2. The Construction of Dendrimers
Most syntheses of dendrimers involve the repetitious alternation of a growth reaction and an activation reaction. Often, these reactions have to be performed at many sites on the same molecule simultaneously. Clearly, the reactions must be very 'clean' and high yielding for the construction of large targets to be feasible. Many dendrimer syntheses rely upon traditional reactions, such as the Michael reaction,[6,7,12] or the Williamson ether synthesis,[1,2] whilst others involve the use of modern techniques and chemistry, such as solid-phase synthesis,[3,4] organotransition-metal[5,6] chemistry, organosilicon[7,8] chemistry, organo-phosphorus[9,10] chemistry, or other contemporary organic methodologies.[11] The choice of the growth reaction dictates the way in which branching is introduced into the dendrimer. Branching may either be present in the building blocks as is more often the case or it can be created as a function of the growth reaction, as is the case with the PAMAMs and the poly(propylene imine)s. For details of the chemistry employed in the production of dendrimers, there are many comprehensive works[12] which can be referred to by the reader.
2.1. 'Divergent' Dendrimer Growth
The synthetic methodology employed in the early dendrimer syntheses came to be known as the 'divergent' approach. This name comes from the way in which the dendrimer grows outwards from the core, diverging into space. A schematic representation of divergent growth is shown in Scheme 3. Starting from a reactive core, a generation is grown, and then the new periphery of the molecule is activated for reaction with more monomers. The two steps can be repeated. The divergent approach is successful for the production of large quantities of dendrimers since, in each generation-adding step, the molar mass of the dendrimer is doubled. Very large dendrimers have been prepared in this way, but incomplete growth steps and side reactions lead to the isolation and characterisation of slightly imperfect samples.[13] Divergently grown dendrimers are virtually impossible to isolate pure from their side products. The synthetic chemist must rely on extremely efficient reactions in order to ensure low polydispersities.
2.2. 'Convergent' Dendrimer Growth
The 'convergent' approach was developed[14] as a response to the weaknesses of divergent syntheses. Convergent growth begins at what will end up being the surface of the dendrimer, and works inwards by gradually linking surface units together with more monomers (Scheme 4). When the growing wedges are large enough, several are attached to a suitable core to give a complete dendrimer. The advantages of convergent growth over divergent growth stem from the fact that only two simultaneous reactions are required for any generation-adding step. Most importantly, this protocol makes the purification of perfect dendrimers simple.
There are also certain other advantages associated with convergent growth. The growth reactions do not have to be so stringently efficient, and it becomes possible to introduce subtle engineering into the dendritic structure. This principle will be examined in detail in the next section. Convergent syntheses are not without their own shortcomings, however. The number of steps required to build up a large structure is not reduced compared with the divergent approach, yet a great deal more starting material is required. The convergent methodology also suffers from low yields in the synthesis of large structures. Dendritic wedges of higher generations encounter serious steric problems in the reactions of their 'focal points'.[15]
2.3. 'Hypercores' and 'Branched Monomers'
In the early 1990s, the area of synthetic methodology in dendrimer research was led by the Fréchet group at Cornell University. After the development of the convergent approach, their efforts focussed on the acceleration of dendrimer syntheses. The outcome of this research was the demonstration (Scheme 5) of 'hypercores'[16] and 'branched monomers'.[17] These methods involve the pre-assembly of oligomeric species which can then be linked together to give dendrimers in fewer steps or higher yields.
Hypercores and branched monomers allow the chemist to devise synthetic strategies that are more convergent in the classical synthetic sense of the word. An interesting comparison of convergent, divergent, and hypercore synthesis in the preparation of phenylacetylene dendrimers was attempted by Moore,[29] but solubility problems in the divergent steps made the convergent approach favourable.
2.4. 'Double Exponential' and 'Mixed' Growth
The most recent fundamental breakthrough in the practice of dendrimer synthesis has come with the concept and implications of 'double exponential'[18] growth (Scheme 6). Double exponential growth, similar to a rapid growth technique for linear polymers,[19] involves an AB2 monomer with orthogonal protecting groups for the A and B functionalities. This approach allows the preparation of monomers for both convergent and divergent growth from a single starting material. These two products are reacted together to give an orthogonally protected trimer, which may be used to repeat the growth process again. The first use of a versatile building block in this way was performed by Shinkai,[20] although the full potential of the idea was not realised. Double exponential-type methodology has also been employed by Fréchet,[21] Sharpless,[22] Roy,[23] and Schlüter.[24]
The strength of double exponential growth is more subtle than the ability to build large dendrimers in relatively few steps. In fact, double exponential growth is so fast that it can be repeated only two or perhaps three times before further growth becomes impossible. The double exponential methodology provides a means whereby a dendritic fragment can be extended in either the convergent or the divergent direction as required. In this way, the positive aspects of both approaches can be accessed without the necessity to bow to their shortcomings.
2.5. Other Accelerated Growth Techniques
Other advances made in the area of fast dendrimer growth have been less versatile. Of particular interest is the very elegant 'two-step' approach developed by Fréchet.[25] In this approach, two different monomers are used so as to avoid the need for an activation step between growth steps (Scheme 7). The two monomers are of AB2 and CD2 constitutions, where A and D react to form a bond under conditions where B and C are stable, and B and C react to form a bond under conditions where A and D are stable. Convergent or divergent growth is theoretically possible using this method. The problems which are likely to be encountered arise from the difficulty in finding a set of reactions which conform to the above criteria, as well as the usual dendrimer requirements of very 'clean', high yields at every step in the reaction sequence.
Scheme 7
Recent advances in the synthesis of low polydispersity hyperbranched polymers[26] has promoted interest in their dendrimer-like properties. It has been found that polymers with a degree of branching[27] (Dbr) of over ca. 60% can display some behaviour more akin to dendrimers than to their linear cousins. Although hyperbranched polymers with a Dbr as high as 90% have been prepared by stepwise polymerisation techniques, they cannot be considered true dendrimers. Hyperbranched polymers are of considerable industrial interest as a consequence of the ease of their synthesis. However, they will never attain the degree of architectural control or monodispersity which the stepwise synthesis of dendrimers offers to the synthetic chemist.