Accelerating corrosion inhibitor discovery through computational routes: a case of naphthalene 1-thiocarboxamide

Corrosion inhibitors are critical in industrial applications such as oil and gas well acidizing, where steels are exposed to strong acids at elevated temperatures and pressures. Conventional design of corrosion inhibitors is largely empirical, costly, and time-consuming, often requiring years and hundreds of molecules to identify a single suitable inhibitor. While computational methods have been used mainly to rationalize mechanisms of known inhibitors, their application to discover new molecules is limited. This work employs first-principles DFT as a rapid, rational screening route to discover effective inhibitors for steels in acidic media. The study focuses on naphthalene 1-thiocarboxamide (NTC), identified through quantum chemical descriptors as a promising candidate, and validates its performance through explicit DFT adsorption/coverage modeling and experimental gravimetric and electrochemical testing in HCl at both room temperature and 60 °C.

Computational screening and quantum chemistry: NTC and comparator sulfur-containing molecules were built in Avogadro 1.2.0 and gas-phase geometries optimized with NWChem 6.8 using B3LYP/6-311++G**. Convergence criteria: energy cutoff 0.0027 meV, force cutoff 0.027 meV Å−1. Quantum chemical descriptors (EHOMO, ELUMO, Egap = ELUMO − EHOMO) were computed to assess reactivity and electron donation/acceptance tendencies.

DFT adsorption and coverage studies: Adsorption of NTC on Fe (001) was modeled using spin-polarized plane-wave DFT in Quantum ESPRESSO 6.7 with PBE-GGA and Grimme DFT-D3 dispersion. Plane-wave cutoffs: 35 Ry (wavefunctions), 300 Ry (charge density). Bulk Fe optimized with Monkhorst-Pack 8×8×8 k-point mesh. Slab: 5×4 surface supercell with 5 Fe layers, relaxed to forces < 0.0025 eV Å−1. Coverage studies used a larger 6×8 supercell to accommodate 1/48, 2/48, 3/48, and 4/48 monolayer (ML) coverages. Due to large cell sizes, Γ-point sampling was used for slab relaxations and adsorbate calculations. Multiple initial NTC orientations (flat and vertical) were considered. Adsorption energies were computed as ΔEads = Ecomplex − (Eslab + N Emol). Electron density difference (EDD) and projected density of states (PDOS) analyses probed bonding; Bader charge analysis (PAW, charge density cutoff 800 Ry; Henkelman group algorithm) quantified charge transfer. Adsorption surface free energy per area, γads = (n(ΔEads − μ))/A, was evaluated versus molecular chemical potential μ to assess stability across coverages.

Experimental materials and sample preparation: NTC (>97%, CAS 20300-10-1; Sigma-Aldrich) was used as received, dissolved in tetrahydrofuran. Mild steel coupons (IS2062 grade; composition provided) were ground with 80–600 grit emery papers, degreased with acetone, and stored in a desiccator.

Electrolyte and conditions: Tests were performed in 1 M HCl at ~27 °C; additional weight-loss tests at 60 °C (for 1 mM NTC). Exposed area for electrochemical tests: 1 cm² in 300 mL electrolyte, ambient atmosphere.

Weight-loss tests: NTC concentrations of 0.05, 0.2, 0.7, and 1 mM (27 °C), and 1 mM (60 °C) were evaluated. Coupons were immersed for 2 h, rinsed with distilled water, air-dried, and weighed. Inhibition efficiency (IE) was computed as IE(%) = [1 − (corrosion rate with inhibitor / corrosion rate without inhibitor)] × 100. Adsorption isotherms (Langmuir, Temkin, Frumkin) were examined; Langmuir best fit. From Kads, ΔGads was calculated via ΔGads = −RT ln(55.5 Kads).

Potentiodynamic polarization: Measurements were performed with an Ametek K0235 flat cell (Pt mesh counter electrode; Ag/AgCl (Vycorr tip) reference) using an Ametek PARSTAT 1000 potentiostat. Scan rate 1 mV s−1 from −250 mV to +250 mV versus OCP. Corrosion current density (Icorr) and Tafel slopes (βa, βc) were obtained with VersaStudio; IE(%) = [1 − (Icorr with inhibitor / Icorr without inhibitor)] × 100. Polarization resistance Rp was computed using the Stern–Geary equation: Rp = (βa βc) / [2.303 Icorr (βa + βc)].

Electrochemical impedance spectroscopy (EIS): AC perturbation 10 mV over 105 to 10−1 Hz. Data were fitted using ZSimpWin with equivalent circuits R(QR) and R(Q(R(QR))) as appropriate. Charge transfer resistance Rct and constant phase element parameters (Qct-Y0, n) were extracted. IE(%) from EIS was computed as IE(%) = [1 − (Rct with inhibitor / Rct without inhibitor)] × 100.

• Quantum chemical screening: Among several sulfur-containing molecules benchmarked against a known VCI (11-mercaptoundecanoic acid), NTC exhibited the lowest Egap (4.04 eV; EHOMO −5.89 eV; ELUMO −1.85 eV), suggesting superior reactivity and inhibition potential.
• Adsorption geometry and energetics: Flat orientations of NTC on Fe (001) chemisorb strongly via Fe–C/N/S bonds; vertical orientations adsorb primarily via Fe–S. Conf1 (flat) yielded the strongest adsorption with |ΔEads| = 4.54 eV, stronger (~17%) than trans-cinnamaldehyde (|ΔEads| = 3.89 eV) reported previously. Average bond lengths indicate covalent character: Fe–C 2.16 Å, Fe–N 2.18 Å, Fe–S 2.25 Å (close to sums of covalent radii). PDOS shows significant hybridization, strongest for Fe–S, weaker for Fe–C, and weakest for Fe–N. EDD maps reveal charge accumulation between C/N/S and surface Fe atoms, consistent with covalent bonding and induced sp2→sp3 rehybridization in the adsorbed molecule.
• Charge transfer: Bader analysis for the most stable adsorption conformation shows net charge transfer of ~1.74 e from Fe surface to NTC. S atom receives the largest charge (−0.193 e), while C atoms receive on average ~−0.135 e; N receives −0.077 e, corroborating dominant roles of S and C in bonding.
• Coverage effects: Increasing coverage from 1/48 to 4/48 ML decreases |ΔEads| only slightly (~0.2 eV; ~4.4%), indicating weak lateral repulsions and favoring compact film formation. Adsorption surface free energy γads versus μ identifies 4/48 ML as most stable at μ > −4.15 eV, supporting monolayer formation under realistic conditions.
• Weight-loss tests (1 M HCl): At 27 °C, corrosion rate reduced from 1.1081 to 0.0307 cm yr−1 at 1 mM (IE 97.23%). At 60 °C, corrosion rate reduced from 16.4233 to 0.2398 cm yr−1 at 1 mM (IE 98.54%). Bubble evolution due to HER diminished with increasing NTC concentration; absent at ≥0.2 mM, consistent with compact layer formation.
• Adsorption isotherm: Langmuir isotherm best describes NTC adsorption on mild steel. From Kads, ΔGads = −22.16 kJ mol−1 (negative and in the mixed physisorption/chemisorption range), though chemisorption nature is more reliably confirmed by DFT bonding metrics (bond lengths, EDD, PDOS).
• Potentiodynamic polarization: Icorr decreased from 1788 μA cm−2 (blank) to 2.93 μA cm−2 at 1 mM (IE 99.84%); corresponding Rp increased to 9561 Ω cm2. At 0.7 mM, Icorr 7.50 μA cm−2 (IE 99.58%). Mixed-type inhibition observed, with changes in cathodic HER behavior above 0.2 mM and anodic behavior differing between low (0.05 mM) and higher concentrations.
• EIS: Nyquist plots show increasing semicircle diameters with concentration. Rct increased substantially up to 6187 Ω cm2 at 1 mM, indicating outstanding protection, aligning with weight-loss and polarization results.

The DFT-based screening successfully identified NTC as a strong inhibitor by correlating low ELUMO and Egap with high adsorption propensity. Explicit adsorption calculations demonstrated strong, covalent chemisorption of NTC in flat orientations on Fe (001), with extensive Fe–S/C/N interactions, significant charge transfer, and orbital hybridization—mechanistic features consistent with robust surface blocking. Coverage simulations suggested minimal lateral repulsion and thermodynamic preference for high coverage (up to 4/48 ML), supporting the formation of a compact protective monolayer, which is essential for mitigating corrosion. Experimental validation in 1 M HCl corroborated the computational predictions: large reductions in corrosion rate and current density at low concentrations (≤1 mM), high charge transfer resistances, and a Langmuir adsorption isotherm reflecting weak lateral interactions. The mixed inhibition character and concentration-dependent changes in cathodic and anodic branches indicate that surface coverage and bonding modulate both HER and anodic dissolution pathways. Overall, the integrated computational–experimental approach directly addresses the research goal of accelerating inhibitor discovery and mechanistic understanding, demonstrating that NTC is highly effective for mild steel in acidic environments at room and elevated temperatures.

This work presents a DFT-driven workflow for rapid identification and validation of corrosion inhibitors, exemplified by the discovery of naphthalene 1-thiocarboxamide (NTC) for mild steel in hydrochloric acid. Quantum chemical descriptors flagged NTC as highly reactive; explicit adsorption and coverage DFT studies established strong chemisorption and compact monolayer formation on Fe (001). Comprehensive experimental validation (weight-loss, polarization, EIS) confirmed exceptional inhibition efficiencies at low dosage (1 mM) and at both 27 °C and 60 °C. The study demonstrates that first-principles modeling can effectively complement and accelerate experimental screening and mechanistic elucidation in inhibitor design. Future work may extend this framework to broader chemical spaces, other metal substrates and environments, and incorporate high-throughput computations and data-driven models to further streamline discovery.

• Candidate space: Only a limited set of sulfur-containing molecules was screened; other promising chemistries were not explored.
• Model assumptions: DFT simulations used PBE-GGA with D3 dispersion, finite slab models, Γ-point sampling for large cells, and specific coverages; these approximations may affect quantitative adsorption energies.
• Environment scope: Studies focus on Fe (001) and 1 M HCl for mild steel at ~27 °C and 60 °C; performance in other acids, concentrations, temperatures, flow conditions, or alloy compositions was not evaluated.
• Isotherm interpretation: While ΔGads from Langmuir fitting suggests mixed adsorption, the authors note that ΔGads alone is not a reliable discriminator between physisorption and chemisorption; conclusions rely on combined computational and experimental evidence.