Journal of Luminescence 174 (2016) 6–10

Contents lists available at ScienceDirect

Journal of Luminescence
journal homepage: www.elsevier.com/locate/jlumin

Full Length Article

Effects of carbon on optical properties of ZnO powder
Nguyen Tu, K.T. Nguyen n, D.Q. Trung, N.T. Tuan, V. Nam Do, P.T. Huy n
Nano Optoelectronic Laboratory (La Nopel), Advanced Institute for Science and Technology (AIST), Hanoi University of Science and Technology (HUST),
01 Dai Co Viet Street, Hanoi 10000, Vietnam

art ic l e i nf o

a b s t r a c t

Article history:
Received 18 March 2015
Received in revised form
21 January 2016
Accepted 25 January 2016
Available online 29 January 2016

We report on C-doped ZnO, with different weight percentages of dopant, prepared by a high-energy ball
milling method. The annealing conditions with temperature of 800 °C and in argon environment appear
to be the optimal conditions for producing good quality crystals as well as pure UV emission. XRD and
FTIR analysis indicate the substitution of C for Zn. In addition, Raman spectroscopy suggests a disordered
graphitic layer covering the crystals. Photoluminescence investigation reveals the continuous quenching
of visible region upon increasing C concentration and the intensity ratio between defect-related and UV

emission can be as negligible as 0.02. The passivation of surface defects and the creation of a nonradiative recombination pathway by carbon integration are proposed as possible origins of the suppression.
& 2016 Elsevier B.V. All rights reserved.

Keywords:
Pure UV emitter
ZnO crystals
Carbon doping
High-energy ball milling
Photoluminescence quenching

1. Introduction
Emission properties of ZnO have been at the center of ZnO
research, besides synthesis methods and growth kinetics [1–3],
due to its potential applications in optoelectronic devices, particularly ultraviolet-emitting devices. Emission properties of ZnO
have been at the center of ZnO research due to its potential
applications in optoelectronic devices, particularly ultravioletemitting devices [4]. However, besides UV emission, photoluminescence spectra of ZnO typically contain other bands in the
visible region which are attributed to defect-related emission [5]
even though the exact type of defect is not conclusively established. The most commonly cited hypothesis for the green emission is the electronic transition from conduction band edge to
oxygen vacancies (VO) or zinc vacancies (VZn) [6–9]. The yellow
luminescence band is attributed to excess oxygen [10], lithium
impurities [11], or hydroxyl group [12]. The orange/red emission is
due to electronic transitions from zinc interstitials (Zni) to oxygen
interstitials (Oi) [13–16]. Nonetheless, surface defects [17] and
defect complexes [18, 19] are also considered to explain the origin
of visible emissions.
Particularly, carbon doping in ZnO has been extensively studied,
apart from a large number of reports on p-type conduction or ferromagnetism properties, the luminescence properties of carbondoped ZnO has also aroused much scientific interest [18–21].
n

Corresponding authors. Tel.: þ 84 436230435; fax: þ84 436230293.

E-mail addresses: (K.T. Nguyen),
(P.T. Huy).
/>0022-2313/& 2016 Elsevier B.V. All rights reserved.

Despite of the fact that up to now not many studies were focused on
the optical properties of carbon-doped ZnO, the reported data on
luminescence of this material are rather complicated and far from
unambiguous. It was reported that blue emission was observed from
the carbon modified ZnO particles [22]. Carbon impurities were also
found to enhance the blue/violet emission band from the ZnO thin
film fabricated by pulse laser deposition [23]. Green luminescence in
carbon-doped ZnO films or in ZnO powder annealed with carbon
black was also reported [19–20]. Additionally, a broad emission band
that covers the whole visible range from 400 to 800 nm was
observed from a hollow microtube–nanowire structure made of
carbon-doped ZnO [24]. Moreover, the possibility to tune the optical
bandgap of ZnO was also demonstrated for the C-implanted ZnO thin
films [21].
In this work, we report on the suppression of the defect-related
emission by C doping in ZnO crystals. XRD measurement shows a
single phase hexagonal wurtzite structure of C-doped ZnO without
any detectable secondary phase. The introduction of C by a ballmilling method causes the blueshift of UV band and helps to
suppress visible emission of ZnO crystals. The intensity ratio
between UV and visible emission gets its maximum value with the
annealing temperature of 800 °C. Adding carbon might help to
reduce the number of defect centers, passivate surface defects in
ZnO crystals, or might create non-radiative recombination pathways that can compete with decay channels resulting in visible
photoluminescence. Above this optimal temperature, carbothermic reaction happens and creates more defects in the crystal,
bringing the ratio back down.



N. Tu et al. / Journal of Luminescence 174 (2016) 6–10

2. Experimental
Commercial ZnO (Merck, 99.99%) and Carbon (99.9%) powders
were mixed together with different weight ratios (2, 3, and 4%).
After coarse grinding for 1 h, the mixture was grounded further by
high-energy planetary ball milling (Restch PM 400) with the speed
of 200 rpm for 60 h, and then annealed in Argon gas for 2 h and at
different temperatures from 200 to 1000 °C. The crystal structure
of the powders was investigated by X-ray diffraction (Bruker D8
Advance XRD). The surface morphology was characterized by
ultra-high resolution scanning electron microscopy (Jeol JSM7600F). Chemical bonds and vibrational frequencies were analyzed
by Fourier transform infrared (Nicole FTIR 6700) and Raman
(Horiba Jobin-Yvon LabRAM HR Raman) techniques. Emission
spectrum of the samples was collected by using photoluminescence spectroscopy (Horiba Jobin-Yvon Nanolog). All
measurements were carried out at room temperature. The chemical bonding structure of obtained catalysts was demonstrated
by X-ray photoelectron spectroscopy (XPS, VG Scientific, ESCALAB250 at 15 kV, 15 mA). Measurements were performed using a
monochromatic Al Ka X-rays (1486 eV) source.

3. Results and discussion
Fig. 1 shows FESEM images of the powders before and after
milling with 4% C and post annealing at different temperatures.
The source ZnO material is in the form of rods and particles with

7

diameters from 0.1 to around 1 μm. After grinding by high-energy
planetary ball milling for 60 h and subsequent post annealing,
small particles aggregate to form big particles. Higher temperature

leads to the formation of bigger particles, which are in order of
micrometer size (i. e. after 2 h annealing at 1000 °C in pure Ar gas
environment).
Fig. 2 shows X-ray diffraction patterns of ZnO and C-doped ZnO
powders. The patterns of the initial ZnO and ZnO doped with 4% C
at various temperatures are shown in Fig. 2a. The diffraction peaks
are broad at temperatures lower than 800 °C, indicating low
crystallinity. This is held true for other C concentrations as well.
The XRD patterns of ZnO with different levels of C doping and
annealed at the same temperature of 800 °C are presented in
Fig. 2b. All the diffraction peaks can be indexed as hexagonal
wurtzite structure. There is no measurable secondary phase or
impurity peak, which indicates that C dopant is well integrated
into the host. The main peaks ((100), (002), (101)) are shown in the
inset of Fig. 2b. The peaks shift to higher diffraction angles with
increasing C content, except 4% C, in comparison with those of ZnO
source. The lattice constants are calculated and tabulated in
Table 1. They are smaller in C-doped ZnO compared to ZnO source.
Carbon doping in ZnO can happen by the substitution of C for
either O or Zn or by taking the interstitial sites of C [25]. The
decreasing of lattice parameters when C is integrated suggests that
the smaller C4 þ (0.15 Å) ions substitute for the larger Zn2 þ (0.74 Å)
ions in the ZnO lattice. At 4% C doping concentration, the diffraction peaks, however, do not continue to shift to higher angles but
the opposite side, suggesting that C dopant takes interstitial sites

Fig. 1. FESEM images of original ZnO (a) and ZnO: 4% C after high-energy planetary ball milling for 60 h and post annealing at 600 °C (b), 800 °C (c), and 1000 °C (d) for 2 h in
pure Argon gas environment.


8


N. Tu et al. / Journal of Luminescence 174 (2016) 6–10

besides Zn replacement, leading to a slight expansion of the host
lattice compared to 3% C doping.
The substitution of C for Zn is also suggested by FTIR spectra
shown in Fig. 3. Both the undoped and doped ZnO samples exhibit
three absorption bands at 3446 cm À 1, 2361 cm-1, 1637 cm À 1
which are attributed to O–H stretching vibration of H2O, the
O¼ C¼ O bending vibration of CO2 in air, and the C ¼O bond
vibrations [26–28], respectively. The band at 430 cm À 1 is assigned
to Zn–O bond in undoped ZnO [29, 30] and it shifts to higher
frequencies at 440 cm À 1, 454 cm À 1, and 480 cm À 1, respectively
for 2, 3, and 4% C-doped ZnO. Carbon atom is lighter than zinc
atom, so the replacement could result in the stretching frequency
upshift.
In order to confirm that the incorporation of C is the culprit of
the shift, we have repeated the same milling and annealing processes but on original ZnO only and its XRD patterns at different
states are shown in Fig. 2c. The crystallinity is drastically improved
at 800 °C which is indicated by sharp diffraction peaks. However,
no significant peak shift has been observed, differentiating Cdoped ZnO from original ZnO. The integration of C results in lattice
contraction.
The ZnO:C system is further investigated by Raman spectroscopy and the results are shown in Fig. 4. The spectra show two
new peaks at 1324 and 1594 cm À 1 in comparison to source ZnO.
The Raman mode at 1594 cm À 1 is a characteristic of graphitic

Fig. 3. FTIR spectra of ZnO source material and ZnO:C (2, 3, and 4%C) annealed at
800 °C. The peak corresponding to ZnO bond shifts to higher wavenumber with
increasing C concentration.


Fig. 2. XRD patterns of ZnO source material and ZnO: 4% C at different annealing
temperatures (a), ZnO:C (2, 3, 4%) at the same annealing temperature of 800 °C. The
inset shows XRD patterns of the three peaks ((100), (002), and (101)), which shift to
the higher angles with C doping.

Table 1
Lattice constant a and c of ZnO and C-doped ZnO powders with different C doping
concentrations.
The lattice constant a and c

a (Å)

c (Å)

ZnO
2%C
3%C
4%C

3,279
3,268
3,262
3,265

5,249
5,232
5,224
5,230

Fig. 4. Raman spectra of C-doped ZnO exhibit characteristic peaks of graphitic

material.


N. Tu et al. / Journal of Luminescence 174 (2016) 6–10

9

Fig. 6. Photoluminescence spectra of ZnO source material and ZnO: 4%C annealed
at various temperatures (a). The inset shows the dependence of the intensity ratio
between UV and visible regions on annealing temperature. The evolution of PL
spectra upon C integration at the annealing temperature of 800 °C (b). The inset
presents the shift of the UV peak toward lower wavelength with increasing C
doping.

Fig. 5. XPS spectra of 4% C-doped ZnO powders annealed at 800 °C in argon: Zn 2p
core level spectrum (a). C 1s core level spectrum (b). O 1s core level (c).

materials and the mode at 1324 cm À 1 is attributed to defects in
graphitic structure. Since intensities of the two bands are comparable, ZnO crystals appear to be covered by a disordered graphitic layer. This layer may play a role in changing the emission
spectrum of ZnO crystals as presented in Fig. 6. The formation of
the graphitic shell around the ZnO core indicates that the actual
dopant concentration is less than the initial doping concentration,
i.e. smaller than 2%, 3% or 4% for corresponding experimental
conditions.

In order examine the stoichiometry and the chemical bonding
in C-doped ZnO powders, XPS measurements were carried out for
the powder doped with 4% C annealed at 800 °C. XPS data for the C
1s, O 1s and Zn 2p binding energy regions is shown in Fig. 5. The
Zn 2p spectrum in Fig. 5a shows a doublet with the binding

energies at 1020.4 eV and 1043.5 eV which correspond to the
reported binding energies of the Zn 2p3/2 and Zn 2p1/2 states,
respectively. The difference between the binding energy of Zn 2p3/
2 and Zn 2p1/2 states is found to be 23.1 eV and is in accordance
with the standard reference value [21]. Both the binding energies
and their difference value infer that Zn ions are in þ2 oxidation
state. The XPS spectrum for the C 1s is shown in Fig. 5(b). Fitting
the spectrum with Gaussian-Lorentzian functions, it can be deconvoluted into three peaks centered at binding energies of 284.7,
285.9 and 288.9 eV. The intense peak at binding energy of
284.7 eV could be attributed to C–C bonds of the “free carbon”
(graphite or carbon contamination). This is in good agreement
with Raman spectra and confirms the presence of graphitic shell.
The peak at 288.9 eV is assigned to carbonate species [31]. It has
been reported that if C substitutes for Zn then the C 1s binding
energy shifts to a slightly higher value [32–33]. Thus, the
appearance of the peak centered at 285.9 eV indicates the successful doping of carbon into ZnO crystal by substituting for zinc,
in good agreement with the XRD and FTIR results as shown Fig. 2
and Fig.3. One of the possibilities of the substitution is the formation of O–C–O bond. The O 1s XPS spectrum as shown in Fig. 5
(c) can be de-convoluted into two peaks at 529.4 eV and 530.8 eV.
The peak at 529.4 eV is assigned to O2 À ions of Zn–O bonds in


10

N. Tu et al. / Journal of Luminescence 174 (2016) 6–10

Wurtzite structure with Zn2 þ in hexagonal coordination and the
peak at 530.8 eV can be attributed to Zn–Vo and Zn–O–C bonds
because C-dopant and Vo usually induce blue shift of Zn–O bond
[31,33]. The above results and the disappearance of the oxygen

vacancies peak at higher binding energy suggest that C atoms
incorporated in ZnO crystals may work to passivate defects as such
oxygen vacancies in the host, leading to the reduction of defectrelated emission.
Fig. 6 shows photoluminescence spectra of undoped and Cdoped ZnO samples annealed in argon at different temperatures
from 200 to 1000 °C. The PL spectrum of the starting material
(Fig. 6a, black curve) consists of two emission features: a sharp
peak at around 383 nm and a wide band in visible region that are
attributed to near band edge (NBE) and defect-related emissions,
respectively [5]. Upon C integration, the evolution of PL spectra at
different annealing temperatures is shown in Fig. 6a. The intensity
ratio between the UV and visible regions is an important quantity
to characterize the emission spectrum purity. For all C concentrations, the intensity ratio keeps increasing until 800 °C
(Fig. 6a inset) where the visible region is remarkably reduced. In
addition, the ratio at 800 °C is proportional with carbon concentration. The quenching upon C doping at 800 °C is shown in
Fig. 6b. This indicates that the defect level is suppressed with C
addition. By both diffusing in and forming a graphitic cover of ZnO
crystals, C integration might help to reduce the number of defect
centers, passivate surface defects [34], or provide a non-radiative
recombination channel that is competitive to decay channels
resulting in visible emission, leading to the defect-related PL
quenching. When the annealing temperature increased to 1000 °C,
the ratio is reduced and the visible emission reappears again. The
enhancement is probably due to carbothermic reaction in ZnO–C
system [35], that creates more defects (Zn and O vacancies) in the
host crystals. The temperature of 800 °C appears to be the optimal
annealing temperature at which the system shows good crystallinity (from XRD and SEM measurements) as well as the suppression of the defect-related emission.
In addition to the quenching of visible emission, the NBE peak
shifts to lower wavelength upon the increasing of C doping level
and annealing at 800 °C as shown in Fig. 6b inset. The peak position is observed at 383, 382, 380, and 379 nm for ZnO source and
ZnO doped with 2, 3, and 4% C, respectively. Since the particle size

of the samples is much larger than the exciton Bohr radius of ZnO,
which is around 10 nm [36,37], the quantum confinement cannot
explain the blue shift. When ZnO is doped with carbon, the dopant
and their complexes can act as donors, providing excess carriers
which occupy impurity energy levels close to the conduction band
edge. When the concentration of excess electrons is high enough,
they start to occupy states above the conduction band edge,
shifting up the lowest available energy level for electrons in the
valence band to be excited to. Therefore, C dopant can cause
degenerate doping and make the bandgap appear to be larger,
known as Burstein–Moss effect [38], resulting in the blue shift of
UV peak as indicated in the inset of Fig. 6b. The higher the dopant
level is, the lower wavelength the UV peak position can be.

4. Conclusions
In conclusion, C-doped ZnO powder with good quality single
crystals has been prepared successfully by high-energy ball milling
technique and with appropriate annealing conditions, i. e. at
800 °C and in argon environment. XRD and FTIR measurements
indicate the substitution of C for Zn. Nonetheless, Raman analysis
suggests a graphitic layer covering ZnO crystals. The incorporation
of C can happen in both ways, diffusing in and forming a graphitic

structure on the surface. The integration of C shows strong effects
on PL spectrum of ZnO powder. Besides the slight blueshift of UV
band, the defect-related emission is drastically quenched and the
quenching is proportional with dopant level, making C doping an
effective method to produce almost pure UV emission from ZnO
crystals. The C incorporation might help to decrease defect concentration, passivate the surface, and/or provide a non-radiative
recombination pathway, resulting in the suppression of

visible band.

Acknowledgments
This work was supported by the National Application-Oriented
Basic Research Program under Project DTDL.05-2011-NCCB. The
authors thank Prof. Nagahiro Saito (Nagoya University) and the
JSPS Core-to-Core program “Establishment of Educational Hub on
Biomas-based Materials Research for Green Mobilty” for supporting the XPS measurements.

References
[1] K. Biswas, N. Varghese, C.N.R. Rao, J. Mater. Sci. Technol. 24 (4) (2008) 615.
[2] K. Biswas, B. Das, C.N.R. Rao, J. Phys. Chem. C 112 (2008) 2404.
[3] C.N.R. Rao, S.R.C. Vivekchand, K. Biswas, A. Govindaraj, Dalton Trans. (2007)
3728.
[4] I.S. Jeong, J.H. Kim, S. Im, Appl. Phys. Lett. 83 (2003) 2946.
[5] A.B. Djurisic, Y.H. Leung, Small 2 (2006) 944.
[6] Q.X. Zhao, P. Klason, M. Willander, H.M. Zhong, W. Lu, J.H. Yang, Appl. Phys.
Lett. 87 (2005) 211912.
[7] K. Vanheusden, C.H. Seager, W.L. Warren, D.R. Tallant, J.A. Voigt, Appl. Phys.
Lett. 68 (1996) 403.
[8] F.K. Shan, G.X. Liu, W.J. Lee, B.C. Shin, J. Appl. Phys. 101 (2007) 101.
[9] P.F. Carcia, R.S. McLean, M.H. Reilly, J.G. Nunes, Appl. Phys. Lett. 82 (2003) 1117.
[10] J. Qiu, X. Li, W. He, S.J. Park, H.K. Kim, Y.H. Hwang, J.H. Lee, Y.D. Kim, Nanotechnology 20 (2009) 155603.
[11] D. Li, Y.H. Leung, A.B. Djurisic, Z.T. Liu, M.H. Xie, S.L. Shi, S.J. Xu, W.K. Chan,
Appl. Phys. Lett. 85 (2004) 1601.
[12] K.H. Tam, C.K. Cheung, Y.H. Leung, A.B. Djurisic, C.C. Ling, C.D. Beling, S. Fung,
W.M. Kwok, W.K. Chan, D.L. Philips, L. Ding, W.K. Ge, J. Phys. Chem. B 110
(2006) 20865.
[13] C.H. Ahn, Y.Y. Kim, D.C. Kim, S.K. Mohanta, H.K. Cho, J. Appl. Phys. 105 (2009)
089901.

[14] M. Gomi, N. Oohira, K. Ozaki, M. Koyano, Jpn. J. Appl. Phys. 42 (2003) 481.
[15] N.H. Alvi, K.U. Hasan, O. Nur, M. Willander, Nanoscale Res. Lett. 6 (2011) 130.
[16] N. Tu, N.T. Tuan, N.V. Dung, N.D. Cuong, N.D.T. Kien, P.T. Huy, N.V. Hieu, D.
H. Nguyen, J. Lumin. 156 (2014) 199–204.
[17] I. Shalish, H. Temkin, V. Narayanamurti, Phys. Rev. B 69 (2004) 24540.
[18] Y.H. Lu, Z.X. Hong, Y.P. Feng, S.P. Russo, Appl. Phys. Lett. 96 (2010) 091914.
[19] L.T. Tseng, J.B. Yi, X.Y. Zhang, G.Z. Xing, H.M. Fan, T.S. Herng, X. Luo, M. Ionescu,
J. Ding, S. Li, AIP Adv. 4 (2014) 067117.
[20] J. Hu, H.J. Chen, Appl. Phys. Lett. 101 (2007) 124902.
[21] P. Kumar, H.K. Malik, K. Asokan, EPL 110 (2015) 67006.
[22] Z.G. Wang, X.T. Zu, S.Z. Yang, L.M. Wang, J. Mater. Sci. 41 (2006) 3729.
[23] R.J. Mendelsberg, J. Kenedy, S.M. Durbin, R.J. Reeves, Curr. Appl. Phys. 8 (2008)
283.
[24] X. Liu, H. Du, X.W. Sun, B. Liu, D. Zhao, H. Sun, Cryst. Eng. Commun. 14 (2012)
2886.
[25] S. Sakong, P. Kratzer, Semicond. Sci. Technol. 26 (2011) 14038.
[26] S. Muthukumaran, R. Gopalakrishnan, Opt. Matls. 34 (2012) 1946.
[27] A. Sharma, B.P. Singh, S. Dhar, A. Gondorf, M. Spasova, Surf. Sci. 606 (2012) L13.
[28] D. Ramimoghadam, M.Z.B. Hussein, Y.H. Taufiq-Yap, Chem. Cent. J. 7 (2013) 71.
[29] B. Zhou, Y. Wu, L. Wu, K. Zou, H. Gai, Physica E 41 (2009) 705.
[30] Z.R. Khan, M.S. Khan, M. Zulfequar, M.S. Khan, Mater. Sci. Appl 2 (2011) 340.
[31] H. Pan, J.B. Yi, L. Shen, R.Q. Wu, J.H. Yang, J.Y. Lin, Y.P. Feng, J. Ding, L.H. Van, J.
H. Yin, Phys. Rev. Lett. 99 (2007) 127201.
[32] S. Kaciulis, Surf. Interface Anal. 44 (2012) 1155.
[33] S. Akbar, S.K. Hasanain, M. Abbas, S. Ozcan, B. Ali, S.I. Shah, Solid State Commun. 151 (2011) 17.
[34] R.R. Devarapalli, J. Debgupta, V.K. Pillai, M.V. Shelke, Sci. Rep. 4 (2014) 04897.
[35] B.S. Kim, J.M. Yoo, J.T. Park, J.C. Lee, Matls. Trans. 9 (2006) 2421.
[36] V. Noack, A. Eychmuller, Chem. Mater. 14 (2002) 1411.
[37] V.A. Fonoberov, A. Balandin, Appl. Phys. Lett. 85 (2004) 24.
[38] K.J. Kim, Y.R. Park, Appl. Phys. Lett. 475 (2001) 78.