# Large-Scale Solid-Phase Peptide Synthesis (SPPS): Sizing Columns, Hydrodynamics, and Solvent Optimization for GLP-1 Blockbusters

# Engineering the Multi-Ton Production of Complex Therapeutic Peptides (Semaglutide, Tirzepatide) in Compressible Resin Beds

# Executive Summary

The explosive clinical demand for GLP-1 receptor agonists (such as Semaglutide [31 amino acids] and Tirzepatide [39 amino acids]) has fundamentally rewritten pharmaceutical manufacturing economics. Historically, therapeutic peptides were niche products manufactured in 50–200 L batch synthesizer columns producing 5 to 50 kilograms annually. Today, commercial API campaigns require metric tons per annum, placing unprecedented hydrodynamic and chemical engineering burdens on industrial Solid-Phase Peptide Synthesis (SPPS) equipment.

Unlike small-molecule batch syntheses that scale in standard agitated glass-lined or Hastelloy vessels, large-scale SPPS is governed by packed-bed chromatography and axial flow mechanics within compressible, swelling polymeric matrices.

Key engineering hurdles in multi-kilogram to ton-scale SPPS include:

1. Compressible Bed Hydrodynamics: Polymeric resins (polystyrene crosslinked with 1% DVB or poly(ethylene glycol) (PEG) grafted beads) behave as viscoelastic porous media. Flow-induced drag causes bed compaction, steep axial pressure gradients, and catastrophic column wall channeling.
2. Extreme Process Mass Intensity (PMI): Traditional SPPS consumes 3,000 to 15,000 kg of hazardous organic solvents (principally DMF, NMP, and DCM) per single kilogram of purified peptide drug substance.
3. Mass Transfer vs. Reaction Kinetics: Deprotection (Fmoc cleavage via piperidine) and coupling reactions (activation via carbodiimides, oxyma, and phosphonium/uronium salts) must achieve $\ge 99.5\%$ conversion per cycle across 30 to 40 consecutive cycles to prevent unmanageable truncated deletion sequences.
4. Final Cleavage Safety & Thermal Management: Bulk resin cleavage using concentrated Trifluoroacetic Acid (TFA) cocktails with silane and thiol scavengers generates severe exotherms, volatile off-gassing (isobutylene, $CO_2$), and potential thermal runaways.

This technical engineering guide provides the mathematical transport models, column sizing equations, green solvent transition frameworks, and a fully worked industrial scale-up case study for large-scale GLP-1 manufacturing.

# 1. Resin Matrix Mechanics: Swelling, Porosity, and Compressibility

In SPPS, the nascent peptide chain grows covalently attached to an insoluble polymeric support matrix. The two dominant resin platforms across commercial API facilities are:

• Microporous Crosslinked Polystyrene (PS-1% DVB): High mechanical rigidity, but exhibits solvent-dependent swelling factors ($S_v = V_{swollen} / V_{dry} \approx 3.0\text{ to }6.5\text{ mL/g}$).
• Pure Poly(ethylene glycol) (ChemMatrix) or PEG-Grafted Polystyrene (Tentagel): Exceptional solvation and macro-porosity for hydrophobic sequences, but prone to high viscoelastic deformation and severe bed compression at modest superficial velocities.

Dry Resin Bead (100 µm) ──+ Solvent / Reagent Plume──► Swollen Matrix (250 µm) ──+ Drag Force (ΔP)──► Compressed Bed (Pore Collapse)
[Rigid Sphere]                                          [Hydrated Swollen Gel]               [Asymmetric Flux & Channeling]

# 1.1 Resin Bed Volume Dynamics Across Chain Elongation

As the 31- or 39-mer peptide backbone extends, the mass of peptide on the bead increases dramatically. For an initial resin substitution level $L_0$ (in $\text{mmol amino acid/g dry resin}$), the instantaneous substitution $L_n$ after $n$ coupling steps is:

Where:

• $L_0$: Initial functional loading of the bare resin ($\text{mmol/g}$)
• $MW_i$: Molecular weight of the protected amino acid derivative ($\text{g/mol}$)
• $MW_{water}$: Water or leaving group mass eliminated during amide bond formation ($18.02\text{ g/mol}$)

At cycle 30 for a GLP-1 analog, the peptide mass fraction often exceeds $60\text{–}75\%$ of the total bead weight, altering the internal pore volume, viscoelastic elasticity modulus ($E_b$), and fluid hydrodynamic resistance.

# 2. Axial Hydrodynamics: The Compressible Ergun-Blake-Kozeny Model

In rigid packed beds, pressure drop is accurately modeled by the classical Ergun equation:

Where:

• $\Delta P$: Bed pressure drop ($\text{Pa}$)
• $L$: Bed height under flow ($\text{m}$)
• $\varepsilon$: Interstitial bed void fraction (dimensionless, typically $0.35 - 0.42$ for uncompressed beds)
• $\mu$: Solvent dynamic viscosity ($\text{Pa}\cdot\text{s}$)
• $u_s$: Superficial fluid velocity ($u_s = Q / A_{col}$, in $\text{m/s}$)
• $d_p$: Mean swollen bead particle diameter ($\text{m}$)
• $\rho$: Solvent density ($\text{kg/m}^3$)

In industrial SPPS, flow is strictly laminar ($\text{Re}_p < 2.0$), rendering the inertial kinetic energy term negligible ($1.75 \rho u_s^2 / d_p \to 0$). However, because peptide synthesis resins are compressible gels, the local bed porosity $\varepsilon(z)$ is not constant—it collapses axially toward the bottom flow distributor screen under the cumulative fluid drag force.

# 2.1 The Compressible Bed Darcy-Ergun Formulation

The local bed permeability $K_p(z)$ decreases exponentially as axial stress $\sigma_z(z)$ builds:

Where $c_m$ is the empirical bed compressibility index ($\text{bar}^{-1}$).

Bed Height (z)
┌──────────────────────────────┐ Top Piston (Low Compression: ε ≈ 0.40)
│    •   •   •   •   •   •   • │
│   •   •   •   •   •   •   •  │
│  •   •   •   •   •   •   •   │ Mid-Bed (Moderate Compression: ε ≈ 0.32)
│ •  •  •  •  •  •  •  •  •  • │
│••••••••••••••••••••••••••••••│ Bottom Retaining Screen (Severe Compaction: ε ≈ 0.18, ΔP Spikes!)
└──────────────────────────────┘

# 3. Industrial Synthesis Column Architecture: Axial vs. Dynamic Axial Compression (DAC)

Industrial commercial SPPS synthesizers (50 L to 1,500 L column vessels) require specialized mechanical configurations to maintain uniform flow distribution without bed fluidization or channeling:

# 4. Solvent Intensity & Green Chemistry Transition in GLP-1 Manufacturing

Traditional SPPS relies extensively on European REACH-restricted substances:

• $N,N$-Dimethylformamide (DMF) and $N$-Methyl-2-pyrrolidone (NMP): Reproductive toxic (Repr. 1B) dipolar aprotic solvents.
• Dichloromethane (DCM): Volatile organohalogen carcinogen utilized for washing and deprotection.

In commercial GLP-1 campaigns producing 1,000 kg of API, traditional DMF/DCM protocols generate over 8,000,000 kg of hazardous liquid waste.

# 4.1 Green Dipolar Aprotic Replacements

Conventional:   [ DMF / NMP ] ───► Reproductive Toxin, High ETP/Incineration Load
                      │
                      ▼
Green Sustained: [ 2-MeTHF / CPME ] or [ DMSO / EtOAc Blends ] or [ Cyrene / γ-Valerolactone ]
                 • Lower Viscosity → 40% Lower Bed ΔP
                 • Clean Solvent Stripping & 90% Recycle Loop

# 4.2 Process Mass Intensity (PMI) & Solvent Recycling Economics

The pharmaceutical industry standard metric for manufacturing sustainability and raw material efficiency is Process Mass Intensity (PMI):

$\text{PMI} = \frac{\sum m_{\text{raw materials}} + \sum m_{\text{solvents}} + \sum m_{\text{water}}}{m_{\text{isolated pure peptide API}}}$

In conventional linear SPPS using single-pass DMF/DCM protocols, the commercial PMI ranges between $6,500\text{ and }15,000\text{ kg mass / kg API}$. By integrating a closed-loop solvent recovery skid with 2-MeTHF and green wash mixtures:

# 5. Cleavage Reactor Design & Exotherm Management

Following completion of the 31- or 39-amino-acid automated assembly, the resin-bound protected peptide undergoes global deprotection and acidolytic cleavage.

The standard cleavage cocktail comprises:

# 5.1 Reaction Hazards & Thermodynamics

Cleavage of acid-labile side-chain protecting groups (such as $t$-butyl ethers, $t$-Boc, Trityl, Pbf) is fiercely exothermic:

• Specific Heat of Reaction ($\Delta H_{cleav}$): $-120\text{ to }-180\text{ kJ per mole of protecting group}$.
• Off-Gassing Kinetics: Massive generation of volatile gaseous isobutylene ($C_4H_8$) and $CO_2$:

   Cleavage Reagent Inlet (TFA / TIS / Water at 5°C)
                │
                ▼
  ┌───────────────────────────┐
  │  [GLR Cleavage Reactor]   │ ──► Off-Gas Vent: Isobutylene + CO2 (Requires Scrubber!)
  │   - Controlled -10°C      │
  │     Cryo-Jacket Cooling   │
  │   - Rushton / PBT Mixed   │
  │     Slurry Suspension     │
  │                           │
  └─────────────┬─────────────┘
                ▼ Bottom Quench Valve
  ┌───────────────────────────┐
  │  Precipitation Crystallizer│ ◄── Cold Methyl tert-Butyl Ether (MTBE / CPME) at -15°C
  │  (Peptide Slurry Out)     │
  └───────────────────────────┘

# 6. Worked Industrial Case Study: Sizing an 800 L Commercial GLP-1 Column

# 6.1 Process Specifications

• Target Molecule: Semaglutide intermediate (31-mer backbone).
• Batch Production Target: $40.0\text{ kg}$ cleaved crude peptide per synthesis cycle.
• Resin Selected: Aminomethyl-polystyrene with Rink Amide Linker ($100\text{–}200\text{ mesh}$, $d_p = 100\mu\text{m}$ dry, $210\mu\text{m}$ swollen).
• Initial Functional Loading ($L_0$): $0.45\text{ mmol/g dry resin}$.
• Solvent Platform: 2-MeTHF / Ethyl Acetate green wash cocktail ($\mu = 0.65\text{ mPa}\cdot\text{s}$, $\rho = 860\text{ kg/m}^3$).

# 6.2 Step-by-Step Engineering Sizing Calculations

# Step 1: Total Resin Mass Required

Target peptide moles ($MW_{peptide} \approx 4,113\text{ g/mol}$):

Accounting for an average cumulative coupling yield of $85\%$ across the 31 cycles:

Required dry bare resin mass:

# Step 2: Final Swollen Bed Volume at Cycle 31

At the end of sequence assembly, the combined mass of the resin-bound peptide is:

With an empirical swollen specific volume $v_{swollen} \approx 5.5\text{ L/kg}$:

Adding a $25\%$ hydrodynamic headspace safety margin for dynamic axial compression piston stroke:

# Step 3: Column Diameter and Superficial Velocity Limits

To prevent exceeding the critical bed pressure drop ($\Delta P_{max} \le 3.5\text{ bar}$), the superficial velocity is capped at $u_s \le 1.2\text{ cm/min}$ ($2.0 \times 10^{-4}\text{ m/s}$).

Selecting an internal column diameter $D_{col} = 800\text{ mm}$ ($A_{col} = \pi/4 \times (0.8\text{ m})^2 = 0.5026\text{ m}^2$):

• Bed Height at Full Elongation ($L$):

• Volumetric Flow Rate ($Q$):

# Step 4: Pressure Drop Verification via Laminar Darcy-Ergun Relation

Using the compressed bed void fraction at maximum velocity ($\varepsilon \approx 0.32$, $d_p = 2.1 \times 10^{-4}\text{ m}$):

# Step 5: 31-Cycle Sequence Automation & Time Breakdown

For a commercial 31-amino-acid GLP-1 chain, cycle time predictability is vital to prevent resin fouling:

• Total Synthesis Campaign Duration: $31\text{ cycles} \times 62\text{ min} = 1,922\text{ min} \approx \mathbf{32.0\text{ hours continuous automated run}}$.

# 7. Operational & Scale-Up Engineering Checklist

1. Elastomer Selection for Wetted Seals: Standard EPDM and Viton seals degrade rapidly in NMP, TFA, and DCM. Specify FFKM (Kalrez 6375 / Chemraz) for all dynamic piston O-rings.
2. Fritted Screen Aperture Sizing: The retaining bottom mesh aperture must be sized to $d_{mesh} \le 0.4 \times d_{p,dry}$ (e.g., $30\text{–}40\mu\text{m}$ for a $100\mu\text{m}$ nominal resin) to prevent bead extrusion into the effluent lines while minimizing boundary flow resistance.
3. Automated UV Detection of Deprotection Kinetics: Install inline variable-wavelength UV flow cells ($301\text{ nm}$ for dibenzofulvene-piperidine adducts) to enable real-time feedback-controlled washing, reducing solvent consumption by up to $35\%$.
4. Thermal Pre-Conditioning: Reagents and wash solvents should be fed through a plate heat exchanger to maintain bed temperature at an optimal $35^\circ\text{C} \pm 2^\circ\text{C}$—accelerating coupling kinetics without inducing racemization of sensitive residues (His, Cys).