Dr. Alexey Ignatchenko Research Group

Cross-Selectivity
    We have done a theoretical analysis of kinetics of a bimolecular reaction between reagents A and B when both of them can have a dual role. Usually, this is not the case for reactions between a reducer and an oxidizer, nucleophile and electrophile, etc., for which the role of each reagent is well defined and the cross-selectivity is equal 100%. A general theory of the cross-selectivity in bimolecular reactions has yet to be developed.
Below, we will discuss criteria for having a high selectivity to the cross-product AB as opposed to the selectivity to symmetrical products AA and BB of a bimolecular reaction, A + B. The symmetry can be broken by polarizing roles of two components of the bimolecular reaction in opposite directions. If one reagent is a better electrophile, the other one must be a better nucleophile for the cross-selectivity to exceed the statistical binomial distribution.
Problem statement. In the cross-condensation of two acids A and B there are four products formed, BeAc, AeBc, AeAc, and BeBc by combinatorial paring of the enolic and carbonyl components (indicated by sub-indexes e and c). What are the products distribution based on the reactivity of the enolic and carbonyl components and the molar ratio of A to B? Specifically, we are interested in the condition of having the highest selectivity to the cross-condensation versus self-condensation of acids A and B.

Solution. We built a kinetic model based on the bimolecular reaction rate equations with dimensionless concentrations and length for making BeAc, AeBc, AeAc, and BeBc ketones.

                                                                                 (16a)

                                                                                 (16b)

                                                                                       (16c)

                                                                                       (16d)

The system of differential equations (16) with the boundary conditions,  and, for acid fractions B and A changing with the conversion, was solved numerically to obtain the product concentrations,  depending on coefficients , , , and . The integral selectivities, , were obtained by  normalization:

                                                                                                           (17)

We have mathematically proven, that the sum of the distribution functions and , which represent the cross-ketone, deviate positively from the sum of binomial probability functions , i.e.   in two cases, if

                                                                                                                   (18a)

or if

                                                                                                                   (18b)


Under all other conditions, a negative deviation was found, i.e. .

The significance. According to the stated criterion, the highest cross-selectivity will be observed when acids A and B are separated in their roles, i.e. one is preferentially serving as the enolic component, while the other is the preferred carbonyl component. In our experimental work we have identified catalysts which can discriminate two acids in their roles in the reaction mechanism.


The mathematical proof is provided below.
Statement. 
Distribution functions and , which represent the cross-ketone, deviate positively from the binomial probability functions and  in two cases, if

                                                                                                                               (A.1a)

or if

                                                                                                                               (A.1b)

Analysis and proof by algebra methods.

The probability functions for making AeAc, AeBc, BeAc, and BeBc ketones respectively, on the basis of the binomial statistical approach are

                                                                                                                     (A.2a)

                                                                                                                (A.2b)

                                                                                                                (A.2c)

                                                                                                                                (A.2d)

where  and , respectively.

Coefficients, , , , and , are assigned to transform functions  into .

                                                                                                    (A.3a)

                                                                                               (A.3b)

                                                                                               (A.3c)

                                                                                                               (A.3d)

Coefficients and represent acids A and B, respectively. Subindexes 1 and 2 represent the enolic and the carbonyl components, respectively. Coefficients e  , and  have the meaning of the relative activity of acid B vs. acid A as the enolic or as the carbonyl component, respectively.

The adjusted distribution functions  are obtained through  normalization and represent the differential selectivities to the four products.

                                                                                                                         (A.4)

The denominator in eq. (A.4) can be transformed into a shorter form:

=

=                              

                                                                   (A.5)

After dividing the numerator and the denominator by  each  function is transformed into a new form:

(A.6a)

(A.6b)

(A.6c)

(A.6d)

The relation between functions  and  is analyzed separately under four different terms. Results are summarized in Table A.1.


 

Table A.1. The relation between  and  functions under four terms.

Terms

AA

BB

AB

1)

 

 

 

2)

 

 

3)

 

 if 

 for any  

if  

 If

 

true for any  if  

 when

If  ,

 for any .

If , then  for any .

 

 

 

If

 

 

4)

 

Function , term 1) , or . In this case

is true, because if  then . Similarly, , therefore the denominator is less than 1, but the whole fraction is greater than 1.

i.e.  under term 1.

Function , term 2) , or .  Under this term, it can be proven than simply by changing the sign in all equations obtained under term 1) above, i.e., because , therefore, , and , therefore the denominator is greater than 1, but the whole fraction is less than 1.

Therefore, under term 2.

Function , term 3) , or ; and term 4) , or .

 After the next transformation, the criteria for a negative deviation of  from the binomial value is:

Or, it can be further transformed into the next form with the change of the inequality sign:

Because   , the next form can be obtained:

 ,

or, 

Further transformations provide the simplified criteria:

in the final form:

         (A.7)

where   .

For a positive deviation of  from the binomial value the sign of the inequality (A.7) is reversed:

(A.8)

Under the terms 3 and 4, function  can deviate from the binomial value in both a positive and negative way depending on  and parameters  and . Condition for a region of a negative deviation, independent of , is 

Criteria for a positive deviation,  independent of x can be found from the condition:

Which can be transformed into

Function , term 1) , or .

Similarly,

If each multiplier is less than one, then the product is also less than one, i.e. 

Function , term 2) , or .  Changing the sign in all equations under term 1 proves that  under term 2.

Function , term 3) , or ; and term 4) , or .

A quick solution for the negative deviation of from the binomial value can be obtained by switching acids A and B in equation A.7. Parameters  and  are replaced by their reciprocal values, and  is replaced with .

The last equation can be modified to:

To find when the negative deviation does not depend on , it is set to  and equation (A.9b) is transformed into  , or

                                                                                                (A.10)

Equation (A.10) is the criteria for having a negative deviation of , i.e. 

For a positive deviation, , regardless of , equation (A.9b) is solved for 

Functions , general remarks for a positive deviation. To preserve the unity of  a negative deviation of  automatically means a positive deviation of 

Condition for having positive deviation of functions  from the binomial value can be obtained by combining equations A.3b, A.3c, A.6b, and A.6c

which gives after transformation

(A.11)

A sufficient, but not necessary condition for the solution of eq. (A.11) is given when each multiplicand is greater than one: 

or 

or 

(A.12)

A more broad solution for eq. (A.11) depending on  is graphically illustrated in Fig. A.1.

 

a) b)

Fig. A.1. Graphical illustration for the positive deviation of the cross-selectivity,  from binomial functions for , represented by values above 1.0 in eq. (A.11) as a) a multi-colored 3D-graph (other than blue), and b) a projection in (e,c) coordinates (red colored area).

Functions , term 1) , or . Because , the first condition in (A.12), is not true.

Functions , term 2) , or . Because , the second condition in (A.12), is not satisfied.

Functions , terms 3) and 4). Under these terms eq. (A.11) can be satisfied by eq. (A.12), which is sufficient, but not necessary.

 

The behavior of functions  under terms 1-4 is illustrated by graphs in Figures A.2-A.5. Parameters  and  for graphs in Figures A.2 and A.3 are chosen so as to demonstrate the symmetry of graphs with respect to the choice of acid, A and B, i.e.,  and . Parameters under term 3 are those found for TK catalyst (Fig. A.4). An example of parameters e and c in Fig. A.5 demonstrates the case when cross-selectivity can change from positive to negative deviation depending on the composition of acids’ mixture.

Fig. A.2. Comparison of functions  to under term 1, e < 1, c < 1

 

 

Fig. A.3. Comparison of functions  to under term 2, e > 1, c > 1

 

Fig. A.4. Comparison of functions  to under term 3, e=0.14 < 1, c=3.92 > 1

 

 

Fig. A.5. Comparison of functions  to under term 4, e=1.9 > 1, c=0.9 < 1

 

Everything discussed above applies for the differential selectivities. In order to get the integral selectivities, a law for the reactants concentration change along the length of the flow reactor has to be known. However, for any specific ratio of the two acids changing along the reactor length, the specified criteria (Table A.1) are still valid for calculating the differential selectivities. Reactor model for TK catalyst (Fig. A.6) shows selectivity graphs for all products closely matching those theoretically calculated for the same parameters and  (Fig. A.4).

Fig. A.6. Integral selectivities for the reactor model with TK catalyst at temperature of 425 °C.


Conclusion.  In summary, it has been shown that when
     , then , but . On the other hand, if


   , then , but 

If both conditions are satisfied, then .

 

This allows and  to exceed binomial functions .

 

The combined condition,  requires either , or    

 

 

For references, please, cite this work as:

Ignatchenko, A.V.; Deraddo, J.S.; Marino, V.J.; Mercado, A. Cross-Selectivity in the catalytic ketonization of carboxylic acids. Applied Catal. A: General, 498, 10-242015.