Thermodynamic Interaction between Transition Metals and Simulated

5 Thermodynamic Interaction between Transition Metals and Simulated Auto Downloaded by UNIV OF MASSACHUSETTS AMHERST on May 29, 2018 | https://pubs.acs.org Publication Date: August 1, 1975 | doi: 10.1021/ba-1975-0143.ch005

Exhaust HOWARD D. SIMPSON Union Oil Co. of California, Union Research Center, Brea, Calif. 92621

A computer study of the equilibrium concentrations of the major constituents in auto exhaust, and of the equilibrium relations of various transition metals with these constituents, provides some interesting insights into the state of chemical combination expected for metals used as NO catalysts. The state of chemical combination of most metals is not affected appreciably by changes in the air/fuel ratio in the normal range. The sulfates of many metals can form in auto exhaust under nearly all normal conditions. Sulfides and oxides can form in much of the normal range. Thus, sulfating, sulfiding, and oxidation can all contribute significantly to catalyst deactivation. x

' T ' h e elimination of N O * from automobile exhaust is a major objective of industrial scientific research. One method of eliminating this pollutant is catalysis. The fundamental purpose of using a catalyst is to cause slow chemical reactions to reach equilibrium rapidly. The effectiveness of a catalyst can be assessed by measuring the extent to which equilibrium has been approached in the system i n which it is being used. In studies connected with any catalytic process, it is therefore extremely useful to determine in advance, if possible, the equilibrium concentrations of the various reactants and products. In catalyst design and development, it is also very useful to have an idea about the states of chemical combination expected for the candidate catalyst metals. Thus, knowledge about the equilibrium state of metals as well as about the concentrations of the appropriate substances in the system of interest is necessary. These objectives were pursued in the Union O i l Co.'s NO^ catalyst program. 39 McEvoy; Catalysts for the Control of Automotive Pollutants Advances in Chemistry; American Chemical Society: Washington, DC, 1975.

40

C A T A L Y S T S

F O RT H E C O N T R O L

1500 ι ι

O F A U T O M O T I V E

1000 ι—ι—ι—Γ

P O L L U T A N T S

500 ° F

τ

4 2

Downloaded by UNIV OF MASSACHUSETTS AMHERST on May 29, 2018 | https://pubs.acs.org Publication Date: August 1, 1975 | doi: 10.1021/ba-1975-0143.ch005

Ο -2

CM Λ

*

^ c

-6

-10|

Cp(RuS ) =1 2

Cp(RuS ) 2

-14

12

10

(«Κ" )

4

Figure

1.

Comparative for RuS

J07T

2

+

20

18

16

14

10 /T

1

values 2H 2

of i w [ P /P Ru + 2 H 2 S ff|8g

->

ffjB

]

vs.

D a t a /rora "Handbook of Chemistry and Physics'' (1) a n d /rom Goldberg and Hepler (2)

Calculation of Equilibrium Constants This work is based on equilibrium constants for four categories of chemical reactions: (a) Reactions expected to occur over a N O * catalyst between some of the constituents i n synthetic automobile exhaust gas, ( b ) Reactions between transition metal sulfides and hydrogen, (c) Reactions between transition metal oxides and hydrogen, and ( d ) The formation of the sulfates of lead, copper, and nickel from the metals and metal sulfides. Almost all the equilibrium constants were calculated by our program based on the equation

McEvoy; Catalysts for the Control of Automotive Pollutants Advances in Chemistry; American Chemical Society: Washington, DC, 1975.

5.

\ηΚτ

ΔΗο RT

AF RT

=

41

Thermodynamic Interactions

S I M P S O N

T

, A(na)\nT R

+

A(nc)7» QR

A(nb)T 2R

,

r

(1)

where K is the equilibrium constant at absolute temperature T; A F is the free energy at that temperature; Δ ( η α ) , Δ ( η 6 ) , and Δ (ne) are the molar summations of the heat capacity equation coefficients for the prod ucts and reactants involved; ΔΗ and C are integration constants which permit the equation to be based on any desired set of reference ΔΗ and AF values; and R is the gas constant. Data were obtained as follows: the reference ΔΗ and ΔΡ values for gaseous reactions and for most of the reactions involving base metals ( 1 ),


T

T

0

1500

1000

500° F T

40

36

32

r-^

CM

28

X 24 20 C p ( R u 0 ) = 14.4 2

16

12

6

8

10

12

14

16

18

20

10 /T < ° K " ) 4

Figure

2.

Comparative

1

values

I 0 V T for Ru0

2

+ 2H

2

of

ΪΛ[Ρ^ /Ρ

-+ Ru +

Data from "Handbook of Chemistry and Physics" Goldberg and Hepler (2)

]

vs.

(1) and

from

0

Η ] 8

2H 0 2


42

CATALYSTS

F O R

O F

A U T O M O T I V E

P O L L U T A N T S

for reactions involving noble metals ( I , 2 ) ; and heat capacity data (3, 4,5). The accuracy of the calculated equilibrium constants for gaseous reactions was good because of the high accuracy of experimentally deter mined thermodynamic data for gases. As a result, the values obtained for reactions such as N O - » % N + 0 , N O + C O -> V2 N + C 0 , and C O + H 0 - » C 0 + H agreed well with the values reported else where (6, 7). Equilibrium constants obtained for reactions involving metals were much less accurate because the experimental data on systems involving solids vary greatly. It is difficult to obtain good thermodynamic data on solids because of nonstoichiometric compositions, surface phases, and partial phase transformations. For example, Nunez et al. (8) found that the heat of formation of C u O depended on mode of preparation, state of subdivision, and previous heat treatment. Annealed, granular C u O was more stable by about 1 kcal/mole than finely divided powder. The effect of these variations on the partial pressure ratios derived from equilibrium constants can be realized by considering the interaction between hydrogen and R u S or R u 0 (Figures 1 and 2, respectively). Heat capacity data for these substances were not found so they were estimated by Kopp's rule (9). The effect of any error incurred in esti mating the heat capacities is reflected by the difference between the curve obtained by using the true estimated (intermediate) value of Cp and those obtained by using half and twice that value (see the sets labeled "Goldberg & Hepler" in Figures 1 and 2). The other type of uncertainty illustrated in the figures is that due to differences in the Δ # and A F values from different sources (cf. the intermediate curve in the Goldberg set and the Handbook curve. Both sources of uncertainty are substantial, but those due to differences in the ΔΗ and ΔΡ values are probably more serious. Uncertainties in the natural logarithm of the partial pressure ratio attributable to these differences are about 4-10 2

2


T H E C O N T R O L

2

2

2

2

2

2

Table I. Simulated Auto Exhaust Used in Union Oil NO# Catalyst Evaluation Unit Content, mole %

Component H o H 0 CO C0 C H e (or C H ) NO N 2

2

2

2

3

2

3

8

0.33 0.35 10.00 2.00 13.00 0.10 0.08 74.14


2

5.

Thermodynamic

S I M P S O N

43

Interactions

Table II. Sets of Independent Chemical Reactions Assumed to Occur Among the Various Components of the Simulated Exhaust Gas in Table I Model

Chemical Reactions

A*

C O + H 0 -> C 0

2

Β·

H + I/2O2

2

2

2


2

+ H

2

H 0

2

C O + H 0 -> C 0

Ρ

+ H

2

H + I/2O2 -> H 0 2

2

C O + H 0 -> C 0 + H C H + 30 -* 3C0 + 3H 0 2

3

I-A

6

2

2

2

2

same as I except 3rd reaction is replaced b y : C H + 9/2Ο2 -> 3 C 0 + 3 H 0

6

3

II

6

2

2

same as I with addition of : N O + C O -> 1/2N + C 0 2

III

same as I with addition of: 2 N 0 + 2 H -> N + 2 H 0 2

IV

2

2

same as I with addition of: 2N0 + 5 H -» 2 N H + 2H 0 2

V

2

3

2

same as I with addition of: 2 N H -> N + 3 H

e

3

2

2

C H , N O , Ν2, a n d O 2 were treated as inerts i n M o d e l A , a n d a l l except O 2 were treated as inerts i n M o d e l B . N O a n d N were treated as inerts i n these runs. 5 0 % of the N O (0.04 mole %) was replaced b y N a n d 5 0 % b y N H i n the i n i t i a l gas composition for this r u n to simulate the operation of a d u a l - f u n c t i o n catalyst. β

3

6

6

2

c

2

3

units. This represents a factor of about 10 -10 in the partial pressure ratio. The data from Goldberg and Hepler (2) indicate that R u S and R u 0 are more stable than would be surmised from the Handbook data ( J ) . Although the uncertainty in dealing with the thermodynamics of noble metals is enormous, the data can still be of value in speculating about the state of chemical combination of these metals in simulated auto exhaust (see below). The Goldberg and Helpler data on noble metals were used throughout this work because it was felt that these were the most dependable values available. Compounds for which Cp values were estimated by Kopp's rule were IrS , PdS, P t O , R u 0 , and R u S . Sources of other data were as follows: equilibrium data for the interaction of N i O with hydrogen (10), data for the interaction of M o S and PbS with hydrogen (4), and the AH for PdS (2). N o value of A F for PdS was available; consequently, it was assumed to equal A f / — (ΔΗ — A F ) . 2

5

2

2

2

2

2

2

P d S

P t s


44

CATALYSTS


500 —ι

700 1

1

500

Initial

3.

1

1100 r—

I

700

1300 1

1

I

1500 ι

1

I

900

I

1100

ι

700 1

Calculated

1

1100

900 1

1

700

1

1

1300 1

900

1

1500 1

1

1100

P O L L U T A N T S

1700 °F ι

°K

CO: Ο, 1 mole % (A/F ratio = 14.2); Δ, 2 mole % (A/F and • , 3 mole % (A/F = 13.7)

500

4.

1


Calculated equilibrium concentrations of major in simulated auto exhaust for Model I

500 —ι

Figure

900

1

_ l

Figure

F O R T H E C O N T R O L

constituents =

13.9);

1700 °F 1

1

°K

equilibrium concentrations of major lated auto exhaust for Model A

constituents

Initial CO level, 2 mole %


in

simu-

5.

Thermodynamic

SIMPSON

45

Interactions

Equilibrium Composition of Simulated Auto Exhaust The normal composition of the simulated exhaust fed to Union Oil's bench scale NO catalyst evaluation unit is listed i n Table I. This com position simulates the auto exhaust obtained with an air/fuel ( A / F ) ratio of 13.9 (11). A t times, catalysts were also evaluated with C O = 1.00 and 3.00 mole % with A / F = 14.2 and 13.7 respectively. This range Downloaded by UNIV OF MASSACHUSETTS AMHERST on May 29, 2018 | https://pubs.acs.org Publication Date: August 1, 1975 | doi: 10.1021/ba-1975-0143.ch005

x

Λ

Λ

500

2. Οι—ι

700

1

1

900

1100

1

1——τ

1

1300

1

1

1500

1

1

1700 ° F

1

1

-he

500

Figure major

700

900

1100 ° Κ

5. Calculated equilibrium concentrations constituents in simulated auto exhaust for

of Model

the Β

Initial CO level, 2 mole %

of C O levels represents the range of A / F ratios encountered i n normal automobile operation. Equilibrium compositions of the simulated exhaust were calculated as a function of temperature for several assumed reaction models. The calculations consisted of solving simultaneously, by the Newton iteration procedure, a set of material balance equations and chemical equilibrium



46

C A T A L Y S T S



P O L L U T A N T S

constraints for the temperature and reactions involved. T h e reaction models considered are summarized i n Table II. A l l the models i n which the reduction of N O was considered pre dicted nearly total elimination of this constituent at equilibrium at a l l temperatures of interest. M o d e l V predicted that at least 9 5 % of any ammonia formed would be decomposed at equilibrium. M o d e l I repre sents fairly well the behavior of the major constituents i n simulated auto exhaust under the influence of an active catalyst. This model is depicted graphically i n Figure 3, for a l l three initial C O levels. T h e reactive hydrocarbon propylene disappeared between 427° and 538°C ( 8 0 0 ° 1 0 0 0 ° F ) . The predicted equilibrium hydrogen concentration at 667°C ( 1 2 5 0 ° F ) with 1 mole % C O initially i n the feed gas was 0.87% ; a value of 0.94% was measured b y mass spectrometry. Most surprising about

6

8

10

12

IOVT

14

16

18

20

(°K" ) 1

Figure 6. Superposition of system ln{? /V ^\ on base metal equilibrium desulfiding curves for exhaust gas containing 20 ppm HS Ht8

H

curves simulated

2

CO:

Ο, 1 mole % (A/F ratio = 14.2); Δ, 2 mole % (A/F 13.9); and • , 3 mole % (A/F = 13.7)

=



5.

S I M P S O N

Thermodynamic

I 6

1

I

8

I

I

10

I

I

12 10 /T 4

Figure noble

47

Interactions

I

I

I

14

I

I

16

I

18

I

I

20

2


r


50

CATALYSTS

πI 6

I

I

8

I


I

10

I

I

I

12 10 T 4

Figure CO:

I


1

1

14 («Κ" )

1

16

1

18

P O L L U T A N T S

1

1

20

1

9. Superposition of system In [Ί? /T?H ] c on noble metal equilibrium deoxiding curves ΗΖ0

2

Ο, 1 mole % (A/F ratio = 14.2); Δ, 2 mole % = 13.9); and •, 3 mole % (A/F = 13.7)

u r v e s

(A/F

curve than on the metal equilibrium curve, the metal w i l l tend to sulfide. A t temperatures where the reverse is true, the metal w i l l tend to desulfide. Such a scheme for base metals is presented in Figure 6. A n analogous system for noble metals is depicted i n Figure 7. These plots demonstrate that, i n general, metals w i l l tend to sulfide at low temperatures and to desulfide at high temperatures. Also, noble metals sulfide nearly as readily as base metals. W i t h the exception of copper, iron, and molyb denum, the C O level ( A / F ratio) had essentially no effect on the de sulfiding temperature. The desulfiding temperatures are summarized i n Table III. Analogously, deoxiding charts can be prepared b y superimposing the system l n [ P o / P H ] function on equilibrium l n [ P o / P n ] curves for various metals. Figure 8 depicts such a chart for base metals, Figure 9 H2

2

H2

2


5.

SIMPSON

Thermodynamic

51

Interactions


Table IV. Predicted Deoxiding Temperatures for Various Metals in Simulated Auto Exhaust Metal

Deoxiding Temperature, °C

Co Cu Fe Mn Ni Zn all noble metals

371 427 and

Thermodynamic Interaction between Transition Metals and Simulated

Recommend Documents