THC Boiling Point 157 °C? Pressure, Vaporization and Substance Data

The existing THC correction is retained

The frequently cited 155–157 °C is not a standard boiling point of THC at ambient pressure. The work by Eyal and colleagues already used on this page assigns approximately 155 °C to a pressure of 0.05 Torr. A value of 425 °C given there for standard pressure is extrapolated; it is not a recommendation for a vaporizer setting. Evaporation is also possible below the boiling point.

Existing supplementary literature outside the pool of twelve: Eyal AM et al. (2023), Vapor Pressure, Vaping, and Corrections to Misconceptions Related to Medical Cannabis’ Active Pharmaceutical Ingredients’ Physical Properties and Compositions, DOI 10.1089/can.2021.0173, in particular Table 3 and the discussion of vapour pressure. This already existing substance-data correction is retained and listed separately. It does not silently expand the study pool of the planned video work.

Evaporation, boiling and decarboxylation

During evaporation, molecules leave a liquid or solid phase and enter the gas phase. This process can occur below a boiling point. Boiling describes a particular state that depends on external pressure. For this reason alone, a reliable boiling point requires the specification of pressure.

Cannabis flower is moreover not a vessel containing a pure substance. It contains a plant matrix and various compounds. Heat must first reach the relevant sites; released substances must then be transported out of the material and through the device pathway.

Decarboxylation, by contrast, is a chemical transformation. In cannabinoid acids, carbon dioxide is cleaved off in this process. Demonstrating extensive conversion does not yet establish that the corresponding neutral compound has arrived in full at the mouthpiece. Lanz investigates conversion and recovery separately. [Lanz2016]

A table that consolidates all these processes into a single column labelled “temperature of effect” loses the critical distinctions. A substance value does not imply a particular mood or a medical effect at precisely that setting.

Five temperature concepts in the device

The concepts are not five interchangeable names for the same number. A sensor can measure its own temperature very precisely and still not be representative of the entire fill. Accuracy at one measurement location and representativeness for the material are different properties.

A concrete example from Gieringer 2004

Gieringer and colleagues placed two thermocouples above and below the sample. At the maximum device setting examined, 155 °C was measured at the upper sample surface and 218 °C at the screen close to the heater. These values belong to this historical set-up and these measurement locations. [Gieringer2004]

The example is not a general error range for today’s vaporizers. It demonstrates instead that within one set-up, different temperatures can be simultaneously relevant. The difference must not be adopted as a blanket display deviation for all devices.

Even an average of both figures would not resolve the problem. It would not be a directly measured temperature of every particle of material. A complete description would require the spatial and temporal progression, not merely a single summarised number.

Why a draw changes the conditions

During a draw, air moves through the set-up and transports heat. The control system can respond to this, but not every system responds equally quickly. Incoming air, heat storage and heating power interact with one another.

Between draws, the state changes again. Hot walls can continue transferring heat whilst other parts cool down. A measurement at idle can therefore give a different picture from a measurement under airflow.

For a comparison it must be established when the value was recorded: directly after the ready signal, during the first draw, after several draws, or after a pause. “Temperature accuracy” without this context is incomplete.

The puffing-machine work by Carrara defines such procedures for the devices examined. This makes results reproducible, but not automatically universal. [Carrara2020]

Resolution, accuracy and repeatability

A display with one-degree increments has a certain setting resolution. It does not prove measurement accuracy of plus/minus one degree within the plant material. These properties are frequently conflated in product texts.

Accuracy concerns proximity to a suitable reference value. Repeatability describes how similar results are when the procedure is repeated under identical conditions. A device can repeatedly deliver a similar progression whilst systematically deviating from another measurement location.

Reliable accuracy specifications require calibrated reference measurements, defined measurement locations and an uncertainty assessment. The earlier blanket claim that most devices deviate by a fixed number of degrees and that premium devices are inherently more accurate is not upheld on the basis of this pool.

Likewise, no supposed own tests on hundreds of devices are claimed. Product data and independent measurements must remain distinguishable as such. A finely resolved user interface is a feature, not a substitute for a test report.

What Lanz and Carrara actually make comparable

Lanz examines cannabinoids in defined sample and device fractions. This makes it possible to ask where the substances were recovered under the respective protocol. The work does not provide a universal temperature menu for all flower types and device categories. [Lanz2016]

Carrara describes a standardised draw regime and the temporal progression of delivery for two devices. In this context, the setting is part of the protocol, not the sole explanation for all differences. The collection method is also examined, because it influences the measured result. [Carrara2020]

A comparison of studies must therefore examine more than their stated Celsius values. Different collectors, loads, air volumes and time windows can alter the results. Even identical device settings are only one component of comparability.

How temperature tables are read meaningfully

A rigorous substance table states the compound, pressure and type of data. A directly measured value must be interpreted differently from an extrapolation. Decomposition and measurement conditions must be taken into account. Missing information should remain visibly absent.

A device table, by contrast, states model, version and control range. If a material temperature was measured, the sensor position and experimental procedure belong with it. Manufacturer range and independently measured range are separate pieces of information.

An efficacy table would be an entirely different matter. It would require data on administered dose, population, and clinical endpoint. A cannabinoid boiling-point table does not meet these requirements. It therefore cannot form the basis of a treatment recommendation for sleep, pain, or any other condition.

Terpenes: No single switch per aromatic compound

With terpenes too, a temperature figure should not be read as an exclusive release point. A substance can be delivered across a range; the composition changes over the course of use. For a real plant material, what matters is the combination of substance properties and transport conditions.

An aroma cannot, moreover, be directly translated into a clinical effect. The mere presence of a compound in the aerosol does not demonstrate its effective concentration at the target site, nor a therapeutic interaction with THC.

The article Vapour Quality explains what chromatograms and receptor-binding tests actually show. A scientific temperature explanation should not use personalised effect animations or invented terpene curves.

Material preparation belongs in the protocol

Grind size, particle distribution, fill mass, and moisture content all influence how the sample is set up. They should therefore be documented whenever two device settings are compared. No universal moisture optimum for all models can be derived from the available pool of data.

The relative humidity of a storage container and the water content of a flower are, furthermore, two different quantities. A percentage figure for the storage environment must not be presented as a directly measured material water content. Confusions of this kind make apparently precise recommendations unreliable.

The device manual specifies what preparation is intended. For one’s own technical comparison, as few factors as possible should be changed simultaneously. The result remains an observation of the chosen setup for as long as no analytical delivery measurement is available.

Why temperature stepping is not a pure temperature test

When the same fill is used successively at increasing settings, not only the temperature changes. Time, prior extraction, and the initial composition of the remaining material also change. A later draw therefore does not take place on the same unaltered sample.

This can be useful as a documented usage sequence, but it does not isolate a pure temperature effect. For a controlled comparison of different settings, comparable fresh loads and an appropriate order of testing would be necessary. The temporal delivery would also need to be taken into account.

A stepping protocol must therefore not be presented as a reliable method for selectively targeting individual active substances or clinical effects. It describes a procedure, not a universal means of dose control.

Residual material and visible signs

The colour of already heated flower can make the process roughly visible. It is not, however, a calibrated determination of the remaining quantity of THC or CBD. Different initial colours, lighting conditions, and uneven heating make interpretation additionally difficult.

Diminishing clouds are likewise not an analytical proof of zero residue. The question of how much of a particular substance remains must be answered by appropriate analysis. A colour-to-milligram table would feign accuracy without validation.

Obvious signs of malfunction or charring are a reason to check the setup in accordance with the manual. They do not, however, retrospectively define a universal safe range for all prior settings.

What a temperature measurement protocol should contain

A comprehensible report begins with the measurement objective. Is the aim to examine the internal control system, the temperature of the incoming air, or the spatial temperature distribution within the material? Only then is a suitable sensor selected and positioned. An arbitrary measurement point does not answer every one of these questions.

The sensor description should state the type, dimensions, calibration, and temporal resolution. A sensor itself requires time to warm up. During rapid transients it may therefore reproduce changes with a delay. Its position can also influence the local airflow or the contact within the material.

The boundary conditions include the environment, initial state, load, and airflow. For a warm repeat run, the state prior to the start must be described. Otherwise, a difference between two curves may simply result from differing residual heat.

The evaluation should not show only a peak value. For many questions, the progression, spatial differences, and repeatability are important. A brief peak, a more sustained plateau, and a mean value over an entire draw are different descriptions.

Finally, an uncertainty statement is required. It can arise from sensor error, placement, repetitions, and sample variation. More decimal places do not automatically make a measurement more precise. A report that prints a figure more precisely than its setup can measure creates spurious accuracy.

Why a schematic curve is not a measurement result

A diagram can illustrate that the set-point, air, and material respond differently over time. As long as no data underlie it, it must not carry apparently empirical degree values, time constants, or error bars. It would need to be explicitly labelled as a schematic.

A mathematical model must also be distinguished from a measured curve. A model can summarise an observed progression and examine relationships. Its parameters are fitted to assumptions and data; they are not automatically new independent measurements.

For the scientific communication planned here, no temperature curves are therefore drawn from an intuitive gut feeling. Where quantitative figures are used, the original data, axes, units, and underlying conditions must be verifiable. The measurement-location table in this article deliberately makes do without invented curve data.

Addendum from the terpene audit

The device and method sources cited above were retained during the consolidation of the current version of the article. The following sections supplement the separate terpene audit with targeted source verification. Neither selection of literature constitutes a systematic comprehensive search.

What terpene data does NIST list?

The following details were cross-checked on 10 September 2026 in the NIST Chemistry WebBook, SRD 69. The linked datasets belong to the respective CAS identity cited. They are substance data, not clinical efficacy studies. The table adopts uncertainties only where the reference reports them.

Precisely the empty fields prevent a table that appears complete but is methodologically inconsistent. Enthalpy of vaporization, vapour-pressure parameters, and boiling point are different quantities. A value in kJ/mol must not become a temperature in °C. Nor should contradictory caryophyllene figures be “resolved” by selecting the most familiar number.

Why does a temperature choice not isolate individual cannabinoids?

The chamber does not contain a series of cleanly separated pure substances. Components sit within a plant matrix that changes as it is heated. The mixture has local temperature differences, and substances already released are carried further by the airstream. A setting can therefore influence the composition without reliably isolating a single molecule.

A claim such as “At this temperature only CBD is delivered” would require appropriate aerosol analysis of the conditions in question. Even such an analysis would initially apply to the conditions studied. It would not yet be a reliable statement about individual uptake or effect.

For the same reason, the terpene articles contain no mapping of temperature to alertness, sleep, or anxiety relief. A receptor hypothesis does not provide a temperature plan. A substance may influence a signalling pathway in the laboratory without the desired clinical effect having been demonstrated during use of a flower.

Why the temperature experiments do not yield a safety table

Pomahacova compared, amongst other conditions, 170, 200 and 230 °C using the then-current Volcano Digital. The measured cannabinoid proportions and yields relate to that collection protocol. The residual condensate mass determined by calculation was not fully characterised chemically or toxicologically. No general safe temperature limit can be derived from this. [Pomahacova2009]

Is there a threshold below which no harmful substances are produced?

The library does not define any such general threshold. A claim such as “below 230 degrees, no pyrolysis products” would require a comprehensive analysis covering all relevant materials, devices, and draw conditions. The early laboratory comparison by Gieringer and the device evaluation by Lanz do not justify this generalisation. Lanz et al. (2016) Gieringer et al. (2004)

For a robust emissions assessment, the substances examined, detection limits, and conditions must be made visible. “Not measured” and “not present” are different statements. Temperature ranges should therefore not be promoted as health safety zones without appropriate evidence.

Frequently asked questions about boiling points

Source status and correction notice

Revised on 10 September 2026. Physical references: the individually linked NIST datasets. Eyal et al. is a supplementary source for the existing temperature article and does not belong to the 24-publication register of the supplied terpene package. The device comparison by Lanz is explained in the linked methods section.

Removed were effect menus, unsubstantiated blanket device deviations, and the conflation of resolution with accuracy. The existing temperature redirects remain references to this article and are not counted as additional scientific articles.

Studies, sources and limits of claims

The study cards describe the respective experimental design and its limitations.

Source review: 10 September 2026. In the targeted pool of twelve primary studies, eleven were checked against relevant full-text passages. For Van Dam 2010, original sections accessible beyond the abstract were also checked; a complete PDF and table review remains outstanding. Further literature cited in the article belongs to separate source reviews. No systematic comprehensive search.