Box Model for Conjugated Molecules

The free-electron box model is far from a historical curiosity confined to textbooks; its core logic — count the delocalized electrons, estimate the delocalization length, predict the gap — remains the fastest back-of-envelope tool for thinking about any new conjugated chromophore. Photographic and textile dye chemists use the qualitative trend daily: extending a polymethine or polyene chromophore by one more vinylene unit reliably red-shifts absorption, a design principle that long predates and now sits comfortably alongside modern time-dependent density functional theory (TD-DFT) calculations used for precise quantitative prediction. Cyanine dyes of the kind analyzed in Example 2 are the workhorse fluorophores in modern biological imaging — Cy3, Cy5, and related near-infrared cyanine dyes used to label antibodies, oligonucleotides, and proteins in fluorescence microscopy and DNA sequencing owe their tunable absorption windows directly to the chain-length dependence this model predicts. Conducting polymers such as polyacetylene and polythiophene, central to organic electronics and organic photovoltaics, are conceptually the "$m\to\infty$" limit explored in Example 3: understanding why their band gaps saturate rather than closing entirely connects directly to the bond-alternation caveat in the traps table, and motivates the more sophisticated Su–Schrieffer–Heeger and Hückel-with-alternating-bonds treatments used in solid-state and polymer chemistry. In photoreceptor biochemistry, the extended conjugated polyene chain of retinal — the chromophore of rhodopsin responsible for vision — is a textbook case where the box-model gap estimate, refined with realistic bond lengths and a proper account of the protonated Schiff base terminus, gives a first quantitative handle on why the visual pigment absorbs where it does, before more accurate multireference quantum chemical methods are brought to bear. Even in a modern computational chemistry course built around Hückel Molecular Orbital Theory or full ab initio methods, the free-electron box model earns its keep as the fastest sanity check: if a proposed new chromophore's conjugation length changes, the box model instantly predicts which direction the color should shift, a first-pass filter before committing to expensive calculations or synthesis.

Ask the tutor: How would the HOMO–LUMO gap formula change for a cyclic conjugated system like benzene, where the box has periodic boundary conditions instead of hard walls?

Picture a linear polyene such as octatetraene, CH₂=CH–CH=CH–CH=CH–CH=CH₂. Draw the carbon backbone as a nearly straight line of eight sp² carbons connected by alternating formal single and double bonds. In reality, because of conjugation, the bonds are not cleanly "single" and "double" — the actual bond lengths are intermediate and roughly equalized, and the π system is a continuous cloud of electron density lying above and below the plane of the carbons, spanning the whole chain. This continuous π cloud is the physical object the box model replaces with a one-dimensional infinite square well: a region where the potential energy is zero, bounded by infinitely high walls just beyond the ends of the chain. Each carbon contributes one 2p orbital and, in the neutral hydrocarbon, one electron to the π system (each double bond of the Lewis structure supplies two π electrons total, and the whole conjugated backbone pools them). These π electrons are treated as free particles confined to the 1D box; they do not interact directly with each other in this crude but useful approximation, and they experience zero potential energy anywhere inside the box and infinite potential energy at and beyond the ends. Because the electrons are indistinguishable fermions, the Pauli exclusion principle allows at most two electrons (spin ↑ and spin ↓) in each of the box's quantized energy levels, so the electrons fill the ladder of levels from the bottom up, exactly as electrons fill atomic orbitals in the aufbau principle. The length of the box, $L$, is not simply the distance between the terminal carbon nuclei. Because the π wavefunction physically extends somewhat beyond the last carbon (into the region of the terminal C–H σ bonds and beyond), a standard refinement adds one extra bond length as an "end correction," half a bond length of overhang on each side. With $N_{\text{bonds}}$ conjugated bonds (both formal single and double) of average length $d$, the working formula for the effective box length is

This is the same $L$ that enters the box energy formula from the particle-in-a-box result; nothing about the quantization itself changes — only the identification of what "$L$" means physically for a real molecule.

For octatetraene, eight π electrons fill levels $n=1$ through $n=4$ (two electrons each), so the HOMO is $n=4$ and the LUMO is $n=5$. The transition drawn as the purple arrow in the figure is exactly the $n=4 \to n=5$ excitation whose energy is computed in the worked examples below. A second useful picture is how the level spacing itself changes as the box lengthens. Because the particle-in-a-box energies go as $n^2/L^2$, both stretching $L$ and increasing $n$ push levels closer together in absolute spacing between adjacent-$n$ neighbors, and the HOMO–LUMO gap shrinks smoothly and monotonically as conjugation length grows — the quantum mechanical root of the empirical rule "longer conjugation, redder color."

The starting point is the ordinary particle-in-a-box energy formula, unchanged from its abstract derivation:

where: $n$ is the quantum number of a π-electron level, $h$ is Planck's constant, $m_e$ is the electron mass, and $L$ is the effective box length of the conjugated chain. Reading it: energy rises as the square of the quantum number and falls as the square of the box length, so both "how high up the ladder" and "how long the wire" control the spacing between levels. The electron count sets which levels are filled. For a chain with $N$ π electrons (an even number, since π bonds always come in pairs of electrons for closed-shell neutral polyenes), the levels $n=1$ through $n=N/2$ are doubly occupied and the level $n = N/2+1$ is the lowest unoccupied one:

where: $N$ is the total number of π electrons in the conjugated system (one per contributing 2p orbital, i.e. one per sp²/sp carbon in the chain). Reading it: doubling the electron count does not simply double the HOMO level — the HOMO index scales linearly with electron number while, crucially, $L$ also grows with chain length, and the two effects together give the overall gap-narrowing trend. The absorption energy is the gap between these two levels:

where: the bracketed difference of squares collapses algebraically to $N+1$ (shown in the derivation below), so the whole gap depends on just two structural facts about the molecule — its electron count $N$ and its box length $L$. Reading it: this is the master formula of the free-electron molecular orbital (FEMO) model; every quantitative prediction in this article traces back to it. Finally, converting the gap into an observable wavelength uses the ordinary photon relation:

where: $c$ is the speed of light and $\lambda$ is the predicted wavelength of the lowest-energy π→π* absorption band. Reading it: $\lambda$ grows as $L^2$ but shrinks only as $1/(N+1)$, and because $L$ itself grows with $N$ (adding a bond adds both an electron and a length increment), the net effect of chain growth is still a strong, monotonic red-shift — visible directly in the trend from Figure 2.

Example 1 — Foundational: hexatriene, done by hand from first principles. 1,3,5-hexatriene, CH₂=CH–CH=CH–CH=CH₂, has $N=6$ π electrons (three double bonds, six contributing carbons) and $N_{\text{bonds}} = 5$ conjugated bonds linking them. Using the average bond length $d = 1.40$ Å = $1.40\times10^{-10}$ m and the end-corrected box length,

With $N=6$ electrons, the HOMO is $n=3$ and the LUMO is $n=4$ (three levels doubly filled: $n=1,2,3$). The gap formula from the derivation gives

Compute the numerator: $h^2 = 4.391\times10^{-67}\ \text{J}^2\text{s}^2$, times 7 gives $3.074\times10^{-66}$. Compute the denominator: $8 \times 9.109\times10^{-31} = 7.287\times10^{-30}$; $L^2 = 7.056\times10^{-19}\ \text{m}^2$; their product is $5.142\times10^{-48}$. Dividing,

Converting to a wavelength via $\lambda = hc/\Delta E$, with $hc = 1.986\times10^{-25}\ \text{J m}$,

The experimental $\lambda_{\max}$ for 1,3,5-hexatriene in solution is close to 258 nm for the strong π→π* band (with vibronic structure spreading it further), so the naive box model overshoots here, predicting a redder absorption than observed — a known systematic feature discussed in the traps section, driven mainly by the fixed uniform bond-length assumption and the neglect of electron repulsion. Even so, the order of magnitude (near-UV) and the qualitative existence of a single dominant transition are both captured correctly.

Example 2 — Applied: a cyanine dye, showing the model's real strength. The box model's best quantitative performance is historically for symmetric cyanine dyes, $\text{R}_2\text{N}^+=\text{CH–(CH=CH)}_p\text{–NR}_2$, because the positive charge is delocalized over the whole conjugated chain including both nitrogen lone pairs, giving a cleaner "wire with electrons at both ends" than a neutral polyene whose π system is not quite uniform. Consider a cyanine with $p=2$ (two vinylene units between the nitrogens), giving a conjugated chain of the two nitrogens plus the intervening carbons. Such a dye has a conjugated system contributing $N = 8$ π electrons (including the nitrogen lone pair electrons entering the π system, following the standard cyanine electron count of $2p+4$) delocalized over $N_{\text{bonds}} = 7$ bonds. Using the same $d=1.40$ Å,

The HOMO is $n=4$, LUMO is $n=5$. The gap:

Numerator: $4.391\times10^{-67} \times 9 = 3.952\times10^{-66}$. Denominator: $7.287\times10^{-30} \times (1.254\times10^{-18}) = 9.140\times10^{-48}$. So

Then

This lands squarely in the blue-absorbing region of the visible spectrum — cyanine dyes of this chain length are indeed vivid orange-to-red dyes in real formulations, because absorbing blue light around 450–460 nm leaves the complementary orange/red light transmitted. Historically, this class of dye is exactly where Kuhn's free-electron model achieved its most celebrated quantitative agreement with experiment, often to within a few percent for the longer cyanines, better than for neutral polyenes, precisely because the delocalized positive charge makes the "flat, zero-potential wire" approximation more realistic.

Example 3 — Competition-level: extracting a general scaling law and testing self-consistency across a homologous series. Consider a general linear polyene series $\text{H(CH=CH)}_m\text{H}$ (so $m=1$ is ethylene, $m=2$ is butadiene, $m=3$ is hexatriene, $m=4$ is octatetraene, and so on), with $N = 2m$ π electrons and $N_{\text{bonds}} = 2m-1$ bonds, so

Substitute into the general gap formula:

As a check on the algebra, examine the large-$m$ limiting behavior. For large $m$, $(2m+1) \to 2m$ to leading order, so

This shows the box model predicts the HOMO–LUMO gap ultimately decays as $1/m$ for very long polyenes — not $1/m^2$ as one might naively guess from the $1/L^2$ dependence alone, because the electron count $N=2m$ also grows linearly with $m$ and partially cancels one power of the length-squared decay. This $1/m$ asymptotic law is a genuine, checkable prediction: plotting $\Delta E$ against $1/m$ for a real polyene series should be closer to linear than plotting $\Delta E$ against $1/m^2$, and this is indeed the qualitative shape seen in tabulated polyene UV data, even though the absolute gap values from the crude model are systematically too small compared to experiment for long chains (the model over-predicts the red-shift once bond-length alternation becomes chemically important, since real long polyenes do not achieve full bond equalization). As an explicit self-consistency check at finite $m$, compare $m=3$ (hexatriene, from Example 1: $\Delta E = 5.978\times10^{-19}$ J) against the general formula directly:

Using $d = 1.40\times10^{-10}$ m, $d^2 = 1.96\times10^{-20}\ \text{m}^2$; $32\, m_e\, d^2 = 32(9.109\times10^{-31})(1.96\times10^{-20}) = 5.714\times10^{-49}$. Then $h^2/(32m_ed^2) = 4.391\times10^{-67}/5.714\times10^{-49} = 7.686\times10^{-19}$ J, and multiplying by $7/9 = 0.7778$ gives $5.978\times10^{-19}$ J — matching Example 1 exactly, confirming the general formula reduces correctly to the specific hexatriene case.