Calibration and Performance of a Selective Catalytic Reduction ( SCR ) Bench Rig for

A laboratory test rig was designed and built to easily test SCR (Selective Catalytic Reduction) technology. Equipped with three 6 kW heaters, connections for liquid N2 and an assortment of test gases, and a connection with an MKS NOx Analyzer, the rig allows for a vast range of SCR test conditions, and can easily be adapted for degreening, aging, and the implementation of other technologies onto the rig. To calibrate the test rig, baseline parameterization of a Cu-based zeolite (aluminosilicate) core was performed, including a temperature sweep and an NO/NO2 concentration sweep, while always maintaining a 1:1 ratio between NOx and NH3 reductant. The catalyst was found to have a peak deNOx efficiency of 99.8% between 250 'C and 300 'C, and maintained 98% efficiency with NO/NO2 < 1, while minimizing pollutant N20 generation. Thesis Supervisor: Victor W. Wong Title: Principal Research Scientist, Manager of Sloan Automotive Laboratory Introduction As it becomes more apparent that greenhouse gases are contributing to global warming, reducing the emission of gases from vehicle exhaust becomes paramount. In the automotive industry, NOx gases produced in most combustion processes as a result of elevated temperature have been shown to contribute significantly to the greenhouse effect. Originally designed for use in heavy industry [6], selective catalytic reduction (SCR) provides a means of converting NOx gases to harmless N2 and H20 components through the use of a reductant. In the last decade, adapting SCR technology for use in the diesel automotive industry has been an important topic of research; several technical barriers prohibit the direct transplant of the technology. In industrial SCR applications anhydrous or aqueous NH 3 is used, both of which require large storage vessels that would be difficult to reduce to the appropriate sizes for automotive applications. Additionally, NH3 or in solution can be extremely toxic and corrosive [4]; therefore, an alternate reductant which can be easily stored on a vehicle platform, can be metered out with relative ease, and most importantly, can be reduced to NH3 in situ to be used as the main reductant of NOx gases is needed. Most studies focus on the use of urea as this reductant. Chemistry of urea decomposition and NOx reduction To obtain the NH3 necessary for the catalytic reduction of NOx, urea decomposes as shown in the following reactions: (NH2)CO NH3 + HNCO (1) HNCO + H20 --+ NH3 + CO2 (2) A necessary condition for these reactions to go to completion is a sufficiently elevated temperature of the exhaust gas [3], as inefficiency results in unhydrolized urea depositing itself on the catalyst surface, blocking reaction sites and decreasing the efficiency of the overall catalytic activity. A major part of catalyst design for SCR applications is determining the temperature ranges at which the urea is fully decomposed, and avoiding the passing of excess uncatalyzed reductant, or slip, of both urea and NH3 through the catalyst. With this in mind, consider the following reactions of NH3 and NOx gases: 4NH3 + 4NO + 02 4N2 + 6 H20 (3) 2NH3 + NO + NO2 -2N2 + 3H20 (4) 8NH3 + 6NO 2 --+ 7N 2 + 12H20 (5) The rate of these reactions and their prevalence are determined in large part by the ratio of NO2/NO gases in the exhaust stream. Reaction (3) dominates when no NO2 is present. Reaction (4) dominates when NO2/NO < 1 and is a fast reaction, while reaction (5) dominates when N0 2/NO > 1, and is a slow reaction. Secondary, pollutant-forming reactions (or reactions that prevent NOx from decomposing to N2 and H20) can also occur: 4NH 3 + 4 0 2 -+ 2N 20 + 6 H20 (6) 2NH 3 + 2NO2 N20 + N2 + 3H20 (7) 4NH 3 + 4 NO + 3 0 2 -4N20 + 6 H20 (8) 4NH3 + 302 2N 2 + 6H20 (9) 4NH3 + 502 4NO + 6 H20 (10) 4NH3 + 702 4N0 2 + 6H20 (11) The actual chemistry on the catalytic surface is complex and can involve the evolution of phosphates, sulfates, and other species in the catalyzed exhaust both from the urea itself and from any impurities in the exhaust of a real engine [7]. However, because the metering of urea into an exhaust stream requires bulky, closed loop controlled equipment, simulated exhaust gas is often used and NH 3 is injected into this mixture to establish the baseline behavior of new catalysts. Thus, the above formulae provide a useful theoretical basis for analyzing the baseline behavior of catalysts tested in this manner. Test Rig Specifications Much of the work of this thesis focuses on the creation of a practical test rig for the testing of the baseline performance parameters of SCR catalysts. As shown in Figure 1, the developed rig is well suited for this goal, but can also be easily adapted to run different types of catalyst de-greening procedures. Fig 1. Catalyst bench rig schematic [8] Because of the high space velocities that many catalyst testing procedures implement (i.e. above 30k hr -1), a dedicated, insulated input line is used to draw nitrogen gas from a 250 L liquid N2 dewar pressurized at 230 psig. This input line runs through a 6 kW Sylvania heater, which operates at 208 V and can draw up to 60 amperes of current. This heater is used to heat the N2 gas to about room temperature before it enters the exhaust gas-mixing manifold. Five input ports are used to feed cylinder gases into the mixing manifold, and both these ports and the insulated line have flow meters used to control the volume of gas per unit time being mixed into the system. Just downstream of the mixing manifold are two additional 6 kW Sylvania heaters in series, which are used to heat the exhaust mixture to the desired testing temperature. An Athena Series 16C and an Extech 48VTR temperature controller are used in conjunction with a power relay to control the heaters, with thermocouples placed just downstream of the N2 input line and just upstream of the catalyst section. An additional thermocouple was placed just downstream of the catalyst section, but was not used for temperature control of the es~r~~ ~_Ea~B exhaust gas. The catalyst section itself has sampling ports controlled through ball valves just upstream and downstream of the position of the catalyst. These sampling ports lead to an MKS Multigas 2030 FTIR (Fourier transform infrared spectroscopy) NOx analyzer capable of discerning all major exhaust gas species (with the exception of diatomic species, although for 02 there is an built-in dedicated sensor) at a maximum rate of one sample per second. Once exhaust goes through the catalyst section, it is carried through a gate valve that is used to generate a pressure differential to ensure proper gas flow, and subsequently through an exhaust port that removes the gases from the system. Experimental Catalyst preparation: For the initial test rig experiment, a proprietary copperbased formulation on a zeolite (hydrated aluminosilicate) matrix was obtained through suppliers at Ford. From the original sample, a 2.5" diameter, 5" long piece was cored with a diamond bit. The catalyst was characterized by 400 cpsi (cells per square inch). Testing parameters and procedures: The base exhaust mixture used in all experiments was 14% 02, 1500 ppm NH3, and the balance in N2. Before each test procedure, the system was purged with pure N2 gas at 1500 C for 20 minutes. Two distinct experiments were carried out. To analyze NOx conversion as a function of catalyst inlet temperature, a temperature sweep scheme was performed. Outlet composition data was taken until the outlet compositions reached a steady state, starting at 150 oC catalyst inlet temperature and increasing the temperature by 50 'C for each data point, up to 500 'C. With each successive temperature increment, data was also taken from the catalyst inlet to ensure that the desired gas composition of 750 ppm NO and 750 ppm NO 2 was maintained at steady state. The second procedure focused on the effects of NOx composition itself. Keeping the catalyst inlet temperature at a nominal 300 oC, 1500 ppm of NOx was mixed into the exhaust stream at a ratio of 0.5 NO/NO 2. After recording catalyst outlet data at steady state, the ratio of NO/NO 2 was increased with steady state data taken for each additional concentration as shown in Table 1, always maintaining the total NOx levels at 1500 ppm. As in the first procedure, catalyst inlet data was taken between each data point to ensure that the desired exhaust gas composition was maintained. NO/NO 2 [NO] ppm [NO 2] ppm 0.5 50