A known addition produces only 60% of the signal seen in clean solvent. Did extraction lose 40% of the analyte, or did the sample matrix reduce the detector response? That single comparison cannot separate the two explanations.
Comparing additions made before and after extraction can distinguish preparation recovery from a matrix response effect under controlled conditions. This guide works through three matched preparations, a cancellation case and the limits of the conclusion. It is a data-interpretation lesson, not a laboratory procedure or a universal acceptance standard.
THE GOLD STANDARD OF CHEMISTRY
Choose the question you need to answer
Locate the measurement stages · Chapters 1–2. Define the spike timing and matched comparisons.
Separate and challenge the effects · Chapters 3–5. Calculate recovery, response and cancellation.
Practise and continue · Chapters 6–8. Diagnose changed cases and choose further reading.
Open the chapter contents · Try the transfer checks · The Gold Standard of Chemistry
Chapter contents
Locate the measurement stages
Separate and challenge the effects
3. Calculate three ratios from one consistent signal set
Practise and continue
6. Use changed signals to diagnose the stage
The analyte is the substance being measured. The matrix consists of the other sample components. An extraction aims to transfer analyte into a form suitable for measurement, but some analyte can be lost during preparation. The final extract can also change the instrument’s response to whatever analyte reaches it.
Those effects occupy different stages. Losing half the analyte before measurement and detecting all the analyte with half the response can produce the same final signal. A neat calibration curve in clean solvent does not reveal which process occurred in a sample.
A spike is a known addition of analyte. Where it enters the process determines which stages it experiences. A spike added only to a completed extract cannot test how much analyte would have survived extraction.
Before interpreting any percentages, draw the sequence: original material, extraction and cleanup, final volume, instrument response. Mark the spike’s entry point. This simple diagram often exposes a claim that the available control never tested.
Call preparation A a known amount in neat solvent. Call B the same amount added to an analyte-free matrix extract after preparation. Call C the same amount added to a comparable analyte-free matrix before the full preparation process.
For this teaching model, the spike amount would yield the same final concentration in all three if no preparation loss occurred. Final volumes, injected volumes and other instrument conditions are matched. Background signals are negligible or have been appropriately accounted for, and response is locally linear.
A versus B tests how the final matrix changes response. B versus C tests preparation recovery while comparing signals in a comparable final matrix. A versus C combines both stages. If the final matrices or concentrations differ substantially, these simple ratios may no longer isolate their intended effects.
The established bioanalytical literature describes this comparison strategy. Real method validation also considers matrix sources, concentration levels, repeatability, selectivity and the stated measurement purpose. Our labels organise the reasoning without prescribing one validation design for every technique.
| Preparation | When analyte is added | Stage represented |
|---|---|---|
| A: neat solvent | No sample extraction | Reference instrument response |
| B: post-extraction spike | After matrix preparation | Response in final matrix |
| C: pre-extraction spike | Before sample preparation | Preparation plus matrix response |
Suppose the background-corrected responses are A = 100, B = 80 and C = 60 arbitrary signal units. The post-extraction matrix factor is B/A = 0.80. Written as a percentage of neat-solvent response, it is 80%, indicating suppression in this model.
Preparation recovery is C/B = 60/80 = 0.75, or 75%. The relevant comparator is B because it has the matrix response experienced by recovered analyte. Calling C/A the extraction recovery would wrongly assign the matrix suppression to preparation loss.
The combined process response is C/A = 0.60, or 60% of the neat-solvent response. In this matched multiplicative model, 0.75 × 0.80 = 0.60. Both a preparation loss and a response change contribute.
Terminology varies between methods and publications. A matrix factor of 80% is not the same label as a matrix effect reported as −20% relative to neat solvent. Write the equation with the label. That prevents a report of 80% matrix factor being mistaken for 80% suppression.
| Quantity | Comparison | Result |
|---|---|---|
| Matrix factor | B/A = 80/100 | 0.80: 20% response suppression |
| Preparation recovery | C/B = 60/80 | 0.75: 75% recovery |
| Combined process response | C/A = 60/100 | 0.60: 60% of neat response |
Now use A = 100, B = 125 and C = 100. Comparing only C with A gives 100%. It would be tempting to conclude that the entire process is unaffected.
The separated comparisons tell another story. B/A = 1.25 means the matrix enhances response by 25%. C/B = 0.80 means preparation recovery is 80%. Their product is 1.00. Loss during preparation and enhancement during measurement cancel in the combined signal.
This cancellation is fragile. If the next sample’s matrix factor is 1.00 while recovery remains 0.80, the combined response becomes 80%. The original apparently perfect result did not establish robustness across sample matrices.
The practical lesson is to retain the stage-specific evidence. A final answer near an expected value can occur for the wrong combination of reasons. Repeated injections of the same final extract might demonstrate stable response while leaving preparation recovery untested.
A freshly added analyte need not behave like analyte already present in the sample. Native residues may be bound, distributed within material, altered during storage or otherwise less accessible. Good recovery of a fresh addition therefore does not automatically establish complete extraction of native analyte.
Likewise, one matrix source does not represent every food, environmental or biological sample. Repeatability within one extract and variation between independently prepared samples answer different questions. Keep those levels separate in the report.
The simple model also assumes suitable background treatment, comparable final matrices and response within a useful range. Interference that contributes its own signal, changing final volumes or detector saturation can invalidate a ratio interpretation. Investigate those conditions rather than treating every discrepancy as extraction loss.
Do not apply an automatic recovery correction to unknown samples solely because one teaching ratio can be calculated. Whether correction is appropriate depends on the validated method, how representative the controls are and the uncertainty introduced. Acceptance criteria belong to the specified analytical purpose and method, not to a universal percentage chosen here.
Transfer check: suppression without preparation loss
Let A = 100, B = 50 and C = 50. The matrix factor is 0.50 and preparation recovery is 1.00. The 50% combined response does not establish a 50% extraction loss. Explain why B is the necessary reference.
Transfer check: preparation loss without a matrix effect
Let A = 100, B = 100 and C = 70. The matrix factor is 1.00; recovery is 0.70. These data support a preparation-loss interpretation under the model, while preserving the other assumptions and uncertainty.
Transfer check: a dilution mismatch
Suppose the pre-extraction preparation has twice the final volume of B but the same added amount. Its nominal final concentration is halved before any loss occurs. Can you still call C/B a recovery estimate? Not without accounting for that difference and checking the resulting matrix and response conditions.
Practical evidence check
Give a partner A = 200, B = 160 and C = 120. Ask for the three ratios and the stage each addresses. Then show the cancellation case from Chapter 4. They should explain why a 100% combined response and 100% preparation recovery are different statements.
Does post-extraction spiking measure extraction recovery?
By itself, no. The addition bypasses extraction. It can help assess final-matrix response when compared with an appropriately matched neat reference.
Does 100% combined response prove that nothing was lost?
No. Preparation loss and signal enhancement can cancel. Separate controls are needed to test those stages.
Is an internal standard enough to solve every problem?
No. Its suitability and point of addition matter. A standard added after extraction cannot experience earlier losses, and a different substance may not track every effect on the analyte.
Can these equations be applied to every analytical instrument?
Only when the preparation and response assumptions are appropriate. The cited primary work concerns HPLC–MS/MS bioanalysis; other methods need their own justified controls.
What should accompany a reported recovery?
Give the spike timing, reference comparison, concentration level, matrix, replicate structure, uncertainty and equation. A percentage without that context is difficult to interpret.
A reviewable result keeps the three preparation sets visible and names what each comparison tests. It also states which stages the controls bypass, where matching could fail and what remains unknown about native analyte.
Use the existing recovery and matrix-effect introductions for the foundation. The standard-addition guide develops a different calibration question, while commutability asks whether a material behaves appropriately across measurement procedures. Those related tools address different weaknesses; none should replace the specific stage comparison needed here.
Subject and learning map
The Gold Standard of Chemistry
Education: learning that transfers into life
How X Works: connected subject hub
Foundations of analytical evidence
Recovery testing: interpret a known addition
Matrix effects: suppression and enhancement
Sample preparation: preserve the measurement target
Analytical Chemistry: the wider measurement map
Different controls for different uncertainties
Blank corrections and honest titration results
Commutability: when a calibration material misleads
Standard addition when spike volume changes
Internal standards: compare signals with an appropriate reference
Authoritative sources
Sources checked on 21 October 2026. Numerical cases and practice tasks are original teaching examples, not observations from a published experiment or official statistics.
Matuszewski, Constanzer and Chavez-Eng: Strategies for the Assessment of Matrix Effect in Quantitative Bioanalytical Methods Based on HPLC–MS/MS — primary analytical research, Analytical Chemistry 75 (2003), 3019–3030; distinguishes matrix response from recovery.
European Union Reference Laboratories: Studies on the Extractability of Incurred Pesticides — investigates extraction of incurred residues, a distinct evidence question from recovery of a fresh spike.