Identifying Symmetry in Common Molecules

Correctly identifying symmetry elements is the unglamorous first step behind an enormous range of modern chemistry. In vibrational spectroscopy, whether a stretching or bending mode is infrared-active, Raman-active, both, or silent depends entirely on which irreducible representation it transforms as within the correct point group — and that classification is impossible without first getting the element list right, exactly as demonstrated for water, ammonia, and benzene above; the mutual exclusion rule for centrosymmetric molecules like benzene or trans-dichloroethylene (no mode is simultaneously IR- and Raman-active) is a direct consequence of the presence of $i$ found in Worked Examples 2 and the benzene derivation. In molecular orbital theory, symmetry-adapted linear combinations of atomic orbitals — the basis for understanding benzene's aromatic $\pi$ system, ammonia's lone pair hybridization, or water's frontier orbitals in acid–base chemistry — require the same element inventory as an input before any Reducible and Irreducible Representations machinery can act on it. Crystallographers refining X-ray diffraction data use point-group symmetry to reduce the number of independent structural parameters that must be fit, directly speeding up structure solution for everything from small-molecule pharmaceuticals to metal–organic frameworks. Computational chemistry packages run an automated version of exactly this element-search procedure as a preprocessing step before every geometry optimization or frequency calculation, both to exploit symmetry for computational savings and to correctly label vibrational and electronic states. In materials science, the presence or absence of an inversion center in a candidate crystal structure — the same test performed by hand on trans-dichloroethylene and benzene above — determines whether a material can be piezoelectric or ferroelectric, since those properties are forbidden by symmetry in any centrosymmetric structure. And in pharmaceutical chemistry, recognizing that a chiral drug candidate's point group contains no improper elements at all ($S_n$, $\sigma$, or $i$) is precisely the symmetry-based test for optical activity used routinely when screening candidate molecules.

Ask the tutor: Once I have the full element list for a molecule like benzene or staggered ethane, what is the exact flowchart procedure for turning that list into its formal point-group label?

Every rigid, static molecular geometry can be tested against five kinds of symmetry operations, each generated by a corresponding element. The identity $E$ does nothing and every molecule "has" it trivially — its only role is to make the mathematics of the point group close (every group needs an identity element). A proper rotation $C_n$ turns the molecule by $360^\circ/n$ about an axis and leaves it superimposed on itself; the axis of highest $n$ found in a molecule is called the principal axis and conventionally drawn as $z$. A mirror plane $\sigma$ reflects every atom through a plane and maps the molecule onto itself; planes are labeled $\sigma_v$ (containing the principal axis), $\sigma_h$ (perpendicular to it), or $\sigma_d$ (containing the axis but bisecting the angle between adjacent $C_2$'s). The inversion center $i$ sends every point $(x,y,z)$ to $(-x,-y,-z)$ through a single center, usually — but not always — an atom. Finally the improper rotation $S_n$ combines a $C_n$ rotation with a reflection through a plane perpendicular to that axis; you already know how $S_n$ behaves algebraically, so here the task is simply to notice when a molecule's own geometry hides one, often "for free" once a $C_n$ and a perpendicular $\sigma_h$ both exist, since $S_n = \sigma_h \cdot C_n$. Water is the cleanest teaching case. Its two O–H bonds define a plane, and the H–O–H angle is bent at roughly $104.5^\circ$, far from linear. Because the two hydrogens are chemically identical, there is exactly one $C_2$ axis: it bisects the H–O–H angle and passes through the oxygen. Rotating the molecule $180^\circ$ about this axis swaps H$_a$ with H$_b$ — indistinguishable atoms trade places, so the molecule looks unchanged. There is no $C_3$, $C_4$, or higher axis because the molecule has no such higher rotational repetition; a bent triatomic simply cannot support one. Two mirror planes emerge: the plane containing all three nuclei (call it the molecular plane) reflects the molecule into itself trivially since every atom already lies in that plane, and a second plane perpendicular to the first, also containing the $C_2$ axis and the oxygen, swaps the two hydrogens exactly as the $C_2$ does. There is no inversion center — inverting through the oxygen would send each hydrogen to a point on the opposite side where there is no atom, since the molecule is bent rather than linear. No improper $S_n$ axis appears either, beyond the trivial $S_1 = \sigma$ and $S_2 = i$ identifications, because there is no $\sigma_h$ perpendicular to the $C_2$ to combine with it. Ammonia extends the pattern to three-fold symmetry. The nitrogen sits above the plane of three hydrogens in a shallow pyramid, with an H–N–H angle near $107^\circ$. The principal axis is a $C_3$ running from nitrogen straight down through the centroid of the three hydrogens; rotating by $120^\circ$ cycles H$_a \to$H$_b \to$H$_c \to$H$_a$, and doing it twice ($C_3^2$, a $240^\circ$ rotation) cycles them the other way — both operations leave the pyramid superimposed on itself. Three vertical mirror planes exist, each containing the $C_3$ axis, the nitrogen, and exactly one hydrogen, reflecting the other two hydrogens into each other. There is no horizontal mirror plane, because reflecting through a plane perpendicular to $C_3$ at the nitrogen would try to swap the pyramidal "up" arrangement with a "down" arrangement that does not exist in this single pyramidal geometry. Consequently there is no $S_3$ improper axis either (no $\sigma_h$ to pair with the $C_3$), and no inversion center, for the same reason water lacks one: pyramidal, non-planar, non-linear shapes generically break $i$. Benzene is where every element type shows up at once, rewarding careful bookkeeping. The six carbons and six hydrogens lie in a single plane, arranged in a perfect hexagon. The principal axis is $C_6$, perpendicular to the ring plane, threading through its center; rotating $60^\circ$ cycles every carbon into the position of its neighbor and is a genuine symmetry operation because all six C–H units are identical. Every $C_6$ carries with it the full ladder $C_6, C_6^2 (=C_3), C_6^3 (=C_2), C_6^4 (=C_3^2), C_6^5$. Perpendicular to this principal axis lie six distinct $C_2$ axes, in two sets of three: three pass through opposite pairs of carbon atoms (and the C–H bonds attached to them), and three pass through the midpoints of opposite C–C bonds. The molecular plane itself is a mirror plane, and because it is perpendicular to the principal $C_6$ axis, it is classified as $\sigma_h$. There are also six vertical planes containing the $C_6$ axis and perpendicular to the ring plane: three contain pairs of opposite atoms ($\sigma_v$) and three bisect opposite bonds ($\sigma_d$). At the exact center of the ring sits an inversion center, since every atom has a partner directly opposite through the centroid. Finally, because a genuine $C_6$ axis coexists with a perpendicular $\sigma_h$, the improper rotations $S_6$ and $S_3$ are both present automatically, generated as $\sigma_h C_6$ and $\sigma_h C_3$ respectively.

The order of a rotation axis is defined by the smallest angle of rotation that superimposes the molecule on itself:

where: $n$ is the order of the axis (an integer $\ge 2$ for any nontrivial axis) and $\theta$ is the rotation angle that leaves the molecule indistinguishable from its starting orientation.

Reading: to find $C_n$ axes in an unknown molecule, look for the largest $n$ for which rotating by $360^\circ/n$ maps every atom onto a chemically identical atom; that $n$ fixes the principal axis and, implicitly, every lower divisor axis it contains (a $C_6$ axis automatically contains $C_3$ and $C_2$ as $C_6^2$ and $C_6^3$). The improper rotation composition rule, which you have already met, is what explains why benzene's $S_6$ and $S_3$ "appear for free":

where: $S_n$ is rotation by $360^\circ/n$ followed by reflection through a plane perpendicular to that same axis, and the two operations on the right commute because the axis and the perpendicular plane share full rotational symmetry about that axis.

Reading: whenever a molecule is confirmed to have both a $C_n$ axis and a $\sigma_h$ perpendicular to it, the corresponding $S_n$ is guaranteed to exist — you never need to search for it independently once those two ingredients are both present. The count of symmetry-related mirror planes containing a principal $C_n$ axis follows directly from the axis order itself, because reflections and rotations generate each other around that axis:

where: $n$ is the order of the principal axis, and the total count of vertical-type planes (split between $\sigma_v$ through atoms and $\sigma_d$ through bond midpoints, or vice versa) equals $n$ whenever the molecule's substituent pattern supports a complete set.

Reading: ammonia's $C_3$ carries exactly $3$ vertical planes; benzene's $C_6$ carries exactly $6$ (three $\sigma_v$ through atom pairs, three $\sigma_d$ through bond midpoints) — this is not a coincidence but a direct consequence of the axis order. The perpendicular $C_2$ count obeys an analogous rule for axes of order $n \ge 3$ with a horizontal symmetry element present:

where: $n$ perpendicular two-fold axes intersect the principal $C_n$ axis at right angles, spaced by $360^\circ/n$ around it, whenever the molecule's geometry is "sandwich-like" or planar-hexagonal enough to support them (this is what elevates a $C_n$-only group to a dihedral-type arrangement).

Reading: benzene's six in-plane $C_2$ axes (three through atoms, three through bond midpoints) are the direct geometric expression of this rule with $n = 6$.

Example 1 — Foundational: hydrogen peroxide, H$_2$O$_2$, skew (non-planar) conformation. Take gas-phase H$_2$O$_2$ in its actual equilibrium geometry: the two O–H bonds are twisted out of coplanarity by a dihedral angle of about $111.5^\circ$ (an intentionally non-round, non-special number, since the exact experimental value varies with method but sits reliably between $90^\circ$ and $180^\circ$, never at a symmetric $180^\circ$ planar limit). Set up coordinates with the O–O bond along $z$. Test for a $C_2$ axis bisecting the O–O bond and perpendicular to it, lying in the plane that bisects the dihedral angle: rotating $180^\circ$ about this axis swaps O$_a \leftrightarrow$O$_b$ and, because the dihedral twist is symmetric about this bisecting plane, also swaps H$_a \leftrightarrow$H$_b$ correctly onto real atom positions — the skew geometry is exactly built to make this true. This $C_2$ is confirmed as the only rotation axis; there is no $C_3$ or higher (only two oxygens and two hydrogens are present, incapable of three-fold periodicity). Test for a mirror plane containing this $C_2$ axis and the O–O bond: since the dihedral angle is not $0^\circ$ or $180^\circ$, H$_a$ and H$_b$ sit on opposite "sides" in a genuinely twisted, chiral-like sense, and no mirror plane maps the molecule onto itself — reflecting would require reversing the handedness of the twist, landing on a geometrically distinct (mirror-image) conformer rather than the original. So there is no $\sigma$ of any kind. With no mirror plane, there is no possibility of $\sigma_h$, hence no $S_n$ can be built from the $C_2$, and testing inversion directly (send $(x,y,z)\to(-x,-y,-z)$ centered at the O–O midpoint) also fails, because the twisted, one-handed arrangement of the two O–H bonds is not centrosymmetric. The complete element list is therefore just $\{E, C_2\}$ — the sparse limiting case that will later be recognized as point group $C_2$, useful precisely because it shows that a real molecule can have only a bare rotation axis and nothing else.

Example 2 — Applied: trans-1,2-dichloroethylene, ClHC=CHCl (the trans isomer, planar alkene). This molecule is fully planar: both carbons, both chlorines, and both hydrogens lie in one plane, with the two chlorines on opposite sides of the C=C double bond (that is what "trans" means structurally). Search for the highest rotation axis: a $C_2$ perpendicular to the molecular plane, passing through the midpoint of the C=C bond, rotates the molecule $180^\circ$ and swaps C$_a\leftrightarrow$C$_b$, Cl$_a\leftrightarrow$Cl$_b$ (on the opposite corners, consistent with trans geometry), and H$_a\leftrightarrow$H$_b$ — every swap lands on a genuine like atom, confirming a valid $C_2$ axis perpendicular to the plane. No higher-order axis exists since there are only two of each atom type. Test the molecular plane itself as a mirror: every atom already lies in it, so it is trivially a valid $\sigma$; because it contains the $C_2$ axis found above (the axis is perpendicular to the plane, but wait — check carefully: the axis pierces the plane rather than lying in it, so this plane is perpendicular to the $C_2$, making it $\sigma_h$ by the standard convention that the principal axis defines "vertical" versus "horizontal"). Test for a second, perpendicular mirror plane containing the $C_2$ axis and standing perpendicular to the molecular plane: reflecting through this plane would need to swap C$_a\leftrightarrow$C$_b$ while leaving each atom's up/down position in the molecular plane fixed, but the actual trans substituent pattern (Cl up-left paired with H down-right, and vice versa) is not symmetric under that particular reflection — it would send Cl$_a$'s position to a spot where only H sits, so this candidate plane fails. So only one mirror plane exists, the molecular plane itself, correctly classified $\sigma_h$. Test inversion through the C=C bond midpoint: $(x,y,0)\to(-x,-y,0)$ sends Cl$_a$ (say at upper-left) to the lower-right position, which is exactly where Cl$_b$ sits in the trans arrangement — inversion succeeds. With a genuine $C_2$ and a genuine $\sigma_h$ confirmed, the product rule guarantees $S_2 = \sigma_h C_2$, but $S_2$ is by definition identical to $i$, which has already been found independently, consistent rather than redundant information. The full element list is $\{E,\, C_2,\, i,\, \sigma_h\}$, matching what is later named point group $C_{2h}$ — worth contrasting directly with the cis isomer, whose chlorines on the same side destroy both the $C_2$'s atom-matching and the inversion center while leaving only a single in-plane $\sigma_v$ and a perpendicular $C_2$, giving $C_{2v}$ instead. The lesson: cis/*trans* isomerism is precisely a symmetry-element difference, not just a geometric label.

Example 3 — Exam/competition level: the staggered conformation of ethane, C$_2$H$_6$. Freeze ethane in its lowest-energy staggered conformation (as opposed to eclipsed) and locate every element, being careful about the axes that are easy to miss. Place the C–C bond along $z$. Rotating $120^\circ$ about $z$ cycles each methyl group's three hydrogens among themselves ($H_{1a}\to H_{1b}\to H_{1c}\to H_{1a}$ on the front carbon, and independently the same on the back carbon), a genuine $C_3$ axis, with $C_3^2$ also valid. Now hunt for perpendicular $C_2$ axes — the step most commonly missed. Because the conformation is staggered, the front methyl's hydrogens sit rotated exactly $60^\circ$ relative to the back methyl's hydrogens (an offset, not a coincidence — it is the literal definition of staggered). Consider an axis perpendicular to $z$, passing through the C–C bond midpoint, oriented so that it bisects the angle between a front hydrogen and its nearest back hydrogen. A $180^\circ$ rotation about this axis swaps the two carbons with each other and simultaneously swaps each front hydrogen with the correctly staggered back hydrogen sitting at the complementary angle — every atom lands on a genuine like atom. By the three-fold symmetry of the methyl groups, there are three such perpendicular $C_2$ axes, spaced $60^\circ$ apart (not $120^\circ$, because each axis must thread between a front-H and a back-H pair, and the staggered offset places three such bisecting directions around the full $360^\circ$ circle). Next, test for a $\sigma_h$ perpendicular to $z$ at the C–C midpoint: this would need to swap the front carbon with the back carbon while leaving each hydrogen's azimuthal angle unchanged, but a $\sigma_h$ reflection does not rotate positions — it would require the front and back hydrogen sets to already sit at identical azimuthal angles, true only for the eclipsed conformer, not the staggered one. So staggered ethane has no $\sigma_h$. Test for vertical planes instead: a plane containing the $z$-axis and one front hydrogen, reflecting the other two front hydrogens into each other, would need to simultaneously map the back methyl's hydrogens onto themselves correctly — but the $60^\circ$ stagger means such a plane through a front-H does not pass through any back-H nor bisect the back set symmetrically; direct substitution shows the reflected back-hydrogen position lands at an azimuthal angle with no atom present. So staggered ethane has no $\sigma_v$ either, and no $\sigma_d$ of the ordinary kind. Instead, test dihedral-type improper rotation directly: an $S_6$ operation (rotate $60^\circ$ about $z$, then reflect through the perpendicular plane at the C–C midpoint) sends the front carbon to the back carbon's position (the $180^\circ$ combined effect... actually verify stepwise: $60^\circ$ rotation moves each front-H to a new azimuthal angle exactly matching a back-H's angle after reflection flips $z\to -z$, swapping front and back). Careful azimuthal bookkeeping confirms this composite operation does map every atom onto a genuine like atom, so $S_6$ is present (with $S_6^5$ its inverse partner), even though neither a bare $\sigma_h$ nor a bare $C_6$ individually exists — the hallmark of a "hidden" improper axis, exactly the subtlety flagged as a common trap below. Finally, test inversion at the bond midpoint: this is equivalent to $S_2$, and since $S_6^3 = S_2 = i$ by the power rule for improper axes, inversion is present too. The complete element list for staggered ethane is $\{E,\, 2C_3,\, 3C_2,\, i,\, 2S_6,\, ...\}$ (technically $S_6, S_6^5$, and the $S_2\equiv i$ already counted) — no $\sigma$ at all — which is later recognized as the point group $D_{3d}$, a case specifically chosen because it is the single most common place students wrongly guess "must have mirror planes" and miss the improper axis instead.

As a limiting case, imagine flattening ammonia's pyramid: shrink the H–N–H angle deformation to zero so nitrogen sits exactly in the plane of the three hydrogens. At that limit a new $\sigma_h$ appears (the plane itself), and the group jumps discontinuously in richness from $C_{3v}$-type behavior (no $\sigma_h$) to the planar $D_{3h}$-type behavior (with $\sigma_h$, extra $C_2'$ axes, and $S_3$) — a useful check that the absence of planarity in real ammonia is exactly why it lacks these extra elements, and confirms the derivation above did not silently assume planarity anywhere.

Molecular symmetry as a formal chemical tool traces back to the marriage of nineteenth-century crystallographic group theory with early twentieth-century quantum mechanics, but the person most responsible for making point-group symmetry a standard working tool in the chemist's toolkit — rather than a purely mathematical curiosity — was the physicist and crystallographer whose classification of space groups underlies all of modern structural symmetry analysis.