PERSpECTIvES

In this Perspective, we critically

A comparison of RAFT and ATRP evaluate ATRP and RAFT, revealing key
differences and similarities that offer a

methods for controlled radical fundamental understanding of the CRP

methodologies and enable one to select the

polymerization ideal polymerization method for a desired

synthetic target or application. In addition,
we also highlight recent developments
Nghia P. Truong   , Glen R. Jones   , Kate G. E. Bradford   , Dominik Konkolewicz    throughout the mechanistic comparison
and other sections.
and Athina Anastasaki
Abstract | Reversible addition–fragmentation chain-​transfer (RAFT) polymerization Reaction components and mechanistic
comparison
and atom transfer radical polymerization (ATRP) are the two most common controlled
The ATRP and RAFT methods subject the
radical polymerization methods. Both methods afford functional polymers with a chosen monomer to chemical species
predefined length, composition, dispersity and end group. Further, RAFT and ATRP under certain polymerization conditions.
tame radicals by reversibly converting active polymeric radicals into dormant chains. The identity of the reagents used and
However, the mechanisms by which the ATRP and RAFT methods control chain parameters like reaction time, concentration
growth are distinct, so each method presents unique opportunities and challenges, and temperature (as well as other stimuli like
electrochemical potential and light) afford
depending on the desired application. This Perspective compares RAFT and ATRP by us plenty of experimental variables to adjust.
identifying their mechanistic strengths and weaknesses, and their latest synthetic We first describe the reagents involved in the
applications. two methods.

The concept of macromolecules, or the end group cannot be reactivated to Reagents and processes common to
polymers, came 100 years ago with continue polymerization. As a result, control conventional RAFT and ATRP
Staudinger’s pioneering work1 and can since over composition, topology, dispersity In the literature, it often appears that the
be found in almost every material in daily and end-​group functionality is limited, choice between RAFT and ATRP is made
life, including packaging, clothes, computers, thereby, limiting the usefulness of the arbitrarily or according to experience with
smartphones and medicine. Conventional products. Reversible deactivation radical one or the other protocol, rather than how
radical polymerization, also known as polymerization (RDRP) methodologies, effective they can be. Indeed, although
free radical polymerization (FRP), is one also referred to as controlled radical the mechanisms of RAFT and ATRP
of the most popular methods to synthesize polymerization (CRP), were developed to are different, they share many aspects in
commodity polymers and enable many of address these limitations. CRP strategies common. Similar to FRP, both reactions
these applications. In FRP, the prevalence control crucial molecular characteristics, operate on alkenes or substituted alkenes
of irreversible termination and transfer such as chain length, distribution of linked together by free radical addition
processes, and the absence of mediating monomers along the backbone, dispersity and propagation. Common monomers are
species to control the polymerization, and end-​group fidelity, among others. vinyl species M–CH = CH2 (here, M can
means that polymer growth affords a Additionally, CRP proceeds under relatively be any functional group) and substituted
statistical distribution of chain lengths, over mild conditions and is amenable to a broad derivatives, which end up forming the
which we have only limited control. The scope of vinyl monomers with different bulk of the polymer, with control over
heterogeneity in chain lengths corresponds functionalities. The two most popular CRP the reaction being afforded by adjusting the
to the width of the molecular weight methodologies are atom transfer radical reagents and mechanisms. Both reactions
distribution (MWD), which is typically polymerization (ATRP)2,3 and reversible can be run in a solvent to dissolve all the
quoted as dispersity Ð, the quotient of the addition–fragmentation chain-​transfer components (particularly the activator and
weight average molecular weight (Mw) (RAFT) polymerization4, and these were deactivator in ATRP) and reduce viscosity.
to the number average molecular weight first reported in 1995 and 1998, respectively. Alternatively, the neat reaction mixtures
(Mn). Due to our lack of control over FRP, The two CRP methods are ubiquitous, would predominantly comprise the liquid
the products typically have fairly high being the subject of over 24,000 publications alkene and would have some advantages
dispersities (Ð ~ 1.5–2), in contrast to the in the past two decades. The methods have in terms of higher concentrations and
many biological processes that produce many differences and, perhaps contrary polymerization rate. As we noted above,
perfectly uniform biomacromolecules, to common belief, many similarities that incontrast to FRP, both reactions feature an
such as DNA (Ð = 1). In FRP, all chains have revolutionized polymer chemistry equilibrium between dormant and active
undergo irreversible termination or transfer, and afforded products that were previously chains. This ensures continuous growth
such that, once a polymer stops growing, inaccessible5,6. and uniformity of polymer chains, while

Nature Reviews | CheMiSTRy volume 5 | December 2021 | 859

0123456789();:
Perspectives

suppressing the impact of unavoidable an initiating radical R• and the oxidized metal with a halogen atom X. The activator and
radical termination and transfer processes7. complex (the deactivator). Additionally, the deactivator complexes switch the polymers
Thus, intermittent activation of dormant activator must sustain polymerization by between active and dormant states.
chains to radicals occurs in both ATRP and continuing to transfer halogens from dormant
RAFT. Irreversible termination and transfer chains to activator complexes throughout ATRP initiation. The activator LCuiX
still occur during the ATRP and RAFT the polymerization. In ATRP, this halogen abstracts X from the initiator R–X (in
activation cycles. However, living chains atom transfer occurs through an inner-sphere principle, the halogen X on the activator and
derived from an ATRP initiator or a RAFT electron transfer (ISET)8, in which the that in the alkyl halide need not be the same),
chain-​transfer agent (CTA) greatly exceed activator and (macro)initiator are associated forming the initiating active radical R• and
the concentration of radicals and dead during the atom transfer. The deactivator the deactivator LCuiiX2. Realizing uniform
chains derived from irreversible transfer or is required to return the halogen X to the chain initiation requires that all initiators be
termination reactions. terminus of an active polymeric radical to activated rapidly and at a much higher rate
convert it into a dormant state and reform than propagation. In this way, each active
Conventional ATRP reagents the activator. As with activation, deactivation initiating radical R• will ideally react with
Let us now consider some of the components occurs through ISET8, with the radical and a first monomer molecule before longer
unique to ATRP. The initiator, most deactivator being associated during transfer. polymer chains start to form. These two
commonly an alkyl halide (denoted R–X), The ligand L chosen is typically a criteria can be met if sufficient activation and
starts the polymerization and offers the R chelating N-​donor, and tuning this is an deactivation are occurring, as it is possible
terminus from which polymer chains grow. important means to modulate the activity of that LCuiiX2 can quickly cap the initiating
Although metal-​free examples exist, ATRP the complex and its redox potentials9. It also radical R•, preventing it from adding too
typically involves a transition metal (usually solubilizes the metal species in the solvent many monomers in one activation cycle.
Cu) complex of an N-​donor ligand L, with or monomer. Described below is a typical To assist initiation, the initiator must be
the coordination sphere being such that the mechanism of ATRP conducted using such carefully chosen to match the activity of the
metal centre has two readily accessible a ligand L and Cu, with the complex either monomer and metal complex10.
oxidation states. When the complex is in the being preformed or generated in situ.
lower oxidation state, we call it an activator, Main ATRP equilibrium. When all radicals
and when in the higher oxidation state Conventional ATRP mechanism R• have reacted with at least one monomer
(for example, LCuiiX2), it is the deactivator. In ATRP, control over the molecular weight and only macroinitiators are present, we
The activator triggers polymerization by distribution occurs through reversible now have the main ATRP equilibrium
abstracting the halogen X from R–X to give deactivation (Fig. 1a) of the polymeric radical (Fig. 1a). Control over the MWD is possible

a Sufficient activation and deactivation b Sufficient degenerative chain transfer

ATRP initiation RAFT pre-equilibrium

M M
M
Activation
+ + + +
Deactivation

Initiator Activator Active Deactivator Active Dormant Intermediate Dormant Active

Activation S S S S S S
R X + LCuIX R + X LCuIIX Pm + R Pm R Pm + R
Deactivation
M M Z Z Z M

ATRP main equilibrium RAFT main equilibrium

M M M
Activation
+ + + +
Deactivation

Dormant Activator Active Deactivator Active Dormant Intermediate Dormant Active

Activation S S S S S S
Pn X + LCuIX Pn + X LCuIIX Pn + Pm Pn Pm Pn + Pm
Deactivation
M M Z Z Z M

Fig. 1 | Mechanisms of ATRP and RAFT. a | The reversible deactivation chain-​transfer agent (CTA, RʹS(S)CZ) with the macroCTA PnS(S)CZ, thus,
mechanism is involved in initiation and later on in the active–dormant allowing the polymer chain Pn to exist in dormant and active states. A sim-
equilibrium for a polymer chain Pn. The simplified representation of the plified mechanism of the main reversible addition–fragmentation
main atom transfer radical polymerization (ATRP) equilibrium involves chain-transfer (RAFT) equilibrium involves RAFT between the two polymer
transfer of a halogen atom X to a propagating polymer chain Pn. Overall, chains Pn and Pm. This method relies on sufficiently fast degenerative chain
it is important that activation and deactivation occur to sufficient extents. transfer — fast interconversion of the macroCTAs P nS(S)CZ and
b | During initiation, degenerative transfer interconverts the PmS(S)CZ.

860 | December 2021 | volume 5 www.nature.com/natrevchem

0123456789();:
 Perspectives

because of the reversible redox reaction that in which the activator is continuously (re) fragmentation of an initiator (for example,
enables efficient switching between active generated in situ through the reduction of azobisisobutyronitrile) and react with
and dormant species. In a similar vein to LCuiiX2. This can be achieved through a free monomer molecules to form propagating
initiation, the activator LCuiX now abstracts radical initiator as part of initiators for con- radicals Pn•. Radical generation occurs
X from a dormant macroinitiator (Pn–X, tinuous activator regeneration (ICAR), continuously throughout the reaction to
where n is the degree of polymerization) to a reducing agent for activators (re)generated compensate for termination.
form a propagating active radical Pn•, and the by electron transfer (A(R)GET)16, Cu0
deactivator LCuiiX2. In the reverse reaction, for supplemental activator and reducing RAFT pre-​equilibrium. Radicals generated
the radical can either be rapidly deactivated agent (SARA)17,18, and light, current and from the initiator enter a RAFT pre-​
by LCuiiX2 to yield the original dormant mechanical forces in photoinduced ATRP equilibrium, which quantitatively converts
halide-​capped macroinitiator (whence no (photoATRP)19–21, electrochemical the small-​molecule CTAs to oligomeric
polymer growth occurs) or can react with a ATRP (eATRP)22 and mechanically macroCTAs. Thus, initiator•, perhaps after
monomer (to enable polymer growth) before induced ATRP (mechanoATRP)23, respec- adding monomer units to give macroradical
deactivation occurs again to give a longer tively. The advantages of these advanced Pn•, adds to the small-​molecule CTA RʹS(S)
macroinitiator. Therefore, ATRP involves methods include the use of low catalyst CZ and forms the RAFT intermediate
only low radical concentrations, which loadings, the possibility for spatiotem- radical RʹS(PnS)C•Z, which fragments into
results in termination and propagation both poral control24,25 and higher end-​group PnS(S)CZ and a new chain Rʹ•. The role
being slow. The key to the ATRP equilibrium fidelity at higher monomer conversions26. of the small-​molecule CTA is to deactivate
is the sufficiently high deactivation rate of As we noted above, in addition to con- the propagating radical by transferring the
polymeric radicals, as well as the sufficiently ventional ATRP, there are also metal-​free radical from an active chain (Pn•) to a new
high activation rate of polymer chains, approaches in which the metal complexes chain (Rʹ•). Rʹ• then reacts with monomers
with both processes eventually being part are replaced by unsaturated organics such (in an ideal scenario with less than one
of a dynamic equilibrium. A high enough as 10-​phenylphenothiazine27 and diphenyl monomer, on average, per activation cycle)
deactivation rate ensures that each chain dihydrophenazine28. We now consider RAFT, to yield Pm• (refs6,31).
exists as an active radical for only a very which is always a metal-free reaction.
short period of time for each activation Main RAFT equilibrium. When all the
event and can, therefore, only react with a RAFT reagents small-​molecule CTA RʹS(S)CZ is consumed,
limited number of monomer molecules — The RAFT process requires an initiator to we get establishment of the RAFT main
just one in an ideal scenario. The sufficient provide a continuous source of radicals. equilibrium (Fig. 1b), in which the polymeric
activation rate also ensures that all chains Following this comes involvement of radical Pm• adds to the macroCTA PnS(S)
can be activated multiple times within the the CTA, typically a thiocarbonylthio CZ, where n and m are likely of a similar
experimental time frame. In combination, or thiocarbonylsulfanyl compound like value if a narrow MWD is desired. Again,
the sufficient activation and deactivation the dithioester RʹS(S)CZ (Fig. 1b), which via an intermediate radical, fragmentation
enable uniform growth10,11. In contrast facilitates chain transfer and enables efficient affords a new dormant macroCTA PmS(S)
to a common misconception, products switching between active and dormant CZ and an active chain Pn•. Therefore, RAFT
with broader MWDs yet high end-​group species. Analogous to ATRP, it is necessary relies on promoting chain transfer over
fidelity can also be obtained by lowering to match the activity of the CTA with propagation and termination, while the
the deactivator concentration within a that of the monomer to realize controlled radical concentration, in principle, remains
suitable range12. polymerization. The Z substituent in the constant throughout the polymerization.
CTA modulates RAFT reactivity and ensures The newly formed Pn• either reacts with a
Irreversible termination in ATRP. efficient addition of the propagating radical limited number of monomer molecules or
Despite the controlled nature of ATRP, to the CTA, by stabilizing the intermediate adds to a macroCTA. The key to controlling
termination can and will occur during radical. The CTA also has an organic Rʹ this reaction is sufficiently fast macroradical
the activation cycles (when radicals are group, which serves as the initiating end deactivation through addition to the
formed), mainly through radical coupling from which the majority of living polymers macroCTA, and a sufficient rate of activation
or disproportionation13. However, the grow. For efficient initiation, the radical Rʹ• through intermediate fragmentation. In
mediating species LCuiiX2 ensures that must be at least as good a leaving group as combination, this process leads to efficient
the radical concentration is sufficiently the propagating chain Pn•. It should be noted exchange between dormant macroCTA
low so the living chains greatly outnumber that vinyl-​terminated macromonomers species and active radicals. If control over
the dead ones. Surprisingly, the activator obtained from catalytic chain-​transfer the MWD is desired (for example, an
complex can also contribute to the polymerization can also be used as CTAs in application calls for a given high value of Ð),
irreversible termination of radicals via S-​free RAFT29,30. mixtures of low-​activity and high-​activity
LCuii–Pn organometallic intermediates, CTAs can be used without compromising
although this is often negligible14,15. Conventional RAFT mechanism the end-​group fidelity32.
To reiterate, the control over the MWD in
Alternative approaches for driving ATRP. RAFT is provided by a degenerative transfer Irreversible termination in RAFT. As with
Although conventional ATRP affords good mechanism that allows efficient switching FRP and ATRP, termination events in
control over polymer distributions, tradi- between active and dormant species. RAFT take the form of disproportionation
tional methods require near-​stoichiometric (for example, H atom abstraction between
loadings of activator (often referred to as RAFT initiation. Radical generation in two radical chains to afford one alkene and a
the ‘catalyst’). To circumvent this, parts per RAFT occurs the same way as it does in FRP saturated alkane), chain transfer or coupling.
million (ppm) methods have been developed — initiator• radicals are generated through However, the success of RAFT does not rely

Nature Reviews | CheMiSTRy volume 5 | December 2021 | 861

0123456789();:
Perspectives

on the termination rate itself being lowered, In photoinduced electron/energy transfer a method affords high-​molecular-​weight
as is the case for ATRP. Instead, efficient RAFT (PET-​RAFT)37,38, one photoexcites polymers to reflect the robustness of the
RAFT reactions see chain transfer competing a photocatalyst that relaxes through method. The highest molecular weight
with and dominating both termination and electron or energy transfer to generate Pn• possible is limited by the efficiency of the
propagation. Importantly, the number of from the macroCTA. Higher end-​group propagation and reversible deactivation
terminated chains can be predicted by the fidelity and spatiotemporal control are processes relative to side reactions such as
amount of free radical initiator used. the two most important advantages of irreversible termination and transfer. In addi-
these methods. In addition, electrochemical tion, high-​molecular-​weight polymers have
Alternative approaches to drive RAFT. methods (eRAFT)39 and sono-​RAFT40 use attracted considerable attention, owing to
Although initiators are most commonly electrical currents and mechanical forces desirable properties like high chain entan-
activated simply by heating, it is also in solution to split molecules into radicals glement and large domain size. Their utility
possible to use species activated by light, that can enter the RAFT equilibrium. The in various applications has been reviewed
electric current, mechanical forces or fundamental differences and similarities elsewhere41 and we, instead, consider
enzymatic processes33 to provide radicals between conventional ATRP and fundamental aspects here.
for RAFT. Photoiniferter RAFT is the conventional RAFT are listed in Table 1. Acrylates are arguably the easiest to
simplest photochemical process and involves polymerize by either ATRP or RAFT.
a light source that homolytically cleaves Monomer scope Their propagation rate constant kp is
the Pn–S(S)CZ bond in the macroCTA ATRP and RAFT are amenable to many typically high42, enabling us to make
to generate Pn• and the new radical •S(S)CZ monomer classes, the most widely investi- high-​molecular-​weight products under
(ref.34). In this process, the RAFT agent gated being acrylates, acrylamides, meth- mild conditions, including at room
itself acts as both the initiator and the acrylates and styrene (Fig. 2a). To identify temperature. For instance, Percec et al.
CTA. Although this concept was already optimal conditions for different monomer reported the synthesis of poly(methyl
mentioned in the first patent on RAFT classes, we focused on those that best afford acrylate) (PMA, Mn = 1,420,000, Đ = 1.15)
polymerization35, it has gained attention high-​molecular-​weight polymers, which are through a Cu0-​mediated approach17. The
recently because it allows the synthesis the most challenging materials to synthesize. record PMA weight was obtained using
of high-​molecular-​weight polymers34,36. Thus, we considered the ease with which RAFT (Mn = 5,520,000, Đ = 1.10) by means

Table 1 | A comparison of RAFT and ATRP

Consideration RAFT ATRP
How does reversible Degenerate transfer: Pn• + CTA–Pm ⇌ CTA–Pn + Pm• Atom transfer: Pn–X + activator ⇌ Pn• + deactivator
deactivation occur?
How are radicals Initiator2 → 2Initiator˙ R–X + LCuiX ⇌ R• + LCuiiX2
generated?
Macroradical Similar [Pn•] to conventional steady-​state radical Low [Pn•] due to activation–deactivation equilibrium. Low
concentration [Pn•] polymerization. Termination is not curbed; instead, chain rate of termination
transfer competes with both termination and propagation
Theoretical degree of DPRAFT = [Monomer]/[CTA] DPATRP = [Monomer]/[R–X]
polymerization (DP)
Obtaining narrow As well as sufficient deactivation, CTA must be consumed early As well as sufficient deactivation, R–X must be consumed
molecular weight in the reaction: rCT > rpropagation early in the reaction: rinitiation > rpropagation
distributions
Ensuring predictable CTA must be fully consumed for theoretical and experimental R–X must be fully consumed for theoretical and
molecular weights molecular weights to agree experimental molecular weights to agree
Maximizing The number of dead chains in conventional RAFT is End-​group fidelity enhanced by selecting a catalyst that
end-​group fidelity determined by number of initiators that have decomposed. favours deactivation over activation. This lowers [Pn•] and
Using low [Initiator2] gives higher end-​group fidelity rtermination
Ease of Simpler to implement, uses fewer components (monomer, More components: monomer, ATRP initiator, activator,
implementation radical initiator and CTA). The activity of the monomer must be deactivator and ligand. The activity of the monomer has to
matched with the CTA. Reactions are similar to conventional be matched with the activity of the initiator and ligand
radical polymerization with added CTA
Reaching high Radical generation continues until all radical initiator is Radical generation continues to occur even at 100%
monomer consumed. If 100% conversion is achieved but initiator is still conversion, potentially impacting end-​group fidelity
conversions present, then further termination can occur
pH tolerance Typically not tolerant to basic environments due to Few examples under acidic conditions due to protonation
degradation of CTAs at high pH of N-​donor ligands at low pH
Practical downsides Polymers are often yellow or pink in colour. CTAs often have a Polymers are often blue/green to brown in colour. Residual
bad odour, and polymers can require purification to avoid this metal species may be problematic for some applications
End group S Z X
R n–1 R n–1
M M S M M

ATRP, atom transfer radical polymerization; CTA, chain-​transfer agent; r, reaction rate; RAFT, reversible addition–fragmentation chain-​transfer.

862 | December 2021 | volume 5 www.nature.com/natrevchem

0123456789();:
 Perspectives

a More activated monomers Less activated monomers

Common monomers Acidic Coordinating

monomers monomers
Me/H

R R O O N
O O O N O O O N O OH O R O
N
R R R R

Acrylates Acrylamides Styrene Methacrylates Methacrylamides (Meth)acrylic acid Acrylonitrile Vinyl ketones Vinyl acetate N-vinyl pyrrolidone

RAFT:
C136 C134 C149 C236 C252 C231 C254 C356 C463 C465

ATRP:
L1, I117 L1, I246 L1, I348 L2, I447 L3, I651 L2, I567 L2, I753 L4, I261

b CTAs/RAFT agents c ATRP ligands d ATRP initiators

HO2C S S HO2C S S N O O O
N
Br Cl Br
S N N EtO MeO MeO
C 4H 9 N N
N N N
C1 C2 L1 L2 I1 I2 I3

O Cl Br
N N N Br
S S MeO S Br
N N HO
N N N O O
S O S N N
n-C12H25 OEt
C3 C4 L3 L4 I4 I5 I6 I7

Fig. 2 | Common monomers and reagents used in RAFT and ATRP. denote the commonly used RAFT and ATRP agents for that monomer class.
a | Representative monomer classes are shown, along with their amenability b | Different chain-​transfer agents (CTAs) give rise to intermediates that are
to reversible addition–fragmentation chain-​transfer (RAFT) and atom trans- stabilized, to different degrees, by extensive π-​delocalization, facilitating
fer radical polymerization (ATRP). A rating of four indicates that polymer­ different rates of chain transfer. c | ATRP makes extensive use of N-​donor
ization is highly controlled and is reported in numerous literature examples. chelating ligands to form redox-​active Cu complexes. d | The initiators used
On the other hand, a zero rating indicates that the technique is presently are typically alkyl halides that form resonance-​stabilized radicals on
not suited to polymerizing the monomer. Abbreviations below the ratings dissociation.

of a photoiniferter36. Although suitable give an inactive alcohol and aquation/ Đ = 1.11 at lower conversion)47. A similar
CTAs (Fig. 2b) and other parameters can dissociation of the LCuiiX2 deactivator can approach has also been used to make the
see RAFT give much higher molecular give free halide and copper species, which highest molecular weight polystyrene
weight polymers, both methods are are incapable of deactivating radicals43,44. (PS, Mn = 1,100,000, Đ = 1.24) yet reported48.
considered excellent in polymerizing To address these issues, one frequently uses With RAFT, the photoiniferter strategy also
acrylic moieties. Acrylamides also have slow (re)generation of LCuiX activators, alkyl generates PMMA with fairly high molecular
high kp values, and the highest molecular chloride initiators R–Cl (Fig. 2d) and high weight (Mn = 1,370,000, Đ = 1.35)36. RAFT
weight acrylamide-​based polymer to date deactivator or halide salt concentrations45. emulsion polymerization approaches have
is poly(dimethylacrylamide), with an Despite these improvements, reports on the proved most efficient for the polymerization
impressive Mn = 8,570,000 and Đ = 1.17 polymerization of acrylamides by ATRP of styrene (Mn = 1,043,000, Đ = 1.39)49,
(ref.34). Polymerization of acrylamides are more scarce than for RAFT, with the although pure PS homopolymer was not
by ATRP is more challenging, mostly highest molecular weight using ATRP synthesized because a lower molecular
due to side reactions associated with the being poly(hydroxyethyl acrylamide), with weight macroCTA was required to sustain
catalyst (Fig. 2c) and halogen end group Mn = 440,000 and Đ = 1.23 (ref.46). Lastly, we the emulsion producing PS block copolymer,
(Fig. 2d). ATRP reactions with ppm catalyst note that monomers such as methacrylates of which the macroCTA is the first block.
concentrations perform better with high- and styrene have lower kp values and seem What one notices is that synthesizing
activity catalysts (such as Cu complexes to give higher molecular weight products materials of high molecular weight
of L1 and L2) to ensure the deactivator is with ATRP, though RAFT also remains requires adaptations of the conventional
present in sufficiently high concentrations. very efficient. ATRP or RAFT methods. However, the
Conducting ATRP in aqueous solution The modified ATRP method ‘activators conventional methods are amenable to
is, unfortunately, complicated by several generated by electron transfer’ (AGET), polymerizations of methacrylamide and
obstacles. For example, the high activity when conducted under high pressure acrylonitrile, but, as we see, the molecular
of the LCuiX activators in H2O leads to (6  kbar) affords high-​molecular-​weight weights are not as high as those above. A Cu
excessive termination and high [Pn•], the poly(methyl methacrylate) (PMMA, pre-disproportionation protocol affords
alkyl halide can undergo hydrolysis to Mn = 1,900,000, Đ = 1.32 or Mn = 850,000, poly(hydroxypropyl methacrylamide)

Nature Reviews | CheMiSTRy volume 5 | December 2021 | 863

0123456789();:
Perspectives

a O O 1.2
O R1 O R2 O
Br Br Br 1.0
O O
O O O

w(log M) (a.u.)
x x y
CuI/Cu0 CuII/Cu0 0.8
O O O O O O 0.6
R1 R1 R2 0.4
0.2
O
R3 0
R1 = Me 1.5 2.0 2.5 3.0 3.5 4.0
O
log M (g mol–1)
R2 = nBu O
R4
PMA
R3 = Et O PMA-PBA-PEA
R4 = (CH2)2C(H)(Et)nBu O PMA-PBA-PEA-P2EHA-PEA
R3
PMA-PBA
O PMA-PBA-PEA-P2EHA
O PMA-PBA-PEA-P2EHA-PEA-PBA
R2
O

b
O
S S
HO 10 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
N O N O SnBu
N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O N O
O O O O O O O O O O O

PNAM10-b-PNAM3-b-PDEA3- 1.0 2.0

w(log M) (normalized)

8,000 Mn,SEC
b-PNAM3-b-PDMA3-b-PNAM3-
b-PDEA3-b-PNAM3-b-PDMA3- 0.8 Mn,th. 1.8

Mn,SEC,THF (g mol–1)
b-PDEA3-b-PNAM3-b-PDMA3- 6,000 Đ
0.6 1.6
b-PDEA3-b-PNAM3-b-PDMA3-

Đ
b-PDEA3-b-PNAM3-b-PDMA3- 4,000 1.4
0.4
b-PDEA3-b-PNAM3-b-PDMA3
0.2 2,000 1.2

0 0 1.0
200 2,000 20,000 0 2 4 6 8 10 12 14 16 18 20
Molecular weight (g mol ) –1
Number of blocks

c
A B C D E
O O O O O O O O O O
tBu nHex nDodecyl
CH2CF3 PEG O
O Br
PEG = (CH2CH2O)8–9Me O m n o p q
Br
O
O O O O O O O O O O
Sequential monomer addition tBu nHex nDodecyl
CH2CF3 PEG

1.0 Ni,theor = 24
0.8 A
dRI response
Normalized

0.6 B
C
0.4 D
E
0.2
0
10 12 14 16
Retention time (min)
d

PBMA275-b-PBzMA100-b-PtBMA100-b-PiBMA100-b-PHMA100-b-PEMA100-b-PMMA 100
MacroRAFT
1st PBMA275
2nd PBzMA100
N
O 3rd P/BMA100
S S 4th P/BMA100
HO C12H25
60 10 275 100 100 100 100 100 100 5th PHMA100
S
HO O O O O O O O O O O O O O O O O O 6th PEMA100
7th PMMA100
Ph
R group from R X Z group from R′S(S)CZ
3 4 5 6
log M (g mol–1)

864 | December 2021 | volume 5 www.nature.com/natrevchem

0123456789();:
 Perspectives

◀ Fig. 3 | Synthesis of complex multiblock copoly­mers with ATRP and RAFT. a | High-order In ATRP, multiblocks can be prepared by
poly(acrylate) multiblock copolymers are avail­able using Cu0-​mediated radical polymerization76. ‘low-​ppm’ methods, such as photo-​ATRP75
b | Sequence-controlled poly(acrylamide)s can form in rapid and quantitative one-pot aque­ous revers- or Cu0-​mediated76 strategies, predominately
ible addition–fragmentation chain-transfer (RAFT) polymerization78. c | ABCDE multiblock copolymers using acrylates as model monomers (Fig. 3a),
with hydrophobic, hydrophilic and semi-​fluorinated segments can be made in one pot using
with the highest number of acrylic blocks
light-mediated atom transfer radical polymerization (ATRP)75. d | Poly(methacrylate) multiblock
copolymers with different ester groups have been made using macroRAFT-​mediated emulsion poly­
achieved being 11 (Đ = 1.19)77. Although
merization80. Part a adapted with permission from ref.76, ACS. Part b adapted from ref.78, Springer reaching a similar level of control with
Nature Limited. Part c adapted with permission from ref.75, Wiley. Part d adapted with permission RAFT was considered challenging due
from ref.80, ACS. to the necessity of adding additional
free radical initiator at each step, using
high kp monomers (acrylamides) in H2O
(PHPMA, Mn = 99,200; Đ = 1.23)50,51, By contrast, while acidic solutions aid affords impressive icosablock copolymers
while an equally well-​controlled RAFT RAFT, it is hampered when run in very (20-​block species)78 (Fig. 3b). ATRP has
reaction affords a lighter product slate basic conditions, under which typical CTAs not been able to give similarly high block
(Mn = 35,800, Đ = 1.26)52. The controlled (Fig. 2b) are degraded31,69. numbers, perhaps because aqueous
polymerization of acrylonitrile through acrylamide polymerization is subject to side
activators regenerated by electron transfer Block copolymers reactions under these conditions43 and the
(ARGET) ATRP16 affords a relatively narrow Both RAFT and ATRP have been used to deactivator gradually accumulates because
MWD (Mn = 161,300, Đ = 1.18) albeit after a make block copolymers, structures that are of the low concentration of terminated
reaction time of 288 h (ref.53). In comparison, routinely used in bulk and solution self- chains, eventually leading to cessation of
RAFT gives polyacrylonitrile more quickly, assembly, surfactants and other applications70. polymerization79. Nevertheless, ATRP holds
without a large broadening of the MWD The vast majority of block copolymers the record for the most versatile multiblock
(Mn = 405,100, Đ = 1.32)54. consist of monomers of the same family copolymer consisting of hydrophobic,
For all other monomer families, RAFT (for example, poly(acrylate)-​b-poly(acrylate), hydrophilic and semi-​fluorinated segments75
has a clear advantage over ATRP (Fig. 2a). poly(methacrylate)-​b-poly(methacrylate)) (Fig. 3c), while RAFT dominates in producing
In particular, monomers such as vinyl or with decreasing radical stability (for multiblocks with the highest number of
ketones, which coordinate to metal ions like example, poly(methacrylate)-​b-​poly(styrene) blocks and number of blocks each derived
Cuii/i, are presently impossible to polymerize or poly(methacrylate)-​b-​poly(acrylate)). for a different monomer family80,81 (Fig. 3b,d).
using ATRP55. By contrast, RAFT affords In the latter case, the radical intermediates Interestingly, recently established low-​ppm
poly(vinyl methyl ketone) at high involved in the second block are less stable ATRP methods have substantially enhanced
conversions and relatively low dispersity than those involved in the first71. Thus, the its synthetic potential for block copolymers,
(Mn = 41,400, Đ = 1.27)56–58. Very-​low-​activity order of monomer addition is crucial for while RAFT methods such as PET-​RAFT
non-​conjugated monomers, such as efficient block copolymer synthesis, with or photoiniferter RAFT have not yet been
vinyl acetate, N-​vinyl pyrrolidone and block copolymers being synthesized from the explored for the synthesis of high-​order
N-​vinyl carbazole, are extremely challenging most stable/active macroradical first, towards multiblocks37.
to polymerize by ATRP. This is mainly least stable/active as the last block. If one
attributed to the lack of sufficiently active wishes to prepare a material with the reverse End-​group modification and removal
ATRP catalysts, although recent progress order, one must turn to elegant methods that The high ‘livingness’ achieved in ATRP
to overcome this has produced new ligands involve either halogen exchange72 (ATRP) and RAFT (evidenced, for example, by
that afford highly active Cu complexes. or iniferter strategies (RAFT)73. Based on linear plots of first-​order kinetics and
Nevertheless, the polymerization of vinyl its mechanism, ATRP is expected to pro- Mn versus conversion) offers additional
acetate has yet to be reported59,60 and reports duce purer block copolymers compared opportunities through end-​group
of poly(N-​vinyl pyrrolidone) are limited61. with RAFT, because, in the latter case, the modification and removal strategies
Unfortunately, side reactions make the free radical initiator continu­ously generates (Fig. 4). Post-​polymerization modification
polymerization of these monomers through radicals throughout the polymerization, is required whenever polymerization
RAFT also not ideal62, but the method is such that high-​order blocks will be contami- and functionalization chemistries are
a controlled means of making poly(vinyl nated by a small amount of homopolymer74. incompatible82. End-​group modification
acetate) (Mn = 38,200, Đ = 1.23)63 and RAFT polymers are also expected to have methods are routinely used for subsequent
poly(N-​vinyl pyrrolidone) (Mn = 16,900, lower α-​end-​group purity, since the free chemistry and enable applications such
Đ = 1.25)64,65, and even polyethylene66. radical initiator adds new α-​terminal groups as bioconjugate synthesis83, surface
Conducting ATRP in acidic solutions is throughout the polymer­ization. To investi- functionalization84 and cyclic polymer
challenging because the N-​donor ligands gate whether ATRP or RAFT can yield better synthesis85. By contrast, end-​group removal
used to bind Cu are also good Brønsted block copoly­mers, it is necessary to examine strategies are often used to remove colour
bases. Moreover, if carboxylic acids are recent developments in sequence-controlled (for example, from RAFT polymers)69
present, the conjugate bases can displace the multiblock copolymers, where the or increase the stability of the resulting
halido ligands X− from the metal catalyst. monomers are added sequentially at materials (for example, those made by
This makes the polymerization of acidic near-quantitative conversions (>95%). These ATRP)86. In ATRP, nucleophilic substitution
monomers very difficult, although notable multiblock copoly­mer syn­theses require is perhaps the most popular end-​group
steps to overcome this have already been exceptional polymerization efficiency and modification strategy. The high reactivity
implemented, with high polymerization livingness, and can highlight the strengths of the alkyl halide end group enables
rates and alkyl chloride initiators offering and limitations of a given polymerization fast and quantitative reactions between
good results (Mn = 23,700, Đ = 1.25)67,68. approach. the alkyl halide and any thiol through

Nature Reviews | CheMiSTRy volume 5 | December 2021 | 865

0123456789();:
Perspectives

a End-group removal End-group functionalization b End-group removal End-group functionalization

SH S
R R n–1
H R R′
n–1 R n–1
n–1
M M Δ RNH2 M M nBu R′SH M M
M M 3SnH
Thermolysis Aminolysis to thiol Initiator2
Hydrodehalogenation: Thiol–halogen
radical-induced nucleophilic substitution
Initiator RAFT AIBN reduction ATRP
R n–1 Initiator2 Air
S Z OH X R R
M M R R R R
n–1 PPh3 n–1 n–1 n–1
Radical addition M M S M M M M M M
fragmentation Radical coupling
coupling Initiator2 H2 NaN3
[4+2]
Pd/C
H H atom Z
R S H
n–1 donor N3
R R n–1
M M n–1 R n–1
M M S M M
Radical-induced M M
reduction Hydrodehalogenation: Huisgen
Hetero-Diels–Alder chain-end reduction cycloaddition

Fig. 4 | There are diverse methods to remove or functionalize an end group. a | The thiocarbonyl group from reversible addition–fragmentation
chain-​transfer (RAFT) can undergo several different reactions. b | Likewise, the halogen X resulting from atom transfer radical polymerization (ATRP) is
readily replaced by H or another nucleophile.

thio–bromo ‘click’ chemistry, providing CZ terminus can also be converted to an a spin-​off of the CSIRO, and specific RAFT
for a wide substrate scope87,88. Another alcohol –OH by a quantitative two-​step agents are produced industrially by Solvay,
widely used nucleophilic substitution process100. Taken altogether, ATRP and Arkema and Lubrizol. Importantly, the
strategy is azidation, whereby the halogen RAFT present complementary chemistries, first CSIRO and Rhodia patents reached
at a terminus is replaced with an azido depending on the targeted functionality. the end of their enforceable life in 2018,
group by nucleophilic displacement with Despite these developments, the alkyl so we expect the industrial applicability of
NaN3 (refs89,90). The azido group serves halide or thiocarbonyl end groups may RAFT to increase. Currently, >1,000 patents
as a versatile precursor for subsequent be considered undesirable for certain have been filed by various companies,
Huisgen cycloadditions (alkyne–azide click applications, which have motivated the including DuPont, Solvay, Arkema, Lubrizol
reactions)91. A third nucleophilic substitution development of the end-​group removal and L’Oréal. For instance, one of the very
methodology is the acid–bromo reaction, strategies we have described. After an ATRP first commercial applications of RAFT
which can generate macromonomers reaction, the alkyl halide end groups can afforded a paint, developed by Dulux
from ATRP polymers92. Other useful but be dehalogenated by using either Pd-​C/H2 Australia, with improved flow, adhesion,
less frequently used strategies include the or nBu3SnH (ref.101). Following RAFT, the abrasion resistance and durability. Another
amine–bromo reaction (although suffering end groups can be removed by thermolysis, commercial example is Lubrizol’s Asteric
from side reactions), the reaction with radical-​induced reduction and radical range of viscosity modifiers that consist of
phosphines (quantitative) and radical addition–fragmentation coupling69. methacrylic star polymers. Similarly, ATRP
transformation approaches in which the has also attracted interest from industry,
ATRP end group can be replaced by end Industrial implementation with over 1,500 patents filed to date by a
groups resulting from RAFT (–S(S)CZ) or RAFT has, in principle, an easier pathway wide range of companies, including Arkema,
nitroxide-​mediated polymerization (–ONR2, to industrial implementation, mainly due to L’Oréal, BYK, Evonik, Dionex, PPG, Mitsui,
>90% efficiency)93,94. RAFT’s resemblance with conventional Ciba, Unilever and Kaneka. The latter was
The end group of RAFT products are FRP102. RAFT uses identical components the first company to industrialize ATRP
most commonly modified by reaction to FRP, apart from the RAFT agent, such on a metric ton scale in a process that gave
with nucleophiles69. The thiocarbonylthio that no major issues are expected when polyacrylates with crosslinkable end groups.
group –S(S)CZ can be converted to a thiol scaling up RAFT to reactors presently The same company uses ATRP to produce
–SH in the presence of primary amines, used for FRP. Initial obstacles to RAFT’s eight polymers with controlled molecular
secondary amines, OH− or reducing agents. further implementation include the cost weight, low dispersity and high end-​group
The thiol can undergo a range of very and limited availability of RAFT agents, the functionality. These materials outperform
efficient reactions, including thiol–ene, need to chemically remove the end group standard silicones in adhesives and coatings.
thiol–yne, thiol–halide, thiol–isocyanate and the foundation intellectual property At first, scaling up ATRP was challenging
and thiol–Michael click reactions95, but patent by the Commonwealth Scientific and due to the high catalyst concentrations
care should be taken to avoid oxidative Industrial Research Organisation (CSIRO). required and the sensitivity of the catalysts
disulfide formation. The hetero-​Diels– Another patent, on xanthate designs as to even trace amounts of O2. However, the
Alder reaction is another powerful click CTAs, is held by the French National Centre development of low-​ppm methods such
technique, and, here, the RAFT end group for Scientific Research (Centre national as ARGET or ICAR ATRP has addressed
can serve as a dienophile96,97. This reaction de la recherche scientifique, CNRS) and these issues and seen several companies,
is particularly useful for the synthesis of Rhodia Chimie (now Solvay). At present, including Kaneka Corporation, PPG
block, grafted and star polymers, as well industrial quantities (tonnes) of RAFT Industries, Thermo Fisher Scientific and
as for surface modification98,99. The –S(S) agents are available from Boron Molecular, ATRP Solutions, successfully scale up ATRP.

866 | December 2021 | volume 5 www.nature.com/natrevchem

0123456789();:
 Perspectives

Perhaps the most vexing remaining barrier of monomers and molecular weights12,117,118. atom transfer radical polymerisation: monomodal
distributions with ppm copper concentration. Angew.
to further commercialization is the cost to In a similar fashion, tacticity control for CRP Chem. Int. Ed. 58, 13323–13328 (2019).
performance ratio of the polymer products. products is rather limited119,120 and requires 13. Ribelli, T. G. et al. Disproportionation or combination?
The termination of acrylate radicals in ATRP.
Polymers made by CRP clearly have specific monomers/reagents. To further Macromolecules 50, 7920–7929 (2017).
advantageous characteristics and functions, aid both reproducible polymer synthesis 14. Ribelli, T. G., Wahidur Rahaman, S. M., Krys, P.,
Matyjaszewski, K. & Poli, R. Effect of ligand structure
but this performance comes with added and commercialization, we should also on the Cuii–R OMRP dormant species and its
financial costs, suggesting that RAFT and further investigate efficient automation and consequences for catalytic radical termination
in ATRP. Macromolecules 49, 7749–7757 (2016).
ATRP polymers are best suited to high-​value machine-​learning strategies121. Last but not 15. Xie, G. et al. Benefits of catalyzed radical termination:
products. Readers interested in these costs least, recyclable RDRP methodologies that high-​yield synthesis of polyacrylate molecular
bottlebrushes without gelation. Macromolecules 51,
and industrial implementations are well can quantitatively regenerate the starting 6218–6225 (2018).
served by a recent review on the subject102. materials (that is, monomers, RAFT 16. Matyjaszewski, K. et al. Diminishing catalyst
concentration in atom transfer radical polymerization
agents and ATRP initiators) have been with reducing agents. Proc. Natl Acad. Sci. USA 103,
Future outlook underdeveloped122 and should be a future 15309–15314 (2006).
17. Percec, V. et al. Ultrafast synthesis of ultrahigh molar
RAFT and ATRP have been used in many focus aiming at a circular polymer economy. mass polymers by metal-​catalyzed living radical
applications and developments, such as Addressing these challenges will further polymerization of acrylates, methacrylates, and vinyl
chloride mediated by SET at 25 °C. J. Am. Chem. Soc.
polymer–protein bioconjugation83,103–106, expand the scope of materials available by 128, 14156–14165 (2006).
cell engineering107,108, synthesis of CRP and enable further commercialization. 18. Anastasaki, A. et al. Cu(0)-mediated living radical
polymerization: a versatile tool for materials synthesis.
polymeric nanoparticles with controlled Nghia P. Truong   1,3, Glen R. Jones   1,3, Chem. Rev. 116, 835–877 (2016).
characteristics109–111 and 3D printing112,113. Kate G. E. Bradford   2, Dominik Konkolewicz   2 ✉ 19. Anastasaki, A. et al. Copper(ii)/tertiary amine synergy
in photoinduced living radical polymerization:
Introducing external stimuli19,20,22,23,27,28, such and Athina Anastasaki1 ✉
accelerated synthesis of ω-​functional and α,ω-​
as light, has further expanded the application 1
Laboratory for Polymeric Materials, Department of heterofunctional poly(acrylates). J. Am. Chem. Soc.
136, 1141–1149 (2014).
pool, creating additional opportunities. Materials, ETH Zürich, Zürich, Switzerland.
20. Ribelli, T. G., Konkolewicz, D., Bernhard, S. &
Both methods have the potential to 2
Department of Chemistry and Biochemistry, Miami Matyjaszewski, K. How are radicals (re)generated
University, Oxford, OH, USA. in photochemical ATRP? J. Am. Chem. Soc. 136,
continue transforming polymer chemistry. 13303–13312 (2014).
However, to realize their potential and allow 3
These authors contributed equally: Nghia P. Truong, 21. Konkolewicz, D., Schröder, K., Buback, J., Bernhard, S.
Glen R. Jones. & Matyjaszewski, K. Visible light and sunlight
further commercialization, we first need to photoinduced ATRP with ppm of Cu catalyst. ACS
✉e-​mail: [email protected];
overcome several challenges. For example, Macro Lett. 1, 1219–1223 (2012).
[email protected] 22. Magenau, A. J. D., Strandwitz, N. C., Gennaro, A. &
neither ATRP nor RAFT offer full control Matyjaszewski, K. Electrochemically mediated atom
https://doi.org/10.1038/s41570-021-00328-8
over polymerization (give low Ð with high transfer radical polymerization. Science 332, 81–84
Published online 18 October 2021 (2011).
end-​group fidelity) of important commodity 23. Mohapatra, H., Kleiman, M. & Esser-​Kahn, A. P.
monomers (for example, ethylene66 and 1. Staudinger, H. Über polymerisation. Ber. Dtsch. Mechanically controlled radical polymerization
initiated by ultrasound. Nat. Chem. 9, 135–139
propylene), sterically hindered moieties114 Chem. Ges. 53, 1073–1085 (1920).
2. Kato, M., Kamigaito, M., Sawamoto, M. & (2017).
(for example, ethyl α-​ethylacrylate), Higashimura, T. Polymerization of methyl methacrylate 24. Leibfarth, F. A., Mattson, K. M., Fors, B. P., Collins, H. A.
cyclic monomers or monomers that yield with the carbon tetrachloride/dichlorotris- & Hawker, C. J. External regulation of controlled
polymerizations. Angew. Chem. Int. Ed. 52, 199–210
(triphenylphosphine)ruthenium(ii)/methylaluminum
conjugated polymers115. In addition, bis(2,6-di-​tert-butylphenoxide) initiating system: (2013).
the synthesis of high-​molecular-​weight possibility of living radical polymerization. 25. Poelma, J. E., Fors, B. P., Meyers, G. F., Kramer, J. W.
Macromolecules 28, 1721–1723 (1995). & Hawker, C. J. Fabrication of complex three-​
polymers consisting of lower reactivity 3. Wang, J.-S. & Matyjaszewski, K. Controlled/“living” dimensional polymer brush nanostructures through
monomers like vinyl acetate is currently not radical polymerization. atom transfer radical light-​mediated living radical polymerization. Angew.
polymerization in the presence of transition-​metal Chem. Int. Ed. 52, 6844–6848 (2013).
possible. High-​molecular-​weight materials complexes. J. Am. Chem. Soc. 117, 5614–5615 26. Wang, Y. et al. Improving the “livingness” of ATRP by
are also cost-​effective because they require (1995). reducing Cu catalyst concentration. Macromolecules
4. Chiefari, J. et al. Living free-​radical polymerization 46, 683–691 (2013).
very small amounts of catalyst, initiator or by reversible addition–fragmentation chain transfer: 27. Treat, N. J. et al. Metal-​free atom transfer radical
RAFT agent for a given mass of product. the RAFT process. Macromolecules 31, 5559–5562 polymerization. J. Am. Chem. Soc. 136, 16096–16101
(1998). (2014).
For ATRP, the development of thermally 5. Matyjaszewski, K. & Xia, J. Atom transfer radical 28. Theriot, J. C. et al. Organocatalyzed atom transfer
activated metal-​free catalysts would be polymerization. Chem. Rev. 101, 2921–2990 (2001). radical polymerization driven by visible light. Science
6. Moad, G., Rizzardo, E. & Thang, S. H. Living radical 352, 1082–1086 (2016).
highly beneficial and complement present polymerization by the RAFT process. Aust. J. Chem. 29. Moad, C. L., Moad, G., Rizzardo, E. & Thang, S. H.
light-​driven strategies27. For RAFT, the 58, 379–410 (2005). Chain transfer activity of ω-​unsaturated methyl
7. Konkolewicz, D., Krys, P. & Matyjaszewski, K. methacrylate oligomers. Macromolecules 29,
development of new S-​free RAFT agents Explaining unexpected data via competitive equilibria 7717–7726 (1996).
with higher chain-​transfer rate constants and processes in radical reactions with reversible 30. Engelis, N. G. et al. Sequence-​controlled methacrylic
deactivation. Acc. Chem. Res. 47, 3028–3036 multiblock copolymers via sulfur-​free RAFT
and a wider monomer scope29,30 may also (2014). emulsion polymerization. Nat. Chem. 9, 171–178
be an interesting direction, in particular, 8. Lin, C. Y., Coote, M. L., Gennaro, A. & Matyjaszewski, K. (2017).
Ab initio evaluation of the thermodynamic and 31. Perrier, S. 50th anniversary perspective: RAFT
for applications where traces of odour electrochemical properties of alkyl halides and radicals polymerization — a user guide. Macromolecules 50,
and colour are undesirable. Designing and their mechanistic implications for atom transfer 7433–7447 (2017).
radical polymerization. J. Am. Chem. Soc. 130, 32. Whitfield, R., Parkatzidis, K., Truong, N. P., Junkers, T.
inexpensive universal catalytic/synthetic 12762–12774 (2008). & Anastasaki, A. Tailoring polymer dispersity by
systems that controllably give polymers 9. Tang, W. et al. Understanding atom transfer radical RAFT polymerization: a versatile approach. Chem 6,
polymerization: effect of ligand and initiator structures 1340–1352 (2020).
from diverse monomer classes would also be on the equilibrium constants. J. Am. Chem. Soc. 130, 33. Zhang, B. et al. Enzyme-​initiated reversible addition–
very useful. Early work on switchable RAFT 10702–10713 (2008). fragmentation chain transfer polymerization.
10. Tang, W. & Matyjaszewski, K. Effects of initiator Macromolecules 48, 7792–7802 (2015).
agents could assist efforts in this direction116. structure on activation rate constants in ATRP. 34. Carmean, R. N., Becker, T. E., Sims, M. B. &
Despite recent developments, accessing Macromolecules 40, 1858–1863 (2007). Sumerlin, B. S. Ultra-​high molecular weights via
11. Fang, C. et al. Mechanistically guided predictive models aqueous reversible-​deactivation radical polymerization.
polymers of truly tunable dispersities has not for ligand and initiator effects in copper-catalyzed Chem 2, 93–101 (2017).
been achieved, as we cannot, at will, change atom transfer radical polymerization (Cu-ATRP). 35. Rizzardo, E. & Moad, G. Synthesis of dithioester chain
J. Am. Chem. Soc. 141, 7486–7497 (2019). transfer agents and use of bis(thioacyl) disulfides
from getting perfectly uniform polymers to 12. Whitfield, R., Parkatzidis, K., Rolland, M., Truong, N. P. or dithioesters as chain transfer agents. Patent
much higher dispersity materials for a range & Anastasaki, A. Tuning dispersity by photoinduced WO/1999/005099 (1999).

Nature Reviews | CheMiSTRy volume 5 | December 2021 | 867

0123456789();:
Perspectives

36. Carmean, R. N. et al. Ultrahigh molecular weight 60. Ribelli, T. G. et al. Synthesis and characterization protein conjugation “pegylation”. J. Am. Chem. Soc.
hydrophobic acrylic and styrenic polymers through of the most active copper ATRP catalyst based on 126, 13220–13221 (2004).
organic-​phase photoiniferter-​mediated polymerization. tris[(4-dimethylaminopyridyl)methyl]amine. J. Am. 84. Pauloehrl, T. et al. Adding spatial control to click
ACS Macro Lett. 9, 613–618 (2020). Chem. Soc. 140, 1525–1534 (2018). chemistry: phototriggered Diels–Alder surface (bio)
37. Xu, J., Jung, K., Atme, A., Shanmugam, S. & Boyer, C. 61. Lu, X. et al. Controllable synthesis of functionalization at ambient temperature. Angew.
A robust and versatile photoinduced living poly(N-vinylpyrrolidone) and its block copolymers Chem. Int. Ed. 51, 1071–1074 (2012).
polymerization of conjugated and unconjugated by atom transfer radical polymerization. Polymer 85. Laurent, B. A. & Grayson, S. M. An efficient route to
monomers and its oxygen tolerance. J. Am. Chem. Soc. 48, 2835–2842 (2007). well-​defined macrocyclic polymers via “click” cyclization.
136, 5508–5519 (2014). 62. Harrisson, S. et al. RAFT polymerization of vinyl J. Am. Chem. Soc. 128, 4238–4239 (2006).
38. Allegrezza, M. L. & Konkolewicz, D. PET-​RAFT esters: synthesis and applications. Polymers 6, 86. Altintas, O. et al. ATRP-​based polymers with modular
polymerization: mechanistic perspectives for future 1437–1488 (2014). ligation points under thermal and thermomechanical
materials. ACS Macro Lett. 10, 433–446 (2021). 63. Ding, C. et al. Photocatalyst-​free and blue light-​ stress. Polym. Chem. 6, 2854–2868 (2015).
39. Lorandi, F. et al. Toward electrochemically mediated induced RAFT polymerization of vinyl acetate at 87. Anastasaki, A., Willenbacher, J., Fleischmann, C.,
reversible addition–fragmentation chain-​transfer ambient temperature. Macromol. Rapid Commun. 36, Gutekunst, W. R. & Hawker, C. J. End group
(eRAFT) polymerization: can propagating radicals 2181–2185 (2015). modification of poly(acrylates) obtained via ATRP:
be efficiently electrogenerated from RAFT agents? 64. Stace, S. J., Moad, G., Fellows, C. M. & Keddie, D. J. a user guide. Polym. Chem. 8, 689–697 (2017).
Macromolecules 52, 1479–1488 (2019). The effect of Z-​group modification on the RAFT 88. Huang, X., Nakagawa, S., Li, X., Shibayama, M.
40. McKenzie, T. G., Colombo, E., Fu, Q., Ashokkumar, M. polymerization of N-​vinylpyrrolidone controlled by & Yoshie, N. A simple and versatile method for
& Qiao, G. G. Sono-​RAFT polymerization in aqueous “switchable” N-​pyridyl-functional dithiocarbamates. the construction of nearly ideal polymer networks.
medium. Angew. Chem. Int. Ed. 56, 12302–12306 Polym. Chem. 6, 7119–7126 (2015). Angew. Chem. Int. Ed. 59, 9646–9652 (2020).
(2017). 65. Guinaudeau, A., Mazières, S., Wilson, D. J. & 89. Lutz, J.-F., Börner, H. G. & Weichenhan, K. Combining
41. An, Z. 100th anniversary of macromolecular science Destarac, M. Aqueous RAFT/MADIX polymerisation ATRP and “click” chemistry: a promising platform
viewpoint: achieving ultrahigh molecular weights of N-​vinyl pyrrolidone at ambient temperature. toward functional biocompatible polymers and polymer
with reversible deactivation radical polymerization. Polym. Chem. 3, 81–84 (2012). bioconjugates. Macromolecules 39, 6376–6383
ACS Macro Lett. 9, 350–357 (2020). 66. Dommanget, C., D’Agosto, F. & Monteil, V. (2006).
42. Asua, J. M. et al. Critically evaluated rate coefficients Polymerization of ethylene through reversible 90. Adzima, B. J. et al. Spatial and temporal control of the
for free-​radical polymerization, 5. Macromol. Chem. addition–fragmentation chain transfer (RAFT). alkyne–azide cycloaddition by photoinitiated Cu(ii)
Phys. 205, 2151–2160 (2004). Angew. Chem. Int. Ed. 53, 6683–6686 (2014). reduction. Nat. Chem. 3, 256–259 (2011).
43. Jones, G. R. et al. Copper-​mediated reversible 67. Fantin, M., Isse, A. A., Venzo, A., Gennaro, A. & 91. Kolb, H. C., Finn, M. G. & Sharpless, K. B. Click
deactivation radical polymerization in aqueous media. Matyjaszewski, K. Atom transfer radical polymerization chemistry: diverse chemical function from a few good
Angew. Chem. Int. Ed. 57, 10468–10482 (2018). of methacrylic acid: a won challenge. J. Am. Chem. Soc. reactions. Angew. Chem. Int. Ed. 40, 2004–2021
44. Averick, S. et al. ATRP under biologically relevant 138, 7216–7219 (2016). (2001).
conditions: grafting from a protein. ACS Macro Lett. 1, 68. Lorandi, F. et al. Atom transfer radical polymerization 92. Renterghem, L. M. V. et al. Design and use of organic
6–10 (2012). of acrylic and methacrylic acids: preparation of acidic nanoparticles prepared from star-​shaped polymers
45. Simakova, A., Averick, S. E., Konkolewicz, D. polymers with various architectures. ACS Macro Lett. with reactive end groups. J. Am. Chem. Soc. 130,
& Matyjaszewski, K. Aqueous ARGET ATRP. 9, 693–699 (2020). 10802–10811 (2008).
Macromolecules 45, 6371–6379 (2012). 69. Willcock, H. & O’Reilly, R. K. End group removal 93. Wager, C. M., Haddleton, D. M. & Bon, S. A. A simple
46. Appel, E. A., Del Barrio, J., Loh, X. J., Dyson, J. & and modification of RAFT polymers. Polym. Chem. 1, method to convert atom transfer radical polymerization
Scherman, O. A. High molecular weight polyacrylamides 149–157 (2010). (ATRP) initiators into reversible addition fragmentation
by atom transfer radical polymerization: enabling 70. Bates, C. M. & Bates, F. S. 50th anniversary chain-​transfer (RAFT) mediators. Eur. Polym. J. 40,
advancements in water-​based applications. J. Polym. perspective: block polymers — pure potential. 641–645 (2004).
Sci. A Polym. Chem. 50, 181–186 (2012). Macromolecules 50, 3–22 (2017). 94. Kulis, J., Bell, C. A., Micallef, A. S., Jia, Z. &
47. Kwiatkowski, P. et al. High molecular weight 71. Keddie, D. J. A guide to the synthesis of block Monteiro, M. J. Rapid, selective, and reversible
polymethacrylates by AGET ATRP under high pressure. copolymers using reversible-​addition fragmentation nitroxide radical coupling (NRC) reactions at ambient
Macromolecules 41, 1067–1069 (2008). chain transfer (RAFT) polymerization. Chem. Soc. Rev. temperature. Macromolecules 42, 8218–8227
48. Mueller, L. et al. Synthesis of high molecular weight 43, 496–505 (2014). (2009).
polystyrene using AGET ATRP under high pressure. 72. Peng, C.-H., Kong, J., Seeliger, F. & Matyjaszewski, K. 95. Harvison, M. A. & Lowe, A. B. Combining RAFT radical
Eur. Polym. J. 47, 730–734 (2011). Mechanism of halogen exchange in ATRP. polymerization and click/highly efficient coupling
49. Truong, N. P., Dussert, M. V., Whittaker, M. R., Macromolecules 44, 7546–7557 (2011). chemistries: a powerful strategy for the preparation
Quinn, J. F. & Davis, T. P. Rapid synthesis of ultrahigh 73. Easterling, C. P., Xia, Y., Zhao, J., Fanucci, G. E. & of novel materials. Macromol. Rapid Commun. 32,
molecular weight and low polydispersity polystyrene Sumerlin, B. S. Block copolymer sequence inversion 779–800 (2011).
diblock copolymers by RAFT-​mediated emulsion through photoiniferter polymerization. ACS Macro 96. Inglis, A. J., Sinnwell, S., Stenzel, M. H. &
polymerization. Polym. Chem. 6, 3865–3874 (2015). Lett. 8, 1461–1466 (2019). Barner-Kowollik, C. Ultrafast click conjugation
50. Zhang, Q. et al. Aqueous copper-​mediated living 74. Gody, G., Zetterlund, P. B., Perrier, S. & Harrisson, S. of macromolecular building blocks at ambient
polymerization: exploiting rapid disproportionation The limits of precision monomer placement in chain temperature. Angew. Chem. Int. Ed. 48, 2411–2414
of CuBr with Me6TREN. J. Am. Chem. Soc. 135, growth polymerization. Nat. Commun. 7, 10514 (2009).
7355–7363 (2013). (2016). 97. Hansell, C. F. et al. Additive-​free clicking for polymer
51. Raus, V. & Kostka, L. Optimizing the Cu-​RDRP of 75. Anastasaki, A. et al. One-​pot synthesis of ABCDE functionalization and coupling by tetrazine–norbornene
N-(2-hydroxypropyl) methacrylamide toward multiblock copolymers with hydrophobic, hydrophilic, chemistry. J. Am. Chem. Soc. 133, 13828–13831
biomedical applications. Polym. Chem. 10, 564–568 and semi-​fluorinated segments. Angew. Chem. Int. Ed. (2011).
(2019). 56, 14483–14487 (2017). 98. Nebhani, L. et al. Efficient surface modification
52. Pan, X. et al. Effect of solvents on the RAFT 76. Soeriyadi, A. H., Boyer, C., Nyström, F., Zetterlund, P. B. of divinylbenzene microspheres via a combination of
polymerization of N-(2-hydroxypropyl) methacrylamide. & Whittaker, M. R. High-​order multiblock copolymers RAFT and hetero Diels–Alder chemistry. Macromol.
Eur. Polym. J. 115, 166–172 (2019). via iterative Cu(0)-mediated radical polymerizations Rapid Commun. 29, 1431–1437 (2008).
53. Dong, H., Tang, W. & Matyjaszewski, K. Well-​defined (SET-​LRP): toward biological precision. J. Am. Chem. Soc. 99. Inglis, A. J., Sinnwell, S., Davis, T. P., Barner-​Kowollik, C.
high-​molecular-weight polyacrylonitrile via activators 133, 11128–11131 (2011). & Stenzel, M. H. Reversible addition fragmentation
regenerated by electron transfer ATRP. 77. Anastasaki, A. et al. Photoinduced sequence-​control chain transfer (RAFT) and hetero-​Diels–Alder
Macromolecules 40, 2974–2977 (2007). via one pot living radical polymerization of acrylates. chemistry as a convenient conjugation tool for access
54. Niu, S. et al. Synthesis of high molecular weight and Chem. Sci. 5, 3536–3542 (2014). to complex macromolecular designs. Macromolecules
narrow molecular weight distribution poly(acrylonitrile) 78. Gody, G., Maschmeyer, T., Zetterlund, P. B. & 41, 4120–4126 (2008).
via RAFT polymerization. J. Polym. Sci. A Polym. Chem. Perrier, S. Rapid and quantitative one-​pot synthesis 100. Dietrich, M. et al. Facile conversion of RAFT
51, 1197–1204 (2013). of sequence-​controlled polymers by radical polymers into hydroxyl functional polymers:
55. Mittal, A., Sivaram, S. & Baskaran, D. Unfavorable polymerization. Nat. Commun. 4, 2505 (2013). a detailed investigation of variable monomer
coordination of copper with methyl vinyl ketone in 79. Fischer, H. The persistent radical effect: A principle and RAFT agent combinations. Polym. Chem. 1,
atom transfer radical polymerization. Macromolecules for selective radical reactions and living radical 634–644 (2010).
39, 5555–5558 (2006). polymerizations. Chem. Rev. 101, 3581–3610 101. Lunn, D. J., Discekici, E. H., Read de Alaniz, J.,
56. Lee, I.-H., Discekici, E. H., Anastasaki, A., (2001). Gutekunst, W. R. & Hawker, C. J. Established
de Alaniz, J. R. & Hawker, C. J. Controlled radical 80. Clothier, G. K. K. et al. Exploitation of the nanoreactor and emerging strategies for polymer chain-​end
polymerization of vinyl ketones using visible light. concept for efficient synthesis of multiblock copolymers modification. J. Polym. Sci. A Polym. Chem. 55,
Polym. Chem. 8, 3351–3356 (2017). via macroRAFT-​mediated emulsion polymerization. 2903–2914 (2017).
57. Cheng, C., Sun, G., Khoshdel, E. & Wooley, K. L. ACS Macro Lett. 8, 989–995 (2019). 102. Destarac, M. Industrial development of reversible-​
Well-defined vinyl ketone-​based polymers by reversible 81. Martin, L., Gody, G. & Perrier, S. Preparation of deactivation radical polymerization: is the induction
addition–fragmentation chain transfer polymerization. complex multiblock copolymers via aqueous RAFT period over? Polym. Chem. 9, 4947–4967 (2018).
J. Am. Chem. Soc. 129, 10086–10087 (2007). polymerization at room temperature. Polym. Chem. 6, 103. Theodorou, A. et al. Protein–polymer bioconjugates
58. Reeves, J. A., Allegrezza, M. L. & Konkolewicz, D. Rise 4875–4886 (2015). via a versatile oxygen tolerant photoinduced
and fall: poly(phenyl vinyl ketone) photopolymerization 82. Gauthier, M. A., Gibson, M. I. & Klok, H.-A. Synthesis controlled radical polymerization approach.
and photodegradation under visible and UV radiation. of functional polymers by post-​polymerization Nat. Commun. 11, 1486 (2020).
Macromol. Rapid Commun. 38, 1600623 (2017). modification. Angew. Chem. Int. Ed. 48, 48–58 104. De, P., Li, M., Gondi, S. R. & Sumerlin, B. S.
59. Enciso, A. E. et al. p-​Substituted tris(2-pyridylmethyl) (2009). Temperature-​regulated activity of responsive
amines as ligands for highly active ATRP catalysts: 83. Tao, L., Mantovani, G., Lecolley, F. & Haddleton, D. M. polymer–protein conjugates prepared by grafting-​
facile synthesis and characterization. Angew. Chem. α-​Aldehyde terminally functional methacrylic polymers from via RAFT polymerization. J. Am. Chem. Soc.
Int. Ed. 59, 14910–14920 (2020). from living radical polymerization: application in 130, 11288–11289 (2008).

868 | December 2021 | volume 5 www.nature.com/natrevchem

0123456789();:
 Perspectives

105. Bontempo, D., Heredia, K. L., Fish, B. A. & 113. Zhang, Z., Corrigan, N., Bagheri, A., Jin, J. & Boyer, C. self-optimizing flow reactors. Angew. Chem. Int. Ed.
Maynard, H. D. Cysteine-​reactive polymers A versatile 3D and 4D printing system through 58, 3183–3187 (2019).
synthesized by atom transfer radical polymerization photocontrolled RAFT polymerization. Angew. Chem. 122. Coates, G. W. & Getzler, Y. D. Chemical recycling to
for conjugation to proteins. J. Am. Chem. Soc. 126, Int. Ed. 58, 17954–17963 (2019). monomer for an ideal, circular polymer economy.
15372–15373 (2004). 114. Convertine, A. J., Benoit, D. S., Duvall, C. L., Nat. Rev. Mater. 5, 501–516 (2020).
106. Panganiban, B. et al. Random heteropolymers Hoffman, A. S. & Stayton, P. S. Development of a novel
preserve protein function in foreign environments. endosomolytic diblock copolymer for siRNA delivery. Acknowledgements
Science 359, 1239–1243 (2018). J. Control. Release 133, 221–229 (2009). A.A. gratefully acknowledges ETH Zürich for financial support.
107. Niu, J. et al. Engineering live cell surfaces with 115. Hedir, G. G. et al. Functional degradable polymers by N.P.T. acknowledges the award of a DECRA Fellowship from
functional polymers via cytocompatible controlled xanthate-​mediated polymerization. Macromolecules the ARC (DE180100076). D.K. acknowledges the National
radical polymerization. Nat. Chem. 9, 537–545 47, 2847–2852 (2014). Science Foundation under grant no. DMR-1749730 and the
(2017). 116. Benaglia, M. et al. Universal (switchable) RAFT agents. Robert H. and Nancy J. Blayney Professorship.
108. Kim, J. Y. et al. Cytocompatible polymer grafting J. Am. Chem. Soc. 131, 6914–6915 (2009).
from individual living cells by atom-​transfer 117. Lawrence, J. et al. A versatile and scalable strategy Author contributions
radical polymerization. Angew. Chem. Int. Ed. 55, to discrete oligomers. J. Am. Chem. Soc. 138, N.P.T., G.R.J., K.G.E.B., D.K. and A.A. co-​w rote the
15306–15309 (2016). 6306–6310 (2016). manuscript. Key concepts were developed by A.A. and D.K.
109. Warren, N. J. & Armes, S. P. Polymerization-​induced 118. Holerca, M. N. et al. Monodisperse macromolecules
self-​assembly of block copolymer nano-​objects via by self-​interrupted living polymerization. J. Am. Competing interests
RAFT aqueous dispersion polymerization. J. Am. Chem. Soc. 142, 15265–15270 (2020). The authors declare no competing interests.
Chem. Soc. 136, 10174–10185 (2014). 119. Lutz, J.-F., Neugebauer, D. & Matyjaszewski, K.
110. Yan, J. et al. Solution processable liquid metal Stereoblock copolymers and tacticity control in Peer review information
nanodroplets by surface-​initiated atom transfer radical controlled/living radical polymerization. J. Am. Chem. Nature Reviews Chemistry thanks G. Moad, C. Barner-Kowollik
polymerization. Nat. Nanotechnol. 14, 684–690 Soc. 125, 6986–6993 (2003). and the other, anonymous, reviewers for their contribution to
(2019). 120. Shanmugam, S. & Boyer, C. Stereo-, temporal and the peer review of this work.
111. Ramírez-​García, P. D. et al. A pH-​responsive chemical control through photoactivation of living
nanoparticle targets the neurokinin 1 receptor in radical polymerization: synthesis of block and gradient Publisher’s note
endosomes to prevent chronic pain. Nat. Nanotechnol. copolymers. J. Am. Chem. Soc. 137, 9988–9999 Springer Nature remains neutral with regard to jurisdictional
14, 1150–1159 (2019). (2015). claims in published maps and institutional affiliations.
112. Jung, K. et al. Designing with light: advanced 2D, 3D, 121. Rubens, M., Vrijsen, J. H., Laun, J. & Junkers, T.
and 4D materials. Adv. Mater. 32, 1903850 (2020). Precise polymer synthesis by autonomous © Springer Nature Limited 2021

Nature Reviews | CheMiSTRy volume 5 | December 2021 | 869

0123456789();: