(752b) Kinetics and Reactor Design for Pyrolytic Treatment of Petroleum-Contaminated Soils

Authors: 
Zygourakis, K., Rice University
Gao, Y., Rice University

Scope and Significance

While marine oil spills from offshore platforms or tankers attract most of the public attention, the vast majority of oil spills (about 98%) occur on land. Every year about 10-25 million gallons of petroleum products are spilled mostly from pipelines and fixed facilities. Petroleum hydrocarbons pose long term threats to groundwater quality, inhibit the germination and root elongation rates of plants and are extremely toxic to important soil microbes. Moreover, polycyclic aromatic hydrocarbons (PAH) present in petroleum can damage the immune systems of aquatic organisms and wildlife.

Current remediation methods for petroleum-contaminated soils are either relatively slow or have unintended consequences in the form of soil damage and high-energy usage. Even worse, some processes such as aerobic bioremediation can activate toxic hydrocarbons and transform them to more noxious byproducts such as PAH derivatives. For all these reasons, there is still a pressing need for more efficient and sustainable remediation of petroleum-contaminated soils.

Pyrolysis is receiving increasing attention as an on-site remediation approach because of its potential to rapidly and reliably remove total petroleum hydrocarbons (TPH) with lower energy requirements and better post-treatment soil fertility than other ex situ thermal remediation approaches [1-3]. Pyrolytic treatment of contaminated soils at 420oC with only 15-min residence time in a continuous reactor reduced TPH by 99.9%, lowered the total PAH concentration by almost 95%, and restored soil fertility to 98% of the clean soil level [3]. These results clearly suggest that pyrolysis has the potential for improved ecosystem restoration following remediation compared to traditional thermal technologies. However, the same study also revealed potential tradeoffs between pyrolytic treatment intensity, soil detoxification efficacy and soil fertility restoration. Specifically, treatment at 470oC for 15- or 30-min reduced soil fertility to 51% and 39% of the clean soil level, which was only marginally higher that the fertility of the contaminated soil.

The existence of an optimal treatment intensity for contaminated soils underscores the need for mechanistic insight into the pyrolysis process and for a deeper understanding of how the detoxification of contaminated soil and the restoration of soil fertility are affected by (a) the properties of both the soil and the petroleum crude, and (b) the operating conditions of pyrolysis reactors.

Kinetics and Reactor Design for Pyrolytic Treatment of Petroleum-Contaminated Soils

Taking the first steps towards a rational design of pyrolysis reactors, we have combined mathematical techniques and advanced analytical measurements to:

  1. Develop a full kinetic model for the pyrolytic treatment of soils contaminated with petroleum hydrocarbons. The new model describes all thermally-induced transformations of soil components (e.g. clays, carbonates, or SOM), the evaporation of light hydrocarbons and the pyrolysis of the heavier hydrocarbon fractions.
  2. Develop a mathematical model describing the operation of continuous reactors (rotary kilns) that will be used for ex situ treatment of petroleum-contaminated soils. This dynamic model will then be used to estimate the optimal (a) pyrolysis temperature, (b) reactor residence time, and (c) gas flow rates that will achieve a specified TPH reduction at the minimum energy cost for any given throughput of contaminated soil.

Two clean soils with different compositions were spiked with oil crudes to prepare petroleum-contaminated soils. Thermogravimetry with evolved gas analysis (TG-IR and TG-MS) was used first to identify the mineral transformations taking place as the clean soils were heated in an anoxic atmosphere. The same analytical techniques were applied to study the two distinct processes occurring during the pyrolytic remediation of contaminated soils. Desorption of the lighter hydrocarbons is the dominant process for temperatures between 150 and 350oC. Thermal cracking reactions dominate in the 350-500oC range releasing pyrolysis products (like hydrogen, methane, or olefins) and leading to the formation of a carbonaceous solid material (char) that remains in the treated soil [2]. It is important to note that hydrocarbon desorption and pyrolysis overlap with the temperature ranges over which soil mineral transformations take place.

To develop our kinetic model, we assumed that the soil consists not only of inert components but also of “reacting pseudo-components” (e.g. clays, carbonates, SOM) that release either water or carbon dioxide as the soil is heated. Weight losses due to water or carbon dioxide release accounts for more than 95% of the weight loss observed during heating of clean soils. Contaminated soils contained additional pseudo-components that account for the weight losses observed due to the desorption of light hydrocarbons and the pyrolysis of the heavy petroleum fractions to yield gaseous products and char (coke). Using the distributed activation energy approach (DAE) approach, each pseudo-component was assumed to consist of a large number of species that react with activation energies distributed according to Gaussian probability density functions (PDF). To obtain the kinetic parameters, an optimization problem was formulated by fitting model predictions to experimental thermogravimetric data. Solving this nonlinear constrained optimization problem with an interior-point algorithm, the expected activation energies and variances of all pseudo-components were estimated, together with the preexponential factors for all the reactions and the initial mass fractions of all pseudo-components.

Excellent fits to the experimental data were obtained using four reacting pseudo-components for the clean soils. Three pseudo-components were required to model clay dehydration and dehydroxylation processes that occur over a wide range of temperatures and generate different multi-peak water release patterns for each soil. One pseudo-component was enough to model carbon dioxide release from either carbonate decomposition reactions or soil organic matter (SOM) pyrolysis. Two additional pseudo-components were required to model hydrocarbon desorption and pyrolysis.

The kinetic models were then used to simulate the operation of non-isothermal rotary kiln reactors that will be used for ex situ treatment of petroleum-contaminated soils. To preserve anoxic conditions, the reactors were heated externally while a stream of nitrogen gas flowed through the rotary kiln to sweep the desorbing hydrocarbons and pyrolysis products. A finite stage or mixing cell model with separate solid and gas phases was used to describe the operation of the pyrolysis reactor.

Model predictions were compared to experimental data we obtained with a continuous rotary kiln reactor [3]. The TPH reductions predicted by the model agreed very well with the experimental data for all the conditions considered in our earlier study: pyrolysis temperatures, 370, 420 and 470oC; residence times 15, 30 and 60 min; solid flow rates 7, 14, and 25 lb/h.

More importantly, however, our model was able to explain the trade-offs observed in our earlier study between pyrolysis treatment intensity, soil detoxification and soil fertility restoration [3]. The clay components of the soil used in [3] undergo significant dehydration between 350 and 450oC, a range that overlaps with that of pyrolysis reactions. As contaminated soil moves through the reactor zone kept at 420oC, the reaction rate for the heavy hydrocarbon fraction is fast enough to achieve almost complete conversion within 15 or 30 min. But, the same conditions are not severe enough to completely remove the water from the clay components of the soil. If the soil is treated at 470oC, on the other hand, the clay components are almost completely dehydrated, causing irreversible soil damage and loss of fertility.

By emphasizing the importance of soil composition and by raising the possibility of trade-offs between detoxification and fertility of treated soils, this study has significantly advanced our fundamental understanding of the multiple and concurrentphysicochemical processes taking place during the pyrolytic treatment of petroleum-contaminated soils. Theoretical kinetic and reactor models are essential for designing pyrolysis reactors optimized for specific petroleum-soil systems, enabling robust operation and facilitating acceptance of this novel technology by regulatory agencies and stakeholders.

References

  1. Vidonish, J. E.; Zygourakis, K.; Masiello, C. A.; Gao, X.; Mathieu, J.; Alvarez, P. J., Environmental Science & Technology 2016, 50, (5), 2498-2506.
  2. Vidonish, J. E.; Alvarez, P. J. J.; Zygourakis, K., Industrial & Engineering Chemistry Research 2018, 57, (10), 3489-3500.
  3. 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.