Isolation and X-ray characterization of palladium–N complexes in the guanylation of aromatic amines. Mechanistic implications

Summary In the context of palladium-catalyzed guanylation of anilines herein, we have been able to characterize and isolate bis(anilino) and bis(guanidino)Pd(II) complexes using reaction conditions under which stoichiometric amounts of palladium salts are used. Characterization of these palladium complexes strongly supports a mechanistic proposal for the catalytic guanylation of anilines using PdCl2(NCCH3)2 as catalyst that involves the intermediacy of these Pd(II) complexes.


Isolation and X-ray characterization of palladium-N complexes in the guanylation of aromatic amines. Mechanistic implications
Introduction N-Arylguanidines are important compounds with interesting biological activities [1,2] such as fungicides [3] and also in supramolecular chemistry as complementary partners of carboxylate and nitro groups [4][5][6][7]. Some of these guanidines are commercially used as antifouling agents in marine paints and in the formulation of protective surface coatings [8][9][10]. N-Arylguanidines can be obtained by aniline insertion into the corresponding carbodiimide [11][12][13][14][15][16][17][18]. This nucleophilic addition can be efficiently catalyzed by palladium salts [19], such as PdCl 2 or Pd(OAc) 2 in homogenous phase. Also recently we have reported that palladium nanoparticles supported on magnesia can be a solid catalyst for this process [20]. Working with PdCl 2 (NCCH 3 ) 2 in dichloromethane we were able to isolate two types of palladium complexes with iodoaniline and guanidine, respectively, (see Scheme 1) that give some clue about the reaction mechanism of the catalytic process.

Results and Discussion
In order to provide further support to the mechanistic proposal for the C-N insertion promoted by palladium(II) suggested by us [20], in the present report we describe the study of palladium-catalyzed guanylation of three additional anilines (1a-c) with N,N'-diisopropylcarbodiimide (2). For these reactions we have been able to characterize three palladium complexes of the type of bis(anilino)Pd(II) (3a-c) as well as three palladium complexes of the type of bis(guanidino)Pd(II) (4a-c) (Scheme 2) whose structures have been characterized by singlecrystal X-ray structural analysis, as well as to obtain evidence for the intermediacy of these complexes in the catalytic process.
Overall the present data reinforce the previous proposal for the mechanism of aniline guanylation.
When a stoichiometric (2:2:1) mixture of anilines 1a-c and carbodiimide 2 with PdCl 2 (MeCN) 2 is stirred at 60 °C in CH 2 Cl 2 , evolution at initial reaction times of a solid precipitate is observed (Scheme 2). Filtration of these precipitates and subsequent washing with CH 2 Cl 2 renders three solids whose combustion analysis is in accordance with the percentages expected for dichlorobis(anilino-ĸN)palladium(II) (3a-c) (see Supporting Information File 1, experimental section). IR spectra of complexes 3a-c show the characteristic absorption peaks due to the coordinated anilines 1a-c (see Supporting Information File 1, Figures S1-S3); these are compatible with the proposed structure for these intermediates. Complex 3a derived from 1a has been recently characterized by single-crystal X-ray diffraction [21] showing similar coordination to trans-dichlorobis(4- iodoaniline-ĸN)palladium(II) complex recently published by us [20]. Compounds 3a-c were also characterized by solid state 13 C NMR spectroscopy that gave spectra showing carbon peaks compatible with the proposed structure (see Supporting Information File 1, Figures S4-S6).
After prolonging the reaction time, the initially evolved precipitate undergoes dissolution indicating that it has been transformed under the reaction conditions. At this stage, filtering of the transparent orange-red (4a,b) and red (4c) solutions followed by subsequent addition of ethyl ether or toluene and slow solvent evaporation at ambient temperature allows the formation of crystals with suitable quality for a crystallographic diffraction study. The structures of these intermediates solved by X-ray analysis showed that these compounds corresponding to trans-dichlorobis[arylguanidino-ĸN(aryl)]palladium(II). Figure 1, Figure 2 and Figure 3 present ORTEP views of complexes 4a-c as well as selected views along some crystallographic axes (see Tables S1-S3 and also Figures S7, S11 and S15 in Supporting Information File 1, and for full details of the crystallographic data see Supporting Information File 2).
Besides X-ray crystal structure analysis, palladium complexes 4a-c were also characterized by NMR spectroscopy, ESIMS and combustion analysis (see experimental section in Supporting Information File 1). 1 Figures S10, S14 and S19).   Structure of guanidine 5a was confirmed by single-crystal X-ray analysis, Figure 4 shows the corresponding ORTEP for compound 5a as well as some views of the crystal packing (see also Supporting Information File 1, Table S4 and Figure S20 and for full details of crystallographic data see Supporting Information File 2). Beside X-ray crystal analysis of guanidine 5a, guanidines 5a-c were also characterized by 1 H, 13 Figure S23 and Figure S26 in Supporting Information File 1).
Overall the information obtained from the experiments performed in the presence of a large palladium excess, in which two kinds of palladium complexes have been detected and isolated, with the formation of guanidines under conditions in which a catalytic amount of palladium is present, allows us to make reasonable mechanistic proposals. Thus, upon contacting anilines 1a-c and PdCl 2 (MeCN) 2 , a rapid formation of dichlorobis(anilino-ĸN)palladium(II) (3a-c) complexes should take place. These palladium complexes will interact with the N,N'-diisopropylcarbodiimide (2) giving rise to the dichlorobis(guanidino-ĸN)palladium(II) (4a-c) complexes. When Pd is used in catalytic amounts, cleavage of this bis(guanidino-ĸN) complex by aniline will form another dichlorobis(anilino-ĸN)palladium(II) (3a-c) completing one cycle and liberating guanidines 5a-c as free products of this catalytic reaction with high yields and selectivities (see Scheme 4). In this mechanism the rate determining step will be the attack of dichlorobis(anilino-ĸN)palladium(II) (3a-c) to the N,N'-diisopropylcarbodiimide (2) (Scheme 4).
In this process, coordination of nitrogen to palladium should strongly reduce the nucleophilicity of the corresponding nitrogen atom and, therefore, the attack at the carbodiimide would be significantly slowed down compared to free uncoordinated aniline. However, this negative effect of palladium coor- dination to aniline should be overcompensated by coordination of carbodiimide to palladium in close proximity to aniline in fixed geometry as indicated in the presumed intermediates 6a-c. There are precedents in the literature [22] showing that palladium(II) can interact weakly with accumulated carbodiimide bonds, but this type of complex is typically very labile and difficult to isolate under the reaction conditions due the presence of an aniline excess, and for this reason we have been unable to isolate labile intermediates 6a-c. This preassociation between complexes 3a-c and carbodiimide 2 to form intermediates 6a-c would make easier the key step of C=N insertion leading to guanidines (Scheme 4).
Alternatively, it can be also envisioned that the acidity of hydrogen atoms bonded to nitrogen in intermediates 3a-c increases sufficiently to protonate the nitrogen atom of the carbodiimide that subsequently would be activated to accept the nucleophilic attack of the resulting anilide anion or aniline (see Scheme 4). This mechanism would be similar to that accepted for peptide-bond formation mediated by carbodiimides [23][24][25][26]. It can also be possible that these two steps, i.e., protonation and nucleophilic attack occur in a quasi-concerted manner around the intermediates 6a-c.

Conclusion
In conclusion, we show the possibility to isolate and characterize palladium complexes by performing some reactions using large amounts of palladium salts. The structures of these complexes shed light onto the reaction mechanism of the palladiumcatalyzed reaction. In this case, we have applied this methodology to isolate and characterize bis(anilino)-and bis(guanidino)palladium complexes that are proposed to be reaction intermediates, together with the still not isolated aniline-carbodiimide palladium complex 6, in the mechanism of the guanylation of anilines. Our study opens the way to apply a similar methodology to study the reaction mechanism of other catalytic reactions.

Supporting Information
Supporting Information File 1 Experimental details of preparation, isolation and full characterization of new palladium compounds 3a-c, 4a-c as well as guanidine compounds 5a,b, including IR, NMR, ESIMS and GC-MS spectra for new compounds.