The Neutral-Loss Spectrum

An interactive explainer for raw GC-MS and LC-MS data: what a neutral-loss scan is, how the hardware produces it, and how to read one without fooling yourself.

The 60-second version

When an ion fragments in a tandem mass spectrometer it breaks into a charged piece (which the instrument can see) and a neutral piece (which it cannot — no charge, no steering, no detection). The neutral piece is only revealed by the mass gap between precursor and fragment: the neutral loss, Δm.

A neutral-loss spectrum is a filtered spectrum: on a triple-quadrupole, Q1 and Q3 are scanned together, always Δm apart. A compound reaches the detector only if it loses exactly Δm in the collision cell. The output looks like an MS¹ spectrum (intensity vs. precursor m/z) but contains only members of one structural class — e.g. Δm = 98 Da shows only things that lose H₃PO₄ → phosphopeptides; Δm = 176 Da → glucuronide metabolites.

1 · Why “neutral” losses are invisible

Mass spectrometers manipulate and detect charged particles with electric and magnetic fields. A neutral fragment keeps the precursor's velocity, drifts out of the ion beam, and is pumped away — it never produces a peak. You only ever infer it from arithmetic:

[M+H]⁺ at m/z 305  →  fragment⁺ at m/z 261 (detected)   +  CO₂ · 44 Da (neutral — invisible, inferred from 305 − 261)

The gap of 44 Da is a fingerprint: something in that molecule was a carboxyl group. Different chemistries give characteristic gaps — water (18), ammonia (17), CO₂ (44), SO₃ (80), H₃PO₄ (98), glucuronic acid (176)… That turns a mass difference into a class test.

Analogy: a security checkpoint that only counts travelers who shed exactly one 18 kg suitcase in the tunnel. You never weigh the suitcase — you weigh people before and after, and the 18 kg difference tells you who was carrying one.

GC-EI note: in classical GC-MS the molecular ion M⁺• and its fragments are all born in the ion source, so one EI spectrum already contains these gaps (M−15, M−18, M−29…). Reading them is neutral-loss analysis — see the interactive explorer in section 5.

2 · The triple-quadrupole neutral-loss scan — simulator

Below is a virtual Q–q–Q. Q1 scans precursor mass M upward; q2 is a collision cell filled with N₂/Ar; Q3 scans in lock-step at M − Δm. Watch the readouts under Q1 and Q3 stay exactly Δm apart. Colored dots are ions; gray dots are neutral fragments (invisible to a real detector). Ions of a compound only enter q2 when the scan reaches their m/z — that's Q1 doing its filtering job.

Pick a neutral loss above, then press ▶ Start scan.

What to watch for

  • Q1 and Q3 move together, glued Δm apart. Nothing else about the hardware changes during the scan.
  • In q2, each precursor fragments. Only the fragments whose mass equals the current Q3 setting (= precursor − Δm) survive Q3 and hit the detector (flash). Wrong-mass fragments are deflected into the rods.
  • Analyte C has no labile neutral group — it is abundant in the full scan (top panel) but never appears in any neutral-loss spectrum. Absence of a peak ≠ absence of the compound.
  • Analyte D and the phosphopeptide lose two different neutrals — they show up in two different NL spectra.
  • Peak height ∝ abundance × fragmentation yield, not abundance alone. NL spectra are screening tools, not quantitative ones.
Singly charged ions are used here for clarity. For charge state z, the m/z offset is Δmass/z — a doubly charged phosphopeptide losing H₃PO₄ (98 Da) shifts by 49 m/z, a classic gotcha (see quiz!).

3 · How to read a neutral-loss spectrum

  • X-axis = precursor m/z (same axis as an MS¹ scan). Y-axis = detector counts.
  • A peak at m/z M asserts two things at once: a precursor of mass M exists and it ejects Δm under CID → the molecule contains the corresponding functional group.
  • It is a class filter on a dirty matrix: e.g. a 176 Da NL scan of urine extract shows glucuronide metabolites and hides almost everything else.
  • It tells you nothing about where in the molecule the loss came from, and gives no full fragment fingerprint — follow up hits with product-ion scans or HRMS.
  • QqQ scans run at unit resolution: isobaric coincidences cause false positives. A peak at M could be an unrelated M′ that happens to satisfy M′ − fragment′ = Δm. Confirm.
  • Intensity depends on collision energy and fragmentation yield — don't compare abundances across compound classes.

4 · Common neutral losses worth memorizing

Δm (Da)Neutral speciesTypical compound classWhere you meet it
17.03NH₃amines, amino acids, peptidesLC-ESI & CI
18.01H₂Oalcohols, carboxylic acids, peptideseverywhere — GC-EI & LC-ESI
~28CO (27.99) / C₂H₄ (28.03)aldehydes, ketones / ethyl groupsGC-EI
32.04CH₃OHmethyl esters, methoxy compoundsGC-EI
44.00CO₂free carboxylic acids, carbamatesLC-ESI, GC-EI
56.06C₄H₈ (isobutene)Boc-protected aminesLC-ESI (med-chem)
63.96SO₂sulfonamides, sulfonesLC-ESI
79.96SO₃sulfate conjugates, sulfonatesLC-ESI (metabolite ID)
97.98H₃PO₄phosphopeptides (pSer/pThr), phosphate metabolitesLC-ESI (phosphoproteomics)
129.04C₅H₇NO₃ (pyroglutamic acid)glutathione (GSH) adductsLC-ESI (reactive-metabolite screening)
162.05C₆H₁₀O₅ (hexose)O-glycosides, flavonoid glycosidesLC-ESI (natural products)
176.03C₆H₈O₆ (glucuronic acid)glucuronide conjugatesLC-ESI (drug metabolism)

5 · The same idea, two different worlds

GC-MS

  • Most GC-MS systems are single-quadrupole EI — there is no collision cell and no Q3, so a hardware NL scan doesn't exist (you'd need a GC-QqQ).
  • Instead, neutral losses live inside every EI spectrum: read the gaps below the molecular ion. M−15 (•CH₃), M−18 (H₂O), M−29 (•C₂H₅/CHO), M−31 (•OCH₃), M−43, M−60 (acetic acid from acetates), M−90 (TMS-OH from silylated analytes).
  • In chemical ionization, [M+H]⁺ often loses H₂O, NH₃, etc. — same arithmetic.
  • On high-res GC-Orbitrap/Q-TOF systems you can compute constant-neutral-loss views in software.

LC-MS

  • ESI + triple quadrupole is the classic home of the NL scan: 98 (or 49 for z = 2) for phosphopeptides, 80 for sulfates, 176 for glucuronides, 129 for GSH adducts, 44 for acids.
  • Use it to screen a class out of a filthy matrix, then characterize hits with product-ion scans.
  • Q-TOFs and Orbitraps can't link-scan; you acquire DDA/DIA and mine constant neutral losses in software (MS-DIAL, mzMine, Compound Discoverer, MassHunter…).
  • Mostly qualitative — quantitation is done by MRM instead.

GC-EI explorer — the “other” neutral-loss spectrum

Hover the peaks of this stylized EI spectrum (molecular ion M⁺• = 158). Then flip the axis: re-plotting every peak at M − m/z converts an ordinary spectrum into a neutral-loss spectrum — the software meaning of the term you'll meet in data-processing tools.

6 · Check yourself

Score: 0/5

7 · Going deeper

Charge states: why Δm/z = Δmass ÷ z matters

Quadrupoles select by m/z, not mass. A neutral loss of 98 Da appears as an offset of 98 for z = 1, 49 for z = 2, 32.7 for z = 3. Phosphopeptide NL methods on QqQ instruments therefore run two linked scans (98 and 49) to catch singly and doubly charged precursors. In EI-GC-MS everything is z = 1, so the gap equals the mass directly.

Where NL scan sits among the four QqQ scan modes
ModeQ1q2Q3What you get
Product ionfixed on MCIDscansfull fragment fingerprint of one precursor
Precursor ionscansCIDfixed on fall precursors that produce fragment f
Neutral lossscansCIDscans, locked at Q1 − Δmall precursors that lose Δm — a structural class
MRM / SRMfixed on MCIDfixed on fone specific transition — the quantitation workhorse
Practical tuning tips (LC-QqQ)
  • Collision energy is class-specific: too low → no fragmentation, no signal; too high → secondary losses (e.g. phosphopeptides losing H₃PO₄ and H₂O) and weaker signal. Optimize on a standard.
  • Scan speed trades against duty cycle and sensitivity; don't outrun your chromatographic peak (≥ 10 points across it).
  • Unit resolution → confirm hits by product-ion scan or accurate mass. Isobaric coincidences are the main artifact.
  • Common paired strategy: NL survey scan → automatically trigger product-ion scans on detected peaks (information-dependent acquisition).
Common GC-EI neutral losses (read as M⁺• minus fragment)
Gap (Da)LostSuggests
15•CH₃methyl branch; ubiquitous in TMS derivatives
17•OHalcohols, acids
18H₂Oalcohols (often with M−15 companion)
29•C₂H₅ / •CHOethyl chains / aldehydes
31•OCH₃methyl esters, methoxy
35/36•Cl / HClchlorinated compounds (check 3:1 isotope)
43CH₃CO• / C₃H₇•methyl ketones / propyl chains
60CH₃COOHacetate esters, alditol acetates (sugar analysis)
90TMS-OHsilylated –OH groups (derivatized GC-MS)