Performance Studies and Correlation between Vehicle- and Rapid- Aged Commercial Lean NOx Trap Catalysts.
Carbon dioxide (C[O.sub.2]) is an important greenhouse gas emitted by natural processes and human activities. After power generation, burning fossil fuels to transport goods and people is the second largest contributor to global C[O.sub.2] emissions. Therefore, as a countermeasure to global warming, low-carbon dioxide emission vehicles have been investigated by promoting lean-burn engines and high combustion efficiencies. Nevertheless, the main constraint for this approach is related to NOx emission control. Thermal NOx is formed in the combustion process when nitrogen reacts in an environment of high concentration of oxygen and elevated temperatures. Hence, emission standards have been established to ensure air quality standards and protect the human health by setting quantitative limits on the permissible amount of specific air pollutants. The evolution of emission standards leads to more and more stringent emission limits, which requires the development of efficient NOx abatement materials, technologies and strategies for lean-burn vehicles.
Selective catalytic reduction (SCR) and lean NOx traps (LNTs) are currently the most popular NOx reduction technologies implemented by car manufacturers. High NOx conversion efficiencies can be achieved with the SCR system, however it requires that the temperature at the urea dosing position is above 180 [degrees]C, limiting the effectiveness of the system at low engine operating temperatures. This technology also has some cost and space requirements, such as storage tank, urea injection and control system, which limits its implementation in small vehicles. On the contrary, LNTs are a low cost and simple technology that do not require any additional devices and it has a better low-temperature NOx reduction performance compared to SCR system . LNTs store NOx emissions during the lean-burning phase. Subsequently, before NOx slip becomes significant, the nitrous oxides are released and reduced to [N.sub.2] over the precious metal during a rich combustion phase. The engine control unit is able to manage the lean/rich transitions by changing engine parameters such as fuel injection quantity and timing.
An efficient LNT is formulated with a NOx storage component, alkali or alkaline earth metals, and precious metals impregnated on a high surface area support. Particularly, the first generation of LNT was composed by noble metals, barium and alumina . Subsequently, several studies have been focused on the development of more efficient NOx abatement materials. It is well known that modern commercial LNT formulations generally contain noble metals, Ba, Ce, Zr, Al and many high temperature stabilizers [3, 4, 5, 6, 7, 8]. Platinum, palladium, and rhodium are the most common noble metals used and they are responsible for the catalytic oxidation and reduction of the exhaust emission species. Moreover, ceria is widely known as an oxygen storage support, however it can also contribute in the NOx storage-reduction process as a supplement to the Ba NOx storage component . Furthermore, ceria enhances the LNT sulfur tolerance due to its ability of trapping sulfur species, thus reducing their accumulation on the Ba NOx storage component . Lastly, the addition of zirconium to ceria is reported to inhibit the sintering of ceria and enhance the thermal durability of the catalyst .
Although lean NOx traps have been studied extensively, sulfur poisoning still remains one of the major challenges for this technology. Even with the improved sulfur tolerance and lower fuel sulfur levels, sulfur poisoning still leads to a significant reduction of NOx storage capacity and conversion efficiency of the LNT. As a consequence, LNTs have to perform periodic desulfurization strategies, in which the catalyst temperature exceeds 650 [degrees]C in order to recover its storage capacity . The high temperatures required increase the sintering of precious metals, and storage and support materials. In addition, the interactions between barium and [gamma]-[Al.sub.2][O.sub.3] support will decrease to form less efficient storage media [H, 12]. However, new formulations show significant improvements in terms of thermal stability to tolerate the desulfation conditions. In addition, desulfation strategies have been enhanced to achieve fast and efficient sulfur removal under real driving conditions [13, 14].
Moreover, the deactivation mechanisms of a particular catalyst depends on the operating conditions which it has been exposed to . For vehicles running mainly in urban environments, the catalyst might be subjected to several interrupted desulfation events due to short transients and low temperature working conditions. Thus, the poisons can accumulate and form chemical compounds with the active catalytic materials, reducing the catalyst effectiveness. On the contrary, for vehicles running in extra-urban driving conditions, the catalysts can be exposed to high temperatures for extended periods of time, thus high thermal degradation could be the primary cause of catalyst deactivation in this case . Both extreme working modes are considered by car manufacturers for the development of the catalysts in order to guarantee the emission performance in the projected life of the vehicle. Thus, long and expensive durability tests have to be performed to assess the catalyst operation after the aging process. Hence, the implementation of alternative methods are being studied by car manufacturers and legislators to reduce durability testing time and cost.
In fact, different accelerated aging procedures have been developed to replicate the catalyst deterioration under real-world operating conditions, achieving an equivalent emission performance in a reduced time. The processes can be performed on a road, vehicle chassis dynamometer, engine rig or in an oven to thermally stress the catalyst. Since the thermal stress procedures do not include all the effects of the normal real use (abnormal deposition of ash from engine-oil for instance ), they need to be properly correlated with on-road aged catalysts in order to have a meaningful demonstration of catalyst durability and fulfil the performance requirements. However, because of the dominant effect of the thermal aging, the most common durability evaluations are based on high temperatures aging and thermal stress methods. There is one study where Benramdhane et al.  examined thermal and accelerated vehicle aging on LNTs and observed similar impact on their functionalities. Ottinger et al.  studied the thermal aging of fully-formulated LNTs in the flow reactor at different aging temperatures and found different degradation mechanisms depending on the aging temperature. Moreover, Ruetten et al.  presented an alternative method of accelerating the catalyst aging by increasing the aging temperatures through modified burner technique.
However, to our knowledge, there are no investigations that focus on the deterioration of a 100000 km vehicle-aged LNT catalyst and studied its correlation with accelerated aging methods. Hence, this study describes Volvo Cars's continuing efforts to explore the deactivation mechanisms affecting the performance and durability of commercial lean NOx traps through engine bench and vehicle chassis dynamometer experiments. Moreover, this investigation intends to establish a proper correlation between vehicle-aged and rapid-aged LNT catalysts, which is critical for cost effectiveness when evaluating new catalyst formulations.
This study has been performed with three commercial LNT catalysts with the same formulation. The catalysts are supported on 1.5 l standard honeycomb metallic substrates with a cell density of 300 cpsi (cells per square inch). The washcoat is deposited on the substrate channels, composed by a modern commercial LNT formulation with 119 g/[ft.sup.3] of precious metal loading (Pt:Pd:Rh = 95:19:5, see also Table 1).
Throughout this investigation, the vehicle aging and oven aging methods are evaluated and compared between them. The vehicle aging method provides realistic conditions, however the process requires long durability testing time. Thus, the oven aging approach has been used as an accelerated aging method with the purpose of reducing the total duration of the test by thermally stressing the catalyst in a high temperature environment.
A LNT catalyst was aged for 100000 km in a vehicle chassis dynamometer exposed to real conditions in order to mimic the effect of sulfur and thermal exposure under real driving conditions. The aging was performed in a 2 liter Volvo XC90 diesel vehicle, using standard fuel with 10 ppm of sulfur content. The method consisted of firstly driving a sequence of Standard Road Cycles (SRC)  up to a total distance of 60000 km. This cycle can be used by car manufacturers in order to demonstrate the durability of the pollution control devices as it is established in the US and European emission legislations . However, this cycle might not replicate the real driver behavior as the Common Artemis Driving Cycle (CADC), being developed based on a statistical analysis of European real world driving patterns . There are some differences between SRC and CADC, such as the CADC cycle is shifted to lower operating temperatures compared to SRC, which might lead to a reduced thermal deterioration of the catalyst. Moreover, CADC incorporates more transient operating modes than the SRC which might lead to several interrupted desulfation events, which affect the chemical poisoning. In addition, real driving patterns from Volvo Cars users were analyzed, suggesting that a modified Artemis cycle mimic more precisely the real driving behavior. Therefore, the last 40000 km were driven following a modified Artemis cycle, which consisted of four driving schedules: two times the urban cycle, one rural road cycle and one motorway cycle. In addition, vehicle emission cycles were performed along the vehicle aging testing in order to follow the catalyst performance and degradation.
The oven-aged catalyst underwent an accelerated aging procedure corresponding to 160000 km by subjecting the catalyst to 800 [degrees]C during 24 hours in a furnace under a flow of air. The suitable temperature, gas exposure time and gas composition were obtained from the internal database of Volvo Cars.
Engine Bench Testing
In this investigation, fresh, vehicle-aged and oven-aged catalysts were evaluated in the engine bench by carrying out several performance tests. The systems were tested in close-couple position in a 2-liter Volvo diesel engine attached to an AVL dynamic dynamometer. The emission species were measured before and after the LNT brick by means of a Horiba MEXA-7500DEGR system. A quantum-cascade laser (QCL) was also used to detect ammonia emissions and other species after the lean NOx traps catalyst. The exhaust gases and substrate temperatures were measured along the catalyst axis to monitor the temperature patterns during the regeneration and NOx storage events, as it is illustrated in Figure 1.
The LNT temperature was calculated as the average substrate temperature measured by the thermocouples disposed along the LNT brick (see Equation 1).
[T.sub.LNT] = [[T.sub.3] + [T.sub.8] + [T.sub.10]]/3 (1)
Prior to testing, the fresh LNT sample was first degreened in the engine bench by performing several soot and sulfur regeneration events until the activity of the catalyst has been stabilized. This procedure was also implemented on the vehicle- and oven-aged samples before carrying out the performance tests. It was performed for the vehicle-aged catalyst to ensure the sulfur trapped removal, while it was carried out for the oven-aged to re-activate the catalyst after 24 hours in an oxidant environment. In this way, the variability of the final state of the samples is minimized after the different aging procedures.
Firstly, the catalyst performance has been evaluated in terms of NOx storage at different working conditions. The NOx storage test consisted of a DeNOx or NOx purge event followed by a lean period until the NOx conversion has decreased below 50 % of efficiency. The DeNOx event was composed by a sequence of rich periods of 15 seconds (X = 0.94) up to two lambda breakthrough were reached. The lambda breakthrough refers to the point in which the downstream air-to-fuel ratio intercepts the upstream air-to-fuel ratio of the catalyst. The engine operating points were chosen based on representative situations of real driving behavior. They include different inlet exhaust gas temperatures ranging from 180 [degrees]C to 450 [degrees]C and space velocities between 35000 [h.sup.-1] and 125000 [h.sup.-1]. The composition of the exhaust gases depends on the fuel injected and the gas mass flow required in order to reach the desired working condition. Figure 2 depicts the engine-out emissions and the LNT substrate temperature for four different space velocities. The water and oxygen concentrations are assumed to be similar over the different space velocities. THC and CO engine-out emissions decrease by increasing the operating temperature. However, NOx emissions rises as the LNT temperature increases, although the N[O.sub.2] / NOx ratio is reduced. Additionally, note that the engine-out gas composition vary in a limited range by keeping the LNT substrate temperature constant, hence it is assumed comparable among different space velocities.
Secondly, some representative working points were selected in order to evaluate the NOx purge capabilities of the LNT at different engine operating conditions. The test involved a series of cycles of rich and lean periods until the NOx conversion rate is stable as it is shown in Figure 3. Since the regenerated NOx quantity cannot be measured directly, it is assumed that the NOx stored amount is equal to the NOx purged quantity when the steady conversion rate is reached. Thus, the regenerated NOx amount is computed by the difference between NOx LNT-in and LNT-out during a full rich-lean cycle. In contrast to the previous test, the lambda value and the duration of rich periods vary and the lean periods were determined based on fixed periods of time.
It is observed that during rich periods hydrocarbons (THC) and carbon monoxide (CO) slips through the LNT. Particularly, for instance during cold start periods when the catalyst is less effective due to the relatively low temperature operation. Therefore, THC/CO light-off is important to investigate, especially its performance degradation with aging under different engine operating conditions. Note that THC/CO light-off represents the temperature at which the catalyst reaches 50 % of THC and CO conversion respectively. The test consisted of executing three different engine operating points with two fuel injection ramps of 15 minutes from 100 % to 0 % of THC and CO conversion efficiencies and vice-versa, as it is illustrated in Figure 4.
Vehicle Emissions Testing
The performance of the catalysts were also investigated in a vehicle chassis dynamometer by running some vehicle emission cycles, such as New European Driving Cycle (NEDC), Worldwide harmonized Light duty driving Test Cycle (WLTC) and Common Artemis Driving Cycle (CADC). They were conducted in a 2-liter Volvo XC90 diesel vehicle following the procedures and requirements established by the European legislation. The pre-conditioning consists of running the same driving cycle which was successively used for the emission tests. Before the emission tests were carried out, the vehicle remained between 8 to 24 hours in the soak area in order to cool down the engine and reach stable conditions.
In terms of emissions, the engine-out and post-catalyst pollutants were continuously measured with a Horiba MEXA-7100D system. The tailpipe diluted emissions were correspondingly collected in bags and analyzed after the cycle was finalized. Note that a Selective Catalytic Reduction (SCR) system was placed in underfloor position downstream of the LNT since the complete exhaust system was mounted in the vehicle for this test, therefore the tailpipe bag emissions were affected by the SCR performance as well. However, in this study the results are focused on the LNT.
RESULTS AND DISCUSSION
Durability Testing - Vehicle-Aged Catalyst
Figure 6. Aging time of the vehicle-aged catalyst exposed to high operating temperatures (above 450 [degrees]C). LNT temperature [[degrees]C] Artemis special front 0.50 3.96 9.49 3.62 0.60 0.16 0.03 0.00 Artemis special mid 0.44 0.48 5.97 9.49 2.00 0.15 0.00 0.00 Artemis special back 0.44 0.38 3.78 10.28 3.50 0.18 0.00 0.00 SRC front 1.08 1.74 5.73 7.17 4.65 1.46 0.17 0.00 SRC mid 1.07 0.84 3.04 7.22 7.22 2.80 0.33 0.01 SRC back 1.19 0.82 2.04 7.11 8.27 3.12 0.18 0.00 Note: Table made from bar graph.
The catalyst performance and degradation was monitored throughout the entire aftertreatment aging process by running periodically vehicle emission cycles. The lean NOx trap conversion efficiency as a function of the driving distance is presented in Figure 5. The results shown are the average NOx, THC and CO conversion efficiencies of a series of WLTC cycles carried out each 30000 km. The vehicle was equipped with the complete aftertreatment system; it means the LNT and Diesel Particulate Filter (DPF) in close-couple position, while the SCR was mounted downstream in underfloor position. It is important to mention that, due to a software update made at 10000 km after the vehicle emission cycles, the first point in Figure 5 is not taken into consideration for the analysis of this test. For this reason, an improvement in the LNT NOx conversion efficiency at 40000 km can be observed, since the engine calibration used was different compared to the first driving cycles.
The deterioration of the overall vehicle-aged catalyst activity is clearly seen in Figure 5 as the driving distance increases. The catalyst was exposed to slightly more than 36 hours of operating temperatures above 550 [degrees]C (see Figure 6) during whole test up to 100000 km, particularly during DPF regenerations and desulfation events, which might lead to loss of catalytic surface area due to precious metals sintering. In addition, the support is also subjected to sintering if the catalyst is exposed to high temperatures, resulting in decreased surface area . Moreover, the catalyst was subjected to sulfur species formed as by-products of the combustion of organic sulfur compounds present in the diesel fuel. The experiments in the study by Rohr et al.  confirm that S[O.sub.2] is the most abundant sulfur specie in the lean period. The S[O.sub.2] is thereafter oxidized over the noble metal sites and is subsequently absorbed on the base metal supports such as Ce[O.sub.2] and [Al.sub.2][O.sub.3] [22, 23]. Although desulfation events were periodically performed to recover the storage performance, some of the sulfur compounds may be difficult to remove, reducing the overall storage capacity and conversion efficiency of the LNT.
Moreover, the temperature profile of the aging cycles during the DPF regeneration and desulfation events are compared in Figure 6. It is observed that the back of the catalyst was exposed to higher temperatures compared to the front, this is expected to result in larger deterioration in the back of the catalyst. This is consistent with the results reported by Epling et al. . The background for this is that the temperature increases, by the exothermic reactions occurring between the reductants and the oxygen stored. The heat generated is then convectively and conductively transported . Thus, the combination of the exothermic reactions and the heat transported will progressively increase the gas temperature along the axial direction (relative to the gas flow) as it moves further down.
Engine Bench Testing
The engine bench allows the evaluation of specific properties of automotive catalysts by running the designed experiments under controlled conditions. Likewise, the catalyst is subjected to real working situations such as exhaust gas flow rates, compositions and temperatures. Hence, the performance of the fresh and aged catalysts were assessed in the engine bench by running the NOx storage capacity, NOx purge capabilities and THC/CO light-off experiments.
The NOx storage as a function of the LNT substrate temperature and space velocity is presented in Figure 7. It was determined by the cumulative difference between inlet and outlet NOx mass flow in the lean working period. A substantial degradation of the catalyst activity is obvious by comparing the fresh and aged catalysts. The deterioration ranges between 35 % and 85 %, and being more pronounced in conditions of low working temperatures. Similar trend is also observed in Figure 8, which shows the N[O.sub.2] /NOx ratio after the LNT brick. At temperatures between 200 [degrees]C and 350 [degrees]C, the NO to N[O.sub.2] conversion is significantly affected after exposing the catalyst to the aging methods. In the work by Toops et al. , it is observed that the NOx storage is significantly larger at low temperatures compared to high temperatures when N[O.sub.2] is used as an inlet gas. However, when NO is used, the NOx storage is limited at low temperatures , which is likely related to low rate of NO oxidation. Therefore, our results indicate that because the NO oxidation is significantly decreased for the aged catalysts (see Figure 8) and a similar trend is observed for the NOx storage (see Figure 7), one of the reasons for the reduced NOx trap efficiency could be due to the decreased NO oxidation activity of the aged catalysts.
Even though the oven-aged catalyst shows slightly larger NOx storage values compared to the vehicle-aged catalyst in Figure 7, there is no significant difference. This minor variance can be attributed to the exposure of the vehicle-aged catalyst to sulfur species in the exhaust stream, which could affect the NOx storage capacity of the catalyst as discussed previously. Likewise, the thermal load of both samples might also be slightly different because the oven-aged catalyst was treated for a period of time equivalent to a driving distance of 160000 km and this was done in dry conditions, while the vehicle-aged catalyst was exposed to high water levels at high temperatures. Although there are several reasons, the NOx storage performance of the oven-aged catalyst is well correlated to the normal real use of the vehicle-aged catalyst driven for 100000 km.
Furthermore, the NOx storage performance of the catalysts is also affected by the exhaust flow rate. Figure 9 presents the NOx storage for a fixed NOx traps efficiency as a function of the space velocity at 300 [degrees]C, which represents one of the operating temperatures with the greatest performance. The three catalysts show a reduction of 30 % to 35 % of total NOx storage by increasing the space velocity from 35000 [h.sup.-1] to 85000 [h.sup.-1]. The NO to N[O.sub.2] conversion as a function of the space velocity follows a similar trend as well, which can be seen in Figure 10. These results indicate that there is a correlation between the NO oxidation and NOx storage, as discussed in the previous section.
NOx Purge Capabilities
The results presented in this section summarize the trend observed in the NOx purge capabilities test carried out on 160 diverse engine operating points and under different lean/rich cycle timings and lambda conditions. The results shown were determined from the last lean/rich cycle when the NOx conversion rate was stable as described previously. The results in Figure 11 shows the cycling conducted at 2500 rpm and 15 mg/st ([T.sub.LNT] = 280 [degrees]C) using 280 s of lean and 5 - 8 s of rich period. The lambda value was also varied between 0.92 and 0.97. Under such conditions, the NOx conversion efficiency demonstrates that the fresh catalyst performs better than the aged samples and its performance improves significantly by increasing the duration of the rich periods. Moreover, the increase of the concentration of reductants in the exhaust stream also leads to greater NOx adsorption and reduction, however it seems to contribute in a lesser extent than the increase of the duration of rich periods. Similar effect of the increase of the reductant concentration on the NOx conversion is also shown in Figure 12, a cycling performed at 2500 rpm and 30 mg/st ([T.sub.LNT] = 450 [degrees]C) using 60 s lean and the same conditions concerning the lambda and rich periods variability. However, in this case, it is not observed any notable improvement of the NOx conversion by increasing the duration of the rich periods.
This trend was also observed by Di Giulio et al.  where they found that the rate of NOx release and reduction take place as a function of the operating temperature. The low working temperatures result in low rates for NOx release and reduction which can be compensated by the extended reduction periods , as it is the case in Figure 11.
Although the NOx storage test showed a similar performance and a comparable oxidation activity between the aged catalysts, the oven-aged catalyst presents a greater NOx conversion efficiency compared to the vehicle-aged catalyst. Ohtsuka  examined various noble metals activity for the selective catalytic reduction of N[O.sub.2], and observed a high reduction activity of N[O.sub.2] to [N.sub.2] for palladium and high oxidation activity of NO to N[O.sub.2] for platinum. Moreover, Gelin et al.  showed that the palladium is more pronounced to sulfur poisoning than platinum. Thus, we propose that because of the exposure of the vehicle-aged sample to sulfur species, the overall reduction activity has been reduced due to the deactivation of palladium sites by the formation of highly stable surface palladium sulfate species . Instead, the oxidation activity of NO to N[O.sub.2] remained unaffected by subjecting the vehicle-aged catalyst to sulfur compounds due to higher sulfur resistance of the platinum sites, keeping the NOx storage capacity at the same level as for the oven-aged.
Turning the focus on the N[H.sub.3] production, an increase in the formation of N[H.sub.3] is observed (see Figure 11 and Figure 12) as the length of the rich period and the reductant concentration increase. It is well known that N[H.sub.3] is formed over the noble metal sites under conditions of high [H.sub.2] / NO ratios (H / NO > 2.5 ), which is directly proportional to the reductant concentration for a certain operating temperature. Moreover, several studies have indicated that N[H.sub.3] is observed with a slight delay after the lean to rich transition [29, 30, 31, 32]. This is related to the [H.sub.2] consumption in the reaction front by the NOx and oxygen stored in the catalyst . This implies that extending the duration of the rich period leads to larger NOx conversion as mentioned previously, however the N[H.sub.3] conversion is reduced, which results in higher levels of ammonia slip . Additionally, the aged catalysts exhibited significantly larger net N[H.sub.3] production compared to the fresh sample. Because the amount of reductant were the same for all three samples during the rich period, the [H.sub.2] / NO ratio will depend on the reductant consumption. In fact, it has been reported extensively in the literature that the N[H.sub.3] formation decreases by increasing the NOx and oxygen stored in the catalyst [34, 35, 36]. Particularly, Ji et al.  studied the effect of aging on the NOx storage and regeneration, and mentioned two probable reasons concerning the influence of the oxygen storage capacity (OSC) in the LNT on the N[H.sub.3] production. Firstly, the OSC of the catalyst is reduced as it becomes aged . Since less reductant is consumed by the oxygen stored during the rich periods, the higher the concentration of [H.sub.2] will be in the gas front, promoting the N[H.sub.3] formation. Similarly, the second reason is based on the post-reaction of the N[H.sub.3] formed at the front of the catalyst, which reacts downstream with the stored oxygen. Since the OSC is reduced in aged catalysts, less N[H.sub.3] reacts at the back of the catalyst, increasing the N[H.sub.3] slip. Moreover, the NOx storage capacity decreases as the catalyst ages due to fewer NOx storage sites, which might result in an extended NOx storage-reduction zone [33, 36]. Therefore, larger amounts of N[H.sub.3] slips without being oxidized at the back of the catalyst since the increased length of the NOx storage-reduction zone reduces the downstream OSC zone .
Besides the NOx and oxygen stored, the dispersion of the precious metals might also play an important role on N[H.sub.3] formation . Before the stored NOx spills over onto the noble metals and reacts with [H.sub.2], it must diffuse to reach the interface between the noble metal and the storage compound. Lower dispersion of the precious metals leads to larger diffusion distance, which results in lower rates of NOx transported reducing the N[H.sub.3] production. When the diffusion of the NOx stored is slower than the [H.sub.2] feed rate, the noble metals becomes covered by [H.sub.2], and the reaction performed, as the stored NOx is transported, will preferentially form N[H.sub.3] .
Previous interpretations of the net N[H.sub.3] production are considerably consistent with the trend observed in Figure 11 and Figure 12 by comparing the vehicle- and oven-aged with the fresh catalyst. However, a substantial difference of the N[H.sub.3] formed between the aged catalysts is also observed. It has been reported in the literature that the water gas shift (WGS) reaction and the formation of surface isocyanates represent the two most probable pathways to N[H.sub.3] formation when CO is used as a reductant . Adams et al.  demonstrated that palladium stands out as a promising compound for passive SCR solutions because of its wide range of working temperatures and its high WGS activity. It is possible, as discussed earlier, that Pd sites may be deactivated on the vehicle-aged catalyst by the sulfur species, and thus might affect the occurrence WGS reaction and hence the N[H.sub.3] formation rate.
Moreover, the [N.sub.2]O profile matches those of the NOx conversion efficiency (see Figure 11 and Figure 12), which is consistent with the work reported by Ji et al. . High [N.sub.2]O levels were observed with the fresh catalyst, while its selectivity decreased by aging the samples. The vehicle-aged catalyst exhibited a lower [N.sub.2]O selectivity than the oven-aged probably due to the deactivation of the noble metals previously discussed. The reason that the duration of the rich period does not have any substantial effect on the [N.sub.2]O formation might be because it is formed over the precious metal sites in the first part of the rich regime .
The [N.sub.2]O is also known to be formed immediately after the rich to lean transition as it is shown in Figure 13. Choi et al.  inferred that this behavior is because of the deposition of certain species formed by the reaction between reductive species with residual surface NOx. The decomposition of those inhibiting surface species occurs when the reductive species are removed from the exhaust stream, giving rise to the formation of [N.sub.2]O and [H.sub.2]O.
The THC and CO light-off temperature as a function of the average space velocity during the fuel injection ramp is depicted in Figure 14. The exhaust mass flow rates range between 30000 [h.sup.-1] and 62000 [h.sup.-1]. It is important to mention that the engine-out gas composition varies among the three different space velocities presented, therefore they cannot be compared among them. The HC light-off is observed to occur at higher temperatures than the CO, which is consistent with the results presented previusly by Nievergeld et al. . The CO light-off took place around 160 [degrees]C for the fresh catalyst and slightly above 170 [degrees]C for the aged catalysts. On the contrary, the HC light-off exhibits a larger deterioration compared to the CO light-off after exposing the samples to the aging methods. The HC light-off was about 175 [degrees]C for the fresh catalysts, while the aged catalysts showed 50 % of THC conversion around 215 [degrees]C as an average temperature among the three cases.
Gonzalez-Velasco et al.  studied the thermal aging effect of different PGM formulations on their catalytic activity and reported an important increase of the light-off temperature for the Pd-Pt-Rh catalysts. In fact, this was also observed in our results. Several studies have attributed this behaviour to the loss of the PGM activity, and the sintering of ceria and precious metals, which reduces the interaction of both components and limits the oxidation of the THC and CO species [43, 44].
Vehicle Emission Testing
In order to evaluate the catalyst performance and degradation under real driving conditions, vehicle emission tests were performed in the vehicle chassis dynamometer. Figure 15 illustrates the LNT conversion efficiency as a function of the emission species over the NEDC, WLTC and CADC. The overall CO conversion efficiency was significantly large for all three cycles, ranging from 91.4 % to 98 %. The fresh catalyst exhibited the largest conversion, followed by the oven-aged sample. Similar trend is also observed with the THC species; however, it shows slightly lower conversion compared to CO (in the range of 83.3 % and 95 %). On the contrary, poor NOx conversion was observed for CADC and WLTC cycles, while relatively better performance is presented for NEDC. It should be noted that these are the emissions after the LNT only, and after this catalyst, an SCR unit is placed. Furthermore, the influence of the aging methods on the NOx species was noticeable over the NEDC, in which the oven-aged and vehicle aged catalysts exhibited a reduction of 25 % and 33.4 % of NOx conversion efficiency respectively compared to the fresh sample.
NEDC has been criticized in several investigations [45, 46, 47] for covering a limited range of real driving conditions because of its smooth acceleration profile and underloaded vehicle operation, which do not replicate the actual on-road driving behaviors. WLTC and CADC were instead developed based on real world in-use data with more frequent and harder accelerations, and higher vehicle speeds compared to NEDC, which properly represent the real-life driving patterns. However, NEDC offers high repeatable results and optimal operating conditions for the evaluation of different aftertreatment systems. Thus, the different LNT samples are compared in Figure 16A, which shows the cumulative post-LNT emission profiles of CO, THC and NOx species using NEDC as a reference cycle. Note that the NOx purge events were performed twice during the cycle after approximately 870 s and 1075 s (see Figure 16B). The fresh and oven-aged catalysts present similar CO trends, in which 81 % and 70 % of the overall post-LNT CO emissions were respectively registered in the first 140 s of the cycle, demonstrating the importance of warming up the system as quickly as possible to the proper operating temperatures. The vehicle-aged catalyst has a smaller share of CO emissions during the first 140 s (36 %) due to a significant contribution of CO slip during the DeNOx events. Because of the NOx purges were performed at relatively low temperatures, these results might prove that the palladium sites were significantly affected by the sulfur species, since Pd is well known to have a greater low-temperature WGS activity than the Pt sites . Moreover, the CO can be consumed by the reaction with the stored oxygen, thus these findings could also suggest lower OSC of the vehicle-aged catalyst in comparison with the oven-aged sample. It is important to mention that even though a significant difference is observed in Figure 16A by comparing the cumulative post-LNT CO emissions between the vehicle-aged, and the fresh and oven-aged catalysts, the overall CO conversion efficiency is clearly high (see Figure 15) due to high CO engine-out emission levels.
Likewise, the fresh catalyst exhibits significantly lower overall THC emissions than the aged samples probably due to larger number of active sites and greater oxygen storage capacity. Moreover, the aged catalysts behave clearly similar throughout the NEDC, however a minor difference is observed during the NOx purge events, in which larger amount of THC slips from the vehicle-aged sample compared to the oven aged likely due to lower reducing capabilities of the catalyst.
Turning the focus on the NOx emissions, the results are substantially consistent to the findings discussed previously in the NOx storage and NOx purge capabilities evaluations. The fresh catalyst stands out for its high NOx storage efficiency during the urban driving part of the NEDC, whereas the aged catalysts show lower NOx storage and analogous trends as it was presented in Figure 7. The results in Figure 16B focuses on the NOx species during the extra-urban driving cycle, in which the DeNOx events were performed. This figure illustrates high NOx slip for the vehicle-aged catalyst during the NOx purge events, which results in poor reducing capabilities and high overall NOx emissions probably because of the deactivation of the Pd sites.
Correlation between Accelerated Aging Method and Vehicle-Aged Catalysts
In the present investigation, a number of engine bench and vehicle chassis dynamometer tests have been performed in order to characterize the performance and deactivation of a commercial LNT catalyst exposed to vehicle aging, with the purpose of determining to what extend it could be replicated in an accelerated aging method. Significant deterioration was observed in Figure 7 by comparing the NOx storage between the fresh and aged catalysts. This behavior could be attributed to lowered NO oxidation capabilities after exposing the catalysts to the aging methods, due to loss of the noble metal activity. There was a very good correlation between the aged catalysts, indicating that they were probably exposed to similar thermal loads. Similar trend was also observed in Figure 16 during the urban driving part of the NEDC, which demonstrates similar NOx storage limitations for the aged catalysts under real driving conditions.
On the contrary, the vehicle-aged catalyst showed a larger NOx reduction degradation compared to the oven sample (see Figure 11 and Figure 12). In addition, significant difference was also observed in terms of N[H.sub.3] and [N.sub.2]O formation between the aged samples. This behavior was attributed probably due to the deactivation of Pd sites by the presence of sulfur species in the exhaust stream during the aging test. The results obtained in the engine bench were well correlated to those illustrated in Figure 16, particularly during the extra-urban driving part of the NEDC. Vehicle-aged catalyst showed larger amounts of NOx slip during the DeNOx events compared to the oven-aged sample. Therefore, in this context, by poisoning the oven-aged sample in the engine bench, the deactivation can be improved towards a better correlation of the LNT NOx storage-reduction activity between the vehicle-aged and oven-aged catalysts.
SUMMARY AND CONCLUSIONS
The catalytic activity of a vehicle aged commercial LNT has been studied throughout this investigation with the purpose of determining its catalytic degradation and establishing a correlation with respect to an accelerated aging method. The vehicle-aged catalyst was driven in the vehicle chassis dynamometer for 100000 km, while the rapid-aged catalyst was treated at 800 [degrees]C for a period of time equivalent to a driving distance of 160000 km. A significant degradation of the catalytic activity was exhibited in the performed experiments. The NOx storage deterioration ranges between 35 % and 85 %, being more pronounced in conditions of low operating temperatures. Poor NO oxidation activity was suggested to be one reason for this behavior.
Additionally, the deterioration was also shown in the NOx purge test. Poor NOx conversion efficiency was observed for the aged catalysts, attributed to a slower kinetics of NOx adsorption and reduction. Likewise, high [N.sub.2]O levels were observed with the fresh catalyst, while its selectivity decreased by aging the samples. On the contrary, the aged catalysts exhibited significantly higher N[H.sub.3] production than the fresh sample. This was attributed to the fact that less reductant was consumed by the NOx and oxygen stored during the rich periods, which leads to higher concentration of [H.sub.2] in the gas front, promoting the N[H.sub.3] formation. Moreover, increasing the reductant concentration and extending the regeneration periods lead to higher [H.sub.2] / NO ratios and thereby rises the amount of N[H.sub.3] formed.
The THC/CO light-off has also been influenced by exposing the catalysts to the aging methods. An increase of the light-off temperature probably because of the loss of the noble metals activity was observed. In particular, larger deterioration was observed for the THC species than the CO. The results were clearly consistent with those acquired in the vehicle chassis dynamometer under real driving conditions.
Moreover, the oven aging method proposed in this investigation is definitely a good approach to mimic the long-term catalytic activity in which the thermal stress applied is able to replicate NO oxidation capabilities and the loss of the noble metals activity of a 100000 km vehicle-aged catalyst. However, the LNT NOx reduction capabilities remain to improve by poisoning the oven-aged sample in the engine bench with a high sulfur-content fuel to promote the chemical deactivation of the Pd sites with a reduced testing time and a realistic spatial poison distribution.
Finally, detailed characterization studies of the catalysts are planned to be performed at Chalmers University of Technology by carrying out flow reactor experiments and using some advance analytical techniques (SEM, TEM, etc.), which includes the measure of the oxygen storage capacity loss, noble metal dispersion loss, loss of N[O.sub.x] storage etc.
[1.] Johnson, T., "Vehicular Emissions in Review," SAE Int. J. Engines 9(2):1258-1275, 2016, doi:10.4271/2016-01-0919.
[2.] Ocampo, F., Harle, V., Ohtake, N., Rohe, R. et al., "Innovative Hybrid Rare Earth and Barium Materials with Enhanced Properties for NOx Storage Applications," SAE Int. J. Engines 8(3):1136-1143, 2015, doi:10.4271/2015-01-0988.
[3.] Choi, J., Partridge, W., Lance, M., et al., "Nature and spatial distribution of sulfur species in a sulfated barium-based commercial lean NOx trap catalyst," Catalysis Today. 151(3-4):354-361, 2010, doi: 10.1016/j.cattod.2010.01.016.
[4.] Benramdhane, S., Millet, C., Jeudy, E., et al., "Impact of thermal and vehicle aging on the structure and functionalities of a lean NOx-trap," Catalysis Today. 176(1):56-62, 2011, doi:10.1016/j.cattod.2011.03.049.
[5.] Chan, D., Gremminger, A., and Deutschmann, O., "Effect of Hydrothermal Aging on Physical and Chemical Properties of a Commercial NOx-Storage Catalyst," Topics in Catalysis. 56(1):293-297, 2013, doi:10.1007/s11244-013-9969-5.
[6.] DiGiulio, C., Pihl, J., Choi, J., et al., "N[H.sub.3] formation over a lean NOx trap (LNT) system: Effects of lean/rich cycle timing and temperature," Applied Catalysis B: Environmental. 147:698-710, 2014, doi: 10.1016/j.apcatb.2013.09.012.
[7.] Dupre, J., Bazin, P., Marie, O., et al., "Understanding the storage function of a commercial NOx-storage-reduction material using operando IR under realistic conditions," Applied Catalysis B: Environmental. 160-161:335-343, 2014, doi: 10.1016/j.apcatb.2014.05.027.
[8.] Dupre, J., Bazin, P., Marie, O., et al., "Effects of temperature and rich-phase composition on the performance of a commercial NOx-Storage-Reduction material," Applied Catalysis B: Environmental. 181:534-541, 2016, doi:10.1016/j.apcatb.2015.08.033.
[9.] Ji, Y., Fisk, C., Easterling, V., et al., "NOx storage-reduction characteristics of Ba-based lean NOx trap catalysts subjected to simulated road aging," Catalysis Today. 151(3-4):362-375, 2010, doi:10.1016/j.cattod.2009.12.009.
[10.] Ozawa, M., "Role of cerium-zirconium mixed oxides as catalysts for car pollution: A short review1," Journal of Alloys and Compounds. 275-277:886-890, 1998, doi:10.1016/S0925-8388(98)00477-0.
[11.] Ottinger, N., Nguyen, K., Bunting, B., Toops, T. et al., "Effects of Rapid High Temperature Cyclic Aging on a Fully-Formulated Lean NOx Trap Catalyst," SAE Int. J. Fuels Lubr. 2(1):217-228, 2009, doi:10.4271/2009-01-0634.
[12.] Pieta, I., Garcia-Dieguez, M. Larrubia, M., et al., "Nanofiber Alumina Supported Lean NOx Trap: Improved Sulfur Tolerance and NOx Reduction," Topics in Catalysis. 56(1):50-55, 2013, doi:10.1007/s11244-013-9928-1.
[13.] Rohr, F., Kattwinkel, P., Kreuzer, T., Miiller, W. et al., "NOx-Storage Catalyst Systems Designed to Comply with North American Emission Legislation for Diesel Passenger Cars," SAE Technical Paper 2006-01-1369, 2006, doi:10.4271/2006-01-1369.
[14.] Rohr, F., GriBtede, I., Sundararajan, A., and Miiller, W., "Diesel NOx-Storage Catalyst Systems for Tier 2 BIN5 Legislation," SAE Technical Paper 2008-01-0766, 2008, doi:10.4271/2008-01-0766.
[15.] Kumar, S. and Sardesai, P., "Durability Evaluation of Spark-Ignition Engine Three Way Catalytic Converters," SAE Technical Paper 2001-26-0012, 2001, doi:10.4271/2001-26-0012.
[16.] Shao, H., Plaatje, A., and Meffert, M., "Proof-of-Principle Investigation into the Use of Custom Rapid Aging Procedures to Evaluate and Demonstrate Catalyst Durability," SAE Int. J. Fuels Lubr. 3(2):963-981, 2010, doi:10.4271/2010-01-2269.
[17.] Ruetten, O., Pischinger, S., Kipper, C., Weinowski, R. et al., "Catalyst Aging Method for Future Emissions Standard Requirements," SAE Technical Paper 2010-01-1272, 2010, doi:10.4271/2010-01-1272.
[18.] European Parliament, "Establishing a framework for the approval of motor vehicles and their trailers, and of systems, components and separate technical units intended for such vehicles," 2007.
[19.] Andre, M., "The ARTEMIS European driving cycles for measuring car pollutant emissions," Science of The Total Environment. 334-335:73-84, 2004, doi:10.1016/j.scitotenv.2004.04.070.
[20.] Heck, R., Farrauto, R., and Gulati, S., "Catalytic Air Pollution Control," (John Wiley & Sons, Inc., 2009), 79-99, ISBN: 9781118397749.
[21.] Rohr, F., Peter, S., Lox, E., et al., "On the mechanism of sulphur poisoning and regeneration of a commercial gasoline NOx-storage catalyst," Applied Catalysis B: Environmental. 56(3):201-212, 2005, doi:10.1016/j.apcatb.2004.09.011.
[22.] Auvray, X., and Olsson, L., "Effect of Enhanced Support Acidity on the Sulfate Storage and the Activity of Pt/[gamma]-[Al.sub.2][O.sub.3] for NO Oxidation and Propylene Oxidation," Catalysis Letters. 144(1):22-31, 2013, doi:10.1007/s10562-013-1145-9.
[23.] Kylhammar, L., Carlsson, P., Ingelsten, H., et al., "Regenerable ceria-based SOx traps for sulfur removal in lean exhausts," Applied Catalysis B: Environmental. 84(1-2):268-276, 2008, doi:10.1016/j.apcatb.2008.04.003.
[24.] Epling, W., Yezerets, A., and Currier, N., "The effect of exothermic reactions during regeneration on the NOx trapping efficiency of a NOx storage/reduction catalyst," Catalysis Letters. 110(1):143-148, 2006, doi:10.1007/s10562-006-0102-2.
[25.] Toops, T., Smith, D., and Partridge, W., "NOx adsorption on Pt/K/[Al.sub.2][O.sub.3]," Catalysis Today. 114(1):112-124, 2006, doi: 10.1016/j.cattod.2006.02.006.
[26.] Ohtsuka, H., "The selective catalytic reduction of nitrogen oxides by methane on noble metal-loaded sulfated zirconia," Applied Catalysis B: Environmental. 33(4):325-333, 2001, doi:10.1016/SO926-3373(01)00190-4.
[27.] Gelin, P., and Primet, M., "Complete oxidation of methane at low temperature over noble metal based catalysts: a review," Applied Catalysis B: Environmental. 39(1):1-37, 2002, doi:10.1016/SO926-3373(02)00076-0.
[28.] Clayton, R., Harold, M., and Balakotaiah, V., "Selective catalytic reduction of NO by [H.sub.2] in [O.sub.2] on Pt/BaO/[Al.sub.2][O.sub.3] monolith NOx storage catalysts," Applied Catalysis B: Environmental. 81(3-4):161-181, 2008, doi:10.1016/j.apcatb.2007.11.038.
[29.] Pihl, J., Parks, J., Daw, C., and Root, T., "Product Selectivity During Regeneration of Lean NOx Trap Catalysts," SAE Technical Paper 2006-01-3441, 2006, doi:10.4271/2006-01-3441.
[30.] Tonkyn, R., Disselkamp, R., and Peden, C., "Nitrogen release from a NOx storage and reduction catalyst," Catalysis Today. 114(1):94-101, 2006, doi:10.1016/j.cattod.2006.02.005.
[31.] Lietti, L., Nova, I., and Forzatti, P., "Role of ammonia in the reduction by hydrogen of NOx stored over Pt-Ba/[Al.sub.2][O.sub.3] lean NOx trap catalysts," Journal of Catalysis. 257(2):270-282, 2008, doi: 10.1016/j.jcat.2008.05.005.
[32.] Bartova, S., Mracek, D., Koci, P., et al., "Ammonia reactions with the stored oxygen in a commercial lean trap catalyst," Chemical Engineering Journal. 278:199-206, 2015, doi:10.1016/j.cej.2014.09.115.
[33.] Ji, Y., Easterling, V., Graham, U., et al., "Effect of aging on the NOx storage and regeneration characteristics of fully formulated lean NOx trap catalysts," Applied Catalysis B: Environmental. 103(3-4):413-427, 2011, doi:10.1016/j.apcatb.2011.02.005.
[34.] Nova, I., Castoldi, L., Lietti, L., et al., "How to control the selectivity in the reduction of NOx with [H.sub.2] over Pt-Ba/[Al.sub.2][O.sub.3] Lean NOx Trap catalysts," Topics in Catalysis. 42(1):21-25, 2007, doi:10.1007/s11244-007-0144-8.
[35.] Ji, Y., Choi, J., Toops, T., et al., "Influence of ceria on the NOx storage/reduction behavior of lean NOx trap catalysts," Catalysis Today. 136(1-2):146-155, 2008, doi:10.1016/j.cattod.2007.11.059.
[36.] Easterling, V., Ji, Y., Crocker, M., et al., "Application of spaciMS to the study of ammonia formation in lean NOx trap catalysts," Applied Catalysis B: Environmental. 123-124:339-350, 2012, doi: 10.1016/j.apcatb.2012.05.002.
[37.] Epling, W., Yezerets, A., Currier, N., Hess, H. et al., "Spatially-Resolved Thermal Degradation Induced Temperature Pattern Changes along a Commercial Lean NOX Trap Catalyst," SAE Int. J. Fuels Lubr. 3(1):723-732, 2010, doi:10.4271/2010-01-1214.
[38.] Clayton, R., Harold, M., Balakotaiah, V., et al., "Pt dispersion effects during NOx storage and reduction on Pt/BaO/[Al.sub.2][O.sub.3] catalysts," Applied Catalysis B: Environmental. 90(3-4):662-676, 2009, doi: 10.1016/j.apcatb.2009.04.029.
[39.] Bhatia, D., Harold, M., and Balakotaiah, V., "Modeling the effect of Pt dispersion and temperature during anaerobic regeneration of a lean NOx trap catalyst," Catalysis Today. 151(3-4):314-329, 2010, doi: 10.1016/j.cattod.2010.02.055.
[40.] Adams, E., Skoglundh, M., Folic, M., et al., "Ammonia formation over supported platinum and palladium catalysts," Applied Catalysis B: Environmental. 165:10-19, 2015, doi:10.1016/j.apcatb.2014.09.064.
[41.] Choi, J., Partridge, W., Pihl, J., et al., "Spatiotemporal distribution of NOx storage and impact on N[H.sub.3] and [N.sub.2]O selectivities during lean/rich cycling of a Ba-based lean NOx trap catalyst," Catalysis Today. 184(1):20-26, 2012, doi:10.1016/j.cattod.2011.11.007. '
[42.] Nievergeld, A., Selow, E., Hoebink, J., et al., 1997. "Simulation of a catalytic converter of automotive exhaust gas under dynamic conditions." In Studies in Surface Science and Catalysis, edited by Froment, G. and Waugh, K., 449-458. Elsevier.
[43.] Gonzalez-Velasco, J., Botas, J., Ferret, R., et al., "Thermal aging of Pd/Pt/Rh automotive catalysts under a cycled oxidizing-reducing environment," Catalysis Today. 59(3-4):395-402, 2000, doi:10.1016/S0920-5861(00)00304-7.
[44.] Rogemond, E., Frety, R. Perrichon, V., et al., "Evolution of the ceria surface area of Pt/Rh ceria-alumina catalysts submitted to hydrothermal aging at 1273 K," Applied Catalysis A: General. 156(2):253-265, 1997, doi:10.1016/S0926-860X(97)00040-9.
[45.] Weiss, M., Bonnel, P., Hummel, R., et al., "On-Road Emissions of Light-Duty Vehicles in Europe," Environmental Science & Technology. 45(19):8575-8581, 2011, doi:10.1021/es2008424.
[46.] Sileghem, L., Bosteels, D., May, J., et al., "Analysis of vehicle emission measurements on the new WLTC, the NEDC and the CADC," Transportation Research Part D: Transport and Environment. 32:70-85, 2014, doi:10.1016/j.trd.2014.07.008.
[47.] Yang, L., Franco, V., Mock, P., et al., "Experimental Assessment of NOx Emissions from 73 Euro 6 Diesel Passenger Cars," Environmental Science & Technology. 49(24):14409-14415, 2015, doi:10.1021/acs.est.5b04242.
Jesus Emmanuel De Abreu Goes Diesel Exhaust Aftertreatment systems Volvo Car Corporation
Gunnar Engellaus vag VAKHD2N SE-405 31 Gothenburg, Sweden T +46 (0) 723 865633
This study is a collaboration between Volvo Car Corporation and Competence Centre for Catalysis at Chalmers University of Technology. We gratefully acknowledge the financial support from Volvo Car Corporation and the Swedish Foundation for Strategic Research (ID15-0030).
CADC - Common Artemis Driving Cycle
CO - Carbon monoxide
DMode - Engine operating mode
DPF - Diesel Particulate Filter
LNT - Lean NOx Trap catalyst
NEDC - New European Driving Cycle
NOx - Mono-nitrogen oxide
OSC - Oxygen storage capacity
QCL - Quantum-cascade laser
SCR - Selective Catalytic Reduction
SRC - Standard Road Cycle
THC - Total unburned hydrocarbons
WLTC - Worldwide harmonized Light duty driving Test Cycle
Jesus Emmanuel De Abreu Goes
Volvo Car Corporation
Chalmers University of Technology
Malin Berggrund, Annika Kristoffersson, Lars Gustafson, and Mikael Hicks
Volvo Car Corporation
Table 1. Overview of the catalysts tested Samples Fresh catalyst Vehicle-aged catalyst Oven-aged catalyst Washcoat composition Modern commercial LNT formulation Precious metal loading 119 g/[ft.sup.3] (Pt:Pd:Rh = 95:19:5) Substrate Metallic Dimension 124 x 124 x 128 mm (V= 1.546 1)