Corrosion Science 51 (2009) 2098–2106

Contents lists available at ScienceDirect

Corrosion Science

journal homepage: www.elsevier .com/locate /corsc i

Dry and wet lab studies for some benzotriazole derivatives as possiblecorrosion inhibitors for copper in 1.0 M HNO3

K.F. Khaled a,*, Mohammed A. Amin b

a Electrochemistry Research Laboratory, Chemistry Department, Faculty of Education, Ain Shams University, Roxy, Cairo 11711, Egyptb Chemistry Department, Faculty of Science, Ain Shams University, Abbassia, Cairo 11566, Egypt

a r t i c l e i n f o

Article history:Received 1 April 2009Accepted 29 May 2009Available online 7 June 2009

Keywords:A. CopperB. In silicoB. PolarizationB. EISC. Acid inhibition

0010-938X/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.corsci.2009.05.038

* Corresponding author. Tel.: +20 227650485; fax:E-mail address: khaledrice2003@yahoo.com (K.F. K

a b s t r a c t

With contrast to the traditional techniques of identifying and synthesizing new corrosion inhibitors inwet lab, a prior dry-lab process, followed by a wet-lab process is suggested by using cheminformaticstools. Quantum chemical method is used to explore the relationship between the inhibitor molecularproperties and its inhibition efficiency. The density function theory (DFT) is also used to study the struc-tural properties of three selected benzotriazole derivatives namely, benzotriazole-1-carboxamide (BCA),1H-benzotriazole-1-acetonitrile (BAN) and benzotriazole-1-carbonyl chloride (BCC) in aqueous phase. Itis found that when the benzotriazole derivatives adsorb on the copper surface, molecular structure influ-ences their interaction mechanism. The inhibition efficiencies of these compounds showed a certain rela-tionship to highest occupied molecular orbital (HOMO) energy, Mulliken atomic charges and Fukuiindices. A wet lab study has been carried out using weight loss, Tafel polarization and impedance mea-surements to evaluate their inhibition performance in 1.0 M HNO3 solutions at 25 �C. Adsorption takesplace by a direct chemisorption on the exposed copper surface, while it most probably occurs via hydro-gen bonding on the oxidized surface. BCA was the most effective among the tested inhibitors, while BCCwas less effective than BAN. Results obtained from dry-lab process are in good agreement with thoserecorded from wet-lab experiments.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Copper is a metal that has a wide range of applications due to itsgood properties. It is used in electronics, for production of wires,sheets, tubes, and also to form alloys. The use of copper corrosioninhibitors in acid solutions is usually to minimize the corrosion ofcopper during the acid cleaning and descaling. The possibility ofthe copper corrosion prevention has attracted many researchersso until now numerous possible inhibitors have been investigated[1–3].

The inhibiting action of these inhibitors is attributed to theiradsorption to the metal/solution interface. It has been observedthat adsorption depends mainly on certain physico-chemical prop-erties of the inhibitor group. Like functional groups, aromaticity,electron density at the donor atoms and p-orbital character ofdonating electrons and also the presence of hetero-atom such asN, O and S, as well as multiple bonds in their molecular structure,are assumed to be active centers of adsorption [4].

Traditionally, scientists identify new corrosion inhibitor mole-cules either by fiddling with existing inhibitors or by testing hun-dreds of compounds in laboratory, which is a laborious, expensive

ll rights reserved.

+20 222581237.haled).

and time-consuming process. Thus, it became necessary to developnew inhibitors which behave like corrosion inhibitors in less time,by in silico approach [5]. As computing power has grown exponen-tially, this approach to research often referred to as in silico (as op-posed to in vitro), has amassed more attention especially in thearea of cheminformatics.

DFT (density functional theory) methods have become verypopular in the last decade due to their accuracy that similar toother methods in less time and with a smaller investment fromthe computational point of view [6]. In agreement with the DFT,the energy of the fundamental state of polyelectronic systemscan be expressed through the total electronic density, and in factthe use of the electronic density instead of the wave function forthe calculation of the energy constitutes the fundamental base ofDFT [6]. The local reactivity of the molecules was analyzed throughan evaluation of the Fukui indices [7]. These are a measurement ofthe chemical reactivity, as well as an indicative of the reactive re-gions and the nucleophilic and electrophilic behaviour of the mol-ecule. The Fukui function, f ð~rÞ, is defined as the derivative of theelectronic density, qð~rÞ, with respect to the number of electrons,N, at a constant external potential, tð~rÞ

f ð~rÞ ¼ @qð~rÞ@N

� �tð~rÞ

ð1Þ

K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106 2099

If the effects of relaxation associated with the addition or removalof electronic charges are not considered, then:

qþð~rÞ � qLUMOð~rÞ ð2Þq�ð~rÞ � qHOMOð~rÞ ð3Þ

where qLUMOð~rÞ is the density of the lowest unoccupied molecularorbital and qHOMOð~rÞ the density of the highest occupied molecularorbital [8].

The condensed Fukui functions [7] are found by taking the finitedifference approximations from Mulliken population analysis ofatoms in benzotriazole derivatives, depending on the direction ofthe electron transfer

fþk ¼ qkðN þ 1Þ � qkðNÞ ðfor nucleophilic attackÞ ð4Þf�k ¼ qkðNÞ � qkðN � 1Þ ðfor electrophilic attackÞ ð5Þ

f ok ¼

qkðN þ 1Þ � qkðN � 1Þ2

ðfor radial attackÞ ð6Þ

where qk is the gross charge of atom k in the molecule.The objective of this work is to identify three selected benzotri-

azole derivatives, namely benzotriazole-1-carboxamide (BCA), 1H-benzotriazole-1-acetonitrile (BAN) and benzotriazole-1-carbonylchloride (BCC) in dry-lab process as possible corrosion inhibitorsfor copper in 1.0 M HNO3. Geometric structure, total negativecharge on the molecule, highest occupied molecular orbital (EHO-

MO), lowest unoccupied molecular orbital (ELUMO), dipole moment(l), charges on atoms, total energy and reactivity of the moleculeswas analyzed through an evaluation of the Fukui indices. Theseparameters have been calculated using DFT (density function the-ory) methods. From the results obtained in the dry lab, we couldidentify the studied molecules as possible corrosion inhibitors forcopper in 1.0 M HNO3 in our laboratory, which investigated exper-imentally by weight loss, Tafel polarization and electrochemicalimpedance spectroscopy, EIS, measurements.

2. Experimental

2.1. Weight loss measurements

The structures of the heterocyclic compounds studied are pre-sented below:

All the investigated compounds were obtained from AldrichChemical Co., they were put in 1.0 M HNO3 (Fisher Scientific) with-out pre-treatment at concentrations of 10�4 M, 5 � 10�4 M, 10�3 Mand 5 � 10�3 M. The weight loss measurements were carried out ina 250 ml capacity glass beaker placed in a thermostat water bath.The used copper coupons (Puratronic, 99.999%) had a rectangularform (length = 2.5 cm, width = 2 cm, thickness = 0.05 cm). Priorto all measurements, the coupons were etched in a 7.0 M HNO3

solution for 15 s (the nitric acid etching method provided a freshand active (oxide-free) copper surface [9]), then washed thor-oughly with bi-distilled water, degreased with acetone, washedagain with bi-distilled water, and finally dried at roomtemperature.

The coupons were weighed and suspended in 100 ml of an aer-ated 1.0 M HNO3 solution containing benzotriazole derivatives at

the desired concentrations for 36 h exposure period of time at 25± 1 �C. At the end of the tests, the coupons were taken out, washedwith bi-distilled water, degreased with acetone, washed again withbi-distilled water, dried, and then weighed using an analytical bal-ance (precision: ±0.1 mg). Three measurements were performed ineach case and the mean value of the weight loss has been reported.

2.2. Electrochemical measurements

Copper specimens from Johnson Mattey (Puratronic, 99.999%)were mounted in Teflon. An epoxy resin was used to fill the spacebetween Teflon and copper electrode. The circular cross sectionalarea of the copper rod exposed to the corrosive medium, that usedin electrochemical measurements was (0.28 cm2). The electro-chemical measurements were performed in a typical three-com-partment glass cell consisted of the copper specimen as workingelectrode (WE), platinum counter electrode (CE), and a saturatedcalomel electrode (SCE), as the reference electrode. The counterelectrode was separated from the working electrode compartmentby fritted glass. The reference electrode was connected to a Luggincapillary to minimize IR drop. Solutions were prepared from bi-dis-tilled water of resistivity 13 MX cm.

Prior to each experiment, the surface pre-treatment of theworking electrode was performed by mechanical polishing (usinga polishing machine model POLIMENT I, BUEHLER POLISHER) ofthe electrode surface with successive grades of emery papers upto 4/0 grade, cleaned with acetone, washed with bi-distilled waterand finally dried.

The electrode potential was allowed to stabilize for 60 min (thetime necessary to reach a quasi-stationary value for the open cir-cuit potential; known from the OCP vs time plots, data not in-cluded here) before starting the measurements. All experimentswere conducted at 25 ± 1 �C. The electrolyte solution was madefrom 1.0 M HNO3 (Fisher Scientific) and bi-distilled water. Theinhibitors (Aldrich Chemical Co.) used without any pre-treatment.Measurements were performed using Gamry Instrument Potentio-stat/Galvanostat/ZRA. This includes a Gamry framework systembased on the ESA400, Gamry applications that include dc105 fordc corrosion measurements, EIS300 for electrochemical impedancespectroscopy measurements along with a computer for collectingdata. Echem Analyst 4.0 Software was used for plotting, graphingand fitting data.

The potentiodynamic current–potential curves were obtainedby changing the electrode potential automatically from (�0.30up to 0.10 VSCE) at a scan rate of 1.0 mV s�1. EIS measurementswere carried out in a frequency range from 10 kHz to 10 mHz withan amplitude of 5 mV peak-to-peak using ac signals at open circuitpotential.

2.3. Theoretical study (FI calculations)

The FI calculations were performed using Materials Studio[10,11] DMol3 version 4.3.1, a high quality quantum mechanicscomputer program (available from Accelrys, San Diego, CA). Thesecalculations employed an ab initio, local density functional (LDF)method with a double numeric polarization (DNP) basis set and aBecke–Perdew (BP) functional. Dmol3 use a Mulliken populationanalysis [12].

3. Results and discussion

3.1. Dry lab study

Highest occupied molecular orbital energy level (HOMO), as ameasure of electron donating ability of a molecule, explains the

Table 1Molecular properties of the three benzotriazole derivatives calculated with DFT method in aqueous phase.

Total energy ET/a.u. HOMO/eV LUMO/eV DE/eV g = DE/2/eV l/debyes

BCA �559.81 �6.076 �2.820 3.256 1.628 0.475BAN �522.93 �6.570 �2.700 3.870 1.935 3.54BCC �968.76 �6.812 �2.8541 3.958 1.979 4.794

Table 2Calculated Mulliken atomic charges and Fukui functions for the three benzotriazole derivatives.

Atom qN qNþ1 qN�1 fþk f�k f ok

BCA N(1) �0.235 �0.2200 �0.282 0.015 0.047 0.031N(2) �0.104 0.0350 �0.164 0.139 0.060 0.100N(3) �0.224 �0.1080 �0.306 0.116 0.082 0.099N(12) �0.437 �0.4090 �0.573 0.028 0.136 0.032O(11) �0.425 �0.3520 �0.549 0.073 0.124 0.032

BAN N(1) �0.199 �0.1740 �0.260 0.025 0.061 0.043N(2) �0.187 �0.0450 �0.259 0.142 0.072 0.107N(3) �0.245 �0.1480 �0.347 0.097 0.102 0.099N(12) �0.177 �0.1020 �0.266 0.075 0.089 0.077

BCC N(3) �0.233 �0.2360 �0.272 �0.003 0.039 0.018N(4) �0.020 0.1100 �0.087 0.130 0.067 0.099N(5) �0.177 �0.0560 �0.251 0.121 0.074 0.098O(2) �0.321 �0.2120 �0.377 0.109 0.056 0.083Cl(12) �0.027 0.1040 �0.133 0.131 0.106 0.119

2100 K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106

adsorption on metallic surfaces by the way of delocalized pair of pelectrons. A high value expresses intrinsic electron donating ten-dency to an appropriate acceptor, i.e. any molecule with lowerEHOMO and empty molecular orbital. On the contrary, the energyof the lowest unoccupied molecular orbital, ELUMO, signifies theelectron receiving tendency of a molecule. Accordingly, the differ-ence between ELUMO and EHOMO energy levels, the dipole moment, l,the electronic charge on nitrogen, oxygen and chlorine atoms, andthe total energy of inhibitor molecule (ET) were determined. Thedipole moment, l, is another way to obtain data on the electronicdistribution in a molecule and is one of the properties used tradi-tionally to discuss and rationalize the structure and reactivity ofmany chemical systems [13].

Calculations of the HOMO–LUMO gap (DE = ELUMO � EHOMO)[14] were done as well as the global hardness, g, that is approxi-mated as DE/2, and can be defined under the principle of chemicalhardness and softness (HSAB) [6,13]. Both parameters also provideinformation about the reactive behaviour of molecules and pre-sented in Table 1. With the purpose of having a wider knowledgeabout the local reactivity of the benzotriazole derivatives, the Fu-kui indices, FI, for each atom in the molecules have been calculated.An analysis of the Fukui indices along with the distribution ofcharges and the global hardness provides a more complete schemeof the reactivity of the studied molecules [15]. The calculated Fukuiindices for all the charged species (N + 1 and N � 1) as well as theneutrals (N) of each of the three benzotriazole derivatives are pre-sented in Table 2. For simplicity, only the charges and Fukui func-tions over the nitrogen (N), oxygen (O) and chlorine (Cl) atoms arepresented. It is well known that the phenomena of electrochemicalcorrosion appear in aqueous phase, and for this reason it is neces-sary to include in the computational calculations of the solventeffect.

In a similar way, it is important to take into account the effectsthat can appear as much in the geometric properties as in the elec-trical ones. Dmol3 includes certain COSMO [16] controls, which al-low for the treatment of solvation effects. The COnductor-likeScreening MOdel (COSMO) is a continuum solvation model inwhich the solute molecule forms a cavity within the dielectric con-tinuum of permittivity, e, that represents the solvent. The solvent

molecules are modeled as a continuous of uniform dielectric con-stant (e = 78.39 for water) and the solute is placed in a cavity with-in it [14]. The results for the total energies, HOMO and LUMOorbitals energies, DE, g and l calculated for each benzotriazolederivative in solution phase are presented in Table 1. All thesecomputational calculations have been performed in the aqueousphase. The optimized geometries of the benzotriazole derivativesas well as HOMO and LUMO molecular orbitals are shown in Fig. 1.

A Mulliken population analysis performed for benzotriazolederivatives showed that (BCA: N1, N2, N3, O11, N12), (BAN: N1,N2, N3, N12) and (BCC: N3, N4, N5, Cl12, O2), respectively, havea larger electronic density which could make them nucleophiliccenters when interacting with copper surfaces [15]. In Fig. 1, thereis a representation of the HOMO orbitals calculated in the presenceof a solvent (water). It is indicated from Fig. 1 that HOMO is locatedin the benzotriazole as well as the functional groups attached tothis ring. This would indicate that the preferred actives sites foran electrophilic attack are located within the region around thenitrogens belonging to the benzotriazole ring. In Table 1, a seriesof properties calculated for each one of the derivatives in aqueousphase are presented. One of these properties is the global hardness,g, calculated as DE/2, which is a parameter that gives importantinformation about the reactive behaviour of the molecule. In Table1, the derivative BCC exhibits the maximum hardness value,whereas the BCA presents the minimum value of hardness. There-fore, the expected reactivity order for the derivatives in aqueousphase is BCA > BAN > BCC. The close values of chemical hardness,g, indicate the potential use of these derivatives as possible corro-sion inhibitors with more potential for BCA derivatives over theother.

An analysis of the Fukui indices shown in Table 2, in which onlythe largest values are presented, it is possible to observe that incase of each one of the derivatives, the nitrogen atoms in the ben-zotriazole ring (N1, N2, N3) are the most susceptible sites for elec-trophilic attacks. These sites present the highest values in f�k whichrange from 0.039 to 0.102. In BCA there exists, in addition, otherprobable sites susceptible to electrophilic attack: N12, O11, withvalues of f�k equal to 0.136 and 0.124, respectively. Nevertheless,in BCA, the atom N12, demonstrated to be much more reactive

Fig. 1. Molecular orbital plots as well as active sites for electrophilic, nucleophilic and radial attack on benzotriazole derivatives.

K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106 2101

than the nitrogen atom in the benzotriazole ring, which indicatesthat the electrons pertaining to the N12 are more able to interactwith the copper surface [15]. As far as sites prone to nucleophilicattacks, in all case the N2 (N4 in case of BCC) atom presented thehighest values of fþk falling in a range of 0.130–0.142.

In order to support results obtained form the dry-lab processand gain more insight regarding the inhibition mechanism of thesecompounds, several traditional electrochemical techniques, includ-ing weight loss, Tafel extrapolation and impedance measurements,

were implied here to study the inhibition performance of thesecompounds towards copper corrosion in 1.0 M HNO3 solutions.

3.2. Wet lab study

3.2.1. Weight loss measurementsWeight loss measurements provide the most reliable results

concerning the efficiency of a given inhibitor compound, so thatthe corresponding corrosion data obtained from this approach

Table 3Electrochemical parameters, inhibition efficiencies and corrosion rates obtained from Tafel polarization and weight loss measurements for copper in 1.0 M HNO3 solutions inabsence and presence of various concentrations of the three benzotriazole derivatives at 25 �C.

[Inhib]/M jcorr � 106/A cm�2 ba/V dec�1 Corrosion rate (tTafel)lm y�1 nTafel (%) Corrosion rate (w)/mg cm�2 h�1 nw (%)

Blank 12.00 0.106 0.140 – 0.34 –BCA 10�4 6.35 0.105 0.074 47.12 0.16 52.78

5 � 10�4 2.64 0.107 0.031 78.01 0.07 80.3510�3 2.26 0.104 0.026 81.15 0.05 85.635 � 10�3 0.94 0.106 0.011 92.14 0.02 93.54

BAN 10�4 7.22 0.104 0.084 39.82 0.20 42.225 � 10�4 3.48 0.105 0.041 71.03 0.09 74.4810�3 2.87 0.106 0.033 76.09 0.07 79.765 � 10�3 1.80 0.104 0.021 84.99 0.05 86.51

BCC 10�4 7.79 0.103 0.091 35.07 0.21 39.585 � 10�4 4.06 0.104 0.047 66.14 0.10 71.5510�3 3.37 0.107 0.039 71.90 0.08 75.365 � 10�3 2.03 0.106 0.024 83.07 0.05 85.04

-0.30

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

-7 -6 -5 -4 -3

βa = 0.106 V dec-1

Extrapolated anodicTafel line

log (j / A cm-2)

E / V

(SC

E)Anodic Tafel region

An extrapolationstarting from E corr

jcorr = 12 x 10-6 A cm-2

One decade

0.106 V

Cathodic Tafel regionis absent

Fig. 2. Determining corrosion current and anodic Tafel slope, via extrapolation ofthe anodic Tafel region, for copper in 1.0 M HNO3 solution at a scan rate of1.0 mV s�1 and 25 �C.

2102 K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106

service conditions more accurately than the data obtained withany other test. A series of weight loss measurements was carriedout in a 1.0 M HNO3 free solution and in solutions containing dif-ferent concentrations of benzotriazole derivatives. Corrosion ratesof copper coupons, designated here as w, expressed inmg cm�2 h�1, as well as the inhibition efficiency, nw (%), were cal-culated over the exposure time period, based on Eq. (7), and listedin Table 3:

nwð%Þ ¼ ½ðWo �WÞ=Wo� � 100 ð7Þ

where Wo and W are the weight losses without and with benzotri-azole derivatives, respectively.

The standard deviation for these parameters was 1.0%. It followsfrom the data of Table 3 that, in all cases, the corrosion rate is sup-pressed, and therefore the corrosion inhibition strengthened, withincrease in inhibitor concentration. This trend may result from thefact that adsorption and surface coverage increases with the in-crease in inhibitors, concentration; thus the surface is efficientlyseparated from the medium [16].

It is possible that these inhibitors are first adsorbed on the elec-trode surface, impeding corrosion either by merely blocking, aswill be discussed, the reaction sites (anodic and cathodic) or byaltering the mechanism of the anodic and cathodic processes[17]. The results presented in Table 3 indicate that all tested ben-zotriazole derivatives inhibit the corrosion of copper in 1.0 MHNO3 solutions to an extent depending on the concentration andtype of the introduced inhibitor. This is clearly seen from the inhi-bition efficiency values depicted in Table 3. Increasing the concen-tration of each benzotriazole derivative enhances the inhibitionefficiency up to a maximum value of 93.54% at 5 � 10�3 M BCA,86.51% at 5 � 10�3 M BAN, and 85.04% at 5 � 10�3 M BCC. Theseresults, in agreement with theoretical study, clearly indicate thatBCA is better inhibitor than BAN and BCC.

3.2.2. Tafel polarization measurementsFig. 2 presents typical cathodic and anodic polarization curves

recorded for copper after 1.0 h of immersion in 1.0 M HNO3 solutionat a scan rate of 1.0 mV s�1 at 25 �C. It is obvious from Fig. 2 that theelectrode without inhibitor treatment, presented a monotonic in-crease of current with potential, and displayed a linear region from�0.03 to 0.10 V (SCE). The linear relation between potential, E, andlog j in this active dissolution region gives an apparent Tafel slope ofabout 106 mV dec�1. Corrosion current density, jcorr, was calculated,as will be fully discussed, via extrapolation of the anodic Tafel re-gion to the corrosion potential, Ecorr, inspect Fig. 2. The recorded jcorr

value by this method was 12 � 10�6 A cm�2. This linear relationsuggests that no passive films are formed on the copper surface.Consequently, copper may directly dissolve in 1.0 M HNO3 solu-

tions. The Pourbaix diagram for copper–water system [18], pre-sented in Fig. 3, indicates that copper is corroded to Cu2+ in HNO3

solutions, and no oxide film is formed to protect the surface fromcorrosion. Copper dissolution is thus expected to be the dominantreaction in HNO3 solutions.

Fig. 4 shows cathodic and anodic polarization curves recordedfor copper in 1.0 M HNO3 without and with different concentra-tions of the three selected benzotriazole derivatives at a scan rateof 1.0 mV s�1 at 25 �C. By comparing polarization curves in the ab-sence and presence of various concentrations of the three inhibi-tors, it is observed that, in all cases, increase in inhibitorconcentration reduces both the anodic and cathodic current densi-ties. In addition, no definite shift in corrosion potential, Ecorr, wasobserved, therefore values of Ecorr are not needed to be recordedhere. These results therefore reveal that, the individual presenceof these three benzotriazoles in HNO3 solutions inhibits both theanodic and cathodic processes (mixed-inhibitors). Based on theseresults, the action of these compounds may be related to adsorp-tion and formation of a barrier film on the copper surface, imped-ing both cathodic and anodic sites.

It is well known that the corrosion of copper in nitric acid in-volves reduction of the nitrate ion following Eqs. (8) and (9):

NO�3þ3Hþþ2e� ! HNO2þH2O ð8ÞNO�3 þ 4Hþ þ 3e� ! NOþ 2H2O ð9Þ

-0.30

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

-7 -6 -5 -4 -3

2

5 43

1

5 4 3 2

log (j / A cm -2)

E / V

(SC

E)

1(a)

-0.30

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

-6 -5 -4 -3

2

5 4 3

1

5 4 3 2

log (j / A cm -2)

E / V

(SC

E)

1(b)

-0.30

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

-6 -5 -4 -3

25 4 3

1

5 4 3 2

log (j / A cm -2)

E / V

(SC

E)

1(c)

Fig. 4. Anodic and cathodic Tafel polarization curves recorded for copper in 1.0 MHNO3 solutions in the absence and presence of various concentrations of the threebenzotriazole derivatives at a scan rate of 1.0 mV s�1 at 25 �C. (a) BCA; (b) BAN; (c)BCC. (1) Blank; (2) 10�4 M; (3) 5 � 10�4 M; (2) 10�3 M; (2) 5 � 10�3 M.

Fig. 3. Potential-pH equilibrium diagram for the system, copper–water, at 25 ± 1 �C[18,19]. The dash regions and the equilibrium components corresponding to thePourbaix diagram in HNO3.

K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106 2103

However, in aerated nitric acid solutions, dissolved oxygen may alsobe reduced on copper surface and this will enable some corrosion totake place [19–21]. It is a good approximation to ignore the hydro-gen evolution reaction and only consider oxygen reduction in theaerated nitric acid solutions at potentials near the corrosion poten-tial [22]. Cathodic reduction of oxygen can be expressed either by adirect 4e� transfer, Eq. (10):

O2þ4Hþþ4e� ! 2H2O ð10Þ

or by two consecutive 2e� steps involving a reduction to hydrogenperoxide first, Eq. (11), followed by a further reduction, according toEq. (12) [23]:

O2þ2Hþþ2e� ! H2O2 ð11ÞH2O2þ2Hþþ2e� ! 2H2O ð12Þ

The transfer of oxygen from the bulk solution to the copper/solutioninterface will strongly affect rate of oxygen reduction reaction, de-spite how oxygen reduction takes place, either in 4e� transfer ortwo consecutive 2e� transfer steps. Dissolution of copper in nitricacid is described by the following two continuous steps:

Cu� e� ¼ CuðIÞads ðfast stepÞ ð13ÞCuðIÞads � e� ¼ CuðIIÞ ðslow stepÞ ð14Þ

where CuðIÞads is an adsorbed species at the copper surface and doesnot diffuse into the bulk solution [22,24–28]. The dissolution of cop-per is controlled by the diffusion of soluble Cu(II) species from theouter Helmholtz plane to the bulk solution, see more details inthe impedance studies (Section 3.2.3).

It has been shown that in the Tafel extrapolation method, theuse of both the anodic and cathodic Tafel regions is undoubtedlypreferred over the use of only one Tafel region [29]. However,the corrosion rate can also be determined by Tafel extrapolationof either the cathodic or anodic polarization curve alone. If onlyone polarization curve is used, it is generally the cathodic curvewhich usually produces a longer and better defined Tafel region.Anodic polarization may sometimes produce concentration effects,due to passivation and dissolution, as well as roughening of thesurface which can lead to deviations from Tafel behaviour.

The situation is quite different here; the anodic dissolution ofcopper in aerated 1.0 M HNO3 solutions obeys, as previously men-

tioned, Tafel’s law. The anodic curve is, therefore preferred over thecathodic one for evaluation of corrosion currents, jcorr, by the Tafelextrapolation method, see Fig. 2. However, the cathodic polariza-tion curve deviate from the Tafel behaviour, exhibiting a limitingdiffusion current, may be due to the reduction of dissolved oxygen.

2104 K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106

Accordingly, there is an uncertainty and source of error in thenumerical values of the cathodic Tafel slopes calculated by thesoftware. This is the reason why values of the cathodic Tafel slopes,calculated from the software, are not included here.

Table 3 also collects the values of the corrosion current densi-ties, jcorr, and the anodic Tafel slope, ba, associated with polariza-tion measurements for copper in 1.0 M HNO3 solutions withoutand with various concentrations of the three tested benzotriazolederivatives at 25 �C. The inhibition efficiency, nTafel (%), values, alsopresented in Table 3, were calculated from jcorr values based on Eq.(15).

nTafelð%Þ ¼ ½ðjocorr � jcorrÞ=jcorr� � 100 ð15Þ

where jocorr and jcorr are corrosion current densities in absence and

presence of benzotriazole derivatives, respectively. All the recordedjcorr values are converted into the corrosion rate, t, see Table 3, inmm y�1 (millimeter per year; the penetration rate of corrosionthrough a metal) using the expression [30]:

mðlm y�1Þ ¼ 3280� jcorr � ðM=ndÞ ð16Þ

where M is the atomic weight of Cu (63.55 g), n is the number ofelectrons transferred in the corrosion reaction (n = 2) and d is thedensity of Cu (8.92 g cm�3).

Data presented in Table 3 reveal that the jcorr and t values, in allcases, decrease (corresponding to an increase in nTafel(%) values) inthe presence of these compounds, and this decrease in jcorr and t

Fig. 5. Nyquist and Bode plots recorded for copper in 1.0 M HNO3 solutions without andcorrosion potentials at 25 �C. (a) BCA; (b) BAN; (c) BCC. (1) Blank; (2) 10�4 M; (3) 5 � 1

values enhances with increase in concentration of these com-pounds. It is obvious that, under the same conditions, the corrosioncurrents, and hence the rates of corrosion, decrease in the order:BCA < BAN < BCC. This sequence again reflects, as confirmed fromtheoretical study and weight loss measurements, the increasedability of BCA to inhibit nitric acid corrosion of copper as comparedto BAN and BCC. This is clearly seen from the highest efficiency re-corded for BCA (92.14% at a concentration of 5 � 10�3 M, inspectTable 3).

Upon addition of inhibitors, it is obvious that the slopes of theanodic (ba) Tafel lines remain almost unchanged, giving rise to anearly parallel set of anodic lines, and an almost parallel cathodicplots results too. Thus the adsorbed inhibitors act by simple block-ing of the active sites for both anodic and cathodic processes. Inother words, the adsorbed inhibitor decreases the surface areafor corrosion without affecting the corrosion mechanism of copperin these solutions, and only causes inactivation of a part of the sur-face with respect to the corrosive medium [31–33].

3.2.3. Electrochemical impedance spectroscopy, EISThe impedance behaviour of copper, after 1.0 h of immersion in

1.0 M HNO3 solutions without and with various concentrations ofthe three derivatives, was studied. Measurements were conductedat the respective corrosion potentials at 25 �C; results obtained aredepicted in Fig. 5. The recorded impedance spectra are, in all cases,characterized by the appearance of two not clearly resolved timeconstants related to the capacitive loops observed in the Nyquist

with different concentrations of the three benzotriazole derivatives at the respective0�4 M; (2) 10�3 M; (2) 5 � 10�3 M.

K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106 2105

plots at low frequencies (LF). The (HF) capacitive loop refers to anactivation-controlled reaction, while the low frequency (LF) oneconcerns a diffusion-limited process through a layer of corrosionproducts. This latter process can be ascribed to, as previously indi-cated in Section 3.2.2, the reduction of the dissolved oxygen atpH 6 2.9 [34–38]. Impedance spectra measured in Fig. 5a exhibiteda Warburg impedance at (LF), indicating that the corrosion processinvolved the transport of reactants from the bulk of solution to thecopper/solution interface or transport of soluble corrosion prod-ucts from the interface to the bulk solution in the early stages ofcorrosion [39].

The increase in the size of the capacitive loop with the additionof benzotriazole derivatives shows that a barrier gradually formson the copper surface. The barrier is probably related to, as willbe discussed, the formation of an inhibitor surface film on the elec-trode surface. The increase in the capacitive loop size enhances, ata fixed inhibitor concentration, following the order: BCC < BAN <BCA, confirming the highest inhibitive influence of BCA.

Fig. 5b and c present the impedance diagrams for the sameexperimental data in Bode (log Z vs. log f) and Bode-phase format(log f vs. phase) in the presence of different concentrations of inhib-itor solutions. From the Bode plots (Fig. 5b), the total impedanceincreases with inhibitor concentration. Fig. 5c shows the continu-ous increase in the phase angle shift, obviously correlating withthe increase of inhibitor adsorbed on Cu surface.

3.3. Inhibition mechanism

Adsorption of benzotriazoles can be described by two maintypes of interactions: physical adsorption and chemisorption. Ingeneral, physical adsorption requires the presence of both the elec-trically charged surface of the metal and charged species in solu-tion. A chemisorption process, on the other hand, involves chargesharing or charge-transfer from the inhibitor molecules to the me-tal surface to form a coordinate type of a bond. This is possible inthe case of a positive as well as a negative charge of the surface.The presence of a transition metal, having vacant, low-energy elec-tron orbitals (Cu+ and Cu2+ in our case) and of an inhibitor withmolecules having relatively loosely bound electrons or hetero-atoms with a lone pair of electrons is necessary [40].

In aqueous acidic solutions, our three benzotriazoles derivativesexist either as neutral molecules or in the form of cations. Thethree benzotriazoles may be adsorbed on the metal surface inthe form of neutral molecules involving the displacement of watermolecules from the metal surface and sharing of electrons betweenthe nitrogen atoms and the metal surface (chemical adsorption).

Following literature, two mechanisms have been proposed toexplain the high inhibiting efficiency recorded for benzotriazole,designated as BTAH. One mechanism attributes it to the formationof Cu(I)BTA protective film on the surface, which inhibits the ano-dic dissolution reaction, that is,

CuþðaqÞ þ BTAHðaqÞ ¼ CuðIÞBTAðsÞ þHþðaqÞ ð17Þ

The other mechanism postulates an adsorbed layer of BTAH, that is,

CuðsÞ þ BTAH ¼ Cu :BTAHðadsÞ ð18Þ

where Cu: BTAH(ads) refers to BTAH adsorbed on the copper surface.Under oxidizing conditions, this adsorbed species can be oxidized togive the complex Cu(I)BTA [16], that is,

Cu : BTAH ¼ CuðIÞBTAþHþ þ e� ð19Þ

Eq. (19) indicates that the complex Cu(I)BTA is favored at more ano-dic potentials in less acidic media, while the adsorbed species isformed in more acidic media under cathodic potentials. Benzotria-zole derivatives in this study, namely BCA, BAN and BCC also pres-

ent a remarkable inhibition performance, with BCA being the bestone. The high inhibition efficiency of BCA is due to the presenceof other probable sites susceptible to electrophilic attack, namelyN12 and O11 as well as low HOMO energy, low DE and lowhardness.

It was reported that the presence of benzotriazole derivativesinduced the formation of semiconductive copper oxides [40]. Thiswas possibly responsible for the improvement of corrosion resis-tance. The type of intermediates that formed on Cu surface in mo-lar nitric acid can be explained according to the potential-pHdiagram in Fig. 3. It is seen that a stable Cu2O is formed only inacidic solution of pH over 2. Nevertheless, since the proton H+ iscontinuously consumed to the hydrogen gas during the Cu dissolu-tion, the pH value near the Cu electrode in molar nitric acid instan-taneously jump from 1 to greater values. Thus, it is acceptable tothink that Cu2O can be metastably formed on the Cu surface evenat open circuit potential in molar nitric acid [1].

The presence of Cu2O may facilitate adsorption via H-bond for-mation. Another possible mechanism, therefore may be adsorptionassisted by hydrogen bond formation between protonated andunprotonated N atoms in the molecule and the oxidized surface(Cu2O) species. The latter should be more prevalent for protonatedN atoms, because the positive charge on N is conductive to the for-mation of hydrogen bonds. Unprotonated N atoms may adsorb bydirect chemical adsoprtion or by hydrogen bonding to a surfaceoxidized species. The extent of adsorption by the respective modesdepends on the nature of the metal surface. The adsorption layeracts as an additional barrier to the corrosive attack and enhancesthe performance of the passive layer as a result.

The criteria for inhibitor selection can also be inferred fromabove considerations. A good inhibitor must have strong affinityfor the bare metal atoms. The requirement is different in the pres-ence of the oxide species, a passive oxide film is formed on theelectrode surface, where hydrogen bond formation accounts formost of the inhibition action. An effective inhibitor is one thatforms hydrogen bonds easily with the oxidized surface. A compar-ison of the inhibition performance of BCA with BAN and BCCshowed, as previously proved here using theoretical, chemicaland electrochemical methods, that BCA was the most effective cor-rosion inhibitor among the three inhibitors used.

These findings could be further explained on the basis that inpresence of dissolved oxygen (the role of dissolved oxygen in oxideformation will be presented and fully discussed in a separate pa-per), where the metal surface is oxidized, the ability of a benzotri-azole derivative to provide corrosion inhibition is related to itstendency to form hydrogen bonds with the oxide species on themetal surface. Such capability should be proportional to the num-ber of NH linkages in the molecule. It is quite evident from thechemical structure of BCA molecule that it has the highest numberof NH bonds (due to the presence of an extra –NH2 in the amidegroup, which is not present in the BAN and BCC structures). There-fore, BCA possesses the most tendency to form hydrogen bondswith an oxidized surface. These results confirmed the importanceof hydrogen bonding in effective corrosion inhibition in presenceof oxide species.

4. Conclusion

A strategic algorithm for the design of good corrosion inhibitoris developed. With contrast to the traditional technique of identi-fying and synthesis of new corrosion inhibitors in wet lab, a priordry-lab process, followed by wet-lab process is suggested by usingcheminformatics tools. Three benzotriazole derivatives were se-lected and tested computationally as possible corrosion inhibitors.It is found that all of them have corrosion inhibition properties.

2106 K.F. Khaled, M.A. Amin / Corrosion Science 51 (2009) 2098–2106

Experimental investigations, including chemical and electrochem-ical measurements, showed that these derivatives reduce mark-edly the copper corrosion in HNO3 solutions, and this reductionin corrosion rates enhances with concentration of these com-pounds. Polarization studies show that this class of compoundsacts as mixed-type inhibitors.

References

[1] K.F. Khaled, Experimental and atomistic simulation studies of corrosioninhibition of copper by a new benzotriazole derivative in acid medium,Electrochim. Acta 54 (2009) 4345.

[2] T. Kosec, D.K. Merl, I. Milošev, Impedance and XPS study of benzotriazole filmsformed on copper, copper–zinc alloys and zinc in chloride solution, Corros. Sci.50 (2008) 1987.

[3] K. Rahmouni, N. Hajjaji, M. Keddam, A. Srhiri, H. Takenouti, The inhibitingeffect of 3-methyl 1,2,4-triazole 5-thione on corrosion of copper in 3% NaCl inpresence of sulphide, Electrochim. Acta 52 (2007) 7519.

[4] L.M. Rodriguez-Valdez, A. Martinez-Villafane, D. Glossman-Matnik, CHIH-DFTtheoretical study of isomeric thiatriazoles and their potential activity ascorrosion inhibitors, J. Mol. Struct. THEOCHEM 713 (2005) 65.

[5] M.G.V. Satyanarayana, V. Himabindu, Y. Kalpana, Muttineni Ravi Kumar, K.Kumar, Part 1: In-silico studies as corrosion inhibitor and its experimentalinvestigation on mild steel in wet-lab, J. Mol. Struct. THEOCHEM, in press,doi:10.1016/j.theochem.2009.01.005.

[6] J. Andrés, J. Beltran, Química Teórica y Computacional, Universitat Jaume I,Castellón de la Plana, España, 2000.

[7] W. Yang, W.J. Mortier, The use of global and local molecular parameters for theanalysis of the gas-phase basicity of amines, J. Am. Chem. Soc. 108 (1986)5708.

[8] F. Méndez, M. Galván, A. Garritz, A. Vela, J. Gázquez, Local softness andchemical reactivity of maleimide: nucleophilic addition, J. Mol. Struct. 277(1992) 1981.

[9] Y. Feng, W. Teo, K. Siow, Z. Gao, K. Tan, A. Hsieh, Corrosion protection of copperby a self-assembled monolayer of alkanethiol, J. Electrochem. Soc. 144 (1997)55.

[10] B. Delley, An all-electron numerical method for solving the local densityfunctional for polyatomic molecules, J. Chem. Phys. 92 (1990) 508.

[11] B. Delley, From molecules to solids with the DMol3 approach, J. Chem. Phys.113 (2000) 7756.

[12] R.S. Mulliken, Electronic population analysis on LCAO-HOMO molecular wavefunctions, J. Chem. Phys. 23 (1955) 1833.

[13] V.S. Sastri, Corrosion Inhibitors – Principles and Applications, Wiley,Chichester, England, 1998.

[14] J. Foresman, �. Frisch, Exploring Chemistry with Electronic Structure Methods,Gaussian Inc., Pittsburgh, PA, 1996.

[15] J. Cruz, L.M. Martínez-Aguilera, R. Salcedo, M. Castro, Reactivity properties ofderivatives of 2-imidazoline: an ab initio DFT study, Int. J. Quant. Chem. 85(2001) 546.

[16] T. Zhao, G. Mu, The adsorption and corrosion inhibition of anion surfactants onaluminium surface in hydrochloric acid, Corros. Sci. 41 (1999) 1937.

[17] M. Abdallah, E.A. Helal, A.S. Fouda, The effect of some phthalimide derivativeson corrosion behavior of copper in nitric acid, Corros. Sci. 48 (2006) 1639.

[18] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE,Houston, TX, 1975.

[19] G. Quartarone, G. Moretti, T. Bellomi, G. Capobianco, A. Zingales, Using indoleto inhibit copper corrosion in aerated 0.5 M sulphuric acid, Corrosion 54(1998) 606.

[20] W.D. Bjorndahl, K. Nobe, Copper corrosion in chloride media, effect of oxygen,Corrosion 40 (1984) 82.

[21] R. Schumacher, A. Muller, W. Stockel, An in situ study on the mechanism of theelectrochemical dissolution of copper in oxygenated sulphuric acid: anapplication on the quartz microbalance, J. Electroanal. Chem. 219 (1987) 311.

[22] W.H. Smyrl, in: J.O’M. Bockris, B.E. Conway, E. Yeager, R.E. White (Eds.),Comprehensive Treatise of Electrochemistry, vol. 4, Plenum Press, New York,1981, p. 116.

[23] P. Jinturkar, Y.C. Guan, K.N. Han, Dissolution and corrosion inhibition ofcopper, zinc, and their alloys, Corrosion 54 (1984) 106.

[24] E. Mattsson, J.O’M. Bockris, Galvanostatic studies of the kinetics of depositionand dissolution in the copper + copper sulphate system, Trans. Faraday Soc. 55(1959) 1586.

[25] G.G.O. Cordeiro, O.E. Barcia, O.R. Mattos, Copper electrodissolution mechanismin a 1M sulphate medium, Electrochim. Acta 38 (1993) 319.

[26] D.K.Y. Wang, B.A.W. Coller, D.R. Macfarlane, A kinetic model for the dissolutionmechanism of copper in acidic sulfate solutions, Electrochim. Acta 38 (1993)2121.

[27] R. Caban, T.W. Chapman, Statistical analysis of electrode kineticsmeasurements-copper deposition from CuSO4-H2SO4 solutions, J.Electrochem. Soc. 124 (1977) 1371.

[28] T. Hurlen, G. Ottesen, A. Staurset, Kinetics of copper dissolution and depositionin aqueous sulphate solution, Electrochim. Acta 23 (1978) 23.

[29] E. McCafferty, Validation of corrosion rates measured by the Tafelextrapolation method, Corros. Sci. 47 (2005) 3202.

[30] Electrochemistry and Corrosion – Overview and Techniques, Application NoteCORR-4, EG and G, Princeton Applied Research, USA.

[31] G. Quartarone, T. Bellomi, A. Zingales, Inhibition of copper corrosion by isatinin aerated 0.5 M H2SO4, Corros. Sci. 45 (2003) 722.

[32] J.W. Schltze, K. Wippermann, Inhibition of electrode processes on copper byAHT in acid solutions, Electrochim. Acta 32 (1987) 823.

[33] T.N. Anderson, M.H. Ghandehari, H. Eryng, A limitation to the mixed potentialconcept of metal corrosion, J. Electrochem. Soc. 128 (1975) 1580.

[34] C. Deslouis, B. Tribollet, G. Mengoli, M.M. Musiani, Electrochemical behaviourof copper in neutral aerated chloride solution. II. Impedance investigation, J.Appl. Electrochem. 8 (1988) 384.

[35] Y. Feng, W.K. Teo, K.S. Siow, K.L. Tan, A.K. Hsieh, The corrosion behaviour ofcopper in neutral tap water. Part I: corrosion mechanisms, Corros. Sci. 38(1996) 369.

[36] Y. Feng, W.K. Teo, K.S. Siow, K.L. Tan, A.K. Hsieh, Corrosion mechanisms andproducts of copper in aqueous solutions at various pH values, Corrosion 53(1997) 389.

[37] S. Magaino, Corrosion rate of copper rotating-disk-electrode in simulated acidrain, Electrochim. Acta 42 (1997) 377.

[38] G. Brunoro, A. Frignani, A. Colledan, C. Chiavari, Organic films for protection ofcopper and bronze against acid rain corrosion, Corros. Sci. 45 (2003) 2219.

[39] H. Ma, S. Chen, B. Yin, S. Zhao, X. Liu, Impedance spectrospic study of corrosioninhibition of copper by surfactants in the acidic solutions, Corros. Sci. 45(2005) 867.

[40] J.G.N. Thomas, in: Proc. 5th Europ. Symp. on Corrosion Inhibitors, Ann. Univ.Ferrara, Italy, 1980, 1981, pp. 453.