Fragmentation Mechanisms and Diagnostic Losses

Fragmentation mechanism reasoning is not a historical curiosity confined to EI-MS of small organics; it is the working logic behind tandem mass spectrometry (MS/MS) used throughout modern biochemistry and pharmaceutical analysis. In proteomics, collision-induced dissociation of peptide ions fragments preferentially at the amide backbone to generate b- and y-ion series, and the reasons those bonds break preferentially (charge-directed, mobile-proton mechanisms analogous in spirit to alpha cleavage) are taught using exactly the same cation/radical-stability logic developed here for small molecules. In drug metabolism studies, LC-MS/MS fragmentation patterns of a parent drug and its metabolites are compared peak-by-peak, and a metabolite that shows the same diagnostic loss (say, loss of 44 for a carboxylic acid decarboxylation-type fragmentation, or loss of 18 for dehydration) as the parent compound is strong structural evidence that the modification occurred at a site remote from that fragmenting bond. Environmental and forensic mass spectrometry relies on exactly the same reasoning to identify unknown contaminants or drugs of abuse from fragmentation libraries: a spectrum showing m/z 91 (tropylium), a loss of 43 (isopropyl or acetyl), and a loss of 18 (water) together sharply constrains candidate structures before any comparison to a reference library is even made. Petrochemical quality control still uses alpha-cleavage and McLafferty patterns to determine branching patterns in hydrocarbon mixtures rapidly, without full chromatographic separation. And in the classroom, this fragmentation-mechanism toolkit is the natural bridge into combining-spectroscopic-methods, where a mass spectral fragmentation pattern, an infrared carbonyl stretch, and an NMR splitting pattern are triangulated together — mass spectrometry supplying the molecular formula and connectivity clues that infrared and NMR alone cannot.

Ask the tutor: How do I decide, from a real unknown spectrum, whether a given neutral loss came from alpha cleavage or from a McLafferty rearrangement when both could in principle give the same nominal mass loss?

Electron ionization (EI) removes an electron from whichever molecular orbital has the highest energy — generally a nonbonding lone pair (on O, N, S, halogens) if one exists, otherwise a π orbital, and only as a last resort a σ bond. This ordering, $n > \pi > \sigma$ in ease of ionization, is the single most useful fact for predicting where the radical cation's "hole" sits, because that is where fragmentation initiates. Once the hole is created, the ion has an odd number of electrons and a formal positive charge distributed (in a localized picture) mostly at the site of ionization. Two things can then happen in a unimolecular fragmentation: a bond can break homolytically, sending one electron to each fragment (this is the arrow-pushing convention using single-barbed "fishhook" arrows, and it generates a new radical plus keeps the charge on one piece), or a bond can break heterolytically, with both electrons going to one fragment — but heterolytic cleavage of a bond within a radical cation is mechanistically different from ordinary closed-shell heterolysis, because one of the two "pushed" species already carries the radical. In practice, most simple cleavages in mass spectrometry are described as: the radical cation's odd electron (or an adjacent lone pair) assists departure of a group as a radical, while the charge stays on the fragment best able to stabilize a cation — usually via resonance, hyperconjugation, or an adjacent heteroatom lone pair. This is exactly the same cation-stability logic used for $S_N1$ solvolysis and carbocation rearrangements: tertiary > secondary > primary, benzylic/allylic (resonance) beats all simple alkyl cations, and an adjacent lone pair (oxocarbenium, iminium, thionium formation) beats everything else. The difference in mass spectrometry is that the other product is a neutral radical, not a nucleophile-trapped closed-shell molecule, and the whole process happens in the gas phase, unimolecularly, with no solvent to stabilize anything — so intrinsic (gas-phase) stability differences are often even larger than their solution-phase counterparts.

Notice the asymmetry that the figure is built to expose: alpha cleavage of an unsymmetrical ketone can in principle break either C–C bond flanking the carbonyl, but it overwhelmingly breaks the one that ejects the larger, more substituted alkyl radical, because that leaves the charge on the acylium ion $\text{R–C}{\equiv}\text{O}^+$, which is stabilized by full oxygen lone-pair donation into the empty carbon p-orbital regardless of which R is retained — so the radical stability (not the cation) actually decides which bond breaks, since the cationic fragment (acylium) is comparably stabilized either way when the same carbonyl remains attached. For 2-hexanone this means the m/z 43 acylium ($\text{CH}_3\text{C}{\equiv}\text{O}^+$) dominates over the alternative m/z 85 fragment, because forming a primary methyl radical is disfavored relative to forming a more stable secondary/primary n-butyl radical pathway complexities aside — in practice the more substituted C–C bond cleaves preferentially, and both empirical rules ("more substituted radical leaves" and "more stable cation forms") point the same direction for most real ketones.

The tropylium picture is the cleanest illustration of why gas-phase cation stability can look different from naive intuition: the initially formed benzylic cation $\text{PhCH}_2^+$ (still a six-membered aromatic ring with an exocyclic $+$) undergoes ring expansion to the seven-membered tropylium cation, which distributes its positive charge over seven equivalent carbons and is aromatic (Hückel $4n+2$, $n=1$, 6 π electrons) rather than merely benzylic. This is why essentially every alkylbenzene with a chain long enough shows a prominent m/z 91 peak, independent of chain length or branching — the initial fragmentation site (Cα–Cβ bond, "benzylic cleavage") is common to all of them, and the destination cation is the same aromatic sink.

Every fragmentation obeys strict mass and charge bookkeeping. If the molecular ion has mass $M$ and charge $+1$, and it splits into a detected cationic fragment of mass $m^+$ and a neutral (radical or closed-shell) fragment of mass $m_0$, then

where:

• $M$ = nominal mass of the molecular ion (equal, to within the mass of one electron, to the molecular weight from the formula),
• $m^+$ = m/z of the observed daughter ion (assuming singly charged, the overwhelmingly common case),
• $m_0$ = mass of the neutral fragment lost (never detected directly, only inferred).

Reading a spectrum is therefore an exercise in solving

and matching $m_0$ against a short list of chemically reasonable neutral losses (Table below), each with its own mechanism. The second governing relationship is qualitative but just as rigorous: among the accessible fragmentation pathways from a given radical cation, the dominant one is the one whose transition state (or, for simple cleavages with no real barrier beyond the bond-breaking itself, whose product) has the lowest energy, which almost always tracks cation stability

where:

• more substituted, resonance-stabilized, or heteroatom-adjacent carbons make better cations,
• more substituted or resonance-stabilized carbons also make better radicals (radical stability order tracks cation stability order: 3° > 2° > 1° > methyl, benzylic/allylic ≈ best of all),
• so bonds that can simultaneously deposit charge on one stabilized fragment and the radical on the other stabilized fragment cleave fastest.

Because both cation and radical stability point the same direction (more substituted wins for both), you rarely need to separately optimize them — the qualitative rule "cleave to expose the most stable possible cation, ejecting whatever radical results" almost always agrees with "cleave off the most stable possible radical, leaving whatever cation results."

Example 1 — Foundational: 2-methylbutan-2-ol, deducing the base peak from alpha cleavage. The alcohol $(\text{CH}_3)_2\text{C(OH)CH}_2\text{CH}_3$ has molecular formula $\text{C}_5\text{H}_{12}\text{O}$, so

Ionization removes an electron from an oxygen lone pair, placing the radical cation character on O. Alpha cleavage requires breaking a C–C bond at the carbon bearing the OH (the alpha carbon, here the tertiary carbinol carbon), ejecting an alkyl radical and leaving an oxocarbenium-stabilized cation. Three different alpha C–C bonds are available at this tertiary carbinol carbon: two equivalent C–CH$_3$ bonds and one C–CH$_2$CH$_3$ bond. Cleaving a methyl group ejects $\text{CH}_3^{\bullet}$ (mass 15) and leaves

Cleaving the ethyl group ejects $\text{CH}_3\text{CH}_2^{\bullet}$ (mass 29) and leaves

Both cations are resonance-stabilized oxocarbenium ions (the oxygen lone pair conjugates into the empty p-orbital on the former carbinol carbon in each case), so the deciding factor is which radical is easier to expel: ethyl radical is a more stable (better hyperconjugated, marginally more substituted-adjacent) radical than methyl, but critically there are two methyl groups statistically available versus one ethyl, and more importantly loss of the larger group is empirically and mechanistically favored in alpha cleavage of alcohols because the transition state benefits from greater radical stabilization with increasing substitution at the radical carbon itself — here methyl vs. ethyl only differs by one carbon, but the well-established rule ("largest group lost preferentially in alpha cleavage") predicts the m/z 59 peak (loss of ethyl) should be the dominant one, and it is: the base peak of 2-methyl-2-butanol's EI spectrum sits at m/z 59, confirming loss of the larger (ethyl) group. As a limiting-case check: if instead the alcohol were symmetric, e.g., 3-ethylpentan-3-ol with three identical ethyl groups on the carbinol carbon, alpha cleavage would give only one possible fragment mass regardless of which bond breaks, since all three groups lost are identical — a useful sanity check that the general two-pathway analysis above correctly collapses to a single answer when the substituents become equivalent.

Example 2 — Applied: distinguishing pentan-2-one from pentan-3-one by McLafferty rearrangement. Both ketones share the molecular formula $\text{C}_5\text{H}_{10}\text{O}$, $M = 5(12)+10(1)+16 = 60+10+16=86$, and cannot be told apart by molecular ion mass alone — exactly the situation where fragmentation mechanism becomes the decisive tool. Pentan-2-one, $\text{CH}_3\text{C}(=\text{O})\text{CH}_2\text{CH}_2\text{CH}_3$: numbering out from the carbonyl carbon along the propyl chain gives Cα (CH$_2$), Cβ (CH$_2$), Cγ (CH$_3$, terminal methyl). The γ-hydrogens exist on this terminal methyl, so a six-membered-ring McLafferty transfer is geometrically accessible. The rearrangement transfers a γ-H to the carbonyl oxygen and cleaves the Cα–Cβ bond, ejecting neutral ethylene ($\text{CH}_2{=}\text{CH}_2$, mass 28) and leaving the enol cation $\text{CH}_3\text{C(OH)=CH}_2^{+\bullet}$, mass

Pentan-3-one, $\text{CH}_3\text{CH}_2\text{C}(=\text{O})\text{CH}_2\text{CH}_3$: this ketone is symmetric, with an ethyl group on each side of the carbonyl. Numbering out either side: Cα (CH$_2$), Cβ (CH$_3$). There is no Cγ on either side — the chain is too short to reach a γ-hydrogen — so no McLafferty rearrangement is possible for pentan-3-one. Its EI spectrum shows alpha cleavage products (loss of ethyl radical, mass 29, giving the acylium $\text{CH}_3\text{CH}_2\text{C}{\equiv}\text{O}^+$ at m/z 57) but conspicuously lacks any m/z 58 McLafferty peak. The diagnostic conclusion: a strong m/z 58 peak (loss of 28 from $M=86$) signals pentan-2-one; its complete absence, combined with a strong m/z 57 (loss of 29), signals pentan-3-one. This is a clean illustration of how a missing peak is as diagnostic as a present one, because the mechanism (need for a γ-hydrogen reachable in a six-membered ring) is a hard geometric requirement, not a statistical tendency.

Example 3 — Exam/competition level: multistep deduction for an unknown ester, C$_7$H$_{14}$O$_2$. An EI-MS shows: $M^{+\bullet}$ at m/z 130 (weak), a peak at m/z 88 (loss of 42 from M), a peak at m/z 61, and a strong peak at m/z 43. Molecular formula $\text{C}_7\text{H}_{14}\text{O}_2$ gives

consistent with the observed weak molecular ion (esters often show weak $M^{+\bullet}$ because the initially formed radical cation fragments rapidly). Degrees of unsaturation:

consistent with one C=O (an ester carbonyl, no ring or additional unsaturation), so the compound is a saturated aliphatic ester. Work the losses one at a time, each tied to a mechanism: Loss of 42 (130 → 88). A loss of exactly 42 mass units from an ester molecular ion, arriving at m/z 88, is the classic signature of a McLafferty rearrangement on the alkyl (alcohol-derived) side or alkoxy-side elimination of an alkene, but more diagnostically, loss of 42 from an ester frequently corresponds to McLafferty loss of propene ($\text{C}_3\text{H}_6$, mass 42) from the acid-derived alkyl chain, requiring a γ-hydrogen counted from the carbonyl carbon on the acyl side — analogous exactly to the 2-hexanone case worked above, but now with the acyl chain playing the role of the ketone's alkyl chain and an –OR group replacing one methyl. This tells us the acid portion of the ester has at least three carbons past the carbonyl carbon with an accessible γ-H, i.e., the acyl group is derived from an acid of the form $\text{R'CH}_2\text{CH}_2\text{CH}_2\text{C(=O)O–}$ or similar (butanoate or longer, or a branched equivalent placing a CH at the γ position). Peak at m/z 43. A fragment at m/z 43 from an ester is very commonly the acylium ion $\text{CH}_3\text{C}{\equiv}\text{O}^+$ (from an acetate ester, i.e., the acid portion is acetic acid, $\text{CH}_3\text{CO–}$) or the isopropyl cation $(\text{CH}_3)_2\text{CH}^+$ (from alpha cleavage on the alkoxy side of an isopropyl ester) or $\text{C}_3\text{H}_7^+$ generally. Given that we already deduced (from the loss of 42) that the acid side needs a three-carbon chain past the carbonyl for a McLafferty γ-H, the acyl group is unlikely to be the two-carbon acetate; instead, the m/z 43 peak is best assigned to alpha cleavage that ejects the alkoxy oxygen's alkyl radical and leaves an acylium — or, working the arithmetic directly: if the ester is butyl acetate-like in reverse substitution (i.e., a pentyl or hexyl acetate), check against the formula. Testing hexyl formate or, more systematically, matching $\text{C}_7\text{H}_{14}\text{O}_2$ with an acyl chain long enough for the loss of 42: propyl butanoate, $\text{CH}_3\text{CH}_2\text{CH}_2\text{C}(=\text{O})\text{O–CH}_2\text{CH}_2\text{CH}_3$, fits — the acid-derived butanoyl chain has Cα, Cβ, Cγ (the terminal methyl of the butanoyl group) giving exactly the required γ-H for McLafferty loss of propene (42) from the acyl side, reproducing the 130 → 88 transition. Alpha cleavage at the ester oxygen (breaking the O–CH$_2$ bond of the propoxy group, ejecting propyl radical, mass 43, or retaining it) is also available; specifically, alpha cleavage that breaks the acyl C–O bond, keeping the electrons to form the acylium $\text{CH}_3\text{CH}_2\text{CH}_2\text{C}{\equiv}\text{O}^+$, gives mass

not matching 61 or 43 directly, so we instead assign m/z 43 to the propoxy-side cleavage forming the propyl cation $\text{C}_3\text{H}_7^+$ (mass $3(12)+7(1)=43$) via loss of the neutral acyl radical $\text{CH}_3\text{CH}_2\text{CH}_2\text{C(=O)O}^{\bullet}$ (mass $130-43=87$). The m/z 61 peak corresponds to protonated propanoic acid–type fragment $\text{CH}_3\text{CH}_2\text{C(OH)}{=}\text{OH}^+$ arising from McLafferty-type or inductive cleavage retaining the carbonyl and both oxygens with the propyl portion of the acyl chain: $\text{C}_2\text{H}_5\text{COOH}_2^+$-type ion at mass $2(12)+5(1)+12+2(16)+1 = 24+5+12+32+1=74$ (not 61) — refining, the m/z 61 ion is more simply $\text{CH}_3\text{CH}_2\text{C(OH)}_2^+$ or the common ester-alpha-cleavage acylium-plus-oxygen ion $\text{C}_2\text{H}_5\text{OCH}{=}\text{OH}^+$ type fragment at $2(12)+5(1)+16+12+1(1)+16+1=$ consistent with $\text{C}_2\text{H}_5O\text{–CH=OH}^+$, mass $24+5+16+12+1+16+1=75$ — since exact assignment of minor secondary ions requires care, the pedagogically load-bearing deductions are the two robust ones: DoU = 1 confirms an ester, and the loss of 42 mechanistically requires a γ-H three carbons into the acyl chain, which together with the m/z 43 alpha-cleavage ion (propyl cation from the propoxy group) converge on propyl butanoate as the consistent structural assignment. The pedagogical point of this third example is procedural: build the molecular formula and DoU first (from Isotope Patterns and Molecular Formula Determination), then explain the largest, most diagnostic neutral loss with a named mechanism (here, McLafferty), and only then use smaller confirmatory peaks to fix remaining ambiguity — exactly the workflow a real unknown-structure problem demands.

Fragmentation mechanism analysis as a discipline owes its systematic form to Fred McLafferty, whose 1956–1959 work identifying the six-membered-ring hydrogen-transfer rearrangement in ketone and ester mass spectra gave physical organic chemists a mechanistic handle on peaks that earlier workers had cataloged empirically without explanation. McLafferty's insight was to recognize that a rearrangement, not just simple bond homolysis, could dominate a fragmentation pathway, and that its strict geometric requirement (a γ-hydrogen reachable through a six-membered transition state) made it a powerfully predictive — and disprovable — mechanistic hypothesis, exactly the kind of falsifiable structural claim that turned mass spectral interpretation from pattern-matching into mechanistic reasoning.