JOURNAL or CATALYSIS 148, 427-437 (i994)
`
`Kinetic Studies of Reduction of Nitric Oxide with Ammonia on
`Cu“-Exchanged Zeolites
`
`Takayuki Komatsu, Makoto Nunokawa, Il Shik Moon,‘ Toshiya Takahara,
`Seitaro Namba} and Tatsuaki Yashima
`
`Department of Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo 152, Japan
`
`Received May I7, 1993; revised April 18, I994
`
`The selective catalytic reduction of nitric oxide with ammonia
`in the presence of oxygen has been studied on Cu“ ion-exchanged
`zeolite catalysts. In the case of Cu“-exchanged ZSM-5, the reac-
`tion proceeded selectively at lower temperatures, where nitric oxide
`and ammonia reacted with one-to-one stoichiometry to produce
`nitrogen. At higher temperatures, an oxidation of ammonia with
`oxygen occurred concomitantly to reduce the conversion of nitric
`oxide. The formation rate of nitrogen was first-order with respect
`to the partial pressure of nitric oxide, nearly half-order to that of
`oxygen, and zero-order to that of ammonia. The apparent activa-
`tion energies were almost the same for all Cu—ZSM-S examined
`having different Cu“-exchange levels and Si/Al atomic ratios.
`The ideality of Cu“-exchange ascertained from the stoichiometry
`of ion-exchange for the exchange levels less than 200% suggested
`the atomic dispersion of copper species. The specific activity per
`Cu“ ion increased with increasing ion-exchange level from 29 to
`195% and also with decreasing the Si/Al atomic ratio from 71 to
`23. The specific activities of various Cu—ZSM-S were found to
`depend on the concentration of the Cu“ ions; i.e., the higher the
`copper concentration, the higher the specific activity. The active
`copper species were proposed to be paired Cu“ species in view
`of the relation between the specific activity and the copper concen-
`tration.
`o 1994 Academe Press, Inc.
`
`INTRODUCTION
`
`Copper is one of the promising elements which have
`the ability to activate nitric oxide for the reduction with
`ammonia or hydrocarbons and also for the decomposition
`into nitrogen and oxygen. Seiyama et al. (1) have exam-
`ined various transition metal cation exchanged Y zeolites
`for the selective catalytic reduction (SCR) of nitric oxide
`with ammonia in the absence of oxygen and found that
`Cu2*—NaY is the most active catalyst. They also exam-
`
`’ Present address: Department of Chemical Engineering, College of
`Engineering, Suncheon National University, Suncheon. Chonnam,
`Korea 540-742.
`2 Present address: Department of Materials, The Nishi-Tokyo Univer-
`sity. Uenohara-machi, Kitatsuru-gun, Yamanashi 409-01. Japan.
`
`ined the effects of coexisting gases (2). The presence of
`oxygen enhanced the activity, while H20 and S02 had
`poisoning effects. Cu“-exchanged mordenite has also
`been reported (3, 4) to be active for the SCR of nitric oxide
`with ammonia in the presence of oxygen. The activity
`was much higher than that of Fe3*- or proton—exchanged
`mordenite. Iwamoto er al. (5) have found that Cu-NaY
`is also active for the decomposition of nitric oxide into
`nitrogen. The decomposition activity was very low for
`Cu-NaY with Cu“-exchange level less than 40%, but
`above 40% it increased markedly with Cu2*—exchange
`level up to 73%.
`ZSM-5 is an unique zeolite, that is, Cu“ cations can be
`introduced by an ion-exchange method with the exchange
`level exceeding 100%, while a similar excess in the ex-
`change level has been reported for Cu—Y (6). Although
`the definitive mechanism for the overexchange in ZSM-
`5 is not clarified yet, Cu“-exchanged ZSM-5 has been
`studied intensively as a catalyst for the removal of nitro-
`gen oxides. Cu“-exchanged ZSM-5 zeolites were re-
`cently found to have the highest activity so far docu-
`mented for the catalytic decomposition of nitric oxide into
`nitrogen and oxygen (7). The formation of nitrogen took
`place at the Cu“-exchange level of 40% or more and the
`yield of nitrogen increased with Cu” * -exchange level even
`when the exchange level exceeded 100% (8). Li and Hall
`(9) have also measured the activity for the decomposition
`of nitric oxide on Cu—ZSM-S having various Cu loadings
`and found that the turnover frequency calculated from
`the total amount ofCu increased with increasing the Cu“ —
`exchange level from 72 to 114%. The same catalyst,
`Cu—ZSM-S, has been found to be active for the reduction
`
`of nitric oxide using hydrocarbons as reductants (10-15).
`When ethene was used as a reductant (l2), the activity
`of Cu-ZSM-5 depended on the Cu“-exchange level; that
`is, the activity increased with increasing the exchange
`level, went through the maximum at around 100% ex-
`change level, and then decreased, which was different
`from the dependence in the decomposition of nitric oxide
`
`427
`
`002i-9517/94 $6.00
`Copyright © 1994 by Academic Press, Inc.
`All rights of reproduction in any form reserved.
`
`Exhibit 2030.001
`
`Exhibit 2030.001
`
`

`
`428
`
`KOMATSU ET AL.
`
`as mentioned above. The reduction of nitric oxide with
`
`Procedure
`
`some oxygenate compounds,
`acetoaldehyde, was
`also
`catalysts (16).
`We have reported (17) that H—ZSM-5 is acive for the
`selective reduction of nitric oxide with ammonia in the
`
`such as propanol and
`studied
`on Cu-ZSM-5
`
`presence of an excess amount ofoxygen, which is a similar
`result to the same reaction on H—mordenite reported ear-
`lier (18). In this study, we applied Cu“-exchanged zeolite
`catalysts to the same reaction. Our main purpose is to
`clarify the effect of copper concentration in zeolites on the
`activity for the selective reduction of nitric oxide based on
`the kinetic results.
`
`EXPERIMENTAL
`
`Catalysts
`
`ZSM-5 zeolites were synthesized hydrothermally (19)
`to have Si/Al atomic ratios of 23, 45, and 71. They were
`transformed into H—ZSM—5 by the usual
`ion-exchange
`method using NH4NO, aqueous solutions followed by
`the calcination in air at 773 K. Na—mordenite (Nippon
`Kagaku Kogyo, Si/Al = 6.4) and Na—Y (Tosoh, Si/Al =
`3.6) were transformed into the proton forms by a similar
`ion-exchange method. Dealuminated H—mordenite (To-
`soh, Si/Al = 67) was used as supplied. Cupric ions were
`loaded on these zeolites by the ion-exchange method using
`aqueous solutions (pH = 5) of copper acetate, Cu(CH3
`COO)2, followed by washing with pure water, drying in
`air at 403 K for 12 h, and calcination in air at 773 K for
`4 h to obtain Cu-ZSM-5 (Cu—Z), Cu—mordenite (Cu—M),
`and Cu—Y. Cu“-loading was controlled by changing the
`temperature of the ion-exchange (298 or 343 K), the con-
`centration of copper acetate (2.0 X 10*‘ or 4.0 X l0’5
`mol cm’3), the number of exchange-decantation cycles
`and the exchange time. Though these procedures did not
`provide the strict reproducibility in the loading of copper,
`we varied the loading by trial and error from 0.29 to 2.0
`wt% for Cu—Z, from 0.64 to 3.7 wt% for Cu—M, and from
`1.4 to 5.2 wt% for Cu—Y. These loadings correspond to
`the Cu“-exchange levels of 29—l95% (Cu—Z), 56—l36%
`(Cu—M), and 17-65% (Cu—Y), respectively. Omitting the
`washing after Cu“ ion-exchange from the procedure
`above, which was a kind of impregnation, resulted in
`higher copper loadings on ZSM-5 up to 2.8 wt%, corre-
`sponding to 275% apparent exchange level. The concen-
`trations of copper and aluminum in zeolites were deter-
`mined by atomic absorption spectrophotometry. The
`catalysts were pressed and broken into grains of 20-28
`mesh. We
`express
`the
`catalysts
`as Cu(x)—Z(y),
`Cu(x)-M(y), and Cu(x)—Y(y), where x and y represent
`the percentage degree of Cu2*ion—exchange and the Si/
`Al atomic ratio of zeolites, respectively.
`
`The reduction of nitric oxide with ammonia was carried
`
`out using a conventional flow reaction system under atmo-
`spheric pressure with a quartz reactor of 17-mm i.d. A
`specific amount of catalyst put in the reactor was heated
`in flowing helium from 298 to 773 K in 1 h and kept at
`the temperature for 1 h to remove adsorbed water. Then
`a mixture of NO, NH3, and 02 was fed with helium carrier
`onto the catalyst at a specific reaction temperature. The
`standard concentrations of reactants were 0.10% for both
`
`NO and NH3 and 2.0% for 02. The weight of catalyst
`(0.027—0.76 g) and the total flow rate (2.1 X 10“—4.0 X
`10‘ cm3 h" were adjusted to get W/F of 1.0 X 10‘°—
`2.5 x I0‘-‘g h cm'3 (GHSV of2.9 X 104-7.] x 105 h").
`The effluent gas was analyzed by GC (Tyran M-200) with
`Molecular Sieve 5A and Plot Q columns for N2, N20,
`and 02, by an NO_, analyzer (Yanaco ECL-77A) with
`a chemiluminescent detector for NO and N02, and by
`colorimetry using Nessler’s reagent for ammonia. Unre-
`acted ammonia was collected by passing the effluent
`through 2 X l0’5 mol cm‘3 hydrochloric acid before it
`was introduced into the NO, analyzer.
`
`RESULTS AND DISCUSSION
`
`We carried out the reaction on various Cu“-exchanged
`zeolites. To study the effect of copper concentration,
`ZSM-5 is an useful zeolite because the aluminum concen-
`
`tration can be varied widely by the hydrothermal synthe-
`sis without any severe dealumination treatments and be-
`cause the apparent Cu“-exchange level can be varied up
`to 200%. Therefore, we first concentrated on the results
`obtained for Cu-ZSM-5. The reaction was carried out on
`
`Cu—Z having various Cu2*—exchange levels and various
`Si/Al ratios at temperatures of 373-873 K. At these tem-
`peratures, the activity and selectivity of these catalysts
`did not change for more than 48 h of process time. Reac-
`tion products containing an N—atom were N2, N02, and
`N20. The amount of N02 was negligibly small for all the
`catalysts. Figures la and lb show the effect of the reaction
`temperature for Cu(79)—Z(45) and Cu(147)—Z(45), respec-
`tively, obtained at W/F of 3.3 X 10" g h cm‘3. Though
`these catalysts had different Cu“-exchange levels, below
`and above 100%, the curves showed a similar dependence
`on the reaction temperature. Figure 1c shows the results
`obtained on H—ZSM-5 (Si/Al = 45). The conversion of
`nitric oxide was much lower than that on the Cu—Z cata-
`
`is clear that the reduction of nitric oxide on
`It
`lysts.
`Cu-ZSM-5 was catalyzed by the copper ions exchanged
`into ZSM-5. Protons themselves and a trace amount of
`
`iron impurity, if any, are not responsible for the activity
`of Cu-ZSM-5. A possibility for the protons to affect the
`activity of copper ions was examined by comparing the
`
`Exhibit 2030.002
`
`Exhibit 2030.002
`
`

`
`REDUCTION OF NO WITH NH; ON Cu-ZEOLITES
`
`429
`
`100
`
`100
`
`80
`
`60
`
`40
`
`20
`
`
`
`Conversion,Yield(1)
`
`600
`
`700
`
`800
`
`900
`
`Reaction temperature (K)
`
`100
`
`80
`
`
`
`
`
`Conversion.Yield(X)
`
`80
`
`60
`
`40
`
`aoo
`
`800
`760
`600
`500
`Reaction temperature (K)
`
`9004
`
`
`20
`
`
`60
`
`
`
`Conversion.Yield(2) 20
`
`40
`
`T7300
`[woo
`Reaction temperature (K)
`FIG. 1. Reduction of nitric oxide with ammonia in the presence of oxygen on Cu(79)—-Z(45) (a). Cu(147)—Zl45) (b), and l-l—Z(45) (c). Conversion
`of nitric oxide (:1) and ammonia (A) and yield of nitrogen (0) were measured at W/F = 3.3 X l0“’ g h cm”. Concentrations of reactants were
`0.10% (both NO and NH3) and 2.0% (03) in He carrier.
`
`500
`
`activity of Cu3*(99%)-exchanged H—ZSM-5 (Si/Al = 45)
`with that of Cu“(l00%)-exchanged Na—ZSM-5 (Si/Al =
`45). NO—conversions and N2-yields were 32.4 and 34.9%
`for the former and 35.5 and 34.0% for the latter, indicating
`that the protons on the cation sites do not have any sig-
`nificant effect on the reduction of nitric oxide catalyzed
`by the copper ions.
`The possibility for the decomposition of nitric oxide
`into nitrogen and oxygen was examined under the same
`reaction conditions as used in the SCR of nitric oxide
`
`without feeding ammonia on the same Cu—Z catalysts
`used in Figs. Ia and lb. Nitric oxide reacted with oxygen
`at 473 K and above. The conversion of nitric oxide in-
`
`creased with the temperature and reached the maximum
`at 673 K, then decreased at higher temperatures because
`of the equilibrium of the oxidation of NO into N02 (20).
`The maximum conversions on these catalysts were about
`25%. The N-containing product was almost exclusively
`N02; N2 was only detected in a negligible amount at any
`temperature. Iwamoto et al. (8) studied the effect of coex-
`
`isting oxygen on the decomposition of nitric oxide using
`Cu—ZSM-5 catalysts with various Cu“-exchange levels.
`They found that the decomposition was inhibited by oxy-
`gen and that the conversion of nitric oxide into N2 was
`only 5% on Cu—ZSM-5 of 152% ion-exchange level under
`the following reaction conditions: T = 773 K; NO, 0.10%;
`03, 0.50%; W/F = 8.3 X l0‘5 g h cm”. In our case,
`W/F was one order of magnitude smaller and the partial
`pressure of O: was four times higher. Therefore,
`it
`is
`reasonable that
`the decomposition of nitric oxide on
`Cu—ZSM-5 could be neglected under our reaction condi-
`tions. Kinetic studies by Li and Hall (9) also revealed the
`inhibiting effect of oxygen on the decomposition of nitric
`oxide on Cu(l66%)—ZSM-5. They found that the rate of
`decomposition fell off at the inverse of (1 + K[O2]"2),
`where K, the equilibrium constant for O2 adsorption, in-
`creased with decreasing reaction temperature. This again
`supports the idea that, in our case, the inhibition by oxy-
`gen is serious at as low as 573 K to observe the negligible
`decomposition.
`
`Exhibit 2030.003
`
`Exhibit 2030.003
`
`

`
`430
`
`KOMATSU ET AL.
`
`In the case of the reduction of nitric oxide with ammonia
`
`on Cu(79)—Z(45), the amount of N20 formed as well as
`that of NO2 were negligible. The lack of N20 formation
`is in contrast with the results obtained with platinum foil
`catalysts (21), where N20 was produced in a N2/N20 ratio
`of about 2 at 573 K from the mixture of NO, NH3, and
`02. As shown in Fig. la, the conversion of NO, the con-
`version of NH3, and the yield of N2 gave almost the same
`values at 623 K or lower. These observations indicate
`
`that the reduction of nitric oxide occurs selectively at the
`lower temperatures with the following overall stoichi-
`ometry:
`
`4N0 + 4NH, + 02-» 4N2 + 6H2O.
`
`[1]
`
`This is the same reaction stoichiometry that has been
`proposed for the SCR of nitric oxide on vanadia—titania
`(22), H—mordenite (3), and Cu—NaY (2).
`As shown in Figs.
`la and lb, NO—conversion, NH,-
`conversion, and N2-yield were different from each other
`at higher temperatures. The NH,-conversion was the
`highest, reaching almost 100% at above 700 K, while the
`NO~conversion was the lowest. Therefore, the selectivity
`for the formation of nitrogen, Eq. [1], decreased at higher
`temperatures. This could be explained by side reactions
`between NH; and O2:
`
`4NH, + 30,-» 2N2 + 6H2O,
`
`4NH, + 50,-» 4N0 + 6H2O.
`
`[2]
`
`[3]
`
`These side reactions lead to the difference in the conver-
`
`sions and yield, i.e., the high NH3—conversion and the
`low NO—conversion compared with the N2-yield. In order
`to confirm the contribution of these reactions, the oxida-
`tion of ammonia with oxygen was carried out under the
`same reaction conditions as for the SCR without adding
`nitric oxide in the feed gas. Figure 2 shows the change
`in NH3-conversion with the reaction temperature ob-
`tained
`on
`Cu(79)—Z(45),
`Cu( l47)—Z(45),
`and
`Cu(l78)—Z(45). In all cases, N-containing products were
`N2, N0, and N20. As the yields of both NO and N20 were
`less than 1%, the conversion curves in Fig. 2 represent the
`yields of N2. Therefore, it is confirmed that the oxidation
`reaction, Eq. [2], is catalyzed by Cu—ZSM—5 and that it
`decreases the selectivity for the formation of nitrogen,
`Eq. [1]. If the two reactions, Eqs. [1] and [2], take place
`simultaneously, the difference between NH3-conversion
`and NO—conversion should correspond to the amount of
`consumed ammonia by reaction [2], where two molecules
`of N2 are produced from four molecules of NH3. There-
`fore, the yield of N2 should be at the middle of the two
`conversion values. The results in Fig. 1 at higher tempera-
`tures roughly followed this relation, which indicates the
`concomitant oxidation of ammonia into nitrogen, Eq. [2].
`
`100 ..
`
`YieldofN2(1)
`
`IVC
`
`40
`
`o 600
`Reaction temperature (K)
`
`700
`
`800
`
`900
`
`FIG. 2. Oxidation of ammonia on Cu(79)—Z(45) (O), Cu(l47)-Z(45)
`(A), and Cu(l78)—Z(4S) (V). A mixture of ammonia (0. 10%) and oxygen
`(2.0%) was fed with helium carrier at W/F = 3.3 x 10‘° g h cm".
`
`is clear that the activities of
`it
`1 and 2;
`From Figs.
`Cu(147)—Z(45) both for the SCR of nitric oxide (T § 573
`K) and for the oxidation of ammonia were higher than
`those of Cu(79)—Z(45). Higher activities of the excessively
`Cu“-exchanged (>l00%) ZSM—5 were also reported for
`the decomposition of nitric oxide (8, 9). In order to clarify
`the effect of the extent of Cu“-exchange on the SCR
`activity, kinetic studies were carried out. Figure 3 shows
`the effect of the concentration of reactants on the rate of
`
`N2 formation on Cu(147)—Z(45) measured at 573 K. At
`this temperature, the rate of N2 formation represents the
`rate of the SCR of nitric oxide, Eq. [1], because NO-
`conversion and NH3-conversion were almost the same as
`shown in Fig.
`lb. The concentration of the specific re-
`actant was varied while keeping those of other reactants
`equal to the standard values (NO, 0.10%; NH3, 0.10%;
`and 02, 2.0%). From Fig. 3, it is apparent that the reaction
`rate depended positively on the concentration of nitric
`oxide (a) and oxygen (c). On the other hand, the concen-
`tration of ammonia exhibited a different effect, that is,
`
`the rate depended on NH3-concentration only at very
`low concentrations. At around the standard concentration
`
`(0.10%), the rate was almost independent of NH2-concen-
`tration, which indicates that the rate is zeroth—order in
`
`NH3—concentration. This suggests that most of the surface
`of Cu—ZSM-5 is covered with ammonia under the stan-
`dard conditions.
`
`The change in O2-concentration before and after the
`passage through the catalyst bed must be negligibly small
`under the standard conditions because the reactant mix-
`
`ture contains an excess amount of oxygen compared with
`nitric oxide and ammonia. Therefore, the reaction rate
`
`can be expressed only by the partial pressure of nitric
`oxide. Figure 4 shows the change in N2-yields with W/
`F measured at 573 K with Cu—Z having various Cu“-
`
`Exhibit 2030.004
`
`Exhibit 2030.004
`
`

`
`REDUCTION OF NO WITH NH; ON Cu-ZEOLITES
`
`431
`
`YieldofN2(X)
`
`80
`
`l\)AO‘ooo
`
`H/F (x 10-6 g n cm-3)
`
`FIG. 4. Change in the yield of nitrogen with W/F on Cu(29)—Z(45)
`(O), Cu(l20)—Z(71)
`(V), Cu(79)—Z(45)
`(O). Cu(79)—Z(23)
`(A). and
`Cu(l78)—Z(45) (0) measured at 573 K with the reactant containing NO
`(0.l0%), NH; (0.10%). and O; (2.0%).
`
`reaction rate (r) can be expressed as
`
`r = k[NO] = k’[O2]'"[NO],
`
`[4]
`
`where k and k’ are the rate constants. The value of k can
`
`be obtained from the slope of the first-order plot (Fig. 5).
`A similar first—order plot with respect to NO-concentra-
`tion was obtained for each data in Fig. 3c. Though each
`plot only had one point, the good linearity shown in Fig.
`5 made it reasonable to elucidate the k value from each
`
`slope. As the relation between k and k’
`pressed as
`
`is simply ex-
`
`k = /<'[02l’".
`
`[5]
`
`-ln(1-X/100)
`
`W/F (x 10‘5 9 h cm‘3)
`
`FIG. 5. First-order plot for the data of Fig. 4. Symbols represent
`the same catalysts as in Fig. 4.
`
` o
`
`F"
`
`0.05
`
`0.1
`
`0.15
`
`o.é
`
`Concentration of NO or NH3 (76)
`
`
`
`RateofN2formation(moig"n‘l)
`
` o
`
`30
`
`4
`
`"0
`
`50
`
`1
`
`(N2
`
`Concentration of 02 (2)
`FIG. 3. Effects of the concentrations of nitric oxide (a), ammonia
`(b), and oxygen (c) on the rate of nitrogen formation measured with
`Cu(l47)—Z(45) at 573 K and W/F = 3.3 X l0‘°g h cm‘). Concentration
`of a specific reactant was varied keeping the others constant at 0.10%
`(NO or NH3) or 2.0% (03).
`
`exchange levels and Si/Al ratios. W/F was varied by
`changing the weight of catalyst (0.044—0.76 g) and the
`total flow rate (2.1 X I0‘—4.0 X 10‘ cm3 h"). Curvatures
`were found for most of the catalysts, indicating that the
`reaction took place out of the differential conditions. The
`results were then treated assuming that the reaction rate
`had first-order dependence on the concentration of nitric
`oxide. The first-order plots using an integrated equation
`are shown in Fig. 5. Linear relations were obtained for
`all the catalysts, which confirms that the SCR of nitric
`oxide with ammonia in the presence of excess oxygen is
`the first-order reaction with respect to the concentration
`of nitric oxide under our reaction conditions. The first—
`
`order dependence on NO-concentration has been also
`reported (23) for the SCR of nitric oxide with ammonia
`on Cu“-exchanged mordenite.
`The reaction order with respect to the pressure of oxy-
`gen was treated as follows. From the above results, the
`
`
`
`Exhibit 2030.005
`
`Exhibit 2030.005
`
`

`
`432
`
`KOMATSU ET AL.
`
`60
`
`40
`
`20
`
`
`
`.0
`
`100
`
`200
`
`300
`
`Aooarent Cu3+-exchange level (2)
`
`‘T
`
`3
`-aX
`>.
`anL.
`8g,
`C
`.24-’
`as
`‘.3
`ID
`t,’
`anE
`Q)L
`(U
`
`Q 2
`
`FIG. 7. Relation between the apparent activation energy and Cu“-
`exchange level for Cu-Z(23) (A), Cu—Z(45) (O), and Cu—Z(7l) (V).
`
`plained by the similarity in the nature of active species
`between these two catalysts with the only difference in
`number or concentration of the active sites. A similar
`
`activation energy (49 kJ mol‘ ') was reported for the SCR
`of nitric oxide with ammonia in the presence of oxygen
`on the V205 catalyst (24), where V5*=O species on V205
`was responsible for the reduction activity. In the case of
`copper catalysts, Cu=O species can hardly generate at
`the exchange site of zeolites or on the surface of copper
`oxide particles. Therefore, the reaction mechanism proba-
`bly is different from that on the V205 catalyst though the
`activation energies were nearly the same.
`The homogeneous nature of the active site of various
`Cu—Z indicated by the similarity in the activation energy
`may result from the homogeneity in the Cu“-exchange
`behavior even above 100% exchange levels. We studied
`on the details of Cu“-exchange first by exchanging Na*
`in ZSM-5 by Cu“. The ion-exchange was carried out for
`Na—ZSM-5(Si/Al = 46) at 303 K using 1.7 X 10‘5 mol
`cm” aqueous solution of copper acetate (pH = 5) with
`varying the exchange period to get various copper load-
`ings. Figure 8a shows the relation between the amount
`of incorporated Cu“ and that of remaining Na" in ZSM-
`5, both expressed by the atomic ratio based on the amount
`of aluminum. The amount of Na" (Na*/Al) decreased
`almost
`linearly with increasing the amount of Cu“
`(2Cu2*/Al) with a slope of — 1 (dotted line) up to 2Cu2*/
`AI = 0.7. This may be explained by the simple exchange
`between two Na* ions and one Cu“ ion and also by that
`between two Na* and a pair of Cu(0H)‘” and H". The
`latter ion-exchange has been suggested for Cu—ZSM-5 (8,
`26) and Cu—Y (27-29). Further incorporation of Cu“ took
`place without a large decrease in Na*/Al, which did not
`change significantly above 2Cu2"/Al = 1.3. This phenom-
`ena can be explained by the exchange between one H*
`
`Exhibit 2030.006
`
`we can estimate the reaction order, m, with respect to
`O2-concentration from the slope of the In k vs In [02] plot
`(Fig. 6). The slope of this line was 0.60, which suggests
`that the SCR of nitric oxide is roughly half order with
`respect to the partial pressure of oxygen. In the case of
`the same reaction on V205 (24), the reaction rate increased
`with 02 concentration similarly to Fig. 3c except for the
`negligible increase at above 1% of O2 and the rate was
`zero order with respect to NH3-concentration and first
`order in NO-concentration. Arakawa et al. (25) have stud-
`ied the reduction of nitric oxide with ammonia in the
`
`absence of oxygen on Cu“-exchanged NaY at 383 and
`413 K. They found that the reaction rate was first-order
`in NO-concentration and nearly half—order in NH3-con-
`centration. Their results were obtained at lower tempera-
`tures and higher W/F than those used in our work. There-
`fore, the reaction mechanism may not be the same, which
`results in the different reaction order with respect to
`NH3—concentration.
`From the slope of each line in Fig. 5, the first-order
`rate constant, k, was obtained for each Cu—Z at 573 K.
`
`Apparent activation energies at 543-603 K were calcu-
`lated from Arrhenius plots using the k values obtained in
`a similar manner. Figure 7 shows the relation between
`the apparent activation energy and the Cu“-exchange
`level for Cu—Z with various Si/Al ratios. Although some
`variations existed in the activation energies, they did not
`show any significant dependence on the exchange level
`or on the Sil Al ratio of ZSM-5 expressed by the different
`symbols. It is suggested that the catalytic property of an
`individual active site is almost the same for all the Cu—Z
`
`catalysts tested. As shown in Figs. la and lb, the maxi-
`mum conversion of nitric oxide was the same for
`
`Cu(79)-Z(45) and Cu(l47)—Z(45) with a shift of the tem-
`perature at the maximum conversion. This may be ex-
`
`Ink
`
`-5
`
`22022-4
`
`-3
`
`in P(0z)
`
`FIG. 6. Relation between ln It’ and In P(O3) measured with
`Cu(l47)—Z(45) at 573 K and W/F = 3.3 X l0“’ g h cm”.
`
`Exhibit 2030.006
`
`

`
`REDUCTION OF NO WITH NH, ON Cu-ZEOLITES
`
`433
`
`QJ
`.‘;'
`'8(UN...
`:2'F\
`~ *2E-U-
`o
`-0-’C
`=
`E
`
`o
`
`0.5
`
`
`
`o
`
`1
`
`1.
`
`2
`
`in zeolite (2cu3"/A1)
`Amount of Cu?‘
`Ion-exchange behavior of ZSM-5 in the forward exchange
`FIG. 8.
`of Na—Z(46) by Cu“ (a) and the reverse exchange of Cu(l61)—Z(46) by
`Na‘ (b). Exchange was carried out at 303 K with 1.7 X l0‘5 mol cm‘3
`solutions of copper acetate (a) and sodium chloride (b).
`
`and one Cu(OH)*, where H* has already been incorpo-
`rated into ZSM-5 at the initial stage of the ion-exchange.
`Therefore, the exchange between two Na* and a pair of
`Cu(OH)* and H* should occur at the initial stage.
`To ascertain the ideality of this exchange and to check
`the formation of unexchangeable copper aggregates, the
`reverse exchange was carried out with Cu(l61)—Z(46) and
`an aqueous solution of sodium chloride (1.7 X 10'5 mol
`cm”) in a similar manner to the forward exchange men-
`tioned above. In Fig. 8. the relation between the amount
`of remaining Cu“ (2Cu2*/Al) and that of incorporated
`Na* (Na*/Al) in ZSM-5 is also shown. Na*/Al increased
`with decreasing 2Cu3*/Al, finally following the broken
`line with a slope of — 0.5. This broken line may represent
`the exchange between one Na* and one Cu(OH)* ap-
`proaching the complete exchange into Na—ZSM-5. The
`relatively steep increase in Na*/Al at the initial stage of
`the reverse exchange (2Cu“/Al from 1.6 to 1.3) may be
`explained by the exchange between two Na* and a pair
`of Cu(OH)* and H*; this exchange apparently corre-
`sponds to the stoichiometric exchange between two Na*
`and a Cu? *. We do not know the reason why this exchange
`takes place preferably to the one-to-one exchange indi-
`cated by the broken line. There should be no exchangeable
`protons remaining when the forward exchange has been
`achieved completely to get 200% exchange level. In the
`case of Cu(l61)—Z(46) used here, however, there could
`remain the exchangeable H * in the amount corresponding
`to ca. 40% exchange level. Although the reverse exchange
`did not proceed inversely through the same path as that
`of the forward exchange, the above results indicate that
`all the copper ions exist in Cu—Z as exchangeable counter
`cations. Therefore, copper species in the excessively ex-
`changed Cu—Z are dispersed atomically and would not
`form aggregates within the loading of 2Cu2*/Al < 2. The
`X-ray diffraction patterns of Cu—Z after the calcination,
`in fact, did not show the presence of large CuO particles.
`In order to clarify the active copper species for the SCR
`of nitric oxide, the effect of copper concentration on the
`
`activity of Cu“ ions was investigated. The specific activ-
`ity, k(Cu), was calculated by dividing the rate constant,
`k, obtained from the first-order plot (Fig. 5) by the total
`number of Cu ions in the catalyst. Figure 9 shows the
`relation between k(Cu) and the copper loading expressed
`by the corresponding Cu“-exchange level. In the case
`of Cu—Z having the same Si/Al ratio, k(Cu) increased
`with increasing Cu“-exchange level up to 200%, as indi-
`cated by three curves. The increase in k(Cu) was the most
`remarkable for Cu—Z(23), where k(Cu) increased from
`0.5 X 10'” to 3.2 X 10"5 cm3 h" (Cu-atom)” within
`a 110% increase in the exchange level. It should be noted
`that k(Cu) increased monotonously above the exchange
`level of 100% for Cu—Z(45) and Cu—Z(7l) up to 200%. In
`the case of Cu—NaY, Williamson and Lunsford (30) have
`studied the effect of Cu2 "-exchange level on the reduction
`of nitric oxide with ammonia in the absence of oxygen.
`They found that the activity increased linearly with the
`exchange level from 6.5 to 44%, indicating that the specific
`activity per Cu“ did not change. Seiyama et al. (1) have
`also studied the same reaction system and obtained a
`linear relation between the conversion of nitric oxide and
`
`the Cu“-exchange level from 7 to 79%. This is consistent
`with the results of Williamson and Lunsford (30) men-
`tioned above. The reason why the dependence of the
`specific activity in our system on the Cu2 *-exchange level
`was different from that in their system may be the differ-
`ence in the parent zeolites or the presence of oxygen in
`our system.
`In the case of Cu—ZSM-5, Iwamoto et al. have also
`reported the effect of Cu“-exchange level on the decom-
`position of nitric oxide (8) and on the reduction of nitric
`oxide with ethene (12). In the former reaction, the conver-
`sion of nitric oxide gave S-shaped dependence, where the
`
`U.)
`
`NJ
`
`k(Cu)(x10"'5cm3h'lCu-atoirl)
`
`0
`
`I
`
`21
`
`300
`
`Apparent cu?‘-exchange level
`
`(X)
`
`FIG. 9. Effect of Cu“-exchange level on the specific activity of
`Cu“. The activity was measured at 573 K on Cu—Z(23) (A), Cu—Z(45)
`(O), and Cu—Z(7l) (V) with the reactant mixture of NO (0.10%), NH,
`(0.10%). and 03 (2.0%).
`
`Exhibit 2030.007
`
`Exhibit 2030.007
`
`

`
`434
`
`KOMATSU ET AL.
`
`conversion of nitric oxide increased sharply above the
`exchange level of 40%. Though the increase in activity
`at the exchange level above 70% was not clear because
`the conversion approached 100%, at least in the lower
`exchange levels, specific activity ofCu ions for the decom-
`position of nitric oxide increased with increasing the ex-
`change level. Li and Hall (9) also reported that the specific
`activity (expressed by turnover frequency) for the decom-
`position of nitric oxide increased with copper exchange
`level from 72 to 114% on Cu—ZSM—5. These are similar
`
`phenomena as observed in our system. In the case of the
`reduction of nitric oxide with ethene (12), the conversion
`into nitrogen increased with Cu“-exchange level, but
`reached the maximum at around the exchange level of
`100%. The cause of this difference in the effect of Cu
`
`loading on the two reactions is not clear so far and needs
`further investigation.
`As shown in Fig. 9, when the Cu“—exchange level ex-
`ceeded 200%, the specific activity of Cu—Z(45) decreased
`significantly. Above 200% exchange level, Cu“ ions
`could not be located as countercations on cation exchange
`sites. These excess Cu“ ions may form an aggregate such
`as a small particle ofCuO. The decrease in specific activity
`indicates that copper species in such an aggregate are
`inactive or much less active than Cu“ present as a
`counter cation. In addition, such CuO species may block
`the pores of ZSM-5, which may also result in the decrease
`in activity.
`It is obvious from Fig. 9 that there is another factor
`which affected the activity; it is the Si/Al ratio. At the
`same exchange level, the lower the Si/Al ratio, the higher
`the specific activity. The lower Si/Al ratio corresponds
`to the higher aluminum concentration and also to the
`higher concentration of copper ions at the same Cu“-
`exchange level. Therefore, we considered the concentra-
`tion of copper ions to be a critical factor which governs
`the specific activity. As shown by the open symbols in
`Fig.
`l0, the k(Cu) values in Fig. 9 were plotted against
`the concentration of copper ions in Cu—Z except for
`Cu(265)—Z(45) and Cu(275)—Z(45), which could contain
`the aggregated copper species as mentioned above. In
`contrast with the results in Fig. 9, all the k(Cu) values
`followed the same relation independent ofthe Si/Al ratio,
`which was roughly expressed by a single curve in Fig.
`10. Accordingly, it is concluded that the specific activity
`is governed by the concentration of copper ions in
`Cu—ZSM—5.
`
`Next, we examined the catalytic properties of Cu“-
`exchanged mordenite and Y-zeolite to clarify the effect
`of zeolite structure on the specific activity of Cu“. Figure
`11 shows the effect of reaction temperature on the SCR
`of nitric oxide with ammonia using Cu(95)—M(67) (a) and
`Cu(l7)—Y(3.6) (b) as catalysts at W/F of 7.0 X l0“’ and
`6.0 x 10" g h cm‘3, respectively.
`In the case of
`
`LA)
`
`I\)
`
`k(Cu)(x10-15cm3n-1cu-atom-1) Q_n
`
`4
`Copper concentration (x 10' mo} 9")
`
`FIG. 10. Effect of copper concentration on the specific activity of
`Cu“. The activity was measured at 573 K on Cu-Z(23) (A), Cu—Z(45)
`(O), Cu—Z(7l) (V), Cu—M(6.4) (A), Cu—M(67) (V). and Cu—Y(3.6) (C).
`
`Cu(95)-M(67), the activity arose above 500 K and in-
`creased