Probing dense packed limits of a hyperbranched polymer through ligand binding and size selective catalysis

Article

pubs.acs.org/Macromolecules

Probing Dense Packed Limits of a Hyperbranched Polymer through

Ligand Binding and Size Selective Catalysis

Adam Ellis and Lance J. Twyman*

Department of Chemistry, University of Sheﬃeld, Brook Hill, Sheﬃeld, South Yorkshire S3 7HF, U.K.

S

* Supporting Information

ABSTRACT: In the area of dendritic chemistry (hyper-

branched polymers and dendrimers) it is often generalized that

dendrimers are the molecule of choice for smart, selective, or

technical applications involving encapsulation or controlled/

selective environments. This is despite the fact that hyper-

branched polymers (HBP)s are generally easier and cheaper to

synthesize, making them more amenable to large-scale

applications. Dendrimers have been successful in these

applications by virtue of a dense packed or dense shell limit.

This paper describes the synthesis of a series of narrowly dispersed HBPs possessing binding and catalytic cores with a high and

uniform loading. Subsequent binding experiments clearly demonstrated the existence of a dense packed limit with respect to

polymer molecular weight and ligand size. A series of catalytic experiments were also performed in an attempt to exploit these

molecules and their dense packed limit to the area of shape/size selective catalysisan area where dendrimers have previously

been used with celebrated success. However, although we were able to show the existence of a dense packed limit, we were

initially unable to demonstrate any selectivity based on substrate size or shape. Nevertheless, further studies into core branching

motif and multiplicity eventually enabled us to obtain a series of HBPs capable of perturbing the shape/size selectivity of a simple

oxidation reaction involving two alkenes. Speciﬁcally, we were able to demonstrate a 3.5-fold shift in chemoselectivity toward a

smaller alkene of lower reactivity. These results compare favorably with those obtained using dendrimers and allow us to

conclude that, with careful thought regarding core design, HBPs are indeed capable of being applied to technical/smart

applications involving controlled and selective environments.

closed shell properties, are switched on or oﬀ as we move from

one particular generation to the next. This level of precision

makes it relatively easy to control local environment (electronic

and steric) and results in high selectivity and control in a range

of applications.⁶A particularly good exempliﬁer of how this

controlled structure can be exploited was described by Suslick

and Moore.⁷Suslick and Moore showed how porphyrin cored

dendrimers could be used to mimic the reactivity of

cytochrome P450⁸and catalyze the oxidation of various alkenes

in a shape/size selective way. The key experiment involved

taking a 1:1 mixture of the two substrates and catalyzing the

reaction using the porphyrin cored dendrimers. The two

substrates used were a large/fat substrate with a high relative

reactivity and the second smaller/thinner substrate possessing

much lower relative reactivity. When catalyzed by a dendrimer,

the results showed that selectivity shifted toward the smaller

substrate reactivity, despite it possessing a lower inherent

reactivity (compared to the larger substrate). This change in

selectivity was by virtue of the dendrimer’s structure, which

created a steric barrier around the porphyrin and through which

the smaller substrate could pass through more easily.

INTRODUCTION

■

Although hyperbranched polymers (HBPs) can be applied to a

range of applications,¹it is generally perceived that the most

high-end or technical applications are reserved for dendrimers.²

We wanted to challenge this idea and determine whether or not

this generalization was valid. We were particularly interested in

investigating applications that exploit the use of dense shell/

packing³properties when applied to selective encapsulation,

release, or catalysis. To do this, we need to identify the speciﬁc

structural and design requirements necessary for HBPs to

operate at a similar level of performance when applied to the

same high-end or technical applications normally reserved for

dendrimers. These structural and design properties are

controlled by the synthetic methods used to construct the

relevant macromolecule (dendrimer or HBP).

Dendrimers are synthesized using stepwise methods.⁴

Although this is often tedious and time-consuming, it does

lead to a series of well-deﬁned structures known as generations.

These molecules are monodisperse with regards to molecular

weight, structure, and degree of branching (100% by virtue of

complete and controlled reactionsi.e., all branching points

have reacted). Furthermore, as each dendrimer possesses a

unique molecular weight and therefore size, a dendrimer series

can be considered quantized with respect to its generations.⁵As

such, the internal environment, including any packing and

Received: June 27, 2013

Revised: August 1, 2013

A

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Conversely, and despite being inherently more reactive, the

larger substrate had a more diﬃcult journey to the central

porphyrin catalyst. These result demonstrated the ability of a

core functionalized dendrimer and, more speciﬁcally, how

packing around a catalytic core could be exploited to inﬂuence

selectivity. If HBPs were to act in a same way to the dendrimers

described by Suslick and Moore, they would require a similar

level of control with respect to their structure. In particular, we

would need a polymer that possessed a steric or dense packed

limit with respect to molecular weight. However, HBPs are

synthesized using a much simpler procedure involving a

branching monomer and a single-step random polymerization

procedure, generating a mixture of molecules diﬀering in

molecular weight, branching architecture, and 3D structure.⁹

The resulting structures are poorly deﬁned, and any change in

internal steric or electronic environment (i.e., dense shell/

packing properties) is gradual and poorly deﬁned with respect

to size or molecular weight. It is this lack of control regarding

the internal environment that ultimately leads to poor

performance in selective encapsulation, release, or catalytic

applications. Nevertheless, we proposed to investigate whether

or not a well-deﬁned hyperbranched polymer could be applied

as enzyme mimetics^6,10capable of replicating the same levels of

shape and size selectivity demonstrated by the dendrimers

reported by Suslik and Moore.

The speciﬁc aims of the work described in this paper are

threefold. Our initial aim was to obtain a narrowly dispersed,

pseudo-generational series of HBPs possessing a catalytic/

binding core in each and every polymer molecule. Having

achieved this, we would then determine whether or not (and at

what point) an observable and well-defined steric or dense

packed limit existed. The steric properties of the HBP series

would be assessed by studying ligand binding to the central

core. Finally, catalytic experiments would be carried out either

side of any steric limit to test for possible shape/size selectivity.

If these aims could be achieved, then we would have

demonstrated that, in principle, HBPs can be applied to the

more technical and high-end applications usually reserved for

dendrimers.

eﬀects observed with dendrimers. As a consequence of the

controlled synthetic procedure used to construct dendrimers, it

is trivial to place a catalytic group at the core.¹²Speciﬁcally, the

core can be used to start or end the synthetic sequence,

ensuring total incorporation (after puriﬁcation). This is not the

case for hyperbranched polymers, where the core is often

distributed unevenly across the full molecular weight range or is

concentrated in a particular molecular weight fraction.¹³For

any study involving core incorporation and isolation, it is

essential that the core distribution is controlled. Although it is

desirable for each and every polymer molecule to possess a

core, it is not essential. However, it is paramount that the core

is distributed evenly across the polymers molecular weight

distribution, as this allows us to compare a series of HBPs that

diﬀer in molecular weight (pseudo-generations).¹⁴In a

conventional one-pot hyperbranched polymer synthesis carried

out under kinetic conditions, an even distribution of core

molecule cannot be guaranteed, with core units being

distributed randomly across the molecular weight range (or

not incorporated at all). Despite this diﬃculty there are a

number of methods that have been reported for core

incorporation.¹⁵In our case we have used reversible reaction

conditions to control core incorporation with respect to

distribution and ﬁnal molecular weight. A reversible reaction

is dominated by thermodynamics and will result in a statistical

distribution of core units across all molecular weights. If the

conditions are such that the equilibrium is on the side of the

products, then in principle a core molecule will be incorporated

in each and every polymer molecule. We approached this

problem by applying a reversible/equilibrium procedure using a

transesteriﬁcation procedure.

We elected to use 3,5-diacetoxybenzoic acid 1¹⁶as the AB₂

monomer as it generates the same repeat unit used by Suslick

and Moore when synthesizing their shape/size selective

catalytic dendrimer.⁷As such, the electronic environment

created within our HBP will be very similar to that generated in

Suslick’s dendrimer. The AB₂monomer 1 was synthesized from

3,5-dihydroxybenzoic acid by reaction with acetic anhydride

(Scheme 1).^15,16Tetraacetoxyphenylporphyrin 3 was selected

RESULTS AND DISCUSSION

Scheme 1. Synthesis of 3,5-Diacetoxybenzoic Acid

■

Synthesis of Tetra(acetoxyphenyl)porphyrin Cored

Hyperbranched Polymers. Because of the lack of control

regarding the synthesis of HBPs, dense packing will be

dependent on a number of diﬀerent and variable properties.

These include molecular weight, which is eﬀectively a measure

of size, which in turn controls the packing density. However,

this is not necessarily going to aﬀect dense packing in the same

way observed for dendrimers and is due to the reduced degree

of branching that occurs with HBPs.¹¹This reduction in

branching results in a more ﬂexible and open structure with less

geometric constraints. As such, any dense packing is likely to

occur at a higher molecular weight for HBPs in comparison to

dendrimers. Another important parameter to be considered is

polydispersity. In a polydisperse solution there will be a mix of

polymers with a range of molecular weights and sizes;

inevitably, some of these will be below the dense packing

limit. This is not an issue for dendrimers, whose controlled

stepwise synthesis ensures a monodisperse structure. A ﬁnal

problem relates to the level of core or catalyst incorporation.

This is probably the most signiﬁcant consideration when

comparing the properties of a hyperbranched polymer and a

dendrimer, particularly when trying to mimic the generational

as the central porphyrin unit as it contains four acetoxy groups,

each of which can react with the carboxylic acid group on the

AB₂repeat units (via a reversible transesteriﬁcation process).

The porphyrin 3 was synthesized using 4-acetoxybenzaldehyde

2, which was obtained after acetylation of 4-hydroxybenzalde-

hyde using acetyl chloride and triethylamine (Scheme 2).¹⁷

Scheme 2. Synthesis of 4-Acetoxybenzaldehyde

B

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

The porphyrin was then synthesized as shown in Scheme 3,

using the methods developed by Rothemund¹⁸and Adler and

Longo.¹⁹

HBP that possessed a p-nitrophenol core with a 100% level of

incorporation.³Speciﬁcally, 3,5-diacetoxybenzoic acid 1 was

polymerized with porphyrin 3 in a 20:1 molar ratio using

diphenyl ether as solvent (Scheme 4). The mixture was placed

in a round-bottom ﬂask and heated to 225 °C for 45 min. The

temperature was then lowered to 180 °C and the pressure

reduced to 5 mmHg for 4 h. Placing the ﬂask under reduced

pressure allows the removal of the acetic acid byproduct,

driving the equilibrium in favor of the product. Once the

reaction time was complete the crude mixture was dissolved in

reﬂuxing tetrahydrofuran and precipitated into a large excess of

cold methanol to give the porphyrin cored polymer, TAPP-

HBP 6 in 65% yield by mass. The presence of porphyrin in the

polymer sample was immediately apparent from the reddish-

brown color of the product and was substantiated by ¹H NMR

results, which showed small coincident sharp and broad

resonances at chemical shifts corresponding to the porphyrins

aromatic protons at 8.89 ppm. This suggested a mixture of

“free” and incorporated porphyrins with similar evidence

provided by GPC. In addition to a broad peak for the polymer,

a second peak was seen at the low molecular weight end of the

chromatogram. This peak was small accounting for 6% of the

total peak area. GPC analysis using a ﬁxed wavelength UV

detector at 418 nm indicated that this peak was indeed from

free porphyrin and at the same time demonstrated the presence

Scheme 3. Synthesis of TAPP 3, Zn-TAPP 4, and Fe-TAPP 5

Having obtained the building blocks, the next step was to

synthesize the porphyrin cored hyperbranched polymer. As

previously mentioned, a reversible strategy was used to ensure

the porphyrin core was evenly distributed across all molecular

weights. We have previously used this strategy to synthesize a

Scheme 4. Synthesis of Porphyrin Cored Hyperbranched TAPP-HBP 6 and Metalated Derivatives

C

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

of porphyrin groups in the polymer itself. Unfortunately, due to

insolubility of the porphyrin in methanol, it was not possible to

separate “free” porphyrin from the polymer through

puriﬁcation of the polymer using the established trituration

procedure. Thus, chromatography using an alumina column

was used to separate the free porphyrin from the rest of the

polymer mixture. Subsequent analysis conﬁrmed removal of

free porphyrin. GPC now showed a single broad trace covering

molecular weights greater than that of the free porphyrin, and

the sharp peaks corresponding to unincorporated porphyrin

were no longer present in the ¹H NMR. As well as the

porphyrin resonances in the aromatic region, a strongly

deshielded peak at minus 2.03 ppm was also seen for the NH

protons. Along with a visible Soret band in the UV spectra,²⁰

this is a diagnostic characteristic associated with all porphyr-

ins.²¹The resulting polymer sample possessed a molecular

weight of 11 700 and polydispersity of 1.90. Mass spectrometry

was carried out on the bulk polymer sample, although problems

with disproportionate ionization meant that peaks were not

observed above ∼2000 mass units. The resulting spectrum did

however demonstrate full incorporation of the porphyrin core

for the lower molecular weight fractions. Peaks were observed

at 178 mass unit intervals, with each peak corresponding to

polymers consisting of n monomer residues and a single

porphyrin core (674 + n178). Peaks from cyclized and

“coreless” polymers were not observed. Although mass

spectrometry provided direct evidence for 100% core

incorporation, it was only true for the low molecular weight

fractions (mass spectrometry on the higher molecular weight

fractions was inconclusivesee above). To prove 100%

incorporation across all molecular weights, the polymer needed

to be fractionated into a series of diﬀerent sizes and the level of

core incorporation determined for each (using a non-mass

spectrometric method). Fractionation was also important with

respect to achieving the main aims of this work, speciﬁcally

ensuring a series of narrowly dispersed “pseudo” dendrimers

where potential dense packing properties could be observed.

The polymer was therefore fractionated using preparative GPC

in the form of a column packed with SX-1 Biobeads.²²

Table 1. Minimum Core Incorporation Calculated from

GPC and UV Analysis

a

min

max

fraction M_n(GPC) M_n(UV) minimum core incorporation (%)

bulk

8500

3000

14250

5000

59

60

58

61

62

1

2

3

4

5500

9500

11000

15100

18500

24500

a

min

M_nby GPC is a function of the minimum number of polymer

max

molecules, and M_nby UV is a function of the minimum number

porphyrin units. Thus, dividing M_nby M_ngives the minimum

min

max

level of core incorporation. These are expressed as percentages.

the desired even distribution for the core molecule.

Furthermore, having already proven 100% incorporation for

the low molecular weight fractions (via mass spectrometry), it

follows that the bulk polymer sample must also possess a core

incorporation approaching 100%.

The two protons meta to the carboxyl function exist in a

number of diﬀerent environments and resonate as a series of

1

signals between 8.03 and 7.83 ppm in the H NMR spectrum.

There are three well-deﬁned peaks at 7.50, 7.35, and 7.20 ppm

corresponding to the para protons, which are present in three

distinctive environments. These are the dendritic, linear, and

terminal repeat units; integrating these peaks and applying

́

Frechet’s equation returned a value of 50% for the degree of

branching. After polymerization the porphyrin’s pyrrolic

resonance could be seen as a singlet at 8.89 ppm. Although

only slightly shifted from the position of the same proton peak

in the starting material, a split or unsymmetrical peak pattern

was not observed. A similar situation was also observed for the

porphyrin’s aromatic peak at 8.24 ppm. As such, there is no

clear evidence for unreacted acetoxy groups (on the porphyrin)

and conﬁrms that all four of the porphyrin’s acetoxy groups

have reacted with monomer. However, the degree of

polymerization from each site will vary dramatically, and a

variety of structures will exist. If we consider two extremes, they

are: (a) A polymer with four large and roughly equal HBP

sections (one from each of the pophyrin’s acetoxy groups).

This structure is similar to that of a core functionalized

dendrimer. (b) A polymer with one very large and dominant

HBP section (with a degree of polymerization equivalent to 4

times that observed for each HBP section in (a) above). This

structure is more like that observed for dendrons (Figure 1).

The extent of core incorporation was quantiﬁed by

comparing the polymer’s molecular weight determined using

the polymer’s bulk property (i.e., GPC) with the molecular

weight obtained from a core group analysis (i.e., UV of the

porphyrin core). Because of the way these values are calculated,

it is unlikely that these two values will be identical. For example,

it is well-known²³that GPC calibrated using linear standards

underestimates molecular weight of a globular hyperbranched

polymer. As such, the molecular weight cannot be lower than

.

¹⁵On

min

that obtained using GPC and can be referred to as M_n

the other hand, it may be possible that molecular weights

determined using a core unit analysis may be overestimated.

This is due to the assumption that ALL polymer molecules will

possess a core unit. If this is not the case, then polymers

without core units will eﬀectively contaminate the mixture and

increase the relative proportion of repeat units relative to core

units. Therefore, in the case of core analysis the molecular

weights obtained represent a maximum possible value and are

Figure 1. For a single molecular weight there are a number of possible

structures with respect to polymerization. Two extreme structures are

shown above: (left) a tetrafunctionalized polymer possessing four

roughly equivalent HBP sections; (right) a monofunctionalized

polymers with a large HBP section (di- and trisubstituted structures

as well as a number of other nonsymmetrical structures are also

possible).

15

max

referred to as M_n

.

The data obtained for all polymer

fractions are shown in Table 1. The apparent level of

incorporation for each fraction can be obtained by dividing

M_n^minby M_n^max; the results are also shown in Table 1. The ratio

of the two molecular weights indicates a level of incorporation

around 60% and is consistent across all fractions, demonstrating

D

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Nevertheless, due to the same back folding properties observed

for dendrons,²⁴the porphyrin units will remain encapsulated by

the polymer matrix, and a very similar environment will exist

for all possible structures (of similar molecular weight).

as the ﬁrst was not completely reliable or reproducible

(extremely dependent on vacuum and temperature); fractiona-

tion would also reduce the polydispersity. Therefore, the bulk

hyperbranched polymer was resynthesized as described above

using a 40:1 molar ratio of AB₂monomer 1 and the porphyrin

unit 3. Zinc was inserted into the porphryin core using excess

zinc acetate at room temperature in DCM. Puriﬁcation was

achieved by ﬁltration to remove unreacted zinc acetate followed

by reducing the volume of solvent by rotary evaporation before

trituration into excess methanol to yield the zinc inserted

hyperbranched polymer Zn-TAPP-HBP 7 in an overall yield of

62% by mass (Scheme 4). Conﬁrmation of successful insertion

was provided by ¹H NMR, in which the highly shielded peak at

minus 2.03 ppm was absent; this corresponds to the inner

protons that are removed upon metal insertion. Further

conﬁrmation of successful functionalization was provided by

UV/vis spectrometry in which the four Q-bands of the starting

material at 515, 548, 592, and 648 nm were replaced by two

resonances at 548 and 589 nm. For control experiments zinc

was also inserted into the free base porphyrin 3 yielding the

zinc porphyrin Zn-TAPP 4, using a similar procedure (Scheme

3).

Although it was not possible to control the molecular weight

precisely via the core/monomer ratio, it was possible to

inﬂuence the molecular weight by altering this ratio in favor of

whatever product was desired. As described above, when using

a 20:1 ratio of AB₂monomer relative to the porphyrin core a

polymer with an M_nof approximately 8500 was obtained. If the

ratio was doubled (to 40:1), then a polymer with roughly

double the molecular weight was obtained (M_nof around 18

500). While this ratio could be used to favor a certain

approximate molecular weight, the temperature, reaction time,

and the quality of vacuum were paramount. For example, if the

vacuum was poor, a relatively small polymer would be

produced regardless of the monomer/core ratio. This was

exempliﬁed when a polymer with a very low M_nvalue of 2000

(and a PD of around 4.0) was obtained when using a 20:1 ratio.

This was much smaller than polymers obtained from similar

experiments (using the same 20:1 ratio) and was probably due

to a poor vacuum resulting in ineﬃcient removal of the acetic

acid byproduct. However, the product was returned to a

reaction ﬂask containing solvent (diphenyl ether), and the

mixture was heated back up to 180 °C and a vacuum of 5

mmHg applied. When the product was analyzed, an M_nvalue

similar to that recorded from previous experiments was

obtained (around 9000). These experiments conﬁrm the

reversible nature of the reaction and how this can be used to

generate hyperbranched polymers that possess speciﬁc and

functional units with very high levels of incorporation. As

described above, this is a key requirement when comparing the

properties of hyperbranched polymers with those of

dendrimers.

Prior to performing the UV/vis titrations, it was necessary to

fractionate and purify the zinc functionalized polymer.

Preparative size exclusion chromatography was employed in

the form of a glass column (3 cm diameter) using SX-1

Biobeads and DCM as the eluent. This process allowed (500

mg of) the polydisperse bulk hyperbranched polymer to be

separated into more discrete molecular weights with lower

polydispersities, and the data are shown in Table 2.

Table 2. Pseudo-Generational Series of the Zn-TAPP-HBP

a

7

theoretical values for

equivalent dendrimer

experimentally determined values for

fractionated Zn-TAPP-HBP-5

Probing Dense Packing of the Zn-TAPP-HBP through

Ligand Binding Studies. After conﬁrming the reversibility

and high/even levels of core incorporation of the hyper-

branched polymer system, investigations were conducted to

determine any site isolation or dense packing properties of the

resulting polymers. Speciﬁcally we wanted to determine

whether there was a sudden break in binding properties,

similar to that observed for dendrimers at their dense packed or

dense shell limit,³or a linear or nonspeciﬁc relationship

between binding and molecular weight, as is the case for simple

linear polymers.²⁵The presence of a porphyrin at the center

allows for direct monitoring of access to the core through UV

binding experiments. This can be achieved using zinc

metalloporphyrin cored polymers and studying the binding of

various nitrogen ligands, enabling any trends to be identiﬁed

relative to ligand size. In addition, analysis of binding aﬃnity to

a particular polymer fraction with respect to the diﬀerent sized

ligands would indicate whether the hyperbranched system was

capable of excluding larger ligands while allowing access to

smaller substratesa requirement for future size selective

catalysis experiments. However, the ﬁrst requirement was to

obtain a pseudo-generational series of hyperbranched polymers

possessing a zinc metalloporphyrin core. One approach would

be to carry out a number of polymerizations using an increasing

monomer to core ratio. Alternatively, a bulk metalloporphyrin

polymer could be synthesized and then fractionated using the

chromatographic method described above and separate the

polymer into the required pseudo-generational series of

hyperbranched polymers. We opted for this second method

MW

repeat

(GPC-

repeat

units

pseudo-

MW

units generation fraction

M_n)

generation

1450

2874

4

12

1.0

2.0

3.0

4.0

5.0

1

2

3

4

5

6

7

1650

2150

5

8

1.0

1.5

3.0

3.5

4.0

4.5

5.0

5722

28

5900

29

45

62

75

91

11418

23522

60

8750

128

11500

13900

16800

a

Half-generations were assigned to hyperbranched polymers when the

number of terminal groups and the molecular weights fell between

those for the full generations. Values for the equivalent dendrimer are

also shown for comparison.

It is clear from the table that early generations are very close

in molecular weight; this placed a limit on the fractionation

procedure. The narrow molecular weight divergence at low

generations is irrelevant for quantized dendrimers but does

create a diﬃculty for nonquantized poly dispersed hyper-

branched polymers. By taking the molecular weight and

estimating the number of repeat units for each fraction, we

were able to obtain a series of pseudo-generations. These

values, along with those corresponding to the equivalent

dendrimers, are shown in Table 2. Inevitably, overlap exists for

some of the lower molecular weight fractions due to the similar

molecular weight values moving from one generation to

another (this is less of a problem for higher generations as

E

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

the divergence of molecular weights becomes more marked

between each generation). Consequently, a judgment was made

as to which pseudo-generation each fraction belonged. This

dictated that some fractions are not included in the data set,

and fractions with ambiguous molecular weights were not used;

only fractions with molecular weights corresponding closely to

their pseudo-generation are included. Half-generations were

assigned to hyperbranched polymers when the number of

repeat units and the molecular weights were between the

theoretical full generations (G1.5, G3.5, and G4.5).

Once the zinc inserted polymer had been fractionated and

characterized, the next step was to probe whether or not this

polymeric system was capable of dense packing, core isolation,

and size selectivity. Therefore, binding experiments were

carried out in the form of UV/vis titrations using pyridine

and other pyridine derivatives.²⁶Binding experiment would be

carried out using the fractionated hyperbranched polymers and

a series of ligands with diﬀerent molecular volumes or size. If

dense packing was observed at diﬀerent molecular weights for

this series of ligands, this could indicate a selective capability of

these hyperbranched polymers with respect to size. The speciﬁc

pyridyl ligands used must diﬀer in size and must be free of

groups in the ortho positions which may block the interaction.

Based on these criteria, the following ligands were chosen:

pyridine, 3,5-lutidine, and 3-phenylpyridine. These are shown

along with their space-ﬁlling models in Figure 2. The titrations

Figure 3. Plots showing how pyridine binding aﬃnity changes with

respect to polymer molecular weight for the Zn-TAPP-HBP 7. All data

normalized for binding to the core porphyrin (Zn-TAPP 4) and also

normalized relative to each ligand. (For reference, the K_avalues were

8600, 11 400, and 4950 M⁻¹for pyridine, 3,5-lutidine, and 3-

phenylpyridine, respectively.)

a superior polar environment, with the additional support of

favorable π−π stacking interactions,²⁷creating a partition eﬀect.

A similar eﬀect on microenvironment has been reported for

polyether dendrimers.²⁸The pattern is the same for all ligands;

at an M_nvalue of around 8500 the binding constants become

smaller and continue to decrease in a linear fashion as the

polymer’s molecular weight increases. This is clear and obvious

evidence of dense packing. The molecular weight at which

dense packing occurs can be determined by placing a best ﬁt

line through the two obvious linear regions of each plot. The

molecular weight at which they cross is the dense packing limit.

For all ligands the dense packed limit occurs at an M_nvalue of

around 7200 ( 500) or between G3.0 and G3.5. Interestingly,

this is the same point at which dense packing often occurs for

dendrimers.²⁹However, due to the polydispersity of hyper-

branched polymers, any dense packed limit is likely to occur

over a much broader range than that observed for the quantized

dendrimers. For hyperbranched polymers it is probably more

appropriate to describe dense packing limits with respect to a

molecular weight range. Therefore, in the case of binding to our

hyperbranched polymer ZnTAPP-HBP 7, dense packing occurs

between molecular weights (M_n) of 6000 and 8000 (for the

ligands studied). It is interesting to note that despite the

increase in size for the largest ligand (3-phenylpyridine), the

dense packed limit is the same. This initially seemed to suggest

that shape or size selectivity does not take place. However, if

the binding is studied in more detail, it is noticeable that K_a

does not plummet to zero and reasonable binding was possible

for all ligands across the molecular weight range studied.

However, it is also noticeable that the rate of change in K_awith

respect to molecular weight is greater for the larger 3-

phenylpyridine ligand. Speciﬁcally, when comparing K_afor

the porphyrin core with that obtained for the pseudo-fourth-

generation polymer studied, we observe a 11−12% reduction in

binding for the pyridine and 3,5-lutidine ligands. This contrasts

to a much larger 29% reduction in binding strength for the 3-

phenylpyridine ligand, which is signiﬁcant and does suggest an

Figure 2. Pyridyl ligands used to probe steric, electronic, and dense

packing properties of the porphyrin HBPs. These ligands diﬀerence in

size, electronics, and hydrophobicity.

were performed using a dichloromethane solution of Zn-TAPP-

HBP 7 and titrating in 10−20 μL aliquots of a pyridyl ligand

solution and recording a UV/vis spectrum after each addition.

The changes in absorbance for the bound peak were plotted

against the concentration of ligand added, and the binding

constant, K_a, was determined by ﬁtting the data points to a 1:1

binding isotherm.²⁶A number of repeats were carried out for

each M_n(pseudo-generation), and a value for K_awas

determined for each. These were then averaged with respect

to molecular weight. Plotting the results with respect to

generation would involve the use of a nonlinear scale in the x-

axis, with equal distances between each generation. However,

when moving across a generational series the molecular weight

diﬀerences get larger as the generations get higher and a

generational plot would not necessarily provide an accurate

picture of events. Therefore, our binding results are plotted

with respect to M_n(Figure 3).

This plot shows a small increase in the binding constant from

the free porphyrin core (eﬀectively G0) to the G3 polymer with

molecular weight 5900. Although this could be due to error, it

is also possible that the polymer is starting to generate a good

environment for the pyridine ligands, presumably by providing

F

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Table 3. Binding Data of Pyridyl Ligands to Zn-TAPP-HBP 7

a

ligand

pyridine

K_aporphyrin core (M⁻¹

)

K_aof HBP G4 (M_n∼ 11 500) ( M⁻¹

)

onset of dense packing (M_n)

reduction in binding (%)

8600 ( 400)

11400 ( 850)

4950 ( 350)

4700 ( 200)

6900 ( 500)

1000 ( 75)

7200 ( 500)

7100 ( 500)

7200 ( 500)

11 ( 5)

12 ( 7.5)

29 ( 3)

3,5-lutidine

3-phenylpyridine

a

Calculated by comparing the K_avalues obtained from the porphyrin core unit (Zn-TAPP 4) and the pseudo-fourth-generation Zn-TAPP-HBP 7

(M_n11 500).

element of shape and/or size selectivity with respect to ligand

size and the polymer’s molecular weight; the data are

summarized in Table 3.

substrate 9 to occur equally well with a catalytically

functionalized polymer above or below its dense packed limit.

Conversely, we might predict that the larger cyclic alkene

substrate 11 would struggle to reach the catalytic core of a

hyperbranched polymer whose molecular weight was above the

dense packed limit, resulting in a slower reaction and a lower

yield. We proposed to study these reactions using the same

method reported by Suslick and Moore,⁷speciﬁcally using

iodosylbenzene as the oxygen source and a metalloporphyrin as

the catalyst. Although zinc porphyrins were used to study and

probe binding, they are not suitable as oxidation catalysts.

There are a number of suitable metals that can be inserted into

a porphyrin;³⁰in this case iron (Fe(III)) was used as the

products are robust and make good oxidation catalysts.³¹We

therefore set about synthesizing Fe(III) metalloporphyrin³²

cored hyperbranched polymers above and below the dense

packed limit as well as aiming for a polymer with a molecular

weight close to the dense packed limit. For control and

reference experiments we would also required the simple

Fe(III) metalloporphyrin 5, which was synthesized as shown in

Scheme 3.

Insertion of iron into free base porphyrin 3 was achieved by

reaction with a large excess of iron(II) chloride in the presence

of 2,6-lutidine.³³After reﬂuxing for 4 h under nitrogen, the

mixture was exposed to air as it cooled, oxidizing the Fe(II) to

Fe(III). After work-up and column chromatography, the mass

spectra showed a peak corresponding to the target molecule,

but also a series of peaks above the molecular ion. This reaction

was repeated a number of times, but on every occasion an

identical mass spectrum was obtained. It was assumed the extra

peaks were due to an unknown species strongly ligated to the

metal center. Therefore, the porphyrin was dissolved in

dichloromethane and washed with 1 M hydrochloric acid.

The organic phase was collected, the solvent removed and the

product reanalyzed. On this occasion the EI mass spectrum

contained a single dominant peak of the molecular ion and no

peaks of a higher value. Further authentication for the product

Fe-TAPP 5 was provided by UV/vis analysis, which showed a

reduction in the number of Q-bands from four to two, a

diagnostic change for a metal inserted porphyrin. Because of

the paramagnetic iron, substantial broadening of the peaks

occurred in the ¹H NMR spectrum.³⁴With the porphyrin itself

metalated, it was also necessary to insert iron into the

porphyrin cored hyperbranched polymer. The polymerization

was repeated using the unmetalated porphyrin and metal

insertion performed after puriﬁcation (Scheme 4). The

synthesis was performed in a similar manner to that described

above for the free porphyrin, using iron(II) chloride and 2,6-

lutidine; however, after reﬂuxing TAPP-HBP 6 for 4 h the

solution was ﬁltered through a short plug of silica to remove

unreacted iron chloride, washed with dilute HCl, and

precipitated into excess methanol to give Fe-TAPP-HBP 8 in

good yield. GPC analysis carried out after the acid wash

returned an M_nvalue of 12 100, which was almost identical to

Probing Shape and Size Selectivity of the Fe-TAPP-

HBP via Catalysis. The binding data presented above

conﬁrmed the porphyrin cored hyperbranched polymers

possess a dense packed limit at a molecular weight around

8000. The data also suggest the polymers were capable of size

selective binding. That is, at a molecular weight below 8000 a

relatively open structure persists, and this allows unhindered

access to the core, whereas at higher molecular weights (above

8000) the polymer possesses a tighter more compact structure

and access to the core is restricted. This eﬀect is more

pronounced for the bigger polymers and larger ligands.

Therefore, assuming all other things are equal, it should be

possible for smaller substrates to access the core in preference

to larger substrates. We now wish to exploit this eﬀect and

study the possibility of size selective catalysis. Suslick and

Moore demonstrated the selective capability of similar

dendrimers using a mixture of linear and cyclic octenes.¹⁹In

an attempt to make a valid comparison between dendrimers

and hyperbranched polymers, we decided to study the same

catalytic reaction (Scheme 5), that is, the catalytic epoxidation

of a linear alkene 9 and a cyclic alkene 11, in isolation and

together in equal ratios.

Scheme 5. Iron Porphyrin Catalysed Oxidation of Small/

Thin Alkene 9 and a Larger/Fat Alkene 11

The speciﬁc substrates selected were cyclooctene, which is a

relatively large substrate, and 1-octene, which is a small (thin)

linear alkene. The linear substrate has a terminal alkene, which

is relatively electron poor and therefore much less reactive than

the cyclic alkene, which is more electrophilic. This increase is

due to the higher level of substitution and increased sigma

donation, leading to a more electron-rich alkene compared to

the equivalent linear system. This is an ideal scenario when

trying to study the eﬀect of size on reactivity, speciﬁcally when

trying to encourage the reaction between a catalyst and a small

but unreactive substrate, in the presence of a more reactive but

larger substrate. To test this, reactions were carried out on both

substrates using catalytically functionalized hyperbranched

polymers below and above their dense packed limit. If a size/

shape selective reaction is possible, we would expect any

reaction using the smaller (but unreactive) linear alkene

G

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

the M_nof TAPP-HBP 6 (11 700), conﬁrming that the structure

remained intact during metalation. As before, UV/vis analysis

showed that the number of Q-bands had been reduced from

four to two, conﬁrming iron insertion. Because of the size of the

polymer, which eﬀectively shields the central iron, line

Now in a position to study the catalytic properties of the

hyperbranched polymers, the individual oxidation of each

alkene with each polymer fraction (or porphyrin core) was

studied. The reaction for the linear alkene 9 is shown in

Scheme 7. It is important to ensure that all catalyst/porphyrin

concentrations are equivalent when studying a series of

polymers. Therefore, the concentration of all solutions with

respect to the amount of porphyrin was checked by UV

spectroscopy and adjusted if necessary, to ensure equivalent

porphyrin concentrations for all polymer fractions (see

Experimental section in the Supporting Information). As well

as the dominant epoxide products, a number of other oxidation

products can also form, either directly or after initial formation

of the epoxide. Therefore, the reactions were monitored by

following the reduction product, that is, the conversion of

iodosyl benzene to iodobenzene (which is a measure of all

oxidation reactions catalyzed by the porphyrin).³⁶A second

series of experiments would study the product yields when each

polymer (and core) was reacted with an equimolar mixture of

both alkenes. This set of experiments would indicate whether or

not the hyperbranched polymers could inﬂuence a shape/size

selective reaction. Gas chromatography was used to quantify

the outcome of the catalyzed oxidation reactions by monitoring

the epoxide and the iodobenzene; however, it was ﬁrst

necessary to calibrate the instrument using the alkene starting

materials, the epoxide products, the iodobenzene side product,

and an internal standard (n-dodecane). An initial blank

experiment indicated that iodosylbenzene reacted with the

porphyrin in the absence of alkene (producing iodobenzene).

However, it has been shown that if a large excess of alkene is

used, then this side reaction is suppressed (eﬀectively saturating

the rate of alkene oxidation).³⁷A second blank experiment was

carried out to ascertain if the alkenes could react with the

oxygen source in the absence of porphyrin catalyst. The GC

results showed that although oxidation took place, the yields

were extremely low and not at a level that would eﬀect the

results from the catalytic experiments.

1

broadening in the H NMR was signiﬁcantly reduced, and

some structural information remainedthe most important of

which was the disappearance of the free base porphyrin NH

peak at −2.03 ppm, conﬁrming metal insertion.

The bulk Fe-TAPP-HBP 8 was fractionated using a Biobead

column, and the molecular weights of each fraction were

analyzed by GPC. Fractions were required with Mn values

above and below the dense packed limit, as well as one with an

Mn reasonably close to the dense packed limit (previously

shown to be somewhere between the G3.0 and G3.5 pseudo

generation). Table 4 below shows the fractions selected for the

Table 4. Theoretical Values for Pseudo-Generations vs the

Values for Obtained Fractions of Fe-TAPP-HBP 8

a

molecular weight

theoretical obtained

repeat units

theoretical

pseudo-generation

obtained

0 (Fe-TAPP)

901

3038

901

3200

0

12

28

60

0

13

25

63

2

3

4

5888

5400

11588

12100

a

Rounded to the nearest integer.

catalytic experiments along with data regarding the number of

repeat units and the corresponding pseudo generation. The

core unit would be used as reference and control catalyst during

the future epoxidation and catalytic experiments and is also

included in the table. Iodosylbenzene 13, was used as the

oxygen source for the catalytic reactions and was synthesized as

shown in Scheme 6. The synthesis is trivial and iodosylbenzene

Scheme 6. Synthesis of Iodosylbenzene 13

With the instrument calibrated and the blank experiments

performed, investigation into the activity of the porphyrin

catalyst could begin. The yield of oxidation products when

using just the core porphyrin unit (i.e., pseudo-generation G =

0) were 74% and 96% for the linear and cyclic alkenes,

respectively.³⁸This indicates that the cyclic alkene is more

reactive, which is consistent with previous results and the

inherent diﬀerences in reactivity between a mono- and

disubstituted alkene.³⁹To make comparisons easier, the yields

of the linear and cyclic species are normalized relative to the

value obtained using the reference porphyrin (Fe-TAPP 5, i.e.,

the G = 0 pseudo-generation) and plotted against pseudo-

generation (Figure 4).⁴⁰

was prepared by the stirring iodobenzene diacetate in 3 M

sodium hydroxide for 45 min.³⁵The solid obtained was

collected and thoroughly washed with water followed by

chloroform, if left to stand for a few days the product can

decompose (and is particularly sensitive to prolonged exposure

to light). However, as the product was easy to synthesis and

characterize, a fresh batch was prepared each time a series of

catalytic and epoxidation reactions were performed (a control

reaction was always carried out to test the activity of the oxygen

source).

The results for the linear alkene show that the activity of G2,

G3, and G4 remained similar to that of the free porphyrin G0

(within the 10% error). However, for the cyclic alkene, there

appears to be a steady decrease in yield and therefore activity.

Scheme 7. Oxidation of Alkene 9 to the Epoxide 10 and Other Oxidation Products

H

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

no shape or size selectivity, whereas values higher than unity

indicate a decreased ratio and therefore higher selectivity

toward the less reactive linear alkene.

It is clear from the graph that no selectivity was observed,

and the hyperbranched polymer system displayed selectivity

similar to that of the free porphyrin for all molecular weights.

The free porphyrin, G0, displayed a preference for cyclooctene;

this corresponds directly with the relative reactivity of the two

alkenes. This demonstrates that although able to catalyze the

reaction, the product ratio remains in line with the relative

reactivity of the two alkenes. This suggests that the catalytic

porphyrin core is not sterically crowded enough to prevent

access of the larger substrate and slow down the reaction. This

is despite the fact that binding studies clearly indicated size

selectivity. Therefore, when considering the catalytic and

binding studies together, we can conclude that although a

steric barrier is generated, it does not prevent the substrate

reaching the catalytic core and reacting over the time scale of

the experiment. To slow down substrate access, an increased

steric barrier is required. Clearly the problem is not completely

related to the polymers molecular weight, and simply making

the polymer bigger will not necessarily constitute a solution.

An alternative way to increase sterics around the porphyrin

core is to modify the core branching architecture. In this case

the polymer is initiated from the 4-position of the porphyrin’s

phenyl ring, and polymerization occurs away from the core.

This initiation motif unavoidably leaves a volume of space

surrounding the porphyrin core, despite the occurrence of

dense packing. Therefore, sterics may not be the only issue but

also the free volume surrounding the active site. As such, it may

be more eﬀective to consider steric crowding around the core

as well as steric eﬀects related to polymer size. To do this would

require the initial position of polymeric growth to be less

remote from the porphyrin core but still be symmetrically

arranged as for TAPP 3. This is shown schematically in Figure

6, where it can be seen how the polymer grows away from the

porphyrin when initiated from the single 4-position. However,

an increased steric environment is possible if the initiation or

growth point is moved one substitution position closer to the

porphyrin. This could be achieved symmetrically if the polymer

backbone started from the 3- and 5-positions on the phenyl

Figure 4. Relative yields of all oxidation products obtained from the

oxidation of 1-octene (9) and cyclooctene (11) using catalytic

experiments using Fe-TAPP-HBP 8 as the catalyst. Yields quoted

relative to those obtained using the G = 0 pseudo-dendrimer (which

were 74% and 96%, respectively).

For the larger G4 pseudo-generation there is an approximate

24% reduction in yield. As such, it suggests that there may be

some element of selectivity when these oxidations are catalyzed

by this polymer system. To conclusively determine whether or

not selectivity was possible, a mixed experiment using both

alkenes in a 1:1 ratio was carried out. This experiment was

performed identically to the activity experiments described

above, except for the simultaneous addition of 1 mmol of each

alkene. Because iodobenzene is produced in both reactions (i.e.,

oxidation of linear and cyclic alkenes), the reactions were

followed by measuring the yield of the corresponding linear and

cyclic epoxides, 10 and 12, respectively (Scheme 5). The results

are shown in Figure 5 and the data normalized with respect to

Figure 5. Selectivity of epoxide products from a mixed experiment

using the Fe-TAPP-HBP system 8. Yields normalized with respect to

the inherent reactivity diﬀerence between the linear (terminal) alkene

9 and the more reactive cyclic alkene 11. Therefore, values higher than

1.00 indicate a shift in reactivity toward the less reactive linear alkene.

the free porphyrin as before. In addition, a further normal-

ization was carried out to take into account the inherent

reactivity diﬀerences of the cyclic and linear alkenes. As such,

values obtained for the linear system were normalized to take

into account the increased reactivity of the cyclic alkene.

Therefore, a value of unity represents a 14:1 ratio and indicates

Figure 6. Schematic showing how the point of branching can aﬀect the

steric environment around the polymers central catalytic porphyrin.

Although two HBP phenyl substituents are shown, the polymers are

likely to be a mix of mono-, di-, tri-, and tetrafunctionalized systems

(each possessing one or two HBP arms).

I

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Scheme 8. Synthesis of Tetra(3,5-diacetoxyphenyl)porphyrin 16 and Its Metalated Derivatives 17 and 18

ring. In fact, this is exactly the same substitution pattern used by

Suslick and Moore on their work on porphyrin cored

dendrimers.⁷Thus, it was postulated that this alternative

porphyrin core may be capable of providing the correct steric

environment required for selectivity.

this was achieved using a standard demethylation procedure.³⁹

Speciﬁcally, porphyrin 14 was reacted with an excess of boron

tribromide, followed by water. The product obtained after

puriﬁcation showed that the methoxy peak at 55.6 ppm in ¹³C

NMR was no longer present. Removal of the methoxy group

was conﬁrmed by ¹H NMR, which showed that the

corresponding methoxy signal was absent. In addition, a large

OH peak was observed at 3425 cm⁻¹in the IR spectrum. This

procedure produced the desired tetra(3,5-dihydroxyphenyl)-

porphyrin 15 in high purity and a yield approaching 90%. The

ﬁnal synthetic step was acetylation; this was carried by reﬂuxing

porphyrin 15 in acetic anhydride for 6 h. Excess acetic

anhydride and acetic acid byproduct were removed via vacuum

distillation, and the target tetra(3,5-diacetoxyphenyl)porphyrin

Synthesis of Tetra(3,5-diacetoxyphenyl)porphyrin

Cored Hyperbranched Polymers. The more substituted

target porphyrin 16 could in principle be synthesized in a

manner analogous to that used to construct the simple TAPP

core 3. Unfortunately the required 3,5-diacetoxybenzaldehyde

was not easily available, and its acetylated precursor was

prohibitively expensive. However, 3,5-dimethoxybenzaldehyde

could be used to assemble tetra-3,5-dimethoxyphenylporphyrin

14, which could in turn be demethylated and ﬁnally acetylated

to give the target porphyrin, tetra-3,5-diacetoxyphenylporphyr-

in, tetra-(3,5-DAPP) 16 (Scheme 8). Tetra(3,5-

dimethoxyphenyl)porphyrin 14 was synthesized via the same

Rothemund and Adler−Longo procedure used previously;¹⁸

equal molar quantities of 3,5-dimethoxybenzaldehyde and

pyrrole were reﬂuxed for 30 min and after work-up and

puriﬁcation gave 3,5-dimethoxyphenylporphyrin 14 in 13%

yield. The UV/vis absorption spectrum conﬁrmed the presence

of the porphyrin displaying the characteristic Soret band at

420.5 nm and four Q-bands at 514, 548, 588, and 648.5 nm.

1

16 was obtained in 64% yield after puriﬁcation. H NMR

analysis showed a large singlet corresponding to three

hydrogens from on the newly introduced acetoxy group at

2.42 ppm. Mass spectrometry gave a molecular ion peak at

1079 which corresponds to that required for the target

porphyrin.

The next step was to repeat the polymerization using this

alternate porphyrin core. The reaction is shown in Scheme 9

and was carried out using the same synthetic and puriﬁcation

methods described for TAPP-HBP 6. After precipitation from

tetrahydrofuran and methanol, followed by puriﬁcation using

an alumina column, tetra(3,5-diacetoxyphenyl)porphyrin cored

hyperbranched polymer 19 (tetra-(3,5-DAPP-HBP)) was

recovered in good yield as a red/brown solid. SEC analysis

was carried out using RI and UV detection, with each trace

showing a single peak that was directly super imposable on the

other. This conﬁrms that the porphyrin was incorporated

1

The H NMR spectrum contained peaks corresponding to the

β-pyrrolic protons at 8.96 ppm. The ortho and para aromatic

protons resonated at 7.43 and 6.93 ppm, respectively, and a

large singlet at 3.99 ppm was assigned to the methoxy protons.

Finally, a peak corresponding to the highly shielded inner NH

hydrogens could be seen at minus 2.81 ppm. The second stage

was to demethylate the octamethoxy-functionalized porphyrin;

J

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Scheme 9. Synthesis of 3,5-(Diacetoxyphenyl)porphyrin Cored Hyperbranched Polymer Tetra-(3,5-DAPP)-HBP 19 (left) and a

Representative Hyperbranched Wedge (right)

Table 5. M_n, Polydispersity, and Number of Repeat Units for Hyperbranched Polymer Zn-3,5-DAPP-HBP 20

M_n

polydispersity

no. of repeat units

M_n

polydispersity

no. of repeat units

1142

3100

3700

5000

6100

6800

1.00

1.81

1.82

1.72

3.12

2.74

0 (core unit)

8000

9000

2.46

2.33

2.14

1.99

1.87

40

46

55

61

66

13

16

24

30

34

10500

11800

12500

evenly across all molecular weights.¹⁵Analysis returned an M_n

value of 8100 and a polydispersity of 2.64. ¹H NMR conﬁrmed

porphyrin incorporation, showing the characteristic deshielded

peak from the NH protons at minus 2.03 ppm. As with the 4-

substituted HBP system, both SEC and NMR indicated very

small peaks corresponding to unincorporated porphyrin

(equivalent to just 2% of the total product and is completely

gone after the fractionation step). Mass spectrometry showed a

series of peaks separated by 178 mass units, each corresponding

to n monomer residues and a single porphyrin core (744 +

n178). Peaks from cyclized and “coreless” polymers were not

observed.

reaction was complete after 30 min, after which unreacted zinc

acetate dihydrate was removed by ﬁltration, followed by ﬂash

chromatography to give Zn-tetra-(3,5-DAPP) 17 in 68% yield.

Structural conﬁrmation of Zn-tetra-(3,5-DAPP) 17 was

1

provided by H NMR, in which the highly shielded peak at

minus 2.94 ppm was no longer present and UV/vis analysis,

which showed the number of Q-bands had been reduced from

four to two. Mass spectrometry showed only a single peak of

the molecular ion with M/z of 1141. Metalation of tetra-(3,5-

DAPP)-HBP 19 to give Zn-tetra-(3,5-DAPP)-HBP 20 is shown

in Scheme 9 and was performed in an identical manner to that

described above for the porphyrin core 16, with the exception

that ﬁltration and trituration (from THF into methanol) was

used in place of chromatography for puriﬁcation. UV/vis

spectroscopy showed a strong Soret band at 421.5 nm and that

the number of Q-bands had been reduced from four to two.

Probing Dense Packing of Zn-3,5-DAPP-HBP through

Ligand Binding Studies. To probe binding ﬁrst required zinc

insertion into the synthesized porphryin 16. This was carried

out using zinc acetate dihydrate as shown in Scheme 8. The

K

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

SEC analysis conﬁrmed the hyperbranched polymer remained

intact after the metalation reaction. The polymer was then

fractionated using size exclusion chromatography in the form of

a Biobead column; the results are shown in Table 5. As

described above for the TAPP-HBP system, UV analysis on the

fractions conﬁrmed that the core was evenly distributed across

all molecular weights.¹⁵On this occasion, smaller fractions were

obtained, enabling more data points to be collected. This also

meant that molecular weight diﬀerences between fractions

would be smaller. This allowed a more accurate analysis of

binding data and eﬀects, particularly the onset of dense packing.

As the diﬀerences between fractions was smaller, there was a

large number of fractions between each pseudo-generation, and

there was no longer any value in assigning pseudo-generations

to the fractions (i.e., between pseudo-generation 3.0 and 4.0

there are ﬁve fractions).

The ﬁrst few hyperbranched polymers exhibited an increase in

K_awith respect to the zinc metalated porphyrin core 17. This

increase is again due to the internal surroundings of the

hyperbranched polymer, which generate an optimum polarity

and electronic environment for the ligands, in turn leading to

an increased concentration of ligand and a higher binding

aﬃnity. As the molecular weight increases further, steric

hindrance takes over and starts to overwhelm any positive

electronic eﬀects. The point at which this starts to take is eﬀect

is the dense packed limit for this system and is roughly similar

for each ligand. This point can be determined by plotting best

ﬁt lines through the two distinct (linear) regions of each

binding plot. The dense packed limit can be found where these

line cross and extrapolating to the M_nvalue on the x-axis. The

dense packed limit for each ligand, along with the reduction in

binding aﬃnities are shown in Table 6.

As before, dense packing was probed by studying the binding

of various pyridyl ligands to the fractions of Zn-tetra-(3,5-

DAPP)-HBP 20 (as well as to the core unit Zn-tetra-

(3,5,DAPP) 17). The same three ligands were used, and to

aid comparison, the data for all ligands are plotted on a single

graph. The data have been normalized relative to the binding

constants obtained for the porphyrin core as well as being

normalized relative to each ligand (K_avalues for each ligand are

diﬀerent); the plots are shown in Figure 7.

On this occasion there does seem to be a relationship

between the M_nvalue that dense packing occurs and the size/

shape of the ligand. As the size of the ligand increases, this

corresponds to a decrease in the point at which dense packing

takes place. Admittedly, these diﬀerences are small, and it could

be argued these diﬀerences are within experimental error.

However, the experiments were repeated a number of times,

and the same trend was observed on every occasion. Therefore,

although it is unwise to deﬁne a precise molecular weight to the

dense packing limit, there is a clear trend regarding molecular

weight and ligand size. As observed for the more open Zn-

TAPP-HBP 7, binding is signiﬁcantly reduced for the larger

ligands at higher molecular weight. Furthermore, the extent to

which binding is reduced is much greater than that observed for

the previous HBP system (i.e., a 42% reduction in binding

occurs between the largest ligand and the polymer with M_n11

800). These binding results are encouraging with respect to any

size or shape selective reactions catalyzed at the core of these

polymers.

Probing Shape and Size Selectivity of the Fe-3,5-

DAPP-HBP via Catalysis. The binding data clearly show

access to the core is restricted at a lower molecular weight for

the larger ligands and suggest the 3,5-substituted polymers may

have an ability for size selective catalysis. To this end the

catalytic potential again probed using the alkene oxidation

catalysis as a probe, As before, this requires the free porphyrin

and the polymers to be functionalized with iron and then

fractionated. Inserting iron into the core was achieved by

reﬂuxing the starting material tetra-(3,5-DAPP) 16 with

iron(II) chloride and 2,6-lutidine in THF for 4 h (Scheme

8). Subsequent to an acid work-up and puriﬁcation by column

chromatography, the Fe(III) tetra(3,5-diacetoxyphenyl)-

porphyrin, Fe-tetra-(3,5-DAPP) 18, was obtained in 75%

yield. Mass spectrometry displayed a single peak corresponding

to the molecular weight of the iron inserted porphyrin complex

(MH⁺1133). The UV/vis spectrum showed the Soret band

and two Q bands (at 418, 513, and 562 nm), conﬁrming the

Figure 7. Plot showing ligand binding aﬃnities vs molecular weight for

the Zn-(3,5-DAPP)-HBPs 20. All data normalized for binding to the

core porphyrin (Zn-tetra-(3,5-DAPP) 17) and also normalized relative

to each ligand. (For reference the K_avalues were 14 000, 15 000, and

12 000 M⁻¹for pyridine, 3,5-lutidine, and 3-phenylpyridine,

respectively.)

It is immediately evident the pattern displayed is very similar

to the pattern observed for the Zn-TAPP-HBP 7 system

described earlier, with each ligand having two distinct regions.

Table 6. Binding Data of Pyridyl Ligands to Zn-Tetra-(3,5-DAPP)-HBP 20

a

ligand

pyridine

K_aporphyrin core (M⁻¹

)

K_aof HBP with M_n11 800 (M⁻¹

)

onset of dense packing (M_n)

reduction in binding (%)

14000 ( 1750)

15000 ( 1750)

12000 ( 1500)

10500 ( 1200)

9000 ( 1000)

4200 ( 500)

7800 ( 500)

6800 ( 500)

6200 ( 500)

25 ( 5)

21 ( 5)

42 ( 5)

3,5-lutidine

3-phenylpyridine

a

Calculated by comparing the K_avalues obtained from the porphyrin core unit 16 and the Zn-3,5-DAPP-HBP fraction with M_nclosest to 11 500

(allowing comparison with the pseudo-fourth-generation Zn-TAPP-HBP system discussed above).

L

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

iron insertion. Insertion of iron into the bulk hyperbranched

polymer 19 was carried out in analogous fashion, with

puriﬁcation achieved by trituration from THF into cold

methanol, giving Fe(III) tetra(3,5-diacetoxyphenyl)porphyrin

cored hyperbranched polymer 21 (Fe-(3,5-DAPP)-HBP) as

shown in Scheme 9. Once more, as for Fe-TAPP-HBP 8, line

1

broadening in the H NMR was reduced relative to that of the

free porphyrin alone. Crucially, the spectrum showed the

disappearance of the highly shielded inner porphyrin protons,

conﬁrming the metal insertion was a success. Further

conﬁrmation was provided by UV/vis spectrometry which

showed a reduction from four Q-bands at 514.5, 548.5, 587.5,

and 645 nm to two at 511 and 569.5 nm. SEC analysis of the

functionalized polymer conﬁrmed that it had remained intact

during the procedure (M_nremained at 8100). The bulk

polymer 21 was then fractionated as previously described, and

the molecular weights were determined by GPC analysis.

Although suﬃcient material for the catalysis study (plus

repeats) was required, it was attempted to maximize the

number of data points by collecting relatively small fractions.

The molecular weight obtained for each fraction along with the

theoretical values calculated for the equivalent dendrimers are

shown in Table 7. The data show good agreement between the

Figure 8. Relative yields of all oxidation products obtained from the

oxidation of 1-octene 9 and cyclooctene 11 using catalytic experiments

using Fe-3,5-DAPP-HBP 21 as the catalyst. Yields quoted relative to

those obtained using the G = 0 pseudo-dendrimer (which were 76%

and 90%, respectively).

of the larger cyclic alkene 11 shows that there is initially a small

decrease in activity for each polymer as size is increased.

However, there is a notable decrease in yield for the G3.5 and

G4 polymers, where the activity drops to 76% and 72% of the

value obtained for the core unit. Therefore, in both cases a

point is reached when access to the catalytic core becomes

more hindered, either through dense packing of the polymer

backbone/shell or dense packing speciﬁcally around the central

porphyrin core. Signiﬁcantly, this point occurs a generation

earlier when the larger alkene substrate 11 is used. Overall, the

catalytic data concur well with the binding evidence that

showed access to the porphyrin core is being restricted as the

pseudo-generation increases, and this restriction is more

pronounced for larger species. This was once again encouraging

with regard to possible shape and size selectivity reactions.

To probe the possibility of a selective reaction, a mixed

experiment was carried out using a 1:1 mixture of the linear

alkene 9 and the cyclic alkene 11. This experiment was

performed as described for the Fe-TAPP-HBP system

described above, and once again the reactions were followed

by measuring the yield of epoxides, 10 and 12 (Scheme 5). The

data are again normalized with respect to core porphyrin (Fe-

tetra-(3,5-DAPP) 18) and the reactivity diﬀerence between the

linear and cyclic alkenes taken into account as before (Figure

9). Although the data appear encouraging and the graph does

indicate a very small shift in selectivity (toward the less reactive

linear species), the change is less than 10%; therefore, once

again it may be argued the data are within the error limits of the

experiment. It had been hoped that moving the growth/

initiation points to the 3,5-position on the phenyl rings on the

porphryin core would reduce the space around the porphyrin

core suﬃciently to achieve selectivity, and with respect to the

zinc metalated polymers and ligand binding this was indeed the

case. However, no deﬁnitive shape or size selectivity could be

observed when using the iron metalated polymers to

simultaneously catalyze the epoxidation of the two alkenes.

Although only a modest change in selectivity was observed

when the initiation/growth points were moved to the 3,5-

positions, the results are an improvement over those obtained

using the 4-substituted system studied earlier. In a ﬁnal eﬀort to

improve selectivity, we proposed to move the initiation/growth

Table 7. Theoretical Values for Pseudo-Generations vs the

Values Obtained for Fractions of Fe(III) 3,5-DAPP-HBP

a

21

a

molecular weight

theoretical obtained

repeat units

theoretical

pseudo-generation

obtained

0 (Fe-HBP-17)

1133

3983

1133

3500

0

15

0

13

1.5

2.5

3.0

3.5

4.0

8271

7750

37

11134

16862

22590

11500

16500

21500

56

59

83

86

120

117

a

Half-generations were assigned to hyperbranched polymers when the

number of terminal groups and the molecular weights fell between

those for the full generations.

values obtained for the fractionated polymer and the theoretical

values. The obtained molecular weights are all close enough to

ensure valid comparison using pseudo-generations.

The catalytic reactions were carried out using the same

method and concentrations described for Fe-TAPP-HBP 8,

again using UV spectroscopy to ensure that the porphyrin

concentrations were constant for all polymer fractions tested.

The experiments independently studied the eﬀect of catalyst

size on the oxidation reactions of the linear alkene 9 and the

cyclic alkene 11 (Figure 8). The yields of all oxidation products

for each polymer catalyst are normalized with respect to the

porphyrin core, Fe-tetra-(3,5-DAPP) 18. As expected, the

porphyrin cored hyperbranched system was capable of

catalyzing the oxidation of both alkenes. For the linear alkene

9 the activity of G1 to G3.5 never drops below 90% of the value

obtained for the core unit. This is within the 5% error limits,

and we can conclude that the reaction remains fairly constant

for these polymeric catalysts. However, there is a decrease in

activity for the larger G4.0 polymer, where the recorded yield

has dropped to around 82% of the value obtained for the core

unit. Therefore, it can be concluded that the catalytic reaction

involving the smaller linear alkene 9 occurs without hindrance

until the largest polymer is used (M_n> 21 000). The oxidation

M

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

using an analogous method to that of the substituted 3,5-

substituted diacetoxyphenylporphyrin, tetra(3,5-DAPP) 16.

This dictated that the synthesis must begin with 2,6-

dimethoxyporphyrin, which was synthesized from the corre-

sponding 2,6-dimethoxybenzaldehyde, which, in turn was

obtained by reacting 1,3-dimethoxybenzene with butyllithium

and dimethylformamide (Scheme 10).⁴⁰After puriﬁcation, 2,6-

Scheme 10. Synthesis of 2,6-Dimethoxybenzaldehyde 22

dimethoxybenzaldehyde 22 was isolated as a pale yellow solid

in a 90% yield. The ¹H NMR spectrum shows a highly

deshielded peak corresponding to the aldehyde proton at 10.46

ppm. The three aromatic protons are also easily distinguished

in the spectrum in the 6.55−7.44 ppm range as well as a large

singlet at 3.88 ppm corresponding to the six protons from the

methoxy hydrogens.

Figure 9. Selectivity of epoxide products from a mixed experiment

using the Fe-tetra-(3,5-DAPP)-HBP 21. Yields normalized with

respect to the inherent reactivity diﬀerence between the linear

(terminal) alkene 9 and the more reactive cyclic alkene 11. Therefore,

values higher than 1.00 indicate a shift in reactivity toward the less

reactive linear alkene.

Although it was possible to synthesize the 2,6-substituted

porphyrin using the procedure described previously, the yield

was very low (<1%). In an eﬀort to improve the yield, the

synthesis was repeated using the low concentration approach

developed by Lindsey.⁴⁰This involved the use of the Lewis acid

boron triﬂuoride diethyl etherate to synthesize a porphyri-

nogen, followed by oxidation with 2,3-dichloro-5,6-dicyano-1,4-

benzoquinone (DDQ) as shown in Scheme 11. After column

chromatography the desired porphyrin was isolated in 9% as a

points one position further toward the porphyrin ring to the 2-

and 6-positions. This substitution motif places the acetoxy

functionalities directly above and below the plane of the

porphyrin. As such, a steric environment around the catalytic

core would develop as soon as the polymer started to grow.

This is shown schematically in Figure 10.

Synthesis of Metalated Tetra(2,6-diacetoxyphenyl)-

porphyrin. The 2,6-substituted porphyrin 25 was synthesized

1

purple crystalline solid (Scheme 11). The H NMR spectrum

of 23 contained a singlet at 8.65 ppm, from the β-pyrrolic

protons, a triplet at 7.66 ppm from the para-aromatic hydrogen,

a doublet at 6.94 ppm from the meta-aromatic protons, a large

singlet at 3.47 ppm from the methoxy protons, and a shielded

peak at minus 2.52 ppm from the inner NH protons. This

combined with the UV/vis spectrum showing four Q-bands at

513, 543.5, 589, and 650.5 nm and a Soret band at 417.5 nm

conﬁrmed the synthesis was a success. The methyl groups of 23

were removed using boron tribromide (as previously described

for 15) to give the 2,6-dihydroxy derivative 24.⁴¹Successful

deprotection was conﬁrmed by ¹³C NMR, which showed loss

of the methoxy peak at 56.1 ppm, and by mass spectroscopy,

which indicated a molecular ion at 855 Da.

The introduction of the required acetoxy groups was carried

out using the same method described above for the 3,5-

functionalized system. Puriﬁcation using column chromatog-

raphy gave the porphyrin 25 in 65% yield. Mass spectrometry

indicated a molecular ion at 1079 Da, providing strong evidence

that the product had been formed. The product’s structure was

ﬁnally conﬁrmed by ¹H NMR which displayed the characteristic

pyrrolic and shielded inner protons of the porphyrin at 8.85

ppm and minus 2.94 ppm, respectively, and a large singlet

corresponding to the 24 acetoxy methyl protons at 1.11 ppm.

Insertion of zinc to generate the metalated porphyrin 26 was

carried out using by reﬂuxing the free base porphyrin with zinc

acetate dihydrate in chloroform for 30 min. (Compared to the

4- and 3,5-substituted porphyrins, higher temperatures are

required to insert zinc into the the 2,6-porphyrin system.) After

work-up and puriﬁcation, the UV/vis spectrum of the crude

reaction mixture now displayed only two Q-bands, indicating

Figure 10. Schematic showing how steric crowding around the

porphyrin core is more acute for the 2,6-substituted system (below)

compared to the 3,5-substituted system (top). Moving the initiation/

branching point one position closer to the porphyrin ring immediately

eﬀects the steric environment around the catalytic porphyrin. Although

two HBP phenyl substituents are shown, the polymers are likely to be

a mix of mono-, di-, tri-, and tetrafunctionalized systems (each

possessing 1 or 2 HBP arms).

N

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Scheme 11. Synthesis of the 2,6-Functionalized Porphyrins 23, 24, 24, and 26

successful insertion. Further conﬁrmation of zinc insertion was

conﬁrmed by mass spectrometry, which showed only a single

peak corresponding to the functionalized porphyrin at 1141 Da

and disappearance of the highly shielded inner porphyrin

(2,6-DAPP)) in a yield of 40%. Although the spectrum was

broadened (due to the paramagnetic iron), ¹H NMR

spectroscopy showed a large singlet at 1.55 ppm corresponding

to the acetoxy protons on the product. Furthermore, the mass

spectrum displayed only one peak at 1133 Da corresponding to

the molecular ion.

1

protons in the H NMR (previously seen at minus 2.94 ppm).

Insertion of iron proved to be more problematic. Following

the same method as for 5 and 18 and simply reﬂuxing the

porphyrin precursor with Fe(II) chloride and 2,6-lutidine did

not yield any of the metalated product. This is due to the

increased steric hindrance imposed by the eight acetate

substituents. In an eﬀort to circumvent this diﬃculty, the

reaction was reﬂuxed for 24 h; disappointingly, this did not

solve the problem. The use of Fe(CO)₅and iodine has

previously been used to metalate “twin coronet” porphyrins

bearing four binaphthalene bridges.⁴²Unfortunately, after

several attempts, we were still unable to insert iron into the

2,6-functionalized porphyrin. Therefore, the only option was to

reduce the steric hindrance; consequently, the iron must be

inserted into the porphyrin prior to the acetylation.

Accordingly, the less hindered 5,10,15,20-tetrakis(2,6-dihydrox-

yphenyl)-21H,23H-porphyrin 24 was metalated using iron(II)

chloride in methanol at 50 °C for 3 h, as shown in Scheme

12.³¹Puriﬁcation using column chromatography gave the

Synthesis of Zinc Metalated Tetra(2,6-

diacetoxyphenyl)porphyrin Cored Hyperbranched Poly-

mers. This polymer was synthesized using exactly the same

procedure as that previously used to construct the two polymer

systems discussed above. Speciﬁcally, 2,6-diacetoxyphenylpor-

phyrin was reacted with a 20 M excess of the AB₂monomer,

3,5-diacetoxybenzoic acid, as shown in Scheme 13. After

puriﬁcation the polymer was isolated in good yield and

1

possessed an M_nof 11 306. UV and H NMR conﬁrmed

porphyrin incorporation and the successful synthesis of tetra-

(2,6-DAPP)-HBP 29. The UV/vis spectrum showed an intense

band at 417 nm and Q-bands at 511, 539, 586, and 653 nm.

Zinc was inserted into the corresponding polymer by reﬂuxing

the polymer and excess zinc acetate in CH₂Cl₂. The crude

reaction mixture was then ﬁltered to remove unreacted zinc

acetate, the solvent was then removed, and the product

precipitated into methanol from reﬂuxing chloroform. GPC

analysis showed the zinc metalated polymer 30 had a very

similar molecular weight (11 710) to the free base polymer 29

(11 306) subsequent to metal insertion. This conﬁrms that zinc

insertion does not damage the polymer.

Scheme 12. Synthesis of 2,6-Diacetoxyphenylporphyrin Iron

Complex 28

Successful insertion of zinc was conﬁrmed by UV/vis

spectroscopy, which showed two Q-bands at 546.5 and 590.5

1

nm. Furthermore, the H NMR spectrum showed that the

highly shielded peak at minus 2.94 ppm was absent. The bulk

polymer was fractionated using size exclusion chromatography

and analyzed by GPC; the data are shown in Table 8 (during

fractionation the very small amounts of residual unincorporated

porphyrin were removed; equivalent to around 5% of the total

product). Once again, monitoring UV absorption with respect

to M_n/concentration conﬁrmed that the core was evenly

distributed across all molecular weights.¹⁵Dense packing was

once again probed by studying the binding strengths of various

pyridyl ligands to the fractions of Zn-tetra-(2,6-DAPP)-HBP

30. The same three ligands were used and to aid comparisons,

and the data normalized relative to the binding constants

obtained for the porphyrin core unit Zn-tetra-(2,6-DAPP) 26,

as well as being normalized relative to each ligand; the plots are

shown in Figure 11. Although data were not available for a low

molecular weight polymer, the plots obtained for each ligand

are very diﬀerent to those observed for Zn-TAPP-HBP 7 and

Zn-(3,5-DAPP)-HBP 20. For the Zn-(2,6-DAPP)-HBP 30,

steric hindrance appears to be immediate (over the molecular

desired metalated porphyrin 27 in an excellent 98% yield. The

ﬁrst indication the reaction had been a success was provided by

UV/vis absorption spectroscopy, which showed a signiﬁcantly

broadened peak when compared with the unfunctionalized

porphyrin 24. Final conﬁrmation was provided by the mass

spectrum which displayed only a single peak corresponding to

the successfully functionalized porphyrin at 796 Da. The next

synthetic step was to introduce acetoxy functionality; this was

performed as before using acetic anhydride under reﬂux for 6 h.

The acetic acid side product and excess solvent were then

removed under vacuum and the crude product puriﬁed via ﬂash

chromatography giving the acetylated porphyrin 28 (Fe-tetra-

O

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Scheme 13. Synthesis of Tetra-(2,6-DAPP)-HBP 29, Zn-Tetra-(2,6-DAPP)-HBP 30, and Fe-Tetra-(2,6-DAPP)-HBP 31

Table 8. M_n, Polydispersity, and Number of Repeat Units for

Hyperbranched Polymer Zn-2,6-DAPP-HBP 30

no. of repeat

units

pseudo-

generation

M_n

polydispersity

1142 (Zn-2,6-DAPP)

0

5900

3.74

2.93

2.53

2.55

27

43

57

68

2.0

2.5

3.0

3.5

8800

11300

13225

weights studied) and increase with polymer molecular weight.

Interestingly, the binding strength decreases in an almost linear

manner up to an M_nvalue of 11 300, after which binding

appears to decrease at a slower rate.

The binding data for each ligand are shown in Table 9.

Overall, there is a very large reduction in binding aﬃnities for

all ligands, with values between 72 and 79% being recorded for

all three. Within error the values obtained are equivalent, with

no ligand being hindered any more or less than the any other.

Furthermore, with respect to core binding, we notice that

binding aﬃnity for each ligand was similar to the values

obtained for the Zn-tetra-(3,5-DAPP)-HBP 20. This suggests a

similar steric environment around the central metal, at least

with respect to the ligands studied. However, when comparing

the binding for the Zn-tetra-(3,5-DAPP)-HBP 20 and Zn-tetra-

(2,6-DAPP)-HBP 30 polymer systems, it is clear that a

considerable steric barrier is imposed on the latter. Conversely,

a signiﬁcantly improved binding environment is generated by

the Zn-tetra-(3,5-DAPP)-HBP 20. Speciﬁcally, when compar-

ing the two polymer systems we observe a 35−50% drop in

binding for the Zn-tetra-(2,6-DAPP)-HBP 30 with M_nof 5900

Figure 11. Plot showing ligand binding aﬃnities vs molecular weight

for the Zn-2,6-DAPP-HBP 30. All data normalized for binding to the

core porphyrin (Zn-2,6-DAPP 26) and also normalized relative to

each ligand. (For reference, the K_avalues were 14 900, 13 900, and 12

100 M⁻¹for pyridine, 3,5-lutidine, and 3-phenylpyridine, respectively.)

(with respect to the simple porphyrin core). However, for the

equivalent Zn-tetra-(3,5-DAPP)-HBP 20 with an M_nof 6100,

we observe a 20−40% increase in binding aﬃnity when

compared to its simple porphyrin core. This implies that

although dense packing is not a factor for the Zn-tetra-(2,6-

DAPP) porphyrin core 26, it must be occurring very early on

within the polymer series. For the polymer with molecular

weight 11 300 (i.e., the HBP with a molecular weight closest to

the analysis used for the original TAPP polymer system), the

P

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Table 9. Binding Data of Pyridyl Ligands to Zn-Tetra-(2,6-DAPP)-HBP 30

a

ligand

pyridine

K_aporphyrin core (M⁻¹

)

K_aof HBP with M_n11 300 (M⁻¹

)

onset of dense packing (M_n)

reduction in binding (%)

14900 ( 1500)

13900 ( 1500)

12100 ( 1500)

3100 ( 350)

3900 ( 450)

2900 ( 350)

immediate

79 ( 9)

70 ( 8)

75 ( 8)

3,5-lutidine

3-phenylpyridine

a

Calculated by comparing K_avalues obtained from Zn-2,6-DAPP 26 and the Zn-2,6-DAPP-HBP fraction with M_nclosest to 11 500 (allowing

comparison with the other HBP systems described above).

Table 10. Theoretical Values for Pseudo-Generations vs the Values for Obtained Fractions of Fe-2,6-DAPP-HBP 31

molecular weight

theoretical

repeat units

pseudo-generation

obtained

theoretical

obtained

0 (Fe-DAPP-28)

1133

2000

0

0.5−1

3

1846 (G = 0.5), 2557 (G = 1.0)

4 (G = 0.5), 8 (G = 1.0)

4−8

59

11134

11500

56

reduction in binding is much greater, with a 70−79% reduction

in binding being observed for Zn-tetra-(3,5-DAPP)-HBP 20.

Synthesis of the Iron Metalated Tetra(2,6-

diacetoxyphenyl)porphyrin Cored Polymers (Fe-Tetra-

(2,6-DAPP)-HBP) and Catalytic Studies. For the catalytic

studies an iron metalated tetra(2,6-diacetoxyphenyl)porphyrin

cored polymer would be required. Previously, we have

metalated the hyperbranched polymers prior to fractionation.

However, due to the diﬃculties in inserting iron into the 2,6

substituted porphyrin, it was decided that an alternative

polymerization procedure would be required. First, the metal

would be inserted into the porphyrin and then the polymer-

ization performed. The polymerization procedure is shown in

Scheme 13. Iron metalated 2,6-diacetoxyphenyl porphyrin 28

was polymerized under vacuum with excess 3,5-diacetoxyben-

zoic acid. After initial puriﬁcation a reddish/brown solid was

obtained. Size exclusion chromatography revealed the bulk

polymer had an M_nof 11 700 and PD of 2.73. Although the

structures of the G0.5 and the G1.0 dendrimers using a simple

modeling package (Chem-3D), we noticed that the G1.0

dendrimer does not necessarily possess a signiﬁcantly larger

size/volume, despite having twice the number of repeat units.

As such, both would elute with a similar retention time/

molecular weight. The M_nvalue of the lower molecular weight

fraction obtained after preparative GPC was around 2000 and is

therefore between the theoretical values of the G0.5 and G1.0

polymers. However, based on the size exclusion and molecular

volume arguments given above, it is probably closer in structure

to the larger G1.0 pseudo-dendrimer.

Probing Shape and Size Selectivity of the Fe-2,6-

DAPP-HBP via Catalysis. The catalytic reactions were carried

out using the same method described for Fe-TAPP-HBP 8 and

Fe-tetra-(2,6-DAPP)-HBP 31 as described above. Before

carrying out the mixed experiment, each substrate was once

again studied individually. As before, the yields of all oxidation

products determined for each polymer catalyst are normalized

with respect to the porphyrin core, Fe-tetra-(2,6-DAPP) 28

(Figure 12). It was noticed the yields for Fe-tetra-(2,6-DAPP)

28 were much lower than those obtained using the

corresponding 3,5- and 4-substituted Fe-porphyrin cores 5

and 18. For both the linear and cyclic substrates, a 20%

1

peaks in the H NMR were relatively broadened (due to the

paramagnetic iron), there was enough resolution to conﬁrm

that the reaction had been successful.

The previous catalytic experiments using the 3,5-function-

alized porphyrin derived polymers indicated that any selectivity

occurred at high molecular weights. Coupled with the binding

data obtained for the 2,6-functionalized porphyrin polymers

described above, which showed that binding decreased linearly

with increasing molecular weight, it was decided that nothing

would be gained by separating the bulk polymer into a larger

series of pseudo-generation polymers. Therefore, it was decided

to use just three main fractions: the G0 or core unit, a high

molecular weight polymer, and a low molecular weight

polymer. As before, this was achieved using preparative size

exclusion chromatography, and the data are shown in Table 10.

The larger fraction obtained had an M_nvalue of around 11 000

and was assigned as the G3.0 pseudo-dendrimer. The smaller

fraction had an M_nof 2000, and assigning a pseudo-generation

to this fraction proved slightly more diﬃcult. This is due to the

fact that only very small changes in molecular volume occur

between the smaller pseudo-generation polymers. As such, it is

hard to obtain an accurate M_nvalue using size exclusion

chromatography.²³This is particularly true for tetra-(2,6-

DAPP)-HBP and derivatives, where initial growth occurs over

the porphyrin ring, rather than out and away from the

porphyrin (as is the case for the 3,5- and 4-substituted

systems). As such, there is no signiﬁcant change in the

molecules molecular volume. For example, when comparing the

Figure 12. Relative yields of all oxidation products obtained from the

oxidation of 1-octene 9 and cyclooctene 11 using Fe-2,6-DAPP-HBP

31 as the catalyst. Yields of all oxidation products for the G = 0

pseudo-dendrimer were 62% and 77% (for 1-octene and cyclooctene,

respectively).

Q

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

reduction in yield was observed. For the linear alkene, the yields

dropped from 77% to 62%, while yields dropped from 96% to

77% for the larger cyclic alkene. In both cases, this reduction in

yield is due to the increased steric environment around the Fe-

2,6-DAPP core unit. Interestingly, the activity of the G1 and G3

pseudo-dendrimers when catalyzing the oxidation of the linear

alkene 9 was actually higher than that recorded for the simple

Fe-2,6-DAPP core system (the G0 dendrimer). This is despite

the observed decrease in binding aﬃnity observed during the

binding studies described above. Although this was not

necessarily expected, it can be explained by considering the

local environment provided by the pseudo-dendrimers and the

diﬀerence in structure between the pyridine ligands used to

probe the structure and the linear alkene substrate 9. The

binding aﬃnity for the G1 and G3 pseudo-dendrimers was

much lower than that obtained for the core G0 porphyrin,

conﬁrming a more sterically demanding environment for these

polymers (when binding pyridine). However, the linear alkene

9 is much narrower than the bulky pyridine ligands and can ﬁt

more easily within a polymer-generated cavity surrounding the

catalytic porphyrin unit. As such, we conclude that within the

hyperbranched polymers there is an optimum environment for

the oxidation species and the linear substrate, which results in

higher relative yields. Regarding the results obtained using the

cyclic alkene (Figure 12), it is evident that the pseudo-

generation dendrimers catalyze the reaction with a consistent

level of activity over the 20 min duration of the reaction. Again,

this is despite the reduction in binding aﬃnities described

above, and although there is a reasonably consistent catalytic

environment surrounding the porphyrin core, the environment

is more optimal for the linear system. Nevertheless, it is clear

the reactivity of both alkenes is dependent on a balance

between steric hindrance, binding aﬃnity, and reaction times

(related to binding and reaction rates). These diﬀerences have

led to signiﬁcance variations in the reaction behavior of the

linear alkene 9 when compared to the cyclic alkene 11. As such,

some element of shape and size selectivity may be possible for

the Fe-2,6-DAPP-HBP system. Selectivity was probed in the

same manner as the other porphyrin cored systems described

above.

Mixed experiments using a 1:1 mixture of the linear alkene 9

and the cyclic alkene 11 were performed in the presence of

iodosyl benzene and catalyst (the Fe-2,6-DAPP-HBPs; G0, G1,

and G3), and reactions were monitored by measuring the yield

of epoxides, 10 and 12 (Scheme 5). Reactivity diﬀerences were

taken into account, and the data were normalized with respect

to the core porphyrin (Fe-tetra-(2,6-DAPP 28); the results are

shown in Figure 13. As before, a value greater than 1.0 indicates

a shift in favor of the less reactive, but smaller/thinner linear

alkene.

It is immediately apparent from the graph that the Fe-tetra-

2,6-DAPP-HBP system 31 is capable of shape/size selectivity.

After normalization, the G1 pseudo-dendrimer showed more

than twice the selectivity of the Fe-tetra-(2,6-DAPP) core (G0

porphyrin) 28. The increase is even larger for the G3 pseudo-

dendrimer, which exhibits a 3.5-fold increase in selectivity

relative to the G0 or Fe-2,6-DAPP core system 28. To ensure

that selectivity was genuine and not an isolated result, many

repeats were performed (as they were for the simpler catalytic

studies using either the linear or cyclic alkene). Therefore, this

sterically hindered polymer displays a remarkable shift in

selectivity toward the less reactive, but smaller linear octene.

This selectivity is due to the restricted access imposed on the

Figure 13. Selectivity of epoxide products from a mixed experiment

using the Fe-tetra-(2,6-DAPP)-HBP 31. Yields normalized with

respect to the inherent reactivity diﬀerence between the linear

(terminal) alkene 9 and the more reactive cyclic alkene 11. Therefore,

values higher than 1.00 indicate a shift in reactivity toward the less

reactive linear alkene.

larger cyclic alkene. Overall, this shows that the Fe-tetra-(2,6-

DAPP)-HBP 31 is capable of behaving like a dendrimer and

displaying reaction selectivity based on size.

CONCLUSIONS

■

At the outset of this work we wanted to determine whether or

not a hyperbranched polymer could be applied to some of the

more demanding technical applications usually associated with

dendrimers. We were particularly interested in targeting core

functionalized systems and applications involving selective

encapsulation, regioselectivity, and/or chemoselectivity. After

some modiﬁcation to the core unit, we were successful in this

respect. Our initial target was to try and reproduce the size and

shape selective oxidation of simple alkenes catalyzed by

porphyrin cored dendrimers.

A reversible transesteriﬁcation procedure using 4-

(tetraacetoxphenyl)porphyrin as the core unit and 3,5-

diacetoxybenzoic acid as the monomer generated a porphyryirn

cored HBP. Fractionation produced a series of “pseudo

dendrimers” ranging from G0 to G5 with well controlled

molecular weights, low polydispersities, and an even level of

core incorporation. Metalation of the central porphyrin

followed by binding studies identiﬁed a signiﬁcant drop in

binding strength at M_nvalues greater than 8000 Da. This

provided convincing evidence for the existence of a dense

packed limit or structure, which provides a steric barrier capable

of restricting access to the central porphyrin. Unfortunately,

when used to catalyze the oxidation of two alkenes diﬀering in

size and reactivity, no change in product distribution was

observed. This lack of selectivity was attributed to diﬀerences in

the rate of substrate access, relative to the rate of reaction for

each substrate. That is, although the HBPs possessed a steric

barrier capable of slowing down and restricting substrate access,

it was not suﬃcient enough to aﬀect the overall rate of the two

reactions (oxidation of the small/thin alkene and the large/fat

alkene). We suspected that the problem lay in part with the

initial branching geometry from the central catalytic unit, where

initial growth occurred at the 4/para-position and branching is

directed away from the core, resulting in a relatively open core

environment and a more densely packed external structure. A

more sterically demanding 3,5-substituted porphyrin core was

subsequently used to generate a second series of HBPs. Binding

R

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

studies again identiﬁed a dense packed limit at a molecular

weight around 8000 Da. Furthermore, the same data indicated

that the larger polymer fractions possessed a more sterically

crowded environment that was more sensitive to ligand size

and geometry (when compared to the 4-substituted system).

This was encouraging and was backed up by initial catalytic

studies showing how the relative yields of oxidation product

obtained from the larger alkene were smaller than those

obtained using the smaller alkene. This suggested that shape

and size selectivity might be possible when oxidizing both

alkenes at the same time, and although the results did show an

improvement in selectivity, it was very small. In a ﬁnal attempt

to impose steric demand at the core, a 2,6-disubstituted

porphyrin unit was used. The branching motif would force the

resulting polymer structure to grow directly over and around

the catalytic porphyrin core. Binding studies on this more

sterically demaning HBP system revealed a very diﬀerent

binding proﬁle to that obtained using the other porphyrin

cored systems. In the case of the 2,6 system, a linear decrease in

binding was observed as molecular weight increased, which

indicates a severe steric environment with any dense packed

limit being immediate and strong. Catalytic studies carried out

independently on the two alkenes revealed higher than

expected yields in both cases, which suggests that the polymer

environment (around the core) might be helping the catalytic

process. This eﬀect seemed to be particularly strong for the

linear alkene, where increased relative yields were observed

with respect to molecular weight. When the two alkenes were

reacted together with the catalysts, a marked switch in relative

selectivity was observed. When yields were compared to those

obtained using the nonpolymeric catalysts, the inherently less

reactive linear substrate was 3−4 times more reactive relative to

the larger bulkier cyclic alkene (after normalization with respect

to the inherent background reactivities). We suggest that this

change in selectivity is due to the core’s branching motif, which

encourages growth over and around the reactive porphyrin,

rather than away from the core, as was the case with the other

core units.

ACKNOWLEDGMENTS

■

We thank the EPSRC for providing funds through the DTA

scheme to support a studentship.

REFERENCES

■

(1) (a) Liu, J.; Zhong, Y.; Lam, J. W. Y.; Lu, P.; Hong, Y.; Yu, Y.; Yue,

Y.; Faisal, M.; Sung, H. H. Y.; Williams, I. D.; Wong, K. S.; Tang, B. Z.

Macromolecules 2010, 43, 4921. (b) Zhu, S.-W.; Shi, W. F. Polym.

Degrad. Stab. 2002, 75, 543. (c) Kirkorian, K.; Ellis, A.; Twyman, L. J.

Chem. Soc. Res. 2012, 41, 6138.

(2) (a) Klajnert, B.; Peng, L.; Cena, V. Dendrimers in Biomedical

Applications; Royal Society of Chemistry: Cambridge, 2013.

(b) Reymond, J. L.; Darbre, T. Org. Biomol. Chem. 2012, 10, 1483.

(c) Wen, S.; Li, K.; Cai, H.; Chen, Q.; Shen, M.; Huang, Y.; Peng, C.;

Houa, W.; Zhuc, M.; Zhangb, G.; Shi, X. Biomaterials 2013, 34, 1570.

(d) Isea1, R.; Hoebeke1, J.; Mayo-García, R. Am. J. Bioinf. Comput.

Biol. 2013, 1, 1.

(3) (a) de Gennes, P. G.; Hervet, H. J. Phys. Lett. 1983, 44, 351.

(b) Lescanec, R.; Muthukumar, M. Macromolecules 199.0, 23, 2280.

(c) Pandita, D.; Santos, J. L.; Rodrigues, J. O.; Peg

̂

o, A. P.; Granja, P.

L.; Tomas, H. Biomacromolecules 2011, 12, 472.

́

(4) Newkome, G. R.; Mooreﬁeld, C. N.; Vogtle, F. Dendritic

̈

Molecules; VCH: Weinheim, 1996.

(5) Tomalia, D. A. New J. Chem. 2012, 36, 264.

(6) Wallace, M.; Ellis, A.; Twyman, L. J. Chem. Commun. 2013, 49,

8063.

(7) (a) Bhyrappa, P.; Young, J. K.; Moore, J. S.; Suslick, K. S. J. Am.

Chem. Soc. 1996, 118, 5708. (b) Bhyrappa, P.; Young, J. K.; Moore, J.

S.; Suslick, K. S. J. Mol. Catal. A: Chem. 1996, 113, 109.

(8) (a) Pulkkinen, J.; Zachar, P.; Borek, V. A.; Laatikainen, R.; Kral,

V. J. Mol. Catal. A: Chem. 2004, 219, 21. (b) Schenning, A. P. H. J.;

Hubert, D. H. W.; Feiters, M. C.; Nolte, R. J. M. Langmuir 1996, 12,

1572. (c) Niedercorn, F.; Ledon, H.; Tkatchenko, I. New J. Chem.

1988, 12, 897. (d) Ostovic, D.; Bruice, T. C. J. Am. Chem. Soc. 1989,

111, 6511. (e) Mashiko, T.; Dolphin, D.; Nakano, T.; Traylor, T. G. J.

Am. Chem. Soc. 1985, 107, 3735. (f) Schenning, A. P. H. J.; Spelberg, J.

H. L.; Hubert, D. H. W.; Feiters, M. C.; Molte, R. J. M. Chem.Eur. J.

1998, 4, 871.

(9) (a) Yates, C. R.; Hayes, W. Eur. Polym. J. 2004, 40, 1257.

(b) Gao, C.; Yan, D. Prog. Polym. Sci. 2004, 29, 183. (c) Kim, Y. H. J.;

Polym. Sci., Part A: Polym. Chem. 1998, 36, 1685. (d) Grayson, S. M.;

́

Frechet, J. M. J. Chem. Rev. 2001, 101, 3819.

If we compare these results to those obtained by Suslick and

Moore using analogous dendrimers, we can conclude that

HBPs can indeed be applied as genuine alternatives to

dendrimers in a number of advanced and high-tech

applicationsparticularly those requiring site isolation, encap-

sulation, and release. Notwithstanding this generic claim, it is

important to appreciate that simply “bulking up” the polymer is

not enough and that it is equally important to consider how the

polymer’s architecture and core branching motif will aﬀect the

speciﬁc environment directly around any active site/core unit.

(10) (a) Liu, L.; Breslow, R. J. Am. Chem. Soc. 2002, 124 (18), 4978.

(b) Liu, L.; Rozenman, M.; Breslow, R. J. Am. Chem. Soc. 2002, 124

(43), 12660. (c) Kirby, A. J. Angew. Chem., Int. Ed. 1996, 35, 706.

(11) Holter, D.; Burgath, A.; Frey, H. Acta Polym. 1997, 48, 30.

̈

(12) King, A. S. H. K.; Martin, I.; Twyman, L. J. Chem. Soc. Res. 2002,

31, 69.

(13) Bharathi, P.; Moore, J. S. Macromolecules 2000, 33, 3212.

(14) Zheng, X.; Oviedo, I. R.; Twyman, L. J. Macromolecules 2008,

41, 7776.

(15) (a) Gittins, P. J.; Alston, J.; Ge, Y.; Twyman, L. J.

Macromolecules 2004, 37, 7428. (b) Twyman, L. J.; Ge, Y. Chem.

Commun. 2006, 1658. (c) Gittins, P. J.; Twyman, L. J. J. Am. Chem.

Soc. 2005, 127 (6), 1646.

ASSOCIATED CONTENT

■

(16) Turner, S. R.; Voit, B. I.; Mourey, T. H. Macromolecules 1993,

26, 4617.

S

* Supporting Information

Synthesis and characterization of all compounds. This material

is available free of charge via the Internet at http://pubs.acs.org.

(17) (a) Hulsman, N.; Medema, J. P.; Bos, C.; Jongejan, A.; Leurs, R.;

Smit, M. J.; de, E. I. J. P.; Richel, D.; Wijtmans, M. J. Med. Chem. 2007,

50, 2424. (b) Liao, C.-C.; Wang, C.-S.; Sheu, H.-S.; Lai, C. K.

Tetrahedron 2008, 64, 7977.

(18) Rothemund, P. J. Am. Chem. Soc. 1935, 57, 2010.

(b) Rothemund, P. J. Am. Chem. Soc. 1936, 58, 625.

(19) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.;

Assour, J.; Korsakoff, L. J. Org. Chem. 1967, 32, 476.

(20) Soret, J. L. C. R. Chim. 1883, 97, 1267.

AUTHOR INFORMATION

■

Corresponding Author

*E-mail l.j.twyman@she ﬃ eld.ac.uk (L.J.T.).

Notes

The authors declare no competing ﬁnancial interest.

(21) Caughey, W. S.; Kosk, W. S. Biochemistry 1962, 1 (5), 923.

S

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

(22) Friley, B. K.; Phelps, J. B.; Kincaid, J. R. J. Chromatogr. 1983,

258, 310. (b) Gilbert, J.; Shepherd, M. J.; Wallwork, M. A. J.

Chromatogr. 1985, 320, 361.

Naruta, Y. Chem.Eur. J. 2003, 9, 862. (b) Matsu-ura, M.; Tani, F.;

Naruta, Y. J. Am. Chem. Soc. 2002, 124, 1941.

(23) (a) Matthews, O. A.; Shipway, A. N.; Stoddart, J. F. Prog. Polym.

́

Sci. 1998, 23, 1. (b) Frechet, J. M. J.; Hawker, C. J. Compr. Polym. Sci.,

2nd Suppl. 1996, 71. (c) Twyman, L. J.; Beezer, A. E.; Mitchell, J. C. J.

Chem. Soc., Perkin Trans. 1 1994, 4, 407.

(24) (a) Chow, H.-F.; Mak, C. C. J. Org. Chem. 1997, I62, 5116.

(b) Mak, C. C.; Chow, H.-F. Macromolecules 1997, 30, 1228.

(25) (a) Xiao, F.; Sun, X.; Wu, X.; Zhao, J.; Luo, Y. J. Organomet.

̌

Chem. 2012, 713, 96. (b) Etrych, T.; Subr, V.; Jir

̌

Strohalm, J.; Sírova,

́

̌

M.; Ríhova,

́

B.; Ulbrich, K. J. Controlled Release 2012, 164, 346.

(c) Huang, M.; Khor, E.; Lim, L. Y. Pharm. Res. 2004, 21, 344.

(26) UV titrations were performed by monitoring changes in the

porphyrin’s Soret band, which appeared at 418 ( 0.5 nm) for all

fractions studied. For information on binding analysis and techniques,

see : Schneider, , H. J.; Yatsimirsky, , A. Principles and Methods in

Supramolecular Chemistry; Wiley: Chichester, 2000,.

(27) Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112,

5525.

(28) (a) Hawker, C. J.; Wooley, K. L.; Frec

Soc. 1993, 115, 4375. (b) Piotti, M. E.; Rivera, F., Jr.; Bond, R.; F.

Hawker, C. J.; Frechet, J. M. J. J. Am. Chem. Soc. 1999, 121, 9471.

́

het, J. M. J. J. Am. Chem.

́

(29) (a) Boris, D.; Rubinstein, M. Macromolecules 1996, 29 (22),

7251. (b) Jansen, J. F. G. A.; Meijer, E. W.; de Brabander-van den

Berg, E. M. M. J. Am. Chem. Soc. 1995, 117, 4417. (c) Burnett, J. L.;

King, A. S. H.; Martin, I. K.; Twyman, L. J. Tetrahedron Lett. 2002, 43,

2431.

(30) Sanders, J. K. M. S; Bampos, N.; Clyde-Watson, Z.; Darling, S.

L.; Hawley, J. C.; Kim, H. J.; Mak, C. C.; Webb, S. J.; Axial

Cordination Chemistry of Metalporphyrins. In The Porphyrin

Handbook; Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic

Press: San Diego, 2000; Vol. 3, Chapter 15.

(31) We have also investigated to use manganese porphyrins but

found the polymeric products to be unstable and therefore unreliable.

(32) Adler, A. D.; Longo, F. R.; Kampas, F.; Kim, J. J. Inorg. Nucl.

Chem. 1970, 32, 2443.

(33) Twyman, L. J.; Ellis, A.; Gittins, P. J. Macromolecules 2011, 44,

6365.

(34) Otting, G. J. Biomol. NMR 2008, 42, 1.

(35) Dauban, P.; Sanier

Soc. 2001, 123, 7707.

̀

e, L.; Tarrade, A.; Dodd, R. H. J. Am. Chem.

(36) Lindsay, S. J. R.; Sleath, P. R. J. Chem. Soc., Perkin Trans. 2 1982,

1009.

(37) Collman, J. P.; Brauman, J. I.; Hampton, P. D.; Tanaka, H.;

Bohle, D. S.; Hembre, R. T. J. Am. Chem. Soc. 1990, 112, 7980.

(38) The yields of just the epoxide 10 were 12% and 40% for the

linear and cyclic alkene, respectively.

(39) Clayden, J.; Greeves, N.; Warren, S. Organic Chemistry; Oxford

University Press: Oxford, 2012; Chapter 19.

(40) Experiments were also performed on the bulk polymer and the

yields and selectivities were similar to those obtained using the core

porphyrins. See: Zheng, X.; Oviedo, I. R.; Twyman, L. J. Macro-

molecules 2008, 41, 7776.

(41) Huang, L.; Chen, Y.; Gao, G.-Y.; Zhang, X. P. J. Org. Chem.

2003, 68, 8179.

(42) (a) Wagner, R. W.; Lindsey, J. S. J. Am. Chem. Soc. 1994, 116,

9759. (b) Wagner, R. W.; Johnson, T. E.; Lindsey, J. S. J. Am. Chem.

Soc. 1996, 118, 11166. (c) Prathapan, S.; Johnson, T. E.; Lindsey, J. S.

J. Am. Chem. Soc. 1993, 115, 7519.

(43) (a) Adams, R.; Bockstahler, T. E. J. Am. Chem. Soc. 1952, 74,

5346. (b) Haight, A. R.; Bailey, A. E.; Baker, W. S.; Cain, M. H.; Copp,

R. R.; DeMattei, J. A.; Ford, K. L.; Henry, R. F.; Hsu, M. C.; Keyes, R.

F.; King, S. A.; McLaughlin, M. A.; Melcher, L. M.; Nadler, W. R.;

Oliver, P. A.; Parekh, S. I.; Patel, H. H.; Seif, L. S.; Staeger, M. A.;

Wayne, G. S.; Wittenberger, S. J.; Zhang, W. Org. Process Res. Dev.

2004, 8, 897.

(44) (a) Tani, F.; Matsu-ura, M.; Ariyama, K.; Setoyama, T.;

Shimada, T.; Kobayashi, S.; Hayashi, T.; Matsuo, T.; Hisaeda, Y.;

T

dx.doi.org/10.1021/ma401344d | Macromolecules XXXX, XXX, XXX−XXX

Article Doi

Probing dense packed limits of a hyperbranched polymer through ligand binding and size selective catalysis

DOI: 10.1021/ma401344d

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ma401344d

Full text of DOI:10.1021/ma401344d

Products guided by the article

R&D Labs maybe for 142382-43-2

Relevant to this article

Hot Product