Ziegler-Natta Polymerization

Why Ziegler-Natta matters

• Made polyolefins economic. Pre-1953, polyethylene came only from ICI's 1500 bar / 200 °C / free-radical process — high-energy, low-density (LDPE), branched. Ziegler's atmospheric-pressure TiCl4/AlEt3 system in 1953 dropped pressure 100x and produced linear, high-density HDPE with 50-100% higher tensile strength. By 2025 global polyethylene output is ~120 Mt/yr; HDPE alone is ~60 Mt/yr.
• Enabled an entirely new commodity polymer — polypropylene. Before Natta's 1954 isotactic-PP demonstration, propylene had no useful polymer; atactic PP is a tacky low-strength gum. Isotactic PP at >95% tacticity is now ~80 Mt/yr, used in auto bumpers, fibers, films, packaging, and rigid containers. Tm = 165 °C, tensile strength 30-40 MPa, modulus ~1.5 GPa.
• Stereoregularity is a tunable parameter. Internal donor (phthalate, succinate, 1,3-diether) plus external donor (alkoxysilane like cyclohexylmethyldimethoxysilane) selects for which Ti coordination geometries survive. Modern fourth-generation catalysts make iPP with mmmm pentad fraction >97%, giving consistent crystallinity and processing.
• Productivity at industrial scale. First-generation TiCl4/AlEt3 made ~1 kg polymer per gram of Ti — too little; the polymer needed expensive deashing to remove residual chloride. Fourth-generation MgCl2-supported catalysts hit ~10^5 kg/g Ti, leaving residual titanium below 1 ppm and chloride below 10 ppm. No deashing — polymer goes straight to extrusion.
• Pioneered single-site catalysis as a concept. ZN's 'active sites' were the first transition-metal heterogeneous polymerization centers, opening the door to metallocenes (Kaminsky 1980), late-transition-metal Brookhart catalysts (1995), and post-metallocenes — a multi-billion-dollar specialty-polymer industry.
• Drives gas-phase, slurry, and solution reactor design. Spheripol (slurry/loop), Unipol (gas-phase fluidized bed), Innovene (gas-phase stirred), and Borstar (slurry+gas-phase combined) are licensed reactor architectures all built around ZN catalyst chemistry. Single trains hit 600 kt/yr.
• Enabled HDPE pipe, blow molding, fibers, and EPDM rubber. Linear HDPE for water and gas pipe (PE100 grade); HMW-HDPE for blow-molded jerry cans; isotactic PP fiber for carpets and nonwovens; EPDM rubber (ethylene-propylene-diene terpolymer) for automotive seals; ethylene-vinyl-acetate-like LLDPE plastomers for food films.

Common misconceptions

• Ziegler-Natta is a single catalyst. It is a family. The original 1953 TiCl4/AlEt3 system is unsupported, low-productivity, and obsolete. Modern industrial 'ZN catalysts' are MgCl2-supported TiCl4 with phthalate or 1,3-diether internal donors and silane external donors. Newer 'late ZN' catalysts replace phthalate with succinate or diether to comply with REACH restrictions on phthalate plasticizers.
• Heterogeneous = poorly defined. Industrial ZN catalysts have multiple distinct active site populations (3-5 in typical iPP catalysts), each with characteristic stereoselectivity and chain-termination kinetics. The resulting polymer has a broad MW distribution (Mw/Mn ~5-8) and a tacticity distribution. This is why metallocenes — single-site, narrow MWD — never displaced ZN entirely; broad MWD improves melt processing in extruders and blown-film lines.
• The aluminum cocatalyst is just an alkylator. Et3Al has multiple roles: (1) alkylates Ti-Cl to Ti-Et (the actual chain-initiating bond), (2) reduces Ti(IV) to Ti(III), the active oxidation state, (3) scavenges trace impurities (water, oxygen) that poison the catalyst, (4) participates in chain transfer. Productivity and stereoselectivity depend on Al/Ti ratio, typically 100:1 to 300:1.
• You can polymerize any monomer with ZN. Polar monomers (vinyl chloride, vinyl acetate, acrylates) kill the catalyst — their Lewis-basic heteroatoms bind Ti irreversibly. Branched alpha-olefins (isobutylene) cannot insert; they require cationic catalysts. Cyclic monomers need ROMP (Grubbs catalysts) or metallocenes.
• HDPE and LDPE are made by Ziegler-Natta. HDPE yes; LDPE no. LDPE is still made by the original ICI 1933 high-pressure free-radical process at ~1500 bar / 200-300 °C, giving short and long chain branching that ZN linear catalysts cannot produce. LLDPE (linear LDPE, with controlled comonomer content) is the ZN-style replacement; it has linear backbone plus alpha-olefin (1-butene, 1-hexene, 1-octene) short branches and is now ~30 Mt/yr.
• Hydrogen is added to terminate chains. Hydrogen acts as a chain transfer agent, ending one chain by hydrogenolysis (Ti-polymer + H2 -> Ti-H + H-polymer) and starting a new one. It is the primary MW control knob in industrial reactors; doubling H2 partial pressure roughly halves Mn. Without hydrogen the polymer would have MW > 10^6 and be unprocessable.

Mechanism: Cossee-Arlman cycle on a chloride edge

The Cossee-Arlman mechanism (proposed 1964, refined since) explains chain growth at a single Ti center on the catalyst surface. The active site is an octahedral Ti(III) at a step or edge of the MgCl2 lattice, coordinated by four Cl atoms (lattice plus terminal), one alkyl group (the growing polymer chain), and one vacant coordination site. The cycle has two steps: (1) Coordination — an alpha-olefin docks at the empty site through its pi-system, with the substituent (methyl for propylene) oriented by the chiral pocket of surrounding chlorides and donor molecules. (2) Migratory insertion — the alkyl group migrates into the C=C double bond via a four-membered transition state, breaking the Ti-alkyl bond, forming a new C-C bond, and lengthening the chain by two carbons. The net effect: one olefin enters, one bond forms, and the empty site reappears on the opposite face of the metal. Each turnover takes roughly 10⁻⁴ s; one active site grows a chain to 10,000+ carbons in seconds.

Stereoselectivity comes from the chiral environment around the Ti. In iPP catalysis, the propylene's methyl group is forced toward one specific face by the steric bulk of the surrounding chlorides and an external silane donor. Each insertion installs a stereocenter with the same configuration as the previous one — this is 1,2-insertion with isotactic regularity. Statistical models (the 'enantiomorphic site' model) describe the >97% mmmm pentad as resulting from the catalyst site itself selecting prochirality, with occasional defects when the chain end inverts insertion direction. Roughly one defect per 100-300 monomer insertions sets the upper limit on isotacticity for industrial catalysts.

Chain termination is dominated by chain transfer to monomer (gives a vinyl-terminated polymer + a Ti-Et that starts the next chain), chain transfer to hydrogen (gives a saturated polymer + Ti-H + H2 that starts the next chain via Ti-H + olefin), and beta-hydride elimination (gives a vinyl polymer + Ti-H). Hydrogen is the main industrial control: in an HDPE slurry-loop reactor at 80 °C and 4 bar ethylene + 0.05-0.5 bar hydrogen, doubling H2 from 0.1 to 0.2 bar reduces Mn from ~120 kg/mol to ~60 kg/mol. Comonomer (1-butene, 1-hexene) added at 0.5-15 mol% controls density and crystallinity by introducing short branches that disrupt PE crystallization — moving from 0.96 g/cm³ HDPE to 0.92 g/cm³ LLDPE.

Ziegler-Natta vs metallocene vs free-radical polyethylene/polypropylene

Catalyst family	Active center	Pressure / Temp	MWD (Mw/Mn)	Products	Industrial scale
1st-gen Ziegler-Natta	TiCl4 + AlEt3, unsupported	1-10 bar / 50-80 °C	5-12 (very broad)	HDPE, isotactic PP (with deashing)	Obsolete after ~1965
4th-gen ZN (MgCl2-supported)	TiCl4 on MgCl2 + phthalate/1,3-diether	5-40 bar / 60-90 °C	4-8	HDPE, LLDPE, iPP, EPDM	~150 Mt/yr globally
Phillips chromium	CrO3 on silica, no cocatalyst	20-50 bar / 100-150 °C	10-30 (very broad)	HDPE pipe, blow-mold grades	~10 Mt/yr
Metallocene (Kaminsky)	Cp2ZrCl2 + MAO	5-30 bar / 70-90 °C	~2 (narrow)	mLLDPE plastomers, syndiotactic PP, COC	~10 Mt/yr
Free-radical (ICI 1933)	O2 or peroxide initiator, no catalyst	1200-3000 bar / 150-300 °C	5-25 (with long-chain branching)	LDPE only	~25 Mt/yr
Brookhart late-TM	Ni or Pd α-diimine + MAO	1-30 bar / 25-80 °C	~2-3	Hyperbranched PE, polar copolymers (lab/specialty)	Specialty (<0.1 Mt/yr)

Applications and case studies

• Borealis Borstar combined slurry-gas reactor. Two-stage process: a slurry-loop reactor with low H2 makes high-MW homopolymer, then a gas-phase fluidized bed with higher H2 and added comonomer adds the lower-MW copolymer fraction. Result: bimodal HDPE with PE100 pipe-grade strength (long-term hoop stress >10 MPa) at 600 kt/yr per train. Used by Borealis Schwechat, INEOS Cologne, and licensed widely.
• LyondellBasell Spheripol process for iPP. Two slurry-loop reactors in series, MgCl2-supported ZN catalyst, propylene as polymerization solvent. Produces isotactic homopolymer, random copolymer (1-5% ethylene), and impact copolymer (rubber phase added in gas-phase finishing reactor) at 350-500 kt/yr per train. Largest single-train installations at SABIC, Sinopec, Reliance.
• Automotive bumpers and dashboards. Heterophasic PP impact copolymers — iPP matrix with 15-25% ethylene-propylene rubber dispersed phase — are the standard car-bumper material, displaced steel and reaction-injection-molded urethane in the 1990s. Toyota, VW, GM each consume >100 kt/yr at the OEM level.
• EPDM rubber for door seals and roofing. Ethylene-propylene-diene terpolymer (60% ethylene, 35% propylene, 5% ENB diene) is made by ZN catalysts in solution at 50-80 °C; the diene gives sulfur-vulcanizable double bonds. Global EPDM is ~1.5 Mt/yr; Dow Nordel and ExxonMobil Vistalon are dominant brand-grades.
• Polyolefin fibers for nonwovens. iPP from ZN catalysts is melt-spun into fibers for carpet face yarn (~3 Mt/yr), spunbond/meltblown nonwovens (diapers, surgical masks, hygiene products — the COVID-19 N95 mask supply chain depends on iPP meltblown), and rope/twine. Tacticity >97% is required for adequate fiber tenacity (~5 g/denier).