Ionization Methods: Hard vs Soft

A chemist hands you two mass spectra of the same unknown drug metabolite. They look nothing alike. One is a forest of fragment peaks with barely a trace of the intact molecule. The other shows a single dominant peak at the molecular mass and almost nothing else. Both spectra are correct. The difference is not the molecule but the method used to turn it into a gas-phase ion. This is the central lesson of ionization choice in mass spectrometry. The ionization source is not a neutral doorway into the analyzer. It is an active participant that decides how much internal energy the ion carries away. That energy decides whether the ion survives intact or shatters. Ionization methods split broadly into two philosophies, conventionally called hard and soft. Hard methods, typified by electron ionization (EI), slam a beam of ~70 eV electrons into vapor-phase molecules. This rips out an electron and deposits far more energy than a single bond can absorb. The result is extensive, reproducible fragmentation that acts like a structural fingerprint. Soft methods, typified by electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI), transfer only a few electron-volts or less. So the molecular ion (or a protonated/deprotonated version of it) survives largely intact. It appears as a base peak with little fragmentation. Neither approach is "better." They answer different questions. A working analytical chemist must know which question they are asking before choosing a source. The reason this matters extends across the whole discipline of mass spectrometry. It also applies to its hyphenated cousins (GC-MS, LC-MS, MALDI-TOF imaging). A gas chromatograph delivers small, thermally stable, volatile analytes. For these, EI's harsh fragmentation is a feature. Fragment patterns are cataloged in libraries like the NIST database. They are used for unambiguous identification. A liquid chromatograph, by contrast, delivers large, fragile biomolecules. These are often thermally labile: proteins, oligonucleotides, intact glycoconjugates. For these, EI would simply destroy the very molecular-weight information you need. ESI's gentleness is what makes LC-MS of a 66 kDa protein possible at all. The same is true for a small peptide. The instrument itself is described generically in The Mass Spectrometer: Components and Workflow. It is a common chassis. Ionization is the interchangeable engine that determines what kind of spectrum comes out the other end. This article builds the conceptual scaffolding. It explains what "hard" and "soft" mean physically, in terms of energy deposited per ionization event. It shows how that connects to the internal-energy distribution of the resulting ion population. It also shows how that distribution controls the survival probability of the molecular ion against unimolecular fragmentation. We will look at the physical picture with two annotated energy diagrams. We will work through the governing energetics with real numbers. These include ionization energies, appearance energies, and a semi-quantitative treatment of why a 70 eV electron beam over-shoots so badly. Then we will apply the ideas to three realistic worked examples. These span a textbook EI spectrum, a pharmaceutical LC-MS identification, and an exam-style problem comparing appearance energies. This groundwork is what later articles build directly upon: Electron Ionization and the Molecular Ion, Electrospray Ionization, and MALDI.

Key takeaways

Ionization methods are classified by how much excess internal energy they deposit in the newly formed ion, not by the chemistry of the analyte itself.
Hard ionization (e.g., EI at 70 eV) deposits energy far above the ionization energy (IE), driving extensive, reproducible fragmentation useful for structure elucidation and library matching.
Soft ionization (e.g., ESI, MALDI, CI) deposits energy close to the minimum needed to form an ion, so the molecular/quasi-molecular ion dominates the spectrum.
The survival of M⁺• (or [M+H]⁺) is governed by the competition between the rate of unimolecular fragmentation and the residence time in the source/analyzer — a kinetic, not just thermodynamic, criterion.
Hard sources suit small, volatile, thermally robust molecules coupled to GC; soft sources suit large, fragile, non-volatile molecules coupled to LC or deposited as solids (MALDI).
Fragmentation is not "noise" — in hard ionization it is deliberately harvested as a structural fingerprint; in soft ionization it is deliberately suppressed to preserve the molecular-weight signal.
Key idea: hard vs. soft ionization is a spectrum of excess internal energy deposited per ion, and that energy — not the analyzer — is what decides whether you see a shattered molecule or an intact one.

Where this lives today

The hard/soft distinction is not a historical curiosity; it is the daily operating logic of every mass spectrometry laboratory. GC-MS instruments running 70 eV EI remain the workhorse for environmental analysis (pesticide residues, PAHs), forensic toxicology, and petrochemical fingerprinting, precisely because reproducible fragment patterns allow automated library searching against hundreds of thousands of reference spectra — a small, volatile, thermally stable molecule is exactly the regime where hard ionization's destructive tendencies become an identification asset rather than a liability. At the other extreme, ESI-LC-MS/MS is the backbone of modern pharmaceutical bioanalysis, proteomics, and metabolomics: intact protein mass determination, peptide mapping, and quantitative drug-metabolite assays in plasma all depend on ESI's gentle, multiply-charged ion formation to preserve molecular-weight information for species from a few hundred daltons up to intact antibodies exceeding 150 kDa. MALDI-TOF, meanwhile, dominates clinical microbiology (rapid bacterial identification from whole-cell protein fingerprints) and is the enabling technology behind imaging mass spectrometry, which maps the spatial distribution of lipids, drugs, and metabolites directly across a tissue section without prior extraction — a feat that would be impossible if the ionization step destroyed the analyte's molecular identity. Modern instruments increasingly blur the line deliberately: ESI or MALDI sources are now routinely coupled to collision cells that perform on-demand, tunable "hard" fragmentation (CID, higher-energy collisional dissociation HCD, or electron-transfer dissociation ETD) after a soft ionization step, letting an operator dial in exactly how much internal energy to add and when. This hybrid strategy — gentle formation, controlled fragmentation on command — is the direct engineering descendant of the energy-deposition logic developed here, and it underlies virtually every tandem MS/MS structural elucidation workflow used in modern drug discovery, glycoproteomics, and lipidomics.

:::quiz[In the RRK-type rate expression k(E) = A[(E−E₀)/E]^(s−1), what happens to k(E) as the internal energy E approaches the threshold E₀ from above?]

k(E) approaches infinity
k(E) approaches A regardless of E₀
k(E) approaches zero ✓
k(E) becomes negative, indicating no reaction

Ask the tutor: How does the choice between hard and soft ionization change what appears in an MS/MS collision-induced dissociation spectrum compared to the original ionization step?

The physical picture

Every ionization event begins with a neutral molecule M in the gas phase (or, for desorption methods, at a surface) and ends with a charged species that a mass analyzer can manipulate. The two things that vary from method to method are (1) how the charge is introduced — electron removal, electron capture, proton transfer, adduct formation — and (2) how much extra energy beyond the strict minimum is dumped into the resulting ion. It is this second quantity, the excess internal energy $E_{int}$ , that separates "hard" from "soft." In electron ionization, a heated filament emits electrons that are accelerated through a potential difference of about 70 V and collide with the gas-phase analyte. The ionization energy of a typical organic molecule is only 8–11 eV, so a 70 eV electron carries roughly six to nine times the energy strictly needed to knock out an electron. The excess does not disappear — a fraction of it is transferred into the internal (vibrational, rotational) energy of the newly formed radical cation M⁺•. Because chemical bonds have dissociation energies typically in the 3–5 eV range, an ion carrying several eV of excess vibrational energy has more than enough energy to break several bonds sequentially, and it does so on a timescale of nanoseconds to microseconds while still inside the source and flight path. This cascade of unimolecular dissociations is what produces the rich fragment-ion pattern seen in classic EI mass spectra. Soft methods are engineered to avoid this energy dump. Chemical ionization (CI) uses a reagent gas (e.g., methane, ammonia) ionized first; the reagent ions then transfer a proton to the analyte in a much lower-energy, closer-to-thermoneutral process, typically depositing only 1–2 eV of excess energy. Electrospray ionization avoids gas-phase electron bombardment altogether: a charged liquid droplet, formed by applying a few kilovolts across a fine capillary, shrinks by solvent evaporation and Coulombic fission (Rayleigh instability) until bare, multiply protonated analyte ions are ejected into the gas phase near room temperature, with little vibrational excitation. MALDI embeds the analyte in a solid UV-absorbing matrix; a nanosecond laser pulse ablates a small volume, and the matrix — not the analyte — absorbs almost all of the photon energy, later transferring a proton to the analyte in the resulting hot plume, but far more gently than a 70 eV electron would.

Figure

Energy deposited per ionization event: hard vs. soft methods, drawn on a common internal-energy axis

The key structural feature of that diagram is the width and position of the internal-energy distribution relative to the dissociation thresholds of the weakest bonds. A narrow distribution sitting just above the ionization energy (soft ionization) leaves almost every ion below its fragmentation threshold — it survives to the detector as an intact molecular or quasi-molecular ion. A broad distribution extending well past several bond dissociation energies (hard ionization) guarantees that a large fraction of ions have enough energy to fragment, often more than once, before detection.

Figure

From neutral molecule to detected ion: two competing pathways after ionization

The core ideas

The whole conceptual apparatus reduces to a competition between two rates: the rate of unimolecular dissociation of the excited ion, $k(E)$ , and the inverse of the time the ion spends in the source and early flight path, $1/\tau$ . This is the qualitative content of the Rice–Ramsperger–Kassel–Marcus (RRKM) picture of unimolecular reactions, simplified here to the level needed for a foundational treatment.

E_{int} = E_{deposited} - IE

where:

$E_{deposited}$ is the energy delivered by the ionizing agent (electron beam, photon, proton transfer) to the molecule during the ionization event, in eV;
$IE$ is the adiabatic ionization energy of the neutral molecule, the minimum energy needed to remove an electron (or, for protonation processes, the analogous proton affinity threshold), in eV;
$E_{int}$ is the excess internal (vibrational + rotational) energy left in the resulting ion, available to drive fragmentation.

Reading it: $E_{int}$ is simply "energy in minus the toll paid to make the ion in the first place." A soft method tries to make $E_{deposited}$ only slightly larger than $IE$ ; a hard method massively overpays.

k(E) = A \left(\frac{E - E_0}{E}\right)^{s-1}

where:

$k(E)$ is the microcanonical unimolecular rate constant for a specific fragmentation channel with critical (threshold) energy $E_0$ , in s⁻¹;
$A$ is a frequency factor of order $10^{13}$ – $10^{16}$ s⁻¹, reflecting the "attempt frequency" of the bond-breaking vibration;
$E$ is the total internal energy of the ion (here, essentially $E_{int}$ from above), in eV;
$E_0$ is the appearance-energy threshold for that particular fragmentation, in eV;
$s$ is the effective number of vibrational degrees of freedom over which the energy is statistically distributed (this is the simplified RRK form of RRKM theory).

Reading it: $k(E)$ is vanishingly small when $E$ is only barely above $E_0$ , and rises steeply once $E$ exceeds $E_0$ by even a modest margin — this steepness is why a narrow window of extra energy above threshold is enough to turn a stable ion into a fast-fragmenting one, and why controlling $E_{int}$ so tightly is the whole game of ionization-method design.

Derivation AWhy 70 eV electrons in EI guarantee fragmentation, quantitatively

Goal. Show that the excess internal energy typically deposited by a 70 eV electron beam is comfortably larger than the weakest bond dissociation energies in an organic molecule, so that $k(E) \gg 1/\tau$ for many channels — explaining why EI spectra are fragment-rich rather than molecular-ion-rich.

Step 1 — energy budget of the electron impact event. A 70 eV electron does not transfer all 70 eV to the molecule; ionization is a collision in which only the energy exceeding the ionization energy can, in principle, appear as internal excitation of the cation, since the ejected electron carries away kinetic energy too. The maximum possible internal excitation is

E_{int}^{max} = E_{beam} - IE

For a molecule with $IE = 9.0$ eV struck by the nominal 70 eV beam, $E_{int}^{max} = 70 - 9.0 = 61$ eV — enormous compared to any single bond. In practice the ionization efficiency curve (cross-section vs. electron energy) peaks broadly between roughly 50 and 100 eV, and the actual internal-energy distribution of the ion population $P(E)$ is broad, typically spanning several eV up to a few tens of eV, because most collisions are "glancing" and transfer far less than the kinematic maximum. The important point is not the exact shape of $P(E)$ but that its mean comfortably exceeds typical bond dissociation energies of 3–5 eV.

Step 2 — comparing to fragmentation thresholds. Suppose the weakest bond in the molecule (e.g., an α-cleavage next to a heteroatom) has an appearance energy of $E_0 = 1.5$ eV above the ionization energy (a realistic value for many C–C bonds adjacent to carbonyls or amines). Then even an ion that received only, say, 4 eV of excess internal energy (a small fraction of the available 61 eV maximum) sits at

E - E_0 = 4.0 - 1.5 = 2.5 \text{ eV above threshold}

Using the RRK rate expression with a typical $A = 10^{14}\ \text{s}^{-1}$ and $s = 20$ effective oscillators (a mid-sized organic ion),

k(E) = 10^{14}\left(\frac{2.5}{4.0}\right)^{19} \approx 10^{14} \times (0.625)^{19}

Evaluating $(0.625)^{19} \approx 8 \times 10^{-5}$ gives

k(E) \approx 10^{14} \times 8\times10^{-5} \approx 8 \times 10^{9}\ \text{s}^{-1}

Step 3 — comparing to the residence time. The ion spends roughly $10^{-6}$ s in a typical EI source before extraction (source residence time $\tau \sim 1\ \mu s = 10^{-6}$ s). The dissociation probability while in the source scales with $k\tau$ :

k\tau \approx (8\times10^{9}\ \text{s}^{-1})(10^{-6}\ \text{s}) = 8\times10^{3} \gg 1

Since $k\tau \gg 1$ , essentially every ion that receives even a modest 4 eV of excess energy fragments via this channel long before reaching the detector. This is the quantitative reason why EI, with its broad multi-eV internal-energy distribution, produces heavily fragmented spectra: the rate of dissociation vastly exceeds the rate of extraction for nearly all accessible channels.

Limiting check. If instead the excess energy were only $E = 1.6$ eV (barely above the 1.5 eV threshold), $(E-E_0)/E = 0.1/1.6 = 0.0625$ , and $(0.0625)^{19}$ is astronomically small (~ $10^{-24}$ ), giving $k(E) \approx 10^{-10}\ \text{s}^{-1}$ — a lifetime far longer than the age of the universe, so the ion would survive. This confirms the model's internal consistency: it is specifically the broad, high-energy tail of EI's internal-energy distribution — not the mere fact of ionization — that drives fragmentation, and a soft method that keeps $E$ close to $E_0$ leaves ions essentially unreactive on the mass-spectrometric timescale.

Go deeper · Derivation B · 2 minSurvival fraction as a simple kinetic competition — connecting to soft ionization

Derivation BSurvival fraction as a simple kinetic competition — connecting to soft ionization

Goal. Derive an expression for the fraction of molecular ions that survive to the detector without fragmenting, given an internal-energy-dependent rate constant $k(E)$ and a fixed instrument residence time $\tau$ , and use it to show why soft ionization methods yield a dominant molecular-ion peak.

Step 1 — first-order kinetics for a single ion. Treat the disappearance of intact M⁺• (or [M+H]⁺) by unimolecular fragmentation as pseudo-first-order once formed, with rate constant $k(E)$ depending on how much internal energy that particular ion happens to carry:

\frac{d[M^{+\bullet}]}{dt} = -k(E)\,[M^{+\bullet}]

Integrating over the residence time $\tau$ that the ion spends between formation and detection,

\frac{[M^{+\bullet}]_\tau}{[M^{+\bullet}]_0} = e^{-k(E)\tau}

This ratio is the survival fraction $S(E)$ for an ion formed with internal energy $E$ .

Step 2 — averaging over the internal-energy distribution. A real ionization event does not produce ions with a single $E$ ; it produces a distribution $P(E)\,dE$ of internal energies. The overall fraction of molecular ions observed intact is the population-weighted average

S = \int_0^\infty P(E)\, e^{-k(E)\tau}\, dE

This integral is the mathematical statement of "hard vs. soft": $S$ depends entirely on how much weight $P(E)$ places at energies where $k(E)\tau \gtrsim 1$ .

Step 3 — the two limiting regimes. For soft ionization, $P(E)$ is narrow and centered just above $IE$ , so that for essentially all ions $E - E_0 < 0$ or is a tiny positive number; $k(E)\tau \ll 1$ for the entire distribution, and

S \approx \int_0^\infty P(E)\,(1 - k(E)\tau)\,dE \approx 1 - \tau\langle k(E)\rangle \approx 1

because $\langle k(E)\rangle \tau \ll 1$ throughout the populated range — nearly all ions survive, giving a spectrum dominated by the molecular or quasi-molecular ion. For hard ionization, $P(E)$ extends well past $E_0$ for many channels, so for a large fraction of the population $k(E)\tau \gg 1$ and $e^{-k(E)\tau} \approx 0$ ; only the low-energy tail of $P(E)$ (ions that happened to receive very little excess energy) contributes to $S$ , so

S \ll 1

and the spectrum is dominated by fragment ions instead, with only a small residual molecular-ion peak (sometimes vanishingly small, as in highly branched hydrocarbons).

Limiting-case check. As $\tau \to 0$ (an instrument with an extremely short source residence time, e.g., a fast time-of-flight extraction), $e^{-k(E)\tau} \to 1$ for any finite $k(E)$ , so $S \to 1$ regardless of how hard the ionization is — even EI would show a strong molecular ion if extraction were instantaneous. This correctly captures a subtlety that real instrument design must handle: fragmentation is a kinetic competition between $k(E)$ and $1/\tau$ , not solely a property of the ionization method, though in practice real EI sources have $\tau$ long enough (∼microseconds) that hard ionization methods reliably show heavy fragmentation.

Worked examples

Example 1 (foundational) — reading a classic EI spectrum of butanoic acid. Butanoic acid, CH₃CH₂CH₂COOH, has a molecular weight of 88.11 g/mol (monoisotopic mass 88.0524 Da) and an ionization energy of about 10.17 eV. Under 70 eV EI, the literature spectrum shows the molecular ion M⁺• at m/z 88 with only modest relative intensity (a few percent of base peak), and a dominant base peak at m/z 60, corresponding to loss of C₂H₄ (28 Da) via a McLafferty rearrangement transferring a γ-hydrogen to the carbonyl oxygen with cleavage of the Cα–Cβ bond, expelling ethylene and leaving the enol of acetic acid as the charged fragment. Checking the arithmetic: $88 - 28 = 60$ , consistent with loss of C₂H₄ (mass $2\times12 + 4\times1 = 28$ ). A further well-known fragment appears at m/z 73, corresponding to loss of •CH₃ ( $88 - 15 = 73$ ), and at m/z 27 (C₂H₃⁺, vinyl cation-like fragment) and m/z 41 (C₃H₅⁺). None of this is arbitrary: each loss corresponds to breaking a specific bond whose appearance energy is low enough that, given the broad internal-energy distribution from 70 eV electron impact (Derivation 1), a large fraction of the ion population has $E > E_0$ for that channel. The near-absence of an intense M⁺• (relative intensity often under 10%) is the signature of hard ionization: the molecule's own weakest bonds (Cα–Cβ next to the carboxyl) have appearance energies low enough, and the EI internal-energy distribution broad enough, that $k(E)\tau \gg 1$ for most ions, exactly as the survival-fraction formula predicts.

Example 2 (applied) — LC-ESI-MS identification of a drug metabolite. A pharmaceutical laboratory analyzes a plasma extract by LC-MS using positive-ion electrospray. The chromatogram shows a peak at retention time 6.3 min whose mass spectrum (full scan, m/z 100–600) shows essentially one dominant peak at m/z 302.1 and a much smaller companion peak at m/z 324.1, with no other significant ions. Interpretation using soft-ionization logic: the base peak at m/z 302.1 is the protonated molecular ion [M+H]⁺, implying a neutral monoisotopic mass of

M = 302.1 - 1.0073 = 301.1\ \text{Da (to one decimal)}

using the mass of a proton, 1.0073 Da. The smaller peak at m/z 324.1 is the sodium adduct [M+Na]⁺, since

324.1 - 302.1 = 22.0

and the mass difference between a sodium adduct and a proton adduct is $22.99 - 1.01 = 21.98 \approx 22.0$ Da, matching within instrumental mass accuracy (typically a few ppm to a few mDa on a quadrupole or time-of-flight instrument). Because ESI is soft, no fragment ions accompany the parent — the internal-energy distribution deposited during droplet desolvation and ion ejection is far too narrow and too close to threshold to drive dissociation, consistent with $S \approx 1$ from the survival-fraction argument. To get structural information (e.g., confirm a hydroxylation site on the metabolite), the analyst must deliberately re-introduce excess energy downstream via collision-induced dissociation (CID) in a tandem MS/MS experiment — effectively borrowing the "hard ionization" fragmentation logic on demand, at a stage separate from the initial soft ionization event. This separation of "make the ion gently, then fragment it on purpose, later, if you want structure" is the entire operating principle of soft-ionization LC-MS/MS workflows.

Example 3 (exam-level) — using appearance energies to rank fragmentation intensity. A problem gives two candidate molecules, A and B, both with ionization energy 8.90 eV, examined by 70 eV EI. Molecule A has an appearance energy for its dominant fragmentation channel of $AE_A = 10.10$ eV; molecule B has $AE_B = 12.40$ eV. Using the RRK-type rate expression with $A_{factor} = 5\times10^{13}\ \text{s}^{-1}$ , $s = 15$ , and assuming both ions are formed with the same typical internal energy $E = 13.0$ eV (a representative value drawn from the EI internal-energy distribution, well within its broad tail), estimate which molecule shows the more intense fragment ion relative to its molecular ion, and by roughly what factor. First find the threshold energy for fragmentation relative to $IE$ : since appearance energies are usually quoted relative to the neutral, the excess energy over threshold for each channel is

E - E_0 = E - (AE - IE) \quad\text{is not needed here; treat } E_0 = AE - IE \text{ as the channel threshold above the ion's zero of energy.}

So $E_{0,A} = AE_A - IE = 10.10 - 8.90 = 1.20$ eV and $E_{0,B} = AE_B - IE = 12.40 - 8.90 = 3.50$ eV. With the ion internal energy $E = 13.0 - 8.90 = 4.10$ eV above the IE reference point (i.e., $4.10$ eV of vibrational excess available to drive fragmentation),

k_A(E) = (5\times10^{13})\left(\frac{4.10 - 1.20}{4.10}\right)^{14} = (5\times10^{13})(0.707)^{14}

Evaluating $(0.707)^{14}$ : since $0.707 = 1/\sqrt2$ , $(0.707)^{14} = (1/\sqrt2)^{14} = 2^{-7} = 1/128 \approx 7.81\times10^{-3}$ , so

k_A(E) \approx (5\times10^{13})(7.81\times10^{-3}) \approx 3.9\times10^{11}\ \text{s}^{-1}

k_B(E) = (5\times10^{13})\left(\frac{4.10 - 3.50}{4.10}\right)^{14} = (5\times10^{13})(0.1463)^{14}

Evaluating $(0.1463)^{14}$ : taking logarithms, $\log_{10}(0.1463) \approx -0.835$ , so $14 \times (-0.835) = -11.69$ , giving $(0.1463)^{14} \approx 2.0\times10^{-12}$ . Then

k_B(E) \approx (5\times10^{13})(2.0\times10^{-12}) \approx 0.10\ \text{s}^{-1}

With a source residence time $\tau \approx 1\times10^{-6}$ s, $k_A \tau \approx 3.9\times10^{11}\times10^{-6} = 3.9\times10^{5} \gg 1$ (fragment essentially all ions of A), while $k_B \tau \approx 0.10\times10^{-6} = 1\times10^{-7} \ll 1$ (almost no ions of B fragment via this channel at this internal energy). The ratio of rate constants,

\frac{k_A}{k_B} \approx \frac{3.9\times10^{11}}{0.10} \approx 4\times10^{12}

shows that even a modest 2.3 eV difference in appearance energy (1.20 eV vs. 3.50 eV above IE) — smaller than the energy of a single visible-light photon times a few — translates into a dissociation-rate difference of twelve orders of magnitude, because the rate constant depends on a high power (here, the 14th) of the fractional excess energy. This is exactly why appearance energies measured to a few tenths of an eV by electron-impact ionization-efficiency curves are diagnostic: small differences in threshold energy produce enormous, experimentally distinguishable differences in fragment-ion abundance, which is the basis of using fragmentation patterns as structural fingerprints in EI-MS libraries.

Limiting cases, scope & common traps

Scenario	Expected behavior	Why
Small, volatile, thermally stable analyte + EI	Rich fragmentation, weak or absent M⁺•	Broad $P(E)$ well above bond thresholds; $k(E)\tau \gg 1$ for many channels
Large, thermally labile biomolecule + EI	Analyte decomposes before/without vaporizing; EI often not applicable at all	EI requires gas-phase neutral; nonvolatile, thermally fragile molecules cannot survive the inlet, regardless of $E_{int}$
Same biomolecule + ESI or MALDI	Dominant [M+nH]ⁿ⁺ or [M+H]⁺, minimal fragmentation	Narrow $P(E)$ near threshold; $k(E)\tau \ll 1$
Soft-ionized ion subjected to CID in MS/MS	Deliberate, controlled fragmentation reappears	Extra energy purposefully added downstream of ionization; same kinetic competition, different stage
Extraction/residence time $\tau \to 0$	Even "hard" ionization would show intact M⁺•	Survival fraction $S \to 1$ as $\tau \to 0$ regardless of $E_{int}$ (kinetic limiting case)
Chemical ionization (CI)	Intermediate: [M+H]⁺ often dominant, some fragmentation	Excess energy (~1–2 eV) smaller than EI but larger than ESI/MALDI

The chemist behind it

The hard/soft ionization distinction as a working principle owes its clearest articulation to work spanning several decades and several instrumental lineages, but if one figure deserves the "physicist behind it" treatment for the broader energy-deposition framework, it is Henry (Nino) Rosenstock and, jointly, the RRKM theorists whose statistical treatment of unimolecular decay underlies the $k(E)$ curve used throughout this article. Rosenstock, working with quasi-equilibrium theory (QET) — essentially RRKM theory applied specifically to mass spectrometric fragmentation — provided the first rigorous statistical framework explaining why electron-impact mass spectra show reproducible, energy-dependent fragmentation patterns rather than a chaotic jumble, tying together electron-impact energetics, appearance energies, and unimolecular rate theory into a single coherent picture. This quasi-equilibrium framework is precisely the machinery invoked (in simplified RRK form) in the derivations above.

Origins · 1 minHenry M. Rosenstock and how this was discovered

Henry M. Rosenstock

1926–2007 · Applied statistical unimolecular-reaction-rate theory to electron-impact mass spectrometry, founding the quasi-equilibrium theory (QET) of mass spectra

Rosenstock's 1952 paper with Wallenstein, Wahrhaftig, and Eyring — "Absolute Rate Theory for Isolated Systems and the Mass Spectra of Polyatomic Molecules" — adapted the RRKM statistical theory of unimolecular reactions (itself building on transition-state theory pioneered by Henry Eyring) to the specific problem of predicting relative abundances of fragment ions in electron-impact mass spectra. QET treats the excited molecular ion as a quasi-equilibrated system that partitions its excess internal energy statistically among all accessible vibrational modes before choosing a fragmentation pathway, exactly the picture underlying the $k(E)$ expression used throughout this article. His work, done at the National Bureau of Standards, gave mass spectrometrists a physically grounded reason to trust that EI fragmentation patterns are not idiosyncratic noise but the reproducible output of well-defined statistical kinetics — the theoretical backbone that makes spectral-library matching in modern GC-MS trustworthy.

History

- 1918 — Arthur Dempster builds an early mass spectrograph using electron-impact ionization, establishing electron bombardment as a practical ionization method.

- 1952 — Rosenstock, Wallenstein, Wahrhaftig, and Eyring publish the quasi-equilibrium theory (QET) of mass spectra, providing the first rigorous statistical link between internal energy and fragment-ion abundance.

- 1966 — Munson and Field introduce chemical ionization (CI), the first widely adopted "softer" alternative to standard 70 eV EI, using ion–molecule proton-transfer reactions.

- 1984 — John Fenn and coworkers demonstrate electrospray ionization of large intact biomolecules, the technique later recognized with a share of the 2002 Nobel Prize in Chemistry.

- 1987 — Koichi Tanaka reports soft laser desorption of large proteins, contributing to the development of MALDI alongside Franz Hillenkamp and Michael Karas, also recognized in the 2002 Nobel Prize.

What exactly counts as "hard" versus "soft" — is there a strict numeric cutoff?

There is no official numeric boundary; the terms describe a continuum of excess internal energy deposited per ionization event. Conventionally, methods depositing several eV to tens of eV above the ionization energy (70 eV EI is the classic example) are called hard, because that energy comfortably exceeds typical bond dissociation energies (3–5 eV) and drives extensive fragmentation. Methods depositing at most a few eV, and often under 1 eV of true excess (ESI, MALDI, and to a lesser extent CI), are called soft, because the resulting ions mostly sit below fragmentation thresholds. Chemical ionization is often described as "intermediate," illustrating that the categories are qualitative regions on a continuous energy axis rather than a binary switch.

If soft ionization gives so little structural information, why is it preferred for large biomolecules?

Because for a 50 kDa protein, the single most valuable number is often the intact molecular mass itself — confirming identity, detecting post-translational modifications, or verifying correct disulfide bonding requires an intact mass measurement, not a fragment pattern. EI cannot even deliver the analyte to the gas phase intact (proteins are non-volatile and thermally fragile, decomposing long before vaporizing), so hard ionization is not a milder-vs-harsher choice here — it is simply inapplicable. Soft ionization (ESI, MALDI) is the only route to an intact gas-phase ion at all, and structural detail, when needed, is obtained afterward via a deliberate, controlled fragmentation step (CID/HCD/ETD) in tandem MS/MS.

Does "soft" ionization mean zero fragmentation, ever?

No. Even ESI and MALDI show some fragmentation for labile analytes (e.g., glycosidic bonds in oligosaccharides, or facile loss of a labile phosphate group), because the internal-energy distribution, while narrow and low relative to EI, is not a delta function at exactly the ionization threshold. Source conditions such as cone voltage in ESI or laser fluence in MALDI can add extra, "in-source" internal energy and induce partial fragmentation even without a dedicated collision cell — this is sometimes an unwanted artifact and sometimes deliberately exploited (in-source CID) for extra structural information.

Why does electron energy matter so much — wouldn't a lower-energy electron beam in EI make the technique softer?

It does, in principle, and this is exactly what "low-voltage EI" (e.g., 12–15 eV rather than 70 eV) exploits: reducing the electron energy toward the ionization energy narrows the internal-energy distribution and increases the relative abundance of the molecular ion, at the cost of reduced overall ionization efficiency (fewer ions formed per unit time, since the ionization cross-section falls off steeply near threshold). The standard 70 eV value became a de facto standard specifically because the ionization cross-section for most organic molecules is near its broad maximum there, giving high sensitivity, reproducible fragmentation patterns for library matching — a deliberate trade of "softness" for signal and reproducibility.

How does the appearance energy (AE) of a fragment connect to bond dissociation energy?

The appearance energy is the minimum electron-beam (or photon) energy at which a particular fragment ion is first observed, and by Hess's-law-style bookkeeping it equals the ionization energy of the parent plus the energy required for the specific bond-breaking/rearrangement step that produces that fragment, plus any additional activation barrier beyond simple bond dissociation (relevant for rearrangement channels like McLafferty transfer). Because $AE - IE$ isolates the fragmentation-specific energy requirement, comparing $AE - IE$ values across candidate structures — as in the worked exam example — lets chemists rank which fragmentation channels should dominate a spectrum, and appearance-energy measurements historically provided some of the first quantitative bond-energy data for gas-phase ions.

Is MALDI "harder" or "softer" than ESI?

Both are firmly in the soft category relative to EI, but MALDI typically deposits somewhat more internal energy than ESI because the ionization event involves a hot, rapidly expanding plume generated by nanosecond laser ablation of the matrix, and the analyte can pick up excess vibrational energy from collisions within that plume before it fully cools. This is why MALDI spectra sometimes show modest in-source fragmentation or matrix-related adducts that are rarer in a well-tuned ESI spectrum, even though both techniques reliably preserve the intact molecular ion as the dominant species for most analytes.

Does the hard/soft classification apply only to positive-ion methods?

No — the same energy-deposition logic applies symmetrically to negative-ion techniques. Negative-ion chemical ionization via electron capture, for instance, is exquisitely soft (electron capture by an electrophilic molecule can deposit almost no excess energy, sometimes producing spectra with essentially only [M]⁻• or [M−H]⁻), while dissociative electron attachment or negative-ion EI-type processes can be comparatively hard, generating anionic fragments. The hard/soft framework is about the magnitude and distribution of excess internal energy relative to fragmentation thresholds, a criterion that is charge-sign-agnostic.