DOE / PC94218-8

KINETICS AND MECHANISMS OF NOX - CHAR REDUCTION

E. M. SUUBERG (PRINCIPAL INVESTIGATOR) W.D. LILLY (STAFF) I. AARNA (Ph.D. STUDENT)

DIVISION OF ENGINEERING BROWN UNIVERSITY PROVIDENCE, RI 02912 TEL. (401) 863-1420

QUARTERLY TECHNICAL PROGRESS REPORT 1 MAY, 1996 - 31 JULY, 1996

PREPARED FOR: U. S. DEPT. OF ENERGY PITTSBURGH ENERGY TECHNOLOGY CENTER P.O. BOX 10940 PITTSBURGH, PA 15236

DR. RODNEY GEISBRECHT TECHNICAL PROJECT OFFICER

"US/DOE Patent Clearance is not required prior to the publication of this document"

DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

1. Introduction

The emission of nitrogen oxides from combustion of coal remains a problem of

considerable interest, whether the concern is with acid rain, stratospheric ozone chemistry,

or "greenhouse" gases. Whereas earlier the concern was focused mainly on NO (as a

primary combustion product) and to a lesser extent NO2 (since it is mainly a secondary

product of combustion, e.g. see ref. l), in recent years the emissions of N20 have also

captured considerable attention2-8, particularly in the context of fluidized bed combustion,

in which the problem appears to be most acute. The research community has only recently

begun to take solid hold on the N20 problem. This is in part because earlier estimates of

the importance of N20 in combustion processes were clouded by artifacts in sampling

which have now been resolvedg. This project is concerned with the mechanism of

reduction of both NO and N20 by carbons.

It was recognized some years ago that NO formed during fluidized bed coal

combustion can be heterogeneously reduced in-situ by the carbonaceous solid intermediates

of combustionlO. This has been recently supplemented by the knowledge that

heterogeneous reaction with carbon can also play an important role in reducing emissions

of N202p6,7, but that the NO-carbon reactions might also contribute to formation of

N202,*. The precise role of carbon in N20 reduction and formation has yet to be

established, since in one case the authors of a recent study were compelled to comment that

"the basic knowledge of N20 formation and reduction still has to be improved"8 The same

can be said of the NO-carbon system.

Also of importance in the context of "practical" application of the reported NO-

carbon reaction kinetics is how other gases, particularly CO, C02,02 and H20 might

affect these kinetics. There have been only few systematic studies concerning these effects.

Interest in the NO- and N2O-char reactions has been significant in connection with

both combustor modeling, as well as in design of post-combustion NOx control strategies.

As in the case of the NO-char reactions, the reaction of N20 with char is probably too slow

1

to be of significance in dilute particle phase, short residence time, pulverized coal

combustion environments3. The suggestion has been made that the reactions could still be

important within the pore structure of the coals, even in a pulverized firing environmentl1.

The possibility of rebuming combustion gases in the presence of fresh coal or char also

exists.

The above chemical processes are, however, unquestionably important in the lower

temperature, slower reaction rate regime of fluidized beds8. Of course, it is also the lower

temperatures of fluidized bed systems that lead to release of greater amounts of N20 from

these systems, since the N20 destruction processes have higher activation energies than do

formation processes7. Therefore, there remains a significant incentive for studies of these

reactions associated with developing better control strategies associated with fluidized bed

technologies.

Beyond the applicability of this chemistry in fluidized beds, there is interest in

developing new post-combustion processes to control NOx emissions. The possibility of

using carbons in the role of catalysts for the catalytic DeN%-type processes has been

exploredl2. Their possible roles as catalyst supports has also been e~aminedl3,1~. The

use of activated carbons for NO removal has been ~tudied~2?~5,16. And as noted above,

the use of carbons, with various kinds of catalytic promoters, has been suggested as

holding some promise for lowering the useful temperature range of the reduction processes

into that of interest for post-combustion processing17,1g,19,20. Interestingly, it was even

suggested a few years ago that even spent oil shale, which contains char in a largely

limestone matrix, could be an effective material for reduction of N&1,22.

2. Experimental

During this quarter, the primary focus of activity continued to be the effects of other

non-inert gases on the rates of NO reduction by carbon. Earlier, we had explored the effect

2

of CO on this reaction, and found that this carbon oxide enhanced the rate. The same

general experimental strategy was employed here to look at the effects of C a on the

kinetics of the NO-carbon reaction.

A quartz packed bed reactor tube of 4 mm internal diameter and 500 mm overall

length was used. A bed of 20-200 mg (in a predetermined length of 1-30 mm) of char

particles, held in place with quartz wool, was located at the center of the tube. The reactor

was heated by an electrical tube furnace and a chromel-alumel thermocouple was placed

inside the reactor tube for temperature measurements.

To test the importance of NO reduction by the quartz reactor tube and the quartz

wool, blank runs were conducted by passing NO mixtures over the bed materials at

different temperatures. The results showed no significant NO reduction (4%, invariant

with temperature). The same was not true when ordinary glass wool was used in the

reactor. In this case, a few percent NO reduction was observed.

The reactor was outgassed prior to each run by a one hour vacuum pumping at

room temperature. This was followed by thermal surface cleaning (at 1173K for 1-2 hours

in He) to remove surface oxides. Nitric oxidehelium mixtures of the desired concentration

levels were prepared using a KIN-TJ2K precision calibration system. The desired NO/He or

NO/C02 mixtures (10-300 ppm of NO in He or C02) were obtained by controlling the

flow rate of helium or C@, and both the absolute temperature of the permeation membrane

and the NO partial pressure in KIN-TEK calibration system. In experiments in which the

effect of C@ concentration was explored, the desired CqZ concentration (0.5-10% of C02

in NO/He mixture) was obtained by controlling the flow rate of C@.

During the reactivity measurements, the product gases were continuously analyzed

for NO and NO2 using a chemiluminescence analyzer. The other product gases, such as

C 0 2 and CO, can be analyzed by FTIR, but this was done only during the C02

gasification experiments.

3

The TA Instruments thermogravimetric analyzer was used for the C02 gasification

studies. The temperature and sample mass were recorded as a function of time. Products

formed during the gasification experiments can be analyzed using a Bomem FTIR

spectrometer, fitted with a multipass gas cell. Samples of 30-50 mg were dispersed on a

circular platinum pan with a large, flat surface and raised sides, resulting in thin particle

beds. The sample bucket was located in the heated zone of the TGA furnace, and a K-type

thermocouple was placed about 5mm from the sample.

The gasification reactions were performed in a flow of C o m e mixture, with a

purge gas flow rate of 100 mL/min. The reactant gas mixture contained 2% CQ in He.

Temperature programmed reaction (TPR) was carried out for all samples, at a linear heating

rate of 10 "C/min to a maximum temperature of 1o00"C.

Specific surface areas of char samples were determined by the N2 BET method at

77 K. A standard flow-type adsorption device (Quantasorb) was used for the

measurements. prior to surface area analysis, all samples were outgassed in a flow of

nitrogen at 573 K for 3 hours.

The carbonaceous solids studied were resin char, graphite and Wyodak coal char.

The resin char (1 80-29Opm) samples were derived from phenol-formaldehyde resins made

in-house. The coal char (150-212pm) was also prepared in-house from the Wyodak sample

of the Argonne Premium Coal Sample Program. The chars were prepared by a two hour

pyrolysis, in inert gas, at a temperature of approximately 1273 K. The graphite for the

packed bed measurements (180-29Opm) was prepared from graphite rods. None of these

samples was treated any further, except for surface cleaning.

3. Results

During the last year or so, the main focus was on NO reduction by different

carbonaceous materials (such as graphite, resin char and Wyodak coal char); however, the

effect of coexisting gases such as CO, C02, 0 2 and H20 is not yet fully established.

4

Generally, and CO appear to enhance the rate of reduction of NO. During this quarter,

the effect of C02 on reduction of NO was investigated. It was found that the reduction of

NO was inhibited by the addition of CO2, where the effect was largest with graphite and

insignificant with respect to resin char and Wyodak coal char.

Figure 1 shows the kinetics of the NO graphite and NO/CO2 graphite reactions. In

this case, the mixture contained 60 to 100 ppm of NO in roughly 2% C02, balance helium.

Both sets of results (for C02-containing and CO2-free gas) show a "break" at around

8OO0C, where a significant change in the mechanism of reduction of NO occurs. Similar

activation energies were found in the different temperature regimes in both cases.

From Figure 1, it can clearly be seen that the NO-reduction is strongly inhibited by

the addition of C02. The reaction rate in the NO/C02 graphite system is, on average,

lower by a factor of 1.7. Those results imply that C02 is not an inert gas in this system.

Figure 2 shows the NO reduction rate as a function of initial C02 pressure at

80O0C, for graphite. Clearly the rate of NO reduction depends on CO2 partial pressure and

it appears as though the reactivity goes through a minimum at around 1.5 P a . The

concentration dependence is very strong at lower concentrations, and weak at higher

concentrations.

The Arrhenius plot for the NO-Wyodak coal char and NO/C02-Wyodak coal char

reactions is given in Figure 3. In contrast to the situation with graphite, in this case, there

is almost no difference in reactivities in the presence or absence of C02, when C02 partial

pressure is below 2 kPa. The inhibiting effect of C@, however, becomes obvious when

its partial pressure is around 100 kPa. Thus the extent of inhibition of the carbon-NO

reaction depends in some way upon the nature of the carbon itself.

In addition to the steady rate experiments discussed above, some transient

experiments were performed with the Wyodak char. The results obtained with different

concentrations of CO2 at 800°C are shown in Figure 4. From this figure, it can be seen

that with increasing C02 partial pressure, the time to "NO breakthrough" decreases, and

5

therefore the "pseudo-steady state" is achieved at different times for different C02 partial

pressures. The pseudo-steady state is obtained sooner, the higher the C02 partial pressure.

As already implied by the results in Figure 3, the 2% C02 case gives a steady reactivity that

is comparable to that in NO-helium mixtures. The steady state, however, suggests no

effect of low concentrations of C@, but the transient reveals a significant effect. It appears

as though sites capable of reducing NO are more quickly destroyed in the presence of Cq2,

even if the ultimate reactivity is largely unaffected. This could indicate the existence of two

different types of sites in Wyodak; one might be the type that dominates in graphite, and the

other is of a different kind, relatively less affected by C02, and dominates the steady state

reactivity in Wyodak char. This, at the moment, is little more than speculation.

To establish whether the effect of Co;! merely has to do with a competition between

C02 and NO for the same adsorption sites, an additional experiment was performed in

which Wyodak coal char was first treated with 2 kPa C02 in He for 30 minutes, followed

by introduction of NO to the reactor. As can be seen, the initial treatment with C02 did not

alter the transient behavior of the reaction system, relative to its behavior in inert gas-NO

mixtures. Thus, the gasification of the carbon in C@ does little to the nature of the sdace

dative to the main NO reaction. This could mean that the C02 and NO must be present at

the same time in order for C02 to influence the reduction of NO. This seemingly implies

that there is a competition for active sites between NO and C02, but that empty sorption

sites simply left behind by surface cleaning are not those sites. Another possibility is that at

8OoC, the active sites that are filled during C02 gasification desorb upon removal of Cm,

and that the surface is essentially again virgin surface. This would explain why the

transients for pure NO on oxidized and unoxidized surface are similar. The case in which

C02 is fed continuously at 2% level would result in filling up of active sites more quickly,

and hence, the shorter time to breakthrough.

The results with resin char are in some respects similar to those from Wyodak coal

char. The inhibiting effect of C02 is insignificant at 2% concentrations (see Figure 5).

6

Figure 6 shows the NO reduction rate as a function of initial C02 pressure at 800"C, for

the same resin char. In this case, the rate of NO reduction does not depend strongly on

C02 partial pressure, even at C02 partial pressures as high as 101 kPa.

Clearly the role of C02 gasification by itself must be understood to begin to clarify

the role of C02 in inhibiting the NO reduction. To this end, experiments were performed

with all three carbon materials, in order to establish their reactivities to C02 itself.

Temperature Programmed Reactivity, TPR, experiments were performed with a linear

heating rate of 10°C/min in an atmosphere of pure C02. Figure 7 shows that the results in

terms of fractional mass loss. The behaviors of all three samples during C02 gasification

are different from those of the other two. Figure 8 emphasizes the difference, showing the

mass spectrometrically determined yields of CO during the runs of Figure 7.

During the C02 gasification of Wyodak coal char the evolution of CO started at

around 200°C and the main gasification started at 650°C. For resin char the low

temperature CO evolution started at 450°C and the main gasification at 800°C. During the

C02 gasification of graphite, low temperature evolution of CO is almost missing and the

main gasification starts at temperatures as high as 900°C. This indicates the different nature

of the active sites involved in C@ gasification in each of the three cases.

4. Discussion

Only the NO reduction on graphite was seen to be sensitive to the presence of C02

at low concentrations. The reasons for the differences between different materials remain

unclear. Above, it was speculated that there might be a role of two different types of sites in

the case of the Wyodak coal char. Other possibilities cannot., however, be ruled out. For

example, both resin char and Wyodak coal char have a similar porous structure, whereas

graphite is a highly non-porous material. Perhaps there is some role of a transport

limitation. Another possible factor is the concentration of CO in the products of the

7

gasification reactions themselves. The more reactive char materials produce much more CO

than does the graphite. Since the CO promotes NO reduction via a carbon surface catalyzed

route, the Cq;! inhibition could be masked in systems in which the CO reaction plays a

bigger role. This latter possibility will be further considered below.

To understand the inhibiting effect of C02 on the reduction of NO, it is necessary

to consider the Cq;! gasification reaction. The C02 gasification reaction has been studied

extensively over the last fifty year~~3-27. Ergun23 proposed a so called single-site oxygen

exchange mechanism for C@ gasification:

co, + c*.A’-.co k-1 + C(0) c(o)*co+ c*

where C* is a vacant carbon active site, and C(0) is carbon-oxygen surface complex. The

first step (Rl) in the mechanism is the reversible oxygen-exchange reaction and the second

step is the desorption step which generates new active sites for gasification. From the

above mechanism, the reaction rate can be expressed as: k; * c*-P

k, k,

co2 r = h. Pco* +b. P,, +1

where C* is the concentration of active sites, and Pco2 and Pco are the partial pressures

of C02 and CO, respectively. From the proposed mechanism, it can be seen that during

Cq;! gasification, some CO is produced and oxygen surface complexes are formed. These

oxygen complexes can play a significant role in the reduction of NO, as the earlier work in

this laboratory has shown.

As an alternative to the above, Koenig et al.28929 proposed a mechanism involving

a two site adsorption and dissociation of C@ :

c0*+2c*-*c** k-3

c * * v * c ( o ) + C(C0)

c ( c o ) ~ ~ c * + c o C(O)+CO

8

where C* and C** are a single active surface site and a two-site surface complex,

respectively, and C(0) and C(C0) are surface complexes. The f is t two steps (R3) and

(R4) represent the formation of surface complexes and steps (R5) and (R6) are the

desorption steps.

Both mechanisms show a formation of CO and surface complexes during the (2%

gasification, as is required by the data. The difference between mechanisms is that in the

two site oxygen exchange mechanism, the formation of two types of surface complexes are

proposed. As noted above, we earlier showed that addition of even small amounts of CO

(100 ppm) increases the reduction of NO significantly, and of course CO is also a product

of the C02 gasification reaction. What is less clearly established is how the presence of

surface oxides from C02 gasification affects the reaction rate, though the results of Figure

4 imply that in certain regimes they may have little effect. The biggest unknown is the

effect of differences in the active sites. These are only qualitatively represented in the above

mechanisms, and perhaps represents the biggest gap in our knowledge of these reaction

systems. The difference in active sites has shown itself in the results of TPR in C02.

At the moment, we feel that the most probable reason why we did not see any effect

of low concentration C02 inhibition in the case of Wyodak coal char and resin char is the

low temperature evolution of CO. In the last quarterly report we showed that even small

amounts of CO (50 ppm) increases the reduction of NO significantly. If the evolution of

CO from some types of active sites starts below the temperature of the main gasification

process, then it could influence the rate of reduction of NO. Only when C@ concentrations

get very high does the inhibition effect, due to active site blockage, overtake the CO

reaction effect. These results show that the interpretation of fixed bed reactor experiments is

very difficult because the reaction products formed at the beginning of the bed act as a

reactants at the end of the bed.

9

5. Plans for the Upcoming Quarter

Experiments will be continued, using the packed bed reactor. The NO reduction by

CO on quartz surfaces and on carbonaceous materials will be studied further during the

next quarter, since these kinetics have not yet been fully explored. As noted above, the

issues with respect to C02 have also not yet been fully resolved, and will be the subject of

further work. Then attention will be turned to the influence of other added gas components

on NO reduction rate. These added components will include those normally expected to be

present in combustion products, in addition to CO and C02, including low levels of 02,

and moderate levels of H20. The influence of oxygen on the NO-carbon reaction will

receive particular attention. In addition, the mechanistic implications of the results obtained

to date will be examined.

Hopefully, the quantitative routines of TGA/FTIR system will be fixed and some

experiments will be performed with this system.

6. References 1. Johnson, G.M., Smith, M.Y., Mulcahy, M.., 17th Synp. ( Int.) on Comb,, Combustion Institute, 647

(1979). 2. desoete, G.G. 23rd Svmp. (Int.) on Combustion, The Combustion Institute, 1257 (1990). 3. Aho, M., Rantanen, J., and Linna, V., m, 69,957 (1990). 4. Aho, M. and Rantanen, J., M, 68, 586 (1989). 5. Houser, T., McCarville, IvL, and Zhuo-Ying, G., m, 67,642 (1988).

6. Wojtowicz, W, Pels, J,, Moulijn, J., Roc. 1991 Int. Con€. Coal Sc i*, p 452, Butterworth, Oxford,

1991. 7. Wojtowicz, M., Oude Lohuis, J., Tromp, P., Moulijn, J., Int. Cod. Fluid. Bed Comb,, 1013, 1991. 8. h a n d , L., Lecher, B., and Andersson, S., Energv and Fuels, 5,815 (1991). 9. Muzio, L. and Kramlich, J., GeoD hvs. Res, LetL , 15,1369 (1988).

10. Pereira, E, Beer, J., Gibbs, B. Hedley, A., 15th SvmD. Int.) on Co mb,, Combustion Institute, 1149 (1975). 11. Hahn, W. and Shadman, E, Comb. Sci. and Tech,, 30,89 (1983).

10

12. Hjalmarsson, A.M., NOx Control Technologies for Coal Co mbustion, EA Coal Res. Rept.

IEACR/24, 1990. 13. Inui, T., Otowa, T., and Takegami, Y., I BrEC Prod. Res. Dev, 21,56 (1982). 14. Stegenga, S., Mierop, A., devries, C., Kapteijn, F., and Moulijn, J., Poc. 19th Biennial Con f. oq

carbon, p. 74, The American Carbon Society, 1989. 15. Mochida, I., Ogaki, M., Fujitsu, H., Komatsubara, Y., and Ida, S., Ew!, 64,1054 (1985). 16. Richter, E., Kleinschmidt, R., Pilarczyk, E., Knoblauch, K., and Juntgen, H., Thermoch. Ac$, 85,

311 (1985). 17. Lai, C.-K. S., Peters, W. A. and Longwell, J. P., Energv & Fuels, 2,586(1988). 18. Il&in-G6mez, M., Linares-Solano, A, Salinas-Martinez , C., Calo, J.M., Energv and Fuels, 7,146

(1993). 19. Yamashita, H., Tomita, A.,Yamada, H., Kyotani, T., and Radovic, L, Enerpv and Fuels, 7,85 (1993). 20. Okuhara, T. and Tanaka, K., J. Chem. Soc., Faradav Tms.-l, 82,3657 (1986). 21. Taylor, R.W. and Moms, CJ., ACS Div. Fuel Chem. Prept, 29(3), 277 (1984).

22. Personal Communication, Prof. I. Opik, May, 1992. 23. Ergun, S., J. Phys. Chem, 60,460, (1956). 24. Walker, PL. Jr., Pentz, L., Biederman, D.L. and Vastola, FJ., Carbon, 15,165, (1977). 25. Hiittingen, K., and Nill, J.S., Carbon, 28,4,457, (1990). 26. Meijer, R., Ph.D. Thesis, University of Amsterdam, The Netherlands, (1992). 27. Zhang, L., Ph.D. Thesis, Brown University, USA, (1996). 28. Koenig, P.C., Squires, R.G., and Laurendeau, N.M., Carbon, 23,5,531, (1985). 29. Koenig, P.C., Squires, R.G., and Laurendeau, N.M., E&& 65,412, (1986).

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government, Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsi- bility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark.

I manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom- mendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

1 1

1

0.5

0

-0.5

-1

-1.5

-2

-2.5

\ \

1

0.8 0.85 0.9 0.95 1 lOOO/T E-']

1.05 1.1

Figure 1. The effect of CO2 on the kinetics of the NO-graphite reaction.

0.22

0.2

0.1 8

0.1 6

0.1 4

0.1 2

0.1

Graphite T= 800°C

[NO] = 80ppm 0

0 1 2 3 4 5

Figure 2. The effect of C02 partial pressure on the NO reduction by graphite. All tests were performed at 800°C in a fixed bed reactor.

4

3

2

1

0

-1

I " " I " " l " " l " " l " " l " " I " " - Wyodak+NO _+ Wyodak+N0/2%C02 e - Wyodak+N0/99.99%C02 -

- -

i , * , , l , , , , l , , , , l , I , , ! , , , , l , , , , l , , , ,

0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 lOOO/T K-l]

Figure 3. The effect of C02 on the kinetics of the NO-Wyodak coal char reaction.

1.1

1

0.9

0.8

0.7

0.6

0.5

I ' 1- - - NO Reaction

Gasification by C02 followed by NO Reaction

\ - N0/2% C 0 2 Reaction - - - - N0/99.9!9% C 0 2 Reaction

- I - \

- I - I \ - I -

\ - i - I - 1 - 1 - 1 - 1 - 1

I I 1 I

- -- \ - -- - 4 - - I

\ - 1

I I I I

-

- _ _ _ _ _ - - - - - - - - - - \ 0

- /

0 \ \ / c -

0.4 I , , , i , , , l , , , l , , , l , , , , ,

0 20 40 60 80 100 120 140 Time [min]

Figure 4. The effect of C02 partial pressure on the transient behaviour of NO reduction by Wyodak coal char. All tests were pezformed at 800°C in a fixed bed reactor.

n E 6 e3

*

d

2

1.5

1

0.5

0

-0.5

-1

-1 .5

-2 0.85 0.9 0.95 1 1 .os 1.1

lOOO/T [K-lJ

Figure 5. The effect of C02 on the kinetics of the NO-resin char reaction.

5

4.5

4

3.5

3

2.5

2

a a

a

1.5

1 1 10 100

p c02 WaI 1000

Figure 6. The effect of C02 partial pressure on the NO reduction by resin char. All tests were performed at 800°C in a fixed bed reactor.

1.1

1

0.9

0.8

0.7

0.6

0.5

0.4 0

Pcoz= 2 kPa

200 400 600 Temperature ["C]

800 1000

Figure 7. Fractional remaining mass as a function of temperature during the gasification with C02. TPR experiments were performed with linear heating rate of 10 deg/min to a maximum temperature of 1OOO"C.

r

5 cs W

1000 I

800

600

400

-

-

-

200 -

0 -

Wyodak coal char Graphite

- Resin char 1 - - - - _ - -

[CO ] = 2% 2 0

I

I : I -

I

I

I

f -

200 400 600 Temperature ['C]

800 1000

Figure 8. TPR CO evolution profiles for different samples. TPR tests were performed with linear heating rate of 10 deg/min to a maximum temperature of 1000°C.