Mass Analyzers: Resolution and Mass Accuracy

A chemist measures a peak at m/z 400.1652. They must decide between two options. Is this C₂₅H₂₆NO₃⁺ or C₂₄H₂₄N₃O₂⁺? These two ions differ in mass by only 4.5 millimass units. That is roughly one part in ninety thousand. Yet assigning the wrong one derails a structure elucidation. It can misidentify a metabolite, or invalidate a forensic case. The instrument's ability to make that call is not a vague notion of "good sensitivity." It is quantified precisely. It carries the same rigor a chemist brings to a titration curve or a calibration line. This precision comes through two numbers: resolving power and mass accuracy. This matters because every downstream claim in mass spectrometry rests on one thing. A molecular formula, an isotope pattern, a fragment assignment — all depend on one thing. They depend on how faithfully the analyzer converts an ion's mass-to-charge ratio into a measured signal. A The Mass Spectrometer: Components and Workflow produces ions. But it is the analyzer that decides how finely those ions can be sorted. It is the detector and data system that decide something else. They decide how precisely the resulting peak position can be read back as a number. Confusing "high resolution" with "high accuracy" is one of the most common errors students make. The two are genuinely independent axes of instrument performance. This is much as precision and accuracy are independent in any measurement. The big picture is that every mass analyzer trades off four figures of merit. These are traded off against one another. They are: resolving power (how well it separates two close masses), mass accuracy (how close the measured mass is to the true mass), m/z range (the window of masses it can transmit), and scan speed (how fast it can sweep or record that window). No analyzer wins on all four simultaneously. A quadrupole is fast and rugged but modest in resolving power. A Fourier-transform ion cyclotron resonance (FT-ICR) instrument delivers exquisite resolving power. But it scans slowly and is expensive to run. Understanding these trade-offs matters greatly, as do the equations that quantify them. It is what lets a chemist choose the right instrument for a question. It also lets them correctly interpret the number that comes out. This article builds the vocabulary and the mathematics of analyzer performance from first principles. It covers how resolving power is defined and measured. It explains why mass accuracy is reported in parts-per-million. It shows how the two connect through peak width and m/z. It also shows how real analyzers compare to one another. These range from unit-resolution quadrupoles to the multi-million-resolving-power FT-ICR. It stacks them up on a single, comparable scale. It sets up the deeper treatments of the Quadrupole Analyzer, the Time-of-Flight Analyzer, the Magnetic Sector and Ion Trap, and the practice of High-Resolution and Accurate-Mass mass spectrometry.

Key takeaways

Resolving power $R = m/\Delta m$ quantifies how well an analyzer separates two nearby masses; it is dimensionless and always quoted at a stated m/z.
$\Delta m$ can be defined by 10% valley (two equal peaks touch at 10% height) or by full width at half maximum (FWHM) — the two give different numeric values for the same physical peak.
Mass accuracy is the fractional error between measured and true mass, reported in parts-per-million (ppm); it depends on calibration quality, not just on resolving power.
High resolving power helps reveal closely spaced peaks, but it does not by itself guarantee low mass error — a well-resolved peak can still be miscalibrated.
Every analyzer trades resolving power, m/z range, and scan speed against each other; quadrupoles favor speed and robustness, TOF and FT-based instruments favor resolving power and mass accuracy.
Reporting a mass to four decimal places is meaningless unless the instrument's resolving power and calibration can support that many significant figures.
Key idea: resolving power tells you whether two peaks are distinguishable at all; mass accuracy tells you whether the number you read off a peak is trustworthy — and neither one implies the other.

Where this lives today

Resolving power and mass accuracy specifications drive nearly every purchasing and method-development decision in a modern analytical laboratory. In pharmaceutical impurity profiling, regulatory guidelines increasingly expect accurate-mass confirmation (typically < 5 ppm) of degradation products and process impurities, which requires an Orbitrap or QTOF rather than a unit-resolution quadrupole, precisely because isobaric degradation products routinely differ by only a few millimass units. In proteomics and metabolomics, high resolving power (routinely 60,000–500,000 depending on instrument and acquisition mode) lets researchers distinguish isotope fine structure and resolve near-isobaric peptides or metabolites directly in a complex biological matrix without complete chromatographic separation, dramatically shortening analysis time while preserving specificity. In environmental and forensic toxicology, non-targeted screening for unknown contaminants or novel psychoactive substances leans on high mass accuracy to propose molecular formulas for compounds with no prior reference spectrum — an application in which resolving power (separating the analyte peak from background matrix ions) and mass accuracy (constraining candidate formulas via the accurate-mass search described in the derivations above) work together as the two-part diagnostic. In isotope-ratio and nuclear forensics work, magnetic-sector instruments still deliver the combination of very high mass accuracy and excellent linear dynamic range needed to detect trace isotopic anomalies (for example uranium enrichment signatures) that would be invisible to lower-resolution analyzers. And in petroleomics, where a single crude-oil sample can contain tens of thousands of distinct elemental compositions across a narrow mass range, only ultra-high-resolving-power FT-ICR instruments (R > 500,000, sometimes exceeding 1,000,000) can resolve and assign the resulting dense forest of overlapping peaks — a direct, industrial-scale demonstration of why the resolving-power/mass-accuracy framework developed here is not an academic abstraction but the operating logic of real instruments used every day in drug development, clinical diagnostics, and materials characterization.

Ask the tutor: How do I convert a manufacturer's 10%-valley resolving-power spec into an equivalent FWHM value when comparing two mass spectrometers from different vendors?

The physical picture — thorough prose

Every mass analyzer performs the same conceptual task: it converts an ion's mass-to-charge ratio, m/z, into some physically measurable quantity — a flight time, a cyclotron frequency, a stable trajectory through an oscillating field, a radius of curvature in a magnetic field — and the data system inverts that relationship to report m/z back to the user. The peak that appears in a mass spectrum is never infinitely narrow. It has a finite width set by the physics of the analyzer: energy and angular spread of the ion beam, field imperfections, space-charge effects, detector response time, and how the raw signal is digitized and centroided. Two figures of merit describe how well this conversion works. The first, resolving power, asks: how close together in mass can two ions be and still appear as two distinguishable peaks rather than one blob? The second, mass accuracy, asks a completely different question: once you have a peak, how close is its measured centroid to the ion's true, theoretical mass? A useful mental image is a spectrophotometer's slit width versus its wavelength calibration — narrow slits (high resolving power) let you separate two close absorption lines, but if the wavelength dial is miscalibrated, every reading is shifted regardless of slit width. The same logic applies here: resolving power is about peak shape and separation; mass accuracy is about peak position and calibration. The practical consequence is that these two numbers must be reported and read independently. A high-resolving-power instrument (FT-ICR, Orbitrap) can still deliver poor mass accuracy if the calibration drifts between an external standard run and the sample run. Conversely, even a low-resolving-power quadrupole can achieve reasonable accuracy for an isolated, well-separated peak if it is properly and recently calibrated — though its resolving power will never let it distinguish isobaric interferences. Analytical chemists exploit both quantities together: resolving power to confirm you are looking at a single chemical species (not an unresolved doublet), and mass accuracy to assign a molecular formula with confidence.

Figure

Two adjacent peaks at m/z 400 and 400.02: the 10%-valley definition of resolving power

Figure

FWHM definition on a single peak, and how R changes with m/z at constant Δm

The figures show the two most common ways resolving power is measured in practice. The 10% valley definition considers two peaks of equal height and asks how close together they can be while the valley between them still dips to 10% of the peak height (a convention historically tied to magnetic-sector instruments and to the ability of a human eye or simple algorithm to recognize two distinct maxima). The FWHM definition, now standard for TOF, Orbitrap, and FT-ICR instruments, uses a single isolated peak and reports $R = m/\Delta m_{50\%}$ , where $\Delta m_{50\%}$ is the full width of the peak at half its maximum height. These are not numerically interchangeable: for a Gaussian peak shape, the 10%-valley resolving power for two equal, adjacent peaks corresponds to roughly 1.8–2× the separation implied by a single peak's FWHM, so a resolving power quoted as "10,000 (10% valley)" is a different physical claim from "10,000 (FWHM)." Always check which convention a manufacturer's specification sheet uses before comparing instruments.

The core ideas — governing equations

Resolving power (dimensionless, always stated at a particular m/z):

R = \frac{m}{\Delta m}

where: $m$ is the mass-to-charge ratio at which the measurement is made, and $\Delta m$ is either the FWHM of a single peak or the mass difference between two equal peaks whose valley falls to 10% of peak height (state which). Reading: resolving power is a ratio, so the same absolute peak width $\Delta m$ corresponds to a higher $R$ at higher m/z — an analyzer that resolves 0.01 Da at m/z 400 is quoted as $R = 40{,}000$ , but that same 0.01 Da width at m/z 1000 is $R = 100{,}000$ . This is why manufacturers must always state the m/z at which a resolving-power specification applies.

Mass accuracy, expressed as a signed or absolute error in parts-per-million:

\text{error (ppm)} = \frac{m_{\text{measured}} - m_{\text{true}}}{m_{\text{true}}} \times 10^{6}

where: $m_\text{measured}$ is the experimentally observed centroid mass and $m_\text{true}$ is the theoretical monoisotopic (or exact) mass computed from the proposed formula. Reading: mass accuracy is a relative error, reported in ppm precisely because absolute mass errors scale with m/z — a 4 mDa error is 10 ppm at m/z 400 but only 4 ppm at m/z 1000. "High-resolution accurate-mass" (HRAM) work typically targets errors below 5 ppm (often < 2–3 ppm on well-calibrated Orbitrap or FT-ICR instruments), which is what allows confident discrimination among candidate elemental formulas.

Number of resolvable formula candidates, an important derived relationship connecting mass accuracy to formula assignment, uses the fact that the number of chemically plausible formulas within a mass window grows roughly linearly with the window width and with mass itself:

N_{\text{candidates}} \approx k\, m \,(\text{error window, ppm})\times 10^{-6}

where: $k$ is an empirical density-of-formulas constant (typically 5–20 formulas per Da per unit mass in the 300–600 Da range for CHNOPS compounds), $m$ is the nominal mass, and the error window is the total ± mass tolerance used for formula matching. Reading: this relationship is why sub-5-ppm accuracy at moderate mass (a few hundred Da) usually narrows candidate formulas to a handful, while accuracy of only 50–100 ppm (typical of a poorly calibrated low-resolution instrument) can leave dozens of indistinguishable candidates — resolving power alone cannot fix this if the calibration itself is off.

Derivation AFrom peak width to resolving power: the FWHM route

Goal: show explicitly how a measured peak's shape parameters combine into the resolving-power number reported on an instrument specification sheet, and why the definition must specify FWHM vs. 10% valley.

Step 1. Model the analyzer's response to a single m/z value as approximately Gaussian in the m/z domain (a good approximation for TOF and FT-based instruments after appropriate signal processing):

I(m) = I_0 \exp\!\left[-\frac{(m-m_0)^2}{2\sigma^2}\right]

Here $m_0$ is the true centroid and $\sigma$ is the standard deviation of the peak in mass units.

Step 2. Find the full width at half maximum by solving $I(m)=I_0/2$ :

\exp\!\left[-\frac{(\Delta m/2)^2}{2\sigma^2}\right] = \frac{1}{2}

Taking the natural log of both sides,

\frac{(\Delta m/2)^2}{2\sigma^2} = \ln 2

so that

\Delta m_{\text{FWHM}} = 2\sigma\sqrt{2\ln 2} \approx 2.3548\,\sigma

Step 3. Substitute into the resolving-power definition:

R_{\text{FWHM}} = \frac{m_0}{2.3548\,\sigma}

This shows resolving power is inversely proportional to the standard deviation of the peak — any physical process that broadens $\sigma$ (space charge, poor ion-beam energy focusing, insufficient transient length in FT instruments) directly lowers $R$ .

Step 4. Now derive the 10%-valley criterion for two equal, adjacent Gaussian peaks of separation $\delta$ (centered at $m_0$ and $m_0+\delta$ , each with the same $\sigma$ ). By symmetry the valley minimum sits at the midpoint $m_0+\delta/2$ . The combined intensity there is

I_{\text{valley}} = 2I_0\exp\!\left[-\frac{(\delta/2)^2}{2\sigma^2}\right]

Setting $I_\text{valley} = 0.10\,I_0$ (peak height normalized to $I_0$ , so total at the peak maximum ≈ $I_0$ since the neighboring peak contributes negligibly at its own center):

2\exp\!\left[-\frac{\delta^2}{8\sigma^2}\right] = 0.10

\frac{\delta^2}{8\sigma^2} = \ln(20) \approx 2.996

\delta \approx \sigma\sqrt{8 \times 2.996} \approx 4.89\,\sigma

Step 5. Compare the two Δm values for the same underlying peak shape ( $\sigma$ fixed):

\frac{\Delta m_{\text{10\% valley}}}{\Delta m_{\text{FWHM}}} = \frac{4.89\,\sigma}{2.35\,\sigma} \approx 2.08

This confirms that the 10%-valley definition reports a larger Δm (hence a lower numerical R) than the FWHM definition applied to the same physical peak — roughly a factor of two apart. A spec sheet quoting "R = 10,000, 10% valley" describes peaks that are meaningfully broader (in the FWHM sense) than a spec sheet quoting "R = 10,000, FWHM." This factor-of-two trap is one of the most common sources of confusion when comparing analyzer types across manufacturers.

Go deeper · Derivation B · 2 minPropagating resolving power and calibration error into a molecular-formula decision

Derivation BPropagating resolving power and calibration error into a molecular-formula decision

Goal: connect resolving power (can we even see two peaks separately?) and mass accuracy (is the number trustworthy?) to the real analytical decision of assigning a molecular formula, using an asymmetric, realistic example set up algebraically before plugging in numbers.

Step 1. Suppose two candidate ions have masses $m_1$ and $m_2$ with $m_2 > m_1$ , mass difference $\delta = m_2-m_1$ . The analyzer can only report them as two distinct peaks if its resolving power at that m/z, $R_\text{inst}$ , satisfies

R_\text{inst} \gtrsim \frac{m_1}{\delta}

(using the FWHM convention here; for equal-height doublets replace with the 10%-valley form derived above). This is a necessary separability condition — it says nothing yet about whether the reported mass is accurate.

Step 2. Independently, suppose the instrument has a systematic calibration offset $\epsilon$ (in ppm) plus a random noise contribution $\sigma_\epsilon$ (also in ppm, one standard deviation) from shot-to-shot centroiding jitter. The measured mass for an ion of true mass $m$ is

m_\text{measured} = m\left(1 + \frac{\epsilon + z}{10^6}\right), \qquad z \sim \mathcal{N}(0,\sigma_\epsilon^2)

where $z$ is a normally distributed random error term with standard deviation $\sigma_\epsilon$ .

Step 3. For a formula assignment to be reliable at some confidence level (commonly ±3 standard deviations, i.e., 99.7%), the total allowed mass window must satisfy

|\epsilon| + 3\sigma_\epsilon \le \text{tol}_\text{ppm}

where $\text{tol}_\text{ppm}$ is the tolerance used when searching a formula database. If the instrument is resolving-power-limited (Step 1 fails), no calibration procedure can rescue the assignment, because the two ions are not even separately observed — their signals physically overlap into one composite, mass-shifted peak.

Step 4. Combine both conditions into the two-part diagnostic that a working analytical chemist actually applies:

\underbrace{R_\text{inst} \ge \dfrac{m_1}{\delta}}_{\text{resolving-power test: are they separable?}} \quad\text{AND}\quad \underbrace{|\epsilon| + 3\sigma_\epsilon \le \text{tol}_\text{ppm}}_{\text{accuracy test: is the number trustworthy?}}

Step 5 — worked numeric check. Take $m_1 = 400.1652$ (C₂₅H₂₆NO₃⁺) and $m_2 = 400.1707$ (C₂₄H₂₄N₃O₂⁺), an isobaric pair separated by $\delta = 0.0055$ Da (5.5 mDa) — an asymmetric, realistic pair chosen so neither mass is a round number. The needed resolving power is

R_\text{inst} \ge \frac{400.1652}{0.0055} \approx 72{,}757

An Orbitrap operated at $R = 120{,}000$ (FWHM, at m/z 200, scaling roughly as $1/\sqrt{m/200}$ so at m/z 400 it delivers about $R \approx 120{,}000/\sqrt{2} \approx 84{,}900$ ) clears this bar with modest margin — the two ions would appear as separable, though not widely separated, peaks. A quadrupole at unit resolution ( $R \approx 400/1 = 400$ at this mass) fails by a factor of roughly 180 — it would report a single, unresolved, mass-shifted blend, and no amount of recalibration fixes that. This is the central lesson: resolving power sets what is possible; accuracy then determines whether the possible measurement is trustworthy.

Worked examples

Example 1 — Foundational: computing R and comparing to a spec sheet. A time-of-flight instrument records a peak for reserpine's protonated molecule, $[\text{M+H}]^+$ at nominal m/z 609. The peak profile, read directly from digitized data, has a full width at half maximum of $\Delta m = 0.0122$ Da centered at $m = 609.2812$ . Compute the resolving power (FWHM convention):

R = \frac{m}{\Delta m} = \frac{609.2812}{0.0122}

R \approx 49{,}941 \approx 5.0\times10^4 \ (\text{FWHM, at } m/z\ 609)

Now suppose the manufacturer's data sheet lists "R = 50,000 (10% valley) at m/z 609." Using the factor derived above ( $\Delta m_{10\%} \approx 2.08\,\Delta m_{\text{FWHM}}$ for equal Gaussian peaks), the FWHM-equivalent resolving power implied by that spec is

R_{\text{FWHM,equiv}} \approx 2.08 \times 50{,}000 \approx 104{,}000

So this particular instrument, as actually measured (R ≈ 50,000 FWHM), performs below what a naive comparison to a "50,000, 10%-valley" spec would suggest — the measured FWHM resolving power is roughly half of the FWHM-equivalent of the quoted spec. This is exactly the kind of apples-to-oranges comparison a chemist must catch: always convert to a common convention (usually FWHM, since that's now the field standard) before comparing two analyzers' numbers.

Limiting case check: if the peak were perfectly symmetric and infinitely narrow ( $\Delta m \to 0$ ), $R \to \infty$ — an idealization no real analyzer reaches, since finite ion-beam energy spread, field imperfections, and detector bandwidth all impose a floor on $\Delta m$ .

Example 2 — Applied: distinguishing two adducts in a complex mixture. A metabolomics sample gives a chromatographic peak whose mass spectrum, on a quadrupole time-of-flight (QTOF) instrument at $R = 35{,}000$ (FWHM, at m/z 300), shows a peak apparently at m/z 301.1414. Two candidate assignments are proposed:

Candidate A: $\text{C}_{14}\text{H}_{21}\text{N}_2\text{O}_4^{+}$ , exact mass 297.1496 (protonated, $[\text{M+H}]^+$ ) — actually let us instead directly use the sodium adduct problem, which is common in practice.
Candidate A: $[\text{M+H}]^+$ for a compound with formula $\text{C}_{15}\text{H}_{21}\text{N}_2\text{O}_4^{+}$ , monoisotopic mass 309.1496 — again mismatched; set up cleanly below.

To keep the arithmetic transparent and asymmetric, use this concrete pair: the observed ion at m/z 301.1414 might be $\text{[M+H]}^+$ of $\text{C}_{16}\text{H}_{21}\text{N}_2\text{O}_3^{+}$ (exact mass 301.1390) or $\text{[M+Na]}^+$ of a related compound $\text{C}_{15}\text{H}_{22}\text{N}_2\text{O}_3\text{Na}^{+}$ (exact mass 301.1523). The mass difference between these two candidates is

\delta = 301.1523 - 301.1390 = 0.0133 \text{ Da}

Required resolving power to separate them as two peaks:

R_\text{needed} = \frac{301.1390}{0.0133} \approx 22{,}640

The QTOF's actual resolving power at this mass, $R = 35{,}000$ , exceeds this requirement, so the two candidates would in principle appear as separable peaks — meaning if only one peak is observed experimentally, only one of the two ions is actually present (not an unresolved doublet). The observed centroid, 301.1414, is compared against each candidate's exact mass to compute ppm error: For candidate 1 ( $\text{C}_{16}\text{H}_{21}\text{N}_2\text{O}_3^{+}$ , 301.1390):

\text{error} = \frac{301.1414 - 301.1390}{301.1390}\times 10^6 \approx 8.0 \text{ ppm}

For candidate 2 (sodium adduct, 301.1523):

\text{error} = \frac{301.1414 - 301.1523}{301.1523}\times 10^6 \approx -36.2 \text{ ppm}

Given a typical QTOF accuracy specification of ±5 ppm under good calibration, candidate 1 (8.0 ppm) is marginal but plausible if calibration has drifted slightly, while candidate 2 (−36.2 ppm) is essentially excluded. This demonstrates the accuracy test operating exactly as derived: resolving power confirmed the peaks could be separated (so we trust we're looking at a single species), and mass accuracy then discriminated between chemically reasonable formulas.

Limiting case check: if calibration were perfect ( $\epsilon = 0$ ) and only counting-statistics noise remained, both ppm errors above would shrink toward the intrinsic noise floor $3\sigma_\epsilon$ ; a 36 ppm discrepancy would never be explained by noise alone on a properly functioning QTOF (whose noise floor is typically 1–3 ppm), confirming candidate 2 is wrong on physical, not just statistical, grounds.

Example 3 — Exam/competition level: resolving power needed across a mass range, and scan-speed trade-off. An FT-ICR instrument must resolve a doublet that recurs across a chromatographic run at three different nominal masses due to different charge states of the same underlying multiply protonated peptide: m/z 542.2871 (2+), m/z 361.1939 (3+), and m/z 271.1477 (4+), with each doublet separated from an isobaric interference by a constant absolute mass difference of 3.5 mDa in each case (a reasonable approximation since the interfering species differs from the peptide ion by a fixed elemental substitution whose mass difference is charge-state-independent when expressed per charge... but here we treat the raw m/z difference as given directly, i.e. already charge-corrected in this problem for simplicity). (a) Compute the resolving power required at each charge state:

R_{2+} = \frac{542.2871}{0.0035} \approx 154{,}940

R_{3+} = \frac{361.1939}{0.0035} \approx 103{,}198

R_{4+} = \frac{271.1477}{0.0035} \approx 77{,}471

Notice the required resolving power decreases substantially as m/z decreases (higher charge state), because Δm is held fixed while $m$ shrinks — a direct consequence of $R=m/\Delta m$ . This means the 2+ charge state is the "hardest" of the three to resolve, even though it's a smaller absolute Δm requirement relative to typical unit-resolution instruments; it demands nearly double the resolving power of the 4+ state. (b) FT-ICR resolving power scales inversely with observation time (the length of the recorded time-domain transient) and, for a given magnet, with $1/m$ (because cyclotron frequency $\omega_c = zeB/m$ decreases with $m$ , and frequency resolution scales as $1/T_\text{acquire}$ while mass resolution near a given $m$ scales as $R \propto \omega_c \cdot T_\text{acquire} \propto T_\text{acquire}/m$ ). Suppose the instrument achieves $R = 200{,}000$ at m/z 400 using a 0.768 s transient. Using $R \propto T_\text{acquire}/m$ , the transient time needed to hit the required $R_{2+} = 154{,}940$ at m/z 542.2871 is found by scaling:

\frac{R_{2+}}{R_{400}} = \frac{T_{2+}/542.2871}{T_{400}/400}

T_{2+} = T_{400}\times\frac{R_{2+}}{R_{400}}\times\frac{542.2871}{400}

T_{2+} = 0.768 \times \frac{154{,}940}{200{,}000}\times \frac{542.2871}{400}

T_{2+} = 0.768 \times 0.7747 \times 1.3557 \approx 0.807 \text{ s}

So roughly an 0.81 s transient is required — longer than the reference 0.768 s transient, because although the target resolving power (154,940) is less than the reference (200,000), the higher m/z of this ion (542.3 vs. 400) partially offsets that, since resolving power at fixed transient time degrades with increasing mass. This is the direct trade-off between resolving power and scan speed that defines FT-ICR practice: chasing higher resolving power, or working at higher m/z, always costs acquisition time, which in turn limits how many mass spectra per second can be recorded during a fast chromatographic elution — a real bottleneck when a peptide elutes over only a few seconds and many MS/MS events must be interleaved.

Limiting case check: as $T_\text{acquire} \to \infty$ , $R \to \infty$ in this idealized scaling (real FT-ICR performance eventually saturates due to ion cloud coherence loss/collisional damping, but the qualitative trend — longer transient, higher R, slower scan rate — holds and is the reason FT-ICR is rarely used for very fast, wide-range survey scans compared to TOF).

Limiting cases, scope & common traps

Situation	What happens	Why
Very high resolving power, poor calibration	Sharp, well-separated peaks, but centroid systematically off by tens of ppm	Resolving power and calibration accuracy are independent instrumental properties
Low resolving power (unit-resolution quadrupole)	Two isobaric ions merge into one peak with a shifted, intensity-weighted centroid	Δm of the two species is smaller than the instrument's peak width
Comparing "R = 10,000" between two vendors without checking convention	Apparent tie, but one instrument may have ~2× narrower peaks than the other	10%-valley and FWHM conventions differ by a factor of ~2 for equal Gaussian doublets
Reporting a mass to 4 decimal places from a unit-resolution instrument	Meaningless precision; last two digits are noise	Significant figures in a mass measurement must be supported by the instrument's actual resolving power and noise floor
Increasing FT-ICR transient time indefinitely to chase resolving power	Diminishing returns; eventually limited by ion coherence decay and duty cycle	Real transients decay (damping), and scan speed drops as $1/T$ , hurting chromatographic compatibility
Using resolving power alone to assign a molecular formula	Under-constrained; multiple formulas can fit if calibration/accuracy is poor	Formula assignment requires both adequate R (separability) and adequate ppm accuracy (trustworthy centroid)

Common trapMisconception: "high resolution" and "high mass accuracy" are the same thing

Students frequently treat these as synonyms because both are marketed together on modern Orbitrap and FT-ICR literature. They are not. Resolving power is a statement about peak shape — whether two close masses appear as one blob or two distinct maxima. Mass accuracy is a statement about peak position relative to a calibration. An instrument can have enormous resolving power (say $R = 500{,}000$ ) and still report a mass 50 ppm off true value if the calibration is stale, contaminated, or affected by space-charge shifts. Conversely, a modest-resolving-power instrument, properly and freshly calibrated on an isolated peak, can report quite accurate masses — it simply cannot separate closely spaced isobars in the first place. Always ask both questions separately: "could these two species even be resolved?" and "is this number properly calibrated?"

The chemist behind it

The modern vocabulary of resolving power and mass accuracy owes its precision-instrument lineage to Francis William Aston, whose early mass spectrographs first demonstrated that isotopes of the same element have distinct, measurable masses — a discovery that itself depended on achieving resolving power sufficient to separate neon-20 from neon-22. Aston's 1919 mass spectrograph, refined through several iterations, reached resolving powers on the order of a few hundred to a thousand, enough to resolve isotopes differing by a single mass unit at modest atomic mass — a landmark performance for the era, achieved through careful velocity and direction focusing rather than any digital calibration standard.

Origins · 1 minFrancis William Aston and how this was discovered

Francis William Aston

1877–1945 · Cambridge, Cavendish Laboratory, early 20th century

Aston trained originally in chemistry and brewing science before turning to physics under J. J. Thomson at the Cavendish Laboratory. Working with Thomson's positive-ray apparatus, Aston recognized that the parabolic traces produced by ionized gas streams in combined electric and magnetic fields could, with proper focusing, be sharpened into resolvable lines corresponding to distinct isotopic masses. His key innovation — velocity focusing, arranging the electric and magnetic sectors so that ions of the same mass but slightly different velocities converge to the same point on a photographic plate — is the direct conceptual ancestor of the double-focusing magnetic-sector instruments still used today for the highest-precision mass measurements. Aston used his mass spectrograph to measure the masses of over 50 elements' isotopes and formulated the "whole number rule," recognizing that all isotopic masses are close to whole numbers when oxygen-16 is fixed as a reference, with small deviations he correctly attributed to nuclear binding energy — a direct experimental link between his resolving-power figure of merit and the physics of mass-energy equivalence. He was awarded the Nobel Prize in Chemistry in 1922 for this work.

History

- 1919 — Aston builds his first mass spectrograph and resolves neon isotopes ²⁰Ne and ²²Ne, establishing that resolving power, not just detection sensitivity, is the defining figure of merit for isotope discovery.

- 1922 — Aston awarded the Nobel Prize in Chemistry for isotope discovery via mass spectrography and formulation of the whole-number rule.

- 1934–1937 — Aston's third mass spectrograph achieves resolving power sufficient to detect mass defects at the level of a few parts in 10,000, directly supporting nuclear binding-energy calculations.

- 1953 — the quadrupole mass filter is introduced by Wolfgang Paul, trading resolving power for compactness, speed, and robustness — the origin of the modern low-resolution, high-throughput analyzer class.

- 1974 — Fourier-transform ion cyclotron resonance mass spectrometry is demonstrated by Comisarow and Marshall, eventually delivering resolving powers exceeding 10⁶, redefining the practical ceiling for mass accuracy in complex-mixture analysis.

Why is resolving power always quoted at a specific m/z instead of as one universal number?

Because $R = m/\Delta m$ is a ratio, and $\Delta m$ (the physical peak width in mass units) is not constant across the mass range for most analyzers. In FT-based instruments, resolving power at fixed acquisition time scales roughly as $1/m$ , so it is intrinsically higher at low m/z and lower at high m/z. In TOF instruments, resolving power tends to be more stable across a moderate mass range because it derives from timing precision rather than frequency. Because the relationship between $R$ and $m$ differs by analyzer type, a single bare number like "R = 60,000" is ambiguous unless paired with the m/z at which it was measured — reputable specification sheets always state this, e.g., "R = 60,000 at m/z 400 (FWHM)."

Can I improve mass accuracy just by increasing resolving power?

Not directly. Mass accuracy depends primarily on how well the instrument is calibrated — the correspondence between measured signal (flight time, frequency, trajectory) and true mass — plus the noise floor in the centroiding process. Higher resolving power can indirectly help accuracy by preventing an unresolved neighboring peak from dragging the centroid of your peak of interest off-center (a real and common source of error called "peak coalescence" or interference-induced mass shift). But if your peak is already isolated, adding resolving power without improving calibration will not shrink the ppm error; you need better internal or external mass calibration, more stable ion optics, or lock-mass correction.

What does "10% valley" resolving power mean physically, and why 10%?

It describes two adjacent peaks of equal height and asks how close together (in mass) they can be while the dip between them ("valley") drops to only 10% of the peak height. The 10% threshold is a historical convention from the era of photographic-plate and early electronic detection on magnetic-sector instruments, chosen because a 10% valley is reliably distinguishable by eye or simple threshold algorithms from a single, unresolved peak. It remains in use, particularly for double-focusing sector instruments, but modern software-driven analyzers (TOF, Orbitrap, FT-ICR) more often report FWHM resolving power from a single isolated peak, since automated peak-picking algorithms work more naturally with that definition.

Why does a quadrupole mass analyzer have such comparatively low resolving power?

A quadrupole mass filter selects ions by stability of trajectory in an oscillating quadrupolar electric field, typically operated at "unit resolution," meaning it can separate m/z 400 from m/z 401 but not finer distinctions within that unit. This is a deliberate design point, not a fundamental physical ceiling — quadrupoles can be tuned for higher resolving power at the cost of transmission (fewer ions successfully traverse the filter), so they trade resolving power for speed, robustness, sensitivity, and low cost. This makes them extremely well suited to targeted quantitation (e.g., triple-quadrupole methods monitoring known transitions) where the identity of the ion is already known and only quantitative signal matters, rather than to non-targeted, high-resolution structure elucidation.

If I report a mass to five decimal places, does that automatically mean my measurement is accurate to that many digits?

No — this is one of the most consequential traps in reporting mass spectrometry data. The number of decimal places you display has nothing to do with the number of decimal places your instrument can actually support, which is set jointly by resolving power (can the peak even be measured as an isolated, well-shaped feature) and by the noise floor and calibration stability of the specific acquisition. A unit-resolution quadrupole reporting "400.16523" is fabricating precision the hardware cannot deliver; an Orbitrap under poor calibration reporting the same number of digits may likewise be over-stating its trustworthy precision. Significant figures in a mass measurement must be justified the same way they would be in any titration or gravimetric result — by the instrument's demonstrated, calibration-verified performance, not by how many digits the software happens to print.

How do resolving power, mass accuracy, m/z range, and scan speed trade off against each other in practice?

No analyzer maximizes all four simultaneously, and instrument choice reflects which trade-off matters for a given experiment. Quadrupoles offer fast scan speed and robustness at low resolving power and modest accuracy. Time-of-flight analyzers offer wide m/z range, good resolving power, and fast full-spectrum acquisition, at accuracy typically in the low single-digit to ~5 ppm range. Orbitraps offer very high resolving power and sub-ppm-to-few-ppm accuracy but sacrifice raw scan speed because achieving high resolving power requires a longer transient (time-domain signal), directly limiting spectra-per-second. FT-ICR pushes resolving power and accuracy to the extreme (often exceeding 10⁶ and sub-ppm) at the cost of the slowest scan speed and highest instrument cost/complexity among common analyzers. Magnetic-sector instruments occupy a niche of extremely high accuracy and stability for isotope-ratio work, at the cost of size, scan speed, and operational complexity.

Does higher resolving power always mean I need less sample?

Not necessarily, and sometimes the reverse. Achieving higher resolving power in FT-based instruments often requires longer ion accumulation and longer transient acquisition times, during which signal from a finite, decaying ion population is recorded — for very dilute samples, this can actually help (integrating over a longer, cleaner transient improves the signal-to-noise ratio of the eventual Fourier transform peak) but for space-charge-sensitive instruments, too many ions accumulated for a long high-resolution scan can itself degrade both resolving power and mass accuracy through ion-ion Coulombic interactions. Sample amount, ionization efficiency, and analyzer duty cycle interact with the resolving-power setting in ways that must be optimized experimentally, not assumed from the resolving-power specification alone.