale. Stars form in the densest parts of the ISM, which later contributes to molecular clouds and returns matter and energy to the ISM through planetary nebulae, stellar winds, and supernovae. This interaction between stars and the ISM affects how quickly a galaxy uses up its gas and how long it can form new stars. Voyager 1 entered the ISM on August 25, 2012, becoming the first human-made object from Earth to reach it. Scientists will study interstellar plasma and dust until the mission ends in 2025. Its twin, Voyager 2, entered the ISM on November 5, 2018.
Interstellar matter
Table 1 shows the properties of the components in the interstellar medium (ISM) of the Milky Way.
In 1969, Field, Goldsmith, and Habing proposed a model called the static two-phase equilibrium model to explain the observed properties of the ISM. Their model included two phases: a cold, dense phase (temperature less than 300 K), made of neutral and molecular hydrogen clouds, and a warm intercloud phase (temperature about 10 K), made of rarefied neutral and ionized gas. In 1977, McKee and Ostriker added a third phase, a dynamic phase of very hot gas (temperature about 10 K) that had been heated by supernovae. This hot gas fills most of the ISM’s volume. These phases represent temperatures where heating and cooling balance to reach stable equilibrium. Their work became the foundation for research over the next 30 years. However, scientists still do not fully understand the proportions of these phases or their subdivisions.
The basic physics of these phases can be understood by studying hydrogen, since it is the most abundant component of the ISM. The different phases are roughly in pressure balance across most of the Milky Way’s disk. Regions with higher pressure expand and cool, while regions with lower pressure compress and heat. Since pressure (P) equals the product of particle density (n), Boltzmann’s constant (k), and temperature (T), hot regions (high T) usually have low particle density (n). Coronal gas has such low density that particle collisions are rare, so little radiation is produced. This means energy is lost slowly, allowing the temperature to remain high for millions of years. In contrast, when the temperature drops to about 10 K, with higher density, protons and electrons recombine to form hydrogen atoms, emitting photons that remove energy from the gas, causing rapid cooling. This process would create the warm neutral medium. However, OB stars are so hot that some of their photons have energy greater than 13.6 eV, enough to ionize hydrogen. These photons ionize neutral hydrogen atoms, creating a dynamic balance between ionization and recombination. Gas near OB stars becomes mostly ionized, with a temperature of about 8,000 K, until the point where all ionizing photons are absorbed. This boundary marks the edge between the warm ionized and warm neutral medium.
OB stars and cooler stars emit many photons with energy below the Lyman limit (13.6 eV), which pass through the ionized region with little absorption. Some of these photons have enough energy (>11.3 eV) to ionize carbon atoms, forming a region of ionized carbon (C II) outside the hydrogen ionization front. In dense regions, the size of this region may be limited by the number of available photons, but often these photons can travel through the neutral phase and are absorbed only in the outer layers of molecular clouds. Photons with energy greater than 4 eV can break apart molecules like H₂ and CO, creating a photodissociation region (PDR), which is similar to the warm neutral medium. These processes contribute to heating the warm neutral medium. The distinction between the warm and cold neutral medium arises from a range of temperatures and densities where rapid cooling occurs.
The densest molecular clouds have much higher pressure than the average interstellar pressure because they are held together by gravity. When stars form in these clouds, especially OB stars, they convert surrounding gas into the warm ionized phase, increasing the temperature by several hundred degrees. Initially, the gas remains at molecular cloud densities, creating a high-pressure region called an H II region. The large pressure causes the ionized gas to expand outward from the molecular gas (a process called a Champagne flow). This expansion continues until the molecular cloud is fully evaporated or the OB stars die after a few million years. When OB stars explode as supernovae, they create blast waves that heat the gas to the coronal phase (supernova remnants, SNR). These remnants expand and cool over millions of years until they return to the average ISM pressure.
Most studies of the ISM focus on spiral galaxies like the Milky Way, where most of the ISM mass is confined to a thin disk. This disk has a scale height of about 100 parsecs (300 light years), compared to a typical disk diameter of 30,000 parsecs. Gas and stars in the disk orbit the galactic center at speeds of about 200 km/s. This orbital motion is much faster than the random motions of atoms in the ISM, but since the gas moves coherently, its average motion does not directly affect the ISM’s structure. The vertical scale height of the ISM is determined by a balance between the gravitational pull of the disk’s stars and the pressure of the gas. Farther from the disk plane, the ISM is mainly in the low-density warm and coronal phases, which extend several thousand parsecs above and below the disk. This galactic halo or "corona" also contains significant magnetic fields and cosmic ray energy.
The rotation of galaxy disks affects ISM structures in several ways. Because the angular velocity decreases with distance from the center, features like giant molecular clouds or magnetic field lines that span different radii are stretched in the tangential direction by differential rotation. This stretching is opposed by interstellar turbulence, which randomizes structures. Spiral arms form due to density waves in the disk, causing gas to compress and expand. Visible spiral arms are regions of maximum density, where compression often triggers star formation in molecular clouds, leading to many H II regions along the arms. The Coriolis force also influences large-scale ISM features.
Irregular galaxies, such as the Magellanic Clouds, have ISMs similar to spirals but with less organization. In elliptical galaxies, the ISM is mostly in the coronal phase because there is no coherent disk motion to support cold gas far from the center. Instead, the ISM’s scale height must match the galaxy’s radius, which explains the lack of current star formation in ellipticals. Some elliptical galaxies show evidence of small disk components with ISM properties similar to spirals, located near their centers. Lenticular galaxies have ISM properties that are intermediate between spirals and ellipticals.
Very close to the centers of most galaxies (within a few hundred light years), the ISM is greatly altered by the central supermassive black hole. For example, the Milky Way’s ISM near its center is studied in detail, and extreme cases are found in other galaxies with active galactic nuclei. The rest of this article focuses on the ISM in the disk plane of spiral galaxies, far from the galactic center.
Astronomers describe the ISM as turbulent, meaning the gas has random motions that are consistent over many spatial scales. Unlike typical turbulence, where fluid motions are much slower than the speed of sound, the ISM’s bulk motions are
Heating and cooling
The ISM is usually not in a stable energy state. Collisions between particles create a pattern of speeds known as the Maxwell–Boltzmann distribution. The temperature used to describe interstellar gas is called the "kinetic temperature," which represents the temperature at which particles would have the observed speed pattern if the gas were in a stable energy state. However, the interstellar radiation field is much weaker than in a stable energy state; it is most often similar to the light from an A star (surface temperature of about 10,000 K) but spread out over a large area. Because of this, energy levels within atoms or molecules in the ISM are rarely filled according to the Boltzmann formula (Spitzer 1978, §2.4).
The temperature of gas in the ISM depends on its temperature, density, and ionization state. Different processes for heating and cooling the gas affect its temperature.
Grain heating through energy exchange is important in supernova remnants, where temperatures and densities are very high.
Gas heating caused by collisions between gas and dust grains is most common in dense areas of giant molecular clouds. Far infrared radiation can travel far through these regions because the gas is not very thick. Dust grains absorb this radiation and transfer heat to the gas during collisions. A measure of how efficiently this energy is transferred is called the accommodation coefficient: α = (T₂ − T) / (T_d − T), where T is the gas temperature, T_d is the dust temperature, and T₂ is the temperature of the gas particle after a collision. This value was measured by (Burke & Hollenbach 1983) as α = 0.35.
Observations of the ISM
Even though the interstellar medium (ISM) has very low density, photons from the ISM are visible in almost all parts of the electromagnetic spectrum. In fact, the optical band, which astronomers used until the 20th century, is where the ISM is least noticeable.
Ionized gas emits energy across a wide range through a process called bremsstrahlung. For gas in the warm phase (10 K), this is mostly detected in microwaves, while bremsstrahlung from gas at million-kelvin temperatures is prominent in soft X-rays. Many spectral lines are also produced, including those important for cooling mentioned earlier. One of these, a forbidden line from doubly-ionized oxygen, gives many nebulae their green color in visual observations. This line was once mistaken for a new element called nebulium. Spectral lines from highly excited hydrogen atoms are detectable at infrared and longer wavelengths, including radio recombination lines. Unlike optical lines, radio recombination lines are not blocked by dust, allowing them to trace ionized regions across the Galaxy. At high temperatures, coronal gas emits a different set of lines because atoms lose more electrons.
The warm neutral medium produces most of the 21-cm line emission from hydrogen detected by radio telescopes. Atomic hydrogen in the cold neutral medium also contributes to this emission and to absorption of photons from background gas (called H I self-absorption, or HISA). While not important for cooling, the 21-cm line is easily observed with high detail, providing the clearest view of the warm neutral medium.
Molecular clouds are detected through spectral lines caused by changes in the rotational energy of small molecules, especially carbon monoxide (CO). The most common line is at 115 GHz, which corresponds to a change in angular momentum from the 1 to 0 quantum state. Hundreds of other molecules have been identified, each with many lines, allowing scientists to study physical and chemical processes in molecular clouds. These lines are most common at millimeter and submillimeter wavelengths. Hydrogen gas (H₂), the most common molecule in molecular clouds, is usually not directly observable because it stays in its lowest energy state except during rare events like interstellar shock waves. Some regions, called "dark gas," contain molecular hydrogen but lack detectable CO lines because CO molecules are broken apart. These regions are identified by the presence of dust grains with no matching gas emission.
Interstellar dust grains release absorbed starlight energy as emission that resembles blackbody radiation in the far infrared, matching dust temperatures of 20–100 K. Very small dust grains, similar to fragments of graphene bonded to hydrogen atoms (polycyclic aromatic hydrocarbons, or PAHs), emit many spectral lines in the mid-infrared, around 10 microns. Nanometer-sized grains can spin rapidly when struck by ultraviolet photons, and the dipole radiation from these spinning grains is believed to cause anomalous microwave emission.
Cosmic rays create gamma-ray photons when they collide with atomic nuclei in ISM clouds. Some cosmic-ray electrons collide with photons in the interstellar radiation field and the cosmic microwave background, increasing photon energy to X-rays and gamma-rays through a process called inverse Compton scattering. Due to the galactic magnetic field, charged cosmic-ray particles move in spiral paths. For electrons, this motion produces synchrotron radiation, which is very bright at low radio frequencies.
Radiowave propagation
Radio waves are influenced by the properties of plasma in the interstellar medium (ISM). Radio waves with frequencies below about 0.1 MHz cannot travel through the ISM because their frequency is too low compared to the plasma frequency of the ISM. At higher frequencies, the plasma causes radio waves to bend, with the bending decreasing as frequency increases. This bending depends on the number of free electrons in the ISM. Changes in the number of free electrons cause a phenomenon called interstellar scintillation, which makes distant radio sources appear larger. This effect becomes less noticeable as frequency increases. The bending of radio waves also causes delays in the arrival times of pulses from pulsars and fast radio bursts at lower frequencies. This delay depends on the total number of free electrons along the path of the radio waves, known as the dispersion measure (DM). DM helps scientists map ionized gas in the Galaxy and estimate distances to pulsars, as more distant pulsars have larger DM values.
A second effect is Faraday rotation, which changes the direction of polarization in radio waves that are polarized in a straight line, such as those from synchrotron radiation. Faraday rotation depends on both the number of free electrons and the strength of magnetic fields, making it a tool for studying magnetic fields in space.
The ISM is mostly transparent to radio waves, allowing clear observations through the Galaxy. However, some radio waves are blocked in specific cases. Strong radio emission lines, such as those from carbon monoxide at millimeter wavelengths or neutral hydrogen at 21 cm, can become opaque, hiding the interiors of the emitting clouds. These clouds remain transparent at other wavelengths. Dense ionized regions emit radio waves through a process called thermal bremsstrahlung. At short wavelengths, like microwaves, these regions are mostly transparent. However, their brightness increases with wavelength, and they become opaque at longer wavelengths. At meter wavelengths, H II regions appear as dark spots against the bright background of synchrotron radiation from the Galaxy. At decameter wavelengths, the entire galactic plane is blocked, and the longest radio waves (1 km) can only travel 10–50 parsecs through the Local Bubble. The wavelength at which a region becomes opaque depends on its emission measure (EM), which is the total squared electron density along the path of the radio waves. Very dense regions, such as ultra-compact H II regions, can block radio waves at centimeter wavelengths.
Although radio waves above 10 GHz are blocked by Earth’s atmosphere, the ISM is much less dense than Earth’s atmosphere. This allows radio waves to travel through the Galaxy more easily, even though the ISM is mostly transparent to radio waves, especially microwaves.
History of knowledge of interstellar space
The word "interstellar" (meaning between the stars) was first used by Francis Bacon in the context of an old idea that stars were part of a fixed, spherical structure. Later, in the 17th century, when scientists began to believe stars were scattered across endless space, they debated whether this space was completely empty or filled with an imagined fluid called aether, as suggested by René Descartes in his theory about how planets move in swirling patterns. Although Descartes' theory was later replaced by Newton's physics, the idea of an invisible aether was reintroduced in the 19th century as a medium for light waves. For example, in 1862, a writer described how light waves caused vibrations in the aether that filled space between stars.
In 1864, William Huggins used a technique called spectroscopy to study a nebula and discovered it was made of gas. Huggins had a private observatory with an 8-inch telescope, which included a lens made by Alvan Clark. His telescope was specially designed for spectroscopy, allowing him to make important discoveries.
Around 1889, Edward Barnard began using deep sky photography to study the Milky Way. He noticed dark areas that looked like "holes" in the galaxy. At first, he thought these were similar to sunspots, but by 1899, he suggested they might be caused by dark clouds of gas and dust blocking light from stars behind them. These dark areas are now called dark nebulae, and many were recorded in Barnard's Catalogue. In 1904, Johannes Hartmann used a large telescope to study the binary star Mintaka. He observed that the absorption line of calcium appeared very sharp and did not change with the star's motion. Hartmann concluded that the gas causing the absorption was not in the star itself but in a separate cloud of matter along the line of sight to the star. This discovery marked the beginning of studying the interstellar medium.
In 1909, Slipher confirmed the presence of interstellar gas, and by 1912, he also confirmed the existence of interstellar dust. In 1919, Mary Lea Heger discovered interstellar sodium by observing absorption lines from the star Delta Orionis and Beta Scorpii.
Viktor Ambartsumian later proposed that interstellar matter exists in the form of clouds. In 1936, Beals studied calcium absorption lines from stars Epsilon and Zeta Orionis and found double, uneven patterns in the spectra. These patterns showed that the absorption lines came from multiple clouds of gas moving at different speeds. Because each cloud moved at a different speed (toward or away from Earth), the absorption lines appeared shifted in color (blue or red) due to the Doppler Effect. These findings proved that interstellar matter is not spread evenly but exists in separate clouds.
Pickering (1912) noted that the interstellar medium might be similar to the aether but that its absorption patterns suggested it was made of gas. Around the same time, Victor Hess discovered cosmic rays—high-energy particles from space. This led scientists to wonder if these particles also filled space between stars. In 1913, Kristian Birkeland suggested that space might be filled with electrons and ions, with material from stars and galaxies scattered throughout the universe.
Thorndike (1930) stated that the space between stars could not be completely empty because charged particles from the Sun and other stars likely filled it.
In September 2012, NASA scientists found that polycyclic aromatic hydrocarbons (PAHs), when exposed to interstellar conditions, change into more complex organic molecules. These changes may explain why PAHs are rarely detected in cold, dense regions of space.
In February 2014, NASA updated a database to track PAHs in the universe. Scientists believe more than 20% of the universe’s carbon may be linked to PAHs, which might have formed shortly after the Big Bang and are found near stars and planets.
In April 2019, scientists using the Hubble Space Telescope confirmed the presence of large, complex molecules called buckminsterfullerene (C60), or "buckyballs," in interstellar space.
In September 2020, evidence showed that solid water ice, mixed with silicate grains, exists in interstellar space within cosmic dust.