RESEARCH ARTICLE | ENGINEERING

Threading-­the-­Needle: Compatibilization of HDPE/iPP blends

with butadiene-­derived polyolefin block copolymers
Liyang Shena,1 , Gabriela Diaz Gorbeaa,1 , Evan Danielsona , Shuquan Cuib , Christopher J. Ellisona,2 , and Frank S. Batesa,2

Edited by Jeffrey Moore, University of Illinois at Urbana-­Champaign, Urbana, IL; received January 24, 2023; accepted June 30, 2023
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

Management of the plastic industry is a momentous challenge, one that pits enormous
societal benefits against an accumulating reservoir of nearly indestructible waste. A Significance
promising strategy for recycling polyethylene (PE) and isotactic polypropylene (iPP),
constituting roughly half the plastic produced annually worldwide, is melt blending for Plastic waste constitutes a rapidly
reformulation into useful products. Unfortunately, such blends are generally brittle and growing environmental dilemma,
useless due to phase separation and mechanically weak domain interfaces. Recent studies with about half the current and
have shown that addition of small amounts of semicrystalline PE-­iPP block copolymers projected production of these
(ca. 1 wt%) to mixtures of these polyolefins results in ductility comparable to the pure vital materials centered on
materials. However, current methods for producing such additives rely on expensive polyethylene (PE) and isotactic
reagents, prohibitively impacting the cost of recycling these inexpensive commodity polypropylene (iPP). Recycling
plastics. Here, we describe an alternative strategy that exploits anionic polymerization
polyolefins is stymied by the
of butadiene into block copolymers, with subsequent catalytic hydrogenation, yielding
technical and financial challenges
E and X blocks that are individually melt miscible with PE and iPP, where E and X are
poly(ethylene-­ran-­ethylethylene) random copolymers with 6 wt% and 90 wt% ethyl- associated with separating these
ethylene repeat units, respectively. Cooling melt blended mixtures of PE and iPP con- materials in waste streams,
taining 1 wt% of the triblock copolymer EXE of appropriate molecular weight, results making melt mixing and
in mechanical properties competitive with the component plastics. Blend toughness is reformulation of the plastics into
obtained through interfacial topological entanglements of the amorphous X polymer new products an attractive
and semicrystalline iPP, along with anchoring of the E blocks through cocrystallization option. Retention of the essential
with the PE homopolymer. Significantly, EXE can be inexpensively produced using mechanical properties in phase
currently practiced industrial scale polymerization methods, offering a practical approach separated blends of PE and iPP
to recycling the world’s top two plastics. requires the use of interfacially
polyolefin | block copolymer | recycling
active block copolymers. This
report exposes an economically
Plastics contribute indispensable value to virtually all aspects of modern life, exemplified viable approach to accomplishing
by: i) high-­density polyethylene (HDPE) used in artificial hips, bullet proof vests, and this goal based on the
natural gas piping (1); ii) food packaging of all types (2); and iii) modern windmill blades polymerization and subsequent
made from thermoset polymer composites (3). Global production in 2021 of the two catalytic hydrogenation of
largest volume plastics, PE and isotactic polypropylene (iPP), was 107 million metric tons inexpensive butadiene, leading
(MMT) and 75.6 MMT, respectively, representing over half of the total synthetic polymer to block copolymers that can
market; the combined output of these polyolefins is estimated to be 230 MMT by 2029
be produced in existing
(4, 5). These inexpensive and remarkably versatile plastics offer many societal benefits, yet
industrial facilities.
contribute to a rapidly growing reservoir of nearly indestructible waste (6, 7), which ends
up in landfills and the environment, including the oceans (8). Together PE and iPP com-
prise more than two-­thirds of the average discarded plastic waste stream (9). Several
approaches for addressing this crisis are being explored (10–12), such as chemical upcycling
into feedstock chemicals (e.g., hydrogenolysis) (13–15), and reintroduction of polyolefins
into the economy through reprocessing (16–18). However, sorting polyolefin products
in a recycle stream by flotation or optical methods is largely ineffective due to similar
densities and virtually identical chemical structures (both are saturated hydrocarbons) Author contributions: L.S., G.D.G., C.J.E., and F.S.B.
designed research; L.S., G.D.G., E.D., and S.C. performed
(19). Combining most PE and iPP products for reuse through melt blending results in research; L.S., G.D.G., E.D., C.J.E., and F.S.B. analyzed
phase separation leading to brittle and essentially useless materials due to poor interfacial data; and L.S., G.D.G., C.J.E., and F.S.B. wrote the paper.
adhesion (20–22), obviating an otherwise attractive approach to recycling. The authors declare no competing interest.
Recent advances in organometallic chemistry have afforded access to PE-­iPP multiblock This article is a PNAS Direct Submission.
copolymers, shown to be remarkably effective at compatibilizing the associated commercial Copyright © 2023 the Author(s). Published by PNAS.
This article is distributed under Creative Commons
homopolymers at low concentrations (i.e., ≤1 wt%) (23, 24). Simple melt blending fol- Attribution-­NonCommercial-­NoDerivatives License 4.0
lowed by molding at elevated temperatures and cooling to ambient conditions leads to (CC BY-­NC-­ND).
solid specimens with physical properties comparable to the pristine homopolymer com- 1
L.S. and G.D.G. contributed equally to this work.
ponents, e.g., equivalent elastic modulus and high ductility with tensile strain at break εb 2
To whom correspondence may be addressed. Email:
≈ 300 to 600%; unmodified blend specimens containing 70% PE fail at εb < 20%. This cellison@umn.edu or bates001@umn.edu.
This article contains supporting information online at
strategy is significantly more efficient than competing blend modifiers that rely on com- https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.​
mercially available random copolymers such as ethylene-­propylene rubber [(25–27) and 2301352120/-­/DCSupplemental.
ethylene propylene diene rubber (28, 29)] or Dow Chemical’s INFUSE and INTUNE Published August 14, 2023.

PNAS 2023 Vol. 120 No. 34 e2301352120 https://doi.org/10.1073/pnas.2301352120 1 of 9

 olefin block copolymers (OBCs), which require larger loadings referred to as 1,4-­PB (polybutadiene). Tetrahydrofuran (THF)
(i.e., >5 wt%) to achieve comparable blend toughness. Lin et al. is added to the living polymer solution at a molar concentration
(30, 31) report the compatibilization of PE/iPP blends using a of [THF]: [Li] = 200:1 followed by addition of more butadiene
series of INFUSE OBCs. Adding 10 wt% OBC enhanced the monomer at 20 °C to produce a second 1,4-­PB/1,2-­PB statistical
tensile properties of the blends but did not fully recover the duc- copolymer block with ≈90% 1,2-­PB content, denoted 1,2-­PB.
tility of the neat PE or iPP. Wolff et al. (32) compared the mechan- Following complete conversion of the monomer to polymer, a
ical properties of a HDPE/iPP 30/70 blend compatibilized with stoichiometric amount of dichlorodimethylsilane ([Cl-­Si(CH3)2-­
INTUNE and INFUSE, and neither OBC produced tensile prop- Cl]:[Li] = 1:2) is used to couple the living diblock copolymer
erties comparable to the neat iPP at 5 wt% loading. However, chains leading to predominately 1,4-­PB-­block-­1,2-­PB-­block-­1,4-­
living block polymerization of ethylene and propylene requires PB triblock copolymer. (Alternatively, termination of the living
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

one hafnium-­pyridyl amido molecular complex per macromole- 1,4-­PB or 1,4-­PB-­block-­ 1,2-­PB chains with acidic methanol
cule, which likely makes this strategy uneconomical (23, 24, 33). results in homopolymer or diblock copolymer, respectively).
The effectiveness of PE-­iPP multiblock copolymers in recycling Saturation of the block copolymers with hydrogen at 100 °C using
PE and iPP has been attributed to two factors: 1) thermodynamic a macroporous Pt/SiO2 catalyst (or various other homogeneous
compatibility of the two block types with the homologous polyole- catalysts) generates the desired products, EXE and EX, where the E
fins leads to melt state interfacial activity in the phase separated block contains 1.5 ethyl branches per 100 backbone carbon atoms
blends; and 2) cocrystallization of the blocks with PE and iPP and the X blocks are statistical copolymers with about 90 wt%
homopolymers binds the two phases together, eliminating interfa- ethylethylene and 10 wt% ethylene repeat units. [Here, we note
cial debonding during deformation, the principal mode of failure that PB can be efficiently saturated using various heterogeneous
in the unmodified composites. Here, we demonstrate comparable and homogeneous catalysts, including those associated with
toughening of PE and iPP blends after mixing small amounts (ca. large-­scale industrial practice (34)]. Representative size exclusion
1 wt%) of a poly(ethylene)-­block-­poly(ethylene-­ran-­ethylethylene)- chromatography (SEC) and proton NMR (1H NMR) traces of
­block-­poly(ethylene) triblock copolymer, denoted EXE, with the the unsaturated PB block copolymers are provided in SI Appendix,
same commercial polyolefin homopolymers studied previously Figs. S1 and S3, respectively. We detect little homopolymer
(23, 24). EXE is obtained through anionic polymerization of buta- in the specimens (SI Appendix, Fig. S1), moreover, any such
diene followed by catalytic hydrogenation, processes currently prac- homopolymer would simply disperse into the PE domain during
ticed on an industrial scale. These findings expose a mechanism of melt processing and have no influence on the blend properties. The
interfacial strengthening, dubbed threading-­ the-­
needle, which coupling efficiency results in approximately 80 wt% of triblock
combines topological entanglement of iPP melt-­compatible and copolymer as determined using SEC, as shown in SI Appendix,
amorphous X center blocks with PE cocrystallizable terminal E Fig. S2, consistent with literature values (35–37). 1H NMR
blocks. This strategy eliminates the need to employ expensive orga- and SEC traces of the saturated products (SI Appendix, Figs. S4
nometallic catalysts, thus bringing recycling of polyolefins through and S5) demonstrate complete hydrogenation (>99%) without
blending closer to practical reality. degradation. The molecular characteristics of the saturated block
copolymers are summarized in Table 1. All three block copolymers
Results listed in Table 1 are microphase separated in the melt state up to
240 °C as evidenced through dynamic mechanical spectroscopy
Synthesis. The synthetic scheme for producing EXE triblock measurements (SI Appendix, Figs. S7 and S10).
copolymers is shown in Fig. 1. Initiation of butadiene with sec-­
butyllithium in cyclohexane at 40 °C leads to a polymer containing Blend Morphology. Polymer blends containing commercial
94 mol% 1,4-­butadiene and 6 mol% 1,2-­butadiene addition, HDPE (referred to as PE) and iPP, both provided by the Dow

Li

Li

Li
1-x x n 1-y y m

Si
1-x xn 1-y ym y 1-y m x 1-x n

Si
xn 1-y ym y 1-y m x 1-x n
1-x

Fig. 1. Synthetic scheme for the preparation of EXE triblock copolymer, where the fraction of ethylethylene repeat units in the E blocks (red) is x ≈ 0.06 and in
the X blocks (blue) is y ≈ 0.9.

2 of 9 https://doi.org/10.1073/pnas.2301352120 pnas.org
 Table 1. Molecular characteristics of EX and EXE block copolymers
wt% ethylethylene
Sample* Mn (kDa)‡ Đ (Mw/Mn) fX ¶ in X block Tg, X block (°C) # Tm, E block (°C) #
E91X93 184 1.08 50 86 −30 103
E65X88E65† 176 (218)§ 1.20 40 90 −28 104
E28X34E28† 74 (90)§ 1.18 38 91 −28 107
*
Subscripts x and y in Ex and Xy indicate number-­average molecular weights in kg/mol.
†
Subscripts based on the molecular weight of the PB diblock copolymer prior to coupling. The molecular weight of the triblock copolymer is calculated based on the molecular weight of
the diblock (SI Appendix).
‡
Absolute Mn based on SEC using universal calibration with polystyrene standards, 1H NMR, and corrected for the addition of hydrogen (SI Appendix).
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

§
Based on the molecular weight of the diblock prior to coupling. The lower molecular weight reflects about 80% coupling efficiency as shown in SI Appendix, Fig. S2.
¶
Volume fraction of X block based on 1H NMR, and assuming equal block melt densities.
#
Based on DSC measurements.

Chemical Co., along with 0.5 to 5 wt% EXE or EX, were prepared of compatibilization. Substituting the E28X34E28 for E65X88E65
using a recirculating twin screw microcompounder operated at or replacing the triblock with a relatively high molecular weight
190 °C followed by molding into 0.5-­mm-­thick films at 180 °C diblock (E91X93) at the same 1 wt% loading, results in significantly
and subsequent cooling (~20 °C/min) to room temperature. Blend reduced blend toughness (εb < 40%). Fig. 3B illustrates the role
morphologies were characterized with atomic force microscopy of block copolymer concentration on tensile toughness for the
(AFM) and scanning electron microscopy (SEM). Fig. 2 illustrates three additives. Considerable ductility (εb > 200%) is obtained
representative AFM images obtained from blends of PE and iPP with just 0.5 wt% of E65X88E65, increasing to about 800% at a
containing either 70% or 30% PE (PE:iPP = 70:30 or 30:70, concentration of 5 wt%, roughly equivalent to that of pure PE.
respectively) without (unmodified) and containing 1 wt% of Both E28X34E28 and E91X93 exhibit modest benefits (εb ≈ 200%)
E91X93, E28X34E28, or E65X88E65. Phase contrast is derived primarily at 3 wt% and about twice this value at 5 wt% loadings.
from the difference in modulus between the two components, Fig. 4A summarizes the mechanical properties of iPP continu-
where the softer PE and stiffer iPP materials appear darker and ous (30:70 PE:iPP) blends with 1 wt% of each of the three block
lighter, respectively. Melt mixing results in droplet morphologies, copolymers, along with the uncompatibilized mixture. Here again,
where the block copolymer modified blends consistently show the unmodified mixture exhibits a reduced ductility (εb ≈ 50%)
smaller domain sizes and narrower size distributions compared relative to the pure polyolefins; we attribute the increase in εb
to the neat blends (SI Appendix, Figs. S11 and S12). These results compared to the unmodified 70:30 blend (Fig. 3A) to the disper-
evidence localization of block copolymer at the phase separated sion of the more compliant PE as particles in a stiffer iPP matrix.
domain interfaces, which reduces the interfacial tension facilitating And as before, adding just 1 wt% of the E65X88E65 triblock copol-
droplet breakup during mixing. Minor variations in the domain ymer results in a strain at break (εb ≈ 550%) intermediate to the
dimensions between the PE continuous and iPP continuous pure homopolymers. As shown in Fig. 4B, the dependence of the
blends can be attributed to the well-­established effects associated strain at break on block copolymer concentration for the 30:70
with differences in the homopolymer viscosities (38). blends resembles what appears in Fig. 3B, with a couple of notable
exceptions: Diblock copolymer E91X93 promotes significantly
Mechanical Properties. Representative stress–strain curves greater toughness in the 30:70 blend at a concentration of 1 wt%,
obtained from neat PE and iPP homopolymers, and PE:iPP and E65X88E65 affords nearly twice the strain at break at 0.5 wt%
(70:30) blends with and without 1 wt% block copolymer, are as was obtained with the 70:30 PE:iPP mixture.
shown in Fig. 3A. Both homopolymers are tough, exhibiting Overall, the most striking finding of this work is that triblock
strains at break of εb ≈ 450% (iPP) and εb ≈ 750% (PE). Blending copolymer E65X88E65 performs as well in toughening the same
these two polyolefins together results in a brittle plastic with εb ≈ PE:iPP blends as the fully semicrystalline PP60PE80PP75PE90
10% (Fig. 3 A, Inset), similar to previously published results (23, tetrablock copolymer reported earlier (23).
24). Adding just 1 wt% of the E65X88E65 triblock copolymer leads Brittle fracture in uncompatibilized polyolefin blends has been
to full recovery of ductility with εb ≈ 600%, a mechanical signature attributed to interfacial failure when such composites are subjected

Neat E91X93 E28X34E28 E65X88E65

A B C D
70:30
(PE:iPP)

E F G H
30:70
(PE:iPP)

Fig. 2. Representative AFM phase images of neat PE/iPP blends (A and E), and blends containing 1 wt% E91X93 (B and F), E28X34E28 (C and G), and E65X88E65 (D and H). The
Top and Bottom rows depict PE:iPP compositions of 70:30 and 30:70, respectively. Light and dark domains correspond to the iPP and PE, respectively. Reduction
in the domain sizes with the addition of block copolymer is indicative of interfacial activity.

PNAS 2023 Vol. 120 No. 34 e2301352120 https://doi.org/10.1073/pnas.2301352120 3 of 9

 40
A (Here we note that most commercial processing involves relatively
rapid cooling rates, dT/dt > 20 °C/min). We conducted a series of
PE: iPP (70:30)+1% block copolymer
differential scanning calorimetry (DSC) experiments to establish
neat iPP the crystallization behavior of the PE and iPP homopolymers as
30 a function of cooling rate.
E65X88E65 As illustrated in SI Appendix, Fig. S20, both homopolymers crys-
Stress (MPa)

tallize (Tc) and melt (Tm) at slightly lower temperatures when cooled
neat PE more rapidly: i) dT/dt = 1 °C/min, Tc,PE = 119.9 °C, Tm,PE =
20
133.1 °C, and Tc,iPP = 128.5 °C, Tm,iPP = 164.2 °C; ii) dT/dt =
E91X93 20 °C/min, Tc,PE = 117.0 °C, Tm,PE = 130.2 °C, and Tc,iPP = 117.8 °C,
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

30

E28X34E28 Tm,iPP = 161.4 °C. These trends are also reflected in the percent
20
10 PE: iPP (70:30) crystallinity: 63% vs. 60% for PE and 58% vs. 52% for iPP at the
10 slower and faster cooling rates, respectively. We attribute the ductile
0 to brittle transition associated with iPP as the cooling rate is lowered
0
0 10 20
to the increased crystallinity.
0 200 400 600 800 In the Discussion section (see below), we propose a molecular
Strain (%) scale mechanism to explain the extraordinary toughness
B imparted to PE:iPP blends by the addition of small amounts of
1000
E28X34E28 the E65X88E65 triblock copolymer. A key aspect of this hypothesis
E65X88E65 relies on cocrystallization of the E blocks with the PE homopol-
800 E91X93
neat PE
ymer in the vicinity of the interfacial region between the phase
PE:iPP (70:30) separated domains. In order to probe such cocrystallization, two
Strain at break (%)

1,4-­PB homopolymers were synthesized and hydrogenated as

600 described in SI Appendix. We refer to these as E35 and E65 where
neat iPP

400
A
200

0
0 1 2 3 4 5 6
Block copolymer concentration (wt%)

Fig. 3. (A) Representative stress–strain curves for neat PE and iPP

homopolymers, and 70:30 blends containing 1 wt% block copolymer. The
Inset in A shows the stress–strain response of the uncompatibilized blend. (B)
Strain at break, εb, of compatibilized 70:30 PE:iPP blends as a function of block
copolymer concentration. Error bars represent one-­SD.

to large deformations. As demonstrated by the SEM image in

Fig. 5A, iPP particles appear to be cleanly separated from the PE
matrix along the fracture surface in a failed 70:30 PE:iPP blend
specimen (εb ≈ 10%). In sharp contrast, the addition of 1 wt% B
1000
E65X88E65 completely suppresses interfacial failure during tensile E28X34E28

deformation resulting in drawing of both the matrix and particle E65X88E65

E91X93
domains without delamination up to the point of fracture (εb ≈ 800 neat PE
PE:iPP (30:70)
600%) as seen in Fig. 5B.
Strain at break (%)

PEiPP

600
Crystallization. The physical properties of semicrystalline
polyolefins are intimately connected with the detailed molecular neat iPP
configurations associated with chain-­ folded crystalline 400
morphologies. These in turn are determined by several factors,
including the number of chain defects (e.g., the fraction of
branches in PE or stereochemical irregularity in iPP) and the 200
crystallization temperature, which is influenced by the rate of
cooling from the melt state. The latter point was evaluated with
0
the PE and iPP homopolymers by preparing rapidly (20 °C/min) 0 1 2 3 4 5 6
and slowly (1 °C/min) cooled specimens, followed by tensile Block copolymer concentration (wt%)
testing. The PE material exhibited a modest reduction in strain
at break at the slower cooling rate (from εb ≈ 750% to εb ≈ 550%) Fig. 4. (A) Representative stress–strain curves for neat PE and iPP
(SI Appendix, Fig. S19A). However, when cooled at 1 °C/min, homopolymers, and 30:70 blends containing 1 wt% block copolymer. The
Inset in A shows the stress–strain response of the uncompatibilized blend. (B)
the iPP plastic became brittle, failing at εb ≈ 10% in contrast to Strain at break, εb, of compatibilized 30:70 PE:iPP blends as a function of block
εb ≈ 500% when cooled at 20 °C/min (SI Appendix, Fig. S19B). copolymer concentration. Error bars represent one-­SD.

4 of 9 https://doi.org/10.1073/pnas.2301352120 pnas.org
 from the mixed melt state, with some degree of segregation in
the solid state after slow cooling.
We performed analogous solidification experiments with the
70:30 PE:iPP blend containing 1 wt% E65X88E65, guided by the
thermal behavior determined for the two homopolymers. Fig. 6
highlights the mechanical properties obtained following 4 different
cooling and heating procedures. A tough blend (εb ≈ 600%) is
obtained in the limit of fast cooling (20 °C/min) from the melt
(180 °C) as described earlier. Reducing the cooling rate to 1 °C/min
has a profound impact on the product, decreasing the strain at
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

break to εb ≈ 25%. A third thermal history involved cooling from

180 °C at 1 °C/min (slow cool) to 125 °C followed by cooling at
20 °C/min (fast cool) to room temperature, designated “slow
cooled iPP/fast cooled PE” in Fig. 6. This protocol, designed to
produce slowly crystallized iPP particles dispersed in a PE melt at
125 °C, with subsequent crystallization of PE (and E) while cooling
rapidly, results in a tough material, albeit with a somewhat reduced
strain at break, εb ≈ 350%. In a fourth procedure, the entire blend

A
200
Slow cool - 1 °C/min
Fast cool - 20 °C/min
180

Temperature (C) 160

140

Fig. 5. SEM images obtained from cryo-­fractured cross-­sections of (A) PE:iPP 120
(70:30) blend following failure in tension at εb ≈ 10%, and (B) PE:iPP (70:30) slow cooled iPP
slow cooled PE
blend containing 1 wt% E65X88E65 after failure at εb ≈ 600%. 100

80 fast cooled iPP fast cooled iPP

the molecular weights were chosen to be similar to the E blocks fast cooled PE slow cooled iPP slow cooled PE
fast cooled PE
in the two triblock copolymers (Table 1). DSC measurements 60
show that these polymers have about 36% crystallinity with 0 20 40 60 80 100
Time (min)
melting temperatures of Tm,E = 108 to 110 °C, following cooling
from the melt at 20 °C/min, i.e., slightly higher than the melting B
temperatures of the E blocks in the pure block copolymers 40
PE:iPP (70:30)+1% E65X88E65
(Table 1). The difference in percent crystallinity and melting
temperature Tm,PE -­ Tm,E ≈ 21 °C is attributable to the 1.5 ethyl
branches per 100 backbone carbon atoms in the E polymer,
30
which can result in segregation during melt crystallization in Fast cool
mixtures with the PE homopolymer. DSC experiments with (20 °C/min)
Stress (MPa)

iPP fast cool (20 °C/min)

blends of E35 and E65 containing PE demonstrate that these PE slow cool (1 °C/min)
polymers largely cocrystallize when cooled from the melt at 20
20 °C/min, as evidenced by a depression of the principle melting iPP slow cool (1 °C/min)
temperature to Tm = 126 to 128 °C, and reduction in the percent Slow cool PE fast cool (20 °C/min)
crystallinity to 48 to 50%, in mixtures containing 50% of either (1 °C/min)
E polymer; a minority fraction (<50%) of E segregates during 10
solidification as indicated by a broad second melting endotherm
at 108 to 110 °C in these blends (SI Appendix, Fig. S17).
Mixtures containing 25% E and 75% PE exhibit almost no
evidence of segregation after crystallization. Cooling these mix- 0
tures slowly (1 °C/min) somewhat increases segregation between 0 200 400 600
Strain (%)
E and PE during crystallization as evidenced by a slightly higher
principal peak melting temperature (129 to 130 °C) and a more Fig. 6. (A) Schematic illustration of four thermal processing strategies, where
distinct lower melting temperature (≈ 108 °C) endotherm in slow cool ≅1 °C/min and fast cool ≅20 °C/min. 1) Slow cool iPP/slow cool PE
the 50/50 PE/E mixtures (SI Appendix, Fig. S18). Wignall et al. from 180 to 30 °C (red/black lines). 2) Fast cool iPP/fast cool PE from 180 to
30 °C (teal/blue lines). 3) Slow cool iPP/fast cool PE; slow cool from 180 to
reported that slowly cooling induces the formation of separate 125 °C (red line) followed by fast cool from 125 to 30 °C (red line). 4) Fast cool
crystalline phases in HDPE/low-­density PE blends based on iPP/slow cool PE; fast cool from 180 to 100 °C (teal line) followed by heating
small-­angle neutron and X-­ray scattering and DSC measure- from 100 to 135 at 10 °C/min (teal line), isothermal annealing for 5 min at
135 °C (teal line) then slow cooling from 135 to 30 °C. (B) The representative
ments (39). These experiments show that, to a considerable stress–strain curves obtained from PE: iPP (70:30) blends with 1 wt. % E65X88E65
extent, E35 and E65 cocrystallize with PE when cooled rapidly following different cooling recipes from the melt state.

PNAS 2023 Vol. 120 No. 34 e2301352120 https://doi.org/10.1073/pnas.2301352120 5 of 9

 was solidified by fast cooling from 180 °C to 100 °C (i.e., well spectroscopy (DMS). Both pure EXE triblocks and the EX diblock
below the crystallization temperature of both PE and iPP), then display nonterminal low-­frequency elastic (G′) and loss (G′′)
was heated to 135 °C for 5 min (leaving fast cooled crystalline iPP dynamic moduli (SI Appendix, Figs. S7 and S10) indicative of
dispersed in a PE melt), followed by slow cooling to room tem- microphase separation up to 240 °C. Transport of the block copol-
perature. This process, referred to as “fast cooled iPP/slow cooled ymers to the interface between the phase separated PE and iPP
PE” in Fig. 6, results in a relatively brittle plastic with εb ≈ 30%. domains during melt blending is therefore driven by thermody-
These surprisingly different materials are discussed further in the namic compatibility of E and PE, and X and iPP.
following section in the context of the “threading-­the-­needle” Interfacial activity of the EXE and EX block copolymers serves
mechanism. two purposes: 1) It reduces the interfacial tension between PE and
iPP in the melt state, similar to the behavior of a surfactant in
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

Discussion contact with water and oil; and 2) the localization of block copol-
ymers at domain interfaces provides varying degrees of adhesion
Addition of a small amount (ca. 1 wt%) of an appropriately pre- between the solid (semicrystalline) domains. Reduced interfacial
pared EXE triblock copolymer to mixtures of commercially avail- tension is evident in the reduction in domain sizes relative to the
able PE and iPP leads to remarkably tough materials, offering a unmodified mixtures apparent in the AFM images found in Fig. 2;
promising approach to recycling these leading plastics. At first well-­established theory and experiments anticipate this tradeoff
blush, this finding is surprising because the X block, which is an between particle size and interfacial tension under the influence
amorphous polymer with a glass transition temperature Tg,X = of shearing and extensional flows in the two-­phase liquid systems
−30 °C, cannot cocrystallize with iPP. The results presented in the (45, 46). To a reasonable approximation, diblock and triblock
previous section offer important clues as to the mechanism respon- copolymers should function similarly in this respect, where com-
sible for this intriguing and enabling behavior. peting factors set the optimal molecular weight M; increasing M
Polyolefin thermodynamics involves purely dispersive van der leads to enhanced steric stabilization but promotes micelle forma-
Waals interactions, hence there is little if any heat of mixing asso- tion, while reducing M ultimately results in miscibility of the block
ciated with PE and iPP melt blends; note that the densities of copolymer with the homopolymers (47, 48). This aspect of EXE
these chemical isomers in the melt state are nearly identical. and EX compatibilization of PE and iPP warrants additional
Previous theoretical and experimental research has attributed the research. Here, we focus on interfacial adhesion, which is respon-
phase separation of polyolefins to the excess entropy of mixing sible for achieving tough blends.
associated with differences in the conformational states of satu- EXE triblock and EX diblock copolymers will be configured at
rated hydrocarbon polymers with differing branch types and com- the domain interfaces in the melt state as sketched in Fig. 7A. For
positions. According to the conformational asymmetry theory the triblock molecular architecture, the X block will form a loop
proposed by Bates and Fredrickson (40–42), miscibility between that extends into the liquid iPP domain. Assuming Gaussian sta-
different polyolefins can be obtained by matching the associated tistics, the radius of gyration of a flexible polymer is Rg = b(N/6)1/2,
statistical segment lengths at a common segment volume. Random where N is the degree of polymerization and b is the statistical
copolymers of ethylene and ethylethylene, containing about 90 segment length (40, 41). For the X block (and iPP based on a
wt% of the branched repeat units, denoted X, are well established 4-­carbon repeat unit), bX = biPP = 0.58 nm (43), hence the X loop
to be melt-­miscible with iPP (i.e., the Flory–Huggins interaction will project roughly D = 2Rg into the iPP domain; for E65X88E65
parameter χ ≅ 0) owing to matched statistical segment lengths and E28X34E28, DX ≈ 19 nm and 12 nm, respectively. Similarly,
defined at a common segment volume (43). Following earlier work the E blocks will project into the PE melt roughly DE ≈ 22 nm
by Chaffin et al. (44), we designed the EXE and EX block copol- and 15 nm, respectively, based on bE = 0.80 nm (also with a
ymers with this in mind. A key factor in applying these com- 4-­carbon repeat unit) (40, 41). Once formed, iPP chains will
pounds to compatibilization of PE and iPP is the segregation of rapidly penetrate (“thread”) the X loop (“needle”) in order to main-
the E and X blocks, which was verified using dynamic mechanical tain constant density, creating a state of entanglement with the

Fig. 7. Schematic illustration of proposed mechanisms for block copolymer interfacial behavior in the (A) melt state, and (B–D) solid state. Fast cooling from the
melt results in cocrystallization of E and PE, and entanglement of X with the amorphous portion of semicrystalline iPP (B and C). A triblock architecture leads to
topologically entrained X blocks and high interfacial strength (C). Diblocks present entangled but unconstrained X chains that can disengage from the interface
under stress (B). Slow cooling (D) results in semicrystalline E blocks segregated from the PE morphology leading to poor interfacial strength.

6 of 9 https://doi.org/10.1073/pnas.2301352120 pnas.org
 homopolymer. (For the iPP melt considered here, the melt molec- chain-­folded and bridging portions of the amorphous region of
ular diffusion (reptation) time is τr < 1 s). Estimating the X loop the semicrystalline iPP, creating topological crosslinks that stitch
configurational volume as VX = (4/3)πRg3 = 3.1 × 103 nm3 with a the block copolymer to the solid iPP domain. We anticipate that
melt density of ρ = 0.9 g/cm3 yields roughly 10 to 20 iPP chains this mechanism will be operative at considerably lower EXE
within a loop volume for E65X88E65, based on Mn, iPP = 1 × 105 g/mol molecular weights (see below). However, the E block must be
and assuming the relevant metric is the number average molecular considered as well.
weight. Mn provides a conservative estimate of the needle size; the Fig. 6 also reveals that slowly cooling the blends leads to poor
X blocks are relatively narrow in dispersity so that this distinction ductility. Most tellingly, fast cooling to 100 °C, with crystallization
is not very important. More significant are the molecular weights of PE and iPP, followed by melting of the PE (but not the iPP),
and dispersity of the iPP chains in estimating how many of these and then slow crystallization of PE renders a material with εb ≈
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

polymers are entrained in a needle. We believe the most relevant 30%, which is almost indistinguishable from blends cooled con-
molecular weight for this purpose is Mn, as we calculate the num- tinuously from 180 °C to room temperature at ~1 °C/min.
ber of iPP polymers contained in an X block coil volume. (Clearly Clearly, the critical step of E and PE cocrystallization near the
these are crude calculations intended to provide a qualitative sense domain interface is rate dependent (50). As mentioned previously,
of the proposed mechanism). Since the entanglement molecular the extent of cocrystallization between E and PE will be smaller
weight for iPP (and the X blocks) is Me = 6,300 g/mol (49), the with slow cooling compared to fast cooling from the mixed melt
homopolymer chains associated with the X loop will be fully state. Moreover, there appears to be an E block molecular weight
entangled. Maintaining the entropically favorable Gaussian coil dependence. We believe the poor performance of E28X34E28 at low
configuration necessitates penetration of many iPP chains within concentrations (Figs. 3 and 4) is due to the reduction from 65
the associated coil volume. Upon cooling, nucleation and growth kDa to 28 kDa in the E block molecular weight. These results
of a chain-­folded semicrystalline morphology will require local point to the need to better understand how E and PE engage
separation of the iPP and X chains, since the latter cannot crys- during crystallization in the vicinity of the domain interfaces.
tallize. We speculate that the X loops will be entrained by the loops Overall, the interfacial activity of the EXE triblock is not
and bridging portions of the amorphous part of the semicrystalline dependent on the blend composition, i.e., 70:30 PE:iPP vs. 30:70
iPP structure, creating topological constraints that bind the PE:iPP, in both limits acting as a macromolecular surfactant that
E65X88E65 triblock copolymer to the iPP domain (Fig. 7 C and lowers the interfacial tension while mechanically coupling the two
D). An EX diblock architecture does not afford the same type of domains through cocrystallization of E and PE, and topological
entrainment since the X block is immobilized only at one end (i.e., entanglement of X and iPP. The curious behavior of the E91X93
the EX block junction in Fig. 7B) and can escape confinement by diblock copolymer in the 30:70 PE:iPP blends (Fig. 4) warrants
wiggling out of the iPP entanglements in response to an applied an additional comment. At 1 wt% loading, the fast-­cooled mixture
load. However, the diblock will be interfacially active as evidenced exhibits εb ≈ 325%, yet this block copolymer imparts no added
by the reduction in domain size shown for the blends containing toughness beyond the pure homopolymer blend at 0.5 wt%. We
E91X93 (Fig. 2 B and F). Such interfacial activity will not interfere suspect this reflects an increased E block molecular weight, com-
with the “threading the needle” mechanism associated with the bined with the role played by embedding lower modulus PE par-
EXE triblock copolymer containing minor amounts (ca. 20%) of ticles in a stiffer iPP matrix, clearly reflected in the unmodified
uncoupled diblock copolymer. mixture (Fig. 4 A, Inset). Presumably, interfacial failure occurs
In order to support a stress across the interface, the E blocks through retraction and delamination of the X blocks, which are
also must be bound to the semicrystalline PE domains. Here, we probably highly entangled in the amorphous region of the semic-
invoke cocrystallization of E and PE, shown above to occur in rystalline iPP domain.
blends of the two homopolymers. We speculate that upon rapid Finally, we reflect on the threading-­the-­needle concept in the
cooling a fraction of the E blocks are immobilized through crys- light of a recent report of remarkable toughness obtained by add-
tallization with the homopolymer. This may involve actual mixing ing low concentrations of dihydroxy PE (HO-­PE-­OH) to blends
of crystalline E and PE stems within common lamellae (Fig. 7 B of PE and poly(ethylene terephthalate) (PET) (51). Mixing just
and C) or could result from entrained chain folded loops and 0.5 wt% of HO-­PE-­OH with molecular weights ranging from 1
bridges between separate crystalline E lamellae and PE lamellae. to 20 kDa to a 80:20 PET:PE blend and melt processing at 270 °C
What is essential is that a critical fraction of the E blocks are fixed for 5 min produced composites with εb ≈ 300%; the unmodified
within the PE domain in order to withstand the forces created at blends were characterized by εb < 50%. This processing procedure
the domain interface during deformation. drives the formation of PET-­PE-­PET triblock copolymers through
The most essential feature of the proposed toughening mecha- transesterification at the interface between the two homopolymers.
nism is that iPP chains “thread-­the-­needle” formed by the looping Ostensibly, this finding seems analogous to what is reported here.
X block in the melt state, which becomes topologically trapped However, the underlying “hook-­ and-­
clasp” mechanism for
upon crystallization of the homopolymer. Anchoring of the E PET:PE blends is decidedly different than threading-­the-­needle.
blocks then results in interfacial adhesion. Support for this hypoth- Facile reaction of a terminal HO-­PE-­OH hydroxyl group with a
esis is provided by the experimental results, especially in the limit PET chain generates a diblock copolymer that is “hooked” to the
of low concentrations of block copolymer (≤1 wt%). narrow (<1 nm) interface associated with the strongly phase sep-
Here, we focus on the findings shown in Fig. 6. Superior duc- arated blend. Proximity of the hydroxyl moiety of the resulting
tility (εb > 300%) was obtained when PE crystallization occurred PET-­PE-­HO diblock to additional PET at the interface facilitates
under fast cooling conditions, regardless of whether the iPP was a second transesterification reaction leading to a PET-­PE-­PET
cooled quickly or slowly. This implies that the threading-­the-­needle triblock copolymer. The resulting “clasping” process captures PE
mechanism is not rate dependent. Disengagement of iPP during homopolymer chains, and subsequent crystallization of PE and
crystallization could occur if folding individual chains into grow- PET upon cooling provides interfacial adhesion. A Mn = 1 kDa
ing lamellae drew them out of the X loop. However, this would PE chain will have VPE ≈ 15 nm3 with roughly 85% of the space
require collapse of the X loop, which would be entropically costly. available for clasping high molecular weight PE homopolymer.
We believe that iPP crystallization captures the X loop in multiple Unlike threading-­the-­needle, the hook-­and-­clasp mechanism does

PNAS 2023 Vol. 120 No. 34 e2301352120 https://doi.org/10.1073/pnas.2301352120 7 of 9

 not require transporting block copolymer to the interface, nor high-­temperature SEC in trichlorobenzene at 135 °C using an Agilent PL-­220
diffusion of PE homopolymer; the clasping step is driven by the instrument equipped with a refractive index detector to confirm a lack of chain
transesterification reaction, greatly enhanced by proximity to the degradation, and 1H NMR spectroscopy was performed with a Bruker HD500
interface created by diblock formation. As a consequence, even instrument at 90 °C using deuterated toluene solutions to establish the extent
highly compositionally asymmetric triblocks, can efficiently com- of polymer saturation.
patibilize PE and PET; preformed PET-­PE-­PET triblock copoly- Blend Preparation and Tensile Test. Blends of PE, iPP, and 0.5 to 5 wt%
mer would not be expected to be as effective (52). block copolymers (weight fraction based on the weight of neat PE/iPP blends)
were prepared using a recirculating 5 mL DSM Xplore twin-­screw microcom-
Conclusion pounder, with mixing for 8 min at 130 rpm at 190 °C. The blended materials
were molded into 0.5-­mm-­thick films at 180 °C employing a pressure of 4 MPa
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

Repurposing PE and iPP through blending requires economically for 5 min with a Carver hot press. Unless otherwise stated, cooling water was
tractable approaches to combining these plastics without sacrific- used for quenching (~20 °C/min). Dumbbell-­shaped tensile bars were prepared
ing mechanical properties. We have demonstrated that adding with a die cutter (ASTM D1708, 5-­mm gauge width, 22-­mm gauge length). All
only 1 wt% of EXE triblock copolymers, containing PE and iPP tensile tests were conducted at room temperature (22 °C) using an Instron
melt miscible E and X blocks, generates ductile blends with a 5966 Universal Testing System operated at a crosshead speed of 22 mm/min
strain at break of εb ≈ 600%. The synthetic methods for obtaining (100%/min strain rate).
this relatively inexpensive block copolymer, anionic polymeriza-
tion of butadiene followed by catalytic hydrogenation, are cur- AFM. The morphology of neat and compatibilized PE:iPP blends were imaged

rently practiced on an industrial scale. Mechanically superior using AFM in dynamic mode (AFM; Bruker Nanoscope V Multimode 8, Digital
blends result from interfacial localization of the EXE triblock Instruments Santa Barbara, CA open-­loop system). Smooth imaging surfaces were
obtained on pressed and annealed films using a cryo-­ultramicrotome (Leica UC6)
copolymer during melt mixing, which leads to X chains that are
operated at −120 °C, first using a glass knife to create a cutting face, followed by
topologically entrained with semicrystalline iPP and E blocks that
sectioning of 500-­nm-­thick slices with a diamond knife (Diatome), which were
cocrystallize with the PE homopolymer upon rapid cooling. This mounted on a silicon wafer. Samples were scanned in the repulsive regime using
strategy brings the possibility of recycling PE and iPP through an n-­type silicon tip cantilever (resonant frequency = 166 Hz, spring constant =
blending and reuse closer to feasibility. 2 N/m, and radius = 8 nm). Captured images were processed using Gwyddion
2.56 open-­source software to level the data, align rows, correct scarring, and
Materials and Methods adjust the contrast via histogram. Details regarding data handling are provided
in SI Appendix.
Materials. Homopolymers iPP (H314-­02Z, Mn = 100 kg/mol, Ð = 4.1, MFI =
2.0 g/10 min at 230 °C with 2.16 kg) and HDPE (DMDA-­8904, Mn = 22 kg/mol, Fractography. Dogbone tensile specimens were cryo-­fractured in liquid
Ð = 3.8, MFI = 4.4 g/10 min at 190 °C with 2.16 kg) were both obtained from nitrogen, and the resulting cross sections were examined using a JEOL 6500
the Dow Chemical Company. A series of E-­X block copolymers were synthesized field emission SEM with 2 kV accelerating voltage and approximately 10-­mm
using sequential anionic polymerization followed by catalytic hydrogenation; working distance. Specimens were affixed with carbon tape to a 90-­degree pin
detailed descriptions of the procedures have been reported previously (53). stub mount and sputter coated with a 5-­nm-­thick platinum conducting layer
Cyclohexane (HPLC, Fisher Scientific) and THF (HPLC, Fisher Scientific) were before imaging.
purified by passing through activated alumina columns. Butadiene (≥99%,
Sigma-­Aldrich) was twice distilled from n-­butyllithium (2.5 M in hexanes, Sigma-­ Thermal Analysis. Glass transition, melting, and crystallization temperatures
Aldrich). Cyclohexane was first added to the reactor under an argon atmosphere, were determined using DSC. 5 to 10 mg of sample was sealed in an aluminum
followed by sec-­butyllithium (sBuLi, 1.4 M in cyclohexanes, Sigma-­Aldrich) and pan and loaded in a TA Q1000 DSC instrument under a nitrogen atmosphere at
butadiene, which was allowed to react for 8 h at 40 °C. An aliquot of the living gas flow rate 50 mL/min. Calorimetry data was acquired while heating or cooling
polymer was taken to determine the average molecular weight and dispersity. the specimen at 10 °C/min.
Then the reactor was cooled to 20 °C, and THF was added at a concentration Rheology. Bulk rheological data were acquired for the three saturated block copol-
of [THF]:[Li] = 200:1, followed by additional butadiene. After an additional ymers (E91X93, E28X34E28, and E65X88E65) using an ARES-­G2 rheometer (Thermal
8 h, an aliquot of the diblock was removed from the reactor for analysis, and Analysis Instruments, New Castle, DE) under nitrogen gas purge employing an
a stoichiometric amount of dimethyldichlorosilane (≥99.5%, Sigma-­Aldrich) 8-­mm parallel plate geometry and a 0.5-­mm gap. Frequency sweeps spanning
was added ([Cl]:[Li] = 1:1) and allowed to react for 3 d at room temperature. 0.1 to 100 rad/s at a constant strain amplitude of 2% were conducted from 120
PB homopolymers and diblock copolymers were prepared by terminating living to 240 °C in 20 °C increments with 10 min between measurements for temper-
polymers following the first and second stages using degassed methanol. The ature equilibration. Master curves, referenced to 180 °C, were prepared using
product was precipitated in cold methanol and dried under vacuum at 40 °C time–temperature superposition.
to constant weight.
PB polymers were dissolved in isooctane (≥99%, Fisher Scientific) at a con- Data, Materials, and Software Availability. All data are included in the man-
centration of 5 g/L and hydrogenated to E homopolymers, EX diblock, and EXE uscript and supporting information. Data files are openly available in the Data
triblock copolymers, in a high-­pressure reactor operated at 100 °C with 500 psi Repository for the University of Minnesota (DRUM) at https://conservancy.umn.
of H2 over a SiO2 supported Pt catalyst [1:5 catalyst-­to-­polymer (w/w)] for 24 h edu/drum (56).
(54). Catalyst was removed by hot filtration and the product was precipitated in
cold methanol and dried under vacuum at 100 °C to constant weight. ACKNOWLEDGMENTS. Funding for this work was provided by the Center
The molecular weight and dispersity of the PB compounds were determined for Sustainable Polymers, a NSF-­supported Center for Chemical Innovation
using room temperature SEC at a concentration of 3 to 5 mg/mL with THF as the (CHE-­1901635). Parts of the work were carried out in the Characterization Facility,
mobile phase and calibrated with polystyrene standards. The eluent flow rate University of Minnesota, which receives partial support from the NSF through the
is 1 mL/min, and the sample injection volume is 100 μL. The Mark–Houwink MRSEC (Award Number DMR-­2011401) and the NNCI (Award Number ECCS-­
parameters used for the universal calibration are KPB = 2.52 × 10−2 mL/g, 2025124) programs.
αPB = 0.727, KPS = 8.63 × 10−3 mL/g, and αPS = 0.736 (55). The composition
of the 1,4-­PB and 1,2-­PB blocks were determined by 1H NMR spectra obtained
Author affiliations: aDepartment of Chemical Engineering and Materials Science, University
from 10% (w/w) CDCl3 solutions at 30 °C using a Bruker HD500 NMR spec- of Minnesota, Minneapolis, MN 55455; and bDepartment of Chemistry, University of
trometer. The PB precursors and the hydrogenated products were examined by Minnesota, Minneapolis, MN 55455

8 of 9 https://doi.org/10.1073/pnas.2301352120 pnas.org
 1. K. Patel, S. H. Chikkali, S. Sivaram, Ultrahigh molecular weight polyethylene: Catalysis, structure, 31. Y. Lin, G. R. Marchand, A. Hiltner, E. Baer, Adhesion of olefin block copolymers to polypropylene
properties, processing and applications. Prog. Polym. Sci. 109, 101290 (2020). and high density polyethylene and their effectiveness as compatibilizers in blends. Polymer 52,
2. S. Billiet, S. R. Trenor, 100th anniversary of macromolecular science viewpoint: Needs for plastics 1635–1644 (2011).
packaging circularity. ACS Macro Lett. 9, 1376–1390 (2020). 32. P. Wolff, A. Dickert, W. P. Kretschmer, R. Kempe, iPP/PE multiblock copolymers for plastic blend
3. L. Thomas, M. Ramachandra, Advanced materials for wind turbine blade–A review. Mater. Today Proc. recycling synthesized by coordinative chain transfer polymerization. Macromolecules 55,
5, 2635–2640 (2018). 6435–6442 (2022).
4. Statista, Polyethylene market volume worldwide 2015–2029 (12 January 2023). 33. K. Klimovica et al., Compatibilization of iPP/HDPE blends with PE-­g-­iPP graft copolymers. ACS Macro
5. Statista, Polypropylene global market volume 2015–2029 (12 January 2023). Lett. 9, 1161–1166 (2020).
6. J. Nakatani, T. Maruyama, Y. Moriguchi, Revealing the intersectoral material flow of plastic 34. N. K. Singha, S. Bhattacharjee, S. Sivaram, Hydrogenation of diene elastomers, their properties and
containers and packaging in Japan. Proc. Natl. Acad. Sci. U.S.A. 117, 19844–19853 (2020). applications: A critical review. Rubber Chem. Technol. 70, 309–367 (1997).
7. K. L. Law et al., The United States’ contribution of plastic waste to land and ocean. Sci. Adv. 6, 35. L. J. Fetters, B. H. Meyer, D. McIntyre, The effect of diblock and homopolymer impurities on the
eabd0288 (2020). morphology of triblock polymers. J. Appl. Polym. Sci. 16, 2079–2089 (1972).
8. Y. Peng, P. Wu, A. T. Schartup, Y. Zhang, Plastic waste release caused by COVID-­19 and its fate in the 36. D. N. Schulz, A. F. Halasa, Anionic polymerization initiators containing protected functional groups.
Downloaded from https://www.pnas.org by PENNSYLVANIA STATE UNIVERSITY PSU ELECTRONIC RESOURCES UNIT on September 26, 2023 from IP address 128.118.7.115.

global ocean. Proc. Natl. Acad. Sci. U.S.A. 118, e2111530118 (2021). II. J. Polym. Sci. Polym. Chem. Ed. 15, 2401–2410 (1977).
9. O. US EPA, Advancing sustainable materials management: 2018 fact sheet (2018) (23 August 2022). 37. J. C. Cho et al., Microstructure effect in the coupling reactions of polymeric organolithium
10. H. Mangold, B. von Vacano, The frontier of plastics recycling: Rethinking waste as a resource for compounds using chlorosilanes. J. Polym. Sci. A Polym. Chem. 36, 1743–1753 (1998).
high-­value applications. Macromol. Chem. Phys. 223, 2100488 (2022). 38. S. Jose et al., Phase morphology, crystallisation behaviour and mechanical properties of isotactic
11. J. M. Garcia, M. L. Robertson, The future of plastics recycling. Science 358, 870–872 (2017). polypropylene/high density polyethylene blends. Eur. Polym. J. 40, 2105–2115 (2004).
12. L. Dai, R. Ruan, S. You, H. Lei, Paths to sustainable plastic waste recycling. Science 377, 934–934 (2022). 39. G. D. Wignall et al., Morphology of blends of linear and long-­chain-­branched polyethylenes in the
13. C. Jehanno et al., Critical advances and future opportunities in upcycling commodity polymers. solid state: A study by SANS, SAXS, and DSC. Macromolecules 28, 3156–3167 (1995).
Nature 603, 803–814 (2022). 40. F. S. Bates, G. H. Fredrickson, Conformational asymmetry and polymer-­polymer thermodynamics.
14. Z. Xu et al., Cascade degradation and upcycling of polystyrene waste to high-­value chemicals. Proc. Macromolecules 27, 1065–1067 (1994).
Natl. Acad. Sci. U.S.A. 119, e2203346119 (2022). 41. F. S. Bates, M. F. Schulz, J. H. Rosedale, K. Almdal, Correlation of binary polyolefin phase behavior
15. G. W. Coates, Y. D. Y. L. Getzler, Chemical recycling to monomer for an ideal, circular polymer with statistical segment length asymmetry. Macromolecules 25, 5547–5550 (1992).
economy. Nat. Rev. Mater. 5, 501–516 (2020). 42. G. H. Fredrickson, A. J. Liu, F. S. Bates, Entropic corrections to the flory-­huggins theory of polymer
16. Z. O. G. Schyns, M. P. Shaver, Mechanical recycling of packaging plastics: A review. Macromol. Rapid blends: Architectural and conformational effects. Macromolecules 27, 2503–2511 (1994).
Commun. 42, 2000415 (2021). 43. P. A. Weimann et al., Phase behavior of isotactic polypropylene−Poly(ethylene/ethylethylene)
17. S. Yin et al., Mechanical reprocessing of polyolefin waste: A review. Polym. Eng. Sci. 55, 2899–2909 random copolymer blends. Macromolecules 30, 3650–3657 (1997).
(2015). 44. K. A. Chaffin, F. S. Bates, P. Brant, G. M. Brown, Semicrystalline blends of polyethylene and isotactic
18. D. Jubinville, E. Esmizadeh, S. Saikrishnan, C. Tzoganakis, T. Mekonnen, A comprehensive review of polypropylene: Improving mechanical performance by enhancing the interfacial structure. J. Polym.
global production and recycling methods of polyolefin (PO) based products and their post-­recycling Sci. B Polym. Phys. 38, 108–121 (2000).
applications. SM&T 25, e00188 (2020). 45. C. W. Macosko et al., Compatibilizers for melt blending: Premade block copolymers.
19. B. D. Vogt, K. K. Stokes, S. K. Kumar, Why is recycling of postconsumer plastics so challenging? ACS Macromolecules 29, 5590–5598 (1996).
Appl. Polym. Mater. 3, 4325–4346 (2021). 46. S. Lyu, T. D. Jones, F. S. Bates, C. W. Macosko, Role of block copolymers on suppression of droplet
20. A. Graziano, S. Jaffer, M. Sain, Review on modification strategies of polyethylene/polypropylene coalescence. Macromolecules 35, 7845–7855 (2002).
immiscible thermoplastic polymer blends for enhancing their mechanical behavior. J. Elastomers 47. A. Adedeji, S. Lyu, C. W. Macosko, Block copolymers in homopolymer blends: Interface vs micelles.
Plast. 51, 291–336 (2019). Macromolecules 34, 8663–8668 (2001).
21. J. W. Teh, A. Rudin, J. C. Keung, A review of polyethylene–polypropylene blends and their 48. J. L. Self et al., Linear, graft, and beyond: Multiblock copolymers as next-­generation compatibilizers.
compatibilization. Adv. Polym. Technol. 13, 1–23 (1994). JACS Au 2, 310–321 (2022).
22. K. A. Chaffin, J. S. Knutsen, P. Brant, F. S. Bates, High-­strength welds in metallocene polypropylene/ 49. A. Eckstein et al., Determination of plateau moduli and entanglement molecular weights
polyethylene laminates. Science 288, 2187–2190 (2000). of isotactic, syndiotactic, and atactic polypropylenes synthesized with metallocene catalysts.
23. J. M. Eagan et al., Combining polyethylene and polypropylene: Enhanced performance with PE/iPP Macromolecules 31, 1335–1340 (1998).
multiblock polymers. Science 355, 814–816 (2017). 50. M. J. Galante, L. Mandelkern, R. G. Alamo, The crystallization of blends of different types of
24. J. Xu et al., Compatibilization of isotactic polypropylene (iPP) and high-­density polyethylene (HDPE) polyethylene: The role of crystallization conditions. Polymer 39, 5105–5119 (1998).
with iPP–PE multiblock copolymers. Macromolecules 51, 8585–8596 (2018). 51. A. J. Zervoudakis et al., Dihydroxy polyethylene additives for compatibilization and mechanical
25. L. D’Orazio et al., Effect of the addition of ethylene-­propylene random copolymers on the properties recycling of polyethylene terephthalate/polyethylene mixed plastic waste. ACS Macro Lett. 11,
of high-­density polyethylene/isotactic polypropylene blends: Part 1—Morphology and impact 1396–1402 (2022).
behavior of molded samples. Polym. Eng. Sci. 22, 536–544 (1982). 52. K. Nomura et al., Multiblock copolymers for recycling polyethylene–Poly(ethylene terephthalate)
26. L. D’Orazio, R. Greco, E. Martuscelli, G. Ragosta, Effect of the addition of EPM copolymers on the mixed waste. ACS Appl. Mater. Interfaces 12, 9726–9735 (2020).
properties of high density polyethylene/isotactic polypropylene blends: II. Morphology and 53. S. Ndoni, C. M. Papadakis, F. S. Bates, K. Almdal, Laboratory-­scale setup for anionic polymerization
mechanical properties of extruded samples. Polym. Eng. Sci. 23, 489–497 (1983). under inert atmosphere. Rev. Sci. Instrum. 66, 1090–1095 (1995).
27. M. Louizi, V. Massardier, P. Cassagnau, Contribution of high-­shear processing to the 54. S. P. Ertem et al., Hydrogenolysis of linear low-­density polyethylene during heterogeneous catalytic
compatibilization of (PP/EPR)/PE ternary blends. Macromol. Mater. Eng. 299, 674–688 (2014). hydrogen-­deuterium exchange. Macromolecules 53, 6043–6055 (2020).
28. T. Nedkov, F. Lednický, Morphologies of polyethylene–ethylene/propylene/diene monomer particles 55. C. Jackson, Y.-­J. Chen, J. W. Mays, Size exclusion chromatography with multiple detectors: Solution
in polypropylene-­rich polyolefin blends: Flake structure. J. Appl. Polym. Sci. 90, 3087–3092 (2003). properties of linear chains of varying flexibility in tetrahydrofuran. J. Appl. Polym. Sci. 61, 865–874
29. M. Hemmati, H. Nazokdast, H. Shariat Panahi, Study on morphology of ternary polymer blends. I. (1996).
Effects of melt viscosity and interfacial interaction. J. Appl. Polym. Sci. 82, 1129–1137 (2001). 56. L. Shen et al., Data for “Threading-­the-­Needle: Compatibilization of HDPE/iPP blends with
30. Y. Lin, V. Yakovleva, H. Chen, A. Hiltner, E. Baer, Comparison of olefin copolymers as compatibilizers butadiene-­derived polyolefin block copolymers”. The Data Repository for University of Minnesota.
for polypropylene and high-­density polyethylene. J. Appl. Polym. Sci. 113, 1945–1952 (2009). https://hdl.handle.net/11299/255641. Deposited 31 July 2023.

PNAS 2023 Vol. 120 No. 34 e2301352120 https://doi.org/10.1073/pnas.2301352120 9 of 9