NMR and MRI physics · Intermediate

J-coupling: why peaks split (doublet, triplet, quartet)

Written by SignalsLab

Early edition: these lessons are being reviewed.

Content

Summary

J-coupling is the through-bond interaction between neighbouring nuclear spins that splits one NMR resonance into a multiplet. A nucleus with n equivalent neighbouring spin-1/2 nuclei is split into n+1 lines with Pascal's triangle intensities. The spacing between lines, J in hertz, is independent of the magnetic field.

Intuition

Picture a proton as a tiny compass needle. Its neighbour, a few bonds away, is also a tiny magnet that can point with or against the applied field. Electrons in the bonds carry this information across, so the first proton feels a slightly stronger or weaker local field depending on the neighbour's orientation. The two orientations are almost equally likely, so half of the molecules in the sample resonate a little higher and half a little lower, giving two lines: a doublet. With two neighbours there are four combinations: up-up, up-down, down-up, down-down. The two mixed ones have the same effect, so the middle line is twice as tall: a triplet 1:2:1. Three neighbours give 1:3:3:1, a quartet. The splitting is a fixed energy interaction, so it is measured in Hz and does not grow when you use a stronger magnet, unlike chemical shift in Hz.

Key equations

Nlines=n+1
(1)

Number of lines from n equivalent neighbouring spin-1/2 nuclei

Δν=J
(2)

Spacing between adjacent lines of a multiplet equals the coupling constant in Hz

E=JmAmX
(3)

Energy shift from coupling depends on the spin orientations of the two nuclei (m = ±1/2)

Δδ(ppm)=J(Hz)/ν0(MHz)
(4)

Splitting in ppm shrinks as spectrometer frequency rises

Derivation

  1. Take spin A coupled to one spin X. X has orientation m = +1/2 or -1/2, with almost equal populations.
  2. The coupling adds an energy term J m_A m_X to the system, so A's transition frequency becomes ν0 + J/2 or ν0 − J/2 depending on X.
  3. Two equally populated frequencies separated by J give a 1:1 doublet.
  4. Add a second equivalent neighbour X2. The combinations are ↑↑ (+J), ↑↓ and ↓↑ (0), ↓↓ (−J), so the lines sit at ν0+J, ν0, ν0−J with weights 1:2:1.
  5. For n neighbours, the number of ways to get k spins up is the binomial coefficient C(n,k), producing n+1 lines with Pascal's triangle intensities.

Worked example

Problem

In ethyl groups, the CH2 protons appear as a quartet with lines at 4.130, 4.118, 4.106 and 4.094 ppm on a 400 MHz spectrometer. Find J and the number of neighbouring protons.

  1. Line spacing = 4.130 − 4.118 = 0.012 ppm.
  2. Convert to Hz: 0.012 ppm × 400 MHz = 4.8 Hz... use exact: 0.012 × 400 = 4.8 Hz.
  3. A quartet has 4 lines, so n + 1 = 4 and n = 3 neighbours, matching the CH3 group.
  4. On a 100 MHz instrument the spacing would still be 4.8 Hz but 0.048 ppm.

Answer J ≈ 4.8 Hz (a typical value is about 7 Hz in real ethyl groups, so these figures are illustrative), with 3 neighbouring protons.

Common mistakes

  • Thinking J grows with magnetic field strength; it is constant in Hz, only ppm spacing shrinks.
  • Counting lines as n instead of n+1.
  • Expecting equal intensities in triplets and quartets rather than 1:2:1 and 1:3:3:1.
  • Expecting splitting between equivalent protons in the same group; they do not split each other.
  • Forgetting that coupling to non-spin-1/2 nuclei, or to far away protons, follows different rules.

Animation

Animation coming soon

The animated version of this lesson is being prepared for the web. In the Signals AI app, lessons are drawn step by step.

Code

scene.py(64 lines)
import sys
from math import comb
import numpy as np
from manim import *
from blocks import *

BG = "#0e1117"


def multiplet(n, J=1.0, nu0=0.0):
    """Return line positions and intensities for n equivalent spin-1/2 neighbours."""
    ints = [comb(n, k) for k in range(n + 1)]
    pos = [nu0 + (k - n / 2) * J for k in range(n + 1)]
    # one step per neighbour: list of (positions) after splitting
    steps = [[nu0]]
    cur = [nu0]
    for _ in range(n):
        cur = [p + s * J / 2 for p in cur for s in (-1, 1)]
        steps.append(cur)
    return pos, ints, steps


class Main(Scene):
    def construct(self):
        self.camera.background_color = BG
        TitleCard(self, "J-coupling", "Why peaks split")
        J = 1.0
        scale = 1.3
        rows = []
        names = ["singlet (n=0)", "doublet (n=1)", "triplet (n=2)", "quartet (n=3)"]
        ys = [1.8, 0.6, -0.6, -1.8]
        for n in range(4):
            pos, ints, _ = multiplet(n, J)
            lines = VGroup()
            for p, h in zip(pos, ints):
                x = -2.5 + p * scale
                lines.add(Line([x, ys[n], 0], [x, ys[n] + 0.22 * h + 0.1, 0],
                               color=C_ACCENT, stroke_width=6))
            lab = Text(names[n], font_size=24, color=WHITE)
            lab.move_to([3.6, ys[n] + 0.25, 0])
            rows.append((lines, lab))
            base = Line([-5.3, ys[n], 0], [0.3, ys[n], 0], color=C_DIM, stroke_width=2)
            self.add(base)
        StepCaption(self, "Each neighbouring spin splits a line by J")
        for n in range(4):
            lines, lab = rows[n]
            self.play(Create(lines), FadeIn(lab), run_time=0.9)
        StepCaption(self, "Intensities follow Pascal: 1:1, 1:2:1, 1:3:3:1")
        self.wait(1.5)
        StepCaption(self, "n neighbours give n+1 lines, spaced by J")
        self.wait(1)


if __name__ == "__main__" and "--selftest" in sys.argv:
    for n in range(5):
        pos, ints, steps = multiplet(n)
        assert len(pos) == n + 1
        assert sum(ints) == 2 ** n
        assert len(steps[-1]) == 2 ** n
        assert abs(np.mean(pos)) < 1e-9
    assert multiplet(2)[1] == [1, 2, 1]
    assert multiplet(3)[1] == [1, 3, 3, 1]
    assert abs(multiplet(1)[0][1] - multiplet(1)[0][0] - 1.0) < 1e-9
    print("ok")

The Manim scene code that draws this lesson's animation in the Signals AI app (Python, Manim Community).

ccBYNC© 2026 SignalsLab. Lesson text shared as is under CC BY-NC 4.0.