Catalyst activity or stability: The dilemma in Pd-catalyzed polyketone synthesis

Dalton

Transactions

View Article Online

View Journal | View Issue

PAPER

Catalyst activity or stability: the dilemma in

Pd-catalyzed polyketone synthesis†

Cite this: Dalton Trans., 2013, 42, 14583

Francesco Amoroso,^a,bEnnio Zangrando,^aCarla Carfagna,^cChristian Müller,^d

Dieter Vogt,^eMohamed Hagar,^fFabio Ragaini*^gand Barbara Milani*^a

A series of Pd-complexes containing nonsymmetrical bis(aryl-imino)acenaphthene (Ar-BIAN) ligands, charac-

terized by substituents on the meta positions of the aryl rings, have been synthesized, characterized and

applied in CO/vinyl arene copolymerization reactions. Crystal structures of two neutral Pd-complexes have

been solved allowing comparison of the bonding properties of the ligand. Kinetic and mechanistic investi-

gations on these complexes have been performed. The kinetic investigations indicate that in general

ligands with electron-withdrawing substituents give more active, but less stable, catalytic systems,

although steric eﬀects also play a role. The good performance observed with nonsymmetrical ligands is at

least in part due to a compromise between catalyst activity and lifetime, leading to a higher overall pro-

ductivity with respect to catalysts based on their symmetrical counterparts. Additionally, careful analysis of

the reaction proﬁles provided information on the catalyst deactivation pathway. The latter begins with

the reduction of a Pd(^II) Ar-BIAN complex to the corresponding Pd(0) species, a reaction that can be

reverted by the action of benzoquinone. Then the ligand is lost, a process that appears to be facilitated by

the contemporary coordination of an oleﬁn or a CO molecule. The so formed Pd(0) complex immediately

reacts with another molecule of the initial Pd(^II) complex to give a Pd(I) dimeric species that irreversibly

evolves to metallic palladium. Mechanistic investigations performed on the complex with a nonsymmetri-

cal Ar-BIAN probe evidence that the detected intermediates are characterized by the Pd–C bond trans to

the Pd–N bond of the aryl ring bearing electron-withdrawing substituents. In addition, the intermediate

resulting from the insertion of 4-methylstyrene into the Pd–acyl bond is a ﬁve-member palladacycle and

not the open-chain η³-allylic species observed for complexes with Ar-BIANs substituted in ortho position.

Received 31st May 2013,

Accepted 19th July 2013

DOI: 10.1039/c3dt51425k

www.rsc.org/dalton

nitrogen atoms, have been known since 1969.¹However, only in

1992 the first straightforward synthetic procedure was reported

Introduction

by Elsevier.²Afterwards, the main application of these mole-

cules as ancillary ligands in homogeneous catalysis was brought

to the attention of a wider scientific community by Brookhart’s

discovery that the relevant nickel and palladium complexes

generate extremely eﬃcient catalysts for polyethylene synthesis.³

This finding had an impressive impact, in the field of catalysis

for polymerization, resulting in a huge development of ligands

of this family.⁴The same palladium complexes were also

revealed to be able to catalyze the copolymerization of ethylene

with polar vinyl monomers, a highly challenging reaction.^5–7

Ar-BIANs applied as ancillary ligands in catalysts for

polymerization reactions share some common features: (i) the

two nitrogen atoms bear identical aryl rings; (ii) these aryl

rings are almost invariably substituted in the ortho and/or para

positions; (iii) the nature and the number of substituents

aﬀect the yield, the selectivity,^4,8and the microstructure of the

polymer⁸and the relative amount of comonomers inserted

into the polymer chain in both α-olefins⁹and in ethylene/polar

vinyl monomer copolymerization.^6,10,11

α-Diimine ligands, having a 1,4-diaza-1,3-butadiene or an

acenaphthene skeleton (Ar-BIAN) and aryl rings on the imine

^aDipartimento di Scienze Chimiche e Farmaceutiche, Università di Trieste, Via Licio

Giorgieri 1, 34127 Trieste, Italy. E-mail: milaniba@units.it; Fax: +39 040 5583903;

Tel: +39 040 5583956

^bDipartimento di Chimica, Fisica e Ambiente, Università di Udine, Via del

Cotonificio 108, 33100 Udine, Italy. E-mail: francesco.amoroso@uniud.it;

Fax: +39 0432 558803; Tel: +39 0432 558755

^cDipartimento di Scienze Biomolecolari, Università di Urbino, Piazza Rinascimento 6,

61029 Urbino, Italy. E-mail: carla.carfagna@uniurb.it; Fax: +39 0722 303311;

Tel: +39 0722 303312

^dInstitut für Chemie und Biochemie-Anorganische Chemie, Freie Universität Berlin,

14195 Berlin, Germany. E-mail: c.mueller@fu-berlin.de; Tel: +49 30 838 54004

^eSchool of Chemistry, University of Edinburgh, Edinburgh EH9 3JJ, Scotland, UK.

E-mail: d.vogt@ed.ac.uk; Tel: +44 (0)131 651 7767

^fChemistry Department, Alexandria University, Ibrahimia, Alexandria 21321, Egypt

^gDipartimento di Chimica, Università di Milano, Via C. Golgi 19, 20133 Milano,

Italy. E-mail: fabio.ragaini@unimi.it; Fax: +39 02 50314405; Tel: +39 02 50314373

†Electronic supplementary information (ESI) available. CCDC 942156 and

942157. For ESI and crystallographic data in CIF or other electronic format see

DOI: 10.1039/c3dt51425k

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14583

View Article Online

Paper

Dalton Transactions

Palladium complexes with ortho substituted Ar-BIANs copolymerization reactions, it might indicate the validity of a

(^oAr-BIANs) were applied as model compounds to unravel the more general principle.

mechanism of the polymer chain growth during the CO/alkene

With the aim to unravel the nature of the diﬀerent catalytic

copolymerization.¹²Some of these complexes were also tested behaviors of complexes with symmetrical and nonsymmetrical

as precatalysts in the CO/4-Me-styrene copolymerization (4-Me- Ar-BIANs we have chosen the CO/vinyl arene copolymerization

styrene = 4-methylstyrene): low productivity (15 g CP per g Pd; as the model reaction, and on palladium complexes with

g CP per g Pd = grams of copolymer per gram of palladium) or selected ligands (Fig. 1) we have performed: (1) kinetic investi-

inactivity was observed depending on the substituents on the gations; (2) the related modelization; (3) in situ NMR mecha-

ortho positions. Nevertheless, polyketones made of isotactic nistic studies on the reactivity of the complexes with the

stereoblocks of opposite relative configuration were obtained comonomers. In addition, with these purposes the series of

despite the C_2vsymmetry of the ligands.^13–15

nonsymmetrical ligands has been extended to a new com-

For the first time we introduced nonsymmetrical Ar,Ar′- ponent, ligand 3, characterized by one aryl ring substituted in

BIANs characterized by diﬀerent aryl rings substituted in meta meta and para positions with methoxy groups and the other

position on the two N-donors.¹⁶The organometallic palladium aryl ring having fluorine atoms in the meta positions.

complexes, [Pd(CH₃)(CH₃CN)(^mAr-BIAN)][PF₆], with symmetri-

cally and nonsymmetrically meta substituted Ar-BIANs (^mAr-

BIANs), were applied as catalysts in CO/vinyl arene copolymeri-

zation. We found that moving the substituents on the aryl

rings of Ar-BIAN from the ortho to the meta positions resulted

in an enhancement of productivity of more than two orders of

Results and discussion

Synthesis and characterization of the new ligands and the

related Pd-complexes

Ligands 3 and 7 are reported here for the first time. The syn-

thetic methodology for ligand 7 is based on the well known

condensation reaction of acenaphthenequinone with 3,5-

magnitude, and a productivity value of 3.70 kg CP per g Pd

was reached.¹⁷Polyketones with an atactic microstructure and

an M_wof 47 000 (M_w/M_n= 2.0) were produced. Both pro-

ductivity and molecular weight were aﬀected by the nature and

difluoro aniline in the presence of ZnCl₂as a templating

agent.^19,20

the number of substituents on the ^mAr-BIANs. The catalysts

with the nonsymmetrical Ar,Ar′-BIAN are more productive than

those having the corresponding symmetrical ligands and lead

to polyketones of higher molecular weight.

The synthesis of the new nonsymmetrical Ar,Ar′-BIAN 3 was

performed by the transimination reaction between the zinc

derivative of ligand 7 and 3,4,5-trimethoxy aniline (Scheme 1),

similar to what was previously described by some of us for

other nonsymmetrical Ar,Ar′-BIAN derivatives.¹⁶The reaction

proceeds quickly and the mixed ligand could be isolated in

52% yield after chromatographic purification. It should be

noted that previously mixed ligands had only been obtained

by this methodology when the starting compound had two tri-

fluoromethyl groups on each aryl ring, according to the idea

The positive eﬀect of nonsymmetrical Ar,Ar′-BIAN ligands

has been very recently demonstrated even in the ethylene/

methyl acrylate cooligomerization: the catalyst with the non-

symmetrical ligand showed a higher productivity and a higher

content of the inserted polar monomer with respect to those

with the related symmetrical ligands.¹⁸

These catalytic results suggest that the subtle unbalance of

the electronic and steric properties of the N-donor atoms of

the Ar-BIAN ligands has a positive eﬀect on the catalyst per-

formance and, thanks to the similarity between these two

Fig. 1 The studied Ar-BIANs and their numbering scheme.

14584 | Dalton Trans., 2013, 42, 14583–14602

Scheme 1 Synthetic pathway for ligand 3.

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

electron-withdrawing groups (Scheme 2). According to the

¹H NMR spectra the trans to cis ratio was 3 : 1 and 5 : 1 for 3a

and 3b, respectively.

Single crystals suitable for X-ray analysis were obtained for

complexes 1a and 3a by slow diﬀusion of diethyl ether into a

CD₂Cl₂solution of the complexes kept at 4 °C for one week.

The crystal structure of the complex [Pd(CH₃)Cl(5)] with the

symmetrically substituted ^mAr-BIAN with CF₃groups in meta

position has already been reported,²⁴thus allowing a compari-

son of its structural features with those of 1a and 3a.

In the crystals of complexes with the nonsymmetrical Ar,

Ar′-BIANs 1 and 3, only the trans isomer was detected in the

unit cell (Fig. 2, Table 1), in agreement with the NMR charac-

terization in solution that showed this species to be the pre-

vailing isomer. This confirms our previous report regarding

the complex [Pd(CH₃)Cl(8)], where 8 is the nonsymmetrical Ar,

Ar′-BIAN characterized by one aryl ring substituted by CH₃groups

in ortho position and the other ring having CF₃substituents

in meta.¹⁸

In the complexes the palladium ion attains the expected

square planar coordination geometry. In agreement with the

trans influence of the Pd–CH₃fragment, in both complexes the

Pd–N1 bond distance trans to it is remarkably longer than the

other Pd–N bond length. If a comparison of coordination

bond distances is limited to complexes 1a, 3a and 5a (all

Scheme 2 Synthetic pathway for the monocationic complexes 1b–7b and the

numbering scheme for both neutral and monocationic derivatives.

that electron-withdrawing substituents on the starting Ar-BIAN having ^mAr-BIAN), the Pd–N1 bond length is 2.225(8), 2.205(5)

should render transimination easier. The present result indi- and 2.227(4) Å, respectively. This indicates that ligand 3

cates that even less electron-withdrawing groups can be appears to be more strongly bonded to palladium, as a result

eﬃciently used in this protocol.

of lower steric hindrance and electron-withdrawing eﬀect of

The monocationic Pd complexes, [Pd(CH₃)(CH₃CN)- F with respect to CF₃. This is also supported by the corres-

(Ar-BIAN)][PF₆] 1b–7b, applied as precatalysts in the model ponding values measured for the Pd–N2 bond distance,

reaction were synthesized by halogen abstraction on the corres- trans to chloride, which is the shortest among the three

ponding neutral derivatives, [Pd(CH₃)Cl(Ar-BIAN)][PF₆] 1a–7a, (2.042(5) Å for 3a vs. 2.063(7) and 2.073(5) Å for 1a and 5a,

following the literature methodology (Scheme 2).^17,21

respectively).

Ligands 3 and 7 and the related palladium complexes were

Another similarity in the structural features of the studied

characterized in solution by NMR spectroscopy. For Ar-BIANs, complexes is concerned with the orientation of the substituted

(E,E) and (E,Z) isomers are possible according to the relative phenyl rings with respect to the BIAN plane. The dihedral

configuration of the aryl rings with respect to the CvN imine angle between the BIAN plane and the aryl ring on the side of

bonds. In general, only the E,E isomer is observed for sym- the Pd–Cl bond falls in the range 52.5°–64.5°, while that on

metrical Ar-BIANs,¹⁹but a variable amount of the E,Z isomer the side of the Pd–CH₃bond is 76.4°–80.06° indicating that

is always found in the case of nonsymmetrical Ar,Ar′-BIAN the aryl ring is almost orthogonally oriented with respect to

derivatives^16–18,20and for alkyl-BIANs.^22,23For both ligands 3 the BIAN plane. These values indicate a greater conformational

and 7, the ¹H NMR spectrum recorded at room temperature in freedom on the Pd–Cl side than on the Pd–CH₃part, reason-

CDCl₃and CD₂Cl₂, respectively, indicated the presence in solu- ably due to the less steric hindrance of chloride with respect to

tion of the E,E isomer only. Though this was expected for the that of the methyl group.

latter, it is surprising for the former. The NMR characteri-

For the complexes with the nonsymmetrical Ar,Ar′-BIANs,

zation of complexes 7a and 7b was in line with data reported in the preference, both in solution and in the solid state, towards

the literature for analogous compounds with symmetrically the trans isomer is in agreement with the literature data on

substituted Ar-BIANs.^12,17,18In the case of complexes 3a and palladium complexes with bidentate nitrogen-donor ligands

3b, bearing the nonsymmetrical ligand 3 on palladium, in having electronically nonequivalent N atoms^21,25–28and with

analogy with what we reported for complexes 1a, 2a and 1b, the general trend that the weakest ligand (here the nitrogen

2b,¹⁷cis and trans isomers were observed in solution diﬀering atom bearing the fluorine substituted aryl ring) is found trans

in the relative position of the Pd–CH₃fragment with respect to to the stronger trans-labilizing ligand (here the methyl group).

the two halves of the ligand. We define as trans the isomer The opposite preference towards the cis isomer was observed

having the Pd–CH₃fragment, or more generally the Pd–C in the case of high steric hindrance around the palladium

bond, trans to the Pd–N bond of the aryl ring substituted with centre.²⁹

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14585

View Article Online

Paper

Dalton Transactions

Fig. 2 The ORTEP drawing (thermal ellipsoids at the 35% probability level) of complexes 1a, 3a and 8a.¹⁸

productive catalyst among those tested (3.70 kg CP per g Pd),

while in the case of 4-Me-styrene the highest productivity was

reached with the catalyst obtained from complex 1b (2.36 kg

CP per g Pd). For both alkenes, the catalyst with ligand 1

showed a productivity remarkably higher than that of the cata-

lysts with the related symmetrically substituted ^mAr-BIANs, 4

and 5. The same trend was observed for the CO/styrene copoly-

merization catalyzed by 2b with respect to 5b and 6b, while in

the CO/4-Me-styrene copolymerization the productivity

obtained with 2b was slightly lower than that obtained with 6b

(1.64 kg CP per g Pd vs. 1.95 kg CP per g Pd).¹⁷Moreover, in

the CO/styrene copolymerization no formation of inactive palla-

dium metal was observed, whereas in the case of CO/4-Me-

styrene copolymerization, catalyst decomposition was some-

times observed even in the presence of 1,4-benzoquinone.¹⁷

To have a better understanding of the catalytic behavior of

Table 1 Selected bond lengths and angles for complexes 1a and 3a

1a

3a

Lengths (Å)

Pd–N(1)

Pd–N(2)

Pd–C(1)

Pd–Cl(1)

2.225(8)

2.063(7)

2.011(10)

2.294(3)

2.205(5)

2.042(5)

2.020(6)

2.279(2)

Angles (°)

N(1)–Pd–N(2)

N(1)–Pd–C(1)

N(1)–Pd–Cl(1)

N(2)–Pd–C(1)

N(2)–Pd–Cl(1)

C(1)–Pd–Cl(1)

78.3(3)

171.6(3)

99.2(2)

93.3(4)

177.3(2)

89.2(3)

78.7(2)

172.1(3)

100.56(15)

93.9(3)

177.17(17)

87.0(2)

Kinetic investigations and modeling

General outcome. We reported that the monocationic com- these complexes some kinetic investigations were performed

plexes 1b, 2b, 4b–6b generated active catalysts for the CO/ by taking advantage of an automated parallel reactor (AMTEC

styrene and the CO/4-Me-styrene copolymerization yielding the SPR16), allowing us to monitor the CO uptake during the co-

corresponding polyketones with an atactic microstructure.¹⁷polymerization reactions. All the catalytic tests were carried out

Both the productivity and the molecular weight of the syn- in 2,2,2-trifluoroethanol (TFE) in the presence of 1,4-benzo-

thesized copolymers were aﬀected by the nature of the substi- quinone (BQ; [BQ]/[Pd] = 5), at T = 30 °C, under a preferential CO

tuents. Even though some diﬀerences in the order of pressure of 5 bar, and with a ratio [alkene]/[Pd] = 22 000 for the

productivity of the catalysts were observed for the two alkenes runs involving 4-methylstyrene and 25 000 for those with

and the eﬀect of the ligand nature was more pronounced in styrene. The large excess of the vinyl arene used ensures

the case of styrene than for 4-methylstyrene, for both alkenes conditions of pseudo-zero order with respect to the alkene

the catalysts with the nonsymmetrical Ar,Ar′-BIANs 1 and 2 even for a longer reaction time.

were found to be the most productive within the series.¹⁷

In particular, with styrene, ligand 2 generated the most ligand bonded to palladium, within the first 24 hours of

As a general comment, depending on the nature of the

14586 | Dalton Trans., 2013, 42, 14583–14602

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

reaction some catalysts show an almost linear dependence of reaction. Afterwards the activity of the catalyst with ligand 1

CO uptake with time while others deactivate to a much higher decreases and becomes lower than that of the catalyst with

extent. This catalytic behavior is diﬀerent from that found by ligand 4 (Fig. 3, right).

us for analogous palladium complexes with ligands belonging

When the same comparison is made among the catalysts

to the family of 1,10-phenanthroline²¹that over a range of 90 h with the nonsymmetrical ligand 2 and those with the related

showed a linear dependence of CO uptake with time, in agree- symmetrical ligands 5 and 6, the active species containing

ment with the stronger coordinating capability of phenanthro- ligand 6 was found to be much more productive than that

line ligands with respect to that of Ar-BIANs.²⁰

having the related nonsymmetrical Ar,Ar′-BIAN 2 in the CO/

The comparison of the CO uptake curves of precatalyst 1b 4-Me-styrene copolymerization (Fig. 4, left).

with those of complexes 4b and 5b having the corresponding

The lower productivity of 2b with respect to that of 6b

symmetrical ^mAr-BIANs probes evidences that for both alkenes might be related to the presence of the CF₃groups on one of

the catalyst with ligand 5 is deactivated within the first hour of the aryl rings of 2 that might have a negative eﬀect for steric

reaction, while for the catalyst with ligand 1 deactivation is reasons. This hypothesis is supported by the investigation of

limited and the complex with 4 is the only one that does not the catalytic behavior of complex 3b, where the nonsymmetri-

show any deactivation process for at least 24 hours (Fig. 3). cal Ar,Ar′-BIAN 3 is analogous to 2 but with fluorine substitu-

With 4-Me-styrene, the active species containing the nonsym- ents in place of CF₃groups (Fig. 5).

metrical ligand 1 is remarkably more productive than the cata-

By comparing the CO uptake curves of 3b, 6b and 7b it is

lysts with the two symmetrical counterparts (Fig. 3, left), while evident that, in analogy with the trends involving 5b, the sym-

for styrene the same trend is observed in the first 14 hours of metrical ligand with the electron-withdrawing substituent

Fig. 3 CO/vinyl arene copolymerization: eﬀect of Ar-BIAN ligand on CO uptake. Catalyst precursor: [Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆] (1b, 4b, 5b). Reaction con-

ditions: n_Pd= 0.106 × 10⁻⁵mol, T = 30 °C, TFE V = 5 mL, P_CO= 5 bar, t = 24 h, [BQ]/[Pd] = 5, vinyl arene V = 3 mL, [styrene]/[Pd] = 25 000, [4-Me-styrene]/[Pd] =

22 000. Left: CO/4-Me-styrene; right: CO/styrene copolymerization.

Fig. 4 CO/vinyl arene copolymerization: eﬀect of Ar-BIAN ligand on CO uptake. Catalyst precursor: [Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆] (2b, 5b, 6b). Reaction con-

ditions: see Fig. 3. Left: CO/4-Me-styrene; right: CO/styrene.

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14587

View Article Online

Paper

Dalton Transactions

Fig. 5 CO/vinyl arene copolymerization: eﬀect of Ar-BIAN ligand on CO uptake. Catalyst precursor: [Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆] (3b, 6b, 7b). Reaction con-

ditions: see Fig. 3. Left: CO/4-Me-styrene; right: CO/styrene.

fluorine, 7, generates the least productive species for both

alkenes. The catalyst with the nonsymmetrical ligand 3 is

found to be less productive than that with the corresponding

symmetrical methoxy-substituted Ar-BIAN 6 for the CO/styrene

copolymerization (Fig. 5, right), whereas in the CO/4-Me-

styrene copolymerization, the catalyst with 3 shows a remark-

ably higher productivity than the active species with the sym-

metrically substituted ligand 6 (Fig. 5, left). Thus, at least for

the 4-Me-styrene the replacement of the CF₃groups with fluor-

ine atoms had a remarkable positive eﬀect, in agreement with

the lower value of both the Hammett σ constant (electronic

eﬀect) and the Taft parameter (steric eﬀect) of fluorine with

respect to CF₃.³⁰

Set up of the kinetic modeling. Although the aforemen-

tioned diﬀerences in total productivity are well defined, a

closer examination of the reaction profiles evidences that the

reason for these behaviors is far from clear-cut. In most cases

productivity inversions are noted over time with diﬀerent cata-

lysts and some of those showing the lowest final productivity

are indeed the most active in the early stages of the reaction,

suggesting that catalyst deactivation plays a relevant role in the

final productivity order.

In order to gain a better understanding of the kinetic

outcome of the diﬀerent reactions and of the productivity

inversions mentioned above, we engaged in a mathematical

modeling of the reaction profiles that could take into account

both the chain propagation and the catalyst deactivation. The

followed approach is not a standard one and several problems

needed to be solved. In order to keep the article text as

compact as possible, only a general description of the method

employed is reported, whereas more details can be found in

the ESI.†

Scheme 3 Simpliﬁed reaction pathways for kinetic modeling.

concentration of both reagents is constant during the reaction

and the rate depends only on the active palladium con-

centration, which decreases with time because of catalyst

decomposition:

rate ¼ k₁½Pd^IIꢀ

ð1Þ

where k₁is the apparent first order kinetic constant with

respect to the active catalyst concentration.

Experimentally, the actual reaction rate is measured by the

CO consumption (mmol h⁻¹), but in order to have the kinetic

constants expressed in the usual (time)⁻¹units (for first order

kinetics) it is better to divide the CO absorption (in mmol)

by the reaction solution volume (V = 8 mL for all reactions).

That is:

rate ¼ d½ðCO abs=mmolÞ=ðV=mLÞꢀ=dt

ð2Þ

n

m

¼ k_p½Pd^IIꢀ½olefinꢀ ðP_CO

Þ

¼ k₁½Pd^IIꢀ

where [Pd^II] depends on time.

As far as catalyst decomposition is concerned, it is well

recognized that in the present catalytic system this is due to

palladium(^II) reduction, as supported by the fact that benzo-

quinone retards such deactivation by oxidizing back the palla-

A

general reaction pathway evidencing the catalyst

decomposition is shown in Scheme 3. A palladium(^II) species

catalyzes the copolymerization reaction with a kinetic constant dium(0) complex (processes associated with k_d1and k_−d1in

k_p(the kinetic order with respect to the reagents has been

left unspecified for the moment). Since the olefin is present

Scheme 3) and that metallic palladium is observed sometimes

after complete catalyst deactivation. There is little doubt that

in large excess and CO is continuously replenished, the the initial step of decomposition is the reduction of a single

14588 | Dalton Trans., 2013, 42, 14583–14602

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

palladium(II) complex to palladium(0). Although several possi-

bilities exist for such a process, it is almost surely described by

a first order kinetics with respect to palladium(^II). What occurs

later is less obvious. Several papers have been devoted to the

kinetics of metallic nanoparticles formation, but they usually

focus on later stages of the reaction.^31,32As far as the early

stages of decomposition are concerned, we can identify two

general reaction pathways.

In the first pathway (path A in Scheme 3), the palladium(0)

complex decomposes by an irreversible unimolecular pathway,

e.g. an irreversible loss of the nitrogen ligand, eventually giving

metallic palladium. Alternatively, the palladium(0) complex can

react with a palladium(^II) complex to aﬀord an inactive palla-

dium(^I) dimer (path B in Scheme 3). Precedents exist for such a

process.^33–35Very recently some of us identified a series of

complexes of general structure [Pd^IL²X(μ-CO)]₂(L²= 1,10-

phenanthroline, 2,2′-bipyridine or their substituted analogues,

X = halide, carboxylate, –C(O)OMe) as intermediates in the

reduction of palladium(^II) complexes by CO and water or alco-

hols under conditions close to those employed in the presently

studied system³⁶and the formation of the corresponding

Ar-BIAN complexes may as well occur easily.³⁷

½Pd^IIꢀ_t¼ ½Pd^IIꢀ₀e^ꢁk2t

where [Pd^II]₀and [Pd^II]_tare the palladium(II) concentrations at

time zero and t respectively.

From eqn (4) and (2), eqn (5) is derived, which leads to

eqn (6) after integration.

ð4Þ

rate ¼ d½ðCO absÞ=Vꢀ=dt ¼ k₁½Pd^IIꢀ₀e^ꢁk2t

ðCO absÞ=V ¼ ðk₁=k₂Þ½Pd^IIꢀ₀ð1 ꢁ e^ꢁk2t

ð5Þ

ð6Þ

Þ

where [Pd^II]₀= 1.32 × 10⁻⁴M.

Eqn (6) is of the type Y = a(1 − e^−bX), a Stirling type function

that can be modeled by a suitably modified fitting function in

the program Origin, to give the best a and b values, from

which k₁and k₂are immediately calculated.

For the second order decomposition pathway, an analogous

procedure leads to the equations:

2

d½Pd^IIꢀ=dt ¼ ꢁk₃½Pd^IIꢀ

½Pd^IIꢀ_t¼ ½Pd^IIꢀ₀=ð1 þ k₃½Pd^IIꢀ₀tÞ

rate ¼ d½ðCO absÞ=Vꢀ=dt ¼ k₁½Pd^IIꢀ₀=ð1 þ k₃½Pd^IIꢀ₀tÞ ð9Þ

ð7Þ

ð8Þ

Modeling the complete system without knowing the benzo-

quinone concentration vs. time profile is not possible in

general, even if the steady state approximation is applied to

ðCO absÞ=V ¼ ðk₁=k₃Þlnð1 þ k₃½Pd^IIꢀ₀tÞ

ð10Þ

Eqn (10) corresponds to a logarithmic function that can

the Pd(0) complex. However, in those cases in which de- again be modeled in Origin.

activation is limited, it can be approximated that benzoqui-

Fitting by integration of the 4-Me-styrene experiments run

none concentration is essentially constant during the reaction. under 5 bar CO. The most extensive series of experiments

Under these conditions paths a and b can be, respectively, both with styrene and 4-methylstyrene were run under 5 bar

approximately described by a first and a second kinetics with CO and at [BQ]/[Pd] = 5. Since deactivation was clearly lower

respect to palladium(^II) concentration (Scheme 4). Note that with the latter substrate, we will start analyzing the corres-

the exact identity of the inactive species formed during the ponding set of data. The plots of CO absorption vs. time for

decomposition is not known and diﬀers between the two path- selected reactions are reported in Fig. 3–5, while the corres-

ways. However, this is not relevant at all to the kinetic of the ponding plots for the complete set of experiments presently

copolymerization reaction as experimentally measured.

The first order decomposition is described by eqn (3)

and (4):

discussed are reported in the ESI (Fig. S1†).

Even from a quick examination of Fig. 6 it is clear that eqn

(10) fits very well the data in the whole time region, whereas

eqn (6) tends to underestimate catalyst decomposition (which

is proportional to the second derivative of the CO absorption

plot; see later) at the beginning of the reaction and to overesti-

mate it later. The initial underestimation is especially signifi-

cant. In fact, if the assumption that the benzoquinone

concentration is approximately constant was wrong, then cata-

lyst decomposition should be slower than that calculated

during the first part of the reaction, when benzoquinone is

more abundant. This would result in the model giving an even

worse fitting of the data and cannot justify the observed discre-

pancies. Thus it is clear that the process associated with path

a must be discarded as a relevant deactivation mode for palla-

dium, whereas the goodness of the fit obtained by modeling

path b strongly supports its operation. It is worth noting that

such a good fit was obtained with only two floating para-

meters, which would be unlikely to occur if the physical basis

of the model were wrong.

d½Pd^IIꢀ=dt ¼ ꢁk₂½Pd^IIꢀ

ð3Þ

Eqn (10) could be used to fit the data of five of the reaction

profiles with an excellent approximation (adj. R²> 0.999), but

Scheme 4

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14589

View Article Online

Paper

Dalton Transactions

Fig. 6 Data for the copolymerization of 4-Me-styrene under 5 bar CO ([BQ]/[Pd] = 5) with ligand 2, 3,5-(CF₃)₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN. (a) Fitting by eqn (6)

(ﬁrst order decomposition). (b) Fitting by eqn (10) (second order decomposition). Fitting functions are shown as dashed lines (---).

Table 2

k₁and k₃values calculated from eqn (10) and k₁’ and k₃’ values calculated as detailed in the ESI for reactions run with 4-Me-styrene under 5 bar CO ([BQ]/

[Pd] = 5)

Ligand

Σσ/2

k₁/h⁻¹

k′ /h⁻¹

k₃/h⁻¹M⁻¹

k′ /h⁻¹M⁻¹

1

3

[3,5-(CH₃)₂C₆H₃]₂-BIAN (4)

[3,4,5-(CH₃O)₃C₆H₂]₂-BIAN (6)

Ph₂-BIAN

3,5-F₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN (3)

3,5-(CH₃)₂C₆H₃;3,5-(CF₃)₂C₆H₃-BIAN (1)

3,5-(CF₃)₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN (2)

[3,5-F₂C₆H₃]₂-BIAN (7)

−0.14

−0.03

0

0.325

0.36

0.415

0.68

0.86

46

45

182

144

151

123

122

188

553

1.27 × 10²

1.33 × 10²

7.85 × 10²

2.05 × 10²

2.85 × 10²

1.40 × 10³

1.12 × 10³

2.51 × 10³

3.35 × 10⁴

149

173

125

105

2.79 × 10²

5.48 × 10²

6.09 × 10²

1.36 × 10³

[3,5-(CF₃)₂C₆H₃]₂-BIAN (5)

gave significant deviations in the remaining three cases fitting procedure described in the previous paragraph,

(ligands [3,5-(CF₃)₂C₆H₃]₂-BIAN 5, [3,4,5-(CH₃O)₃-C₆H₂]₂-BIAN 6, although k₁and k′ should ideally be the same. A comparison

1

[3,5-F₂C₆H₃]₂-BIAN 7), where catalyst decomposition is more between the k₁and k′ values in Table 2 for corresponding reac-

1

evident and the assumption that benzoquinone concentration tions shows that the agreement is excellent in three out of five

does not vary is clearly untenable. The values of k₁and k₃best cases ([3,5-(CH₃)₂C₆H₃]₂-BIAN, 4: k₁= 46, k′ = 45; Ph₂-BIAN:

1

fitting the data for these five cases are reported in Table 2.

k₁= 149, k′ = 144; 3,5-(CH₃)₂C₆H₃;3,5-(CF₃)₂C₆H₃-BIAN, 1:

1

A discussion of the k₁and k₃trends specific for this set of k₁= 125, k′ = 123), but it is acceptable even for the other two,

1

data can be found in the ESI.† However, we have developed a thus validating the mathematical approach employed.

mathematical approach to obtain the desired data even for the

Plots of the first and second derivatives of CO absorption of

three reactions that did not give a good fitting with the pro- representative reactions are shown in Fig. 7 and 8. Plots for all

cedure described above and only the results of the full set of the reactions are presented in the ESI (Fig. S6 and S7†).

experiments will be described here.

The new approach, described in detail in the ESI,† is based the more active systems deactivate more quickly and also con-

on the following considerations: firms that electron-withdrawing substituents give more active

Even a qualitative inspection of derivative plots reveals that

(1) Since the CO pressure is kept constant and the olefin systems. The activity and deactivation rates of the catalyst with

concentration is high enough that it can be considered con- [3,5-(CF₃)₂C₆H₃]₂-BIAN, 5, as ligand, are striking. Although on

stant during the reaction (pseudo zero order kinetics with a long timescale, 5b gave worse conversion (see Fig. 3), from

respect to both CO and olefin), the reaction rate is pro- Fig. 7 it appears that at short reaction times it aﬀords by far

portional only to the active palladium concentration. The reac- the more active catalytic system, although the corresponding

tion rate as

diﬀerentiating the CO absorption curve.

a

function of time can be obtained by deactivation rate is also extremely high (Fig. 8).

In previous papers some of us have shown that the logar-

(2) Given what was said above, the opposite of the derivative ithm of the relative coordinating strength of substituted

of the active catalyst concentration with respect to time, corres- Ar-BIAN ligands shows a good correlation with the Hammett σ

ponding to the second derivative of the CO absorption plot, is constants.^20,38When two substituents were present, the corre-

the rate of decomposition of the catalytic system.

lation was maintained if the sum of the two individual σ was

Developing these concepts, a series of parallel values for employed, whereas ligands having diﬀerent substituents on

the k₁and k₃values could be determined and are also pres- the two aryl rings followed the same trend if the arithmetic

ented in Table 2. We named the values calculated in this way average of the Hammett σ constants for the two rings was

k′ and k′ to distinguish them from those calculated by the employed in the correlation.¹⁶In general, this corresponds to

1

3

14590 | Dalton Trans., 2013, 42, 14583–14602

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

Fig. 9 Plot of ln k’ vs. Σσ/2 for the copolymerization reactions of 4-Me-styrene

1

with CO (5 bar).

even for the reactions of unsubstituted styrene to be discussed

in the following and [3,4,5-(CH₃O)₃C₆H₂]₂-BIAN is that which

deviates more from the correlation even with the latter sub-

strate. It should be considered that the methoxy group has

strong and opposing inductive and resonance eﬀects and the

Σσ value appears to aﬀord a worse quantification of the elec-

tronic eﬀects of the trimethoxyphenyl group than for other

substituents. It should also be recalled that the catalytic reac-

tions were run in trifluoroethanol, a solvent that gives strong

hydrogen bonds. The methoxy oxygen is a likely donor for

such hydrogen bonds and their formation would result in a

decrease of the electron-donating properties of the trimethoxy-

phenyl group. As a matter of fact, the data for trimethoxy-sub-

stituted ligands statistically deviate from the general plot in

the direction expected for a more electron-poor ligand.

Fig. 7 1^stderivative plot for CO/4-Me-styrene copolymerization. Catalyst pre-

cursor: [Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆] (1b, 4b, 5b). Reaction conditions: see

Fig. 3.

The quantitative agreement between k′ and k₃values is

3

worse than in the case of k₁, but this was expected because

now even the approximations involved in the determination of

the second derivative are involved. However, the trend is the

same for both series of values and a logarithmic plot of k′ vs.

3

Σσ/2 is shown in Fig. 10 (the corresponding plot of k₃vs. Σσ/2

is presented in the ESI†).

From the plots (Fig. 10 and S4†) it is evident that the pres-

ence of electron-withdrawing substituents leads to catalytic

Fig. 8 2^ndderivative plot for CO/4-Me-styrene copolymerization. Catalyst

precursor: [Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆] (1b, 4b, 5b). Reaction conditions:

see Fig. 3.

systems that deactivate at a faster rate (higher k₃and k′ values)

3

and an approximately linear correlation with Σσ/2 exists. Again

the methoxy-substituted ligands are those that statistically

employing for both symmetrical and nonsymmetrical ligands deviate more from the correlation.

the value of Σσ/2 (where the sum is extended to all substituents

on the two rings), which is also shown in Table 2.

Fitting of the data of the styrene experiments run under

5 bar CO. A series of experiments paralleling those just

A logarithmic plot of k′ with respect to Σσ/2 is shown in described, but employing unsubstituted styrene as the sub-

1

Fig. 9.

strate was also performed. The plots of some of the reactions

From the plot it appears that as a general rule electron-with- were already presented in Fig. 3–5. The plots of all reactions

drawing groups lead to more active systems, although a clean are reported in the ESI (Fig. S10†). The experiment using 3,5-

linearity of ln k′ vs. Σσ/2 is not observed.

(CF₃)₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN, 2, as ligand gave some

1

Ligand 6 with six methoxy groups is that which deviates problems and has been discarded.

more strongly from the value expected based on its Σσ/2 and

The data were subjected to the same mathematical treat-

steric hindrance. In general, ligands having an aryl ring with ment as that described for the 4-Me-styrene reactions. Given

three methoxy groups are those that fit less the correlation the negative results previously obtained with the first order

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14591

View Article Online

Paper

Dalton Transactions

Fig. 11 Plot of ln k’ vs. Σσ/2 for the copolymerization reactions of styrene with

CO (5 bar). The straight line corresponds to the best ﬁt to the data points,

1

Fig. 10 Plot of ln k’ vs. Σσ/2 for the copolymerization reactions of 4-Me-

styrene with CO (5 bar).

3

excluding those for [3,4,5-(CH₃O)₃C₆H₂]₂-BIAN, 6, and [3,5-F₂C₆H₃]₂-BIAN, 7.

decomposition model, only the fit to the second order one was

performed. The same three ligands that had failed to give data

series that could be fitted by eqn (10) in the case of 4-Me-

styrene also gave a poor fit in the present case, which leaves

only four data points for which very good results were

obtained. The calculated values of k₁and k₃are reported

in Table 3 and are graphically represented in Fig. S11 and

S12 (ESI†).

To get information even from the remaining three reac-

tions, the same interpolation/diﬀerentiation procedure as that

described for 4-Me-styrene was then applied. The corres-

ponding derivative plots are shown in Fig. S13 and S14 (ESI†).

From these plots, the k′ and k′ values can be calculated as

1

3

described for the 4-Me-styrene reactions. Results are also pres-

Fig. 12 Plot of ln k’ vs. Σσ/2 for the copolymerization reactions of styrene with

CO (5 bar). The straight line corresponds to the best ﬁt to all data points.

3

ented in Table 3 and logarithmic plots of k′ and k′ vs. Σσ/2

1

3

are shown in Fig. 11 and 12, respectively. As for the case of

4-Me-styrene, only the most complete data series is discussed

here, but a complete discussion of the results of the two data

series can be found in the ESI.†

In general the data follow the same trend as that evidenced

for the 4-Me-styrene experiments, that is, electron-withdrawing

substituents lead to more active but less stable, catalytic

systems.

As in the case of 4-Me-styrene, there is good agreement

between corresponding k₁and k′ values and worse agreement

1

between k₃and k′ values, but the general trends are main-

A more interesting comparison can be made between the

corresponding k₁(or k′ ) and k₃(or k′ ) values for the two sub-

3

tained. The k′ value for [3,5-F₂C₆H₃]₂-BIAN, 7, appears to be

1

3

abnormally low, which may be due to extensive decomposition

before the measurement of the consumed CO was started, as

better discussed in the ESI.†

strates. In a previous paper in which some of the same ligands

employed here had been tested, it had been noted that a

higher productivity was obtained with unsubstituted styrene,¹⁷

Table 3 k₁and k₃values calculated from eqn (10) and k₁’ and k₃’ values calculated as detailed in the ESI for reactions run with styrene under 5 bar CO ([BQ]/

[Pd] = 5)

Ligand

Σσ/2

k₁/h⁻¹

k′ /h⁻¹

k₃/h⁻¹M⁻¹

k′ /h⁻¹M⁻¹

1

3

[3,5-(CH₃)₂C₆H₃]₂-BIAN (4)

[3,4,5-(CH₃O)₃C₆H₂]₂-BIAN (6)

Ph₂-BIAN

3,5-F₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN (3)

3,5-(CH₃)₂C₆H₃;3,5-(CF₃)₂C₆H₃-BIAN (1)

[3,5-F₂C₆H₃]₂-BIAN (7)

−0.14

−0.03

0

0.325

0.36

0.68

0.86

208

234

503

213

157

358

85

4.94 × 10²

2.08 × 10²

8.88 × 10²

2.16 × 10³

1.01 × 10³

6.09 × 10²

4.17 × 10³

1.67 × 10⁴

231

153

425

1.13 × 10³

1.11 × 10³

3.05 × 10³

[3,5-(CF₃)₂C₆H₃]₂-BIAN (5)

422

14592 | Dalton Trans., 2013, 42, 14583–14602

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

Table 4 k₁, k’₁, k₃and k₃’ values for reactions run at diﬀerent CO pressures with diﬀerent ligands (styrene as the substrate, [BQ]/[Pd] = 5)

Ligand

3 bar

5 bar

10 bar

3,5-F₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN (3)

k₁

77

73

153

157

k′

125

1

k₃

3.12 × 10²

2.16 × 10²

124

1.11 × 10³

1.01 × 10³

425

k′

1.03 × 10³

3

3,5-(CH₃)₂C₆H₃;3,5-(CF₃)₂C₆H₃-BIAN (1)

k₁

k′

k₃

k′

3

125

358

1

1.07 × 10³

3.66 × 10²

3.05 × 10³

6.09 × 10²

contrary to what was generally reported in the literature.^39–43

The new, more extensive, series of data supports this obser-

vation, but also shows that the higher activity is also associated

with a faster deactivation, since not only k₁, but also k₃values

are consistently higher in the case of styrene. To the best of

our knowledge, an active role of the olefin in the deactivation

process has not been proposed before. We will discuss this

aspect in more detail later, after having reported the rest of the

results.

The faster deactivation is responsible for the less good

agreement between k₁and k′ and k₃and k′ values, since all

1

3

approximations become larger.

Fig. 13 CO/styrene copolymerization: eﬀect of reaction time and a higher

[BQ]/[Pd] ratio. Catalyst precursor: [Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆]. Reaction

conditions: n_Pd= 0.106 × 10⁻⁵mol, T = 30 °C, TFE V = 5 mL, P_CO= 10 bar, t =

70 h, [BQ]/[Pd] = 40, styrene V = 3 mL, [styrene]/[Pd] = 25 000.

Eﬀects of CO pressure. To investigate the eﬀect of the

carbon monoxide pressure, selected reactions were performed

under either 3 or 10 bar CO. The fitted values are presented in

Table 4, together with the corresponding k′ and k′ values, cal-

1

3

culated as in the other cases. Note that the values for the reac-

tions at 5 bar, already reported in Table 3, are shown again contrary to what was previously observed under the same con-

here for an immediate comparison.

ditions, but with [BQ]/[Pd] = 5. The rate decrease that is

The reactions under 5 bar are faster than those under 3 bar, observed in all cases after about 3.2 mmol CO have been

but a further increase in pressure leads to a decrease in absorbed is due to the polyketone precipitation in the reaction

activity.⁴⁴The additional information we get from the numeri- mixture with the switch of the catalytic system from homo-

cal simulations is that the decrease in productivity on going to geneous to a mixture of homogeneous and heterogeneous

10 bar is not due to a faster catalyst deactivation in the last species, each of them displaying its own kinetic behavior. This

case. Indeed even k₁or k′ decreases at 10 bar and moreover phenomenon was previously pointed out in both the CO/ethy-

1

the values of k₃and k′ increase on going from 3 to 5 bar, but lene⁴⁷and CO/vinyl arene copolymerizations.⁴³Thus, this

3

are of the same order of magnitude for reactions run at 5 or decrease cannot be described by our models and we have

10 bar. This indicates that the rate of deactivation is higher under limited the modeling to the first 24 h of reaction. Due to the

5 rather than 3 bar of CO, but no faster deactivation takes increased catalyst stability, good fitting by the second order de-

place at higher CO pressures. Thus, the decrease in activity in activation mode was obtained in all cases, but in order to have

the latter case is due to the inhibiting role of CO. This is a a direct comparison with data with the lower benzoquinone

phenomenon well known for this reaction and the value of CO amount, k′ and k′ values were also calculated. All constant

1

3

pressure at which this eﬀect becomes evident depends on the values are presented in Table 5. Good agreement exists

catalyst nature, if dicationic⁴⁵or monocationic²¹precatalysts between the two calculation approaches, except for the k′

3

are applied, and on the nature of the ancillary ligand bonded value for 3,5-(CH₃)₂C₆H₃;3,5-(CF₃)₂C₆H₃-BIAN, which appears

to palladium.^21,46

to be overestimated.

Eﬀect of a longer reaction time and a larger amount of

Only one direct comparison can be made between reaction

benzoquinone. The eﬀect of a longer reaction time was investi- runs working at [BQ]/[Pd] ratios of 5 and 40 respectively, but

gated by running selected reactions for 70 h instead of 25 h, otherwise under exactly the same conditions (styrene, 10 bar

under 10 bar CO. Since it was expected that deactivation would CO) as that involving 3,5-F₂C₆H₃;3,4,5-(CH₃O)₃C₆H₂-BIAN. The

occur, a higher benzoquinone amount ([BQ]/[Pd] = 40) was k′ value marginally increased (from 125 to 188; values in

1

also added (Fig. 13).

Tables 4 and 5, respectively) upon an increase in the benzoqui-

From Fig. 13 it is evident that the addition of a larger none amount, which may be due to the approximations in

amount of benzoquinone strongly stabilizes the catalytic neglecting the deactivation before the gas consumption starts

system. Deactivation is very low during the first 24 hours, to be measured. However, it is k′ that is most aﬀected,

3

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14593

View Article Online

Paper

Dalton Transactions

result in either positive or negative eﬀects depending on the

Table 5 Values of k₁, k₃, k₁’ and k₃’ for reactions run with styrene, under 10 bar

CO and with [BQ]/[Pd] = 40

position of the group with respect to the metal center. For

o

instance, Ni(^II) complexes with symmetrical Ar-BIANs substi-

Ligand

Σσ/2 k₁/h⁻¹k′ /h⁻¹k₃/h⁻¹M⁻¹k′ /h⁻¹M⁻¹

1

3

tuted on one ortho position only with a CF₃group catalyzed

ethylene polymerization, leading to oligomers instead of poly-

mers for steric reasons, but with higher TOF than the catalyst

with the related CH₃substituted ligand for electronic eﬀects.⁵⁰

Instead, when the CF₃group is in a proper remote position of

the ligand, as is the case for nickel-phenoxyiminato catalysts,

an enhancement of catalyst activity and thermal stability and

of molecular weight of the produced polyethylene was

observed. These eﬀects were attributed to C–H⋯F–C inter-

actions between the hydrogen atom on the β-C atom of the

growing polymer chain and the fluorine atom of the ligand,

rather than to inductive eﬀects.⁵¹In the present investigation

the introduction of CF₃groups to meta positions of the aryl

rings has a detrimental eﬀect on catalyst stability, despite the

positive eﬀect on catalyst activity.

This kinetic investigation highlights that among the cata-

lysts with the nonsymmetrical ligands, that with ligand 3 is the

most active and the slowest to decompose for CO/4-Me-styrene.

The same is valid for CO/styrene but in this case the catalyst

with ligand 1 is more active (Tables 2 and 3). By analyzing all

the data as a whole, it is clear that the catalyst with ligand 3

represents the best compromise between the activity and rate

of deactivation. By supposing that catalyst deactivation pro-

ceeds through ligand dissociation, the observed slow de-

activation is in agreement with the higher stability of the

active species thanks to the shorter Pd–N bond distances

observed in the solid state in the neutral palladium derivative.

Further details of the catalyst decomposition pathway. The

faster catalyst deactivation observed with ligands bearing elec-

tron-withdrawing groups may have diﬀerent origins, including

the formation of complexes bearing two Ar-BIAN ligands and

the comproportionation reaction of monomeric palladium(^II)

and palladium(0) complexes also having a bound ligand.

However, the fact that all data point to a ligand dissociation

step during the early (kinetically relevant) stages indicates that

a deactivation process of the kind described should be a

minor pathway during catalyst deactivation and the ligand

must be lost at some stage, even if it may be regained at a later

stage.

Ph₂-BIAN

0

125

124

188

107

157

92

170

3,5-F₂C₆H₃;3,4,5-

(CH₃O)₃C₆H₂-BIAN (3)

3,5-(CH₃)₂C₆H₃;3,5-

(CF₃)₂C₆H₃-BIAN (1)

0.325 177

0.36 159

167

265

438

decreasing from 1.03 × 10³to 170. This variation is in agree-

ment with the literature about the role of the oxidant that

aﬀects catalyst stability but not catalytic rate.^48,49

General trends in the activity vs. time plots. The fitting pro-

cedure allows us to estimate the values of the rate constants

for both the propagation step of the polymer chain and the

decomposition pathway of the catalyst for the two vinyl arenes

and to correlate them to the nature of the Ar-BIAN bonded to

palladium (Tables 2 and 3). The CO uptake curves with time

highlight that all the studied Ar-BIAN catalysts were found to

be more active in CO/styrene than in CO/4-Me-styrene copoly-

merization. This trend is in agreement with the productivity

data obtained in the batch copolymerization reactions,¹⁷but is

the opposite of that found in the case of the phenanthroline-

containing catalysts.^21,43The kinetic analysis shows that the

higher activity of the present catalysts in the CO/styrene copoly-

merization is also associated with a faster deactivation, since

not only k₁, but also k₃values are consistently higher than in

the case of 4-Me-styrene (Tables 2 and 3). The values of the

kinetic constants highlight that in the series of catalysts with

the symmetric Ar-BIANs, those with electron-donating substi-

tuents show the lowest propagation and decomposition rate,

while the ligands symmetrically substituted with CF₃or F yield

the most active catalysts, but also show the highest de-

activation rate, in particular in the CO/4-Me-styrene copolymeri-

zation. These trends are in agreement with simple

considerations about the coordination chemistry of the two

comonomers: electron-donating substituents increase the elec-

tron density on palladium, disfavoring the nucleophilic attack

of the alkene and increasing the π-back donation to CO, result-

ing in a stronger bond to the metal center. This in turn

enhances its competing role, the combination of these eﬀects

being a slow reaction rate. On the other hand ligands with

electron-donating substituents are more strongly bonded to

palladium and, since the decomposition pathway requires

ligand dissociation, the catalysts are more stable. In contrast,

electron-withdrawing substituents decrease the electron

density on the metal center favoring the attack of the alkene

and decreasing the π-back donation to CO. The combination

of these eﬀects is a higher reaction rate. On the other hand

ligands with electron-withdrawing substituents are more

weakly bonded to palladium resulting in less stable catalysts.

These considerations are correct from a general point of view;

nevertheless, in peculiar cases the explanation is more

complex. Indeed, the introduction of CF₃groups to the ancil-

lary ligands of catalysts for polymerization was reported to

In addition, one has to take into account the positive role of

benzoquinone in retarding catalyst decomposition. Benzoqui-

none is known to act as an oxidant towards palladium(0) com-

plexes and the mechanism of its action has been

investigated.⁵²In addition, complexes of the kind [(N–N)Pd

(BQ)] (N–N = 2,2′-bipyridine, Ar-BIAN) are known and can be

intermediates in the palladium(0) reoxidation.^53–55Oxidation

even of palladium(^I) dimers by benzoquinone has been con-

sidered to be a possibility,⁵⁶but, to the best of our knowledge,

such a process has never been experimentally observed. The

fact that catalyst deactivation follows a second order kinetics

with respect to palladium suggests that such a dimer oxidation

either does not occur at all or is sensibly slower than the oxi-

dation of palladium(0) and only occurs in a few cases.

14594 | Dalton Trans., 2013, 42, 14583–14602

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

Scheme 5 Alternative pathways for ligand loss.

Otherwise a higher order in palladium should be observed for (COD-OCH₃)]⁺(COD-OCH₃is the ligand resulting from the

the decomposition. Thus we will not consider palladium(^I) attack of a methoxide anion onto coordinated 1,5-cycloocta-

reoxidation as a possibility in the following.

diene), featuring a cationic palladium(^II) complex similar to

As far as ligand dissociation is concerned, we have two some complexes that are formed as intermediates in the

main alternatives. In the first (path a in Scheme 5), the ligand present catalytic polymerization, has ρ_BIAN= −3.47. The slope

is lost from a palladium(^II) compound, to yield a less stabilized of the log k₃or k′ vs. Hammett σ constant plots shown in the

3

and more easily reducible species. The ligand (L in the previous figures (once corrected for the fact that decimal logar-

Scheme) substituting Ar-BIAN may be a solvent molecule, CO ithms are used in the calculation of ρ_BIANinstead of natural

or the olefin. P is the growing polymer chain and may also play ones) shows some scatter (1.50, 1.89, 1.08, 1.31, 0.86, 1.47,

the role of a ligand.

respectively, for the plots in Fig. S4, 10 (= S9), S12, 12 (= S16),

Alternatively, reduction can precede ligand substitution, as S21 and S23†), as expected given the small number of data

shown in path b. If this is the case, it may be advanced that points in some cases, but the average value is 1.35 (1.15 if only

ligand loss would generate a palladium(0) species that would the more precise k₃values are considered). The value is posi-

aggregate at a rate much faster than that with which it can be tive, contrary to those previously collected, but it should be

reoxidized by benzoquinone. Indeed it should be realized that remembered that now we are looking at a ligand dissociation

the Ar-BIAN ligand renders the complex more electron-rich, at process, whereas the previous values referred to a ligand

the same time occupying two coordination sites. The first coordination process, so that the sign is reversed. If the abso-

feature should favor oxidation, whereas the second should lute values are compared, it clearly emerges that the present

inhibit aggregation with respect to a complex lacking such a process is associated with the ligand loss from very electron-

ligand.

rich complexes. This clearly excludes any palladium(^II)

Such a kind of alternative is present in most, if not all cata- complex, especially if positively charged, and indicates that the

lytic cycles in which palladium is involved, but we are not Ar-BIAN detachment is occurring from a complex being more

aware of any case in which experimental data have been pro- electron-rich than [Pd(Ar-BIAN)(DMFU)]. Given the compounds

vided that allow discriminating among the two. Both pathways available in solution, by far the only likely candidates for this

are consistent with the kinetic data of any single reaction and type of compound are [Pd(Ar-BIAN)(olefin)] (olefin = styrene or

cannot be distinguished looking at them independently.

4-methylstyrene) or [Pd(Ar-BIAN)(CO)_n] (n = 1, 2). Palladium(0)

At this stage we can analyze the trends in k₃or k′ vs. the complexes with nitrogen ligands and olefins lacking electron-

3

Hammett σ constant on a more quantitative basis. We have withdrawing substituents have not been reported in the litera-

already mentioned that the coordination strength of Ar-BIAN ture. Although, to the best of our knowledge, ligand substi-

ligands is correlated with the σ value of the substituents.²⁰tution by an olefin has never been proposed as a relevant step

More precisely, a linear correlation is observed between the in palladium catalysts deactivation, limited precedents exist

logarithm of the equilibrium constant for the substitution of for it. Vrieze and co-workers have reported that substituted di-

Ph₂-BIAN by a generic [RC₆H₄]₂-BIAN and the σ value of azabutadienes (R-DAB) could be substituted by an olefin in

R. When two substituents are present, the correlation is main- [Pd⁰(R-DAB)(olefin)] complexes⁵⁷and theoretical calculations

tained if the two σ values are summed up. Nonsymmetrical Ar, by Elsevier and co-workers indicate that [Pd(R-DAB)(ethylene)]

Ar′-BIAN fitted the common plot if the average of the σ values is thermodynamically unstable with respect to [Pd(ethylene)₃]

for the two diﬀerent rings was employed in the correlation.¹⁶and R-DAB.⁵⁸

What is most important to the present discussion is that the

It should be noted that a substitution of an Ar-BIAN ligand

slope of these linear plots (ρ_BIAN) is sensitive to the Lewis bound to a palladium(0) complex by an olefin as a key step in

acidity of the metal fragment³⁸and is generally lower for com- catalyst deactivation is also consistent with the observed

plexes in which the palladium atom is zerovalent with respect higher instability of the catalytic system when styrene rather

to those in which it has the formal oxidation state two. The than 4-methylstyrene is employed as the substrate. In fact it

lowest sensitivity to electronic eﬀects that could be measured has long been known that electron-poor olefins coordinate

(ρ_BIAN= −1.57) corresponds to the series [Pd(Ar-BIAN)(DMFU)] better to palladium(0) complexes than those lacking this feature

(DMFU = dimethyl fumarate). By comparison the series [Pd(Ar- and some quantitative data on this trend have been recently

BIAN)(CH₃)(Cl)] has ρ_BIAN= −2.71 and the series [Pd(Ar-BIAN)- gained.⁵⁹Note that exactly the opposite would be expected if

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14595

View Article Online

Paper

Dalton Transactions

Ar-BIAN loss were occurring from a palladium(II) intermediate, whereas 1a and 1d complexes are a mixture of trans and cis

thus further supporting the conclusions reached. isomers in the ratio 10 : 1 and 3 : 1, respectively, in the NMR

Palladium(0) carbonyl complexes with nitrogen ligands are spectrum of 1c only the signals of the isomer characterized by the

also unknown and only metallic palladium is invariably acetyl group trans to the CF₃-substituted aryl ring are observed.

obtained when palladium(^II) complexes with nitrogen ligands

Starting from complex 1d the intermediates resulting from

are reduced in the presence of CO. Loss of Ar-BIAN from a the migratory insertion reactions of the two comonomers were

[Pd(Ar-BIAN)(CO)_n] complex is consistent with the higher k₃and investigated. When 1d is treated with one equivalent of 4-Me-

k′ measured at 5 bar CO with respect to those at 3 bar and styrene in CH₂Cl₂at −20 °C the five-member palladacycle 1e is

3

may also be consistent with the ρ_BIANvalue observed, although obtained as a dark red solid. Analogous palladacycles were

a direct comparison with a reference compound is not avail- formed from the reaction of 4-Me-styrene with palladium com-

able. Our data cannot distinguish between loss of the ligand plexes having ligands belonging to the family of phenanthro-

from a palladium–olefin or a palladium–CO complex. The lines,⁶⁰bisoxazolines⁴²and diisopropyl diazabutadiene.⁶¹

o

diﬀerent deactivation rates observed in the case of the reac- Instead, in the case of complexes with Ar-BIANs the reaction

tions run in styrene with respect to those run in 4-methyl- with 4-Me-styrene did not lead to the palladacycle, but to an

styrene suggest that the former kind of complex is at least in part intermediate involving an η³-allyl bond between Pd and the

involved, although deactivation originating from the carbonyl inserted styrene unit.⁶²This result clearly correlates the

complex is likely to also occur as a competitive pathway.

diﬀerent catalytic behavior of complexes with ortho and meta-

substituted Ar-BIANs to the nature of the intermediates

formed after the vinyl arene migratory insertion: when the

migratory insertion leads exclusively to the η³-allyl intermedi-

ate, low or negligible catalytic activity is observed, while when

it yields the palladacycle the copolymerization reaction pro-

ceeds much faster. Analogous observations were proposed to

be at the origin of the inactivity of palladium complexes with

diphosphine ligands (P–P) in the CO/vinyl arene copolymeriza-

tion, in contrast to the activity of the related complexes with

the phosphino–phosphite ligand BINAPHOS, leading to the

metallacycle intermediate.^63,64In the latter case it was demon-

strated that the CO insertion into the Pd–allyl bond is much

less eﬃcient than into the Pd–alkyl bond.⁶⁵

Reactivity studies

Catalytic and kinetic investigations show that Pd complexes

containing nonsymmetrical ligands generate better perform-

ing catalysts than those with symmetrical Ar-BIANs. To get

some insights into the nature of this phenomenon we investi-

gated by means of NMR spectroscopy the reactivity of complex

1d, having the nonsymmetrical ligand 1, with both comono-

mers, CO and 4-Me-styrene (Scheme 6). The Pd-carbonyl-

methyl complex 1d represents the active species and it was

obtained in high yield as a red solid through reaction of the

corresponding Pd-acyl-chloride species 1c with a stoichio-

−

metric amount of NaBArF. In complex 1d, in place of the PF₆

The observation in the NMR-2D-NOESY spectrum of 1e of a

strong cross-peak between H_aof the styrene unit and H_orthoof

the CH₃-substituted aryl ring, together with the lack of such

interactions with the H_orthoof the CF₃-substituted aryl ring,

probes evidence that only the isomer characterized by the Pd–C

counterion, the less coordinating tetraarylborate is present to

favor alkene insertion in the stoichiometric reactions. In turn,

1c was isolated as a dark green solid by dissolving 1a in

dichloromethane previously saturated with carbon monoxide

at −20 °C (Scheme 6). It is straightforward to note that,

Scheme 6 Synthesis of the active species 1d and reaction pathway for the ﬁrst steps of the copolymerization reaction. While 1f and 1g have been detected, the

intermediate 1e has been isolated.

14596 | Dalton Trans., 2013, 42, 14583–14602

This journal is © The Royal Society of Chemistry 2013

View Article Online

Dalton Transactions

Paper

bond trans to the Pd–N bond of the CF₃-substituted aryl ring was observed for both palladium complexes with amine–

is formed. This compound is the result of two migratory imine ligands,²⁷with hybrid P–N ligands⁴⁸and with

insertions: the migratory insertion of CO into the Pd– the phosphino–phosphite (P–PO) molecule BINAPHOS.⁶⁷

CH₃bond followed by that of 4-Me-styrene into the Pd– A mechanistic hypothesis involving cis/trans isomerization was

acyl bond which takes place with secondary regiochemistry, invoked.

exclusively.

Very recently, both in the polyethylene synthesis and in the

In the next step the reactivity of 1e with CO was studied by ethylene/methyl acrylate copolymerization catalyzed by Pd-

bubbling the gaseous reagent into the NMR tube containing a complexes with phosphino–sulfonate (P–O) ligands the cis/

CD₂Cl₂solution of 1e at −30 °C. The complete formation of trans isomerization was indicated as an important step of the

the Pd–acyl–carbonyl intermediate, 1f, was observed within catalytic cycle.^68,69In all the mentioned examples the preferen-

1

10 min. Some relevant features of the H NMR spectrum of 1f tial polymer chain coordination site is determined by an elec-

are: (i) only the signals of one species are detectable; (ii) the tronic diﬀerentiation of the two donor atoms, i.e. hard and

signal of H_ais shifted from 3.51 in 1e to 4.61 ppm in 1f due to soft for the P–O and N_sp3–N_sp2ligands, or trans influence in

the deshielding eﬀect of the adjacent CO group; (iii) in the the hybrid P–N catalysts. For the catalysts with the bis(imine)

2D-NOESY spectrum a cross-peak is present between H_aand Ar-BIANs, the topic of the present investigation, the diﬀerence

H^14′,18′; (iv) the resonances of protons H_b,b′of the CH–CH₂frag- in electronic features of the two N-donor atoms should be less

ment do not exhibit a diamagnetic shift of the type that was pronounced than that in the other two series. Nevertheless,

observed in the case of the complexes with ^oAr-BIANs, it appears to be suﬃcient for driving the preferential co-

suggesting that the CH₂group in the present case is not inside ordination of the polymer chain.

the shielding cone of the aromatic ring of the ligand; (v) two

signals are observed for the methyl groups of the aryl ring,

while in complexes 1a–1e only one singlet was observed, indi-

cating that in 1f the rotation of the aryl ring is hindered. This

Conclusions

NMR analysis shows that 1f is the Pd–acyl–carbonyl species Palladium complexes with α-diimine ligands having an ace-

with the growing polymer chain oriented towards the CH₃- naphthene skeleton and aryl rings on the imine nitrogen

substituted aryl ring (Scheme 6).

atoms have been widely applied as homogeneous catalysts for

The solution of 1f was treated with 1 equivalent of 4-Me- (co)polymerization reactions. In the past we demonstrated that

styrene under a CO atmosphere. After 4 days at −18 °C no the best performing catalysts within this ligand family, both

signals of 1f and free vinyl arene are present, while signals of in the CO/vinyl arene and ethylene/methyl acrylate co-

the new species 1g are observed. The complexity of the ¹H polymerization, are characterized by nonsymmetrically substituted

NMR spectrum of the double olefin insertion product 1g Ar,Ar′-BIANs. The library of the Ar,Ar′-BIANs has now been

suggests the presence of two diastereoisomeric complexes with extended to a new molecule characterized by a meta-F-substituted

an RR and RS (or SS and SR) configuration of the CH–Ph aryl ring and a meta,para-methoxy-substituted ring.

stereogenic centers, corresponding to like and unlike dyads in

The kinetic study together with its modelization reported in

the polymeric chain. This result is in agreement with the this paper allows us to quantitatively distinguish between cata-

atactic stereochemistry of the synthesized polyketones and lyst activity and its decomposition rate and to mainly correlate

indicates that no enantioface discrimination takes place since them to the electronic properties of the substituents on the

the insertion of the second molecule of vinyl arene. Finally, in aryl rings. Catalysts having symmetrical Ar-BIANs substituted

the spectrum of 1g the region of the signals of the aromatic with electron-donating groups show the lowest propagation

protons is crowded due to the signals of the two inserted and decomposition rate, whereas those having symmetrical

4-methylstyrene molecules; thus the NOE experiments are not Ar-BIANs substituted with electron-withdrawing groups yield the

clear-cut as it was for 1f in indicating that the growing polymer most active catalysts, but with the highest deactivation rate.

chain is trans to the CF₃-substituted aryl ring.

Thus, the best performing catalysts are those with the nonsym-

Thanks to this NMR investigation it was possible to detect metrical Ar,Ar′-BIANs and, among them, the active species

the intermediates formed at the beginning of the copolymeri- with the new ligand 3 represents the best compromise between

zation reaction and to demonstrate that when the ancillary activity and rate of deactivation.

ligands are nonsymmetrical molecules there is a preferential

When studying a catalytic system one usually tries to

coordination site for the growing polymer chain, which is trans employ experimental conditions that minimize catalyst

to the N atom with lowest Lewis basicity. Thus, each time this decomposition and quantitative kinetic studies are usually per-

is not the isomer resulting from the migratory insertion, an formed only under conditions in which decomposition is neg-

isomerization process takes place before the next insertion ligible. In this work we have quantitatively investigated the

reaction proceeds. For the present complexes there is a behavior of a series of catalysts that deactivate to a variable

site selective coordination of the growing polymer chain. This and sometime high extent. This allowed us to gain important

phenomenon is well known as back-skip in the catalytic information on the catalyst deactivation pathways, which are

synthesis of isotactic polypropene.⁶⁶In the case of CO/alkene probably much more general. The deactivation was shown to

copolymerization the site selective coordination of the alkene involve the following steps:

This journal is © The Royal Society of Chemistry 2013

Dalton Trans., 2013, 42, 14583–14602 | 14597

View Article Online

Paper

(1) Palladium reduction from +2 to zero oxidation

Dalton Transactions

Ligands 1, 2, and 4–6 and the related palladium complexes

state. This step can be reverted easily by reaction with were synthesized by following literature procedures.^{17,19–21,38}

benzoquinone.

Synthesis of ligands 3 and 7

(2) Chelating ligand loss, likely due to displacement by the

olefin or CO. Reoxidation after ligand dissociation appears not

Ligand 3. The procedure is adapted from ref. 38. To a

to compete eﬀectively with the next step in the deactivation.

Schlenk flask were added ZnCl₂(3,5-F₂C₆H₃-BIAN), obtained as

(3) The resulting palladium(0) complex reacts with another described below (0.500 g, 0.93 mmol), 3,4,5-trimethoxyaniline

palladium(^II) unit to generate a palladium(I) dimer that cannot (174.0 mg, 0.95 mmol), and dry methanol (60 mL). The sus-

be oxidized back under the reaction conditions and either pension was stirred under dinitrogen for 2 hours and, then it

accumulates or irreversibly evolves towards metallic palladium. was evaporated in vacuo and the obtained solid was washed

These results suggest two strategies to increase catalyst life. with a hexane–toluene 1 : 1 mixture (10 mL). The remaining

Since ligand loss is an equilibrium reaction, this should be solid was dissolved in dichloromethane (25 mL) in a separ-

shifted to the reactants side by adding an excess of free nitro- ating funnel and decomplexed by shaking with a saturated

gen ligand in solution, resulting in catalyst stabilization. In solution of potassium oxalate (10 mL). The organic layer was

addition, if ligand loss is favored by simultaneous olefin separated and washed with water (3 × 10 mL). The organic

coordination, decreasing olefin concentration may also result layer was then dried over sodium sulfate, filtered and evapor-

in a more stable catalytic system, although activity may also ated in vacuo. The product was purified by column chromato-

decrease. These strategies will be investigated in the future graphy on silica (eluent hexane–triethylamine 9.5 : 0.5). The

and will be the topic of a future study.

nonsymmetrical ligand was eluted first (220 mg, 52% yield),

followed by some [3,4,5-(CH₃O)₃C₆H₂]₂-BIAN.

Elemental analysis: C, 70.74; H, 4.40; N, 6.11. Found:

C, 70.51; H, 4.50; N, 6.09.

Experimental

Materials and methods

δ_H(400 MHz; CDCl₃, 298 K), 7.98 (d, J = 8.2 Hz, 1H, H⁵or

H⁸), 7.96 (d, J = 8.2 Hz, 1H, H⁸or H⁵), 7.47 (pst, J = 7.8 Hz, 2H,

[Pd(OAc)₂] was a gift from Engelhard Italia and was used as H⁴, H⁹), 7.14 (d, J = 7.3 Hz, 1H, H¹⁰), 6.99 (d, J = 7.2 Hz, 1H,

received. All the solvents were purchased from Sigma-Aldrich H³), 6.72 (pst, J = 9.0 Hz, 1H, H¹⁶), 6.66 (d, J = 8.0 Hz, 2H, H¹⁴,

and used without further purification for synthetic, spectro- H¹⁸), 6.38 (s, 2H, H^14′, H^18′), 3.94 (s, 3H, p-OCH₃), 3.83 (s, 6H,

scopic and catalytic purposes, with the only exception of the m-OCH₃). δ_C(75 MHz, CDCl₃, 298 K) 162.7 (C^1,12), 154.7

dichloromethane used for the synthesis of complexes, which (C^15′,17′), 147.7 (C¹³or C⁷), 131.7 (C⁶), 130.2 (C⁵or C⁸), 129.7

was purified through distillation over CaH₂under an argon (C⁵or C⁸), 128.4 (C⁴or C⁹), 128.2 (C⁴or C⁹), 124.8 (C¹⁰), 124.5

atmosphere and was used freshly distilled. Carbon monoxide (C³), 102.1 (d, J²

= 27.0 Hz, C^14,18), 99.9 (C¹⁶), 95.8 (C^14′,18′),

C–F

(grade 4.7) was supplied by Linde Gas for kinetic investigations 61.62 (p-OCH₃), 56.6 (m-OCH₃). Eight quaternary carbons were

and by Air Liquide (CP grade 99.99%) for mechanistic studies. not detected. ¹⁹F{¹H} NMR (376 MHz, CDCl₃, 298 K), δ, ppm,

NMR spectra of ligands and complexes were recorded using a −108.2 (s, F).

Jeol EX 400 and a Varian 500 spectrometer at 298 K and at 400

Ligand 7. The procedure is adapted from ref. 20. To a

and 500 MHz, respectively; NMR studies for mechanistic inves- Schlenk flask under a dinitrogen atmosphere were added

tigations were performed using a Bruker-AC200 spectrometer. acenaphthenequinone (313.0 mg, 1.72 mmol), dry ZnCl₂

The resonances were referenced to the solvent peak versus (606.5 mg, 4.45 mmol), and AcOH (4.7 mL). The suspension

TMS: CDCl₃at δ 7.26 (¹H) and δ 77.0 (¹³C), CD₂Cl₂at δ 5.32 was heated to about 50–60 °C and 3,5-F₂C₆H₃NH₂(521.6 mg,

(¹H). ¹³C NMR spectra of polyketones were recorded in 4.04 mmol) was then added. The solution was reflux heated

1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) with a small amount for 45 min and then filtered on a Büchner funnel while hot.

of CDCl₃for locking purposes at 125.68 MHz and referenced The solid was washed on the filter with diethyl ether to help

at δ 77.0. IR spectra of monocationic complexes were recorded remove the acetic acid and dried in vacuo to give ZnCl₂(7)

in Nujol using a Perkin Elmer System 2000 FT-IR.

(733.3 mg, 1.36 mmol, 79.1% yield). Care must be taken to

NMR spectra for ligands 3 and 7 were recorded using eliminate any residue of the acetic acid employed as the

Bruker Avance 300-DRX or Avance 400-DRX spectrometers. solvent in the synthesis, as this has a strong negative eﬀect on

Chemical shifts (ppm) are reported relative to TMS. The ¹H the eﬀectiveness of the separation in the following step. Note

NMR signals of the compounds described in the following also that the addition of toluene to the solution mixture is gen-

have been attributed by COSY and NOESY techniques. Assign- erally required when anilines bearing trifluoromethyl groups

ment of the resonance in ¹³C NMR was made using the APT are employed, to make precipitation of the complex easier.²⁰

pulse sequence and HSQC and HMBC techniques. CAUTION: However, with difluoroaniline this addition was not necessary.

when NMR spectra of Ar-BIAN ligands containing electron-

ZnCl₂(7) (717.1 mg, 1.33 mmol) was then placed in a flask

withdrawing substituents are recorded in CDCl₃, it is essential in the air and suspended in CH₂Cl₂(65 mL). A solution of

that the solvent is purified by passing it over a short column K₂C₂O₄(1.48 g, 8.14 mmol) in water (9 mL) was then added

of basic alumina to eliminate acidic impurities. Otherwise, and the biphasic mixture vigorously stirred for 15 min. The

partial hydrolysis of the ligand is observed.

phases were separated and the organic layer was washed with

14598 | Dalton Trans., 2013, 42, 14583–14602

View Article Online

Dalton Transactions

Paper

water (2 × 10 mL), dried with Na₂SO₄, filtered and evaporated ^mOCH_3,M,m), 2.20 (3H, s, Pd–NCCH_3,M), 2.17 (3H, s, Pd–

to dryness aﬀording the analytically pure 7 (456.7 mg, NCCH_3,m), 1.00 (3H, s, Pd–CH_3,M), 0.87 (3H, s, Pd–CH_3,m).

1.13 mmol, 85.0% yield). Elemental analysis: C, 47.05; H, 3.42; N, 5.49. Found:

Elemental analysis: C, 71.30; H, 2.99; N, 6.93. Found: C, C, 46.98; H, 3.44; N, 5.35.

71.12; H, 3.15; N, 6.84.

[Pd(CH₃)(CH₃CN)(7)][PF₆] (7b). Yield: 72%. Yellow solid.

δ_H(400 MHz; CDCl₃, 298 K) 7.99 (2H, d, J_H–H= 8.3 Hz,

δ_H(500 MHz; CD₂Cl₂, 298 K) 8.26 (2H, t, H^5,8), 7.65 (2H, dt,

H^4,9), 7.37 (1H, d, H³), 7.05–6.90 (6H, m, H^{14,14′,16,16′,18,17′}), 6.83

H^5,8), 7.50 (2H, pt, J_H–H= 7.5 Hz, H^4,9), 7.05 (2H, d, J_H–H

=

7.2 Hz, H^3,10), 6.73 (2H, tt, J_H–F= 9.0 Hz, J_H–H= 2.2 Hz, H¹⁶), 6.67 (1H, d, H¹⁰), 2.21 (3H, s, Pd–NCCH₃), 0.95 (3H, s, Pd–CH₃).

(4H, dd, J_H–F= 7.9 Hz, J_H–H= 2.2 Hz, H^14,18) ppm. δ_C(100 MHz,

Elemental analysis: C, 45.56; H, 2.55; N, 5.90. Found:

CDCl₃, 298 K) 169.9 (dd, J¹

= 247.2, J³

= 1.4 Hz, C^15,17), C, 45.44; H, 2.45; N, 5.85.

C–F

161.7 (C^1,12), 153.8 (d, J³

= 1.2 Hz, C¹³), 142.1 (C⁷), 131.4

Reactivity studies. Isolation of intermediates of the copoly-

merization reaction.

C–F

(C⁶), 129.9 (C^5,6), 127.6 (C^2,11), 123.9 (C^3,10), 101.7 (dd, J²

=

C–F

19.4, J⁴

= 7.4 Hz, C^14,18), 99.7 (d, J²

= 25.4 Hz, C¹⁶). δ_F

[Pd(COCH₃)(Cl)(1)] (1c). Yield: 92%. Dark green solid.

Compound 1a (100 mg, 0.153 mmol) was dissolved at

−20 °C under CO atmosphere in 2.5 mL of dichloromethane

previously saturated with CO. The solution was slowly warmed

to 25 °C, stirred for 1 h, and then filtered through celite to

remove traces of metallic Pd. A dark-green solid was obtained

after evaporation of the solvent under vacuum which was

washed with hexane (3 × 5 mL) and dried, yielding compound

1c (96 mg, 0.141 mmol, 92%).

C–F

(376 MHz; CDCl₃, 298 K) −108.4 (pt, J = 8.0 Hz).

Synthesis of Pd complexes

All syntheses were performed using standard vacuum-line and

Schlenk techniques under an argon atmosphere and at room

temperature, according to the published procedures.^19,21

[Pd(CH₃)Cl(Ar-BIAN)] (3a, 7a)

General synthesis. [Pd(CH₃)Cl(cod)] (201.5 mg, 0.76 mmol)

was kept in a Schlenk flask and dissolved in CH₂Cl₂(3 mL). A

solution of the ligand (0.84 mmol) in CH₂Cl₂(3 mL) was

added. The reaction mixture was protected from light, stirred

at r.t. for 1 h, and then concentrated at half volume under

reduced pressure. Upon the addition of diethyl ether a red

solid was obtained.

δ_H(200 MHz, CD₂Cl₂, 298 K) 8.17 (2H, dd, H^5,8), 7.98 (1H, s,

H¹⁶), 7.87 (2H, s, H^14,18), 7.55 (2H, dt, H^4,9), 7.14 (2H, m, H^16′

and H³), 6.93 (s, 3H, H^14′,18′and H¹⁰), 2.43 (6H, s, –(CH₃)₂–Ar),

2.24 (3H, s, CH₃–CO). δ_13C(50.33 MHz, CD₂Cl₂, 298 K) 220.26

(CO–Pd), 144.61 (s, ipso–C–Ph–CF₃), 140.00 (s, ipso–C–Ph–

CH₃), 132.17, 131.34 (s, CH^5,8), 129.40 (s, CH¹⁶), 128.97, 128.45

(s, CH^4,9), 126.08 (s, CH³), 125.21 (s, CH¹⁰), 122.33 (s, CH^14,18),

120.95 (s, CH^16′), 117.98 (s, CH^14′,18′), 32.65 (s, CH₃CO), 21.11

(s, CH₃–Ar).

[Pd(CH₃)Cl(3)] (3a). Yield: 89%. Dark red solid.

Elemental analysis: C, 54.65; H, 3.77; N, 4.55. Found: C,

54.76; H, 3.83; N, 4.75.

[Pd(CH₃)(CO)(1)][BArF] (1d). Yield: 91%. Red solid.

δ_H(400 MHz; CD₂Cl₂; 298 K; M = major, m = minor)

8.19–8.10 (m, H^5,8_M,m), 7.59–7.51 (m, H^4,9_M,m), 7.46 (1H, d,

Compound 1c (90 mg, 0.132 mmol) was reacted with Na-

[BArF] (117 mg, 0.132 mmol) at −40 °C in 2.5 mL of dichloro-

methane. The reaction mixture was slowly warmed to 0 °C,

and the solution was filtered through celite to remove NaCl.

Then solvent was evaporated in vacuum, and the resulting

solid was washed with hexane (3 × 5 mL). A 182.38 mg sample

of the red compound 1d (0.121 mmol, 91%) was collected.

δ_H(200 MHz, CD₂Cl₂, 298 K) 8.25 (2H, dd, H^5,8), 8.01 (1H, s,

H¹⁶), 7.75 (10H, brs, H^14,18, Ar′–Ho), 7.57 (6H, dt, H^4,9, Ar′-Hp),

7.28 (2H, brs, H^16′and H³), 7.08 (2H, brs, H^14′,18′), 6.64 (1H, s,

H¹⁰), 2.46 (6H, s, –(CH₃)₂–Ar), 1.31 (3H, s, CH₃–Pd, major-

trans), 1.07 (3H, s, CH₃–Pd, minor-cis). δ_13C(50.33 MHz,

CD₂Cl₂, 298 K) 174.38 (s, CO–Pd, trans), 172.77 (s, CO–Pd, cis),

H¹⁰_m), 7.25 (1H, d, H³_M), 6.93–6.84 (m, H^14,16,18

and H¹⁰_M),

M,m

6.79 (1H, d, H³_m), 6.66 (2H, s, H^14′,18′_m), 6.49 (2H, s, H^14′,18′_M),

3.89 (s, ^pOCH_3,M,m), 3.84 (s, ^mOCH_3,M,m), 0.91 (3H, s, Pd–CH_3,M),

0.80 (3H, s, Pd–CH_3,m).

[Pd(CH₃)Cl(7)] (7a). Yield: 87%. Red solid.

Elemental analysis: C, 53.50; H, 2.69; N, 4.99. Found: C,

53.90; H, 2.58; N, 4.98.

δ_H(500 MHz; CD₂Cl₂; 298 K) 8.18 (2H, m, H^5,8), 7.58

(2H, m, H^4,9), 7.28 (1H, d, H³), 7.01–6.83 (7H, m,

H^{14,14′,16,16′,18,18′,10}), 0.85 (3H, s, Pd–CH₃).

[Pd(CH₃)(CH₃CN)(Ar-BIAN)][PF₆] (3b, 7b)

General synthesis. A solution of AgPF₆(0.401 mmol) in

CH₃CN (1 mL) was added to a solution of the neutral complex

(0.349 mmol) in CH₂Cl₂(2 mL). The solution was stirred at r.t.

for 30 min, and then AgCl was filtered over celite; the solution

was then concentrated to half volume under vacuum. Upon

the addition of diethyl ether the product precipitated as a

solid.

[Pd(CH₃)(CH₃CN)(3)][PF₆] (3b). Yield: 77%. Orange solid.

δ_H(400 MHz; CD₂Cl₂; 298 K; M = major, m = minor)

8.25–8.19 (2H, m, H^5,8_M,m), 7.66–7.57 (m, H^4,9_M,m), 7.47 (1H, d,

H¹⁰_m), 7.33 (1H, d, H³_M), 7.06–6.99 and 6.89 (m, H^14,16,18_M,m),

6.84 (1H, d, H¹⁰_M), 6.81 (1H, d, H³_m), 6.63 (2H, s, H^14′,18′_m),

6.48 (2H, s, H^14′,18′_M), 3.90, 3.88, 3.85 (s, ^pOCH_3,M,mand

1

161.73 (q, J_CB= 49.8 Hz, Ar′-Cipso), 147.57, 147.27 (s, CvN),

141.26 (s, ipso–C–Ph–CH₃), 141.04 (s, ipso–C–Ph–CF₃), 134.71

(s, Ar′–Co), 134.20 (s, CH^5,8), 133.71 (s, CH^16′), 128.70 (q, J_CF

=

2

1

28.33 Hz, Cm–Ar′), 126.35 (s, CH^4,9), 124.84 (q, J_CF= 272.28

Hz, CF₃), 124.05, 123.26 (s, CH^14,18), 121.78 (s, CH¹⁰), 121.43

(s, CH¹⁶), 117.95 (s, CH^14′,18′), 117.50 (s, Ar′–Cp), 21.29 (s, CH₃–

Ph), 12.09 (s, CH₃–Pd, trans), 9.11 (s, CH₃–Pd, cis).

[Pd(CH(p-CH₃-Ph)CH₂C(O)CH₃)(1)][BArF] (1e). Yield: 83%.

Dark red solid.

p-Methylstyrene (13.00 μL, 0.1 mmol) was added to a solu-

tion of 1d (150 mg, 0.099 mmol) in 2.5 mL of dichloromethane

cooled to −20 °C. After 20 min the solvent was evaporated

Dalton Trans., 2013, 42, 14583–14602 | 14599

View Article Online

Paper

Dalton Transactions

under vacuum, and the resulting solid was washed with

hexane (2 × 4 mL). A 132.46 mg sample of the dark-red com-

pound 1e (0.083 mmol, 83%) was collected. This complex is

stable in solution only for a few hours.

Table 6 Crystallographic data for compounds 1a and 3a

1a

3a

Empirical formula

Fw

Crystal system

Space group

a (Å)

C

₂₉H₂₁ClF₆N₂Pd

C₂₈H₂₃ClF₂N₂O₃Pd

615.33

Monoclinic

P2₁/c

14.707(4)

14.627(3)

12.192(3)

103.82(3)

2546.8(11)

4

1.605

0.882

1240

1.43–27.05

29 141

5547

0.0870

2614

334

0.901

0.0463, 0.0928

0.486, −0.463

δ_H(200 MHz, CD₂Cl₂, 298 K) 8.13 (2H, dd, H^5,8), 8.11 (3H,

m, H¹⁶and H^14,18), 7.74 (8H, s, Ar′–Ho), 7.56 (6H, s, H^4,9and

Ar′–Hp), 7.24 (2H, m, H^16′and H³), 6.97 (2H, s, H^14′,18′), 6.88

(2H, m, CHo–Ph), 6.59 (1H, s, H¹⁰), 6.55 (2H, s, CHm–Ph), 3.51

(1H, dd, CHa), 2.87 (1H, d, CHb), 2.47 (6H, s, –(CH₃)₂–Ar), 2.10

(3H, s, CH₃(CO)), 1.98 (1H, d, CHb′), 1.89 (3H, s, CH₃–Ph).

[Pd(C(O)CH( p-CH₃-Ph)CH₂C(O)CH₃)(CO)(1)][BAr′₄](1f ).

Bubbling CO at −30 °C for 10 min into a CD₂Cl₂solution

(0.6 mL) of 1e (60 mg) resulted in the formation of 1f. This

complex is stable in solution only for a few hours.

653.33

Monoclinic

P2₁/c

12.138(4)

24.329(5)

8.973(4)

94.82(4)

2640.4(16)

4

b (Å)

c (Å)

β (°)

V (Å³)

Z

D calcd (g cm⁻³

)

1.644

0.868

1304

μ (Mo-Kα), (mm⁻¹

F(000)

)

θ range (°)

1.67–27.10

28 892

5799

0.1192

2762

353

0.916

0.0647, 0.1449

0.793, −0.665

No. of reflns collcd

No. of indep reflns

R_int

δ_H(200 MHz, CD₂Cl₂, 250.9 K) 8.16 (2H, dd, H^5,8), 8.10 (1H,

s, H¹⁶), 7.87 (2H, s, H^14,18), 7.77 (8H, s, Ar′–Ho), 7.64 (2H, dt,

H^4,9), 7.56 (s, 4H, Ar′–Hp), 7.23 (2H, m, H^16′and H³), 7.18 (4H,

m, CH–Ph), 6.95 (2H, s, H^14′,18′), 6.86 (1H, s, H¹⁰), 4.64 (1H, t,

CHa), 2.81 (1H, d, CHb), 2.76 (d, 1H, CHb′), 2.36 (3H, s, CH₃–

Ph), 2.27, 2.13 (6H, s, –(CH₃)₂–Ar) 1.96 (3H, s, CH₃(CO)).

[Pd(C(O)(CH(p-CH₃-Ph)CH₂)₂C(O)CH₃)(CO)(1)][BAr′₄](1g). p-

Methylstyrene (4.3 μL, 0.1 mmol) was added into a CD₂Cl₂

solution (0.6 mL) of 1f (50 mg, 0.033 mmol) cooled to −30 °C.

No. of reflns (I > 2σ(I))

No. of refined params

Goodness-of-fit (F²)

R₁, wR₂(I > 2σ(I))^a

Residuals (e Å⁻³

)

^aR₁= ∑||F_o| − |F_c||/∑|F_o|, wR₂= [∑w (F_o²− F_c²)²/∑w (F_o²)²]^1/2

.

The solution was left for 16 h at −18 °C to yield compound 1g. least-squares method based on F²with all observed reflec-

This complex is stable in solution only for a few hours.

tions.⁷¹All the calculations were performed using the WinGX

δ_H(200 MHz, CD₂Cl₂, 243 K) 8.13 (2H, dd, H^5,8), 7.89 (3H, System, Ver. 1.80.05.⁷²Crystal data and details of refinement

brs, H^14,16,18), 7.78 (10H, brs, H^4,9, Ar′–Ho), 7.49 (4H, s, Ar′– for 1a and 3a are summarized in Table 6.

Hp), 7.49–7.41 (2H, m, H^16′and H³), 7.30–6.85 (11H, m, CH–

Ph, H^14′,18′, H¹⁰), 4.25–3.75 (2H, m, CHc and CHd), 3.50–3.10

Acknowledgements

(2H, m, CHa and CHe), 2.80–2.50 (2H, m, CHb and CHf),

2.50–2.00 (9H, m, –(CH₃)₂–Ar, CH₃–Ph), 1.84 (3H, s, CH₃(CO)).

Collaborative research was carried out by D.V. and B.M. in the

framework of the COST D40 project of the EC that is also

acknowledged for the financial support of an STSM of F.A. to

Eindhoven University of Technology. This work was supported

by MIUR (PRIN no. 2007HMTJWP_002; PRIN no.

2008A7P7YJ_003). We thank Dr F. Ferretti and A. Caselli for

help in the synthesis and characterization of ligands 3 and 7

and Dr P. Natanti for the characterization of complexes 1a–1g.

Engelhard Italia is gratefully acknowledged for a generous gift

of [Pd(AcO)₂]. Fondazione CRTrieste is also gratefully acknowl-

edged for the generous donation to the Dipartimento di

Scienze Chimiche e Farmaceutiche of a Varian 500 NMR

spectrometer.

CO/vinyl arenes copolymerization reactions at 3–10 bar

Catalytic experiments were performed simultaneously in the

parallel autoclave system AMTEC SPR16, equipped with

pressure sensors and a mass-flow controller suitable for moni-

toring and recording gas uptakes throughout the reactions.

Four stainless steel autoclaves (12 mL) of the AMTEC SPR16

were flushed with argon. A solution of the Pd catalyst and 1,4-

benzoquinone in trifluoroethanol (20 mL) was prepared and

the reactors were charged each with 5 mL of the catalyst solu-

tion as well as 3 mL of the alkene comonomer. The atmos-

phere was exchanged with carbon monoxide (gas exchange

cycle 1) and the reactors were pressurized to the desired value.

After heating to the desired temperature (30 °C), the final

pressures were adjusted and kept constant throughout the

experiment. The carbon monoxide uptake was monitored and

recorded automatically.

References

1 I. Matei and T. Lixandru, Chem. Abstr., 1969, 70, 3623.

2 R. Van Asselt and C. J. Elsevier, Organometallics, 1992, 11,

1999.

X-ray crystallography

Collection data of 1a and 3a were carried out at room tempera-

ture using a Nonius DIP-1030H system equipped with Mo-Kα

radiation (λ = 0.71073 Å). Cell refinement, indexing and

scaling of the data sets were carried out using Denzo⁷⁰and

Scalepack.⁷⁰The structures were solved by direct methods and

subsequent Fourier analyses⁷¹and refined by the full-matrix

3 L. K. Johnson, C. M. Killian and M. Brookhart, J. Am. Chem.

Soc., 1995, 117, 6414.

4 S. D. Ittel, L. K. Johnson and M. Brookhart, Chem. Rev.,

2000, 100, 1169.

5 L. K. Johnson, S. Mecking and M. Brookhart, J. Am. Chem.

Soc., 1996, 118, 267.

14600 | Dalton Trans., 2013, 42, 14583–14602

View Article Online

Dalton Transactions

Paper

6 S. Mecking, L. K. Johnson, L. Wang and M. Brookhart, 30 J. March, Advanced organic chemistry, John Wiley & Sons,

J. Am. Chem. Soc., 1998, 120, 888. Inc., 1992.

7 A. Nakamura, S. Ito and K. Nozaki, Chem. Rev., 2009, 109, 31 C. Besson, E. E. Finney and R. G. Finke, J. Am. Chem. Soc.,

5215. 2005, 127, 8179.

8 D. Pappalardo, M. Mazzeo, S. Antinucci and C. Pellecchia, 32 M. A. Watzky and R. G. Finke, J. Am. Chem. Soc., 1997, 119,

Macromolecules, 2000, 33, 9483. 10382.

9 J. Kiesewetter and W. Kaminsky, Chem.–Eur. J., 2003, 9, 33 P. H. M. Budzelaar, P. W. N. M. Van Leeuwen, C. F. Roobeek

1750. and A. G. Orpen, Organometallics, 1992, 11, 23.

10 C. S. Popeney, D. H. Camacho and Z. Guan, J. Am. Chem. 34 M. Tromp, J. R. A. Sietsma, J. A. van Bokhoven, G. P. F. van

Soc., 2007, 129, 10062.

11 L. Guo, H. Gao, Q. Guan, H. Hu, J. Deng, J. Liu, F. Liu and

Q. Wu, Organometallics, 2012, 31, 6054.

Strijdonck, R. J. van Haaren, A. M. J. van der Eerden,

P. W. N. M. van Leeuwen and D. C. Koningsberger, Chem.

Commun., 2003, 128.

12 R. van Asselt, E. E. C. G. Gielens, R. E. Rülke and 35 M. Portnoy, F. Frolow and D. Milstein, Organometallics,

C. J. Elsevier, J. Chem. Soc., Chem. Commun., 1993, 1203. 1991, 10, 3960.

13 G. Bellachioma, B. Binotti, G. Cardaci, C. Carfagna, 36 F. Ragaini, H. Larici, M. Rimoldi, A. Caselli, F. Ferretti,

A. Macchioni, S. Sabatini and C. Zuccaccia, Inorg. Chim.

Acta, 2002, 330, 44.

14 B. Binotti, C. Carfagna, C. Zuccaccia and A. Macchioni,

Chem. Commun., 2005, 92.

15 B. Binotti, G. Bellachioma, G. Cardaci, C. Carfagna,

C. Zuccaccia and A. Macchioni, Chem.–Eur. J., 2007, 13,

1570.

16 M. Gasperini, F. Ragaini, E. Gazzola, A. Caselli and

P. Macchi, Dalton Trans., 2004, 3376.

P. Macchi and N. Casati, Organometallics, 2011, 30, 2385.

37 The possibility that the irreversible step of the decompo-

sition is the reaction of two palladium(0) complexes should

also be considered. However, palladium(0) complex con-

centrations should be very low at any moment and this

process is quite unlikely to play a relevant role in the

present system. From a kinetic point of view, a process of

this kind would be distinguishable from that in path B

only if the concentration/time profile for benzoquinone

were also known, which is not the case, and so we will

ignore it in the following.

17 A. Scarel, M. R. Axet, F. Amoroso, F. Ragaini, C. J. Elsevier,

A. Holuigue, C. Carfagna, L. Mosca and B. Milani, Organo-

metallics, 2008, 27, 1486.

38 M. Gasperini and F. Ragaini, Organometallics, 2004, 23, 995.

18 A. Meduri, T. Montini, F. Ragaini, P. Fornasiero, 39 Note that the same “volume” amount of the two olefins

E. Zangrando and B. Milani, ChemCatChem, 2013, 5, 1170.

19 R. Van Asselt, C. J. Elsevier, W. J. J. Smeets, A. L. Spek and

R. Benedix, Recl. Trav. Chim. Pays-Bas, 1994, 113, 88.

20 M. Gasperini, F. Ragaini and S. Cenini, Organometallics,

2002, 21, 2950.

was employed. Given the higher M_Wand molar volume of

the methyl-substituted substrate, the molar concentration

of the latter was actually a little lower than that of styrene

during the reactions. However, the reactivity diﬀerence is

much larger than that attributable to such a small dilution

eﬀect and must have a diﬀerent origin.

21 J. Durand, E. Zangrando, M. Stener, G. Fronzoni,

C. Carfagna, B. Binotti, P. C. J. Kamer, C. Müller, 40 M. Brookhart, M. I. Wagner, G. G. A. Balavoine and

M. Caporali, P. W. N. M. van Leeuwen, D. Vogt and

B. Milani, Chem.–Eur. J., 2006, 12, 7639.

22 F. Ragaini, M. Gasperini, P. Parma, E. Gallo, N. Casati and

P. Macchi, New J. Chem., 2006, 30, 1046.

23 M. Hagar, F. Ragaini, E. Monticelli, A. Caselli, P. Macchi

and N. Casati, Chem. Commun., 2010, 46, 6153.

24 J. Zhou, H. Sun, K. Harms and J. Sundermeyer, Z. Anorg.

Allg. Chem., 2008, 634, 1517.

H. A. Haddou, J. Am. Chem. Soc., 1994, 116, 3641.

41 S. Bartolini, C. Carfagna and A. Musco, Macromol. Rapid

Commun., 1995, 16, 9.

42 C. Carfagna, G. Gatti, D. Martini and C. Pettinari, Organo-

metallics, 2001, 20, 2175.

43 A. Scarel, B. Milani, E. Zangrando, M. Stener, S. Furlan,

G. Fronzoni, G. Mestroni, S. Gladiali, C. Carfagna and

L. Mosca, Organometallics, 2004, 23, 5593.

25 A. Bastero, A. Ruiz, C. Claver, B. Milani and E. Zangrando, 44 The experiment at 10 bar with 3,5-(CH₃)₂C₆H₃;3,5-

Organometallics, 2002, 21, 5820.

(CF₃)₂C₆H₃-BIAN was also performed. It is not included in

the discussion because the data are of poor quality.

However, even if quantitative data are inaccurate, the reac-

tion rate was without doubt slower than that of both the

reactions run under 3 and 5 bar with the same ligand.

26 A. Scarel, J. Durand, D. Franchi, E. Zangrando, G. Mestroni,

B. Milani, S. Gladiali, C. Carfagna, B. Binotti, S. Bronco

and T. Gragnoli, J. Organomet. Chem., 2005, 690, 2106.

27 M. R. Axet, F. Amoroso, G. Bottari, A. D’Amora, E. Zangrando,

F. Faraone, D. Drommi, M. Saporita, C. Carfagna, P. Natanti, 45 B. Milani, G. Corso, G. Mestroni, C. Carfagna, M. Formica

R. Seraglia and B. Milani, Organometallics, 2009, 28, 4464. and R. Seraglia, Organometallics, 2000, 19, 3435.

28 A. Bastero, C. Claver, A. Ruiz, S. Castillon, E. Daura, C. Bo 46 M. Sperrle, A. Aeby, G. Consiglio and A. Pfaltz, Helv. Chim.

and E. Zangrando, Chem.–Eur. J., 2004, 10, 3747. Acta, 1996, 79, 1387.

29 A. D’Amora, L. Fanfoni, D. Cozzula, N. Guidolin, 47 W. P. Mul, A. W. van der Made, A. A. Samaardijk and

E. Zangrando, F. Felluga, S. Gladiali, F. Benedetti and

B. Milani, Organometallics, 2010, 29, 4472.

E. Drent, in Catalytic synthesis of copolymers and ter-

polymers, ed. A. Sen, Dordrecht, 2003.

Dalton Trans., 2013, 42, 14583–14602 | 14601

View Article Online

Paper

Dalton Transactions

48 A. Gsponer, T. M. Schmid and G. Consiglio, Helv. Chim. 60 M. Brookhart, F. C. Rix, J. M. DeSimone and J. C. Barborak,

Acta, 2001, 84, 2986. J. Am. Chem. Soc., 1992, 114, 5894.

49 J. Durand, A. Scarel, B. Milani, R. Seraglia, S. Gladiali, 61 B. Binotti, C. Carfagna, G. Gatti, D. Martini, L. Mosca and

C. Carfagna and B. Binotti, Helv. Chim. Acta, 2006, 89,

1752.

50 D. P. Gates, S. A. Svejda, E. Onate, C. M. Killian,

C. Pettinari, Organometallics, 2003, 22, 1115.

62 C. Carfagna, G. Gatti, L. Mosca, A. Passeri, P. Paoli and

A. Guerri, Chem. Commun., 2007, 4540.

L. K. Johnson, P. S. White and M. Brookhart, Macro- 63 K. Nozaki, H. Komaki, Y. Kawashima, T. Hiyama and

molecules, 2000, 33, 2320. T. Matsubara, J. Am. Chem. Soc., 2001, 123, 534.

51 M. P. Weberski, C. Chen, M. Delferro, C. Zuccaccia, 64 B. Milani, F. Paronetto and E. Zangrando, J. Chem. Soc.,

A. Macchioni and T. J. Marks, Organometallics, 2012, 31,

3773.

52 H. Grennberg, A. Gogoll and J. E. Baeckvall, Organo-

metallics, 1993, 12, 1790.

53 R. A. Klein, C. J. Elsevier and F. E. Hartl, Organometallics,

1997, 16, 1284.

54 R. A. Klein, P. Witte, R. van Belzen, J. Fraanje, K. Goubitz,

M. Numan, H. Schenk, J. M. Ernsting and C. J. Elsevier,

Eur. J. Inorg. Chem., 1998, 319.

Dalton Trans., 2000, 3055.

65 J. A. Iggo, Y. Kawashima, J. Liu, T. Hiyama and K. Nozaki,

Organometallics, 2003, 22, 5418.

66 L. Resconi, L. Cavallo, A. Fait and F. Piemontesi, Chem.

Rev., 2000, 100, 1253.

67 K. Nozaki, N. Sato, Y. Tonomura, M. Yasutomi, H. Takaya,

T. Hijama, T. Matsubara and N. Koga, J. Am. Chem. Soc.,

1997, 119, 12779.

68 D. Guironnet, L. Caporaso, B. Neuwald, I. Gottker-Schnet-

mann, L. Cavallo and S. Mecking, J. Am. Chem. Soc., 2010,

132, 4418.

69 S. Noda, A. Nakamura, T. Kochi, L. W. Chung, K. Morokuma

and K. Nozaki, J. Am. Chem. Soc., 2009, 131, 14088.

55 B. Milani, A. Anzilutti, L. Vicentini, A. Sessanta o Santi,

E. Zangrando, S. Geremia and G. Mestroni, Organometallics,

1997, 16, 5064.

56 E. Drent and P. H. M. Budzelaar, Chem. Rev., 1996, 96, 663.

57 K. J. Cavell, D. J. Stufkens and K. Vrieze, Inorg. Chim. Acta, 70 Z. Otwinowski and W. Minor, in Macromolecular Crystallo-

1981, 47, 67.

graphy, part A, ed. C. W. Carter Jr. and R. M. Sweet,

New York, 1997, p. 307.

71 G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystal-

logr., 2008, 64, 112.

58 A. M. Kluwer, C. J. Elsevier, M. Bühl, M. Lutz and

A. L. Spek, Angew. Chem., Int. Ed., 2003, 42, 3501.

59 B. V. Popp, J. L. Thorman, C. M. Morales, C. R. Landis and

S. S. Stahl, J. Am. Chem. Soc., 2004, 126, 14832.

72 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.

14602 | Dalton Trans., 2013, 42, 14583–14602

Article Doi

Catalyst activity or stability: The dilemma in Pd-catalyzed polyketone synthesis

DOI: 10.1039/c3dt51425k

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c3dt51425k

Full text of DOI:10.1039/c3dt51425k

Products guided by the article

R&D Labs maybe for 1020254-65-2

Relevant to this article

Hot Product