Current turnaround: 21 days from receipt of your samples. Due to high demand, HPLC-UV peptide testing and Certificates of Analysis are taking longer than usual.

Independent Research Testing Center · 875 Innovation Drive, Cambridge, MA 02139, USA For all independent testing inquiries, email admin@novagenlaboratories.com
Analytical methods

Limit of Detection and Limit of Quantitation

Two numbers that rarely appear on a certificate but govern everything that does. Limit of detection and limit of quantitation define the floor of what a method can see, and they determine whether a "not detected" result means anything useful at all.

Novagen Analytical Labs

The limit of detection (LOD) and the limit of quantitation (LOQ) are the two numbers that define the floor of what an analytical method can see. LOD is the lowest concentration at which an instrument can reliably distinguish an analyte signal from background noise. LOQ is the lowest concentration at which that signal can be measured with acceptable accuracy and precision. When a certificate of analysis reports a result as "not detected", it is making a statement bounded by the LOD of that method. It is not saying the analyte is absent. It is saying the instrument produced no signal above the threshold this method can distinguish from noise. Those two things are not the same, and the difference between them matters every time a certificate is used to make a decision about a sample.

What LOD and LOQ actually define

Every analytical instrument produces a signal. It also produces noise: random fluctuations in the baseline that have nothing to do with the analyte being measured. The question at low concentrations is whether a small signal from the analyte can be distinguished from that noise with enough confidence to mean anything.

The LOD answers the first question: can we see it? At the LOD, by convention, the analyte signal is approximately three times the magnitude of the baseline noise. That 3:1 ratio provides enough statistical separation to say, with reasonable confidence, that something is there. It does not provide enough information to say how much of it is there.

The LOQ answers the second question: can we measure it? At the LOQ the signal is approximately ten times the noise. At that ratio, measurement precision and accuracy are sufficient for the number to be reported as a quantity with meaning. The LOQ is higher than the LOD. The gap between them is a region where detection is possible but reliable quantitation is not.

Between the LOD and the LOQ, a laboratory is in an uncertain position. A signal exists. Whether to report it as a trace, an estimate, or simply as "detected but not quantifiable" depends on the method and the purpose of the analysis. A certificate that collapses this into a single "not detected" entry is not wrong, but it requires the reader to understand what detection threshold is implied.

Three ways to calculate them

ICH Q2(R1), the standard guideline for pharmaceutical analytical-method validation, describes three approaches to establishing LOD and LOQ. They give different numerical results for the same method, which is why validation reports must state which approach was used.

Signal-to-noise approach. The analyst injects progressively lower concentrations of a known standard and measures the ratio of the analyte peak height to the peak-to-peak noise in the adjacent baseline. The LOD is set at the concentration producing a signal roughly three times the noise. The LOQ is set at the concentration producing a signal roughly ten times the noise. This approach is straightforward for chromatographic methods where baseline noise is visible and measurable.

Calibration-curve approach. The analyst constructs a calibration curve from several low-concentration standards and calculates the standard deviation of the response at those concentrations together with the slope of the curve. The LOD and LOQ are then calculated from the formulas:

  • LOD = 3.3 × sigma / S
  • LOQ = 10 × sigma / S

where sigma is the standard deviation of the response and S is the slope of the calibration line. This is the approach most commonly cited in method validation documents for pharmaceutical impurity testing. It has the advantage of being grounded in actual calibration data rather than a visual judgement about where noise ends and signal begins.

Blank-based approach. The analyst runs a series of replicate blank measurements, typically ten or more, and calculates the standard deviation of those responses. The LOD is set at the concentration corresponding to the mean blank signal plus three standard deviations. The LOQ is set at the mean blank plus ten standard deviations, converted to concentration units using the calibration slope. This approach is used where the blank itself produces a measurable response, as can occur with complex matrices.

Each approach is accepted under ICH Q2(R1). A method validated by one approach will not necessarily give the same LOD or LOQ as the same method validated by another. Neither result is wrong. They are answers to slightly different questions about the same underlying data.

Why "not detected" is not the same as "not present"

A result of "not detected" on a certificate means the instrument produced no signal above the method's detection threshold in the sample submitted. The substance may still be present at a concentration below that threshold. A method with a LOD of 0.05% cannot see anything below 0.05%. If a contaminant is present at 0.03%, the certificate will read "not detected". The certificate is accurate. The substance is still there.

This is not a flaw in analytical testing. It is a physical reality. No instrument can detect infinitely small concentrations. What matters is whether the LOD of the method is low enough to be meaningful for the decision being made. For a certificate to say something useful about a particular impurity, the method's LOD must sit below the concentration that matters for that impurity. If it does not, "not detected" is a statement about the instrument, not about the sample.

The same logic applies to identity testing and to any other measurement reported on a certificate. A certificate describes what was measured in the sample submitted, with the precision and sensitivity the method provides, and nothing beyond that. The LOD and LOQ are the numbers that define exactly what "nothing beyond that" means in practice.

See also the discussion of impurity peaks and what gets counted in our note on peptide impurities and where they come from.

How method sensitivity sets the floor

The LOD and LOQ of a chromatographic method are not fixed properties of the analyte. They are properties of the method as operated under specific conditions. Several practical factors determine where the floor sits.

Detector type. A UV detector measures absorbance at a chosen wavelength. A fluorescence detector measures emission after excitation and can be orders of magnitude more sensitive for compounds that fluoresce. A mass spectrometer can achieve lower detection limits still. The choice of detector is the single largest determinant of how low the method can go.

Detection wavelength. For UV detection, sensitivity depends on how strongly the analyte absorbs at the chosen wavelength. Measuring at or near the analyte's absorption maximum produces the strongest signal for a given concentration and therefore the lowest LOD. Measuring away from the maximum reduces signal strength and raises the detection floor. This is one reason different laboratories using nominally similar methods may report different LODs for the same compound.

Injection volume. Increasing the volume of sample injected onto the column delivers more analyte to the detector. This raises the signal without changing the noise floor, effectively lowering the LOD and LOQ. There are practical limits: too large an injection overloads the column, distorts peaks, and degrades the accuracy of area measurements. Within those limits, injection volume is a straightforward lever for improving sensitivity.

Baseline noise. Noise is affected by instrument condition, mobile phase purity, column cleanliness, and ambient temperature stability. A noisy baseline raises the threshold at which a signal can be distinguished from background. Routine instrument maintenance and mobile phase preparation practices directly affect where the LOD sits in everyday operation.

All of this means that when comparing LOD and LOQ values across different laboratories or different method versions, the conditions under which those limits were established must be known. A lower LOD from one laboratory does not necessarily mean better analytical capability. It may reflect a different injection volume, a different wavelength, or a different approach to calculating the limit.

The relationship between peak area measurement and how the method converts signal to a reported percentage is covered in why purity is measured by peak area.

Reporting thresholds and the certificate

A certificate of analysis does not report everything the instrument could theoretically detect. It reports what the method was validated to report reliably, above a defined reporting threshold. The relationship between that threshold and the LOQ is direct: the method's LOQ must sit at or below the reporting threshold the analysis is intended to cover.

In pharmaceutical impurity testing, reporting thresholds are set by guidelines that specify the minimum impurity level requiring disclosure. For the method to support those thresholds, its LOQ must be at or below them. If a guideline requires reporting impurities at or above 0.05% and the method's LOQ is 0.10%, impurities between 0.05% and 0.10% will not appear on the certificate regardless of whether they are present. The method cannot reliably quantify them.

This has a direct consequence for reading a certificate. The total impurity figure reported is the sum of impurities detected and reliably quantified above the reporting threshold. It is not a complete inventory of everything present in the sample. What appears below the LOQ, or below the reporting threshold, simply does not get counted. A lower total impurity figure on a certificate with a high LOQ may reflect method sensitivity as much as sample purity.

This is one of the more important distinctions between laboratories using different methods. Two certificates for the same compound, both reporting 98% purity, may describe very different situations if their LOQs differ. The certificate with the more sensitive method has seen more of the sample. The reasons why different laboratories report different purity values for the same compound are explored in more detail in why two labs report different purity for the same vial.

The region between LOD and LOQ

Between the LOD and the LOQ is a concentration range where measurement becomes unreliable. A signal is visible. The instrument is responding to something. But the precision and accuracy at that level are insufficient for the number to be reported as a reliable quantitative result.

Different laboratories handle this region differently. Some report the result as "detected, not quantifiable". Some apply the value with a qualifier flag. Some report "not detected" with the understanding that this means below the LOD rather than below the LOQ. The handling is a methodological choice and it affects what a certificate means.

When a certificate is being used to assess whether a particular impurity falls below a threshold that matters, the question to ask is not only what the reported value is but where the LOD and LOQ of the method sit relative to that threshold. If the threshold is at 0.1% and the LOQ is at 0.15%, the method cannot confirm compliance with that threshold by measurement alone. The number being reported as "not detected" or as a small percentage below threshold carries less weight than it appears to.

What this means when you read a certificate

A certificate of analysis reports what was measured in the sample submitted. The LOD and LOQ are the parameters that define the limits of that measurement. They are not always stated on the certificate face, but they govern every entry on it.

When a certificate reports "not detected" for a particular analyte, that result has a meaning bounded by the LOD. It means the instrument saw no signal above the detection threshold of this method, in this sample, on this run. It does not mean the analyte is absent from the sample at all concentrations. It does not mean a more sensitive method would reach the same conclusion.

When a certificate reports a purity figure, that figure counts everything the method could reliably quantify above the LOQ and the reporting threshold. Anything below those levels is not in the number. A high purity figure from a low-sensitivity method and a high purity figure from a high-sensitivity method are not the same statement about the sample, even when the numbers match.

The honest reading of a certificate requires knowing the method's limits as well as its results. The LOD and LOQ are where those limits sit.

Novagen publishes a verifiable register of certificates at novagenlaboratories.com. Enquiries about method parameters, including detection and quantitation limits for specific analyses, can be directed to admin@novagenlaboratories.com. Current quoted turnaround is 21 days from receipt of samples.

Common questions

What is the difference between limit of detection and limit of quantitation?

The limit of detection (LOD) is the lowest concentration at which an instrument can reliably distinguish a signal from background noise. The limit of quantitation (LOQ) is higher: it is the lowest concentration at which that signal can be measured with acceptable accuracy and precision. At the LOD you can say something is there. At the LOQ you can say how much of it is there. Between the two, a result exists in an uncertain middle ground where detection is possible but reliable measurement is not.

Does "not detected" mean a substance is absent?

No. "Not detected" means the instrument produced no signal above the method's detection threshold. The substance may still be present at a concentration below the LOD. The statement is about the sensitivity of the method, not about the composition of the sample. A more sensitive method, or a larger injection volume, or a different detector, may find what the first method missed.

How is the limit of detection calculated?

ICH Q2(R1), the standard analytical-method-validation guideline, describes three approaches. The signal-to-noise approach sets the LOD at the concentration where the analyte signal is roughly three times the baseline noise. The calibration-curve approach uses the formula LOD = 3.3 × sigma / S, where sigma is the standard deviation of the response at low concentrations and S is the slope of the calibration curve. The blank-based approach measures the standard deviation of replicate blank responses and sets the LOD at the mean blank plus three standard deviations, converted to concentration units using the calibration slope. Each approach gives a different numerical result for the same method, which is why validation reports must state which approach was used.

What is the signal-to-noise ratio for LOQ?

By convention, the LOQ is set at the concentration where the analyte signal is approximately ten times the baseline noise, written as S/N of 10:1. This is roughly three times the LOD established by the 3:1 signal-to-noise criterion. The 10:1 ratio is accepted in ICH Q2(R1) for chromatographic methods because at that level a measurement typically meets the precision and accuracy requirements needed for quantitative reporting.

Why does the reporting threshold on a certificate depend on the LOQ?

A laboratory can only report a number it can measure reliably. The LOQ must sit at or below the reporting threshold the method is intended to cover. If an impurity guideline requires reporting at 0.05%, the method's LOQ must be at or below 0.05%. If the LOQ is 0.1%, the method cannot support that reporting requirement and impurities between 0.05% and 0.1% will simply not appear on the certificate, not because they are absent but because the method cannot see them reliably at that level.

Can the LOD and LOQ be improved without changing the analytical technique?

Yes, within limits. Using a larger injection volume increases the mass of analyte reaching the detector and raises the signal without changing the noise floor, lowering both limits. Selecting a UV detection wavelength closer to the analyte's absorption maximum increases response. Using a more sensitive detector, such as a fluorescence detector for compounds that fluoresce, can reduce the noise floor directly. Each of these changes the method, however, and the new limits must be validated separately. LOD and LOQ are properties of a specific method as run on a specific instrument under specific conditions, not fixed properties of the analyte.