Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486 DOI 10.1007/s40195-015-0222-z

Corrosion Behaviors of Pure Titanium and Its Weldment in Simulated Desulfurized Flue Gas Condensates in Thermal Power Plant Chimney Zheng-Bing Wang • Hong-Xiang Hu • Chun-Bo Liu • Huai-Ning Chen • Yu-Gui Zheng

Received: 14 August 2014 / Revised: 30 September 2014 / Published online: 28 January 2015 Ó The Chinese Society for Metals and Springer-Verlag Berlin Heidelberg 2015

Abstract The corrosion behaviors of pure titanium and its weldment welded by tungsten inert gas (TIG) welding in simulated desulfurized flue gas condensates in thermal power plant chimney were investigated using potentiodynamic polarization, electrochemical impedance spectroscopy (EIS) and immersion tests. The effects of heat input and shielding gases on the corrosion behavior of the welded titanium were also studied. Grain coarsening and Widmansta¨tten structure were found in both the fusion zone and the heat-affected zone. The welded titanium exhibited active–passive behavior in the simulated condensates. Both the polarization curves and EIS measurements confirmed that TIG welding process with different parameters had few effects on the corrosion behavior. It was proved that the microstructure changes were not the key material factors affecting the corrosion behavior of pure titanium under the test conditions, while the oxide film had remarkable effect on improving the corrosion resistance. KEY WORDS:

Pure titanium; Tungsten inert gas welding; Simulated desulfurized flue gas condensates; Corrosion behavior; Microstructure; Oxide film

1 Introduction The emission of acidic gases such as SO2 resulting from the combustion of coal in thermal power plants can cause a

Available online at http://link.springer.com/journal/40195 Z.-B. Wang  H.-X. Hu  Y.-G. Zheng (&) Key Laboratory of Nuclear Materials and Safety Assessment, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China e-mail: [email protected] C.-B. Liu Center for Wear and Corrosion Resistant Materials, Huadian Zhengzhou Mechanical Design Institute Co., Ltd., Zhengzhou 450015, China H.-N. Chen Laboratory of Properties and Reliable Joining of Welded Joints, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China

serious air pollution. In present years, a technology termed as flue gas desulfurization (FGD) has been widely used to reduce the emission of SO2, which involves the desulfurization of the flue gas generated during combustion. The limestone–gypsum wet FGD is one of the most popular processes. The flue gas after desulfurization contains a lot of vapor whose temperature may fall below its dew point. As a result, condensates with high concentrations of acids can form on the surface of the chimney lining, which can cause serious corrosion to the chimney lining materials [1, 2].Therefore, the selection of the material with high corrosion resistance concerns with the safe operation and service lifetime of FGD systems. Titanium exhibits excellent corrosion resistance in many acidic aqueous solutions, which is recommended as one of the candidate materials for anticorrosion linings of chimneys by International Committee on Industrial Chimneys [3]. In China, titanium has been used for linings of chimneys in some thermal power plants. Welding must be used during the construction of metallic anti-corrosion linings. However, titanium and its

123

478

alloys are easy to absorb harmful gases (such as oxygen, hydrogen and nitrogen) due to the high chemical activity, which can lead to poor properties [4].Many efforts have been made to solve the welding problems of titanium. Tungsten inert gas (TIG) welding [4–9], electron beam welding [4], laser welding [4] and friction stir welding [10] have been used successfully in welding of titanium and its alloys.TIG welding is the most common method used in titanium for industrial applications due to its low capital cost and comparatively simple and portable equipment [7, 11]. It is one of the fusion weldings, during which the metals in fusion zone (FZ) melts and the as-cast structure forms in the subsequent cooling process, while the metal in heat-affected zone (HAZ) suffers heat treatment. As a result, the structure of the weld metal (WM) is different from that of the base metal (BM), and their properties are different consequently. Corrosion resistance is one of the concerned aspects in the selection of the materials used for chimney linings. However, the corrosion behavior of pure titanium in the desulfurized flue gas condensates has not been investigated so far. What is more, the effect of TIG welding on the corrosion behavior also seems to be unclear. Blasco-Tamarit et al. [5, 12], Shankar et al. [6] and Orsi et al. [8] have found that, for pure titanium, the TIG welding seems unable to deteriorate the corrosion resistance apparently in oxidizing acidic media and salt solutions. But both the WM and BM are in spontaneous passivation rather than active state in the test solutions in their researches. Therefore, it still needs to be clarified whether the TIG welding has the same effect on titanium in the simulated desulfurized flue gas condensates. In addition, it was found in practice that the welding parameters, such as welding current and filler rod, which are associated with welding heat input, may deviate from the set values due to the effect of the welders or the equipments. Meanwhile, compared with the back of welds, only the front can be shielded with the inert gas because of the structure restriction of the inner chimney wall. All these factors above may affect the anti-corrosion performance of the TIG welded pure titanium. Relevant researches mainly focused on the mechanical properties [7, 13–15], while only Shankar et al. [6] have paid an attention to the corrosion behavior. So it is valuable for both fundamental researches and practical applications to investigate the effect of welding parameters on corrosion behaviors of TIG welded pure titanium. In the present paper, the corrosion behaviors of pure titanium and its TIG weldment with different parameters in the simulated desulfurized flue gas condensates are investigated using potentiodynamic polarization, electrochemical impedance spectroscopy (EIS) and immersion tests combined with scanning electron microscope (SEM). The material factors that affect the corrosion resistance of welded titanium in the simulated condensates are discussed.

123

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

2 Experimental The TIG welding was conducted on commercial grade 2 titanium plate with a dimension of 200 mm 9 150 mm 9 1.5 mm, and the filler rod used was grade 2 titanium wire. The nominal chemical composition and the microstructure of the as-received grade 2 titanium plate are shown in Table 1 and Fig. 1, respectively. The microstructure consists of equiaxed a grain with grain size about 10 lm and a small amount of intergranular b, which indicates that the annealing temperature exceeds the a ? b transition temperature [16]. Overlap welded joints were obtained using Dc pulse TIG welding process with argon arc welding machine (Miller Syncrowave 351). The detailed welding parameters are listed in Table 2. The front of the weld was welded with shielding gas (high purity argon) compared with the back. To assess the microstructures of the welds, the samples containing BM, FZ and HAZ were sectioned and prepared by standard metallographic techniques finishing with a mixture of 0.06 lm non-crystallizing colloidal silica and H2O2 (volume ratio about 7:3). The polished samples were cleaned ultrasonically using acetone and then etched with Kroll’s reagent (10 mL HF, 5 mL HNO3 and 85 mL H2O). Optical microscope (MEF-4) was used to observe the microstructures.

Table 1 Nominal chemical composition (wt%) of as-received grade 2 titanium plate used for TIG welding C

Fe

H

N

O

Ti

0.08

0.20

0.015

0.03

0.18

Bal.

Fig. 1 Optical microstructure of the base metal of grade 2 titanium

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

479

Table 2 Detail process parameters of Dc pulse TIG welding Process

Current (A)

Pulse frequency (Hz)

Filler rod diameter (mm)

Welding speed (mm min-1)

Torch gas flow (L min-1)

A

70/10

5

1

150

12

B

75/10

5

1.2

150

12

C

75/10

Unpulsed

1.2

150

12

Table 3 Explanation of the symbol used in electrochemical tests Symbol

Sample

BM

Base metal

A(B,C)-FZ-f

FZ sample with shielding gas welded with Parameter A(B,C)

A(B,C)-FZ-b

FZ sample without shielding gas welded with Parameter A(B,C) HAZ sample with shielding gas welded with Parameter A(B,C)

A(B,C)-HAZ-f A(B,C)-HAZ-b

HAZ sample without shielding gas welded with Parameter A(B,C)

The samples used for electrochemical and immersion tests were cut from BM, FZ and HAZ, respectively, and the explanation of the symbol used in electrochemical tests is presented in Table 3. After polishing and cleaning, the samples were embedded in epoxy resin with the working surface of about 0.5 cm2. The working electrodes were first polished to No. 800 with a series of silicon carbide abrasive papers and then ultrasonically cleaned with ethanol and water before electrochemical tests. Besides, some of the working electrodes were tested with original surface after welding. All electrochemical measurements were performed in a three-electrode cell using an EG&G Princeton Applied Research model 2273 Potentiostat/Galvanostat. The counter electrode was a Pt foil, and all potentials were measured against a saturated calomel electrode (SCE) connected to the cell via a Luggin probe. Potentiodynamic polarization was conducted from -0.3 V versus open circuit potential (OCP) to 2 V at a sweep rate of 1 mV s-1. EIS measurements were carried out at OCP using a sinusoidal potential perturbation of 10 mV in a frequency range from 10,000 to 0.01 Hz. Before potentiodynamic polarization and EIS measurements, the OCP was monitored for 1 h when the potential value became reasonably stable. All electrochemical measurements were at least triplicated. Fitting of the polarization curves and EIS data was performed with CView 3.3d software and Zsimpwin software, respectively. Because titanium does not exhibit Tafel behavior in the active state, only the cathodic Tafel fitting was conducted to determine the corrosion current density (Icorr) by the intersection of the cathodic Tafel line with

Table 4 Chemical composition (wt%) of the simulated desulfurized flue gas condensates H2SO4

HCl

HNO3

H3PO4

HF

H2O

2.03

0.031

0.03

0.012

0.0092

Bal.

the open circuit potential (potential of zero current in the potentiodynamic curves or Ecorr). The immersion tests were conducted for 24 h with the samples polished to No. 2000. The corrosion morphology was observed using SEM (FEI Inspect F). All measurements were carried out in naturally aerated simulated desulfurized flue gas condensates without stirring at room temperature. The chemical compositions of the condensates were based on those obtained from the chimney of a power plant in China and as shown in Table 4. The pH value of the simulated condensates was approximately 0.78. 3 Results and Discussion 3.1 Microstructures Figure 2 shows the low magnification optical microstructure of TIG welded pure titanium with different parameters. It can be seen that significant grain coarsening has occurred in HAZ immediately adjacent to BM and the grains become larger toward the center of FZ. Compared with pulsed welding (sample A and B), unpulsed welding (sample C) can cause higher welding heat input [13]. Therefore, the degree of the grain coarsening for sample C is the largest. The heat input was lower for sample B than that for sample A due to its bigger diameter of filler rod. As a result, the degree of the grain coarsening of sample B is the smallest, which is medium for sample A. Figure 3 shows the high magnification optical microstructures of FZ and HAZ of TIG welded pure titanium with different parameters. Both FZ and HAZ of three kinds of the samples consist of Widmansta¨tten structure. Larger acicular of serrated a grains are evident in FZ of sample A and C because of higher heat input (Fig. 3a, c), while non-typical acicular or serrated microstructures can be observed in the FZ of sample B and the HAZ of all samples due to lower heat input (Fig. 3b, d, f).

123

480

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

Fig. 2 Low magnification optical microstructures of TIG welded pure titanium: a Parameter A; b Parameter B; c Parameter C

Fig. 3 a–c optical microstructures of FZ samples welded with Parameter A, B and C, respectively; d–e optical microstructures of HAZ samples welded with Parameter A, B and C, respectively

123

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

481

Fig. 4 Potentiodynamic polarization curves of BM, FZ and HAZ samples in naturally aerated simulated desulfurized flue gas condensates: a Parameter A, b Parameter B, c Parameter C

During welding, the metal in FZ is heated to molten state and solidified subsequently and quickly, resulting in coarsened grains and as-cast microstructures that consist of acicular a grains surrounded by prior b grain boundaries for pure titanium (Fig. 3a–c), known as Widmansta¨tten structure [6]. In contrast, the metal in HAZ undergoes heat treatment causing the grain coarsening and a ? b ? a phase transformation. This process leads to the formation of prior b grains with acicular a substructures [16, 17], as shown in Fig. 3d–f. So the phase compositions and the grain sizes of FZ and HAZ are very different from those of BM because of the effect of welding thermal cycle (Figs. 1, 3). However, the phase compositions and the grain sizes between FZ and HAZ are similar in spite of undergoing different thermal cycles. Similarly, the phase compositions are similar in the same position of three kinds of the samples although the welding heat inputs are different, while the only distinction is the grain size. It indicates that the heat input can affect the grain size rather than the phase composition under the test conditions. Besides, although the shielding gases can prevent the welds from oxidation

and adsorption of harmful gases, few differences in microstructures can be found between the front and back of the welds (Fig. 2). Table 5 Electrochemical parameters obtained from the polarization curves Sample

Ecorr (mV)

Icorr (lA cm-2)

A-FZ-f

-667 ± 16

9.0 ± 0.8

A-FZ-b

-677 ± 26

9.6 ± 0.6

A-HAZ-f

-672 ± 9

10.7 ± 1.2

A-HAZ-b

-673 ± 11

10.8 ± 1.1 11.8 ± 2.3

B-FZ-f

-667 ± 21

B-FZ-b

-673 ± 20

8.8 ± 0.4

B-HAZ-f

-649 ± 8

10.2 ± 0.9

B-HAZ-b

-667 ± 15

9.1 ± 0.2

B-FZ-f

-689 ± 34

10.3 ± 1.6

B-FZ-b

-672 ± 17

11.4 ± 1.6

B-HAZ-f

-683 ± 40

9.4 ± 1.7

B-HAZ-b

-689 ± 43

10.9 ± 2.9

BM

-669 ± 2

9.9 ± 0.5

123

482

3.2 Corrosion Behavior of Pure Titanium and its Weldment 3.2.1 Potentiodynamic Polarization Test The potentiodynamic polarization curves of BM and welded pure titanium with different parameters in naturally aerated simulated desulfurized flue gas condensates are shown in Fig. 4. Apparent anodic peak and passivation region can be observed for all curves, suggesting that titanium and its weldment show typical active–passive behaviors. The contents of H2SO4, HCl, HNO3 and H3PO4 in the simulated condensates are low enough that titanium exhibits passivation behavior and high corrosion resistance [18, 19]. However, HF in the simulated condensates can

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

seriously deteriorate the corrosion resistance of titanium when the content exceeds a critical value related to pH value [20]. The effect of fluoride ions on the corrosion behavior of pure titanium in 0.05 mol/L H2SO4 has been investigated, and the results indicated that the critical value of the fluoride concentration was 0.0005 mol/L (about 0.001 wt%) at pH value of 1, which would decrease with lowering pH value [20]. Therefore, the critical value of fluoride content would be lower than 0.001 wt% at pH value of 0.78 for the simulated condensates. It can be seen from Table 4 that the fluoride content in the simulated condensates is 0.0092 wt%, which is higher than the critical value at pH value of 0.78. As a result, the active behaviors for titanium and its weldment are mainly caused by high acidity and HF in the simulated condensates.

Fig. 5 Nyquist plots of BM, FZ and HAZ samples in naturally aerated simulated desulfurized flue gas condensates: a Parameter A, b Parameter B, c Parameter C

123

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

483

Table 6 Parameters of the equivalent circuit obtained by fitting the experimental results of EIS Sample

Rs (X cm2)

Cdl (lF cm-2)

n

Rct (X cm2)

Cf (lF cm-2)

n

Rf (X cm2)

v2

BM

2.952

551.1

0.93

492.1

24,320

0.97

1,099

1 9 10-4

A-FZ-f

2.732

821.8

0.89

475.5

29,420

1

891.2

4 9 10-4

A-FZ-b

2.697

615.5

0.92

495.4

24,020

0.96

865.6

2 9 10-4

A-HAZ-f

2.849

568.5

0.92

480.1

24,070

1

A-HAZ-b

2.764

562.1

0.93

497.4

21,930

0.99

B-FZ-f

2.575

637.1

0.91

525.3

31,580

0.98

1,021

5 9 10-4

B-FZ-b

3.171

604.1

0.92

506.7

20,560

0.99

1,207

5 9 10-4

B-HAZ-f

2.170

595.7

0.93

597.7

28,580

1

1,081

4 9 10-4

B-HAZ-b

2.271

661.7

0.91

643.4

29,030

0.96

1,208

4 9 10-4

C-FZ-f

1.762

397.0

0.91

514.5

22,550

1

1,105

9 9 10-4

C-FZ-b C-HAZ-f

2.093 1.944

416.1 328.2

0.92 0.91

505.5 553.5

27,810 18,810

0.98 1

993.6 1,392

9 9 10-4 1 9 10-3

C-HAZ-b

2.297

345.3

0.89

592.3

32,360

0.97

931.3

3 9 10-4

Fig. 6 Polarization resistance calculated using Eq. (2) based on the data in Table 6

Meanwhile, it can be seen from Fig. 4 that there are few differences between the polarization curves of BM and welded samples with different welding parameters. Table 5 presents the electrochemical parameters obtained by fitting the polarization curves to clarify the effect of TIG welding. Obviously, Icorr is similar to one another for BM, FZ and HAZ samples, which demonstrates that welding process has few effects on the corrosion behavior of pure titanium in active state under the test conditions. What is more, the influences of heat input and shielding gases on the corrosion behavior can be neglected in this study. This phenomenon is in accordance with that observed by other researchers in their solutions [5, 6, 8]. 3.2.2 EIS Measurements In addition to the potentiodynamic polarization curves, EIS measurements were conducted to further confirm the effect of TIG welding on the corrosion behavior of pure titanium.

1,108 965.3

3 9 10-3 2 9 10-4

The results are shown in Fig. 5. The Nyquist plots of the welded samples with different parameters and BM are similar to each other, indicating the similar corrosion behavior. Two capacitive loops can be observed for all samples. The loop at higher frequency corresponds to the charge transfer resistance and that at lower frequency represents the film resistance [21]. An equivalent circuit, the insert in Fig. 5, was used to fit the EIS spectra, in which Rct and Cdl correspond to the charge transfer resistance and electric double-layer capacitance, respectively, Rf and Cf to the resistance and capacitance of the film, respectively, and Rs to the resistance of the solution. A constant phase element (CPE) was used to replace the capacitance due to the distribution of relaxation times resulting from different degrees of heterogeneities at the electrode surface [22]. For simplicity, the symbols of Cdl and Cf will be used instead of CPE. The fitted parameters are listed in Table 6. Polarization resistance is usually used to evaluate the corrosion resistance, which can be determined from the EIS data [23]: RP ¼ ðZF Þx¼0 ;

ð1Þ

where ZF stands for the impedance of the circuit. For the equivalent circuit shown in Fig. 5, Rp can be expressed as follows when x = 0 [23]: RP ¼ Rct þ Rf :

ð2Þ

Based on the parameters in Table 6, Rp of different samples can be calculated using Eq. (2) and the results are shown in Fig. 6. There are few differences between the Rp values of all samples, which is in line with the results of potentiodynamic polarization (Fig. 4) and Nyquist plots (Fig. 5). It is proved that the TIG welding processes with different parameters have few effects on the corrosion behavior of titanium under the conditions of this article.

123

484

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

Fig. 7 Corrosion morphology of BM and weldment with different parameters after immersion for 24 h in the simulated condensates: a pure titanium; b Parameter A; c Parameter B; d Parameter C

3.2.3 Correlation of Corrosion Behavior with Microstructures Generally, welding process can affect the corrosion behavior by changing the microstructures of the metal. It has been found that the grain size for pure titanium [24] and phase composition for titanium alloy [10, 25] can influence the corrosion behavior. First, the grain sizes of BM, FZ and HAZ are quite different from one another (Figs. 1, 2). Balyanov et al. [26] found that ultrafinegrained pure titanium (300 nm) showed superior corrosion resistance compared with coarse-grained titanium (7 lm), while other researchers [27, 28] found opposite results that refining grain to nanoscale or submicro-level was harmful to the corrosion resistance of pure titanium. In present study, the grain sizes are larger than 10 lm (Figs. 1, 2). According to the investigation carried by Shankar et al. [6] and Blasco-Tamarit et al. [12], the grain size in this level has few effects on the corrosion behavior, which is

123

different from the case of nanograins [26–28]. Therefore, it is reasonable to conclude that grain size is not the main material factor affecting the corrosion behavior of pure titanium under the test conditions. Second, the microstructures of TIG welded pure titanium consist of acicular a phase and prior b phase (Fig. 3). Atapour et al. [10, 25] found that the corrosion resistance of Ti-6Al-4 V would be destroyed due to the fine acicular a/b microstructure and alloying element partitioning caused by welding. Similarly, Codaro et al. [29] considered the a/b interfaces in Ti-6Al4 V as the preferential dissolution locations that could increase the corrosion rate. Figure 7 shows the corrosion morphology of BM and weldments with different parameters. All samples exhibit uniform corrosion without any preferential dissolution, in spite of the existence of acicular a/b microstructure as shown in Fig. 3. The characteristic of uniform corrosion is in line with the behavior of pure titanium in fluoride-containing solutions [30]. It demonstrates that acicular a/b microstructure has few effects on

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

485

Fig. 8 Potentiodynamic polarization curves of the polished HAZ samples (as-polished samples) and the unpolished samples after welding (aswelded samples): a Parameter A, b Parameter B, c Parameter C

the corrosion behavior of pure titanium under the test conditions, confirming that the phase composition is not the key material factor affecting the corrosion behavior of titanium. Therefore, the similar corrosion behaviors of pure titanium and its weldment with different parameters in the simulated condensates can be interpreted reasonably based on the analyses above that neither the grain size nor the acicular a/b microstructure is the main factor dominating the corrosion resistance under test conditions. 3.3 Effect of the Oxide Film on Corrosion Behavior of Welded Titanium It has been found that the oxide film formed on the surface of titanium provides resistance to uniform corrosion in many oxidizing acidic media and some diluted reducing acidic media [31]. Oxidation can occur inevitably for titanium during TIG welding even shielded with inner gases. Therefore, it is meaningful to investigate the effect of the oxide film on the corrosion behavior of welded titanium. Figure 8 shows the polarization curves of HAZ

samples with polished surfaces (as-polished samples) and unpolished surfaces after welding (as-welded samples). The as-welded samples with different welding parameters exhibit passivation behaviors, and the current densities are much smaller. It indicates that their corrosion resistances are superior to those of as-polished samples that exhibit active–passive behaviors, due to the oxide film formed. In addition, it has been found that the critical value of the fluoride concentration depends on the extent of interaction of fluoride ions with oxide films [20]. The fluoride ion will not interact with the entire film until the fluoride concentration exceeds the critical value [20]. Accordingly, it is reasonable to speculate that the critical value of the fluoride concentration will be higher in the case of thicker oxide film, because it needs more fluoride ions to interact with the film. Usually, the oxide film formed during welding is thicker than that formed in air or via spontaneous passivation (approximately 3 nm [20] ). From Fig. 8, the critical value of the fluoride concentration for as-welded samples in this study must be higher than 0.0092 wt% (Table 4). However, the accurate critical value of the fluoride

123

486

Z.-B. Wang et al.: Acta Metall. Sin. (Engl. Lett.), 2015, 28(4), 477–486

concentration relates to the thickness of the oxide film and its physical properties. It needs to be confirmed by further investigation. Besides, the passivation region is shorter for the as-welded sample of process C than those of the other parameters (Fig. 8), indicating poorer protectiveness of the oxide film. It is probably because that the larger heat input of process C degrades the structure of the oxide film. But the detailed relationship between welding parameters and oxide film structure still needs to be investigated systematically. On the whole, it is favorable for the corrosion resistance of welded titanium to keep the oxide film formed during welding undestroyed.

4 Conclusions (1)

(2)

(3)

(4)

Grain coarsening and Widmansta¨tten structure can be observed in both FZ and HAZ of TIG welded titanium. The grain size was larger for higher heat input. Pure titanium and its TIG weldment showed active– passive behaviors in the simulated desulfurized flue gas condensates with high acidity and containing a certain amount of fluoride ions. Variations of the grain size and phase composition caused by welding were not the key material factors influencing the corrosion behavior of pure titanium. Moreover, the effect of the heat input and shielding gases on the corrosion behaviors could be neglected for three welding parameters in this study. The oxide film formed during welding could improve the corrosion resistance of pure titanium significantly, which was the dominant material factor affecting the corrosion behavior of pure titanium.

Acknowledgments This work was supported by the National Natural Science Foundation of China (No. 51131008).

References [1] J. Bordzilowski, K. Darowicki, Anti-Corros. Method Mater. 45, 388 (1998) [2] K. Chandra, V. Kain, G.K. Dey, J. Fail. Anal. Prev. 11, 466 (2011) [3] International Committee on Industrial Chimneys, Model Code for Steel Chimneys (Germany, 1999) [4] Y.L. Qi, J. Deng, Q. Hong, L.Y. Zeng, Mater. Sci. Eng., A 280, 177 (2000)

123

[5] J.G. Anton, E. Blasco-Tamarit, A. Igual-Munoz, D. GarciaGarcia, Corros. Sci. 49, 1000 (2007) [6] A.R. Shankar, G. Gopalakrishnan, V. Balusamy, U.K. Mudali, J. Mater. Eng. Perform. 18, 1116 (2009) [7] A. Short, Mater. Sci. Technol. 25, 309 (2009) [8] I.A. Orsi, L.B. Raimundo, O.L. Bezzon, M.A. de Arruda Nobilo, S.E. Kuri, C.A.D. Rovere, V.O. Pagnano, J. Prosthodont. 20, 628 (2011) [9] X.L. Gao, L.J. Zhang, J. Liu, J.X. Zhang, Mater. Sci. Eng., A 559, 14 (2013) [10] M. Atapour, A.L. Pilchak, G.S. Frankel, J.C. Williams, Metall. Mater. Trans. A 41A, 2318 (2010) [11] G. Lu¨tjering, J.C. Williams, Titanium, 2nd edn. (Springer, Manchester, 2003), p. 193 [12] E. Blasco-Tamarit, A. Igual-Munoz, J. Garcia Anton, D.M. Garcia-Garcia, Corros. Sci. 51, 1095 (2009) [13] S. Sundaresan, G.D.J. Ram, G.M. Reddy, Mater. Sci. Eng., A 262, 88 (1999) [14] B.H. Choi, B.K. Choi, J. Mater. Process. Technol. 201, 526 (2008) [15] B. Holmes, M. Gittos, S. Maddox, A. Bahrami, Welding inspection and performance of joints in titanium. Paper presented at the NACE international corrosion 2013 conference, Orlando, 17–21 March 2013 [16] J. Elmer, J. Wong, T. Ressler, Metall. Mater. Trans. A 29, 2761 (1998) [17] S. Lathabai, B.L. Jarvis, K.J. Barton, Mater. Sci. Eng., A 299, 81 (2001) [18] H.H. Uhlig, R.W. Revie, Uhlig’s corrosion handbook (Wiley, New York, 2011), pp. 867–870 [19] J. Vaughan, A. Alfantazi, J. Electrochem. Soc. 153, B6 (2006) [20] Z.B. Wang, H.X. Hu, C.B. Liu, Y.G. Zheng, Electrochim. Acta 135, 526 (2014) [21] T.P. Cheng, J.T. Lee, W.T. Tsai, Electrochim. Acta 36, 2069 (1991) [22] K. Juttner, Electrochim. Acta 35, 1501 (1990) [23] C.N. Cao, J.Q. Zhang, An introduction to electrochemical impedance spectroscopy (Science Press, Beijing, 2002), pp. 45–48. (in Chinese??) [24] M. Hoseini, A. Shahryari, S. Omanovic, J.A. Szpunar, Corros. Sci. 51, 3064 (2009) [25] M. Atapour, A. Pilchak, G.S. Frankel, J.C. Williams, Corros. Sci. 52, 3062 (2010) [26] A. Balyanov, J. Kutnyakova, N.A. Amirkhanova, V.V. Stolyarov, R.Z. Valiev, X.Z. Liao, Y.H. Zhao, Y.B. Jiang, H.F. Xu, T.C. Lowe, Y.T. Zhu, Scr. Mater. 51, 225 (2004) [27] H. Garbacz, M. Pisarek, K.J. Kurzydlowski, Biomol. Eng. 24, 559 (2007) [28] E.K. Sevidova, A.A. Simonova, Surf. Eng. Appl. Electrochem. 47, 162 (2011) [29] E.N. Codaro, R.Z. Nakazato, A.L. Horovistiz, L.M.F. Ribeiro, R.B. Ribeiro, L.R.O. Hein, Mater. Sci. Eng., A 341, 202 (2003) [30] C.S. Brossia, G.A. Cragnolino, Corros. Sci. 46, 1693 (2004) [31] S.Y. Yu, J.R. Scully, C.M. Vitus, J. Electrochem. Soc. 148, B68 (2001)