Introduction

One of the important mineral acids that is widely used in many applications, including well acidizing, water treatment, chemical cleaning, and acid pickling, is hydrochloric acid (HCl)1,2,3,4,5,6,7,8. The selection of cost-effective materials to handle this acid requires extreme care and detailed engineering. The presence of certain impurities, such as ferric salts, cupric salts, and chlorine, in the acid and/or a high level of aeration amplifies the oxidizing power of the solution, leading to accelerated corrosion damage9. In the petroleum and gas industry, the exposure of materials to acidic environments is more common and frequent than to neutral or alkaline environments10,11,12,13. This necessitates exploring options and efficient techniques to mitigate and control the corrosion of the different types of steel as they constitute a large fraction of the metallic materials that are exposed to acidic media. Corrosion inhibitors are widely utilized to mitigate corrosion risks. For example, Umoren14 investigated the corrosion inhibition of polypropylene glycol for X60 pipeline steel in 15% HCl. The corrosion inhibition efficiency (IE%) was found to be 90% at 55 °C using 1000 ppm of polypropylene glycol. Ituen et al.15 explored the corrosion inhibition of N-acetyl cysteine (NAC) as a base inhibitor compound with different additives of 5-hydroxytryptophan (5-HTP), glutathione, potassium iodide (KI), and polyethylene glycol in 15% HCl solution for different grades of steel (mild steel, J55, and X80) at elevated temperatures. The optimum composition showed a corrosion protection efficiency of 97% at 90 °C on X80 steel. Ansari et al.16 synthesized two pyrazolone derivatives (PZ-1 and PZ-2) to study their corrosion inhibition in 15% HCl solution for N80 steel. They found that PZ-1, which contains an additional methyl group, showed a better corrosion IE% of 93% at 35 °C in comparison to 85.5% for PZ-2. Yadav et al.17, prepared two carbohydrate compounds namely BIHT and MIHT as green inhibitors for N80 steel in the aforementioned HCl concentration. It was revealed that the chemisorbed inhibitor (BIHT) had the highest IE% of 94.8% at 30 °C. Unfortunately, some of the effective corrosion inhibitors used to mitigate corrosion are highly toxic. Increased environmental awareness and the development of regulations have imposed restrictions on the use of such inhibitors. Additionally, the safe disposal of corrosion inhibitors after use or the treatment of contaminated streams is critical and usually defined as a step in all chemical treatment programs, which adds to the corrosion control cost. The environmental and safety concerns related to corrosion inhibition processes have encouraged researchers to explore alternatives that are eco-friendly and offer acceptable inhibition efficiency, especially in acidizing treatment processes, which cost the state of Qatar, in addition to chemical treatment applications, a direct expenditure close to 8 billion USD per year18. Therefore, focus on natural and nature-based products is increased to produce so-called “green inhibitors”19,20,21,22. However, the harsh acidic and high-temperature environments during drilling and stimulation processes in the oil and gas wells limit the full utilization of such green corrosion inhibitors. The development of green corrosion inhibitors that withstand severe acidic corrosive environments and cog the carbon steel corrosion, particularly in well acidizing, is highly desired, especially in the carbonate formations found in the state of Qatar and many other countries.

In this work, laurhydrazide N′-propan-3-one (LHP) is tested as an inexpensive and highly efficient green inhibitor for the corrosion of mild steel (MS) in 16% HCl solution (5 M) at various temperatures using the potentiodynamic polarization (PDP) and electrochemical impedance spectroscopy (EIS) techniques. In addition, several surface characterization methods are used, such as scanning electron microscopy (SEM) coupled with an energy-dispersive X-ray spectroscopy (EDX) unit, atomic force microscopy (AFM), and X-ray photoelectron spectroscopy (XPS). Furthermore, the kinetics and thermodynamic parameters are measured and/or calculated. A geometry optimization of LHP is also performed and the time-dependent density functional theory was used to calculate the electronic absorption spectra. Furthermore, frequency calculations are also performed on the optimized geometry.

Experimental Work

Materials and Material Preparation

The elemental analysis of the MS (MS) used in this work, shown in Table 1, was performed using the ARL 3460 optical emission spectrometer (ThermoFisher Scientific, Waltham, MA, USA).

Table 1 Elemental composition analysis for MS.

Coupons of equal size (1.5 × 1.5 × 0.5 cm3) were cold-cut from a MS plate. They were polished using SiC emery papers from 250 down to 4000 grit, washed with ethanol for 10 min in an ultrasonic bath, then degreased with acetone for 1 min, followed by rinsing with ethanol and finally with deionized (DI) water. Finally, they were dried using air. The corrosive acidic solution that was used in this work was prepared by diluting analytical HCl from 36 to 16% (5 M) using DI water. The chemical structure of LHP (purchased from Shanghai Dejun Technology Co., Ltd, Shanghai, China) is shown in Fig. 1.

Figure 1
figure 1

Laurhydrazide N′-propan-3-one chemical structure.

Electrochemical Measurements

A 250 ml double-jacketed three-electrode corrosion cell was used, in which a graphite rod, MS coupon (only 0.5 cm2 is exposed to the corrosive solution), and saturated calomel electrode were the counter, working and reference electrodes, respectively. SCE was used with a Luggin capillary to minimize the IR drop. The temperature of the cell was controlled 20 and 80 °C using a thermostat water circulator (Julabo F12, Seelbach, Germany). A thermometer was used to monitor the temperature of the electrolyte before and during the experiments. A Reference 300 GAMRY (Warminster, PA, USA) was the potentiostat used to measure the EIS and PDP curves. The GAMRY measurement software packages include EIS300 for EIS and DC105 for corrosion analysis. The free corrosion potential (EOCP) was stabilized before any electrochemical testing by placing it in the solution for 30 min. EIS was performed within a frequency range of 1 × 10−1 to 1 × 105 Hz, with a 5 mV AC peak-to-peak amplitude. PDP curves were measured between −250 and +250 mV with respect to EOCP in the more noble direction. The sweep rate was always 0.3 mV s−1. Four concentrations of the corrosion inhibitor were used: 92, 185, 277, and 370 μmol L−1.

Surface Analysis

Surface analysis has a vital role in characterizing the surface morphology and studying the effect of the inhibitor and its interaction with the substrate. Immersion tests were conducted to quantify the effect of the corrosive acidic medium on the MS electrode and to study the surface topography before and after the addition of the LHP inhibitor. Three steel coupons were ground and then polished to a mirror-like finish using alumina suspensions of different particle size. The coupons were immersed in the LHP -free 5 M HCl solution for 24 h at 25 °C. The same procedure was repeated, but with the addition of 370 μmol L−1 of LHP. The morphology of the two samples was examined and compared using SEM (FEI NOVA NANOSEM 450, Hillsporo, OR, USA), typically operated with an acceleration voltage of 20 kV coupled with an EDX unit. The adsorbed inhibitors on the MS were analyzed using XPS (AXIX Ultra DLD, Kratos, UK), employing a monochromatic Al Kα X-ray source. AFM (MFP-3D, Asylum Research, Goleta, CA, USA) was used in the non-contact tapping mode in air for determining the surface roughness.

Computational

The geometry optimization was performed using the wb97xd density functional method and 6-311++ g(d, p) as a basis set23. The time-dependent density functional theory was utilized to calculate the electronic absorption spectra (using 30 excited states) and the frequency calculations were also performed on the optimized geometry. All these simulations were performed using the Gaussian 09 software package24. The electronic absorption spectra were collected using a UV–Vis peak half-width of 0.15 eV. The visualizations were created using the Chemcraft software package25. The molecular simulations were created using the adsorption locator tool in Materials Studio26. Adsorption Locator identifies possible adsorption configurations by carrying out Monte Carlo searches of the configurational space of the substrate-adsorbate system as the temperature is slowly decreased using simulated annealing. In our calculation we implement the universal force field with fine quality. The surfaces were constructed from pure iron metal followed by a 100, 110, or 111-orientation surface cleavage with a thickness of two layers and a 30 Å vacuum slab. A super cell with dimensions of 10 × 10 was utilized to measure the adsorption energy and the geometry of the adsorbed molecules. The GaussView 5.0 software package was further applied in visualizing the graphical isosurfaces of the electron density.

Results and Discussion

Potentiodynamic Polarization Studies

Figure 2 illustrates the PDP curves at different temperatures for MS in 5 M HCl at various LHP concentrations. The values of Ecorr, cathodic (βc) and anodic (βa) Tafel slopes besides the corrosion current density (icorr) are listed in Table 2. In addition, the IE% and surface coverage, θ in Table 2 are calculated using Equations (1) and (2), respectively27.

$$IE \% =(\frac{{i}_{1}-{i}_{2}}{{i}_{1}})\times 100$$
(1)
$$\theta =\frac{IE \% }{100\,}$$
(2)

where i1 and i2 are the corrosion current densities in the absence and presence of LHP, respectively. Furthermore, the Stern–Geary equation, Equation (3), is utilized to calculate the polarization resistance, Rp28:

$${R}_{{\rm{p}}}=\frac{{\beta }_{{\rm{c}}}\,{\beta }_{{\rm{a}}}}{2.303\,{i}_{{\rm{corr}}}({\beta }_{{\rm{c}}}+{\beta }_{{\rm{a}}})}$$
(3)
Figure 2
figure 2

PDP curves of MS in 5 M HCl before and after the addition of different concentrations of LHP at (a) 20, (b) 40, (c) 60, and (d) 80 °C.

Table 2 PDP parameters at various temperatures for MS in 5 M HCl at different concentrations of LHP.

Table 2 reveals the direct relation between the inhibitor concentration and the reduction in the rate of corrosion. It can be noticed that by increasing the inhibitor concentration, the corrosion current density (icorr), decreases. However, the corrosion current density (icorr), increases by increasing the temperature. A maximum IE% of 94% is attained at 20 °C using 370 µmol L−1 of LHP. Elevating the temperature to 80 °C decreases the IE% to 61%. The parallel anodic and cathodic Tafel lines and their slopes suggest that the cathodic reaction is primarily activation-controlled, i.e. cathodic and anodic currents were moved to lower values at the same potential with increasing the inhibitor concentration at elevated temperatures. This behavior indicates that both the anodic dissolution and cathodic reactions were suppressed, i.e., LHP is a mixed-type inhibitor. Moreover, an 85 mV shift in Ecorr, in the more or less noble directions after the addition of an inhibitor, the inhibitor is classified as an anodic or cathodic inhibitor, respectively29. Otherwise, the inhibitor is considered as mixed type category, i.e., it suppresses both reactions30,31,32. That is, the explored compound inhibit the anodic dissolution of the metal surface in the 5 M HCl in addition to the cathodic reaction.

Figure 2 shows shoulders in the anodic curves at high inhibitor concentrations, at all temperatures except for 80 °C. This can be attributed to the alteration of the surface area covered with the corrosion inhibitor, because of the rearrangement of the adsorbed inhibitor molecules on the electrode surface and/or the destabilization or delamination of the thin protective layer formed over the substrate. Additionally, a change in the adsorption or desorption rate of the inhibitor molecules or a local change in the inhibition mechanism for the anodic reaction33 could be reasons for the appearance of these shoulders.

EIS Studies

EIS is widely employed to provide useful information about the kinetics and mechanisms of electrochemical systems34,35,36,37. Figure 3 shows a typical one time constant equivalent electrical circuit that is used to analyze all the measured EIS data in this study. The used one time constant equivalent circuit indicate that the adsorbed inhibitor forms a monolayer on the MS surface38.

Figure 3
figure 3

Equivalent circuit used to fit the measured impedance spectra for MS in 5 M HCl.

It consists of a charge transfer resistance (Rct), solution resistance (Rs), and constant phase element (CPE) used to describe the non-ideal behavior of the double layer, which is mainly attributed to non-uniform surface coverage and/or surface roughness.

The impedance of the CPE is analyzed using Equation 439:

$${Z}_{Q}={[{Y}_{0}{(jw)}^{n}]}^{-1}$$
(4)

where ZQ represents the CPE impedance (Ω cm−2), Y0 is the CPE constant, ω is the frequency in rad s−1, and the values of n range between 0 and 1 and define the divergence from capacitance linearity. When n = 1, Y0 is equivalent to that of an ideal capacitor. When n = 0, Y0 is equivalent to that of a resistor.

The measured (dotted lines) and fitted (solid lines) impedance spectra shown in Fig. 4 are for MS in a LHP-free 5 M HCl solution at 20, 40, 60, and 80 °C. It is clear that the semicircles decrease in size as the temperature increases.

Figure 4
figure 4

EIS: (a) Nyquist plots and (b) Bode plots for MS in a LHP – free 5 M HCl at various temperatures.

Figures 5 and 6 show the Nyquist and Bode plots, respectively, of MS in a 5 M HCl solution with various concentrations of LHP at (b) 20, (c) 40, (d) 60, and (e) 80 °C. It can be noticed from the Nyquist plots that the diameter of the semicircle increases as the corrosion inhibitor concentration increases at different temperatures. The low impedance modulus (Z) increases in line with the increase in the LHP concentration due to the increased amount adsorbed of LHP on the metallic substrate. It can be seen from the Bode plots, that the values of phase angle (θ), for the inhibited samples are higher than that of the uninhibited MS at elevated temperature. The increased values of the phase angle for the specimens in the inhibited solutions indicated that the metallic surface significantly becomes smooth due to the formation of protective layer by the adsorbed inhibitors over the MS surface.

Figure 5
figure 5

EIS Nyquist plots MS in 5 M HCl with various concentrations of LHP at (A) 20, (B) 40, (C) 60, and (D) 80 °C.

Figure 6
figure 6

EIS Bode for MS in 5 M HCl in the presence of 92, 185, 277, and 370 µmol L−1 of LHP at (a) 20, (b) 40, (c) 60, and (d) 80 °C. The symbols are the measured data and the solid lines are the fittings using the equivalent circuit shown in Fig. 3.

Table 3 shows the EIS parameters derived from Figs 46 using the equivalent circuit shown in Fig. 3. The IE% is calculated using Equation (5)40:

$$IE \% =(\frac{{R}_{{\rm{ct}}2}-{R}_{{\rm{ct}}1}}{{R}_{{\rm{ct}}1}})\times 100$$
(5)

where Rct2 and Rct1 represent the charge transfer resistances in the inhibited and uninhibited test solutions, respectively. θ is calculated using Equation 2.

Table 3 EIS parameters at various temperatures for MS in 5 M HCl with various concentrations of LHP.

Rct, Y0, and n are used to calculate the metal solution interface double layer capacitance (Cdl) using Equation (6)41:

$${C}_{{\rm{dl}}}={({Y}_{0}{{R}_{ct}}^{1-n})}^{1/n}$$
(6)

Table 3 reveals that Rct increases as LHP concentration is raised. This is attributed to the adsorbed molecules of LHP on the MS forming a barrier film, which restricts the accessibility of Cl ions to the metal surface. On the contrary, the double layer capacitance (Cdl) gradually decreases as the inhibitor concentration increases. The decrease in the Cdl is primarily caused by (i) the increase in the thickness of the double layer thickness (δ) because of the adsorption of the LHP onto the MS surface and/or (ii) the decrease in the dielectric constant (ε) due to the replacement of water molecules by the LHP ones as indicated by the Helmholtz equation42:

$${C}_{{\rm{dl}}}=\frac{{\rm{\varepsilon }}\,{{\rm{\varepsilon }}}_{0}\,A}{\delta }$$
(7)

where A represents the cross-sectional area of the electrode, and ε0 and ε represent the dielectric constants of air and water, respectively.

As the IE% is directly proportional to Rct2, it also increases as the LHP concentration increases and decreases as the temperature is raised, as seen in Table 3.

It is worth noting that n has the highest values in LHP-free solution at any temperature, and decreases as the LHP dosage increases at a given temperature. In addition, the higher the temperature is, the higher is the value of n at any concentration of LHP. The values of n (approaching unity), indicate that the CPE is getting closer to the ideal capacitor behavior. It is noteworthy that the EIS parameters summarized in Table 3, are consistent with the Tafel analyses shown in Table 2.

Adsorption Isotherm and Thermodynamic Calculations

Adsorption isotherms are used extensively to illustrate and characterize the interaction between the applied corrosion inhibitor and a metallic substrate.

Langmuir is found to be the best isotherm by far that fits the measured experimental data. The Langmuir isotherm relates Cinh and θ as follows43:

$$\frac{{C}_{{\rm{inh}}}}{\theta }=\frac{1}{{K}_{{\rm{ads}}}}+{C}_{{\rm{inh}}}$$
(8)

where Cinh is the LHP concentration and Kads is the adsorption equilibrium constant, which can be obtained from the intercept of the plots shown in Fig. 7.

Figure 7
figure 7

Langmuir adsorption plots at various temperatures for MS in 5 M HCl.

Figure 7 shows the relation between Cinh/θ and Cinh at various temperatures. The standard Gibbs free energy change of adsorption (\({\rm{\Delta }}{G}_{{\rm{ads}}}^{o}\)) can be readily obtained using Equation (9) after getting the constants of adsorption (Kads) at different temperatures from the (Cinh/θ)-intercepts of the plots in Fig. 7.

$${K}_{{\rm{ads}}}=\frac{1}{55.5}{e}^{-\frac{{\rm{\Delta }}{G}_{{\rm{ads}}}^{{\rm{o}}}}{RT}}$$
(9)

where R equals 8.314 j mol−1 K−1, T is the temperature, and the 55.5 is the number of moles of water in 1 liter20.

By plotting ln Kads versus T−1, as shown in Fig. 8, a straight line is obtained, which follows the van’t Hoff equation19:

$$\mathrm{ln}\,{K}_{{\rm{ads}}}=\frac{-{\rm{\Delta }}{{H}^{0}}_{{\rm{ads}}}}{RT}+\frac{{\rm{\Delta }}{{S}^{0}}_{{\rm{ads}}}}{R}$$
(10)

where \({\rm{\Delta }}{S}_{{\rm{ads}}}^{0}\) and \({\rm{\Delta }}{H}_{{\rm{ads}}}^{0}\) are change in the entropy and standard enthalpy of adsorption, respectively.

Figure 8
figure 8

ln Kads vs. T−1 for the LHP at the MS surface in 5 M HCl.

By determining \({\rm{\Delta }}{H}_{{\rm{ads}}}^{0}\) from Equation (10), \({\rm{\Delta }}{S}_{{\rm{ads}}}^{0}\) is calculated utilizing either Equation (10)44 or Equation (11).

$${\rm{\Delta }}{G}_{{\rm{ads}}}^{0}={\rm{\Delta }}{H}_{{\rm{ads}}}^{0}\,-\,T{\rm{\Delta }}{S}_{{\rm{ads}}}^{0}$$
(11)

Table 4 lists the \({\rm{\Delta }}{H}_{{\rm{ads}}}^{0}\), \({\rm{\Delta }}{S}_{{\rm{ads}}}^{0}\), Kads, and \({\rm{\Delta }}{G}_{{\rm{ads}}}^{o}\) values for the adsorption of LHP at the MS surface in 5 M HCl.

Table 4 Thermodynamic parameters derived from Figs 7 and 8.

The positive values of \({\rm{\Delta }}{S}_{{\rm{ads}}}^{0}\) indicate an increase in the entropy due to the adsorption of LHP on the MS surface. This is attributed to the exothermic nature of the adsorption process, as seen from the negative values of \({\rm{\Delta }}{H}_{{\rm{ads}}}^{0}\). The high values for the constant of adsorption (Kads), particularly at 25 °C, for the studied LHP corrosion inhibitor indicate strong adsorption on the MS substrate. Large Kads values suggest a strong adsorption tendency and hence a better inhibition performance. This is explained by the presence of π-electrons in the inhibitor’s inherent molecular structure.

The negative values of \({\rm{\Delta }}{G}_{{\rm{ads}}}^{0}\) are aligned with the spontaneity of the LHP adsorption on the MS surface (as -RT ln Kads has negative values too). The values of \({\rm{\Delta }}{G}_{{\rm{ads}}}^{0}\) are usually interpreted in relation to the nature of the adsorption process: whether it is physisorption or chemisorption. Generally, if the values of \({\rm{\Delta }}{G}_{{\rm{ads}}}^{0}\) are more than −20 kJ mol−1, then the physisorption mechanism is favored, whereas if the values of \({\rm{\Delta }}{G}_{{\rm{ads}}}^{0}\) are −40 kJ mol−1 or lower, then the adsorption process is chemisorption.

As seen in Table 4, the obtained \({\rm{\Delta }}{G}_{{\rm{ads}}}^{0}\) ranges from −33.1 to −34.8 kJ mol−1, which is between −20 and −40 kJ mol−1. Thus, the process cannot be classified as chemisorption or physisorption. Rather, it is a mix of chemisorption and physisorption.

Effect of Activation Energy and Temperature on the Corrosion Rate

The inhibition mechanism and efficiency are directly influenced by the activation energy (Ea). The rate of most chemical reactions tends to increase as the temperature increases. The effect of temperature on the corrosion rate of MS can be evaluated using the Arrhenius equation:

$$\mathrm{log}\,CR=\,\mathrm{log}\,A-\frac{\,{E}_{{\rm{a}}}}{2.303\,RT}$$
(12)

where CR is the corrosion rate expressed in terms of icorr at a specific temperature (T), (Ea) is the activation energy, R = 8.314 J mol−1 K−1, and A is the Arrhenius constant, which is affected by the metal type and electrolyte composition.

The Arrhenius plots of the MS before and after the addition of the corrosion inhibitor are given in Fig. 9, with a linear regression rate that is close to unity.

Figure 9
figure 9

Arrhenius plots for MS in 5 M HCl with various concentrations of LHP.

The calculated Ea are listed in Table 5. Ea increases as the inhibitor concentration is raised. The higher value of Ea is referred to the formed barrier film of LHP molecules at the MS surface.

Table 5 Activation energy (Ea), regression coefficient (r2), ∆H*, and ∆S* for MS in 5 M HCl with various concentrations of LHP.

The entropy of activation (∆S*) and the enthalpy of activation (∆H*) due to the dissolution of MS in 5 M HCl are calculated using the Arrhenius equation33:

$$CR=\frac{RT}{Nh}{\exp }^{\frac{{\rm{\Delta }}{S}^{\ast }}{R}}{\exp }^{\frac{-{\rm{\Delta }}{H}^{\ast }}{RT}}$$
(13)

where h is Planck’s constant and N is Avogadro’s number.

From Fig. 10, ∆H* and ∆S* are obtained from the slope of (∆H*/2.303R) and the intercept of [log (R/Nh) + (∆S*/2.303R)] are used to calculate respectively. The average difference between Ea and ∆H* is approximately 2.6 kJ mol−1 for each test, which is almost the value of RT (2.63 kJ mol−1). This indicates that the dissolution of MS in this environment is a unimolecular reaction.

Figure 10
figure 10

Transition-state plots for MS in 5 M HCl with various concentrations of LHP.

The thermodynamic parameters of LHP inhibitor were compared with some other reported inhibitors in 5 M HCl as shown in Table 6. It is worth to mention that the reported inhibitors herein, exhibited a mixed type of adsorption.

Table 6 Comparison of the energetic parameters of the investigated inhibitor (LHP), and some other reported inhibitors in 5 M HCl.

Surface Topography and Characterization

SEM analysis

Figure 11 depicts SEM micrographs that show the morphology of the MS surface after immersion for 24 h in (a) a LHP -free 5 M HCl and (b) 5 M HCl with 370 µmol L−1 of LHP at 20 °C. Figure 11a shows that the examined surface is heavily corroded. However, for the immersion in the presence of 370 µmol L−1 of LHP, the corrosion looks more uniform.

Figure 11
figure 11

SEM surface analysis micrograph for (a) MS immersed in 5 M HCl and (b) after the addition of 370 µmol L−1 of LHP at 20 °C for 24 h.

EDX is conducted to examine the presence of nitrogen on the corroded surfaces in HCl with 370 µmol L−1 of the LHP at 20 °C. The presence of nitrogen on the metallic substrate is confirmed to be 1.9%, which demonstrates the adsorption of LHP on the MS surface.

XPS analysis

The XPS survey of the adsorbed inhibitor and the high-resolution XPS analysis of C, O, and N are shown in Fig. 12 for MS after immersion in 5 M HCl with 370 µmol L−1 of LHP O at 20 °C. The C 1 s is deconvoluted into three different peaks. A peak at 284.7 eV is related to the C-C of the adsorbed inhibitor45. In addition, a second peak at 288.1 eV is credited to the existence of -C=O groups46,47 and a third one at 286.2 eV for the C-N of the adsorbed LHP on the protected MS surface48. The deconvolution of the O 1 s spectrum yields three peaks. The first at 529.6 eV is related to O2−, which is mainly is associated with oxygen atoms bonded to Fe2O349. The second one at 531.5 eV is attributed to the OH of FeOOH50. The third peak observed at 533.5 eV could be ascribed to the existence of oxygen in the adsorbed water4. On the contrary, the deconvolution of the N 1 s spectrum peak results in two peaks around 399.8 and 402.3 eV, which are mainly referred to the N atoms bonded to the MS surface (N-Fe) and to the protonated nitrogen atoms of the hydrazine group, respectively17. XPS confirms the presence of the adsorbed LHP corrosion inhibitor on the metal surface.

Figure 12
figure 12

(a) XPS survey scan composition of the MS immersed in 5 M HCl with 370 µmol L−1 of LHP at 20 °C for 24 h and the profiles of (b) C 1 s, (c) O 1 s, and N 1 s.

Atomic force microscopy analysis

AFM is a powerful technique that has been widely used to explore the effect of inhibitor on the surface roughness and topography of the metal surface in aggressive media33,51,52,53. Figure 13 shows three-dimensional images of two MS samples that had been ground using SiC to 4000 grit and immersed in 5 M HCl for 24 h with and without 370 µmol L−1 of LHP. The average roughness of a MS surface in the absence and presence of LHP in 5 M HCl is approximately 43.4 and 16.4 nm, respectively. The appreciable decrease in the surface roughness after the addition of LHP demonstrates the good inhibition performance of LHP in 5 M HCl.

Figure 13
figure 13

AFM images for (a) the polished MS surface, (b) after immersion in 5 M HCl and (c) after the addition of LHP for 24 h at 20 °C.

Structural Geometrical Parameters of the Optimized Molecule

The optimized geometry of the studied LHP results in the structure shown in Fig. 14. The optimized geometry shows no imaginary frequencies, which suggests the stability of the converged structure.

Figure 14
figure 14

Optimized geometry of LHP using the wb97xd/6-311++ g(d, p) level of the calculations.

The asymmetric and symmetric vibrational bands of C=O are located at 1812 and 1827 cm−1, respectively. The vibrational stretching bands of the two N-H groups are located at 3604 and 3658 cm−1, respectively. The aliphatic straight chain with 10 C atoms has similar C-C bonds with an average value of 1.526 Å for the C-C bonds between C1 and C10, whereas the C10-C11, C11-C12, C13-C14, and C13-C15 bond lengths are 1.536, 1.511, 1.532, and 1.514 Å, respectively. Both C15=O and C12=O have the same bond length of 1.21 Å. The C15-N2 bond length is slightly shorter than the C12-N1 bond length, by approximately 0.015 Å, with values of 1.376 and 1.391 Å, respectively. The N1-N2 bond length is 1.376 Å. The bond angles C11-C12-O1, O1-C12-N1, N1-C12-C11, C13-C15-O2, O2-C15-N2, and N2-C15-C13 are 123.7°, 119.3°, 116.8°, 123.3°, 122.1°, and 114.6°, respectively, suggesting the sp2 hybridizations whereby C=O bonds take up more space and consequently slightly decrease the N-C-C bond angle. The dihedral angles O2-C15-N2-N1 and O1-C12-N1-N2 are −1.44° and −165.35°, suggesting more deviation than planarity on the side of the longer aliphatic chain.

The natural bond order charge analysis showed that the most negative atoms are O1 and O2 with values of −0.606 and −0.618, followed by N1 and N2 with values of −0.481 and −0.472, respectively. The most positive atoms are C12 and C13, with charges of 0.712 and 0.686, respectively. All the other C atoms in CH2 have a partial negative charge of approximately −0.4 that is neutralized by a positive charge of approximately 0.2 on each H atom. The terminal C in CH3 carries a larger partial negative charge of approximately −0.577, because it is surrounded by three H atoms each, which have a positive charge of approximately 0.19. The nonlocalized negative charges suggest that more than one interaction site with the metal surface is involved, as opposed to the possibility of a single active site in the adsorption process. The total dipole moment is 5.7294 D, directed approximately above the C9 atom and perpendicular to the C8-C9 bond, which further suggests the non-localized distribution of atomic charges.

Molecular simulation

The molecular simulation of LHP on different pure iron surfaces showed approximately identical adsorbed molecular geometries to that shown in Fig. 15. The whole molecule is adsorbed on the surface, with bond lengths of 4.9, 3.2, 2.9, and 3.1 Å for methyl, CH2, O, and N, respectively, as measured relative to the nearest Fe atom on the surface. The molecular simulation suggested that adsorption sites are not localized on one atom, and that the part with the heteroatoms are much more strongly adsorbed than aliphatic hydrocarbon. Upon comparing the geometry of the isolated LHP with the adsorbed one, the main geometrical changes are localized in C15-N2, C12-N1 and N1-N2 bond lengths with 1.359, 1.374, and 1.412 Å, respectively. Both C15-N2 and C12-N1 have shortened by 0.017 Å, which results in elongation of N1-N2 by 0.035 Å. Furthermore, the calculation showed rigid adsorption energies of −2.914, −2.635, and −1.907 kcal mol−1 for iron surfaces with orientations of 100, 110, and 111, respectively. Although the values of the adsorption energy are not very different, the adsorption energy values may lead to the conclusion that the 100-iron surface will have the best coverage of the three surfaces, and that the 111 surface will have the weakest coverage.

Figure 15
figure 15

Adsorption of LHP on the 100 Fe surface; bond lengths are in Å.

Summary and Conclusion

The LHP eco-friendly corrosion inhibitor has shown extraordinary performance at normal and elevated temperatures in inhibiting the corrosion of MS in 5 M HCl. The efficiency was found to increase as the inhibitor concentration increased. When the testing temperature increased to 80 °C, a slight decline in efficiency was noted, due to the increased rate of iron dissolution in the acidic brine and the expected desorption of the attached inhibitor from the metallic substrate. An inhibition efficiency of more than 94% was successfully achieved upon the addition of the candidate inhibitor to a concentration of 370 µmol L−1. The relationship between the surface coverage and inhibitor concentration is consistent with Langmuir’s adsorption isotherm. LHP is found to be a mixed type inhibitor. The adsorption of the inhibitor on the steel surface was proven to be chemi–physisorption, which is confirmed by the calculated standard Gibbs free energy (\({\rm{\Delta }}{G}_{{\rm{ads}}}^{0}\)). The optimized geometry shows no imaginary frequencies, which suggests the stability of the converged structure. In addition, the molecular simulation of LHP proved that the 100-iron surface will have better coverage compared to the 110 surface, while the 111 surface will have the weakest coverage. AFM analyses revealed a significant decrease in the surface roughness under an inhibited condition compared to immersion with no inhibitor added. Additionally, EDS and XPS revealed the presence of nitrogen on the MS, which validates the adsorption of LHP onto the surface. Applications of this inhibitor for the mitigation of acidic corrosion in the petroleum and gas industry will add value in terms of enabling safer operations and contributing to significant cost savings. Further qualification testing of the candidate inhibitor under dynamic conditions is recommended to achieve greater assurance and gain more confidence in the inhibition capabilities.