Luận án tiến sĩ: Design and synthesis of branched polymer architectures for catalysis

The Effect of Macromolecular Architecture in Nanomaterials: A Comparison of Site Isolation in Porphyrin Core Dendrimers and Their Ilsomeric Linear AnaloQU€S.... Effects of Polymer Archit

Design and Synthesis of Branched Polymer Architectures for Catalysis

ByBrett Anthony HelmsB.S (Harvey Mudd College) 2000

A dissertation submitted in partial satisfaction of the

requirements for the degree ofDoctor of Philosophy

inChemistry

in theGRADUATE DIVISION

of theUNIVERSITY of CALIFORNIA at BERKELEY

Committee in charge:

Professor Jean M J FréchetProfessor Dirk TraunerProfessor David Schaffer

Spring 2006

UMI Number: 3228353

Copyright 2006 by Helms, Brett Anthony

All rights reserved.

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

®UMI

UMI Microform 3228353 Copyright 2006 by ProQuest Information and Learning Company All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, MI 48106-1346

Design and Synthesis of Branched Polymer Architectures for Catalysis

Copyright 2006

By

Brett Anthony Helms

Design and Synthesis of Branched Polymer Architectures for Catalysis

By

Brett Anthony Helms

Doctor of Philosophy in ChemistryUniversity of California, BerkeleyProfessor Jean M J Fréchet, Chair

Polymers have been broadly applied in synthesis applications for decades

in part due to their macromolecular nature, which facilitates product recovery,

and their multivalency, which enables high loading Both heterogeneous and

homogeneous polymer supported reagents, catalysts, scavengers, etc are

known and have widespread use in industry for the production of fine chemicals

It was clear, however, from the onset of their implementation in chemicalsynthesis that functional polymers displayed unique behavior in solution that was

not observed for their small molecule counterparts In fact, the reactivity ofmoieties bound to the support was regulated by steric accessibility,

mircoenvironment, proximal confinement and other properties that wereassociated with the polymer As these effects became enumerated, it was

thought that they could be controlled through more precisely constructedmaterials However, only recently have such methods for controlled polymer

synthesis become available Particularly useful have been the families ofbranched polymers: for example, dendrimers, dendronized polymers, starpolymers, hyperbranched polymers, and shell crosslinked nanoparticles Thesemacromolecules have proven effectiveness in advanced applications, includinglight harvesting, vaccine and drug delivery, bioimaging, molecular transport, etc.

In the realm of catalysis, they offer unprecedented opportunity for furtherdevelopment Their globular nature is reminiscent of enzymes, whose functionand efficiency are only now being approached by these systems

In Chapter 1, the “dendritic effect” as it applies to catalysis applications

is addressed The benefits of catalyst placement throughout this perfectlybranched polymer architecture are outlined In Chapter 2, a series of porphyrin-cored poly(benzyl ethers) is described The light harvesting efficiency of thepolymers is evaluated as a function of polymer topology (i.e linear or branched).Chapter 3 presents the synthesis and application of the first organocatalyticdendronized polymer This molecule acts as a molecular concentrator forsubstrates undergoing chemical transformation This phenomenon is revisited inChapter 4 with a related pair of dendrimer catalysts, which differ in eitherarchitecture or catalyst nanoenvironment Their impact on the catalytic properties

of the material is described In Chapter 5, a one-pot reaction cascade employingacid- and amine-containing star polymers is presented Through precisechemical synthesis, these otherwise opposing reagents are confined to thesterically restricted cores of the macromoecules and are thus _ spatiallysegregated in solution, thereby preventing their mutual deactivation Finally, a

facile method for the preparation of dendronized polymers via “click chemistry” isdescribed in Chapter 6.

approved:

a TJanuayy 22,200¢Chair Date

Table of Contents ACKNOWIECGEMENL 0 ce en ELE EEE Eee EEE EEE EEE Kế nà nà kết iii Chapter 1 The Dendrimer Effect in CatalySiS eee renee nhu ho 1

lntroducfiOn - Een nh hà bà kh kh 2 Commonly Used Dendrimer Platforms for Macromolecular Catalysts 4 Core-Functionalized Dendrimers in Cafalysis 8 Peripherally Modified Dendrimers in CatalysSis 34 Conclusions and OuflOOK c ST eee eee nh be Đà bà kho 50 REFEFENCES 0 EEE ern rene etter ites 52

Chapter 2 The Effect of Macromolecular Architecture in Nanomaterials: A Comparison

of Site Isolation in Porphyrin Core Dendrimers and Their Ilsomeric Linear

AnaloQU€S EE nà ĐK EEE Ere Enea Ki cà 59

0wss00ei 0 eee eeeeenenn nee EEE EEE ene Ete neta ret bettas 60 EXperimental 0.0 cece cece ee ee cece eens ee ee TT TK Km ng net 61 Results and DiscUssion ST «BE BE Bề kh ta 80 ConcluSÌOn -c c nnnn nen KĐT ki Ki kh 102 Acknowledgemerl -.cQ QQnnnnn nnn n nh HH 103 Ref©r©enC©S ere EEE ELLE LEELA kh gi hà cà eS 103

Chapter 3 Dendronized Cyclocopolymers with a Radial Gradient of Polarity and TheirUse as Catalysts in a Difficult Esterification c.cccScnnnnnnnề nhe hao 111

INTPOGUCTION 0 UE EOE EOE EEE net kh 112 ExperimenfalL ‹ re SE nà nh TK EEE EE EEE 113 Results and DisCUSSỈOn -c ee enn ene kh ch 118 Acknowledgemen c SH Ene ern EE nh ke tee 121 Supporting lnforrmafiOn HT eee HT nh kh ru 121

Chapter 4 Effects of Polymer Architecture and Nanoenvironment in Acylation

Reactions Employing Dendritic 4-(Dialkylamino)pyridine Catalysts 130

EXperimentl 0c cece cece eter e tenner rte enn EEE EEE EEE eae Ett ng 134 Results and Discussion nh nh nh ko bà Hy 142 ConclUSỈOn SH nền TT nh nà ch nh kh kg 145 Acknowledgement ccQnnnnnn Tnhh nh nh kh ch 146 Supporting lInformatfion c nền nh nh nh nà nas 146 Ref©erenC©S ch nen Ki BE nh nh kh EES 162

Chapter 5 One-Pot Reaction Cascades Using Star Polymers with Core-Confined

Catalytic GFOUDS - cence SH TT nh nh Km kh Thì in ee een etes 166

JaixeeI9ie:|eaEưdddiiiiiiad(dÁđú 167 ExperimenfalL - TS nn nn nh nh nh nh ni bon ki kinh ng 169 Results and DisCuSSỈOn rr nh nh bờ 175

The experience I’ve gained here at Berkeley is in no small part due to thegenerosity and guidance of Professor Jean M J Fréchet His leadership, depth anddrive are a constant source of inspiration for me | have learned from him that asuccessful yet aggressive approach to innovation and creative expression is a delicatebalance between individuality and humility For this, | will always be grateful

To my colleagues who helped make this all possible, | would like to extend mymany thanks: Prof Craig Hawker, Scoobie Mynar, Amanda Murphy, Cathy Liang, YuXie, Steve Guillaudeu, Zack Fresco, Claire Pitois, Lisa Marcaurelle, John Klopp, ShawnBurdette, Ray Thibault, Jeff Pyun, Cameron Lee, Sarah Goh, Stephanie Standley, Kevin

Sivula, Christine Luscombe, Will Dichtel, Stefan Hecht, and countless others And to the

little bulls, Andrew Cramer and Canadian Dan You all contribute more than what is

asked to our community, and we are better for it

To my family, | thank you for your patience, love and support for all my

endeavors, fruitful or otherwise We will grow old together and it will be grand

Finally, | would like to thank my extended family: those who have dared tread the

same path as | these last 5 years (or longer) Houston and Taylor deserve merit badgesfor their stamina in this respect Glendon and Jeremy, for taking me down new paths.Patrick and Tony for always leading me down the wrong one And the Claremont kids forreturning me down old ones: Sage, Elise, Dallas, Vida, Addy, Minna, Annie, Brett, Sara,

Govil, Yumé, Yumi, Leika, Cho, Heidi, Ju Young, Pete, Kimmy, Mikel, Andy You all have

my respect, admiration and love.

This thesis is dedicated to my Great Grandfather Joseph

Manuel Gomes Jr on the occasion of his 90" birthday

After you have exhausted what there is in business,

politics, conviviality, and so on - have found that none

of these finally satisfy, or permanently wear - what

remains? Nature remains

Walt Whitman

| took a deep breath and listened to the old bray of my

heart l am 1am lam

Sylvia Plath

Chapter 1.

The Dendrimer Effect in Catalysis.

Abstract

The immobilization of catalytic groups at the core of dendrimers or at their

periphery gives rise to unique properties that affect rates of reaction, substrateactivation or selectivity, etc When advantageous, these properties can be classified

as a positive dendritic effect Positive dendritic effects can arise from site isolation,transition state stabilization and/or dielectric effects in the case of core-modifieddendrimers, while peripherally modified dendrimers usually benefit from stericcrowding or cooperativity for catalytic residues at the polymers surface In this

review, the appearance of positive dendritic effects from the literature will be

highlighted as well as prospects for future work in the field

In the interest of “green chemistry” and atom economy, molecular catalysisremains an active area of research in both academia and industry, primarily for theproduction of fine chemicals used in agribusiness, drug manufacture, organic

electronics, etc.” Solution-based methodologies rely not only on the discovery and

optimization of small molecule catalysts for a given transformation, but also onefficient means to separate the desired product(s) from the catalyst and anybyproducts in the reaction mixture While this process can be facilitated by usingheterogeneous catalysts, i.e catalysts immobilized onto organic or inorganic solidsupports such as silica or crosslinked polymeric beads, there are several drawbacks

to this approach that generally limit their widespread application to batch processes.These drawbacks include nonuniformity in catalyst structure and microenvironmentonce it is immobilized onto the support material, slow diffusion of substrates therein,

catalyst leaching, and lower overall activities than the homogeneous system.Š Solid

polymeric supports also exhibit limited swelling ability in certain solvents Efforts toovercome these disadvantages have primarily been directed at using soluble

polymers for catalyst immobilization.* In this manner, the desirable features of

homogeneous catalysis, such as comparable reaction kinetics and mass transfer,can be maintained while the macromolecular nature of the material provides aconvenient means of purification and, in some cases, recyclability (vida infra)

Both linear polymers and various families of branched polymers have beenused as soluble macromolecular supports for reagents and catalysts While the

former are, in general, more readily available, they can suffer from poor loading

capacity It is common for only one catalyst moiety to be appended to the end of alinear polymer, such as monomethoxy-poly(ethylene glycol) (MeO-PEG) Forexample, MeO-PEGzooo-catalyst conjugates carry a loading of only 0.2 mmol catalystper gram of polymer while branched polymers typically carries between 5 to 25

mmol g1 when the degree of branching is greater than 50%.° In the case of

dendrimers, where the degree of branching is 100%, the highest possible loadingcan be achieved Certainly, in the literature, there has been extensive work tooptimize the catalytic activity of functional dendrimers so as best represent that for

the small molecule, but with the inherent retention ability that accompanies

immobilization on macromolecular supports There are both covalent®’ and noncovalent®"° approaches described in the literature and many have been used in

continuous flow membrane reactors (CFMRs) with high efficacy.2""'? The

dendrimers in these studies showed a high degree of independence and stability forthe catalytic groups at the polymer surface even at high loadings for the larger

dendrimers

In some cases, however, high catalyst loading and comparable reactivity to

the parent catalyst are not the paramount concerns Rather that immobilization ofthe catalyst onto a polymer is expected to give rise to certain properties that are notpossible otherwise Such properties might include enhanced stability through stericisolation, cooperativity through proximal confinement of reactive groups, or catalyst

tuning through dielectric effects exerted by its nanoenvironment These properties

are uniquely afforded by systems employing branched polymers, particularly

dendrimers, since their syntheses generally produce well-defined materials with a

high degree of structural control and placement of functional groups throughout the

macromolecule.''“?Š In this review, we will discuss the unique role of polymer

topology in controlling the performance of soluble polymers in homogeneouscatalysis with an emphasis on dendritic polymer architectures The properties arisingfrom catalyst incorporation onto these macromolecules can be described as a

“dendritic effect.” This term has been generally invoked in the literature to explainphenomena that arise with increasingly larger generations of dendrimers In theexamples that follow, special attention is given to those wherein a positive dendriticeffect has been observed From these examples, several common themes begin toemerge and thereby comprise the first comprehensive and qualitative description of

the dendrimer effect in catalysis

Commonly Used Dendrimer Platforms for Macromolecular Catalysts

Polymer topology, i.e linear or branched architecture, has profoundimplications in catalysis It determines the number and spatial arrangement ofimmobilized catalyst groups in solution The chemical nature of the support also

affects the rate of reaction through dielectric effects and through its interaction withthe substrate(s) and product(s), thereby affecting mass transfer These effects havebeen noted in the literature since the first Merrifield resins were used in solid-phasesynthesis applications in the 1960s Until now, however, the effects have been hard

to quantify without reliable and precise chemical means of controlling polymermicrostructure and topology through organic synthesis

Recently, chemical approaches to control polymer topology have reached a

9 tưanê

pinnacle Convergent’ and divergen syntheses of dendrimers are well known,

1821 Controlled

as are slow-addition protocols for hyperbranched polymers

syntheses of star polymers***° have also been put forth, and dendronized linear polymers***° have benefited from the methods developed for their spherical

dendrimer counterparts Each of these architectures has advantages anddisadvantages for catalysis — particularly in how perfect, or monodisperse, theresulting structure is and how easily the material can be prepared and characterized.For example, a high polydispersity has negative implications on characterizingcatalyst activity, due to structural nonuniformity, and generally prevents theapplication of the macromolecular catalyst in CFMRs The high-throughput nature ofthe polymer synthesis is also an important factor when considering itsimplementation in an industrial process

Because of their near structural perfection, dendrimers are arguably the most

well-suited of the branched polymer architectures for catalysis applications.Dendrimers experience known hierarchical order in their conformation at various

Figure 1 Commonly encountered catalyst placement on dendritic

polymer supports: (a) core-functionalized dendrimers; (b) dendrimers

stages of their growth, eventually reaching a globular conformation at the higher

generations.°°** These prescribed changes in structure with increasing generation

provide a convenient basis for rationalizing the catalytic behavior for a series of

related dendrimer catalysts This would be near impossible with other branchedpolymer architectures Other advantages of dendrimers include their monodispersity,

their uniformity in catalyst structure, and their precise syntheses As a result of theiriterative synthesis, catalytic groups may be incorporated at any point resulting in afunctional macromolecule with a single known structure A survey of dendrimer typescommonly used in catalysis applications is shown in Figure 1 Catalyst residueshave been appended to the structures where they are typically found in the

literature: either at the core, at each branch point, or at the periphery Positive

dendritic effects arising from each of these catalyst arrangements are known and will

herein be discussed

Apart from architecture, the chemical make-up of the dendritic polymer can

affect various aspects of catalysis (vida infra) There are several dendritic platformsthat have found widespread use in _ catalysis applications (Figure 2):

polyamidoamines** (PAMAM), polypropylene imines3° (PPI), polybenzyl ethers”Ê (Fréchet-type), polyaliphatic esters,?” polycarbosilanes*® and polyester amides*®

(Newkome-type) These platforms are employed when synthetically convenient,

when reaction conditions allow, and often when there might be some degree ofparticipation from the polymer to catalysis

wre 2 Lo ¬ a g 8 tà 3 mm ana pi ư"

Figure 2 Structures of dendrimers commonly used in catalysis: (a) PAMAM;

(b) PPI; (c) polybenzyl ether; (d) polyaliphatic ester; (e) polycarbosilane; and

(f) polyester amide

Core-Functionalized Dendrimers in Catalysis

The steric crowding of reactive core moieties upon dendritic encapsulationremains one of the more challenging obstacles to overcome in catalysis In general,slower rates of reaction with increasing dendrimer generation are observed whencore-confined catalysts are so isolated from the reaction medium A quantitativetreatment of this phenomenon was recently reported by the groups of Cossio and

Lopez.*° In their work, a single tertiary amine catalyst for the Henry (nitroaldol)

reaction was encapsulated by a series of dendrimers of different generation (GO to

G2) and with different degrees of branching (A-B, A-Bạ, and A-Bạ) Dendrimercatalysts 1, 2 and 3 derived from A-Bạ monomer types are shown in Scheme 1

Pseudo-first-order rate constants, Kops, for the reaction between benzaldehyde and2-nitroethanol were measured for these families of dendrimers via in situ FT-IRmeasurements From these data, they were able to show that the catalytic ability of

a single active site in this reaction decreased with higher generation and higherdegree of branching For example, the ko»; values for dendrimers 1, 2 and 3 were

12.11, 1.89 and 1.03 x 10 sf, respectively The authors further characterized the

structure-activity profiles for these dendrimers using molecular dynamics

simulations According to their results, there was a linear relationship between therelative collision rates of the dendrimers to a value generated from their “reagentaccessible surface” and their molecular weight In addition, the slope of that line,when compared to that for ideal behavior in which the polymer does not participate

in any way in the reaction, were used to determine any (non)cooperative effects

brought about by encapsulation The negative linear departure from ideal behavior

observed in this system clearly indicated the deleterious kinetic effects thataccompany dendritic encapsulation when there is no direct participation by the

polymer backbone in the catalytic cycle or to substrate binding These results may

offer insight to so-called “negative dendrimer effects” in the literature.*1'8

Scheme 1 Henry reaction between benzaldehyde and 2-nitroethanolusing single site dendritic tertiary amine catalysts

While steric shielding of reactive groups at the core of dendrimers usuallycarries negative kinetic consequences, there are several reports where the globularpolymer shell is advantageous The incorporation of monodentate ligands such aspyridines and phosphines into the interior of dendrimers has, in general, been found

to enhance the stability of transition metal complexes derived thereof in thepresence of air and/or other agents that might otherwise lead to catalyst

deactivation.*““ In some cases, catalysts at the cores of these dendrimers suffer a

concomitant loss of activity due to steric crowding of the active site However, incases where stability is the dominant factor in determining activity, dendronizationbecomes an extremely attractive approach to achieve maximum activity from theexpensive metals Due to their strong bond with metals, N-heterocyclic carbene(NHC) ligands are very desirable and have been shown to be extremely versatile intheir applicability to a variety of organic transformations.*? Rhodium(l) complexesbearing NHC-ligands derived from their imidazolium salts have been used inhydrosilylation reactions of ketones under mild conditions, although these reactionsare usually limited by factors such as catalyst stability and activity Tsuji ef al.reported the use of dendritic NHC-ligands in the Rh(I)-catalyzed hydrosilylation of

acetophenone and cyclohexanone (Scheme 2).°° A positive dendritic effect was

observed in that the yield increased with increasing generation of the Rh(I) complex.This effect was attributed to the folding of the dendrimer around the active site thatled to greater stability and hence greater overall turnover during the course of thereaction

Scheme 2 Hydrosilylation of ketones by dendritic NHC-Rh(l) catalysts.

While the previous example invoked a single dendritic ligand in the catalystsystem, it is also common for multiple ligands to be present at an active metalcenter In those instances, various equilibria are involved when the catalyst isformed in situ from metal precursors and excess ligand Indeed, several catalystspecies may be present in solution, and because of their different ligandenvironments, they will have different activities It is also conceivable that core-

functionalized dendritic ligands would alter these equilibria due to excluded volumeeffects brought about by the assembly of multiple ligands to the metal center;because of their smaller size, such a phenomenon would not be experienced withthe parent ligand For example, nickel complexes bearing P,O-ligands, such as o-phenylphosphinophenols, have found wide-spread industrial utility, including the

Shell Higher Olefin Process.°' These reactions are usually carried out in nonpolar

solvents such as toluene, although there is a drive toward more environmentallybenign media In polar solvents like alcohols or water, however, there is a strongtendency for the formation of a bis(P,O)nickel complex 5, which is inactive (Scheme3) In their work, van Leeuwen ef a/ constructed a second generation dendritic P,O-ligand 7 based on a carbosilane platform that was effective in preventing undesirable

bis(P,O)nickel complexes in toluene.°* As a consequence of this site isolation, the dendritic ligand 7 (TOFayg = 7700 h”) outperforms the parent ligand 6 (TOFavg =

3600 h”) in ethylene oligomerization in this solvent and results in higher yields of

oligoethylene The site isolation offered by the dendritic ligand 7 is mitigated in

methanol in that bis(P,O)Ni complexes are observed However, the second ligand appears to be labile at higher reaction temperatures, which is not the casewith the 6 and its bis(P,O)Ni complex As a result, the dendritic ligand gives rise to

P,O-significantly more active nickel catalyst (TOFavg = 3242 h'') for the oligomerization in

methanol than the parent ligand set which did not give rise to any higher olefins

Scheme 3 Ethylene oligomerization using P,O-Ni complexes Formation of

a bis(P,O)-Ni adduct results in catalyst deactivation

Apart from site isolation, the globular dendritic architecture around acatalytically active site is known to bind guest molecules Many early reports of thisphenomenon focused on dye molecules as reporter probes, although substratemolecules can be treated similarly when catalysis applications are considered.Often, the active site of an enzyme consists of a hydrophobic binding pocket forsubstrates in close proximity to moieties that perform the transformation For years,this simple model has been applied to hydrophobic, sometimes amphiphilic,dendrimers with mixed success As described above, slower kinetics often resultfrom the encapsulation of reactive groups within the dendrimer due to stericcongestion, however there are notable exceptions: for instance, when substrate

binding ability and nanoenvironment effects are paramount to achieving reactivity Inthis respect, binding pockets with sufficient size will manifest only in the largerdendrimers A nice example of this phenomenon was reported recently by Zhang et

al with a series of Fréchet-type poly(benzyl ether) dendrimers with a diselenide core

that demonstrate generation-dependent glutathione peroxidase (GPx) activity

(Scheme 4).* In their catalytic cycle, hydrogen peroxide is reduced by benzenethiol

in the presence of the diselenide dendrimers Substrate binding was greater with

increasing generation (Kz = 16.4, 39.4 and 252.7 MT for G1, G2 and G3,

respectively), as were the initial rates for peroxide reduction (vp = 4.07, 8.19 and

2431.20 uM min” for G1 to G3, respectively) In addition, the third generation

catalyst performed around 1400 times faster than Ebselen, an antioxidant previously

studied as a GPx mimic Higher activities for the dendrimers was achieved upon the

introduction of small amounts of water to the reaction medium, which resulted instronger substrate binding to the dendrimer due to the hydrophobic effect This studypoints to the importance of nanoenvironment effects in dendrimer catalysts wheresubstrate binding requires larger dendrimers and whose increasingly hydrophobicenvironment with size provides a more suitable reaction medium

Substrate binding can also be achieved through specific molecularinteractions, such as ion pairing Highly efficiencient, light-driven hydrogen evolutionfrom water was demonstrated by using poly(phenylene ethynylene) (PPE) bearingcarboxylate-terminated poly(benzyl ether) dendritic side groups, G1 to G3, as the

photosensitizer.“ Greater encapsulation of the conjugated backbone with increasing

generation of dendritic appendages led to the suppression of undesirable

self-quenching of the photoexcited state of PPE Methyl viologen (MV**), a positively

charged electron acceptor, was found to aggregate along the negatively chargedsurface of the dendronized PPE, thereby generating a spatially segregated donor-acceptor supramolecular complex (Figure 3) Time-resolved fluorescence

spectroscopy showed that the fluorescence quenching rate constant for the third

generation dendronized linear polymer (kg = 1.2 x 10'° M1 s”) was much greater

than most diffusion controlled rate constants, consistent with the high rate of electron

transfer (key = 9.3 x 10° s†) that would be afforded by the supramolecular assembly Upon excitation of 8 in the presence of a mixture of MV”, triethanolamine (as a

sacrificial electron donor), and colloidal PVA-Pt, hydrogen evolution took place with

an overall efficiency of 13%, an order of magnitude better than prior systems usingcommon organic dyes Specifically, comparative studies with several referencesensitizers showed that spatial isolation of the conjugated backbone and its long-

range z-electronic conjugation, along with electrostatic interactions on the exterior

Figure 3 Carboxylate-terminated dendronized PPE with surface bound

methylviologens (MV2") used in photodriven hydrogen evolution.

surface, play important roles in achieving such efficient photosensitized waterreduction.

While the prominent features of site-isolation have thus far been described in

detail, no particular mention has been made as to the particular nanoenvironment

generated by the dendritic polymer and how that might affect catalysis Dielectriceffects on catalytic reactions are well documented, as are microenvironment effects

in catalysis applications involving soluble and insoluble polymer supports Certainlythose exerted by a soluble polymer on the nanoscale should behave similarly Anearly example on the subject of nanoenvironment effects in catalysis was reported

by Fréchet’s group In their work, third and fourth generation unimolecular dendriticreverse micelles were prepared with a nonpolar corona resembling the reactionsolvent and either ester or alcohol functionality at predetermined locationsthroughout the interior (Figure 4).°° These dendrimers were specifically designed tocatalyze reactions in which a nascent positive charge is developed during thetransition state One such reaction included a simple unimoleculardehydrohalogenation of 2-iodo-2-methylheptane, which proceeds via an E1elimination mechanism Dendrimers with polyol interiors outperformed those withpolyester cores In addition, yields were significantly improved using the larger G4dendrimer One of the most striking features of this catalyst system was its ability toachieve high turnover even at very low catalyst loadings Up to 99% conversion was

observed in the elimination reaction using as little as 0.01 mol% of the G4

polyol-cored catalyst The authors attributed this result to a “concentrator effect” wherebythe relatively polar interior of the amphiphilic dendrimer generated a radially-directed

polarity gradient in the system that favored substrate accumulation within thedendritic nanoenvironment, which was more suitable for the E1 elimination than thereaction medium, hexane Higher generation dendrimers showed better efficacy forthe reaction due to a larger internal reaction volume that could accommodate moresubstrate molecules In addition, the nonpolar alkene product had a greater affinityfor the hydrophobic surrounding medium thereby achieving the necessary turnover.The authors have described this latter phenomenon as a “catalytic pump.”

These concepts, a “concentrator effect” to accumulate substrates near acatalytically active site and a “catalytic pump” to prevent product inhibition, have

been broadly applied to other dendrimer catalysts with amphiphilic designs Hecht et

al reported the first results on light-driven catalysis at the core of the dendrimers via

photosensitization.°° In their design, a ‘Oz-sensitizing benzophenone core was

incorporated into globular dendrimers having a hydrophobic interior and hydrophilic

surface exposed to the polar solvent In this manner, nanoscale photoreactors were

created and applied to the oxidation of cyclopentadiene (CP) with dioxygen (Scheme

5) The nonpolar core was designed to enable both substrate accumulation for CP

as well as longer lifetimes for photogenerated-!Oa To complete the catalytic cycle, the endo-peroxide cycloadduct 9 derived from 1O; and CP was reduced in situ with

thiourea to cis-2-cyclopentene-1,4-diol Since the diol product 10 had a greateraffinity for the surrounding medium, vacant sites for other nonpolar cyclopentadienesubstrates were generated and a high rate of turnover was achieved Indeed, theauthors reported rapid conversions to 10 within 1h using only 0.1 mol% of catalyst,although photobleaching of the dendrimer (approximately 10% over 1h) and CP-dimerization prevented a complete reaction Significantly greater yields wereobtained with increasingly larger dendrimer photocatalysts: G1 to G3 catalysts 11,

12 and 13 gave the diol in overall yields of 15%, 35% and 50%, respectively Inaddition, a benzophenone model compound 14 with a hydrophilic triethyleneglycolenvironment gave less than 10% of diol 10 These data point to the importance of asufficiently large hydrophobic interior to promote adequate substrate binding andaccumulation in these systems employing amphiphilic dendrimers

Core 1O; Q

3Core* 3O; we Mã lài,hài my

9 40 ISC

Scheme 5 Light-driven photooxidation of cyclopentadiene by dendritic

benzophenone-cored 1O; sensitizers.

Other amphiphilic dendritic catalysts from the literature have taken strict

advantage of the ability to concentrate functionality within an amphiphilic dendrimerwhile also looking at the kinetic consequences of the steric shield at the periphery ofthe dendrimer Kaneda eí ai functionalized the periphery of a third generation PPIdendrimer with either C+o or Cig chains and then quaternized the interior with Mel to

afford lipophilic tetraalkyl ammonium iodide dendrimers."” These dendrimers were

used as Lewis-base catalysts (via iodide ions) for the Mukaiyama-Aldol reaction of methoxy-2-methyl-1-(trimethylsilyloxy)propene 15 with various aldehydes in toluene.Their results showed that dendritic iodides were remarkably more effective thanother small molecule iodide sources, such as tetraalkylammonium salts TBAI andTHAI The authors argued that the higher activity of the dendrimers over TBAI and

1-THAI stemmed from the high polarity of the nanoenvironment that encapsulated the

iodide catalysts They noted that these reactions are promoted to a greater extent inpolar solvents like DMF Thus the concentration of multiple cationic charges withinthe nanoscopic confines of the dendritic interior uniquely functioned to stabilize thereactive anionic intermediate In addition, they were able to demonstrate that

dendrimer 16 with a more crowded periphery (i.e with Cig chains) did not perform aswell as 17 with Cio chains For the reaction of 15 with benzaldehyde, dendrimer 17

reached 98% yield of the silylated aldol while dendrimer 16, under the same reactionconditions, gave only 32% Their work clearly delineated the opportunistic features

of concentrating the catalyst in a small reaction volume within the core of the

dendrimer, where the dielectric constant of this discrete nanoenvironment gave

enhanced reactivity, while stressing that sufficient access to that environment wasalso important

OTMS O OH

cat HCI

Me So ——* Me R

RCHO H;O 15

Mukaiyama-Aldol reaction of

1-methoxy-2-methyl-1-(trimethylsilyloxy)propene with various aldehydes

The control of reaction selectivity through dielectric effects exerted by the

dendrimer nanoenvironment has also been observed by Kaneda’s group in catalyzed Heck reactions and allylic aminations Their catalysts were prepared using

Pd-a self-Pd-assembly Pd-approPd-ach from decPd-anoyl-terminPd-ated PPI dendrimers, G2 to G4, Pd-and

4-diphenylphosphinobenzoic acid as ligand for the metal center (Scheme 7).°° The

acid-amine ion pair ensured specific catalyst placement exclusively throughout theinterior of the dendrimer Subsequent treatment of the supramolecular construct with[PdCl(C3Hs)]2 generated the active catalyst 18 In Heck reactions betweeniodobenzene and n-butyl acrylate, rate increases with increasing generation were

observed, while in the absence of the dendrimer no catalysis was observed Asimilar reaction between 1,4-diiodobenzene and n-butyl acrylate with the G4dendrimer catalyst gave a selectivity for the mono-Heck adduct (mono:di = 92:8)while the catalyst system derived from only 19 showed little selectivity (mono:di =45:55) These experiments strongly support the notion that catalysis was occuringinside the dendrimers In contrast with the rate acceleration with increasinggeneration seen with Heck reactions, the allylic amination reaction of cinnamylmethyl carbonate with morpholine showed decreasing activity with increasinggeneration However, linear to branched ratios were higher for larger dendrimercatalysts (/b = 9.1 for G4) and significantly higher than that for 20 (//b = 5.1) It isknown that solvent polarity affects //b ratios, with more polar solvents giving higher

values (b = 14.1 in DMSO) Thus a significant nanoenvironment effect was exerted

by the polymer support in that the charged interior where the catalysts reside was

responsible for the higher //b ratios

Scheme 7 Pd-catalyzed allylic amination by dendrimer-bound Pd-z-allyl

complexes and model compounds

The first systematic treatment of the specific roles of macromoleculararchitecture and nanoenvironment in the catalytic properties of dendritic polymers

was reported by Helms et ai for a related pair of dendrimers and a dendronized

linear polymer containing 4-(dialkylamino)pyridines.°° The catalytic properties of

these materials were investigated in the context of acylation reactions employingsterically demanding tertiary alcohols as substrates Amphiphilic Fréchet-type benzylether and aliphatic ester G3 dendrimers 21 and 22, respectively, were prepared from

a common trivalent core containing three 4-(dimethylamino)pyridine (DMAP) analogswhile a polyester dendronized linear polymer 23 containing 4-(pyrrolidino)pyridines(PPY) along the backbone was also prepared Catalysis experiments clearly

indicated that nanoenvironment played the dominant role in determining the activity

of the polymer catalysts, with the polyester platform being superior to the benzylether (Figure 5) It was noteworthy that model compounds DMAP and PPY were

only marginally effective for the transformation under the reaction conditions

employed while the dendritic catalysts were more competent These results wereconsistent with the “concentrator effect” previously described for amphiphilic

dendrimers in that the alcohol substrates partitioned preferentially into the more

polar core of the dendrimers from the nonpolar reaction medium (i.e cyclohexane);the reaction then benefited from a high local concentration of substrate near the

catalyst In addition, the nonpolar ester product of the reaction would diffuse backinto the solvent thus giving rise to a “catalytic pump” to achieve turnover.Interestingly, polymer architecture played little or no role in affecting rates ofcatalysis in that both the dendrimer and the dendronized linear polymer based onthe polyester platform showed comparable reaction kinetics With respect tomolecular transport and catalysis, this represented the first comparative study of theeffect of architecture and nanoenvironment using structurally similar dendriticmaterials Since then, other work has appeared in the literature to this effect with Ru-

BINAP dendrimers and dendronized linear polymers.°9.5!

literature, this has often been referred to as shape selectivity Early work by Moore

and Suslick with dendritic manganese porphyrins were among the first

shape-selective catalysts.°*°? Pioneering work along this line has also been conducted by Crooks and co-workers using dendrimer encapsulated nanoparticles.~ In each of

these reports, there is a common thread that relates how substrate flexibility affectsthe interaction with the polymer catalyst and what are the kinetic ramifications that

arise from that interaction In a recent paper from Thayumanavan and coworkers,the kinetic behavior of various guest molecules with different degrees of stericconstraint were investigated using distance-dependent excited-state quenching

through photoinduced electron transfer.°° Anthracene-cored dendrimers were

prepared from Fréchet-type dendrons up to the fourth generation (Figure 6) Since,trialkylamines are known to quench the photoexcited state of certain chromophoresthrough electron transfer, core accessibility of the dendrimers could be readilymeasured through the Stern-Volmer efficiency of these molecules to quench theexcited state of anthracene Various trialkylamines of different sizes and rotationaldegrees of freedom were used as quenchers, including triethylamine (Et3N),

N,N,N,N"tetramethylethylenediamine (TMEDA),

NM,N,N,N,N",N"-hexamethyl-tris(2-aminoethyl)amne (TREN-Mea), diazabicycloociane (DABCO), and dimethylaminoadamantane (ADM-NMe;) Each of these amine-based quenchersgave smaller bimolecular quenching rate constants, ky, with increasing dendrimer

WN,N-generation; for example, kạ values for EtsN were 3.5, 3.2, 2.9 and 2.0 x 10° M' s7

for G1 to G4 anthracene-cored dendrimers This is consistent with the greater site

isolation afforded by increasing larger dendritic shells at larger generations The

authors also found that geometrically constrained molecules have better access todendritic cores compared to their more flexible counterparts To that end, kg valuesfor TMEDA (CaH:aN;) and DABCO (CaH:¿N¿) may be compared given their similar

molecular composition but different geometry Quenching constants using the G4

dendrimer were measured at 1.7 x 10° MT s” for TMEDA and 11.3 x 10° MT s" for

DABCO, indicating the more flexible TMEDA molecule has less core access These

results are consistent with other reports in the literature and provide a methodicaland fascinating probe of shape selectivity for substrates using dendritic catalysts.

R3N Quenchers Sy

¬N

Cok Cáa `

-rEtgN TMEDA TREN-Meạ DABCO ADM-NMe;

Figure 6 Quenching of photoexcited anthracene at the cores ofdendrimers by various amines

The question whether to functionalize a dendrimer with catalytic groups at theperiphery or at the interior is often a difficult one Prospects for catalyst loading forboth approaches have been described, as have some of the benefits Sometimes

there are other effects, usually unforeseen, that affect catalysis and are aconsequence of the directed catalyst placement on the dendrimer In a related body

of work from the Ford group, the effects of dendritic quaternary ammonium salts onthe unimolecular decarboxylation of 6-nitrobenzisoxazole-3-carboxylate wereinvestigated using dendrimers with either peripheral or internal functionalization Theperipherally modified dendrimer 26 was prepared by exhaustive methylation of asecond generation polyether dendrimer with 36 primary amine groups at thesurface.°® While the binding constant measured for the dendrimer comparesfavorably with those measured for analogous micelles, various polyelectrolytes andeven latexes, the observed rate constant for decarboxylation of 24 at [24]o < 10” M was only 20 times faster in the presence of < 1.0 mg mL” of 26 than in water alone(i.e Kcat/Kwater = 20) whereas those for other systems typically fall in the range ofKeat/Kwater = 150 — 9,000 for micelles and even higher in the latexes (Kcat/Kwater >21,000 for lipophilic poly[(styrylmethyl)tributylammonium] ion latexes) Thus, whilethe dendrimer did provide a suitable motif for substrate binding, the solvent-exposedenvironment at the surface was too hydrated to destabilize the carboxylate forsubsequent reaction

In later work, the Ford group employed fully quaternized PPI dendrimers G2

to G4 whose periphery had been modified with both Cg chains and triethyleneglycolmonomethylethers to impart hydrophobicity at the substrate-binding interface whilemaintaining aqueous solubility (Scheme 9).°’ These catalysts showed markedimprovement over the previously studied 26 For example, the second and fourthgeneration dendrimers gave Kcat/Kwater values of 323 and 564, respectively In