Prerequisites: Signal response versus concentration, and the basic idea of calibration. Calibration versus tuning and reading chromatograms supply useful preparation.
Your task
For each suspicious result, name the comparison that would change your interpretation. This lesson develops a control-selection habit, not a universal batch sequence. All worked responses are synthetic. Acceptance limits, control frequency and corrective actions must come from the applicable method, project requirements and laboratory procedure.
1. A blank, a QC and an internal standard do different jobs
A blank asks about unwanted signal when the target contribution is intended to be absent. A quality-control material asks whether a defined measurement performs acceptably against an assigned value or other established expectation. An internal standard is a reference substance measured within an individual preparation or injection. These roles complement each other; passing one does not replace the others.
| Control | Useful question | What it cannot establish alone |
|---|---|---|
| Solvent / instrument blank | Is signal present in the introduced solvent and instrument path? | Cleanliness of the complete preparation process |
| Method / process blank | Do preparation reagents, vessels or handling contribute signal? | Equivalent behavior in every actual sample matrix |
| Blank after a high sample | Is later signal consistent with retained material? | The physical source of the carryover |
| Independent QC / check standard | Does a known material meet the preselected criterion? | Performance at every level, matrix and time |
| Matrix spike / matrix QC | Does an added amount behave acceptably in this matrix and process? | Recovery of every native analyte form |
| Internal standard | How does a reference response behave within this sample? | Freedom from all analyte-specific interferences |
Describe a blank by what actually happened to it. “Solvent blank injected through the autosampler” is more useful than “blank 1.” A no-injection run, a solvent injection and an extracted blank do not interrogate identical paths. Record whether an internal standard was added; “blank” does not always mean that every channel should be empty.
An independently prepared check can challenge an error shared by the calibrators. Reinjecting one calibrator is useful for a different question, such as repeat response, but is not independent evidence against an error in that calibrator's assigned concentration.
Internal-standard correction depends on suitable behavior relative to the analyte. Also record when it was added: a substance added after extraction has not experienced earlier extraction losses. Review its absolute response as well as the analyte-to-standard ratio. A plausible ratio can conceal two unusually weak signals.
2. Place controls where the question arises
A control has both a preparation history and a position in the analytical sequence. A process blank prepared with a batch but injected after a highly concentrated sample can contain contributions from preparation and from carryover. Its label alone cannot distinguish them.
For planning practice, annotate a proposed sequence with three questions:
- Before the unknowns: what evidence establishes the initial background and acceptable response?
- Across the run: which controls can reveal a change after the initial check? Put the required checks where the method specifies, including relevant concentration levels and matrices.
- After a high challenge: what measures residual signal before it could affect a subsequent low sample?
This is a reasoning framework, not permission to replace a prescribed order. A passing start-of-run QC cannot demonstrate that response stayed acceptable later. A final passing QC cannot prove that an earlier isolated problem never happened.
Check: a process blank is high only when injected after a high standard. What should you compare?
Compare its behavior at a clean starting point with a separately prepared process blank and a solvent blank in a controlled sequence allowed by the method. Preserve the original files. The comparison can separate competing explanations better than renaming the original result “preparation contamination.” Changing both preparation and injection order at once would make attribution harder.
3. Work through a carryover investigation
Use the same response definition and processing rules for every entry below. Assume no intentional change in acquisition or injection volume. “High sample” describes this exercise, not a validated concentration class.
| Order | Material | Response |
|---|---|---|
| 1 | Pre-run solvent blank | 2 |
| 2 | Separately prepared process blank | 3 |
| 3 | High sample | 10,000 |
| 4 | Solvent blank A | 120 |
| 5 | Solvent blank B | 30 |
| 6 | Solvent blank C | 7 |
State the leading hypothesis, one explanation the data weaken, and one conclusion they cannot support.
Reveal the investigation logic
The elevated response after the high sample, declining through successive blanks, is consistent with carryover or a memory effect. The low earlier process blank weakens the hypothesis of substantial contamination in that particular preparation blank. It does not clear all other preparations.
The sequence cannot locate the source in the needle, valve, tubing, column, spray chamber or source. Some of those components apply only to particular instruments. It also cannot establish that response 7 meets the actual blank criterion.
The denominator changes the practical meaning
The first post-sample blank is 120 / 10,000 × 100 = 1.2% of the high response. Suppose a separate low-level standard gives response 50. The same blank is 120 / 50 × 100 = 240% of that low response. Both ratios are arithmetically true; neither is an acceptance criterion or a concentration correction.
A small percentage of a very high sample can still matter for a subsequent low result. ICH M10, section 3.2.6, illustrates this principle in its own bioanalytical scope by assessing carryover after the upper calibration standard and relating it to low-level analyte and internal-standard responses. Do not import that framework's limits into an unrelated assay.
Choose the next test to distinguish hypotheses. Repeated blanks establish a time pattern; they do not by themselves reveal a component to replace. If the approved investigation changes a wash or rinse condition, document the single change and compare equivalent high-challenge/blank sequences. Instrument-specific servicing and solvent choices belong to trained operation and the manufacturer's procedure.
4. Clean blanks do not establish recovery
A matrix may change response without producing a target-like peak in a blank. Preparation can also lose analyte. These are different questions from contamination. Matuszewski, Constanzer and Chavez-Eng's original study distinguishes matrix effects from extraction recovery in quantitative HPLC-MS/MS.
Consider three synthetic, matched preparations containing the same nominal analyte amount: a neat solution gives 1000; a blank matrix extract spiked after extraction gives 600; and matrix spiked before extraction gives 450. Assume matched final volumes, injection conditions and stability, negligible endogenous response, and suitable comparability of the preparations.
Calculate what each comparison tests
Post-extraction / neat = 600 / 1000 = 60%: under the stated assumptions, response in the matrix extract is lower. Pre-extraction / post-extraction = 450 / 600 = 75%: this isolates an apparent extraction-recovery comparison. Pre-extraction / neat = 45% combines both effects and should not be called extraction recovery alone.
These are explanatory ratios, not evidence that a method passes. Repeatability, concentration dependence and relevant matrix variation still matter.
Adding a dissolved spike does not guarantee that it behaves like native analyte bound within the original sample. A good spike result therefore answers a bounded question. Likewise, a single matrix lot cannot establish performance in every specimen. ICH M10 section 7.3 describes the pre- versus post-extraction comparison and emphasizes consistent recovery rather than requiring recovery to be exactly 100%.
5. Let the method govern the response to a failure
Before reporting affected results, locate the applicable criterion and required action. Record:
- The observation: control identity, preparation batch, sequence position, analyte channel and raw response.
- The comparison: criterion, units, denominator and procedure version; distinguish a diagnostic concern from a formally failed check.
- The scope: which samples could be affected by preparation, sequence position or response drift?
- The disposition: the prescribed investigation, reanalysis, qualification or rejection, with supporting evidence and review.
Do not silently subtract the blank, delete the failed injection or keep reinjecting until one result looks satisfactory. Preserve the initial evidence and justify any repeated analysis. EPA Method 6020B, section 9.7, explicitly cautions against subtracting a single method-blank result from associated samples in its elemental ICP-MS application.
Final check: blanks are low, but matrix QCs are low and internal-standard responses have changed. Is the batch cleared?
No. Low blanks address unwanted signal, while the QC and internal standard raise response or preparation questions. Inspect the calibration, standard preparation, matrix behavior and timing of the change. Apply the actual method's acceptance and corrective-action rules. The information given does not justify either reporting the whole batch or rejecting it wholesale.
Before you move on
Finish this sentence for every control: “This tests ___, at this point in the process, but does not establish ___.” That makes both a passing result and a failure more useful.
Primary references
- US EPA Method 6020B: sections 4.6–4.8 and 9 discuss matrix effects, memory, preparation contamination and QC within elemental ICP-MS.
- ICH M10, Bioanalytical Method Validation and Study Sample Analysis: sections 3.2.6 and 7.3 provide scoped carryover and recovery examples.
- Matuszewski et al., Analytical Chemistry 75 (2003), 3019–3030: original matrix-effect assessment research; DOI 10.1021/ac020361s.
Next step
Return to reading chromatograms to compare sample and blank with explicit processing settings. Use calibration versus tuning for independent response checks, and where ions come from to understand why source and matrix conditions matter.