- Email: [email protected]
Removal of NOx pollutant by catalytic combustion reactions
REMOVAL OF NO~ POLLUTANT BY CATALYTIC COMBUSTION REACTIONS F. SOLYMOSI AND J. KISS Gas Kinetics Research Group, The University, Szeged, Hungary t~iaetic measurements have been made for the oxidation of C0, C~H~, H2 and natural gas by nitric oxide on SnO~/CrzO3 and on AhOa/Cr~O~ catalysts. The interaction of NO with catalysts was examined by adsorption, infrared spectra and electric conductivity measurements. The adsorption isot3aermsof NO on SnO~/CrzO~ catalysts are of the Freundlich type in the pressure range from 3 to fi0 Tort. Eva/nation of the NO chemisorp~ion rate by the Elovich equation yielded two linear segments with a distinct break between ~Amm.The instantaneous adsorption of NO on the partially reduced She2 was larger than that of the unredoced sample. The rate of adsorption was ton times higher. A very large increase in the amount of adsorbed NO was obtained by adding Cr~O~ to SnO2. Electric conductivity measurements during NO ad.qorption revealed that both negatively and positively charged NO are present on pure and doped SnO~. The catalytic reactions on pure SnO~ occurred only above 360~ Small amounts of Cr20~ in the SnO~ lowered the reaction temperatares by 100 200r The effideney of SnOz/CrzOa exceeded that of AhOs/Cr~Oa. It is suggested that on the partially reduced Sn0~ the Sn3+ions are the active catalytic sites for NO and tha~ nitric oxide dissociates upon chemisorption. The slowest step of the raduc~ion of NO is assumed to be the oxidation of reduced centers on the catalyst, which involves the formation of nitrogen molecules. The high activity of catalysts containing chromium oxide is attributed to the chromium ions incorporated into the surface layer of the carrier which can easiJy be reduced by fnels and rapidly reoxidized by NO. Introduction In recent yearn, concern over our environmcnt has led to substantial action ia all h~dustrialized countries. A major environmental concern is that of air pollution. Catalytic combustion processes play an important role in combating environmental pollution caused by chemical plants and antemobiles. Thanks to the vast amount of basic research on the oxidation of CO arid hydrocarbons, the development of CO-CH catalysts for car exhaust gases is in an advanced stage. The state of development of NO catalysts, however, is far less advanced and requires more fundamental research. The simplest way for the removal of nitric oxide formed in various combustion reactions would be its catalytic decomposition. Unfortunately the decomposition rates of nitric oxide on different types of catalysts axe much too low to be of practical use)." The catalytic reduction of nitric oxide with CO offers an attractive means
for eliminating this pollutant from car exhaust. The study of this reaction has been the subject of many recent papers but only a few of them are concerned with the reaction mechanism,s,4 It was found that the oxidation and reduction of the catalyst surfaces play an important role in the catalytic reaction. Therefore our attention turned to catalyst systems that especially favor the redox reactions. A prominent catalyst in this respect is SnO~ containing a small amount of added chromium orSde. The prescott work was mldertaken to examine the interaction of nitric oxide with these catalysts and to investigate the kfiletics and mechanism of the catalytic reduction of nitric oxide by different reduch~g agents. Experimental Malerials CrzOa was reagent grade material from Reanal. It was fired a~ 700~ ~'or 5 h. A1203 was from
1233
COMBUSTION SYSTEMS
1234
Degnssa. The Al2Or-Cr~O~ catalyst was prepared by impregnating alumina with an aqueous solution of Cr(~. The gel was dried in air for 16 h at 120~ For catalytic investigations, the powder was calcined for 8 hours at 600~ SnO~ was obtained by the action of HNO3 on metallic Sn. It was dried at 120~ and heated at 350~ for 3 h and at 500~ for 5 h. The doping of the SnO~ with Cr2Oa was effected in the following way~.6:Cry03 was added to an aqueous suspension of Sn(h and after a sufficiently long period of stirring, the homogeneous suspension was dried and heated at 350~ for 5 h. Final sintering was performed at 900~ for 5 h in air. The nitric oxide (Matheson Company) was commercially pure (99%). Ethylene was a product of Fluka Gmbh. Both gases were purified by bulb-to-bulb distillation before use. Carbon monoxide was prepared in the laboratory by reaction between formic acid and sulphuric acid a t 83~ For the catalytic studies, small pellets (1.5 ram diam and 1.5 mm long) were made. A fixed amount of catalyst (1.5 g, 8-10 small pellets) was used in all experiments.
Methods of Measurements For the kinetic investigations, a closed circulation system (volume 277 ml) was used. The reactions were followed by measuring the pressure of the reacting gases. A Hewiett-Packard gas chromatograph connected to the reactor was used to analyze the reaction products. Conductivity measurements during the adsorption a n d catalytic reaction were performed in the same reactor, using pellets of 8 ram diameter and 6-7 nun long. An ac bridge was used for these measurements. The active oxygen content of the catalysts was determined b y the Bunsen method. Specific surface areas were calculated from the adsorption of nitrogen a t liquid nitrogen temperature. Adsorption measurements were performed in a Sartorius microbalance. Infrared spectra were taken with a double beam I R spectrophotometer (Speetromom 2000). The discs of AhOs/CraO~ were made by pressing the finely divided powder dried at 120~ to a thickness corresponding to 30 m g / m ~. The sample was degassed in vacuo at 10-4 TorT. Further treatment is described in the text. R e s u l t s and D i s c u s s i o n
Characterization of the Catalysts SnO~ is an n-type semiconductor. Its electric conductivity is due to donor levels t h a t arise from
defects such as impurity cations and anion vacancies.7,s The incorporation of a small amount of Cr2Oa into SilOs in air significantly changes the physical properties of SnO2.~'6 On the other hand, the chromium ions also undergo a noteworthy change; part of them is oxidized to a higher valence and stabilized in the surface layer of SnO~. A remarkable feature of higher valence chromium in the surface layer of Sn0~ is t h a t after its reduction with fuels it can easily be reoxidized even at low temperatures. Characteristic data for different catalysts used in this work are summarized in Table I.
Adsorption Measurements The adsorption of NO has been investigated on AhOa/C~08 catalyst,9 but no studies have been made on Sn02. The characteristic feature of the adsorption of NO on diamagnetic oxides is its extreme slowness; the establishment of the adsorption equilibrium is very difficult. The adsorption of NO on SnO2 activated at 400*C in vacuum is also slow, in spite of the fact t h a t there is no experimental evidence for a substantial energy of activation. Adsorption isotherms a n d rates for NO were measured at 25-100~ The data conform well to the Freundiich isotherms in the pressure range covered, as shown by a linear relationship between the logarithmic amount of adsorbed NO, q, and the logarithmic pressure, p. The Freundlich isotherm is defined by
q = cpl/n where the coefficients n and c are functions of temperature. The points of intersection of isotherms show t h a t the amount of NO required for a monolayer (q~) is 7.46 ~moles/g. The preSsure which is necessary to achieve monolayer coverage is 74.2 Torr. The average value of H,,, which is a measure of the heat of adsorption, is 6.9 kcal/mole. Similar low values have been obtained b y Shelef, et aL, on transitional metal oxides.0-a The rate of adsorption of NO was evaluated b y the integrated form of the Elovich equation q=
(2.3/~) ]og(~+~)-(2.3/a) 1og~
where tom 1/an is an integration constant. The Elovieh plots consist of two linear segments with a distinct break between them. A possible explanation of the break is t h a t two distinct populations of adsorption sites exist on the sur-
REMOVAL OF NO= POLLUTANT
1235
TABLE I Characteristic data for different catalysts
Catalyst
Cr2Oa content mole%
Highest calcination temperature ~
Surface area m2/g
Active oxygen %
Average valency of chromium*
Crd:)~ Sn02 SnO~ SnOs SnOz Sn02 AlsOs A12Oa
100 -0.01 O.1 0.5 1.0 1.0 5
700 900 900 900 900 900 900 600
30.4 5.81 0.12
0.0224 -0.0015 0.01 0.043 0.082
3 -4.5 3.95 3.81 3.77
1.0243
5
5.58
8.84 11.87 4.64 70.0
* Calculated on the basis of active oxygen content. t This sanlple of A120,was a Reanal product. The doping with Cr2Os was similar to the case of SnO,.
TABLE II Parameters for the adsorption of gases at 1O0~ on SnO2-bCr~O, catalysts 8nOs al El~nole/g] -~ NO O2 CO C~HI
7.5 1. 856 1.77 4.025
SuOz+ 1%Cr=O~ q0 t.~rnole/g 1.1 2. 875 0.4 0.714
~1 [~mole/g']-1
q0 ~anole
1.15 3. 072 0.621 2.3
3.3 3.25 1.66 0.71
q0ffitheamotmt of instantaneousadsorption,al is the slope of the first segment of the Elovlch plots. face. The temperature coefficient of the adsorption is positive; both the czl and a2 decrease with temperature. On the partially reduced Sn02 (with CO at 400~ for 60 rain) the instantaneous adsorption of N O was almost six times larger than that of adsorption on unreduced samples. In addition the rate of adsorption at i00~ was about ten times higher. The chemisorption of N O on unreduced SnOs is reversible, whereas on SnO~ reduced with fuels it was somewhat irreversible. In the latter case it is assumed that mainly oxygen remains on the surface, while N O is reduced to Ns0. A very large increase in the adsorbed amount of N O and in the rate of chemisorption was achieved by incorporating CrtOs (I mole%) into the surface layer of Sn0s (Table II). Similar measurements were carried out for the
adsorption of other gases, e.g., CO, 02, and CsH4. The histantaneous adsorption of CO is commensurable with that of N0. The Elovieh plots also consist of two intersecting straight line segments. While al decreases with temperature, a~ increases with T. If we compare the adsorption of NO and O~ on Sn02, we find that both the extent of instantaneous adsorption of O~ and the rate of subsequent adsorption are always larger than for NO. With the increase of adsorption temperature, this difference becomes greater. The initial adsorption of ethylene on SnOa below 100~ has the highest value of all the gases (fuels) studied so far, it is however hardly affected by the increase of temperature or by incorporating chromium into Sues. Accordingly, near the reaction temperature, the adsorption of 0s and NO greatly exceeds that of ethylene. Data for
1236
COMBUSTION SYSTEMS
observed between 1000-2000 cm-I when NO is added to alumina or silica are due to the presence of transitional metal impurities) ~,15 Unfortunately all our efforts to make transpareut discs from SnO2/Cr20a have been mtsueessful, restricting our infl'ared studies to the A1203/ Cr20a system. When NO was added to AhOa/Cr~O~, pretreated in CO (initial pressure 10 Torr) at 4O0~ for 1 h, two large bands appeared at 1865 and 1740 em-1 (Fig. 1). The latter is perceptibly composed of a combination of separate bands at 1743 and 1738 cm-1. Both bands have been observed recently during the adsorptioo of NO oa Al~Oz/Cr~Oaactivated in ethylene2 ~ Based orl the assignments of NO bands observed on transition metal oxidesy ,J7 the band at 1743 and 1738 em-1 3000 2000 1800 1600 1400 1200 cm"1 was attributed to NO cnvalently bound to Cr ~+ and Cr~+ sites, respectively. The absorption band Fzo. 1. Infr~ed epeetra of NO on AhO#Cr~Os. at ~ 1 8 6 0 cmy2 was assigned to NO chelrJsorbed Adsorption of NO (5 Torr) was carried out, at 25~ by a double bond ionic link (Cr~+--N+--O) to for 1 h, thee the sample was evacuated and heated chromium ions Cr <:+n+. However, the observaat diffe~'en~bemperatures for 30 rain. All the spectra tion that tl~e intensities of both bands increased were taken at 25~ The temperatures of heat treatupon reduction at high temperatures--which rement were the following: 1, 25~ 2, 100~ 3, 150~ suited in extensive formation of Cr~+ o~ the 4, background spectra. s u r f a e + led to the assumption that these bands are due to NO adsorbed on Cr~+ ionsY adsorption of different gases on SnO~+Cr20a Further study is evidently necded to identify catalysts are shown in Table II. the adsorption sites and t.o choose between the The eoadsorption of NO+fuel and the effect alteruatives. The marked difference found beof preadsorbed NO on the subsequent adsorption tween the adsorptive properties of strougly and of fuel were studied at 100~ on SnO~ activated weakly reduced catalysts containing chromium hi vacuo at 400~ It appears that the pre- supports the idea that the lower valence adsorbed nitric oxide promoted the adsorption of chromium ions are the primary adsorption sites both carbon monoxide and ethylene. The ratio for NO39 of the araount of adsorbed NO and that of In furtber study, it appeared that the band adsorbed CO (promoted) approaches the value at 1860 em-~ decreased in intensity with temof 1, pointing to the existence of a surface perature and disappeared completely at 200~ complex comprising 1 mole NO mad 1 mole CO. The band at 1740 cm-~ did not change up to Promoted adsorption was also found on SnO~ 150~ subsequently the intensity decreased and using a mixture of O~ and C~1~.TM However, on above 300~ t}ds band was no longer discernible. Al~Os/Cr~O~ catalysts no promoted adsorption With a gas raixtllre of N O q - C 0 ( I : I ) , only the was observed in either case. bands due to adsorbed NO appeared on the spectra., although the adsorption of CO in the absence of NO yielded its elmraoteristic bands. In]tared Spectroscopic Measurements* The intensity of the bands due to adsorbed Infrared studies have been reported of adsorp- NO was slowly reduced when CO was added to tion of NO on various oxides ~a~ and a great the sample ~t 25-100~ and depending on the number of bands have been observed representing temperature, the bands were completely elimiadsorbed NO species ranging from (NO)~ dimer nated. The rate of removal of the band at 1860 to adsorbed NO + and NO- funs. There are how- em~ was somewhat higher than that of the band ever substautiat contradictions un several im- at 1740 em-I. I t is intel~ting that the bands due portant devils in the various papem. 1~ now to adsorbed CO appeared neither d t a ~ g nor seems certain that the transition metal ions are after the eompletiun of the surface process. the primary adsorption centers and that bands Taking into account that adsorbed NO is much more strongly bom~d than the adsorbed CO it * Infrared studies were made in cooperation with nmy be assumed that the disappearance of these Dr. J. R~k6. bands is due to the reaction between the adsorbed
RE~OVAL OF 1~O. FOLLU'rAI~r~
x I00.~ _.~
9
500000
T'~C
"~m~cuotion
NO+CO
t NO
0
i
6
[~,a. 2. "I'Ibeefl'e~ o( ~O, CO, =maiNO~UO mi~omm om the eJie(~ic r,e,~vit.w of ~
- ~ 0 and ~ CO. When fl~ sample was evacuated after the removal of NO Imnds a n d expost~ a t r i a to N O aX 25~ no ~ of tL~eb ~ d ~ al 1860 a n d 1740 cm "~ was ~ e e d . Th~ i n ~ t h a t the surface o m t e ~ for t l ~ of NO Uhat yield hands a t 1860 and 17~0 r have Imm deaetivated (toddled) d m ~ the m d a r e vmction.
A ~ n
C o ~ Mamm~at* o / N O ~ d CO
~
$ u r p r i s h ~ y little i n f m ~ d J e n is available vetoing the electronic htemcVion of N O wiLh s o l i ~ and. to o ~ knowledge, ~ eondue~ivity mp,~urermm~ Imve not bc~n used y e t t o e ~ i u a t e the m ~ ~ N O a n d semi~ondu~dng The low ~ poteul~ o(NO (9~V) make~ p o ~ b l e the converdon d 0~e nitric m i d e m o t ~ c ~ to the p 0 ~ i v e a ~ w ~ m ioa [ N O * ~ I t can also m ~ y b ~ e o v ~ a a dedzem to be c~m~erted to a . e g ~ i v e n i t m ~ ion ~ O - ~ as ~hew i~ no evidm~e of a n y Imrrler t o a t t a ~ h m m k The iom~r limit to elecizon affinity of gaseous NO is 0.65 eV?*
~timae~ at
l~.me 2 sho~ that ~ of NO i n ~ tl~ ele~ m ~ t ; . ; t ~ d aclivated SaO~. The t l ~ same dix,~dion Ulm~ the ~ of XO. a This m a y indieate t h a t the d.mdsm'ptlon of N O on ~*-type oYM~ is a n aeeeptm" ~
Howev~, e ~ a e d . g the m ~ t s p e d m ~ o p ~ inof K o r m m ~ on the a d s o r ~ i ~ of N O o n ZnO, it h also p u ~ t h ~ $.O~ m ~ m ~ ] m ldgh v a m u m a t 400*C is capable d r o t a t i n g N O to N~O a t 100"C, a e t ~ d i n g to the equathm 2~O-~-=
N~O-t-O-c,,B~.
(2)
This p m c e ~ could also eontn'lmte 1o the incr~se m the d e m i c r e d ~ i ~ y d ~ h ~ h t e ~ s f i n g to o b s e r ~ t h a t p u m ~ off the ~ O from the r m d i o ~ ~eil a t the mine t m n p e r ~ u ~ eau,ed a furtive ~ t ira=mine h the e k e t ~ res~,~iLy ~f Sm~. This belmvior ~ a s also observed for SnO2 doped ~ith 0.1 or 0.5
COMBUSTION SYSTEMS
1238
rates
W 7
3o~
6
/~a.2Otorr
4 3
x
/
,/ /
/
Y:/'" //
t.O5~
~
aTPco=20t~ 395~ [ p,.
395"C /
/PNo'2Otorr , =
H-o=orl __.__l,
,~" x-
~ o
~ 340"C -~--~0 "20t~ ~
,os'c
,--.O-20,orr
10 20 30 z,O 50 60PNO10 20 30 LO 50 60Pfuels FIG. 3. The effects of the partial pressure of the reacting gases on the rates of NO-fuels reactions.
mole% Cr20a. A tentative explanation is that, in addition to the above mode of adsorption, parS of the NO may be chemlsorbed by gi~,'ing its unpaired electron to the oxide surface to form an apparently positive chemisorbed ion N'O~ = NO+(oh~)+e0-.
(3)
This kind of NO adsorption partly compensates the result of the electron acceptor adsorption of NO EEq. (1)]. This adsorption is reversible or much weaker than the former, so that upon evacuation of the sample at 100~ NO+ desorbs from the surface N O + ( ~ ) + e o - = NO(0)
(4)
and as a result the electric resistivity of SnO2 increases, The electric resistivity of Sn02 also increased when a mixture of NO and CO (1:1) was introduced onto the sample, in spite of the fact that CO alone decreases the resistivity of Sn02 CO(a)= CO(o~.=)++e,-.
(5)
Kinetic Measurements The reaction of CO with NO on SnO~ activated at 400~ in vaeuo proceeded at a measurable rate only above 360~ After a slight decrease, the activity of the catalyst remained constant. With a steichiometric composition, a steady state activity at this temperature was attained after 40--60 rain. When CO was in excess, the reduction of nitric oxide to nitsogen was practically complete; the amount of nitrous oxide fotu~d remained below 1%. A slight activity of Sn02 was also experienced at 180-360~ but the conversion reaction was very low and the catalyst surface rapidly became poisoned. The reaction was zero order with respect to CO, and first order with respect to NO. The activation energy was 36.6 keal/mole, and thus cotmiderably larger than that of the C0-r reaction.2t The incorporation of a small amount of Cr2Oa into the surface layer of Sn02 markedly increased the efficiency of Sn02. On a catalyst containing 1% chromium oxide, the reaction between CO and NO proceeded at temperatures as low as 150~ The value of the activation energy
R E M O V A L OF NO= P O L L U T A N T
1239
400,~
~ 0,01
~
Oil
o
-
NO+C~H,,
0,5
mole%
-=-
CrzO=
i
-
FIG. 4. Values of reaction tenlperatures at which the rate constant of the reactions is l X I 0 -s rain -z m -=.
decreases with the increase of the amount of CreOa in Sn%. The catalyticreaction between H~ mad N O on pure tin dioxide was measured at 320--360~ It was somewhat faster than the N 0 - C O reaction under similar conditions. The formation of N H a was detected even in the presence of a slight excess of He (He:NO mole ratio 2:1). Pure tin dioxide catalyzed the reduction of N O with ethylene only above 390~ The sample doped with I % chromium oxide lowered the temperature range of beth reactions by about 100-150~ Figure 3 shows the rate of the reduction of NO with different fuels as a function of the partial presenre of the reacting gases. The catalytic efficiency of SnO~ containing Cr~O~ has been te~ted in the reduction of NO with natural gas containing CH4 (91.1%), C2I-I4 (1.46%), Carte (0.72v/o), C~tIl0 (0.54%), C~Hn (0.12%), CO2 (2.74%), Ne (3.37%). Undoped Sn(~ catalyzed the reaction only above 450~ When 1% Cr20a was incorporated into SnO~ the oxidation reaction between natural gas and NO occurred around 300~ In the presence of excess natural gas, the reduction of N O to N2 was complete at this temperature.
In order to compare the effect of the chromium content of Sn02 on the reduction of NO with different fuels, the reaction temperatures, a t which the rate constants of the reactions have a value of 1X 10-a: axe shown in Fig. 4 as function of the chromium oxide content of the tin dioxide. The addition of Cr20a to Sn(h exhibited the largest effect in the case of N O + C O reaction. For comparison, the reduction of NO has been investigated on pure Cr~O, a n d on A120a/Cr~08 catalysts. The latter was found to be one of the most active oxides in the reaction between NO and CO? Kinetic d a t a for the reduction of NO are collected in Table III. These data show that SnO~fCr20a has the highest activity among the catalysts studied in this work in the low temperature reduction of nitric oxide.
A Possible Mechanism of Reduction of NO with CO We m a y attempt now to formulate a mechanistic picture of the reaction mechanism. It seems certain t h a t different active centers are responsible for the catalytic reactions on pure Sn02
(X)MBU~IIOlg TABLE I ~ ] ~ m ~ e &m~ fro- t]~e cat~dytir reduetim~ ,~ NO with d ~ f f ~ m fue~
NO-(X) vema~oe Temq)et'tum~ ~ ~ ~ o t n ~ ~o 1~O R e ~ a ~ u ovd~ to CO k . . , . ~i ~$0"C E, kmL~ol~
405440 1 0 ~XIO ~ 36.6
NO-r xeaetion Tmq~atune ~ =C R m o ~ ord~ ~ N~O R e m i m ~ l k ~ to C~H~
~ O Z 7.4XI0~
9 l;~
170-20~ t 0 1.18X10 ~ 18.2
215-300 l 0 .~04XI0 ~ 3~.5
250-310 o 1
31~ 1 0
~)-26~
33.8
5.8• -~ 13.3
380-4~
37,0-t30
~50-~100
&X I 0 ~
1•
~
4-7 X 10-~
I-~7X10 - ~
2-~ X l O ~'~
19.6
14.1
17.1
17.~
:~0-380 i @ l.S,t xz0-',* 1~.~
1~o-~o
E, k g ~
NO-I~ Temperatare hinge, ~(~ R e m i ~ ovd~ to ~ O l h s m i ~ arder ~o H~ ~ ~ g, ~
md m 8.xO~-Cr~ ~
tlutt i
We m a y a ~ D e
tl~ c . ~ d u = ~ q ~ i $aO~ ~ z + is t ~
o~ our o h s w ~ f i n -
~
the ~ i t ~
N O is ~ , , ~ ~gaeu it is ~ the e s t s l y ~ sm-f~cs m m~S=ed s n d ~ NO. ~ ,~m~dw~=l~v~ ~ ;rag line ~ d N O ~l, , m ; r NO ~
1 0 ~.~xnr ~ LS.~
m iJl~ ~
~ tl~
the ti~ the the
I~.~L whenms the ~ o~ an e t e e U ~ to mliboud~ ~.1 ~ N O tends to ~ k t ~ N - O bond, it ~ I d a u l ~ to I s ~ t/~t o~dstioa Q / t ~ ~ l y s t surfa~ ~ d~ ta
~mh~d~ by
for the r e d u t ~ of XO b~r CO C~[ r ~ e s ~ the s t ~ e s i ~ ~ l + ) :
dmr~
M-t- N ( ~ - M + N O "-
smdPm~e.The f , ~ n , , , u ~
M+NO--~- M---~M-- Nq-M+O -
(gs~
d - fis electron o ( S n ~ ion t o N O
8,,~-I-N~NO
-
M§
(6)
whereto the .i~rom.ium ~ N O ~, ~ formed d ~ ~o tln~ t n m s ~ d lhe olM ~ from NO
M - - NO-~M--*M-- N-]-M~O -M - - N -[- NO--~M~'O--4-~'t
Sn'++NO~S~NO ]~e~ r
~.
Ca)
eondu~ memumm~ d s o mt l v ~ NO+ ~s mue.h ~ weskl3v Ineki on than N O - ions. Sinee the f o . ~ ,~ NO* m s~n~it,d by the ~ of
M+O--bGO-~M§ The f
~
(Sb) (gb) (10) 01)
d N20 m y ooctw in ~ e v m z ~ u M--~+ NO-,M+ NO.
(~)
REMOVAL OF NO= POLLUTANT
ug 20
co, to,
1241
Reduction and oxidation of SnOz
aw~
?,.V~e~acuation
10
/J8"C
150 torrCO ,
-I
50 torrO=
o/
evacuation
"r''/" /
2o
~;o NO*CO " ' - -
"-~
II
evacuation/~ 60
"~50 torr NO i
|
50
|
100
150
i
200
rn]n
Fro. 5. Interaction of SnO~ with NO, CO, NO+CO, and O2q-CO gas mixtures under reaction conditions at 400~ measured with a sensitive microbaJance. The weight changes of the catalyst during the interaction with different gases axe shown as ordinates.
An alternative to reactions (10) and (11) is that CO reacts with the adsorbed nitrogen atom
M-N+ C~M+NC0
-
M+N CO--i- NO-~M-f- N~-]-CCh
(13) (14)
yielding a surface isoeya~late group, which could be a possible intermediate of the catalytic reaction. Isecyanate species have been detected by infrared spectroscopy during the reaction of nitric oxide with carbon monoxide on the surface of noble metal catalystsl~,~aand also on supported CuO.4 The slowest step of the reaction of NO with CO is very probably the oxidation of active sites on the catalyst. From the study of the interaction of the reacting mixture with the catalyst during the catalytic process, it appeared t h a t the oxidation of CO with NO takes place on a weakly
reduced surface (Fig. 5). As against that, the oxidation of CO with 02 occurs on an oxidized surface, in agreement with the results of kinetic studies? 1 In the ease of Sn02 containing Cr20,, the catalytic reaction proceeded at a very low temperature, where pure Sn02 did not exhibit any catalytic effect. We propose that the active catalytic sites in SnO~-Cr~0~ catalyst are the lower valence (Crz+) ions located in the surface layer of SnO~. As already shown these ehronfium centers can prolaote the ehcmisorption of NO and can be easily oxidized with NO to higher valence states. A possible driving force for this surface oxidation is t h a t CrO~ and Sn02 both have l~util structure, and t h a t Cr(h is stabilized in the Sn02 lattice. The slowest step of the catalytic reaction is the same as prcviously, i.e., the oxidation of reduced centers, as carbon monoxide reduces the higher valence chromium in a very fast reaction.~1 We assume that this redox mechanism also operates in the reaction of NO with other fuels.
COMBUSTION SYSTEMS
1242
REFE~NCF~
13. l.a~-.~, L. H.: Infrared 8pe~ra of Adsorbed 8pec~, Academic Press, (1966). 1. SHEI~I~, M. AND KUMMZ]g,J. T. : Chem. Eng. 14. T]emlmN, A. AND Rozv, L.: A ~ Prog. Syrup. Sedee 67, 74 (1971). Confr. ln~. CaWJyse, p. 183j Techip, 1961. 2. D w ~ R , F. G.: Catsl. Rev. 6e, 261 (1972). 15. P ~ N s , N. D.: Proc~KnCs of tlte g~fl& l~.~r3. S~mx~F, M., Ox,ro, K., A.~D GANDHI, ~,i.: na~ona/ Confre~ on ~ a t ~ s , North-Honaad J. CataL I~, 361 (1968). Pub]., 1973. 4. Lo~'vo~r J. W. ~.~-v B ~ , A. T.: J. CataL 81, le. Ex~Y, D. D., Rocaffi~mR, C. H., ANy Scvm~L, .T2, 96 (1973). M. S.: J. Chem. Soc. Faraday Tra~. I 89, 660 5. SOL~'MOS~, F. AND BLosSom, T.: Proceedinqe (1973). of ~he Second Inter~ional Conference on Spa~ 17. Rozv, L. M. ANy/t~.=~rRuv, A. V.: B/ementa~ En~nee~'nq, p. 145, D. Roidsl Publ., 1970. P&~opracee~ ~n Mo/ecu/e~, (English t~ansl~6. Bozs6, F.: Ph.D. Thesis, Unlvere~ty of Szeged tion), p. 2~0, Consultant~ Bureau, 1968. (1978). 18. PznI, J. B.: Paper presented at the National 7. KoaNffi-% g. E.: J. Phys. Chem. Solids eS, ,Meeting of American Chemical Society, Dallas, 1557 (1962). Apr. 1973. 8. S~drrr, R., MO~,3cT, J. A., xN~ Bo~.~m,g 19. SOL,Ore, F. aND RAsx6, J.: unpublished N. F.: J. Phys. Chem. Solids ~5, 1465 (1964). reeul@. 9. Oq'ro, K. ~ . ~ Sar~X~F, M.: J. Cstal. ~ , 226 20. SrocKwJ~, J. A. D., COMp~ON,1L N., Hvawr, (I~9). G. S., ~ v RzmH~v,t, P. W.: J. Chem. Phys. 10. Sn~x~, M. ~ v 0"~o, K.: J. Catal. 18, 184 50, 2176 (1969). (197o). 21. SoLruoel, F., KJBS, J., AND BOzS6, F.: Paper 11. GAtc~,v, ~-~ S. ~,,~D Sw~z.ms,, M.: J. Catsl. prmented at Euchem Conference on the Role ~J, 241 (1972). of C a t a l y ~ in Problem of Pollution, 8antaader, 12. ~OLYMOelI, F. AND BOZS~, F.: 8~y~l~q~'um on July, 1973. the Mechanfim~s of Hydrocarbon Reaction#, 22. Koa~t'u, G. AND ~N~Rlg, H.: ~ r . Bul~sng88. Siofok (Hungary). Publishing IIotme of the 77, 85 (1973). Hm3gaxian Academy of Sciences, 1975 in pre~8. 23. UN~NV, M. L.:J. Phys. Chem. 77, 1952 (1973).
for eliminating this pollutant from car exhaust. The study of this reaction has been the subject of many recent papers but only a few of them are concerned with the reaction mechanism,s,4 It was found that the oxidation and reduction of the catalyst surfaces play an important role in the catalytic reaction. Therefore our attention turned to catalyst systems that especially favor the redox reactions. A prominent catalyst in this respect is SnO~ containing a small amount of added chromium orSde. The prescott work was mldertaken to examine the interaction of nitric oxide with these catalysts and to investigate the kfiletics and mechanism of the catalytic reduction of nitric oxide by different reduch~g agents. Experimental Malerials CrzOa was reagent grade material from Reanal. It was fired a~ 700~ ~'or 5 h. A1203 was from
1233
COMBUSTION SYSTEMS
1234
Degnssa. The Al2Or-Cr~O~ catalyst was prepared by impregnating alumina with an aqueous solution of Cr(~. The gel was dried in air for 16 h at 120~ For catalytic investigations, the powder was calcined for 8 hours at 600~ SnO~ was obtained by the action of HNO3 on metallic Sn. It was dried at 120~ and heated at 350~ for 3 h and at 500~ for 5 h. The doping of the SnO~ with Cr2Oa was effected in the following way~.6:Cry03 was added to an aqueous suspension of Sn(h and after a sufficiently long period of stirring, the homogeneous suspension was dried and heated at 350~ for 5 h. Final sintering was performed at 900~ for 5 h in air. The nitric oxide (Matheson Company) was commercially pure (99%). Ethylene was a product of Fluka Gmbh. Both gases were purified by bulb-to-bulb distillation before use. Carbon monoxide was prepared in the laboratory by reaction between formic acid and sulphuric acid a t 83~ For the catalytic studies, small pellets (1.5 ram diam and 1.5 mm long) were made. A fixed amount of catalyst (1.5 g, 8-10 small pellets) was used in all experiments.
Methods of Measurements For the kinetic investigations, a closed circulation system (volume 277 ml) was used. The reactions were followed by measuring the pressure of the reacting gases. A Hewiett-Packard gas chromatograph connected to the reactor was used to analyze the reaction products. Conductivity measurements during the adsorption a n d catalytic reaction were performed in the same reactor, using pellets of 8 ram diameter and 6-7 nun long. An ac bridge was used for these measurements. The active oxygen content of the catalysts was determined b y the Bunsen method. Specific surface areas were calculated from the adsorption of nitrogen a t liquid nitrogen temperature. Adsorption measurements were performed in a Sartorius microbalance. Infrared spectra were taken with a double beam I R spectrophotometer (Speetromom 2000). The discs of AhOs/CraO~ were made by pressing the finely divided powder dried at 120~ to a thickness corresponding to 30 m g / m ~. The sample was degassed in vacuo at 10-4 TorT. Further treatment is described in the text. R e s u l t s and D i s c u s s i o n
Characterization of the Catalysts SnO~ is an n-type semiconductor. Its electric conductivity is due to donor levels t h a t arise from
defects such as impurity cations and anion vacancies.7,s The incorporation of a small amount of Cr2Oa into SilOs in air significantly changes the physical properties of SnO2.~'6 On the other hand, the chromium ions also undergo a noteworthy change; part of them is oxidized to a higher valence and stabilized in the surface layer of SnO~. A remarkable feature of higher valence chromium in the surface layer of Sn0~ is t h a t after its reduction with fuels it can easily be reoxidized even at low temperatures. Characteristic data for different catalysts used in this work are summarized in Table I.
Adsorption Measurements The adsorption of NO has been investigated on AhOa/C~08 catalyst,9 but no studies have been made on Sn02. The characteristic feature of the adsorption of NO on diamagnetic oxides is its extreme slowness; the establishment of the adsorption equilibrium is very difficult. The adsorption of NO on SnO2 activated at 400*C in vacuum is also slow, in spite of the fact t h a t there is no experimental evidence for a substantial energy of activation. Adsorption isotherms a n d rates for NO were measured at 25-100~ The data conform well to the Freundiich isotherms in the pressure range covered, as shown by a linear relationship between the logarithmic amount of adsorbed NO, q, and the logarithmic pressure, p. The Freundlich isotherm is defined by
q = cpl/n where the coefficients n and c are functions of temperature. The points of intersection of isotherms show t h a t the amount of NO required for a monolayer (q~) is 7.46 ~moles/g. The preSsure which is necessary to achieve monolayer coverage is 74.2 Torr. The average value of H,,, which is a measure of the heat of adsorption, is 6.9 kcal/mole. Similar low values have been obtained b y Shelef, et aL, on transitional metal oxides.0-a The rate of adsorption of NO was evaluated b y the integrated form of the Elovich equation q=
(2.3/~) ]og(~+~)-(2.3/a) 1og~
where tom 1/an is an integration constant. The Elovieh plots consist of two linear segments with a distinct break between them. A possible explanation of the break is t h a t two distinct populations of adsorption sites exist on the sur-
REMOVAL OF NO= POLLUTANT
1235
TABLE I Characteristic data for different catalysts
Catalyst
Cr2Oa content mole%
Highest calcination temperature ~
Surface area m2/g
Active oxygen %
Average valency of chromium*
Crd:)~ Sn02 SnO~ SnOs SnOz Sn02 AlsOs A12Oa
100 -0.01 O.1 0.5 1.0 1.0 5
700 900 900 900 900 900 900 600
30.4 5.81 0.12
0.0224 -0.0015 0.01 0.043 0.082
3 -4.5 3.95 3.81 3.77
1.0243
5
5.58
8.84 11.87 4.64 70.0
* Calculated on the basis of active oxygen content. t This sanlple of A120,was a Reanal product. The doping with Cr2Os was similar to the case of SnO,.
TABLE II Parameters for the adsorption of gases at 1O0~ on SnO2-bCr~O, catalysts 8nOs al El~nole/g] -~ NO O2 CO C~HI
7.5 1. 856 1.77 4.025
SuOz+ 1%Cr=O~ q0 t.~rnole/g 1.1 2. 875 0.4 0.714
~1 [~mole/g']-1
q0 ~anole
1.15 3. 072 0.621 2.3
3.3 3.25 1.66 0.71
q0ffitheamotmt of instantaneousadsorption,al is the slope of the first segment of the Elovlch plots. face. The temperature coefficient of the adsorption is positive; both the czl and a2 decrease with temperature. On the partially reduced Sn02 (with CO at 400~ for 60 rain) the instantaneous adsorption of N O was almost six times larger than that of adsorption on unreduced samples. In addition the rate of adsorption at i00~ was about ten times higher. The chemisorption of N O on unreduced SnOs is reversible, whereas on SnO~ reduced with fuels it was somewhat irreversible. In the latter case it is assumed that mainly oxygen remains on the surface, while N O is reduced to Ns0. A very large increase in the adsorbed amount of N O and in the rate of chemisorption was achieved by incorporating CrtOs (I mole%) into the surface layer of Sn0s (Table II). Similar measurements were carried out for the
adsorption of other gases, e.g., CO, 02, and CsH4. The histantaneous adsorption of CO is commensurable with that of N0. The Elovieh plots also consist of two intersecting straight line segments. While al decreases with temperature, a~ increases with T. If we compare the adsorption of NO and O~ on Sn02, we find that both the extent of instantaneous adsorption of O~ and the rate of subsequent adsorption are always larger than for NO. With the increase of adsorption temperature, this difference becomes greater. The initial adsorption of ethylene on SnOa below 100~ has the highest value of all the gases (fuels) studied so far, it is however hardly affected by the increase of temperature or by incorporating chromium into Sues. Accordingly, near the reaction temperature, the adsorption of 0s and NO greatly exceeds that of ethylene. Data for
1236
COMBUSTION SYSTEMS
observed between 1000-2000 cm-I when NO is added to alumina or silica are due to the presence of transitional metal impurities) ~,15 Unfortunately all our efforts to make transpareut discs from SnO2/Cr20a have been mtsueessful, restricting our infl'ared studies to the A1203/ Cr20a system. When NO was added to AhOa/Cr~O~, pretreated in CO (initial pressure 10 Torr) at 4O0~ for 1 h, two large bands appeared at 1865 and 1740 em-1 (Fig. 1). The latter is perceptibly composed of a combination of separate bands at 1743 and 1738 cm-1. Both bands have been observed recently during the adsorptioo of NO oa Al~Oz/Cr~Oaactivated in ethylene2 ~ Based orl the assignments of NO bands observed on transition metal oxidesy ,J7 the band at 1743 and 1738 em-1 3000 2000 1800 1600 1400 1200 cm"1 was attributed to NO cnvalently bound to Cr ~+ and Cr~+ sites, respectively. The absorption band Fzo. 1. Infr~ed epeetra of NO on AhO#Cr~Os. at ~ 1 8 6 0 cmy2 was assigned to NO chelrJsorbed Adsorption of NO (5 Torr) was carried out, at 25~ by a double bond ionic link (Cr~+--N+--O) to for 1 h, thee the sample was evacuated and heated chromium ions Cr <:+n+. However, the observaat diffe~'en~bemperatures for 30 rain. All the spectra tion that tl~e intensities of both bands increased were taken at 25~ The temperatures of heat treatupon reduction at high temperatures--which rement were the following: 1, 25~ 2, 100~ 3, 150~ suited in extensive formation of Cr~+ o~ the 4, background spectra. s u r f a e + led to the assumption that these bands are due to NO adsorbed on Cr~+ ionsY adsorption of different gases on SnO~+Cr20a Further study is evidently necded to identify catalysts are shown in Table II. the adsorption sites and t.o choose between the The eoadsorption of NO+fuel and the effect alteruatives. The marked difference found beof preadsorbed NO on the subsequent adsorption tween the adsorptive properties of strougly and of fuel were studied at 100~ on SnO~ activated weakly reduced catalysts containing chromium hi vacuo at 400~ It appears that the pre- supports the idea that the lower valence adsorbed nitric oxide promoted the adsorption of chromium ions are the primary adsorption sites both carbon monoxide and ethylene. The ratio for NO39 of the araount of adsorbed NO and that of In furtber study, it appeared that the band adsorbed CO (promoted) approaches the value at 1860 em-~ decreased in intensity with temof 1, pointing to the existence of a surface perature and disappeared completely at 200~ complex comprising 1 mole NO mad 1 mole CO. The band at 1740 cm-~ did not change up to Promoted adsorption was also found on SnO~ 150~ subsequently the intensity decreased and using a mixture of O~ and C~1~.TM However, on above 300~ t}ds band was no longer discernible. Al~Os/Cr~O~ catalysts no promoted adsorption With a gas raixtllre of N O q - C 0 ( I : I ) , only the was observed in either case. bands due to adsorbed NO appeared on the spectra., although the adsorption of CO in the absence of NO yielded its elmraoteristic bands. In]tared Spectroscopic Measurements* The intensity of the bands due to adsorbed Infrared studies have been reported of adsorp- NO was slowly reduced when CO was added to tion of NO on various oxides ~a~ and a great the sample ~t 25-100~ and depending on the number of bands have been observed representing temperature, the bands were completely elimiadsorbed NO species ranging from (NO)~ dimer nated. The rate of removal of the band at 1860 to adsorbed NO + and NO- funs. There are how- em~ was somewhat higher than that of the band ever substautiat contradictions un several im- at 1740 em-I. I t is intel~ting that the bands due portant devils in the various papem. 1~ now to adsorbed CO appeared neither d t a ~ g nor seems certain that the transition metal ions are after the eompletiun of the surface process. the primary adsorption centers and that bands Taking into account that adsorbed NO is much more strongly bom~d than the adsorbed CO it * Infrared studies were made in cooperation with nmy be assumed that the disappearance of these Dr. J. R~k6. bands is due to the reaction between the adsorbed
RE~OVAL OF 1~O. FOLLU'rAI~r~
x I00.~ _.~
9
500000
T'~C
"~m~cuotion
NO+CO
t NO
0
i
6
[~,a. 2. "I'Ibeefl'e~ o( ~O, CO, =maiNO~UO mi~omm om the eJie(~ic r,e,~vit.w of ~
- ~ 0 and ~ CO. When fl~ sample was evacuated after the removal of NO Imnds a n d expost~ a t r i a to N O aX 25~ no ~ of tL~eb ~ d ~ al 1860 a n d 1740 cm "~ was ~ e e d . Th~ i n ~ t h a t the surface o m t e ~ for t l ~ of NO Uhat yield hands a t 1860 and 17~0 r have Imm deaetivated (toddled) d m ~ the m d a r e vmction.
A ~ n
C o ~ Mamm~at* o / N O ~ d CO
~
$ u r p r i s h ~ y little i n f m ~ d J e n is available vetoing the electronic htemcVion of N O wiLh s o l i ~ and. to o ~ knowledge, ~ eondue~ivity mp,~urermm~ Imve not bc~n used y e t t o e ~ i u a t e the m ~ ~ N O a n d semi~ondu~dng The low ~ poteul~ o(NO (9~V) make~ p o ~ b l e the converdon d 0~e nitric m i d e m o t ~ c ~ to the p 0 ~ i v e a ~ w ~ m ioa [ N O * ~ I t can also m ~ y b ~ e o v ~ a a dedzem to be c~m~erted to a . e g ~ i v e n i t m ~ ion ~ O - ~ as ~hew i~ no evidm~e of a n y Imrrler t o a t t a ~ h m m k The iom~r limit to elecizon affinity of gaseous NO is 0.65 eV?*
~timae~ at
l~.me 2 sho~ that ~ of NO i n ~ tl~ ele~ m ~ t ; . ; t ~ d aclivated SaO~. The t l ~ same dix,~dion Ulm~ the ~ of XO. a This m a y indieate t h a t the d.mdsm'ptlon of N O on ~*-type oYM~ is a n aeeeptm" ~
Howev~, e ~ a e d . g the m ~ t s p e d m ~ o p ~ inof K o r m m ~ on the a d s o r ~ i ~ of N O o n ZnO, it h also p u ~ t h ~ $.O~ m ~ m ~ ] m ldgh v a m u m a t 400*C is capable d r o t a t i n g N O to N~O a t 100"C, a e t ~ d i n g to the equathm 2~O-~-=
N~O-t-O-c,,B~.
(2)
This p m c e ~ could also eontn'lmte 1o the incr~se m the d e m i c r e d ~ i ~ y d ~ h ~ h t e ~ s f i n g to o b s e r ~ t h a t p u m ~ off the ~ O from the r m d i o ~ ~eil a t the mine t m n p e r ~ u ~ eau,ed a furtive ~ t ira=mine h the e k e t ~ res~,~iLy ~f Sm~. This belmvior ~ a s also observed for SnO2 doped ~ith 0.1 or 0.5
COMBUSTION SYSTEMS
1238
rates
W 7
3o~
6
/~a.2Otorr
4 3
x
/
,/ /
/
Y:/'" //
t.O5~
~
aTPco=20t~ 395~ [ p,.
395"C /
/PNo'2Otorr , =
H-o=orl __.__l,
,~" x-
~ o
~ 340"C -~--~0 "20t~ ~
,os'c
,--.O-20,orr
10 20 30 z,O 50 60PNO10 20 30 LO 50 60Pfuels FIG. 3. The effects of the partial pressure of the reacting gases on the rates of NO-fuels reactions.
mole% Cr20a. A tentative explanation is that, in addition to the above mode of adsorption, parS of the NO may be chemlsorbed by gi~,'ing its unpaired electron to the oxide surface to form an apparently positive chemisorbed ion N'O~ = NO+(oh~)+e0-.
(3)
This kind of NO adsorption partly compensates the result of the electron acceptor adsorption of NO EEq. (1)]. This adsorption is reversible or much weaker than the former, so that upon evacuation of the sample at 100~ NO+ desorbs from the surface N O + ( ~ ) + e o - = NO(0)
(4)
and as a result the electric resistivity of SnO2 increases, The electric resistivity of Sn02 also increased when a mixture of NO and CO (1:1) was introduced onto the sample, in spite of the fact that CO alone decreases the resistivity of Sn02 CO(a)= CO(o~.=)++e,-.
(5)
Kinetic Measurements The reaction of CO with NO on SnO~ activated at 400~ in vaeuo proceeded at a measurable rate only above 360~ After a slight decrease, the activity of the catalyst remained constant. With a steichiometric composition, a steady state activity at this temperature was attained after 40--60 rain. When CO was in excess, the reduction of nitric oxide to nitsogen was practically complete; the amount of nitrous oxide fotu~d remained below 1%. A slight activity of Sn02 was also experienced at 180-360~ but the conversion reaction was very low and the catalyst surface rapidly became poisoned. The reaction was zero order with respect to CO, and first order with respect to NO. The activation energy was 36.6 keal/mole, and thus cotmiderably larger than that of the C0-r reaction.2t The incorporation of a small amount of Cr2Oa into the surface layer of Sn02 markedly increased the efficiency of Sn02. On a catalyst containing 1% chromium oxide, the reaction between CO and NO proceeded at temperatures as low as 150~ The value of the activation energy
R E M O V A L OF NO= P O L L U T A N T
1239
400,~
~ 0,01
~
Oil
o
-
NO+C~H,,
0,5
mole%
-=-
CrzO=
i
-
FIG. 4. Values of reaction tenlperatures at which the rate constant of the reactions is l X I 0 -s rain -z m -=.
decreases with the increase of the amount of CreOa in Sn%. The catalyticreaction between H~ mad N O on pure tin dioxide was measured at 320--360~ It was somewhat faster than the N 0 - C O reaction under similar conditions. The formation of N H a was detected even in the presence of a slight excess of He (He:NO mole ratio 2:1). Pure tin dioxide catalyzed the reduction of N O with ethylene only above 390~ The sample doped with I % chromium oxide lowered the temperature range of beth reactions by about 100-150~ Figure 3 shows the rate of the reduction of NO with different fuels as a function of the partial presenre of the reacting gases. The catalytic efficiency of SnO~ containing Cr~O~ has been te~ted in the reduction of NO with natural gas containing CH4 (91.1%), C2I-I4 (1.46%), Carte (0.72v/o), C~tIl0 (0.54%), C~Hn (0.12%), CO2 (2.74%), Ne (3.37%). Undoped Sn(~ catalyzed the reaction only above 450~ When 1% Cr20a was incorporated into SnO~ the oxidation reaction between natural gas and NO occurred around 300~ In the presence of excess natural gas, the reduction of N O to N2 was complete at this temperature.
In order to compare the effect of the chromium content of Sn02 on the reduction of NO with different fuels, the reaction temperatures, a t which the rate constants of the reactions have a value of 1X 10-a: axe shown in Fig. 4 as function of the chromium oxide content of the tin dioxide. The addition of Cr20a to Sn(h exhibited the largest effect in the case of N O + C O reaction. For comparison, the reduction of NO has been investigated on pure Cr~O, a n d on A120a/Cr~08 catalysts. The latter was found to be one of the most active oxides in the reaction between NO and CO? Kinetic d a t a for the reduction of NO are collected in Table III. These data show that SnO~fCr20a has the highest activity among the catalysts studied in this work in the low temperature reduction of nitric oxide.
A Possible Mechanism of Reduction of NO with CO We m a y attempt now to formulate a mechanistic picture of the reaction mechanism. It seems certain t h a t different active centers are responsible for the catalytic reactions on pure Sn02
(X)MBU~IIOlg TABLE I ~ ] ~ m ~ e &m~ fro- t]~e cat~dytir reduetim~ ,~ NO with d ~ f f ~ m fue~
NO-(X) vema~oe Temq)et'tum~ ~ ~ ~ o t n ~ ~o 1~O R e ~ a ~ u ovd~ to CO k . . , . ~i ~$0"C E, kmL~ol~
405440 1 0 ~XIO ~ 36.6
NO-r xeaetion Tmq~atune ~ =C R m o ~ ord~ ~ N~O R e m i m ~ l k ~ to C~H~
~ O Z 7.4XI0~
9 l;~
170-20~ t 0 1.18X10 ~ 18.2
215-300 l 0 .~04XI0 ~ 3~.5
250-310 o 1
31~ 1 0
~)-26~
33.8
5.8• -~ 13.3
380-4~
37,0-t30
~50-~100
&X I 0 ~
1•
~
4-7 X 10-~
I-~7X10 - ~
2-~ X l O ~'~
19.6
14.1
17.1
17.~
:~0-380 i @ l.S,t xz0-',* 1~.~
1~o-~o
E, k g ~
NO-I~ Temperatare hinge, ~(~ R e m i ~ ovd~ to ~ O l h s m i ~ arder ~o H~ ~ ~ g, ~
md m 8.xO~-Cr~ ~
tlutt i
We m a y a ~ D e
tl~ c . ~ d u = ~ q ~ i $aO~ ~ z + is t ~
o~ our o h s w ~ f i n -
~
the ~ i t ~
N O is ~ , , ~ ~gaeu it is ~ the e s t s l y ~ sm-f~cs m m~S=ed s n d ~ NO. ~ ,~m~dw~=l~v~ ~ ;rag line ~ d N O ~l, , m ; r NO ~
1 0 ~.~xnr ~ LS.~
m iJl~ ~
~ tl~
the ti~ the the
I~.~L whenms the ~ o~ an e t e e U ~ to mliboud~ ~.1 ~ N O tends to ~ k t ~ N - O bond, it ~ I d a u l ~ to I s ~ t/~t o~dstioa Q / t ~ ~ l y s t surfa~ ~ d~ ta
~mh~d~ by
for the r e d u t ~ of XO b~r CO C~[ r ~ e s ~ the s t ~ e s i ~ ~ l + ) :
dmr~
M-t- N ( ~ - M + N O "-
smdPm~e.The f , ~ n , , , u ~
M+NO--~- M---~M-- Nq-M+O -
(gs~
d - fis electron o ( S n ~ ion t o N O
8,,~-I-N~NO
-
M§
(6)
whereto the .i~rom.ium ~ N O ~, ~ formed d ~ ~o tln~ t n m s ~ d lhe olM ~ from NO
M - - NO-~M--*M-- N-]-M~O -M - - N -[- NO--~M~'O--4-~'t
Sn'++NO~S~NO ]~e~ r
~.
Ca)
eondu~ memumm~ d s o mt l v ~ NO+ ~s mue.h ~ weskl3v Ineki on than N O - ions. Sinee the f o . ~ ,~ NO* m s~n~it,d by the ~ of
M+O--bGO-~M§ The f
~
(Sb) (gb) (10) 01)
d N20 m y ooctw in ~ e v m z ~ u M--~+ NO-,M+ NO.
(~)
REMOVAL OF NO= POLLUTANT
ug 20
co, to,
1241
Reduction and oxidation of SnOz
aw~
?,.V~e~acuation
10
/J8"C
150 torrCO ,
-I
50 torrO=
o/
evacuation
"r''/" /
2o
~;o NO*CO " ' - -
"-~
II
evacuation/~ 60
"~50 torr NO i
|
50
|
100
150
i
200
rn]n
Fro. 5. Interaction of SnO~ with NO, CO, NO+CO, and O2q-CO gas mixtures under reaction conditions at 400~ measured with a sensitive microbaJance. The weight changes of the catalyst during the interaction with different gases axe shown as ordinates.
An alternative to reactions (10) and (11) is that CO reacts with the adsorbed nitrogen atom
M-N+ C~M+NC0
-
M+N CO--i- NO-~M-f- N~-]-CCh
(13) (14)
yielding a surface isoeya~late group, which could be a possible intermediate of the catalytic reaction. Isecyanate species have been detected by infrared spectroscopy during the reaction of nitric oxide with carbon monoxide on the surface of noble metal catalystsl~,~aand also on supported CuO.4 The slowest step of the reaction of NO with CO is very probably the oxidation of active sites on the catalyst. From the study of the interaction of the reacting mixture with the catalyst during the catalytic process, it appeared t h a t the oxidation of CO with NO takes place on a weakly
reduced surface (Fig. 5). As against that, the oxidation of CO with 02 occurs on an oxidized surface, in agreement with the results of kinetic studies? 1 In the ease of Sn02 containing Cr20,, the catalytic reaction proceeded at a very low temperature, where pure Sn02 did not exhibit any catalytic effect. We propose that the active catalytic sites in SnO~-Cr~0~ catalyst are the lower valence (Crz+) ions located in the surface layer of SnO~. As already shown these ehronfium centers can prolaote the ehcmisorption of NO and can be easily oxidized with NO to higher valence states. A possible driving force for this surface oxidation is t h a t CrO~ and Sn02 both have l~util structure, and t h a t Cr(h is stabilized in the Sn02 lattice. The slowest step of the catalytic reaction is the same as prcviously, i.e., the oxidation of reduced centers, as carbon monoxide reduces the higher valence chromium in a very fast reaction.~1 We assume that this redox mechanism also operates in the reaction of NO with other fuels.
COMBUSTION SYSTEMS
1242
REFE~NCF~
13. l.a~-.~, L. H.: Infrared 8pe~ra of Adsorbed 8pec~, Academic Press, (1966). 1. SHEI~I~, M. AND KUMMZ]g,J. T. : Chem. Eng. 14. T]emlmN, A. AND Rozv, L.: A ~ Prog. Syrup. Sedee 67, 74 (1971). Confr. ln~. CaWJyse, p. 183j Techip, 1961. 2. D w ~ R , F. G.: Catsl. Rev. 6e, 261 (1972). 15. P ~ N s , N. D.: Proc~KnCs of tlte g~fl& l~.~r3. S~mx~F, M., Ox,ro, K., A.~D GANDHI, ~,i.: na~ona/ Confre~ on ~ a t ~ s , North-Honaad J. CataL I~, 361 (1968). Pub]., 1973. 4. Lo~'vo~r J. W. ~.~-v B ~ , A. T.: J. CataL 81, le. Ex~Y, D. D., Rocaffi~mR, C. H., ANy Scvm~L, .T2, 96 (1973). M. S.: J. Chem. Soc. Faraday Tra~. I 89, 660 5. SOL~'MOS~, F. AND BLosSom, T.: Proceedinqe (1973). of ~he Second Inter~ional Conference on Spa~ 17. Rozv, L. M. ANy/t~.=~rRuv, A. V.: B/ementa~ En~nee~'nq, p. 145, D. Roidsl Publ., 1970. P&~opracee~ ~n Mo/ecu/e~, (English t~ansl~6. Bozs6, F.: Ph.D. Thesis, Unlvere~ty of Szeged tion), p. 2~0, Consultant~ Bureau, 1968. (1978). 18. PznI, J. B.: Paper presented at the National 7. KoaNffi-% g. E.: J. Phys. Chem. Solids eS, ,Meeting of American Chemical Society, Dallas, 1557 (1962). Apr. 1973. 8. S~drrr, R., MO~,3cT, J. A., xN~ Bo~.~m,g 19. SOL,Ore, F. aND RAsx6, J.: unpublished N. F.: J. Phys. Chem. Solids ~5, 1465 (1964). reeul@. 9. Oq'ro, K. ~ . ~ Sar~X~F, M.: J. Cstal. ~ , 226 20. SrocKwJ~, J. A. D., COMp~ON,1L N., Hvawr, (I~9). G. S., ~ v RzmH~v,t, P. W.: J. Chem. Phys. 10. Sn~x~, M. ~ v 0"~o, K.: J. Catal. 18, 184 50, 2176 (1969). (197o). 21. SoLruoel, F., KJBS, J., AND BOzS6, F.: Paper 11. GAtc~,v, ~-~ S. ~,,~D Sw~z.ms,, M.: J. Catsl. prmented at Euchem Conference on the Role ~J, 241 (1972). of C a t a l y ~ in Problem of Pollution, 8antaader, 12. ~OLYMOelI, F. AND BOZS~, F.: 8~y~l~q~'um on July, 1973. the Mechanfim~s of Hydrocarbon Reaction#, 22. Koa~t'u, G. AND ~N~Rlg, H.: ~ r . Bul~sng88. Siofok (Hungary). Publishing IIotme of the 77, 85 (1973). Hm3gaxian Academy of Sciences, 1975 in pre~8. 23. UN~NV, M. L.:J. Phys. Chem. 77, 1952 (1973).