Corrosion Science 53 (2011) 1948–1953

Contents lists available at ScienceDirect

Corrosion Science

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

Effects of precipitates on the electrochemical performance of Al sacrificial anode

Junguang He a,b,⇑, Jiuba Wen b, Xudong Li a

a School of Materials Science and Engineering, Lanzhou University of Technology, Lanzhou 730050, Chinab School of Materials Science and Engineering, Henan University of Science and Technology, Luoyang 471003, China

a r t i c l e i n f o

Article history:Received 9 October 2010Accepted 9 February 2011Available online 19 February 2011

Keywords:A. AluminiumB. TEMB. PolarizationB. EISC. Anodic dissolution

0010-938X/$ - see front matter � 2011 Elsevier Ltd.doi:10.1016/j.corsci.2011.02.016

⇑ Corresponding author at: School of Materials ScienUniversity of Technology, Lanzhou 730050, China. Tel.379 64230597.

E-mail address: he.ellen@163.com (J. He).

a b s t r a c t

The influence of precipitates on the electrochemical performance of Al–Zn–In–Mg–Ti–Si sacrificial anodewas investigated by the TEM observation and electrochemical measurements. The results indicate thatthe shape and size of precipitates in the alloys has great impact on the electrochemical performance.The anodes with rod-like precipitates are easily corroded along grain boundaries, resulting in the low cur-rent efficiency caused by serious grain loss. In comparison, the anodes with spherical or discal precipi-tates have high current efficiency and even corrosion morphology. The precipitates with a size ofabout 400 nm are conducive to improve the electrochemical performance of anodes.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Al–Zn–In based anodes has been widely used as the sacrificialanode materials for cathode protection of steel objects in seawaterin China due to its high current efficiency and a relatively even at-tack morphology [1]. It has been documented that the main precip-itated phase is MgZn2 particles in Al–Zn–In–Mg–Ti alloy [2].Andreatta et al. pointed that the solution heat treatment and theaging could change the size, the composition and the distributionof MgZn2 particles [3]. Clark found the variation of MgZn2 particlesin shape during the aging process [4]. The usual precipitation se-quence of MgZn2 particles can be summarized as: solid solution—GP zone—metastable g0—stable g (MgZn2) [5].

The formation and the distribution of precipitates in the alu-minium matrix have been promoted as a key factor in achievinga good surface activation of the anode [6]. Therefore, the precipi-tates have significant impact on the properties of aluminium sacri-ficial anodes [7,8]. The effect of MgZn2 particles on the corrosionmorphology, especially the intergranular corrosion susceptibility,Frankel et al. considered that the size of MgZn2 particles at theboundaries and their interparticle distance play an important role[9–11].

The electrochemical behavior of Al–Zn–In alloys has been inves-tigated previously [12–14]. Most of them basically focused on theeffect of trace element on the electrochemical properties of anode,whereas that for precipitates is paid little attention. As regards to

All rights reserved.

ce and Engineering, Lanzhou: +86 379 64231846; fax: +86

the activation of Al–Zn–In alloys, Bessone et al. pointed that theactivation degree attained in Al anode depends on the amount ofIn3+ present at the interface [15,16]. Graver et al. proposed thatthe activation of aluminium in chloride solution was a result ofsegregation of indium to the alloy surface, and has proved the factof segregation of indium to the aluminum surface by annealed at300 �C [17].

In the present study, the solution treatment and the aging processwere used to change the size, morphology and distribution of theprecipitates. To avoid segregation problems, the external part ofsamples was removed after aging process. Effects of precipitates onthe electrochemical properties of aluminium sacrificial anode wereinvestigated by TEM observation, current efficiency, potentiodynam-ic polarization and electrochemical impedance spectroscopy tests.

2. Experimental

2.1. Material preparation

Pure aluminium, zinc, indium, magnesium, titanium and silicon(>99.9%) were used to produce the necessary alloy. The nominalcomposition of the experimental alloy is 5Zn–0.03In–1Mg–0.05Ti–0.14Si–Al (wt.%). The raw materials were cut, dried andweighed the required amount of materials and melted in a corun-dum crucible in ZGJL0.01-4C-4 vacuum induction furnace underargon shield at 760 �C for about 5 min. Then, the homogeneousmelt was poured into a preheated metal mold with dimensionsof U 20 � 140 mm.

The ingots were solution treated at 470 �C for 1 h in an anneal-ing furnace followed by water quenching at room temperature, and

J. He et al. / Corrosion Science 53 (2011) 1948–1953 1949

then underwent the various aging treatments listed in Table 1.After aging, the external part of samples was eliminated to avoidsegregation problems and they were further processed into cylin-ders with a size of ø16 � 28 mm for current efficiency measure-ments and ø11.3 � 5 mm for electrochemical tests. The samplesfor electrochemical tests were covered with an epoxy resin leaving1 cm2 exposed area in contact with the solution. The exposed sur-face was ground successively on silicon carbide papers up to 2000grit, polished using 1.5 lm diamond paste, rinsed with acetone andthen dried in air.

2.2. Anode current efficiency

The current efficiency tests of anodes after different treatmentswere performed by a three electrodes system at room temperature.Sample, mild steel and a saturated calomel electrode (SCE) wereused as anode, cathode and reference electrode, respectively. Thesurface area of the anode was 14 cm2. The surface area ratio ofthe test anode to cathode was 1:60. The anode and cathode wascoupled together using a copper wire and impressed an anode cur-rent density of 1 mA/cm2 for a period of 10 days in artificial seawa-ter with an initial pH 8.2 according to a standard procedure (GB/T17848-1999) [18]. After the tests, the samples were cleaned using68% HNO3 solution to clear away the corrosion products adheringon its surface. The sample weights before and after immersion inthe galvanic system were measured after cleaning. The anode cur-rent efficiency was calculated by means of the weight loss accord-ing to following formulas.

g ¼ Q=Q 0

Q ¼ C=W

Here, g is the current efficiency, Q is the actual current capacity,Q0 is the theoretical current capacity, C is the total quantity ofcharge passed through the electrochemical circuit and W is the to-tal weight loss.

At least three experiments were done for each aging treatmentmaterial in view of the reproduction quality of the data.

2.3. Electrochemical measurements

The electrochemical measurements were carried out with athree electrodes system by CHI660C electrochemical workstation.The corrosion medium was 3.5 wt.% NaCl solution. A platinumsheet with a large area was used as counter electrode, SCE servedas reference electrode, and the sample as working electrode. Whenthe open-circuit potential (OCP) became stable after the sampleswere immersed in 3.5 wt.% NaCl solution for 4000 s, electrochem-ical impedance spectroscopy (EIS) was recorded with an excitationvoltage of 5 mV in the scan frequency range of 10 kHz–0.1 Hz.Then, the polarization curve was carried out at a constant voltagescan rate of 1 mV/s. The scan scope was set from �1.3 to�0.7 V(SCE). The electrochemical experiments were repeated totest reproducibility.

Table 1Various aging treatment processes of Al–Zn–In–Mg–Ti–Si alloy (wt.%).

Samples Process

T12 Aging at 120 �C for 24 hT16 Aging at 120 �C for 24 h, and re-aging at 160 �C for 30 hT25 Aging at 120 �C for 24 h, and re-aging at 250 �C for 30 hT30 Aging at 120 �C for 24 h, and re-aging at 300 �C for 30 h

2.4. Microstructure analysis

The precipitates distribution in the samples was analyzed witha transmission electron microscope (TEM, JSM-2010). The samplesfor TEM observation were thinned to approximately 50 lm usingmechanical methods and were cut into circular thin foils 3 mmin diameter. Then they were prepared by twin jet-polishing in25% HNO3 and 75% CH3OH solution cooled to �30 �C with liquidnitrogen at 20 V. The morphologies of the samples were character-ized using digital camera.

3. Results and discussion

3.1. Microstructure

Fig. 1a–d shows the TEM bright field image within the grains ofsamples T12, T16, T25 and T30, respectively. In the case of T12sample, many fine spherical precipitates of MgZn2 particles wereobserved as shown in Fig. 1a. The distribution of precipitate parti-cles in a-Al matrix was dispersive. The overall concentration ofMgZn2 particles was very high and their size was only a few nano-meters much smaller than that for grains. Compared with T12, theparticles density was lower and the particles size was larger forT16 sample indicated in Fig. 1b. However, the distribution of pre-cipitate particles remained dispersive. Besides, some precipitatesparticles changed from spheres to rods. The size of spherical parti-cles was about 10 nm, whereas that of rod-like particles wasapproximately 7 � 20 nm. With the increasing of re-aging temper-ature, the density and volume fraction of precipitates decreasedobviously, and the proportion of rod-like particles in precipitatesincreased by degree. In addition, the rod-like particles extendedand coarsened gradually. The size of rod-like particles was approx-imately 50 � 200 nm when the re-aging temperature reached250 �C as shown in Fig. 1c. However, the rod-like precipitate parti-cles were almost disappeared, replaced by the disk-like or block-shaped particles with a dimension of approximately 400 nm whenthe re-aging temperature was 300 �C presented in Fig. 1d. The re-sult further confirms that the rod-like precipitate was not an equi-librium phase but rather a transition phase, which is consistentwith the results reported by Clark [4].

Fig. 2 displays the TEM bright field images of the area aroundgrain boundaries of T12 and T25 samples. The spherical particleswith about 25 nm in size distributed discontinuously on the grainboundaries in T12 sample as seen in Fig. 2a. The size of sphericalparticles was at least five times more than that of the grain interior.On the contrary, the coarse rod-like particles with the length ofabout 500 nm inconsecutive scattered on the grain boundaries inT25 sample indicated in Fig. 2b. This may be due to the coarseningand changing of the precipitates. The intergranular corrosion resis-tance of aluminium alloys is strongly affected by the precipitates atthe grain boundaries [19]. The rod-like particles on the grainboundaries has the high susceptibility of corrosion resistance.

3.2. Corrosion morphology

Fig. 3 presents the surface features of corroded samples thatwere cleaned using a solution of 68% HNO3 after immersing in arti-ficial seawater for 10 days at current density of 1 mA/cm2. The sur-face of T12 sample in Fig. 3a was rugged, and the steep boundariesof raised regions could be seen, which indicates that the particleshad dropped out from the sample surface. However, some regionsremain relatively bright, implying non-corrosion attack. This isconsidered to be due to the following two factors. First, the precip-itates in the grain interior for T12 sample were too small to be acti-vated because the thickness of passive film was larger than the size

Fig. 1. TEM bright field micrographs within grain of samples: (a) T12, (b) T16, (c) T25 and (d) T30.

Fig. 2. TEM bright field micrographs of region around grain boundary of samples: (a) T12 and (b) T25.

1950 J. He et al. / Corrosion Science 53 (2011) 1948–1953

of precipitate particles [20]. Second, the precipitates with highdensity in the grain boundary were susceptible to the intergranularcorrosion [3,21].

Similar surface features were founded for T16 sample shown inFig. 3b and T25 sample indicated in Fig. 3c. A great quantity of cor-rosion pits and many non-corrosion regions were observed on thesurface of samples T16 and T25. Moreover, serious intergranularcorrosion could be found in the samples interior after the sampleswere cut transversely. This may be attributes to the larger rod-likeprecipitates particles appeared in T16 and T25 samples comparedwith T12 sample, especially at the grain boundary part.

A flat surface and shallow corrosion pits were observed for T30sample as indicated in Fig. 3d, which implying the uniform corro-sion. This indicates that the discal or blocky precipitates are bene-

ficial to improve the dissolution uniformity of aluminiumsacrificial anodes.

3.3. Electrochemical performance

Table 2 lists the electrochemical properties of samples T12, T16,T25 and T30. The current efficiency of T12, T16 and T25 was 91.3%,84.9% and 87.8%, respectively, whereas the current efficiency ofT30 sample was 95.8%. Moreover, the samples T12, T16 and T25exhibited the similar OCP values approximately �0.89 V(SCE),while the OCP value of T30 was ca. �1.08 V(SCE). The T30 samplepresents the highest current efficiency and the lowest OCP valuesamong all samples. This is probably caused by the fact that theshape and dimension of precipitates particles in T30 sample were

Fig. 3. Corrosion surface photographs of samples after immersed in artificial seawater for 10 days: (a) T12, (b) T16, (c) T25 and (d) T30.

Table 2Electrochemical performances of Al–Zn–In–Mg–Ti–Si alloy with different heattreatments.

Samples OCP (V vs SCE) Capacity (A h kg�1) Efficiency (%)

T12 �0.8960 2615 91.3T16 �0.8915 2434 84.9T25 �0.8990 2515 87.8T30 �1.0832 2746 95.8

J. He et al. / Corrosion Science 53 (2011) 1948–1953 1951

different from that in T12, T16 and T30 samples, since the compo-sition of samples was not changed by aging treatment.

Fig. 4 exhibits the variation curves of close circuit potential(CCP) values versus time during the samples immersed in artificialseawater for 10 days at the current density of 1 mA/cm2. The CCPvalues of T12 sample was about �0.87 V(SCE) at the few days ofbeginning. The CCP shifted to more negative with the increasingof immersion time and reached to �1.03 V(SCE) after the 5 daysimmersion. Compared with T12 sample, the samples T16 and T25revealed more positive potential. Whereas, the CCP values of T30sample was about �1.05 V(SCE), and had the small fluctuation.More negative and stable CCP values of T30 sample implies theuniform corrosion of the anode, corresponding to high anode effi-ciency caused by the uniform consumption of the anode.

As indicated above, the T30 sample had the highest current effi-ciency, most negative OCP and CCP values and the uniform corro-sion morphology. Therefore, it can be concluded that T30 samplehas the best electrochemical performance.

Fig. 4. CCP vs time plot of samples in artificial seawater: N T12, . T16, j T25 and d

T30.

3.4. Polarization curves

Fig. 5 shows the potentiodynamic polarization curves of sam-ples T12, T16, T25 and T30 immersed in 3.5 wt.% NaCl solutionfor 4000 s. The corrosion potential (Ecorr) of the T30 sample was�1.078 V(SCE), while, that of the others was about �0.915 V(SCE).In addition, polarization curve of T30 sample appeared an obviousactivation platform, at which the passive current density increasedrapidly, indicating the initiation of pitting corrosion. The pittingpotential of T30 sample shifted to more negative, which indicatedthat T30 sample was susceptible to pitting corrosion resulting inthe even corrosion morphology [12,22]. This is in accord with thesurface features of the T30 sample after immersed in artificial sea-water for 10 days at current density of 1 mA/cm2 in Fig. 3d. It canbe speculated that the precipitates with moderate-size spherical ordiscoid in the alloy is beneficial to improve the electrochemicalperformance of aluminium sacrificial anodes.

3.5. Electrochemical impedance spectra

The influence of precipitates on the corrosion behavior of Al–Zn–In–Mg–Ti–Si alloy was investigated with EIS measurementsat OCP as shown in Fig. 6. The Nyquest plots for the samples T12,T16 and T25 in Fig. 6a presented a single capacitance loop in thewhole frequency range, which indicates that the system can be de-scribed by a single time constant model. Conversely, the plot forT30 sample was comprised by a capacitance arc at high frequencyand an inductive loop at the lower frequency. The presence of two

Fig. 5. Potentiodynamic scanning curves of samples in 3.5% NaCl solution.

Fig. 6. Eis patterns of samples in 3.5% NaCl solution: N T12, . T16, j T25 and d T30.

1952 J. He et al. / Corrosion Science 53 (2011) 1948–1953

time constants in the spectra can be observed from the asymmetryof the phase angle at low frequencies as shown in Fig. 6c. Accordingto the Bode plots of |Z| versus frequency shown in Fig. 6b, it wasfound that the |Z| for T12 sample was the maximum value, andthen it reduced gradually with the increase of aging temperature.

The EIS plot for samples T12, T16 and T25 can be simply inter-preted using the equivalent circuit shown in Fig. 7a and b showsthe equivalent circuit for T30 sample. Where, Rs is the solution

Fig. 7. The equivalent circuits of different samples in 3.5% NaCl solution: (a) T12,T16 and T25; (b) T30.

resistance. Rt represents the charge transfer resistance of the inter-facial reaction and CPE is a constant phase angle element whichstands for the electric double layer capacity. However the doublelayer capacitor does not behave ideally due to the roughnesscaused by the corrosion product layer. The CPE was used to com-pensate for the non-homogeneity in the system [23,24]. The CPEis defined by Y0 and n (0 6 n 6 1). n = 1 corresponds to a capacitor,n = 0.5 Warburg impedance, n = 0, a resistance and n = �1, aninductance. R1 and L characterize the inductive loop at low fre-quency. The equivalent circuit element values obtained by ZSimp-Win software are listed in Table 3.

The Rs listed in Table 3 was always very low, so the corrosionrate was determined by Rt. The Rt of T12 sample was1.56 � 104 X cm2, which implies that passive film has occurred inthe surface of alloy. Then it reduced gradually with the increaseof aging temperature, which indicates that the alloy was activated.The Rt for T30 sample was about 2300 X, the alloy was activatedand the dissolution corrosion rate was slow, so its current effi-ciency increased. The inductive loop in low frequency has been re-lated in the past with the relaxation of adsorbed species [25,26], inparticular with chloride adsorption in the case of aluminium pit-ting. The moderate pitting is a favorable factor for the activation al-loys. As indicated above, it can be concluded that the moderate sizeof capacitive loop and inductive loop corresponds to a high currentefficiency and more uniform corrosion morphology.

4. Conclusions

In the present study, the effects of precipitates on the electro-chemical performance of Al–Zn–In–Mg–Ti–Si sacrificial anodes

Table 3Parameters of the equivalent elements in equivalent circuit for different samples.

Samples Rs

(X cm2)CPE(X�1 cm2 s�1)

n(0 6 n 6 1)

Rt

(X cm2)L(H cm2)

R1

(X cm2)

T12 5.48 1.25 � 10�5 0.91 1.56 � 104 – –T16 5.65 1.18 � 10�5 0.94 8848 – –T25 5.22 4.91 � 10�5 0.79 6079 – –T30 3.72 1.46 � 10�5 0.94 2339 121 1423

J. He et al. / Corrosion Science 53 (2011) 1948–1953 1953

were investigated. Main results obtained from this study are asfollows:

(1) The anode with rod-like precipitates is easily corroded alongthe grain boundaries, which results in the severe shedding ofgrain caused by the intergranular corrosion and reduces thecurrent efficiency. The anodes with spherical or discal pre-cipitates have the better electrochemical properties.

(2) The precipitate particles with moderate size are conduciveto improve the electrochemical performance of anodes,whereas the anode with a very fine precipitation was notbeneficial to the activation.

(3) The T30 sample with approximately 400 nm disk-shapedprecipitates has the highest current efficiency, most negativeOCP and CCP values and the uniform corrosion morphology.

Acknowledgement

This work was supported by Technology Creative programmerof Henan for Excellent Talents (Grant No. 094200510019) and theNatural Science Foundation of Henan Province, China (Grant No.092300410132).

References

[1] GB/T 4948-2002, Sacrificial Anode of Al–Zn–In Series Alloy, China StandardPress, Beijing, 2002.

[2] J.L. Ma, J.B. Wen, X.W. Lu, Activation effects of g precipitates in the Al–Zn–In–Mg–Ti sacrificial anode alloys, Special Casting & Nonferrous Alloys 30 (2010)18–21.

[3] F. Andreatta, H. Terryn J.H.W. de Wit, J.H.W. de Wit, Corrosion behaviour ofdifferent tempers of AA7075 aluminium alloy, Electrochimica Acta 49 (2004)2851–2862.

[4] J.B. Clark, Transmission electron microscopy study of age hardening in a Mg-5wt.% Zn alloy, Acta Materialia 13 (1965) 1281–1289.

[5] G. Sha, A. Cerezo, Kinetic Monte Carlo simulation of clustering in an Al–Zn–Mg–Cu alloy (7050), Acta Materialia 53 (2005) 907–917.

[6] A. Barbucci, G. Cerisola, G. Bruzzoneb, A. Saccone, Activation of aluminiumanodes by the presence of intermetallic compounds, Electrochimica Acta 42(1997) 2369–2380.

[7] N. Birbilis, M.K. Cavanaugh, R.G. Buchheit, Electrochemical behavior andlocalized corrosion associated with Al7Cu2Fe particles in aluminum alloy7075–T651, Corrosion Science 48 (2006) 4202–4215.

[8] J.F. Li, Z.Q. Zheng, S.C. Li, W.J. Chen, W.D. Ren, X.S. Zhao, Simulation study onfunction mechanism of some precipitates in localized corrosion of Al alloys,Corrosion Science 49 (2007) 2436–2449.

[9] Zhijun Zhao, G.S. Frankel, The effect of temper on the first breakdown inAA7075, Corrosion Science 49 (2007) 3089–3111.

[10] D. Wang, Z.Y. Ma, Z.M. Gao, Effects of severe cold rolling on tensile propertiesand stress corrosion cracking of 7050 aluminum alloy, Materials Chemistryand Physics 117 (2009) 228–233.

[11] J.F. Li, Z.W. Peng, C.X. Li, Z.Q. Jia, W.J. Chen, Z.Q. Zheng, Mechanical properties,corrosion behaviors and microstructures of 7075 aluminium alloy withvarious aging treatments, Transactions of Nonferrous Metals Society of china18 (2008) 755–762.

[12] J.L. Ma, J.B. Wen, The effects of lanthanum on microstructure andelectrochemical properties of Al–Zn–In based sacrificial anode alloys,Corrosion Science 51 (2009) 2115–2119.

[13] H. Sina, M. Emamy, M. Saremi, A. Keyvani, M. Mahta, J. Campbell, The influenceof Ti and Zr on electrochemical properties of aluminum sacrificial anodes,Materials Science and Engineering A 431 (2006) 263–276.

[14] J.L. Ma, J.B. Wen, The corrosion behaviour of Al–Zn–In–Mg–Ti alloy in NaClsolution, Corrosion Science 52 (2010) 534–539.

[15] J.B. Bessone, D.O. Flamini, S.B. Saidman, Comprehensive model for theactivation mechanism of Al–Zn alloys produced by indium, Corrosionscience 47 (2005) 95–105.

[16] C.B. Breslin, L.P. Friery, The synergistic interaction between indium and zinc inthe activation of aluminium in aqueous electrolytes, Corrosion science 36(1994) 231–238.

[17] Brit Graver, Antonius T.J. van Helvoort, Kemal Nisancioglu, Effect of heattreatment on anodic activation of aluminium by trace element indium,Corrosion science 52 (2010) 3774–3781.

[18] GB/T17848-1999, Test Methods for Electrochemical Properties of SacrificialAnodes, China Standard Press, Beijing, 2002.

[19] F. Andreatta, M.M. Lohrengel, H. Terryn J.H.W. de Wit, J.H.W. de Wit,Electrochemical characterisation of aluminium AA7075–T6 and solution heattreated AA7075 using a micro-capillary cell, Electrochimica Acta 48 (2003)3239–3247.

[20] P.L. Cabot, J.A. Garrido, E. Perez, A.H. Moreira, P.T.A. Sumodjot, W. Pround, EISstudy of heat-treated Al–Zn–Mg alloys in the passive and transpassivepotential region, Electrochimica Acta 40 (1995) 447–454.

[21] A. Zinelinski, J. Chrzanowski, M. Warmuzek, A. Gazda, E. Jezierska, Influence ofretrogression and re-aging on microstructure, mechanical properties andsusceptibility to stress corrosion cracking of an Al–Zn–Mg alloy, Materials andcorrosion 55 (2004) 77–87.

[22] Y.S. Huang, X.T. Zeng, X.F. Hu, Corrosion resistance properties of electrolessnickel composite coatings, Electrochimica Acta 49 (2004) 4313–4319.

[23] A. Venugopal, V.S. Raja, AC impedance study on the activation mechanism ofaluminium by indium and zinc in 3.5% NaCl medium, Corrosion Science 39(1997) 2053–2065.

[24] J. Bessone, C. Mayer, K. Jiittner, W.J. Lorenz, AC-impedance measurements onaluminium barrier type oxide films, Electrochimica Acta 28 (1983) 171–175.

[25] J.B. Bessone, D.R. Salinas, C.E. Mayer, M. Ebert, W.J. Lorenz, An EIS study ofaluminium barrier-type oxide films formed in different media, ElectrochimicaActa 37 (1992) 2283–2290.

[26] Di Li, Yangling Hu, Baolan Guo, Study on the evaluation of localized corrosionof 2024T3 aluminum alloy with EIS, Materials Science and Engineering A 280(2000) 173–176.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具