Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article
Demystifying Ziegler-Natta Catalysts: the Origin of Stereoselectivity
Antonio Vittoria, Anika Meppelder, Nicolaas Friederichs, Vincenzo Busico, and Roberta Cipullo
ACS Catal., Just Accepted Manuscript • Publication Date (Web): 26 May 2017
Downloaded from http://pubs.acs.org on May 26, 2017

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 26 ACS Catalysis

1
2
3
4 Demystifying Ziegler-Natta Catalysts: The Origin of
5
6
7
8 Stereoselectivity
9
10
11
12 Antonio Vittoriaa, Anika Meppelderb, Nic. Friederichsb, Vincenzo Busicoa,*, Roberta Cipulloa,*
13
14
15 a
Laboratory of Stereoselective Polymerizations (LSP), Department of Chemical Sciences, Federico
16
17
II University, Via Cintia, 80126 Naples, Italy
18
19 b
20 SABIC, NL-6160 AH Geleen, Netherlands
21
22
23
24
25 ABSTRACT. Industrial Ziegler-Natta catalysts for polypropylene production are complex
26
27 formulations with a reputation of ‘black boxes’. In the present paper, we report the results of an
28
29
30
extensive investigation of the three latest commercial generations, carried out with advanced High
31
32 Throughput Experimentation tools and methods. The thus obtained database of structure/properties
33
34 relationships, of extraordinary width and depth, provided a high-definition picture of the screened
35
36 systems, enabling us to highlight important details of their inner workings. In particular, the delicate
37
38
relationship between surface coverage and lateral steric pressure on the stereoselectivity of the
39
40
41 catalytic species, as well as the role of chemisorbed donor and Al-alkyl species on said parameters
42
43 for different formulations, were revealed.
44
45
46 KEYWORDS. Ziegler-Natta Catalysts; Isotactic Polypropylene; Internal Donor; External Donor;
47
48 Stereoselectivity.
49
50
51
52
INTRODUCTION
53
54 Isotactic polypropylene (i-PP) and Ti-based Ziegler-Natta (ZN) catalysts have parallel success
55
56 stories. With an annual consumption of more than 60 M tons, i-PP has become the second largest-
57
58
59 volume polymer on the market after polyethylene.1 Progress in catalysis was also dramatic: activity
60 1
ACS Paragon Plus Environment
 ACS Catalysis Page 2 of 26

1
2
3 boosted from few kg to several tons of polymer per g of Ti, and stereoselectivity from moderate to
4
5 almost perfect.2-4
6
7
8 The first two catalyst generations, consisting of crystalline TiCl3 in a layered modification3,4, were
9
10 relatively simple. Stereoselectivity was the consequence of a peculiar crystal lattice with chirotopic
11
12 Ti both in the bulk and on the side edges of the structural layers. The latter, in particular, exposed
13
14
15
linear arrays of enantiomorphous Ti atoms amenable to Cl/R metathesis with an Al-alkyl
16
17 compound.3-5 Elegant experimental6 and computational7,8 studies highlighted the surface constraints
18
19 on the thus formed Ti-alkyls, conformationally locked at the first C-C bond so as to define chiral
20
21 pockets in which the two propene enantiofaces could be discriminated at the insertion step.4,9
22
23
24 Supported homologues with greatly improved performance, but also much more complex
25
26 formulations, were introduced in the 1970s.2-4 MgCl2 was serendipitously identified as the best-
27
28 working support for TiCl4 (the Ti precursor of largest use).2 The addition of certain donor
29
30
31 compounds as powerful stereoselectivity enhancers (Table 1), at the precatalyst preparation stage
32
33 (‘Internal Donors’, IDs) or in combination with the AlR3 activator (‘External Donors’, EDs), was
34
35 also a trial-and-error achievement.2,3
36
37
38 Table 1. Typical formulations and performance ranges of MgCl2-supported ZN catalysts for i-PP
39
40
production.2,3
41
42
43 Generation Internal Donor External Donor Productivity(a) XS(b) Mw/Mn
44 Aromatic
45 Third Ethylbenzoate 0.5-0.8 3-5 6-9
monoester
46
47 Fourth Dialkylphthalate Alkoxysilane 1-2 1-5 6-8
48 2,2’-dialkyl-1,3- None or
Fifth >2 2-5 4-6
49 dimethoxypropane Alkoxysilane
50 Sixth Dialkylsuccinate Alkoxysilane 1-2 1-5 >8
51 (a) 3 -1 (b)
10 kg(PP) g(Ti) . Xylene-Soluble Fraction in wt.-%.
52
53
54
Unraveling the cross-interactions between the components of the different formulations that
55
56
57 followed one another at intervals of approximately 10 years2,3,9 is a long-standing open problem,
58
59 which generated the catalysts’ reputation of ‘black-boxes’. The similarity between the crystal
60 2
ACS Paragon Plus Environment
Page 3 of 26 ACS Catalysis

1
2
3 lattices of MgCl2 and TiCl3 originated the idea of an epitaxial relationship between at least part of
4
5 the TiClx adsorbates and the MgCl2 substrate.2,3,9,10 Rather than representing a constructive input,
6
7 this hypothesis triggered decades of flawed mechanistic speculations, including that of a
8
9
10 competition between TiCl4 and the donors for selective chemisorption on the support.2,3,9,10
11
12 MgCl2(10l) crystal terminations, exposing pentacoordinated Mg, have long been claimed to host
13
14 Ti2Cl8 adducts that, once activated by an AlR3, would mimic the active sites of authentic TiCl3
15
16 catalysts. MgCl2(110) terminations, in turn, featuring tetracoordinated Mg and as such more acidic,
17
18
were postulated to be preferred targets for donor binding, and home to non-stereoselective sites
19
20
21 only. The successful introduction, in the 1990s, of 2,2-dialkyl-1,3-dimethoxypropanes as a class of
22
23 IDs especially prone to chelate tetracoordinated Mg, thus supposedly hampering TiCl4 interaction
24
25 with MgCl2(110) facets2,3,11, was presented as a compelling demonstration of the hypothesis, and
26
27 even an achievement of molecular design.11,12
28
29
30 It was only several years later that more critical analyses of the experimental data and Quantum
31
32 Mechanics modeling studies disproved the concept of MgCl2(10l) and MgCl2(110) as ‘good’ and
33
34
35
‘bad’ surfaces, respectively. As a matter of fact, it became impossible to ignore the unambiguous
36
37 evidence that donor molecules have a direct and specific impact on polymer microstructure13, and
38
39 therefore are, if not part of the catalytic species, at least at non-bonded contact with them.9 Several
40
41 independent Density Functional Theory (DFT) calculations, in turn, concluded that Ti2Cl8
42
43
adsorbates on MgCl2(10l)-like edges are not stable14,15, and lately that TiCl4 chemisorption is only
44
45
46 feasible in mononuclear form on MgCl2(110)-like edges.16,17
47
48 The current view is that the role of donors in MgCl2-supported ZN catalyst systems is twofold:
49
50
51 i) Stabilize the primary particles by strong chemisorption, lowering their surface energy.18-21
52
53 Mg/donor mole ratios in the range of 10 to 20 are not unusual2,3, which points to lateral dimensions
54
55
56
of the structural layers of only few unit cells18,22, and values of surface area in excess of 150 m2 g-1
57
58 (unattainable for binary MgCl2/TiCl4 particles because TiCl4 adsorption is too weak16).
59
60 3
ACS Paragon Plus Environment
 ACS Catalysis Page 4 of 26

1
2
3 ii) Impart the necessary steric hindrance to the inherently chiral but otherwise too open catalytic
4
5 species, very much alike ancillary ligands in molecular catalysts.9 A qualitative model for this
6
7 function was proposed by some of us in Ref23, and still accounts for all known facts, including the
8
9
10 complex stereoblock microstructure of (part of) the polymers.9
11
12 Although in principle one single donor might exert both functions, in most cases ID and ED roles
13
14
15
are differentiated (Table 1).2,3,9 Surface stabilization is important already at the stage of primary
16
17 particle formation; this usually entails chlorination of a Mg precursor compound with excess TiCl4
18
19 in the presence of the ID, which requires that the latter compounds are mutually unreactive. For the
20
21 modification of the catalytic species, on the other hand, a proper steric demand, the preference for
22
23 chemisorption on Mg rather than Ti, and the lack of reactivity with AlR3 are equally important
24
25
26 conditions. Some well-functioning IDs (e.g. dialkylphthalates) react irreversibly with the AlR3
27
28 activator, and are extracted from the solid catalyst during polymerization2,3,24-26; therefore, they
29
30 need to be replaced by an ED. By far the most widely used EDs are sterically demanding
31
32 alkoxysilanes2,3; these are poorly reactive with AlR3 compounds, but do react with TiCl4, which
33
34
35
prevents their use as IDs.
36
37 While the above general picture is sound, what is still missing is an adequate understanding of the
38
39
details that would enable true catalyst design. In particular, how the catalytic species look like and
40
41
42 what determines their diverse behaviors in the different catalyst generations remain largely
43
44 unanswered questions. Ironically, formulations that work well experimentally are not trivial to
45
46 validate by means of computational modeling (as we shall see in the following sections), whereas
47
48 some published models claiming a high stereoselectivity involve compositions that are not used in
49
50
51 industrial practice.17,27 In this sense, the aforementioned black-box perception is justified.
52
53 In this paper, we present the results of an extensive and thorough study of fourth-, fifth- and sixth-
54
55
56
generation ZN catalysts (Table 1), used in combination with AlEt3 and an array of alkoxysilane EDs
57
58 with large structural diversity. The investigation, carried out with advanced High Throughput
59
60 4
ACS Paragon Plus Environment
Page 5 of 26 ACS Catalysis

1
2
3 Experimentation (HTE) tools and methods28, consisted of two parts. In a first part, catalyst
4
5 performance (in terms of polymerization behavior and polymer properties) was determined using a
6
7 fully automated secondary screening platform with 48 mini-reactors (Freeslate PPR48), integrated
8
9
10 with a polymer characterization workflow including Gel Permeation Chromatography (GPC),
11 13
12 analytical Crystallization Elution Fractionation (A-CEF), and C NMR spectroscopy. In a second
13
14 part, another HTE platform (Freeslate Core Module) was used to follow the evolution in
15
16 composition of the solid catalysts under conditions closely mimicking those of application. The
17
18
resulting database of structure/properties relationships, of extraordinary width and depth for a single
19
20
21 investigation, provided a high-definition picture of the screened systems enabling us to highlight for
22
23 the first time important details of their inner workings. In particular, we explored the delicate
24
25 relationship between surface coverage at saturation and lateral steric pressure on the
26
27 stereoselectivity of the catalytic species, and clarified the roles of chemisorbed donors and Al-alkyls
28
29
30 on said parameters for different formulations (i.e. catalyst surface distributions); this is very
31
32 important for further progress. In the latter respect, it should be recalled that a recent REACH ban
33
34 on dialkylphthalates for toxicity issues has generated a growing market demand for their
35
36 replacement1; considering that fourth-generation ZN systems are the working horses of i-PP
37
38
industry2, the question is of high relevance and calls for urgent attention.
39
40
41
42 EXPERIMENTAL SECTION
43
44 The composition of the four (pre)catalysts, prepared according to Ref.29, can be found in Table 2.
45
46
47 The eight alkoxysilane EDs, in turn, are listed in Table 3; ED1 was purchased from Gelest Inc.,
48
49 ED6 was custom-made, whereas all others were obtained from Evonik AG.
50
51
52
Table 2. Compositions of the screened (pre)catalysts.
53
54 Code Internal Donor (ID) Ti (wt%) Mg (wt%) ID (wt%)
55
C1 Dibutylphthalate 2.0 18.6 11.5
56
57 C2 2,2-Diisobutyl-1,3-dimethoxypropane 2.7 18.4 13.2
58 C3 2,2-Dimethyl-1,3-dimethoxypropane 2.1 19.1 9.3
59 C4 2,3-Diisopropyldiethylsuccinate 2.4 19.2 9.5
60 5
ACS Paragon Plus Environment
 ACS Catalysis Page 6 of 26

1
2
3 Table 3. The set of screened alkoxysilane EDs.
4
5
6 ED1 ED2 ED3 ED4
7 Dimethyldimethoxy- Propyltrimethoxy- Propyltriethoxy- Cyclohexylmethyldimethoxy-
8
9
10
11
12 ED5 ED6 ED7 ED8
13 Diisobutyldimethoxy- Diethylaminotriethoxy- Diisopropyldimethoxy- Dicyclopentyldimethoxy-
14
15
16
17
18
19
20
21 All semi-batch propene polymerization experiments in heptane slurry were run in a Freeslate
22
23
Parallel Pressure Reactor setup with 48 reaction cells (PPR48), fully contained in a triple MBraun
24
25
26 glove-box under Nitrogen. The cells, with a working volume of 5.0 mL, featured an 800 rpm
27
28 magnetically coupled stirring, and individual on-line reading/control of temperature, pressure,
29
30 monomer uptake and uptake rate. The setup and the operating protocol were described in full detail
31
32 in Ref28. Polymerization conditions were as follows: T = 70°C; p(C3H6) = 4.5 bar; p(H2) = 0.20 bar;
33
34
35 [Al]/[Ti] = 160; [ED]/[Al] = 0, 0.025, 0.050, 0.10, 0.20; t = 30 min. The p(C3H6)/p(H2) ratio was set
36
37 at a value resulting into average polymer molecular weights in the typical commercial range. All
38
39 experiments were performed in duplicate, for a total of 132 pairs.
40
41
42 The polymerization products were characterized by high-temperature Gel Permeation
43
44 Chromatography (GPC) with a Freeslate Rapid-GPC setup; analytical Crystallization Elution
45
46 Fractionation (A-CEF) with a Polymer Char setup; quantitative 13C NMR with a Bruker DRX 400
47
48
spectrometer equipped with a high-temperature cryoprobe accommodating 5 mm OD tubes and a
49
50
51 pre-heated robotic sample changer (see Ref28 and Supporting Information (SI) for detailed operating
52
53 protocols and conditions).
54
55
56 A second High Throughput Experimentation platform, namely a Freeslate Extended Core Module
57
58 (XCM), also contained in a triple MBraun glove-box and enabling the robotic handling of solids,
59
60 6
ACS Paragon Plus Environment
Page 7 of 26 ACS Catalysis

1
2
3 liquids and slurries according to fully automated weighing, dispensing and reaction protocols, was
4
5 utilized to determine the composition of the four solid catalysts following the reaction with
6
7 AlEt3/ED mixtures or AlEt3 alone in heptane slurry under conditions mimicking those of
8
9
10 polymerization (albeit in the absence of the monomer). For this study, aliquots of 15-20 mg of solid
11
12 (pre)catalyst in a matrix of 24×8 mL pressure-tight vials were suspended in 3.8 mL of heptane,
13
14 added with predefined amounts of AlEt3(/ED) solutions ([Al]/[Ti] = 25, [ED]/[Al] = 0.10), allowed
15
16 to react at 70°C under magnetic stirring (800 rpm) for 30 min, and quenched by rapid cooling to
17
18
19 -15°C. After centrifugation (1400 rpm) in a Savant SPD121P centrifugal drying station, the vials
20
21 were opened and re-positioned in the reaction deck, where the supernatants were removed
22
23 robotically by aspiration through a needle. The solid phases were washed twice with 3.6 mL
24
25 aliquots of heptane (applying the same sequence of decantation and supernatant aspiration), once
26
27
28
with 3.6 mL of pentane, and finally dried under vacuum at 50°C for 10 h in the aforementioned
29
30 drying station. Each dry solid phase was then dissolved in 1.00 mL of methanol-d4; 0.60 mL were
31
32 analyzed by 1H NMR to determine ID and ED contents, whereas 0.40 mL were dried again,
33
34 mineralized, and analyzed by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-
35
36 OES) for Mg, Ti and Al. Quantitative 1H NMR analyses were performed with a Bruker Avance
37
38
39 DRX 400 spectrometer operating at 400 MHz. Acquisition conditions were: 5 mm probe;
40
41 acquisition time, 3.0 s; relaxation delay, 5.0 s; pulse angle, 90°; spectral width, 10 ppm; 8 transients.
42
43 Resonance assignment was based on the literature, and preliminary 1H NMR characterizations of
44
45 the neat donor molecules. Quantitative determinations were based on peak integration against that
46
47
48 of an aliquot of acetonitrile added as internal standard (methyl peak at δ = 2.05 ppm downfield of
49
50 TMS). ICP-OES analyses were carried out using an Agilent 700-series spectrometer, on water
51
52 solutions of the solid phases treated in sequence with 2.0 mL of concentrated H2SO4, 2.0 mL of
53
54 concentrated HNO3, and (when needed) 2.0 mL of 30 vol% H2O2 (total time 16 h). The
55
56
57 spectrometer was calibrated using commercial standard solutions (metal concentrations in the 1-100
58
59 ppm range).
60 7
ACS Paragon Plus Environment
 ACS Catalysis Page 8 of 26

1
2
3 RESULTS AND DISCUSSION
4
5
6 The catalyst systems
7
8 The four ZN catalysts selected for this study (Table 2) belong in the three latest generations of
9
10
11
commercial relevance (Table 1).2,3 Catalyst C1 (ID = dibutylphthalate) is a widely used fourth-
12
13 generation representative.2,3 Catalysts C2 (ID = 2,2-diisobutyl-1,3-dimethoxypropane) and C3 (ID
14
15 = 2,2-dimethyl-1,3-dimethoxypropane) are members of the fifth generation; the former found
16
17 industrial application, whereas the latter is poorly stereoselective11,12 but very interesting from a
18
19
mechanistic standpoint, as we shall see in following sections. Catalyst C4 (ID = 2,3-
20
21
22 diisopropyldiethylsuccinate) was chosen as an example of the sixth generation, the most recent and
23
24 also the least described in the scientific literature.2 Alkoxysilane EDs are employed with all three
25
26 generations (Table 1), even though fifth-generation systems may also be used without.2,3,11 The ED
27
28 set screened in the present study (Table 3) included dimethoxy-, trimethoxy- and triethoxysilanes
29
30
31 bearing linear, branched and cyclic substituents with different steric demand; three of them
32
33 (namely, ED1-ED3) were selected as ‘minimal structures’ for comparative purposes, whereas the
34
35 remaining five (ED4-ED8) are applied commercially.
36
37
38 Polymerization screening
39
40
All 132 duplicate pairs of slurry propene homopolymerization experiments were run under the same
41
42
43 conditions (T = 70°C, p(C3H6) = 4.5 bar, p(H2) = 0.20 bar, [Al]/[Ti] = 160, t = 30 min), except for
44
45 the [ED]/[Al] ratio that was varied stepwise ([ED]/[Al] = 0, 0.025, 0.050, 0.10, 0.20). Catalyst
46
47 deactivation was always negligible, and polymerization kinetics could be simply quantified in terms
48
49 of average catalyst productivity (Rp, in kg(PP) g(catalyst)-1 h-1). Polymer molecular weight,
50
51
52
crystallinity and stereosequence distributions were determined by Rapid-GPC, analytical
53
54 Crystallization Elution Fractionation (A-CEF), and quantitative 13C NMR spectroscopy, all applied
55
56 in HTE mode with protocols specifically implemented for use downstream of the PPR48 platform
57
58 (i.e., applicable to polymer amounts of 0.1-0.2 g, representing the typical yields of PPR48 mini-
59
60 8
ACS Paragon Plus Environment
Page 9 of 26 ACS Catalysis

1
2
3 reactors).28 In particular, the Amorphous Fraction (AF) measured by A-CEF was used as a
4
5 replacement of the Xylene-Soluble (XS) Fraction, difficult to obtain rapidly and reliably in PPR48
6
7 scale; the two methods were shown to correlate nicely.28 The concentration of stereodefects in the
8
9
10 isotactic fraction, in turn, was obtained from the 13C NMR spectra of raw samples, by measuring the
11
12 fractional amount of the mmmrrmmm nonad in the methyl region.9,28 For selected PP samples, the
13
14 mrrm pentad and rrrrrr heptad were also quantified to reveal the presence of stereoblock chains.9,23
15
16
17 The full results of the polymerization screening are reported in Tables S1-S4 (SI). An excerpt for all
18
19 catalyst systems at [ED]/[Al] = 0.10 is given in Table 4 (A-D) and Figures 1, 2. The three catalyst
20
21 generations are known to yield polymers with different and characteristic Mw/Mn ranges (Table
22
23 1)2,3; our results are in line with that.30 The propensity to undergo ED modification was also
24
25
26 idiosyncratic: very high for catalyst C1, moderate for catalyst C4, almost negligible for catalysts C2
27
28 and C3. We will comment extensively on the overall results in general, and on the latter aspect in
29
30 particular, in a subsequent section.
31
32
33 Table 4. Experimental results for catalyst C1-C4-AlE3/EDx at [ED]/[Al] = 0.10.
34
35
A) Catalyst C1
36
37 Rp Mn Mw AF Tel,max [mmmrrmmm] [mrrm] [rrrrrr]
EDx Mw/Mn
38 (Kg g-1 h-1) (KDa) (KDa) (%) (°C) (%) (%) (%)
39 None 7.2 17 103 6.1 15.2 113.2 0.90 2.1 1.3
40 7.9 18 103 5.7 11.9 113.0 0.92 2.0 1.3
41 ED1 6.3 29 175 6.0 6.3 112.9 1.12 2.0 0.4
42 5.0 29 175 6.0 7.3 113.1 1.06 1.9 0.5
43 ED2 5.0 42 222 5.3 3.3 115.8 0.48 1.0 0.3
44 5.5 29 229 7.8 3.5 115.9 0.49 0.9 0.2
45 ED3 5.2 30 180 6.0 4.8 115.5 0.49 0.9 0.4
46 6.3 26 145 5.5 4.4 115.6 0.48 1.0 0.3
47 ED4 8.3 42 257 6.1 4.3 116.2 0.41 1.0 0.3
48 6.4 32 284 8.8 3.8 116.2 0.47 1.0 0.3
49 ED5 7.2 38 251 6.6 4.7 116.3 0.46 0.8 0.3
50 8.3 41 242 5.9 3.6 116.3 0.40 0.8 0.3
51 ED6 4.8 25 198 7.9 5.2 116.4 0.43 1.0 0.4
52 5.1 25 177 7.2 5.5 116.4 0.43 0.9 0.4
53 6.7 40 319 8.0 3.2 117.4 0.30 0.7 0.2
ED7
54 7.1 39 313 8.0 3.6 117.5 0.29 0.6 0.2
55 8.7 45 393 8.8 3.0 117.9 0.26 0.6 0.2
ED8
56 7.6 46 399 8.7 3.0 117.8 0.29 0.6 0.2
57
58
59
60 9
ACS Paragon Plus Environment
 ACS Catalysis Page 10 of 26

1
2
3
4
5
6
7 B) Catalyst C2
8 Rp Mn Mw AF Tel,max [mmmrrmmm] [mrrm] [rrrrrr]
EDx Mw/Mn
9 (Kg g-1 h-1) (KDa) (KDa) (%) (°C) (%) (%) (%)
10 None 11.1 27 124 4.6 3.9 114.0 0.75 1.3 0.1
11 10.7 33 162 4.9 4.4 113.7 0.67 1.3 0.1
12 11.2 32 157 4.9 3.5 113.5 0.66 1.1 0.2
ED1
13 8.8 33 173 5.3 4.1 113.6 0.65 1.1 0.1
14 7.8 32 169 5.3 2.8 114.3 0.56 1.0 0.1
ED2
15 6.5 32 164 5.2 2.4 114.2 0.55 1.0 0.1
16 10.4 34 185 5.5 3.6 114.4 0.65 1.0 <0.1
ED3
17 8.9 32 169 5.3 3.6 114.1 0.57 1.0 0.2
18 10.1 35 157 4.5 3.7 114.1 0.52 1.0 0.1
19 ED4
10.1 35 161 4.6 2.9 114.2 0.53 1.0 0.1
20 9.2 35 194 5.6 3.1 114.1 0.58 1.0 0.1
21 ED5
7.1 26 144 5.6 3.4 113.9 0.55 1.0 0.1
22
ED6 9.5 34 179 5.2 3.7 113.5 0.56 1.0 0.1
23
9.6 34 205 6.0 4.4 112.8 0.47 0.9 0.2
24
ED7 10.6 29 166 5.8 3.6 114.0 0.55 1.0 0.1
25
5.9 33 168 5.1 4.9 113.8 0.58 1.0 0.1
26
ED8 10.8 38 189 5.0 3.6 114.6 0.60 1.0 0.2
27
10.0 33 180 5.5 3.8 114.3 0.53 0.9 0.1
28
29
30
31
32
33
34
35 C) Catalyst C3
36 Rp Mn Mw AF Tel,max [mmmrrmmm] [mrrm] [rrrrrr]
EDx Mw/Mn
37 (Kg g-1 h-1) (KDa) (KDa) (%) (°C) (%) (%) (%)
38 6.0 19 87 4.7 19.4 109.7 1.63 3.5 0.8
None
39 5.9 28 105 3.7 18.2 110.0 1.53 3.0 0.6
40 4.7 21 104 4.9 14.8 110.5 1.33 2.6 0.6
ED1
41 4.0 22 108 5.0 15.0 110.4 1.35 2.5 0.4
42 3.5 23 101 4.3 12.7 111.1 1.33 2.4 0.5
ED2
43 3.8 21 109 5.1 10.1 111.2 1.17 2.4 0.5
44 4.5 22 122 5.6 12.3 110.9 1.24 2.4 0.5
45 ED3
4.7 18 96 5.4 13.2 110.9 1.24 2.4 0.5
46 5.0 22 110 5.0 13.6 111.2 1.18 2.4 0.5
47 ED4
5.0 20 99 4.9 12.9 111.0 1.22 2.5 0.5
48
ED5 4.5 24 114 4.7 12.5 110.9 1.28 2.4 0.5
49
4.9 22 105 4.7 12.4 110.9 1.39 2.5 0.5
50
ED6 4.1 23 110 4.7 12.7 111.1 1.29 2.3 0.6
51
4.5 28 114 4.1 12.3 111.4 1.16 2.1 0.4
52
ED7 4.1 21 103 4.9 14.5 111.4 1.28 2.4 0.5
53
4.8 25 108 4.3 13.3 111.7 1.19 2.5 0.5
54
ED8 4.8 23 113 5.0 12.5 110.9 1.14 2.4 0.5
55
56 4.3 23 112 4.9 13.2 110.8 1.15 2.5 0.5
57
58
59
60 10
ACS Paragon Plus Environment
Page 11 of 26 ACS Catalysis

1
2
3
4
5
6
7 D) Catalyst C4
8 Rp Mn Mw AF Tel,max [mmmrrmmm] [mrrm] [rrrrrr]
EDx Mw/Mn
9 (Kg g-1 h-1) (KDa) (KDa) (%) (°C) (%) (%) (%)
10 4.0 16 103 6.5 9.9 115.0 0.70 1.6 0.6
None
11 4.8 15 124 8.1 9.6 115.2 0.78 1.5 0.7
12 5.1 16 134 8.3 8.5 115.2 0.78 1.6 0.6
13 ED1
4.9 17 135 7.8 9.0 115.1 0.81 1.6 0.7
14 4.8 18 139 7.7 5.0 115.3 0.53 1.1 0.3
15 ED2
5.1 20 152 7.6 5.1 115.8 0.50 1.0 0.3
16
ED3 5.2 18 167 9.1 5.9 115.9 0.57 1.1 0.6
17
5.7 19 143 7.7 5.6 115.8 0.53 1.1 0.5
18
ED4 6.3 20 182 9.3 5.6 116.1 0.54 1.1 0.5
19
6.3 19 180 9.5 6.0 116.2 0.51 0.9 0.4
20
ED5 5.9 20 175 9.0 6.0 115.9 0.52 0.9 0.5
21
6.9 20 168 8.6 5.7 116.3 0.51 0.9 0.4
22
23 ED6 5.1 18 155 8.7 7.1 116.3 0.45 0.9 0.4
24 5.1 18 153 8.6 6.3 116.4 0.48 1.0 0.4
25 ED7 6.0 22 174 8.0 5.9 116.7 0.40 1.0 0.3
26 6.2 20 171 8.6 5.1 116.7 0.42 0.9 0.3
27 ED8 7.2 23 274 12.1 5.5 117.2 0.37 0.8 0.3
28 6.1 19 250 13.2 5.1 117.2 0.39 0.9 0.3
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59 Figure 1. Normalized fraction of the mmmrrmmm nonad9,28 for the i-PP samples of Table 4.
60 11
ACS Paragon Plus Environment
 ACS Catalysis Page 12 of 26

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Figure 2. Overlay of the A-CEF elution curves for representative polymers obtained with catalysts
24
25
26 C1 (A), C2 (B), C3 (C), and C4 (D) in combination with different AlEt3(/EDx) mixtures.28
27
28
29 Catalyst composition screening
30
31
In parallel with the polymerization screening, we determined the evolution of catalyst composition,
32
33
34 in terms of Ti, Al, ID and ED contents, following the reaction of all four catalysts with AlEt3(/ED)
35
36 in heptane slurry under conditions similar to those of application (see Experimental Section). For
37
38 this study, we selected two EDs largely differing in steric bulk, namely dimethyldimethoxysilane
39
40 (ED1) and dicyclopentyldimethoxysilane (ED8); the [ED]/[Al] ratio was set at a value of 0.10 (i.e.,
41
42
43
the same used for the experiments in Table 4). From the results, reported in Figure 3 and Table S5
44
45 (SI), the following facts should be noted:
46
47 [a] Interaction with AlEt3 resulted into a substantial ID extraction for catalysts C1 (≈60%) and C4
48
49
(30%). With AlEt3/ED combinations the process went even further (≈80% ID extraction for C1,
50
51
52 ≈60% for C4), and was associated with an extensive ID/ED exchange.
53
54 [b] The IDs of catalysts C2 and C3, on the other hand, were retained for over 85% upon exposure to
55
56 AlEt3 as well as AlEt3/ED. In the latter case, only a modest amount of ED was adsorbed.
57
58
59
60 12
ACS Paragon Plus Environment
Page 13 of 26 ACS Catalysis

1
2
3 [c] For a given catalyst, the incorporation of ED1 and ED8 (in mol per mol of Mg) was practically
4
5 the same.
6
7 [d] A massive chemisorption of Al species was observed for all systems, including those where the
8
9
10 ID was retained.
11
12 [e] All catalysts underwent a significant (20-50%) loss of Ti.
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47 Figure 3. Evolution of catalyst composition upon reaction with AlEt3(/ED) (see text and Table S5).
48
49
50
Facts [a] and [b] are the aggregate of several concurring processes, on which the previous literature
51
52 shed light only in part.2,3,10,11,24-26 ID extraction by the AlR3 may follow from an irreversible
53
54 chemical reaction, or the formation of a strong Lewis acid-base adduct. Based on preliminary
55
56 results of a wider (and still ongoing) kinetic study (SI, Table S6), we concluded that the former is
57
58 the case for dibutylphthalate (for reduction by AlEt3 at 70°C in toluene solution we measured a
59
60 13
ACS Paragon Plus Environment
 ACS Catalysis Page 14 of 26

1
2
3 first-order kinetics with a value of kobs = (3.45±0.05)×10-3 s-1), but not for 2,3-diisopropyldiethyl-
4
5 succinate (with kobs = (1.95±0.15)×10-5 s-1). 1,3-Dimethoxypropanes, in turn, do not react with AlR3
6
7
8 compounds, but rapidly form adducts with them in solution.2,3 According to the first seminal papers
9
10 on fifth-generation ZN catalyst systems11,12, IDs with bulky substituents on C-2 (like e.g. 2,2-
11
12 diisobutyl-1,3-dimethoxypropane in C2) are strongly bound to MgCl2(110) terminations, and their
13
14 extraction by AlR3 is marginal; on the other hand, less sterically demanding homologues (like 2,2-
15
16
17
dimethyl-1,3-dimethoxypropane in C3) would adsorb much more weakly, because in a large
18
19 fraction of low-energy conformers the two O atoms are too far apart to chelate tetracoordinated Mg
20
21 (O∩O distance >3 Å). Our results (Figure 3 and Table S5) do not endorse such a discrimination; as
22
23 a matter of fact, the ID of catalyst C3 turned out to be as extensively and firmly bound to the
24
25
26 catalyst surface as that of catalyst C2 (which is also in line with recent DFT calculations21).
27
28 We will come back to fact [c] at a later stage. Regarding fact [d], our ICP-OES determinations were
29
30
31
not informative on the chemical nature of the Al adsorbates. TiCl4 reduction by AlR3 compounds is
32
33 known to proceed with the formation of AlR3-xClx species (in particular, AlEt2Cl).2-4 In the case of
34
35 catalyst C1, AlEt2OBu was also formed, as a reduction product of the dibutylphthalate ID (See SI).
36
37 Last but not least, fact [e] confirms the rather weak chemisorption of TiCl4 onto MgCl216; we
38
39 verified that the fraction of TiClx lost to the liquid phase had negligible activity in propene
40
41
42 polymerization under the conditions of this study, as was demonstrated by means of propene
43
44 polymerization tests on the filtrates.
45
46
47 Donor structure/properties analysis
48
49
50 Let us now examine the impact of individual donor structures on catalyst performance. As noted
51
52 above, the screened EDs modulated very effectively the stereoselectivity of systems C1-AlEt3/EDx
53
54 and, to a lower extent, of C4-AlEt3/EDx ones (Tables 4-A, 4-D, S1, S4 and Figures 1, 2). It is
55
56
plausible to trace this finding to the observed ID/ED exchange (Figure 3 and Table S5), that should
57
58
59 result into a significant fraction of catalytic species in C1 and C4 with neighboring ED molecules.
60 14
ACS Paragon Plus Environment
Page 15 of 26 ACS Catalysis

1
2
3 We note at this point that the x numeral in the EDx identification codes of the screened
4
5 alkoxysilanes was assigned ex-post, in such a way that a higher x corresponded to a higher
6
7 stereoselectivity within the C1-AlEt3/EDx series (Table 4-A). Notably, a very similar ordering
8
9
10 turned out to hold for the C4-AlEt3/EDx series too (Table 4-D and Figure 1). In a first
11
12 approximation, the steric crowding at the Si atom, that is to say next to the surface once the ED
13
14 molecules get adsorbed, also grows with growing x (Table 3). This correlation is less obvious than
15
16 it may appear; in fact, it suggests that for all alkoxysilanes in the set similar chemisorption modes
17
18
ended up with comparable degrees of coverage at saturation (in mol per mol of Mg) for the
19
20
21 available surfaces of each given catalyst20,31; the adsorption data of ED1 and ED8 (Figure 3) are
22
23 compatible with such an assumption. If the hypothesis holds, then the lateral steric pressure
24
25 experienced by the catalytic species, and hence their stereoselectivity according to the model of
26
27 Ref23 (see below), should grow with growing alkoxysilane steric demand, and attain a characteristic
28
29
30 plateau value for each ED once surface saturation is reached; looking at Tables S1 and S4, this
31
32 seems indeed to occur around [ED]/[Al] ≈ 0.05.
33
34
35 What does not seem to fit in the picture, on the other hand, is that systems C1-AlEt3 and C4-AlEt3
36
37 turned out to be slightly more stereoselective than C1-AlEt3/ED1 and C4-AlEt3/ED1 (Tables 4-A,
38
39 4-D and Figures 1, 2). Our explanation is that the chemisorbed Al species (Figure 3) surrogated the
40
41 ED as ID replacements. AlEt3 and AlEt2Cl are strong Lewis acids, known for their self-dimerization
42
43
44 equilibria32,33; in the monomeric state, they can form hetero-dinuclear adducts with Al-Cl-Mg and
45
46 Al-Cl-Ti bridges.2-4 Al-alkyl binding to TiCl3 with formation of doubly-bridged Al-[(µ-Cl)(µ-Et)]-
47
48 Ti moieties is strong34,35, and likely one of the reasons for the low concentration of active Ti
49
50 measured in Quenched-Flow studies.36,37 As far as binding to MgCl2 is concerned, a recent DFT
51
52
study concluded that AlEt3-xClx chemisorption (x = 0, 1) on MgCl2(104) facets is exergonic17,
53
54
55 which is in line with the results in Figure 3. Based on our polymerization data (Figures 1, 2), the
56
57 ability of adsorbed Al-alkyls to enhance catalyst stereoselectivity is similar or even slightly higher
58
59 than that of small EDs (like e.g. ED1), but much poorer than for best-in-class EDs (e.g., ED7 or
60 15
ACS Paragon Plus Environment
 ACS Catalysis Page 16 of 26

1
2
3 ED8). It should be noted that the amount of chemisorbed Al on catalysts C1 and C4 was lower in
4
5 the presence of an alkoxysilane (Figure 3), which we interpret as evidence for a competition; as a
6
7 matter of fact, according to computational modeling data, alkoxysilanes prevail over AlEt3-xClx
8
9
10 species for adsorption on plain MgCl2 crystal terminations.17,20,27
11
12 A completely different picture emerged for systems based on catalysts C2 and C3, whose 1,3-
13
14
15
dimethoxypropane IDs were the dominant donors in the adsorbate pool, leaving very limited room
16
17 for ED action (Tables 4-B, 4-C, S2, S3, S5, and Figures 1-3). Based on conventional wisdom, such
18
19 ZN systems should be the easiest to interpret; as a matter of fact, the strong preference of their IDs
20
21 for chemisorption on MgCl2(110) terminations11,12,31 is expected to determine the least
22
23 differentiated surface environment (which indeed is consistent with the comparatively narrow
24
25
26 molecular weight distribution (MWD) of the produced polymers). Yet, our computational modeling
27
28 studies indicated that 1,3-dimethoxypropanes on plain MgCl2(110) facets, irrespective of the steric
29
30 bulk of the alkyl substituents on C-2, cannot get close enough to adjacent TiCl2R catalytic species
31
32 to make them highly stereoselective in propene insertion. The data in Figure 3 and Table S5,
33
34
35
demonstrating unexpectedly that large amounts of Al species were chemisorbed on both catalysts
36
37 C2 and C3, despite the presence of the ID and with little (catalyst C2) or practically no (catalyst
38
39 C3) evidence of competition with the ED when used, can provide a solution to this puzzling
40
41 problem, as will be illustrated in the next section.
42
43
44 Improving the computational models of catalytic species
45
46
47 Let us now discuss how to interpret the phenomenological picture of the previous sections in terms
48
49 of suitable models of ZN catalytic species. Different approaches have been reported to quantify the
50
51
52
number of such species. Simulations of PP MWDs as summations of Schulz-Flory functions ended
53
54 up with a minimum of four components38; however, this method cannot discriminate between
55
56 chemical and physical effects on the MWD, and the possibility of overdetermined solutions is high
57
58 because the poor resolution and limited precision of MWD data complicate the evaluation of model
59
60 16
ACS Paragon Plus Environment
Page 17 of 26 ACS Catalysis

1
2
3 significance. A more robust approach, in our opinion, is based on the statistical analysis of high-
4
13
5 resolution C NMR stereosequence distributions.9,23 This identified three basic types of
6
7 stereosequences, namely highly isotactic, weakly isotactic (‘isotactoid’), and syndiotactic9,23; the
8
9
10 plausible assumption of a corresponding number of distinct families of catalytic species was
11
12 translated into the three-site model of Ref23. Here we propose an updated version, assuming that all
13
14 catalytic species are mononuclear Ti(III) surface adducts with the structure of Figure 4, in line with
15
16 the indications of recent Raman39 and high-resolution ESR40 studies, and of the latest DFT
17
18
calculations.16,17 The first coordination sphere of Ti is octahedral and C2-symmetric (like in
19
20
21 crystalline TiCl3)9; steric hindrance in the second coordination sphere, on the other hand, can vary.
22
23 Assuming a Cossee-type chain migratory insertion mechanism5,9, highly isotactic chain propagation
24
25 requires that the active sites are sterically constrained at two diagonal octants out of the four where
26
27 the first C-C bond of the growing polymer chain can be located in the 1,2 propene insertion TS, thus
28
29
30 locking chain conformation in the desired chiral orientation and ensuring site (pseudo)homotopicity
31
32 (Figure 4-A).9 Such a condition can be met when adsorbates with adequate bulk (vide infra) occupy
33
34 the adjacent surface just at the limit still allowing fast monomer access to the Ti center with the
35
36 favored enantioface, and are under strong lateral pressure by the neighboring co-adsorbate pool,
37
38
freezing diffusion phenomena41,42 or even hindered conformational motions. Should said steric
39
40
41 pressure fade, the enantioselectivity will decrease, because the conformational constraints on the
42
43 growing chain will weaken, and the chiral active pockets become too loose. Depending on the
44
45 extent of said fading, and whether only one octant or both octants is/are involved, chain propagation
46
47 will deteriorate to weakly isotactic (Figure 4-B), or even chain-end-controlled syndiotactic (Figure
48
49
50 4-C).9,23 In case of a dynamic character of the interested surfaces, stereoblock chains may form.9,23
51
52 The distribution of the three basic cases of Figure 4 (i.e. close/close, close/open, open/open octants)
53
54
55
is a function of the adsorbate pool. In the previous sections, we gave evidence that said pool
56
57 includes not only donors, but also Al-alkyls. Looking at the recent literature17,27, it appears that
58
59 models of TiCl4 adsorbates at defective locations of MgCl2(104)-like edges exposing
60 17
ACS Paragon Plus Environment
 ACS Catalysis Page 18 of 26

1
2
3 tetracoordinated Mg are not incompatible with the hypothesis of an effective steric modification by
4
5 adjacent alkoxysilane EDs (even if explicit calculations are still pending).
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23 Figure 4. Updated three-site model for ZN catalysts (see text and Ref.23). Mg and Cl atoms are
24
25
26 colored in violet and green, respectively. The large spheres in red highlight the two active Ti sites
27
28 according to the Cossee insertion mechanism; those in light blue, the surface Mg sites where the
29
30 presence of an adsorbate would hinder one of the two octants (in light grey) where the first chain C-
31
32 C bond could be located. Chain propagation is predicted to be highly isotactic in case (A); weakly
33
34
isotactic in case (B); chain-end-controlled syndiotactic in case (C).
35
36
37
38 A case where the cooperation of Al-alkyls is required, on the other hand, seems that of fifth-
39
40 generation catalysts. Figure 5-A shows a computational model (for full details see SI) of a portion
41
42
43 of plain MgCl2(110) edge accommodating a TiCl4 unit and an adjacent 2,2-dimethyl-1,3-
44
45 dimethoxypropane molecule in the minimum energy structure; it is evident on inspection, and was
46
47 confirmed by calculation, that the two co-adsorbates are too far apart to give rise to a catalytic
48
49 species falling under the case of Figure 4-A. Figure 5-B shows the same fragment with an
50
51
additional AlEt2Cl molecule chemisorbed in between the two aforementioned adsorbates; the
52
53
54 calculated free energy of adsorption was ∆Gads = -7.9 kcal mol-1. Notably, even small alkoxysilane
55
56 molecules like ED1 were estimated to be too bulky to effectively compete with the AlEt2Cl moiety
57
58 for chemisorption at that specific surface vacancy. For propene insertion at the catalytic species
59
60 18
ACS Paragon Plus Environment
Page 19 of 26 ACS Catalysis

1
2
3 formed by alkylation and reduction of the TiCl4 precursors in Figure 5-A and 5-B, we calculated
4
5 ∆Gre/si ≈ 0 and 1.5 kcal mol-1, respectively; the latter is in good agreement with experiment (Table
6
7
4-C). Conformational interlocking of ID and AlEt2Cl enhanced stereorigidity; with bulkier 1,3-
8
9
10 dimethoxypropanes this can only be more severe (calculations are running). The presence of ED
11
12 molecules at distal surface locations can also contribute to enforce the necessary lateral steric
13
14 pressure (Figures 1-3).
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32 Figure 5. (A) DFT model of adjacent TiCl4 and 2,2-diisobutyl-1,3-dimethoxypropane co-adsorption
33
34 on a MgCl2(110) edge. (B) The same after the adsorption of an AlEt2Cl molecule (see text). Color
35
36 key: Mg/Violet; Ti/Light grey; Al/Pink; Cl/Green; O/Red; C/Dark grey.
37
38
39
40 CONCLUSIONS
41
42
In this paper, we re-visited the polymerization behavior of the three latest generations of industrial
43
44
45 MgCl2-supported ZN catalysts for i-PP production with a comprehensive integrated HTE approach.
46
47 The extensive structure/properties database implemented under rigorously controlled experimental
48
49 conditions enabled us to analyze and compare catalyst behaviors at an unprecedented level of detail;
50
51 as a result, we highlighted the roles of all individual catalyst components, and how surface coverage
52
53
54 and lateral steric pressure on the active sites can be modulated by means of co-adsorbed donor and
55
56 Al-alklyl molecules for different surface distributions. A full molecular description of these systems
57
58 is still difficult to imagine, if only because we are far from sorting out quantitatively the competing
59
60 19
ACS Paragon Plus Environment
 ACS Catalysis Page 20 of 26

1
2
3 interactions between all components of successful formulations. As long as this is the case,
4
5 deterministic catalyst design remains unfeasible. On the other hand, the working principles of such
6
7 formulations are now understood, and searching for novel/better ones is possible with a
8
9
10 combination of HTE and statistical modeling tools (e.g. of QSAR type).
11
12 Concerning in particular the implementation of new donors, the pros and cons of the two viable
13
14
15
options (i.e., one single donors able to play the two key functions of surface stabilization and active
16
17 site modification, vs function-specific ID/ED pairs2,3) can be better evaluated in the light of our new
18
19 results. The single-donor solution is undoubtedly more elegant, and in principle more simple from
20
21 the catalyst application standpoint. However, donor design is extremely demanding, because the
22
23 aforementioned functions call for different (albeit not incompatible) requirements; combining all of
24
25
26 them without compromising on ultimate performance is virtually impossible, as is demonstrated by
27
28 the 1,3-dimethoxypropane case history. Moreover, the approach is difficult to implement in HTE
29
30 mode, because it includes the stage of precatalyst synthesis, involving a harsh chemistry with
31
32 several lengthy kinetically controlled steps.2,3 The ID/ED pair solution, in turn, looks more
33
34
35
complicated on paper, because it entails two independent searches instead of one, and the
36
37 optimization of additional process variables (in particular, [Ti]/[ED] and [Al]/[ED] ratios as a
38
39 function of polymerization temperature). Yet, compared with the single-donor approach, each of the
40
41 two searches is simpler, because the number of performance criteria to be met by the donor is
42
43
lower. Moreover, extra-functions can be introduced more easily; as an example, ID design can be
44
45
46 aimed to steer the formation of MgCl2(10l)-like or MgCl2(110)-like facets in case chemisorption is
47
48 favored on one of the two31, with interesting opportunities for e.g. polymer MWD modulation. The
49
50 ED search, in turn, can be carried out with HTE methodologies28, or even be unnecessary because it
51
52 is unlikely that a novel ID can lead to a (pre)catalyst with unprecedented MgCl2 surfaces, which
53
54
55
implies that existing alkoxysilanes should always represent an option. A superior versatility might
56
57 be claimed as a further advantage of the ID/ED philosophy; diversifying the performance of one
58
59 single (pre)catalyst by means of a broad ED portfolio looks appealing in fundamental terms,
60 20
ACS Paragon Plus Environment
Page 21 of 26 ACS Catalysis

1
2
3 although plant engineers who know better about the drawbacks of transitions in huge continuous
4
5 production lines may have a different opinion.
6
7
8 As a final remark, we note that in the present paper we did not consider two very important and
9
10 closely related properties of ZN systems for i-PP, namely regioselectivity and H2 response.2,3,9,37,43
11
12 We have recently implemented HTE protocols for determining such properties for the first time; the
13
14
15
results will be presented and discussed in an independent article.
16
17
18 AUTHOR INFORMATION
19
20
Corresponding Authors
21
22
23 *E-mail: busico@unina.it, rcipullo@unina.it
24
25
26 Notes
27
28 The Authors declare no competing financial interest.
29
30
31
ASSOCIATED CONTENT
32
33
34
35 Supporting Information. Chemicals specifications. Full results of the polymerization screening.
36
37 Full results of the catalyst composition screening. Kinetic data for the reaction of dibutylphthalate
38
39 and 2,3-diisopropyldiethylsuccinate with AlEt3 at 70°C. Computational modeling details. This
40
41 material is available free of charge via the Internet at http://pubs.acs.org.
42
43
44
45
ACKNOWLEDGMENTS
46
47 The authors are grateful to Emanuele Breuza and Andrea Correa for the computational modeling
48
49
results of Figure 5, and Peter H. M. Budzelaar for useful discussions.
50
51
52
53
54
55
56
57
58
59
60 21
ACS Paragon Plus Environment
 ACS Catalysis Page 22 of 26

1
2
3 REFERENCES
4
5
6 (1) Ali, S., Catal. Rev. 2014, 27, 7-14.
7
8
9 (2) Cecchin, G.; Morini, G.; Piemontesi, F.; Seidel, A., Kirk-Othmer Encyclopedia of Chemical
10
11 Technology. Wiley Interscience: New York, 2007; Vol. 26.
12
13
14 (3) Pasquini, N., Ed., Polypropylene Handbook (2nd Ed.). Hanser Publishers: Munich, Germany,
15
16 2005.
17
18
19
(4) Busico, V., In Giulio Natta and the Development of Stereoselective Propene Polymerization
20
21 in Polyolefins: 50 years after Ziegler and Natta I.; Kaminsky, W. Ed.; Springer: Heidelberg, 2013;
22
23 Vol. 257, 37-58.
24
25
26 (5) Arlman, E. J.; Cossee, P. J. Catal. 1964, 3, 99-104.
27
28 (6) Zambelli, A.; Sacchi, M. C.; Locatelli, P.; Zannoni, G. Macromolecules 1982, 15, 211-212.
29
30
31 (7) Corradini, P.; Barone, V.; Fusco, R.; Guerra, G. Eur. Polym. J. 1979, 15, 1133-1141
32
33
34 (8) Corradini, P.; Guerra, G.; Fusco, R.; Barone, V. Eur. Polym. J. 1980, 16, 835-842.
35
36
(9) Busico, V.; Cipullo, R. Prog. Polym. Sci. 2001, 26, 443-533.
37
38
39 (10) Corradini, P.; Barone, V.; Fusco, R.; Guerra, G. Gazz. Chim. Ital. 1983, 113, 601-607.
40
41
42 (11) Albizzati, E.; Giannini, U.; Morini, G.; Galimberti, M.; Barino, L.; Scordamaglia, R.
43
44 Macromol. Symp. 1995, 89, 73-89.
45
46
47 (12) Scordamaglia, R.; Barino, L., Macromol. Theory Simul. 1998, 7 (4), 399-405.
48
49
50
(13) See, e.g.: Morini, G.; Albizzati, E.; Balbontin, G.; Mingozzi, I.; Sacchi, M. C.; Forlini, F.;
51
52 Tritto, I., Macromolecules 1996, 29 (18), 5770-5776.
53
54
(14) Seth, M.; Margl, P. M.; Ziegler, T., Macromolecules 2002, 35, 7815-7829
55
56
57 (15) Boero, M.; Parrinello, M.; Weiss, H.; Hüffer, S., J. Phys. Chem. A 2001, 105, 5096-5105.
58
59
60 22
ACS Paragon Plus Environment
Page 23 of 26 ACS Catalysis

1
2
3 (16) D’Amore, M.; Credendino, R.; Budzelaar, P. H. M.; Causá, M.; Busico, V. J. Catal. 2012,
4
5 286, 103-110
6
7
8 (17) Credendino, R.; Liguori, D.; Fan, Z.; Morini, G.; Cavallo, L. ACS Catal. 2015, 5, 5431-
9
10 5435.
11
12
13 (18) Busico, V.; Causà, M.; Cipullo, R.; Credendino, R.; Cutillo, F.; Friederichs, N.; Lamanna,
14
15 R.; Segre, A.; Van Axel Castelli, V. J. Phys. Chem. C 2008, 112, 1081-1089
16
17
18 (19) Credendino, R.; Pater, J. T. M.; Correa, A.; Morini, G.; Cavallo, L. J. Phys. Chem. C 2011,
19
20 115, 13322-13328
21
22
23 (20) Capone, F.; Rongo, L.; D’Amore, M.; Budzelaar, P. H. M.; Busico, V. J. Phys. Chem. C
24
25 2013, 117, 24345-24353
26
27
28 (21) Kuklin, M. S.; Bazhenov, A. S.; Denifl, P.; Leinonen, T.; Linnolahti, M.; Pakkanen, T. A.
29
30 Surf. Sci. 2015, 635, 5-10.
31
32
(22) D’Amore, M.; Thushara, K. S.; Piovano, A.; Causà, M.; Bordiga, S.; Groppo, E., ACS Catal.
33
34
35 2016, 6, 5786-5796.
36
37
(23) Busico, V.; Cipullo, R.; Monaco, G.; Talarico, G.; Vacatello, M.; Chadwick, J. C.; Segre, A.
38
39
40 L.; Sudmeijer, O. Macromolecules 1999, 32, 4173-4182.
41
42 (24) Busico, V.; Corradini, P.; De Martino, L.; Proto, A.; Savino, V.; Albizzati, E., Makromol.
43
44
45 Chem. 1985, 186, 1279-1288.
46
47 (25) Busico, V.; Corradini, P.; Demartino, L.; Proto, A.; Albizzati, E., Makromol. Chem.,
48
49
50
Macromol. Chem. Phys. 1986, 187, 1115-1124.
51
52 (26) Noristi, L.; Barbè, P. C.; Baruzzi, G., Makromol. Chem. 1991, 192, 1115-1127.
53
54
55 (27) Correa, A.; Credendino, R.; Pater, J. T. M.; Morini, G.; Cavallo, L. Macromolecules 2012,
56
57 45, 3695-3701.
58
59
60 23
ACS Paragon Plus Environment
 ACS Catalysis Page 24 of 26

1
2
3 (28) Busico, V.; Cipullo, R.; Mingione, A.; Rongo, L. Ind. Eng. Chem. Res. 2016, 55, 2686-
4
5 2695.
6
7
8 (29) WO 2007 134851 to SABIC.
9
10
11
(30) For third-generation catalysts, see e.g.: (a) Zhou, Q.; Zheng, T.; Li, H.; Li, Q.; Zhang, Y.;
12
13 Zhang, L.; Hu, Y. Ind. Eng. Chem. Res. 2014, 53, 17929-17936. (b) Kissin, Y. V.; Zhou, Q.; Li, H.;
14
15 Zhang, L. J. Catal. 2015, 332, 156-163.
16
17
18 (31) Correa, A.; Piemontesi, F.; Morini, G.; Cavallo, L. Macromolecules 2007, 40, 9181-9189.
19
20 (32) Shreve, A. P.; Muelhaupt, R.; Fultz, W.; Calabrese, J.; Robbins, W.; Ittel, S. D.
21
22
23 Organometallics 1988, 7, 409-416.
24
25 (33) Ehm, C.; Antinucci, G.; Budzelaar, P. H. M.; Busico, V. J. Organomet. Chem. 2014, 772-
26
27
28 773, 161-171.
29
30 (34) Bahri-Laleh, N.; Correa, A.; Mehdipour-Ataei, S.; Arabi, H.; Haghighi, M. N.; Zohuri, G.;
31
32
Cavallo, L. Macromolecules 2011, 44, 778-783
33
34
35 (35) Kumawat, J.; Gupta, V. K.; Vanka, K. Chemcatchem 2016, 8, 1809-1818.
36
37
38 (36) Chammingkwan, P.; Thang, V.; Terano, M.; Taniike, T. Top. Catal. 1989, 57, 911-917.
39
40
41 (37) Yu, Y.; Busico, V.; Budzelaar, P. H. M.; Vittoria, A.; Cipullo, R. Angew. Chem. Int. Ed.
42
43 2016, 55, 8590-8594.
44
45
46
(38) Kissin, Y. V.; Chadwick, J. C.; Mingozzi, I.; Morini, G. Macromol. Chem. Phys. 2006, 207,
47
48 1344-1350.
49
50
(39) Brambilla, L.; Zerbi, G.; Piemontesi, F.; Nascetti, S.; Morini, G. J. Phys. Chem. C 2010,
51
52
53 114, 11475-11484.
54
55 (40) Morra, E.; Giamello, E.; Van Doorslaer, S.; Antinucci, G.; D'Amore, M.; Busico, V.;
56
57
58 Chiesa, M. Angew. Chem., Int. Ed. 2015, 54, 4857-4860.
59
60 24
ACS Paragon Plus Environment
Page 25 of 26 ACS Catalysis

1
2
3 (41) Credendino, R.; Pater, J. T. M.; Liguori, D.; Morini, G.; Cavallo, L. J. Phys. Chem. C 2012,
4
5 116, 22980-22986.
6
7
8 (42) Credendino, R.; Liguori, D.; Morini, G.; Cavallo, L. J. Phys. Chem. C 2014, 118, 8050-
9
10 8058.
11
12
13 (43) Chadwick, J. C.; van der Burgt, F.; Rastogi, S.; Busico, V.; Cipullo, R.; Talarico, G.; Heere,
14
15 J. J. R. Macromolecules 2004, 37, 9722-9727.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 25
ACS Paragon Plus Environment
 ACS Catalysis Page 26 of 26

1
2
3
4
5
6
7
8
9 TOC Graphics
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 26
ACS Paragon Plus Environment