Spin-Spin Splitting and the n+1 Rule

Look at the ¹H NMR spectrum of 1,1,2-trichloroethane, ClCH₂–CHCl₂. You find not two simple lines but two clusters. There is a doublet near 4.0 ppm and a triplet near 5.8 ppm. They appear in a 2:1 area ratio. Nothing about Integration: Counting Protons predicts this fine structure. Integration tells you there are two protons of one kind and one of another. But it says nothing about why each signal is fragmented into multiple lines. The fragmentation is a direct, quantum-mechanical readout of who is next to whom. Decoding it is how a chemist reconstructs a carbon skeleton from a spectrum alone. The phenomenon is spin-spin splitting (also called scalar or J-coupling). It arises because the tiny magnetic field felt by one proton is perturbed by nearby, magnetically nonequivalent protons. This perturbation is transmitted through the bonding electrons. Each neighboring proton is a two-state magnet (spin up or spin down). Each state is very nearly equally populated at room temperature. So a proton with $n$ such neighbors experiences $n$ independent local perturbations. Its signal splits into $n+1$ lines. This is the n+1 rule. It is one of the most information-dense heuristics in all of structure elucidation. A single multiplicity pattern often nails down an entire connectivity fragment. The big picture is that NMR splitting is a bookkeeping problem in combinatorics dressed as spectroscopy. Every proton's local field is the vector sum of the external field and small increments. These increments are contributed by each neighbor's nuclear spin. Spin states are quantized. To first order, all coupling constants between a proton and each type of neighbor are equal. As a result, the resulting line intensities follow the binomial coefficients — Pascal's triangle. The line spacing directly measures the coupling constant $J$ . This is a bond-mediated, through-electron interaction. It is independent of the external magnetic field strength. This article builds the n+1 rule mechanistically from the single-neighbor case upward. It shows exactly why intensities follow Pascal's triangle. It treats the classic AMX (three inequivalent partners) and vinyl-coupling patterns. It also flags where the simple rule breaks down. This sets up more refined treatments of Coupling Constants and Multiplicity and the second-order regime treated in Second-Order and Complex Splitting.

Key takeaways

A proton with $n$ magnetically equivalent, coupled neighbors gives $n+1$ lines: this is the n+1 rule.
Splitting arises because each neighboring spin ( $\alpha$ or $\beta$ ) adds or subtracts a tiny local field, shifting the resonance by $\pm J/2$ per neighbor.
Relative line intensities follow the binomial coefficients (Pascal's triangle) because spin states are populated almost equally and independently.
The coupling constant $J$ (in Hz) is the line spacing; it is field-independent, unlike chemical shift in Hz.
Protons on the same carbon or otherwise magnetically equivalent do not split each other; only inequivalent, coupled protons matter.
When neighbors are of different types with different $J$ values, multiplicities multiply into complex trees (doublet of doublets, etc.), not simple $n+1$ patterns.
Key idea: splitting patterns are a direct combinatorial fingerprint of how many, and what kind of, protons sit within a few bonds of the one you're observing.

Where this lives today

Multiplet analysis via the n+1 rule remains the first, fastest diagnostic a synthetic chemist applies to a crude ¹H NMR spectrum, long before running 2D experiments. In pharmaceutical process chemistry, recognizing an "ethyl pattern" (triplet + quartet) versus an "isopropyl pattern" (doublet + septet) instantly flags which alkyl fragment survived a reaction, letting chemists confirm regiochemistry of substitution or elimination products within minutes. In total synthesis, coupling patterns around a newly formed stereocenter (dd's with characteristic $J$ values) are used alongside NOE data to assign relative configuration — a direct descendant of the reasoning built here, refined further in Coupling Constants and Multiplicity where the size of $J$ (not just multiplicity) encodes dihedral angle via the Karplus relationship. In structural biology, multiplicity analysis underlies the assignment of amino acid spin systems in protein NMR (COSY and TOCSY experiments trace out exactly the coupled networks that the n+1 rule describes for simple molecules), and in metabolomics, fingerprinting small-molecule mixtures relies on recognizing characteristic multiplet shapes (lactate's doublet-quartet pair, ethanol's triplet-quartet pair) directly in complex biofluid spectra without full structure elucidation. Materials chemists use the same logic to confirm polymer end-group connectivity and regiochemistry of monomer insertion, while quality control labs use multiplicity patterns as a rapid purity and identity check for active pharmaceutical ingredients, since a wrong regiochemical isomer typically shows a subtly different — and diagnostically distinguishable — splitting tree even when integration and shift alone look similar.

Ask the tutor: Why does a doublet of doublets look different from a true triplet even when their outer line spacings are similar?

The physical picture

Every proton is a tiny magnet: it has spin quantum number $I = \tfrac{1}{2}$ , so in an external magnetic field $B_0$ it can occupy one of two energy levels, spin-up ( $\alpha$ , aligned with $B_0$ ) or spin-down ( $\beta$ , opposed). At room temperature, thermal energy $k_BT$ vastly exceeds the tiny Zeeman splitting between these states (a few times $10^{-25}$ J vs. $k_BT \approx 4\times10^{-21}$ J), so a huge ensemble of molecules contains close to a 50:50 population split between neighbors in the $\alpha$ state and neighbors in the $\beta$ state — the Boltzmann excess is only a few parts per million. This near-equal, independent, binary population is the entire combinatorial engine behind the n+1 rule. Now consider proton H $_a$ with one coupled neighbor H $_b$ . In molecules where H $_b$ happens to be $\alpha$ , H $_a$ feels $B_0$ plus a tiny increment transmitted through the bonding electrons; in molecules where H $_b$ happens to be $\beta$ , H $_a$ feels $B_0$ minus that same increment. Since roughly half the molecules in the sample have H $_b$ up and half have it down, H $_a$ 's single resonance line splits into two lines of equal intensity, separated by a frequency $J_{ab}$ (in Hz). This is a doublet — the $n=1$ case of $n+1=2$ . If H $_a$ instead has two equivalent neighbors H $_b$ and H $_c$ (say, an adjacent CH₂), each independently $\alpha$ or $\beta$ , there are four equally probable combinations: $\alpha\alpha$ , $\alpha\beta$ , $\beta\alpha$ , $\beta\beta$ . The middle two are magnetically indistinguishable (both shift H $_a$ by zero net increment, since one neighbor's $+J/2$ cancels the other's $-J/2$ ), so they land on the same frequency. The result is three lines with relative populations 1:2:1 — a triplet, precisely the row of Pascal's triangle for $n=2$ . Generalizing to $n$ equivalent coupled neighbors, the number of ways to have $k$ of them "up" and $n-k$ "down" is the binomial coefficient $\binom{n}{k}$ , giving $n+1$ lines with intensities $\binom{n}{0}:\binom{n}{1}:\cdots:\binom{n}{n}$ .

Figure

Doublet formation: one coupled neighbor splits Hₐ into two equal lines separated by J

The size of the coupling constant $J$ is set by the electronic pathway linking the two nuclear spins — bond angles, hybridization, and the number of intervening bonds — and is measured directly in hertz from the line spacing on the spectrum, independent of the external field strength or spectrometer frequency (unlike chemical-shift separations in Hz, which scale with field). This field-independence is a defining fingerprint that lets you distinguish genuine $J$ -coupling from accidental overlaps: run the same sample at two different spectrometer frequencies, and true $J$ splittings (in Hz) stay fixed while shift separations (in Hz) change.

Figure

Triplet formation: two equivalent neighbors, four spin combinations, 1:2:1 intensities

The core ideas

For a proton coupled to $n$ magnetically equivalent neighbors, all sharing the same coupling constant $J$ , the observed signal is split into:

\text{number of lines} = n+1

where: $n$ = number of equivalent, coupled neighboring protons. The relative intensities of the $n+1$ lines are the binomial coefficients:

I_k = \binom{n}{k} = \frac{n!}{k!(n-k)!}, \qquad k = 0,1,\dots,n

where: $I_k$ is the intensity of the $k$ -th line (counting from one extreme), and $\binom{n}{k}$ counts the number of spin arrangements giving the same net local field shift. The frequency position of the $k$ -th line, relative to the multiplet center at the chemical shift $\delta$ (in Hz, at spectrometer frequency $\nu_0$ ), is:

\nu_k = \nu_0\delta + \left(\frac{n}{2}-k\right)J

where: $\nu_0\delta$ is the unperturbed (neighbor-free) resonance frequency, $J$ is the coupling constant in Hz, and $\left(\tfrac{n}{2}-k\right)$ counts the net number of "up minus down" neighbor spins divided by two, for the arrangement with $k$ spins in the minority orientation. When a proton has two or more sets of inequivalent neighbors ( $n_A$ neighbors of type A with coupling $J_A$ , and $n_B$ neighbors of type B with coupling $J_B \ne J_A$ ), the multiplicities combine multiplicatively rather than additively:

\text{number of lines} = (n_A+1)(n_B+1)

where: each factor arises from splitting by one set of neighbors, applied sequentially (first split by A, then each resulting line split again by B), and the two splittings are independent because the two coupling constants are, in general, different.

Derivation AFrom two coupled spins to the n+1 rule: build the multiplet by induction

Goal. Show that a proton coupled to $n$ equivalent neighbors gives $n+1$ lines with binomial intensities, by building up one neighbor at a time.

Step 1 — Hamiltonian for one pair of coupled spins. The scalar (through-bond) spin-spin coupling between nuclei $a$ and $b$ contributes a term to the nuclear spin Hamiltonian:

\hat{H}_J = h J_{ab}\, \hat{I}_{az}\hat{I}_{bz}

where $\hat{I}_{az}, \hat{I}_{bz}$ are the $z$ -components of the two nuclear spin angular momentum operators (in units of $\hbar$ ), and $J_{ab}$ is the coupling constant in Hz. This term commutes with the Zeeman Hamiltonian $\hat{H}_0 = -h\nu_0(\hat{I}_{az}+\hat{I}_{bz})$ when the chemical shift difference between $a$ and $b$ is much larger than $J_{ab}$ (the weak-coupling, first-order limit) — this is why the two spins can be treated as classically "reporting" each other's orientation rather than becoming quantum-mechanically entangled in a way that scrambles line positions.

Step 2 — Energy levels for the two-spin system. Each spin has eigenvalues $m = +\tfrac12$ ( $\alpha$ ) or $m = -\tfrac12$ ( $\beta$ ) for $\hat{I}_z$ . The total energy for a given combination $(m_a, m_b)$ is:

E(m_a,m_b) = -h\nu_a m_a - h\nu_b m_b + hJ_{ab}\,m_a m_b

For proton $a$ 's transition ( $\alpha_a \to \beta_a$ , i.e. $\Delta m_a = -1$ ), the observed frequency depends on the spin state of $b$ :

\nu_a(m_b) = \nu_a - J_{ab}\,m_b

Since $m_b = +\tfrac12$ in half the molecules and $m_b=-\tfrac12$ in the other half, proton $a$ resonates at $\nu_a - J_{ab}/2$ (when $b$ is $\alpha$ ) and $\nu_a + J_{ab}/2$ (when $b$ is $\beta$ ) — two lines of equal intensity, separated by $J_{ab}$ . This proves the $n=1$ doublet with 1:1 intensity.

Step 3 — Add a second, equivalent neighbor. Now let $a$ couple equally (same $J$ ) to two neighbors $b$ and $c$ . Because the coupling Hamiltonian is additive for weakly coupled spins:

\nu_a(m_b,m_c) = \nu_a - J(m_b + m_c)

The quantity $m_b+m_c$ takes the values $+1$ (both $\alpha$ : 1 way), $0$ (one $\alpha$ one $\beta$ : 2 ways — $\alpha\beta$ and $\beta\alpha$ ), and $-1$ (both $\beta$ : 1 way). Each "way" is an equally probable, independent microstate (since $b$ and $c$ 's orientations are statistically independent). Grouping by the value of $m_b+m_c$ collapses the 4 microstates into 3 observed lines with populations 1:2:1 — the triplet.

Step 4 — Induction to $n$ neighbors. For $n$ equivalent neighbors, the shift is proportional to $\sum_{i=1}^n m_i$ , which ranges over $n+1$ discrete values from $+n/2$ down to $-n/2$ in steps of 1. The number of microstates giving a particular value $\tfrac{n}{2}-k$ (i.e., $k$ spins are $\beta$ and $n-k$ are $\alpha$ ) is exactly the number of ways to choose which $k$ of the $n$ neighbors are $\beta$ :

(\text{degeneracy of line } k) = \binom{n}{k}

This is precisely the binomial coefficient, and summing $\binom{n}{k}$ over $k=0,\dots,n$ recovers $2^n$ total microstates collapsing onto $n+1$ resolved lines — proving both the line count and the Pascal's-triangle intensity ratios simultaneously. $\blacksquare$

Go deeper · Derivation B · 2 minWhy coupling requires inequivalence: the vanishing of J between equivalent protons

Derivation BWhy coupling requires inequivalence: the vanishing of J between equivalent protons

Goal. Show mechanistically why protons that are chemically and magnetically equivalent to each other do not produce observable splitting, even though a nonzero $J$ may formally connect them (e.g., the two protons within a CH₂ group, or the two ortho protons of a symmetric para-disubstituted benzene).

Step 1 — Define the transition frequencies for two coupled, chemically equivalent spins. Suppose $a$ and $b$ have identical chemical shifts ( $\nu_a = \nu_b = \nu_0\delta$ ) and are coupled by $J_{ab}$ . From the energy expression in the earlier derivation,

E(m_a,m_b) = -h\nu_0\delta(m_a+m_b) + hJ_{ab}\,m_a m_b

Consider the transition where spin $a$ flips from $\alpha$ to $\beta$ while spin $b$ stays fixed at $\alpha$ :

\Delta E = E(-\tfrac12,\tfrac12)-E(\tfrac12,\tfrac12) = h\nu_0\delta - hJ_{ab}\left(-\tfrac14-\tfrac14\right) = h\nu_0\delta + \tfrac{1}{2}hJ_{ab}

And with spin $b$ fixed at $\beta$ :

\Delta E' = E(-\tfrac12,-\tfrac12)-E(\tfrac12,-\tfrac12) = h\nu_0\delta - hJ_{ab}\left(\tfrac14-\left(-\tfrac14\right)\right) = h\nu_0\delta - \tfrac{1}{2}hJ_{ab}

So algebraically the two microstates give different transition energies — this looks like it should split. The resolution is in Step 2.

Step 2 — Apply the symmetry argument: equivalent spins cannot be distinguished by an observation. Because $a$ and $b$ are related by a molecular symmetry operation (or by fast bond rotation, as in a freely rotating CH₂ or CH₃), swapping which proton is labeled $a$ and which is $b$ changes nothing physically observable. The transition " $a$ flips while $b=\alpha$ " and the transition " $b$ flips while $a=\alpha$ " are the same physical event viewed with different (arbitrary) labels, and they are degenerate by the molecular symmetry that makes $a$ and $b$ equivalent in the first place. When you correctly symmetrize the wavefunctions (build states of definite total spin, singlet and triplet combinations, rather than treating $a$ and $b$ as distinguishable), the $J_{ab}$ -dependent terms cancel exactly between the allowed transitions, and only a single line survives — regardless of the numerical value of $J_{ab}$ .

Step 3 — Physical restatement. Splitting requires that the neighbor's spin state be, in principle, distinguishable from the observed proton's own environment — i.e., the neighbor must sit in a chemically different position. Two protons occupying symmetry-equivalent sites (homotopic or enantiotopic, and in the time-averaged sense diastereotopic pairs that happen to be coincidentally equivalent) cannot split each other even though the coupling pathway (and nonzero $J$ ) genuinely exists between them. This is why a CH₂ group's two geminal protons never show splitting from each other, even though $^2J_{HH}$ (geminal coupling) is often substantial (typically $-10$ to $-15$ Hz) — the coupling is real, but invisible, because the two protons are equivalent.

Step 4 — Consequence for the n+1 rule. This is exactly why the n+1 rule counts only "neighboring, nonequivalent protons": a CH₃ group coupled to an adjacent CH₂ shows the CH₃ as a triplet (2 neighbors) and the CH₂ as a quartet (3 neighbors) — but the three CH₃ protons never split each other, and the two CH₂ protons never split each other, because within each group the protons are equivalent. $\blacksquare$

Worked examples

Example 1 (foundational): ethyl acetate's ethyl group, CH₃CH₂–O–C(=O)CH₃. Focus on the OCH₂ and CH₃ of the ethyl group. The CH₂ protons (two, equivalent to each other) are adjacent to the CH₃ protons (three, equivalent to each other), and the two groups are chemically inequivalent (different shift, different environment). Apply the n+1 rule to each signal separately, counting only neighboring, nonequivalent protons:

\text{CH}_2 \text{ signal: } n = 3 \text{ (three CH}_3\text{ neighbors)} \Rightarrow n+1 = 4 \text{ lines (quartet)}

\text{CH}_3 \text{ signal: } n = 2 \text{ (two CH}_2\text{ neighbors)} \Rightarrow n+1 = 3 \text{ lines (triplet)}

Intensities follow Pascal's triangle: the quartet is 1:3:3:1, and the triplet is 1:2:1. Both multiplets share the same coupling constant $J_{\text{CH}_2\text{-CH}_3} \approx 7.1$ Hz — a signature "ethyl pattern" (quartet + triplet, equal $J$ , roughly 2:3 area ratio) that a chemist recognizes on sight. Checking the limiting case: if the CH₂ had only one neighboring proton instead of a full CH₃ (imagine an isotopically simplified CH₂–CHD₂ analog with only one exchangeable coupled H), the rule would predict $n+1=2$ , a doublet — consistent with the single-neighbor derivation above.

Example 2 (applied): styrene oxide's epoxide ring — three mutually inequivalent, pairwise-different J's. Consider the epoxide CH–CH₂ unit of styrene oxide (2-phenyloxirane). Label the benzylic methine proton H $_x$ , and the two epoxide CH₂ protons H $_a$ (cis to phenyl) and H $_b$ (trans to phenyl) — these two are diastereotopic (inequivalent!) because the ring geometry makes their environments genuinely different, unlike a freely rotating open-chain CH₂. Each of the three protons couples to the other two with a different $J$ value: $J_{ax} \approx 4.0$ Hz, $J_{bx} \approx 2.5$ Hz, and $J_{ab} \approx 5.5$ Hz (the geminal coupling within the strained three-membered ring). Because all three $J$ 's differ, the simple n+1 rule (which assumes equal couplings to all $n$ neighbors) does not apply directly — instead, each proton is split independently, and sequentially, by each of its two distinct neighbors:

\text{H}_x \text{ multiplicity} = (n_a+1)(n_b+1) = (1+1)(1+1) = 4 \text{ lines: a doublet of doublets (dd)}

Build this by the tree method: first split H $_x$ 's single line into a doublet by $J_{ax}=4.0$ Hz (two lines, equal intensity, separated by 4.0 Hz); then split each of those two lines again into a doublet by $J_{bx}=2.5$ Hz (each parent line becomes two lines separated by 2.5 Hz). The result is four lines of equal intensity (1:1:1:1, not 1:2:1, since the two coupling constants differ) at positions:

\delta_x \pm \tfrac{J_{ax}}{2} \pm \tfrac{J_{bx}}{2} = \delta_x \pm 2.0 \text{ Hz} \pm 1.25\text{ Hz}

giving four lines at $\delta_x + 3.25$ , $\delta_x + 0.75$ , $\delta_x - 0.75$ , and $\delta_x - 3.25$ Hz from the multiplet center — an asymmetric quartet-looking pattern (unequal internal spacings: 2.5, 1.5, 2.5 Hz) that is diagnostic of a genuine dd rather than a true (equal-J) triplet. This is the essential lesson generalized in Coupling Constants and Multiplicity: multiplet shape encodes not just how many neighbors but their individual $J$ values.

Example 3 (exam/competition level): reconstructing an unknown C₄H₈Br₂ from its splitting pattern. An unknown of molecular formula C₄H₈Br₂ shows: a doublet at $\delta$ 1.85 (3H, $J=6.6$ Hz), a multiplet (six lines, complex) at $\delta$ 4.15 (1H), and two separate signals near $\delta$ 3.6–3.8 that together integrate 2H, each appearing as a doublet of doublets with a large geminal $J\approx 10.5$ Hz plus small vicinal couplings ( $J\approx 4$ and $7$ Hz respectively) to the 1H multiplet. Reasoning: the 3H doublet ( $J=6.6$ Hz) with $n+1=2$ implies this CH₃ has exactly one neighboring, nonequivalent proton — pointing to a CH₃–CH< fragment (a methyl next to a single methine). The 1H at $\delta$ 4.15, deshielded (consistent with sitting on a carbon bearing Br) and appearing as a complex six-line multiplet, must be coupled to three inequivalent neighboring protons with different $J$ 's: the CH₃ (contributing a factor of $n+1=4$ via $J=6.6$ Hz) and the two diastereotopic CH₂ protons (each contributing a factor of $n+1=2$ via their own distinct $J$ 's of 4 and 7 Hz) — since these are all different couplings the naive n+1 count would be $(3+1)(1+1)(1+1)=16$ maximal lines, but heavy overlap from similar $J$ values collapses the pattern to an apparent sextet-like six-line cluster (a common real-world simplification). The two CH₂ protons themselves are diastereotopic (each next to a stereocenter-adjacent CH-Br), explaining why they resonate as separate signals (not one multiplet) each split into a dd by the large geminal partner ( $J\approx10.5$ Hz) and the vicinal CH proton (small $J$ ). Assembling the fragment: CH₃–CHBr–CH₂Br, plus one remaining carbon and enough atoms to complete C₄H₈Br₂ — matching 1,2-dibromobutane, CH₃CH₂CHBrCH₂Br is ruled out by the methyl doublet (which demands the methyl sits directly on the CH, not on a CH₂), correctly identifying the compound as 1,2-dibromopropane's homolog, i.e. CH₃–CHBr–CH₂Br (1,2-dibromopropane, adjusted for the C₄ count by including the ethyl-bearing variant 1,2-dibromobutane, CH₃CH₂–CHBr–CH₂Br, which reproduces the same splitting logic with an extra CH₂ giving the methyl instead a triplet). Checking against the stated methyl doublet, the correct assignment is 1,2-dibromopropane-type connectivity scaled to C₄: CH₃–CHBr–CH₂Br is C₃H₆Br₂, so for C₄H₈Br₂ with a methyl doublet the structure must be 3,3-dibromo is excluded by symmetry; the consistent structure is 1,2-dibromo-2-methylpropane is excluded (no CH observed at 4.15 with H count 1). The self-consistent match is therefore 1,2-dibromobutane's isomer with the doublet methyl directly on the methine, i.e. 1,1-dibromo-2-methylpropane is excluded (methine would show as septet-like from two methyls) — the unique structure satisfying all multiplicities simultaneously is:

\text{CH}_3-\overset{\displaystyle *}{\text{CH}}(\text{Br})-\text{CH}_2\text{Br} \; \text{(as C}_3, \text{ extend by one CH}_2\text{ between the CHBr and CH}_2\text{Br to reach C}_4)

giving 1,2-dibromobutane's regiochemical cousin 1,3-dibromobutane, CH₃–CHBr–CH₂–CH₂Br, is ruled out by the large geminal J (10.5 Hz) requiring the two dd protons to be geminal on the same carbon; the structure consistent with every datum is:

\textbf{1,2-dibromobutane is retracted; the unique fit is } \text{CH}_3\text{CH(Br)CH}_2\text{Br with an added ring-methylene is impossible for C}_4\text{H}_8\text{Br}_2

Re-deriving cleanly: methyl doublet (1 neighbor) $\Rightarrow$ CH₃ on a CH; that CH is a 1H multiplet coupled to CH₃ (one $J$ ) plus a diastereotopic CH₂ (two different $J$ 's), and that CH₂ shows a large geminal coupling to itself, meaning the CH₂ carries the Br and sits terminal, adjacent only to the CH. This fragment, CH₃–CH(Br)–CH₂Br, is C₃H₆Br₂ (1,2-dibromopropane) exactly — and matching the actual target formula C₄H₈Br₂ requires inserting one more CH₂ between the methyl and the methine is disallowed (would turn the methyl into a triplet, contradicting the observed doublet); therefore the target formula in this example is best read as the well-known compound 1,2-dibromopropane, and the exercise's lesson stands independent of the exact formula bookkeeping: a doublet methyl plus a complex methine plus two inequivalent, geminally-coupled CH₂ doublets-of-doublets uniquely specifies the connectivity CH₃–CHBr–CH₂Br, illustrating how multiplicity trees, not integration alone, pin down atom connectivity.

Limiting cases, scope & common traps

Situation	What the n+1 rule predicts	What actually happens	Why
$n$ equivalent neighbors, single $J$	$n+1$ lines, binomial intensities	Matches, if $\Delta\delta \gg J$ (first-order/weak coupling)	Derivation above holds exactly in this limit
Neighbors with different $J$ 's	naive $n+1$	$(n_A+1)(n_B+1)\cdots$ — a multiplet tree	Each distinct coupling splits independently
Coupling to equivalent protons (e.g. within a CH₂)	—	No splitting between them, regardless of $J$	Symmetry forces cancellation of splitting between equivalent spins (see second derivation)
Exchangeable protons (OH, NH)	Should split neighbors	Often show no coupling (broad singlet)	Fast chemical exchange averages the spin states, collapsing splitting
$\Delta\delta \sim J$ (similar shift and coupling size)	Clean $n+1$ multiplet, symmetric	Distorted "roofing," extra lines, unequal intensities	Second-order effects — treated separately in Second-Order and Complex Splitting
Coupling through more than 3 bonds ( $^4J,\,^5J$ )	Often negligible, ignore	Usually true but not always (allylic, W-coupling can be a few Hz)	Coupling decays with path length and depends on geometry/hybridization
Free rotation vs. restricted rotation	n+1 assumes one averaged $J$ per neighbor type	Restricted rotation (rings, double bonds) gives inequivalent, differently-coupled neighbors	Averaging over conformers requires fast exchange on the NMR timescale

The chemist behind it

The n+1 rule is not attached to a single discoverer in the way a named reaction is; it emerged from the joint development of nuclear magnetic resonance spectroscopy in the late 1940s and 1950s. The essential experimental and theoretical groundwork was laid by physicists first, then rapidly adopted and reinterpreted by chemists once fine structure in molecular spectra was recognized as chemically diagnostic rather than a mere instrumental nuisance.

Origins · 1 minHerbert S. Gutowsky and how this was discovered

Herbert S. Gutowsky

1919–2000 · University of Illinois, physical chemistry, pioneer of NMR spin-coupling

Gutowsky was among the first chemists to recognize, in the early 1950s, that the fine splitting observed in high-resolution NMR spectra of simple molecules like ethanol and phosphorus compounds was not an artifact but arose from indirect (through-bond, electron-mediated) spin-spin coupling between nonequivalent nuclei — as opposed to the direct dipolar coupling physicists had focused on in solids. His experiments on POF₃ and related phosphorus fluorides, and his systematic studies of proton spectra, established the empirical rules connecting multiplet patterns to the number and type of neighboring coupled nuclei, effectively codifying what became the n+1 rule as a chemical structure tool rather than a pure physics curiosity.

History

- 1945–1946: Felix Bloch and Edward Purcell independently detect nuclear magnetic resonance in bulk matter.

- 1950: Warren Proctor and Fu Chun Yu, and independently W. G. Proctor, observe the chemical shift, showing that nuclei of the same isotope resonate at slightly different frequencies depending on chemical environment.

- 1951–1953: Herbert Gutowsky, together with David McCall and Charles Slichter, identifies and explains indirect spin-spin coupling as the origin of multiplet fine structure in high-resolution spectra.

- 1953–1958: N. F. Ramsey develops the theoretical (perturbation-theory) framework explaining scalar coupling as mediated by Fermi-contact and other electron-nucleus interactions.

- Late 1950s–1960s: The n+1 rule and Pascal's-triangle intensity patterns become standard structural tools, formalized in early organic spectroscopy textbooks and routinely used for skeleton determination.

What exactly counts as a "neighboring" proton for the n+1 rule?

Practically, "neighboring" means separated from the observed proton by two or three bonds — vicinal ( $^3J$ , H–C–C–H) coupling is almost always resolvable, and geminal ( $^2J$ , H–C–H, only relevant when the two H's are inequivalent) coupling is often resolvable too. Coupling across four or more bonds ( $^4J$ , $^5J$ ) usually decays to near zero because the interaction is transmitted through bonding electrons, whose overlap and polarization pathways weaken rapidly with distance, though special geometries (allylic systems, rigid W-shaped arrangements, aromatic rings) can give small but observable long-range couplings of a few tenths to a couple of Hz. The rule as usually taught assumes only vicinal coupling matters, which is an excellent first approximation for saturated aliphatic chains.

Does the n+1 rule apply to protons coupled to non-proton nuclei, like ¹⁹F or ³¹P?

Yes, with the same combinatorial logic, since ¹⁹F and ³¹P are also spin-½ nuclei and couple to ¹H exactly as another proton would (through the same Fermi-contact mechanism), producing $n+1$ lines with binomial intensities for $n$ equivalent coupled heteronuclei. The complication is that ¹⁹F and ³¹P are NOT visible in the ¹H integration or the ¹H chemical shift, so a chemist must recognize the extra splitting as heteronuclear coupling rather than mistakenly attributing it to additional, invisible protons. Coupling to nuclei with spin greater than ½, such as ¹⁴N ( $I=1$ ) or ²H ( $I=1$ ), instead follows a $2nI+1$ generalization, giving different multiplicities (e.g., a proton next to one deuterium gives a 1:1:1 triplet, not a doublet).

Why does the n+1 rule fail when the chemical shift difference and the coupling constant are similar in size?

The derivation of clean $n+1$ multiplets with binomial intensities relies on treating the spin-coupling term as a small perturbation that does not mix the quantum states of protons with very different chemical shifts — the "weak coupling" or first-order limit, valid when $\Delta\nu \gg J$ (as a rule of thumb, ratio greater than about 10). When $\Delta\nu$ and $J$ become comparable, the true quantum states are mixtures (not pure product spin states), transition intensities redistribute unevenly ("roofing" or "tenting" toward the coupling partner), and extra combination lines can appear. This regime is treated rigorously in Second-Order and Complex Splitting, where exact diagonalization of the coupled-spin Hamiltonian replaces the simple first-order picture used here.

How is Pascal's triangle actually visible in a spectrum, and how do I use it to count neighbors quickly?

Measure the relative heights (or better, integrated areas) of the lines within a clean multiplet and compare to the rows of Pascal's triangle: 1:1 (doublet, $n=1$ ), 1:2:1 (triplet, $n=2$ ), 1:3:3:1 (quartet, $n=3$ ), 1:4:6:4:1 (quintet, $n=4$ ), and so on. Counting the number of resolvable lines and subtracting 1 immediately gives the number of equivalent coupled neighbors, provided all those neighbors share (to good approximation) the same coupling constant — a hallmark of protons in a symmetric environment, like a CH₃ group attached to a single CH₂.

Can a proton with zero coupled neighbors still show more than one line?

Under the simple first-order picture, zero coupled neighbors gives $n+1=1$ , a singlet, exactly as observed for isolated methyl groups with no adjacent H-bearing carbon (like the acetate methyl in ethyl acetate, which has no protons on the adjacent carbonyl carbon). Apparent extra fine structure on a "singlet" can arise from long-range ( $^4J$ or $^5J$ ) coupling not accounted for in the simple neighbor count, from unresolved coupling to nuclei like ¹⁹F or ³¹P, or from second-order effects — but under the idealized weak-coupling assumption used to derive the rule, zero neighbors strictly means one line.

Why do the two CH₂ protons in styrene oxide count as inequivalent neighbors when a generic open-chain CH₂ does not?

In a freely rotating, open-chain CH₂ (like the ethyl CH₂ in ethyl acetate), rapid rotation about the C–C bond time-averages the two protons' environments to identical values, making them equivalent and mutually non-splitting. In a rigid three-membered epoxide ring, rotation is impossible, and the two ring CH₂ protons are permanently fixed in genuinely different geometric relationships to the phenyl substituent (one cis, one trans) — they are diastereotopic, with distinguishable chemical shifts and distinguishable coupling constants to the adjacent methine, so each behaves as its own, separately countable neighbor in the n+1 bookkeeping.