1. Introduction
This account discusses the discovery and the development of self-assembling and self-organizable dendrons, dendrimers, and dendronized polymers. It also briefly illustrates similarities and differences between covalent dendrimers and supramolecular dendrimers by stressing the advantages and disadvantages exhibited by each of them. This article provides a personal view of the corresponding author (VP) on this discovery. As the main author (VP), I would like to state that our laboratory never intended to make this discovery. Therefore, brief stories that provided this accidental discovery are presented here. We were happy when the corresponding author received the invitation from Anne-Marie Caminade to write an account on dendrimers in this special issue, and we decided to tell the story of this discovery. Two reasons led to this decision: (a) the French scientific community played a very important role in this discovery; (b) the most important paper that led to this finding was also published in this journal. Finally, I would like to make a few comments about the rules I (VP) employed in writing this report. Searching by current electronic systems is very efficient. However, this search could generate misleading results if authors did not use the same words in their title and abstract that were employed during the electronic search. Therefore, we used the rules of the late Professor Herbert Morawetz when he wrote the most accurate and influential history of polymers in his book Polymers. The Origins and Growth of a Science [1]. In the Preface to his book, Herbert Morawetz made the following important statement:
I found out quite quickly that there is no substitute for reading every reference cited—second-hand citations are incredibly unreliable. Unfortunately, the insistence on citing only publications I had read has led undoubtedly to some neglect of the Russian and Japanese contributions; for this I apologize but it was a price I had to pay if I was to cite only what I knew in detail.
He also listened to recollections of many polymer scientists from different areas of research. One example was Herman Mark. On talking about Mark, Herbert Morawetz stated the following:
Throughout our association of more than thirty years I had never seen so much of him. Listening to his recollections was frequently a revelation; his memory of events lying back sixty years or more never ceased to amaze me.
In addition to the two statements made by Herbert Morawetz, I would like to add that an electronic search of the literature can be not only extremely efficient but also incredibly unreliable if different names are used for a certain structure. For example, in the case of the discovery of covalent dendrimers, on which I have written a brief article recently [2], different names in addition to the name dendrimer were used by the four laboratories that pioneered this field. Therefore, an electronic search of the literature would have failed to find three of the four discoverers of covalent dendrimers. Only a careful and complete reading of the literature enables documenting discovery events. Covalent dendrimers were discovered by synthetic organic chemists who published their work in chemistry journals and in patents. However, credit was given to all of them regardless of the different names they used. Supramolecular dendrimers were discovered at the interface among several different disciplines including molecular biology, synthetic biology, chemistry, physics, crystallography, and liquid crystals. Most of the scientists working on covalent dendrimers are not involved in these fields. Therefore, the literature on liquid crystals was not of interest to scientists working on covalent dendrimers. Incidentally, our laboratory was and is active in a large number of research fields including liquid crystals. Therefore, this account tries to bridge all disciplines that influenced the discovery process reported here.
2. The inspiration
During the fall of 1982, my colleagues from the Department of Physics at Case Western Reserve University in Cleveland, Ohio, invited me to attend a lecture given by Professor Aaron Klug of the Laboratory of Molecular Biology from Cambridge University [3, 4]. His lecture was followed by a French dinner, which I was also invited to attend. Professor Aaron Klug had obtained his PhD in physics at Cambridge University in the UK. After his PhD, he started to work with Rosalind Franklin in the Department of Physics led by J. D. Bernal at Birkbeck College, University of London, on the structure of viruses (Figure 1) [5]. Rosalind Franklin had made important contributions to the determination of the double helix of DNA and to the elucidation of the structure of viruses [5, 6, 7, 8, 9, 10]. Originally, I was hesitant to attend a lecture given by an expert in physics rather than chemistry. However, the invitation to the French dinner together with the lecturer convinced me to attend his presentation. A few days after this event, Professor Aaron Klug received the Nobel Prize for Chemistry for
his investigations of the three-dimensional structure of viruses and other particles that are combinations of nucleic acids with proteins and for the development of the crystallographic method of electron microscopy. [4]
“We aim to understand the functioning of molecules through their structure”, stated Aaron Klug while explaining his work [3]. The fact that Aaron Klug was a physicist and received the Nobel Prize in Chemistry was quite intriguing to me, so I decided to read his Nobel lecture [4, 11]. Almost immediately, I became inspired by the self-assembly and self-organization of viruses. I decided to develop synthetic systems that could, at least in a primitive way, mimic the structure(s) and the self-assembly of viruses. I saw this as a fascinating new avenue for research in my laboratory. However, for several years, I struggled to find a viable approach to tackle this challenge, which involved synthesizing unknown non-biological molecules that behaved like biological molecules and therefore self-assemble and self-organize into helical rod-like and globular virus-like supramolecular architectures. Natural viruses were fascinating to us since they were known to form various liquid crystal lyotropic phases [12, 13, 14, 15, 16]. Helical rod-like viruses were known to self-organize into columnar hexagonal lyotropic liquid crystalline phases while globular viruses generate cubic periodic arrays. Crick and Watson stated that it is a striking fact that all viruses are either rods or spheres [17]. The structural characterization of crystalline and liquid crystalline periodic arrays of viruses was provided by X-ray and by other diffraction methods.
Rosalind Franklin (left), TMV model (center), and Aaron Klug (right). © Elliott & Fry and the National Portrait Gallery. © MRC Laboratory of Molecular Biology.
3. The scientific failure leading to the discovery
During the early fall of 1987, Professor Alfred Saupe from Kent State University visited my office on the recommendation of our mutual friend Professor Helmut Ringsdorf from the University of Mainz. Professor Ringsdorf had suggested that Saupe reach out to me for assistance in solving a synthetic scientific problem. Saupe presented to me a paper by Jacques Malthête and his collaborators [18] from College de France who had reported a molecule constructed from combining a half-disc with a rod (Scheme 1, right). Malthête et al. had claimed that this molecule exhibits a monotropic biaxial nematic (Nb) liquid crystal phase. Malthête’s structure was inspired by a concept advanced by the theoretical physicist Chandrasekhar (Scheme 1, left) [19, 20]. A monotropic phase refers to a phase that is metastable and is observed only during cooling. Saupe discovered the first Nb liquid crystal phase that was lyotropic and difficult to investigate [21]. Normal nematic (N) molecules undergo translational and rotational motions while Nb molecules are expected to display a more complex combination of motions. Saupe was keen to transform the monotropic phase of the Malthête et al. molecule into an enantiotropic state in order to be able to explore and elucidate its physical properties including its motions.
The thermotropic biaxial nematic liquid crystal is composed of a half-disc and rod-like entities as proposed and demonstrated by Chandrasekhar (see structure on left side). The structure from Malthête et al. is on the right side [18, 20].
We were just about to publish together with Professor Andrew Keller from Bristol University in the UK some phase diagrams that explained the correlation between the molecular structure of liquid crystals and the transformation of monotropic and even virtual liquid crystal phases into the enantiotropic phase [22]. Based on this concept, I asked my junior graduate student Jim Heck to attach a methacrylate group at the end of the rod side of Malthête’s molecule expecting that the resulting polymethacrylate, or any other polymer backbone, will accomplish the monotropic to enantiotropic transformation of Malthête’s Nb phase (Scheme 2).
Synthesis of poly(4-11-methacryloylundecan-1-yloxy)-4’-[3,4,5-tri-(p,n-dodecan-1-yloxbenzyloxy)benzoate]biphenyl) (top); synthesis of 4-(11-undecan-1-yloxy)-4′{3,4,5-tri[p-(n-dodecan-1-yloxy)-benzyloxy]benzoate}biphenyl side groups (middle); hydrogen-bonded structure of 3,4,5-tri-(p-n-dodecan-1-yloxybenzyloxy)benzoic acid (bottom).
As seen from the top part of Scheme 2, we have shown the half-disc part of Malthête’s molecule to be a first-generation self-assembling dendron or as we will discuss later a “mini-dendron”, which by H-bonding self-organizes a supramolecular disc-like structure forming a column [23, 24]. Supramolecular columns form a hexagonal periodic array. A combination of differential scanning calorimetry (DSC), thermal optical polarized microscopy, and X-ray diffraction was employed to demonstrate that by H-bonding, this self-assembling dendron forms a disc-like structure assembling a supramolecular columnar dendrimer, which self-organizes a columnar hexagonal liquid crystalline periodic array. This will be discussed later. The experiments shown in Scheme 2 along with numerous studies on libraries of closely related molecules to Malthête’s compound and their polymers did not yield the expected transformation from monotropic Nb to enantiotropic Nb (Schemes 2 and 3). All these compounds self-organized to form columnar hexagonal periodic arrays. In fact, the Nb phase was shown by Saupe and others from the Liquid Crystal Institute at Kent University to be a monotropic smectic phase. Since we do not publish negative results about other laboratories, we did not publish this information. However, two different laboratories made this correction [25, 26]. A reinspection of the first two lab notebooks of Jim Heck was made during the writing of this article. This inspection indicated that from January 1, 1988 to February 28, 1989 he investigated numerous libraries of Malthête-like, Malthête-related, and fragments of Malthête’s molecules (Schemes 2–5). The combination of methodologies mentioned previously revealed that libraries of Malthête-like molecules formed a supramolecular coat around the polymer backbone, assembling into a helical column or a supramolecular column if no polymer backbone was attached to the building block.
Polysiloxanes containing Malthête’s side group and variants of a Malthête-like side group.
Representative examples of self-organizable molecules based on the modified disc-like part of Malthête’s molecule.
Synthesis of 2-vinyloxyethyl 3,4,5-tris[4-(n-dodecanyloxy)benzyloxy]benzoate based corresponding polymer.
Scheme 3 illustrates that kinked or bent-rod-like moieties attached to the half-disc of Malthête’s structure or even two simplified half-discs assembled on their own (not shown) or when attached to polymer backbones self-organize into columnar hexagonal liquid crystalline periodic arrays. Surprisingly, removing the rod-like moiety of Malthête’s molecule and replacing it with monoethylene glycol, diethylene glycol, triethylene glycol, and tetraethylene glycol containing an OH or methoxy groups at their end generate by complexation with metal triflates or without the same self-organized columnar hexagonal periodic array. The presence of a single H-bond mediates this self-organization process. Replacing Malthête’s half-disc with a simplified half-disc without benzyl ether groups provided the same periodic arrays. Replacing the oligooxyethylene fragments from Malthête’s half-disc or from the simplified half-disc with crown ethers benzo-15-crown-5 and 15-crown-5 produced the same periodic arrays after complexation with metal triflates. Representative results with triethylene glycol and with crown ethers are summarized in Scheme 4.
The structures from Scheme 4 are in fact the first examples of self-assembling amphiphilic dendrons [27, 28, 29, 30, 31, 32, 33, 34, 35]. Jim Heck was highly disappointed by these results but I (VP) became enthusiastic, as the resulting polymers exhibited a three-dimensional structure resembling the helical rod-like structure of the tobacco mosaic virus (TMV) (Figure 1). These were the first results on self-organizable dendrons and dendronized polymers. At that time, covalent dendrimers were thought to have a spherical shape. A different shape was reported in our communication from the 1989 ACS meeting at Miami Beach, Florida [23], in five additional ACS presentations [27, 36, 37, 38, 39] up to 1992, and five publications between 1990 and 1991 [24, 40, 41, 42, 43]. In all these presentations and publications, we referred to these groups as tapered-shaped moieties, tapered side groups, tapered side chains, or hemiphasmidic mesogens rather than tapered self-assembling or self-organizable dendrons. In 1991 [43] and 1992, powder X-ray diffraction experiments allowed us to name the supramolecular columns generated in the absence of the rod-like part of the dendronized polymer with Malthête’s molecule, TMV-like assembly periodic arrays. These periodic arrays were the first to be generated by living cationic polymerization of vinyl ethers containing Malthête’s half-disc group (Scheme 5) [44, 45].
Divergent and convergent synthesis of covalent dendrimers [46, 47, 48, 49, 50, 51, 52].
Living cationic polymerization of vinyl ethers [53, 54, 55] and other living polymerizations [56, 57, 58, 59, 60, 61] were routinely employed in our laboratory to elucidate molecular-structure phase-transition properties in other complex systems including in liquid crystal polymers. Moreover in 1992, we introduced the name liquid crystalline dendrimeric polymer for a library of higher generation self-organizable dendrimers [62], demonstrating that with proper molecular architecture, dendrimers can adopt a variety of conformations including columnar and willow-like rather than only the commonly expected spherical shape [63, 64]. Branched columnar architectures were also elaborated at the same time [65]. These experiments demonstrated that simple self-organizable dendrimers of first and other lower generations can form complex molecular architectures that are not achievable by conventional covalent dendrimers. In 1992, Jim Heck defended his Master of Science thesis titled “Supramolecular Polymers” [66]. Most probably, this was the first thesis with this title in the entire world. Few years later in January 1995, Jim Heck defended his PhD thesis titled “Self-Assembly of Tubular Supramolecular Architectures via a Combination of Endo- and Exo-Recognition Processes” [67]. In his PhD thesis, he employed self-assembling dendrons to self-organize into functional columnar assemblies. During the early 1990s, the corresponding author of this paper was invited to give numerous invited and plenary lectures around the world. The first four brief review articles on this topic including the first on TMV-like assemblies were published in 1994 [29, 68, 69, 70]. Seven more brief reviews were published in the next two years, and in some of them, the name dendrimer is part of their title [28, 71, 72, 73, 74, 75, 76]. The only proof that covalent dendrimers such as PAMAM may also exhibit a rod-like shape was provided in a landmark publication by Naylor, Goddard, Kiefer, and Tomalia, which suggested, by molecular mechanics simulations, an ovoidal highly asymmetric shape in generations 1–4 followed by a transition to a nearly spherical shape in generations 5–7 [77]. Definitive proof for “helical” columns was provided by oriented-fiber X-ray diffraction experiments conducted with my colleague John Blackwell and his postdoc Sergei N. Chvalun, which were reported in 12 publications starting in 1994 [74, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88]. These studies showed that the half-disc of Malthête’s molecule was not a half-disc but rather one-eighth of the cross-section of a supramolecular helical column regardless of whether it was attached to a polymer backbone or not.
The experiments demonstrated that these first-generation dendrons, synthesized through a convergent rather than divergent approach (Scheme 6, top) even before the name convergent synthesis (Scheme 6, bottom) was coined [2], self-assemble into helical columns, which in turn self-organize just like TMV into helical columnar hexagonal periodic arrays (Figure 2). The helical pitch of the supramolecular columns self-assembled from both dendronized polymers and from tapered dendrons was also discussed in these publications.
Supramolecular helical columns from self-assembling dendrimers (top) and from self-organizable dendronized polymers (bottom).
The libraries of self-organizable dendrons and dendronized polymers from Schemes 4 and 5 and from references associated with them concluded that helical columns induce a helical conformation in an atactic polymer backbone, resembling the case of mRNA in TMV (Figure 2).
These publications encouraged both our laboratory and the broader scientific community to recognize that supramolecular dendrimers can adopt a variety of shapes or conformations in addition to the widely accepted spherical shape. The name “TMV-like” was included in publications and in review titles [29, 44]. A microsegregation of the aliphatic part from the aromatic fragment of self-assembling dendrons was demonstrated to be responsible for the mechanism of self-assembly and self-organization (Figure 3) [89].
(a) Structures of 1,2-bis{3,4,5-tris[S(-)-2-methylbutan-lyloxy]benzamido)ethane and l,2-bis[3,4,5-tris(n-dodecan-l-yloxy)benzoyloxy]ethane (top); (b) schematic representation of the columns in the 𝜱h phase (top view), where the inner core is composed of fixed width by aromatic groups and linkers and an outer sheath is composed of molten alkyl chains of uniform density (bottom left); (c) linear dependence of the column diameter (DCol) or hexagonal lattice parameter ($a_n^2$, DCol = a) on n (n = 1–18); (d, e) schematic views of the columns of two different models forming a tilted angle. Each parallelepiped represents one DA-n molecule, and the connecting vertical lines represent the hydrogen bonds.
This demonstration provided access to systematic developments of self-organizable dendrons, dendrimers, and dendronized polymers. Moreover, different functionalities were incorporated at the apex of self-assembling dendrons, which led to the development of functional supramolecular columns and their self-organizable periodic arrays. This was the time when electronic literature searches began to be used. Unfortunately, the absence of terms such as self-assembling, self-organizable dendron, dendrimer, and dendronized polymer in the titles of our very early publications may have made it difficult for laboratories to find our and other laboratories’ work in specialized literature by electronic searches except for those specifically interested in the liquid crystallinity of covalent or supramolecular dendrimers. Nevertheless, we believe that the words taper, tapered, tapered molecule, tapered building block, and half-disc clearly indicate the structure of a dendron. Therefore scientific curiosity should have provided access to the chemical structure investigated just like in the case of covalent dendrimers. What else could a tapered shape be generated from other than a dendron [90, 91, 92]?
4. The discovery of self-organizable spherical supramolecular dendrimers
In 1997, after many years of very complex and time-consuming structural analysis [93, 94, 95], spherical supramolecular dendrimers having the perfection to self-organize into cubic periodic arrays, as spherical viruses do [17] and as originally predicted by Jim Heck only from DSC and optical microscopy, were reported. When Jim Heck predicted many years in advance the spherical shape of the supramolecular dendrimers from 1997 [94], I showed the structure to Professor Andrew Keller who was visiting our department. He responded by saying, “Fire this student, since when an organic chemist can predict the crystal structure of organic molecules?” I did not fire Jim. Instead, together with Goran Ungar, a former postdoc of Andrew Keller, we determined the structure of Jim’s molecules and demonstrated the spherical shape of the supramolecular dendrimers, which were self-organized from conical dendrons just as Jim had predicted. This challenging structural analysis used a combination of X-ray diffraction, electron density maps, reconstruction of the experimental X-ray data with molecular models, isomorphic replacement, transmission electron microscopy, and scanning force microscopy. My colleague Steven Hudson together with Sergei Sheiko, who was working in Martin Moeller’s laboratory at the University of Ulm (Germany) at the time, also contributed to this pioneering work confirming Jim’s original prediction. This cubic periodic array shown in Figure 4 corresponds to the Frank–Kasper A15 phase. This phase was discovered in 1958 in metals by Sir Charles Frank [96, 97], the head of the laboratory at Bristol University in the UK, with whom Professor Andrew Keller worked. Shortly thereafter, the tetragonal or Frank–Kasper 𝜎 [98] and the soft quasicrystal known as liquid quasicrystal (LQC) [99, 100] phases were also discovered in our laboratory. These discoveries laid the foundation for the field of soft quasicrystals and soft Frank–Kasper phases in other forms of self-organized soft matter [101] including in block copolymers [102].
(a) Synthetic supramolecular system with cylindrical and spherical shapes; (b) supramolecular shape change during radical polymerization of conical dendronized monomers, illustrating the quasiequivalence concept; the SFM images of (c) spherical and (d) cylindrical assemblies. Adapted from Ref. [93]. Copyright © 1998, Macmillan Magazines Ltd.
At the time we discovered the first Frank–Kasper A15 phase in supramolecular dendrimers, we had a joint grant on dendrimers with several other scientists including W. A. Goddard from Caltech. Ref. [94] appeared in the 1997 JACS issue no. 7 [94], which is published every Wednesday prior to our meeting on Mondays. Each of us comes in the morning of the meeting and starts our presentations at 9:00 am. Bill Goddard from Caltech was the first speaker and I was second. To my shock, he gave a lecture on the work published in Ref. [94] explaining by computer simulation how and why the A15 Frank–Kasper phase would self-organize from our self-assembling dendrons. I started my presentation by asking Bill Goddard how he could do this work so quickly since at that time JACS and all other ACS journals were delivered as hard copies and this issue would not be at Caltech on Wednesday but on Thursday or Friday. “I put all my group to work on this simulation from the minute we have seen the paper until I boarded the plane on Sunday night to Washington”, he replied. I asked him if he could predict what other primary self-organizable dendron structures would generate new periodic arrays? His reply was very simple: “If somebody can say that he can predict the periodic array in advance without having X-ray data on the structure, he is not your friend!” Since 1997, I kept asking Bill every couple of years the same question. His answer was the same as in 1997: “Progress is being made but we are not yet there!!!” It required another 7 years to complete the work of his Washington presentation, which was published in JACS in the year 2004 [103]. In a review article on quasicrystals, Tomonari Dotera [101] explains the discovery and the name in the following manner:
In his speech at the conference banquet for the 9th International Conference on Quasicrystals at Ames, Iowa, in 2005, notable polymer chemist Virgil Percec coined the term soft quasicrystal, when two decades had already passed since the revelation of quasicrystals [104, 105]. Until the advent of quasicrystals in a supramolecular dendrimer system [99] followed by a star polymer system [106], quasicrystalline materials had been isolated in one category of materials alone, namely, synthetic intermetallic systems [107].
Following the discovery in dendrimers, soft Frank–Kasper phases and quasicrystals have been discovered in block copolymers [102, 108, 109, 110], giant surfactants [111, 112, 113, 114], lipids [115, 116, 117, 118, 119, 120, 121], surfactants [122, 123, 124, 125], nanocrystals [126], DNA particles [127], sugar–polyolefin block copolymers [128], and in many other soft materials [101]. Work on soft quasicrystals and Frank–Kasper phases was accompanied by a substantial amount of theoretical and simulation research [103, 129, 130, 131, 132], creating a new and independent field of research. “It is a striking fact that almost all small viruses are either rods or spheres” [17], and it is even more striking that almost all supramolecular dendrimers are either helical columns or spherical helices. By drawing an analogy with the concept of quasiequivalence from molecular biology [133, 134, 135], we demonstrated that conical dendrons can also be quasiequivalent, allowing for a conformational change from conical to tapered under a wide range of conditions [2].
5. The most important similarities and differences between covalent and supramolecular dendrimers
When comparing covalent with supramolecular dendrimers, we observe that the multiplicity at the core of the spherical structure in supramolecular dendrimers can be much larger than the maximum value of 4 possible for covalent dendrimers. The largest multiplicities observed in our laboratory for hollow and non-hollow micellar or vesicular spherical supramolecular dendrimers were 480 and less [136], and 770 [137]. The 770 multiplicity provides a supramolecular dendrimer of ultrahigh molar mass equal to 1.78 × 106 g/mol, which is in the molar mass range of ribosomes [137]. In the case of supramolecular dendrimers, we define multiplicity as the number of dendrons self-assembling a molecular object regardless of its structure. The maximum multiplicity of conventional covalent dendrimers like PAMAM is 4 although 6 was reported for covalent dendrimers containing cyclotriphosphazene and cyclotriveratrylene cores and 8 for covalent dendrimers containing phthalocyanine, cyclotetraveratrylene, triphenylene, and other cores. At the same time, the steric hindrance present on the surface of covalent dendrimers is less pronounced in supramolecular dendrimers. These two dissimilar features, larger conical multiplicity in the core and reduced surface steric hindrance, determine the structures and physical properties that differentiate many physical properties of covalent dendrimers from supramolecular dendrimers. Regardless of the mechanism of self-assembly, most supramolecular dendrimers adopt either columnar or spherical shapes similar to viruses and thus exhibit helical chirality (Figure 5) [137]. Covalent dendrimers do not exhibit helical chirality. Self-organizable dendrimers can be characterized in the same way as biological systems by a combination of X-ray diffraction, electron density maps, and circular dichroism. This combination is not feasible for covalent dendrimers that do not self-organize.
Structural and retrostructural analyses of libraries of supramolecular dendrimers and dendronized polymers. Adapted from Ref. [137]. Copyright © 2009, American Chemical Society.
6. Discovery of the molecular design principles of supramolecular dendrimers
The combination of methodologies available for the characterization of self-organizable dendrimers endowed us with an accelerated modular–orthogonal approach to the design and characterization of libraries of supramolecular dendrimers. This allowed determination of the molecular structure of self-organizable dendrons, dendrimers, and dendronized polymers, which self-organize into helical columns or spherical helices. The ovoidal shape of the covalent dendrimer PAMAM during its first four generations indicated by computer simulation [77] was demonstrated by our laboratory to be columnar in supramolecular dendrimers. By analogy with the case of PAMAM, supramolecular dendrimers become spherical at higher generations [138, 139]. This was demonstrated to be a general concept available in many libraries of self-organizable dendrons and dendrimers [136, 137, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183]. The primary structure of the self-organizable dendron is responsible for the shape of the supramolecular dendrimer and dendronized polymer as a function of the degree of polymerization. An example illustrating the principles of this concept is shown in Figure 4.
Libraries obtained by increasing the generation number were named “generational” while those produced by deconstruction were denoted “deconstruction” libraries. Both types of libraries were employed in this process. Figure 5 summarizes the combined structural and retrostructural analyses of these libraries, which illustrates the conformation of the primary structure required to generate periodic arrays constructed from hollow and non-hollow helical columns and spherical helices. Since they are helical, it means that all supramolecular assemblies are in fact chiral but racemic or by spontaneous deracemization, homochiral, even if they are self-organized from achiral building blocks. These results are not unexpected since both helical rod-like and spherical viruses are also chiral although they are generated from chiral building blocks [137].
A nanoperiodic table of self-organizable dendrons was generated by employing these principles [137]. Once a nanoperiodic table of self-organizable dendrons, dendrimers, and dendronized polymers is available (Figure 5), the molecular design of a wide range of functions becomes feasible. However, it should be noted that not all self-assembling dendrons form supramolecular dendrimers that are self-organizable into periodic or quasiperiodic arrays [91, 184]. It is important to state that unless supramolecular dendrimers self-organize into periodic arrays, their structural analysis by X-ray diffraction is not possible. More complex columnar structures including bundles of columns were self-organized from what we called at a certain time self-organizable mini-dendrons or mini-monodendrons [185, 186]. We explained in one of our publications that self-assembling mini-dendrons have extraordinary synthetic capabilities for the elaboration of new structural motifs from larger generations of dendritic building blocks. The role of these self-assembling mini-dendrons is analogous to that of simple peptides used in the understanding of the molecular engineering involved in the assembly of more complex and larger proteins or of maquettes used by sculptors and architects to appreciate various aspects of full-size objects [185, 187]. Vesicular columns and vesicular spheres are now achievable by employing mini-dendrons as maquettes (Figure 5).
7. From discovery to a brief discussion of the development and impact on other research fields
We mention only briefly the development of numerous functions and applications generated by employing self-organizable dendrons, dendrimers, and dendronized polymers. The first application was to inspire artists to paint the supramolecular structure of dendrimers. An example is shown in Figure 6, which illustrates the structure of an A15 supramolecular dendrimer discovered in our laboratory. Additional examples refer to ionic and electronic [188, 189, 190] conductors, mimics of the transmembrane protein aquaporin [191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203], which provided an entirely new field of research for water purification [204, 205], supramolecular orientational memory effect, which predicts the structure of novel rod-like helical assemblies [206, 207, 208, 209, 210, 211], mediating the reactivity of functional groups including conjugated polymer chains by eliminating their intramolecular electrocyclization followed by chain cleavage [212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223], molecular machines [224, 225], amphiphilic Janus dendrimers that self-assemble into dendrimersomes, which are employed in the delivery of drugs [226, 227, 228, 229, 230, 231, 232, 233], amphiphilic Janus glycodendrimers that self-assemble into glycodendrimersomes mimicking the glycans of biological membranes [234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247], deracemization in crystal state [248], the cogwheel mechanism of self-organization of homochiral supramolecular systems [249, 250, 251, 252, 253, 254], and last but not least, multifunctional sequence-defined ionizable amphiphilic Janus dendrimers [255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265], which coassemble with mRNA into dendrimersome nanoparticles with predictable dimensions. They enable targeted delivery of mRNA for therapeutics and vaccines. It is also important to recognize that similar to covalent dendrimers, the discovery of self-organizable dendronized polymers, dendrons, and dendrimers was not accompanied immediately by the name dendrimer. Additionally, in parallel, self-organizable dendronized biological macromolecules such as peptides [192, 266, 267], proteins, and RNA [255] were elaborated [137, 268]. Although the numerous functions mediated by self-organizable dendrons, dendrimers, and dendronized polymers are not enumerated here in detail, it is essential to understand that the mechanisms of self-organization in both biological and synthetic systems are closely related or even identical.
The painting of a spherical supramolecular dendrimer self-organized from conical dendrons inspired by Ref. [139]. Credit: Iektje Meijer Oosterbeek.
8. Conclusions
Accidental discoveries have permanently impacted numerous fields of science including chemistry, and I (VP) am not sure if there is any record about the significance and the outcome of accidental discoveries versus rationally designed research in any field of science. I would like to mention only a few examples referring to the discoveries of crown ethers, cryptands, molecular hosts and guests, molecular recognition, and self-assembly [269, 270, 271]. This selection was made since some of these concepts and building blocks were also employed in the discovery that was discussed here. This account tells the story of the accidental discovery of self-organizable dendrons, dendrimers, and dendronized polymers. I would like to give credit to J.-M. Lehn for providing clarification when referring to self-assembly and self-organization [272]. The main goal of this account is that there is no failure in science since we never know if a correct or an incorrect conclusion of a research and publication will have the largest impact on future scientific developments. In this particular case, the incorrect assignment of a biaxial nematic phase by Malthête’s laboratory impacted in a positive way the discovery of self-organizable dendrons, dendrimers, and dendronized polymers, which by themselves led to the development of numerous additional research fields. We want to thank again Malthête’s laboratory for being incorrect in their assignment of this phase. In fact, the biaxial nematic phase continues to be debated and it could be that we discovered again by accident a biaxial nematic phase in macrocyclic liquid crystals [273]. Our Nb phase seems to be less debated but it could be that it is also less known. However, Malthête’s laboratory made numerous important contributions to the development of supramolecular dendrimers almost at the same time as us, but they received no recognition for their work since they, like us, did not use the name dendrimer [274, 275]. This and other labs’ work may become part of a different story. I would like to apologize for not citing a lot of additional work on supramolecular dendrimers including the spectacular work of Didier Astruc [276], Aida [277], Majoral and Caminade [278, 279, 280], and many others, but the decision to write a personal story of our discovery limited the scope of this account.
Acknowledgments
The author VP thanks many generations of very gifted and hard-working students and postdocs and many collaborators from the US and from around the world who made this story possible. He also thanks Donald A. Tomalia who recommended writing this story to him.
Declaration of interests
The authors do not work for, advise, own shares in, or receive funds from any organization that could benefit from this article, and have declared no affiliations other than their research organizations.
Funding
Financial support by the National Science Foundation (grants DMR-2104554, DMR-1720530, DMR-1807127), the Leonard Case Jr. Chair at Case Western Reserve University, the P. Roy Vagelos Chair at the University of Pennsylvania, and two Alexander von Humboldt Foundation Awards is gratefully acknowledged.


CC-BY 4.0