Mechanism of halogen exchange in ATRP

Article abstract of DOI:10.1021/ma201035u

Detailed mechanistic studies reveal that halogen exchange (HE) in ATRP can occur not only by a radical pathway (atom transfer) but also by an ionic pathway (S_N2 reaction) because Cu^(I)(L)X and Cu ^(II)(L)X₂ complexes contain weakly associated halide anion that can participate in the S_N2 reaction with alkyl halide (ATRP initiator). Both pathways were kinetically studied, and their contributions to the HE process were quantitatively evaluated for seven alkyl halides and three Cu^(I)(L)Cl complexes. Radical pathway dominates the HE process for 3° and 2° alkyl bromides with more active complexes such as Cu ^(I)(TPMA)Cl. Interestingly, ionic pathway dominates for 1° alkyl bromides and less active ATRP catalysts. These studies also revealed that degree of association of alkyl halide anion depends on the structure of copper complexes. In addition, radical pathway is accompanied by the reverse reactions such as deactivation of radicals to alkyl bromides and also activation of alkyl chlorides, reducing the efficiency of halogen exchange.

Full text of DOI:10.1021/ma201035u

ARTICLE
pubs.acs.org/Macromolecules
Mechanism of Halogen Exchange in ATRP
Chi-How Peng, Jing Kong, Florian Seeliger, and Krzysztof Matyjaszewski*
Center for Macromolecular Engineering, Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213,
United States
S Supporting Information
b
ABSTRACT: Detailed mechanistic studies reveal that halogen
exchange (HE) in ATRP can occur not only by a radical
pathway (atom transfer) but also by an ionic pathway (SN2
reaction) because Cu(I)(L)X and Cu(II)(L)X2 complexes con-
tain weakly associated halide anion that can participate in the
SN2 reaction with alkyl halide (ATRP initiator). Both pathways
were kinetically studied, and their contributions to the HE
process were quantitatively evaluated for seven alkyl halides and
three Cu(I)(L)Cl complexes. Radical pathway dominates the
HE process for 3° and 2° alkyl bromides with more active
complexes such as Cu(I)(TPMA)Cl. Interestingly, ionic path-
way dominates for 1° alkyl bromides and less active ATRP
catalysts. These studies also revealed that degree of association of alkyl halide anion depends on the structure of copper complexes.
In addition, radical pathway is accompanied by the reverse reactions such as deactivation of radicals to alkyl bromides and also
activation of alkyl chlorides, reducing the efficiency of halogen exchange.
’ INTRODUCTION
used to chain extend poly(methyl acrylate) macroinitiators
with methyl methacrylate to form block copolymers.10
Atom transfer radical polymerization is a well-established
technique for the synthesis of homo- and block copolymers with
defined compositions, architectures, and functionalities.7,11ꢀ16
The ATRP equilibrium (Scheme 2) regulates the control over
polymerization and is dominated by the carbonꢀhalogen bond
(PmꢀX) homolysis, the redox reaction between two oxidation
states of the transition metal complexes, and the formation of a
copperꢀhalogen bond (XꢀCu(II)). Altering the halogen atom at
the dormant chain end can shift the ATRP equilibrium and affect
the kinetics of ATRP process.9,10,17ꢀ24 The weakly associated
halide anion (Y) in Cu(I)(L)Y and XꢀCu(II)(L)Y species could
participate in a SN2 reaction with alkyl halide (ATRP initiator)
during the block copolymer synthesis by halide exchange. Since it
would form less active alkyl halide before the more active
monomer is added, the benefit of halogen exchange could be lost.
Herein, kinetic parameters of the radical pathway (atom
transfer) and ionic pathway (SN2 reaction) recognized as
potential contributors to the halogen exchange process were
measured for catalyst complexes formed with three ligands with
seven initiators (Scheme 3). This systematic analysis provides
an in-depth understanding of the mechanism of HE process
and a guide to improve the synthesis of block copolymers
by ATRP.
Block copolymers find many applications such as adhesives,1,2
sealants,3,4 and surfactants.5,6 Synthesis of block copolymers
is an important procedure for controlled radical polymeriza-
tion (CRP). Normally, during the synthesis of block copoly-
mers, more active monomers are polymerized before less
active monomers to ensure a near simultaneous growth of the
second block from the first macroinitiator. This order is
generally followed in CRP and, specifically in atom transfer
radical polymerization (ATRP), shows that the following
order of activity: acrylonitrile > methacrylates > styrene ∼
acrylates > acrylamides . vinyl chloride > vinyl acetate.7,8
If the order is reversed, the initially extended chains, which
have a faster reactivation, would grow more rapidly than
the unextended chains, producing polymers with a broad
or bimodal molecular weight distribution, as shown in
Scheme 1A. However, in ATRP, a procedure known as
halogen exchange (HE) allows this order to be altered.7,9,10
In this technique, a macromolecular alkyl bromide is chain
extended with a more reactive monomer in the presence
of a Cu(I)(L)Cl catalyst. Since the ATRP equilibrium constant
for a chloro-(macro)initiator is ∼1 order magnitude smaller
than for a bromo-(macro)initiator,8 the CꢀCl bond formed
upon deactivation of the chain extended radical is reactivated
more slowly than the remaining bromo-(macro)initiator chain
ends. The rate of chain growth of the second block is thus
decreased, leading to increased initiation efficiency from the
bromo-(macro)initiator and to polymers with a low poly-
dispersity, as shown in Scheme 1B. This technique has been
Received: May 6, 2011
Revised:
August 15, 2011
Published: September 14, 2011
r
2011 American Chemical Society
7546
dx.doi.org/10.1021/ma201035u Macromolecules 2011, 44, 7546–7557
|

Macromolecules
ARTICLE
Scheme 1. Chain Extension from Terminally Functionalized Polymers Composed of Less Active Monomers with a More Active
Monomer: (A) Using Regular ATRP; (B) Using ATRP with Halogen Exchange
Scheme 2. ATRP Equilibrium between an Alkyl Halide and
Copper Complex in the Lower Oxidation State and Alkyl
Radical and Copper Complex in the Higher Oxidation State
experiments were performed under pseudo-first-order condi-
tions using a large excess of the alkyl halide and nitroxide. The
reactions were monitored by following the increase of the
UVꢀvis absorbance at λmax of the formed Cu(II) complexes
(λ = 740 nm, ε = 178 L molꢀ1 cmꢀ1 for Cu(II)(bpy)2X2; λ =
740 nm, ε = 220 L molꢀ1 cmꢀ1 for Cu(II)(PMDETA)X2; λ =
950 nm, ε = 200 L molꢀ1 cmꢀ1 for Cu(II)(TPMA)X2 in
acetonitrile). Pseudo-first-order rate constants were obtained
from the exponential increase in the UVꢀvis absorbance by
fitting the single exponential y = A[1 ꢀ exp(ꢀkobst)] + C to the
observed time-dependent Cu(II) absorbance. The activation rate
constants (kact) were then obtained from kact = kobs/[alkyl
halide].25 Figure 1 shows a typical experiment for the measure-
ment of kact for Cu(I)(PMDETA)Cl with MBriB. All other kact
measurements are presented in the Supporting Information
(Figures SI 1ꢀ12).
’ RESULTS AND DISCUSSION
Scheme 4 shows the halogen exchange occurring by both the
radical pathway and the ionic pathway. The radical pathway,
shown in Scheme 4B, is an atom transfer process based on the
accepted ATRP mechanism. Three distinct steps—activation,
halogen interchange on the resulting Cu(II) species, and
deactivation—are involved in this pathway. Because the rate
of activation is generally slower than the rate of halogen inter-
change and deactivation, the activation step should be the rate-
determining step. Therefore, the activation rate constant (kact)
measured by trapping experiment with TEMPO25 has been used
to evaluate the contribution of the radical pathway to the overall
halogen exchange process.
The ionic pathway, shown in Scheme 4C, is based on the SN2
mechanism, and this pathway can be termed as a halide exchange.
In this pathway the bromo-(macro)initiator can be directly
converted to the chloro-(macro)initiator without generation of
a radical and before potential addition of the less active mono-
mer. This would result in a diminished level of control in the
synthesis of block copolymers. The rate constant of ionic path-
way (kionic) determined in the reaction with NBu4Cl has been
used to evaluate the contribution of the ionic pathway to the
overall halogen exchange process. Although halogen exchange
can be performed with several copper(I) halides and copper(I)
pseudohalides, Cu(I)(L)Cl and bromo-initiators were used for
the mechanistic study reported in this article.
Radical Pathway (Atom Transfer). The rate-determining
step of the radical pathway should be the activation process;
therefore, the activation rate constants (kact) were independently
determined by UVꢀvisꢀNIR spectroscopy using trapping ex-
periments with TEMPO. The radicals generated by CꢀX bond
homolysis from the alkyl halides activated by Cu(I)(L)X species
are irreversibly trapped by this stable nitroxide radical to yield the
corresponding alkoxyamines (Scheme 5). This method was
reported before26 and frequently used to measure the kact in
ATRP. Previous studies indicate that TEMPO only performs as a
radical trap, and the measured rate constants were independent
of the concentration of excess TEMPO.25,27ꢀ31 The kinetic
The activation rate constants with Cu(I)(L)Cl species
(kactCuCl) have not been as extensively studied as those with
Cu(I)(L)Br (kactCuBr).25,29ꢀ32 According to the ATRP mechan-
ism (Scheme 2), the halide counterion is not involved in the
activation step, and thus replacing the halide anion in Cu catalysts
from Brꢀ to Clꢀ may not affect the value of kact very strongly.
This assumption was confirmed by the measurement of kact for
selected Cu(I)(L)Cl transition metal complexes with a range of
initiators (Table 1). The values of kact for selected Cu(I)(L)Cl
and different initiators are similar to those for corresponding
Cu(I)(L)Br catalyst complexes. Nevertheless, according to recent
speciation studies, halide anions associate reversibly with copper
complexes and may affect their reactivities.33
Figure 2 shows the similarity of kact for Cu(I)(L)Cl and
Cu(I)(L)Br with PMDETA (Figure 2A) and bpy (Figure 2B)
as ligands. The values for kact for the Cu(I)(TPMA)Cl complex
are too large to be measured by the method used in this study;
therefore, only one initiator was selected to confirm the equiva-
CuCl
CuCl
lence of kact
and kactCuBr. The values of kact
measured
with alkyl chloride are similar to those of kactCuBr, indicating both
alkyl halides are activated by the catalyst with a similar role.
Because kact of Cu(I)(L)Cl is essentially equal to that of Cu(I)-
(L)Br, thus the previously published kact for Cu(I)(L)Br
species25,31 can be adopted for the halogen exchange study. As
kHE and kionic were measured at 27 °C, kact for 27 °C (Figure 3)
was calculated using the values of kact at 22 °C and the Eyring plot
(ln(kact/T) = a(1/T) + b).25,31
To better illustrate the trend of the rate constants, different
colors and symbols were used to label the data points in the
figures. Black, blue, and red symbols represent the rate constants
with primary, secondary, and tertiary alkyl halides, respectively;
7547
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
Scheme 3. Structures of (A) Ligands and (B) Initiators Used in This Kinetic Study of Halogen Exchange in ATRP
Scheme 4. Proposed Radical and Ionic Pathways in Halogen Exchange Process: (A) Overall Halogen Exchange; (B) Radical
Pathway; (C) Ionic Pathway
Scheme 5. Schematic Illustration of the Activation Rate
Constant (kact) Measurement by Irreversible Trapping of the
Formed Radicals with TEMPO
square, triangle, and circle mean the rate constants with ester-,
phenyl-, and cyano-initiators, respectively.
The range of ATRP activation rate constants is wide, spanning
Figure 1. Exponential increase of the absorbance at 950 nm (red line)
and the fitting curve (green line) of the reaction of Cu(I)(PMDETA)Cl
∼6 orders of magnitude (4.9 ꢁ 10ꢀ4ꢀ3.6 ꢁ 102 Mꢀ1 sꢀ1). The
values are mainly affected by the ligand associated with the Cu(I)
catalysts and the functional group attached to the carbon atom of
the CꢀX bond in the initiators.29ꢀ31 With the same initiator, the
ratio of kact with TPMA/PMDETA/bpy is∼1000/50/1 (Figure3,
kact(TPMA) ∼ 1000 ꢁ kact(bpy); kact(PMDETA) ∼ 50 ꢁ
kact(bpy)) because a ligand, which can better stabilize the Cu(II)
state of the catalyst and has a smaller entropic penalty for ligand
rearrangement when transitioning from the Cu(I) to Cu(II) state,
(1.09 ꢁ 10ꢀ3 mol Lꢀ1) with MBriB (3.43 ꢁ 10ꢀ2 mol Lꢀ1) and
TEMPO (1.26 ꢁ 10ꢀ2 mol Lꢀ1) in MeCN at 27 °C.
forms the Cu(I) catalyst with higher activity. For the same Cu(I)
catalyst, initiators with a cyano functional group have a much larger
kact than those with a phenyl or an ester group, following the
combination of polar and resonance effects. The influence of a
steric effect was also observed and shows the following trend: kact
7548
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
(tertiary initiators) > kact (secondary initiators) > kact (primary
initiators). However, compared to the differences caused by the
functional groups, the steric effect has a relatively smaller impact.7
Ionic Pathway (Halide Exchange). Weakly associated halides
in copper complexes can participate in SN2 reaction to displace
the halides from the initiators via an ionic pathway, concurrently
to atom transfer during the halogen exchange. In order to
evaluate the contribution of ionic pathway, the halide exchange
reaction between tetra-n-butylammonium chloride (NBu4Cl)
and alkyl bromides was studied. Tetra-n-butylammonium chlor-
ide, which dissociates well in polar media,34 was used as the
chloride source to react with equivalent amounts of alkyl
bromides (Scheme 6) in order to determine the rate constants
of the second-order kinetic plots. Since alkyl chlorides are
generally much less reactive than corresponding alkyl bromides,8
integrated rate law for an irreversible second-order reaction was
used though the kinetic order of halide exchange has not firmly
been established. A specific example measuring kionic for the
reaction of NBu4Cl with PEBr is shown in Figure 4. The kinetic
plots used to determine the other kionic are in the Supporting
Information (Figures SI 13ꢀ19).
The range of kionic values with NBu4Cl is from 9.7 ꢁ 10ꢀ6 to
5.8 ꢁ 10ꢀ1 Mꢀ1 sꢀ1 (Figure 5). Unlike the activation rate
constants, steric effects dominate the behavior of kionic. The trend
of kionic is (primary initiators) > kionic (secondary initiators) >
kionic (tertiary initiators) and matches the proposed SN2 mechan-
ism. The ester-initiators show a higher kionic than phenyl- and
cyano-initiators, but the influence of adjacent functional groups is
relatively small.
1
for an ionic pathway (kionic). H NMR was used to follow the
increase in the alkyl chloride concentration and the decrease in
alkyl bromide concentration in the reactions of NBu4Cl for
different initiators. The value for kionic was measured as the slope
The rate constants for the ionic pathway (halide exchange),
kionic, of the Cu(II)(L)Cl2 complexes (L = TPMA, PMDETA,
bpy) were evaluated by the same method as above and show the
same trend as kionic of NBu4Cl, in agreement with SN2 mechan-
ism (Figure 6). The kinetic plots used to determine the kionic are
in the Supporting Information (Figures SI 20ꢀ40). Previous
studies demonstrated that one chloride in Cu(II)(L)Cl2 complex
is directly bonded to the Cu(II) center by a CuꢀCl bond, but the
second chloride is not directly coordinated.35,36 This weakly
bonded chloride could dissociate to start the halide exchange
Table 1. Activation Rate Constants (kact) for Selected Cu(I)-
(L)Cl Species (kactCuCl) with Various Initiators in MeCN at
22 °C and the Corresponding kact for Cu(I)(L)Br (kact
CuBr
)
CuBr
CuCl
kact
kact
(Mꢀ1 sꢀ1
)
a
b
entry
L
initiator
(Mꢀ1 sꢀ1
)
1
TPMA
PMDETA
PMDETA
PMDETA
PMDETA
PMDETA
PMDETA
bpy
PECl
1.2 ꢁ 10ꢀ1
7.1 ꢁ 10ꢀ2
3.9
1.6 ꢁ 10ꢀ2
1.3 ꢁ 10ꢀ1
2.4
2.0 ꢁ 10ꢀ1
6.3 ꢁ 10ꢀ2
3.9
1.4 ꢁ 10ꢀ2
1.8 ꢁ 10ꢀ1
2.8
2
ClAN
BrAN
MBrA
MBrP
MBriB
EtBriB
BnBr
3
4
5
6
7
1.8
2.0
8
1.2 ꢁ 10ꢀ3
8.7 ꢁ 10ꢀ2
4.0 ꢁ 10ꢀ3
2.8 ꢁ 10ꢀ1
3.2 ꢁ 10ꢀ2
1.0 ꢁ 10ꢀ3
8.5 ꢁ 10ꢀ2
5.1 ꢁ 10ꢀ3
2.2 ꢁ 10ꢀ1
5.5 ꢁ 10ꢀ2
9
bpy
BrAN
MBrP
BrPN
MBriB
10
11
12
bpy
bpy
bpy
a Values from refs 25 and 29ꢀ31. b [CuCl]0/[L]0/[initiator]0
=
2.74 mM/2.74 mM/59.96 mM (entry 1); 2.83 mM/2.83 mM/
50.99 mM (entry 2); 1.11 mM/1.11 mM/34.56 mM (entry 3);
2.55 mM/2.56 mM/48.59 mM (entry 4); 2.55 mM/2.56 mM/
50.57 mM (entry 5);1.09 mM/1.09 mM/34.30 mM (entry 6);
1.09 mM/1.09 mM/35.34 mM (entry 7); 9.70 mM/19.50 mM/
296.29 mM (entry 8); 4.49 mM/9.00 mM/90.46 mM (entry 9);
2.73 mM/5.44 mM/54.64 mM (entry 10); 3.31 mM/6.65 mM/
65.84 mM (entry 11); 4.68 mM/9.40 mM/96.12 mM (entry 12).
Figures SI 1ꢀ12.
Figure 3. Activation rate constants (kact) for a variety of catalyst
complexes and initiators in MeCN at 27 °C. The values of kact are
shown in Tables 2ꢀ4.
Scheme 6. Model Reaction and the Rate Equation for kionic
Measurement
Figure 2. Plots of activation rate constants measured with Cu(I)(L)Cl vs Cu(I)(L)Br in MeCN at 22 °C: (A) L = PMDETA; (B) L = bpy. The values and
CuBr
conditions are shown in Table 1. The dashed line is the calculated line for kactCuCl = kact
.
7549
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
Figure 4. Measurement of the rate constant for an ionic pathway (kionic) with NBu4Cl (2.05 ꢁ 10ꢀ2 M) and PEBr (2.05 ꢁ 10ꢀ2 M) in MeCN-d3 at
27 °C: (A) time evolution 1H NMR spectra showing the conversion from PEBr to PECl; (B) second-order kinetic plots used to calculate the value of
kionic for PEBr, kionic = 2.61 ꢁ 10ꢀ3 Mꢀ1 sꢀ1
.
Scheme 7. Model Reaction and the Rate Equation for kHE
Measurement
and Clꢀ occurs to a greater extent. The ratio of kionic(Cu(II)
-
(TPMA)Cl2)/kionic(Cu(II)(PMDETA)Cl2)/kionic(Cu(II)(bpy)2Cl2) =
8.6/1.3/1.0 was observed.
Halogen Exchange. The rate constants of overall halogen
exchange, denoted kHE, were measured by following the decrease
of the concentration of the alkyl bromide and increase of the
concentration of alkyl chloride. In these experiments, alkyl
bromides were reacted with 1 equiv of Cu(I)(L)Cl (L = bpy,
PMDETA, TPMA) (Scheme 7), and the reaction was monitored
Figure 5. Rate constants for the ionic pathway (kionic) with NBu4Cl in
MeCN at 27 °C. The values of kionic are shown in Tables 2ꢀ4, and the
conditions are shown in the Supporting Information (Figures SI
13ꢀ19).
1
using H NMR spectra. The slope of the second-order kinetic
plots provides the value for kHE. A specific example of determin-
ing the kHE for reaction between Cu(I)(bpy)2Cl and MBrA is
shown in Figure 7. All other details of the measurements of kHE
can be found in the Supporting Information (Figures SI 41ꢀ61).
The values of halogen exchange rate constants, kHE
,
were largest with Cu(I)(TPMA)Cl, followed by Cu(I)(PMD-
ETA)Cl, and the kHE with Cu(I)(bpy)2Cl were the smallest
(Figure 8). This trend is similar to that reported for the
activation rate constants, kact, and can be correlated with
ATRP activity. However, the trend for the same catalysts
(Cu(I)(PMDETA)Cl, Cu(I)(bpy)2Cl) are somehow different
than that for kact, indicating some contributions of ionic
pathway.
A more comprehensive analysis of the trend for overall kHE
was performed by overlapping the kHE and kact for the same
catalyst with the value of kionic for NBu4Cl (Figures 9ꢀ11). In the
halogen exchange reaction of Cu(I)(bpy)2Cl complex with
different initiators, the kHE of secondary and tertiary initiators
show a similar trend to that of kact, but the trend for kHE with
primary initiators is similar to that for kionic (Figure 9). This
observation indicates that the halogen exchange of Cu(I)-
(bpy)2Cl with secondary and tertiary initiators is dominated by
the radical pathway, while the ionic pathway dominates the
overall halogen exchange with primary initiators. The switch of
dominant pathways for the halogen exchange process is due to
Figure 6. Rate constants of halide exchange (kionic) in MeCN at 27 °C
with NBu4Cl, Cu(II)(TPMA)Cl2, Cu(II)(PMDETA)Cl2, and Cu(II)
-
(bpy)2Cl2 versus initiator. The values of kionic are shown in Table 5,
and the conditions are in Figures SI 13ꢀ40.
process. The values of kionic for Cu(II)(TPMA)Cl2 are very close
to those for NBu4Cl which indicate that one chloride in Cu(II)
-
(TPMA)Cl2 complex is almost as dissociated as that in NBu4Cl.
The value for kionic of the Cu(II)(PMDETA)Cl2 complex are
smaller than those for Cu(II)(TPMA)Cl2 and only slightly larger
than kionic value of Cu(II)(bpy)2Cl2. The smaller values for kionic
mean that the degree of association between [Cu(II)(L)Cl]+
7550
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
Figure 7. Measurement of rate constants of halogen exchange (kHE) with Cu(I)(bpy)2Cl (1.98 ꢁ 10ꢀ2 M) and MBrA (1.98 ꢁ 10ꢀ2 M) in MeCN-d3 at
27 °C: (A) time evolution the 1H NMR spectra showing the conversion from MBrA to MClA; (B) second-order kinetic plots used to calculate the kHE of
MBrA, kHE = 1.13 ꢁ 10ꢀ2 Mꢀ1 sꢀ1
.
Figure 8. Rate constants for halogen exchange (kHE) in MeCN at
27 °C. The values for kHE are shown in Tables 2ꢀ4, and the conditions
are shown in the Supporting Information (Figures SI 41ꢀ61).
Figure 10. Plots of kHE and kact for Cu(I)(PMDETA)Cl overlapped
with kionic for NBu4Cl in MeCN at 27 °C. The values of kHE, kionic, and
k
act are shown in Tables 2ꢀ4.
Figure 9. Plots of kHE and kact for Cu(I)(bpy)2Cl overlapped with kionic
for reaction with NBu4Cl in MeCN at 27 °C. The values of kHE, kionic
,
Figure 11. Plots of kHE and kact for Cu(I)(TPMA)Cl with different
initiators overlapped with kionic for NBu4Cl in MeCN at 27 °C. The
values of kHE, kionic, and kact are shown in Tables 2ꢀ4.
and kact are shown in Tables 2ꢀ4.
the large difference between kact and kionic for primary initiators
(kionic . kact).
In the halogen exchange reaction using Cu(I)(PMDETA)Cl
(Figure 10) and Cu(I)(TPMA)Cl (Figure 11), the contribu-
tion from the radical pathway becomes larger due to an
increased kact, the ionic pathway becomes less significant.
Consequently, the trend for kHE is similar to that for kact with
all initiators, with the exception of BrAN. The kHE of BrAN is
7551
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
Table 2. Summary of kHE and kact for Cu(I)(bpy)2Cl and kionic for NBu4Cl in MeCN at 27 °C with the Linear Combination of
HE = α(I)L ionic + β(I)Lkact
k
k
kHE = α(I)bpykionic + β(I)bpykact
Cu(I)(bpy)2Cl
initiator
%
ionicb
%
radicalc
a
β(I)
bpy
kHE
α(I)
kionic(N(Bu)4Cl)
kact
bpy
MBrA
BnBr
1.1 ꢁ 10ꢀ2
3.0 ꢁ 10ꢀ3
8.0 ꢁ 10ꢀ3
2.8 ꢁ 10ꢀ3
1.8 ꢁ 10ꢀ3
4.9 ꢁ 10ꢀ2
2.2 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
1.9 ꢁ 10ꢀ2
5.8 ꢁ 10ꢀ1
1.3 ꢁ 10ꢀ1
1.7 ꢁ 10ꢀ1
8.7 ꢁ 10ꢀ3
2.6 ꢁ 10ꢀ3
1.2 ꢁ 10ꢀ3
9.7 ꢁ 10ꢀ6
3.2 ꢁ 10ꢀ1
4.3 ꢁ 10ꢀ1
4.3 ꢁ 10ꢀ2
3.9 ꢁ 10ꢀ1
6.4 ꢁ 10ꢀ1
1.7 ꢁ 10ꢀ1
3.0 ꢁ 10ꢀ1
4.9 ꢁ 10ꢀ4
1.3 ꢁ 10ꢀ3
1.1 ꢁ 10ꢀ1
6.7 ꢁ 10ꢀ3
2.8 ꢁ 10ꢀ3
2.9 ꢁ 10ꢀ1
7.2 ꢁ 10ꢀ2
98.60
81.03
39.74
5.81
1.40
18.97
60.26
94.19
97.30
99.95
>99.99
BrAN
MBrP
PEBr
2.70
BrPN
0.05
MBriB
0.01>
a Average βvalue with Cu(I)(bpy)2Cl is 4.2 ꢁ 10ꢀ1 excluding βvalues of BrAN and BrPN. b %ionic=α(I)L ꢁ kionic(NBu4Cl)/kHE. c % radical = β(I)L ꢁ kact/kHE
.
Table 3. Summary of kHE and kact for Cu(I)(PMDETA)Cl and kionic for NBu4Cl in MeCN at 27 °C with the Linear Combination of
HE = α(I)L ionic + β(I)Lkact
k
k
kHE = α(I)PMDETAkionic + β(I)PMDETAkact
Cu(I)(PMDETA)Cl
initiator
%
ionicb
%
radicalc
a
kHE
α(I)
kionic(N(Bu)4Cl)
β(I)
kact
PMDETA
PMDETA
MBrA
BnBr
2.2 ꢁ 10ꢀ2
4.2 ꢁ 10ꢀ2
3.0 ꢁ 10ꢀ2
1.6 ꢁ 10ꢀ1
4.8 ꢁ 10ꢀ2
1.8
2.6 ꢁ 10ꢀ2
2.6 ꢁ 10ꢀ2
2.6 ꢁ 10ꢀ2
2.6 ꢁ 10ꢀ2
2.6 ꢁ 10ꢀ2
2.6 ꢁ 10ꢀ2
2.6 ꢁ 10ꢀ2
5.8 ꢁ 10ꢀ1
1.3 ꢁ 10ꢀ1
1.7 ꢁ 10ꢀ1
8.7 ꢁ 10ꢀ3
2.6 ꢁ 10ꢀ3
1.2 ꢁ 10ꢀ3
9.7 ꢁ 10ꢀ6
3.8 ꢁ 10ꢀ1
5.7 ꢁ 10ꢀ1
5.1 ꢁ 10ꢀ3
8.5 ꢁ 10ꢀ1
3.5 ꢁ 10ꢀ1
7.0 ꢁ 10ꢀ2
2.1 ꢁ 10ꢀ1
1.8 ꢁ 10ꢀ2
6.8 ꢁ 10ꢀ2
5.1
1.9 ꢁ 10ꢀ1
1.4 ꢁ 10ꢀ1
2.6 ꢁ 101
3.1
68.55
8.05
31.45
91.95
BrAN
MBrP
PEBr
14.73
0.14
85.27
99.86
0.14
99.86
BrPN
0.01>
0.01>
>99.99
>99.99
MBriB
6.3 ꢁ 10ꢀ1
a Average β value with Cu(I)(PMDETA)Cl is 4.7 ꢁ 10ꢀ1 excluding β values of BrAN and BrPN. b % ionic = α(I)L ꢁ kionic(NBu4Cl)/kHE. c % radical =
β(I)L ꢁ kact/kHE
.
Table 4. Summary of kHE and kact for Cu(I)(TPMA)Cl and kionic for NBu4Cl in MeCN at 27 °C with the Linear Combination of
HE = α(I)L ionic + β(I)Lkact
k
k
kHE = α(I)TPMAkionic + β(I)TPMAkact
Cu(I)(TPMA)Cl
initiator
%
ionicb
%
radicalc
a
kHE
α(I)
kionic(N(Bu)4Cl)
β(I)
kact
TPMA
TPMA
MBrA
BnBr
1.8 ꢁ 10ꢀ1
1.2
3.9 ꢁ 10ꢀ1
3.6
1.5 ꢁ 10ꢀ1
1.5 ꢁ 10ꢀ1
1.5 ꢁ 10ꢀ1
1.5 ꢁ 10ꢀ1
1.5 ꢁ 10ꢀ1
1.5 ꢁ 10ꢀ1
1.5 ꢁ 10ꢀ1
5.8 ꢁ 10ꢀ1
1.3 ꢁ 10ꢀ1
1.7 ꢁ 10ꢀ1
8.7 ꢁ 10ꢀ3
2.6 ꢁ 10ꢀ3
1.2 ꢁ 10ꢀ3
9.7 ꢁ 10ꢀ6
2.0 ꢁ 10ꢀ1
7.5 ꢁ 10ꢀ1
3.4 ꢁ 10ꢀ3
7.2 ꢁ 10ꢀ1
6.5 ꢁ 10ꢀ1
9.3 ꢁ 10ꢀ2
5.8 ꢁ 10ꢀ1
4.6 ꢁ 10ꢀ1
1.6
1.1 ꢁ 102
5.0
48.33
1.63
51.67
98.38
BrAN
MBrP
PEBr
6.54
93.46
0.04
99.96
1.7
2.6
0.02
99.98
BrPN
3.3 ꢁ 101
3.6 ꢁ 102
0.01>
0.01>
>99.99
>99.99
MBriB
2.3 ꢁ 101
3.9 ꢁ 101
a Average βvalue with Cu(I)(TPMA)Clis5.8ꢁ 10ꢀ1 excluding βvalues of BrAN and BrPN. b %ionic=α(I)L ꢁ kionic(NBu4Cl)/kHE.c %radical=β(I)L ꢁ kact/kHE
.
surprisingly smaller than kact, in comparison to other initia-
tors. This is possibly caused by either the reduction of radical
to carbanion or radical termination reactions. In addition, the
difference between kHE and kact progressively decreases from
Cu(I)(bpy)2Cl to Cu(I)(PMDETA)Cl and to Cu(I)(TPMA)Cl,
indicating that the efficiency of halogen exchange can be
improved by using more active catalysts.
using a linear combination equation: kHE = α(I)L ionic + β(I)Lkact,
k
since the observed halogen exchange should be the sum of
contributions from these two independent pathways. The super-
script (I) refers to Cu(I) species. The values α and β show how
much the halide exchange in the presence of Cu(I)Cl species is
slower in comparison with NBu4Cl and how much halogen
exchange is slower that a pure activation step, respectively. The
results are listed in Tables 2, 3, and 4 for Cu(I)(bpy)2Cl, Cu(I)-
(PMDETA)Cl, and Cu(I)(TPMA)Cl, respectively.
The contributions of both the radical atom transfer and ionic
halide exchange to the overall halogen exchange were evaluated
7552
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
Table 5. Summary of kionic for Cu(II)(L)Cl2 (L = TPMA, PMDETA, bpy) and NBu4Cl with Different Initiators in MeCN at 27 °Ca
b
α(II)
bpy
initiator
NBu4Cl
Cu(TPMA)Cl2
Cu(PMDETA)Cl2
Cu(bpy)2Cl2
α(II)
α(II)
TPMA
PMDETA
MBrA
BnBr
5.8 ꢁ 10ꢀ1
1.3 ꢁ 10ꢀ1
1.7 ꢁ 10ꢀ1
8.7 ꢁ 10ꢀ3
2.6 ꢁ 10ꢀ3
1.2 ꢁ 10ꢀ3
9.7 ꢁ 10ꢀ6
5.5 ꢁ 10ꢀ1
9.3 ꢁ 10ꢀ2
1.6 ꢁ 10ꢀ1
6.1 ꢁ 10ꢀ3
2.2 ꢁ 10ꢀ3
8.0 ꢁ 10ꢀ4
8.1 ꢁ 10ꢀ6
5.6 ꢁ 10ꢀ2
1.2 ꢁ 10ꢀ2
1.6 ꢁ 10ꢀ2
7.9 ꢁ 10ꢀ4
5.8 ꢁ 10ꢀ4
1.0 ꢁ 10ꢀ4
1.4 ꢁ 10ꢀ6
5.1 ꢁ 10ꢀ2
1.2 ꢁ 10ꢀ2
1.5 ꢁ 10ꢀ2
6.4 ꢁ 10ꢀ4
3.6 ꢁ 10ꢀ4
8.8 ꢁ 10ꢀ5
9.9 ꢁ 10ꢀ7
9.5 ꢁ 10ꢀ1
7.2 ꢁ 10ꢀ1
9.4 ꢁ 10ꢀ1
7.0 ꢁ 10ꢀ1
8.5 ꢁ 10ꢀ1
6.7 ꢁ 10ꢀ1
8.4 ꢁ 10ꢀ1
9.7 ꢁ 10ꢀ2
9.2 ꢁ 10ꢀ2
9.4 ꢁ 10ꢀ2
9.1 ꢁ 10ꢀ2
2.2 ꢁ 10ꢀ1
8.3 ꢁ 10ꢀ2
1.4 ꢁ 10ꢀ1
8.8 ꢁ 10ꢀ2
9.2 ꢁ 10ꢀ2
8.8 ꢁ 10ꢀ2
7.4 ꢁ 10ꢀ2
1.4 ꢁ 10ꢀ1
7.3 ꢁ 10ꢀ2
1.0 ꢁ 10ꢀ1
BrAN
MBrP
PEBr
BrPN
MBriB
a Conditions are in Figures SI 13ꢀ40. b Because Cu(L)Cl2 does not activate the alkyl halide, kact for Cu(L)Cl2 should be zero, and thus kHE = kionic(L) =
α(II)L ionic(NBu4Cl) + βkact = α(II)L ionic. Therefore, α(II)L = kionic(L)/kionic(NBu4Cl).
k k
Scheme 8. Model Reactions and Kinetic Parameters for
Predici Simulation of Halogen Exchange
Figure 12. Time-dependent [BrPN] and [ClPN] to estimate the equi-
librium constant of halogen exchange (Kex) with Cu(I)(PMDETA)Cl and
BrPN in MeCN-d3 at 27 °C. [Cu(I)(PMDETA)Cl]0 = 6.3 ꢁ 10ꢀ3 M;
[BrPN]0 = 6.3 ꢁ 10ꢀ3 M; [ClPN]eq = 5.2 ꢁ 10ꢀ3 M; [BrPN]eq = 1.1 ꢁ
10ꢀ3 M; Kex = 5.6.
The values of α(I) and β(I) were estimated by using the
L
L
following principles: (1) both α(I)L and β(I)L should be larger
than 0 and smaller than 1; (2) the value of α(I) should be
L
identical for all initiators with the same Cu(I) species because
the extent of dissociation of chloride from Cu(I)(L)Cl is not
affected by the nature of the initiators; (3) the α(I)L value was
Cu(I)(PMDETA)Cl (Table 3). The range of α(I) values for
Cu(I)(TPMA)Cl was even larger (3 ꢁ 10ꢀ1 > α(IL)
> 0)
TPMA
since kact is 10 times larger than kionic, causing difficulties in
distinguishing the influence of α(I)TPMA. In this case the median
number, 1.5 ꢁ 10ꢀ1, was used for β(I)TPMA estimation (Table 4).
Fortunately, the accuracy of β(I)TPMA is not significantly affected
calculated from the kinetic data for MBrA (kHE, kact, kionic
)
because it has the smallest kact and largest kionic, increasing the
accuracy of the value of α(I)L; (4) the value of β(I)L was then
obtained from the equation of β(I)L = (kHE ꢀ α(I)
Lkionic)/kact
by the less precise α(I)
values because the contribution of
TPMA
once α(I)L was determined. The values of α(I)L and β(I)L were
further used to determine the contribution of ionic pathway
(% ionic in Tables 2ꢀ4) and radical pathway (% radical in
the ionic pathway is relatively small.
α Values. The linear combination equation provides infor-
mation on the relative contribution of the ionic pathway
(α(I)Lkionic/kHE) and the radical pathway (β(I)Lkact/kHE) to the
halogen exchange. Since kionic, measured with NBu4Cl, should
correspond to the halide exchange rate constant for a fully
Tables 2ꢀ4) to the halogen exchange by % ionic = α(I)
ꢁ
L
kionic(N(Bu)4Cl)/kHE and % radical = β(I)L ꢁ kact/kHE, respectively.
The Cu(I)(bpy)2Cl system is discussed as an example. Since
the value of kact for MBrA is quite small (4.9 ꢁ 10ꢀ4 Mꢀ1 sꢀ1
)
dissociated chloride, the α(I) value could be correlated with
L
and the contribution of the radical pathway, β(I)bpykact, is also
small, the value of α(I)bpy can be estimated from kHE = α(I)bpykionic
as 1.90 ꢁ 10ꢀ2. The range for the value of α(I)bpy was found to
the degree of dissociation of a chloride ion from the Cu(I)(L)Cl
catalysts. The ratio of α(I)TPMA/α(I)PMDETA/α(I) = 1.5 ꢁ
bpy
10ꢀ1/2.6 ꢁ 10ꢀ2/1.87 ꢁ 10ꢀ2 indicates that the chloride is
weakly associated with [Cu(I)(TPMA)]+ but interacts much
stronger with [Cu(I)(bpy)2]+. The association between Clꢀ
and Cu(I)(PMDETA)+ is similar to that for Clꢀ and
[Cu(I)(bpy)2]+. This trend is similar to that observed in the
kionic measurements using Cu(II)(L)Cl2 as a chloride provider
(Figure 6 and Table 5).
be 1.90 ꢁ 10ꢀ2 > α(I) > 1.83 ꢁ 10ꢀ2 in order to obtain a
bpy
β(I)bpy value within the range of 0 < β(I)bpy < 1. The median value
for α(I)bpy, 1.87 ꢁ 10ꢀ2, was used to calculate the β(I)bpy values
for each alkyl halide (Table 2). The range of values for
α(I)
α(I)
with Cu(I)(PMDETA)Cl was 3.8 ꢁ 10ꢀ2
>
PMDETA
> 1.4 ꢁ 10ꢀ2, which was larger than that of the
Cu(PI)M(bDpEyT)A2Cl system. The median number for α(I)
,
The trend α(II)TPMA/α(II)PMDETA/α(II)
= 8.1 ꢁ 10ꢀ1
/
PMDETA
bpy
1.2 ꢁ 10ꢀ1/9.4 ꢁ 10ꢀ2 = 8.6/1.3/1.0 is similar to that for
2.6 ꢁ 10ꢀ2, was used to calculate the β(I)
values for
PMDETA
7553
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
α(I)TPMA/α(I)PMDETA/α(I)bpy = 1.5 ꢁ 10ꢀ1/2.6 ꢁ 10ꢀ2/1.9 ꢁ
10ꢀ2 = 8.0/1.4/1.0. Recent speciation studies, evaluating the
stability constants (K) for copper complexes used as catalysts in
ATRP, also support these observations.33
Kex Values. Equilibrium constant of halogen interchange at
Cu(II) center (Kex) was estimated by using the [RBr]eq and
[RCl]eq in the halogen exchange process together with the
reported values of KATRP for Br and Cl systems (eqs 1ꢀ3).31
Since the radical species (R•) is not involved in the halogen
exchange on Cu(II)(L)2+, Kex should be the same for each
initiator. Thus, the halogen exchange of Cu(I)(PMDETA)Cl
and BrPN was used to estimate Kex of Cu(II)(PMDETA)BrCl
(Figure 12), since this system reached equilibrium very quickly.
β Values. The β(I)L values define the efficiency of the halogen
exchange process via atom transfer. It should be equal to 1.0 for
the ideal case when every ATRP activation converts an alkyl
bromide to an alkyl chloride. However, due to the reverse reactions
(back deactivation to an alkyl bromide and activation of a product
alkyl chloride), termination, and side reactions, most of the initiators
show β(I)L values smaller than 1. Values of β for BrAN are much
smaller and indicate side reactions, as discussed previously.
Using eq 3 with KATRPBr = 5.9 ꢁ 10ꢀ7, KATRPCl = 1.7 ꢁ 10ꢀ7
,
[ClPN]eq = 5.2 ꢁ 10ꢀ3 M, and [BrPN]eq = 1.1 ꢁ 10ꢀ3 M, Kex
was calculated as 5.6. Similar values were obtained for other
systems. The value Kex = 5.6 indicates that CuII/L species has
small preference for bonding with Cl stronger than with Br anion,
also in agreement with speciation results.33
Although the β(I) values for the five benzyl and ester
L
initiators fluctuate, a discernible trend exists. The average
β(I) value of Cu(I)(TPMA)Cl is the highest (5.8 ꢁ 10ꢀ1
,
L
Table 4), followed by those of Cu(I)(PMDETA)Cl (4.7 ꢁ
Br
Kex
s
s
R
CuCl þ RBr K
R• þ ðCuꢀBrÞCl
R•
F
ATRP
10ꢀ1, Table 3) and last by Cu(I)(bpy)2Cl (4.2 ꢁ 10ꢀ1
,
F
R
Table 2). The small increase of the average β(I) value from
L
þ ðCuꢀClÞBr
CuBr þ RCl
ð1Þ
ð2Þ
ð3Þ
F
R
Cu(I)(bpy)2Cl to Cu(I)(PMDETA)Cl to Cu(I)(TPMA)Cl in-
dicates the efficiency of halogen exchange can be improved by
using more active ATRP catalysts.
Cl
KATRP
2
2
Br
½RClꢂeq½CuBrꢂeq
½RBrꢂeq½CuClꢂeq
½RClꢂeq
½RBrꢂeq
The contributions of ionic pathway (α(I)
k
L ionic/kHE) and
KATRP
K
Cl
ex
KHE
¼
¼
¼
radical pathway (β(I)
Lkact/kHE) show a more clear trend:
KATRP
with the same initiators, the contribution of radical pathway
increases for more active catalysts (Cu(I)(TPMA)Cl > Cu(I)-
(PMDETA)Cl > Cu(I)(bpy)2Cl); with the same catalysts, the
contribution of radical pathway follows the trend: tertiary alkyl
halides > secondary alkyl halides > primary alkyl halides
(Tables 2ꢀ4, % radical). Radical pathway has the smallest
contribution when MBrA was used but more than 95% con-
tribution with secondary and tertiary alkyl halides, which indi-
cates that more active catalysts and secondary or tertiary
initiators are better substrates for the halogen exchange process.
2
2
Cl
KATRP ½RClꢂeq
Kex
¼
Br
KATRP ½RBrꢂeq
Kinetic Modeling. The influence of activation, deactivation,
and termination on β values was evaluated computationally using
Predici.37 Simulations were performed using rate constants for
Cu(I)(PMDETA)Cl and MBrP available in the literature and
comparing them with those obtained in the study. The model
reactions and kinetic parameters for halogen exchange are shown
in Scheme 8. The following procedure was used to better
understand how different parameters affect β values: (1) time-
dependent [RBr] and [RCl] were obtained by Predici using
various kinetic parameters; (2) conversion was calculated as
[RBr]/([RBr] + [RCl]); (3) second-order kinetic plots of
1/[RBr] ꢀ 1/[RB]0 = (1/[RBr]0) ꢁ ((1/conversion) ꢀ 1) =
kHE ꢁ t provided the values of kHE; (4) β values were calculated
Br
as β = kHE/kact
.
Initially, in the simulations, the ionic pathway (halide ex-
change) was neglected to simplify the system because its con-
tribution to overall halogen exchange is very small (0.14% in
Table 3). Initial concentrations of Cu(L)Cl and RBr were equal
Br
Figure 13. Influence of kex on β values simulated with kinetic
parameters for Cu(I)(PMDETA)Br and MBrP using the model in
Scheme 8. kexCl was varied as well to keep Kex constant.
Figure 14. (A) Model in Predici simulation for halogen exchange including only the forward reactions. (B) Second-order kinetic plots of [MBrP]
performed by Predici simulation of halogen exchange including only the forward reactions to estimate β value.
7554
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
Figure 15. (A) Model in Predici simulation for halogen exchange including only atom transfer mechanism. (B) Second-order kinetic plots of [MBrP]
performed by Predici simulation of halogen exchange including only atom transfer mechanism to estimate β value.
Figure 16. (A) Second-order kinetic plots of [MBrP] performed by Predici simulation of halogen exchange using the model in Scheme 8 to estimate
β value. (B) Simulated β values given by different model reactions for the halogen exchange.
Figure 17. Influence of (A) kdeactBr, (B) kdeactCl, (C) kactCl, and (D) kt to β value simulated with kinetic parameters for MBrP and model shown in
Scheme 8. When one parameter was varied, all other parameters were constant.
to 5 ꢁ 10ꢀ3 M. The reported activation and deactivation rate
constants (kactBr, kactCl, kdeactBr, and kdeactCl) in MeCN at 22 °C31
were used. The experimental β values were recalculated using
kactBr at 22 °C to compare with the simulated ones. A termination
rate constant of 1 ꢁ 109 Mꢀ1 sꢀ1 was used. The rate constants of
halide interchange at Cu(II)(PMDETA)2+ were unknown
although their ratio was experimentally determined (cf. previous
Cl
section) as Kex = kexBr/kex = 5.6. These rate constants were
7555
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
initially assumed as kexBr = 1.0 ꢁ 106 sꢀ1 and kexCl = 1.8 ꢁ 105 sꢀ1
and then subsequently varied to evaluate their influence on β.
’ ASSOCIATED CONTENT
Supporting Information. Experimental section, 1H
The value of kex was varied from 108 to 10ꢀ2 sꢀ1 with a
Br
S
b
NMR spectrum of all alkyl halides, and kinetic plots used to
evaluate kact, kionic, and kHE. This material is available free of
charge via the Internet at http://pubs.acs.org.
constant value Kex = 5.6 in the simulation using kineticꢀp1ara-
meters of MBrP. The β value stayed constant for kexBr > 1 s but
Br
decreased to 3.2 ꢁ 10ꢀ2 only for kex = 1 ꢁ 10ꢀ2 sꢀ1
(Figure 13). The speciation studies33 indicate the dissociation
constants of halides from XꢀCu(II)/L species in the range Kdis ca.
10ꢀ5ꢀ10ꢀ6 M. Assuming that association proceeds with the
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: km3b@andrew.cmu.edu.
diffusion controlled rate kass = 109 Mꢀ1 ꢀ1, the dissociation
s
should occur with the rate constant kdiss = 103 to 104 sꢀ1. Of
course, the halide exchange at Cu(II) center may occur faster,
without full halide dissociation. Regardless, it seems that values of
β < 1 are not caused by slow halide exchange but rather by some
other reactions.
’ ACKNOWLEDGMENT
This research was supported by the National Science Founda-
tion (CHE-10-26060) and the members of the CRP Consortium
at Carnegie Mellon University.
The simulations were started from the simplest model in
which only the forward reactions of halogen exchange were
concerned (Figure 14A). In this case, β value is equal to 1
(Figure 14B), indicating that each activation converts one alkyl
bromide to one alkyl chloride irreversibly. However, when the
reverse reactions were added (Figure 15A), the β value was
decreased from 1 to 0.76 as shown in Figure 15B. Finally, radical
termination was included in the simulation (Scheme 8) and
caused no further difference to β value (Figure 16A). Simulated
β values given by different model reactions are summarized in
Figure 16B and are slightly smaller than the experimental value of
β = 0.85 (Table 3). The small difference may also be due to a
limited accuracy of previously measured kact and kdeact values.
However, some β values in Tables 1ꢀ3 are smaller than 0.76,
and it was of interest to evaluate how various kinetic parameters
can affect β. Therefore, the values of kdeactBr, kdeactCl, kactCl, and kt
were varied one at a time to study the influence of deactivation,
activation, and termination on β value (Figure 17). The following
trends were observed: (1) the larger kdeactBr, the smaller β value
(Figure 17A); (2) the larger kdeactCl, which contributes to the
generation of alkyl chlorides, gives a larger β value (Figure 17B);
(3) the larger kactCl, which regenerates radical from alkyl chlor-
ides, the smaller β value (Figure 17C); (4) kt has essentially no
effect on the β values in a reasonable range of kt values. However,
when radical termination is significant (e.g., in systems with high
KATRP), β could be affected by termination (Figure 17D); (5) β
value is more sensitive to the changes of kdeactBr and kdeactCl and
less sensitive to the change of kactCl and kt.
’ REFERENCES
(1) Simal, F.; Jeusette, M.; Leclere, P.; Lazzaroni, R.; Roose, P.
J. Adhes. Sci. Technol. 2007, 21, 559–574.
(2) Creton, C.; Hu, G. J.; Deplace, F.; Morgret, L.; Shull, K. R.
Macromolecules 2009, 42, 7605–7615.
(3) Matsui, T.; Miwa, Y. J. Appl. Polym. Sci. 1999, 71, 59–66.
(4) Nivasu, V. M.; Reddy, T. T.; Tammishetti, S. Biomaterials 2004,
25, 3283–3291.
(5) Jain, S.; Bates, F. S. Science 2003, 300, 460–464.
(6) Li, W. W.; Min, K.; Matyjaszewski, K.; Stoffelbach, F.; Charleux,
B. Macromolecules 2008, 41, 6387–6392.
(7) Braunecker, W. A.; Matyjaszewski, K. Prog. Polym. Sci. 2007, 32,
93–146.
(8) Lin, C. Y.; Coote, M. L.; Gennaro, A.; Matyjaszewski, K. J. Am.
Chem. Soc. 2008, 130, 12762–12774.
(9) Ando, T.; Kamigaito, M.; Sawamoto, M. Macromolecules 2000,
33, 2819–2824.
(10) Shipp, D. A.; Wang, J.-L.; Matyjaszewski, K. Macromolecules
1998, 31, 8005–8008.
(11) Coessens, V.; Pintauer, T.; Matyjaszewski, K. Prog. Polym. Sci.
2001, 26, 337–377.
(12) Matyjaszewski, K.; Xia, J. Chem. Rev. 2001, 101, 2921–2990.
(13) Tsarevsky, N. V.; Matyjaszewski, K. Chem. Rev. 2007, 107,
2270–2299.
(14) Oh, J. K.; Drumright, R.; Siegwart, D. J.; Matyjaszewski, K. Prog.
Polym. Sci. 2008, 33, 448–477.
(15) Sheiko, S. S.; Sumerlin, B. S.; Matyjaszewski, K. Prog. Polym. Sci.
2008, 33, 759–785.
(16) Gao, H.; Matyjaszewski, K. Prog. Polym. Sci. 2009, 34, 317–350.
(17) Matyjaszewski, K.; Shipp, D. A.; Wang, J. L.; Grimaud, T.;
Patten, T. E. Macromolecules 1998, 31, 6836–6840.
(18) Matyjaszewski, K.; Shipp, D. A.; McMurtry, G. P.; Gaynor,
S. G.; Pakula, T. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 2023–2031.
(19) Matyjaszewski, K.; Paik, H.; Shipp, D. A.; Isobe, Y.; Okamoto,
Y. Macromolecules 2001, 34, 3127–3129.
These simulations indicate that the β value is mainly affected
Br
by kdeactCl and kdeactBr. The larger kdeactCl and the smaller kdeact
provide a β closer to 1. The values of kexBr have no effect on β, if
they are larger than 1 sꢀ1
.
’ SUMMARY
The contributions of a radical pathway (atom transfer) and an
ionic pathway (SN2 reaction) to halogen exchange in ATRP were
quantitatively studied for seven alkyl halides and three Cu(I)-
(L)Cl complexes. Radical pathway dominates the HE process for
3° and 2° alkyl bromides with more active complexes such as
Cu(I)(TPMA)Cl. Ionic pathway, however, becomes important
for 1° alkyl bromides and less active complexes. These studies
also revealed that the dissociation of halide anion from Cu(I)(L)X
and Cu(II)(L)X2 complexes is affected by the structure of copper
complexes. In addition, the reverse reactions such as deactivation
of alkyl bromides and activation of alkyl chlorides in the radical
pathway reduce the efficiency of halogen exchange.
(20) Tang, C.; Kowalewski, T.; Matyjaszewski, K. Macromolecules
2003, 36, 1465–1473.
(21) Huang, J.; Jia, S.; Siegwart, D. J.; Kowalewski, T.; Matyjaszewski, K.
Macromol. Chem. Phys. 2006, 207, 801–811.
(22) Ramakrishnan, A.; Dhamodharan, R. Macromolecules 2003,
36, 1039–1046.
(23) Stoffelbach, F.; Poli, R. Chem. Commun. 2004, 2666–2667.
(24) Poli, R.; Stoffelbach, F.; Maria, S.; Mata, J. Chem.—Eur. J. 2005,
11, 2537–2548.
(25) Seeliger, F.; Matyjaszewski, K. Macromolecules 2009, 42, 6050–
6055.
(26) Goto, A.; Fukuda, T. Macromol. Rapid Commun. 1999, 20,
633–636.
7556
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557

Macromolecules
ARTICLE
(27) Chambard, G.; Klumperman, B.; German, A. L. Macromolecules
2000, 33, 4417–4421.
(28) Matyjaszewski, K.; Paik, H. J.; Zhou, P.; Diamanti, S. J. Macro-
molecules 2001, 34, 5125–5131.
(29) Tang, W.;Matyjaszewski, K. Macromolecules 2006, 39, 4953–4959.
(30) Tang, W.;Matyjaszewski, K. Macromolecules 2007, 40, 1858–1863.
(31) Tang, W.; Kwak, Y.; Braunecker, W.; Tsarevsky, N. V.; Coote,
M. L.; Matyjaszewski, K. J. Am. Chem. Soc. 2008, 130, 10702–10713.
(32) Tang, W.; Nanda, A. K.; Matyjaszewski, K. Macromol. Chem.
Phys. 2005, 206, 1171–1177.
(33) Bortolamei, N.; Isse, A. A.; Di Marco, V. B.; Gennaro, A.;
Matyjaszewski, K. Macromolecules 2010, 43, 9257–9267.
(34) Song, L.; Trogler, W. C. J. Am. Chem. Soc. 1992, 114, 3355–3361.
(35) Pintauer, T.; Matyjaszewski, K. Coord. Chem. Rev. 2005, 249,
1155–1184.
(36) Baisch, U.; Poli, R. Polyhedron 2008, 27, 2175–2185.
(37) Wulkow, M. Macromol. Theory Simul. 1996, 5, 393–416.
7557
dx.doi.org/10.1021/ma201035u |Macromolecules 2011, 44, 7546–7557