GEET Reactor Evidence Review
What is documented, what is suggestive, and what remains unverified.
Verdict: GEET is a documented fuel-pretreater design with plausible conventional heat-exchange and vaporization effects. Published testing is sparse and mixed. It does not establish a unique plasma mechanism, mostly-water fuel, zero emissions, or extraordinary efficiency.
How this review ranks evidence
Evidence is ranked by what it can support. A patent proves that claims were filed and granted; it is not a performance certificate. A community build proves a device can be assembled and may run; without a baseline, fuel-mass balance, load measurement, and uncertainty budget, it does not quantify efficiency. A controlled study can estimate performance only for its tested configuration.
HIGHER WEIGHT
Controlled measurements
Stated load, fuel use, baseline, methods, and limitations.
CONTEXT
Patent and technical records
Useful for design and claim history, not independent validation.
LOWER WEIGHT
Testimonials and demonstrations
Useful leads, but vulnerable to boundary and measurement errors.
The patent: documented apparatus, inventor assertions
US Patent 5,794,601, issued in 1998, describes an exhaust-heated volatilization chamber, a reactor tube with a coaxial rod, counter-current flow, and delivery of the treated vapor to an engine intake. The patent also records Pantone's assertions about alternate fuels and cleaner exhaust.
The patent does not publish a controlled fuel-consumption dataset, uncertainty analysis, chemical composition of the treated gas, or independent replication. Its grant therefore documents the apparatus and legal claims—not the magnitude or mechanism of performance.
The strongest direct publication: mixed and inconclusive
The 2018 paper “Evaluation of a Retrieved Pyrolithic Reactor to Be Used in Small Farms” tested a small generator with a GEET-style pyrolytic reactor under specified electrical loads. It is the clearest directly relevant peer-reviewed source in this archive.
What the underlying thesis reports
- With common automotive gasoline and water, the GEET configuration was not efficient and fuel consumption increased.
- With aviation gasoline, one 25/75 gasoline-water condition reported average efficiency of 8.46%, versus 6.44% with aviation gasoline alone in that setup.
- The authors described high and low results and said more work was required for correct evaluation.
- The thesis explicitly called its results inconclusive.
Primary thesis record: UNIOESTE repository PDF.
Calling this paper merely “promising” omits the negative common-gasoline result and the authors' uncertainty. It supports further controlled investigation, not a general claim that GEET works.
French thesis and builder literature
French builder and thesis archives are historically important. Christophe Martz's 2001 ENSAIS work is frequently cited for large reductions in some measured pollutants and a small fuel-use change on a non-optimized setup. It is a student thesis hosted by a pro-GEET community site, not a replicated journal result, and its measurements should be reported with that limitation.
French “Gillier-Pantone” water-doping systems also differ from Pantone's full fuel-pretreater arrangement. Reports from farmers, municipalities, forums, and television document adoption and claimed outcomes, but community scale does not substitute for standardized before/after testing.
Important distinction: a percentage of water in a bubbler or feed mixture is not the percentage of engine energy supplied by water. Water is not a combustible energy source; fuel vaporization and preferential consumption must be measured separately.
Established neighboring technologies
GEET discussions often point to real research on exhaust-heat recovery, catalytic reforming, and water injection. These fields make parts of the design worth studying, but they do not validate the GEET apparatus by analogy.
- Thermochemical recuperation: a 2023 ACS paper modeled and tested structured, catalyst-coated reformer heat exchangers with selected fuels and controlled temperatures. It shows that waste-heat fuel reforming is legitimate engineering—and that catalyst, kinetics, heat-transfer area, fuel, and operating conditions are essential.
- Water injection: a 2019 review describes knock, combustion-temperature, efficiency, and emissions effects that vary by implementation. Water can change combustion; it does not become net fuel.
Sources: ACS Engineering Auand Energy Conversion and Management.
Evidence matrix
| Claim | Current support | What would strengthen it |
|---|---|---|
| Exhaust heat vaporizes or preheats incoming fuel | Mechanically plausible and consistent with the patent | Temperature map, flow rate, gas composition, and baseline comparison |
| Some configurations change fuel use or emissions | Mixed limited studies and reports | Preregistered load cycles, calibrated fuel mass, full exhaust speciation, replication |
| A unique plasma forms in the passive reactor | Not demonstrated | Direct plasma diagnostics, spectroscopy, field measurements, reproducible conditions |
| Engine derives most energy from water | Not demonstrated and inconsistent with ordinary energy accounting | Complete fuel/water mass balance, calorimetry, gas analysis, independent lab replication |
| Near-zero or “cleaner than air” emissions | Unsupported as a general claim | Certified multi-species testing across load, including CO₂, NOₓ, CO, HC, particulates |
| Several-hundred-percent efficiency improvement | Unverified | Complete energy boundary, uncertainty budget, long-duration independent replication |
A useful test standard
- Define the exact GEET or Gillier-Pantone configuration and freeze it before testing.
- Run matched baseline and modified trials at the same measured shaft or electrical load.
- Measure fuel by mass, water separately, and every external or stored-energy input.
- Measure exhaust temperature, flow, and a full pollutant set—not a single selected gas.
- Publish calibration, raw time-series data, exclusions, and uncertainty.
- Have an independent laboratory rebuild or instrument the setup.