ULTRA-COMPACT FUEL REFORMER FOR CONVENTIONAL AND RENEWABLE FUELS KW TO MW SCALE

Transcription

1 ULTRA-COMPACT FUEL REFORMER FOR CONVENTIONAL AND RENEWABLE FUELS KW TO MW SCALE Saurabh A. Vilekar, Christian Junaedi, Frank Nagy, Curtis Morgan, Dennis Walsh, Rich Mastanduno, Subir Roychoudhury Precision Combustion, Inc. 410 Sackett Point Road North Haven, CT ABSTRACT Commercial success of fuel cells as a viable alternative for clean energy demands robust, compact and dynamic reformer designs. Toward this goal, Precision Combustion, Inc. (PCI) is developing ultra-compact, lightweight, durable and high efficiency fuel processors offering performance and cost advantages over conventional reactors. The reformer core comprises of its patented Microlith substrate coated with proprietary catalyst formulations. Our fuel flexible Auto Thermal Reformer (ATR) has demonstrated stable, coke-free operation at low steam-to-carbon (S/C) ratios with complete fuel conversion and high reforming efficiency and sulfur tolerance to up to 400 ppm w fuel bound sulfur. This paper highlights the state of progress of our ATR along with recent development including balance of component (BOP) towards commercial viability for fuel processors from kw to MW scale. A brief overview of our compact, efficient and moderately contaminant tolerant steam reforming technology is also provided. 1. INTRODUCTION Lack of hydrogen storage and distribution infrastructure makes reforming hydrocarbon fuels to produce hydrogen for fuel cell stacks at point of usage a practical approach for operating fuel cells for many applications, such as remote off-the-grid locations, truck APUs, shipboard, etc. Robust reformers that can operate for thousands of hours with widely available fuels are required to realize this goal. However, these fuels often contain varying amounts of aromatics and sulfur which could lead to coking or poisoning of the reformers or fuel cell stacks. To avoid this, catalytic reformers are required to be sulfur tolerant and produce only trace amounts of higher hydrocarbons. Additionally, avoiding the need for make-up water by operating the reformers at low steam-to-carbon (S/C) ratios, thereby eliminating the need for large water condensers and back-up water supply, is a necessary attribute for deployment of practical APU systems. Precision Combustion, Inc. (PCI) continues to develop integrated reformer systems that are fuel flexible and sulfur tolerant for efficient H 2 production, while operating at low S/C ratios to achieve water neutrality. Three catalytic reforming technologies, namely autothermal reforming (ATR), catalytic waterless partial oxidation (dry CPOX), and steam reforming (CSR) are being developed and optimized for various applications. This paper will focus on the development and maturation of the integrated ATR system. Practicality and performance of water recovery systems will be addressed for water neutral operation using both the water condenser and the tail gas recycle approach. A brief overview of our moderately sulfur tolerant steam reformer technology is also presented. The evaluation of the shortterm and long-term reformer performance results from testing with various fuels provides a measure of the reformer performance that can be expected under realistic conditions with readily available fuels. The results also give valuable insights into the system design and operation strategy of an integrated fuel processor and APU system. 2. OVERVIEW OF MICROLITH TECHNOLOGY The reformer core comprises of a catalytically coated patented Microlith substrate (1) shown in Figure 1. The Microlith substrate consists of a series of short contact time, ultra-short-channel-length, metal meshes with very small channel diameters. While in a conventional honeycomb 1

2 monolith, a fully developed boundary layer is present over a considerable length of the device the very short channel length characteristic of the Microlith substrate minimizes boundary layer buildup, as illustrated by our CFD analysis reported elsewhere (2). This results in remarkably high heat and mass transfer coefficients between the bulk gas and catalyst surfaces compared to other substrates (e.g., monoliths, foams, pellets). The Microlith substrate also provides about three times higher geometric surface area over conventional reactors (e.g., monoliths) with equivalent volume and open frontal area. It packs more active surface area into a small volume, and thus provides more effective fuel conversion for a given pressure drop as well as a higher catalyst utilization. The very low thermal mass of each of the Microlith elements allows rapid heat-up along the reactor to steady operating temperatures. These unique features enable unusually high space velocity in mass transfer-limited operation, and offer the potential for higher selectivity and lower required loadings of precious metal catalysts. As a result, these reactors tend to be ultracompact, lightweight, highly efficient and have rapid transient response. The high sulfur tolerance to be discussed in later sections is a combination of the unique features of the Microlith substrate, effective and rapid temperature control along with our alumina supported Rh-based catalysts prepared according to PCI s proprietary formulations and washcoat/catalyst application procedures. Typical active metal surface area of >50 m 2 /g and metal dispersion of 20-25% are targeted. The catalyst formulations are characterized in-house with a BET apparatus (Quantachrome Autosorb 1-C). Fig. 1: Example of a Microlith substrate assembly. 3. AUTO THERMAL REFORMING OF LIQUID FUEL Current work with Diesel/JP-8 reformers has shown that autothermal reforming is the best low risk choice of technology for transportation applications. We have developed and demonstrated such a reactor that starts up in partial oxidation mode and transitions to ATR mode, operating at low S/C ratios. It uses a limited amount of water that can be recovered from the system exhaust and permits efficient fuel reformation without the deleterious effects of high temperatures and coke formation. Reformer testing with various fuels was performed in a test rig consisting of calibrated fuel and water supply pumps and mass flow controllers, a mass flow controller for air supply, and control and analysis systems for measuring catalyst bed temperatures and gas mixture composition. The control system also allowed for monitoring the inlet stream temperatures before entering the catalyst, the reformate stream temperature, and the catalyst surface temperature in several points along the catalyst bed. The reformate gas composition was measured using a gas chromatograph (GC) with thermal conductivity detectors (Agilent Series 3000 Micro-GC). The fuel conversion and reforming efficiency were determined from mass (i.e., carbon atom) balance calculation using the composition obtained from the GC with N 2 as the internal standard gas. In this study, the fuel conversion to C1 products is defined as the ratio of the total carbon atom present in the product stream as CO, CO 2, and CH 4 to the total carbon atom present in the feed fuel. The reforming efficiency is defined as the ratio between the lower heating values (LHVs) of H 2 and CO in the ATR reformate stream and the LHV of fuel entering the ATR reactor. As part of an effort to evaluate the maximum sulfur level that can be tolerated by the ATR catalyst, a set of 50-hour tests were performed using JP-8 with various levels of sulfur. In order to achieve a higher sulfur level, low-sulfur JP-8 (<5 ppm w S) was doped with dibenzothiophene (DBT) or benzothiophene (BT). JP-8 with 400 ppm w naturally occurring sulfur was also tested in the ATR. Figure 2 shows the fuel conversion to C1 products, the reforming efficiency, and the total H 2 and CO concentration in the reformate stream (mole %, wet basis) observed during these tests when the reactor temperature was maintained at the optimum temperature range by adjusting the O/C ratio (i.e., air flow rate) while keeping the S/C ratio at a constant value of The results show that the ATR catalyst was able to tolerate up to ~250 ppm w DBT doped sulfur JP-8, while maintaining fuel conversion and the reforming efficiency ~100% and 80%, respectively. The performance started to drop when operating the ATR reactor with 1000 ppm w DBT doped sulfur JP-8. Moreover, when comparing the results obtained from operating the ATR reactor with 400 ppm w naturally occurring sulfur JP-8 and with 400 ppm w doped sulfur JP-8 (i.e., 3.3 ppm w sulfur JP-8 doped with either DBT or BT), we concluded that DBT and BT as sulfur dopants are the worst-case scenarios for representing sulfur contamination on catalyst performance. This may be due to DBT and BT being more refractory to conversion to H 2 S versus the wider spectrum of naturally occurring organosulfur compounds. Finally, after the completion of the ATR testing with 1000 ppm w and 2000 ppm w DBT doped sulfur JP-8, the ATR reactor was operated with 2

3 100 Normalized Conv. (%) & Reform η (%), H2+CO mole % (wet basis) ppm 50 ppm 250 ppm 100 ppm 500 ppm 400 ppm "real S" 400 ppm (DBT) S 400 ppm (BT) S Normalized Conv. (%) Normalized Reform. Eff. (%) H2+CO H 2 (mole%, % (wet basis) 1000 ppm 2000 ppm 3.3 ppm 3.3 ppm 3000 ppm Time (hours) Fig. 2: Fuel conversion to C1 products, reforming efficiency (LHV-based), and total H 2 and CO concentration in the reformate stream (mole %, wet basis) as a function of time obtained from all of the 50-hour ATR tests using fuels with various sulfur levels. 3.3 ppm w sulfur JP-8. The results showed that the sulfur poisoning was easily reversible and the ATR was able to recover most of its performance after being exposed to fuels with high sulfur levels. Conv. (%), Efficiency (%) Fuel conversion (%) Reforming efficiency (%) Time (hrs) Fig. 3: Fuel conversion and reforming efficiency (LHVbased) from the 1000-hr test with 400 ppm w naturally occurring sulfur JP-8. After completion of the ATR reformer testing with varied sulfur levels, the ATR reformer with fresh catalyst was tested with 400 ppm w naturally occurring sulfur JP-8 for 1000 hours at an S/C of 0.90 and an average O/C of ~1.0. Figure 3 shows that stable performance was achieved with near complete fuel conversion to C1 products and 80% reforming efficiency (LHV-based) during the steady state. H 2 and CO productions were stable throughout the test. Additionally, average C2s (i.e., ethane and ethylene) and C3s (i.e., propane and propylene) concentrations of less than 20 ppm v and 10 ppm v (wet basis), respectively, were observed during the first 500 hours of the test. The C2 + C3 content increased to about 100 ppm v at the 1000 hour mark. The presence of organosulfur compounds in the liquid fuel thus showed no noticeable effect on the ATR performance during the 1000-hour test. The sulfur concentration in the reformate stream was monitored using a GC/FPD and showed near complete sulfur balance. All the fuel bound sulfur was converted predominantly to H 2 S with trace amounts of COS. Operating the reformer at low S/C ratios allows downstream gas phase desulfurization to less than 1 ppm v (suitable for fuel cells) using expendable or regenerable sorption bed to achieve the desired sulfur removal. This makes for a very simple reforming system avoiding any upstream liquid fuel desulfurization, techniques for which are not as mature as gas phase 3

4 desulfurization is. Continuous effort is on-going to further increase the sulfur tolerance of our reforming catalysts with improved durability. 3.1 Water Neutral Operation Two water recovery methods, namely anode gas exhaust recycle and condensation approach were evaluated and compared for water neutral operation. For the anode gas recycle (AGR) method, the ATR reformer was operated with a simulated AGR mixture composition for 50-hour at S/C of 0.7 and AGR mixture temperature of 400 C. The O/C ratio was kept constant at ~1.24 to maintain the optimum reactor peak temperature throughout the test. The temperature profile of the catalyst bed was stable during the 50-hr test without any temperature excursions or catalyst deactivation due to sintering and coking. Near equilibrium product distribution was achieved when compared to model predictions using ASPEN process simulation. An alternative approach to water recovery using a condenser unit was also demonstrated. Tests were performed to evaluate the maximum achievable water recovery from anode exhaust at different ambient/cooling air conditions and to compare experimental results with the predicted values. Several 3-hr tests were carried out to develop performance maps by varying the anode gas mixture temperature (i.e., C) and the ambient/cooling air temperature (i.e., C). At the end of each test, the water recovery (i.e., the ratio of water collected to the amount of water input) was calculated and the water mass balance was validated. In these tests, the water content in the simulated gas mixture was fixed at 32 mole%. Test results indicated that as the ambient/cooling air was increased from 20 C to 50 C, the water recovery decreased from 93% to 65%. The experimental values were only slightly below the maximum theoretical recovery. ASPEN modeling study of the integrated ATR-SOFC system operating at 75% fuel utilization indicated that 60% water recovery would be sufficient to provide the required amount of water for ATR operation at 5 kw th with Tier II diesel. Therefore, test results validate the feasibility of water recovery via condensation approach for water neutral operation of the ATR reformer. S/C ratio of The O/C ratio was appropriately varied to maintain the desired peak temperature. Stable performance was achieved with complete fuel conversion to C1 products. No higher hydrocarbons (e.g., C2s, C3s) were detected in the ATR reformate. The reforming efficiency (LHV-based) was around 75%. Effort is on-going to perform long term durability testing with different biodiesel blends. Fig. 4: Fuel conversion and reforming efficiency (LHVbased) from the 28-hr test with B50 blend of biodiesel. 3.3 Balance Of Plant (BOP) Development Air in Water in Fuel Flexibility The ATR reformer described herein is capable of operation with a multitude of fuels ranging from liquid (e.g., diesel, gasoline, S8, biofuels, methanol, military logistics fuels including JP-8, and Jet A) to gaseous fuels (e.g., natural gas). Figure 4 shows one such example where the ATR reformer was operated with B50 blend of biodiesel for ~28 hour at Reformate out Ø5.0 Fig. 5: Rendering, photo of thermally integrated 5 kw th ATR + Fuel Injector + Steam generator + Igniter + Sulfur trap. 4

5 To develop integrated fuel processor systems and transition the breadboard design to field applications, it is critical to develop low parasitic BOP, including pumps, blowers, heat exchangers, steam generators, condensers, injectors, igniters and controls. Toward this goal, we have developed integrated auto thermal fuel reforming systems. Figure 5 shows a 5 kw th integrated ATR reformer system that was designed, fabricated, and assembled at PCI. This fuel reformer system consisted of ATR core reformer, fuel injector, igniter, steam generator, and sulfur removal unit, and was fully instrumented with multiple thermocouples, pressure gauges, and gas sampling ports. A 2 L soda bottle is shown next to the prototype to reference the size. The flow through the ATR is indicated by arrows. Water, fuel and air (all at ambient temperatures at startup) are fed to the fuel processor. The ATR is ignited via glow plug and started with ambient fuel and air in CPOX mode. Light-off of the reactor starts at the front of the bed and rapidly propagates towards the back. The heat from the CPOX operation is used to raise steam from the ambient temperature water. Once the steam was available, it is mixed with the preheated air upstream of the reactor and added to the inlet mix to transition the reformer to the ATR mode. This fuel processor has been mated with multiple SOFC stacks and found to permit satisfactory stack operation. The other necessary BOP components (e.g., pumps, blowers, sensors, data acquisition with active feedback control, transducers, etc) are packaged separately in a compact module to account for good flow and heat integration. The integrated fuel processor is capable of auto-start, operating part-load with control logic/algorithm implemented via PC-based interface. The flexible architecture of the Microlith substrates results in a highly modular and scalable reformer design ranging in size from D-cell batteries for 1 kw e fuel cells to 1 gallonsize for 250 kw e fuel cells. We have thus developed integrated fuel processing systems directed to fuel cells ranging in size from 1-10 kw e for mobile applications and up to 250 kw e for shipboard fuel cell auxiliary power units (APUs). One such hardware example is shown in Figure 6 of a 250 kw e fuel processing system consisting of fuel/air/steam injector, ATR, steam generator HEX, and downstream sulfur cleanup. This fuel processor has been tested for multiple hours showing uniform flow distribution, no local hot spots, good temperature control, and equilibrium limited product distribution. 4. BRIEF OVERVIEW OF STEAM REFORMER Our compact heat-integrated steam reformer for syngas generation utilizing low-sulfur fuels is based on tubular reactor design consisting of both the burner and the endothermic steam reformer. The key design novelty is the high heat flux between the closely integrated endothermic and exothermic zones (facilitated by the use of high surface, low pressure drop Microlith substrates) and the flameless catalytic oxidizer instead of flame-stabilized burner. Flameless catalytic burner avoids high flame temperatures, increases thermal uniformity, distribution, durability and control. 3.4 ATR Reformer Scale-up Fig. 7: Heat integrated 1kW e catalytic steam reformer (includes exothermic and endothermic zones). The solid model of the fuel steam reforming prototype developed at PCI is shown in Figure 7. Additional components such as fuel, water pumps and controller are separately packaged, resulting in a manageable and compact system. The technology is scalable into the low MW throughput range. These reformers can be used to generate SOFC-quality reformate or can be coupled with a water-gasshift and CO clean-up reactor train to generate PEM-quality reformate. Integration with a pressure swing adsorber or Fig. 6: 250 kw e integrated fuel processor. 5

6 hydrogen membrane separator will result in production of high purity hydrogen. The heat integrated steam reformer was tested with gas-toliquid (GTL) fuel containing ~2-5 ppm w sulfur and n- dodecane for 500 hours at S/C ratio of 3-4, 1-3 atm pressure and 1-3 kw th fuel input without replacing the catalyst or implementing any regeneration protocol. Stable performance with complete fuel conversion and equilibrium limited product distribution was achieved. Figure 8 shows that stable hydrogen (mole%, dry basis) production was observed over the 500 hour durability testing. Table 1 shows that the experimental product distribution was in reasonable agreement with equilibrium prediction. Short term testing of the steam reformer with methane showed complete fuel conversion to C1 products (Figure 9) with high reforming efficiency. Fig. 9: CH 4 conversion and reformate composition from short-term testing with steam reformer at 1.1 kw th. 5. CONCLUSION Fig. 8: H 2 concentration (mole%, dry basis) in the steam reformer reformate with ~2-5 ppm w sulfur GTL fuel. TABLE 1: STEAM REFORMER EXPERIMENTAL DATA AT 1 ATM VS. THERMODYNAMIC EQUILIBRIUM Exptl. Product Mol %, S/C=3.0, P = 1 atm Equilibrium Mol. %, S/C=3.0, P = 1 atm H CO CO CH LHVbased Eff. (w. CH 4 ) ~119% 115% Reforming sulfur containing fuels to generate fuel (i.e., H 2 and CO) for fuel cells is a practical means for operating fuel cells. We have designed and developed a novel Microlithbased ATR fuel processor system for syngas production that is sulfur tolerant, fuel flexible, and water neutral. The ATR was operated for 1000 hours with 400 ppm w sulfur JP-8 with near complete fuel conversion and reforming efficiency (LHV-based) of ~80%. Limited testing has shown sulfur tolerance up to 3000 ppm w sulfur JP-8. Additionally, a stand-alone syngas generation system, consisting of ATR reformer and complete BOP components, has been developed and tested with satisfactory results for several hundred hours. Our steam reformer utilizes catalytic burner technology for improved control and enhanced durability. Stable equilibrium limited performance was achieved for 500 hours with low sulfur GTL fuel. Moreover, these fuel processors are fuel flexible capable of being operated with liquid and gaseous fuels. Being able to operate the fuel reformer with readily available fuels consisting of moderate sulfur levels can increase the versatility and operating range of the APU systems. 6. ACKNOWLEDGMENT This work was supported by the U.S. Department of Defense (DoD) and by the U.S. Department of Energy 6

7 (DOE). The authors are also grateful to the technical and engineering support groups at PCI. Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of the DoD and DOE. 7. REFERENCES (1) Pfefferle William C., Microlith catalytic reaction system: U.S. Patent 5,051, Sep 24 (2) Zhu Tianli, Walsh Dennis, DesJardins Joel, Roychoudhury Subir, Holcomb Franklin H, Feickert Carl A., Microlith supported sulfur tolerant catalyst for autothermal reforming of E85 fuel: Journal of Fuel Cell Science and Technology 7,