the zinc vacancy (VZn) level. The detected Zn (1 0 1)
peak (Fig. 1) implied that interstitial zinc (Zni) should
be the main defects in ZnO film deposited under low
oxygen pressure (because of insufficiency of oxygen).
The concentration of the interstitial zinc (Zni) and
zinc vacancy (VZn) defects in ZnO films depended
strongly on annealing temperature and oxygen
pressure. In present experiment, the annealing
temperature (550 8C) is higher than the melting
temperature (419 8C) of metal Zn. Melted metal is
in short-range order. It can be considered that there
were considerable interstitial zinc (Zni) and zinc
vacancy (VZn) defects in the liquid metal zinc. After
the oxidation of Zn to form ZnO, because the
oxidation temperature comparing to the melting
temperature of ZnO became very low (TO/
Tm(ZnO) = 823 K/2359 K = 0.37), the diffusion velocity
of zinc atoms or zinc vacancy in ZnO should be
slower. Therefore, the interstitial zinc (Zni) and zinc
vacancy (VZn) defects may be greatly or partly
reserved in ZnO film.
As to the influence of oxygen pressure on the
defects in ZnO film, according to Lin’s report [18],
during oxidation, the variation of these two defects of
Zni and VZn with the oxygen pressure (PO2 ) can be
expressed as:
12
O2 ¼ VX
Zn ¼ OX
O; . . . ; ½VX
Zn/P1=2
O2
(1)
Zni þ 12
O2ðgÞ ¼ ZnZn þ OO; . . . ; ½Zni/P1=2
O2
(2)
As oxygen pressure increases, Eq. (1) indicates that
the concentrations of VZn in ZnO films increases,
while Eq. (2) indicates that the concentrations of
Zni in ZnO films decreases. Experimentally, as oxygen
pressure increases, the decrease of Zni defects and the
increase of VZn defects in ZnO film result in the shift
of the emission peak from 2.91 to 3.0 eV, since the
theoretical predicted energy interval of Zni is 2.9 eV
and that of VZn is 3.06 eV.
It is expected that the existence of Zni in ZnO will
result in volume expansion of ZnO crystal and that of
VZn will result in volume shrinkage. The volume
variation of ZnO crystal can be approximately
predicted by the variations of plane distance of both
(1 0 0) and (0 0 2). Taking the XRD card’s
(PCPDFWIN) plane distance values of (1 0 0) and
(0 0 2) of ZnO as the plane distances of perfect crystal,
i.e. dcard(1 0 0) and dcard(0 0 2), and the values from
XRD analysis of ZnO films as experimental ones,
dexpt(1 0 0) and dexpt(0 0 2), the variation of ratios of
a = dexpt(1 0 0)/dcard(1 0 0) and b = dexpt(0 0 2)/
dcard(0 0 2) as oxygen pressure increases was shown
in Fig. 4. It can be seen that both ratios of a and b are
greater than 1 under the condition of low oxygen
pressure (50 Pa–500 Pa) and they approach to 1 under
the condition of high oxygen pressure (5000–
23,000 Pa). If it is considered that both Zni and VZn
defects all exist in ZnO films, these results indicate
that interstitial zinc (Zni) is dominating at low oxygen
pressure and interstitial zinc (Zni) almost equal to the
Zn vacancy (VZn) at high oxygen pressure. These
results are in agreement with the shift of the PL
emission peak from 2.91 to 3.0 eV when oxygen
pressure increases from low pressure (50–500 Pa) to
high pressure (5000–23,000 Pa).
The defect concentration can also be approximately
estimated. Because of the volume expansion of ZnO
lattice cell caused by Zni and the volume shrinkage by
VZn, the concentration difference of Zni VZn in a
ZnO molecule is:
Zni VZn
ZnO
¼
ðVexpt VcardÞ=VZn
5:5
¼
4
ffiffiffi
3
p
d2ð100Þdcardð002Þða2b 1Þ
5:5VZn
(3)
where Vcard is the ZnO lattice cell volume calculated
from the XRD card and Vexpt the experimental one,
VZn the atom volume of Zn in covalence state. Value
of 5.5 is the number of ZnO molecule in a ZnO