- Email: [email protected]
The effect of substrate temperature on the oxidation behavior of erbium thick films
Vacuum 86 (2012) 1097e1101
Contents lists available at SciVerse ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
The effect of substrate temperature on the oxidation behavior of erbium thick films H.H. Shen a, S.M. Peng b, X.G. Long b, X.S. Zhou b, L. Yang a, X.T. Zu a, * a b
Department of Applied Physics, University of Electronic Science and Technology of China, No. 4, Section 2, North Jianshe Road, Chengdu 610054, China Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics, Mianyang 621900, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 June 2011 Received in revised form 15 September 2011 Accepted 5 October 2011
The effect of substrate temperature on the oxidation behavior of erbium thick films, fabricated by electron-beam vapor deposition (EBVD), was investigated by X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and scanning electron microscopy (SEM). The erbium thick film is black when it is deposited at substrate temperature below 450 C and turns gray at higher substrate temperature in a vacuum pressure of approximately 1.5 106 Torr, which indicates that the thickness of erbium oxide layer formed on the surface of erbium films increases with the decreasing substrate temperature. XPS depth profile results demonstrate that the thickness of the surface erbium oxide layer of erbium film deposited at substrate temperature of 550 and 350 C are about 50 and 75 nm, respectively. The thicker oxide layer at lower substrate temperatures may be attributed to grain size and the dynamic vacuum condition around the substrates. Other possible factors involved in the oxidation behavior are also discussed. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Erbium thick film Substrate temperature Oxidation behavior Grain size
1. Introduction Films of erbium, a rare earth metal, with thickness of 3e6 mm is under consideration to be a candidate material as tritium (T) target of neutron generator (NG) [1]. The critical reasons are because of the property of long time storage in solid phase of its hydride [2], as well as excellent thermal stability of tritium [3] and helium (He) [4] generated by the decay of T in the Er-T-He ternary system. Systematic research on erbium films has been employed for its potential application in the nuclear industry. The erbium oxide layer, formed on the surface of erbium film due to its high susceptibility to oxidation in the deposition process, degrades the thermal dynamic properties for absorbing T by inhibiting T diffusion into the film to react with erbium. Brumbach et al. found that the thermal annealing either with higher temperature or with longer annealing times leads to a degradation of the passive oxide leading to a bulk film more accessible for hydrogen loading [5]. Holloway reported that the electron-beam deposited erbium films had surface oxide layer with thickness ranged from 17 to 19 nm after exposure to air, whereas the hydrided erbium films had oxide thickness of 11e35 nm after annealing at temperatures up to 500 C in a ambient at 1.0 108 Torr [6]. Parish et al. demonstrated that
* Corresponding author. Tel./fax: þ86 28 83202130. E-mail address: [email protected] (X.T. Zu). 0042-207X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2011.10.009
the erbium oxide formation would proceed readily during processing, which may detrimentally contaminate Er(H,D,T)2 films [7]. An additional platinum layer deposited on erbium surface immediately after erbium deposition without air exposure between deposition steps was employed to avoid the formation of erbium oxide compound and improve the capacity of absorbing T [8]. Generally, high temperature annealing would enhance the level of oxidation of metal bulk and film specimens due to their stronger reactivity at high temperature. Thermal oxidation induces a surface smoothening of TieNi thin films and surface roughness of oxidized TieNi films decreases with the increasing oxidation temperature [9]. Roberts et al. believed that via increasing the temperature of the film from 196 C to 183 C, sufficient energy has been provided to allow relatively facile entry of oxygen into the subsurface region [10]. Heating above 327 C for Ni films leads to the diffusion of oxygen atoms into the bulk and the partial reduction of the surface Ni to its metallic state [11]. Grain size can also serve as a significant factor to affect the oxidation behavior of metal materials because the surface energy of the film with small grains is higher than that of big grain film. Above all, eliminating or reducing the possibility of erbium to be oxidized is an effective way to promote its capability as T storage material. However, the mechanism of oxidation behavior of erbium film and the factors influencing the oxidation behavior at lower substrate temperatures are not clear up to now. In this study, erbium films are deposited on the rolled molybdenum substrates with substrate temperatures
1098
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
ranging from 200 C to 650 C, in order to understand the effect of substrate temperature on the oxidation behavior. 2. Experimental procedure A series of 2e4 mm erbium thick films were electron-beam evaporated on the rolled molybdenum (Mo) substrates and mechanically polished-molybdenum (PMo, surface roughness RMS < 5 nm) substrates at substrate temperatures of 200 Ce650 C. Evaporation rates detected by IC5 thin film deposition controller (Inficon, USA) were 5 or 10 nm/s. Before deposition, all substrates were ultrasonically cleaned with ethanol and subsequently in acetone. In addition, the molybdenum holder and molybdenum substrates were outgassed by increasing the temperature from room temperature to 700 C and heating for 2 h at 700 C. The source material erbium bulk (99.95% purity, 25 mm in diameter) were outgassed by melting it using electron-beam to remove the surface oxide layer and possible residual gases absorbed from air. After outgas of substrates and source material, the molybdenum substrates were cooled to the desired temperature for film deposition and the vacuum pressure was better than 3.0 106 Torr. The erbium was then deposited onto the substrates with the direction parallel to the substrate normal. To obtain uniform film, the sample holder was rotated during deposition and the distance between source material and sample holder was w20 cm. The microstructure of the as-deposited erbium thick films were identified by XRD using a X’pert PRO MPD (Panalytical, Holand) with a Cu Ka radiation. The cross-section morphologies of films were characterized using a Apollo 300 FE-SEM (Obducat CamScan, UK). The XPS ion sputtering depth profiling of oxygen and erbium was performed for three representative erbium films with different surface colors by using a ESCALAB 250Xi spectrometer (Thermo Scientific, UK). The X-ray source was a non-monochromatic Al Ka (150 W) lamp at photon energy 1486.6 eV. The residual gas pressure in the analyzing chamber was typically less than 1.8 109 mbar. The specimens were ion sputtered using an argon gun operated at 3 keV to conduct in-depth profile analysis and to determine the thickness of the oxide layer. The argon ion beam with current of 2.1 mA was rastered over a 0.6 mm 0.6 mm area yielding a sputtering rate of w25 nm/min and the etching thickness was measured by Dektak 150 Stylus Profiler (Veeco, USA). 3. Results and discussion The erbium films were deposited onto Mo and PMo substrate at different substrate temperatures of 200 C, 350 C, 450 C, 550 C, and 650 C. For all samples, the color of as-deposited erbium films changes from black to gray with the increasing substrate temperature. The most clear color change occurs between 350 and 450 C. Therefore, only three representative samples (listed in Table 1) with different colors deposited at substrate temperature of 350 and 550 C were characterized by the XPS depth profiling technique. There is no diffraction peak of erbium oxide in the XRD patterns of PMo-350, Mo-550 and PMo550 samples (as shown in Fig. 1), which
Table 1 Sample specifications. Substrate
Substrate temperature ( C)
Surface color
Thickness of oxide layer (nm)
Designation
PMo Mo PMo
350 550 550
Black gray gray
w75 w50 w50
PMo-350 Mo-550 PMo-550
Fig. 1. The XRD patterns of erbium thick films deposited on (a) PMo substrate at substrate temperatures of 350 C and (b) Mo, (c) PMo substrate at 550 C. The structures of hexagonal-close-packed (hcp) and body-centered-cubic (bcc) for erbium (Er) and molybdenum (Mo) are observed in each of the XRD patterns, whereas the diffraction peak of erbium oxide does not appear for any sample.
indicates that the oxide layer is extremely thin in comparison to the thickness of erbium film. Evolution of the Er 4d and O 1s core level spectra as a function of etching time are shown in Fig. 2 for PMo-350, Mo-550 and PMo550 samples. From the Er 4d spectra in Fig. 2(a), (c) and (e), it can be seen that only one broad peak appears at binding energy (BE) of w169.25 eV in the as-deposited erbium films, which is similar to pure Er2O3 in shape and peak position. However, a doublet appears in the Er 4d region (BE z 167.15 and 169.25 eV) just after argon ion sputtering for 30 s, which indicates that the oxide layer has been removed and the metallic erbium is exposed [12]. The O 1s spectra show two peaks (Fig. 2(b), (d) and (f)) in the as-deposited films. One peak at 531.30 eV is corresponding to metal oxide and the other at 533.05 eV is corresponding to the hydroxide [13] which disappears after argon etching for 30 s. The presence of these two peaks indicates that Er2O3 as well as Er(OH)3 are found in the sample surface. Tewell et al. also suggested that the lower energy peak in the O 1s region is attributed to Er2O3 and the higher one is attributed to oxygen from carbonates or alcohols [12]. The intensity of O 1s signal will decrease to zero after etching for 1800 s, 120 s and 240 s for the PMo-350, Mo-550 and PMo-550 samples, respectively. The relative concentration of oxygen/erbium (O/Er) with respect to etching time is displayed in Fig. 3. The relative concentration of erbium and oxygen is equal to the integration area divided by the sensitivity factor. The integration area is obtained by subtracting the shirley background from the original XPS data between 160205 eV and 528e536 eV for Er 4d and O 1s, and the sensitivity factors are 2.0 and 0.67 for the Er 4d and O 1s lines [14]. Fig. 3 shows that the O/Er decreases to zero after 240 s argon etching for PMo350 and 180 s for Mo-550 and PMo-550. The thickness of oxide layer of film deposited at 350 C is thicker than that deposited at 550 C, which is consistent with the surface color change of samples, i.e. the black color suggests that the thicker oxide layer formed during deposition. The argon sputtered spots show the silvery white color of erbium and the thickness of oxide layer were measured immediately by a stylus profiler. In addition, XPS results can also be used to calculate the thickness of oxide layer based on the etching rate (w25 nm/min). The thickness of oxide layer is 75 nm for PMo-350 and 50 nm for Mo-550 and PMo-550,
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
1099
Fig. 2. Evolution of the Er 4d and O 1s core level spectra of erbium film with surface erbium oxide layer as a function of etching time for films deposited on PMo (a and b) substrate at substrate temperature of 350 C and Mo (c and d), PMo (e and f) substrate at 550 C.
consistent with the results obtained from the stylus profiler. The results presented here, the film deposited at 350 C with thicker surface oxide layer than that deposited at 550 C, cannot be easily elucidated by conventional mechanism that annealing at higher temperature would strengthen the oxidation level. Various possible factors contributed to the erbium film oxidation behavior are presented as follows: (I) Although there is no data about the stability of erbium oxide at different temperatures in the diagram of variation of standard free energy changes as a function of temperature for the formation of oxides (i.e. Ellingham diagrams [15]), the enthalpy of formation 6Ho298 for Er2O3 at ambient temperature is 1889 kJ/mol, indicating a strong tendency to oxidize among the highest on the periodic table [16]. Thus, erbium
oxide layer always appears on the surface of erbium film and bulk. (II) The unstable amorphous structure may contribute to the formation of thicker oxide layer at low substrate temperature due to its poor crystallization. However, one can exclude this possibility from the XRD patterns of the completely crystalline erbium as shown in Fig. 1. (III) Hiramatsu et al. found that the sheet resistance of ZnO film was dramatically decreased after post-deposition annealing in vacuum at more than 300 C, while O2 desorbed from the film [17]. The schematic of the dynamic mechanism of erbium film growth and the diffusion path for oxygen from inward film across grain boundaries are illustrated in Fig. 4. The erbium will react with residual gases around the substrate to form erbium oxide or erbium hydroxide deduced from XPS results.
1100
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
Fig. 3. Concentration ratio of oxygen to erbium as a function of etching time for three representative specimens.
At the same time, the oxygen atoms diffuse outward. Therefore, only a thin oxide layer will be detected by XPS depth profiling and the etched spot area shows the silvery white color of erbium rather than the black color of erbium oxide. The schematic presented here can be used to interpret that the surface oxide layer is thinner when the substrate temperature is 550 C, because the diffusion coefficient of oxygen at substrate temperature of 550 C is higher than that of 350 C. (IV) In order to verify whether the oxide layer can be removed by the heat treatment processing, the films deposited at substrate temperature of 350 C were annealed at 550 C for half hour. Erbium can chemisorb residual gases quickly as a getter material and residual gases can be outgassed at high temperature. The annealed samples show the same black surface color as the as-deposited ones. So the oxide layer formed during deposition cannot be removed by the heat annealing. (V) There is a peak at 300 C (as shown in Fig. 5) when outgassing the Mo substrates to 700 C with a heating rate of 30 C/min, indicating the existence of rapid release stage of residual gases from molybdenum. The peak appeared in the heating process, however, is missing during cooling the substrate naturally. It is shown that the substrate has been outgassed
Fig. 4. The hypothetic dynamic processing of the formation of erbium oxide layer, erbium was deposited on Mo substrate with residual gases around it and buried by the following erbium adatom diffused to surface to form erbium oxide upon film growth.
Fig. 5. The outgassing curves of Mo substrate upon heating to 700 C with heating rate of 30 C/min and cooling naturally.
thoroughly after a outgas cycle and the low temperature oxidation behavior (i.e. erbium film is easily oxidized at lower substrate temperature) should not be affected by the outgas of Mo substrates.
Fig. 6. The cross-section morphologies of erbium thick films deposited on PMo substrate at substrate temperatures of 350 C (a) and 550 C (b).
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
1101
with the residual gases for its high reactivity as a getter material. Simultaneously, the erbium with chemisorbed the residual gases will be degassed by the high temperature field created by high substrate temperature (above 450 C) before deposition. In other words, the vacuum condition adjacent to the substrate is correspondingly rigorous when the substrate temperature is relatively higher. The major factor is that the gases adsorbed in erbium are outgassed by high substrate temperature. Therefore, the film deposited at higher temperature is not oxidized significantly as compared with film deposited at low temperature. It could be concluded that the oxidation behavior of erbium thick film is sensitive to the dynamic vacuum condition around the substrate. 4. Conclusions
Fig. 7. Five sets of outgassing curves of erbium bulk upon heating to 550 C with heating rate of 15 C/min and cooling naturally.
(VI) Fig. 6 shows the cross-section morphologies of erbium films deposited at substrate temperatures of 350 and 550 C. Although columnar grains are observed for erbium films that is independent of substrate temperature, the column size increases from w150 nm for 350 C to w350 nm for 550 C. The film with smaller column size deposited at 350 C (as shown in Fig. 6) has higher surface energy and more grain boundaries. The boundary is usually the rapid diffusion path for oxygen, so the film with more grain boundaries is susceptible to be oxidized. Consequently, the column size can be considered as one of the most important factors responsible for the oxidation behavior of erbium film. Furthermore, it should be noted that the dynamic vacuum condition around the substrate during deposition may be the most important reason on leading the films deposited at lower temperatures have relatively thicker surface oxide layer. The vacuum pressure monitored by vacuum gauge is 2.8 106 Torr when the erbium source material is liquation, and changes to a value lower than 1.0 106 Torr once heating of the erbium bulk has stopped, indicating that the residual gases are absorbed by erbium associating with the decreasing temperature of erbium bulk. The change in vacuum pressure can be observed repeatedly upon heating and cooling the erbium bulk. Five sets of outgassing curves are plotted in Fig. 7 for erbium bulk upon heating to 550 C with heating rate of 15 C/min and cooling naturally. As shown in Fig. 7, the vacuum pressure remains constant when the temperature of erbium bulk is below 300 C, whereas significantly increases with the increasing temperature, indicating the existence of starting temperature of decomposition of residual gases from erbium bulk. In addition, the starting temperature (w300 C) of decomposition is still present after 5 times of the degassing processing. Therefore, one can assume that it’s impossible to completely take away the residual gases chemisorbed inside the erbium bulk. The residual gases inside the erbium bulk would be desorbed when the erbium bulk was melting by electron-beam heating. The incoming erbium atoms will react
In summary, the colors of erbium thick films will change from black to gray with the increasing the substrate temperature. It indicates that the thickness of erbium oxide layer formed on the surface of erbium film distinctly decreases with the increasing substrate temperature, which is not consistent to the conventional mechanism. XPS depth profile results demonstrate the thickness of surface erbium oxide layer of erbium film deposited at substrate temperature of 550 and 350 C are nearly 50 and 75 nm, respectively. The grain size and the dynamic vacuum condition around the substrate are thought to be the major factors responsible for the formation of thinner oxide layer at higher temperature. Acknowledgments L. Yang and X. T. Zu are grateful for the support by the National Natural Science Foundation of China - NSAF (Grant No: 10976007) and the Fundamental Research Funds for the Central Universities (Grant No: ZYGX2009J040). S.M. Peng and X.G Long are grateful for the support by the Science and Technology Foundation of CAEP (Grant No: 2009A0301015) and the Major Program of the National Natural Science Foundation of China (Grant No: 91126001). References [1] Dow PA, Briers GW, Dewey MAP, Stark DS. Nucl Instrum Methods 1968;60(3): 293e6. [2] Lundin CE. Trans Metall Soc AIME 1968;242:1161e5. [3] Ferrizz RM. Sandia report, SAND2007e2659. [4] Bond GM, Browning JF, Snow CS. J Appl Phys 2010;107:083514. [5] Brumbach MT, Ohlhausen JA, Zavadil KR, Snow CS, Woicik JC. J Appl Phys 2011;109:114911. [6] Holloway DM. Appl Spectrosc 1973;27(2):95e8. [7] Parish CM, Snow CS, Brewer LN. J Mater Res 2009;24(5):1868e79. [8] Parish CM, Snow CS, Kammler DR, Brewer LN. J Nucl Mater 2010;403(1e3): 191e7. [9] Zhang L, Xie CY, Wu JS. Mater Charact 2007;58(5):471e8. [10] Roberts MW, Wells BR. Surf Sci 1969;15(2):325e32. [11] Jestis JC, Pereira P, Carrazza J, Zaera F. Surf Sci 1996;369(1e3):217e30. [12] Tewell CR, King SH. Appl Surf Sci 2006;253(5):2597e602. [13] Swami GTK, Stageberg FE, Goldman AM. J Vac Sci Technol A 1984;2(2): 767e70. [14] Wagner CD, Davis LE, Zeller MV, Taylor JA, Raymond RH, Gale LH. Surf Interface Anal 1981;3(5):211e25. [15] Ellingham HJT. J Soc Chem Ind 1944;63:125e33. [16] DeHoff RT. Thermodynamics in materials science. New York: McGraw-Hill; 1993. [17] Hiramatsu T, Furuta M, Matsuda T, Li CY, Hirao T. Appl Surf Sci 2011;257(13): 5480e3.
Contents lists available at SciVerse ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
The effect of substrate temperature on the oxidation behavior of erbium thick films H.H. Shen a, S.M. Peng b, X.G. Long b, X.S. Zhou b, L. Yang a, X.T. Zu a, * a b
Department of Applied Physics, University of Electronic Science and Technology of China, No. 4, Section 2, North Jianshe Road, Chengdu 610054, China Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics, Mianyang 621900, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 June 2011 Received in revised form 15 September 2011 Accepted 5 October 2011
The effect of substrate temperature on the oxidation behavior of erbium thick films, fabricated by electron-beam vapor deposition (EBVD), was investigated by X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and scanning electron microscopy (SEM). The erbium thick film is black when it is deposited at substrate temperature below 450 C and turns gray at higher substrate temperature in a vacuum pressure of approximately 1.5 106 Torr, which indicates that the thickness of erbium oxide layer formed on the surface of erbium films increases with the decreasing substrate temperature. XPS depth profile results demonstrate that the thickness of the surface erbium oxide layer of erbium film deposited at substrate temperature of 550 and 350 C are about 50 and 75 nm, respectively. The thicker oxide layer at lower substrate temperatures may be attributed to grain size and the dynamic vacuum condition around the substrates. Other possible factors involved in the oxidation behavior are also discussed. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Erbium thick film Substrate temperature Oxidation behavior Grain size
1. Introduction Films of erbium, a rare earth metal, with thickness of 3e6 mm is under consideration to be a candidate material as tritium (T) target of neutron generator (NG) [1]. The critical reasons are because of the property of long time storage in solid phase of its hydride [2], as well as excellent thermal stability of tritium [3] and helium (He) [4] generated by the decay of T in the Er-T-He ternary system. Systematic research on erbium films has been employed for its potential application in the nuclear industry. The erbium oxide layer, formed on the surface of erbium film due to its high susceptibility to oxidation in the deposition process, degrades the thermal dynamic properties for absorbing T by inhibiting T diffusion into the film to react with erbium. Brumbach et al. found that the thermal annealing either with higher temperature or with longer annealing times leads to a degradation of the passive oxide leading to a bulk film more accessible for hydrogen loading [5]. Holloway reported that the electron-beam deposited erbium films had surface oxide layer with thickness ranged from 17 to 19 nm after exposure to air, whereas the hydrided erbium films had oxide thickness of 11e35 nm after annealing at temperatures up to 500 C in a ambient at 1.0 108 Torr [6]. Parish et al. demonstrated that
* Corresponding author. Tel./fax: þ86 28 83202130. E-mail address: [email protected] (X.T. Zu). 0042-207X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2011.10.009
the erbium oxide formation would proceed readily during processing, which may detrimentally contaminate Er(H,D,T)2 films [7]. An additional platinum layer deposited on erbium surface immediately after erbium deposition without air exposure between deposition steps was employed to avoid the formation of erbium oxide compound and improve the capacity of absorbing T [8]. Generally, high temperature annealing would enhance the level of oxidation of metal bulk and film specimens due to their stronger reactivity at high temperature. Thermal oxidation induces a surface smoothening of TieNi thin films and surface roughness of oxidized TieNi films decreases with the increasing oxidation temperature [9]. Roberts et al. believed that via increasing the temperature of the film from 196 C to 183 C, sufficient energy has been provided to allow relatively facile entry of oxygen into the subsurface region [10]. Heating above 327 C for Ni films leads to the diffusion of oxygen atoms into the bulk and the partial reduction of the surface Ni to its metallic state [11]. Grain size can also serve as a significant factor to affect the oxidation behavior of metal materials because the surface energy of the film with small grains is higher than that of big grain film. Above all, eliminating or reducing the possibility of erbium to be oxidized is an effective way to promote its capability as T storage material. However, the mechanism of oxidation behavior of erbium film and the factors influencing the oxidation behavior at lower substrate temperatures are not clear up to now. In this study, erbium films are deposited on the rolled molybdenum substrates with substrate temperatures
1098
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
ranging from 200 C to 650 C, in order to understand the effect of substrate temperature on the oxidation behavior. 2. Experimental procedure A series of 2e4 mm erbium thick films were electron-beam evaporated on the rolled molybdenum (Mo) substrates and mechanically polished-molybdenum (PMo, surface roughness RMS < 5 nm) substrates at substrate temperatures of 200 Ce650 C. Evaporation rates detected by IC5 thin film deposition controller (Inficon, USA) were 5 or 10 nm/s. Before deposition, all substrates were ultrasonically cleaned with ethanol and subsequently in acetone. In addition, the molybdenum holder and molybdenum substrates were outgassed by increasing the temperature from room temperature to 700 C and heating for 2 h at 700 C. The source material erbium bulk (99.95% purity, 25 mm in diameter) were outgassed by melting it using electron-beam to remove the surface oxide layer and possible residual gases absorbed from air. After outgas of substrates and source material, the molybdenum substrates were cooled to the desired temperature for film deposition and the vacuum pressure was better than 3.0 106 Torr. The erbium was then deposited onto the substrates with the direction parallel to the substrate normal. To obtain uniform film, the sample holder was rotated during deposition and the distance between source material and sample holder was w20 cm. The microstructure of the as-deposited erbium thick films were identified by XRD using a X’pert PRO MPD (Panalytical, Holand) with a Cu Ka radiation. The cross-section morphologies of films were characterized using a Apollo 300 FE-SEM (Obducat CamScan, UK). The XPS ion sputtering depth profiling of oxygen and erbium was performed for three representative erbium films with different surface colors by using a ESCALAB 250Xi spectrometer (Thermo Scientific, UK). The X-ray source was a non-monochromatic Al Ka (150 W) lamp at photon energy 1486.6 eV. The residual gas pressure in the analyzing chamber was typically less than 1.8 109 mbar. The specimens were ion sputtered using an argon gun operated at 3 keV to conduct in-depth profile analysis and to determine the thickness of the oxide layer. The argon ion beam with current of 2.1 mA was rastered over a 0.6 mm 0.6 mm area yielding a sputtering rate of w25 nm/min and the etching thickness was measured by Dektak 150 Stylus Profiler (Veeco, USA). 3. Results and discussion The erbium films were deposited onto Mo and PMo substrate at different substrate temperatures of 200 C, 350 C, 450 C, 550 C, and 650 C. For all samples, the color of as-deposited erbium films changes from black to gray with the increasing substrate temperature. The most clear color change occurs between 350 and 450 C. Therefore, only three representative samples (listed in Table 1) with different colors deposited at substrate temperature of 350 and 550 C were characterized by the XPS depth profiling technique. There is no diffraction peak of erbium oxide in the XRD patterns of PMo-350, Mo-550 and PMo550 samples (as shown in Fig. 1), which
Table 1 Sample specifications. Substrate
Substrate temperature ( C)
Surface color
Thickness of oxide layer (nm)
Designation
PMo Mo PMo
350 550 550
Black gray gray
w75 w50 w50
PMo-350 Mo-550 PMo-550
Fig. 1. The XRD patterns of erbium thick films deposited on (a) PMo substrate at substrate temperatures of 350 C and (b) Mo, (c) PMo substrate at 550 C. The structures of hexagonal-close-packed (hcp) and body-centered-cubic (bcc) for erbium (Er) and molybdenum (Mo) are observed in each of the XRD patterns, whereas the diffraction peak of erbium oxide does not appear for any sample.
indicates that the oxide layer is extremely thin in comparison to the thickness of erbium film. Evolution of the Er 4d and O 1s core level spectra as a function of etching time are shown in Fig. 2 for PMo-350, Mo-550 and PMo550 samples. From the Er 4d spectra in Fig. 2(a), (c) and (e), it can be seen that only one broad peak appears at binding energy (BE) of w169.25 eV in the as-deposited erbium films, which is similar to pure Er2O3 in shape and peak position. However, a doublet appears in the Er 4d region (BE z 167.15 and 169.25 eV) just after argon ion sputtering for 30 s, which indicates that the oxide layer has been removed and the metallic erbium is exposed [12]. The O 1s spectra show two peaks (Fig. 2(b), (d) and (f)) in the as-deposited films. One peak at 531.30 eV is corresponding to metal oxide and the other at 533.05 eV is corresponding to the hydroxide [13] which disappears after argon etching for 30 s. The presence of these two peaks indicates that Er2O3 as well as Er(OH)3 are found in the sample surface. Tewell et al. also suggested that the lower energy peak in the O 1s region is attributed to Er2O3 and the higher one is attributed to oxygen from carbonates or alcohols [12]. The intensity of O 1s signal will decrease to zero after etching for 1800 s, 120 s and 240 s for the PMo-350, Mo-550 and PMo-550 samples, respectively. The relative concentration of oxygen/erbium (O/Er) with respect to etching time is displayed in Fig. 3. The relative concentration of erbium and oxygen is equal to the integration area divided by the sensitivity factor. The integration area is obtained by subtracting the shirley background from the original XPS data between 160205 eV and 528e536 eV for Er 4d and O 1s, and the sensitivity factors are 2.0 and 0.67 for the Er 4d and O 1s lines [14]. Fig. 3 shows that the O/Er decreases to zero after 240 s argon etching for PMo350 and 180 s for Mo-550 and PMo-550. The thickness of oxide layer of film deposited at 350 C is thicker than that deposited at 550 C, which is consistent with the surface color change of samples, i.e. the black color suggests that the thicker oxide layer formed during deposition. The argon sputtered spots show the silvery white color of erbium and the thickness of oxide layer were measured immediately by a stylus profiler. In addition, XPS results can also be used to calculate the thickness of oxide layer based on the etching rate (w25 nm/min). The thickness of oxide layer is 75 nm for PMo-350 and 50 nm for Mo-550 and PMo-550,
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
1099
Fig. 2. Evolution of the Er 4d and O 1s core level spectra of erbium film with surface erbium oxide layer as a function of etching time for films deposited on PMo (a and b) substrate at substrate temperature of 350 C and Mo (c and d), PMo (e and f) substrate at 550 C.
consistent with the results obtained from the stylus profiler. The results presented here, the film deposited at 350 C with thicker surface oxide layer than that deposited at 550 C, cannot be easily elucidated by conventional mechanism that annealing at higher temperature would strengthen the oxidation level. Various possible factors contributed to the erbium film oxidation behavior are presented as follows: (I) Although there is no data about the stability of erbium oxide at different temperatures in the diagram of variation of standard free energy changes as a function of temperature for the formation of oxides (i.e. Ellingham diagrams [15]), the enthalpy of formation 6Ho298 for Er2O3 at ambient temperature is 1889 kJ/mol, indicating a strong tendency to oxidize among the highest on the periodic table [16]. Thus, erbium
oxide layer always appears on the surface of erbium film and bulk. (II) The unstable amorphous structure may contribute to the formation of thicker oxide layer at low substrate temperature due to its poor crystallization. However, one can exclude this possibility from the XRD patterns of the completely crystalline erbium as shown in Fig. 1. (III) Hiramatsu et al. found that the sheet resistance of ZnO film was dramatically decreased after post-deposition annealing in vacuum at more than 300 C, while O2 desorbed from the film [17]. The schematic of the dynamic mechanism of erbium film growth and the diffusion path for oxygen from inward film across grain boundaries are illustrated in Fig. 4. The erbium will react with residual gases around the substrate to form erbium oxide or erbium hydroxide deduced from XPS results.
1100
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
Fig. 3. Concentration ratio of oxygen to erbium as a function of etching time for three representative specimens.
At the same time, the oxygen atoms diffuse outward. Therefore, only a thin oxide layer will be detected by XPS depth profiling and the etched spot area shows the silvery white color of erbium rather than the black color of erbium oxide. The schematic presented here can be used to interpret that the surface oxide layer is thinner when the substrate temperature is 550 C, because the diffusion coefficient of oxygen at substrate temperature of 550 C is higher than that of 350 C. (IV) In order to verify whether the oxide layer can be removed by the heat treatment processing, the films deposited at substrate temperature of 350 C were annealed at 550 C for half hour. Erbium can chemisorb residual gases quickly as a getter material and residual gases can be outgassed at high temperature. The annealed samples show the same black surface color as the as-deposited ones. So the oxide layer formed during deposition cannot be removed by the heat annealing. (V) There is a peak at 300 C (as shown in Fig. 5) when outgassing the Mo substrates to 700 C with a heating rate of 30 C/min, indicating the existence of rapid release stage of residual gases from molybdenum. The peak appeared in the heating process, however, is missing during cooling the substrate naturally. It is shown that the substrate has been outgassed
Fig. 4. The hypothetic dynamic processing of the formation of erbium oxide layer, erbium was deposited on Mo substrate with residual gases around it and buried by the following erbium adatom diffused to surface to form erbium oxide upon film growth.
Fig. 5. The outgassing curves of Mo substrate upon heating to 700 C with heating rate of 30 C/min and cooling naturally.
thoroughly after a outgas cycle and the low temperature oxidation behavior (i.e. erbium film is easily oxidized at lower substrate temperature) should not be affected by the outgas of Mo substrates.
Fig. 6. The cross-section morphologies of erbium thick films deposited on PMo substrate at substrate temperatures of 350 C (a) and 550 C (b).
H.H. Shen et al. / Vacuum 86 (2012) 1097e1101
1101
with the residual gases for its high reactivity as a getter material. Simultaneously, the erbium with chemisorbed the residual gases will be degassed by the high temperature field created by high substrate temperature (above 450 C) before deposition. In other words, the vacuum condition adjacent to the substrate is correspondingly rigorous when the substrate temperature is relatively higher. The major factor is that the gases adsorbed in erbium are outgassed by high substrate temperature. Therefore, the film deposited at higher temperature is not oxidized significantly as compared with film deposited at low temperature. It could be concluded that the oxidation behavior of erbium thick film is sensitive to the dynamic vacuum condition around the substrate. 4. Conclusions
Fig. 7. Five sets of outgassing curves of erbium bulk upon heating to 550 C with heating rate of 15 C/min and cooling naturally.
(VI) Fig. 6 shows the cross-section morphologies of erbium films deposited at substrate temperatures of 350 and 550 C. Although columnar grains are observed for erbium films that is independent of substrate temperature, the column size increases from w150 nm for 350 C to w350 nm for 550 C. The film with smaller column size deposited at 350 C (as shown in Fig. 6) has higher surface energy and more grain boundaries. The boundary is usually the rapid diffusion path for oxygen, so the film with more grain boundaries is susceptible to be oxidized. Consequently, the column size can be considered as one of the most important factors responsible for the oxidation behavior of erbium film. Furthermore, it should be noted that the dynamic vacuum condition around the substrate during deposition may be the most important reason on leading the films deposited at lower temperatures have relatively thicker surface oxide layer. The vacuum pressure monitored by vacuum gauge is 2.8 106 Torr when the erbium source material is liquation, and changes to a value lower than 1.0 106 Torr once heating of the erbium bulk has stopped, indicating that the residual gases are absorbed by erbium associating with the decreasing temperature of erbium bulk. The change in vacuum pressure can be observed repeatedly upon heating and cooling the erbium bulk. Five sets of outgassing curves are plotted in Fig. 7 for erbium bulk upon heating to 550 C with heating rate of 15 C/min and cooling naturally. As shown in Fig. 7, the vacuum pressure remains constant when the temperature of erbium bulk is below 300 C, whereas significantly increases with the increasing temperature, indicating the existence of starting temperature of decomposition of residual gases from erbium bulk. In addition, the starting temperature (w300 C) of decomposition is still present after 5 times of the degassing processing. Therefore, one can assume that it’s impossible to completely take away the residual gases chemisorbed inside the erbium bulk. The residual gases inside the erbium bulk would be desorbed when the erbium bulk was melting by electron-beam heating. The incoming erbium atoms will react
In summary, the colors of erbium thick films will change from black to gray with the increasing the substrate temperature. It indicates that the thickness of erbium oxide layer formed on the surface of erbium film distinctly decreases with the increasing substrate temperature, which is not consistent to the conventional mechanism. XPS depth profile results demonstrate the thickness of surface erbium oxide layer of erbium film deposited at substrate temperature of 550 and 350 C are nearly 50 and 75 nm, respectively. The grain size and the dynamic vacuum condition around the substrate are thought to be the major factors responsible for the formation of thinner oxide layer at higher temperature. Acknowledgments L. Yang and X. T. Zu are grateful for the support by the National Natural Science Foundation of China - NSAF (Grant No: 10976007) and the Fundamental Research Funds for the Central Universities (Grant No: ZYGX2009J040). S.M. Peng and X.G Long are grateful for the support by the Science and Technology Foundation of CAEP (Grant No: 2009A0301015) and the Major Program of the National Natural Science Foundation of China (Grant No: 91126001). References [1] Dow PA, Briers GW, Dewey MAP, Stark DS. Nucl Instrum Methods 1968;60(3): 293e6. [2] Lundin CE. Trans Metall Soc AIME 1968;242:1161e5. [3] Ferrizz RM. Sandia report, SAND2007e2659. [4] Bond GM, Browning JF, Snow CS. J Appl Phys 2010;107:083514. [5] Brumbach MT, Ohlhausen JA, Zavadil KR, Snow CS, Woicik JC. J Appl Phys 2011;109:114911. [6] Holloway DM. Appl Spectrosc 1973;27(2):95e8. [7] Parish CM, Snow CS, Brewer LN. J Mater Res 2009;24(5):1868e79. [8] Parish CM, Snow CS, Kammler DR, Brewer LN. J Nucl Mater 2010;403(1e3): 191e7. [9] Zhang L, Xie CY, Wu JS. Mater Charact 2007;58(5):471e8. [10] Roberts MW, Wells BR. Surf Sci 1969;15(2):325e32. [11] Jestis JC, Pereira P, Carrazza J, Zaera F. Surf Sci 1996;369(1e3):217e30. [12] Tewell CR, King SH. Appl Surf Sci 2006;253(5):2597e602. [13] Swami GTK, Stageberg FE, Goldman AM. J Vac Sci Technol A 1984;2(2): 767e70. [14] Wagner CD, Davis LE, Zeller MV, Taylor JA, Raymond RH, Gale LH. Surf Interface Anal 1981;3(5):211e25. [15] Ellingham HJT. J Soc Chem Ind 1944;63:125e33. [16] DeHoff RT. Thermodynamics in materials science. New York: McGraw-Hill; 1993. [17] Hiramatsu T, Furuta M, Matsuda T, Li CY, Hirao T. Appl Surf Sci 2011;257(13): 5480e3.