vinylogous enolate

In the following compound, the pKb of C1 should be about 24, but it is NaN. If you submit the resonance structure with the charge on C3, the pKa is calculated correctly as 24. So why doesn't the plugin give an appropriate value for C1 in the structure below?

Code:

<?xml version="1.0" ?> <MDocument>   <MChemicalStruct>     <molecule molID="m1">       <atomArray           atomID="a1 a2 a3 a4 a5 a6 a7"           elementType="C C C C O O C"           formalCharge="-1 0 0 0 0 0 0"           x2="-3.994374990463257 -2.454374990463257 -0.9143749904632568 0.6256250095367433 1.3956250095367435 1.3956250095367435 2.9356250095367438"           y2="0.5293750166893005 0.5293750166893005 0.5293750166893005 0.5293750166893005 1.863054138517336 -0.804304105138735 -0.804304105138735"           />       <bondArray>         <bond atomRefs2="a1 a2" order="1" />         <bond atomRefs2="a2 a3" order="2" />         <bond atomRefs2="a3 a4" order="1" />         <bond atomRefs2="a4 a5" order="2" />         <bond atomRefs2="a4 a6" order="1" />         <bond atomRefs2="a6 a7" order="1" />       </bondArray>     </molecule>   </MChemicalStruct> </MDocument>

Hi,








Prior to the pKa calculation you need to take the canonical resonance form.


Improved resonance structure generator available in the latest release.








Jozsi

Jozsi wrote:

Prior to the pKa calculation you need to take the canonical resonance form. Improved resonance structure generator available in the latest release.

OK, so now, when I submit either ^–CH₂CH=O or CH₂=CHO^– to the resonance structrue calculator, and I tell it to give me the canonical form, it gives back ^–CH₂CH=O.





But in a previous post, you said that the "canonical" form was the best description of the compound. In the present case, the best description is CH₂=CHO^–. So either you've changed your definition of "canonical form", or you're no longer giving the canonical form.





Furthermore, I note that the pKb of the O atom of ^–CH₂CH=O is given as -3.5. This value is incorrect. The correct value is 12.0, as given when CH₂=CHO^– is submitted to the pKa calculator. You've essentially just swapped one problem for another.





You need to consider all resonance forms when calculating the pKas of the atoms in a compound. You can't consider just one.





Please see my contributed software for a suggestion on how to solve this problem.

Hi,








[O-]C=C pKa of this molecule 11.96, -4.13





I you take the most stable resonant form ,the so called canonic tautomeric form, then you get this structure





[CH2-]C=O pKa of this molecule 14.50, -6.87











Firstly I want to know whether you get these pKa values or not?




















Jozsi

Jozsi wrote:

[O-]C=C pKa of this molecule 11.96, -4.13 If you take the most stable resonant form ,the so called canonic tautomeric form, then you get this structure [CH2-]C=O pKa of this molecule 14.50, -6.87 Firstly I want to know whether you get these pKa values or not?

Yes, I get these values.





But [CH2-]C=O is not the most stable resonance form of this compound. The most stable resonance form has the negative charge on the more electronegative element, O.





Moreover, I am looking at pKas of individual atoms. For [CH2-]C=O, the O atom is said to have a basic pKa of -6.87. That is incorrect. Its basic pKa is 11.96. Only after the compound is protonated on C does the basic pKa of the O decrease to -6.87. But the whole point of the dynamic pKas is to give the pKas of the compound in the submitted state, not in another protonation state.

Hi,








Ratio of the keto and the enol form depends on the pKa.


pKa keto =14.5


pKa enol = 11.96





Since the keto form is the weaker acid , this is why keto form should be the major resonant form at equlibrium.





Do you agree with this?











Keto group pka -6.87 is really a bug, when carbon is deprotonated then keto group should be become more basic. I agree with this. I will fix.








Jozsi

Jozsi wrote:

Ratio of the keto and the enol form depends on the pKa. pKa keto =14.5 pKa enol = 11.96 Since the keto form is the weaker acid , this is why keto form should be the major resonant form at equlibrium. Do you agree with this?

No. You are confusing the equilibrium between the carbonyl and the enol (neutral species) with the resonance descriptions of the enolate.





Resonance forms are not in equilibrium, because they are not different chemical species. They are alternative descriptions of the same species. The O(-) resonance form is a better description of the ground state of the enolate than is the C(-) form, because the negative charge in the delocalized pi system of the enolate spends more time near the O than it does near the C. Another way of saying it is that the HOMO of the enolate has a larger coefficient at the O atom and a smaller one at the C atom.





It is true that the enolate is more basic at the C atom, but that basicity is due to the greatly lower energy of the carbonyl as opposed to the enol, not because there is a greater negative charge at that C atom in the enolate.

Quote:

Keto group pka -6.87 is really a bug, when carbon is deprotonated then keto group should be become more basic. I agree with this. I will fix.

Good. And conversely, the terminal C atom of CH2=CHO^– should (must!) show a basic pKa of 14.5.





BTW, the pKa calculator still fails to recognize a terminal alkyne as acidic. It really should. It also gives the basic pKa of HCCCH(-)CCH as 47, which I think is too large. Should be more in the mid-30's, I think.

Hi,








Of course, enolate and deprotonated (C-) keto has a common anionic (resonance) form, therefore , ratio of enol and keto form can be calculated.


According to this ratio the keto form is favored over the enol form.





I think this is a life like approxmation.








Jozsi

Jozsi wrote:

Of course, enolate and deprotonated (C-) keto has a common anionic (resonance) form, therefore , ratio of enol and keto form can be calculated. According to this ratio the keto form is favored over the enol form. I think this is a life like approxmation.

Let me try once more. The keto form cannot be favored over the enol form, because there is no keto form separate from an enol form. There is a single enolate, which has a partial negative charge on O and a partial negative charge on C. If you compute (by the Hückel method, for example) the relative charges of the C atom and the O atom, you will find that the partial negative charge on the O atom is of greater magnitude than the partial negative charge on the C atom. Hence, the resonance form with a negative O atom is a better description of the enolate than the resonance form with a negative C atom.





If you don't accept my explanation, I suggest you talk to another organic chemist.





Anyway, I don't really care what you call the canonical resonance form. All I care about is that when I submit either resonance form of the enolate, that you give me basic pKa's for both the C atom and the O atom.

Hi,





If I understand well you want the pKa values of the major resonant structures.








Example 1.


If I submit this molecule: 'CC([O-])=S'


then you want to see these to acidic pKa: 4.43 and 0.62


because this molecule has the next two characteristic resonance forms:


'CC()=O'


'CC([O-])=S'











Example 2.


If I submit this molecule 'CC(=O)\C=C\[CH2-]'


then you want to see the pKa values that are calculated for the next three resonant structures?





'CC(=O)\C=C\[CH2-]'


'CC(=O)[CH-]C=C'


'C\C([O-])=C\C=C'








Is my suspect good?








Jozsi

Yes, you are correct. Each atom has its basic pKa calculated in the resonance structure in which that atom has a negative charge. The basic pKa for the molecule as a whole is the largest of the calculated basic pKas among the individual resonance forms.

And, in a related issue, CC#[NH+] should have an acidic pKa reported for both the C atom and the N atom. Marvin 4.1.6 reports an acidic pKa only for the C atom.

Any progress on making Marvin recognize that C(1) in CH₂=C(O^–)CH₃ should have a basic pKa of around 20?





Also, I notice that for ^–CH₂C(=O)CH₃, the CH₃ group is given an acidic pKa of around 29. I doubt that it would be that acidic if the other side of the ketone is already deprotonated.





These calculations are done with the latest version of Marvin posted on your site.

Jozsi wrote:

Hi, If I understand well you want the pKa values of the major resonant structures. Example 1. If I submit this molecule: 'CC([O-])=S' then you want to see these to acidic pKa: 4.43 and 0.62 because this molecule has the next two characteristic resonance forms: 'CC()=O' 'CC([O-])=S' Example 2. If I submit this molecule 'CC(=O)\C=C\[CH2-]' then you want to see the pKa values that are calculated for the next three resonant structures? 'CC(=O)\C=C\[CH2-]' 'CC(=O)[CH-]C=C' 'C\C([O-])=C\C=C' Jozsi

Hi,





Is this output given in Marvin 4.1.14? Marvin 5.0.0?





-- Bob

Hi,








This output is not available neither in 4.1.14 or in 5.0.0 version.





Jozsi

I hope it will be available soon. I have a workaround, but it is costly computationally.





(If anyone needs the workaround, let me know.)

Hi,


I am also interested in enolate property calculations and so I did Hückel calculations using Marvin 5.0.0. The enolate was opened using the Smiles strings C=C[O-] (left) and [CH2-]C=O (right), see screenshot.


Doing Hückel calculations I get different results for the left hand formula and the right hand formula. This cannot be. The enolat is *one* mesomerism stabilized anion. The results for total charge distribution are also different.


The results for pi energy give 4.3183 beta for the left form and 5.4040 beta for the right form.


What's wrong ?





Hi Bob, I am interested in your workaround !





Regards, Hans-Ulrich

Hi,

Quote:

What's wrong ?

Do you mean that the calculation should not consider the major resonant contributors separately but it is better if I take the most favored resonant contributor and then calculate the Hückel charge for this single form OR what else would you propose?





Thanks.





Jozsi

The results should not, cannot depend on which resonance contributor one draws before doing the calculation.





You will most certainly not arrive at a valid result "if I take the most favored resonant contributor and then calculate the Hückel charge for this single form."





As for how to do the calculation, I believe that you can find the answer to that question in any undergraduate physical chemistry textbook.





By the way, the same stricture applies to acidity and basicity calculations. The results cannot correctly differ among resonance contributors.

Hi,








These are "nice" sentences. But I don't know how to need to combine the resonant structures so that I get one "result".


Do you have any idea about this? Or could you show a reference where such kind of details are given for the Hückel calculations.








Thanks.


Jozsi

Hi,


in "Streitwieser, Molecular Orbital Theory, Wiley" on page 25 there is the sentence: "The mathematical difficulties of the VB method preclude applications to molecules containing heteroatoms (other than carbon) in the pi-bond system". And on page 26 there is: "Neither resonance theory nor its parent VB theory is suited for general quantitativ or semiquantitative calculations and correlations."





I think the main problem is, to translate the different localized Smiles-Strings (and the different formulas corresponding to them) to *one* delocalized pi-System. Of course the problem concerns the pKa, logP etc. and the Hückel calculations.





For the two "resonant" benzene structures you see in the screenshot (one with the double bond between Cl and Br and the other not between) the translation in Marvin to the HMO-modell works well, as it has to be.


In the enolat and similar systems it should work as well.





The problem in the present implementation is, it uses different Hückel parameters for the two "resonant" structures, see the screenshot four posts before this one.





I have an idea how to solve it:


1) Count the number of atoms in the pi-system (natoms).


2) Count the number of electrons in the pi-system (nelectrons).





A) If the nelectrons is equal natoms, then use the standard Hückel-Parameter for one-pi-electron hetero atoms, f.e. Coulomb Integral hO. = 1 beta


(The little point after the O is the pi-electron).





B) If nelectrons is greater then natoms, then use the standard Hückel-Parameter for two-pi-electron hetero atoms, f.e. Coulomb Integral hO: = 2 beta and Resonance Integral kC-O = 0.8 beta.





All other pi-C-atoms have hC = 0 and kC=C = 1.0 beta.





The parameters are standard parameters table 3, Streitwieser.





As I mentioned in an earlier post, the Hückel-Parameters should be open for an OpenScience. http://www.chemaxon.com/forum/ftopic3478.html

The same principle also applies to cases where the number of electrons is fewer than the number of atoms, as in Me₂C=CH-CH₂⁺.





Programs that can properly handle the case of an enolate are abundant. For example, here.

Hi,








HUWagner wrote:

Quote:

I have an idea how to solve it: 1) Count the number of atoms in the pi-system (natoms). 2) Count the number of electrons in the pi-system (nelectrons). ...

I implemented this idea into the Hückel calculator. The new version will be availble in a couple of days.








Jozsi

Hi,


finding the results shown in the screenshot351, I think the problem with mesomeric systems is more general.


Please look at "What resonance is not" in http://en.wikibooks.org/wiki/Organic_Chemistry/Print_version . Resonance is not an "isomerisation process" as mentioned in the paper "A method for calculating of small and large molecules" (BTW, ions are not molecules). And it seems to me, that SMILES formulas cannot describe mesomeric systems other than aromatic systems. (I am not a SMILES expert. The formula is not important for me).


I tried a SMILES formula with small letters (as in the benzene example). It is interesting that MarvinSketch recognizes this as a mesomeric system, but the results for pKa and geometry calculation are wrong (Of course it should be a pKb calculation), see screenshot352.





MarvinSketch should do before any calculation the following:


1) Analyse the system for conjugated mesomeric systems. This could be done by determing the hybrisation af the C atoms as sp3, sp2 or sp. The simplest way is to count the neighbour atoms of this center (The system must of course be "prehydrogenized" as Marvin says it). The Dreiding force field uses different parameter sets for this three types and there is a special parameter set for benzene C-atoms ( J. Phys. Chem. 1990, 94, 8897-8909).


2) Then the HMO and other calculations for the systems are defined and could be done.





These procedures do only make sense for the "organic" atoms CNO and perhaps S, and as attached atoms H, Cl or Br.





I have programmed a software (in FORTRAN) to calculate the UV-Vis spectra of organic colorants. The input is the molecule in xyz cartesian coordinates (with hydrogens !) and charges are given for the total system extra. Then the program detects all pi-systems, it could be more than one. The "saturated" part of the molecule is omitted.


sp2-C and sp-C, aza-N and amino-N, keto-O and ether-O, thio-S and thioether-S, Cl and Br are detected this way. Other atoms are also omitted.


Then the program calculates the ground state and the excited singulet states using the Pariser-Parr-Pople pi-SCF method.





These methods could be used by Marvin also.





Best regards, Hans-Ulrich

HUWagner wrote:

the results for pKa and geometry calculation are wrong (Of course it should be a pKb calculation), see screenshot352.

By default, Marvin calculates pKa and pKb of a neutral compound, regardless of what you submit. For example, if you submit EtO(-), it will tell you the pKa is 17 and the pKb is 0 (or thereabouts). To get the pKa and pKb of what you actually submitted, choose the dynamic (rather than static) method for calculating these values.

Quote:

By default, Marvin calculates pKa and pKb of a neutral compound, regardless of what you submit. For example, if you submit EtO(-), it will tell you the pKa is 17 and the pKb is 0 (or thereabouts).

Of course, see screenshot354.





My point is not how to calculate pKa-values, my point is how to handle mesomeric "resonance" systems. But, perhaps, we should create a new topic.





And the dynamic method for calculating pKa gives for the species which should represent the enolat ion the same result, see screenshot353. The proton is abstracted at a wrong position.





Regards, Hans-Ulrich