See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/280483259

Electrical and optical properties of copper nitride thin films prepared by

reactive DC magnetron sputtering

Article in Acta Physica Slovaca · March 2001

CITATIONS READS

7 212

3 authors:

Milan Mikula Emil Pinčík

Slovak University of Technology in Bratislava Slovak Academy of Sciences
118 PUBLICATIONS 1,097 CITATIONS 147 PUBLICATIONS 684 CITATIONS

SEE PROFILE SEE PROFILE

Dalibor Buc
Slovak University of Technology in Bratislava
40 PUBLICATIONS 362 CITATIONS

SEE PROFILE

All content following this page was uploaded by Emil Pinčík on 27 July 2015.

The user has requested enhancement of the downloaded file.

acta physica slovaca vol. 51 No. 1, 35 – 43 February 2001

ELECTRICAL AND OPTICAL PROPERTIES OF COPPER NITRIDE THIN FILMS

PREPARED BY REACTIVE DC MAGNETRON SPUTTERING

M. Mikula 1
Faculty of Chemical Technology, Slovak University of Technology, 812 37 Bratislava, Slovakia

D. Búc
Faculty of Electrical Engineering and Information Technology, Slovak University of
Technology, 812 19 Bratislava

E. Pinčı́k
Institute of Physics, Slovak Academy of Scioence, 842 28 Bratislava, Slovakia

Received 30 June 2000, in final form 10 October 2000, accepted 7 November 2000

Copper nitride films were deposited by reactive DC magnetron sputtering onto glass and other
substrates in nitrogen atmosphere. Thickness, deposition rate, refractive index and optical
band gap energy were determined from reflection and transmission spectroscopy depending
on the deposition pressure and power. The crystallographic nature of films was investigated
by x-ray diffraction. The films were further characterized by photoelectron spectroscopy
(XPS) and multiple internal reflection (MIR) infrared spectroscopy, where the presence of


Cu N nanocrystalline structure was confirmed. Considerable changes of DC conductivity
and activation energy of the films before and after annealing at 500 C were found.

PACS: 72.20.-i, 78.66.-w, 52.75. Rx

1 Introduction

Technologically important binary nitrides BN, AlN, Si N , WN, TiN known as hard and stable
compounds are rather well characterized. On the other hand covalent metal nitrides as Sn N ,
Cu N and Ni N are less known but potentially important because of their semiconductor proper-
ties and an ability to release nitrogen at higher temperature [1, 2]. They can be locally metalized
by a laser beam [3] which could be utilized in electronic industry, printing technology or in other
fields.
Cu N has a cubic anti-ReO type crystal structure with a lattice constant of 3.815 Å. Cu
atoms do not occupy perfectly the sites in (111) plane, so the crystal structure has many vacant
interstitial sites that can be filled by contaminants with remarkable changes of the electrical and
 Presented at the Workshop on Solid State Surfaces and Interfaces II, Bratislava, Slovakia, June 20 – 22, 2000.
1 E-mail address: mmikula@cvt.stuba.sk


0323-0465 /01 c Institute of Physics, SAS, Bratislava, Slovakia 35
36 M. Mikula et al

optical properties [4]. The metastable phase of Cu N was described by Blucher et al. [5], and
epitaxial films were prepared on different substrates by rf sputtering [4, 6].

Chemically prepared Cu N (reacting CuF with NH F) has the density 5.84 g/cm and a dark
green color. Thermal decomposition of Cu N occurs gradually above 350  C. For the films [2, 7]
it occurs in maximum extent at about 465  C. Electrical, optical and mechanical characteristic of
sputtered Cu N films were studied in [8, 9].
However, the properties of copper nitride films, considering their thermal instability and
sensitivity to contaminants, are not fully understood. In this study, the electrical and optical
properties of sputtered copper nitride films were studied as a function of nitrogen pressure and
discharge power.

2 Experimental

Copper nitride films were deposited using a laboratory-made DC magnetron sputtering system,
with a water cooled Cu target 100 mm in diameter, 2 mm thick and nominally 99.9% pure. The

with a background pressure of 2.10  Pa. Deposition was conducted in pure nitrogen flow up
target-to-substrate spacing was 80 mm and the sputtering chamber was evacuated by oil pumps

to 0.6 cm /s, maintaining constant pressure (from 0.85 to 4.5 Pa), discharge voltage (280 to
400 V) and discharge current (from 0.35 up to 2.5 A), Table I. No external heating was applied
to the holder with different substrates (Silicon wafers, glass, Al foil and some thinner films on
polypropylene foils). Before deposition the target was sputter cleaned to remove oxides and
contaminants.
Table I. Process and material parameters of deposited copper nitrides.

N pressure N flow
Sample
[Pa]

[cm /s]
Power
[W]
Dep. time
[min]
Energy
[kWs]
Thickness
[nm]
Dep. rate
[nm/s/kW]
1 0.85 0.42 125 10 75 230 3.1
2 1.5 0.45 140 9 75 270 3.6
3 2.3 0.47 105 12 75 205 2.7
4 2.3 0.47 210 6 75 160 2.1
5 2.3 0.47 315 4 75 140 1.9
6 3.5 0.54 113 11 75 120 1.6
7 4.5 0.58 300 6 110 155 1.4
8 4.5 0.58 625 4 150 355 2.4
9 4.5 0.58 1000 2 120 320 2.7
The amount of deposited nitride was established (ex situ) by weighting the coated specimens
on a microbalance and locally measuring the thickness by profile-meter Dektak 3030, Veeco
Instruments Inc. and also by interference of light in visible region. The reflection spectra at
two different angles of incident light (18  and 40  ) were used to determine the refractive index
of the nitride films [10] and then the spectra at small incident angle (18  ) for local thickness
determination. Reflection spectra, as well as the absorption ones were measured by a Pye Unicam
spectrophotometer 8800 with reflectance accessories.
Microstructure of the as-deposited films was studied by x-ray diffraction with  scans
using Philips PW 1820 powder diffractometer with CuK ,  nm radiation operating at
 Electrical and Optical Properties of Copper Nitride Thin Films 37

vd [nm/s/1kW]
~100 W
3
1000 W
625 W

2 210 W
315 W

1
1 2 3 4
pN2 [Pa]

Fig. 1. Pressure dependence of the normalized deposition rate with the approximation curve for 100 W of
discharge power (solid symbols). Open symbols and arrows indicate a change of the deposition rate with
the power.

1.6 kW (40 kV, 40 mA) with accuracy 0.015  of 2 . In order to collimate the beam a set of slits
and masks were used. The CuK radiation was attenuated by a Ni foil filter between the x-ray
source and the sample. We used a 0.5  divergence slit and a 2 mm brass mask, to limit the size
of the x-ray beam on the sample. In front of the detector there was a 0.2 mm antiscatter slit and a
0.5  divergence slit followed by a Ge-crystal monochromator. This arrangement eliminates any
K that has passed through the Ni-filter and also reduces fluorescence effects from the sample.
Chemical composition of the films was studied by multiple attenuated total reflection (ATR)
IR technique (also called MIR IR - multiple internal reflection) by spectrophotometer Perkin
Elmer using 45  TlJ/TlBr crystal (KRS-5) in MIR accessory and two samples were tested also
by x-ray photoelectron spectroscopy (XPS) Microlab 310-F system with a non-monochromated
MgK (1254 eV) x-ray source.
DC resistivity of the films was measured by a two electrode method in vacuum within the
temperature scale, from 25 to 170  C, to determine the activation energy of conduction, "! . The
electrodes were deposited onto the nitride film by vacuum evaporation of Al, positioned parallel,
with 8 mm distance and 75 mm long.

3 Results and discussion

The color of as-prepared films was dark reddish-brown (influenced by interference namely in
the thinner case) and their thickness varied between 0.12 and 0.36 # m (Table I). Calculated
deposition rate is normalized to 1 kW of discharge power (Fig. 1). However, it is not fully
independent of deposition power as shown by the arrows. At 100 W of the discharge power
deposition rate decreases with nitrogen pressure above 2 Pa. Below 1.5 Pa the nitridation is
limited, deposition rate achieves a maximum around 1.5 Pa because of increasing incorporation
of nitrogen into the film.
The refractive index , was calculated from the shift of interference pattern extremes of the
38 M. Mikula et al

7,0 15

6,5
10 40°

n 6,0 R%

5 18°
5,5 2.3 Pa
0
5,0 400 500 600
λ , nm
700 800

4,5
4.5 Pa
4,0

3,5

550 600 650 700 750

λ , nm

Fig. 2. Wavelength dependence of the refractive index of copper nitride films deposited on glass at two
different pressures. The inset illustrates the shift of reflection spectra at different angle of incident light (for
sample 4).

5,0
315 W
210 W
4,5
~100 W

n
4,0 625 W
1000 W

3,5

at 700 nm
3,0
1 2 3 4
pN2 [Pa]

Fig. 3. Nitrogen pressure dependence of refractive index at 700 nm of the films prepared at 100 W of
discharge power (the curve) and at higher power (the arrows).

reflection spectra of nitride films deposited on glass and Si substrates by changing the angle of
incident light, Fig. 2. The points calculated from the shifts of the extremes are connected with
the cubic spline approximation curve. The index increases rapidly below 600 nm due to strong
absorption of nitride films. The pressure and discharge power dependence of the index around
700 nm is shown in Fig. 3, where the experimental points responding to films prepared at 100 W
are connected by an approximation curve. (The arrows and open symbols indicate a change of
the index with the deposition power.) The values of $ in the inset of Fig. 2 are just relative
reflectance (measured values). So the values of $ are rather low, considering the high refractive
 Electrical and Optical Properties of Copper Nitride Thin Films 39

T%
3 7
60
8
9
40
9
8
20
2

1/2
(Ahν)
0
500 600 700
λ , nm
7
1

0
1,6 1,8 2,0 2,2 2,4 2,6
hν [eV]

Fig. 4. A plot of %'&)()*,+.-

vs. photon energy, for band gap energy determination. Inset: Transmission
spectra of sample set (7, 8, 9) on glass.

~100 W
2,4
210 W
315 W
direct 625 W
2,2 1000 W
Eg [eV]

~100 W
210 W
2,0 315 W
625 W
1000 W
indirect
1,8

1 2 3 4
pN2 [Pa]

Fig. 5. Optical band gap energies of nitride films prepared at 100 W, for direct and indirect transitions (solid
symbols and curves) vs deposition pressure of nitrogen. Open symbols and arrows indicate the gap change
with the deposition power.

index of the film. Absolute values would be approximately 2.3 times higher.
The band gap energy 0/ was derived from the absorption spectra of nitrides coated on glass
using the Tauc plot for indirect (Fig. 4) an also direct transitions in crystalline semiconductors
[11],
132413576987576,: <;>=?A@CB 138D5 < / 69E
where
8D5 is the photon energy, 2 is the absorbance and FGHJIK
and 2 for the indirect and direct
transitions, respectively. Band gap energies found at different pressure and deposition power
(in Fig. 5, where the points are obtained from Tauc plots, responding to 100 W of deposition
power are connected with the spline approximation curve) are generally higher in comparison
40 M. Mikula et al

80 shifted baseline (Al substrate)

60 Cu3N band

T[%]
40

20

1400 1200 1000 800 600 400

-1
wavenumber, cm

Fig. 6. IR transmission spectra of copper nitride film on Al substrate (sample 4) taken by MIR technique
with KRS-5 crystal. (T is transmitance.)

Fig. 7. XRD pattern of copper nitride film (355 nm) on glass.

with values reported for the Cu N phase (in [4] just for direct and in [9] for the indirect transi-
tions). Higher / can be caused by a presence of other phases than Cu N, amorphous phases
and impurities.
The presence of the Cu N phase in our copper nitride films was confirmed by IR and XRD
spectroscopy. An absorption band around 640 cm ML was clearly found in the IR spectra of all
films except those of samples 1 and 7 (Fig. 6). It is in agreement with and very close to a value
of 643 cm ML of Cu N single band reported by L. Maya [1]. In the spectrum of sample 7 the band
was small, and there was no band observed in sample 1.
The XRD patterns (Fig. 7) clearly exhibited (100) and (200) peaks of Cu N structure reported
 Electrical and Optical Properties of Copper Nitride Thin Films 41

Fig. 8. XRD pattern of copper nitride films (samples 7, 8, 9) on glass.

930 (Cu 2p)

532.4 (O 1s)
200000

399.0 (N 1s)
Intensity (a.u.)

285.2 (C 1s)

sample 3
0
1000 800 600 400 200 0
Binding Energy (eV)

Fig. 9. XPS spectrum of copper nitride on glass (sample 3)

in [4, 7] (d-spacing 3.815 Å. and 1.905 Å., resp.). The narrow singular peaks would indicate
highly textured films. However, the peak 100 lies on a much broader peak (up to 35  ) that is
relatively higher in the case of films prepared at lower pressures and powers. This may imply
that the crystallinity of the films is far from perfect that could be expected when the growth
occurred on an amorphous glass substrate at low deposition temperatures. Films are composed of
crystallites in nanometer scale. High pressure and high deposition rate cause a higher temperature
of substrate and cleaner process so the films of samples 8 and 9 are more ordered (Fig. 8).
The XPS spectrum of sample 3 (Fig. 9) on glass substrate indicates, except Cu, a low content
of N and also a considerable presence of O, and C. The spectrum is taken from a very thin surface
layer (about 6 nm) of the nitride. This can be caused by the instability of copper nitride or by
contamination of films by pump oils and residual oxygen.
42 M. Mikula et al

sample 8
-2 after annealing

-1
Eσ = 0.13 eV

log σ [Ω cm]
-3

before annealing
Eσ = 0.42 eV
-4

0,0027 0,0030 0,0033

1/T[K]

Fig. 10. Electrical conductivity of sample 8 before and after annealing, as a function of inverse temperature.

1000
315 W ~100 W
210 W
100 before annealing 1000 W
625 W
ρ [kΩ cm]

10

315 W
1

210 W 1000 W
~100 W 625 W
0,1 after annealing

0,01
1 2 3 4
pN2 [Pa]



Fig. 11. DC resistivity of the films at 25 C as a function of deposition pressure, before and after annealing
(up to 500 C). (Solid symbols and curves belong to 100 W of deposition power, open symbols and arrows
to higher power.)

The nitride films were rather stable1 6measuring the DC conductivity in the temperature range
(25  C - 170  C). The plot of NO=PRQ JIS is almost linear (see the line in Fig. 10), so the activation
energy of conduction ( "! ) can be calculated by the Arrhenius equation. The samples were
annealed in an air furnace for 20 minutes at 500  C to cause thermal decomposition of films and
to release nitrogen. The changes of the conductivity and its activation energy after annealing
were expressive. DC resistivity at the room temperature has fallen down by several orders of
magnitude. The activation energy decreased also remarkably, however, it remained at a low
positive value. It means the films did not change completely to a metal. Some films, prepared
at lower pressures and higher power (sample 1, 4, 5) were a little destructed after annealing, by
internal stress in the films.
 Electrical and Optical Properties of Copper Nitride Thin Films 43

The changes of DC resistivity (Fig. 11) at room temperature before and after annealing of
the films show that from the point of view of insulator/conductor transformation and of internal
stress, the optimal films are the ones prepared above 3 Pa. (The spline approximation curve in
Fig. 11 connects the experimental points responding to deposition discharge power of 100 W.
The arrows and open symbols indicate the resistivity of films prepared at higher powers than
100 W.)

4 Conclusion

Copper nitride films were deposited by DC magnetron sputtering in pure nitrogen onto glass, Si
Al and polypropylene substrates without any external heating. The thickness, deposition rate,
refractive index and optical band gap energy, determined by weighting and optical reflection and
transmission spectroscopy, depend sensitively on the deposition pressure and power. Sufficient
nitridation occurs in our conditions above 2 Pa. The films were characterized also by XRD, and
MIR IR spectroscopy, where the presence of Cu N nanocrystalline structure was confirmed. X
ray diffraction measurements showed a texture dominated (100) oriented grains. However, XPS
pattern of one sample indicates the presence of oxygen, carbon and other phases. Considerable
changes of DC conductivity and activation energy of the films before and after annealing at
500  C were found. The best results (the highest conductivity changes and compact films without
stress also after annealing) were achieved above 3 Pa. The conductivity is very sensitive to
plasma discharge and gas flow conditions during preparation that must be further precisioned.
Acknowledgments This study was supported by Slovak Scientific Grant Agency, VEGA, Mi-
nistry of Education of the Slovak Republic, No. G 1/6156/99 and 1/7614/20.

References

[1] L. Maya: J. Vac. Sci. Technol. A 11 (1993) 604

[2] L. Maya, M. Paranthaman, F. List, R. J. Warmack: J. Vac. Sci. Technol. A 15 (1997) 2807
[3] M. Asano, K. Umeda, A. Tasaki: Jpn. J. Appl. Phys. 29 (1990) 1985
[4] T. Nosaka, M. Yoshitake, A. Okamoto, S. Ogawa, Y. Nakayama: Thin Solid Films 348 (1999) 8
[5] J. Blucher, K. Bang, B. C. Giessen: Mater. Sci. Eng. A 117 (1989) L1- L3
[6] S. Terada, H. Tanaka, K. Kubota: J. Cryst. Growth 94 (1989) 567
[7] Z. Q. Liu, W. J. Wang, T. M. Wang, S. Chao, S. K. Zheng: Thin Solid Films 325 (1998) 55
[8] T. Murayama, T. Morishita: J. Appl. Phys. 78 (1995) 4104
[9] F. Fendrych, L. Soukup, L. Jastrabı́k, M. Šı́cha, Z. Hubička, D. Chostová, A. Tarasenko, V. Studnička,
T. Wagner: Diamond and Related Materials 8 (1999) 1715
[10] M. Mikula, J. Blecha, M. Čeppan, J. Panák: Czech. J. Phys. 45 (1995) 1125
[11] M. Mikula, M. Čeppan, J. Kindernay, D. Búc: Czech. J. Phys. 49 (1999) 393

View publication stats