Nitrogen doped ZnOfilm grown by the plasma-assisted metal-organic chemical vapor deposition pot

A strong emission coming from A-exciton is observed at 10 and 77 K photoluminescence PL spectra in both samples while deep level transition is hardly observed.. More emission peaks are f

Nitrogen doped ZnOfilm grown by the plasma-assisted

metal-organic chemical vapor deposition Xinqiang Wanga,*, Shuren Yanga, Jinzhong Wanga, Mingtao Lia,

Xiuying Jianga, Guotong Dua, Xiang Liub, R.P.H Changb

a Department of Electronic Engineering, State Key Lab on Integrated Optoelectronics, Jilin University, Changchun 130023,

People’s Republic of China

b Materials Research Center, Northwestern University, Evanston, Il, USA Received 28 December 2000; accepted 5 March 2001

Communicated by M Schieber

Abstract

Nitrogen doped and non-doped ZnOfilms are grown by the plasma-assisted metal-organic chemical vapor deposition (MOCVD) on sapphire X-ray diffraction spectra show that they are both strongly c-oriented while the N-doped sample

is of better crystal quality A strong emission coming from A-exciton is observed at 10 and 77 K photoluminescence (PL) spectra in both samples while deep level transition is hardly observed More emission peaks are found in the PL spectrum of the N-doped sample relative to that of a non-doped one Raman scattering is also performed in back scattering configuration E2mode is observed in both samples while A1(LO)mode can only be found in the N-doped sample, which indicates that high quality of the N-doped ZnOfilm A high resistive ZnOfilm is obtained by nitrogen doping # 2001 Published by Elsevier Science B.V

Keywords: A1 Doping; A3 Metalorganic chemical vapor deposition; B2 Semiconducting materials

1 Introduction

ZnO, a wide gap semiconductor with band-gap

of 3.36 eV at room temperature, is attracting more

attention because of its good optical, electrical and

piezo-electrical properties It has potential uses in

optoelectronical systems such as light emitting

diodes (LEDs) [1], photodetectors [2],

electrolumi-nescence devices and solar cells Recently,

opti-cally pumped ultraviolet (UV) lasing of ZnOfilm

by molecular beam epitaxy (MBE) and pulsed

laser deposition has been reported by several authors [3–5] GaN is known as a good material for the fabrication of optical devices such as LEDs and laser diodes (LDs) ZnOhas not only the same crystal structure as GaN, but also a larger exciton binding energy of 60 meV, which is 2.4 times that

of GaN Furthermore, Yu et al have shown that textured ZnOfilms might have higher quantum efficiency than GaN [6] This indicates that ZnO should be the most potential material to realize the next generation UV semiconductor laser ZnOis also a promising material for surface acoustic wave (SAW) devices, which is very important in this information age However, high resistive or

*Corresponding author Tel +86-431-8922331.

E-mail address: ysr@mail.jlu.edu.cn (X Wang).

0022-0248/01/$ - see front matter # 2001 Published by Elsevier Science B.V.

PII: S 0 0 2 2 - 0 2 4 8 ( 0 1 ) 0 1 3 6 7 - 7

p-type ZnOfilm is needed to realize the devices

mentioned above Nitrogen doping, which has

been successful in fabricating p-type ZnSe [7], is

considered as an effective method to realize high

resistive or p-type ZnOfilm In this paper, a

non-doped ZnOfilm and a nitrogen non-doped one are

deposited by plasma-assisted metal-organic

che-mical vapor deposition (MOCVD) We have

investigated their structural and optical qualities

and found that the nitrogen doped (N-doped) ZnO

film shows better crystal quality and higher

resistivity relative to the non-doped one

2 Experiment

ZnOfilm was grown by MOCVD on (0 0 0 6)

sapphire substrate Di-ethyl zinc (DEZn) and O2

were used as sources High purity Ar was passed

through the DEZn bubbler and saturated with

DEZn vapor to the reactor N2 was used as the

carrier gas The substrate was cleaned in acetone,

methanol, deionized water, and then chemical

etched in H2SO4: H3PO4=3 : 1 for 10 min at

1608C followed by deionized water rinse The

reaction temperature was 6008C and the power of

the plasma source was 900 W The growth rate was

1 mm/h The non-doped sample (sample A) was

grown with O2 flow of 10 SCCM while the

N-doped one (sample B) was deposited with the

flow ratio of O2: N2=1 : 1 The N-doped sample

looked slight red while the non-doped one was transparent

We used SIEMENS D 5005 X-ray diffract-ometer and Rigaku DMAX 2400 X-ray diffraction (XRD) to investigate crystal quality The absorp-tion measurement was performed by UV-3100 SHIMADZU UV-VIS-NIR Recording spectro-photometer with Xe lamp as optical source Raman measurement was taken by RE-NISHAW-Ramascope at room temperature in back-scattering configuration by using 514.5 nm

Ar+laser line excitation with an arriving power of about 70 mW The scattered light was detected by

a water-cooled charge coupled device (CCD) detector The diameter of the laser beam was about 1 mm PL spectrum was measured by a

325 nm He–Cd laser The PL signal from the sample was filtered by a monochromator and picked up by a CCD detector The power arriving

at the sample was about 3 mW with a beam diameter of 200 mm For low temperature mea-surement, the sample was mounted on a closed-cycle refrigerator

3 Results and discussion XRD y22y scan spectra of N-doped (sample B) and non-doped ZnOfilm (sample A) are shown in Fig 1(a) and (b), respectively In both the two spectra, we find a dominant peak at around 2y of

Fig 1 X-ray diffraction spectra of ZnOfilm The o-rocking curve is shown in the inset (a) Non-doped ZnOfilm; (b) N-doped ZnOfilm.

34.68 due to (0 0 0 2) ZnO This shows that the

films are both strongly c-oriented The full-width

at half-maximum (FWHM) of (0 0 0 2) peak of

sample B is 0.1488, which is narrower than that of

sample A, 0.1978 The narrower FWHM implies

that the N-doped film is of better crystal quality

From the statistical result, we inferred that the

length of c-axis was 5.166 (A in sample A while it is

5.181 (A in sample B They are both slightly smaller

than that of bulk ZnOwhose c-axis length is

5.2071 (A, 2y=34.428 This can be ascribed to

tensile stress induced by the deposition process

The o-rocking curves are shown in the inset of

Fig 1 It shows that ZnOgrows in single c-axis

orientation with the c-axis normal to the sapphire

basal plane, indicating a heteroepitaxial

relation-ship of (0 0 0 1)ZnOk(0 0 0 1)sapphire The FWHM

values of o-rocking curve of samples A and B are

0.568 and 0.348, which indicates that the N-doped

sample has smaller mosaicity

Room temperature photoluminescence (RT-PL)

spectra are performed as shown in Fig 2

Ultra-violet (UV) emission, with peak energy positions

of 3.30 and 3.289 eV, is dominantly observed in

samples A and B respectively The FWHM was

87 meV for sample A, which is narrower than the

value 97 meV of sample B The FWHM values of

both samples A and B are higher than that of ZnO

films reported by others using MBE and MOCVD

[8,9] A deep level emission at around 2.513 eV can

be weakly observed in sample A Its enlarged figure is shown inset This deep visible transition is believed to come from oxygen vacancies, inter-stitial zinc or zinc vacancies [10–12] In sample B, the peak position of deep level transition is weakly observed around 2.229 eV In comparison with a previous study [13], the intensity of the deep level transition is much lower The deep level transition shifts to the lower energy side relative to that of sample A, which may be related to nitrogen doping The ratios of the intensity of UV emission (IUV) to that of deep level emission (IDLE) are 193 (sample A) and 136 (sample B), at room tempera-ture, respectively The values are both rather high

in comparison with another reported ratio value of

60 observed in ZnOfilm deposited by molecular beam epitaxy (MBE) [14] and of 1 by MOCVD [15] This high ratio implies that our sample is of high optical quality The relatively small IUV=IDLE

value of sample B may be due to a slight increase

in intrinsic defects in ZnOdue to nitrogen doping

PL spectra under different excitation powers were performed We do not find any shift of the position

of the UV emission peak The integrated PL intensity as a function of the excitation power on the logarithmic scale is plotted as shown in Fig 3 From the figure, it is obvious that the solid line fits well the data shown by the black square in both samples The PL intensity is linearly dependent on the excitation power This indicates that the

Fig 2 Room temperature PL spectrum of ZnOfilm The enlarged deep level transition is shown in the inset (a) Non-doped ZnOfilm; (b) N-doped ZnOfilm.

dominant photoluminescence of our sample

should be the excitonic radiative recombination

at room temperature

Low temperature measurement is performed at

10 and 77 K for further study of the optical

properties Fig 4(a) and (b) correspond to samples

A and B, respectively As shown in Fig 4(a), four

peaks appear at 3.377, 3.370, 3.333 and 3.241 eV,

respectively The dominant peak at 3.377 eV is

ascribed to the A-exciton emission while the peak

at 3.370 eV corresponds to D8X bound exciton

transition The peaks at 3.333 and 3.241 eV are due

to the donor–acceptor pair transition and LO

phonon replica, respectively In the PL spectrum

of 77 K shown in the inset of Fig 4, the peak due

to D8X bound exciton transition is hardly observed The energy positions of other peaks are at 3.373, 3.314 and 3.238 eV, respectively They all shift to the low energy side compared to that at

10 K due to thermal effect From Fig 4(b), we can find two dominant peaks at 3.386 and 3.372 eV and four other peaks with low intensity at 3.324, 3.248, 3.196 and 3.125 eV, respectively The peak

at 3.386 eV is ascribed to A-exciton emission while the peak at 3.372 eV is due to D8X bound exciton transition In comparison with that of sample A, the two peaks shift to high energy level The peak

at 3.324 eV corresponds to donor–acceptor

transi-Fig 3 The room temperature dependence of integrated output intensity on excitation intensity on the logarithmic scale (a) Non-doped ZnOfilm; (b) N-Non-doped ZnOfilm.

Fig 4 Low temperature PL spectra of ZnOfilm (a) Non-doped ZnOfilm; (b) N-doped ZnOfilm.

tion while the peaks at 3.248, 3.196 and 3.125 eV

are due to the LOphonon replicas, respectively

From the 77 K PL spectrum of sample B as shown

in the inset, we find that all the peak positions shift

to the low energy side Six peaks can also be found

at 3.372, 3.360, 3.3137, 3.239, 3.189 and 3.106 eV

In comparison with PL spectrum of sample A, we

find that more peaks are observed and the

dominant peak shifts to the higher energy side

The possible reason is the doping of nitrogen We

find that sample B shows better optical quality

than sample A at low temperature By the way,

deep level emission is hardly observed in the PL

spectra of both samples at low temperature

Raman scattering is performed on both samples

at room temperature in back-scattering

configura-tion as shown in Fig 5(a) and (b) ZnOhas a

hexagonal wurtzite structure and belongs to the

C6n symmetry group In our back-scattering configuration, A1(LO) and E2 are Raman active The peaks at 437.6 cm1 in Fig 5(a) and 437.9 cm1 in Fig 5(b) are ascribed to high frequency E2 mode Since the measure range of our micro-Raman system was from 100 to

4000 cm1, we cannot observe the low frequency

E2mode Raman spectrum of ZnOpowder is also performed as shown in Fig 5(c), in which we observe the peak position of E2 modes lies at 437.4 cm1 The slight discrepancy of the position

of the E2 mode of ZnOfilms and powder shows that our samples are almost free of stress We find the second order Raman spectrum arising from zone-boundary phonons 2-E2(M) at 338 cm1[16]

As shown in Fig 5(a), the peaks at 380, 417.4, 447.7, 575.9 and 749.4 cm1 are ascribed to sapphire substrate Since our ZnOfilm is relatively

Fig 5 Raman spectra of ZnOfilms and ZnOpowder (a) Raman spectrum of non-doped ZnOfilm; (b) Raman spectrum of N-doped ZnOfilm; (c) Raman spectrum of ZnOpowder.

thin and non-doped ZnOand sapphire are both

nearly transparent to visible laser light as shown in

Fig 6, we can clearly observe the Raman peaks of

sapphire In Fig 5(b), the second order Raman

spectrum arising from zone-boundary phonons

2-E2(M) is also observed at 331 cm1 The Raman

peaks from sapphire have a low intensity This is

due to the higher absorption ratio of nitrogen

doped ZnOfilm relative to non-doped one as

shown in Fig 6 We can find a dominant peak at

581 cm1 in Fig 5(b) which is ascribed to A1(LO)

mode We did not find this A1(LO)mode clearly in

non-doped ZnOfilm In the Raman study of GaN,

the intensity of A1(LO) mode increased with the

decrease of carrier concentration [17] We think

that ZnOmay have the same characteristics as

GaN Since the carrier concentration of nitrogen

doped ZnOis more than 1000 times smaller than

that of the non-doped one, which is about

4.0  1017cm3, we expect that the intensity of

A1(LO)should be stronger This may be the reason

why we find A1(LO) mode in N-doped sample but

not in the non-doped sample Both the higher

intensity of A1(LO)mode and the observation of E2

mode and A1(LO) mode coinciding with the

prediction of group theory imply the better crystal

quality of nitrogen doped ZnOfilm In the Raman

spectrum of nitrogen doped ZnOfilm, other peaks

at 274, 508, 641.9, and 857 cm1can be found The

possible reason may be related to nitrogen doping

and further study about this should be going on

The resistivity of non-doped ZnOfilm by four-point probe measurements is 0.65 O cm Due to the high resistivity and low carrier concentration of N-doped film, four-point probe measurements are not reliable However, ohmic contacts are formed

by Al metal on high resistive ZnOfilm, which indicates that the N-doped ZnOfilm is still n-type The resistivity of N-doped ZnOfilm is estimated to

be 5  104O cm by I–V measurement

4 Conclusions High quality N-doped and non-doped ZnO films are successfully deposited XRD spectra show that ZnOfilms are both c-oriented while the N-doped films show better crystal quality with smaller mosaicity relative to non-doped samples

UV emission and deep level transition are ob-served from PL spectrum at room temperature The high value of IUV=IDLE indicates high optical quality We find A band free exciton emission and D8X bound exciton transition from PL spectrum

at 10 K in both samples In Raman spectrum of ZnOfilm, E2and A1(LO)modes are observed in N-doped sample while A1(LO) is hardly observed at non-doped one In comparison with Raman spectrum of ZnOpowder, we find that the peak positions of E2 mode are almost the same indicating that ZnOfilms are almost free of stress High resistive but not p-type ZnOfilm is realized

by nitrogen doping

Acknowledgements This work was supported by NSFC-RGC (No 59910161983) and Jilin Province Science Fund (No 19990518-1)

References [1] Toru Soki, Yoshinori Hatanaka, David C Look, Appl Phys Lett 76 (2) (2000) 3257.

[2] Y Liu, C.R Gorla, S Liang et al., J Electron Mater 29 (1) (2000) 60.

[3] H Cao, Y.G Zhao, H.C Ong et al., Appl Phys Lett 73 (25) (1998) 3656.

Fig 6 Absorption spectrum of ZnOfilm (Solid

line}non-doped ZnOfilm; Dash line}N-line}non-doped ZnOfilm.)

[4] Z.K Tang, G.K.L Wong, P Yu et al., Appl Phys Lett.

72 (25) (1998) 3270.

[5] D.M Bagnall, Y.F Chen, Z Zhu et al., Appl Phys Lett.

70 (17) (1997) 2230.

[6] P Yu, Z.K Tang, G.K.L Wong, M Kawaski, A Ohtono,

H Koinuma, in: M Scheffler, Zimmermann (Eds.),

Physics of Semiconductors, World Scientific, Singapore,

1996, p 1453.

[7] K Ohkawa, A Ueno, O Mitsuyu, J Crystal Growth 117

(1992) 375.

[8] C.R Gorla, N.W Emanetoglu, S Liang et al., J Appl.

Phys 85 (5) (1999) 2595.

[9] Yefan Chen, D.M Bagnall, Hang-jun Koh, et al., J Appl.

Phys 84(7) (1998) 3912.

[10] D.C Reynolds, D.C Look, B Jogni, Solid State Com-mun 101 (1997) 643.

[11] K Vanheusden, W.L Warren, C.H Seager, J Appl Phys.

79 (1996) 7983.

[12] E.G Bylander, J Appl Phys 49 (3) (1978) 1188 [13] K Iwata, P Fons, A Yamada, K Matsubara, S Niki,

J Crystal Growth 209 (2000) 526.

[14] H Kumano, A.A Ashrafi, A Heta et al., J Crystal Growth 214–215 (2000) 280.

[15] S Bethke, K Pan, B.W Wesseis, Appl Phys Lett 52 (1988) 138.

[16] J.M Calleja, M Cardona, Phys Rev B 16 (1977) 3753 [17] N Wieser, M Klose, T Dassow, F Scholz, J Off, J Crystal Growth 189/190 (1998) 661.