📘 Beginner Tips & Guide

Understanding LOD & LOQ — A Beginner's Guide

What Limit of Detection and Limit of Quantitation actually mean, how the regression (calibration curve) method works, what Slope, Intercept, r, R² and σ represent, ICH Q2(R2) acceptance criteria, and the common mistakes that quietly distort the result.

🧪 LOD & LOQ Calculator 📘 Beginner Tips & Guide

Everything You Need to Know About LOD & LOQ

1. What Are LOD and LOQ?

Limit of Detection (LOD) is the lowest concentration of an analyte in a sample that can be reliably detected, but not necessarily quantified with precision. Limit of Quantitation (LOQ) is the lowest concentration that can be measured with acceptable precision and accuracy — it's always higher than the LOD. Both are required validation characteristics under ICH Q2(R2) for any method intended to detect or quantify trace-level analytes, such as impurity or related-substance methods.

🔍
LOD — "Is it there?"
The smallest amount your method can distinguish from background noise — confirms presence, not exact amount.
📏
LOQ — "How much?"
The smallest amount your method can measure reliably enough to report a trustworthy number.

2. The Regression (Calibration Curve) Method

This calculator uses the most widely accepted approach: build a calibration curve of peak area vs. concentration across a series of dilution levels around the expected LOD/LOQ range, then derive LOD and LOQ from the line's slope and the scatter of points around it.

Dilution Factor: entered directly per level, exactly as calculated in your source worksheet
Concentration: Wt. of Std (mg) × Dilution Factor
Slope (S): SLOPE(Avg Area, Concentration)  |  Intercept: INTERCEPT(Avg Area, Concentration)
r: CORREL(Avg Area, Concentration)  |  R²: r²
σ: STEYX(Avg Area, Concentration) — the standard error of the y-estimate
LOD: 3.3 × σ ÷ S   |   LOQ: 10 × σ ÷ S

Each level row uses its own set of area readings — however many reading columns your table currently has (start with 3, add more with ➕ Add Cell if your protocol calls for more replicates) — averaged, against its own calculated concentration. Never lock every row to a single copied-down value, which is a common spreadsheet mistake that quietly distorts the regression.

3. Why σ Is Not the Same as a Simple Standard Deviation

A common beginner mistake is computing σ as the plain standard deviation of the area readings. The correct σ for this formula is the standard error of the y-estimate (STEYX) — it measures how far the actual area readings scatter above and below the fitted regression line, not how spread out the raw readings are from their own average. A tight cluster of readings that still sits far from a straight line will have a small "simple" SD but a large STEYX — and it's the STEYX-based σ that correctly reflects the assay's real noise around the calibration curve.

4. The Signal-to-Noise (S/N) Alternative

For methods where baseline noise is clearly visible on the chromatogram (common for impurity/trace methods), Signal-to-Noise Ratio is an accepted alternative approach: LOD is typically the concentration giving an S/N of 3:1, and LOQ the concentration giving an S/N of 10:1. This calculator records the S/N Ratio you observe at each level alongside the regression data — useful for cross-checking that the regression-based LOD/LOQ lands in a sensible place relative to the S/N trend, and for the printed report.

5. Typical Acceptance Criteria

Exact limits always come from your own validation protocol/SOP, but these are commonly seen in industry:

MetricTypical Limit
System Suitability %RSD≤ 2.0% (standard replicate injections)
Regression Linearity (R²)≥ 0.99 (ICH Q2(R2) benchmark)
Number of Concentration LevelsMinimum 3 (5–6 typical, spanning the expected LOD/LOQ range)
Reference
ICH Q2(R2) — Validation of Analytical Procedures is the internationally recognized guideline defining LOD and LOQ as required validation characteristics, and describing both the regression-based and Signal-to-Noise approaches to determining them.

The calculator itself opens with 5 starter levels labeled 80 / 90 / 100 / 110 / 120 (fully editable) — rename any label, add more levels with ➕ Add Level, or remove one with its ✕. Reading columns work the same way: ➕ Add Cell adds one reading column to every level at once, and the × on a column header removes it from every level at once — just like inserting or deleting a column in Excel.

6. Common Mistakes That Distort LOD/LOQ

⚠️ Using simple SD instead of STEYX for σ
Averaging the standard deviation of raw readings instead of computing the standard error around the fitted regression line gives a σ — and therefore an LOD/LOQ — that doesn't reflect the assay's actual noise. See Section 3 above.
⚠️ Too few concentration levels, or levels bunched too closely
At least 3 levels are mathematically required for the regression; 5–6 well-spaced levels spanning the expected LOD/LOQ range give a far more reliable slope and σ than 3 levels crowded close together.
⚠️ Locking every row to one Dilution Factor value
Each level needs its own, distinct Dilution Factor — copying one row's value down the whole table collapses your calibration curve to a single concentration and silently breaks the regression.
⚠️ Trusting a low R² result
An R² below 0.99 means the calibration curve isn't sufficiently linear — the calculated LOD/LOQ from that regression shouldn't be relied on until the linearity issue (bad readings, wrong dilution, contaminated standard) is resolved.

7. Quick Checklist Before You Calculate

  • Weight of Standard entered correctly.
  • System Suitability readings within your %RSD limit.
  • At least 3 (the calculator starts you with 5, labeled 80/90/100/110/120 — edit these or add/remove levels to match your protocol) complete concentration levels — each with its own Level label, Dilution Factor, S/N Ratio, and every reading cell filled in.
  • No cell left blank — the calculator highlights any missing cell in red and won't calculate until it's filled in (or the whole row/column is removed).
  • Levels reasonably spaced and centered around where you expect the LOD/LOQ to fall.
  • R² reviewed against your acceptance benchmark (≥ 0.99 is the common ICH-aligned figure) before relying on the calculated LOD/LOQ.
Ready to calculate?
Back to the LOD & LOQ Calculator
Enter your System Suitability and concentration-level area readings in the Excel-style table and get your Slope, R², LOD and LOQ instantly.
Open Calculator