The Decision That Gave SyngaPure 99.98% Tar Conversion — And Why Nobody Else Makes It

SyngaPure is the integrated three-stage purification system that gives the X-150 its 99.98% tar conversion. It was built around one architectural decision: control all three stages from a single AI layer instead of running them independently. The choice sounds obvious in retrospect. It is not how the industry works. Almost every commercial gas cleanup system built in the last 30 years uses sequential independent stages, each chasing its own KPI. Zero-X chose the harder path. The result is consistent syngas purity across feedstock variability that would break a sequential system. For an investor, that means a machine that runs 80 days straight on difficult waste without the purification bottleneck that kills competing systems.

Why Now

The bottleneck in small-scale gasification has never been the gasifier. It has been the gas cleanup.

A gasifier can turn waste into syngas reliably. The problem is what comes next. Raw syngas contains tar, hydrogen sulfide, ammonia, and particulates. Run it into an engine without cleaning it first and the engine fouls, the lubricating oil degrades, and the maintenance intervals collapse. Every gas-to-energy project that failed in the field failed not because the gasifier stopped producing gas but because the cleaning system could not keep up with the variability.

This is the problem that has kept distributed gasification stuck at demonstration scale for two decades. A gasifier running on clean wood chips at constant moisture produces consistent syngas. A gasifier running on municipal waste, digestate, or sewage sludge produces syngas that shifts with every batch. The purification system has to handle that shift in real time, or the engine shuts down. The cost of that failure is measured in unplanned downtime, emergency maintenance calls, and missed power purchase commitments.

The X-150 was designed for difficult feedstocks: palm empty fruit bunches, digestate from biogas plants, horse manure, MSW pellets. That means it could not use feedstocks that stay constant. From the start, SyngaPure was designed to handle variability.

The Decision

When Zero-X designed SyngaPure, there were two options.

Option one: three independent stages. A catalytic tar reformer. A sorbent bed for acid gas removal. A particle filter. Each stage is designed to hit its own target. The reformer optimises for tar cracking temperature. The sorbent bed optimises for H2S breakthrough. The filter optimises for pressure drop. None of them see the others' data. When feedstock changes, each stage responds independently, which means they can work against each other. The reformer shifts temperature to handle more tar, which changes the gas composition the sorbent bed sees, which shifts its performance, which changes the load on the filter. The system oscillates.

Option two: one integrated system with a shared AI control layer. The same three physical stages (reformer, sorbent, filter) but one control loop sees all three sets of sensors simultaneously. When feedstock moisture spikes, the AI adjusts reformer temperature, sorbent bed throughput, and filter blowback timing as a coordinated response. The system adapts holistically.

Zero-X chose option two.

The COMETHA numbers show why this matters. The X-150 ran 1,939 hours at the COMETHA site in Paris, processing 16,382 kg of digestate pellets from three different waste streams: MSW, sewage sludge, and horse manure. Over that campaign, SyngaPure maintained 99.98% tar conversion. Not 99% with occasional dips. 99.98% sustained across feedstock transitions that would have sent a sequential system into oscillation.

That 99.98% figure is not the result of a better catalyst or a better sorbent. It is the result of system-level integration. The AI control layer sees the reformer, the sorbent bed, and the filter as one system. When one stage struggles, the other two compensate. The control logic was written to handle variability, not to assume constant conditions.

What It Means

Three consequences follow from this decision. They matter for different audiences in different ways.

For the operator: The X-150 can accept feedstocks no other small-scale gasifier at its price point can handle. Competitors like Spanner Re have sold 500+ units, but they run on clean wood chips only because their purification systems cannot handle the variability of municipal waste or sewage sludge. SyngaPure was built for exactly that variability. The Walzenrost roller grate stops the gasifier from clogging on high-ash feedstocks. The SyngaPure AI control stops the purification system from oscillating when the syngas composition shifts. The COMETHA campaign ran 10+ day continuous cycles without requiring intervention. For a factory producing digestate or a municipality processing sewage sludge, this means the machine runs on the waste you already have, not the clean wood chips you would need to buy.

For the investor: The integrated control architecture is a defensible moat. Sequencing three independent stages and calling it a system is what almost every competitor does. Building one AI layer that controls all three stages holistically took more engineering upfront but produces a system that is harder to replicate. A competitor can buy a reformer, a sorbent bed, and a filter from the same suppliers Zero-X uses. They cannot copy the control logic that makes them work together. The 1,939-hour COMETHA validation de-risks this: 99.98% tar conversion sustained across feedstock transitions is a demonstrated result, not a lab claim. The distributed waste-to-energy market is estimated at EUR 8-15 billion by 2030. Zero-X developed the X-150 for approximately EUR 7 million total. Agnion/Andritz spent EUR 41 million on a comparable heatpipe reformer effort before stopping. The economics of an integrated design show up in the development cost as well as the operating performance.

For the engineer: The AI control layer lowers maintenance costs and is extensible. In a sequential system, each stage drifts independently over time. The operator tunes them one at a time, reacting to symptoms. In an integrated system, the AI detects drift patterns before they cause a failure, and rebalances the stages proactively. Adding a fourth stage (a methanation reactor for e-fuel production, a CO2 capture loop, or a radionuclide capture system) does not require redesigning the control logic. The new stage becomes one more sensor and actuator in the same holistic loop. Zero-X holds patents on radionuclide capture in gasification developed for Fukushima's contaminated biomass. That system plugs into the same AI control framework. The architecture was designed for expansion from day one.

Why It Matters

The gasification industry has spent the last decade chasing better catalysts and better sorbents. Those matter. But the biggest jump in performance came from a control architecture decision, not a chemistry breakthrough. 99.98% tar conversion was achievable because three stages were designed as one system, not three independent boxes bolted together.

That decision is what makes the COMETHA dataset unique. It is the most documented small-scale oxy-steam campaign in the world: 1,939 hours, 33% H2 syngas, near-zero N2, 99.98% tar conversion. Not because the X-150 has better hardware than everyone else. Because the software layer connects the hardware the way a good platform should.

The post that went up on Zero-X's LinkedIn asked: "What's a 'harder path' decision that quietly defined your last project?" The answer for Zero-X was SyngaPure's control architecture. The consequence is a machine that runs 80 days straight on sewage sludge while its competitors still chase steady-state clean wood chips.

SyngaPure is not a product you can buy off the shelf. It is a design philosophy baked into the X-150. And it is the reason the X-150 is ready for commercial deployment while similarly ambitious gasification projects (Agnion/Andritz spent EUR 41 million and stopped) are still trying to make independent stages work together.

If you manage a waste stream that shifts with every batch (digestate, MSW, sewage sludge, palm EFB) or you are evaluating the X-150 for a deployment site, this is the system that makes the economics hold. 99.98% tar conversion is not a lab number. It was measured over 80 days and 16,382 kg of waste in Paris on the same feedstock types you are processing right now. The control architecture decision is the reason that number exists.