Energy & Fuels l996,9, 976-983
976
NO Reduction by Activated Carbon. 6. Catalysis by Transition Metals M. J. Illan-Gbmez, A. Linares-Solano,* and C. Salinas-Martinez de Lecea Department of Inorganic Chemistry, University of Alicante, Alicante, Spain Received March 24, 1995@
Some first series transition metals (Cr, Fe, Co, Ni, and Cu) have been investigated as catalysts of the NO reduction by carbon. An activated carbon was loaded by impregnation from an excess solution of metal nitrate. The NO-carbon reaction was studied in a fixed-bed flow reactor at atmospheric pressure, using two types of experiments: (i) a temperature-programmed reaction (TPR) in a NO/He mixture; and (ii) an isothermal reaction at 300-600 "C. The products of the reactions were monitored allowing detailed oxygen and nitrogen balances. All the metals used catalyze the NO-carbon reaction, causing an important decrease in the activation energy and a substantial shift of the NO reduction curves toward lower temperatures. The TPR curves, the isothermal reactivity data, and the distribution of the products can be explained by means of an oxidatiodreduction mechanism that implies a different oxidation states of the metal and the carbon matrix. The results show that, a t low temperatures, iron, cobalt, and nickel are the most effective, as they are metals able to chemisorb NO dissociatively; at high temperatures the activity is larger for cobalt and copper, metals whose oxides are reduced by carbon at a lower temperature. This order of activity seems t o indicate that, a t low temperatures (2' < 400 "C), the controlling step in the process is the NO chemisorption, whereas at high temperatures (2' > 500 "C), the reduction of the intermediate oxidized metal species is the rate-limiting step.
Introduction In previous studies, a considerable decrease in the temperature for NO reduction by carbon was observed in the presence of catalysts such as potassium, calcium, or iron. The effectiveness of these metals was determined by their capacity for the dissociative chemisorption of N0.1-4 Following this line of research, the present paper analyzes the catalytic activity of some of the first series transition metals in the NO-carbon reaction. As in previous ~ t u d i e s , l -the ~ catalysis of NO reduction by carbon, in the absence of oxygen, is analyzed. The possibility of using transition metals as catalysts of NO-carbon reaction can be inferred from the earliest studies dealing with this r e a c t i ~ n . ~ Thus, , ~ Shelef and Otto5 detected the catalytic effect that carbon impurities have on NO reduction, and Watts6 observed that copper increases the carbon oxidation speed and modifies the product distribution. More recently, Inui et aL7 combining first series transition metals (Ni, Fe, and Co) with lanthanum oxide and precious metals (Pt,Ru, Rh, and Pd), and Yamashita et al.,899 using copper and nickel, have observed a noticeable increase in NO @Abstractpublished in Advance ACS Abstracts, September 15,1995. (1)Illdn-G6mez, M. J.; Linares-Solano, A,; Radovic, L. R.; SalinasMartinez de Lecea, C. Energy Fuels 1995,9,97. (2) I11ln-G6mez, M. J.;Linares-Solano, A,; Radovic, L. R.; SalinasMartinez de Lecea, C. Energy Fuels 1995,9,104. (3)Illln-G6mez, M. J.; Linares-Solano, A.; Radovic, L. R.; SalinasMartinez de Lecea, C. Energy Fuels 1995,9,112. (4) Illln-G6mez, M. J.; Linares-Solano, A,; Radovic, L. R.; SalinasMartinez de Lecea, C. Energy Fuels 1995,9,540. ( 5 )Shelef, M.; Otto, K. J . Colloid Interface Sci. 1989,31,73. (6)Watts, H.J . Chem. SOC.,Faraday Trans. 1958,54, 93. Y. Ind. Eng. (7)Inui, T.: Otowa, T.: Takegami, - Chem. Process Res. Dew. 1982,21,56. ( 8 ) Yamashita, H.; Yamada, H.; Tomita, A. Appl. Catal. 1991, 78, L1.
0887-062495l2509-0976$09.00/0
conversion. On the other hand, the industrial importance of the SCR (selective catalytic reduction), and the need to obtain effective catalysts to remove nitrogen oxides (NO,) a t low temperatures (2' < 200 "C) from the exhausts of both stationary and mobile energy producing devices, have recently led to the use of carbonsupported transition metals for NO reduction with CO or NH3.1°-13 Moreover, transition metals are good catalysts for carbon gasification reaction^.'^-^^ The oxidation state of the metal determines its performance, and the catalyst is only effective when it is in a reduced state.l* Therefore, the facility with which the precursor is reduced by the carbon is an important f a ~ t o r . ~ JThe ~J~ relevance of the initial state of the metal is related to the fact that, in carbon gasification reactions, the catalyst participates in an oxygen transfer mechanism that implies a redox cycle between two metal species with a different oxidation degree.lg (9)Yamashita, H.; Yamada, H.; Kyotani, T.; Tomita, A,; Radovic, L. R. Energy Fuels 1993,7 , 85. (10)Kapteijn, F.; Stenga, S.; Dekker, N. J. J.; Bijsterbosch, J. W.; Moulijn, J. A. Catal. Today 1993,16,273. (11) Singoredjo, L.; Slagt, M.; Wees, J.; Kapteijn, F.; Moulijn, J . A. Catal. Today 1990,7 , 157. (12)Wichterlovl, B.; Sobalil, 2.; Skoklnek, M. Appl. Catal. 1993, 103,269. (13)Blanco, J.; Avila, P.; Fierro, J. L. G. Appl. Catal. 1993,96,331. (14)Walker, P. L., Jr.; Shelef, M.; Anderson, R. A. In Chemistry and Physics ofcarbon; Walker, Jr., P. L., Ed.; Marcel Dekker: New York, 1968;Vol. 4,p 281. (15) Hayes, W. P.; Gasior, S. J.; Forney, A. J. Prepr. Pap.-Am. Chem. Soc., Div Fuel Chem. 1974,131,179. (16)Marsh, H.; Adair, R. R. Carbon 1982,13,327. (17)Holstein, W.; Boudait, M. J. Catal. 1982,75, 337. (18)Yamada, T.; Tomita, A.; Homina, T. Fuel 1982,63,246. (19)Mckee, D. W. In Chemistry and Physics of Carbon; Walker, P. L., Jr., Thrower, P. A., Eds.; Marcel Dekker: New York, 1988;Vol. 16, P 1. (20)Kasaoka, S.;Sakata, Y.; Yamashita, H.; Nishino, T. Int. Chem. Eng. 1981,21,419.
0 1995 American Chemical Society
Energy & Fuels, Vol. 9, No. 6,1995 977
NO Reduction by Activated Carbon The objectives of this study are to investigate the catalytic effect of Cr, Fe, Co, and Ni metals that chemisorb NO dissociatively,21s22and Cu which has been successfully used as a catalyst for the NO reduction with CO or NH3,1°-13 and (b) t o compare their reactivity, using similar experimental conditions, with the results obtained in previous studies with other metals. The joint analysis of all these results will make it possible to reach conclusions on the mechanism of the catalyzed NO-carbon reaction23 and will allow us, in a second stage and with a greater chance of success, to study the much more complex NO-carbon reaction in the presence of oxygen.
Table 1. Metal Content and Nomenclature of Metal-Loaded Activated Carbons
metal Fe cu Cr
co Ni
____
~~~
wt%
4.7 3.0 2.0 3.7 3.5
nomenclature K-UA1-Fe K-UA1-CU K-UA1-Cr K-UA-1-Co K-UA1-Ni
Table 2. Metal Crystal Size from XRD sample crystal size (nm) K-UA1-Cu 7 K-UA1-Co 6 K-UA1-Ni 3
Experimental Section
present in the carbon matrix before the NO reaction and
An activated carbon sample (K-UAl), was chosen for this This high surface area activated carbon (ScoZ= 1790 mz/g and S N=~ 1785 m2/g) is a coal-derived carbon whose activation was accomplished with KOHeZ4The catalysts were introduced by excess-solution impregnation using metal nitrate (10 mL of solutiodg of carbon). The impregnated samples were dried by first bubbling Nz through the solution and then leaving them in a n oven at 110 "C for a period of 12 h. The metal content was determined by atomic absorption spectroscopy. For this purpose, each sample was converted to ash in a muffle furnace at 800 "C over a period of 12 h, and the ashes were subsequently dissolved in a concentrated solution of HCl and analyzed. Due t o the insolubility of Crz03, the chromium was extracted from the sample with a 5/1 mixture of HN03 and HzS04. The identification of the metal species subsequent to a helium-thermal treatment (He at 50 Wmin up to 900 "C) was carried out using XRD (Seifert, JSO Debye-Flex 2002). The kinetics of the NO-carbon reaction was studied at atmospheric pressure in a fured-bed flow reactor (15 mm, i.d.; ca. 300 mg sample) connected to a gas chromatograph (Hewlett Packard Model 5892A). The gaseous products were analyzed using a Porapak Q 80/100 column and a thermal conductivity detector. Two types of experiments were carried out: (i) a temperature-programmed reaction (TPR), consisting of heating the sample at 5 "C/min to 900 "C in a NO/He mixture (0.4% NO, 60 mumin); and (ii) an isothermal reaction at 300, 400, 500, and 600 "C for a period of 2 h. In both types of experiments, the sample was first subjected to an in situ heat treatment in He, a t 50 "C/min to 900 "C for 10 min. In case (i),the temperature was lowered to 20 "C, the reactant mixture was substituted for He, and the TPR experiment was performed. In case (ii), the temperature was lowered to the desired level and the isothermal experiment was initiated by substituting the NO/He mixture for He. In both cases, the maximum amount of carbon consumed at the end of an experiment was less than 5%.
t o determine their crystal size. The K-UA1-Cu, K-UA1-
Results and Discussion Sample Characterization. The metal content of the samples and the nomenclature used are shown in Table 1. Sample characterization by means of XRD following the helium thermal treatment (see Experimental Section) allowed us to identify the metal species (21) Egelhoff, W. J., Jr. The Chemical Physics of Solid Surface and Heterogeneous Catalysis; King, D. A,, Woodruff, D. P., Eds.; Elsevier:
New York, 1982;Vol. 4. (22) Bill, J.; Rybczynski, W.; Seidel, W. Fresenius J . Anal. Chem. 1993,346,241. (23) Ill&n-G6mez, M. J.;Linares-Solano, A.; Radovic, L. R.; SalinasMartinez de Lecea, C. NO Reduction by Activated Carbon. 7. Some Mechanistic Aspects of Uncatalyzed and Catalyzed Reaction. Energy Fuels, submitted for publication. (24)Illan-G6mez, M. J.; Salinas-Martinez de Lecea, C.; LinaresSolano, A,; Calo, J. M. Energy Fuels 1993,7 , 146.
Co, and K-UA1-Ni diffractograms, although obtained after exposing the samples to air, show peaks which can be assigned to metal species; on the contrary, the K-UA1-Fe sample showed the presence of an oxide ( y Fe2O3). The formation of this oxide, presumably due to the contact with air, did not permit the presence of metallic iron to be checked, which has been confirmed from the analysis of the products evolved during TPD and TPR experiments, as discussed later. From these results it follows that, during the thermal treatment prior to the reaction, the initial metal species were reduced by the carbon. These observations agree with the results of Kasaoka et aL20 They also detected peaks corresponding to reduced metals, during a study of the catalytic effect of first series transition metals (Fe, Co, Ni, and Zn) in H 2 O and C 0 2 coal gasification, in the diffractograms of thermally treated samples (N2 up t o 900 "C). The crystal sizes of the various catalysts, calculated by applying the Scherrer equation to the highest intensity peak, are shown in Table 2. In general, fairly small crystal sizes were obtained, suggesting that an acceptable catalyst dispersion was achieved in this high surface area activated carbon. An analysis of the products evolved during the helium thermal treatment ( ( 2 0 2 and CO) complements the results from XRD sample characterization, specially for the K-UA1-Fe sample. The CO and CO2 emission curves for K-UA1-Fe,K-UA1-Co, K-UA1-Ni, and K-UA1Cu show defined peaks with maximums at temperatures near to the theoretical temperature for the reduction of metallic oxides by carbon.25 Therefore, in the K-UA1Fe sample the presence of metallic iron, although undetected by XRD, seems to be confirmed. Interestingly, the K-UA1-Cr diffractogram shows no diffraction peaks, which may be due to several reasons: (i) a high degree of dispersion, in which the average crystal size is below the detection limit of the technique, (ii)the chromium species present not being crystalline, and (iii) bearing in mind that the temperature required for chromium trioxide to be reduced by carbon is high (1250 "C),25the chromium could not have been reduced-at least to a significant extent-to metallic chromium during the thermal treatment performed (He up to 900 "C). NO- Carbon Reaction. Temperature-Programmed Reaction Studies. Temperature-programmed reaction ~
( 2 5 ) Gilchrist, J. D. Extraction Metallurgy; Int. Ser. Mater. Sci.
Technol., Vol. 30; Pergamon Press: London, 1980.
Illan-Gbmez et al.
978 Energy & Fuels, Vol. 9,No. 6, 1995
4 0 1 1
/"'y ,--
_I'
0.6
--K-UAl -K-UA1-Fe K-UAI-Cu -K-UAl-Cr -K-UAI-CO K-UAl-NI
-
L 600
700
800
h
0.4
1
0.2
05
0 100
200
300
400
(TPR) curves for the K-UA1 carbon and the samples containing the different metals are compared in Figure 1. All the curves show the catalytic effect of the metals, which causes a significant decrease both in the starting reaction temperature and the temperature required to reach 100% reduction (2'100). However, the catalytic effect of each metal in these two parameters varies: (i) according to the temperature at which the reaction starts; iron, cobalt, and nickel are the metals that show the highest activity, (ii) the T m values are quite different and depend considerably on the nature of the metal used: 500 "C for copper and cobalt, 520 "C for nickel, 585 "C for chromium, and 600 "C for iron. The 2'100 values observed in this study are similar to those obtained by Inui et al.,7 who used a catalyst in which they combined iron, cobalt, or nickel with lanthanum oxide and platinum, supported on activated carbons. The following observations can be made with regard to the shape of the TPR curves: 1. Chromium and copper do not significantly modify the TPR curve shape of the parent carbon; only a displacement of the curve towards lower temperatures is observed. 2. Cobalt and nickel show a TPR profile different from that of the original K-UA1 carbon; the two readivity steps observed in K-UA1, and discussed in detail p r e v i ~ u s l ycannot , ~ ~ be distinguished in the presence of these two metal catalysts. 3. Iron noticeably modifies the entire TPR profile, showing a maximum NO reduction at low temperatures (around 325 "C), an interval at intermediate temperatures in which a catalyst deactivation is observed, and a final step a t high temperatures (>450 "C), in which iron recovers its activity, This modification of the TPR profiles is similar to that found for potassium as a catalyst of the NO reduction by c a r b ~ n . ~ , ~ , ~ ~ The activation energy for the NO-carbon reaction may also be affected by the presence of metals as was found in a previous study with some of the carbons analyzed that contained metallic i m ~ u r i t i e s . The ~~ apparent activation energies, in the low temperature range (between 200 and 500 "C) have been calculated by application of the Arrhenius equation to the TPR results, assuming a global first-order kinetics with respect to and assuming that the microreactor (26) Suuberg, E. M.; Teng, H.; Calo, J. M. 23rd Symposium (International) Combustion [Proceedings]; The Combustion Institute: Pittsburgh, 1991; p 1199.
600
700
0 900
800
T CC)
T W)
Figure 1. Temperature-programmedreaction (TPR)profiles for NO reduction by activated carbon K-UA1 and transition metal loaded activated carbons.
500
Figure 2. Evolution of products during TPR experiment for iron-loaded carbon K-UA1-Fe. Table 3. Apparent Activation Energy sample
E, (kJ/mol)
sample
E , (kJ/mol)
K-UA1 K-UA1-Fe K-UA1-Cu
86 28 41
K-UA1-Cr K-UA1-Co K-UA1-Ni
59 25 30
operation approximates plug flow behavior. The results, compiled in Table 3, show that the activation energies of carbon samples containing metals are lower than those of the noncatalyzed NO-carbon reaction.24 Iron, cobalt, and nickel, the metals showing the highest activity at low temperature, are those featuring the lowest activation energy values. The analysis of the gases produced from the reaction (NzO, N2, CO, and CO2) and of the reacting gas (NO) during the TPR experiments completes the data supplied by the reduction curves. Figures 2-6 show the gas evolution profiles for each of the transition-metal catalyst studied. Figure 2 shows the gases produced from K-UA1-Fe, which differ greatly from that of the parent free-metal carbon.24 A detailed description was offered in a previous paper.4 In summary, it must be pointed out that four stages of the TPR profile can be described: the first reaction stage (T< 280 "C)consists of a NO dissociative chemisorption. This chemisorption process is characterized by the evolution of NzO and N2, as the only reaction products, and oxygen retention in the catalyst; the second stage (280 < T < 400 "C) is accompanied by N2 release and a loss of catalyst activity; in the third stage (400 < T < 630 "C), C 0 2 release starts (showing an excess compared to N2) and the catalyst recovers its activity; finally, CO evolution starts in the last reaction step (2' > 630 "C) in which the N2 evolution becomes constant. The CO emission curve shows a defined peak at around 800 "C, which is superimposed on the continuous CO release, which is typical of carbon.24This well-defined CO peak (also present in the cobalt sample) strongly suggests that the iron during the TPR is in an oxidized form only at tempergtures lower than 700 "C. Quantification of this CO peak (subtracting the CO corresponding to the metal-free parent carbon) confirms that (i) this peak is related to the reduction of the oxidized iron species by carbon, upon heat treatment, with a mol CO/mol iron ratio close to 1,4and (ii) the iron following the heat treatment in the TPD experiment has more reason to be in its reduced form (as discussed above) because NO is not present. Figures 3 and 4 present the gases produced in the TPR in the presence of cobalt and nickel, respectively.
NO Reduction by Activated Carbon
Energy & Fuels, Vol. 9, No. 6, 1995 979
_. .
0.4 OS 0.2
05
0.2
0
100
200
300
400
500
600
700
800
0 900
n
n
T(W
Figure 3. Evolution of products during TPR experiment for cobalt-loaded carbon K-UA1-Co. n (Irmol/g Us)
n Cmol/g C/s) --CO -NO
O8
o'6 0.4
2
=\
"*COz-Nfi
t
0.8
I/\
-CO
-NO **'COz
-
c
0.6
1
0.4 OS
0.2
0 100
05 0.2
200
300
400
500
600
700
800
0
900
0
100
200
300
400
500
600
700
800
0 900
T CC)
TW)
Figure 4. Evolution of products during TPR experiment for nickel-loaded carbon K-UA1-Ni.
Figure 6. Evolution of products during TPR experiment for chromium-loaded carbon K-UA1-Cr.
N2O production can be observed in the low-temperature region which implies NO dissociative chemisorption. However, if N20 emissions are compared for the samples containing iron, cobalt, and nickel, the smaller production of this product for the two last catalysts suggests their lower chemisorption capacity in this temperature range. Additionally, a displacement of N2 and N2O to a higher temperature is observed which makes them coincide with C02 emission. In other words, the beginning of the reaction, associated with the C02 emission,lS2 overlaps with the N2 and N2O evolutions. Therefore, for cobalt and nickel, the dissociative NO chemisorption process cannot be observed as a separate step from the beginning of the reaction, as occurs with iron and also with p o t a ~ s i u m . ~ , ~ , ~ The performance of cobalt and nickel seems to be identical in principle, which is reasonable considering that these two metals have quite similar chemical behavior. However, a closer look a t these TPR curves shows some differences; nickel seems to be slightly more active than cobalt because it shifts the NO reduction curve and the emission temperatures of C02 and CO toward somewhat lower temperatures. Additionally, if we compare the oxygen balance of these samples, at temperatures lower than 700 "C, the balance is closed for cobalt whereas in the case of nickel the oxygen balance is negative between 280 and 680 "C due to the oxygen retention in the sample. It seems, therefore, that nickel displaces the dissociative chemisorption stage toward higher temperatures, which makes it coincide with the beginning of NO reduction, that is, with the release of C02. This fact, which was observed neither with iron4 nor with potassium,lS2 may be of interest because nickel produces very little N2O and its
catalytic activity is good. At high temperatures (above 700 "C), nickel, contrary t o iron and cobalt, does not show a well-defined CO peak. This indicates that nickel oxide is not present, suggesting that its reduction by the carbon occurs more easily. This in turn implies that the oxygen transfer from the nickel species to the carbon matrix is more favored than in the cases of cobalt and iron species. Considering the data in Table 2, this oxygen transfer could be favored by the smaller crystal size in the sample containing nickel. The gases produced during the W R experiment of the K-UA1-Cu sample, which is shown in Figure 5, is similar to that of the K-UA1 carbon;24NO reduction is accompanied by a release of N2, C02, and CO as reaction products. Nevertheless, the catalytic effect of the metal causes a noticeable displacement of the beginning of the reaction. N2 and C02 releases occur a t a much lower temperature (300 "C compared with 400 "C for the metal-free parent carbon), although the temperature at which CO emission starts is not affected by the presence of copper. The lack of a high-temperature CO peak, explained as in the case of nickel, is more favorable considering the higher propensity of copper species t o be reduced by carbon. Figure 6 corresponds t o the TPR profile of the K-UA1Cr sample. The presence of the metal reduces both the N2 and COZ starting evolution temperature (300 "C compared with 400 "C in K-UA1) and the CO emission temperature which takes place a t a much lower temperature than the one observed for the original carbon (550 "C compared with 650 "C). As mentioned above, XRD and TPD show that chromium has not been reduced to the metallic state to a significant extent during helium thermal treatment, because the temper-
Illan-mmez et al.
980 Energy & Fuels, Vol. 9,No. 6, 1995
ature needed for chromium oxide to be reduced by carbon (1250 0C)25is higher than the maximum temperature reached in such treatment (900 "C). As a result of this, we did not observe, in the low-temperature step, the expected behavior of chromium to chemisorb NO dissociatively (chromium has been classified within the group of metals which chemisorb NO dissociatively at low temperatures).21,22However, the fact that the C02 and CO release temperature is lower than that of carbon seems to indicate that chromium, even in its oxidized form, catalyzes NO reduction by carbon. In summary, two interesting observations can be extracted from the TPR experiments: (i) The magnetic properties of the metals seem to affect the NO-metal interaction because of the presence of an unpaired electron in the NO m ~ l e c u l e . ~ This ~ , ~molecule ~ could interact with unpaired electrons existing on the carbon surface, forming C-NO species. The presence of metals with unpaired electrons (paramagnetic or ferromagnetic) might modify this interaction and hence could explain, in this low-temperature region, that with iron, cobalt and nickel-all ferromagnetic metals-the NO dissociative chemisorption occurs, whereas with copper, a diamagnetic metal, this process was not observed. (ii) The distribution of oxygen-containingproducts ( ( 3 0 2 and CO) during the reaction differs from that observed in the K-UA1 carbon.24 This observation agrees with the results of the gasification reactions of the carbon with 0 2 , C02, and H2O vapor, where the metals also modify the evolution of the carbon gasification products.20 This similarity is reasonable, because the NO-carbon reaction is also a carbon gasification reaction. Isothermal Reaction Studies. Isothermal reaction experiments at 300-600 "C were performed, during 2 h, for all the metal-carbon samples to obtain easily comparable catalytic activity data. This set of experiments, in which the NO reduction and the product evolution profiles are obtained, gives complementary reactivity information to the above TPR data. Figures 7-11 present the variation of the NO reduction activity with time (expressed in pmol of NO reduced per gram of carbon and second). It should be pointed out that the activity curves show a complex dependence on the nature of the catalyst used and temperature, although the experimental conditions are the same for all the metal-carbon samples studied. As can be seen, the activity curves for all the catalysts at 300 and 400 "C exhibit a continuous decrease after which a constant level, different for each metal, is approached asymptotically. The extent of the activity decrease, before reaching the steady state, depends on both the nature of the catalyst and the temperature of the reaction. In this low-temperature range, it is observed that the metals that chemisorb dissociatively NO (Fe, Co, and Ni) show a remarkable activity, which decreases for approximately 20-30 min. This observation in turn confirms both the role of these metals as an acceptor of oxygen from the reduction of NO (see comments below about oxygen balance) and the ability/ inability of these oxidized metal species to transfer oxygen t o the carbon and recover their reduced state. In the low-temperature range NO dissociative chemi(27) Zarifayanz, Y. A.; Kiselev, V. F. Lezhnev, N. N.; Nikitina, 0. V. Carbon 1967,5 , 127. (28) Solbaken, A.; Reyerson, L. H. J. Phys. Chem. 1960,64, 1903.
0.6
0.5
0.4
03
0.2
0.1
n "0
20
40
60
80
100
120
t (min)
Figure 7. Isothermal NO reduction activities as a function of time for iron-loaded carbon K-UA1-Fe.
sorption occurs in the catalysts forming an oxidized species. If these species do not transfer oxygen to the carbon (in other words, if the oxidized catalyst is not reduced by the carbon) the reaction stops. At low temperature the oxygen transfer from the catalyst to the carbon is very slow and causes a noticeable decrease in activity. The period of high activity becomes longer as the reaction temperature increases. On the contrary, those metals that do not show a NO dissociative chemisorption process (Cu and Cr) have almost no catalytic activity for the reduction of NO, confirming that they are poorer oxygen acceptor from NO. Once the steady state is reached, it can be observed that the activity increases with the reaction temperature. Both the increase in activity with the reaction temperature and its evolution with time differ from one metal to another. At steady state, iron shows a very low activity at 300500 "C (Figure 71, whereas at 600 "C it maintains a constant 60% reduction after 2 h of reaction. The main feature for cobalt is the high activity observed at temperatures above 400 "C (Figure 8). In the presence of nickel (Figure 9), the evolution is similar to that observed for the sample containing cobalt. In this case, however, the increase in activity with temperature is less pronounced because both its activity a t low temperature is higher and the activity at 400 and 500 "C is lower than that shown by K-UA1-Co. The activity with copper (Figure 101, at steady state, is very low at 300 and 400 "C but interestingly enough, it is very high at 500 and 600 "C, being practically constant for the 2 h of reaction. Chromium (Figure 11)is only really effective at 600 "C. It is important to recall that at 600 "C chromium and copper are by far the most effective. Analysis of the reaction products evolved during the isothermal experiments complements the information supplied by the activity curves and help to interpret them. Figure 12 (a, b, and c) shows, as an example, the evolution of the reaction products at 400 "C of K-UA1-Co, K-UA1-Ni, and K-UA1-Cu, respectively.
NO Reduction by Activated Carbon 0.6
Energy & Fuels, Vol. 9,No. 6,1995 981
Activity OlmoVg C/s)
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
\
0. I
0
0.1
\
0
20
Activity hmoVg Us)
40
60
80
0 100
120
AI,-
0
20
40
Figure 8. Isothermal NO reduction activities as a function of time for cobalt-loaded carbon K-UA1-Co. Activity @moVg C/s) -3OO'C
-4OO'C
80
100
120
t (mln)
t (min)
0.6
60
1
= S W C -6Oo'C
Figure 10. Isothermal NO reduction activities as function of time for copper-loaded carbon K-UA1-Cu. 0.6
Activity @moVg C/s) -3WOC
0.5
0.5
0.4
0.4
-4OO'C
~'*~500oC -600T
0.3
03
0.1
--- _0
0
--
I
I
I
I
,
20
40
60
80
100
0 120
t (min)
Figure 9. Isothermal NO reduction activities as a function
t (min)
Figure 11. Isothermal NO reduction activities as a function
of time for nickel-loaded carbon K-UA1-Ni.
of time for chromium-loaded carbon K-UA1-Cr.
At this point, it is useful to relate the catalytic activity of the metals studied with their physical and chemical characteristics. An interesting common feature of most of these curves is that the COZ evolution is initially delayed with respect t o the evolution of N2 giving rise to a negative oxygen balance. This behavior is largely observed in the presence of paramagnetic or ferromagnetic metals, iron (not shown in Figure l2I4 and cobalt. Copper almost does not show this delay and nickel only slightly. This C02 delay agrees with previous results reported with potassium182 and iron4 and confirms the above statement that iron, cobalt, and nickel (to a less extent) retain the oxygen produced by the dissociative NO chemisorption. Once C02 evolves, as a result of the reduction of the oxidized metal species by carbon, the catalytic redox cycle is closed and the steady state may be reached. This is confirmed when comparing the
product emission from K-UA1-Co, K-UA1-Ni,and K-UA1Cu with the corresponding activity curves (Figures 8, 9, and 10) because it can be observed that the samples show activity only if they can keep a noticeable COZ emission level which allows the closure of the catalytic cycle. AH a result, copper shows no activity a t this temperature, whereas cobalt and nickel are active. In fact, cobalt is more active than nickel because, at steady state, it maintains a higher COz emission level. Regarding this issue, it has been accepted29 and it is confirmed here that the emission of oxygen-containing products (CO2 and CO) from a carbon oxidation process constitutes a limiting step for any carbon-gas reaction. (29) Radovic, L. R.; Lizio, A. A.; Jiang, H. In Fundamental Issues in Control of Carbon Gasification Reactiuity; Lahaye, J., Ehrburger, P., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1991; p 235.
Illan-G6mez et al.
982 Energy & Fuels, Vol. 9, No. 6, 1995 % NO reduction
,00 "V
0.5
i \ t \ I
60-
-
0.4
0.3
40 2ot
\
0.2
I
0
200
400
Tred 0.1
O
1 j) 0
J
20
40
60 t
,
80
100
80
100
EO
100
I
120
(min)
0.6
0.5
0.4
0.3
0.2
0.1
0
20
40
60 t (min)
0
I
0
20
40
60 t
120
(min)
Figure 12. Evolution of products during isothermal NO reduction at 400 "C of samples: (a)K-UA1-Co; (b) K-UA1-Ni; (c) K-UA1-Cu.
600
800
PC)
Figure 13. Correlation between the activity for NO reduction and the reduction temperature for metal oxides by carbon.
The evolution of COZ allows the catalyst to continue transferring oxygen from NO t o carbon and thus close the redox cycle, generating new active sites on the carbon surface. Although at low reaction temperatures the dissociative chemisorption of NO has to be considered and its contribution increases as the reaction temperature decreases, this must not be seen as the only factor determining the catalyst activity of the metal, because other chemical properties of the metal have t o be considered. Thus, since the emission of oxygen-containing products is the limiting step of the reaction, the reduction temperature of the metal oxide by carbon must be, in principle, a key factor controlling the activity of the metal as a catalyst. To check this, the NO reduction percentage at 500 and 600 "C has been represented in Figure 13 vs the reduction temperature of the metal oxide by carbon, deduced from the Ellingham diagrameZ5An interesting fairly linear relationship between these two magnitudes is observed. Copper is the metal showing the highest activity at 500 and 600 "C because it is the one that can be most easily reduced by carbon. Furthermore, Figure 13 supports previous r e s ~ l t s in ~ which ~ - ~ ~a relationship has been established between the ease of oxide reduction (to metallic or lower oxidation state) and the lattice energy of oxide, or the free energy of oxide formation, and also the fact that all metal oxides that can be reduced by carbon in the 500-700 "C temperature range are effective as catalysts in carbon oxidation.30 As shown in Figure 13, the reduction of the intermediate oxidized metal species (for copper, cobalt, and nickel) cannot be the factor limiting COZ evolution, because the oxides of these three metals are easily reduced (5" .C 400 "C) by carbon. If the metal oxide reduction by carbon is not the limiting step of C02 emission, then the formation of metal oxide, from the oxygen originating in NO dissociative chemisorption on the metal, will be the factor limiting the COZevolution speed. Therefore, the scarce activity of K-UA1-Cu at 400 "C (when compared with K-UA1-Co and K-UA1Nil must be related with the low chemisorption capacity shown by copper. (30)Hemtz, E.A.;Parker, E. Carbon 1966,4,473. (31)Harris, P. S.Carbon 1972,10,643. (32)Mackee, D.W.Carbon 1970,8,623.
NO Reduction by Activated Carbon Activity (mol NO red./mol met.s)*lO -4
Energy & Fuels, Vol. 9, No. 6, 1995 983
Conclusions
A
dL
n 3 o o o c 0400oc
OSOO~C
The study of the catalytic activity of chromium, iron, cobalt, nickel, and copper in the NO carbon reaction has 1000 allowed the following conclusions to be drawn: all the 800 metals studied catalyze the NO-carbon reaction, caus600 ing a noticeable decrease in the activation energy and a substantial shift toward lower temperatures of the NO 400 reduction by carbon. Both dissociative chemisorption 200 I!-! ! with the oxygen acceptor capacity of of NO, associated 0 the metal, and the ease of the reduction by the carbon K Ca Cr Fe Co Ni Cu of the metal oxide formed are important steps. The Figure 14. Isothermal reaction activity at 300,400, and 500 metals with magnetic properties (iron, cobalt, and "C or metal-loaded carbons. nickel) show NO dissociative chemisorption a t low temperatures (300 and 400 "C), with the exception of In summary, the results of this paper allow to state chromium and copper, which do not exhibit such bethat both the dissociative chemisorption of NO (associhavior because the former is in an oxidized form and ated to the oxygen acceptor capacity of the metal) and the second is diamagnetic. At higher temperature (500 the ease of the reduction of the metal oxide formed by and 600 "C), a linear relationship has been found the carbon are important steps and that their relative between the NO reduction percentage and the tempercontributions depends on the reaction temperature. ature a t which the metal oxide can be reduced by Regarding the comparison of transition metals with carbon. The relative contributions of the dissociative alkali and alkaline-earth metals, the specific activity a t process (which increases as the reaction temperature steady state a t 300,400, and 500 "C (expressed as mol decreases) and the ease of the reduction of the oxidized of NO reduced per mol of metals and per second) have metal species by carbon depend on the nature of the been calculated and plotted in Figure 14. The histoinetal and on the reaction temperature. Therefore, a t gram stresses the importance of reaction temperature low temperatures, iron, cobalt, and nickel are the most in the activity sequence for this metal series. At low effective, as they are metals able to chemisorb NO temperatures, potassium, iron, cobalt, and nickel are dissociatively;a t high temperatures they are cobalt and the most effective, as they are metals able to chemisorb copper, metals whose oxides are reduced by carbon at a NO dissociatively; at high temperatures they are potaslower temperature; and chromium is the less active metal. This order of activity seems to indicate that, a t sium, cobalt, nickel, and copper, metals whose oxides low temperatures (T 400 "C), the controlling step in are reduced by carbon a t a lower temperature. This the process is NO chemisorption, whereas a t high order of activity seems to indicate that, at low tempertemperatures (2' > 500 "C), the reduction of the interatures (T< 400 "C), the controlling step in the process mediate oxidized metal species is the rate-limiting step is NO chemisorption, whereas a t high temperatures (T in the reaction sequence. > 500 "C), the reduction of the intermediate oxidized metal species is the rate-limiting step in the reaction Acknowledgment. This study was made possible sequence. It is clear that potassium exhibits a very high by financial support from DGICYT (project AMB92activity all over the temperature range, only exceeded 1032-C02-02), OCICARBON (C-23-435), and by a reby copper at high temperature. This in turn confirms search grant to M.J.I.G. from Caja de Ahorros del that potassium has a high performance for NO dissociaMediterraneo (CAM). tive chemisorption forming species that are easily EF950056S reduced by carbon.1,2
