= Analytical matrix effect
An \b[analytical matrix effect] is the combined influence of sample components other than the <analyte> on a measurement. IUPAC distinguishes the collective effect from an interference attributed to a specifically identified component.<Analytical matrix effect in IUPAC Gold Book>[M03759, definition1: analytical chemistry.]{ref} Effects can alter background, sensitivity or the chemical form reaching a detector; a <mass spectrometry matrix effect> is one important instance.
A local linear model separates two possibilities:
$$
y=b_M+s_Mc,
$$
where $c$ is <analyte> concentration, $b_M$ is a matrix-dependent offset and $s_M$ is matrix-dependent sensitivity. This is a chosen response model over a stated range, not a universal linear law. A blank estimates an offset; it does not automatically determine or restore the slope.
For invented reference parameters $b_0=1$, $s_0=2$, and sample-matrix parameters $b_M=5$, $s_M=1$, a true $c=3$ gives sample response $8$. Inverting the reference calibration gives $(8-1)/2=3.5$, an overestimate, even though the sample-matrix slope is only half the reference slope. Subtracting the matrix blank removes its offset but leaves response $3$; dividing that by the unmatched reference slope gives $1.5$. Neither arithmetic step alone handles both changes.
There is an especially deceptive point: at $c=4$, both models give response $9$. Agreement at one concentration therefore does not establish a common response curve. Compare more than one level when the question concerns sensitivity rather than just one sample indication.
<Matrix-matched calibration> aims to reproduce relevant matrix conditions. Its success depends on which constituents and concentration ranges are actually matched. <Extraction recovery> asks a different question about material retained in preparation, while <analytical blank> measurements constrain particular background contributions. Distinguishing these quantities prevents a single word such as correction from hiding different physical assumptions.
Back to article page
