Hydrogen Production by High Density Polyethylene Steam

Article pubs.acs.org/IECR

Hydrogen Production by High Density Polyethylene Steam Gasification and In-Line Volatile Reforming Gartzen Lopez, Aitziber Erkiaga, Maite Artetxe, Maider Amutio, Javier Bilbao, and Martin Olazar* Department of Chemical Engineering University of the Basque Country UPV/EHU, P.O. Box 644 - E48080 Bilbao, Spain ABSTRACT: Steam gasification (900 °C) of high density polyethylene (HDPE) in a conical spouted bed reactor (CSBR) followed by steam reforming in a fixed bed reactor (Ni commercial catalyst) has been carried out. The effect reforming temperature in the 600−700 °C range has on product yields and gas composition as well as on the amount and nature of the coke deposited on the catalyst has been studied. The reforming leads to a full transformation of C2+ hydrocarbons and tar. The maximum H2 yield is obtained at 700 °C (36 wt % by mass unit of HDPE in the feed, which accounts for 83 wt % of the maximum H2 yield according to stoichiometry), with CH4 conversion being 94% and the coke amount deposited on the Ni catalyst being minimum (3.3 wt % by mass unit of HDPE in the feed). The syngas obtained by reforming at 700 °C has a H2/CO volumetric ratio of 6 and is suitable for hydrogenation reactions and further valorization to produce H2.

1. INTRODUCTION World production of plastics in 2013 was 299 million tons (57 million tons in Europe), with an annual growth of 3.9% from 2012,1 which is a consequence of their use in numerous applications (packaging, building and construction, automotion, electricity and electronics, agriculture, consumer and household appliances) in order to maintain or improve our lifestyle. Nevertheless, plastics also bring about a large number of environmental hazards. In 2012, 25.2 million tons of postconsumer plastic wastes ended up in the waste upstream, with 26% being recovered though recycling and 36% by means of energy recovery processes, but 38% still went to landfill.1 One of the most serious environmental problems associated with plastic landfill is the accumulation of waste plastics in oceans, which accounts for more than 4.8 million tons in 2012.2 Accordingly, there is a growing interest by institutions to promote the implementation of recycling technologies, minimizing the landfill of waste plastics and avoiding the emissions in energy recovery processes. In addition, given that plastics are petroleum-derived materials, its recycling contributes not only to preserve the environment but also to intensify petroleum valorisation. Thermochemical routes (pyrolysis and gasification) have been considered to be the most interesting recycling routes for plastic wastes in order to recover monomers and obtain fuels or high value-added chemicals, and they have good perspectives concerning scaling up.3−5 Thermal or catalytic plastic pyrolysis has been widely studied using several reactor designs and catalysts.6−11 Plastic waste gasification studies are scarce; however this alternative has the advantage of being a more flexible technology, allowing to treat jointly plastic materials of different nature12,13 or plastic materials with biomass or coal.14,15 The syngas obtained may be directly used as fuel in fuel cells or gas turbines,16 or may be used to obtain liquid fuels (via Fischer−Tropsch or methanol synthesis) or hydrogen, which is considered a renewable and sustainable energy source in the long term.17−19 Plastic thermal treatments, such as pyrolysis or gasification, have been widely studied in fluidized bed reactors,7,12,14,20,21 © 2015 American Chemical Society

whose main feature lies in their high heat and mass transfer rates. Nevertheless, the sticky nature of plastic materials may cause agglomeration of particles by the fused plastic and subsequent bed defluidization. The cyclic solid circulation pattern characteristic of the conical spouted bed reactor (CSBR) gives way to the uniform coating of individual sand particles with fused plastic, which favors heat and mass transfer between phases22 and so bed isothermicity. Furthermore, collisions between particles and a high gas velocity in the spout region avoid the formation of solid particle-fused plastic aggregates and subsequent bed defluidization.23 The low segregation is another interesting feature of the CSBR,24 which allows using a catalyst in situ.25 In a previous paper,26 the good performance of the CSBR has been proven in the steam gasification of HDPE, that is, the limitations involving the physical steps prior to devolatilization as well as secondary reactions are minimized. Although the CSBR has the aforementioned advantages compared to other technologies, considerable tar yield has been obtained due to the reduced residence time, which is a serious problem for gas valorisation as fuel or synthesis gas (for the production of hydrocarbons, methanol or dimethyl ether).27 Erkiaga et al.26 used olivine and γ-Al2O3 in situ as a primary measure for tar elimination, but this strategy has a limited effect on tar reduction. In spite of the positive effect observed when temperature and S/P were increased, the yield of tar was still high (1.63 wt %) at 900 °C and with a steam/plastic (S/P) ratio of 2. The tar is mainly formed by the most stable nonsubstituted aromatic hydrocarbons, with benzene and naphthalene being the main compounds. Therefore, reforming of the volatiles is required in order to avoid the problems associated with the tar in the upgrading of the gaseous stream for energy or syngas.28 Received: Revised: Accepted: Published: 9536

July 6, 2015 September 11, 2015 September 16, 2015 September 16, 2015 DOI: 10.1021/acs.iecr.5b02413 Ind. Eng. Chem. Res. 2015, 54, 9536−9544

Article

Industrial & Engineering Chemistry Research

ground and sieved to 0.7−2 mm. Moreover, previously to reforming reaction the catalyst has been subjected to an in situ reduction process at 710 °C for 4 h under a 10% vol. H2 stream. Table 1 shows the NiO content of the catalyst and its physical properties, which have been measured by N 2

The objectives of this study lie in the minimization of the tar yield obtained by continuous HDPE gasification and production of a stream rich in H2. Therefore, a two-step reaction system has been proposed (Figure 1): (i) steam

Table 1. Metal Content and Physical Properties of the Catalyst

a

catalyst

NiO content, wt %

SBET, m2g−1

Vporous, cm3g−1

dporous, Å

G90LDP

14a

19

0.04

122

Provided by the supplier.

adsorption−desorption (Micromeritics ASAP 2010) (Figure 2). As observed, it is a mesoporous material, with an average Figure 1. Diagram of the two-step process.

gasification of HDPE in a conical spouted bed reactor, which is especially suitable for avoiding the limitations involving the physical steps; (ii) in-line catalytic steam reforming of the volatiles in a fixed bed reactor on a Ni catalyst. The use of two reactors in-line increases the efficiency of the catalyst for tar reduction compared to in situ treatment, given that the optimum operating conditions (temperature, residence/space time) may be used in each reactor and the whole catalytic bed is available for reforming the tar contained in the stream leaving the gasification reactor. Ni catalysts are the most used reforming catalysts, mainly because of their lower cost than noble metals (Pd, Pt, Ru, and Rh), which are more active for breaking C−C bonds and more stable.29 Furthermore, Ni catalysts are more active for water− gas shift (WGS) reaction at low temperature.30 Accordingly, they have been used in the reforming of the volatile products obtained by plastic pyrolysis by changing their composition (Ni content, supports and promoters) in order to increase the yield and selectivity to H2 and their stability by minimizing coke deposition.31−35

Figure 2. N2 adsorption−desorption isotherm of the catalyst.

pore diameter of 122 Å and a hysteresis characteristic of mesoporous materials in the N2 adsorption−desorption isotherm. The values of BET surface area and pore volume are rather low. Figure 3 shows the temperature-programmed reduction (TPR) curve of the catalyst measured by AutoChem II 2920 Micromeritics, which enables to establish the catalyst reduction temperature prior to use. The TPR shows a main peak at 550 °C associated with NiO reduction, which interacts with αAl2O3. Besides, another peak is observed at 700 °C, which is probably related to NiAl2O4 according to the composition

2. EXPERIMENTAL SECTION 2.1. Materials. The high density polyethylene (HDPE) was provided by Dow Chemical (Tarragona, Spain) in the form of chippings (4 mm), with the following properties: average molecular weight, 46.2 kg mol−1; polydispersity, 2.89 and density, 940 kg m−3. The carbon content in the polymer is approximately 86% and that of hydrogen 14%. The higher heating value, 43 MJ kg−1, has been determined by differential scanning calorimetry (Setaram TG-DSC-111) and isoperibolic bomb calorimetry (Parr 1356). The olivine used in the gasification reactor has been provided by Minelco (Lulea, Sweden). The material has been ground and sieved to the desired particle diameter, 0.35−0.4 mm. Moreover, it has been calcined at 900 °C for 10 h prior to use in order to enhance its reactivity in the tar cracking reaction. The surface properties have been measured by N2 adsorption− desorption (Micromeritics ASAP 2010). Calcined olivine has a limited porosity, with a surface of only 0.18 m2 g−1. The catalyst used in the reforming reactor (G90LDP catalyst) is a commercial reforming catalyst, with a chemical formulation based on NiO, CaAl2O3 and Al2O3, which has been provided by Süd Chemie. The catalyst has the form of perforated rings (19 × 16 mm) with a metallic phase of Ni supported on Al2O3, which is doped with Ca. This catalyst was

Figure 3. TPR curve of the catalyst. 9537

DOI: 10.1021/acs.iecr.5b02413 Ind. Eng. Chem. Res. 2015, 54, 9536−9544

Article


fine olivine particles entrained from the CSBR bed and the soot or char particles formed in the HDPE gasification. The plant is provided with a plastic feeding system, which consists of a vessel equipped with a vertical shaft connected to a piston placed below the material bed. The plastic is fed into the reactor (in the 0.2−5 g min−1 range) by raising the piston at the same time as the whole system is vibrated by an electric engine. A very low nitrogen flow rate introduced into the vessel stops the volatile stream entering the feeding vessel. The pipe that connects the feeding system with the reactor is cooled with tap water to avoid plastic melting and blocking the system. Water has been fed into the CSBR by means of a Gibson 307 pump, which allows a precise measuring of the flow rate, and has been vaporized by means of an 125 W electric cartridge placed inside the forced convection oven. Moreover, N2 is used as fluidizing agent during the heating process and its flow rate is controlled by a mass flow controller that allows feeding up to 20 L min−1. The gases formed in the reforming reactor circulate through a volatile condensation system consisting of a condenser, a Peltier cooler and a coalescence filter, which ensure the total condensation and retention of the tar. The Peltier cooler consists of a 150 mL tank and a refrigerator that lowers the temperature to around 2 °C, thereby efficiently condensing the volatile products. The noncondensable gas fraction leaves the Peltier through its upper part and the liquids are collected in the tank. The condenser is a double-shell tube cooled by tap water. 2.3. Experimental Conditions. The CSBR contains 70 g of olivine (particle diameter in the 0.35−0.4 mm range) and the fixed bed reactor 12.5 g of Ni catalysts (0.7−2 mm). Water flow rate is 3 mL min−1 in all the studied conditions, corresponding to a steam flow rate of 3.73 NL min−1, which is approximately 1.5 times that corresponding to the minimum spouting velocity.24 The runs have been carried out by continuously feeding 0.75 g min−1 of HDPE and using a steam/plastic ratio equal to 4. Previous experiments revealed that lower S/P ratios give way to severe operational problems of fixed bed blocking by coke formation. The gasification of HDPE (first step) has been carried out at 900 °C and the effect of temperature on steam reforming (second step) has been studied in the 600− 700 °C range. All the runs have been performed in continuous mode for several minutes in order to ensure the steady state in the process. Moreover, the runs have been repeated at least three times under the same conditions in order to guarantee reproducibility of the results. The amount of char formed in the gasification step (below 0.5 wt %) has been determined by weighing the mass retained in the cyclone and sintered steel filter, and its yield by monitoring also the amount of plastic fed into the reactor in each run. 2.4. Product Analysis. The volatile stream leaving the reforming reactor has been analyzed online by means of a GC Agilent 6890 provided with and HP-Pona column and a flame ionization detector (FID). The sample has been injected into the GC by means of a line thermostated at 280 °C, once the reforming reactor outlet stream has been diluted with an inert gas. The noncondensable gases have been analyzed in a micro GC (Varian 4900). The amount of the coke deposited on the catalyst has been measured by temperature-programmed oxidation (TPO) in a TGA Q5000TA thermobalance (Thermo Scientific) connected online to a Blazer Instruments mass Spectrometer (Thermo-

given by the provider. As observed, the catalyst is completely reduced at 900 °C, which is characteristic of α-Al2O3 supports. 2.3. Equipment and Reactors. Gasification and steam reforming runs have been carried out by continuously feeding HDPE in a bench scale plant whose scheme is shown in Figure 4. The main elements of the plant are the CSBR and the fixed

Figure 4. Scheme of the laboratory scale plant.

bed reactor arranged in-line. The design of the CSBR is based on previous hydrodynamic studies24 and on the application of this technology to the continuous pyrolysis and gasification of different solid wastes, such as biomass,36−38 plastics26,39,40 and waste tires. 41,42 In addition this technology has been successfully scaled up to 25 kg/h for biomass pyrolysis process.43 The dimensions of the CSBR are height of the conical section, 73 mm; diameter of the cylindrical section, 60.3 mm; angle of the conical section, 30°; diameter of the bed bottom, 12.5 mm, and diameter of the gas inlet, 7.6 mm.26 The reactor is placed inside a 1250 W radiant oven and the lower section of the reactor is a gas preheater filled with an inert ceramic material that increases the surface area for heat transfer, ensuring that steam is heated to the desired reaction temperature. Two K-type thermocouples are located inside the reactor, one in the bed annulus and the other one close to the wall. The fixed bed reactor is a cylindrical stainless steel vessel, with an internal diameter of 38.1 mm and a total length of 440 mm. This reactor is located inside a 550 W radiant oven which provides the energy needed to maintain the reaction temperature, which is controlled by a thermocouple placed in the catalyst bed. Both reactors are located within an oven which is in turn placed in a forced convection oven maintained at 270 °C to avoid the condensation of steam and the volatile stream formed in the gasification reactor, which is fed in-line to the reforming reactor. A high-efficiency cyclone and a sintered steel filter (5 μm) are located between the two reactors in order to retain the 9538


Article


the reforming step (84.27 g[water] 100 g[HDPE]−1 in the gasification step vs 150.90 g[water] 100 g[HDPE]−1 in the reforming step at 600 °C). This index has been determined based on the H2O mass flow rate in the feed and in the product stream, (mH2O)i and (mH2O)o, respectively, and on the HDPE mass flow rate fed into the reactor eq 4.

star) monitoring the masses corresponding to CO and CO2 in order to avoid the disturbance caused by Ni oxidation. The following procedure was applied: (i) signal stabilization at 100 °C with He stream for 30 min; (ii) a ramp of 5 °C min−1 to 800 °C under O2 diluted in He, finally this temperature was maintained for 30 min in order to ensure full coke combustion. Due to the nonhomogeneous distribution of the coke in the fixed bed, the sample for the TPO analysis was taken subsequent to homogeneously mixing the whole catalyst amount. In addition, the nature of the coke has been studied by TEM (transmission electron microscopy) images and SEM (scanning electron microscopy) imagines obtained by measn of Phillips CM200 and JEOL JSM-6400 devices, respectively.

X H 2O =

Table 2. Effect of Reforming Temperature on the Product Yields (g by 100 g of HDPE in the Feed) In the Steam Gasification-Reforming of HDPE; Gasification temperature, 900 °C Gasification (900 °C) − Steam reforming

(S/P = 1) H2 CO CO2 CH4 C2-C4 tar coke H2O reacted

16.4 108.3 19.6 16.1 20.2 1.6 84.3

(S/P = 4) 600 °C

650 °C

700 °C

27.8 44.8 149.2 10.7 0.0 0.0 17.8 150.9

34.3 64.3 178.9 4.8 0.0 0.0 5.8 187.8

36.4 82.1 170.5 1.0 0.0 0.0 3.3 192.2

of final products (by mass unit of HDPE in the feed). The yields of the compounds obtained in the gasification reactor (first step)26 are also set out in the table in order to show the efficiency of the reforming step. It should be noted that the gasification results shown in Table 2 were obtained with an S/P ratio of 1 and in the present study this ratio has been increased to 4, however a limited effect could be expected at high S/P ratios in the gasification step.16,26 As observed, the reforming step performs well, given that C2+ hydrocarbons and tar are completely reformed. Furthermore, it should be noted that the catalytic reforming has a significant effect on the composition of the gas obtained, increasing the yields of H2 and CO2 and decreasing those of CO and CH4, and therefore improving considerably gas composition with respect to gasification. Although the conversion of CH4 is not complete, reaches 94 wt % at 700 °C, that is, its yield decreases from 16.09 wt % to 1.03%. This significant effect of the reforming catalyst is due to the activation of the following reactions by Ni catalyst: steam reforming:(CH 2)n + nH 2O ⇒ 2nH 2 + nCO

(1)

dry reforming:(CH 2)n + nCO2 ⇒ nH 2 + 2nCO

(2)

water−gas shift reaction:H 2O + CO ⇔ H 2 + CO2

(3)

mHDPE

100

(4)

An increase in temperature from 600 to 650 °C gives way to a significant increase in the amount of water reacted. However, above 650 °C there is only a slight increase in the amount of water reacted and a slight decrease in the yield of CO2. This is due to WGS reaction thermodynamic limitation, which is a well-known fact at this temperature in the steam gasification of biomass.44,45 Furthermore, the decrease in the yield of CH4 as temperature is increased is an indication of the favorable effect of temperature on the reforming of CH4. Higher temperatures cause an increase in the yields of H2 and CO and the gasification of the coke deposited on the catalyst. Accordingly, the yields of H2 and CO reach the values of 82% and 36%, respectively, at 700 °C. As observed in Figure 5, the effect of temperature on the composition of the gaseous components is not as significant as

3. RESULTS 3.1. Products Yields and Concentrations. Table 2 shows the effect of temperature in the reforming reactor on the yields

Gasification

(m H2O)i − (m H2O)o

Figure 5. Effect of steam reforming temperature (second step) on gas composition. Gasification temperature (first step), 900 °C.

its effect on the yields of the components. Thus, the concentration of H2 increases slightly as temperature is increased, from 71% at 600 °C to 73% at 700 °C. The increase in the concentration of CO and the decrease in those of CO2 and CH 4 is of around 2%. It is remarkable that CH4 concentration at 700 °C is almost negligible due to the high CH4 reforming efficiency. The concentrations of CO and CO2 at this temperature are 12% and 15.5%, respectively. It is noteworthy that a high H2/CO molar ratio, 6.2, is obtained at 700 °C, which is much higher than the value of 2 recommended for standard syngas applications, such as the production of fuels or chemical products, and it should therefore be further corrected. In addition, a low CO2 concentration (15.5%) is obtained at 700 °C, which makes this syngas especially suitable for hydrogenation processes (in those applications in which CO2 and CO are not a problem)

As a result of the enhancement of reforming and WGS reactions, there is an increase in the amount of water reacted in 9539


Article


the reactor. This reveals that coke deposition is mainly caused by secondary reactions in the catalytic reforming. The high amount of coke deposited on the catalyst (mainly on the surface of the catalyst particles) blocked the fixed bed reactor subsequent to 20−60 min operation, depending on the reforming temperature (high temperature favored coke gasification and gives the possibility to operate for longer reaction times). Consequently, coke deposition hindered the study of the process for long reaction times. This problem may be solved in future studies using two parallel reactors, with one in operation while the catalyst is being regenerated in the other one. A fluidized bed reactor may also delay or avoid reactor blockage by coke deposition, as has been proven by different authors in the bio-oil reforming.52,53 Furthermore, the industrial implementation of the catalytic reforming would require a fluidized bed reactor with continuous catalyst feed and deactivated catalyst removal. A comparison of the aforementioned results with those in the literature shows that the maximum H2 yield obtained, 83 wt % at 700 °C, is similar to that obtained by Czernik et al.54 in the gasification-steam reforming (650−800 °C) of polyethylene (PE) in a microreactor and of polypropylene (PP) in two fluidized bed reactors, the first one at 600−700 °C and the second one at 850 °C, with H2 yield being 80 wt % in both cases. The same authors obtained a H2 yield of 56 wt % using steam/air mixtures as gasifying agent and burning 25 wt % of PP in the feed in order to attain autothermal process. 3.2. Coke Characterization. Based on the aforementioned results, the high coke deposition (with yields of 18, 6, and 3% at 600, 650, and 700 °C, respectively) is a determining factor for establishing the optimum conditions for the reactor design. This fact is a common feature of Ni catalysts used in the reforming of hydrocarbons and oxygenates, but the type of the coke seems to play an even more significant role on deactivation. Thus, the main responsible for catalyst deactivation seems to be the amorphous external coke (encapsulating coke), whereas the filamentous coke, generally the prevailing one, grows toward the outside of the catalyst particle without contributing to the catalyst deactivation, but causing operational problems in the reactor, such as gas flow blockage.51,55 Figure 7 and Figure 8 show SEM and TEM images of the coke deposited on the catalyst at the reforming temperatures of 600, 650, and 700 °C. As observed, an increase in temperature changes the structure of the coke deposited. At 600 °C (Figure 7a and Figure 8a), the amount of coke deposited is the highest, with its nature being filamentous and amorphous. At 650 °C (Figure 7b and Figure 8b), the coke deposited is more structured and more fibers are observed. At 700 °C (Figure 7c and Figure 8c), a less fibrous coke than at 650 °C is observed due to the partial gasification of coke precursors. Besides, as temperature is increased the fibers are shorter and thicker. Furthermore, Ni crystals are observed in the SEM images (Figure 8), which are swept throughout the filamentous coke growth. These results are similar to those obtained by Wu and Williams35 in the pyrolysis-reforming of PP on Ni−Al−Mg catalyst. In order to know the nature and location of the coke deposited, temperature-programmed oxidation (TPO) has been carried out (Figure 9). The peaks reported in the literature at low temperature (