TS on a SN2 reaction



 Dears CCL users:
 	Last week I post a question regarding to a TS on a SN2 reaction. I
 have received some responses and here is the summary:
 	My original question was:
 **********************************
         I need to calculate the TS for a variety of SN2 reactions using
 semiempirical Mopac calculation and I have some problems. I will be very
 greatfull for every help that you can gave me.
         The strategy that I have used is the following:
         I have define a dummy atom in order to fix the geometry of the
 nuclephile and the leaving group as in the following scheme:
         Du..........C
                     |
                     |
         Nu..........C---L
         I have defined the distance between the dummy atom (Du) and the upper
 carbon atom about 10 A and between the nucleophile (Nu) and the lower carbon
 atom about 5 A.
         The Du-C-C-Nu and Du-C-C-L dihedral angles have been defined as 0 and
 180 degrees, respectively.
         I used the symmetry keyword (option 17) in order to varies the C-C-Nu
 bond angle as 180 degrees - C-C-L bond angle during the calculation.
         I believe that with those definitions, the nucleophile, the carbon
 atom and the leaving group (L) must be in straight line when the distances
 Nu-C is decreased at 0.1 or 0.05 A intervale.
         I expect that when the nucleophile is located at a apropiate distance
 the distance C-L begin to increase so that the geometry of higher energy
 could be used to optimized the TS of the reaction, but this does not occur.
 When the nuclephile is about 2.4-2.2 A, the leaving group is inmediately
 ejected to a distance of 3.5 A or more.
         Can anybody with experience in this type of calculations gave me some
 orientations to resolve this problem?.
         Please, e-mail directly to me and I will summarized to the net.
         Sincerely
         A. Entrena
 **************************************
 	And the answers are the following:
 *********************************************
 From: ansu - at - trout.csb.ki.se (Ansuman Lahiri)
 Dear Dr. Entrena
 Did you try the SADDLE keyword in mopac? What I usually do is to
 choose two configurations (in one of which the leaving group is
 bonded and in the other the attacking group is bonded to the central
 group) and then run SADDLE. The approximate transition state obtained
 from SADDLE can then be further refined using TS.
 Hope this helps
 Good luck!
 Ansuman
 *****************************************
 From:  "Anatoli Korkin" <korkin - at - qtp.ufl.edu>
 Hi!
 Fixing C-Nu eagual C-L in optimization, you will approach TS closely.
 Then compute frequencies and use the resulting force field along with
 opt=ts option. But I have feeling, that you guy hardly undertand what
 are you doing and why.
 regards,
 Anatoli Korkin
 ********************************************************
 From:  Alan.Shusterman - at - directory.Reed.EDU (Alan Shusterman)
 --- You wrote:
 I expect that when the nucleophile is located at a apropiate distance
 the distance C-L begin to increase so that the geometry of higher energy
 could be used to optimized the TS of the reaction, but this does not occur.
 When the nuclephile is about 2.4-2.2 A, the leaving group is inmediately
 ejected to a distance of 3.5 A or more.
 --- end of quoted material ---
 I am not sure what Nu/L combinations you are using, but this may be
 reasonable
 behavior.
 First, if either of the reactants or products are charged, then electrostatic
 interactions (ion-dipole) will be very strong and a true transition state may
 not even exist.  Consider chloride attacking methyl chloride (degenerate
 rxn):
 as chloride approaches the energy falls (favorable ion-dipole interaction),
 then sharply rises to give a symmetric TS, then falls again to give the
 product
 ion-dipole complex.  The AM1 transition state is only a kcal/mol or two
 higher
 than the separated reactants.  Now consider hydroxide attacking methyl
 chloride: as hydroxide approaches the energy falls steadily without
 intervention of a transition state.  This reaction is very exothermic, and
 bond
 energy changes plus electrostatic interactions erase the transition state.
 To
 summarize: 1) you may have trouble locating transition states, 2) minima may
 correspond to ion-dipole complexes in which reactants approach to within LESS
 than the sum of the nonbonded radii, 3) TS, when they do exist, should be
 viewed as resonance hybrids of the reactants and products, and therefore
 reasonably short CN and CL bond distances are expected.
 Finally, least motion paths are interesting, but may not tell the full story.
 For example, hydroxide attack on methyl chloride initially gives an
 ion-dipole
 complex in which the chloride is close to the methyl hydrogens of methanol.
 This is probably not the global minimum, and may not even be a local minimum.
 The best structure is probably one in which a hydrogen bond, Cl..HOCH3, holds
 the ion and dipole together.
 Alan Shusterman
 Department of Chemistry
 Reed College
 3203 SE Woodstock Blvd
 Portland, OR  97202
 ************************************
 From:  Hens Borkent <borkent - at - caos.kun.nl>
 Dear Dr. Entrena,
 There is no reason why the leaving group should stay in
 the neighbourhood of the carbon atom; once you're over
 the 'top' (TS), the structure will minimize the C-L
 distance. And moreover, there is no reason why the
 N-C-L angle should be 180 degrees, unless your carbon
 skeleton is symmetrical (methyl, t-butyl).
 The savest thing would be to do a grid search, in which
 you vary the C-N distance in the range 2.0 2.8, and the
 C-L distance something similar, depending on the nature
 of the leaving group. In this way you fix both values
 and a saddle point should appear in the grid. This
 point should be optimized using the TS keyword.
 Have a look at:
 http://www.caos.kun.nl/tutorials/camm/mopac/gsm.html
 Sincerely,
 --
   *****    J.H. (Hens) Borkent, CAOS/CAMM Center,
  *CAOS *   P.O. Box 9010, 6500 GL Nijmegen, The Netherlands
 *   /   *  Tel 0031 24 36 52137  Fax 0031 24 36 52977
  * CAMM*   e-mail: borkent - at - caos.kun.nl
   *****    http://www.caos.kun.nl/staff/borkent.html
 ****************************************************
 From:    Jose Ignacio Garcia <jig - at - qorg.unizar.es>
 He leido tu pregunta en la CCL. Puede haber varios motivos para el comporta-
 miento que describes. En primer lugar, la forma de la superficie de energia
 potencial depente por supuesto de la reaccion, de forma que podria suceder
 que en la reaccion particular que estas estudiando, haya una zona en la que
 el sistema evolucione muy rapidamente. La solucion en este caso es estudiar
 mas en detalle dicha region, modificando la distancia de ataque Nu-C con
 intervalos mas cortos. Algo parecido sucede cuando se comparan las reacciones
 Cl(-) + CH3Cl --> ClCH3 + Cl(-) y
 F(-) + CH3Cl --> FCH3 + Cl(-)
 En el primer caso, se obtiene un diagrama de reaccion "de libro", pero
 en el segundo, dependiendo de la geometria inicial y el intervalo utilizado
 puede obtenerse un diagrama en el que aparentemente la energia disminuye
 siempre, sin que aparezca barrera de activacion. Resulta necesario estudiar
 muy en detalle la region critica para localizar dicha barrera.
 Por otra parte, creo que utilizas demasiadas restricciones en el acercamiento
 de Nu, lo cual puede dar problemas. Ten en cuenta que cuantas mas
 restricciones
 utilices, menos representativa sera la superficie de energia calculada con
 respecto a la "verdadera".
 Nosotros solemos emplear el siguiente esquema:
 Nu ------C-------L    Donde el diedro se fija en 180 grados y los angulos
          |            Nu-C-Du y Du-C-L se fijan en 90 grados. Es estricta
          |            logica, no deberia fijarse ninguna restriccion, pero
         Du            si Nu o L llevan carga neta, la tendencia a dar
                       reacciones "exoticas", como arrancar un proton
 cercano
                       es bastante grande.
 Con las restricciones indicadas, se fuerza a  que el ataque de Nu y la salida
 de L se produzcan a lo largo de una linea recta, como usualmente se asume en
 las SN2.
 Espero que estas observaciones te sean de alguna ayuda.
 Un cordial saludo.
 Jose Ignacio
 --
 --------------------------------------------------------------------------------
 Dr. Jose Ignacio Garcia-Laureiro                     Phone : 34-(9)76-762077
 Departamento de Quimica Organica                                      761210
 Instituto de Ciencia de Materiales de Aragon         Fax   : 34-(9)76-761159
 C.S.I.C.-Universidad de Zaragoza                     e-mail:
 jig - at - qorg.unizar.es
 E-50009 ZARAGOZA (SPAIN)
 jig - at - msf.unizar.es
 jig - at - posta.unizar.es
 --------------------------------------------------------------------------------
 "And all this science I don't understand it's just my job five days a
 week..."
                                                ELTON JOHN - Rocket man
 -------------------------------------------------------------------------
 **********************************************
 	Thanks to all af them for their help.
 	Sincerely
 	A. Entrena