(252b) Integrating Thermal Analysis and Reaction Modeling for Rational Design of Pyrolytic Processes to Remediate Soils Contaminated with Polyaromatic Hydrocarbons

There is a pressing need for reliable, broadly applicable and sustainable remediation of soils and sediments impacted by polyaromatic hydrocarbons (PAHs) and other persistent hydrophobic hazardous pollutants. Whereas numerous site remediation approaches have been developed, many of these technologies are marginally cost-effective or unreliable. For example, bioremediation (including landfarming) is relatively slow and often difficult to accomplish and has the potential to generate more toxic PAH byproducts [1].

We recently demonstrated that pyrolytic treatment of contaminated soils under carefully selected operating conditions can rapidly and reliably remove heavy hydrocarbons (including PAHs) and completely eliminate toxicity to human lung cells, while preserving the fertility of the soil to facilitate ecosystem restoration and re-greening efforts [2,3]. The more volatile hydrocarbons desorb as temperature of the contaminated soil is raised to 350^oC. When pyrolysis temperatures rise above 350^oC, however, heavier hydrocarbons (including PAHs) undergo a cascade of pyrolytic reactions resulting in the release of hydrogen and C₁-C₄ hydrocarbons. The final product of these reactions is a thin layer of carbonaceous material (coke) that coats the particles of the contaminated soil and drastically reduces its toxicity. The advantages of pyrolytic treatment over other thermal treatment methods (like incineration) include significantly lower energy requirements and preservation of soil fertility due to retention of some organic carbon and water holding capacity.

While pyrolytic treatment can be very effective in removing hydrocarbons from contaminated soils to levels well below regulatory standards, the treated soils may exhibit significant differences in fertility, as determined by germination and biomass production metrics. Moreover, heating contaminated soils in air at temperatures above 600^oC (incineration) either drastically reduced or eliminated the fertility of treated soils [4]. Our earlier work also revealed potential tradeoffs between pyrolytic treatment intensity (and associated energy consumption), soil detoxification efficacy, and soil fertility restoration. For example, treating a contaminated soil in an isothermal reactor at 420^oC for 30 min removed 99.9% for total petroleum hydrocarbons and 94.5% for PAHs, while also restoring its fertility to virtually clean soil levels [4]. When the same contaminated soil was treated at 470^oC for 15 and 30 min, however, its fertility dropped to 51% and 39% respectively of the clean soil level, with the latter fertility value being only marginally higher than that of the contaminated soil [ibid.]. But not all soil/contaminant mixtures showed the same behavior. Working with another soil/contaminant mixture, for example, we did not observe a soil fertility “maximum” with increasing treatment temperature [1].

The fact that different soil/hydrocarbon mixtures respond differently to pyrolytic treatment underscores the need for deeper understanding of the processes taking place as contaminated soils are heated in anoxic atmospheres. Specifically, we need to distinguish between (a) thermally induced changes to soil components that affect soil fertility, and (b) contaminant desorption and pyrolysis, which ultimately determines the rate and extent of their removal.

While the primary objective of soil remediation is regulatory compliance, a deeper mechanistic understanding of the interactions between contaminants and soil components will allow us to optimize pyrolytic treatment to achieve multiple objectives like detoxification and soil fertility restoration in addition to contaminant removal. Since the optimal treatment conditions are system-specific, the design of remediation processes should be based on a systematic study of (a) the properties of the soil (e.g., texture and porosity, clay type and content, presence of catalytic metals) and the contamination scenario (type and concentration of contaminants, interactions between contaminants and metal sites, age and weathering of contamination). However, the operating conditions for current remediation processes are often selected using empirical approaches that do not address the complexity of these systems. This underscores the need for novel and comprehensive testing protocols that quantify soil/contaminant interactions, followed by robust data analysis and system modeling that will ultimately lead to the development of guidelines for selecting cost-effective operating conditions.

Integrating Thermal Analysis and Reaction Engineering to Optimize Pyrolytic Remediation

To address this need, we will present a novel approach that integrates advanced analytical measurements and mathematical modeling to rationally design a pyrolytic remediation process for specific soil/contaminant systems. We first use thermogravimetry with online infrared and mass spectrometry (TG-IR and TG-MS) to model the thermal transformations occurring when clean soil is heated in an anoxic atmosphere. By analyzing the patterns of water and carbon dioxide release (the two major products of clean soil pyrolysis), we can determine the temperature ranges over which (a) the clay components of soil release the water held between their aluminosilicate 2D sheets, (b) soil organic matter (SOM) pyrolyzes, or (c) carbonate minerals decompose to release carbon dioxide. We can then obtain the kinetics of soil decomposition by deconvolving the multipeak patterns of water and carbon dioxide release. Deconvolution is achieved by assuming that the multipeak patterns are weighted sums of the decomposition rates of a number of pseudo-components. These pseudo-components are actually mixtures of individual species (clays, SOM, carbonates) each decomposing to release either water or carbon dioxide with different activation energies that are distributed with a Gaussian probability density function (PDF). The full soil decomposition kinetics are obtained by a distributed activation energy (DAE) approach that allows us to estimate (a) the rate constants, (b) the means and variances of each activation energy PDF and (c) the weight fraction of each pseudo-component by solving a minimization problem. We will demonstrate that predictions from a DAE model with four or five pseudo-components can very accurately match the experimental thermogravimetry data [5].

Once we have obtained the full kinetics of the thermal transformations of clean soil, TG-IR and TG-MS are again used to determine how contaminant desorption and pyrolysis affects the decomposition pattern of the contaminated soil. We will demonstrate that contaminant mixtures consisting of light (LH) and heavy hydrocarbons (HH) superimpose two extra peaks to the multipeak decomposition pattern of clean soil. Thus, the kinetics for the pyrolytic treatment of contaminated soil can now be obtained by adding two additional pseudo-components (denoted here by LH and HH) to the four or five pseudo-components used to describe the kinetics of soil pyrolysis. Again, the pre-exponential factors, the distribution of activation energies and the weight fractions of the final number of pseudo-components are obtained by formulating and solving a DAE minimization problem.

The full kinetics are then used to formulate and solve a reactor model that predicts the conversion of the contaminants (including PAHs) and decomposition of soil components as they are heated in a continuous pyrolysis reactor. We first used an isothermal reactor model to compare our model predictions to experimental data obtained with a pilot-scale continuous rotary kiln reactor operating under flowing nitrogen. This electrically heated reactor treated two contaminated soils with different hydrocarbon contents at three temperatures (370, 420 and 470^oC) and three residence times (15, 30 and 60 min). We will present data showing that the isothermal model provides very accurate predictions for both (a) the residual amount of total petroleum hydrocarbons (TPH) in the treated soil and (b) the amount of PAHs remaining at the exit of the isothermal reactor.

Since commercial pyrolysis reactors will operate under non-isothermal conditions, we extended the model to simulate the operation of non-isothermal rotary kiln reactors with co-current and counter-current flow of the nitrogen sweep gas. We will present results to demonstrate how the key operating parameters (solids flow rate, loading, wall temperature, direction and flow rate of sweep gas etc.) affect the removal efficiency of contaminants.

Significance

This study presents the development of a novel framework that integrates thermal analysis methods and reactor modeling to design, scale up and optimize pyrolytic remediation processes. It consists of:

• Workflows of thermal analysis tests (TG-IR, TG-MS) that (a) identify the temperatures where key soil mineral and contaminant transformations occur, and (b) quantify the gaseous products of these transformations.
• A systematic method for deconvolving the complex decomposition patterns of contaminated soils to obtain the full pyrolysis kinetics. (DAE model for clean soil followed by a full DAE analysis of contaminated soil).
• Reactor models for selecting process conditions and guiding multi-objective optimization with a rigorous decision model.

This novel framework has the potential to become an essential tool for remediation professionals by allowing them to scale up and optimize pyrolytic treatment processes for a broad spectrum of soil/contaminant systems using a rigorous methodology that does not rely on “black box” approaches.

References

1. Chibwe, L.; Geier, M. C.; Nakamura, J.; Tanguay, R. L.; Aitken, M. D.; Simonich, S. L.,. Environmental Science & Technology 2015, 49, (23), 13889-98.
2. Vidonish, J. E.; Zygourakis, K.; Masiello, C. A.; Gao, X.; Mathieu, J.; Alvarez, P. J.,Environmental Science & Technology 2016, 50, (5), 2498-2506.
3. Vidonish, J. E.; Alvarez, P. J. J.; Zygourakis, K.,Industrial & Engineering Chemistry Research 2018, 57, (10), 3489-3500.
4. Song, W.; Vidonish, J. E.; Kamath, R.; Yu, P. F.; Chu, C.; Moorthy, B.; Gao, B. Y.; Zygourakis, K.; Alvarez, P. J. J.,Environmental Science & Technology 2019, 53, (4), 2045-2053.
5. Gao, Y., Zygourakis, K., Industrial & Engineering Chemistry Research, 2019, 58, 10829-10843.