SCR for Improved Activity and

Publication Date (Web): March 25, 2013 ... *Phone: 82-54-279-1325. ... Dioxide: Cryotrapping of Reaction Products for Online Monitoring by Mass Spectr...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Combination of Photocatalysis and HC/SCR for Improved Activity and Durability of DeNOx Catalysts Iljeong Heo,†,§ Mun Kyu Kim,† Samkyung Sung,† In-Sik Nam,*,† Byong K. Cho,*,† Keith L. Olson,‡ and Wei Li‡ †

School of Environmental Science and Engineering/Department of Chemical Engineering, Pohang University of Science and Technology (POSTECH), San 31 Hyoja-dong, Pohang 790-784, Korea ‡ General Motors Global Research and Development Center, Warren, Michigan 48090, United States S Supporting Information *

ABSTRACT: A photocatalytic HC/SCR system has been developed and its high deNOx performance (54.0−98.6% NOx conversion) at low temperatures (150−250 °C) demonstrated by using a representative diesel fuel hydrocarbon (dodecane) as the reductant over a hybrid SCR system of a photocatalytic reactor (PCR) and a dual-bed HC/SCR reactor. The PCR generates highly active oxidants such as O3 and NO2 from O2 and NO in the feed stream, followed by the subsequent formation of highly efficient reductants such as oxygenated hydrocarbon (OHC), NH3, and organo-nitrogen compounds. These reductants are the key components for enhancing the low temperature deNOx performance of the dual-bed HC/SCR system containing Ag/Al2O3 and CuCoY in the front and rear bed of the reactor, respectively. The OHCs are particularly effective for both NOx reduction and NH3 formation over the Ag/Al2O3 catalyst, while NH3 and organo-nitrogen compounds are effective for NOx reduction over the CuCoY catalyst. The hybrid HC/SCR system assisted by photocatalysis has shown an overall deNOx performance comparable to that of the NH3/SCR, demonstrating its potential as a promising alternative to the current urea/SCR and LNT technologies. Superior durability of HC/SCR catalysts against coking by HCs has also been demonstrated by a PCR-assisted regeneration scheme for deactivating catalysts.

■

INTRODUCTION

Given that the OHC/SCR can improve the poor deNOx activity of the HC/SCR at low temperatures, an enhanced deNOx performance over a wider temperature range may be achieved by a (HC+OHC)/SCR system, where a mixture of HC and OHC is employed as the reductant.7,9,15−17 When a mixture of ethanol and dodecane is used as the reductant in the (HC+OHC)/SCR over Ag/Al2O3, ethanol and dodecane react with O2 over Ag/Al2O3 to produce enolic and acetate species, respectively, according to reaction 1.7,15 The enolic species are known to be more reactive intermediates than the acetate compounds for low-temperature NOx reduction. Thus, the formation of −NCO species at low temperatures is primarily due to the reaction of the enolic species with surface nitrates on the Ag/Al2O3 surface according to reaction 2, while the −NCO formation at high temperatures is due to the acetate species reaction. The −NCO species formed via reaction 2 subsequently reacts with NOx to produce N2 [reaction 3],

Nitrogen oxides (NOx) emitted from stationary and mobile sources are harmful gaseous pollutants contributing to the formation of acid rain and smog. The selective catalytic reduction of NOx by urea (urea/SCR) and the lean NOx trap (LNT) have been widely accepted as the most efficient deNOx technologies for their relatively high deNOx efficiencies to control the NOx emissions from lean-burn engines such as the diesel engine. However, they still have serious drawbacks; for example, an additional urea tank to be refilled periodically for the urea/SCR and the high cost of noble metals with complex engine control for the LNT.1−5 As a potential alternative without those drawbacks inherent in the urea/SCR and the LNT, the selective catalytic reduction of NOx by fuel hydrocarbon (HC/SCR) has been extensively studied.6−12 However, for successful commercial applications, the low-temperature deNOx performance of the current HC/ SCR technology has much room for improvement. One way to improve the low-temperature deNOx activity of the HC/SCR is to use oxygenated hydrocarbons (OHCs) such as alcohols and aldehydes as reductants (OHC/SCR), injected externally or generated on-board by a reformer such as a plasma reactor.13,14 © 2013 American Chemical Society

Received: Revised: Accepted: Published: 3657

October 13, 2012 March 6, 2013 March 6, 2013 March 25, 2013 dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

Article

described in the Supporting Information (Table S1). The 3.5 wt.% Ag/TiO2, 3.5 wt.% Ag/Al2O3, and 1.5 wt.% Pt/TiO2 catalysts were prepared by the incipient wetness impregnation of active metals on γ-Al2O3 (CATALOX SBa-200, Sasol) and TiO2 (anatase nanopowder, MTI Corp.) supports using AgNO3 (J.T.Baker) and H2PtCl6•6H2O (Alfa) as active metal precursors. The CuCoY catalyst was prepared by the wet ionexchange method at room temperature using Cu(NO3)2•3H2O (0.01 M, Junsei) and Co(NO3)2•6H2O (0.01 M, Alfa) solutions with 10 g of NaY (Zeolyst, CBV 100). The content of Cu and Co was 4.4 and 2.5 wt.%, respectively. The CuZSM5 and CuMOR for the NH3/SCR system were also prepared by the wet ion-exchange method using Cu(CH3COO)2•H2O (0.01M, Merck), ZSM5 (HSZ-830NHA, Tosoh), and MOR (Zeolon 900Na, PQ Co.). The content of Cu in the CuZSM5 and CuMOR was 2.8 and 2.3 wt.%, respectively. The BaY catalyst was prepared by the wet ion-exchange of NaY using an aqueous solution of Ba(CH3COO)2 to obtain 13 wt.% of Ba loading. After the as-prepared catalysts were dried in an oven at 80 °C overnight, all catalysts were calcined in a muffle furnace under static air condition at 500 °C for 5 h. The V2O5/TiO2 (Umicore) and Au/TiO 2 (Gold Corp.) catalysts were commercial products. Monolith Washcoating. Monolithic catalysts were washcoated by the dipping method. The catalyst slurries were prepared by the mixing of catalyst powder and distilled water at a ratio of 7/3 (H2O/catalyst) on a weight basis. The wet catalyst slurries were then mixed with alumina binder (30 wt.% Alumina sol, Aremco) diluted to 3% aqueous alumina sol solution at a ratio of 1/4 (catalyst slurry/alumina binder solution). The catalyst slurries, including the alumina sol, were ground by ball milling with ZrO2 balls for 24 h. The cordierite substrates (600 cpi, Corning Inc.) with square-shaped cells were submerged into the catalyst slurries for coating. The cordierite monoliths coated with the catalyst slurry were blown out by compressed air to remove excess slurry deposited onto the channels of the cordierite and then dried at 150 °C for 30 min. The amount of the catalyst washcoated onto the internal channels of cordierite substrate was controlled by repeating this procedure. After the final washcoating of the catalysts, the prepared monolith reactors were dried in an oven at 80 °C overnight and then calcined at 500 °C for 5 h. The amount of washcoat was 30% on a weight basis for all the monoliths employed in the present study. Reaction System. Ideally, a fluidized-bed reactor can achieve the best photocatalytic activity due to the efficient interaction between photons and catalyst particles under uniform fluidization conditions. However, uniform fluidization of very small catalyst particles (50−100 μm size) in a laboratory reactor was extremely difficult to obtain, especially for hygroscopic particles due to their high propensity for agglomeration. Although it may be the best choice to demonstrate the concept of the photocatalytic HC/SCR system, the fluidized bed is not suitable for practical application in the automotive exhaust after-treatment system. Thus, in order to circumvent this difficulty with uniform fluidization, a photocatalytic reactor (PCR) was designed in the form of a simulated fluidized-bed in order to mimic the high quantum efficiency of UV radiation in an ideal fluidized-bed of catalyst particles (Supporting Information, Figure S1). This PCR design feature can accommodate any need for a specific distribution of catalyst formulations along the axial direction of the PCR. Though the number of honeycomb monolith catalysts in the

where (g) and (ad) denote the gaseous and adsorbed states of the species, respectively.6,7,10,11,15 OHC(ad) + O2 (ad) → CH 2CH−O−/CH3COO−(ad) (1)

−NO3(ad) + CH 2CH−O−/CH3COO−(ad) → −NCO(ad)

(2)

−NCO(ad) + NOx (ad) → N2(g ) + COx(g )

(3)

The (HC+OHC)/SCR catalysts can be deactivated by HC coking on the active sites of the catalysts, when the OHC/HC feed ratio is below the critical value where the catalyst deactivation starts.7,18,19 There is an optimum OHC/HC feed ratio for the maximum deNOx performance of Ag/Al2O3 during the (HC+OHC)/SCR reaction.7 This catalyst deactivation induced by the HC coking is especially relevant when the on-board OHC generator such as a plasma reactor or a photocatalytic reactor cannot provide the necessary OHC/HC feed ratio required to prevent the coking due to its design limitations. Thus, it is desirable to develop an operating strategy for preventing the coke formation during the (HC+OHC)/ SCR reaction over the catalyst for a successful application of this hybrid/SCR system to the diesel aftertreatment system. In this study, a hybrid HC/SCR system with an enhanced deNOx performance comparable to the urea/SCR system has been developed for the aftertreatment of diesel engine exhausts. Based on the concept of in situ photocatalytic generation of both OHCs and organo-nitrogen compounds from fuel hydrocarbons and NOx in the exhaust stream, the hybrid deNOx system consists of a photocatalytic reactor (PCR) and traditional HC/SCR catalysts in a dual-bed reactor.9

■

EXPERIMENTAL SECTION Catalyst Preparation. A total of 8 catalysts were employed in the test of a prototype photocatalytic HC/SCR system, as listed in Table 1. These catalysts were selected on the basis of the suitability of their main catalytic functions in demonstrating the basic concept of a photocatalytic HC/SCR system, as Table 1. Effect of Catalysts on the Performance of the PCR at 150 °C yield of reductants (%) catalyst 2BaY +

TiO2 V2O5/TiO2 Au/TiO2 Ag/TiO2 Pt/TiO2 CuCoY BaY Ag/Al2O3 blank PCR (w/o catalysts)

total OHCsa b

35 (84 ) 60 24 18 2 19 10 16 45

NH3c

NOx conversiond (%)

6 -

43 29 43 55 22 37 55 54 -

a

Yield of OHCs (on C 1 basis) = Conc. of (Alcohols + Aldehydes)outlet/Conc. of Dodecaneinlet. bYield of OHCs in the absence of NO. cYield of NH3 = Conc. of NH3outlet/Conc. of NOinlet. d NOx conversion =1 − Conc. of NOoutlet/Conc. of NOinlet. Feed gas composition: 200 ppm NO, 134 ppm C12H26, 6% O2, 10% H2O, and N2 balance. Total flow rate = 500 cc/min. -: negligible. 3658

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

Article

Figure 1. Schematic flow diagram of hybrid PCR+HC/SCR reactor system.

16,500 h−1 for a corresponding monolith reactor (Supporting Information, Figure S3). The inlet and outlet concentrations of the reaction gas mixture to and from the reactor system were monitored by an FTIR equipped with a gas cell (2 m) heated at 165 °C and maintained at 940 Torr (Nexus 5700, Thermo Electron Co.). The IR spectra for the gaseous species were obtained in the range of 600−4000 cm−1 with a spectral resolution of 0.5 cm−1 and a scan number of 32. The Thermo Scientific TQ Analyst software was used for identification, verification, and quantitative analysis of individual gaseous components by the FTIR.22 Under steady-state reaction conditions, the fluctuations of the FTIR signals measured as a function of time were within ±1.5% and ±5% of average values for NO and NO2, respectively. Included in the Supporting Information Table S2 is the information for detailed FTIR gas calibration along with the IR bands used for the quantitative analysis of gaseous samples.

simulated fluidized-bed reactor can be as many as desired, we used three thin layers of monolith catalysts (22.2 mm o.d. × 6.3 mm length) separated from each other by 25.4 mm and placed in a quartz tube (25.4 mm o.d.) wrapped with an aluminum foil reflector. The first two catalysts were BaY known to be very active for the OHC/SCR at low temperatures, whereas the third one was chosen from the catalysts listed in Table 1.14,16 A UV lamp (6.3 mm o.d., 177.8 mm long, Model 600, Jelight Co.) emitting UV light at λ = 254 and 185 nm was put through the center of the three monolith bricks to distribute the UV light evenly inside the PCR. In this way, all the flow channels of the three monoliths in the PCR were fully exposed to the UV illumination, while 89.3% of the total UV illumination was available for irradiating the monolith catalysts (Supporting Information, Figure S2). The temperature of the inlet and outlet of the PCR was maintained at 150 °C by using J-type thermocouples and a temperature controller. As shown in Figure 1, the hybrid deNOx system comprised the PCR and HC/SCR units. The HC/SCR reactor was fabricated with a dual-bed monolith catalyst (9.53 mm o.d. × 25.4 mm length each) containing Ag/Al2O3 in the front bed followed by CuCoY in the rear bed, placed in an electric furnace. Since n-dodecane had been used as the major component of the simulated diesel fuel and the average carbon number of diesel fuel is close to that of dodecane, the ndodecane (C12H26) was used for NOx reduction as the representative diesel fuel HC.7,9,20,21 The composition of the feed gas to the PCR and/or the HC/SCR reactor was controlled by mass flow controllers (1479A, MKS) for 200 ppm NO, 6% O2, and N2 balance and by a syringe pump (KDS100, KD Scientific) for 134 ppm dodecane heated at 260 °C (C1/ NOx = 8).7,9 Distilled water vapor was introduced into the feed stream through a heated gas bubbler to maintain the water content in the feed stream at 10 vol. %. The gas hourly space velocity (GHSV) of the monolithic HC/SCR reactor was maintained at 16,500 h−1 during the reaction. For the NH3/ SCR, 1.1 cc of Cu zeolite powder (20/30 mesh size) was charged in an aluminum tube (9.53 mm i.d.) placed in an electric furnace. The composition of the feed gas to the NH3/ SCR reactor was 200 ppm NO, 200 ppm NH3, 6% O2, 10% H2O, and N2 balance. During the NH3/SCR, the GHSV for the powder reactor was kept at 100,000 h−1 which is equivalent to

■

RESULTS AND DISCUSSION Performance of PCR for OHC Production and DeNOx. Presented in Table 1 is the effect of catalysts on the PCR’s performance for both OHC production and NOx conversion, in comparison with a blank PCR containing a UV lamp but no catalyst. The inlet temperature of the PCR was kept at 150 °C by electrically heating the feed gas stream. The V2O5/TiO2 PCR exhibits the highest yield of OHC production (60%) among other catalysts tested, while the Ag/Al2O3 PCR is the only one that produces NH3. This observation of NH3 indicates that N-containing intermediate reductants such as amines and other organo-nitrogen compounds, known as precursors for NH3 formation, may be also produced over the Ag/Al2O3 PCR.6−11,23 The Ag/TiO2 and BaY PCRs show the highest NOx conversion (55%), closely followed by the Ag/Al2O3 PCR (54%). The blank PCR (i.e., PCR without catalyst) also produces a substantial amount of OHC (45%) solely by UV radiation, but its NOx conversion activity is negligible. Interestingly, the blank PCR appears to produce more OHCs than the PCRs containing catalysts (PCR-c) except for the V2O5/TiO2 PCR. This can be explained by further reaction of OHCs produced by the UV radiation; they can be further oxidized to CO and CO2 or can participate in the OHC/SCR process for the NOx reduction. The lower OHC production by 3659

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

Article

catalysts downstream of the PCR. Thus, it is clear that the major role of the catalysts in the PCR is 2-fold: to enhance the efficiency of UV radiation for HC oxidation to OHC and COx, thereby minimizing the amount of residual HCs harmful to the HC/SCR catalysts in the downstream; and to promote the OHC/SCR reaction for the enhanced deNOx activity via its reaction with NOx. This can also explain the variation in the OHC formation shown in Figure 2. The Ag/Al2O3 PCR, due to its higher oxidation and deNOx activity, produces much less OHCs but more CO and CO2 than the V2O5/TiO2 PCR. Since the deNOx performance of the conventional HC/SCR catalyst can be enhanced by both OHCs and NHx-containing reductants,6,7,9,10,15−17,23 the V2O5/TiO2 PCR and Ag/Al2O3 PCR as the two representative PCRs were chosen for further tests in the PCR+HC/SCR hybrid system: the V2O5/TiO2 PCR for its superior OHC formation and the Ag/Al2O3 PCR for its formation of NH3 and possibly organo-nitrogen compounds as precursors of NH3. DeNOx Performance of PCR+HC/SCR Hybrid System. Presented in Figure 3 are the deNOx performances of both the

some PCR-c than that by the blank PCR is attributable to the consumption of the produced OHC via its reaction with NOx (i.e., OHC/SCR) and/or O2 over the catalysts. Indeed, the OHC yield of the TiO2 PCR increases from 35% to 84% by excluding NO from the feed gas stream as evidenced in Table 1. On the other hand, the blank PCR showed no deNOx performance, indicating no consumption of OHC as a reductant for the NOx reduction. This can explain why the OHC production by the blank PCR can be higher than that by most of the PCR-c. The PCRs containing BaY, Ag/TiO2, or Ag/Al2O3 exhibit superior deNOx activities but poor OHC yields, while the V2O5/TiO2 PCR exhibits the lowest deNOx activity but the best OHC yield among the PCRs tested, demonstrating a compensatory relation between the OHC yield and the NOx conversion over the PCR-c due to the OHC/SCR reaction (i.e., OHC+NO reaction) over the catalysts. Without UV radiation, both OHC formation and NOx conversion were negligible, regardless of the catalysts in the PCRs. These observations indicate that the deNOx performance of the PCR is induced by the catalysts assisted by the UV radiation in the PCR and clearly demonstrate that the major role of the PCR is to produce OHCs as well as to reduce NOx. Figure 2 compares

Figure 3. DeNOx activities of the PCR+HC/SCR and NH3/SCR catalysts: (□) Ag/Al2O3 PCR+HC/SCR, (Δ) V2O5/TiO2 PCR+HC/ SCR, (◊) HC/SCR only, (■) Ag/Al2O3 PCR only, (▲) V2O5/TiO2 PCR only, (−+−) CuZSM5 by NH3/SCR, (−x−) CuMOR by NH3/ SCR. Feed gas composition for NH3/SCR: 200 ppm NO, 200 ppm NH3, 6% O2, 10% H2O, and N2 balance. GHSV of the monolith reactor for the HC/SCR = 16,500 h−1. GHSV of the powder reactor for the NH3/SCR = 100,000 h−1 corresponding to 16,500 h−1 of the monolith reactor for the NH3/SCR (Supporting Information, Figure S3).

Figure 2. Conversion of reactants and yield of products over PCRs. Conversion of (dodecane or NO) = 1 − Conc. of (NO or dodecane)outlet/Conc. of (NO or dodecane)inlet. Yield of carbonaceous or nitrogenous compounds (on C1 or N1 basis) = Conc. of (individual carbonaceous or nitrogenous products)outlet/Conc. of (dodecane for carbonaceous or NO for nitrogenous compounds)inlet.

Ag/Al2O3 PCR+HC/SCR and the V2O5/TiO2 PCR+HC/SCR hybrid system in comparison with those of the HC/SCR system alone and the NH3/SCR system. At 200 °C, the Ag/ Al2O3 PCR+HC/SCR and the V2O5/TiO2 PCR+HC/SCR systems exhibit 63% and 41% of NOx conversion, respectively. This observation reflects that the NHx-containing compounds including NH3 produced by the Ag/Al2O3 PCR are more efficient reductants than OHCs for NOx reduction over the following HC/SCR catalyst. The NOx conversion performance of the HC/SCR alone employing dodecane itself as the reductant is negligible at 200 °C due to the low reactivity of dodecane for NOx reduction relative to that of OHCs. At 250 °C, the Ag/Al2O3 PCR+HC/SCR and the V2O5/TiO2 PCR +HC/SCR systems show about 99% and 95% of NOx conversions, respectively. These are almost 90% higher NOx conversions than those over the HC/SCR system alone. Listed

the detailed speciation of reaction products in the exit stream from the PCR under three different conditions of the PCR. The discrepancy in the carbon balance between the dodecane conversion in Figure 2 and the total OHC yield in Table 1 is due to the formation of CO, CO2, and light HCs such as propene and propane produced in the PCR, as can be seen in Figure 2. The Ag/Al2O3 PCR and V2O5/TiO2 PCR exhibit higher dodecane conversions than the blank PCR while producing more CO and CO2. The blank PCR produces light HCs such as propene and propane, whereas the Ag/Al2O3 PCR and V2O5/ TiO2 PCR do not. More important to note here is that the large amount of HCs, unconverted dodecane and newly formed light HCs, emitted from the blank PCR can deactivate the HC/SCR 3660

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

Article

vibration, etc. However, we believe there are ways to overcome or circumvent these drawbacks either by a proper design of the after-treatment system architecture or by the use of an ex situ UV system, or by adopting a plasma system in place of the UV, or by a combination thereof. 24 Reaction Pathway over PCR+HC/SCR System. The reaction steps involved in this hybrid system can be described as follows. Nitric oxide (NO) in the feed stream to the PCR is readily oxidized to NO2 by O3 [reaction 5] produced in situ by UV radiation at λ = 185 nm according to reaction 4.25 The nominal O3 production by UV in the blank PCR reaches to 139 ppm and 348 ppm for the 6% and 15% O2 feed concentration, respectively, at a total gas flow rate of 500 cc/min, as shown in Figure 4.26 The blank PCR also confirms an efficient NO

in Table 2 are the conversions of NO and NOx, conversion of NOx to N2, and yields of NO2, N2O, and NH3 during the NO Table 2. Conversion of NO and Yields of NO2, N2O, and NH3 during NO Reduction over the PCR+HC/SCR System system HC/SCR only

Ag/Al2O3 PCR +HC/SCR V2O5/TiO2 PCR+HC/ SCR

temp (°C)

NO (%)

NO2 (%)

N2O (%)

NH3 (%)

total NOxa (%)

calcd N2b (%)

256 303 400 500 200 250 300 200 250

13 62 100 53 63 99 100 42 95

0 0 0 7 0 0 0 0 0

1 1 3 1 4 5 5 4 5

0 1 2 1 0 4 13 1 0

13 62 100 46 63 99 100 42 95

13 60 96 44 59 91 82 38 91

Total conversion of NOx =1 − Conc. of (NO + NO2)outlet/Conc. of NOinlet. bCalculated conversion of NOx to N2 = 1 − Conc. of (NO + NO2 + 2N2O+ NH3)outlet/Conc. of NOinlet. Feed gas composition: 200 ppm NO, 134 ppm C12H26, 6% O2, 10% H2O, and N2 balance. GHSV of the HC/SCR monolith reactor = 16,500 h−1. a

reduction over the PCR+HC/SCR system, as determined by the FTIR spectrometer. The NOx conversion to N2 was not separately determined but estimated from the conversion of NOx and the yields of NO2, N2O, and NH3 by the mass balance of nitrogen. The N2 selectivities of the Ag/Al2O3 PCR +HC/SCR system are 94%, 91%, and 82% at 200 °C, 250 °C, and 300 °C, respectively. The slightly decreased N2 selectivity of the Ag/Al2O3 PCR+HC/SCR system at 300 °C is due to the increased formation of NH3 by the reaction of NO with OHCs over the HC/SCR reactor, since the PCR readily converts dodecane to OHCs.7 As the reaction temperature reaches 400 °C, the HC/SCR system alone can completely reduce NOx with a high N2 selectivity up to 96%. Thus, the PCR can be turned off and bypassed to save energy at high temperatures above 400 °C. Also compared in Figure 3 is the deNOx activity of the PCR +HC/SCR system with that of the NH3/SCR over the CuMOR as well as over the CuZSM5 powder catalyst. The Ag/Al2O3 PCR alone can convert 54% of NOx at 150 °C. The deNOx performance of the current hybrid system is shown to be comparable to that of the NH3/SCR over CuMOR in the temperature range of 150 to 400 °C. On the other hand, the deNOx performance of the NH3/SCR over CuZSM5 at around 200 °C is better than that over our hybrid system, while this trend is reversed above 250 °C, resulting in a comparable overall deNOx performance of the two systems. These findings indicate that the inherent drawbacks of the urea/SCR and LNT may be avoided by adopting the PCR+HC/SCR system as demonstrated in this work. However, the photocatalytic HC/ SCR system has its own shortcomings as well; for example, the additional photocatalytic reactor containing a UV lamp and the electrical energy required to operate it on top of the cost of fuel consumed as the NOx reductant. Based on the C1/NOx feed ratio of 8, the cost of reductants for the photocatalytic HC/ SCR has been estimated to be about 45% more expensive than that for the urea/SCR (Supporting Information, page S7). Additional shortcomings of the PCR system include the adverse effect of the diesel particulate matter and unburned heavy HCs on the efficiency of UV irradiation, the effect of vehicle

Figure 4. Nominal O3 productions and NO conversion to NO2 by oxidation over the blank PCR (UV only). Feed gas composition for O3 production: 6−15% O2 and N2 balance (Total flow rate: 500 cc/min), for NO oxidation: 100−200 ppm NO, 6% O2, 10% H2O and N2 balance (Total flow rate: 500−1,000 cc/min).

oxidation to NO2 by O3, which depends on the NO feed concentration and the total gas flow rate (Figure 4). UV

3O2 (g ) ⎯→ ⎯ 2O3(g )

(4)

NO(g ) + O3(g ) → NO2 (g ) + O2 (g )

(5)

Subsequently, cracking and partial/complete oxidation of HC, as confirmed in Figure 2, proceed due to the bifunctional UV radiation at λ = 254 and 185 nm, respectively [reactions 6 and 7]. The effect of H2O on the OHC production performance of the PCR has been examined in a separate experiment. Results shown in Table 3 reveal that the presence of H2O in the feed stream promotes the OHC production in both the blank and the TiO2 PCR. This beneficial effect of H2O on the OHC production in the PCRs can be explained by the participation of hydroxyl radicals produced according to the reactions 8 and 9 in the partial oxidation of HCs through the free radical reaction mechanism [reaction 10].27 The (HO• +NOx) reaction [reaction 11] appears negligible in the present reaction system, in view of the negligible NOx conversion over the blank PCR as listed in Table 1. On the other hand, the performance of the Ag/Al2O3 PCR for both the OHC yield and the NOx conversion has been negatively affected by the presence of 10% H2O (Table S3, Supporting Information), probably due to the competitive adsorption of H2O with other reactants on active adsorption/reaction sites. This may be part 3661

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

Article

Particularly, the CuCoY in the rear bed of the HC/SCR reactor reduces more NOx through the low temperature SCR of NOx by NH3 [reaction 17] produced by reaction 16. Note that NH3 and −NH2 species can be oxidized as well to N2 over the CuCoY catalyst [reaction 18].30

Table 3. Effect of H2O on OHC Production in the Blank and the TiO2 PCRs type of PCR blank

feed concentration of H2O (%)

yield of OHCs (%)a

0 2.5 10 7 10

36 42 45 72 84

b

TiO2c

NH3/−NH 2(ad) + O2 (ad) → N2(g )

Reactions 4−11 are induced primarily by the UV radiation, while the catalysts in both the PCR and the HC/SCR reactor participate in reactions 12−18. Summarized in Figure 5 are the detailed reaction pathways for the NO reduction in the PCR+HC/SCR system, which is

a Yield of OHCs (on C 1 basis) = Conc. of (Alcohols + Aldehydes)outlet/Conc. of Dodecaneinlet. bBlank PCR contained no catalyst. cTiO2 PCR contained (2BaY+TiO2) catalyst. Feed gas composition: 134 ppm C12H26, 0−10% H2O, air balance. Total flow rate = 500 cc/min.

of the reason behind the lower OHC yield of the Ag/Al2O3 PCR than that of the blank PCR as shown in Table 1. For the same feed concentration of H2O at 10% in Table 1, the greater beneficial effect of H2O on the OHC yield (84%) of the TiO2 PCR compared with that (45%) of the blank PCR can be explained as follows. When the TiO2 catalyst is irradiated by UV light, electrons (e−) and holes (h+) are generated in TiO2.28 On the other hand, surface hydroxyls are formed from the adsorbed H2O on the TiO2 surface. Some surface hydroxyls on TiO2 may then be used for adsorption sites for HCs, while others may interact with e− and h+ to form hydroxyl radicals that can enhance the formation of OHC from HC [reaction 12].29 This may also explain the higher OHC yield of the V2O5/TiO2 PCR than that of the blank PCR. The partially oxidized HCs (i.e., OHCs) by O3, HO•, NO2, and O2 [reactions 6, 10, 12−14] react with NOx (i.e., NO and NO2) to produce organo-nitrogen compounds such as −CN, −NCO, and amines as well as NH3 [reaction 16] via R−NO2 and/or R−ONO [reaction 15], which may further react with NOx to produce N2 over suitable catalysts in the low temperature region [reaction 17].6,7,10 Here the gaseous (g) and adsorbed (ad) states of an identical molecule are assumed to be in thermodynamic equilibrium of adsorption. HCs(g ) + O3(g ) → OHCs(g ) + O2 (g )

(6)

OHC(g ) + O3(g ) → CO(g ) + CO2 (g ) + H 2O(g )

(7)

H 2O(g ) + O3(g ) → 2HO•(g ) + O2 (g )

(8)

Figure 5. Reaction pathways for the deNOx process over conventional HC/SCR (dotted lines) adapted from Eränen and Burch et al.6,10 and over the PCR+HC/SCR hybrid system (both solid and dotted lines).

an extension of the deNOx reaction mechanism of the conventional HC/SCR system (dotted line) proposed by Eränen and Burch et al.6,10 Note in Figure 5 that the solid lines represent reaction pathways induced by the PCR during the early stage of the HC/SCR in the present hybrid system, which are faster processes relative to the concurrent conventional HC/SCR processes (dotted lines). These relatively fast processes enabled by the PCR during the early stage of the HC/SCR are the key factor in lowering the reaction light-off temperature of NOx conversion from 300 °C in the absence of the PCR to 200 °C in the presence thereof, as shown in Figure 2. It is important to note that the enhancement of the HC/SCR reaction by the PCR occurs in the early stage of the HC/SCR process, particularly before the formation of RNO2 and RONO, as can be seen in Figure 5. Once RNO2 and RONO are formed, the subsequent reactions of these species follow the conventional HC/SCR reaction pathways proposed previously,6,10,15,31,32 as shown in Figure 5. That is, the reaction intermediates decompose to −NCO which reacts with H2O to form NH3, followed by its reaction with NOx to form N2. Durability and Regeneration of PCR+HC/SCR System. Given the high deNOx performance of the PCR+HC/SCR hybrid system, it is desirable to investigate the durability of catalysts as well. As shown in Figure 6 (a), the two hybrid systems containing different PCR catalysts, Ag/Al2O3 and V2O5/TiO2, are stable up to 12 h on stream. However, the V2O5/TiO2 PCR+HC/SCR system starts to deactivate there-

UV

H 2O(g ) ⎯→ ⎯ HO•(g ) + H•(g )

(9)

HO•(g ) + HC(g ) → OHC(g )

(10)

HO•(g ) + NOx(g ) → HNOx(g )/HONOx(g )

(11)

HO•(ad) + HC(ad) → OHC(ad)

(12)

HCs(ad) + NO2 (ad) → OHCs(ad) + NO(ad)

(13)

HCs(ad) + O2 (ad) → OHCs(ad)

(14)

OHCs(ad) + NOx(ad) → R−NO2 /R−ONO(ad or g ) (15)

R−NO2 /R−ONO(ad) → −CN/− NCO/− NH 2 /NH3(ad)

(18)

(16)

NOx(ad) + − CN/− NCO/− NH 2 /NH3(ad) → N2(g ) (17) 3662

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

Article

Figure 6. Durability test over the PCR+HC/SCR systems at 250 °C for 24 h on stream: (a) NOx conversion performance of a blank PCR, Ag/Al2O3 PCR, and V2O5/TiO2 PCR coupled with an HC/SCR; (b) conversion of C12H26 and NOx or yield of CO2 over blank PCR+HC/SCR during the regeneration process at 250 °C. GHSV of the HC/SCR reactor = 16,500 h−1. Regeneration of deNOx performance (●) over the blank PCR+HC/ SCR at 250 °C by increasing O2 concentration in the reaction feed stream. Feed gas composition: 200 ppm NO, 134 ppm C12H26, 6% or 15% (regeneration) O2, 10% H2O, and N2 balance. GHSV of HC/SCR monolith reactor = 16,500 h−1.

this work has demonstrated its potential as a promising alternative to the current urea/SCR and LNT technologies. A key factor in the successful demonstration of this new concept of an advanced deNOx process is undoubtedly the unique design and functionality of the PCR. Major results obtained in this study include the following: 1. The PCR (containing catalysts) is very efficient for both OHC formation and NOx reduction due to its dual function: UV-induced OHC formation and NOx conversion over the in situ catalysts. 2. The blank PCR (containing no catalyst) is very efficient for oxidation of both NO to NO2 and HC to OHC; it is inefficient for NOx conversion due to its lack of catalyst. 3. The Ag/Al2O3 PCR (containing BaY+Ag/Al2O3) can produce both OHC and NH3 as reaction intermediates which can be used subsequently for further conversion of NOx in a downstream reactor containing a dual-bed catalyst such as Ag/ Al2O3 (for OHC/SCR) and CuCoY (for NH3/SCR). Building on this successful feasibility study, we expect to make further improvement in the overall efficiency of the PCR +HC/SCR system by optimizing catalyst formulations and operating conditions as well as by adopting more energyefficient photonic devices along with the design optimization of the PCR system in near future. In view of the shortcomings inherent in the present PCR system, further investigation is clearly warranted to make it more practical in the application to the automotive exhaust after-treatment system.

after, while the Ag/Al2O3 PCR+HC/SCR system remains stable up to 24 h. On the other hand, the blank PCR+HC/SCR system shows a rapid decrease of the deNOx activity with the increasing time on stream, due to the deactivation of the HC/ SCR catalyst by HC slipped from the blank PCR in the upstream.16,17 This is consistent with the large amount of residual HCs emitted from the blank PCR, as shown in Figure 2. Although only slight deactivation has been observed over the Ag/Al2O3 PCR+HC/SCR hybrid system during the first 24 h on stream, it is hard to guarantee that this system will maintain its deNOx performance for its long-term operation. Thus, it is highly desirable to develop a method of either preventing the deactivation of the catalyst or regenerating the deactivated catalyst. One way to prevent the deactivation of the HC/SCR catalyst is to burn off the HC coking to CO2 with a high dose of O3 and NO2 that can be generated by increasing the feed concentration of O2 to the PCR as shown in Figure 4, since both O3 and NO2 are stronger oxidizers than O2.25,33 As shown in Figure 6 (a), the blank PCR+HC/SCR system after severe catalyst deactivation by coking begins its regeneration due to the increased O2 feed concentration from 6% to 15% after 13 h on stream. In about 11 h, the NOx conversion is recovered to around 80%, indicating that the rate of the regeneration is about the same as that of the deactivation of the HC/SCR catalysts. The conversion of C12H26 and the yield of CO2 during regeneration of the deactivated HC/SCR catalyst are also increased by changing the feed concentration of O2 from 6% to 15%, as shown in Figure 6 (b). On increasing the O2 feed concentration to 15%, the conversion of C12H26 starts to increase due in part to its increased oxidation, with a concomitant increase of the CO2 yield. This is an indirect evidence for the oxidation of HC coking deposited on the surface of the HC/SCR catalysts, thereby regenerating the deNOx activity of the HC/SCR catalysts as evidenced by the improving NOx conversion. This finding suggests that a desired durability of the PCR+HC/SCR hybrid system can be achieved by optimizing the feed concentration of O2 to the PCR along with the optimum timing of the catalyst regeneration. The new photocatalytic HC/SCR system, a combination of the conventional HC/SCR with photocatalysis, developed in

■

ASSOCIATED CONTENT

S Supporting Information *

List of catalysts and their main functions, simulated fluidizedbed PCR and efficiency of UV irradiation over PCR, equivalent space velocity: monolith vs powder catalyst, calibration of FTIR spectra for quantitative gas analysis, comparison of reductant costs: the photocatalytic HC/SCR vs the urea/SCR, and effect of H2O on OHC production and deNOx activity over Ag/ Al2O3 PCR. This material is available free of charge via the Internet at http://pubs.acs.org. 3663

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664

Environmental Science & Technology

■

Article

Catalytic Materials To Obtain Broad Temperature Coverage for Plasma-Facilitated NOx Reduction. Catal. Today 2004, 89, 143−150. (15) Yu, Y.; He, H.; Feng, Q.; Gao, H.; Yang, X. Mechanism of the Selective Catalytic Reduction of NOx by C2H5OH over Ag/Al2O3. Appl. Catal., B 2004, 49, 159−171. (16) Schmieg, S. J.; Cho, B. K.; Oh, S. H. Selective Catalytic Reduction of Nitric Oxide with Acetaldehyde over NaY Zeolite Catalyst in Lean Exhaust Feed. Appl. Catal., B 2004, 49, 113−125. (17) Wen, B.; Yeom, Y. H.; Weitz, E.; Sachtler, W. M. H. NOx Reduction from Diesel Emissions over a Non-Transition Metal Zeolite Catalyst: Effect of Water in the Feed. Appl. Catal., B 2004, 48, 125− 131. (18) Houel, V.; Millington, P.; Rajaram, R.; Tsolakis, A. Fuel Effects on the Activity of Silver Hydrocarbon-SCR Catalysts. Appl. Catal., B 2007, 73, 203−207. (19) Houel, V.; James, D.; Millington, P.; Pollington, S.; Poulston, S. A Comparison of the Activity and Deactivation of Ag/Al2O3 and Cu/ ZSM-5 for HC-SCR under Simulated Diesel Exhaust Emission Conditions. J. Catal. 2005, 230, 150−157. (20) Som, S.; Longman, D. E.; Luo, Z.; Plomer, M.; Lu, T. Three Dimensional Simulations of Diesel Sprays Using n-Dodecane as a Surrogate. Fall Technical Meeting of the Eastern States Section of the Combustion Institute Hosted by the University of Connecticut, Storrs, CT, Oct 9−12, 2011. (21) Schmieg, S. J.; Blint, R. J.; Deng, L. Control Strategy for the Removal of NOx from Diesel Engine Exhaust using Hydrocarbon Selective Catalytic Reduction. SAE. No. 2008-01-2486 (2008). (22) http://www.thermoscientific.com/ecomm/servlet/ productsdetail?productId=13133056&groupType= PRODUCT&searchType=0&storeId=11152&from=search (accessed March 14, 2013). (23) Kelly, J. F.; Stanciulescu, M.; Charland, J.-P. Evaluation of Amines for the Selective Catalytic Reduction (SCR) of NOx from Diesel Engine Exhaust. Fuel 2006, 85, 1772−1780. (24) Cho, B. K.; Lee, J.-H.; Crellin, C. C.; Olson, K. L.; Hilden, D. L.; Kim, M. K.; Kim, P. S.; Heo, I.; Oh, S. H.; Nam, I.-S. Selective Catalytic Reduction of NOx by Diesel Fuel: Plasma-Assisted HC/SCR System. Catal. Today 2012, 191, 20−24. (25) Puri, I. K. The Removal of NO by Low-Temperature O3 Oxidation. Combust. Flame 1995, 102, 512−518. (26) http://www.jelight.com/ (accessed March 14, 2013). (27) Carra, S.; Santacesaria, E. Engineering Aspects of Gas-liquid. Catalytic Reactions. Catal. Rev. Sci. Eng. 1980, 22, 75−140. (28) Linsebigler, A. L.; Lu, G.; Yates, J. T., Jr. Photocatalysis on TiO2 Surfaces: Principles, Mechanisms, and Selected Results. Chem. Rev. 1995, 95, 735−758. (29) Lin, H.; Long, J.; Gu, Q.; Zhang, W.; Ruan, R.; Li, Z.; Wang, X. In Situ IR Study of Surface Hydroxyl Species of dehydrated TiO2: Towards Understanding Pivotal Surface Processes of TiO2 Photocatalytic Oxidation of Toluene. Phys. Chem. Chem. Phys. 2012, 14, 9468−9474. (30) Cho, B. K.; Baik, J. H. Bimetallic Catalyst for NOx Reduction in Engine Exhaust. United States Patent. Application number 20110200522 (2011). (31) Yu, Y.; Zhang, S.; He, H. Evidence for the Formation, Isomerization and Decomposition of Organo-Nitrite and −Nitro Species during the NOx Reduction by C3H6 on Ag/Al2O3. Appl. Catal., B 2007, 75, 298−302. (32) Yeom, Y. H.; Li, M.; Sachtler, W. M. H.; Weitz, E. A Study of the Mechanism for NOx Reduction with Ethanol on γ-Alumina Supported Silver. J. Catal. 2006, 238, 100−110. (33) Shrivastava, M.; Nguyen, A.; Zheng, Z.; Wu, H.-W.; Jung, H. S. Kinetics of Soot Oxidation by NO2. Environ. Sci. Technol. 2010, 44, 4796−4801.

AUTHOR INFORMATION

Corresponding Author

*Phone: 82-54-279-1325. Fax: 82-54-279-8299. E-mail: [email protected] (B.K.C.). Phone: 82-54-279-2264. Fax: 82-54-279-8299. E-mail: [email protected] (I.-S.N.). Present Address §

Department of Chemical Engineering, University of Michigan, 2300 Hayward St., Ann Arbor, Michigan 48109, USA. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

This work is part of the project, ‘‘National Research Foundation (No. 2012-0008674 ‘‘Emission Control Catalytic System for Next Generation Energy-Efficient Vehicle” and 2012R1A3A2048833 “Center for Ordered Nanoporous Materials Synthesis”) of Korea (NRF) grant” funded by the Korea government (MEST).

(1) Baik, J. H.; Yim, S. D.; Nam, I.-S. Modeling of Monolith Reactor Washcoated with CuZSM5 Catalyst for Removing NO from Diesel Engine by Urea. Ind. Eng. Chem. Res. 2006, 45, 5258−5267. (2) Yim, S. D.; Kim, S. J.; Baik, J. H.; Nam, I.-S. Decomposition of Urea into NH3 for the SCR Process. Ind. Eng. Chem. Res. 2004, 43, 4856−4863. (3) Forzatti, P.; Nova, I.; Tronconi, E. Enhanced NH3 Selective Catalytic Reduction for NOx Abatement. Angew. Chem., Int. Ed. 2009, 48, 8366−8368. (4) Park, S. J.; Ahn, H. A.; Heo, I.; Nam, I.-S.; Lee, J. H.; Youn, Y. K.; Kim, H. J. Hydrotalcite as a Support for NOx Trap Catalyst. Top. Catal. 2010, 53, 57−63. (5) Kim, C. H.; Qi, G.; Dahlberg, K.; Li, W. Strontium-Doped Perovskites Rival Platinum Catalysts for Treating NOx in Simulated Diesel Exhaust. Science 2010, 327, 1624−1627. (6) Burch, R.; Breen, J. P.; Meunier, F. C. A Review of the Selective Reduction of NOx with Hydrocarbons under Lean-burn Conditions with Non-zeolitic Oxide and Platinum Group Metal Catalysts. Appl. Catal., B 2002, 39, 283−303. (7) Kim, M. K.; Kim, P. S.; Baik, J. H.; Nam, I.-S.; Cho, B. K.; Oh, S. H. DeNOx Performance of Ag/Al2O3 Catalyst using Simulated Diesel Fuel−Ethanol Mixture as Reductant. Appl. Catal., B 2011, 105, 1−14. (8) Thibault-Starzyk, F.; Seguin, E.; Thomas, S.; Daturi, M.; Arnolds, H.; King, D. A. Real-Time Infrared Detection of Cyanide Flip on Silver-Alumina NOx Removal Catalyst. Science 2009, 324, 1048−1051. (9) Kim, M. K.; Kim, P. S.; Cho, B. K.; Nam, I.-S.; Oh, S. H. Enhanced NOx Reduction and Byproduct Removal by (HC+OHC)/ SCR over Multifunctional Dual-Bed Monolith Catalyst. Catal. Today 2012, 184, 95−106. (10) Eränen, K.; Klingstedt, F.; Arve, K.; Lindfors, L.-E.; Murzin, D. Y. On the Mechanism of the Selective Catalytic Reduction of NO with Higher Hydrocarbons over a Silver/Alumina Catalyst. J. Catal. 2004, 227, 328−343. (11) Sultana, A.; Haneda, M.; Fujitani, T.; Hamada, H. Influence of Al2O3 Support on the Activity of Ag/Al2O3 Catalysts for SCR of NO with Decane. Catal. Lett. 2007, 114, 96−102. (12) Shimizu, K.-I.; Shibata, J.; Yoshida, H.; Satsuma, A.; Hattori, T. Silver-Alumina Catalysts for Selective Reduction of NO by Higher Hydrocarbons: Structure of Active Sites and Reaction Mechanism. Appl. Catal., B 2001, 30, 151−162. (13) Miessner, H.; Francke, K.-P.; Rudolph, R. Plasma-Enhanced HC-SCR of NOx in the Presence of Excess Oxygen. Appl. Catal., B 2002, 36, 53−62. (14) Rappé, K. G.; Hoard, J. W.; Aardahl, C. L.; Park, P. W.; Peden, C. H. F.; Tran, D. N. Combination of Low and High Temperature 3664

dx.doi.org/10.1021/es304188k | Environ. Sci. Technol. 2013, 47, 3657−3664