= Internal-standard-normalized matrix factor
An \b[internal-standard-normalized matrix factor] divides the <mass spectrometry matrix factor> of an <analyte> by that of an <internal standard>:
$$
\mathrm{MF}_{\rm norm}=\frac{\mathrm{MF}_A}{\mathrm{MF}_{IS}}
=\frac{A_M/A_0}{IS_M/IS_0}.
$$
Here M and0 denote matched matrix and neat-reference responses, with equal corresponding amounts and appropriate blank correction. The number asks whether matrix-dependent response changes cancel in the analyte-to-standard ratio. It does not assert that neither signal is suppressed.
For two invented matrices, let analyte factors be $(0.4,0.8)$ and standard factors also $(0.4,0.8)$. Both normalized factors are1. With standard factors $(0.8,0.4)$ instead, the normalized factors become $(0.5,2.0)$. Normalizing by a poorly matched standard can increase variation rather than remove it. The original method study distinguishes matrix-response magnitude and variability across sources.<Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS MS by B. K. Matuszewski and others>[Abstract, absolute/relative matrix effects and response/recovery distinction.]{ref}
The cancellation can be expressed algebraically. Suppose $\mathrm{MF}_A=k(1+\epsilon_A)$ and $\mathrm{MF}_{IS}=k(1+\epsilon_{IS})$, with small residual deviations. Then
$$
\mathrm{MF}_{\rm norm}\approx1+\epsilon_A-\epsilon_{IS}.
$$
A common multiplier $k$ cancels exactly, while different residual effects remain. To first order, their variance is $\operatorname{Var}(\epsilon_A)+\operatorname{Var}(\epsilon_{IS})-2\operatorname{Cov}(\epsilon_A,\epsilon_{IS})$. Correlation therefore matters; two individually variable signals can yield a stable ratio when they vary together.
This post-preparation comparison does not by itself establish matched <extraction recovery>. If analyte and standard matrix factors are equal but their recoveries are $0.5$ and1, the <internal-standard-normalized process efficiency> is still $0.5$. State when the standard was added and which response comparison was normalized. An <internal standard> is useful through demonstrated shared behavior, not through its label alone.
Back to article page
