Fe whisker growth revisited: effect of Au catalysis for [021̄] oriented nanowires with 100 nm diameter

Abstract

We re-examined the growth of single crystalline Fe whiskers by chemical vapour deposition with regard to modern nanowire (NW) growth techniques. Single Fe NWs (100–300 nm diameter) with a high aspect ratio and unique [021̄] orientation were grown by optimizing the flow of raw materials (FeCl₂ and H₂) and using a Au nanoparticle catalyst and single crystalline substrate. The growth mechanism was investigated by transmission electron microscopy and electron diffraction analysis of the NW-catalyst interface.

---

Introduction

Single crystalline whiskers of Fe were synthesized by the simple reduction of FeCl₂ in a H₂ atmosphere without a gas flow in the 1960s–1970s^1–3 and used to study the dynamics of magnetic domains.⁴ However, Fe whiskers grown via the gas phase have almost been forgotten in the recent research trends of single crystalline nanowires (NWs) grown by chemical vapour deposition (CVD) techniques. This is because the diameter of the Fe whiskers grown by conventional CVD techniques is very large (several μm) and the orientation of the NWs could not be precisely controlled. In addition, the CVD used in 1960s–1970s for fabrication of Fe whiskers is now outdated because the Fe whiskers can only be grown on FeCl₂ powders by simple reduction in the crucible. Furthermore, the simple CVD system makes it impossible to precisely control the growth conditions.

Nevertheless, magnetic NWs, including Fe NWs, have been studied by many researchers due to not only their unique magnetic anisotropy but also to their potential for applications, such as a precursor of permanent magnets, spintronic devices, sensors, etc.^5–8 Although some research on the fabrication of Fe NWs by alternating current electrodeposition has been reported, NWs fabricated by electrodeposition are basically polycrystalline.^9–11 Large number of steps, including fabrication of alumina oxide (AAO) template, electrodeposition, and AAO removal, to yield the freestanding Fe NWs by electrodeposition method is another disadvantage. Nanoimprint-lithography is another technique to fabricate Fe NWs,¹² however, single crystalline NWs cannot be fabricated due to the use of epitaxial growth and subsequent etching process, resulting principally in polycrystalline NWs. The necessity of large equipment and unsuitability for large scale production are also disadvantage of nanoimprint-lithography.

Since the crystallinity of NWs strongly affects their magnetic properties,¹³ NWs should be single crystals. CVD is one of the best techniques to easily obtain single crystal NWs. Another advantage of CVD is to economically fabricate NWs in large quantity, which could realize various NW applications. The fabrication of Fe NWs (whiskers) has not been studied since the 1970s, yet the synthesis of Fe NWs using CVD techniques has become realistic because many kinds of NWs have been intensively investigated by CVD techniques for years.^14–20 In particular, it is well-known that the catalyst and substrate significantly affect the morphology of the NWs. Furthermore, the core–shell structure, which was recently studied due to its possibility for new applications, such as spin valve devices, drug delivery, and passivation of the NW surface, can be easily fabricated if CVD techniques to fabricate Fe NWs are established.^21–23

In this paper, we report the effect of Au nanoparticle catalysis, substrate, and orientation of the substrate. The growth mechanism will be discussed by investigating the morphology, crystal structure, and orientation of the NWs.

Experimental

Fig. 1 shows the CVD system to grow Fe NWs in this study. The furnace has three temperature-controllable zones which are labelled from upstream zone A, zone B, and zone C. FeCl₂·4H₂O (purchased from Kanto Chemical Co., Inc.) without any further purification and substrates were installed in zone B and zone C, respectively. Si and Al₂O₃ were chosen as the substrate. The Si substrates were cleaned by the Radio Corporation of America cleaning method to remove any metal and organic contaminations on the surface.²⁴ The Al₂O₃ substrates were annealed at 1000 °C for 5 hours to clean the surface before placing in the quartz tube. After the substrate were cleaned by these methods, Au nanoparticles (average diameter = 30 nm) dispersed in a citrate solution were spin-coated on the substrate. Au nanoparticles suspension was prepared by the same method in ref. 25. Subsequently, the substrates were annealed at 300 °C for 30 min in an ambient atmosphere to remove the citric acid from the surface of the Au nanoparticles,²⁵ and then the substrates were placed in the quartz tube. Since Fe can partially dissolve in Au,²⁶ it was expected that the Au nanoparticle is a good catalyst, and the Fe NWs grow by the vapour–liquid–solid (VLS) or vapour–solid (VS) mechanism.^17,27 The schematic illustrations of VLS and VS mechanism are shown in ESI (Scheme S1†). The distance between the FeCl₂·4H₂O and the substrates was approximately 8 cm. The flow paths of the reaction gas (3.9% H₂ + 96.1% Ar) and carrier gas (99.99% N₂) are separated from upstream to zone B. H₂ gas flowed through the inner quartz tube with the inner diameter of 4.0 mm, and the carrier gas with FeCl₂ vapour flowed through the outer quartz tube with the inner diameter of 37.5 mm. These two gases were mixed at the end of zone B. FeCl₂ and H₂ molecules in the vapour phase diffuse to the surface of the substrates and the CVD proceeded. The separation of the gas flow path hinders the reduction of the FeCl₂ powder in zone A, which enables a continuous supply of clean FeCl₂ vapour. This flow separation system is crucial methodology to synthesize the Fe NWs with a good repeatability and control.

Fig. 1 Schematic illustration of CVD with gas flow separation system for controllable fabrication of Fe NWs.

The N₂ flow started at the flow rate of 500 sccm after the FeCl₂·4H₂O and the substrates were installed in the quartz tube. H₂ gas was not introduced when the N₂ flow started in the early stage. The FeCl₂·4H₂O was heated at 200 °C for 30 min to remove the crystal water before the CVD was carried out. After removal of the crystal water from the FeCl₂·4H₂O, each zone of furnace was heated to a preset temperature within 2–3 hours. The temperatures of zone A and zone B was 550 °C and 600 °C, respectively, while the temperature of zone C was ranged from 600 °C to 920 °C as the growth condition parameter. When the temperatures of all the zones were attained at the preset temperatures, the H₂ flow started at the flow rate of 25 sccm. The reaction was continued for 5–30 min then the H₂ flow was immediately cut off after the desired reaction time was reached. The furnace was subsequently cooled to room temperature while maintaining the N₂ flow.

The morphology of the Fe NWs was observed by scanning electron microscopy (SEM; JEOL JSM-6500F and JSM-6390LVS). The crystallinity and growth axis of the Fe NW were investigated by X-ray diffraction (XRD; Rigaku RINT2200) and transmission electron microscopy (TEM; JEOL JEM-2010). An energy dispersive X-ray spectrometer (EDS) equipped with SEM was used for the elemental analysis.

Results and discussion

Fig. 2 shows SEM images of the surface of the Al₂O₃ substrate (0001) (C-plane) (a) with Au nanoparticles and (b) without Au nanoparticles after the CVD was carried out. Not only particles, but also a large number of NWs were observed on the Al₂O₃ substrates with Au nanoparticles while there are only prism-shaped particles on that without the Au nanoparticles. The growth temperature and the growth time were 700 °C and 30 min, respectively. NWs can be grown only at temperatures between 600 °C and 850 °C. At temperatures lower than 600 °C, only particles were formed on the substrate even though the Au nanoparticles were spin-coated on the substrates. This is because the reduction reaction proceeds very slowly at such a low temperature. At temperatures higher than 850 °C, the substrates were completely covered by a thin polycrystalline Fe film because the reduction reaction proceeds very rapidly, which means no selective growth at such high temperatures. These results indicate that the Au nanoparticles acted as catalysts to grow the Fe NWs, and the growth temperature must be precisely controlled. It was found that the diameter of the NWs (100–300 nm) was much larger than the size of the Au nanoparticles (30 nm), which implies that the diameter of the Fe NWs was determined by other factors. The average aspect ratio of the Fe NWs on the Al₂O₃ substrate was approximately 45. Although Si substrates spin-coated with Au nanoparticles were also attempted, no Fe NW was formed, which implies that the choice of the substrate was also important (see Fig. S1 in ESI†).

Fig. 2 Surface morphologies after CVD was carried out at 700 °C for 30 min (a) with Au nanoparticles and (b) without Au nanoparticles.

The NWs were examined by XRD and EDS as shown in Fig. 3(a) and (b). The NWs were grown at 700 °C for 30 min on C-plane Al₂O₃ substrates. In the XRD pattern, there are three main peaks which can be assigned to α-Fe 110, 200, and 112. No iron compound, such as Fe₂O₃ and FeCl₂, was observed, which shows that FeCl₂ was completely reduced by H₂, and the surface of the Fe NWs was not oxidized within several hours in air. EDS mapping also showed that the NWs were composed of Fe, and no other element was detected in the NWs, which supports the results of the XRD analysis. In addition, an elemental depth profile of the Fe NW stored in air for 3 days was obtained by Ar etching in the XPS analysis to investigate the oxidation tolerance that is important for application (see Fig. S2 in ESI†). Fe³⁺ was observed on the surface of the NWs. The intensity of Fe³⁺ decreased as the NW was etched deeper into the inside while the intensity of Fe⁰ increased. This result proves only several nm from the surface of the NW was oxidized and Fe completely remained inside. The oxidized Fe at the surface works as the passive state and hinders more oxidation.

Fig. 3 (a) XRD pattern and (b) Fe mapping by EDS of Fe NW grown at 700 °C for 30 min with Au nanoparticles.

Based on the SEM observations, we noticed many NWs with a greater aspect ratio were seen on the edge of the C-plane Al₂O₃ substrate (see Fig. S3 in ESI†). This clearly indicated that the growth rate of the Fe NWs depends on the orientation of the substrate. The M-plane Al₂O₃ substrate, which is perpendicular to the C-plane, was used to improve the aspect ratio of the Fe NWs. Fig. 4(a) and (b) show that the aspect ratio of the NWs grown on the M-plane Al₂O₃ (approximately 300) is greater than that on the C-plane Al₂O₃. An expanded figure of the Fe NW is shown in Fig. 4(b) as an inset.

Fig. 4 Comparison of the aspect ratio of the Fe NWs grown at 700 °C for 30 min on (a) C-plane Al₂O₃ (b) M-plane Al₂O₃.

The early stage of the NW growth was observed to understand the growth mechanism of the Fe NWs in detail. Fig. 5 shows the Fe NW grown at 700 °C for 5 min on the M-plane Al₂O₃ substrate with Au nanoparticles. The Au nanoparticles aggregated and became island-like structures (called a Au island below) because nanoparticles tend to minimize the surface energy.²⁸ The NW growth started at one point on the surface of a Au island as shown in Fig. 5 by the circle. The NW grew in two directions from the starting point and the growth direction was horizontal to the substrate. As seen in Fig. 5, there is no Au particle at the tip of the NW. This illustrates that the growth mechanism of Fe NWs is the self-catalytic VS mechanism, not the VLS mechanism. Secondly, the Fe NWs grown at 700 °C for 30 min on the M-plane Al₂O₃ substrates were investigated. No Au nanoparticles could be found on the surface of the substrate in an SEM image and no Au was detected in the EDS and XPS (not shown here). To confirm the location of Au, a cross section at the centre of the NW was observed by TEM as shown in Fig. 6(a). It was found that there was a thin Au layer (∼20 nm) between the NW and the substrate, which means that the Au islands after aggregation were covered by Fe if the reaction time is long enough. This is why Au was not detected by EDS and XPS. Note that the NW was connected to the Au layer, which indicates that the NW growth started from the surface of the Au island. This result is consistent with the SEM observations.

Fig. 5 Early stage of the NW growth from Au surface.

Fig. 6 (a) Cross-section TEM image of the Fe NW (b) electron diffraction pattern of the Fe NW shown in (a).

Next, an electron diffraction analysis was carried out to determine the direction of the NW growth. Fig. 6(b) shows the diffraction pattern of the Fe NW shown in Fig. 6(a). There are two non-equivalent reciprocal spots, i.e., 200 and 112. The growth axis of the NW must be perpendicular to these two vectors ((200) and (112)), therefore, this diffraction pattern shows that the growth axis of the Fe NW was [021̄]. The electron diffraction pattern of the cross section of the Fe particle was also investigated and two non-equivalent reciprocal spots were assigned to 200 and 110 (see Fig. S4 in ESI†). This result shows that the Fe NWs do not grow in the direction of [001] and the favourable growth axis of the Fe NWs is [021̄]. Combining the results of the XRD and electron diffraction (see Fig. 6(b)), the 200 peak in the XRD pattern originated from the Fe NWs and the other two peaks (110 and 112) in the XRD are mainly from the Fe particles.

Although the [021̄] oriented growth was reported in ref. 3, the mechanism has not been elucidated. It seems that the growth orientation of Fe NWs on FeCl₂ powder (ref. 1–3) is determined by chance, yet we suspect it is determined by the epitaxial orientation of Fe–Au interface in the present experiment. In order to clarify this point, we made cross section of the Fe NWs and examined by TEM.

Furthermore, we noticed in the cross-section TEM image that the Au layer beneath the Fe NW is parallel to the (200) plane of the Fe NW as shown in Fig. 6(a) and (b) (i.e., (200) vector is perpendicular to the Au layer). This result strongly suggests that the epitaxial growth occurred during the first step of the Fe NW growth. To investigate the orientation of the Au surface after aggregation, Au nanoparticles were spin-coated on the M-plane Al₂O₃ substrate and annealed at 700 °C for 30 min. Then an XRD analysis was carried out. The XRD pattern illustrates that some Au islands were oriented to (111) while other Au islands were oriented to (200) as shown in Fig. S5.† Based on the calculated ratio of the lattice constants of Au (a_Au) and Fe (a_Fe), a_Au/a_Fe = 0.408/0.286 = 1.426 ≅ (lattice mismatch is only 0.6%). Taking into account the crystal structures of α-Fe (bcc) and Au (fcc) and the lattice constants, Fe can epitaxially grow on Au 200.^29–31 These facts suggest that the epitaxial growth of Fe occurs on Au (100) during the first stage of the Fe NW growth. On the other hand, according to ref. 32, γ-Fe (fcc) can be grown on Au (111) when the thickness is less than 1 nm. A phase transition from fcc to bcc occurs when thickness is greater than 1 nm, and the orientation of α-Fe is (110). In addition, Fe grown on Au (111) tends to have an island structure due to the large surface free energy of Fe and small surface free energy of Au. The results in ref. 32 are consistent with the formation of the Fe particles and the 110 diffraction peak in the XRD.

The growth mechanism of the Fe NWs is schematically depicted in Scheme 1. While the temperature of the furnace is being raised, the Au nanoparticles aggregate due to the strong affinity with the substrate and minimization of the surface energy. FeCl₂ is reduced to Fe by H₂ (FeCl₂ + H₂ → Fe + 2HCl) on the surface of the Au islands. During the first stage of the reduction of FeCl₂, Fe is dissolved into the Au islands (solubility of Fe into Au at 700 °C is 0.57%).²⁶ Epitaxial growth occurs when the amount of dissolved Fe into Au had saturated. The NW growth starts in the direction of [021̄] once the epitaxial growth occurs on the surface of the Au island. Only axial growth occurred while the NW was growing as seen in Fig. 4(b) and 5. The diameter of the NW grown for 5 min (60 nm: shown in Fig. 5) does not differ very much from that grown for 30 min (120 nm: shown in Fig. 4(b)), while the lengths of the NW grown 5 min and 30 min are 560 nm and 37.6 μm, respectively. This selective axial growth can be explained by the stability of the sidewall and Ehrlich-Schwoebel barrier of the Fe NWs.³³ In other words, Fe atoms would prefer to attach to the tip of the NW than the sidewall. Additionally, adatoms cannot overcome the potential barrier at the step edge in order to move from the tip to the sidewall. Although the most stable surface of α-Fe is (100),³⁴ some researchers have reported a theoretical prediction by the embedded atom method, which concluded some surfaces are stabilized due to the relaxation of atoms at the surface.^35,36 In particular, since the total relaxation (perpendicular and parallel movement of atoms) of the (112) surface of the first layer is the highest among the high index surfaces such as (112), (310), and (210), which show the tendency to become a more symmetric surface structure, the (112) surface would be relatively stable among the metastable surfaces.³⁷ Furthermore, there is the possibility that the (120) surface is more stable than the (100) surface in the H₂ atmosphere due to the surface adsorption.

Scheme 1 Growth mechanism of the Fe NW.

Conclusions

We have fabricated Fe NWs on an Al₂O₃ substrate by CVD techniques. Separating the gas lines of H₂ and FeCl₂ vapour enabled us to continuously supply clean FeCl₂ vapour, which is crucial for the controlled and repeatable Fe NW growth. It was found that Au nanoparticles acted as catalysts and the morphology of the Fe NW strongly depended on the substrate. The M-plane Al₂O₃ was the best substrate to grow the Fe NWs with a high aspect ratio (approximately 300) During the first step of the NW growth, the Fe epitaxially grows on the surface of the Au islands with the orientation of (100). The epitaxial growth of Fe induces the NW growth, and the NWs grow horizontally to the substrate by a self-catalytic VS mechanism. The growth axis of the Fe NWs was determined to be [021̄] based on the XRD and electron diffraction patterns.

Acknowledgements

T. Yanase received financial support from the MEXT (Japan) program “Strategic Molecular and Materials Chemistry through Innovative Coupling Reactions” of Hokkaido University. A part of this work was conducted at “Joint-use Facilities: Laboratory of Nano-Micro Material Analysis”, Hokkaido University, supported by ”Nanotechnology Platform” Program of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Notes and references

1. S. Kittaka and T. Kaneko, Jpn. J. Appl. Phys., 1969, 8, 860.
2. T. Kaneko and S. Kittaka, J. Cryst. Growth, 1977, 42, 171.
3. R. N. Gardner, J. Cryst. Growth, 1978, 43, 425.
4. J. A. Salsgiver, IEEE Trans. Magn., 1970, 6, 741.
5. C. Cheze, L. Geelhaar, A. Trampert, O. Brandt and H. Riechert, Nano Lett., 2010, 10, 3426.
6. K. T. Chan, J. J. Kan, C. Doran, L. Ouyang, D. J. Smith and E. E. Fullerton, Nano Lett., 2010, 10, 5070.
7. L. Mohaddes-Ardabili, J. Zheng, S. B. Ogale, B. Hannoyer, W. Tian, J. Wang, S. E. Lofland, S. R.Shinde, T. Zhao, Y. Jia, L. Salamanca-Riba, D. G. Schlom, M. Wutting and R. Ramesh, Nat. Mater., 2004, 3, 533.
8. S. M. A. Shibli, K. S. Beenakumari and N. D. Suma, Biosens. Bioelectron., 2006, 22, 633.
9. S. Yang, H. Zhu, D. Yu, Z. Jin, S. Tang and Y. Du, J. Magn. Magn. Mater., 2000, 222, 97.
10. X. Wang, C. Li, G. Chen, C. Peng, L. He and L. Yang, Surf. Rev. Lett., 2010, 17, 419.
11. Y. Peng, T. H. Shen, B. Ashworth, X. G. Zhao, C. A. Faunce and Y. W. Liu, Appl. Phys. Lett., 2003, 83, 362.
12. W. Zhang, M. E. Bowden and K. M. Krishnan, J. Appl. Phys., 2013, 113, 17B502.
13. H. Pan, B. Liu, J. Yi, C. Poh, S. Lim, J. Ding, Y. Feng, C. H. A. Huan and J. Lin, J. Phys. Chem. B, 2005, 109, 3094.
14. J. Goldberger, R. He, Y. Zhang, S. Lee, H. Yan, H. J. Choi and P. Yang, Nature, 2003, 422, 599.
15. M. S. Gudiksen, L. J. Lauhon, J. Wang, D. C. Smith and C. M. Lieber, Nature, 2002, 415, 617.
16. L. J. Lauhon, M. S. Gudiksen, D. Wang and C. M. Lieber, Nature, 2002, 420, 57.
17. D. H. Kuo and B. J. Chang, Cryst. Growth Des., 2010, 10, 977.
18. K. Ikejiri, F. Ishizaka, K. Tomioka and T. Fukui, Nanotechnology, 2013, 24, 115304.
19. S. Hong, J. Yeo, W. Manorotkul, H. W. Kang, J. Lee, S. Han, Y. Rho, Y. D. Suh, H. J. Sung and S. H. Ko, Nanoscale, 2013, 5, 3698.
20. J. P. Alper, A. Gutes, C. Carraro and R. Maboudian, Nanoscale, 2013, 5, 4114.
21. K. T. Chan, C. Doran, E. G. Shipton and E. E. Fullerton, IEEE Trans. Magn., 2010, 46, 2209.
22. Y. Li, J. Li, K. Zhang and K. M. Liew, Carbon, 2012, 50, 566.
23. M. A. Zeeshan, S. Pane, S. K. Youn, E. Pellicer, S. Schuerle, J. Sort, S. Fusco, A. M. Lindo, H. G. Park and B. J. Nelson, Adv. Funct. Mater., 2013, 23, 823.
24. W. Kern, J. Electrochem. Soc., 1990, 137, 1887.
25. G. Frens, Nature (London), Phys. Sci., 1973, 241, 20.
26. H. Okamoto, Physica B, 1988, 149, 194.
27. R. S. Wagner and W. C. Ellis, Appl. Phys. Lett., 1964, 4, 89.
28. K. H. Cho and Y. M. Sung, Nanoscale, 2013, 5, 3690.
29. A. M. Begley, S. K. Kim, J. Quinn and F. Jona, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 48, 1779.
30. S. D. Bader and E. R. Moog, J. Appl. Phys., 1987, 61, 3729.
31. S. D. Bader, E. R. Moog and P. Grunberg, J. Magn. Magn. Mater., 1986, 53, L295.
32. D. T. Dekadjevi and B. J. Hickey, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 054108.
33. D. J. Shu, X. Xiong, Z. W. Wang, Z. Zhang, M. Wang and N. B. Ming, J. Phys. Chem. C, 2011, 115, 31.
34. G. Grochola, S. P. Russo, I. Yarovsky and I. K. Snook, J. Chem. Phys., 2004, 120, 3425.
35. M. I. Haftel, T. D. Andreadis and J. V. Lill, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 42, 11540.
36. J. Sokolov and F. Jona, Phys. Rev. B: Condens. Matter Mater. Phys., 1985, 31, 1929.
37. J. Sokolov and F. Jona, Solid State Commun., 1984, 49, 307.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra02966f

---