Improvement of field emission characteristics of tungsten oxide nanowires by hydrogen plasma treatment

Improvement of field emission characteristics of tungsten oxide nanowires by hydrogen plasma treatment

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2008 Europhys. Lett. 84 16001

(http://iopscience.iop.org/0295-5075/84/1/16001)

Download details:

IP Address: 140.116.208.56

The article was downloaded on 09/10/2009 at 07:35

Please note that terms and conditions apply.

The Table of Contents and more related content is available

HOME | SEARCH | PACS & MSC | JOURNALS | ABOUT | CONTACT US

doi: 10.1209/0295-5075/84/16001

Improvement of field emission characteristics of tungsten oxide nanowires by hydrogen plasma treatment

W.-C. Tsai 1 , S.-J. Wang 1(a) , C.-L. Chang 1 , C.-H. Chen 2 , R.-M. Ko 1 and B.-W. Liou ³

1 Institute of Microelectronics, Department of Electrical Engineering, National Cheng Kung University Tainan, 70101, Taiwan, Republic of China

2 Taiwan Semiconductor Manufacturing Company Ltd (TSMC) - Hsinchu 300, Taiwan, Republic of China

3 Department of Electronic Engineering, Wu-Feng Institute of Technology - Chiayi 62153, Taiwan, Republic of China received 1 May 2008; accepted in final form 12 August 2008

published online 18 September 2008

PACS 61.46.Km – Structure of nanowires and nanorods (long, free or loosely attached, quantum wires and quantum rods, but not gate-isolated embedded quantum wires) Abstract – The use of hydrogen plasma (H-plasma) treatment to improve field emission (FE) characteristics of self-synthesized tungsten oxide nanowires (TONWs) is reported. With a H-plasma treatment under a working power of 200 W and a pressure of 500 mtorr for 20 s, improved FE characteristics with a turn-on field (4.7 V/µm at 10 µA/cm

) lower than those of the as-grown case by 23% and a reduction in the effective emission barrier of 0.72 eV were obtained, which is attributed to the reduction in oxygen adsorption, decrease in the wire length and density, and transition of TONWs surfaces from well crystalline into the amorphous phase.

Copyright c
EPLA, 2008

Introduction. – Nanostructured tungsten oxides (WO _x ) have received considerable attention due to their high electrical conductivity, chemical stability and superior electron field emission (FE) characteristics [1,2].

The size, density, composition (i.e., the O/W ratio), and nanostructure of tungsten oxide nanowires (TONWs) play crucial roles in determining the FE performance [1,3–5].

However, control over the oxygen adsorption of W-based nanowires for FE applications is rarely accomplished due to the required harsh reaction conditions. In general, oxygen adsorption might arise from the oxygen conta- mination of source material during wire growth, the residual oxygen in sputtered films or intentionally doped oxygen gas (O ₂ ) during sputtering deposition [6], and oxygen/humidity adsorption of the grown TONWs [7].

Recently, Ar, O ₂ , and H ₂ plasma treatments have been used to improve the FE characteristics of carbon nanotubes (CNTs) by varying the morphologies and removing impurities or defects [8–10]. Hydrogen plasma (H-plasma) treatment has been shown to greatly improve FE because it effectively eliminates impurities and modifies both the structure and surface morphology of CNTs [10]. In this study, H-plasma treatment is used to reduce the amount of oxygen adsorption and to tailor the density and morphologies of TONWs. Improved FE

(a)

E-mail: [email protected]

characteristics are demonstrated and the related reduc- tion in the effective emission barrier height is analyzed and discussed.

Experiments. – In the experiments, pure tungsten (W) films with a thickness of 60 nm were sputter-deposited on n-type Si (100) wafers. The deposition was carried out under an Ar flow rate of 24 sccm, a dc power of 200 W, and a pressure of 7.6 mtorr at room temperature.

For the growth of TONWs, samples were subjected to thermal annealing in a quartz tube furnace at 700 ^◦ C in nitrogen (N ₂ ) ambient for 30 min. Figure 1(a) shows the typical scanning electron microscope (SEM) top and oblique view (inset) images of the surface morphology of samples after thermal annealing. Straight nanowires with a density of around 250 µm ⁻² and length/diameter of around 0.2 µm/20 nm were obtained. The self-growth of TONWs is attributed to phase transition of the W film during thermal annealing, which was found to start from the nucleus sites at the grain boundaries of WO _x nanocrystallites on the surface of samples [6,7]. Samples with grown TONWs then underwent H-plasma treatment in a PECVD system under a pressure of 500 mtorr for 10–60 s. H-plasma was generated from H ₂ with a flow rate of 600 sccm ± 10% at a working power of 200 W.

Results and discussion. – Figures 1(b)–(g) show the

top and oblique view (inset) SEM images of the prepared

W.-C. Tsai et al.

Fig. 1: SEM images of the top and oblique views (inset) of TONWs with and without H-plasma treatment. (a) As-grown, (b) 10 s, (c) 20 s, (d) 30 s, (e) 40 s, (f) 50 s, and (g) 60 s.

samples after H-plasma treatment for 10, 20, 30, 40, 50 and 60 s, respectively. The samples after H-plasma treatment for 10 (fig. 1(b)) and 20 s (fig. 1(c)) show no obvious change in morphology and density as compared with those of the as-grown case. When the treatment time was increased to 30 s (figs. 1(d)), ion-bombardment of H-plasma resulted in a reduction in both the wire length and density by typically 10% and 27%, respectively, compared to the as- grown TONWs. When the treatment time was further increased to 40 and 50 s, an additional reduction in the wire length and density was much more evident. For the 60 s case, the wire length and density drastically decreased by 89% and 60%, respectively, compared to the as-grown TONWs.

Figure 2 shows the TEM image and SAED pattern of an individual TONW obtained from the samples before and after H-plasma treatment for 20 and 30 s. As shown in figs. 2(a) and (d), clear stripes of the lattice plane extend over the whole wire. The corresponding SAED pattern shown in the inset shows that the nanowire has a well crystalline structure. The interplane distance of d-space was determined to be 3.78 ˚ A, indicating that the main crystalline phase of the wire is the W ₁₈ O ₄₉ (010) phase.

This is similar to that prepared by CVD [11] and to that synthesized on tungsten plates [2]. Note that the oxygen contributing to the growth of TONWs might arise from

Fig. 2: TEM images and SAED patterns for a single TONW with and without H-plasma treatment. (a) As-grown, (b) 20 s, and (c) 30 s. (d), (e), and (f) are the corresponding TEM images at high magnification.

oxygen adsorbed on the surfaces of the as-deposited films when they were exposed to open air. After H-plasma treat- ment for 20 and 30 s, as shown in figs. 2(b,c) and 2(e,f) at high magnification, respectively, the TONWs had the same crystalline phase of W ₁₈ O ₄₉ (010). However, it was found that a substantial portion of the wire surface turned into the amorphous phase. Similar results were found in oxygen-plasma treated CNTs reported by Chen [12]. The amorphous surface layer (ASL) increased in thickness with increasing H-plasma treatment time.

Figure 3 shows the surface XPS O1s and W4f energy spectra of the TONWs with and without H-plasma treatment. The deconvolutions of these spectra into Gaussian peaks are shown by dashed curves. As shown in fig. 3(a), the O1s spectra appear around 530–532 eV for all samples. The two deconvoluted Gaussian peaks indicate the existence of WO ₂ (∼ 530.5 eV) and WO ₃ (∼ 531.6 eV) crystalline phases. The intensity of the O1s peak decreases when the H-plasma treatment time increases from 10 to 30 s and then increases when the H-plasma time increases from 40 to 60 s. The O1s peak of the TONWs with H-plasma treatment for 30 s has the lowest intensity, suggesting that 30 s H-plasma treatment yields the most effective oxygen desorption to the surface of TONWs.

For samples with H-plasma treatment for 40, 50, and

60 s, since a large number of TONWs were removed,

the detected O1s peak intensity during XPS analysis

Fig. 3: The XPS spectra of the prepared samples with and without H-plasma treatment. (a) O1s spectra and (b) W4 f spectra. The dashed lines are the deconvoluted Gaussian curves.

might have come from the exposure surface of the thermal-annealed W film itself.

As shown in fig. 3(b), the three deconvoluted Gaussian peaks were at ∼ 31.4 (W4f _7/2 , denoted as peak I), ∼ 33.5 (W4f 5/2 , denoted as peak II), and ∼ 36 eV (W4f _7/2 of WO _x /W, denoted as peak III), respectively [13]. For the as-grown TONWs, the intensity of peak III was higher than those of peaks I and II. For TONWs after H-plasma treatment, the intensities of peaks I and II increased with increasing plasma treatment time, while the intensity of peak III decreased. These results show that H-plasma treatment led to considerable oxygen desorption on the surface of TONWs. For the cases of H-plasma treatment for 40, 50, and 60 s, the intensity of peaks I and II decreased but that of peak III increased, which is consistent with the behavior of the O1s peak.

Figure 4 illustrates the typical current density-electric field intensity (J-E) curves of TONWs with and without H-plasma treatment. Following the Fowler-Nordheim (FN) equation [14,15]:

J = I

α n = C 1 (βE) ² exp

−C 2 φ ^{3 2} βE



, (1)

Fig. 4: The typical J-E curves of TONWs with and with- out H-plasma treatment. The inset shows the FN plots (ln( J/E

) vs. 1 /E) of TONWs after H-plasma treatment.

where I, J, E, φ, and β are the emission current (A), emission current density (A/cm ² ), applied electric field (V/µm), emission barrier height (eV), and field enhance- ment factor, respectively; α is the effective emission area of a single wire in the emitter array [15], n is the wire density (cm ⁻² ), and C 1 (1.4 × 10 ⁻⁶ AV ⁻² ) and C 2 (6.8 × 10 ⁹ (eV) ^−3/2 Vm ⁻¹ ) are constants [14]. The corresponding FN plots are shown in the inset. The FN plots of samples with H-plasma treatment for 10, 20, 30 and 40 s can be represented by nearly straight lines, indi- cating that the electron emission of TONWs is probably dominated by the FN process. However, there was almost no detectable emission current for the as-grown sample and for those with 50 and 60 s H-plasma treatment for E < 5 V/µm. Note that the one with H-plasma treatment sample for 20 s had the best emission characteristics among all the prepared samples. It emitted a current density of 10 µA/cm ² at around 4.7 V/µm which is much lower than that of TONWs (∼ 9.5 V/µm) prepared by annealing of tungsten nitride (W ₂ N) films [16].

In general, FE current is a complex function of the work function, tip shape, diameter, length, and density of the emitter array. All of these factors affect the local field around the tip of TONWs and hence the FE current.

Nanowires with a lower work function, sharper emitter tip, and a better length/density ratio are expected to give better FE characteristics [3,8,12,17,18]. Our experimental results show that the key effects of ion-bombardment of H-plasma on TONWs include a decrease in O/W ratio (the 1st effect) as well as a transition of the crystalline phase from well crystalline to an amorphous structure (the 2nd effect), which reduce the electron emission barrier [4,12,19], and the wire length and density (the 3rd effect).

The first and second effect are expected to improve

FE characteristics because the former would significantly

reduce the resistivity of TONWs [20] and the latter might

reduce the effective emission barrier height by regarding

ASL as a coating layer to TONWs. Similar results have

been reported for ZrO-coated W [21], diamond-coated

W.-C. Tsai et al.

Si and Mo [22], and BN-coated SiC field emitters [23], where a coating layer thickness-dependent reduction in the effective emission barrier height was found.

However, the third effect of H-plasma treatment leads to different wire lengths and spaces between neighboring wires for different ion bombardment times; as a result, it plays an important part in determining the degree of field screening effect and, in turn, strongly affects FE characteristics [24–26]. The FE characteristics of TONWs after H-plasma treatment for 30 s were worse than those of the case of 20 s even with relatively better oxygen desorption. It is supposed that the third effect, which determines the emitter geometry, might have dominated the first and second effects.

To compare results obtained from samples with different H-plasma treatment times more quantitatively, the effect of H-plasma treatment on the reduction in the effective emission barrier height of TONWs was evaluated from the experimental FN plots shown in fig. 4. The intercept of the y-axis and the slope of a specific FN plot, y i

and m i , respectively, were extracted, and the density n i

was estimated from the corresponding SEM image. Then, according to eq. (1), the effective emission barrier of TONWs was calculated using the following equation:

φ i = 0.54

m ² _i × 10 ^y

n i α i

 _1/3

(meV), (2)

where the subscript “i” denotes the sample number. The effective emission area should be determined before the effective emission barrier height can be derived. Based on the fact that, as revealed by the SEM images shown in fig. 1, the diameter and tip shape of the as-grown and H-plasma treated TONWs for 10, 20, 30, and 40 s remain nearly unchanged, accordingly, the same α(= 2.4 × 10 ⁻¹⁰ cm ² ) derived using the as-grown TONWs with a typical barrier height of 6.2 eV [27], denoted as φ 0 , was used for the derivation of φ for these samples for simplicity.

The cases of as-grown, 50, and 60 s were not considered here due to their poor FE performance.

To alleviate other possible uncertainties in the FN equation, the variation in φ, i.e., ∆φ(= φ − φ 0 ), instead of φ itself, was evaluated. The calculated values of ∆φ for the case of 10, 20, 30 and 40 s were −0.01, −0.72,

−1.61, and −0.17 eV, respectively. Note that TONWs with H-plasma treatment for 30 s led to the largest reduction in the effective barrier height. Our results suggest that the improved FE characteristics obtained from TONWs with H-plasma treatment could be ascribed to the reduction in the effective FE barrier height as a whole. The calcu- lated results of ∆φ reflect the first and second effect of H-plasma treatment mentioned previously. It is suspected that the magnitude of ∆φ might be in part related to the thickness of ASL formed on the surface of TONWs.

The same results were observed for Si and Mo field emit- ters coated with various thicknesses of diamond layer [22], leading to a thickness-dependent barrier reduction and

improved FE behavior. An investigation on the weight- ings of the three effects of H-plasma treatment on field emission improvement and on ∆φ is now underway.

Conclusions. – H-plasma treatment under a pres- sure of 500 mtorr and a working power of 200 W was demonstrated to improve the electron FE characteristics of self-synthesized TONWs from sputtering-deposited pure tungsten films. H-plasma treatment on TONWs leads to a decrease in the O/W ratio, a transition of crystalline phase from well crystalline to an amorphous structure, and an apparent reduction in the wire length and density. In addition, H-plasma treatment for 30 s would yield the best oxygen adsorption and the largest effective barrier height reduction. However, possibly due to the field screening effect dominating the effective barrier height reduction, TONWs subjected to H-plasma treatment for 20 s were found to have the best FE characteristics (a turn-on field of about 4.7 V/µm at 10 µA/cm ² and a reduction in the effective emission barrier of 0.72 eV).

∗ ∗ ∗

This work was supported in part by the National Science Council (NSC) of Taiwan, Republic of China, under Contract No. NSC 96-2221-E-006-285-MY3. The authors would like to thank the Advanced Optoelectronic Technology Center and the Center for Micro/Nano Science and Technology, National Cheng Kung University, Taiwan, for equipment access and technical support.

REFERENCES

[1] Li Y., Bando Y. and Golberg D., Adv. Mater., 15 (2003) 1294.

[2] Gu G., Zheng B., Han W. Q., Roth S. and Liu J., Nano Lett., 2 (2002) 849.

[3] Jun Z., Li G., Shao Z. D., Jun C., Jun C. S., Ning S. X., Rusen Y. and Zhong L. W., Appl. Phys. Lett., 87 (2005) 223108.

[4] Kloss A., Motzke T., Grossjohann R. and Hess H., Phys. Rev. E, 54 (1996) 5851.

[5] Elbaum L. K., Ahn K., Souk J. H., Ting C. Y. and Nesbit L. A. , J. Vac. Sci. Technol. A, 4 (1986) 3106.

[6] Chen C. H., Wang S. J., Ko R. M., Kuo Y. C., Uang K. M., Chen T. M., Liou B. W. and Tsai H. Y., Nanotechnology, 17 (2006) 217.

[7] Wang S. J., Chen C. H., Ko R. M., Kuo Y. C., Wong C. H., Wu C. H., Uang K. M., Chen T. M. and Liou B. W. , Appl. Phys. Lett., 86 (2005) 263103.

[8] Choi J. H., Choi S. H., Han J. H., Yoo J. B., Park C.

Y., Jung T., Yu S., Han I. T. and Kim J. M., J. Appl.

Phys., 94 (2003) 487.

[9] Ahn K. S., Kim J. S., Kim C. O. and Hong J. P., Carbon, 41 (2003) 2481.

[10] Ryu K., Kang M., Kim Y. and Jeon H., Jpn. J. Appl.

Phys., 42 (2003) 3578.

[11] Li X. L., Liu J. F. and Li Y. D., Inorg. Chem., 42 (2003) 921.

[12] Chen Z., den Engelsen D., Bachmann P. K., van Elsbergen V., Koehler I., Merikhi J. and Wiechert D. U. , Appl. Phys. Lett., 87 (2005) 243104.

[13] Kojima I. and Kurahashi M., J. Electron Spectrosc.

Relat. Phenom., 42 (1987) 177.

[14] Satyanarayana B. S., Hart A., Milne W. I. and Robertson J. , Appl. Phys. Lett., 71 (1997) 1430.

[15] Hong D., Asiam M., Feldmann M. and Olinger M., J. Vac. Sci. Technol. B, 12 (1994) 764.

[16] Jeon S. and Yong K., Nanotechnology, 18 (2007) 245602.

[17] Li Y. B., Bando Y. and Golberg D., Appl. Phys. Lett., 84 (2004) 3603.

[18] Vila L., Vincent P., Pra L. D. D., Pirio G., Minoux E., Gangloff L., Champagne S. D., Sarazin N., Ferain E., Leger R., Piraux L. and Legagneux P., Nano Lett., 4 (2004) 521.

[19] Bonnet J. P., Nowotny J., Onillon M. and Sikora I. , Oxid. Met., 13 (1979) 273.

[20] Liu J., Wu J. and Zhu S., Bull. Tongji Univ. Shanghai (2001) http://chemweb.tongji.edu.cn.

[21] Fransen M. J., Overwijk M. H. F. and Kruit P., Appl. Surf. Sci., 146 (1999) 357.

[22] Zhimov V. V., Choi W. B., Cuomo J. J. and Hren J. J. , Appl. Surf. Sci., 94/95 (1996) 123.

[23] Tang C. and Bando Y., Appl. Phys. Lett., 83 (2003) 659.

[24] Nilsson L., Groening O., Emmenegger C., Kuettel O., Schaller E. and Schlapbach L., Appl. Phys. Lett., 76 (2000) 2071.

[25] Teo K., Chhowalla M., Amaratunga G. and Milne W. , Appl. Phys. Lett., 80 (2002) 2011.

[26] Bonard J.-M., Weiss N., Kind H., St¨ ockli T., Forr´ o L., Kern K. and Chˆ atelain A. , Adv. Mater., 13 (2001) 184.

[27] Kloss A., Motzke T., Grossjohann R. and Hess H.,

Phys. Rev. E, 54 (1996) 5851.