《Questioning Fundamental Principles of Organic Chemistry》

by Zhongheng Yu (Zhong-Heng Yu)

专著已经在Amazon出版，也上传到Figshare，researchGate，Academia等平台。最近，引起google AI overview 和Google AI Model的密切关注，几乎每天更新对我专著的评论，认为我的专著重新定义了结构理论，对推动有机结构理论的发展，做出创新新的贡献. 根据ChatGPT的建议，在我的博客，发表我专著的前言

PREFACE

For the structural theory established in the early development of organic chemistry, due to the limitation of understanding in that era, the fundamental principles cannot always be the truth. When the development of organic chemistry reaches a certain level, some chemists will need to use new tools (such as high-performance computers) and new research methods to review the fundamental principles and generally accepted theories. 

Fundamental Principles and Basic Causality. Conjugation effect, steric effect and inductive effect are the three basic structural effects in the structural theory of organic chemistry. The conjugation stabilization and the steric hindrance destabilization, as two fundamental principles, have long been used to understand and interpret the relationship between the molecular structure and its properties and to explore and explain the reaction mechanisms. For a conjugated molecule, the conjugation effect and the steric hindrance make the co-planar conformation the most stable and most unstable, respectively. Molecular conformation is the result of a compromise between steric hindrance and conjugation effect, which is the basic causality among molecular conformation, steric hindrance and conjugation effect.

A reasonable and reliable method of localizing geometry (including the localization of the p-molecular orbitals (MOs) and p-electrons) is the prerequisites for the study of the conjugation effect and aromaticity (including the calculation of the conjugation energy, aromatic stabilization energy and anti-aromatic destabilization energy). Our new methods include our 1998 method and its improved version (our 2006 method), and our 2007 method and its improved versions (our 2011 method and our 2014 method). Their establishment and development run through the entire monograph.

Our Questioning Convincing.  To convince the reader that our questioning is reasonable, as the first effort, our questioning began with the x-ray crystal structures and with the theoretical results calculated by Gaussian 98 software package. These experimental and calculation results can be verified in any laboratory.

In the literature, so-called experimental evidence of the conjugation stabilization refers to the difference (-3.9 kcal/mol) in the hydrogenation heat between butadiene and two butene-1 (In this book, the values of the stabilizing and destabilizing energies are expressed as negative and positive values, respectively). If the so-called experimental evidence is not rejected, our entire monograph will be worthless. Therefore, as the second effort (Section 6.01), using the same set of hydrogenation heats, we have emphasized that when trans-2-butene, instead of 1-butene, is used as reference molecule, the conjugation energy of butadiene is 1.9 kcal/mol and is destabilizing. The literature reported that, in the hydrogenation reaction of butadiene, the main intermediate (60%) is tras-2-butene. The so-called experimental evidence is the result of artificial selection of experimental data under the guidance of the subjective wishes. In the literature, the existence of conjugation effect in butadiene and the role of the conjugation in determining the distance of CC single bond were once controversial.

Benzene and cyclobutadiene are typical aromatic and anti-aromatic molecules, respectively, and their anti-aromatic energy (55 kcal/mol) and aromatic stabilization energy (-36 kcal/mol) were experimentally determined. The calculations of these two energies have been the hot topics in the literature. Conjugation energy of cyclobutadiene must be destabilizing. Otherwise, the calculation method is definitely not reasonable. Using our methods, as the third effort, the aromatic and anti-aromatic energies for these two typical molecules are, respectively, -36.4 and 54.9 kcal/mol, and are equal to their corresponding experimental values.

 

Cause of Questioning.  As a part of my PhD dissertation, ten compounds, Ar-N(Me)-(CH=CH)_n-CHO (Ar = phenyl, 2-naphthyl, 2-anthryl; n = 0, 1, 2, 3) were synthesized, and their UV absorption frequencies were measured. Using the homologous linear rule and according to the frequency and shape of UV absorption peaks, the following conclusions can be inferred: for (2-naphthyl)-N(Me)-(CH=CH)_n-CHO, when n = 1, the conjugation between nitrogen electron lone pair and naphthyl group is stronger than when n = 2.  According to the principle of the conjugation stabilization, when n =1, the dihedral angle q between the methylamino group (-MeN-) and the 2-naphthyl group should be smaller than when n = 2.  Out of the expectation, the size order of the experimental twist angle q is: 33.6^o (n = 1) > 31.4^o (n = 2). This abnormal size order is further confirmed by the theoretical dihedral angles q. For a series of molecules (2-naphthyl)-NH-(CH=CH)_n-CHO (n = 1, 2, 3), (a N-methyl group in each of the above molecules is replaced with a hydrogen atom in order to reduce the effect of the steric hindrance on the angle q), the size order of the theoretical twist angles q is as follow: 28.0^o (n = 0) > 9.6^o (n = 1) > 7.3^o (n = 2) (Chapter 1). The conjugation between the nitrogen lone pair and the 2-naphthyl group appears to be destabilization and to be a driving force for distorting molecule.

For N-benzylideneaniline (NBA, Ph-N=CH-Ph) and NBA-like species (Ar-NH=CH-Ar’) (Chapter 2), as shown by the X-ray crystal structures, the twist angle q between Ar- and -N=CH-Ar’ is abnormally large.  For NBA, for example, the experimental twist angle q is in the region of 36^o to 55^o. In the literature, it is attributed to the steric hindrance between the hydrogen atom of the -N=CH-Ar’ group and the ortho-hydrogen atom of the Ar- group, and to the p-p conjugation between the Ar- group and the nitrogen lone pair. The literature opinions were experimentally and theoretically questioned by us (Chapter 2). Our experiments include the following works: synthesizing eight NBA-like species with five- or six-membered heterocycle rings and four substituted N,2,2-triphenyl-ketenimines, and determining their crystal structures.

For each NBA-like species, the functions, E_e(q) = f(q), E_N(q) = f(q) and E(q) = f(q) (E(q) = E_e(q) + E_N(q), and E_N(q) < |E_e(q)|), describe the changes of the total electronic energy E_e and nuclear repulsion E_N with the increasing of twist angle q.  A relaxed-PES (potential energy surface) scan incrementing twist angle q shows the following interesting results: dE_N(q)/dq > 0, dE_e(q)/dq < 0, dE(q)/dq < 0, when 0 < q < q_exp; d²E_N(q)/dq² < 0, d²E_e(q)/dq² > 0 and d²E(q)/dq² > 0 for all twist angle q.  When the angle q is close to the angle q_exp of the crystal structure, dE_N(q)/d(q) = 0 (the total nuclear repulsion, as well as the nuclear repulsion between Ar- and -N=CH-Ar’, is maximized), dE_e(q)/d(q) = 0 (total electronic energy is minimized, and is the most stabilizing), and dE(q)/d(q) = 0 (molecular energy is minimized, and is the most stabilizing). A crowded conformation with a large twist angle is the most stable.

These PES-scan results inevitably lead us to ask the following question: for conjugation and steric hindrance, which effect is the driving force for distorting the molecule from its planar conformation. Due to the presence of the p-s interaction that may be stabilization or possibly destabilization, it is impossible to determine whether certain types of MO interactions (such as π-π and σ-σ interactions) are destabilization or stabilization, only based on the results of the relaxed PES-scans.

Exploring the structural factors that cause the NBA to be substantially distorted should start with the establishment of a new p-s energy decomposition method.

 

Our 2006 Method. As the first feature of our 2006 method (Chapter 3), a multi-step procedure provides a LFMO (absolutely localized fragment molecular orbitals) basis set for a non-planar conjugated molecule such as NBA-like species. In this basis set, p and s LFMOs are completely separated; the LFMOs are absolutely localized on their respective fragments, and each LFMO has the correct electronic occupancy. Then, the conditional single-point calculations, over the LFMO basis set, give a conjugated molecule the four localized electronic states. Based on the difference in the molecular energy between each pair of the localized electronic states, the following interesting results are obtained (Chapter 4 and Chapter 5): (i) p-p and s-s interactions between the fragments Ar- and -N=CH-Ar’ are always destabilization, and they are the driving forces for distorting NBA-like species from its planar conformation. (ii) p-s interaction is also destabilization, and it is resistance to molecular distortion. The conformation with a large twist angle, as well as the maximization of the nuclear repulsion between the Ar- and -N=CH-Ar’ groups, is the result of a compromise between the above three destabilization MO interactions.

The p-s energy decomposition leads us to believe that the fundamental principles and the basic causality are questionable.

 

Our 2007 Method and Its Improved Versions.  In the field of aromaticity, purely theoretical calculation of aromatic stabilization energy has become the goal pursued by theoretical chemists. Our method has achieved this goal (Chapter 6). By conditionally deleting non-diagonal elements of the AO Fock and overlap integral matrices, our 2007 method can provide a full localized GL geometry and a series of the locally delocalized GE-m geometries (m = 1, ……, k). In the optimized GL geometry, all the p MOs are absolutely localized on their respective double bonds. Each optimized GE-geometry arises from the local p interaction between a specific pair of the double bonds in the localized GL geometry. Based on the localized GL geometry and a series of the locally delocalized GE-m geometries, as the biggest feature of our 2007 method, it can provide following two types of p-electron delocalization energies for every conjugated molecule: DE^A = [E(G) – E(GL)] is the molecular difference between the ground state (G) and its localized GL geometry; DE^Am = E(GE-m) – E(GL), is the molecular energy difference between a specific GE-m geometry and the GL geometries, and always (or mostly) DE^Am > 0 (destabilizing).

The greatest success of our 2007 method is the discovery of a new type of additive energy effect, which makes it possible to provide an aromatic molecule with a virtual reference (VR) molecule that has the same configuration and conformation as the aromatic molecule itself.

For a non-aromatic molecule such as polyene and heteroatom-substituted polyene, DE^A > 0 (destabilizing), and  SDE^Am » DE^A = [E(G) – E(GL)], so that energy effects DE^Am are additive. This discovery ensures that our method can be used to accurately calculate the aromatic stabilization energy of all types of aromatic molecules, without needing the help of any empirical and semi-empirical parameters.

For a typical aromatic hydrocarbon molecule, always DE^A < 0 (stabilizing), SDE^Am > 0, |DE^A| > SDE^Am, and (DE^A - SDE^Am) < 0. Due to the additivity, SDE^Am can be considered as the molecular energy difference, [(E(VR) - E(GL)], between the virtual reference (molecule) and localized GL geometry of aromatic molecule itself. In this case, the energy effect difference, DE^A - SDE^Am = [E(G) - E(GL)] - [E(VR) - E(GL)] = E(G) - E(VR) < 0 (stabilizing). Therefore, the difference, DE^A - SDE^Am = E(G) - E(VR), can be defined as the extra stabilization energy (ESE) of an aromatic molecule with respect to its virtual reference molecule. For the virtual reference molecule of an aromatic molecule itself, the p-delocalization energy, DE^A(VR) = [E(VR) - E(GL)] = SDE^Am, can be calculated using the additivity of the energy effects DE^Am, without needing to know its detailed structure and molecular energy E(VR). Therefore, ESE = DE^A - SDE^Am = E(G) - E(VR) can also be considered as the difference, in the p-delocalization energy, between an aromatic molecule and its virtual reference molecule, ESE = DE^A - DE^A(VR). For benzene, for example, DE^A - SDE^Am = E(G) - E(VR) = -39 kcal/mol (using our 2007 method at B3LYP/6-31G* level). -39 kcal/mol can be defined as the ESE of benzene with respect to its virtual cyclohexatriene, and it is close to the experimental value of -36 kcal/mol.

For benzotricyclobutadiene (C₁₂H₆) at B3LYP/6-31G* level (Chapter 11), the central benzene ring, fused to three cyclobutadiene rings, is a real cyclohexatriene ring with alternating single (1.515 Å) and double (1.338 Å) bonds. For this central ring, interestingly, the destabilizing energy effects are as follows (kcal/mol): DE^A = 38.1, SDE^Am  = 40.4, and SDE^A » SDE^Am . In the case of cyclohexatriene, the energy effects DE^Am is indeed additive.

Our 2007 method and its improved versions have been used to reasonably calculate the ESEs of the following various types of aromatic molecules: five- and six-membered ring aromatic molecules, substituted benzenes, polycyclic benzenoid hydrocarbons (PBHs), [N]annulene, and strained aromatic molecules. In the meantime, the substituent effects, including the conjugation and inductive effects, are quantified. Among all the above conjugated molecules except for furan-like species, the ESE/p of the benzene is the largest.

There are multipole candidates for the GL geometry in the case of PBH. We propose the position rule, energy rule and GL sextet rule in order to choose the GL geometry from its candidates.

However, our 2007 method had to be limited to be used at the B3LYP level of theory (Chapter 8).  At the RHF and MP2 levels of theory, for example, the vertical and adiabatic delocalization energies of benzene, obtained from our 2007 method, become destabilizing (at B3LYP level of theory, they are stabilizing).

The improved “our 2007 method” is named "our 2011 method" (Chapter 8). In a localized geometry such as GL, obtained from our 2011 method, the two-electron exchange integrals, <lr|mw>, between the double bonds have been conditionally deleted, in addition to the Fork and overlap integral matrix elements.  For any (planar) conjugated molecule, the electron delocalization energies, including the ESE, vertical delocalization energy (VDE), and adiabatic resonance energy, can be accurately reasonably calculated using our 2011 method at the RHF, DFT and post-SCF levels of theory no matter whether the molecule is non-aromatic, aromatic or anti-aromatic. For benzene at B3LYP/6-31G*, the ESE value is improved from -39 kcal/mol (from our 2007 method) to -36.3 kcal/mol (from our 2011 method). The value of -36.3 kcal/mol is equal to the experimental value of -36 kcal/mol. It is particularly worth emphasizing that when using our 2011 method, the effect of the theoretical level and basis set size on the ESE value of benzene is much smaller than when using our 2007 method. For benzene at (RHF, MP2, LYP, B3LYP, SLATER, BLYP)/(6-31G*, 6-311G**, 6-311G(2d,2p), 6-311G(2df,p), 6-311++G(2df,p) levels, the ranges of the ESE (kcal/mol) are as follows: -34.5 < ESE (our 2014 method) < -39.3, and -31.7 < ESE(2011) < -37.4, and the average values are -35.0 (2011) and -37.0 (2014).

The last version of our method is called our 2014 method. In the GL geometry of benzene optimized using our 2014 method, for example, the MO interaction, the p-electron exchange interaction and the gradients of exchange integrals are all conditionally excluded from the between double bonds. Various energy effects obtained from our 2014 method are more reasonable than from our 2011 method.

 

p-Distortivity.  The destabilization feature of p-delocalization determines its performance of distorting molecular geometry.

For 621 pairs of the double bonds in the GL geometries of various types of the aromatic molecules, always DE^Am > 0, Dr > 0, and Dr can be fitted as a polynomial function of DE^Am. The local conjugation energy DE^Am between a pair of double bonds in the GL geometry is always destabilization, and the single bond distance between two interacting double bonds in the GE-m geometry is always longer than that of the corresponding single bond in the localized GL geometry.

In the literature including the standard textbook of organic chemistry, the CC bond length alternation in the central benzene ring of strained-aromatic molecule is attributed to the angle strain. In the particular localized geometry (PLG) of benzotricyclobutadiene, the p-interaction and p-exchange have been excluded from between the group A (central benzene ring) and the group B (including three annulated small rings). As a result, the difference, (Dr = r_endo - r_exo), between the endo- and exocyclic bond lengths decreases from Dr(G) = 0.177 Å in the ground state (G) geometry to Dr(PLG) = -0.002 Å in the PLG geometry. It is the p-delocalization (rather than angle strain) to distort the central benzene ring.

For 17 strained-aromatic molecules C₆X₃H₃ (X = B, Al, Ga, P, As) and C₆X₆H₆ (X = B, Al, Ga, N, P, As), C₆X₆ (X = N, P, As) and C₆X₆H₆ (X = C, Si, Ge), dDr can be fitted as the polynomial function of DE. The molecular energy difference DE is the p interaction energy between the central ring and the hetero groups (X-X, XH, and X=X), and -82.6 < DE = [E(G) – E(PLG)] kcal/mol < 81.7. Correspondingly, -0.040 < dDr = [Dr(G) - Dr (PLG)] Å < 0.180.  The bond length alternation of the central benzene ring should be attributed to p-delocalization.

For the localized GL geometry of benzene, at the moment of electron delocalization, as shown by the energy decomposition (Chapter 9), the CC double bonds are subjected to a pulling force, the double bond length change dr_a > 0. In the meantime, the CC single bonds are subjected to a compressive force, the single bond length change dr_b < 0. During the deformation of benzene ring from the localized GL geometry to the delocalized ground state geometry(D_6h) due to p-electron delocalization, the total electronic energy difference DE_e = E_e(G) - E_e(GL) > 0 (destabilizing), the nuclear repulsion difference DE_N = E_N(G) – E_N(GL) < 0 (stabilizing), and |DE_N| > DE_e. The nuclear repulsion change plays a predominant role in determining the D_6h geometry. In particular, when r_a = r_b. that is, when all CC bond lengths are equal to each other, the nuclear repulsion is minimized (d(E_N)/dr = 0), leading to the formation of D_6h geometry of benzene. But the role of the nuclear repulsion is the result of p-electron delocalization. In this case, energy criterion and geometric criterion of the aromaticity are well unified and are mutually causal. This is a new understanding of the essence of aromaticity.

 

Aromaticity of Furan-like Species. Due to Diels-Alder and 2,5-addition reactions of furan, furan's aromaticity is controversial.

For the furan-like species, the energy difference ESE = E(G) – E(EG) = DE^A - SDE^Am, and the size order of ESEs at B3LYP/6-31G* level is as follow (from our 2007 method): -53.1 (2-aza-pyrrole) > -49.4 (pyrrole) > -46.5 (imidazole) > -39.3 (furan) > -39.0 (benzene) = -39.0 (2-aza-furan) > -36.3 (oxazole). In this case, the ESE of benzene is not the largest.

According to the definition of extra stabilization energy, the basic characteristics of an aromatic molecule should be as follows: DE^A = E(G) - E(GL) < 0 (stabilizing), SDE^Am > 0 (destabilizing), and ESE = △E^A- SDE^Am < 0 (stabilizing). But, for furan-like species, DE^A > 0, SDE^Am > 0, and DE^A < SDE^Am, so that still (DE^A - SDE^Am) < 0 (stabilizing). Furan-like species are just a class of the molecule whose electron delocalization energy DE^A is less destabilizing than that of its virtual reference molecule. Such type of molecules should not be considered aromatic. There is no comparability between the ESEs of benzene- and furan-like species. In this case, the energy effect difference, (LDE = DE^A - SDE^Am), of the furan-like species is best defined as the less destabilization energy (LDE), rather than ESE.

 

About 4n+2 Rule and [N]annulene.  Whether the 4n+2 rule is always truth is a controversial topic. In the literature, the 4n+2 rule was once used as truth to test the rationality of the empirical and semi-empirical calculation methods.

For [N]annulenes, the theoretical level and basis set size have the influences on the size and sign of the energy effects such as DE^A and DE^Am, which is different from the typical aromatic molecules. Therefore, we are only concerned about the effect of the ring size on the aromaticity and anti-aromaticity, rather than the accurate calculation of their ESE value.

For [N]annulenes from N = 8 to 26, the VDE always conforms to the 4n+2 rule, without exception. When N = 16, the destabilizing VDE per p-electron (VDE/p) for [4n]annulene is quickly becomes constant (about 0.7 kcal/mol_*electron). The stabilizing VDE/p for [4n+2]annulene is slowly approaching the VDE/p (-0.7 kcal/mol_*electron) of [26]annulene.

The CESE for [4n+2]annulene and  [4n]annulene is stabilizing and destabilizing, respectively. In terms of electron delocalization energy, all [4n]- and [4n+2]annulenes can be regarded as close to the corresponding polyenes when N ³ 16.

 

Naked CC Single Bond of Butadiene. The experimental distance (1.454 to 1.468 Å) of the CC single bond of butadiene is shorter than that (1.54 Å) of ethane. In the standard textbook of organic chemistry, these two distances are used as the experimental evidences to support the conjugation stabilization. In the literature, however, which structural factor, conjugation or hybridization, determines the distance of the CC single bond of butadiene is controversary.

In the GL geometry of butadiene optimized using our 2014 method, the C(2)-C(3) single bond is called the naked single bond, and its distance r₂₃(GL) (1.451 Å) at B3LYP/6-31G* level is shorter by 0.006 Å than that r₂₃(G) (1.457 Å) in the ground state geometry. Correspondingly, the molecular energy difference DE = [E(G) - E(GL)] between the ground state geometry and the localized GL geometry is 1.4 kcal/mol (destabilizing), and it is close to the difference (1.9 kcal/mol) in the hydrogenation heat between trans-1,3-butadiene and two trans-2-butene.

For 66 molecules including butadiene-like species 18 X=C(2)-C(3)=Y (X, Y = O, S, Se, NH, PH, AsH) and 48 substituted butadienes X_nC=C(2)H-C(3)H=CY_m (n, m = 1, 2; X, Y = BH₂, AlH₂, GaH₂, NH₂, PH₂, AsH₂, F, Cl, Br), the distance difference DR₂₃ = [r₂₃(G) - r₂₃(GL)] can be fitted as the polynomial function of the molecular energy difference DE, where -0.046 Å < DR₂₃ < 0.040 Å, and -8.1 kcal/mol < DE < 13.0 kcal/mol. The conjugation plays an important role in determining the distance r₂₃(G) of the C(2)-C(3) single bond. In the case of the butadiene-like species and substituted butadienes, whether conjugation energy DE is stabilizing or destabilizing depends on the electronegativity of the heteroatom of double bond in the butadiene-like species and on the electron occupancy (p electron-sufficient and electron-deficient) of the substituent.

In the GL geometry of substituted butadiene, the conjugation between a CC double bond and its substituent(s) is called the adjacent conjugation (the conjugation adjacent to the C(2)-C(3) single bond). When the substituent is an electron-releasing group such as -XH₂ (X = N, P, As), the adjacent conjugation has a large influence on the distance r₂₃(GL) of the naked CC single bond through the s bond(s). But, when the substitute(s) is an electron-withdrawing group such -XH₂ (X = B, Al, Ga), the adjacent conjugation slightly influences the distance of the naked CC single bond.

 

Sincere Gratitude to My Graduate Students. My graduate students came from different universities, and they were all undergraduates majoring in organic chemistry. Few of these students had received any training in quantum chemistry. They worked hard and used their own wisdom. Especially worth mentioning is the talent and contribution of Dr. Peng Bao. In 1991, Mr. Bao was admitted to the Department of Chemistry, Tsinghua University. After graduation, he was admitted to the Graduate School of Lanzhou University, and his research interest was organic synthesis. Three years later, he obtained a master's degree. He then worked in the field of organic synthesis for many years. In 2004, Mr. Bao joined my research group as a PhD student.  Surprisingly, after only a few months of work, Mr. Bao can read and modify the source code of the Gaussian 98 and PC-GAMESS software packages.  It is through Mr. Bao's efforts that my calculation program can be incorporated into the PC-GAMESS program package, which is an important milestone for my research. Therefore, I would like to take this opportunity to express my sincere gratitude to my graduate students.

 

Acknowledgment. This monograph was supported by National Natural Science Foundation of China (Grants: Chemistry 85228, 2880091, 29272070, 29572074, 29872042, 20072041, 20272063 20472088 and 20672119). PC-GAMESS Source code was provided by Professor Alex A. Granovsky, Lomonosov State University, Moscow, Russia. MQAB-80 source code was obtained from Professor Muzhen Liao, Tsinghua University, Beijing.

 

转载本文请联系原作者获取授权，同时请注明本文来自虞忠衡科学网博客。
链接地址：https://wap.sciencenet.cn/blog-94786-1495278.html

上一篇：我专著的Google Overview AI
下一篇：从学术围攻到AI 共鸣：与ChatGPT之间迟来的对话