White Paper · 2026
Preroll emissions exceed the EPA hazard threshold. Two ECDS formats do not.
This white paper compares emissions from two Electronic Cannabis Delivery System (ECDS) formats, a high-potency pod (labeled at 83% THC) and a high-potency cartridge (labeled at 86% THC), against combustion smoke from a representative un-infused cannabis pre-roll (labeled at 23% THC). The comparison is conducted on a per-puff basis, scaled to a daily exposure estimate, and benchmarked against EPA RfC IRIS Chronic values.
0
Pre-Roll · combustion
screening Hazard Index
0.00
Pod ECDS
screening Hazard Index
0.00
Cart ECDS
screening Hazard Index
Hazard Index of 1.0 is the conventional screening-level threshold.
Director, California NORML
"The dangers of vapes have been misrepresented by anti-smoking advocates to suggest their emissions are as hazardous as smoke. They are not." ... "Relative THC potency can't be directly translated into dosage differences." ... "Hopefully, the government's recent rescheduling decision will end obsolete restrictions that have hindered researchers from studying cannabis vape pens that are readily available to millions of U.S. consumers."
Executive Summary
Across nearly every measured analyte for which a benchmark exists, the two ECDS formats produce substantially lower per-puff and per-day emissions than the pre-roll combustion arm. The pre-roll Hazard Index (HI) is approximately 166, driven almost entirely (>99%) by acrolein. By contrast, the Pod ECDS HI is approximately 0.47 and the Cart ECDS HI is approximately 0.82, both below 1, the conventional screening-level threshold.
The pre-roll Hazard Index is approximately 166, driven almost entirely (>99%) by acrolein, which exceeds its EPA chronic comparator by roughly 165-fold. Both ECDS formats remain below the screening threshold of 1.
Formaldehyde remains the dominant residual carbonyl in both ECDS formats. Heavy metals (particularly nickel and chromium) are detected in both ECDS arms, with nickel in the Cart arm (21.17 ng/puff) markedly higher than in the pre-roll (0.47 ng/puff), pointing to a device-materials source. The cannabinoid HPLC dataset reveals a notable Δ9-THC to Δ8-THC isomerization in both ECDS arms (~28% of the emission cannabinoid pool).
Actionable Insight
Eliminate NiChrome heating elements
Read the full executive summary and methodology in the white paper.
Download PDFHypothesis Disposition
Nine hypotheses tested against the measured data
Evidence labels follow the convention specified in the analytical brief. Five hypotheses are directly supported, one is directionally supported, and three could not be tested from the available dataset.
H1 · PRIMARY
ECDS toxicity-weighted emissions index lower than combustion, normalized to delivered cannabinoid dose
Directionally SupportedPer-puff and per-day data show large reductions across nearly all measured toxicants for both ECDS arms vs. pre-roll. A true dose-normalized TWEI is not computable from the provided dataset.
H2 · SECONDARY
PAHs substantially lower in ECDS than combustion
Not TestedPAHs (e.g., benzo[a]pyrene) are not present in the dataset. Measurement gap identified for follow-up.
H3 · SECONDARY
Benzene and toluene lower in ECDS than combustion
SupportedBenzene is ND/<LOQ in both ECDS arms vs. 7,856 ng/puff in pre-roll. Toluene reduced ~93-99%.
H4 · SECONDARY
Carbonyls lower in ECDS than combustion under optimized conditions
Supported with NuanceAcrolein, crotonaldehyde, acetone are ND in ECDS. Formaldehyde reduced ~50% but remains the largest residual carbonyl.
H5 · SECONDARY
Isoprene higher in combustion than ECDS
SupportedPre-roll 15,517 ng/puff vs. Pod 1,313 and Cart 4,944. Cart > Pod suggests temperature modulation.
H6 · SECONDARY
Aerosolization alters cannabinoid ratios via thermal decarboxylation and isomerization
SupportedPre-roll shows THCA-A → Δ9-THC decarboxylation. Both ECDS arms show ~28% Δ9-THC → Δ8-THC isomerization.
H7 · SECONDARY
Heavy metals (Ni, Cr) detectable in ECDS aerosols from device materials
SupportedNi in Cart arm (21.17 ng/puff) is ~45× higher than pre-roll (0.47 ng/puff). Consistent with device-materials source.
H8 · BOUNDARY
Abuse-condition spike in benzene and carbonyls in ECDS
Not TestedWorkbook does not include an abusive-operation arm. Remains a priority for follow-up study design.
H9 · BOUNDARY
Abuse-condition spike in metals in ECDS
Not TestedNo abusive-operation arm; however, elevated baseline Ni in Cart arm under nominal conditions warrants follow-up.
Founder & Principal Consultant, MayThe5th
"The biggest issue is that ECDS are still regulated as a distinct category, often under standards that rely on food science, or political optics rather than inhalation-specific toxicology. Most of all, ECDS are very often confounded with ENDS. Our comparative study is urgently needed because the public debate on 'vaping vs. smoking cannabis' is often driven by anecdotes and incomplete datasets. By introducing the Hazard Index, we are providing a scientifically grounded, transparent way to compare total emissions burden, which is essential for regulators to require emissions-based testing and set consensus standards. Our hope is that this preliminary research will spark more emissions research funding given the current political environment around cannabis."Download Original Research Brief & Hypothesis
Cross-Arm Comparison
Per-day inhalation exposure vs. EPA RfC IRIS Chronic comparator
ng/day · log scale · 22-puff/day intake assumption · hover bars for detail
Acrolein in Pre-Roll
Exceeds its EPA chronic comparator by ~165×. Drives >99% of the pre-roll Hazard Index.
Nickel in Cart ECDS
Approaches its comparator (MOE ≈ 3). Device-materials source, not combustion-derived.
Benzene in ECDS
Non-detect in both ECDS arms vs. 172,832 ng/day in the pre-roll combustion arm.
Cannabinoid Composition
Source material vs. collected emissions
Percent of total cannabinoids · HPLC · hover bars for detail
Key Finding
~28% Δ8-THC isomerization in ECDS emissions
Δ8-THC is non-detect in both source oils but appears at ~28% of emission cannabinoids in both Pod and Cart arms. This Δ9-THC → Δ8-THC thermal isomerization has implications for product labeling and dose-claim accuracy.
Pre-Roll Decarboxylation
THCA-A drops from ~90% of source to ~2% of emissions. Δ9-THC rises from ~6% to ~90%. This is the canonical combustion-driven thermal decarboxylation signature.
Get the complete white paper
Download the complete white paper with detailed methodology and all supporting data.
Download White Paper (PDF)Raw Data
Measured emissions by product arm
Complete per-puff and per-day emissions data for all 16 measured analytes across three product arms. Values are group means across replicates as reported in the dataset.
Assumptions
Daily exposure uses a fixed 22 puffs/day (Cannabis Puffing Regimes). The three product arms (Average Pre-Roll at 23% THC, Average Pod 100% at 83% THC, and Average Cart 100% at 86% THC) are group means across replicates. ND and <LOQ values are treated as zero in downstream HQ/MOE calculations.
CEO, NN Analytics
"It has been a pleasure to explore the toxicant exposure of cannabis vaporizer products compared to their combustible counterparts with a consumer organization like CANORML. While NN Analytics' prior investigations on aerosol toxicology and smoke exposure have been heavily focused on tobacco products, we are thankful to be able to use the FDA's recommended list of Harmful or Potentially Harmful Constituents (HPHCs) to address exposure to cannabis smoke and aerosol. NN Analytics is committed to working with toxicological experts and well-established industry partners to establish which toxicants are important to be monitored, as well as to provide industry guidance as to the acceptable threshold limit values (TLVs) of each toxicant by inhalation. We encourage the application of comparator exposure limits, such as the US EPA RFC and (CAL)OSHA databases, for use in the establishment of such limits, for both combusted and aerosolized cannabis products. This hazard index is mission critical to focused work in the comparison of ECDS to combustible products. NN Analytics is committed to broadening the body of work and scope of future needed research regarding these toxicant exposure levels."Download NN Analytics Lab Report
Comparative Toxicity of Emissions (ECDS) vs. Combustion of Un-infused Cannabis Pre-rolls
Methods Overview
Screening-level hazard characterization
Sample
Three product arms: un-infused cannabis pre-roll (23% THC), high-potency pod (83% THC), and high-potency cartridge (86% THC).
Collect
Fixed square-wave puff regimen at 1,800 mL/min. 22-puff/day intake assumption (Cannabis Puffing Regimes). ISO 3308 termination for pre-rolls.
Quantify
Carbonyls (EPA 8315A), volatiles (USP 467 / EPA 624), cannabinoids (HPLC), heavy metals (ICP-MS). Benchmarked against EPA RfC IRIS Chronic values.
Full methodology, analytical framework, and normalization approach detailed in the white paper.
Download PDFLimitations
What this analysis does not cover
TWEI not formally computed
Cannabinoid data is relative composition, not absolute mass yields per puff.
PAHs not measured (H2)
The single most consequential remaining measurement gap.
Abuse-condition arms not tested
No experimental arm under intentionally abusive conditions.
Fixed 22-puff/day assumption
Real-world consumption varies materially across users.
ND / <LOQ treated as zero
Conservative-toward-zero choice in HQ calculations.
Six analytes lack EPA comparators
Reported descriptively; omission may understate total hazard.
Conclusions
Screening-level support for the central thesis
Both ECDS formats produce substantially lower screening-level hazard than the pre-roll combustion arm. The Hazard Index is below the conventional screening threshold of 1 for both ECDS formats while exceeding it by approximately two orders of magnitude for the pre-roll. Acrolein is the dominant single contributor to the pre-roll hazard; formaldehyde and device-sourced nickel and chromium dominate the ECDS hazard indices.
The Δ9-THC to Δ8-THC isomerization finding (~28% of the emission cannabinoid pool in both ECDS arms) has implications for product labeling and dose-claim accuracy beyond the toxicant-burden framing.
Priority Follow-Up Studies
PAH measurement (EPA Method 610)
Absolute per-puff cannabinoid mass yields
Abuse-condition experimental arm
Δ9-THC → Δ8-THC isomerization in ECDS?
Full discussion, hypothesis-by-hypothesis interpretation, and next steps in the white paper.
Download PDF© 2026 MayThe5th. All Rights Reserved.