View all hidden authors and organizations
Contributed by: Dudley R. Herschbach, August 23, 2021 (submitted for review on April 18, 2021; reviewed by Bhavik R. Bakshi, Jefferson Tester, and Kevin M. Van Geem)
The plastic garbage accumulated in the world's oceans forms a huge "plastic island" in the ocean circulation. Removal of plastic provides an opportunity for our ocean to return to a more pristine state. In order to clean the circulation, ships must collect and store plastic before shipping it to ports thousands of kilometers away. Conversely, marine plastic waste can be converted into shipboard fuel, for example, using hydrothermal liquefaction (HTL) to depolymerize the plastic at high temperatures (300°C to 550°C) and high pressure (250 bar to 300 bar). The resulting depolymerization product, called "blue diesel", has the potential for self-powered cleanup. The purpose of this work is to evaluate the thermodynamic feasibility of the program and its impact on cleanup.
Collecting and removing ocean plastics can reduce their environmental impact; however, ocean cleanup will be a complex and energy-intensive operation that has not been fully assessed. This work examines the thermodynamic feasibility and the subsequent effects of hydrothermal conversion of this waste into fuel for self-powered cleanup. A comprehensive probabilistic exergy analysis shows that hydrothermal liquefaction has the potential to generate enough energy to power the process and the ship performing the cleanup. Self-powered cleanup reduces the number of round trips to ports full of waste and eliminates the need to use fossil fuels for most plastic concentrations. Several cleanup scenarios were simulated for the Great Pacific Garbage Belt (GPGP), equivalent to removing 230 to 11,500 tons of plastic per year; this range corresponds to the uncertainty of the plastic surface concentration in GPGP. The estimated cleanup time mainly depends on the number of fences that can be deployed in GPGP without sacrificing collection efficiency. Self-powered cleanup may be a viable way to remove plastic from the ocean, and the gap in our understanding of the characteristics of GPGP should be resolved to reduce uncertainty.
It is estimated that 4.8 to 12.7 million tons of plastic enter the ocean every year, widely distributed on the ocean surface and in water bodies, deposited in sediments, and accumulated in marine organisms (1⇓ –3). A large number of studies have shown that plastics can cause significant damage to marine life and birds, thus prompting effective mitigation and removal measures (4). Reducing or eliminating the amount of plastic waste generated is critical, especially when the current load may last for years or even decades (1, 5, 6).
As a very obvious part of the comprehensive method of removing plastics from the environment (1, 5, 6), efforts are being made to collect marine plastics from the circulation gathering areas formed by ocean currents (3, 7). The current method of removing plastic from the high seas is to use a ship that must store the plastic on board until it returns to the port (usually thousands of kilometers away) to unload the plastic, refuel, and resupply.
An optimistic assessment of the cleaning time using the harvest-return method shows that it will take at least 50 years to completely remove the plastic(7), and the annual cost is 36.2 million US dollars(8); more conservative estimates indicate that partial removal will take more than 130 years( 7, 9). Decades of cleanup time means that environmental degradation may have reduced existing plastics to microscopic and smaller forms that cannot be harvested until cleanup is complete (1, 4, 9). These considerations highlight the great challenge of removing plastic from the ocean and naturally raise the following question: Is there any way to remove plastic from the ocean faster than degradation?
Some current plastic removal strategies involve accumulation through boom systems, which consist of semi-circular buoys and are equipped with fine nets that extend below the sea surface (7, 10). The location of these fences is such that prevailing water currents carry plastic to the fence and then accumulate there. The current envisaged method is to let a ship steam into the fence system, collect the plastic, then return to the port to unload and refuel, and then resume collection activities.
If the return journey of filling and unloading plastics is cancelled, the time required to recycle plastics can be reduced. In fact, the harvested plastic has an energy density similar to that of hydrocarbon fuels. Using this energy to power ships can eliminate the need to refuel or unload plastics, reducing fossil fuel use and potential cleanup time.
Self-powered harvesting may provide a way to complete cleanup using passive fence collection methods on a time scale smaller than environmental degradation. Unfortunately, with the advancement of technology (7), especially with the accumulation of plastics, cleaning itself is a constantly changing goal. Therefore, a framework is needed to evaluate the impact of self-powered harvesting on cleanup time and fuel usage. When more data becomes available, the framework can be updated.
In order to be valuable, the cleanup framework must be reduced to practice using actual technology. One possible technology to convert plastics into usable fuels is hydrothermal liquefaction (HTL), which uses high temperature (300°C to 550°C) and high pressure (250 bar to 300 bar) to convert plastic into monomers and other suitable Small molecules are used as fuel (11⇓ –13). Even in the absence of a catalyst, the oil yield of HTL is usually greater than 90%, and unlike pyrolysis, the yield of solid by-products (which need to be stored or burned in a special burner) is less than 5% (11⇓ –13 ), which gives HTL some comparative advantages. Ideally, ships equipped with HTL-based plastic conversion systems can provide themselves with fuel and make fuel from recycled materials. The result can be called "blue diesel" to refer to its marine source, in contrast to traditional marine diesel and "green diesel" derived from land-based renewable resources (14).
In order for the HTL method to be feasible, the work produced by the plastic must exceed the work required for the process, and ideally, must exceed the requirements of the ship's engine in order to store fuel for later use during the collection process. Exergy analysis provides a framework to determine the maximum amount of work that a complex process can produce without violating the basic laws of thermodynamics (15). The reliability of exergy analysis depends on the reliability of the data used as input, and the key parameters describing the performance of HTL and the concentration of plastic on the ocean surface are currently uncertain. Therefore, a rigorous and statistically significant analysis of plastic processing on ships must integrate uncertainty (16). Here, the Monte Carlo (MC) simulation method has proven its suitability for similar types of analysis, is a suitable tool to deal with the uncertainties inherent in current applications (17), and allows the integration of new information and data as a further completion Research on ocean surface plastics.
Therefore, the thermodynamic performance of the HTL process on board was evaluated to determine whether (and when) the process can provide enough energy to power itself and the ship. A framework was then developed to assess the impact of plastic conversion on ships on fuel use and cleanup time. The results provide valuable insights into the potential use of marine conversion technology in accelerating the removal of plastics from the ocean, and the framework should prove useful to guide future work in this area.
The first step is to conduct a thermodynamic analysis of the actual configuration of the plastic conversion process. Figure 1 provides a schematic diagram of the proposed plastic conversion system, which includes components that collect and pulverize plastics, remove salt and other impurities, and convert plastic raw materials into blue diesel, which is a kind of energy density and volatile similar For marine diesel (18). Detailed information can be found in section SI.1.1 of the SI appendix and table SI.1. The whole process is independent and can be easily installed on the boat. Here, 40 m is chosen-smaller than the current boat used to remove marine plastic (7), because the converted boat on the boat does not need to store the plastic it collects.
Conceptual design of the process of converting marine plastics into usable fuel based on the ship-borne HTL: process flow chart. The whole process is designed to fit a standard 20-foot container.
The thermodynamic performance of the HTL reactor itself is determined by the oil yield and heating value obtained at a given reaction temperature. The relationship between oil production and temperature depends on the composition of the plastic feed and is measured experimentally. Therefore, the two common pure plastics (2), polypropylene (PP) and polyethylene (PE) and the "mixed feed" based on the typical components of marine plastics (2:1 PE) are simulated: PP) (2). The operating temperature, total conversion rate and product yield of each plastic raw material are taken from previous work (11⇓ –13), and are listed in SI appendix tables SI.2 and SI.3 (SI appendix, SI.1.2) Section). It is worth noting that the performance of the PP/PE mixed feed is particularly promising, because the HTL of this feed produces an oil yield of 85% at a moderate temperature (400 °C) and no solid by-products.
HTL generates chemical energy in a form usable by the engine. All peripheral equipment will reduce the ideal thermodynamic performance. Parasitic losses in the operation of peripheral equipment required for crushing, heating, pumping and separation are included in the system-level analysis (see SI appendix, table SI.1). The performance of the system consisting of the HTL reactor and the ship itself was evaluated to ensure that the ship sails at a so-called normal speed or super slow speed, which is the speed recommended to minimize fuel consumption (19).
Published performance, thermodynamic characteristics, and equipment specifications provide the basis for a comprehensive exergy model of HTL-based processes. However, most of the required parameters are not accurate; projecting them to scale introduces additional uncertainty. Therefore, the uncertainty of key parameters is included in the exergy model using the MC sampling method (20).
After testing the effects of several parameters, the six parameters shown in Table 1 and the overall conversion rate of each plastic were selected (for detailed information on the reaction conditions, please refer to SI Appendix Table SI.2), including those in Uncertainty Degree analysis. Table 1 provides the reasons for choosing the range and related probability distribution function for the uncertainty quantification of all relevant model input variables.
The probability distribution curve of the input variables of the uncertain model
In most cases, the center of uncertain model input variables is their most probable value, following the probability distribution curve described above. Oil production is the only exception. For oil production, the best value published in the literature is selected as the maximum point in the uniform MC probability distribution function. This approach attempts to explain the effects of additives, aging, and contaminants (21), all of which are expected to reduce oil production.
After selecting the model input variables and their ranges (Table 1), the input uncertainty is propagated through the above exergy model to derive the probability distribution curve of the exergy result, which can be probabilistically characterized and predict net exergy consumption or power generation area. Therefore, compared with the standard single-ignition estimation in the traditional method, the derived probability distribution curve and the range of possible thermodynamic performance results can identify opportunities and risks in a more robust and detailed manner.
The MC-based model does not include the uncertainty of plastic loading on the ocean surface, which is a key parameter for control performance. Unfortunately, the plastic load is only a rough estimate and varies by location (2, 5). Therefore, in order to solve the uncertainty of plastic load, we combine the above MC process simulation with traditional plastic load sensitivity analysis by simulating the performance of a series of plastic loads.
Figure 2 shows that when the ship is traveling with full engine power (Figure 2A) and optimized engine power (1/3 engine power) (Figure 2B), the exergy generated by the ship-borne HTL is more probable than the process and the ship itself. Save fuel. The temperature required for the decomposition of PE is higher than that of PP, even though the depolymerization of the two feeds results in similar oil yields, thus explaining the predicted performance difference between PE and PP.
The probability of generating more exergy energy than the HTL process itself and the combination of ship engines consumes, used in (a) 2:1 PE and PP mixture, (b row) PP and (c row) PE used in (A) full engine Power and (B) optimize engine power [1/3 engine power (19)].
Interestingly, the performance curves of PP and PE/PP feeds show similar characteristics to each other. This is a promising finding because the PE/PP mixture best represents the plastic components present in the Great Pacific Garbage Belt (GPGP) (1 ). Based on the previous modelling of the depolymerization of the mixture, a possible explanation for the advantageous performance observed for PE/PP is the autocatalysis caused by the free radicals formed by the pyrolysis of PP (13).
The point in Figure 2 where the probability of net exergy is greater than 90% is of particular importance because it represents a point where the risk is low enough in the presence of modeling uncertainty (22) (SI Appendix, Section SI.1) . 2 contains more details). Figure 2B shows that when the ship is running with optimized engine power, the probability point of reaching 90% is reached for all plastic loads of PP and PE/PP mixtures <25% and <12%. 12% of the plastic loading is within the predicted range collected by the boom placed in the circulation (9). However, the break-even point load is much larger than expected without prosperity, which means that the HTL conversion cannot be self-powered on the high seas.
Figure 2 shows the range of predicted results, which is a result of the current uncertainty levels of HTL thermodynamics and conversion rates. Improved data and forecasting models will reduce the scope of forecast results, thereby reducing the investment risk of this method. Figure 2 also shows that optimizing engine power [1/3 engine power (19)] is necessary for the thermodynamically favorable processing of mixed plastic flows.
The SI appendix, Figure SI.1, provides more detailed information about the exergy consumed by various sub-processes, obtained when the probability of generating net exergy is greater than 90%. In all cases, the ship itself is the main source of exergy consumption, followed by heating the feed to the HTL reactor.
Pyrolysis is also considered a possible technological option for self-powered ocean cleanup. Like HTL, pyrolysis is a thermal depolymerization process that produces oil products that can be used as fuels. Unlike HTL, pyrolysis requires dry feed, which means that fire is needed to dry the ocean plastic stream before pyrolysis. Similarly, the output of solid by-products from pyrolysis is higher than HTL; these solid by-products must be stored on board—which reduces one of the main benefits of onboard conversion—or burned in dedicated burners for heating. On the other hand, HTL requires preheating of the liquid water feed, only a portion of which can be recovered from the product stream. Thermodynamic analysis is required to evaluate these trade-offs.
The SI appendix, Figure SI.2, provides a comparison of HTL and pyrolysis of mixed PE/PP streams. Interestingly, the thermodynamic performance curves of HTL and pyrolysis are very similar to each other, indicating that the differences in drying, heating, and oil yield of the two processes almost cancel each other out (13, 23). Therefore, from a strict thermodynamic point of view, either pyrolysis or HTL may be viable options for shipboard technology. In other words, the yield of pyrolysis oil is significantly lower than that of HTL (65% compared to 85%), which means that the treatment of pyrolysis by-products is much more difficult than HTL (23). These by-products include gas and char, which must be flashed, stored or burned in a solids compatible burner. The space occupied by the pyrolysis system will be larger than that required by the HTL to accommodate the dryer required for pyrolysis. These secondary considerations indicate that HTL is a more promising option for shipboard conversion technology; however, if HTL is difficult to implement, pyrolysis is still a viable option.
Scale is an important consideration. In addition to the basic case assuming a flow rate of 3.6 m3⋅h-1, the second case of a mixed plastic flow with a flow rate of 36 m3⋅h-1 is also considered. The benefit of increasing the exergy output of a fixed-size ship is greater than balancing the increased energy demand of peripheral equipment, shifting the 90% probability point from 30 vol% to 3.3 vol% (SI Appendix, Figure SI.3). Therefore, thermodynamic considerations encourage Scale the process as much as possible within the size, weight, and economic constraints of a given container. This shows a reasonable expansion of the current work: joint optimization of ship and process dimensions and ship durability, with detailed consideration of the space and energy requirements of the crew and all shipboard systems.
The previous section (Thermodynamic Analysis of HTL Plastic Conversion on Ships) predicts that HTL on ships can collect plastics from their own power. These plastics are first collected by booms placed in the natural ocean circulation. Therefore, the impact of HTL conversion on ships is predicted within the specific application of the cleanup framework: the estimated annual removal of plastic from the GPGP located in the central Pacific Ocean (2).
Figure 3A Located in the port of San Francisco, California, away from the GPGP. Figure 3 B and C respectively show the deployment of the boom array and the filling of a single boom due to ocean currents, and Fig. 3D shows the conversion of plastic into a "blue diesel" alternative fuel. As shown in Figure 3, the frame requires the following parameters: the size of each boom, the number of booms, the distance between the booms, and the distance from the boom to the ship's home port.
An overview of the process of removing plastic from GPGP, showing (A) the overall system overview, (B) part of the collection arm system, (C) a single collection arm and (D) the HTL reactor.
The existing boom design (consisting of a series of semi-circular buoys equipped with a fine mesh, extending a few feet below the sea surface) (10) The speed of collecting plastic depends on the amount of incoming plastic and the speed current of the local ocean, such as As shown in Figure 3C. In contrast to methods that involve collecting plastic, storing it on board and returning to port for unloading, onboard conversion does not directly change the values of any of these parameters. However, onboard conversion can reduce the frequency of return trips, and more oil booms can be deployed and emptied each year compared to the collect-storage-return method.
The onboard conversion shows a second potential advantage compared to the collection-storage-return method. Although plastic harvesting is usually done manually to minimize fuel use [workers manually scoop the plastic out of the fence and put it in small bags (7)], the on-board conversion makes the actual automated plastic collection possible, and the ship can pass through the fence. The plastics in the conveyor belt (7) are sent to the conversion process. Then, the continuous collection rate is related to the ship's speed and the concentration of plastic retained by the fence. Because of its efficiency, continuous collection is modeled for the case of on-board conversion. One potential problem with automatic shoveling is the dispersion of plastics. In order to minimize the wake effect and fuel usage, the ship speed during the collection period is set to 0.5 knots (kn), and the collection efficiency is set to 70% to solve the partial dispersion of the ship's wake during the collection period.
Then calculate the amount of plastic that can be removed, which is used to convert the plastic into fuel on the ship. We assume that the booms are deployed at a distance of 25 kilometers from each other. This distance was chosen to minimize the boom-boom interaction that would reduce collection efficiency, because the 25-kilometer spacing corresponds to the largest boom that can be deployed in a space-based GPGP and served by a single ship in a year quantity. Details of all these assumptions and calculations can be found in the SI appendix, table SI.4 and equations. SI.1-SI.7.
Using this framework and corresponding assumptions, when onboard conversion is used, 2,500 fences can be harvested per year. Assume that a year includes up to 240 days, of which there are 16 hour days, to take into account maintenance downtime, weather, sea conditions, and crew vacation time. Using this value as the number of fences that can be harvested each year and the distance between fences and fences, calculate the total amount of plastic that a single ship can remove from GPGP each year under several different values of plastic surface concentration (1, 2). The results are summarized in Table 2, which shows that for the case with the maximum plastic surface concentration value (2,500 g·km−2) (1), 1.2 × 107 kg can be removed. As the plastic surface concentration decreases, the amount of plastic removed decreases, which is the result of the impact of the plastic surface concentration on the fence collection.
Use shipboard HTL to remove plastics with different surface plastic concentrations in GPGP
By eliminating the trip to the port and replacing marine diesel with blue diesel, onboard conversion reduces fuel demand, especially fossil fuel demand, as shown in Table 2. Specifically, for the highest concentration, enough plastic can be collected to produce a fuel with more than 480% that can be stored and used for the trip to and from the GPGP, eliminating the need to use any fossil fuels. Three of the five concentrations produce excess fuel, which indicates that areas with low plastic concentration can be supplemented with fuel from high-concentration areas.
A key assumption in this analysis is that the HTL feed rate and conversion rate are equal to the plastic collection rate, so ships do not need to pause periodically to allow time for plastic conversion. The HTL reactor described in the previous section usually satisfies this assumption. At the high end of the plastic loading range shown in Table 2, the ship will need two to three reactors to match the conversion rate and collection rate, which should be easy to manage.
Table 2 depicts an optimistic picture of GPGP cleanup using onboard HTL conversion, at least if the plastic concentration exists at the higher end of the current estimated range. On the other hand, if the plastic concentration is at the lower end of the estimated range, the cleanup effort is daunting considering the total amount of plastic in the current GPGP; the effects of continuous accumulation make these estimates even less attractive. Likewise, the size of the ship or the rated power of the engine must be carefully selected and its speed controlled in order to make the process thermodynamically beneficial. The influence of the plastic composition in GPGP on the thermodynamics of the HTL process adds additional uncertainty.
Based on the above considerations, it is expected that the use of HTL to convert marine plastic waste into fuel on a ship will generate enough exergy energy to power itself, power the ship, and generate excess fuel for later use, when passing the fence in the circulation Used when pre-concentrating plastics-but not in open oceans or lack of prosperity in circulation. Unfortunately, using the results in Table 2 to calculate the cleanup time will produce margins or exceed the estimated value of the actual completion of a single ship (7 years to 340 years; see SI appendix, equations SI.1-SI.6 and Table SI .5). Table 2 is based on the boom-boom separation distance of 25 kilometers, which corresponds to the maximum number of booms deployed in the GPGP that a single ship can serve each year. Therefore, assuming that no new plastic enters the GPGP during the cleanup operation, use the framework used to generate estimates in Table 2 to evaluate the impact of increasing the number of fences and ships on the estimated GPGP cleanup time.
The net effect of reducing boom separation is to increase the number of booms in the GPGP, thereby reducing cleaning time. The distance between the boom and the boom must be large enough to prevent the upstream boom from negatively affecting the collection efficiency of the downstream boom. The minimum distance to avoid the shadow effect is currently unclear (24), which means that the boom-boom separation distance can be regarded as a parameter in the framework. For a given GPGP area, the distance between the boom and the boom controls the number of booms deployed, thereby controlling the amount of plastic accumulation, and thus the estimated cleaning time.
Figure 4 shows the estimated cleaning time as a function of the distance between the booms. Including the estimated values of the optimistic scenario and the conservative scenario, the difference lies in the current plastic concentration in the GPGP (1). Figure 4 shows that the time required to clear GPGP strongly depends on the boom-to-boom distance, and the estimated time is less than 1 year under the optimistic case of distance ≤10 km. However, Figure 4 also shows that a shorter cleaning time corresponds to a large number of booms deployed, and a trade-off between cost and practicality must be considered. Similarly, the interaction of prosperity and prosperity will definitely affect the performance of certain separation distances, and this effect should be included in future versions of the framework.
According to the high (2,500 g⋅km−2) and low (50 g⋅km−2) concentration estimates of the plastic in the GPGP (1) Estimate the time required to completely clear the GPGP, the deployment distance between the booms is 1 km 50 Kilometers and corresponding number of booms are deployed.
Thermodynamic properties, boom-boom distance, and boom-boom interaction all affect economic and environmental performance. The previous technical and economic analysis showed that a 110-ton/day HTL can produce fuel at a price of approximately US$4 per gallon of gasoline equivalent (25). Considering the fuel savings predicted by self-powered cleanliness, this cost is more competitive than commercial prices. force. It is more suitable for shipboard applications. Based on the published analysis of similar systems and the rigorous and random economic performance evaluation outlined in the SI appendix, the capital cost of a modular conversion system of an appropriate scale (10 tons/day) will be 2.85 million to 3.55 The operating cost is between 890,000 and 990,000 US dollars. These estimates do not include savings associated with reduced fuel purchases, which means that the marginal cost of the conversion system is therefore much lower than the estimated cost of the boom system and the ship itself (8).
The cleanup time estimated here is limited by the number of fences that can be deployed in the GPGP, whose magnitude is similar to the environmental degradation time scale (6), or thousands of fences need to be deployed. Considering the cleaning time and cost, it is best to intercept at a location with high plastic flux to prevent the plastic from reaching the patch in the first place (9). The estuary is a potential location where fencing and collection systems can be deployed more strategically than ocean circulation (9). Similarly, fence systems can be placed to protect particularly fragile or valuable ecosystems, such as breeding grounds. Alternative methods involving mobile collection technology are possible, but these options do not seem to reduce collection time sufficiently to justify the greater technical complexity. More detailed information is provided in the SI appendix.
Current analysis shows that collecting plastic that has entered the environment is at best part of a multi-level approach to reducing plastic damage to the environment. Other key elements include reducing the use of plastics (26), increasing the recyclability of plastics, or increasing the biodegradability of plastics (27). Finally, it should be pointed out that the implementation of plastic waste collection technology, whether based on HTL, pyrolysis or some other technical options, requires further research and a deeper understanding of the issues that follow, including inevitable leakage (28) And engine performance and emission characteristics (29), and the integrity of the fuel delivery system when operating with plastic-derived fuels (30).
This analysis focuses on HTL as a technically deployable method suitable for plastic conversion. However, the results presented here can be reasonably extended to other conversion methods. Enzymatic conversion (31, 32) has been the subject of several recent studies, and is particularly attractive for low-temperature conversion, thereby increasing thermodynamic efficiency and reducing the hazards to crew members. Since HTL has generated enough energy to power the process and ships, the thermodynamic advantages of enzymatic conversion will be gradual rather than transformative. Similarly, the enzymatic conversion rate is slower than HTL, which means that enzyme-based conversion reactions require larger reactor vessels than HTL (31). Therefore, the transformative effect of enzymatic conversion will be to degrade plastics into harmless products in the ocean without having to harvest them. The current analysis shows that it is necessary to study plastics in situ enzymatic decomposition into harmless products.
Converting ocean plastics into fuel will eventually release the carbon they contain as greenhouse gas emissions. In other words, the amount of CO2 released accounts for a small part of the global emissions budget, which is currently about 485 PgC (33). If the system runs continuously for 10 years, the total percentage will be less than 0.02% of the global carbon budget. On the other hand, converting plastic into fuel can eliminate new fossil emissions, while cleaning the ocean and reducing the amount of plastic recycled in land operations. Converting plastic into fuel also eliminates unnecessary congestion in ports and reduces the possibility of offshore oil spills. Unlike petroleum fuels, plastic-derived fuels have a low sulfur content, which means that their combustion does not release sulfur oxides (34, 35). In view of the importance of SO2 in the formation of pollutants (36) and new regulations, this It is an ideal result of the sulfur content in the fuel (37).
The results of this work provide a strong argument for continuing research to promote the current understanding of the factors that affect the depolymerization of ocean plastics in actual systems. Progress is needed in scientific knowledge of the thermodynamics, rate, and product distribution of marine plastics to reduce the uncertainty of the HTL approach. Similarly, the clean-up framework can be improved by filling in the gaps in current estimates of ocean plastic concentrations, quantities, and fluxes to reduce the uncertainty of deployment results. Finally, the results of this work show the huge challenges facing the prospect of cleaning up the ocean and believe that current plastic use and recycling strategies need to be changed. Therefore, the reasonable probability analysis and framework introduced here can be combined with risk analysis to provide profound and reliable information for future decisions and policy responses in this important area.
In short, the accumulation of waste plastics in the world's oceans is an urgent issue that requires corresponding attention. The cleanup work that relies on returning plastics to ports will consume a lot of fuel. The plastic itself contains chemical energy that can offset oil consumption and potentially reduce cleanup time by reducing the number of times the cleanup vessel must return to port to unload and refuel. This work shows that if the surface plastic concentration is greater than 12 vol%, HTL should be able to provide the exergy needed for self-powered cleaning. The estimated cleaning time depends on how quickly the plastic accumulates on the collection fence, and the cleaning time decreases with the number of fences deployed. Economically speaking, the HTL system is a moderate additional cost compared to the cleaning ship and boom system. Based on this promising analysis, future work can evaluate the effects of additives, contaminants and aging on HTL oil production and the suitability of plastic-derived fuels for existing fuel delivery systems and engines.
All the uncertain model input variables were modeled using appropriately selected uniform distributions in the absence of any accumulated operating experience and pertinent historical data. Perform 10,000 iterations.
The SI appendix, Table SI.1 contains a list of equipment and its energy requirements. Salt is removed from the feed to protect the reactor and other materials from corrosion. Use a reverse osmosis system (operating at 99% efficiency) to remove residual salt and reduce the salt concentration in the system to a level compatible with high-grade stainless steel (<1 wt%) (39). In order to increase the efficiency of depolymerization and allow feed to the reactor, the plastic must be shredded at the process entrance (40, 41). The SI appendix table SI.2 shows the reaction conditions met by the equipment of the plastic under study. The shredded plastic is then mixed with desalinated water to a solid content of 10% to 30% by weight, preheated in a heat exchanger, pumped to pressure, and heated to the reaction temperature. The stream leaving the reactor is cooled to form organic and water-rich products (11⇓-13), which are separated in a gravity separator. Residual organics in the water phase are removed in a hydrogen peroxide/ultraviolet (UV) oxidation system before recycling. The HTL product oil (see SI Appendix Table SI.3 for composition) is fed into the engine to generate power and operates at an efficiency of 35% to 40%.
The exergy of each sub-process is calculated separately, and then summed and normalized according to the quality of the plastic in the entrance.
The chemical exergy and combustion exergy of the flow is calculated as the weighted average of the Gibbs free energy of each product including water at standard temperature and pressure and steady state, and based on the enthalpy of combustion and their published yield The weighted averages, respectively. The thermodynamic data of pure plastics and products are taken from the literature (42⇓ ⇓ –45) and the National Institute of Standards and Technology (46), and the yield can be found in the relevant literature (11⇓ –13). When the thermodynamic values (enthalpy and entropy) of the product cannot be found, the correlation based on the carbon number is used to estimate them (see SI appendix, equations SI.8-SI.13). It is assumed that 100% of the oil formed by HTL can be burned directly without upgrading.
Use a 300 nm lamp to simulate the UV sub-process of removing organic matter. For the solubility value of polystyrene products in water, see relevant literature (47⇓⇓-50). More details of this calculation can be found in section SI.4 of the SI appendix.
All research data are included in the SI appendix or Digital WPI online (https://digital.wpi.edu/show/8c97kt62t).
We thank Geoffrey A. Tompsett for his thoughtful comments during the preparation of the manuscript. The US NSF supported this work as part of its 2026 Idea Machine program (Chemical, Biological Engineering, Environmental and Transportation Systems, Early Concept Exploratory Research Award #2032621). ERB's contribution is partly funded by the NSF Graduate Research Scholarship Program, grant number 2038257. Any opinions, findings, conclusions, or suggestions expressed in this material are those of the author and do not necessarily reflect the views of the author of the National Science Foundation.
Author contributions: research designed by MTT; research conducted by ERB and EI; NKK, CMR, and HK-P. Analyze the data; ERB, NKK, MTT, and DRH wrote this paper.
Reviewers: BRB, Ohio State University; JT, Cornell University; and KMVG, Ghent University.
The author declares no competing interests.
This article contains online support information https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2107250118/-/DCSupplemental.
This open access article is distributed under Creative Commons Attribution 4.0 (CC BY).
Thank you for your interest in advertising on PNAS.
Note: We only ask you to provide your email address so that the people you recommend the page to know that you want them to see it, and that it is not spam. We do not capture any email addresses.
Feedback privacy/legal
Copyright © 2021 National Academy of Sciences. Online ISSN 1091-6490. PNAS is a partner of CHORUS, COPE, CrossRef, ORCID and Research4Life.