Gas Sensor Linearity & Linear Range Explained

2026-09-23 Gas Nose 0 Comments

Gas sensor datasheets often use terms such as:

Sensitivity
Linearity
Linear Range
Measurement Range
Maximum Detecting Concentration
Overload

These terms are related, but they do not mean the same thing.

A sensor may have high sensitivity but poor linearity. It may continue producing a signal beyond its linear range. It may even survive gas concentrations far above its normal measurement range without being suitable for accurate measurement there.

The most useful distinction is:

Sensitivity describes the slope of the response. Linearity describes how consistently the response follows a straight-line relationship. Linear range describes the concentration interval over which that relationship remains acceptably linear.

This matters because many gas-concentration calculations assume that a change in sensor output corresponds predictably to a change in concentration.

If that assumption is wrong, a simple one-slope conversion can produce increasingly large errors.

Linearity, Sensitivity and Range: Quick Comparison

Parameter	Main Question
Sensitivity	How much does output change per unit change in gas concentration?
Linearity	How closely does the actual response follow a defined straight-line relationship?
Linear Range	Over what concentration interval is the response sufficiently linear?
Measurement Range	Over what concentration interval is the device specified for measurement?
Maximum / Overload Concentration	What higher exposure can the sensor tolerate or encounter under specified conditions?

A simple way to remember them is:

Sensitivity = slope. Linearity = straightness. Linear range = where that straightness is valid.

What Is Gas Sensor Linearity?

For an ideal linear gas sensor, output can be represented approximately as:

Y = S × C + b

where:

Y = sensor output
S = sensitivity or slope
C = gas concentration
b = zero offset

For example:

0 ppm  → 0 mV
10 ppm → 100 mV
20 ppm → 200 mV
30 ppm → 300 mV

Each 10 ppm increase produces the same 100 mV output increase.

That is an ideal linear relationship.

Real sensors rarely follow a mathematically perfect straight line at every point. In engineering use, linearity describes how closely the measured calibration curve follows the chosen linear reference within a specified range and tolerance.

NI describes sensor linearity as the extent to which the actual measured response departs from an ideal or fitted straight-line response.

See: NI — Sensor Terminology

What Does a Linear Gas Sensor Response Look Like?

A linear sensor produces approximately equal output increments for equal concentration increments.

For example:

Gas concentration:
100 → 200 → 300 → 400 ppm

and:

Sensor output:
1 → 2 → 3 → 4 units

The slope remains approximately constant.

A nonlinear sensor may instead produce:

100 ppm → 2.5 units
200 ppm → 3.5 units
300 ppm → 4.0 units
400 ppm → 4.3 units

The sensor still responds to increasing concentration, but the amount of additional output becomes smaller at higher concentrations.

That does not automatically make the sensor unusable.

It means that a simple constant-slope conversion may no longer be sufficient.

Sensitivity vs Linearity

Sensitivity and linearity are often confused because both describe the concentration-response curve.

Sensitivity

Sensitivity is essentially the slope:

Sensitivity
=
Δ Output
÷
Δ Concentration

Linearity

Linearity describes whether that slope remains sufficiently consistent across the specified interval.

Consider two hypothetical sensors.

Sensor A

0–100 ppm:
approximately 10 nA/ppm across the range

This sensor has a relatively consistent slope.

Sensor B

0–10 ppm:
15 nA/ppm

80–100 ppm:
5 nA/ppm

Sensor B may be very sensitive at low concentration, but its slope changes considerably with concentration.

So:

A highly sensitive sensor is not automatically a highly linear sensor.

For the wider distinction between sensitivity and gas discrimination, see Gas Sensor Sensitivity vs Selectivity: What’s the Difference?.

What Is the Linear Range?

The linear range is the concentration interval over which the sensor’s concentration-output relationship remains within the manufacturer’s or measurement method’s specified linearity criterion.

Suppose the response behaves approximately linearly from:

0–100 ppm

but begins to curve above:

100 ppm

Then the useful linear range may be:

0–100 ppm

This does not necessarily mean the sensor stops responding at 101 ppm.

It means the same simple linear conversion may no longer meet the specified performance requirement beyond that region.

That distinction is important.

The end of the linear range is not automatically the end of all sensor response.

Linear Range vs Measurement Range

These terms are related, but they should not be treated as universal synonyms.

Linear Range

The concentration interval over which the sensor response satisfies the defined linearity requirement.

Measurement Range

The concentration interval over which the sensor or complete instrument is specified to provide usable measurements.

Depending on the product:

linear range may cover most or all of the measurement range
the measurement range may extend into a region that requires nonlinear compensation
a manufacturer may simply specify the response as “linear” throughout the stated range without publishing a separate numeric linear-range value

This is why you should read the actual datasheet rather than assume one universal relationship.

For example, current Winsen electrochemical sensor pages commonly list Measurement Range, Max Detecting Concentration, Sensitivity, Resolution, Output Linearity, and Zero Drift as separate parameters.

A chlorine sensor may be specified with a normal measurement range of 0–10 ppm and a much higher maximum detecting concentration of 100 ppm, while output linearity is listed separately.

See: Winsen ME3-CL2 Electrochemical Chlorine Sensor

Linear Range vs Maximum Detecting Concentration

This is one of the most important datasheet distinctions.

Suppose a sensor states:

Measurement Range: 0–10 ppm
Maximum Detecting Concentration: 100 ppm

It would be wrong to conclude:

The sensor accurately measures from 0 to 100 ppm.

The higher number may describe:

maximum short-term exposure
maximum detectable concentration
overload capability
a region in which response is no longer guaranteed to meet normal accuracy or linearity specifications

The exact meaning depends on the manufacturer.

For example, Winsen’s ME3-CL2 currently lists:

Measurement Range: 0–10 ppm
Max Detecting Concentration: 100 ppm
Output Linearity: linear

These are separate specifications, not one combined 0–100 ppm calibrated measurement range.

The same pattern appears across several current electrochemical sensor pages, including H₂, HF, NO₂ and O₃ models.

This leads to an important engineering rule:

Do not treat a maximum exposure or maximum detecting concentration as a normal quantitative measurement range unless the manufacturer explicitly specifies it that way.

For the broader distinction, see Gas Sensor Measurement Range vs Linear Range vs Overload.

Why Do Gas Sensors Become Nonlinear?

A concentration-response curve can become nonlinear for many different reasons.

The mechanism depends on the sensing technology.

Diffusion Limitation

In diffusion-limited sensors, gas transport through a membrane, capillary or barrier may no longer scale perfectly with external concentration.

Reaction Kinetics

Electrode or catalytic reaction behavior can change as concentration increases.

Surface Saturation

A semiconductor sensing surface may approach a state where additional gas produces progressively smaller changes.

Optical Behavior

Optical gas measurement follows an absorption relationship that may require nonlinear conversion between raw detector signal and concentration.

Oxygen Availability

Catalytic combustion sensors may behave differently when combustible concentration becomes large enough for oxygen availability or thermal balance to matter.

Electronics Saturation

The sensor element may remain responsive while the analog front end, ADC or firmware reaches its own limit.

So when a detector becomes nonlinear, the cause is not always the sensing material itself.

Electrochemical Gas Sensor Linearity

Electrochemical gas sensors often provide a useful near-linear relationship between target-gas concentration and output current over their designed range.

For a diffusion-controlled amperometric sensor, the simplified behavior is:

Gas concentration
↓
Gas flux to working electrode
↓
Electrochemical reaction
↓
Sensor current

Within the intended range:

Current ∝ Concentration

This is one reason electrochemical sensors are widely used for quantitative toxic-gas measurement.

Analog Devices notes that electrochemical gas sensors offer useful output linearity in addition to low power and good sensitivity.

See: Analog Devices — Overcoming the Technical Challenges of Electrochemical Gas Sensing

Current commercial product specifications reflect this as well. Winsen electrochemical CO, H₂, Cl₂, NO₂, HF and other sensor pages separately identify sensitivity, measurement range and output linearity.

However:

“Linear” on a datasheet does not mean zero measurement error at every concentration.

Sensor-to-sensor sensitivity variation, zero offset, temperature, humidity, calibration and electronics still matter.

MOS Gas Sensors Are Often Intrinsically Nonlinear

Metal-oxide semiconductor sensors frequently have a curved concentration-response relationship.

Depending on the sensing material and gas, the relationship may be represented using:

power-law models
logarithmic models
empirical calibration equations
lookup tables

A simplified MOS relationship is sometimes written in forms such as:

Rs / R0 = A × C^(-B)

where:

Rs = sensor resistance in gas
R0 = reference resistance
C = gas concentration
A, B = empirical constants

The important point is not the exact equation.

It is that:

A nonlinear response can still be useful and repeatable when the calibration model correctly describes it.

Nonlinear does not mean inaccurate by definition.

NDIR Sensors and Linearity

NDIR sensors use infrared absorption to infer gas concentration.

The underlying optical relationship is often described through Beer-Lambert-type absorption behavior.

However, the raw detector voltage is not necessarily a straight-line function of gas concentration across the full range.

Practical NDIR systems may use:

logarithmic conversion
polynomial calibration
lookup tables
multi-point calibration
temperature compensation

NIST notes in its calibration guidance that even when an instrument response is nonlinear, a linear calibration model may still provide accurate results if the calibration interval is sufficiently restricted.

See: NIST — Calibration

This is a useful principle for gas sensors generally:

Global nonlinearity does not prevent a sensor from having a useful local linear range.

Catalytic Combustible Gas Sensor Linearity

Catalytic-bead combustible sensors convert oxidation heat into an electrical bridge response.

Across part of the combustible-gas range, response can be treated as approximately proportional to concentration.

At higher concentrations, however, the curve may change because of:

catalyst behavior
thermal effects
oxygen availability
diffusion
bridge characteristics

This is particularly important near or above the normal calibrated range.

A detector designed for:

0–100% LEL

should not be assumed to produce a perfectly proportional signal at every combustible concentration outside its certified or specified operating range.

What Is Nonlinearity Error?

Nonlinearity error describes the deviation between the actual sensor response and a chosen linear reference.

Conceptually:

Nonlinearity Error
=
Actual Output
-
Reference Linear Output

The maximum deviation may be reported as:

percentage of full scale
percentage of reading
absolute output
concentration equivalent

NI notes that linearity is often specified in terms of percentage nonlinearity and that the calculation depends on how the reference line is defined.

Possible reference methods include:

end-point line
best-fit line
least-squares fit
manufacturer-defined reference

Therefore:

Two sensors with the same published “±2% linearity” are not automatically directly comparable unless the calculation method is also comparable.

Linearity Is Not the Same as Accuracy

A perfectly straight response can still be wrong.

Suppose the true concentrations are:

but the sensor reports:

The relationship is extremely linear.

But the slope is wrong.

After correct span calibration, this may be fixable.

The example demonstrates:

A straight calibration curve can still be inaccurate.

Accuracy also depends on:

correct slope
correct zero
reference-gas accuracy
environmental compensation
signal processing

For the distinction between measurement correctness, resolution and detection capability, see Gas Sensor Detection Limit: LOD vs Resolution vs Accuracy.

Linearity Is Not Repeatability

A sensor may repeatedly give:

52 ppm
52 ppm
51 ppm
52 ppm

when the true concentration is:

50 ppm

Its repeatability may be excellent.

Its calibration may still have an offset or slope error.

Similarly, a sensor can be linear but not repeatable if individual measurements fluctuate significantly.

For this distinction, see Gas Sensor Repeatability vs Accuracy: What’s the Difference?.

Linear Range Is Not the Same as LOD

Suppose:

Linear Range: 0–100 ppm

This does not prove that the sensor can reliably distinguish:

0.001 ppm

from zero.

Low-concentration capability depends on:

sensitivity
baseline noise
zero stability
resolution
LOD
environmental effects

A sensor can be linear across a wide range but still have an LOD far above zero.

See Gas Sensor Detection Limit: LOD vs Resolution vs Accuracy.

What Happens Near Saturation?

At sufficiently high concentration, many sensor responses begin to compress.

Conceptually:

Linear response
↓
Increasing curvature
↓
Compression
↓
Saturation / plateau

Gas concentration continues increasing, but sensor output increases less and less.

If a linear conversion is still applied:

Concentration
=
Signal
÷
Sensitivity

the instrument may significantly underestimate the real gas concentration.

This can be especially important in safety systems.

A reading that appears to be:

100 ppm

may not mean the actual environment is only 100 ppm if the sensor or electronics are already saturated.

For the full topic, see Gas Sensor Saturation and Over-Range Explained.

Calibration Curve vs Linear Range

A calibration curve is the mathematical relationship used to convert sensor output into concentration.

It can be:

Linear

y = ax + b

Polynomial

y = ax² + bx + c

Logarithmic

Power Law

Piecewise Linear

Lookup Table

The correct model depends on sensor behavior and required accuracy.

This leads to an important principle:

A sensor does not need to be globally linear to be accurately calibrated.

It needs a calibration model that adequately represents its response over the intended measurement range.

Two-Point Calibration: When Is It Enough?

For a sensor that is sufficiently linear across the application range, a two-point calibration can establish:

zero / intercept
span / slope

For example:

Point 1: 0 ppm
Point 2: 50 ppm

A linear relationship is then fitted between them.

This is practical and widely used for many gas detectors.

But two-point calibration assumes that the response between and around the calibration points is adequately represented by the linear model.

If the response is strongly curved, that assumption can become inaccurate.

When Is Multi-Point Calibration Better?

Gas Sensor Linearization and Calibration Methods

Multi-point calibration is useful when you need to:

verify linearity
characterize curvature
fit a nonlinear model
improve accuracy over a wide range

Example calibration points:

0 ppm
10 ppm
25 ppm
50 ppm
75 ppm
100 ppm

The resulting data may support:

multi-point linear regression
piecewise-linear conversion
polynomial fitting
logarithmic or power-law conversion
lookup-table interpolation

The best approach depends on:

sensor principle
measurement range
required accuracy
computational resources

NIST’s calibration guidance likewise emphasizes that the calibration model should match the actual instrument response.

Two-Point Calibration vs Multi-Point Linearity Testing

These are related but not identical tasks.

Two-Point Calibration

Used to adjust:

offset
slope

Multi-Point Linearity Testing

Used to determine whether intermediate concentrations follow the assumed calibration relationship.

A detector can pass zero and span calibration while still having significant nonlinearity between those points.

For example:

0 ppm → correct
100 ppm → correct
50 ppm → reads 60 ppm

The two endpoints are correct.

The middle of the curve is not.

This is why multi-point testing is valuable when linearity itself matters.

Why R² Does Not Tell the Whole Story

Linear regression often produces an R² value.

A result such as:

R² = 0.999

looks impressive.

But R² alone does not prove that every concentration point meets the application’s allowable error.

Over a wide measurement range, a model can have a very high R² while still showing unacceptable deviation at one important low-concentration region.

Therefore also examine:

residuals
maximum deviation
low-concentration error
full-scale error
application-specific tolerances

High R² supports a good overall fit; it is not a substitute for checking actual measurement error.

Temperature Can Change Linearity

A calibration curve characterized at:

25°C

may not remain identical at:

-20°C

or:

50°C

Temperature can affect:

sensitivity
zero current
diffusion
reaction kinetics
MOS surface chemistry
optical sources and detectors
analog electronics

NI specifically notes that static nonlinearity can be influenced by environmental factors such as temperature and humidity.

See: NI — Sensor Terminology

For gas-specific effects, see How Temperature Affects Gas Sensor Readings.

Humidity Can Change the Calibration Relationship

Humidity can affect:

electrochemical electrolyte behavior
MOS surface adsorption
VOC response
optical measurement conditions

The effect may not be identical at every gas concentration.

As a result, a calibration curve obtained at:

50% RH

may not perfectly represent the sensor at:

90% RH

For the broader topic, see How Humidity Affects Gas Sensors.

Drift Can Change the Calibration Curve

Gas sensor drift can alter either the intercept or the slope.

Zero Drift

The baseline moves.

In a simple linear model:

Y = S × C + b

the b term changes.

Span Drift

Sensitivity changes.

The S term changes.

This means that even a sensor with excellent initial linearity can become inaccurate if its zero or sensitivity changes over time.

See What Is Gas Sensor Drift? Zero Drift, Span Drift & Baseline Drift Explained.

How to Validate Gas Sensor Linearity

A practical linearity test can follow this workflow.

Step 1 — Define the Required Range

Choose the actual concentration interval that matters for the application.

Step 2 — Select Multiple Reference Concentrations

Use certified or appropriately traceable gas standards.

For example:

0%
20%
40%
60%
80%
100% of the intended range

Step 3 — Stabilize the Sensor

Control:

warm-up
bias
temperature
humidity
flow

according to the sensor or instrument specification.

Step 4 — Apply Each Concentration

Record the stabilized output under comparable conditions.

Step 5 — Plot Concentration vs Output

Create the calibration curve.

Step 6 — Fit the Intended Model

For example:

linear regression
piecewise fit
nonlinear model

Step 7 — Examine Residuals

At each point calculate how far the actual response differs from the model.

Step 8 — Compare With the Required Tolerance

The relevant tolerance may be defined in:

ppm
% reading
% full scale
voltage/current units

Step 9 — Verify With Independent Points

If the measurement is critical, validate the model using concentrations not used to construct the fit.

The goal is not to produce the prettiest graph.

It is to determine whether the calibration model is accurate enough for the real measurement problem.

How to Read Gas Sensor Linearity in a Datasheet

When a datasheet says:

Output Linearity: Linear

do not stop reading.

Check the following.

1. Sensitivity

What output change is expected per unit concentration?

2. Measurement Range

Over what concentration interval is the sensor intended to measure?

3. Linear Range

Is a separate linear range specified?

4. Linearity / Nonlinearity Tolerance

Is there a numeric specification such as:

±2% FS

or another error limit?

5. Reference Method

Is nonlinearity measured relative to:

end-point line
best-fit line
another manufacturer-defined line?

6. Maximum Detecting / Exposure Concentration

Do not confuse this with calibrated range.

7. Temperature

At what temperature was the response characterized?

8. Humidity

At what RH were the values measured?

9. Calibration Method

Was the sensor characterized with:

one span point
several concentrations
nonlinear compensation?

10. Typical vs Guaranteed Values

A typical response curve is not the same as a guaranteed maximum specification.

A Practical Datasheet Example

Consider this illustrative example:

Measurement Range: 0–100 ppm
Linear Range: 0–80 ppm
Sensitivity: 10 nA/ppm
Nonlinearity: ±2% FS
Maximum Exposure: 500 ppm

0–80 ppm

The simple linear conversion is specified to remain within the stated linearity criterion.

80–100 ppm

The sensor may still provide a usable measurement, but the manufacturer may specify different error behavior or nonlinear correction.

100–500 ppm

This should not automatically be treated as a quantitative measurement region.

The sensor may simply be able to tolerate or respond to the exposure without immediately failing.

This example is intentionally generic.

Always use the definitions in the specific sensor datasheet.

Linearity and Response Time Are Different

A sensor can be:

very linear
but slow

or:

less linear
but fast

Linearity describes a static concentration-output relationship.

Response time describes how quickly the sensor reaches that relationship after concentration changes.

Do not use one as a substitute for the other.

For T50 and T90, see Gas Sensor Response Time: T50 vs T90 Explained.

Linearity and Hysteresis Are Different

A sensor may follow one curve when concentration increases and another when concentration decreases.

This is hysteresis.

Even if each direction looks approximately linear, the two curves may not overlap.

That creates another source of measurement error that cannot be fully described by a single linearity number.

See Gas Sensor Hysteresis Explained.

Common Gas Sensor Linearity Mistakes

Mistake 1 — Sensitivity = Linearity

False.

Sensitivity is slope. Linearity describes deviation from a straight-line relationship.

Mistake 2 — Measurement Range = Linear Range

Not necessarily.

Read the manufacturer definitions.

Mistake 3 — Maximum Exposure = Measurement Range

False.

A sensor may survive or respond beyond its calibrated measurement range without providing accurate concentration data there.

Mistake 4 — Linear = Perfectly Accurate

False.

A straight response can still have the wrong slope or offset.

Mistake 5 — High R² = Every Point Is Accurate

False.

Check residuals and application-specific error.

Mistake 6 — Two-Point Calibration Proves Linearity

False.

Two points define a line; they do not prove that intermediate points follow it.

Mistake 7 — Nonlinear = Bad Sensor

False.

A well-characterized nonlinear sensor can be highly useful with an appropriate calibration model.

Mistake 8 — Ignoring Saturation

A linear conversion can severely underestimate gas concentration once the response begins to compress.

Mistake 9 — Ignoring Temperature and Humidity

Environmental conditions can change both slope and curve shape.

Mistake 10 — Comparing Linearity Numbers Without Checking the Method

Different reference-line definitions can produce different reported nonlinearity values.

Gas Sensor Linearity Selection Checklist

Before choosing a sensor based on linearity, confirm:

Target gas
Required measurement range
Sensitivity
Linear range
Measurement range
Maximum detecting / overload concentration
Linearity error specification
Reference-line method
Calibration model
Number of calibration points
Temperature conditions
Humidity conditions
Drift
Saturation behavior
Sensor vs complete-module specification
Required accuracy in the actual application

Frequently Asked Questions

What is gas sensor linearity?

Gas sensor linearity describes how closely the sensor’s output follows a specified straight-line relationship with gas concentration over a defined range.

What is the linear range of a gas sensor?

The linear range is the concentration interval over which the concentration-output relationship remains within the defined linearity requirement.

What is the difference between sensitivity and linearity?

Sensitivity describes the slope of the response.

Linearity describes how consistently the response follows a straight-line relationship.

Can a sensor be highly sensitive but nonlinear?

Yes.

A sensor may produce a large signal at low concentration while its slope changes significantly as concentration increases.

What is the difference between linear range and measurement range?

Linear range describes where the response meets a linearity criterion.

Measurement range describes where the sensor or instrument is specified to provide usable measurements.

They may be identical for some products and different for others.

Is maximum detecting concentration the same as measurement range?

No.

A maximum detecting or overload concentration may describe a higher exposure level that is outside the normal quantitative measurement range.

Is a nonlinear gas sensor inaccurate?

Not necessarily.

A nonlinear sensor can provide accurate concentration measurements if the calibration model correctly represents its response.

What causes gas sensor nonlinearity?

Possible causes include:

surface saturation
diffusion limitation
reaction kinetics
optical behavior
oxygen availability
electronics saturation

What is nonlinearity error?

Nonlinearity error is the deviation between the actual sensor response and the selected linear reference curve.

How is gas sensor linearity tested?

Expose the sensor to several known concentrations across the intended range, record output, fit the intended calibration model and evaluate deviations from that model.

Is R² a measure of sensor linearity?

R² can indicate how well a regression model explains the overall data, but it does not by itself establish acceptable linearity error at every point.

What happens outside the linear range?

The sensor may continue responding, but the same simple linear conversion may no longer produce the required accuracy.

Does saturation cause nonlinearity?

Yes.

As the sensor approaches saturation, concentration can continue increasing while output changes progressively less.

Can temperature affect linearity?

Yes.

Temperature can change sensitivity, reaction kinetics, diffusion, semiconductor behavior and electronics.

Can humidity affect linearity?

Yes.

The effect depends on sensor technology and target gas.

Is two-point calibration enough?

It can be sufficient when sensor response is reliably linear over the required range.

Two-point calibration alone does not prove linearity.

When should multi-point calibration be used?

Multi-point calibration is useful when:

the range is wide
the response is nonlinear
high accuracy is required
linearity must be verified

Final Takeaway

Gas sensor linearity is not simply another word for sensitivity.

Remember:

Sensitivity
=
how steep the response curve is

Linearity
=
how closely the response follows the chosen straight-line model

Linear Range
=
where that linear relationship remains acceptably valid

Measurement Range
=
where the sensor or instrument is specified for measurement

Maximum / Overload Range
=
a higher exposure region that should not automatically be treated as a normal quantitative measurement range

A highly sensitive sensor can be nonlinear.

A highly linear sensor can still be inaccurately calibrated.

A nonlinear sensor can still be accurate when the correct calibration model is used.

And a sensor that survives a high gas concentration does not necessarily measure that concentration accurately.

That leads to the most useful engineering rule:

Do not choose a gas sensor because the datasheet simply says “linear.” Check the range, linearity tolerance, sensitivity, calibration method, environmental conditions and behavior near saturation.

References and Further Reading

Author

Gas Nose

8618037022596

Gas Sensor Linearity and Linear Range Explained

Linearity, Sensitivity and Range: Quick Comparison

What Is Gas Sensor Linearity?

What Does a Linear Gas Sensor Response Look Like?

Sensitivity vs Linearity

Sensitivity

Linearity

Sensor A

Sensor B

What Is the Linear Range?

Linear Range vs Measurement Range

Linear Range

Measurement Range

Linear Range vs Maximum Detecting Concentration

Why Do Gas Sensors Become Nonlinear?

Diffusion Limitation

Reaction Kinetics

Surface Saturation

Optical Behavior

Oxygen Availability

Electronics Saturation

Electrochemical Gas Sensor Linearity

MOS Gas Sensors Are Often Intrinsically Nonlinear

NDIR Sensors and Linearity

Catalytic Combustible Gas Sensor Linearity

What Is Nonlinearity Error?

Linearity Is Not the Same as Accuracy

Linearity Is Not Repeatability

Linear Range Is Not the Same as LOD

What Happens Near Saturation?

Calibration Curve vs Linear Range

Linear

Polynomial

Logarithmic

Power Law

Piecewise Linear

Lookup Table

Two-Point Calibration: When Is It Enough?

When Is Multi-Point Calibration Better?

Two-Point Calibration vs Multi-Point Linearity Testing

Two-Point Calibration

Multi-Point Linearity Testing

Why R² Does Not Tell the Whole Story

Temperature Can Change Linearity

Humidity Can Change the Calibration Relationship

Drift Can Change the Calibration Curve

Zero Drift

Span Drift

How to Validate Gas Sensor Linearity

Step 1 — Define the Required Range

Step 2 — Select Multiple Reference Concentrations

Step 3 — Stabilize the Sensor

Step 4 — Apply Each Concentration

Step 5 — Plot Concentration vs Output

Step 6 — Fit the Intended Model

Step 7 — Examine Residuals

Step 8 — Compare With the Required Tolerance

Step 9 — Verify With Independent Points

How to Read Gas Sensor Linearity in a Datasheet

1. Sensitivity

2. Measurement Range

3. Linear Range

4. Linearity / Nonlinearity Tolerance

5. Reference Method

6. Maximum Detecting / Exposure Concentration

7. Temperature

8. Humidity

9. Calibration Method

10. Typical vs Guaranteed Values

A Practical Datasheet Example

0–80 ppm

80–100 ppm

100–500 ppm

Linearity and Response Time Are Different

Linearity and Hysteresis Are Different

Common Gas Sensor Linearity Mistakes

Mistake 1 — Sensitivity = Linearity

Mistake 2 — Measurement Range = Linear Range

Mistake 3 — Maximum Exposure = Measurement Range