Signal, Noise and the Limit of Detection
Every instrument reads something when nothing is there. How much that reading wanders sets the smallest amount that can be claimed, and the difference between detecting and quantifying is a matter of how many multiples of the noise the signal reaches.
What a learner can do afterwards
- Estimates noise from the spread of repeated blank readings
- Calculates a limit of detection and a limit of quantitation and states the multiple used for each
- Explains how averaging repeated scans improves the signal to noise ratio and at what cost
- Explains why a result below the detection limit is reported as such rather than as a number
1 · Read
Every instrument reads something when nothing is there. Run the blank again and again and watch the readings wander. The spread of those repeats, measured as a standard deviation, is your noise. Learn the twin skills: accuracy is nearness to the truth, precision is tightness of the scatter, and good analysis needs both.
Limits turn noise into rules. The detection limit is 3 times the blank noise divided by the calibration slope, and only signals standing that clear count as real. The quantitation limit is stricter at 10 times the noise over the slope. Between them sits a grey zone where something is clearly present but its amount stays guesswork.
Work one example end to end. Ten blanks scatter with a standard deviation of 0.02 signal units, and the slope is 0.5 units per ppm. Three times the noise is 0.06, divided by 0.5 gives a detection limit of 0.12 ppm. Ten times the noise is 0.2, giving a quantitation limit of 0.4 ppm. Averaging repeats shrinks noise by the square root of the count, so four repeats halve it, at the price of extra time.
Report with discipline. Below detection, write below detection limit and never a number. In the grey zone, report detected but below quantitation. A number with uncertainty is reserved for readings above quantitation. No blank record means no honest limit.
Blanks set the noise, multiples of noise set the limits, and limits set the words.
More to read
1.5 Measurement Uncertainty, Accuracy, and Precision
Every measurement carries doubt, and honest chemistry states it. Repeating a reading shows a scatter, and the width of that scatter is the uncertainty. Accuracy tells how near the middle sits to the truth, while precision tells how tight the scatter is.
The detection limit is uncertainty turned into a rule. Only a signal standing several widths clear of the blank scatter counts as real. Learn to see every instrument number as a middle plus a width, and limits stop feeling arbitrary.
3.12: Accuracy and Precision
Accuracy and precision are different skills. A tight cluster far from the true value is precise but wrong, while a loose spray around the truth is accurate on average but useless singly. Good analysis needs both, and the blank tells you where you stand.
Run blanks first and often. Their wander sets the noise floor, and three times that floor marks the smallest claimable signal. No blank record means no honest limit.
8.4 Range and Standard Deviation
Spread has a number: the standard deviation. It averages how far readings stray from their middle, so a bigger value means a noisier blank and a higher detection limit. Range gives a quick feel, but the standard deviation does the formal work.
This is why statistics problems belong in an analysis course. The same arithmetic that summarises class test scores summarises blank readings. Master spread on paper and instrument noise becomes just another dataset.
1.5: Measurement Uncertainty, Accuracy, and Precision
Quantitation demands more than detection: ten widths of noise instead of three. Between the two limits sits a grey zone where something is clearly present but its amount stays guesswork. Reporting rules respect the zones: a number with uncertainty, or a plain below limit flag.
Averaging shrinks the widths by the square root of the repeat count, buying lower limits for extra time. Four repeats halve the noise, sixteen quarter it. Spend repeats where the sample matters most.
2 · Watch
به انگلیسیTake it off screen
Where it sits
Optional connections
This opens up
12 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.