Explaining VOCs, TVOC and the VOC Index

Of all the data points on an air quality monitor, Volatile Organic Compounds (VOCs) are probably the hardest to interpret. While CO2 and particulate matter (PM) have clear safety thresholds, VOCs are far more complex due to the limitations of MOx sensor technology and the nature of the compounds themselves.

If you have ever wondered why your TVOC numbers fluctuate or what they truly indicate, this article is for you. We will break down (or at least try!) what VOCs are, what tVOC is, the science behind the sensors, and how to interpret the TVOC index on your AirGradient monitor. We will also explain how the Learning Time Offset Duration (LTOD) setting in the AirGradient Dashboard may help you keep your data stable.

VOC vs TVOC: Because One Little ‘T’ Can Confuse Everyone

This distinction is almost always skipped, but it’s fundamental to understanding what low-cost sensors are actually measuring.

VOCs (Volatile Organic Compounds)

The term VOCs refers to a broad range of substances that can easily move through the air.

• Volatile means that these substances evaporate readily at ambient or room temperature.

• Organic is a chemical classification indicating that the substances are carbon-based compounds, typically containing carbon–hydrogen (C–H) bonds. Such compounds are commonly (but not only) found in living organisms, including plants and animals. Hence, the name “organic”.

• Compounds simply means that VOCs are individual chemical substances that make up a small part of the air’s overall composition.

They can come from everyday sources like cleaning products, paints, glues, solvents, new furniture, fragrances, candles, cooking, human metabolic emissions, and outdoor infiltration (traffic, industry, fuels, woodsmoke, etc.). Many VOCs also occur naturally: plants emit VOCs, and humans themselves exhale trace amounts of VOCs as part of normal metabolism.

The U.S. EPA lists 189 hazardous air pollutants, many of which are VOCs. But VOCs are not inherently dangerous as a category; “VOC” simply describes a physical behaviour (volatility), not their health impact. Some VOCs pose clear health risks (formaldehyde, benzene). Others are harmless (ethanol from cleaning wipes).

Therefore, VOCs is a balloon term for very harmful but also completely harmless gases.

Low-cost VOC sensors cannot measure how toxic the VOCs they detect are. They only detect that one or more of these substances are present.

If you are looking for a comprehensive technical overview, you can explore the EPA’s Guide to VOCs or the Berkeley Lab’s resources on the subject.

Because there are so many different types of VOCs, identifying each one individually requires massive, expensive laboratory equipment, such as GC-MS (Gas Chromatography-Mass Spectrometry), which simply isn’t practical for home use. To make monitoring accessible, consumer air quality monitors use low cost TVOC sensors instead.

TVOC (Total Volatile Organic Compounds)

As opposed to VOC, which refers to the individual chemicals themselves (e.g. ethanol is a VOC), the term TVOC represents a combined ‘total’ of multiple airborne VOCs present simultaneously.

To better understand what is included in this measurement, consider the classification of VOCs based on their boiling points:

Importantly, a TVOC metric does not tell you which specific VOCs are present or their toxicity. It only tells you whether the total VOC presence is rising or falling. Rather than identifying specific chemicals, which, as mentioned above, requires expensive, laboratory-grade instruments, low-cost VOC sensors provide this aggregated “Total” score to help you track general changes in your environment.

How Low-Cost VOC Sensors Actually Work (MOx Technology)

Most consumer-grade monitors use Metal-Oxide (MOx) sensors for TVOC detection. To understand their data, it is crucial to understand how they function.

A MOx sensor consists of a heated metal oxide surface that measures changes in electrical resistance as gas molecules interact with it. The sensor reacts based on the oxygen content on its surface:

• Reducing gases (most VOCs) consume surface oxygen and decrease resistance.

• Oxidizing gases (like NOx) add oxygen and increase resistance.

Because the sensor simply measures this change in resistance, it sees “activity,” not specific compounds. Some VOCs like ethanol produce a strong response , while other VOCs like the harmful formaldehyde may produce weaker signals.

Because MOx sensors are broadband-reactive:

• They detect patterns and spikes

• They do not quantify VOCs

• They do not distinguish harmful from harmless VOCs

• They are best interpreted qualitatively

Why Measuring Absolute VOC Concentrations Using Low-Cost MOx sensors is Fundamentally Difficult

While this technology allows for affordable monitoring, it introduces four unavoidable limitations:

1. Lack of Specificity: The sensor reacts to all detected gases at once. Seeing a high reading tells you nothing about the composition, source, or toxicity of the gas. It cannot distinguish between harmless cooking scents and harmful chemical off-gassing.

2. Cross-Sensitivity: In the lab, these sensors are usually calibrated to a specific gas, such as ethanol. In the real world, they encounter unpredictable mixtures of acetone, isopropanol, and other chemicals. This creates signal discrepancies; for example, ethanol often produces a much stronger signal than formaldehyde, even if the formaldehyde is more dangerous.

3. Environmental Sensitivity: Factors like temperature, humidity, and airflow can influence the oxidation reactions on the MOx surface, affecting the reading.

4. Sensor Drift: Sensor baselines shift over time, and unlike lab equipment, low-cost sensors cannot be easily recalibrated with reference gases.

Because of these physical realities, any absolute TVOC concentration output (e.g., “500 ppb” or “0.5 mg/m³”) is not scientifically reliable from a low-cost MOx sensor. This is why major manufacturers like Sensirion and Bosch have largely abandoned absolute TVOC outputs as a default in their newest sensors, including the Sensirion SGP41 used in AirGradient devices. Instead, they have replaced them with a relative VOC Index.

Some consumer air quality monitors, such as those from Airthings, display VOC values in parts per billion (ppb). However, it is important to understand that these values are not directly measured absolute concentrations, as other low cost monitors in this price range rely on the same low-cost broadband sensitive VOC sensors, which cannot reliably determine absolute TVOC concentrations.

As a result, there is no robust way to convert their raw signals into true concentration values.
In practice, what happens is that a fundamentally relative sensor signal or index is converted into a number labelled as “ppb.”or using a formula compliant with a specific green building standard. While this may appear more intuitive or precise, it does not reflect the actual concentration of VOCs in the air and can vary widely depending on the chemical mixture present.

Sensirion themselves explicitly state that the SGP4x MOx sensors are not capable of providing absolute VOC concentrations under real-world conditions. Outside of controlled laboratory environments with known gas mixtures, the sensor output is inherently semi-quantitative and should not be interpreted as an absolute VOC concentration.

We examined this limitation in detail in our analysis of the Sensirion SGP41 sensor, where we showed that while the sensor tracks changes in VOC levels well, it cannot reliably measure their absolute magnitude. Converting such signals into ppb does not overcome this limitation, it simply hides it behind a familiar unit: How Accurate is the Sensirion SGP41 TVOC Sensor?

For this reason, absolute VOC values reported by low-cost monitors, regardless of brand, should be treated with caution. This is why AirGradient uses a VOC Index rather than displaying absolute concentration values, focusing on what this sensor technology can do reliably: track relative changes over time.

How the VOC Index Works

The sensors’ raw concentration output is processed using Sensirion’s Gas Index Algorithm. This results in an index-based measurement that focuses on relative changes, rather than absolute values.

The VOC Index describes the current status relative to the sensor’s recent history. To understand this approach, it helps to compare it to the human nose.

The “Nose” Analogy

When you walk into a room from outside, your nose uses the fresh outside air as a baseline (an offset). It immediately tells you if the room smells “stronger” or “cleaner” than that baseline. However, after you stay in the room, your nose adapts to the new smell, and it becomes your new “normal.”

The VOC Index performs a similar calculation. It uses a moving average of the past 24 hours, called the “learning time”, as its offset.

• 100 (Baseline): Represents the average VOC background of the room over the last 24 hours.
• < 100: Indicates the air is cleaner than the recent average.
• > 100: Indicates a rise in VOCs compared to the average.

How to Interpret the TVOC Index

But because the VOC Index is relative, not absolute, a high VOC environment can eventually become the baseline. To explain this in more detail:

What will happen if VOC levels stay high for the past 12 hours?

The sensor will:

• adapt its baseline upward

• treat those high VOC levels as the new “normal”

• drift back toward 100, even though actual VOC concentrations remain high

For example, if you paint a shelf in your living room without proper ventilation, the strong paint smell may linger for days. However, the TVOC sensor will mainly respond at the start of the painting activity and then slowly return to its baseline value (100), even though the actual TVOC concentration in the room remains high.

This is why the sensor learning time is important.

If the learning window is too short (e.g., 12 hours), the monitor may normalize high VOC loads.

Most air-quality monitors lock this learning time to the default (12 hours) and do not expose it to users. This significantly limits the usefulness of the VOC Index in spaces with persistent emissions.

The AirGradient Dashboard exposes this parameter through the Learning Time Offset Duration (LTOD):

Dashboard → General Settings → Calibration → VOC/NOx Index Learning Time Offset Duration

You can choose values between the default 12 hours and the maximum 720 hours.

In practice, the choice is simple:

• Shorter learning times adapt quickly and are useful when a device is moved between rooms or environments.

• Longer learning times keep the baseline stable, ensuring that sustained VOC problems remain visible instead of being normalized away.

This control allows the VOC Index to be tuned to the application, rather than forcing a one-size-fits-all interpretation.

Conclusion

VOCs and TVOC measurements are often the most confusing data points on an air quality monitor, and many of the questions we receive in support come down to the same core issues: What does TVOC actually measure? Why do the numbers change? Why don’t they behave like CO2 or PM? And why don’t we show “ppb” values like Airthings?

We hope this article clarifies things and makes it easier to use TVOC data in a meaningful and realistic way. If you still have questions about, please get in touch with us via our support page. We’re always happy to help clarify how the data from your monitor should be interpreted.