To follow up on the last blog, here's the next move I've made using perplexity.ai

I hope to one day build a an energy efficient and future-proofed biochar kiln for mixed waste and pure feedstocks using Variable Frequency Microwave Assisted Pyrolysis (VFMAP) technology:

Executive Summary

Variable Frequency Microwave Assisted Pyrolysis (VFMAP) presents a transformative paradigm for engineering hierarchical pore architecture in carbon-based biochar electrodes. Unlike fixed-frequency systems locked at 2.45 GHz or 915 MHz, VFMAP can sweep a broadband frequency range (typically 2430–6000 MHz and beyond) to selectively couple with the evolving dielectric state of the feedstock at each thermochemical stage. When this frequency agility is governed by a machine learning (ML) control loop, it becomes possible to encode "feedstock-adaptive frequency tuning" — a closed-loop strategy in which the system continuously reads the dielectric fingerprint of the carbonising mass and adjusts frequency (and power) in real time to steer pore architecture toward a defined electrode-performance target. The concept of "peak synergy" refers to the operating envelope in frequency–temperature–power space where macro-, meso-, and micropore formation processes are simultaneously reinforced, yielding a hierarchical network that maximises both EDLC charge storage (micropore surface area) and electrolyte-ion transport kinetics (mesopore highway), at an electrically continuous graphitic backbone (graphitization degree for conductivity). This report details the physical basis, ML architecture, sensor-feedback design, optimisation targets, and implementation pathway for achieving this peak-synergy state from carbon-based feedstocks.1

1. Physical Foundation: Why Frequency Governs Pore Architecture

1.1 Dielectric Evolution During Pyrolysis

The dielectric properties of a carbon-based feedstock are not static — they evolve dramatically across the three pyrolysis stages, and this evolution is the physical lever that makes feedstock-adaptive frequency tuning meaningful.

Stage 1 — Drying and torrefaction (25–250 °C): Dipolar polarisation of water dominates. The dielectric constant (ε′) is high and highly frequency-dependent; loss tangent (tan δ) peaks as free water rotates in the microwave field. At this stage, higher frequencies (>2.45 GHz) couple more efficiently to water molecules, driving rapid drying.2

Stage 2 — Pyrolysis/volatilisation (250–500 °C): Cellulose, hemicellulose, and lignin progressively decompose. Polar molecules and oxygenated functional groups (–OH, C=O, C–O–C) dominate the dielectric response. Loss factor (ε″) decreases significantly — biomass dielectric constant drops ~60–65% in the 175–500 °C window. Lower frequencies with deeper penetration depth help maintain uniform volumetric heating as the material transitions.3

Stage 3 — Carbonisation and graphitisation (>500 °C): Free electrons and graphitic microcrystal edges become the dominant loss mechanism (interfacial/Maxwell–Wagner polarisation and scattering). The dielectric constant and loss factor sharply increase as biochar forms, particularly above 500 °C. High-conductivity carbon domains exhibit surface-plasmon-like resonance modes whose absorption peak is strongly sensitive to particle radius, crystallite size, and relative density — all tunable via microwave frequency.42

This staged dielectric transformation means that a single fixed frequency is inherently a compromise. A variable-frequency system that tracks the dielectric state can apply the optimal frequency at each moment.

1.2 Frequency-Selective Pore Formation Mechanisms

Three mechanistic pathways link microwave frequency to specific pore-scale outcomes in carbon feedstocks:

1. Selective component heating: Cellulose, hemicellulose, and lignin have distinct dielectric loss spectra. At a frequency window around 5525 MHz, alkali metal ions (K⁺, Na⁺, Ca²⁺) — which are natively distributed throughout biomass ash — absorb microwave energy strongly enough to drive pyrolysis without added susceptors. Because alkali metals are spatially associated with cell-wall structures, their selective heating generates micron-scale temperature gradients that template micropore and mesopore arrays corresponding to the original biomass architecture.1

2. Hotspot engineering: Microwave directional pyrolysis simulations show that hotspots form preferentially at the centre of biochar particles due to uneven field distribution. These hotspots promote local amorphous-to-graphite conversion and pore growth. Frequency controls the spatial distribution of field nodes and antinodes within the reactor cavity; sweeping frequency dithers hotspot positions, averaging them across the particle ensemble for more uniform pore development.5

3. Graphitisation degree control: At the biochar-formation stage, microwave frequency governs the rate of sp³→sp² carbon conversion. Iron-catalysed microwave graphitisation achieves 95.12% graphitisation at 1300 °C compared to 87.67% under conventional heating, and can initiate graphitisation as low as 800 °C. Higher graphitisation yields a higher IG/ID ratio (confirmed to increase from 0.92 to 3.29 in graphite-grade material after microwave treatment), which directly controls electrical conductivity and the capacitance contribution of the EDLC electrode.67

1.3 The Hierarchical Pore Target for Electrode Performance

For supercapacitor/battery electrodes, hierarchical pore structure must balance three pore scales with specific electrochemical functions:89

Research consistently shows that "peak synergy" corresponds to a hierarchical ratio of approximately ~40% mesopores (with mesopore volume ≥ 0.4 cm³/g), micropore volume ≥ 1.0 cm³/g, and BET surface area ideally exceeding 1000–2000 m²/g depending on the feedstock. MAP-derived carbon electrode materials have demonstrated specific capacitances of 592 F g⁻¹ in matched-ion electrolytes, far exceeding conventionally pyrolysed analogues. A hierarchical corn-straw biochar (KOH-activated, MAP-assisted) achieved SSA of 2790.4 m² g⁻¹ and 327 F g⁻¹ at 1 A g⁻¹ with 120,000-cycle stability at 1.6 V. Crucially, MAP generates 47.6% higher SSA and 55.7% higher specific capacitance compared to conventional pyrolysis of the same feedstock, establishing the MAP baseline advantage that VFMAP seeks to further optimise.1011128

2. The VFMAP-ML Architecture for Feedstock-Adaptive Frequency Tuning

2.1 Hardware Prerequisites

Implementing feedstock-adaptive frequency tuning requires replacing magnetron-based fixed-frequency generators with solid-state GaN (gallium nitride) microwave generators. GaN technology offers precise digital controllability of both frequency and phase, operates at low voltage (~50V vs. ~20kV for magnetrons), has lifetimes up to 100,000 hours vs. 4,000–6,000 hours for magnetrons, and can sweep frequency in real time under software control. Real-world deployment has been demonstrated: Scanship's marine microwave pyrolysis systems (Vow ASA) replaced magnetrons with RFHIC 30kW, 900–930 MHz GaN solid-state generators, achieving uniform and consistent heating patterns with feedstock-composition-adaptive frequency tuning. A VFMAP research reactor targeting biochar electrode optimisation should span at minimum 2430–6000 MHz, as demonstrated by the solid-state variable-frequency fixed-bed reactor used in the 5525 MHz corn stalk work.131

2.2 Sensor Array: The Dielectric Fingerprint Loop

The ML system requires real-time feedstock state estimation. The primary sensing modality is in-situ dielectric spectroscopy — measuring ε′ (dielectric constant) and ε″ (dielectric loss factor) across the operating frequency range during pyrolysis. Microwave dielectric spectroscopy (MDS) is an online, compact, non-destructive analytical method based on the rotation of polar molecules in an alternating EM field. The cavity perturbation technique can measure dielectric properties from room temperature to 1400 °C across 400–3000 MHz on ~0.1 cm³ samples with sufficient sensitivity for low-loss biomass.143

An ML-assisted, real-time RSSI (Received Signal Strength Indicator)-based microwave sensor approach can extract dielectric properties at production scale without requiring sample extraction. The sensor array should yield, at minimum, two real-time state variables per frequency band:15

• Loss tangent (tan δ = ε″/ε′): Proxy for current microwave-material coupling efficiency and dominant polarisation mechanism
• Penetration depth: Proxy for volumetric heating uniformity; samples must not exceed the half-power depth to avoid hot spots or uneven heating2

Secondary sensors feeding the ML state vector:

• Gas-phase composition (inline MS/FTIR): CO₂/CO ratio indicates gasification intensity (and therefore macropore formation via C + CO₂ → 2CO mineral-template etching)10
• Temperature field (4D thermal imaging or multi-point thermocouple array): Detects hotspot formation and spatial uniformity
• Biochar electrical conductivity (contactless RF impedance): Real-time proxy for graphitisation degree (ID/IG) and electrode conductivity

2.3 ML Model Architecture: Three-Layer Stack

The ML system for feedstock-adaptive frequency tuning is most effectively implemented as a three-layer hierarchy, each layer operating at a different timescale and using a different ML paradigm.

Layer 1 — Feedstock Characterisation (Offline, Pre-Run)

Algorithm: Gradient Boosting Regression (GBR) or Random Forest (RF) surrogate model.

Before pyrolysis, proximate and ultimate analysis data of the feedstock (volatile matter, fixed carbon, ash content, elemental C/H/N/O/S, cellulose/hemicellulose/lignin ratios) are fed to a pre-trained GBR model. This model predicts the expected dielectric evolution trajectory (the ε′(f,T) and ε″(f,T) response surfaces over the full pyrolysis temperature range) and outputs a nominal frequency programme — a staged frequency schedule matched to the anticipated feedstock dielectric behaviour.

GBR-based models have demonstrated R² values of 0.90–0.95 for simultaneously predicting biochar yield, nitrogen content, and specific surface area (SSA) from feedstock elemental/proximate compositions and pyrolysis conditions. Rough set-based ML models have similarly identified biomass load, microwave power, time, and absorber dosage as the core parameter set for predicting yield and HHV under microwave pyrolysis. The nominal frequency programme from Layer 1 becomes the initial policy for Layer 2.1617

Layer 2 — Real-Time Dielectric Tracking (Online, 0.1–10 s timescale)

Algorithm: Long Short-Term Memory (LSTM) or Bayesian LSTM surrogate model.

During pyrolysis, the sensor array streams real-time dielectric readings. The LSTM model — trained on a database of (dielectric state vector) → (pore structure outcome) mappings built from prior experimental runs — continuously predicts the projected pore structure trajectory based on the current and recent history of ε′, ε″, temperature, and gas phase data. The Bayesian LSTM variant is particularly valuable because it outputs not just a point prediction of SSA, micropore volume, and mesopore ratio, but a credible interval, allowing the optimiser to quantify uncertainty and avoid overconfident control actions.1819

LSTM-based and GRU-based surrogate models for pyrolysis processes have achieved mean absolute error (MAE) values as low as 0.790 for carbonisation process variables in coking/pyrolysis analogue systems, supporting effective deep RL control. The key insight is that dielectric state is itself a proxy for structural state: as microcrystalline graphite domains grow (graphitisation stage), the loss factor increases sharply and the frequency dependence of ε″ shifts from dipolar to conduction-loss character. The LSTM can learn these fingerprint transitions and map them to pore structure evolution.20

Layer 3 — Closed-Loop Frequency Optimisation (Online, every control timestep)

Algorithm: Deep Deterministic Policy Gradient (DDPG) Reinforcement Learning agent with physics-informed constraints, or alternatively Multi-Objective Bayesian Optimisation (MOBO).

This is the core frequency tuning engine. At each control timestep, the RL agent receives:

• The current dielectric state vector (from Layer 2 sensor readings)
• The LSTM's predicted pore structure trajectory and uncertainty bounds
• The current temperature and gas phase state
• The cumulative deviation of the predicted trajectory from the target pore structure (the reward signal)

The agent outputs an action vector: the next frequency setpoint (within the hardware's operational range), the power level, and optionally the sweep modulation pattern (continuous sweep, dithering, stepped). The RL-MPC architecture — which unifies RL-based policy learning, MPC-based real-time optimisation, and digital twin-based process simulation — has been established as the state-of-the-art for nonlinear manufacturing process control. Deep deterministic policy gradient (DDPG) has been shown to outperform PID and simpler controllers in complex pyrolysis-analogue carbonisation environments, particularly under variable feedstock inputs.2120

For the multi-objective pore structure target, a Pareto front-based Bayesian Optimisation approach (e.g., BoTier multi-tiered BO) can simultaneously optimise:22

• (f_1): Maximise BET specific surface area (SSA)
• (f_2): Maximise micropore volume fraction
• (f_3): Maintain mesopore proportion ~40% (the dual-threshold optimal window)12
• (f_4): Maximise graphitisation degree (IG/ID ratio, proxy for electrical conductivity)
• (f_5): Minimise energy input (frequency efficiency)

2.4 Digital Twin Integration

A digital twin of the VFMAP reactor — built on coupled electromagnetic (EM) simulation and pyrolysis kinetics — serves two functions: (a) pre-training the RL agent in simulation before any physical runs (avoiding costly trial-and-error), and (b) providing a real-time parallel simulation state that the control layer can query for predicted consequences of proposed frequency actions before they are executed. Physics-Informed Neural Networks (PINNs) can embed the governing Maxwell equations and Debye polarisation relations into the surrogate model, improving prediction reliability when training data is sparse — a critical advantage at the early stages of building the VFMAP database for a new feedstock.2321

3. The Peak Synergy Concept: Defining the Objective Function

"Peak synergy" is defined as the state in frequency–power–time–temperature space where the three hierarchical pore formation mechanisms are simultaneously reinforced:

1. Macropore seeding (T < 400 °C, high frequency ~5.5 GHz): Selective heating of K⁺ and alkali metal ions at cell-wall junctions initiates macropore templating along original biomass vascular architecture. Target dielectric signature: ε″ response plateau (confirming ion-coupling mode, not dipolar water mode).1

2. Mesopore generation (400–600 °C, tuned intermediate frequency): The cellulose-to-lignin thermal sequence creates mesopore templates. Microwave-coordinated KOH activation at this stage directionally modulates N/O co-doped mesopore networks. The critical mesopore proportion target (~40%) minimises diffusion resistance while maintaining micropore volume. Target dielectric signature: loss factor decreasing through cellulose-decomposition zone, then recovering as lignin aromatises.2412

3. Micropore formation and graphitisation (600–900 °C, frequency calibrated to graphite domain resonance): Gasification reactions (C + CO₂ → 2CO) etch micropores within the graphitic domains; simultaneously, graphitisation of the aromatic skeleton raises electrical conductivity. At this stage, the loss tangent is dominated by conduction loss and surface-plasmon-like modes of graphite crystallites. Microwave frequency needs to match the absorption resonance of the evolving graphite crystallite size — a parameter that shifts as graphitisation proceeds. The MAP process can achieve full graphitisation of amorphous carbon (IG/ID rising from 0.92 to 3.29), but the degree is controllable via frequency and power.74

The peak synergy is reached when the ML system's frequency programme achieves simultaneous progression through all three formation regimes at their optimal rates. An over-aggressive frequency that drives graphitisation too fast (e.g., high-power hotspots) collapses the developing mesopore walls; under-coupling at the mesopore formation stage leaves the pore hierarchy dominated by macropores with insufficient micropore volume for charge storage. The Bayesian MOBO with the five-objective Pareto front (described in Section 2.3) is the mathematical formulation of finding this balance.12

4. ML Training Strategy and Data Architecture

4.1 Building the Training Database

The primary data challenge for VFMAP-ML is that variable-frequency pyrolysis with full pore characterisation feedback is not yet a standardised experimental protocol. The training database must be built strategically:

Phase 1 — Dielectric mapping experiments: Measure ε′(f, T) and ε″(f, T) for the target carbon feedstock from 25 °C to 900 °C across the full frequency range using a cavity perturbation apparatus. These experiments provide the foundational input-space mapping for Layer 1 and Layer 2 models.32

Phase 2 — Pairwise frequency-outcome experiments: Run a Design of Experiments (DoE) matrix of fixed-frequency and stepped-frequency programmes across the temperature stages. For each run, collect BET-SSA, pore size distribution (N₂ adsorption, DFT analysis), Raman spectroscopy (ID/IG), XPS (surface chemistry), and electrochemical performance (CV, GCD, EIS) of the resulting biochar electrode. This generates the (process parameter) → (electrode outcome) mapping that trains the GBR/LSTM surrogate models.

Phase 3 — Active learning: Once the surrogate models are trained on Phase 1–2 data, an active learning loop uses Bayesian optimisation to select the most informative next experiment — prioritising frequency-temperature combinations where model uncertainty is highest. This significantly reduces the total number of experiments needed to map the full Pareto front.2526

4.2 Feature Engineering for the ML Models

For carbon-based feedstocks specifically, the ML input features should include:

Output targets for multi-target GBR (offline) and LSTM (online):

5. Implementation Pathway

5.1 Phase 0 — Digital Twin Construction

Before any physical VFMAP experiments, build a COMSOL Multiphysics (or equivalent) digital twin coupling:

• EM module: Calculates electric field distribution within the reactor cavity at each frequency setpoint using the measured feedstock dielectric properties as boundary conditions
• Heat transfer module: Solves the coupled electromagnetic-thermal problem, predicting temperature distribution including hotspot locations
• Kinetics module: Applies Broido–Shafizadeh or distributed activation energy model (DAEM) for cellulose/hemicellulose/lignin decomposition, outputting volatile evolution and char yield at each temperature/time step

This digital twin enables pre-training the DDPG agent through millions of simulated control steps at zero experimental cost. The PINN-enhanced version can embed Debye polarisation physics directly into the surrogate architecture.2123

5.2 Phase 1 — Dielectric Characterisation

Measure the target carbon feedstock's dielectric properties using cavity perturbation (400–3000 MHz, 25–900 °C). These data directly parameterise the digital twin and provide the foundation Layer 1 nominal frequency programme. Key design rule: above 500 °C the sample dimension must not exceed the microwave half-power depth to prevent hotspot formation.2

5.3 Phase 2 — Surrogate Model Training

Use the DoE experimental matrix (Phase 2 data, above) to train the GBR (offline feedstock characterisation model) and LSTM (online dielectric tracking model). Target accuracy benchmarks are achievable — published ML models for biochar pyrolysis properties have reached R² > 0.90 for SSA and yield, and Bayesian LSTM frameworks have demonstrated reliable surrogate modelling for process engineering systems. ML models for biochar quality achieved >99% prediction accuracy for fixed carbon content and volatile matter content in validated studies.271718

5.4 Phase 3 — RL Agent Deployment

Deploy the DDPG-RL agent in a hardware-in-the-loop configuration: agent → GaN generator frequency command → reactor → sensor array → LSTM state estimator → reward computation → agent update. The reward function encodes the multi-objective pore synergy target:

[ R = w1 \cdot \text{SSA} + w_2 \cdot V{\text{micro}} + w3 \cdot (1 - |f{\text{meso}} - 0.40|) + w4 \cdot (I_G/I_D) - w_5 \cdot E{\text{total}} ]

where the weights (w_i) reflect the priority trade-off for electrode application (e.g., for supercapacitors, (w_1) and (w_2) dominate; for battery anodes, (w_4) becomes more important). The Pareto front generated by MOBO across this five-dimensional objective space maps the achievable trade-off surface, from which the operator selects an operating point corresponding to the desired electrode application.

5.5 Phase 4 — Feedstock Transfer Learning

When the ML system encounters a new carbon feedstock, it does not need to restart from scratch. The dielectric mapping (Phase 1) of the new feedstock is used to compute a dielectric similarity metric against the training database. The GBR Layer 1 model generates a new nominal frequency programme via interpolation. The LSTM is fine-tuned on a small set (10–20 runs) of new-feedstock experiments using transfer learning, dramatically accelerating adaptation. This is the operational meaning of "feedstock-adaptive" — the system adapts its frequency programme to new input materials without full retraining.16

6. Critical Challenges and Open Research Questions

The most consequential open research question is the experimental validation of the frequency–pore-structure transfer function — demonstrating that a specific frequency sequence consistently produces a specific hierarchical pore distribution for a given feedstock class. While selective heating by alkali metals at 5525 MHz has been demonstrated, and microwave-modulated graded pore carbon has been shown to be achievable through microwave/KOH integration, the direct mapping of frequency programme → pore-size distribution at the VFMAP timescale (sub-second frequency changes) has not yet been published for carbon-based electrode feedstocks. This is the primary experimental gap the proposed VFMAP-ML system would fill.291

7. Projected Performance Targets

Based on the literature of MAP-derived hierarchical pore biochar electrodes and the incremental advantage of VFMAP-ML frequency optimisation over fixed-frequency MAP, the following performance envelope is achievable for carbon-based feedstock electrode materials:

8. Conclusion

VFMAP with ML-programmed feedstock-adaptive frequency tuning represents a convergence of three frontier capabilities: solid-state GaN microwave technology enabling precise digital frequency control, in-situ dielectric spectroscopy providing real-time feedstock state estimation, and deep RL/Bayesian optimisation providing the policy engine to navigate the multi-objective pore-structure trade-off space. The "peak synergy" objective — simultaneous maximisation of micropore charge-storage sites, ~40% mesopore ion-transport fraction, and graphitisation degree — is operationally achievable through a three-layer ML stack (GBR offline, LSTM online, DDPG/MOBO real-time) trained on a dielectric-mapped experimental database and pre-trained on a physics-coupled digital twin.

The physical basis is sound: selective frequency coupling to alkali metal ions seeds macropore architecture at ~5.5 GHz; intermediate frequencies drive mesopore formation during the cellulose-to-lignin thermal sequence; and frequency tuning of the graphite crystallite resonance condition controls graphitisation degree at the carbonisation stage. The ML system's role is to find and execute the precise frequency–power–time trajectory through this three-stage transformation that simultaneously satisfies the multi-scale pore structure target — the definition of peak synergy. Carbon-based feedstocks (biochar-derived, lignin, algal carbon, waste polymer carbon) are particularly amenable to this approach because their dielectric evolution is strongly linked to the progressive aromatic condensation that simultaneously governs pore architecture and electrode conductivity, providing a rich and interpretable sensor signal for the feedback loop to act upon.

References

1. Full Length Article Beyond microwave susceptors: Exploring 5525 MHz frequency for efficient biomass pyrolysis - The transition from fossil fuels to renewable energy sources necessitates innovative approaches to b...

2. Dielectric properties of biomass by-products generated from wood and agricultural industries in Finland - PubMed - Knowledge of the dielectric properties (complex permittivities) of biomasses is critical for underst...

3. Microwave dielectric properties of biomass in combustion ...

4. Surface-plasmon-like modes of graphite powder compact in microwave heating - We determine the mechanism of rapid and selective heating of nonmagnetic conductive particles by the...

5. Microwave Pyrolysis for Biochar Design | PDF - Scribd - This document discusses using microwave pyrolysis and simulation to design porous biochar structures...

6. Atomic-Scale Mechanisms in Microwave-Enhanced Iron-Catalyzed ... - Graphite-based materials hold significant applications in metallurgy, electronics, nuclear engineeri...

7. [PDF] Full graphitization of amorphous carbon by microwave heating - The IG/ID ratios of the pristine samples were 0.92 (G60) and 1.54 (KB), and these values increased t...

8. Biochar-based carbons with hierarchical micro-meso ... - The development of supercapacitors with high energy density and power density is an important resear...

9. A critical review on biochar-based engineered hierarchical ...

10. Microwave-assisted pyrolysis for advanced sustainable ... - Driven by global carbon neutrality goals and resource cycling demands, sustainable carbon materials ...

11. Electrolyte ions-matching hierarchically porous biochar ... - Engineering high-performance carbonaceous electrode materials from earth-abundant biomass has attrac...

12. Microwave-engineered biochar unlocks fast, efficient CO₂ capture ... - A research team unveils a microwave-assisted, two-step activation strategy that converts corn straw ...

13. [Case Study] Microwave Pyrolysis with GaN Solid-State ... - RFHIC - RFHIC's GaN solid-state microwave generator had significantly longer lifetimes than magnetrons, and ...

14. Microwave dielectric spectroscopy A versatile methodology for ... - Microwave dielectric spectroscopy (MDS) is an online, compact, non-destructive/invasive, low power a...

15. Machine learning-assisted real-time RSSI-based microwave auto ... - This study presents a machine learning (ML)-assisted, real-time, Received Signal Strength Indicator ...

16. Rough set-based machine learning for prediction of biochar properties produced through microwave pyrolysis - Biomass Conversion and Biorefinery - Biochar, a carbon-rich compound, has a plethora of applications being explored in soil improvement, ...

17. Machine learning predicting and engineering the yield, N content, and specific surface area of biochar derived from pyrolysis of biomass - ## Abstract

Biochar produced from pyrolysis of biomass has been developed as a platform carbonaceou...

18. Bayesian LSTM framework for the surrogate modeling of process engineering systems

19. Bayesian LSTM framework for the surrogate modeling ...

20. Surrogate model-integrated deep reinforcement learning ...

21. Reinforcement Learning and Model Predictive Control for Real-Time ... - Modern manufacturing processes—particularly in chemical, pharmaceutical, food, and advanced material...

22. BoTier: Multi-Objective Bayesian Optimization with Tiered ... - arXiv - We introduce BoTier, a composite objective that can flexibly represent a hierarchy of preferences ov...

23. A Knowledge-driven Physics-Informed Neural Network model - arXiv - Physics informed neural networks (PINNs) to simulate multiple stiff, and semi-stiff ODEs that govern...

24. Microwave-coordinated KOH directionally modulated N/O co-doped porous biochar from Enteromorpha and its structure–effect relationships in efficient CO 2 capture - To develop the CO2 adsorption capacity of Enteromorpha-based biochar, N/O co-doped porous biochar (N...

25. Trading Off Quality and Uncertainty Through Multi-Objective ... - In this paper, we attempt to address the above issue. We directly treat exploitation (reflected by q...

26. [PDF] Trading Off Quality and Uncertainty Through Multi-Objective ... - The primary aim of the BBO algorithm is to balance exploitation and exploration by selecting q trade...

27. Revolutionizing Biochar Production with AI and Pyrolysis - Discover how AI optimizes palm kernel shells into high-quality biochar, enhancing energy and carbon ...

28. Controllable microwave graphitization of carbon-carbon ... - By meticulously adjusting the degree of graphitization to control the ratio of graphite to amorphous...

29. Microwave-modulated graded porous carbon for ...

30. Nitrogen and Sulfur Co-doped Hierarchical Porous Biochar Derived ... - The preparation of high-value biochar as a supercapacitor is a promising way for biomass utilization...

31. Unravelling the properties of microzonal carbon from waste hard rubber by selective thermal transformation via conventional heating and microwave irradiation - Concerns regarding resource limitations and environmental pollution have led humankind towards a cir...

Here's an extract from my latest perplexity.ai search for solid state Microwave Assisted Pyrolysis (MAP) using Variable Frequency Microwave (VFM). This is a MAJOR gamechanger for pyrolysis of waste/waste management using different feedstocks or mixed feedstocks eg. plastic...

Feedstock-adaptive frequency tuning is essentially the closed-loop control layer that makes solid-state microwave generators useful beyond simple power-on/power-off switching. Here's how it works in practice and what's been demonstrated.

The Core Problem It Solves

As biomass pyrolyses, its dielectric properties change continuously and dramatically — the dielectric constant and loss factor of a material like oil palm shell can vary by an order of magnitude between raw feedstock and fully carbonised biochar. A fixed-frequency magnetron at 2.45 GHz deposits energy into whatever the impedance of the cavity happens to be at that moment, with no feedback. If the cavity goes off-resonance as the load changes, reflected power spikes — wasting energy, damaging the source, and producing non-uniform heating.^[1][2]

Solid-state generators solve this with a frequency sweep + reflected power minimisation loop: the generator continuously measures the forward-to-reflected power ratio (the S₁₁ parameter), sweeps frequency within its operating band, and locks to the frequency at which reflected power is lowest — i.e., where the cavity+load system is closest to resonance.^[3][1-1]

How the Control Loop Works

The SAIREM GMS series (a common lab-scale solid-state generator) implements this as follows:^[3-1]

1. Frequency scan across its available bandwidth (e.g., 2400–2500 MHz, or wider for broadband units)
2. Identifies up to two reflected-power minima — corresponding to cavity resonance modes
3. Locks to the optimum minimum and continuously tracks it as the load evolves during pyrolysis
4. Reports real-time frequency, forward power, reflected power, and load impedance via USB/PLC interface

The Tsubaki et al. (2020, Green Chemistry) cylindrical cavity resonator system — which achieved 330 °C s⁻¹ with rice straw at 915 MHz — used exactly this approach: the in situ resonance frequency shift as carbonisation proceeded was monitored in real time, and the observed frequency drift was used as a proxy for the degree of carbonisation itself. This is a significant finding — it means the microwave system can serve simultaneously as heater and process state sensor, without any additional thermocouple or optical pyrometry.^[4]

What "Feedstock-Adaptive" Means in the 5525 MHz Context

In the Scanship case, "feedstock-adaptive" meant that waste composition varies from batch to batch on a cruise ship (food waste, paper, plastics in different ratios), and the system sweeps 900–930 MHz to find the coupling optimum for whatever is currently in the reactor. This is relatively narrow-band adaptation — finding the cavity resonance, not matching a molecular absorption peak.^[5]

True molecular-selective feedstock adaptation — as implied by the 5525 MHz K⁺ approach — would require a generator capable of sweeping from ~2.4 to 6+ GHz. The operating principle is fundamentally different:

In the 5525 MHz study, the frequency was chosen a priori based on dielectric spectroscopy of corn stalk ash chemistry, not dynamically tuned in real time during the run. The "adaptive" element is therefore currently at the design/recipe level — you measure your feedstock's alkali metal content, calculate the optimal frequency, and set it. Dynamic in-process frequency adaptation based on changing K⁺ availability as pyrolysis proceeds has not yet been demonstrated but is a logical next step.^[6]

The Practical Control Architecture

A fully adaptive MAP system combining both layers would look like:

1. Feedstock characterisation (XRF/ICP-MS of ash → alkali metal content → target frequency band selection)
2. VFM generator sweeps within the target band at startup, locks to reflected-power minimum
3. Continuous tracking: as biomass carbonises, K⁺ is mobilised, volatilised, or incorporated into char-phosphate phases — the optimal frequency drifts slightly, and the loop tracks it
4. Process state inference: resonance frequency shift is used as a carbonisation progress signal (as demonstrated at 915 MHz) — potentially enabling closed-loop biochar quality control without external sensors^[4-1]

This architecture is fully implementable with current hardware — the main gap is the control software and process model mapping resonance frequency shift to carbonisation state for different feedstocks. That's a tractable experimental and modelling problem, and one where pyrolysis optimisation and automated workflows would directly apply. ^{[7][8][9][10][11][12][13][14][15][16]}

***

I (not the AI speaking) would suggest that both Machine Learning, a branch of AI, and molecular chemistry of feedstocks for feedstock characterisation, could be used to model the feedstocks. This could integrate with control software running a solid state VFM in order to optimise microwave frequencies/frequency sweeps, during pyrolysis over time for the most energy efficient pyrolysis with the highest biochar yield and possibly altogether avoid using a susceptor, such as recycled biochar from a previous batch, during the pyrolysis process.

There's also game here for microalgae here:

A possible funnel to the point of manufacturing

• Science-Art ('Think')
  • Philosophy of technology
  • Industrial Arts
  • Scientific research
  • Engineering/Mathematics
  • Design for Industry 4.0 where possible
• Build prototype ('Make', with pre-seed funding eg.Gov, philanthropists)
• Test prototype ('Break', using scientific instruments and observation skills - all clearly documented)
• Develop prototype (field testing of prototypes then refine design and more testing until a prototype is ready for commercialisation)
• Commercialise (seed funding, with possible Gov 'dollar for dollar' of Venture Capital/VC)
• Manufacturing
  • Domestic market, with Industry 4.0 regional manufacturing where possible
  • Overseas markets, with 'no border' Industry 4.0, possibly 'Open Source' IP or closed box IP eg.CADs

3 enablers

• Solving a problem with a technology 'Startup'
• Gov policy technology manufacturing priorities
• Gov funding eg. pre-seed (tiered funding for increasing size of small grants aligned to Gov problem priorities with basic vetting) and seed (outsourced expert panel + Gov 'dollar for dollar' of VC)

As the world changes so do the problems that need to be solved, often with technology. Startups can solve many of these technology problems. Regenerative Growth with Appropriate Carbon Removal Technology (ACRT, see web page) for a 'Circular Bioeconomy' can be achieved over the long-term to solve most of our sustainability problems with practically unlimited business game. New business is the enabler along with existing business adaptation.

What do you think?

Hi there. I've just done my first burn on the Flame Cap 'Corrugated Box' V2 Panel Kiln (which has a page on my website for more details).

Here are the results of the first burn:

• Built a hot fire (up to 784degsC measured with a laser thermometer near the end of the burn) and possibly all VOCs released.
• Layering was not perfect and the entire volume of the kiln was not used
• The forestry waste had minimal prior feedstock processing (cut into limb and branch lengths ~2m long but could be cut to 3m long which would be optimised for this kiln and still ergonomic enough to move around) and loaded into the kiln like a dream. Compared to all the feedstock processing I undertook for the KTE and KTR, I wish I had built this kiln years ago!
• The fire was perfectly wind protected on a still day but a high level of caution was used, allowing plenty of distance between the kiln and me in between adding feedstock to minimise the chance of burning my face
• Feedstock was not bone dry (presumably some above 15% moisture content (MC) but plenty of oily leaf from forestry waste was used intermittently throughout the burn which dried and pyrolysed the feedstock during the burn with almost zero smoke
• In kiln quench worked well with biochar mixed during the quench until zero steam and smoke
• Pretty decent volume of biochar produced for the first test burn!

• Microwave Assisted Pyrolysis (MAP): Transistor eg. GaN->Amplifier (@2.45 GHz) ->Tuned power output (with digital granularity) for the optimised dielectric loss tangent interaction->Machine Learning (ML) of unique feedstocks eg. Different categories of granulated plastic, dewatered sewage, solar/process heat dried biomass 'waste' residue, granulated car tyres etc. + co-generation of pyrolysis heat for Stirling engine eg.free piston-> (bio)electricity to self-power the electronics with a biochar/Carbon based supercap to start the process, excess bioelectricity for Redox flow eg.Fe/Redox Flow Desalination (RFD) battery storage + byproducts eg. Plastic monomers for virgin plastic and bio-oil for 'Ultra Low Sulfur Diesel (ULSD), Microalgae biochar (electrodes/electrolyte Carbon feedstock for supercaps, like the one used in this system) and more + Carbon Dioxide Removal (CDR) credits bought on a Carbon Removal Marketplace (CRM) platform

• first pyrolysis unit optimisation: tuned for the 'Microalgae biorefinery' (see previous blog).
• Most microalgae (e.g., Chlorella, Nannochloropsis, Scenedesmus) have cell walls composed mainly of polysaccharides (cellulose‑like β‑glucans, hemicelluloses, sulfated polysaccharides) and sometimes proteins and algaenan, but they lack the lignocellulosic triad of cellulose–hemicellulose–lignin that defines terrestrial biomass (https://www.intechopen.com/chapters/73244)
• a VERY competitive space. Great work is been done in Canada for 'plastic waste'. Many great biochar kilns now commercially available around the Planet at all scales of feedstock 'waste' availability and feedstock characteristics which influence the kiln 'best fit' as well as co-generation requirements
• I believe GaN transistors in GaN amplifiers could be the future for MAP - and energy efficient MAP the future of many 'waste to wealth' scenarios. How to access GaN MAP tech, pioneered by 'Scanship'?

• alternatively, is there a 'better' transistor chemistry that is available via a commercial license - or even better - an 'open source' license?

• failing that - 'flame cap' or 'TLUD' 'Appropriate Carbon Removal Technology' (ACRT) (see page on this site)

It's hard to know where to start an article about change when so much is happening around the world with so much uncertainty in everyone's minds. What is clear now I believe that what was a relatively opaque global fossil energy system has now been exposed for all it's supply network fragility (stores and flows), interdependence of National economies, global inflation relationships and microeconomics, right down to consumer shortages and inflationary bill pain. The relationships between War and fossil fuel have also been observed. All of this creates a cloud of anxiety and uncertainty, though a few points are crystal clear now.

Fossil is an energy source that most of us were sold as a way of floating more boats and build modern economies. It was a house of cards that was built around one single premise - unlimited growth (as opposed to unlimited 'regrowth' - more about that later) and no consequences for the climate, other interdependent Earth systems, ecosystems and Earth stewards. The impacts of fossil energy are now being felt everywhere. Fossil has reached every narrow corner of the globe. Fossil energy failure was unthinkable - until now.

The Iran/Middle East/Strait of Hormuz conflict exposed how much fossil energy was concentrated in one small area, commonly cited as an energy 'chokepoint'. Figures vary, but approximately 20% of world fossil fuel flows pass through this strait. It was no surprise to me, in fact predicted, that if Iran was attacked the cost of fossil would go up. What I didn't predict was how critical the Strait of Hormuz was to the world economy. After a bunch of different AI searches, I learnt that most Countries were underprepared for this outcome of fuel shortages. It's IEA policy that Countries should have 3 months of strategic fossil fuel reserves. Australia had approximately 30 days of refined petroleum products, one operating refinery at Geelong, Victoria, that could refiine crude to meet 12/2025 introduced fuel Sulfur standards, i.e. until it had a large fire, and one other refinery, at Lytton, Queensland, that was in the process of installing a module to remove Sulfur from fossil fuel to also meet those standards. Between 2012-2021, five oil refineries were shut down. We were even prepared to break an embargo on Russian oil by refining it in India and purchasing the fuel. But, we're not the EU or BRICS+ though more empathy for Zelensky would have been nice. I mean, it's not like we left him for dead.  Indeed, Australia was poorly prepared for the world fossil fuel shortage and it could have been sufficiently prepared under better leadership. I don't even like fossil fuels but they are a necessary evil until we can electrify logistics but I would never complain about classic cars and motorcycles.  The ALP has responded and now increased the fossil fuel reserve to 50 days, which is still under target, but a welcome relief given that the Iran War could fire up again and exacerbate the fossil fuel problem. Stalemate - maybe. Maybe the players should initially remain focused on opening up the Strait of Hormuz but it seems Iran is redrawing the maritime map. The energy economic virus has already spread causing global energy and food supply emergencies. Worry about the other stuff later...?

Approximately 85% of global warming emissions are from fossil fuel related products. So, if we want to cool the global climate system, the most logical choice is to 'Phase out' fossil fuel, which could require leaving some sources alone eg.shale oil (dirty), the Arctic (risky and cynical) and the Amazon (pointless). Who wants another oil conflict zone? Avoid, reduce, replace. But replace it with what? This is where it gets interesting. There are a range of industries that rely on fossil fuels either directly or indirectly. Logistics is the main one, but there's also Nitrogen based fertilisers, using LNG as a N source, construction, plastic and even pharmaceuticals and more (fact check). So, the obvious one is to electrify logistics, but this will take a long time and needs to be affordable to majority of EV consumers. Since the war began, EV waiting lists in Oz have only increased. Biosubstitution can be used for N based fertilisers eg. manure and biochar. Hydrocarbon based plastic can be replaced with plant-based bioplastic, such as hemp bioplastic. Pharmaceuticals can also use biosubstitution with bio-based chemicals.

What is becoming clear to me is that the Iran War has a silver lining if it can accelerate the uptake of fossil fuel alternatives out of economic logic, necessity and sustainability. Could this be the changing of the energy guard? I believe it is. I've spent the last 17 years researching biochar, and at this point I'm convinced that all human 'waste' can be pyrolysed and converted to charcoal. If it's from plastic, it's known as 'plastic char. If it's from a biological 'waste' source it is known as 'biochar'. Both 'plastic char' and 'biochar' can be used either natively or in various materials and can even be post-processed to produce 'Advanced Carbon materials' for various current and future applications. Biochar production systems also produce 'waste' heat, that can be used for feedstock drying, space heating, water heating and bioelectricity via heat exchangers and coupled engines, such as Stirling engines and ORCs.

What next? I've worked out an endgame, that I call a 'Regrowth Circular Bioeconomy'. This is all possible with a transition to a plant and biologically based 'bioeconomy', that integrates all of it's 'waste' back into the economy via pyrolysis that produces biochar, which really needs a 'National Pyrolysis Strategy' (with pre-seed and seed ACRT startup funding) to accelerate this transition.

I believe biochar is the key to unlocking future sustainably adapted Civilisation.

The 'Regrowth' is abbreviated for 'regenerative growth' - basically, regenerating the resource base while developing it simultaneously eg.regen agriculture, with practically no limits/unlimited to what could be achieved considering how much work needs to be done. The 'Circular' part suggests a cascade of product and material use, eg.biochar and 'Zero' waste (by it's former definition). It could be a 'System of Systems' (SoS) where standalone circular subsystems/industries form an overall circular system. In my opinion, this economic idea doesn't rule out using minerals and elements to build materials and technology. What I call 'Appropriate Carbon Removal Technology' (ACRT), steel, for eg, is critical to build this tech, such as steel stoves and kilns that produce biochar. In Australia, we have many economic mineral deposits and are lucky to produce some 'Sovereign' steel and have reliable supply chains for importing missing minerals for other steels eg.Corten and even completed steel products eg. 304 tube from Taiwan. Ideally, we could produce all of our steel in Country but may not be possible.

The main bio-based industry I have now in mind is taking a 'biorefinery' approach to using plants. The biorefinery is basically a 'black box' that

• uses plants as inputs and is plant agnostic
• produces plant biomass waste residue at the end of the plant product/extraction cascade
• needs pyrolysis to pyrolyse the biomass 'waste' residue
  • produces heat as a cogeneration product for heat engines that produce bioelectricity to power the biorefinery (with possible excess for battery, RFD storage 

    (which produces potable water during energy storage - useful for populated areas and some Greener H2 electrolysers), grid, microgrid)

  • produces biochar as a cogeneration product to grow the plants eg. build soil/filter water AND used directly for other applications OR post processed for advanced Carbon materials BOTH which can be cascaded toward lower order applications within the 'Circular bioeconomy'

The main plant group candidate for biorefinery I believe is algae, and within that broad group, microalgae, which grows on every continent. There are microalgae species that grow in both seawater and freshwater. Australia is mostly desert, and we're surrounded by ocean so it makes sense to build a network of near-coastal desert seawater driven microalgae production sites, on marginal land, for biorefinery application. I've read research that suggests protein, bioactive compounds for pharmaceuticals, pigments, nutriceuticals, biodiesel, biohydrogen, biogas, bioelectricity and biochar and more can all be extracted from different microalgae species.

The magic trick is to find an endemic microalgae species that can be grown in seawater and refined in an industrial cascade to extract all of the desired economic products from the strain.

Economically and biologically speaking, different microalgae strains contain unique average ratios of lipids, proteins and carbohydrates. They can be cultivated in different ways to tune the desired ratio. There's also the possibility of genetic engineering, which I would consider as a last resort. In the most basic business model, lipids can be extracted for biodiesel and the 'waste' biomass pyrolysed for bioelectricity and biochar. This biochar, with 'Nitrogen self doping', is suitable for electrodes in batteries and supercapacitors. So, both EVs and utility scale batteries and supercapacitors can all benefit. There's also a microalgae biochar Carbon feedstock possibility for solid state electrolytes. Most of the above, including biorefinery, is more or less blue sky R&D, though different technical aspects and products have been commercialised. Maybe this is just a billionaire's playground, but to be honest I haven't costed (or possibly engineered) a biorefinery. I think it's all the additional capex that has sunk first movers in the past. If the pre-seed funding problem can be solved for a prototype, and VC can be raised for seed funding a commercial venture (eg. dollar for dollar from the SA State Gov), then maybe it could be multiple millions to startup - not billions. I think over time, by necessity for some products eg. biodiesel and biochar, different Governments will invest in the concept and technology in the private sector if they see true value in where this direction can lead to. In the meantime, permaculture and biochar integrated systems on the ground I believe is an excellent option, which I would like to progress growing microalgae in small scale photo bioreactors for micro economy.

Future applications. For eg., why not a 'hybrid' methanol (from Direct Ocean Capture of CO2) OR microelectrolysis of H2O for H2 (ISRU) AND Electric Vehicle (EV) (with a Sodium based 3D printed Solid State Battery using Microalgae Biochar electrodes - with 'self N doping' and Potassium permanganate (KMnO4) post-activation)?

Welcome to the 'Age of Biochar'!

I've had an interest for a number of years in the field known as Micro Combined Heat and Power, or more simply, micro CHP. I would add a 'B' to the end representing biochar as a byproduct from a well engineered system. I had an idea for a TLUD micro CHP system using a large 'Navigator Burner' combined with a free piston Stirling Engine (SE) for bioelectricity. The most commonly used Stirling Engines for domestic micro CHP seem to be produced by microgen https://www.microgen-engine.com/

I did some research into a range of free piston Stirling Engines and found that less than 1.0kW output seems to have more engineering problems, with most of the low powered models relegated to academic research and the space industry. I should mention that there are a range of SE platforms to choose from as well as 'Free piston' eg.TASE.

It's an interesting field of research which I recommend having a look at. If you're really keen and can afford a conference, here's the cutting edge of SE R&D: https://21isec.sciencesconf.org/

Here are some specs for the Microgen Engine Corporation (MEC) 1.0kW 'Free piston' Stirling Engine:

• economic at 1.0kW output
• starter kits available
• almost definitely used in the BioGS-1.0 micro CHP system
• almost definitely used in the Oekofen domestic micro CHP system

A little bit more information:
BioGS-1.0
https://www.kiratechnology.com/

• biomass to power a downdraft gasifier->burner->heat + 'Free piston' SE->bioelectricity and top notch Biochar byproduct (for single purpose applications eg. electrodes or Carbon Removal in a Cascade of uses (CRCU))
• could be used as a model for larger microCHP systems
• no external power needed...it can power its own augur plus excess for batteries
• claims >95% overall efficiency but I struggle to believe this due to the ~25% efficient MEC SE probably used in their system

https://www.oekofen.com/en-gb/pellet-heating/

It seems the smoothest biomass material flow is using downdraft gasifier systems but there's definitely an engineering space for using batch mode TLUDs eg.a 4" Navigator Burner which I got an ~2h burn time on with 1kg of wood pellets (see 'Navigator Kitchen (NK) 2026' page), especially for smaller off-grid jobs eg. using the microCHP to charge a solar generator (when the sun aint shining) with heat used to heat an insulated hot water tank when hot water is needed (which may not be 24/7).

Feel free to leave a comment below...

• The conventional wisdom is that fossil fuel is scarce and renewable energy is abundant

• there's also the technology to harvest the renewable energy that generally requires mining either scarce minerals or Earth abundant minerals/elements (or something in between)
• increasingly more renewable energy harvesting technology can be upcycled at the end of its service life
• scarce minerals are not an issue for vertically integrated manufacturing at or near the same location which gives the associated National economy an advantage through taxation (and patriotism) assuming the minerals are fairly taxed and not just exported with no or minimal tax. Trade leverage can also be employed with exceptions eg. Rare Earth minerals are mainly controlled by China both inside and outside the Country
• Earth abundant minerals/elements give every economy an advantage and probably decreases the likelihood of war between them/with us over scarce minerals/elements and energy deposits/reserves eg.oil, gas and coal
• Earth abundant minerals/elements could be mined from air, land and sea eg.Carbon
• Plants are abundant in almost every ecosystem. Microalgae, macroalgae, diatoms, hemp and bamboo could be the key to unlocking the future. Cyanobacteria, the precursor of oxygenic photosynthesis, are everywhere and could be quite handy as well eg. biochar inoculated with Cyanobacteria for additional CO2 removal

Politically in Oz, the ALP is good at redistributing wealth eg. Welfare, medicines, pensioner discounts but bad at taxing it, consuming an increasing larger number of right wing fiscal ideas. They are no longer a 'tax and spend' Party but arguably adequately tax minerals but not fossil fuels eg. The PRRT for oil and gas can be outsmarted with accounting trickery and LNG export royalty taxation is practically non-existent and could be considered as an extraction of National wealth which the people own but don't profit from enough. So, wealth redistribution/spending is still real but the money pot keeps getting smaller relative to inflation, debt repayment and wages (fact check). Earth abundant minerals/elements can be used to manufacture renewable energy technologies eg. Appropriate Carbon Removal Technology (ACRT), which can be researched, designed, built, tested, developed, commercialised and used domestically and exported.

The 'Climate Emergency' is not a new 'mindset', but a dynamic 'State of Mind' which needs to adapt to new ideas that can change the 'Global Climate System' game - which, is not a game for most people and definitely not a game for most, if not all, species. 

So, what about the future?
A Re(generative)growth circular bioeconomy, in my opinion the ideal economy, probably won't be possible without a transition to 'Plant Civilization' and 'Integrated Pyrolysis 'Waste' Management' (with a 'National Pyrolysis Strategy') which may be the only hope for stabilizing the 'Global Climate System' and regrowing sustainable culture/micro economics, where absent, for future generations. 

I believe there are unlimited jobs for 'sustainable adaptation' to climate change since climate change is getting worse, linear energy systems need to be replaced with circular ones, a lot of infrastructure needs to be upgraded and many technologies such as 'Appropriate Carbon Removal Technology' (ACRT) (see web page), could be manufactured.

I think job creation should be looked at on an industry scale aka net jobs created and not get too misty eyed about less manufacturing jobs with greater automation and robotics used in factories for manufacturing. Or, manufacturing could be a once off 'DIY' 'job' or even an artisanal manufacturing cottage industry. In the case of the Biochar industry, Biochar production has both upstream plant industry biomass 'waste', wastewater treatment, sewage treatment, plastic upcycling, stove and kiln manufacturing jobs and downstream biochar application (with possible vertical integration) associated jobs and other industry integrations, such as regenerative agriculture and hard infrastructure (upgrades). The application list is steadily growing. 

I like the self-empowerment attributed to artisanal stove eg.'TLUD' and kiln manufacturing eg.'Flame cap', which is often more logistically efficient than importing tech and creates local jobs, but this is not always possible (materials, tools, skill set limitations) in which case importing is the best and possibly only option. Also, using a local fabricator may or may not be cheaper and possibly not as road tested as a popular 'Off the shelf' (OTS) industrial product which, in more developed countries, could be purchased from a local hardware store. I would love to see the Flame Cap 'Algorithm' V3 Panel Kiln panels available, using 'last minute manufacturing', in Bunnings and Mitre 10 in the future! But - to be completely honest, it's difficult to cock up a 'TLUD' stove, such as a Navigator, or a 'Flame cap' kiln, such as the Kon-Tiki cone kiln, if the correct design principles are used for engineering from the ground up. I'm predicting that the 'Algorithm' V3, and other panel kiln variations, will probably fill the gaps where Kon-Tiki cone kilns can't be built or are not logistically practical.

Some designs, such as the DIY Navigator 'Adapt' V2 TLUD, a combination of OTS components and modded OTS components are used which can save time and be achieved with a basic skill set eg. Grinding. For this tech, artisanal and industrial strategies are both used.

Overall, according to ANZBIG, it is expected that the biochar industry in Australia will be worth AUD$1-5 billion by 2030, which is a wide figure range in my opinion. I would err on something in between if new avenues/sources for pre-seed, seed and commercial funding are found for biochar production technologies. Another factor is the maturation of the 'Carbon Removal Marketplace' (CRM), which is growing exponentially and will drive ACRT demand. If Carbon Dioxide Removal (CDR) credits on various CRM platforms can pay users/Charistas on all ACRT scales, this will provide the foundation for a cooler Planet and create many jobs.

Apologies, guys. Another perplexity search blog but a great concept of 'Peak synergy' which I believe will advance the green and Carbon based electrode industry forward...

Here it is: