Diagnosing LC-MS sensitivity loss means answering one question first: are analyte molecules failing to reach the source, failing to ionise once there, or failing to be detected after ionisation. Each has a different confirming test, and working through them in order is faster than adjusting source parameters and hoping.
The most expensive mistake in this workflow is starting at the instrument. Source voltages, gas flows and collision energies are the last things to change, not the first, because a method that was working did not spontaneously need different tuning. Characterise the pattern of the loss before touching anything.
Step 1: When did the LC-MS sensitivity loss start?
The timing of the decline narrows the cause more sharply than any single test.
| Pattern | Most likely causes | First confirming test |
|---|---|---|
| Gradual over days or weeks | Source contamination, column fouling, detector ageing | Inject a neat standard; compare to the last known-good response |
| Sudden, between sequences | New mobile phase or reagent lot, hardware fault, failed seal | Re-run with reserved known-good mobile phase |
| Only in certain samples | Matrix effects, sample-specific degradation | Compare matrix-matched against neat standard response |
| Present from the first injection | Method transfer error, wrong transition, mis-prepared standard | Verify transitions and re-prepare from a fresh ampoule |
| Within a single run, worsening | Carry-over, phospholipid build-up, autosampler stability | Inject blanks between samples and track the trend |
Record which of these five patterns applies before proceeding. Every step below assumes it is known.
LC-MS PK Peak Integrity Engine
Interrogate carryover, matrix effects, ion suppression, and integration integrity before your PK data goes to QC.
Step 2: Is ion suppression causing the signal loss?
If the loss appears in matrix but not in neat standard, it is a matrix effect and the remaining steps mostly do not apply. Two tests confirm it.
- Matrix versus neat comparison. Spike blank extracted matrix and neat solvent at the same concentration and compare peak areas. A ratio meaningfully below 100% is suppression; above 100% is enhancement.
- Post-column infusion. Infuse analyte into the column effluent while injecting an extracted blank. Dips in the flat baseline mark where in the run matrix components are compromising the source.
Both tests are described in full, along with mechanism, quantification of the effect and what ICH M10 requires for validation, on the ion suppression and matrix effects page. Confirm suppression there before returning to this checklist — if it is the cause, changing source parameters will not fix it.
Step 3: Is the source contaminated?
Source fouling deposits involatile matrix on the sampling cone, ion block and early ion optics, and produces a steady decline rather than a step change. It affects all analytes in a method roughly equally, which distinguishes it from an analyte-specific problem.
Check the cone and ion block visually. Discolouration or visible residue is diagnostic. Clean to the manufacturer’s procedure, then re-inject the same neat standard used in Step 1; recovery of response confirms it.
If cleaning restores sensitivity but it degrades again within days, the cleaning interval is not the problem — the sample preparation is letting too much involatile material into the source. Diverting the void volume and the final high-organic wash to waste is the cheapest intervention.
Step 4: Did a reagent or mobile phase change?
A sudden loss between sequences most often traces to something newly prepared. Water quality, a new solvent lot, a modifier weighed out fresh, or a buffer prepared to the wrong pH will all reduce response, and all present identically.
The fastest isolation is a reserved control: keep one bottle of mobile phase known to have produced acceptable response, sealed and dated. Re-running a failing sequence against it separates reagent problems from everything else in a single injection.
Check pH specifically. A buffer one unit from target changes analyte ionisation state and can move both retention and response substantially without any visible symptom.
Step 5: Is it peak broadening rather than true sensitivity loss?
Peak height and peak area fail differently, and confusing them sends the investigation in the wrong direction.
| Observation | Interpretation | Where to look |
|---|---|---|
| Height down, area unchanged | Peak broadening, not sensitivity loss | Column condition, void, extra-column volume, injection solvent strength |
| Height and area both down proportionally | True response loss | Steps 2, 3, 4 and 6 |
| Area down, retention shifted | Chromatographic change | Mobile phase composition, temperature, column ageing |
| Area down, peak shape distorted | Overload, void, or blocked frit | Guard column, inlet frit, injection volume |
Always compare integrated areas, not heights, when assessing whether response has genuinely fallen.
Step 6: Are the samples or standards degrading?
Pre-analytical loss looks identical to instrumental loss at the detector and is frequently the answer when every instrumental check passes.
Consider stock standard age and storage, freeze-thaw cycles on study samples, adsorption to plastic consumables for basic or highly lipophilic analytes, and autosampler stability during long sequences. A standard prepared fresh from a new ampoule, injected alongside the suspect stock, resolves this in one comparison.
Where the analyte adsorbs to labware, response falls progressively through a sequence and is worst for the samples that sat longest. Silanised vials, added carrier protein or a small percentage of organic in the reconstitution solvent usually recover it.
Where to go next with persistent LC-MS sensitivity loss
If this checklist confirms a matrix effect, the mitigation work — extraction chemistry, chromatographic separation from the suppressor, internal standard choice — is covered on the ion suppression and matrix effects page.
If sensitivity has fallen far enough to threaten the lowest reportable concentration, re-establishing the method’s limit of detection and re-determining the signal-to-noise ratio is required before reporting further results, since both were established under conditions that no longer apply.
Frequently asked questions
How much LC-MS sensitivity loss is normal between services?
There is no universal figure, which is why a tracked known-good standard matters more than an absolute threshold. Establish the response of a reference injection when the method is performing acceptably and trend it. A decline beyond roughly 20-30% warrants investigation even if results still pass system suitability.
Should I clean the source before or after checking for ion suppression?
After. Source cleaning takes the instrument down and destroys the evidence that would have distinguished a contamination problem from a matrix problem. The matrix-versus-neat comparison in Step 2 takes one sequence and costs nothing.
Can retuning recover lost LC-MS sensitivity?
Retuning compensates for drift in ion optics but will not recover response lost to suppression, contamination or degradation. If a method required different tune parameters than it did last month, the underlying reason still needs finding — retuning conceals it rather than fixing it.
References
ICH M10 Bioanalytical Method Validation and Study Sample Analysis, effective 21 January 2023. European Medicines Agency.