Ultramicroscopy 95 (2003) 145–151

Field emission from diamond particles studied byscanning field emission microscopy

Akihiko Watanabea,*, Masahiro Deguchib, Makoto Kitabatakea, Shozo Konoc

aFrontier Carbon Technology Project/JFCC, Center for Advanced Research Projects, 6F, Osaka University, 2-1 Yamada-oka, Suita,

Osaka 565-0871, JapanbAdvanced Technology Research Laboratories, Matsushita Electric Industrial Co., Ltd., Hikaridai 3-4, Seika, Soraku,

Kyoto 619-0237, Japanc Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Katahira 2-1-1, Aoba, Sendai 980-8577, Japan

Received 1 September 2001; received in revised form 11 December 2001

Abstract

Field emission properties from diamond particles (DPs) are studied. The DPs with thin chemically vapor deposited

(CVD) diamond overcoat, dispersed onto metal substrate, essentially exhibit negative electron affinity (NEA). Field

emission, approximately 1mA/cm2 under a macroscopic electric field of 3.5 kV/mm are observed. Microscopic electrical

properties were studied by scanning tunneling microscopy/spectroscopy. Most parts of the DP surface

exhibit narrow gap and p-type characteristics. The localized regions, which have wide gap like bulk diamond

properties, are randomly distributed near the top of DP. The field emission current distribution depicted by scanning

field emission microscopy (SFEM) show that the electron emission is originating from a localized region on the selected

DPs. We found, through SFEM measurement, some favorable field emission spots (‘‘hot spots’’) where measured

emission current is several orders higher than that of the other DPs (‘‘normal spots’’). Field emission spectroscopy

(FES) results suggest that a poorly conducting layer is present along the electron path from the metal electrode to

vacuum.

We propose two models for field emission from ‘‘hot spots’’, which involve two main mechanisms. One is

electron injection from the metal substrate to the DP, which is attributed to the electric field enhancement at intrinsic

non-doped diamond (i-diamond) layer sandwiched between the metal substrate and the surface conductive layer

(p-diamond) of the CVD diamond overcoat on the DP. The other is electron emission at the top site of NEA DP

through the local i-diamond region or the depletion region of the p-diamond, which is caused by the applied electric

field.

r 2002 Elsevier Science B.V. All rights reserved.

PACS: 79.70.+q; 73.90.+f; 07.79.�v

Keywords: Microparticle; Field emission; Morphology; Scanning tunneling microscopy

*Corresponding author. Tel.: +1-81-6-6879-4146; fax: +1-81-6-6879-4147.

E-mail address: nave@cam.hi-ho.ne.jp (A. Watanabe).

0304-3991/03/$ - see front matter r 2002 Elsevier Science B.V. All rights reserved.

PII: S 0 3 0 4 - 3 9 9 1 ( 0 2 ) 0 0 3 1 1 - X

1. Introduction

Diamond is a promising material because of itsunique mechanical and electrical properties.Hydrogen-terminated diamond surfaces havebeen reported to exhibit a negative electron affinity(NEA) [1,2]. Many reports on field emissionproperties of chemically vapor deposited (CVD)diamond thin films are published and find goodfield emission properties with low thresholdvoltage [3–8]. Mechanisms of field emission fromCVD diamond have been studied extensively andsome models are established [9–14]. However, ithas been reported that so-called CVD diamondfilms included diamond, graphite, diamond likecarbon, nanotubes, surface conductive layers, etc.,depending on the growth conditions. Electronemission from the CVD diamond film is affectedby a mixture of these complicated elements. Theorigin of field emission from CVD diamond isexplained in most literatures as follows, (1) con-duction channels and/or grain boundaries areformed, (2) small protuberances which locallyenhance the electric field are formed on thesurface. There is also some convincing experimen-tal work on the effect of NEA [29], but it is notclear in detail how NEA surfaces affect the fieldemission from diamond under these circum-stances. To examine the proposed models, it isnecessary to investigate emission sites in detailwith high spatial resolution. The method using asharp tip, as in scanning tunneling microscopy(STM) and atomic force microscopy, is useful forsuch a high spatially resolved analyses. Someresults were already reported using this techniqueto analyze surface properties of CVD diamond[15–21].In this paper, the field emission from diamond

particles (DPs) with a CVD diamond overcoat isstudied. Samples indicated NEA surface alwaysshow better field emission characteristics than thatof samples with positive electron affinity (PEA)[22–25]. High spatially resolved analyses areperformed using scanning tunneling microscopy/spectroscopy (STM/STS), and scanning field emissionmicroscopy (SFEM). STS, which performedsimultaneously with STM, makes it possibleto study local electric properties of surfaces of

DPs by measuring the tunneling current as afunction of the voltage applied to the sample.SFEM enables investigation of field emissionindividual DP.

2. Experimental details

High-pressure synthetic DPs with an average sizeof approximately 1mm were mixed with organicepoxy to form a DP suspension. A typical DPconcentration in the DP suspension was 0.05 g/cm3.Then, the DP suspension was spin coated ontopolycrystalline tungsten substrates. After the spin-coat process, the DP-seeded substrate was annealedat 2001C for 20min in air in order to evaporate anorganic volatile solvent. Consequently, DPs wereuniformly dispersed on the substrate. Then anundoped thin diamond layer was grown on theDP surfaces by a conventional microwave plasma-assisted CVD method. The CVD growth wasperformed using a source gas of CH4(1.0%)/H2with a microwave power of 300W, substratetemperature of 8501C, and growth time of 15min.The sample was exposed to a hydrogen plasma for5min after the CVD growth to ensure surfacetermination with hydrogen and cooled down in thehydrogen gas atmosphere.The morphology, surface electrical properties

and field emission characteristics of individual DPswere studied using an STM/STS system which wasmodified to allow SFEM. This STM/STS/SFEMsystem was housed in a UHV chamber at apressure of approximately 2� 10–8 Pa. A chemi-cally etched tungsten tip was used as a scanningprobe. The procedures for measuring SFEM ofDP were as follows. At first, a topographic STMimage of the surface was taken with typically0.3 nA tip-to-DP current and with +1.5V appliedto the sample. Then, the tungsten tip was movedupward away from the top site of the DP by apiezo-drive coarse-motion system, and then thevertical height (z-position) of the tip was fixed andscanned laterally (x–y) with constant appliedvoltage to the sample. One can judge the emissionsite on a DP and spatial distribution of emittedelectron from the image obtained by this proce-dure of SFEM. At last, a microscopic field

A. Watanabe et al. / Ultramicroscopy 95 (2003) 145–151146

emission I2V measurement was performed at theemission site, at the chosen fixed distance z.

3. Results and discussion

A photoemission spectrum (UPS) of the above-mentioned sample is shown in Fig. 1. This wasmeasured with He I (21.2 eV) radiation; �6.5Vapplied to the sample during the UPS measure-ment to overcome the retarding potential due tothe work function of the analyzer. The Fermi level(EF) position of the sample was determined from aspecific UPS spectrum of Aquadag [26] placed onthe sample. The position of the valence bandmaximum (EVBM: 1.0 eV below EF) of the samplewas determined by linear extrapolation to zero ofthe high-energy cutoff in the UPS spectrum. Asharp peak is clearly observed at �16.7 eV in low-energy cutoff in the UPS spectrum shown inFig. 1. The energy position of conduction bandminimum (ECBM) final position is calculatedas ECBM ¼ EVBM þ Eg � hn ¼ �1:0þ 5:5� 21:2 ¼�16:7 eV where the band gap of bulk diamond(Eg ¼ 5:5 eV) and the excitation energy(hn ¼ 21:2 eV) is applied. Energy position of thesharp peak in Fig. 1 corresponds to this ECBMenergy position. This correspondence suggests thata large number of secondary electrons accumulateat ECBM and efficiently emit from the diamond

surface without any energy barrier. Thus it isconfirmed that the surface of this sample exhibitsNEA [19,27]. The surface NEA characteristic ofthe sample was also confirmed using secondary-electron spectroscopy [24,25]. Samples indicatedNEA surface always show better field emissioncharacteristics than that of samples with PEA[22–25].Fig. 2 shows STM/STS results of the DP. Local

electrical properties of the DP surfaces aretypically classified into two characteristic region(A) and (B). Region (A) has a narrow gap ofapproximately 1 eV near EF (Fig. 2(a)). Most partsof the DP surface show such local electricalproperties as region (A). Moreover, the STScurrent when positive voltage applied to thesample, is larger than that under negative appliedvoltage in this region (A). This result suggests thatmost parts of the DP surface have p-type electricalcharacteristics (p-diamond), which agrees with thereported surface conductive layer on CVD dia-mond surface [28]. The other region (B) has a widegap like a bulk diamond (i-diamond) (Fig. 2(b)).Such wide gap regions (B) locally exist the top siteof DP (Fig. 2(c)).Results of field-emission observations are shown

in Figs. 3. In the topographic STM image(Fig. 3(a)), several DPs are observed in 4� 4 mm2

image area. The SFEM image shown in Fig. 3(b)was taken on the corresponding image area afterthe tip was moved 1.6 mm upward away from thesurface and �400V was applied to the sample.Even though several DPs are present in this imagearea there are only two bright SFEM regions,where emission current is detected, in Fig. 3(b).Only some particular DPs contribute to electronemission at the applied electric field. The largestcurrent of approximately 150 nA is measured atthe center of the round shaped emission region inFig. 3(b). These results imply that the electronsemit from special localized region on someparticular DP [23].The relationship between the morphology of the

DP and the emission current distribution are moreprecisely characterized by results shown in Fig. 4.The SFEM image in Fig. 4 was taken with the tippositioned at 160 nm upward from the surfaceand �10V applied to the sample. The apex of the

0000EF

EVBM

Em

issi

on

inte

nsi

ty [

arb

. Un

its]

Energy below the Fermi level [eV]−20 −15 −10 −5

Fig. 1. Photoemission electron energy distribution curve from a

DP seeded sample. The surfaces of DPs were covered with CVD

diamond over layers.

A. Watanabe et al. / Ultramicroscopy 95 (2003) 145–151 147

fan-shaped emission current distribution is origi-nated from the top (higher) site of the DP andspreads towards the right part of Fig. 4. Theemission current is not detected when the tip isplaced on the left part of Fig. 4. This resultindicated that the electrons emitted from the top

site or the edge of the DP where the electric fieldbetween tip and DP concentrates and the spatialdistribution of emitted electron is affected by theshape of the DP.It turns out that the field emission properties are

different among DPs as already mentioned above.

−2 −1 0 1 2−10

−10

−20−5

0

5

10

15T

un

nel

Cu

rren

t [n

A]

Tu

nn

el C

urr

ent

[nA

]

−4 −2 0 2 4 6

0

10

ghost

0 ~ 1 V

1 ~ 5 V

5 V ~

Bias Voltage [V] Bias Voltage [V]

(c)

(b)(a)

Fig. 2. (a) Typical STS spectra taken on DP. (b) Spectra, which are rarely measured on the DP, indicate a wide gap at EF: (c)Topographic image taken at Vs ¼ þ1:0V, It ¼ 1:0 nA and S ¼ 3:8� 3:8mm2 (right side image). Right image is characterized by theSTS gap width at EF: The region labeled ‘‘ghost’’ has no physical meaning because the scanning tip cannot follow the shape of the DPin this measurement.

1 µm 0.0

0.4

0.8

1.2

µm

0

5

10

15

nA

20

1 µm (a) (b)

Fig. 3. (a) Topographic image taken at Vs ¼ þ1:5V, It ¼ 0:3 nA and S ¼ 4� 4 mm2. (b) SFEM image of same area of (a) taken under

the condition that the tip was moved 1.6mm upward away from the surface. �400V applied to the sample. The electric field strength isestimated to 250V/mm.

A. Watanabe et al. / Ultramicroscopy 95 (2003) 145–151148

The field emission spot, where the measuredemission current is several orders higher than onother DPs, was sometimes observed in the diodetype field emission images and the field emissionelectron microscopy (FEEM) images [23]. We callsuch a DP ‘‘hot spot’’. The field emission spotsobserved in Figs. 3 and 4 are not ‘‘hot spots’’ but‘‘normal spots’’. Results of microscopic fieldemission I–V measurements are shown in Fig. 5.The curve labeled ‘‘Site 1’’ (normal spot) wasmeasured on the emission site shown in Fig. 3. Theothers are I–V characteristics of ‘‘hot spots’’. Themeasured emission current of ‘‘hot spots’’ and‘‘Site 1’’ are 1� 10�7–2� 10�5 and 1� 10�11Aunder the applied electric field of 200V/mm,respectively. Furthermore, the electric fields, whichare required to obtain an emission current of100 nA, are 100–200V/mm on a hot spot and360V/mm on a normal spot, respectively. Here, weestimate the applied electric field with the tip-surface distance and the applied voltage.Field emission spectroscopy (FES) shows

that the peak positions of the total energydistributions shift almost linearly towards lowerenergies with increasing extraction voltage [24,25].This shift measured by FES implies that thereexists a poorly conducting layer in the fieldemission electron path from metal substrate tovacuum.

0 400 800 1200 16000

200

400

[nm]

[nm

]

160 [nm]

(a)

(b)

Fig. 4. (a) SFEM image superposed on the topographic image

of top region of a DP. The SFEM image surrounded by white

line was obtained under the condition that the tip was moved

160nm upward away from the highest site of DP surface

(�10V applied to the sample). The applied electric field is

estimated to 62.5V/mm. The topographic image was taken atVs ¼ þ1:5V, It ¼ 0:3 nA, and S ¼ 4� 4mm2. (b) Profiles

measured along the broken line in image (a). The solid line is

the profile of the topographic image. The dashed line is the

spatial distribution of emitted electron measured from SFEM

image (arbitrary units).

0 100 200 300 4000.0

5.0x10−6

1.0x10−5

1.5x10−5

2.0x10−5

2.5x10−5

Field [V / µm]

Em

issi

on

Cu

rren

t [A

]

Site 1 (1.6µm)Site 2 (2.0µm)Site 3 (2.4µm)Site 4 (1.6µm)Site 5 (2.4µm)

0 2 4 6 8 10 12 14−45

−40

−35

−30

−25

−20

−15

−10

ln(I

/ F

2 ) [a

rb. U

nit

s]

1000 /F [V / µm]−1

Fig. 5. Microscopic field emission characteristics measured on several DPs (I–V curves) and Fowler–Nordheim (F2N) plots. The tip-

surface distances for measurement are given in brackets in each line. The electric field in I–V curve is calculated by using this distance.

Emission sites ‘‘2–5’’ are ‘‘hot spots’’.

A. Watanabe et al. / Ultramicroscopy 95 (2003) 145–151 149

From these findings, a model for field emissionof the present sample is discussed as follows. Aschematic image of a DP placed on the metalsubstrate is shown in Fig. 6(a). This DP is coveredby a p-type surface conductive layer (p-diamond)of the CVD-grown diamond layer as indicated bySTS data (Fig. 2) and also exhibits NEA surfaceproperties. The inner region of the DP maintainsproperties of seeded DP, namely intrinsic non-doped diamond (i-diamond). The peak positionshift measured by FES implies that there is apoorly conducting layer in the field emissionelectron path. The first candidate of this poorlyconducting layer is the interface DP/metal. Thesurface conductive layer does not exist at theDP/metal interface as shown in Fig. 6(a). Whenthe electric field is applied between diode-typeelectrode in vacuum and metal substrate, themaximum electric field strength is observed at themetal/i-diamond/p-diamond interface region. Thishigh electric field results in electron injection fromthe Fermi level of the metal substrate into theconduction band of i-diamond by tunneling (seethe energy diagram in Fig. 6(a)). The above-described mechanism is the model for electroninjection from the metal electrode to the DP.Next, electron emission from the DP to vacuum

is discussed as follows. Electrons are emitted from

the p-diamond surface followed by above-men-tioned injection from metal substrate to i-diamondlayer. It is considerable that there is a few eVenergy barrier at the interface between i-diamondand p-diamond surface layer. This energy barrieraffects the field emission of the electrons ini-diamond to vacuum. SFEM images in Figs. 3and 4 indicate that electrons are mainly emittedfrom localized regions on the top site that isaffected by the shape of the DP. STS results inFig. 2 suggest that the i-diamond regions locallyexist near the top site of the DP. The electrons caneasily be emitted to vacuum from the top site ofthe NEA DP through the local i-diamond region.We also consider a depletion of the p-diamondlayer by the applied field (Fig. 6(b)). The appliedfield causes a depletion region of the p-diamond ontop of the DP. The formation of the depletionregion is presumably affected by the shape of theDP. The gap between valence and conductionband vanishes at the surface of the depletion layerin our present model.

4. Conclusions

Field emission properties and microscopic char-acteristics of DPs have been studied. The CVD

Fig. 6. (a) Schematic drawing of DP placed on the metal substrate and of energy band diagram along line A–B. (b) Sketch of emission

site on DP and of band diagram along line C–D.

A. Watanabe et al. / Ultramicroscopy 95 (2003) 145–151150

diamond layer formed on the DPs presumablyresults in the formation of an NEA surface. Thesamples with such NEA surfaces always showbetter field emission properties than the sampleswith PEA. Most DP surfaces exhibit narrow gaps(approximately 1 eV) near EF: The DP surfaceshave p-type electric characteristics, which agreewith earlier reported surface conductive layer(p-diamond) on CVD diamond surface. The widegap regions (i-diamond) locally exist near the topsite of the DPs.The emission current distribution observed by

SFEM suggests that the field emitted electronscome from the localized region on top of the DP.The emission site is affected by the shape of the DParound the top site. Moreover, it is found thatsome particular DPs mainly contribute to the fieldemission current. The measured emission currentof ‘‘hot spots’’ is several orders higher than that of‘‘normal spots’’.We propose a model for the field emission from

‘‘hot spots’’, which involves two mechanisms. Oneis electron injection from the metal substrate intothe DP, attributed to the electric field enhancementat a poorly conducting layer between DP andmetal. The other is electron emission on the top ofNEA DPs through a local i-diamond region or adepletion region.

Acknowledgements

The authors wish to thank Prof. Nemanich andhis group for the PEEM/FEEM analyses. Thiswork was supported by the FCT Project, whichwas consigned to JFCC by NEDO.

References

[1] F.J. Himpsel, J.A. Knapp, J.A. Van Vechten, D.E.

Eastman, Phys. Rev. B 20 (1979) 624.

[2] B.B. Pate, Surf. Sci. 165 (1986) 83.

[3] M. Deguchi, M. Kitabatake, H. Kurokawa, T. Shiratori,

New Diamond Front. Carbon Technol. 9 (1999) 144.

[4] C. Wang, A. Garcia, D.C. Ingram, M. Lake, M.E.

Koldesch, Electron. Lett. 27 (1991) 1459.

[5] K. Okano, K.K. Gleason, Electron. Lett. 31 (1995) 74.

[6] K. Okano, T. Yamada, H. Ishihara, S. Koizumi, J. Itoh,

Appl. Phys. Lett. 70 (1997) 2201.

[7] T. Sugino, Y. Iwasaki, S. Kawasaki, Y. Yokota, R.

Hattori, J. Shirafuji, Diamond Relat. Mater. 7 (1998) 677.

[8] Y. Show, F. Matsuoka, M. Hayashi, H. Ito, M. Iwase, T.

Izumi, J. Appl. Phys. 84 (1998) 6351.

[9] Z.-H. Huang, P.H. Cutler, N.M. Miskovsky, T.E. Sulli-

van, Appl. Phys. Lett. 65 (1994) 2562.

[10] C. Bandis, B.B. Pate, Appl. Phys. Lett. 69 (1996) 366.

[11] W. Zhu, G.P. Kochanski, S. Jin, L. Seibles, J. Appl. Phys.

78 (1995) 2707.

[12] F. Lacher, C. Wild, D. Behr, P. Koidl, Diamond Relat.

Mater. 6 (1997) 1111.

[13] V.V. Zhirnov, E.I. Givargizov, P.S. Plekhanov, J. Vac. Sci.

Technol. B 13 (1995) 418.

[14] R. Schlesser, M.T. McClure, B.L. McCarson, Z. Sitar, J.

Appl. Phys. 82 (1997) 5763.

[15] W.N. Wang, N.A. Fox, J.W. Steeds, S.R. Lin, J.E. Butler,

J. Appl. Phys. 80 (1996) 6809.

[16] V.D. Frolov, A.V. Karabutov, V.I. Konov, S.M. Pimenov,

A.M. Prokhorov, J. Phys. D 32 (1999) 815.

[17] V.D. Frolov, A.V. Karabutov, S.M. Pimenov, E.D.

Obraztsova, V.I. Konov, Ultramicroscopy 79 (1999) 209.

[18] O. Gr .oning, O.M. K .uttel, P. Gr .oning, L. Schlapbach,

J. Vac. Sci. Technol. B 17 (1999) 1064.

[19] O. Gr .oning, O.M. K .uttel, P. Gr .oning, L. Schlapbach, J.

Vac. Sci. Technol. B 17 (1999) 1970.

[20] Y.D. Kim, W. Choi, H. Wakimoto, S. Usami, H.

Tomokage, T. Ando, Diamond Relat. Mater. 9 (2000)

1096.

[21] K.F. Turner, Y.M. LeGrice, B.R. Stoner, J.T. Glass, R.J.

Nemanich, J. Vac. Sci. Technol. B 9 (1991) 914.

[22] M. Kitabatake, A. Watanabe, M. Deguchi, Procceedings

of the ADC/FCT’99, 1999, p. 376.

[23] A. Watanabe, M. Deguchi, M. Kitabatake, Diamond

Relat. Mater. 10 (2001) 818.

[24] S. Kono, T. Goto, K. Sato, T. Abukawa, M. Kitabatake,

A. Watanabe, M. Deguchi, Surf. Sci. 493 (2001) 610.

[25] S. Kono, T. Goto, K. Sato, T. Abukawa, M. Kitabatake,

A. Watanabe, M. Deguchi, New Diamond and Frontier

Carbon Technol. 11 (2001) 299.

[26] P. Oelhafen, J.L. Freeouf, J. Vac. Sci. Technol.A 1 (1983)

96.

[27] R.J. Nemanich, P.K. Baumann, M.C. Benjamin, S.W.

King, J. van der Weide, R.F. Davis, Diamond Relat.

Mater. 5 (1996) 790.

[28] M. Deguchi, Y. Mori, N. Eimori, Y. Show, M. Kitaba-

take, A. Hatta, T. Ito, A. Hiraki, T. Izumi, T. Hirao,

Proceedings of the Second International Conference

on Application of Diamond Films and Related Materials,

Vol. 91(8). The chemical Society, New Jersey, 1993,

p. 793.

[29] J.B. Cui, J. Ristein, L. Ley, Phys. Rev. B 60 (2000) 16135.

A. Watanabe et al. / Ultramicroscopy 95 (2003) 145–151 151