The separation starts before injection
A technically excellent separation cannot recover information lost before injection. The result applies to a defined measurand—the quantity intended to be measured—and to the sampling target and conditions used to obtain it. Claims about a field site, production lot, patient, cell population, or time interval are credible only when the collected material represents that target adequately.
Sampling design defines the target, location, timing, pattern, number of increments, replication, controls, preservation, and the decision the data must support. Laboratory preparation then manages heterogeneity, matrix, concentration, compatibility, and stability. When the measurand refers to the sampling target rather than only the material in the vial, sampling and physical preparation can be major contributors to measurement uncertainty.
From target to injected portion
- The sampling plan states what will be collected, where, when, how, and why.
- The laboratory sample is the material delivered to the laboratory.
- A test sample is prepared from it by operations such as drying, grinding, homogenizing, or splitting.
- The test portion is the amount actually taken for one measurement procedure.
Each transition can change composition. Homogenization can reduce subsampling variation, but only if it is appropriate for the analytes and matrix; it can also introduce heat, contamination, volatilization, or chemical change.
Common operations
- filtration and centrifugation;
- homogenization, comminution, and representative subsampling;
- liquid–liquid and solid-phase extraction;
- solid-phase microextraction and headspace sampling;
- precipitation and digestion;
- desalting, buffer exchange, and cleanup;
- derivatization;
- evaporation, dilution, and preconcentration;
- cell lysis and nucleic-acid purification.
Failure modes
- nonrepresentative collection;
- contamination from tools, containers, air, reagents, or handling;
- adsorption and analyte loss on surfaces;
- degradation during storage or transport;
- variable extraction efficiency;
- matrix components that alter separation or detector response;
- blanks and controls that do not follow the real workflow;
- low-biomass signals that approach contamination backgrounds.
Quality controls that follow the workflow
No single control diagnoses every problem. Field and equipment blanks can reveal contamination during collection; transport and procedural blanks follow other parts of the route. Replicate samples help estimate heterogeneity and repeatability. Fortified samples or matrix spikes can test recovery in a particular matrix when chemically appropriate, while reference materials provide an independent comparison when a suitable material exists.
Control samples should experience the relevant operations, not enter only at the final instrument vial. Chain-of-custody records, sample identifiers, timestamps, preservation conditions, preparation history, and instrument sequence together create traceability from result back to collected material.
Connecting preparation to the method
GC often requires volatility-compatible introduction. LC requires solvent and matrix compatibility with the column and detector. CE requires suitable conductivity, pH, and sample-zone conditions. Microfluidic systems have small volume advantages but can be particularly sensitive to particles, bubbles, adsorption, and interfaces.
The right preparation is therefore method-specific. Recovery, selectivity, contamination, throughput, safety, stability, and uncertainty across the complete workflow should be considered together. A high instrumental recovery from the final vial does not establish representative sampling upstream.