Mass Spectrometry

How mass spectrometry works

The trick of mass spectrometry is to convert a sample into gas-phase ions, push them through electric or magnetic fields where their trajectories depend on mass-to-charge ratio, and count them at the far end. The whole journey happens inside a vacuum chamber — typically 10⁻⁶ torr or better — so ions don't collide with stray gas molecules and lose their identity.

Three stages line up in series:

1. Ion source. Adds a charge. Hard sources (electron impact, EI) shower the analyte with energetic electrons and break it into fragments along the way. Soft sources (electrospray, MALDI, chemical ionisation) tickle a charge onto the molecule with minimal damage, leaving the molecular ion intact.
2. Mass analyser. Sorts ions by m/z. Different physics for different analysers — magnetic-field deflection, time-of-flight in a drift tube, oscillation in a radiofrequency quadrupole, image-current detection in an Orbitrap.
3. Detector. Counts arrivals. An electron multiplier amplifies each ion impact into ~10⁶ electrons; modern instruments record one ion per microsecond.

The output is a spectrum: a vertical bar for each m/z value, height proportional to ion abundance. The most intense peak is normalised to 100% and called the base peak; the rightmost peak in a hard-ionisation spectrum (assuming it comes from the parent) is the molecular ion M⁺•.

The five workhorse ionisation methods

Method	Acronym	Sample type	Hard or soft?	Mass range	Typical use
Electron impact	EI	Volatile small molecules (GC-amenable)	Hard — 70 eV	< 1,000 Da	GC-MS, EPA pesticide screens, NIST library matching
Electrospray ionisation	ESI	Polar liquids, biomolecules	Soft	50 Da – 1 MDa	LC-MS for drugs, peptides, intact proteins
Matrix-assisted laser desorption/ionisation	MALDI	Solids; large biomolecules co-crystallised with matrix	Soft	1 kDa – 500 kDa	Peptide mass fingerprinting, microbial ID, MALDI imaging
Chemical ionisation	CI	Volatile, hard-to-ionise neutrals	Soft	< 1,500 Da	Confirming molecular ion when EI is too destructive
Atmospheric-pressure CI	APCI	Mid-polarity small molecules	Soft	< 2,000 Da	LC-MS for non-polar drugs, lipids, steroids
Atmospheric-pressure photoionisation	APPI	Non-polar aromatics	Soft	< 2,000 Da	PAHs, asphaltenes, low-polarity LC-MS

Pick the source by sample, not by analyser. Volatile and thermally stable: EI on a GC-MS. Polar drug in plasma: ESI on an LC-MS. Intact 50 kDa protein on a stainless-steel plate: MALDI on a TOF. Greasy non-polar steroid: APCI. EI is also the only source that produces fragmentation patterns reproducible enough to support a 350,000-entry NIST library — every modern GC-MS still leans on those reference spectra.

m/z arithmetic with adducts

The peaks you see in a soft-ionisation spectrum are almost never the bare molecular mass M. They are M plus or minus some adduct from the mobile phase, and you have to back the adduct out before claiming you've measured molecular mass.

Mode	Adduct	m/z formula	Source
Positive	[M+H]⁺	M + 1.0073	Acidic mobile phase, ESI
Positive	[M+Na]⁺	M + 22.9898	Trace sodium from glassware
Positive	[M+K]⁺	M + 38.9637	Potassium contamination
Positive	[M+NH₄]⁺	M + 18.0344	Ammonium acetate buffer
Negative	[M−H]⁻	M − 1.0073	Acids, phenols, ESI−
Negative	[M+HCOO]⁻	M + 44.9982	Formate buffer in negative mode
Multi-charge	[M+nH]ⁿ⁺	(M + n × 1.0073) / n	Proteins in ESI

Worked example. You see a peak at m/z 285.1230 in positive ESI of a drug candidate. The true molecular mass depends on the adduct. If the protonated ion [M+H]⁺ → M = 284.1157. If sodiated [M+Na]⁺ → M = 262.1332. The 22-Da gap is huge and routinely catches people. Run the same sample in ammonium acetate and watch the dominant peak shift by exactly 17 Da — that's the [M+NH₄]⁺ vs [M+H]⁺ pair, and it confirms which adduct you're seeing.

For a multiply-charged protein, two adjacent peaks in the charge-state ladder solve for both M and n simultaneously. A peak at m/z 1500 with z and a peak at m/z 1429 with z+1 give M ≈ 30,000 Da and z ≈ 20. Software like maximum-entropy deconvolution (MaxEnt) automates this and reports a single mass per protein.

Fragmentation: structure from a spectrum

Hard ionisation (and soft ionisation followed by collision-induced dissociation in MS/MS) breaks the molecular ion into pieces. The fragmentation pattern is reproducible and diagnostic. Common neutral losses tell you what functional groups are present:

• Loss of 15 Da (CH₃): methyl groups, alpha-cleavage from a substituted carbon.
• Loss of 17 Da (OH) or 18 Da (H₂O): alcohols and acids dehydrating.
• Loss of 28 Da (CO or C₂H₄): carbonyl loss or McLafferty rearrangement on a γ-H.
• Loss of 29 Da (CHO or C₂H₅): aldehyde or ethyl-group cleavage.
• Loss of 31 Da (OCH₃) or 45 Da (COOH): methyl ester or carboxylic acid.
• m/z 91 base peak: tropylium cation C₇H₇⁺ — diagnostic for benzyl or alkylbenzene.
• m/z 77 (C₆H₅⁺): phenyl cation — aromatic ring system.
• m/z 105 (C₆H₅C≡O⁺): benzoyl cation — aromatic ketone or ester.

The McLafferty rearrangement is the textbook example: a γ-hydrogen migrates to the carbonyl oxygen via a six-membered transition state, the β-bond cleaves, and you get an enol radical cation plus a neutral alkene. Spotting the McLafferty fragment (m/z 58 in 2-pentanone, m/z 60 in butanoic acid, etc.) instantly diagnoses a carbonyl with a γ-hydrogen.

Mass analysers compared

Analyser	Resolution	Mass accuracy	Speed	Strength
Quadrupole	~1,000	0.1 Da	Fast	Cheap, robust, ideal for SRM/MRM quantitation
Triple quadrupole (QqQ)	~1,000 each	0.1 Da	Fast	Tandem MS for trace quantitation
Time-of-flight (TOF)	20,000 – 60,000	1–5 ppm	Very fast	Full-spectrum acquisition; broad mass range
Orbitrap	100,000 – 500,000	< 1 ppm	Moderate	Highest accuracy on benchtop scale
FT-ICR	> 1,000,000	< 0.1 ppm	Slow	Petroleomics, top-down proteomics
Ion trap	~3,000	0.1 Da	Fast (MSⁿ)	Multi-stage MS — fragment a fragment of a fragment
Magnetic sector	~30,000	1 ppm	Slow	Legacy high-resolution; isotope-ratio measurements

A modern lab tends to own three analysers: a triple quad for routine quantitation (drug levels in blood), a Q-TOF or Orbitrap for unknown identification (which metabolite is this?), and a MALDI-TOF for big biomolecules (microbe identification, protein mass).

Real-world specs and benchmarks

• Forensic fentanyl detection. LC-MS/MS on a triple quadrupole quantifies fentanyl in human plasma down to 0.05 ng/mL — about 50 picograms per millilitre. Fentanyl analogues (carfentanil, acetylfentanyl) are routinely separated and identified at the same level. The 2024 DEA reference method calls for a 5 ng/mL cut-off for postmortem cases, with confirmation by exact-mass HRMS.
• Peptide mass fingerprinting. Trypsin digestion of an unknown protein, MALDI-TOF over m/z 700 – 4000, match the resulting peak list against an in silico digestion of the SwissProt database. A correct ID needs at least four matching peptides; modern tools score by Mowse or Mascot.
• Doping control. WADA-accredited labs run GC-MS and LC-MS/MS on every Olympic athlete's urine. Detection limits for steroids are under 1 ng/mL; the 12-month re-analysis programme retains samples for retroactive testing as analytical methods improve.
• Environmental. EPA Method 525 uses GC-MS to quantify ~80 organic contaminants in drinking water at the 0.1 µg/L level, with calibration curves over four orders of magnitude.
• Clinical proteomics. Top-down ESI-Orbitrap analysis of intact proteins resolves modifications — phosphorylation, glycosylation, oxidation — that bottom-up tryptic digests miss. A 2024 reference workflow sequences a 50 kDa protein in under an hour at 5-ppm mass accuracy.

Tandem MS and quantitation

Tandem mass spectrometry (MS/MS) couples two stages of mass analysis with a fragmentation cell in between. The first analyser selects a precursor ion of a chosen m/z; the collision cell excites it (collision-induced dissociation, CID, with neutral argon gas at ~10 mTorr); the second analyser scans the resulting fragments. Two practical modes dominate:

• Selected-reaction monitoring (SRM/MRM). Watch only one precursor → product transition per analyte. Massively selective, massively sensitive — femtogram detection, four-orders-of-magnitude dynamic range. The standard for drug-of-abuse quantitation.
• Data-dependent acquisition (DDA). Survey scan finds the most intense ions; the instrument auto-fragments the top N. Used in untargeted metabolomics and proteomics.

Common mistakes and pitfalls

• Confusing nominal mass with exact mass. Nominal masses are integers and don't distinguish isobars. Always quote exact mass and compute formula candidates from it.
• Misassigning the molecular ion in EI. When the M⁺• is weak, the largest m/z may be a fragment. Run a soft-ionisation experiment (CI or ESI) to confirm M.
• Ignoring adducts in ESI. A 22-Da or 38-Da gap on the high-mass side flags Na or K. Don't report molecular mass from a sodiated peak.
• Forgetting the nitrogen rule. Odd nominal mass for the molecular ion (no halogens) means an odd number of nitrogens. Useful sanity check on proposed formulae.
• Reading isotope patterns as background. The M+1 and M+2 peaks are real and informative — don't smooth them away.
• Saturating the detector. Above ~10⁷ counts/s an electron multiplier saturates and intensities go non-linear. Dilute heavily concentrated samples or shorten ion-injection time.
• Trusting library matches blindly. NIST EI libraries match by spectral similarity, not chemistry. A 90% match with poor structural sense is worse than a 70% match with correct fragmentation logic — always verify.