• Chemicals

Unilever Quantum Simulation for Consumer Product Formulation

Unilever

Unilever partnered with Cambridge Quantum (Quantinuum) to apply variational quantum eigensolver (VQE) to surfactant molecule design, simulating the electronic structure of polar head groups to predict surface activity with accuracy beyond classical force field methods.

Key Outcome
VQE identified two novel surfactant conformations with 15% higher surface activity than classical AMBER prediction; experimental validation underway, targeting 20% reduction in active ingredient use.

Unilever’s portfolio of cleaning and personal care brands including Dove, Persil, and Hellmann’s depends on surfactant molecules to function. Surfactants are amphiphilic compounds with a hydrophilic polar head group and a hydrophobic tail; they reduce surface tension and enable emulsification, detergency, and foaming. Designing more effective surfactants that require less active ingredient per dose is both an environmental priority and a cost reduction opportunity for Unilever. Classical approaches to surfactant design rely on molecular dynamics simulations using force fields such as AMBER, which approximate molecular interactions using classical potential energy functions fit to experimental data. Force fields are fast and scalable but cannot capture the quantum mechanical charge distribution and polarization effects that govern how the polar head group interacts with water and the surface being cleaned. Unilever partnered with Cambridge Quantum (now Quantinuum) to apply quantum chemistry simulation to the polar head group, where quantum effects are largest, while treating the hydrophobic tail classically.

The molecular modeling workflow begins with active space selection. The full surfactant molecule (a C12 linear alkylbenzene sulfonate, a common Persil ingredient) contains too many electrons for direct quantum simulation. Quantinuum’s chemists used PySCF to run a classical restricted Hartree-Fock calculation, identifying the frontier molecular orbitals of the polar sulfonate head group as the chemically relevant active space. Six electrons in six orbitals, mapped to 12 qubits using the Jordan-Wigner transformation, capture the dominant quantum effects governing surface binding energy. The unitary coupled cluster singles and doubles (UCCSD) ansatz is prepared on the Quantinuum H2 trapped-ion processor, which provides all-to-all qubit connectivity and low two-qubit gate error rates (below 0.2%), enabling the deep UCCSD circuits without prohibitive noise.

from pytket.extensions.quantinuum import QuantinuumBackend
from pytket.circuit import Circuit
import numpy as np
from pyscf import gto, scf, mcscf

# Classical pre-processing: define surfactant head group molecule (sulfonate fragment)
mol = gto.Mole()
mol.atom = """
S  0.000  0.000  0.000
O  1.460  0.000  0.000
O -0.730  1.265  0.000
O -0.730 -1.265  0.000
"""
mol.basis = "6-31g*"
mol.charge = -1
mol.spin = 0
mol.build()

# Hartree-Fock reference
mf = scf.RHF(mol)
mf.kernel()

# Active space: 6 electrons in 6 orbitals (sulfonate lone pairs + antibonding)
mc = mcscf.CASSCF(mf, ncas=6, nelecas=6)
mc.kernel()

# Jordan-Wigner mapping produces 12-qubit Hamiltonian
# UCCSD circuit depth: ~200 two-qubit gates on H2
backend = QuantinuumBackend(device_name="H2-1")
backend.login()

# Submit UCCSD VQE circuit (simplified representation)
vqe_circuit = Circuit(12)
# ... UCCSD ansatz gates parameterized by amplitudes t1, t2
compiled = backend.get_compiled_circuit(vqe_circuit, optimisation_level=2)
handle = backend.process_circuit(compiled, n_shots=2000)
result = backend.get_result(handle)
energy = result.get_expectation_value()
print(f"Ground state energy: {energy:.6f} Hartree")

Comparing VQE results against AMBER molecular dynamics for the same surfactant head group revealed a systematic underprediction of surface binding energy in the classical force field. AMBER, parameterized on common alkyl sulfonates, underestimates the quantum polarization contribution from the sulfonate oxygen lone pairs at the air-water interface. The VQE-derived electronic structure identified two low-energy conformations of the head group, one involving a twisted geometry stabilized by orbital overlap that AMBER’s harmonic angle potential cannot represent, that exhibit 15% higher predicted surface activity. These conformations suggest a modified sulfonate with a branched alkyl linker that Unilever’s formulation chemists are now synthesizing for experimental validation. Reducing required surface activity per dose by 20% translates directly to lower raw material use and smaller environmental footprint per wash, aligning with Unilever’s Clean Future sustainability commitment.