Astrochemistry of Complex Organic Molecules

Complex organic molecules sit at the frontier of interstellar chemistry: they are the largest, most intricate species we can name by their spectral fingerprints, and they carry the chemical enrichment that seeds forming planetary systems. Astrochemists define a COM as a carbon-bearing molecule of six or more atoms — methanol, methyl formate, glycolaldehyde — a threshold chosen because such species demand multi-step synthesis rather than a single gas-phase two-body encounter. Their dense forests of millimetre and submillimetre rotational lines became routinely accessible only with sensitive interferometers such as ALMA, which now resolves individual hot cores and protoplanetary disks. The decisive lesson of the last four decades is that the dominant assembly line is not the gas but the surfaces of cold dust grains. In the dark interiors of molecular clouds, atoms stick to grains and build icy mantles of water, CO, CO₂, and CH₃OH; hydrogen atoms hydrogenate frozen CO step by step into methanol, and UV photons and cosmic rays crack ice molecules into radicals that recombine, on warming, into species no efficient gas-phase route can make. When a protostar ignites, the ices sublimate in order of binding energy — volatile CO near 20 K, tenacious water and methanol near 100 K — reading the grain's stored recipe out into warm gas as the line-rich spectrum of a hot core. Turning those spectra into numbers is a problem in statistical mechanics: the rotational partition function, Boltzmann populations, and, in dense gas, LTE. This article assembles that machinery — freeze-out, desorption, and the rotational-diagram method that converts line fluxes into a column density and an excitation temperature — and shows why COMs found in cold prestellar cores keep the field an open frontier.

Key takeaways

Complex organic molecules (COMs) are carbon-bearing species of six or more atoms — methanol (CH₃OH), methyl formate, glycolaldehyde — identified by their rotational-line spectra.
They form mainly by grain-surface chemistry: gas freezes onto cold dust, CO is hydrogenated to CH₃OH, and radicals freed by UV and cosmic-ray processing recombine into larger species as the ice warms.
The classic sites are hot cores and hot corinos — compact, warm (T≳100 K), dense (n≳10⁶ cm⁻³) regions around forming stars where thermal desorption returns ice-made COMs to the gas.
Detection rests on the Boltzmann distribution: rotational-diagram analysis turns line intensities into a column density N_tot and a rotational temperature T_rot under LTE.
Prebiotic precursors — glycolaldehyde, formamide, glycine — appear toward protostars and in comet 67P, linking interstellar ice chemistry to the material inherited by planet-forming disks.
COMs now turn up in cold (≈10 K) prestellar cores far below desorption temperatures, forcing non-thermal desorption and cold gas-phase routes into the models.
Key idea: COMs are built by hydrogenation and radical recombination in dust-grain ice mantles during the cold cloud phase, then released to the gas by thermal (or non-thermal) desorption; observed rotational-line intensities, analysed via LTE rotational diagrams, yield column densities that constrain gas–grain chemical networks.

Where this lives today

The working frontier is the ALMA line survey. Toward IRAS 16293-2422, the PILS survey resolved a forest of COM lines from methyl formate and dimethyl ether to glycolaldehyde and even methyl isocyanate (CH₃NCO), which shares the C–N backbone of peptide bonds — all analysed with the machinery of this article. On the Solar System side, the ROSINA spectrometer aboard Rosetta measured the COM inventory of comet 67P and found abundances strikingly close to those in hot corinos, plus a first detection of the amino acid glycine — evidence that much of the pre-solar ice chemistry was inherited rather than reprocessed. The sharpest open problem is the cold prestellar core: methanol and acetaldehyde appear in sources like L1544 at ≈10 K, far below any thermal-desorption threshold, forcing the field to quantify non-thermal desorption (cosmic-ray photodesorption, and chemical desorption that ejects a molecule using its own formation energy) and cold gas-phase routes. Laboratory groups measure binding energies and ice-reaction barriers to feed the networks, and JWST is mapping the ice bands in absorption toward embedded protostars — closing the loop between the frozen reservoir and the gas-phase COMs it releases.

The physical picture

Picture a dust grain in a cold core as a laboratory bench held at 10 K. Gas atoms land, stick, and skate across the icy surface; whenever a hydrogen atom meets a frozen CO it adds on, and step by step CO becomes formaldehyde and then methanol. The grain performs the assembly the tenuous gas never could, holding reactants together long enough for slow reactions to complete and acting as a third body that carries away the released energy. The first figure shows this bench and its two acts: cold assembly on the left, and on the right the warm-up in which the star switches the heat on and the ices evaporate in order of how tightly each species is bound.

Figure

The dust grain as a chemical factory. In the cold core H atoms hydrogenate frozen CO to CH₃OH on the ice mantle; when the protostar warms the grain each species desorbs at T_des ≈ (E_bind divided by k_B) divided by 60, so CO leaves near 20 K and methanol near 80 to 100 K

The 'complexity' threshold is really about pathways: a molecule of six or more atoms almost always needs a surface for sequential additions or radical recombination, so a COM in warm gas is evidence of a grain-surface history read out later. Detecting one is a counting problem in disguise: each rotational line samples one energy level, and the Boltzmann distribution ties every level to one temperature. Plot the log of each level's column density against its energy — the rotational diagram — and, under LTE, the points fall on a straight line of slope −1/T_rot whose intercept fixes the total column. The second figure is that diagram for a methanol-rich hot corino.

Figure

Rotational (population) diagram for methanol in a hot corino. Two measured transitions at E_u divided by k_B of 45 K and 115 K give a straight line of slope −1/T_rot, so T_rot = 101 K; the intercept with Z_rot(101 K) ≈ 6×10² yields N_tot ≈ 1.9×10¹⁶ cm⁻²

The core equations

Everything downstream begins with how molecules distribute over rotational levels. In LTE one temperature T_rot sets every level's population through the Boltzmann factor, weighted by the degeneracy and normalised by the partition function.

n_J = n\,\frac{g_J}{Z_{\mathrm{rot}}}\,e^{-E_J/k_B T_{\mathrm{rot}}} \qquad (1)

where: $n_J$ — number density in level $J$ ; $n$ — total number density of the molecule; $g_J$ — degeneracy of level $J$ ; $Z_{\mathrm{rot}}$ — rotational partition function; $E_J$ — energy of level $J$ above ground; $T_{\mathrm{rot}}$ — rotational (excitation) temperature. Reading it: warmer gas spreads population up to higher $J$ , and the exponential decides which lines are bright. For a linear rotor the energies and degeneracies take the rigid-rotor form; real COMs like methanol are asymmetric tops with three rotational constants, but the linear case fixes the scalings.

E_J = h B J(J+1),\qquad g_J = 2J+1 \qquad (2)

where: $B$ — rotational constant in Hz, $B=h/(8\pi^2 I)$ ; $I$ — moment of inertia; $J$ — rotational quantum number; $h$ — Planck constant. Reading it: heavier, more extended molecules have larger $I$ , smaller $B$ , and closely spaced low-frequency lines. Summing the Boltzmann weights over all levels gives the partition function; for $k_BT\gg hB$ the sum collapses to a closed form — linear for a diatomic, $T^{3/2}$ for an asymmetric top.

Z_{\mathrm{rot}} \approx \frac{k_B T_{\mathrm{rot}}}{h B}\ (\text{linear}),\qquad Z_{\mathrm{rot}}\approx\sqrt{\frac{\pi (k_B T_{\mathrm{rot}})^3}{h^3 A B C}}\ (\text{asymmetric top}) \qquad (3)

where: $A,B,C$ — the three rotational constants of an asymmetric-top molecule such as CH₃OH; $k_B$ — Boltzmann constant. Linearising equation (1) along a line of sight gives the working relation of the whole field: the rotational diagram. Each transition becomes one point, and a straight-line fit reads off temperature and column at once.

\ln\!\left(\frac{N_u}{g_u}\right) = \ln\!\left(\frac{N_{\mathrm{tot}}}{Z_{\mathrm{rot}}}\right) - \frac{E_u}{k_B T_{\mathrm{rot}}} \qquad (4)

where: $N_u$ — column density in upper level $u$ ; $g_u$ — upper-level degeneracy; $N_{\mathrm{tot}}$ — total column density of the species; $E_u$ — upper-level energy. The per-line upper-level column comes from the velocity-integrated intensity, assuming optically thin LTE emission — the measurement that feeds each point into equation (4).

N_u = \frac{8\pi k_B \nu^2}{h c^3 A_{ul}}\int T_B\,\mathrm{d}v \qquad (5)

where: $\nu$ — transition frequency; $A_{ul}$ — Einstein spontaneous-emission coefficient; $T_B$ — brightness temperature; $\int T_B\,\mathrm{d}v$ — velocity-integrated line intensity in K km s⁻¹; $c$ — speed of light. The chemistry that fills those levels is the grain-surface hydrogenation of CO — the field's central reaction sequence, running on the ice at 10 K where gas-phase routes are shut.

\mathrm{CO} \xrightarrow{+\mathrm{H}} \mathrm{HCO} \xrightarrow{+\mathrm{H}} \mathrm{H_2CO} \xrightarrow{+\mathrm{H}} \mathrm{CH_3O/CH_2OH} \xrightarrow{+\mathrm{H}} \mathrm{CH_3OH} \qquad (6)

where: H₂CO — formaldehyde, the observable intermediate; CH₃OH — methanol, the anchor COM and parent of the radicals (CH₃O, CH₂OH, HCO) that recombine into larger species. Reading it: two of the four additions cross barriers by quantum tunnelling, which is why the chain runs even at 10 K. Finally, release to the gas obeys the Polanyi–Wigner desorption rate, an Arrhenius law in the binding energy — the switch that turns a cold core into a line-rich hot core.

R_{\mathrm{des}}(T_d) = N_s\,\nu_0\,\exp\!\left(-\frac{E_{\mathrm{bind}}}{k_B T_d}\right) \qquad (7)

where: $N_s$ — surface density of adsorbed molecules (cm⁻²); $\nu_0$ — vibrational frequency, $\sim10^{12}$ – $10^{13}\,\mathrm{s^{-1}}$ ; $E_{\mathrm{bind}}$ — binding energy to the ice/grain; $T_d$ — dust temperature. Reading it: the exponential switches desorption on abruptly over a narrow window set almost entirely by $E_{\mathrm{bind}}$ .

Derivation AFreeze-out timescale onto cold dust grains

The goal is to estimate how quickly gas-phase molecules deplete onto cold grains, since this timescale sets the clock for ice-mantle growth and decides whether there is time to assemble a COM inventory before the core collapses. A molecule of thermal speed $v_{\mathrm{th}}$ is captured whenever it strikes a grain of cross-section $\sigma_d=\pi a^2$ with sticking probability $S$ ; we count those collisions, express the grain abundance through the dust-to-gas ratio, and invert.

Step 1. Write the collision rate per molecule as $\mathcal{R}=n_d\,\sigma_d\,v_{\mathrm{th}}\,S$ , with $n_d$ the grain number density and $v_{\mathrm{th}}=\sqrt{8k_BT/\pi m}$ the mean molecular speed. Why: freeze-out is just kinetic-theory flux — density times area times speed times a sticking probability.

Step 2. Eliminate the grain density via the dust-to-gas mass ratio, $n_d=n_{\mathrm{gas}}\,m_{\mathrm{gas}}\,\mathcal{D}/[(4/3)\pi a^3\rho_d]$ with $\mathcal{D}\approx0.01$ , $\rho_d\approx3\,\mathrm{g\,cm^{-3}}$ , $a\approx0.1\,\mu\mathrm{m}$ ; these fix the total grain cross-section per H nucleus at the standard $\Sigma_d/n_{\mathrm{gas}}\approx10^{-21}\,\mathrm{cm^2}$ . Why: $n_d$ is unobservable, but the dust-to-gas ratio is well measured, turning it into an observable.

Step 3. Invert to $t_{\mathrm{freeze}}=\mathcal{R}^{-1}\propto n_{\mathrm{gas}}^{-1}$ ; with $v_{\mathrm{th}}\approx10^4\,\mathrm{cm\,s^{-1}}$ for CO at 10 K and $S\approx1$ this gives the compact estimate below. Why: the inverse-density scaling is the whole message — denser gas freezes out faster.

t_{\mathrm{freeze}} \approx \frac{1}{n_{\mathrm{gas}}\,(\Sigma_d/n_{\mathrm{gas}})\,v_{\mathrm{th}}\,S}\;\approx\;\frac{10^{9}\,\mathrm{yr}}{n_{\mathrm{gas}}/\mathrm{cm^{-3}}}

Step 4. At $n_{\mathrm{gas}}=10^4\,\mathrm{cm^{-3}}$ , $t_{\mathrm{freeze}}\sim10^5\,\mathrm{yr}$ — comparable to the free-fall time, so the mantle grows while the core contracts; at $10^6\,\mathrm{cm^{-3}}$ it drops to a few thousand years. Why: because freeze-out and collapse timescales coincide, ice chemistry is not a spectator — the reservoir is built exactly when star formation needs it. *As a numerical check: with $\Sigma_d/n_{\mathrm{gas}}=10^{-21}\,\mathrm{cm^2}$ and $v_{\mathrm{th}}=8.7\times10^3\,\mathrm{cm\,s^{-1}}$ , the formula gives $t_{\mathrm{freeze}}=3.6\times10^6$ yr at $10^3\,\mathrm{cm^{-3}}$ and $3.6\times10^5$ yr at $10^4\,\mathrm{cm^{-3}}$ , reproducing the standard $\sim10^9/n_{\mathrm{gas}}$ yr scaling and consistent with observed CO depletion in prestellar cores.*

Go deeper · Derivation B · 1 minDesorption temperature from binding energy

Derivation BDesorption temperature from binding energy

The aim is to derive the grain temperature at which a species' ice mantle rapidly sublimates, which explains why a warming protostellar envelope releases its molecules in a fixed sequence — CO first, water and methanol last — so that a hot-core spectrum is really a thermometer reading of grain composition. Desorption becomes explosive when the residence time on the grain falls to the local dynamical timescale $t_{\mathrm{dyn}}$ ; we set the two equal and solve, exploiting the fact that the exponential makes the answer almost independent of everything but the binding energy.

Step 1. From the per-molecule Polanyi–Wigner rate, the residence time on the grain is $t_{\mathrm{res}}=\nu_0^{-1}\exp(E_{\mathrm{bind}}/k_B T_d)$ . Why: a molecule rattles in its potential well $\nu_0$ times a second and escapes only when a fluctuation clears the barrier $E_{\mathrm{bind}}$ .

Step 2. Define desorption as the point where a molecule no longer stays put over the timescale on which the region evolves: $\nu_0^{-1}\exp(E_{\mathrm{bind}}/k_B T_{\mathrm{des}})=t_{\mathrm{dyn}}$ . Why: desorption is a race between residence and evolution; equating the two defines the threshold cleanly.

Step 3. Take the logarithm and solve, giving the boxed result. Why: with the timescales buried in a logarithm, a factor-of-ten error in $t_{\mathrm{dyn}}$ shifts $T_{\mathrm{des}}$ by a few percent — the estimate is robust.

T_{\mathrm{des}} \approx \frac{E_{\mathrm{bind}}/k_B}{\ln(\nu_0\,t_{\mathrm{dyn}})}

Step 4. With $\nu_0\sim10^{13}\,\mathrm{s^{-1}}$ and $t_{\mathrm{dyn}}\sim10^{13}\,\mathrm{s}$ , $\ln(\nu_0 t_{\mathrm{dyn}})\approx60$ . CO ( $E_{\mathrm{bind}}/k_B\approx1150$ K) gives $T_{\mathrm{des}}\approx19$ K; methanol ( $\approx4800$ K) gives $\approx80$ K, with full desorption completing near 100–130 K. Why: one formula reproduces both the ≈20 K CO snow line and the ≈100 K hot-core threshold from binding energies alone. *As a numerical check: with $\ln(\nu_0 t_{\mathrm{dyn}})=59.9$ , the formula gives CO → 19.2 K, H₂CO → 34 K, CH₃OH → 80 K, and H₂O → 95 K, bracketing laboratory temperature-programmed-desorption measurements.*

Go deeper · Derivation C · 1 minColumn density from a rotational diagram

Derivation CColumn density from a rotational diagram

The purpose is to show how a handful of measured line intensities combine into the two numbers that characterise a COM — its total column density $N_{\mathrm{tot}}$ and its excitation temperature $T_{\mathrm{rot}}$ — the most-used procedure in hot-core astrochemistry and the method behind the 1.9×10¹⁶ cm⁻² of methanol in the second figure. We assume optically thin emission and one excitation temperature for all levels (LTE), convert each line to an upper-level column, and fit a straight line whose slope and intercept carry the two unknowns.

Step 1. For each detected transition, convert the velocity-integrated intensity into an upper-level column density with $N_u = 8\pi k_B \nu^2/(h c^3 A_{ul})\int T_B\,\mathrm{d}v$ . Why: this leaves a pure count of molecules in one quantum level along the line of sight.

Step 2. Plot $y=\ln(N_u/g_u)$ against $x=E_u/k_B$ ; equation (4) predicts $y=\ln(N_{\mathrm{tot}}/Z_{\mathrm{rot}})-x/T_{\mathrm{rot}}$ , so every line contributes one point on a straight locus. Why: dividing by $g_u$ and taking the log turns the Boltzmann exponential into a straight line — the whole trick.

Step 3. Fit the points: the slope is $-1/T_{\mathrm{rot}}$ and the intercept is $\ln(N_{\mathrm{tot}}/Z_{\mathrm{rot}})$ . Evaluate $Z_{\mathrm{rot}}$ at the fitted temperature with the asymmetric-top formula (3) and multiply through to recover the boxed total column. Why: temperature comes from the ladder's shape, column from its height — two numbers from one plot.

N_{\mathrm{tot}} = Z_{\mathrm{rot}}(T_{\mathrm{rot}})\,\left(\frac{N_u}{g_u}\right)e^{E_u/k_B T_{\mathrm{rot}}}

Step 4. Inspect the residuals: real diagrams often curve, and that curvature is data — it diagnoses optical depth, sub-thermal excitation below the critical density, or several temperature components along the sightline. Why: the straight-line assumption is falsifiable, so its failure is a physical readout of the source. *As a numerical check: two methanol lines at $E_u/k_B=45$ and $115$ K with $N_u/g_u=2.0\times10^{13}$ and $1.0\times10^{13}\,\mathrm{cm^{-2}}$ give slope $-1/101\,\mathrm{K^{-1}}$ , so $T_{\mathrm{rot}}=101$ K; with $Z_{\mathrm{rot}}(101\,\mathrm{K})\approx6.2\times10^2$ the intercept yields $N_{\mathrm{tot}}=1.9\times10^{16}\,\mathrm{cm^{-2}}$ .*

Limiting cases

COM chemistry changes regime with temperature and density; the four rows below are the thresholds an astrochemist keeps in mind.

Case	Result	Meaning
Cold freeze-out (T_d ≲ 20 K)	$t_{\mathrm{freeze}}\approx 10^9/(n_{\mathrm{gas}}/\mathrm{cm^{-3}})\ \mathrm{yr}$	Gas depletes onto grains; CO hydrogenation builds methanol ice while the core collapses. Sticking $S\approx1$ , thermal speed $v_{\mathrm{th}}$ , grain cross-section $\sigma_d$ .
Warm-up / radical mobility (20 ≲ T_d ≲ 100 K)	$k_{\mathrm{diff}}=\nu_0\,e^{-E_{\mathrm{diff}}/k_B T_d}$	Radicals (HCO, CH₃O) gain surface mobility and recombine into larger COMs. Attempt frequency $\nu_0\sim10^{12}\,\mathrm{s^{-1}}$ ; diffusion barrier $E_{\mathrm{diff}}\approx0.3$ – $0.5\,E_{\mathrm{bind}}$ .
Thermal desorption / hot core (T_d ≳ 100 K)	$R_{\mathrm{des}}=N_s\nu_0\,e^{-E_{\mathrm{bind}}/k_B T_d}$	Ices sublimate and COMs flood the gas; methanol desorbs near 80–130 K ( $E_{\mathrm{bind}}/k_B\approx4800$ K). This is the line-rich hot-core spectrum.
Cold prestellar surprise (T ≈ 10 K)	$R_{\mathrm{gas}} + f_{\mathrm{nt}}R_{\mathrm{des}}$	COMs seen far below desorption temperatures need non-thermal (photo-, cosmic-ray-, reactive-) desorption plus cold gas-phase radical routes, since thermal $R_{\mathrm{des}}\!\to\!0$ .

Common trapEvery gas-phase COM must have been made in the gas

It was almost certainly not. Two-body gas-phase reactions are far too slow or too barriered at 10–100 K to build methanol and larger species, and radiative association cannot compete at interstellar densities — models underproduce gas-phase CH₃OH by orders of magnitude. The abundances you measure are downstream tracers of an earlier ice chemistry, released by desorption. The mirror-image trap is on the analysis side: do not read a rotational-diagram temperature as the gas kinetic temperature unless the critical-density condition $n\gtrsim n_{\mathrm{crit}}=A_{ul}/\langle\sigma v\rangle$ is met. Below it the levels are sub-thermally excited, the diagram curves, and a straight-line fit biases both $T_{\mathrm{rot}}$ and the column.

Worked examples

1 · The sublimation ladder from binding energies (fully worked)

Using $T_{\mathrm{des}}\approx (E_{\mathrm{bind}}/k_B)/\ln(\nu_0 t_{\mathrm{dyn}})$ with $\nu_0=10^{13}\,\mathrm{s^{-1}}$ and $t_{\mathrm{dyn}}=10^{13}\,\mathrm{s}$ , predict the order and temperatures at which CO, H₂CO, CH₃OH, and H₂O leave the grain, and interpret what an observer sees as the envelope warms.

Step 1. Fix the logarithm once: $\ln(\nu_0 t_{\mathrm{dyn}})=\ln(10^{13}\times10^{13})=\ln(10^{26})=26\ln10=\mathbf{59.9}$ . Because this factor is shared by every species, the desorption temperature is simply $E_{\mathrm{bind}}/k_B$ divided by ≈60.

Step 2. Divide each binding energy through. CO ( $E_{\mathrm{bind}}/k_B=1150$ K): $T_{\mathrm{des}}=1150/59.9=\mathbf{19\ K}$ . Formaldehyde (2050 K): $2050/59.9=\mathbf{34\ K}$ . Methanol (4800 K): $4800/59.9=\mathbf{80\ K}$ . Water (5700 K): $5700/59.9=\mathbf{95\ K}$ .

Step 3. The sequence is CO → H₂CO → CH₃OH → H₂O: weakly bound apolar species first, hydrogen-bonded polar species last. As a protostar heats its envelope, an observer sees CO fill a large cool region while methanol and water light up only in the compact inner core where $T_d\gtrsim100$ K. The emitting-gas composition maps directly onto local dust temperature, and the methanol zone marks the ≈100 K snow-line surface.

2 · Temperature and column of methanol in a hot corino (fully worked)

Two optically thin CH₃OH lines toward a low-mass protostar give, via equation (5), upper-level columns $N_u/g_u = 2.0\times10^{13}\,\mathrm{cm^{-2}}$ at $E_u/k_B=45$ K and $1.0\times10^{13}$ at $115$ K. Find $T_{\mathrm{rot}}$ and $N_{\mathrm{tot}}$ .

Step 1. Form the slope of $\ln(N_u/g_u)$ versus $E_u/k_B$ : $\dfrac{\ln(1.0\times10^{13})-\ln(2.0\times10^{13})}{115-45}=\dfrac{\ln(0.5)}{70}=\dfrac{-0.693}{70}=-9.90\times10^{-3}\,\mathrm{K^{-1}}$ .

Step 2. Identify the slope with $-1/T_{\mathrm{rot}}$ : $T_{\mathrm{rot}}=1/(9.90\times10^{-3})=\mathbf{101\ K}$ — a genuine hot-corino excitation temperature, consistent with thermal desorption of the methanol ice.

Step 3. Intercept from either point: $\ln(N_{\mathrm{tot}}/Z_{\mathrm{rot}})=\ln(2.0\times10^{13})+45/101=31.07$ , so $N_{\mathrm{tot}}/Z_{\mathrm{rot}}=e^{31.07}=3.1\times10^{13}\,\mathrm{cm^{-2}}$ .

Step 4. Evaluate the asymmetric-top partition function at 101 K. With $A,B,C\approx127.6, 24.68, 23.76$ GHz and $k_BT/h=2.08\times10^{12}$ Hz, equation (3) gives $Z_{\mathrm{rot}}\approx6.2\times10^{2}$ . Multiplying through, $N_{\mathrm{tot}}=3.1\times10^{13}\times6.2\times10^{2}=\mathbf{1.9\times10^{16}\ cm^{-2}}$ . T_rot ≈ 101 K and N_tot(CH₃OH) ≈ 1.9×10¹⁶ cm⁻², a typical hot-corino methanol column and exactly the fit drawn in the second figure.

Physics of the portrait

The story of complex organic molecules is the story of astrochemistry becoming quantitative — from the first unidentified optical lines of the 1930s to the ALMA line surveys of today — and no single scientist has shaped its modern, ice-centred synthesis more than Ewine van Dishoeck.

Origins · 1 minEwine van Dishoeck and how this was discovered

Ewine van Dishoeck

b. 1955 · astrochemistry of star- and planet-forming regions

The field she leads was built on a puzzle: the interstellar medium is cold, dilute, and bathed in dissociating radiation, yet chemically rich. Ion–molecule gas-phase chemistry, set out by Eric Herbst and William Klemperer in 1973, explained many small molecules but failed conspicuously for methanol and the larger COMs; the grain-surface answer came from Alexander Tielens and Werner Hagen in 1982, who argued that hydrogenation of frozen CO on dust could build the very species gas-phase chemistry could not. Ewine van Dishoeck, a Dutch chemist-turned-astronomer at Leiden, made this picture quantitative and observational. Beginning with precise photodissociation rates — which regulate which molecules survive in irradiated gas — she mapped the full gas–ice–gas cycle: freeze-out onto grains, surface processing into water and organics, and thermal release in hot cores, tracing water from clouds to disks with Herschel and ALMA. Her work turned astrochemistry from a catalogue of detections into a predictive account of how chemical complexity is inherited by planet-forming disks, earning her the 2018 Kavli Prize. The modern statement — that a gas-phase COM is the read-out of an earlier grain-surface history — is in large part her synthesis.

History

1940

— McKellar identifies CH and CN in optical absorption, the first interstellar molecules

1969

— formaldehyde, the first interstellar organic molecule, detected by radio astronomy

1970

— methanol detected toward Sagittarius B2, pointing to grain-surface chemistry

2000

— glycolaldehyde, the first interstellar sugar-related molecule, found in Sgr B2

2016

— Rosetta/ROSINA detects the amino acid glycine in comet 67P

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., ch. 12 (The Interstellar Medium and Star Formation) · Tielens, The Physics and Chemistry of the Interstellar Medium (2005) · Draine, Physics of the Interstellar and Intergalactic Medium (2011), ch. 31–32 · Herbst & van Dishoeck, Complex Organic Interstellar Molecules, Annu. Rev. Astron. Astrophys. 47, 427 (2009) · van Dishoeck, Astrochemistry of dust, ice and gas, Faraday Discuss. 168, 9 (2014). Key numerical benchmarks used throughout this article: the high-T partition function $Z_{\mathrm{rot}}=k_BT/hB$ (linear) and the $T^{3/2}$ asymmetric-top form; the freeze-out timescale $t_{\mathrm{freeze}}\approx10^9/(n_{\mathrm{gas}}/\mathrm{cm^{-3}})$ yr (≈3.6×10⁶ yr at 10³ cm⁻³, 3.6×10⁵ yr at 10⁴ cm⁻³, using $\Sigma_d/n_{\mathrm{gas}}=10^{-21}\,\mathrm{cm^2}$ ); the desorption estimate $T_{\mathrm{des}}=(E_{\mathrm{bind}}/k_B)/\ln(\nu_0 t_{\mathrm{dyn}})$ with $\ln\approx60$ , giving CO ≈19 K and CH₃OH ≈80 K for $E_{\mathrm{bind}}/k_B=1150$ and 4800 K; and the worked hot-corino fit ( $T_{\mathrm{rot}}=101$ K, $N_{\mathrm{tot}}=1.9\times10^{16}\,\mathrm{cm^{-2}}$ , $Z_{\mathrm{rot}}\approx6.2\times10^2$ ) derived from the two quoted methanol lines. Cometary COM detections are from Rosetta/ROSINA (Altwegg et al.). Physical constants are CODATA.

Prerequisites. Molecular Clouds and Giant Molecular Clouds · Interstellar Dust Grains · Phases of the Interstellar Medium · Interstellar Gas Heating and Cooling · Interstellar Extinction and Reddening. Leads to. Thermal Dust Emission and Infrared Radiation · CO as a Molecular Gas Tracer and the X-Factor · The Jeans Instability and Jeans Mass.

Ask the tutor: Want to walk a real ALMA methanol line list through a rotational diagram, or trace one CO molecule from freeze-out through hydrogenation to hot-core desorption?

What exactly counts as a complex organic molecule in astrochemistry?

By the working definition of Herbst & van Dishoeck, a COM is a carbon-bearing molecule with six or more atoms — a deliberately low bar, with methanol (CH₃OH) as the anchor example. The point is not chemical elaborateness: the six-atom threshold marks the size above which a species cannot generally be assembled in one gas-phase collision and instead needs the multi-step chemistry of grain surfaces. The definition is about formation pathway, not structure, which is why a COM in the gas is read as a fingerprint of an earlier ice-chemistry history.

Why can't gas-phase chemistry make interstellar methanol?

At 10–100 K the two-body gas-phase reactions that would build methanol are endothermic or barriered, and radiative association is too slow at interstellar densities to compete; gas-only networks underproduce CH₃OH by orders of magnitude. The grain solves both problems: it is a third body that carries away the reaction energy and holds reactants adjacent long enough for slow, tunnelling-assisted hydrogenation of frozen CO to run to completion.

What do the slope and intercept of a rotational diagram actually tell me?

Plotting $\ln(N_u/g_u)$ against $E_u/k_B$ gives, under LTE, a straight line. Its slope is $-1/T_{\mathrm{rot}}$ (steeper means colder); its intercept is $\ln(N_{\mathrm{tot}}/Z_{\mathrm{rot}})$ , so evaluating the partition function at the fitted temperature recovers the total column. The shape gives the temperature and the height gives the abundance, from one plot. Curvature is informative too, flagging optical depth, sub-thermal excitation, or multiple temperature components.

Why do ices desorb in a definite order as a protostar warms?

Because desorption follows an Arrhenius law whose threshold $T_{\mathrm{des}}\approx(E_{\mathrm{bind}}/k_B)/\ln(\nu_0 t_{\mathrm{dyn}})$ is set almost entirely by binding energy. Weakly bound apolar CO ( $E_{\mathrm{bind}}/k_B\approx1150$ K) escapes near 20 K; hydrogen-bonded water and methanol ( $\approx4800$ –5700 K) hold on until ≈80–130 K. As the envelope heats from the inside out, molecules sublimate in strict order of binding energy, so the emitting gas maps onto local dust temperature — a hot-core spectrum is a thermometer.

What is the significance of matching COM abundances between protostars and comet 67P?

The complex-organic abundances measured in comet 67P by Rosetta closely resemble those toward hot corinos like IRAS 16293-2422. That argues for chemical inheritance: the inventory built on grains during the cold, pre-collapse phase survived disk formation and was incorporated largely intact into cometary ices rather than rebuilt. Since comets and meteorites deliver such material to young planets, the match ties interstellar ice chemistry directly to the prebiotic ingredients — glycolaldehyde, formamide, glycine — present at a planetary system's birth.

When is the LTE assumption behind rotational diagrams actually valid?

LTE holds when collisions keep the populations Boltzmann-distributed at the gas kinetic temperature, requiring the collider (mostly H₂) density to exceed the critical density $n_{\mathrm{crit}}=A_{ul}/\langle\sigma v\rangle$ . For many COM lines $n_{\mathrm{crit}}\sim10^5$ – $10^6\,\mathrm{cm^{-3}}$ , comfortably met in the ≳10⁶ cm⁻³ gas of hot cores — which is why the method works there. In diffuse gas the upper levels are sub-thermally excited, $T_{\mathrm{rot}}$ underestimates the kinetic temperature, and the diagram curves; one then needs a full non-LTE treatment.