Highly Stereocontrolled Ring-Opening Polymerization of Racemic Alkyl β-Malolactonates Mediated by Yttrium [Amino-alkoxy-bis(phenolate)] Complexes

Article abstract of DOI:10.1002/chem.201600223

Yttrium [amino-alkoxy-bis(phenolate)]amido complexes have been used for the ring-opening polymerization (ROP) of racemic alkyl β-malolactonates (4-alkoxycarbonyl-2-oxetanones, rac-MLA^Rs) bearing an allyl (All), benzyl (Bz) or methyl (Me) latera

Full text of DOI:10.1002/chem.201600223

DOI: 10.1002/chem.201600223
Full Paper
&
Ring-Opening Polymerization
Highly Stereocontrolled Ring-Opening Polymerization of Racemic
Alkyl b-Malolactonates Mediated by Yttrium [Amino-alkoxy-
bis(phenolate)] Complexes
CØdric G. Jaffredo, Yulia Chapurina, Evgueni Kirillov, Jean-FranÅois Carpentier,* and
Sophie M. Guillaume*[a]
Abstract: Yttrium [amino-alkoxy-bis(phenolate)]amido com-
plexes have been used for the ring-opening polymerization
(ROP) of racemic alkyl b-malolactonates (4-alkoxycarbonyl-2-
oxetanones, rac-MLARs) bearing an allyl (All), benzyl (Bz) or
methyl (Me) lateral ester function. The nature of the ortho-
substituent on the phenolate rings in the metal ancillary dic-
tated the stereocontrol of the ROP, and consequently the
syndiotactic enrichment of the resulting polyesters. ROP pro-
moted by catalysts with halogen (Cl, Br)-disubstituted li-
gands allowed the first reported synthesis of highly syndio-
tactic PMLARs (Pr ꢀ 0.95); conversely, catalysts bearing bulky
alkyl and aryl ortho-substituted ligands proved largely inef-
fective. All polymers have been characterized by 1H and
13C{1H} NMR spectroscopy, MALDI-ToF mass spectrometry
and DSC analyses. Statistical and thermal analyses enabled
the rationalization of the chain-end control mechanism.
Whereas the stereocontrol of the polymerization obeyed
a Markov first-order (Mk1) model for the ROP of rac-MLABz
and rac-MLAAll, the ROP of rac-MLAMe led to a chain end-con-
trol of Markov second-order type (Mk2). DFT computations
suggest that the high stereocontrol ability featured by cata-
lysts bearing Cl- and Br-substituted ligands does not likely
originate from halogen bonding between the halogen sub-
stituent and the growing polyester chain.
alkyl b-malolactonates (4-alkoxycarbonyl-2-oxetanones),[3] this
strategy afforded highly alternating poly(b-hydroxyalkanoate)
(PHA)-based copolymers (>95% alternation). Extension of the
concept to mixtures of functional b-lactones enabled the syn-
thesis and post-polymerization modification of copolymers
with alternating chemically differentiated monomer units; the
fragmentation by MS/MS ESI evidenced the alternating unzip-
ping of both repeating units in these original PHA chains.[3]
This effective synthetic strategy relies on the use of highly syn-
dioselective ROP yttrium catalysts, which are able to focus on
the asymmetric center “regardless” of the nature of the lateral
function of the monomer; the approach is so far limited to
chemically closely related monomers. With the aim to extend
the range of applications of this attractive approach, the devel-
opment of highly syndioselective catalysts that are applicable
to a broad array of monomers, along with an understanding of
the sequence control mechanism, both appear mandatory.[4,5]
To this end, we and others have developed complexes
based on oxophilic metals that are highly active and syndiose-
lective catalysts for the ROP of lactones.[6–9] The class of
[amino-alkoxy-bis(phenolate)]yttrium-amido complexes, in the
presence of an alcohol (typically isopropanol) as co-initiator,
has shown remarkable stereoselectivities for the ROP of race-
mic lactide (rac-LA) and racemic b-butyrolactone (rac-BL),
thereby affording highly enriched heterotactic poly(lactic acid)
(PLA) and syndiotactic poly(3-hydroxybutyrate) (PHB) (Pr, prob-
ability of racemic linkage >0.95), respectively.[10] Tuning of the
steric and electronic features of the phenolate ortho-substitu-
Introduction
Control of monomer sequence in polymers is of fundamental
importance. Hence, Nature is able to encode information di-
rectly within the internal structure of macromolecules, as in
proteins or DNA. Strategies to control monomer sequences
and to adapt encoding techniques to synthetic polymers are
of topical interest.[1] One possibility is based on the use of
mixed batches of monomers. In particular, highly controlled se-
quence polymers can be obtained from the stereocontrolled
ring-opening polymerization (ROP) of 50:50 mixtures of two
different chiral monomers of opposite absolute configuration.
As initially developed by Coates, Thomas, and co-workers with
nonfunctional b-lactones (b-butyrolactone and higher alkyl or
fluoroalkyl derivatives)[2] and further enlarged by our group to
[a] Dr. C. G. Jaffredo, Dr. Y. Chapurina, Dr. E. Kirillov, Prof. Dr. J.-F. Carpentier,
Dr. S. M. Guillaume
Institut des Sciences Chimiques de Rennes
Organometallics: Materials and Catalysis Laboratories
UMR 6226 CNRS-UniversitØ de Rennes 1
Campus de Beaulieu, 35042 Rennes Cedex (France)
E-mail: jean-francois.carpentier@univ-rennes1.fr
sophie.guillaume@univ-rennes1.fr
Supporting information and ORCID from the author for this article are
available on the WWW under http://dx.doi.org/10.1002/chem.201600223. It
contains supporting synthetic procedures, H, 13C{1H} and DOSY NMR data,
MALDI-ToF mass spectra, DSC analyses, and statistical analysis of stereose-
quences in PMLARs synthesized from complexes 1, and computational de-
tails including Cartesian coordinates of computed structures.
1
Chem. Eur. J. 2016, 22, 7629 – 7641
7629
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ents in these ligands modulates the metal environment and
consequently the extent of stereocontrol of the sequence of
the repeating units, and ultimately the physicochemical behav-
ior of the recovered polymers. Although bulkiest substituents
basically led to the highest degree of heterotacticity in the
ROP of rac-LA, experimental observations coupled with com-
putational investigations pinpointed the role of both steric and
electronic parameters to reach the highest syndioselectivity in
the ROP of rac-BL.[8d,10]
(rac-MLARs, R=Bz, All, Me; Scheme 1). The most significant
data for the thus prepared homopolymers are gathered in
Table 1, Table 2, and Table 3.
ROP of rac-MLABz was first investigated at a [monomer]0/[cat-
alyst]0 ratio of 100 in the presence of 1 equiv of iPrOH co-initia-
tor at 208C in toluene. All yttrium complexes 1a–e showed
good activity, with complex 1a being the most active (full con-
version within 2 min; that is, TOFꢀ3000 hÀ1; Table 1, entry 3).
To our knowledge, this catalytic system is the most active de-
scribed to date for the ROP of rac-MLABz.[11] The catalyst system
based on 1b offered an activity one order of magnitude lower
(TOF=360 hÀ1, Table 1, entry 7), yet slightly better than that of
systems based on complexes 1c–e under the same conditions
(TOF=40–50 hÀ1, Table 1, entries 9, 12 and 15, respectively).
These reactivity differences likely reflect the major influence of
the fourth group in the metal ancillary; that is, NMe2 in 1a
versus OMe in 1b–e. A quite good agreement between the
theoretical molar mass values (Mn,theo) and the experimental
values determined by NMR (Mn,NMR) was observed with all cata-
lytic systems. The molar mass values determined by SEC
(Mn,SEC) using polystyrene standards were always approximately
double the Mn,NMR values, thus suggesting a correction factor of
ca. 0.5 to account for the difference in hydrodynamic volume
between PMLABz and polystyrene. The dispersities remained
generally narrow (“M =1.06–1.54), indicating a relatively fast
and quantitative initiation along with limited undesirable side
reactions (intermolecular (reshuffling), intramolecular (back-
biting) transesterification reactions, or other transfer reactions).
Nevertheless, the dispersities were found to be quite depen-
dent on the catalytic system, with the alkyl-substituted cata-
lysts leading to lower dispersity values (“M,1a–c =1.06–1.29) as
In this paper, we report on the use of a series of [amino-
alkoxy-bis(phenolate)]yttrium-amido complexes in combination
with iPrOH for the syndioselective ROP of several b-malolacto-
nates, namely racemic benzyl b-malolactonate (rac-MLABz), ra-
cemic allyl b-malolactonate (rac-MLAAll), and racemic methyl b-
malolactonate (rac-MLAMe; Scheme 1), complementing our pre-
Scheme 1. Syndioselective ROP of rac-MLARs mediated by 1a–e/iPrOH sys-
tems (complex 1e was generated in situ from proligand 5 and Y(N(SiH-
Me2)2)3(THF) and not isolated).
compared with the halogen-substituted catalysts (“
=
M,1d–e
1.44–1.54). Hence, the catalyst system based on 1a afforded
the best activity as well as the best control over the molar
mass values. In the absence of alcohol co-initiator, the amido
complexes showed comparable performances to those ob-
served in the presence of 1 equiv iPrOH (Table 1, entries 5, 8,
10, and 13). The good control over the polymerization reaction
by using 1a/iPrOH was further evidenced by the linear varia-
tion of Mn,NMR values with the amount of monomer consumed
for a [rac-MLABz]0/[1a]0 ratio in the range of 25–100 (Table 1,
entries 1–3; Figure 1). This linear increase is in fair agreement
with the Mn,theo values calculated assuming one growing chain
for each catalyst/co-initiator couple, thus highlighting some
living character of the polymerization. On the other hand,
when the monomer loading was increased to [rac-MLABz]0/
[1a]0 =500, no polymer was formed (Table 1, entry 4); this
most likely highlighted the high sensitivity of these yttrium
catalysts to residual impurities present in such a large batch of
polar monomer.[12] Experiments carried out in THF rather than
in toluene failed to lead to any monomer conversion (Table 1,
entry 6 vs. entry 5, respectively). This result can be rationalized
by competitive coordination of THF versus monomer onto the
metallic center. Such behavior has previously been reported in
the ROP of alike polar cyclic ester monomers.[8f]
liminary results on this chemistry.[3] This manuscript provides
comprehensive and detailed results on catalyst comparison,
polymer microanalysis, and mechanism analysis. The catalyst
systems promoted the efficient synthesis of the corresponding
PMLARs, with—quite uniquely in this chemistry—the com-
plexes with halogen-disubstituted ligands affording the high-
est syndiotacticities (Pr >0.95). Some DFT computations aimed
at exploring the relationship between the nature of the sub-
stituent on the catalyst and its syndioselectivity are also report-
ed.
Results and Discussion
Stereoselective ROP of rac-MLABz, rac-MLAAll, and rac-MLAMe
The performance of five [amino-alkoxy-bis(phenolate)]yttrium-
amido complexes (1a–e; complex 1e was generated in situ
from proligand 5 and Y(N(SiHMe2)2)3(THF) and not isolated;
previous work in this series has established that strictly equiva-
lent polymerization performances are obtained upon using
either isolated or in situ generated complexes[3,8d]) featuring
sterically and electronically differentiated ancillaries was inves-
tigated in the ROP of several racemic alkyl b-malolactonates
The 1a/iPrOH system also showed the highest activity for
the ROP of rac-MLAAll (TOFꢀ3000 hÀ1, Table 2, entry 2 vs.
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7630
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Syndioselective ROP of rac-MLABz mediated by 1a–e/iPrOH systems.[a]
[d]
[e]
[f]
[f]
M
[h]
Complex
[MLABz]0/[1]0/[iPrOH]0
Reaction time[b]
[min]
MLABz conv.[c]
[%]
Mn,theo
Mn,NMR
Mn,SEC
“
Pr[g]
Tm
[gmolÀ1
]
[gmolÀ1
]
[gmolÀ1
]
[8C]
1
2
3
4
1a
1a
1a
1a
1a
1a
1b
1b
1c
1c
1d
1d
1d
1e
1e
25:1:1
50:1:1
1
1
2
120
10
24 h
15
2
120
120
75
120
120
60
100
100
100
0
100
0
91
100
96
98
98
80
98
5200
6200
12900
21000
–
10600
19500
45300
–
41100
–
42100
40600
39200
41200
22800
35200
37500
25000
35200
1.13
1.08
1.06
–
1.23
–
1.18
1.26
1.27
1.29
1.45
1.54
1.44
1.46
1.54
0.81
79
74
71
–
72
–
87
86
NO[i]
NO[i]
111
117
114
100
104
10400
20600
–
20600
–
18800
20600
19800
20200
10100
16500
20200
10300
18800
0.80
0.79
–
0.80
–
0.85
0.85
0.68
0.67
>0.95
>0.95
>0.95
0.91
0.92
100:1:1
500:1:1
100:1:0
100:1:0
100:1:1
100:1:0
100:1:1
100:1:0
50:1:1
5
NO[i]
6[j]
7
–
20300
NO[i]
19500
NO[i]
11 900
17600
NO[i]
11 000
17600
8
9
10
11
12
13
14
15
100:1:1
100:1:0
50:1:1
100
91
100:1:1
120
[a] All reactions were performed with [MLABz]0 =1.0m at 208C in toluene. [b] Reaction times were not necessarily optimized. [c] MLABz conversion as deter-
mined by 1H NMR spectroscopic analysis of the crude reaction mixture. [d] Theoretical molar mass value calculated considering one growing polymer
chain per metal center from the relation: Mn,theo =([MLABz]0/[iPrOH]0 ”conv.MLABz ”MMLABz)+MiPrOH when using iPrOH, and ([MLABz]0/[catalyst]0 ”conv.MLABz
”
MMLABz) in the absence of iPrOH, with MMLABz =206 gmolÀ1 and MiPrOH =60 gmolÀ1. [e] Molar mass value determined by H NMR spectroscopic analysis of the
isolated polymer in CDCl3 at 258C, from the resonances of the terminal isopropoxy group. [f] Number-average molar mass value determined by SEC in THF
at 308C versus polystyrene standards (uncorrected Mn values). [g] Pr is the probability of racemic linkages between MLABz units as determined by 13C{1H}
NMR analyses. [h] Melting temperature as determined by DSC (second heating cycle). [i] Not observed. [j] Reaction performed in THF.
1
Table 2. Syndioselective ROP of rac-MLAAll mediated by 1a–e/iPrOH (1:1) systems.[a]
[MLAAll]0/[1]0/[iPrOH]0
Reaction time[b]
[min]
MLAAll conv.[c]
[%]
Mn,theo
Mn,NMR
Mn,SEC
“
Pr[g]
Tm
[d]
[e]
[f]
[f]
M
[h]
Complex
[gmolÀ1
]
[gmolÀ1
]
[gmolÀ1
]
[8C]
1
2
3
4
5
6
1a
1a
1b
1c
1d
1e
50:1:1
1
2
5
120
10
20
100
100
89
100
80
7900
7500
8900
1.22
1.03
1.41
1.15
1.64
1.54
0.81
0.82
0.87
0.68
>0.95
0.95
49
51
80
NO[i]
112
107
100:1:1
100:1:1
100:1:1
100:1:1
100:1:1
15700
14000
15700
12500
10400
14100
14600
15300
13300
9900
29300
20300
22100
23500
19600
100
[a] All reactions were performed with [MLAAll]0 =1.0m at 208C in toluene. [b] Reaction times were not necessarily optimized. [c] MLAAll conversion as deter-
mined by 1H NMR spectroscopic analysis on the crude reaction mixture. [d] Theoretical molar mass value calculated considering one growing polymer
chain per metal center from the relation: Mn,theo =([MLAAll]0/[iPrOH]0 ”conv.MLAAll ”MMLAAll)+MiPrOH, with MMLAAll =156 gmolÀ1 and MiPrOH =60 gmolÀ1. [e] Molar
1
mass value determined by H NMR spectroscopic analysis of the isolated polymer in CDCl3 at 258C, from the resonances of the terminal isopropoxy group.
[f] Number-average molar mass value determined by SEC in THF at 308C versus polystyrene standards (uncorrected Mn values). [g] Pr is the probability of
racemic linkages between MLAAll units as determined by 13C{1H} NMR analysis. [h] Melting temperature as determined by DSC (second heating cycle).
[i] Not observed.
Table 1, entry 3). Lower activities were recorded for systems
based on methoxy-capped complexes 1b, 1d, and 1e (gener-
ated in situ), although these values were significantly higher in
the ROP of rac-MLAAll than in the ROP of rac-MLABz {1b:
TOFMLAAll =1100 hÀ1 vs. TOFMLABz =360 hÀ1 (Table 2, entry 3 vs.
Table 1, entry 7); 1d: TOFMLAAll =480 hÀ1 vs. TOFMLABz =40 hÀ1
(Table 2, entry 5 vs. Table 1, entry 12); 1e: TOFMLAAll ꢀ300 hÀ1
vs. TOFMLABz =46 hÀ1 (Table 2, entry 6 vs. Table 1, entry 15)}. The
catalyst system based on the bulky 1c was clearly the least
active in this series. All the PMLAAlls synthesized featured ex-
perimental molar mass values (Mn,NMR) in good agreement with
theoretical values (Mn,theo), as well as a monomodal SEC trace
and a fairly narrow dispersity (1.03ꢁ“M ꢁ1.64).
MLAAll or rac-MLABz (Table 3 vs. 1, 2, respectively; 1a, TOFꢀ
6000 hÀ1, Table 3, entry 2; 1b, TOFꢀ6000 hÀ1, Table 3, entry 3;
1c, TOFꢀ1200 hÀ1, Table 1, entry 4). On the other hand, an op-
posite trend was observed with halogen-substituted com-
plexes 1d–e, for which complete conversion of MLAMe was
reached within 48 h, whereas only ca. 20–120 min were neces-
sary with rac-MLAAll and rac-MLABz, respectively. Note however
that, in contrast to 1a–c, complexes 1d–e gave insoluble
PMLAMes (suggesting the formation of a highly crystalline poly-
mer, see below), thus affecting the reaction medium and possi-
bly also the reaction kinetics.
The activities of complexes 1a–c in the ROP of methyl b-ma-
lolactonate significantly surpassed those recorded with rac-
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7631
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 3. Syndioselective ROP of rac-MLAMe mediated by 1a–e/iPrOH (1:1) systems.[a]
[d]
[e]
[f]
[f]
[g]
[h]
Complex
[MLAMe]0/[1]0/[iPrOH]0
Reaction time[b]
[min]
MLAMe conv.[c]
[%]
Mn,theo
Mn,NMR
Mn,SEC
“
M
Pr/r
Tm
[gmolÀ1
]
[gmolÀ1
]
[gmolÀ1
]
[8C]
1
2
3
4
5[j]
6[j]
1a
1a
1b
1c
1d
1e
50:1:1
0.5
1
1
5
48 h
48 h
75
5000
4700
17100
13100
13700
NO[j]
7000
24200
18400
18800
NO[j]
1.15
1.22
1.20
1.24
NO[j]
NO[j]
0.81
0.76
0.77
0.42
0.89
0.92
173
60
62
NO[i]
207
212
100:1:1
100:1:1
100:1:1
100:1:1
100:1:1
100
100
100
100[k]
100[k]
13000
13000
13000
13000
13000
NO[j]
NO[j]
[a] All reactions were performed with [MLAM]0 =1.0m at 208C in toluene. [b] Reaction times were not necessarily optimized. [c] MLAMe conversion as deter-
mined by 1H NMR spectroscopic analysis of the crude reaction mixture. [d] Theoretical molar mass calculated considering one growing polymer chain per
metal center from the relation: Mn,theo =([MLAMe]0/[catalyst]0 ”MMLAMe)+MiPrOH, with MMLAMe=132 gmolÀ1 and MiPrOH =60 gmolÀ1. [e] Molar mass value de-
1
termined by H NMR spectroscopic analysis of the isolated polymer in CDCl3 at 258C, from the resonances of terminal isopropoxy group. [f] Number-aver-
age molar mass value determined by SEC in THF at 308C versus polystyrene standards (uncorrected Mn values). [g] Pr/r is the probability of racemic linkages
between MLAMe units consecutive to a racemic diad as determined by 13C{1H} NMR spectroscopic analysis and DSC. [h] Melting temperature as determined
by DSC (second heating cycle). [i] Not observed. [j] Insoluble polymer. [k] No presence of residual monomer was observed in the soluble fraction.
Bz
&
^
Figure 1. Variation of Mn,NMR ( ), Mn,theo (dashed line) and “
(
) of PMLA s
M
synthesized by ROP of rac-MLABz in the presence of the 1a/iPrOH (1:1)
system (generated in situ), as a function of MLABz conversion (Table 1, en-
tries 1–3).
Figure 2. 1H NMR spectrum (500 MHz, CDCl3, 258C) of a PMLAMe prepared
from the ROP of rac-MLAMe mediated by complex 1a in the presence of
1 equiv of iPrOH (Table 3, entry 1; *: signal from residual grease).
Characterization of PMLARs synthesized by ROP of rac-
MLARs promoted by 1a–e/iPrOH systems
All the PHAs synthesized were characterized by 1H and
13C{1H} NMR spectroscopy and MALDI-ToF mass spectrometry
(Figures 2–7). The 1H NMR spectra of the PMLARs clearly
showed signals corresponding to both the MLAR repeating
unit (see the Experimental section) and chain-end groups
(Figure 2, Figures 4, 5). Hence, when iPrOH was used as co-ini-
tiator, all the polymers isolated showed signals for the terminal
isopropoxycarbonyl hydrogen atoms (-OCH(CH3)2 and
-OCH(CH3)2 at d=4.40–3.40 and 1.30–1.15 ppm, respectively).
The 13C{1H} NMR analyses confirmed the nature of the back-
bone and end-capping groups of PMLARs (Figures 3, S5, S6 and
S8 for PMLAAll, PMLABz and PMLAMe, respectively).
The chemical structure of PMLARs was further supported by
MALDI-ToF MS analyses. The spectra recorded from a low molar
mass sample of PMLABz prepared with the 1a/iPrOH system re-
vealed two populations of macromolecules, both featuring a re-
peating unit of m/z 206 gmolÀ1 (MMLABz). The first corresponds
to the anticipated a-isopropoxy,w-hydroxyl and a-isopropox-
y,w-dehydrated (i.e., crotonate) PMLABz chains both ionized by
Figure 3. 13C{1H} NMR spectrum (125 MHz, CDCl3, 258C) of a PMLAMe pre-
pared by ROP of rac-MLAMe with 1a in the presence of 1 equiv of iPrOH
(Table 3, entry 1).
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7632
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
suggests that dehydration of the macromolecular chains occurs
during the MALDI-ToF MS analyses.
Microstructural analysis and syndiotactic enrichment of
PMLARs
Well-defined PMLARs were synthesized by ROP of rac-MLARs by
using complexes 1a–e and iPrOH as co-initiator. The latter yt-
trium catalysts are known to be stereoselective for the ROP of
rac-b-butyrolactone as well as of rac-lactide.[6a,8a,d,13] To confirm
the microstructure of the PMLARs synthesized in the present
work, 13C{1H} NMR analyses were carried out. The methine
region of the 13C{1H} NMR spectra of PMLABz and PMLAAll re-
vealed four signals corresponding to triad stereosequences;
namely, racemo-racemo (rr), meso-racemo (mr), racemo-meso
(rm) and meso-meso (mm) (Figures 8 and 9, respectively). As-
signment of signals in this region was performed based on
previous work on stereoenriched PMLABz,[14] stereocopoly-
mers,[15] as well as by comparison with isotactic and atactic
PMLABzs synthesized from the ROP of (R)-MLABz and rac-MLABz
by using [(BDI)Zn{N(SiMe3)2}]/iPrOH (1:1) as catalytic system, re-
spectively (Figure 8a).[16,17] Four triads were thus observed for
PMLABzs (d=68.70 ppm, rr; d=68.61 ppm, mr; d=68.53 ppm,
rm; d=68.42 ppm, mm; Figure 8b), and PMLAAlls (d=
68.79 ppm, rr; d=68.70 ppm, mr; d=68.68 ppm, rm; d=
68.61 ppm, mm; Figure 9). The larger contribution of the rr
triad, compared with the much smaller ones of the mr, rm and
mm triads, thus supported the predominant syndiotacticity of
the PMLARs (R=All, Bz) formed.
The syndiotactic enrichment of PMLABzs and PMLAAlls was
determined by calculating the probability of racemic linkage
(Pr) through a statistical analysis of the triad distribution. The
methine signals were deconvoluted with the aim to compare
the relative integration value of each triad signal (Figure 10). A
first-order chain-end control (or Markov first-order chain-end
control, Mk1) is defined in terms of probability of a racemic (Pr)
or meso (Pm) linkage.[18] By comparing these predicted values
to those observed in the NMR spectra, the mechanism of ste-
reocontrol and the degree of stereoselectivity in the polymeri-
zation can be determined. For triad stereosequences, the
agreement with an Mk1 mechanism was supported by the cal-
culation of a Bernoullian test factor (B=4(mm)(rr)/[(mr)+(rm)]),
which was close to unity for a perfect first-order chain-end
control. In fact, the latter factors corresponding to the synthe-
sized PMLABzs and PMLAAlls were close to unity (1.02ꢁBꢁ1.34;
Tables S1 and S2), confirming the good agreement with the
Mk1 chain-end mechanism. Analyses of the relative contribu-
tions of the different stereosequences confirmed a Mk1-type
stereocontrol mechanism,[18] as reported for rac-lactide and
rac-b-butyrolactone.[6a,8a,d,13]
Figure 4. 1H NMR spectrum (500 MHz, CDCl3, 258C) of a PMLAAll prepared
from the ROP of rac-MLAAll mediated by complex 1a in the presence of
1 equiv of iPrOH (Table 2, entry 1; *: signal from residual grease).
Figure 5. 1H NMR spectrum (500 MHz, CDCl3, 258C) of a PMLABz prepared
from the ROP of rac-MLABz mediated by complex 1a in the presence of
1 equiv of iPrOH (Table 1, entry 1; Mn,NMR =6200 gmolÀ1; *: signal from resid-
ual grease).
Na+ (Figure 6). This was unequivocally confirmed by the close
match with the isotopic simulations for [(CH3)2CH(COCH2CH-
(COOCH2C6H5)O)nH]·Na+ with, for example, calculated m/z
4206.2 gmolÀ1 versus found m/z 4205.7 gmolÀ1 for n=20, and
[(CH3)2CH(COCH2CH(COOCH2C6H5)O)nHÀH2O]·Na+ with, for ex-
ample, calculated m/z 4188.2 gmolÀ1 versus found m/z
4188.7 gmolÀ1 for n=20 (expansion of Figure 6). The second
population corresponds to a-amido,w-hydroxy and a-amido,w-
dehydrated PMLABz chains, both ionized by H+ (see Figure S7).
The presence of this latter population reflects the occurrence of
amido-initiation, likely due to the use of a slight default of iso-
propanol versus the amido precursor; however, the good
agreement between theoretical (Mn,theo) and experimental
(Mn,NMR) molar mass values of PMLABzs determined from the iso-
propoxycarbonyl terminal group indicates that amido-initiation
remained minimal versus isopropoxide-initiation and suggests
that the latter macromolecules may be overexpressed in these
MALDI-TOF MS analyses. In fact, the MALDI-ToF mass spectrum
of a low molar mass PMLAAll sample synthesized with the same
1a/iPrOH catalytic system displayed a unique population
of macromolecules with a repeating unit of m/z 156 gmolÀ1
(MMLAAll;
Figure 7),
assigned
to
a-isopropoxy,w-
hydroxy or a-isopropoxy,w-dehydrated (i.e., crotonate) PMLAAll
chains ionized by Na+; namely, [(CH3)2CH(COCH2CH-
(COOCH2CHCH2)O)nH]·Na+ with, for example, calculated
m/z 2424.7 gmolÀ1 (vs. found m/z 2424.4 gmolÀ1) and
[(CH3)2CH(COCH2CH(COOCH2CHCH2)O)nHÀH2O]·Na+ with, for
example, calculated m/z 2406.7 gmolÀ1 versus found m/z
2406.4 gmolÀ1 for n=15, as confirmed by the isotopic simula-
tions (expansion, Figure 7). The absence of noticeable crotonate
groups in the NMR spectra of the prepared PMLABz and PMLAAll
Microstructural analysis of PMLAMe proved to be somewhat
more challenging than that of PMLABz and PMLAAll. Indeed,
tetrad stereosequences were observed in the methine region
of the 13C{1H} NMR spectrum of PMLAMe {mmr, d=68.78 ppm;
rrr, d=68.75 ppm; rmm, d=68.71 ppm; rrm, d=68.67 ppm;
mrr, d=68.64 ppm; rmr, d=68.61 ppm; mmm, d=68.58 ppm;
mrm, d=68.56 ppm} (Figure 11). These signals were assigned
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7633
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 6. MALDI-ToF mass spectrum (matrix IAA) of a PMLABz prepared from the ROP of rac-MLABz mediated by the 1a/iPrOH system (Table 1, entry 2). The
major population stands for [iPrO{MLABz}nH]·Na+ macromolecules (refer to simulation b for n=20). The minor population stands for [iPrO{MLABz}nHÀH2O]·Na+
macromolecules (refer to simulation a for n=20). Secondary populations stand for [(SiHMe2)2N{MLABz}nH]·H+ and [(SiHMe2)2N{MLABz}nHÀH2O]·H+.
by comparison with isotactic PMLAMe, for which only the mmm
tetrad was observed (Figure 11, isotactic), along with statistical
analysis. The major contribution of the rrr tetrad for the
PMLAMes synthesized by ROP of rac-MLAMe in the presence of
1a–c/iPrOH systems supported the high syndiotacticity of this
material.
By analogy with the formation of syndiotactic PMLABz and
PMLAAll, the stereocontrol in the ROP of MLAMe was expected
to take place through a chain-end control mechanism. Un-
fortunately, for tetrad stereosequences, the validity of this hy-
pothesis could not be verified from the calculation of a Ber-
noullian test factor. Two statistic models were then investigat-
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7634
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 7. MALDI-ToF mass spectrum (matrix IAA) of a PMLAAll prepared from the ROP of rac-MLAAll mediated by the 1a/iPrOH system (Table 2, entry 1). The
major population stands for [iPrO{MLAAll}nHÀH2O]·Na+ macromolecules (refer to simulation a for n=15). The minor population stands for [iPrO{MLAAll}nH]·Na+
macromolecules (refer to simulation b for n=15).
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7635
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 8. Detail of the methine region of the 13C{1H} NMR spectra (125 MHz,
CDCl3, 258C) of PMLABzs synthesized from ROP of rac-MLABz in the presence
of a) (i) [(BDI)Zn{N(SiMe3)2}]/iPrOH system (Pr =Pm =0.5); (ii) 1b/iPrOH
(Pr =0.85; Table 1, entry 7); (iii) [(BDI)Zn{N(SiMe3)2}]/iPrOH system (Pm > 0.95);
and b) 1c/iPrOH (Pr =0.68; Table 1, entry 9); 1a/iPrOH (Pr =0.79; Table 1,
entry 3); 1b/iPrOH (Pr =0.85; Table 1, entry 7); 1e (in situ generated)/iPrOH
(Pr =0.92; Table 1, entry 15); 1d/iPrOH (Pr > 0.95; Table 1, entry 11).
Figure 11. Detail of the methine region of the 13C{1H} NMR spectra
(125 MHz, CDCl3, 258C) of PMLAMes synthesized from ROP of rac-MLAMe in
the presence of the 1c/iPrOH system (Pr/r =0.43; Table 3, entry 4); 1a/iPrOH
(Pr/r =0.76; Table 3, entry 2); 1b/iPrOH (Pr/r =0.77; Table 3, entry 3); ROP of
(R)-MLAMe mediated by 1a (Pm > 0.95).
ed for the analysis of the stereocontrol mechanism, namely
Markov first-order (Mk1) and Markov second-order (Mk2)
models.[18] These latter models fit with a chain-end control
driven by the last and the penultimate repeating units of the
growing-chain, respectively. The tetrads contribution were
evaluated by deconvolution of the NMR massif corresponding
to the methine region of the 13C{1H} NMR spectrum (d=68.5–
69.0 ppm), as illustrated in Figure 12, according to a previously
reported approach.[3]
Figure 9. Detail of the methine region of the 13C{1H} NMR spectra (125 MHz,
CDCl3, 258C) of PMLAAlls synthesized from ROP of rac-MLAAll in the presence
of the 1c/iPrOH system (Pr =0.68; Table 2, entry 4); 1a/iPrOH (Pr =0.82;
Table 2, entry 2); 1b/iPrOH (Pr =0.87; Table 2, entry 3); 1e (in situ generat-
ed)/iPrOH (Pr =0.95; Table 2, entry 6); 1d/iPrOH (Pr > 0.95; Table 2, entry 5).
Figure 12. Deconvolution of the signals of the methine region of the 13C{1H}
NMR spectrum (125 MHz, CDCl3, 258C) of PMLAMe prepared from the ROP of
rac-MLAMe mediated by the 1a/iPrOH system (Table 3, entry 1) using Lorentz
functions.[16] [a] Observed spectrum; [b] Eight-component signal resolved;
and [c] (i) simulated spectrum using the component peaks in [b], and (ii) dif-
ferential spectrum obtained upon subtracting [c; i] from [a].
Figure 10. a) Methine region of the 13C{1H} NMR spectrum (125 MHz; CDCl3;
258C) of a PMLABz prepared from the ROP of rac-MLABz mediated by the 1b/
iPrOH system (Pr =0.85; Table 1, entry 7); b) Lorentzian deconvolution of the
methine region[19] (values of integrals corresponding to deconvoluted sig-
p
nals). The Pr value was calculated from the following equation: Pr = (rr).[18]
Chem. Eur. J. 2016, 22, 7629 – 7641 www.chemeurj.org
7636
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The experimental data were then fitted with the theoretical
values obtained by using the Mk1 and Mk2 models (Figure 13).
Although both models afforded good agreement between ex-
perimental and theoretical contributions of the rrr tetrad, the
Mk2 model afforded a much better agreement for all the
seven other tetrads’ contribution. The statistical analysis fitting
data for PMLAMes are reported in Table 4, whereas the data for
statistical analyses for other PMLARs are gathered in the Sup-
porting Information (Tables S3–S6). The stereocontrol was
found to be of Mk2 type, indicating that the last two inserted
repeating units of the growing-chain directly impact the inser-
tion of the next monomer.
A consistent trend in stereoselectivity was hence observed for
the ROP of MLABz and MLAAll mediated by complexes 1a–e.
First, the highest stereoselectivities were measured for catalysts
with halogen-disubstituted ligands; that is, 1d (Pr > 0.95;
Table 1, entries 11–13 and Table 2, entry 5, for PMLABz and
PMLAAll, respectively) and 1e (Pr =0.91–0.95; Table 1, en-
tries 14–15 and Table 2, entry 6, for PMLABz and PMLAAll, re-
spectively). The syndiotactic enrichment induced by catalysts
1a (Pr =0.79–0.81; Table 1, entries 1–5 and Table 2, entries 1–2,
for PMLABz and PMLAAll, respectively) and 1b (Pr =0.85–0.87;
Table 1, entries 7–8 and Table 2, entry 3, for PMLABz and
PMLAAll, respectively) remained more modest. A more signifi-
cant decrease of the stereocontrol was calculated for 1c (Pr =
0.67–0.68; Table 1, entries 9–10 and Table 2, entry 4, for PMLABz
and PMLAAll, respectively). The Mk2 model allowed the mea-
surement of syndiotactic enrichment of soluble PMLAMes syn-
thesized from 1a–c. The trend observed for the other two
PMLARs (R=Bz, All) was similarly observed in the ROP of rac-
MLAMe. Indeed, catalyst 1c (Pr/r =0.60; Table 3, entry 4) showed
a lower enrichment than 1b (Pr/r =0.77; Table 3, entry 3) and
1a (Pr/r =0.76–0.81; Table 3, entries 1–2). The PMLAMes synthe-
sized with complexes with halogen-disubstituted ligands 1d–
e proved to be very poorly soluble in common organic sol-
vents. This insolubility hinted at highly crystalline samples,
likely arising from the high regularity of the stereosequences
in the polymer chain. The syndiotactic enrichment of these
PMLAMes was further estimated by DSC thermal analysis (see
below). In this case, 1e (Pr/r ꢂ0.92; Table 3, entry 6) proved to
be slightly most selective than 1d (Pr/r ꢂ0.89; Table 3, entry 5).
These results, similar to those previously reported in the
ROP of rac-BL and rac-LA,[8d,13] point out the key influence of
the ortho-substituents of the amino-alkoxy-bis(phenolate)
ligand on the stereoselectivity of complexes 1a–e.[10] Neverthe-
less, the relationship between the nature of the ortho-substitu-
ent and the stereoselectivity is completely different in the case
of the rac-MLABzs: R1 =CPh3 (1c; Pr =0.68) <CMe2Ph (1a; Pr =
0.79) <CMe2tBu (1b; Pr =0.85) !Cl (1d; Pr > 0.95), than the
trends observed with rac-LA: R1 =Cl (Pr =0.56) <CMe2Ph (Pr =
0.90) <CMe2tBu (Pr =0.94–0.95) <CPh3 (Pr =0.95–0.96),[8d] and
rac-BL: R1 =Cl (Pr =0.42–0.45) <CMe2tBu (Pr =0.62–0.70) <
CMe2Ph (Pr =0.94–0.95) <CPh3 (Pr =0.94).[8d] These differences
are more significant with the bulky and aromatic substituents
CMe2Ph or CPh3 (Pr/rac-LA =0.90 and 0.96; Pr/rac-BL =0.89 and 0.94,
versus Pr/rac-MLABz=0.79 and 0.68, respectively). On the other
hand, catalysts incorporating Cl substituents gave a poor selec-
tivity towards rac-LA and rac-BL,[8d] whereas this behavior is re-
versed in the case of rac-MLARs (Pr/rac-LA =0.56; Pr/rac-BL =0.42, vs.
Pr/rac-MLAR> 0.95).
&
Figure 13. Fitting of the statistical repartition obtained by experiment ( ),
first-order Markov model ( ; RMS error=3.36.10À2) and second-order
*
Markov model ( ; RMS error=0.67.10À2) for a PMLAMe synthesized from the
~
ROP of rac-MLAMe mediated by the 1a/iPrOH system (Table 3, entry 1).
Table 4. Mk1 and Mk2 statistical analyses of tetrads’ contribution in the
methine region of the 13C{1H} NMR spectra of PMLAMe synthesized from
the ROP of rac-MLAMe mediated by 1a–c/iPrOH systems (Table 3, en-
tries 1–4).[a]
Entry
Complex
Markov 1st order
Markov 2nd order
Fitting with
experiment
(RMS error)[20]
Pr
Fitting with
experiment
(RMS error)[20]
Pr/r
1
2
3
4
1a
1a
1b
1c
3.36”10À2
4.16”10À2
5.43”10À2
3.88”10À2
0.79
0.75
0.74
0.59
0.67”10À2
3.33”10À2
4.53”10À2
3.72”10À2
0.81
0.76
0.77
0.60
[a] Mk1 and Mk2 statistical models used as defined by Sepulchre for the
ROP of a lactone containing only one asymmetric center.[18]
The syndioselectivity of yttrium complexes 1a–e was then
evaluated by a statistical treatment of the contribution of the
different triads observed from the methine region of
13C{1H} NMR spectra of the resulting PMLARs, using Markov
first-order model (Mk1) for PMLABz and PMLAAll, and the contri-
bution of the different tetrads, using Markov second-order
model (Mk2) for PMLAMe (see above).[18] These latter treatments
allowed the probability of racemic linkage (Pr) for PMLABz and
PMLAAll, and the probability of linkages between MLAMe units
consecutive to a racemic diad (Pr/r) for PMLAMe to be calculated.
Although quite seldom seen in the literature, highly stereo-
selective aluminum complexes bearing halogen-disubstituted
salan and salen complexes have been reported for the ROP of
rac-LA by Gibson et al.[21] The exact role of the halogen sub-
stituent in the stereocontrol ability of these systems remains
unclear, but electronic contributions to the stereochemistry of
insertion were suggested.[21] In previous investigations on syn-
dioselective ROP of simple b-lactones such as BL, based on the
pioneering work of Rzepa et al. in the ROP of lactide,[22] we
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7637
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
also pinpointed electronic contributions, namely CÀH···p(arene)
interactions involving the acidic methylene hydrogen atoms of
the k2-O,O-b-alkoxybutyrate-type propagating chain and the
aryl rings present on ortho-substituents of the phenolate moi-
eties of the ligands.[8d] This accounted for the clear superiority
of CMe2Ph- and CPh3- ortho-substituted systems towards simi-
larly sterically crowded but purely aliphatic versions such as
the CMe2tBu-substituted system. Such stabilizing CÀH···p(ar-
ene) interactions (estimated as ca. 5–10 kcalmolÀ1) were sup-
ported by DFT computations performed on model species.[8d]
The above model is clearly unsatisfactory for the present
ROP of b-malolactonates, because both CPh2Me- (1a) and
CPh3- (1c)-substituted systems gave poor syndioselectivities.
We therefore explored the possible involvement of CÀX···O=C
halogen bonding[23,24] between ortho-halogen substituents and
oxygen from carbonyl/alkoxy groups of the propagating poly-
(alkoxybutyrate) chain. For this purpose, DFT computations
were performed on an yttrium species bearing an allyl b-alk-
oxybutyrate moiety as a model of the propagating chain.
Mononuclear and dinuclear geometries,1 varying in the posi-
tioning of the k2-coordinated O,O-b-alkoxybutyrate moiety
were explored (Figure 14). However, all of the mononuclear
species fell into a narrow range of energies, and in none of
these were any close contacts (<3 Š) between the allyl b-alk-
oxybutyrate and Cl moieties detected. For the computed dinu-
clear species, such close contacts could be detected but only
in structures of significantly higher relative energies (>8.6 kcal
molÀ1). Similar geometries and relative energies were also com-
puted for similar models bearing CMe2Ph instead of Cl sub-
stituents (see Table S7). Overall, these non-exhaustive DFT com-
putations suggest that halogen bonding is likely not at the
origin of the high stereocontrol featured by systems 1d and
1e. Initially, the dibromo-substituted complex 1e was de-
signed to explore experimentally the influence of the halogen
substituent on the stereoselectivity, by comparison with the di-
chloro-substituted 1d system; bromide is clearly bulkier and
also more prone to halogen bonding than chloride. As
a matter of fact, both 1d and 1e systems lead to close stereo-
selectivities in the ROP of rac-MLABz and rac-MLAAll. Given that
in the bromo- and chloro-substituted catalyst systems 1d and
1e there is a delicate balance between steric and electronic
factors, it remains difficult to deduce which is the actual driv-
ing phenomenon responsible for the improved stereocontrol
provided by these halogen-substituted systems in the ROP of
MLARs.
Figure 14. Schematic representation of model mononuclear (top) and dinu-
clear (bottom) intermediates in the ROP of rac-MLAAll mediated by
[Y(ONOOCl2)] complexes used for DFT computations (BP86-RI/def-TZVP),
showing the different arrangements of the k2-coordinated O,O-b-alkoxybuty-
rate moiety investigated and relative computed energies (see the Support-
ing Information).
a strong influence of the stereoenrichment on the melting
temperature (Tm; Table 1). The Tm values increased with Pr
values from Tm =718C for Pr =0.79 up to Tm =117 8C for Pr >
0.95 (Figure S9; Table 1, entry 12). A melting transition was ob-
served only for the most stereoregular PMLABzs (Pr ꢀ 0.79), il-
lustrating the close stereoregularity/semicrystallinity relation-
ship. The melting temperature of highly syndiotactic-enriched
PMLABzs (Tm =114–1178C) remained significantly lower than
that of isotactic PMLABzs (Tm =180–1908C).[15,25] All the PMLABzs
showed a glass-transition temperature in the range Tg =30–
358C.
Thermal analysis of the PMLARs
Analysis by differential scanning calorimetry (DSC) of the ther-
mal behavior of the synthesized PMLABzs showed, as expected,
1
Complex 1d is dinuclear in the solid state, as established by X-ray diffraction
studies. In toluene or benzene solutions, complexes 1d and 1e feature a com-
plex NMR behavior and well-resolved, informative spectra could be obtained
at low temperature (see the Supporting Information Figure S0 for 1d). DOSY
NMR spectroscopic analysis for 1d in [D8]toluene at room temperature (that
is, under the same conditions as those used for polymerizations) was consis-
tent with a mononuclear structure (see the Supporting Information).
A similar trend was obtained with PMLAAlls. The highest
melting temperatures were observed for the most stereoregu-
lar polymers, Tm =112 8C for Pr > 0.95 (Figure S10; Table 2,
entry 5), whereas no melting temperature was recorded for the
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7638
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
least syndiotactic-enriched PMLAAll (Pr ꢁ0.81). Melting tempera-
tures were also recorded for PMLAMes, only for highly syndio-
tactic-enriched stereosequences (Pr/r >0.76), whereas the high-
est Tm value was reached for insoluble PMLAMes (Tm =2128C;
Table 3, entry 6). These Tm values of syndiotactic PMLAMes are
higher than those of isotactic PMLAMes (Tm =137–1518C).[26] In
the case of the PMLAMe obtained from complex 1d, a second
minor melting temperature was observed at Tm =1978C
(Figure 15). The presence of two Tm values may arise from the
coexistence of two distinct crystalline phases.[26] The concomi-
tant increase of both Tm and melting enthalpy values with the
Pr/r of PMLAMes synthesized from the ROP of rac-MLAMe mediat-
ed by complexes 1a–c was fitted empirically (Table S8). This
model allowed the syndiotactic enrichment of insoluble
PMLAMes to be evaluated (Figure S11–S13). Pr/r values of 0.89
and 0.92 were thus estimated for the PMLAMes obtained using
complexes 1d and 1e, respectively (see above). The validity of
these values was verified by the linear correlation of the melt-
ing temperature with Pr/r values (Figure 16). A glass transition
temperature Tg ꢂ408C was recorded for all PMLAMes.
Conclusions
Highly syndioselective ROPs of racemic alkyl b-malolactonates
(rac-MALRs), mediated by [amino-alkoxy-bis(phenolate)]-yttrium
amido complexes 1a–e in the presence of a co-initiator (iso-
propanol), have been performed. All yttrium complexes were
efficient, providing the best activities recorded for this class of
monomer (TOFꢀ3000 hÀ1), and afforded well-defined PMLARs
(4700 gmolÀ1 ꢁMn ꢁ21000 gmolÀ1) with good control of the
polymerization (“M ꢁ1.55). In contrast to the ROP of other re-
lated monomers, such as rac-LA and even rac-BL, the yttrium
catalysts bearing halogen-disubstituted ligands were the most
stereoselective (Pr ꢀ 0.95). Effective chain-end control mecha-
nisms were attributed to Markov first-order mechanism for the
ROP of rac-MLABz and rac-MLAAll, and to Markov second-order
for the ROP of rac-MLAMe. The high stereoselectivity achieved
with the halogen-substituted catalysts definitively hints at elec-
tronic effects, which are clearly of different nature than those
pinpointed with aryl-substituted catalysts for ROP of the sim-
pler rac-b-butyrolactone.[8d] Although the definitive scenario of
these electronic interactions could not be delineated yet, DFT
computations performed on models of the active center sug-
gest that halogen bonding between the b-alkoxybutyrate
growing polyester chain and Cl(Br) moieties is probably not
the driving force for the experimentally observed high syndio-
selectivity of the catalysts.
Experimental Section
General conditions
All manipulations were performed under inert atmosphere (argon,
<3 ppm O2) using standard Schlenk, vacuum line, and glovebox
techniques. Solvents (toluene, tetrahydrofuran) were freshly dis-
tilled from Na/benzophenone under argon and degassed thor-
oughly by freeze-thaw-vacuum cycles prior to use. Bisphenol pro-
ligands and yttrium amide precursors 1a–d were synthesized ac-
cording to reported methods.[8d] Isopropyl alcohol (Acros) was dis-
tilled over Mg turnings under argon atmosphere and kept over ac-
tivated 3–4 Š molecular sieves. CDCl3 was dried over a mixture of 3
and 4 Š molecular sieves. Racemic benzyl b-malolactonate (rac-
MLABe), benzyl (S)-b-malolactonate ((S)-MLABe), racemic allyl b-malo-
lactonate (rac-MLAAll), and racemic methyl b-malolactonate (rac-
MLAMe) were synthesized from aspartic acid according to reported
procedures.[15,27]
Figure 15. DSC trace (cycle from À50 to 2508C; 108CminÀ1; first heating,
DHm =64.1 JgÀ1) of a PMLAMe obtained from the ROP of rac-MLAMe mediated
by the 1d/iPrOH system (Pr/r =0.89; Table 3, entry 5).
Instruments and measurements
1H (500 and 400 MHz) and 13C{1H} (125 MHz) NMR spectra were re-
corded with Bruker Avance AM 500 and Ascend 400 spectrometers
1
at 258C. H and 13C{1H} NMR spectra were referenced internally rel-
ative to SiMe4 (d=0 ppm) using the residual solvent resonances.
Average molar mass (Mn,SEC) and dispersity (“M =Mw/Mn) values of
the PMLABzs, PMLAAlls and PMLAMes were determined by size-exclu-
sion chromatography (SEC) in THF at 308C (flow rate=
1.0 mLminÀ1
) with a Polymer Laboratories PL50 apparatus
equipped with a refractive index detector and a set of two Resi-
Pore PLgel 3 mm MIXED-D 300”7.5 mm columns. The polymer
samples were dissolved in THF (2 mgmLÀ1). All elution curves were
calibrated with 12 monodisperse polystyrene standards (Mn
^
Figure 16. Validation of the estimated Pr/r
(
) values, obtained by extrapola-
tion of the fitting DHm =7.10À6 e17.992Pr/r, by linear correlation of Pr/r
Tm values of PMLAMes synthesized from the ROP of rac-MLAMe in the pres-
ence of 1a–e/iPrOH systems (fitting: y=354.07xÀ111.21; R2 =0.991).
(
^
) with
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7639
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
range=580–380,000 gmolÀ1); Mn,SEC values of the PMLABzs,
PMLAAlls, and PMLAMes were uncorrected for the possible difference
in hydrodynamic radius versus polystyrene.
The molar mass values of PMLABzs, PMLAAlls, and PMLAMes samples
were also determined by 1H NMR spectroscopic analysis in CDCl3
from the relative intensities of the signals of the PMLABz main-
chain methine hydrogen signals (-OCH(CO2Bz)CH2, d=5.58 ppm),
PMLAAll main-chain methine (-OCH(CO2All)CH2, d=5.52 ppm),
PMLAMe main-chain methine (-OCH(CO2Me)CH2, d=5.50 ppm), and
those of the isopropyl chain-end (-OCH(CH3)2, d=1.24 ppm).
iPrO-PMLAAll-H: H NMR (500 MHz; CDCl3, 258C): d=5.88 (m, 1nH;
CH2CH=CH2), 5.52 (m, 1nH; CH2CH(CO2All)O), 5.31 and 5.25 (d and
d, J=17 and 10 Hz, respectively (geminal proton-proton couplings
were not observed), 2nH; CH2CH=CH2), 4.64 (d, J=5 Hz, 2nH;
CH2CH=CH2), 3.41 (m, 1H; (CH3)CHO), 3.00 (m, 2nH;
CH2CH(CO2All)O), 1.24 ppm (m, 6H; (CH3)CHO) (Figure 6); 13C{1H}
NMR (125 MHz; CDCl3, 258C): d=168.1 and 167.8 (C=O), 131.3
(CH2CH=CH2), 119.0 (CH2CH=CH2), 68.7 (CH2CH(CO2All)O), 66.4
(CH2CH=CH2), 41.8 ((CH3)CHO), 35.4 (CH2CH(CO2All)O), 21.7 ppm
((CH3)CHO) (Figure S6).
1
iPrO-PMLAMe-H: H NMR (500 MHz; CDCl3, 258C): d=5.50 (m, 1nH;
1
Monomer conversions were calculated from H NMR spectra of the
CH2CH(CO2CH3)O), 3.76 (m, 3nH; CH2CH(CO2CH3)O), 3.45 (m, 1H;
(CH3)CHO), 3.01 (m, 2nH; CH2CH(CO2 CH3)O), 1.26 ppm (m, 6H;
(CH3)CHO) (Figure 2); 13C{1H} NMR (125 MHz; CDCl3, 258C): d=
168.7 and 168.3 (C=O), 68.7 (CH2CH(CO2CH3)O), 53.0
(CH2CH(CO2CH3)O), 35.6 (CH2CH(CO2CH3)O), 34.3 ((CH3)CHO),
22.5 ppm ((CH3)CHO) (Figure 3).
crude polymer samples in CDCl3 by using the integration (Int.) ratio
Int.PMLAR/[Int.PMLAR+Int.MLAR] of the methine hydrogen signals
(-OCH(CO2R)CH2: d=5.50–5.58 ppm for polymers and d=4.88 ppm
for monomers).
MALDI-ToF mass spectra were recorded at the CESAMO (Bordeaux,
France) with a Voyager mass spectrometer (Applied Biosystems)
equipped with a pulsed N2 laser source (337 nm) and a time-de-
layed extracted ion source. Spectra were recorded in the positive-
ion mode using the reflectron mode and with an accelerating volt-
age of 20 kV. A THF solution (1 mL) of the matrix (trans-3-indole-
acrylic acid, IAA; Aldrich, 99%) and a MeOH solution of the cation-
ization agent (NaI, 10 mgmLÀ1) were prepared. A fresh solution of
the polymer samples in THF (10 mgmLÀ1) was then prepared. The
three solutions were then rapidly combined in a 1:1:10 v/v
[matrix]/[sample]/[cationization agent] ratio. An aliquot (1–2 mL) of
the resulting solution was deposited onto the sample target and
vacuum-dried.
Computational details
All calculations were performed with the TURBOMOLE program
package using density functional theory (DFT).[28–31] The gradient
corrected density functional BP86 in combination with the resolu-
tion identity approximation (RI)[32,33] was applied for the geometry
optimizations of stationary point. A triple-z zeta valence quality
basis set def-TZVP was used for all atoms.[34] The stationary points
were characterized as energy minima (no negative Hessian eigen-
values) by vibrational frequency calculations at the same level of
theory. The results of calculations (total electronic energies of eight
intermediates and zero-point energy corrections) are presented in
Figure 14 and Table S7.
Differential scanning calorimetry (DSC) analyses were performed
with a Setaram DSC 131 apparatus calibrated with indium, at
a
rate of 108CminÀ1
,
under continuous flow of helium
(25 mLminÀ1), using aluminum capsules. The thermograms were
recorded according to the following cycles: À50 to 508C at
108CminÀ1; 50 to À508C at 108CminÀ1; À508C for 5 min; À50 to
Acknowledgements
200 or 2508C at 108CminÀ1; 200 to 308C at 108CminÀ1
.
The authors gratefully thank the CNRS and the Region Bre-
tagne for supporting part of this research (Ph.D. grant to
Statistical analysis and model fitting were performed by using soft-
ware OriginPro 8 from OriginLab Corporation and Mk1 and Mk2
models defined previously.[18]
1
C.G.J.). We thank Romain Ligny for performing VT H and DOSY
NMR spectroscopic analysis of complex 1d.
General procedure for the ROP of rac-MLAR
Keywords: density functional calculations
·
lactones
·
polymers · ring-opening polymerization · yttrium
In a typical experiment (Table 1, entry 3), a Schlenk flask was
charged in a glovebox with a solution of complex 1a (6.3 mg,
5.82 mmol) in toluene (0.58 mL), then iPrOH (0.50 mL, 5.82 mmol,
1.0 equiv vs. 1a) was added under stirring. After 5 min, rac-MLABz
was added rapidly (120 mg, 0.58 mmol, 100 equiv) and the mixture
was stirred at 208C for the appropriate time. The reaction was
quenched by addition of acetic acid (ca. 10 mL of a 1.6 mol·LÀ1 so-
lution in toluene). The resulting mixture was concentrated to dry-
ness under vacuum and the conversion was determined by
1H NMR spectroscopic analysis of the residue in CDCl3. The crude
polymer was then dissolved in CH2Cl2 (ca. 1 mL) and precipitated
in cold pentane (ca. 5 mL), filtered and dried under vacuum at
458C overnight (typical isolated yield 90–95%). The final polymer
was then analyzed by NMR, SEC, and DSC analyses (Table 1).
iPrO-PMLABz-H: 1H NMR (500 MHz; CDCl3, 258C): d=7.33 (br m,
5nH; C6H5), 5.55 (br m, 1nH; CH2CH(CO2Bz)O), 5.14 (br s, 2nH;
OCH2C6H5), 4.56 (br m, 1H; OCH(CH3)2), 2.94 (br m, 2nH;
CHCH2C(O)O), 1.24 ppm (m, 6H; OCH(CH3)2) (Figure 5); 13C{1H} NMR
(125 MHz; CDCl3, 258C): d=168.2–168.0 (C=O), 135.1 (C8), 128.3–
128.7 (C9–11), 68.8 (C(O)CH2CH(CO2Bz)O), 67.6 (OCH2C6H5), 65.7
((CH3)2CHO), 35.4 (OC(O)CH2CH), 20.9 ppm ((CH3)2CHO) (Figure S5).
[1] J.-F. Lutz, M. Ouchi, D. R. Liu, M. Sawamoto, Science 2013, 341, 1238149.
[2] J. W. Kramer, D. S. Treitler, E. W. Dunn, P. M. Castro, T. Roisnel, C. M.
Thomas, G. W. Coates, J. Am. Chem. Soc. 2009, 131, 16042–16044.
[3] C. G. Jaffredo, Y. Chapurina, S. M. Guillaume, J.-F. Carpentier, Angew.
Chem. Int. Ed. 2014, 53, 2687–2691; Angew. Chem. 2014, 126, 2725–
2729.
[4] J.-F. Carpentier, Angew. Chem. Int. Ed. 2010, 49, 2662–2663; Angew.
Chem. 2010, 122, 2720–2722.
[5] S. M. Guillaume, E. Kirillov, Y. Sarazin, J.-F. Carpentier, Chem. Eur. J. 2015,
21, 7988–8003.
[6] For leading reviews on stereoselective ROP of b-lactones, see: a) C. M.
Thomas, Chem. Soc. Rev. 2010, 39, 165–173; b) J.-F. Carpentier, Macro-
mol. Rapid Commun. 2010, 31, 1696–1705.
[7] R. W. Lenz, Z. Jedlinski, Macromol. Symp. 1996, 107, 149–161.
[8] a) C.-X. Cai, A. Amgoune, C. W. Lehmann, J.-F. Carpentier, Chem.
Commun. 2004, 330–331; b) A. Amgoune, C. M. Thomas, J.-F. Carpenti-
er, Macromol. Rapid Commun. 2007, 28, 693–697; c) N. Ajellal, M.
Bouyahyi, A. Amgoune, C. M. Thomas, A. Bondon, I. Pillin, Y. Grohens, J.-
F. Carpentier, Macromolecules 2009, 42, 987–993; d) M. Bouyahyi, N.
Ajellal, E. Kirillov, C. M. Thomas, J.-F. Carpentier, Chem. Eur. J. 2011, 17,
1872–1883; e) J. S. Klitzke, T. Roisnel, E. Kirillov, O. d. L. Casagrande, J.-F.
Carpentier, Organometallics 2014, 33, 309–321; f) C. G. Jaffredo, M.
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7640
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Schmid, I. del Rosal, T. Mevel, P. W. Roesky, L. Maron, S. M. Guillaume,
Chem. Eur. J. 2014, 20, 14387–14402.
[22] E. L. Marshall, V. C. Gibson, H. S. Rzepa, J. Am. Chem. Soc. 2005, 127,
6048–6051.
[9] P. T. Altenbuchner, A. Kronast, S. Kissling, S. I. Vagin, E. Herdtweck, A.
PÅthig, P. Deglmann, R. Loos, B. Rieger, Chem. Eur. J. 2015, 21, 13609–
13617.
[10] For a recent review on [amino-alkoxy-bis(phenolate)] rare earth com-
plexes and their use in ROP catalysis, see: J.-F. Carpentier, Organometal-
lics 2015, 34, 4175–4189.
[23] a) L. C. Gilday, S. W. Robinson, T. A. Barendt, M. J. Langton, B. R. Mulla-
ney, P. D. Beer, Chem. Rev. 2015, 115, 7118–7195; b) G. Cavallo, P. Me-
trangolo, R. Milani, T. Pilati, A. Priimagi, G. Resnati, G. Terraneo, Chem.
Rev. 2016, 116, 2478–2601.
[24] P. Auffinger, F. A. Hays, E. Westhof, P. S. Ho, Proc. Natl. Acad. Sci. USA
2004, 101, 16789–16794.
[11] C. G. Jaffredo, S. M. Guillaume, Polym. Chem. 2014, 5, 4168–4194.
[12] For this experiment, the monomer was purified by flash chromatogra-
phy followed by three flash distillations. The fourth distillation always
led to degradation of the monomer.
[25] R. A. Gross, Y. Zhang, G. Konrad, R. W. Lenz, Macromolecules 1988, 21,
2657–2668.
[26] C. E. Fernµndez, M. Mancera, E. Holler, J. A. Galbis, S. MuÇoz-Guerra,
Polymer 2006, 47, 6501–6508.
[13] A. Amgoune, C. M. Thomas, S. Ilinca, T. Roisnel, J.-F. Carpentier, Angew.
Chem. Int. Ed. 2006, 45, 2782–2784; Angew. Chem. 2006, 118, 2848–
2850.
[27] a) M.-A. Leboucher-Durand, V. Langlois, P. GuØrin, React. Funct. Polym.
1996, 31, 57–65; b) S. Cammas, I. Renard, V. Langlois, P. GuØrin, Polymer
1996, 37, 4215–4220.
[14] P. Guerin, M. Vert, C. Braud, R. Lenz, Polym. Bull. 1985, 14, 187–192.
[15] P. Guerin, J. Francillette, C. Braud, M. Vert, Makromol. Chem. Macromol.
Symp. 1986, 6, 305–314.
[16] M. Helou, G. Moriceau, Z. W. Huang, S. Cammas-Marion, S. M. Guillaume,
Polym. Chem. 2011, 2, 840–850.
[28] R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kçlmel, Chem. Phys. Lett. 1989,
162, 165–169.
[29] O. Treutler, R. Ahlrichs, J. Chem. Phys. 1995, 102, 346–354.
[30] K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs, Chem. Phys. Lett.
1995, 242, 652–660.
[17] C. G. Jaffredo, J.-F. Carpentier, S. M. Guillaume, Macromolecules 2013, 46,
6765–6776.
[31] TURBOMOLE V6.2 2010, a development of University of Karlsruhe and
Forschungszentrum Karlsruhe GmbH, 1989–2007, TURBOMOLE GmbH,
since 2007; available from http://www.turbomole.com.
[32] M. Sierka, A. Hogekamp, R. Ahlrichs, J. Chem. Phys. 2003, 118, 9136–
9148.
[33] P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 2004, 384,
103–107.
[34] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829–5835.
[18] M. Sepulchre, Makromol. Chem. 1988, 189, 1117–1131. The Mk1 model
expressions for triad sequences are: (mm)=[m]2, (rm)=(mr)=[mr] and
(rr)=[r]2. The probability of racemic linkage in the polymer chain is
p
then defined by Pr = (rr).
[19] Deconvolution performed using OriginPro 8 software from OriginLab
Corporation.
[20] RMS=((S(IobsÀIcalcd)2/8)0.5). The fidelity of the statistical model remained
dependent of the regularity of the polymer analyzed.
[21] a) P. Hormnirun, E. L. Marshall, V. C. Gibson, A. J. P. White, D. J. Williams,
J. Am. Chem. Soc. 2004, 126, 2688–2689; b) P. Hormnirun, E. L. Marshall,
V. C. Gibson, R. I. Pugh, A. J. P. White, Proc. Natl. Acad. Sci. USA 2006,
103, 15343–15348.
Received: January 17, 2016
Published online on April 15, 2016
Chem. Eur. J. 2016, 22, 7629 – 7641
www.chemeurj.org
7641
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim