Optical absorption spectra studies for amorphous Cu2O-Bi2O3 glass system

J O U R N A L OF M A T E R I A L S SCIENCE: M A T E R I A L S IN E L E C T R O N I C S 6 (1995) 409-414

Optical absorption spectra studies for amorphous Cu20-Bi203 glass system M. M. ELKHOLY, Z. I. EL-BADAWY, A. A. EL-ADAWY, L. M. SHARAF EL-DEEN

Physics Department, Faculty of Science, Menoufia University, Shibin EI-Kom, Egypt Infrared absorption spectra were measured in the spectral range 4000-200 cm-1 for the Cu20-Bi=O3 glass system. Strong bands were observed around 436-477, 600-632, 700-715, 810-875, 975-1000, 1550-1610 and 3225-3510 crn -1 which are due to harmonics of the Bi-O-Bi stretching vibration, Cu-O stretching vibrations, O-Bi-O stretching vibrations, O-Bi-O bending vibrations, Bi-O stretching vibrations, free H=O normal-mode bending vibrations and free H20 molecules or OH- ions, respectively. Quantitative justification of these absorption bands shows that the values of the experimental wave number for most recorded absorption bands agree well with the theoretical ones. The optical absorption spectra were recorded for the same glass system in the spectral range 300-700 nm. These records show that the absorption edge has a tail extending towards lower energies. The edge shifts towards lower energies with increasing Cu20 content. This shift is mostly related to the structural rearrangement and the relative concentrations of the glass basic units. By increasing the Cu20 content, the optical energy gap decreases, while the width of the localized states increases.

1. Introduction Optical absorptions in solids and liquids occur by various mechanisms, in all of which the photon energy will be absorbed by either the lattice or by electrons, where the transferred energy is conserved. The lattice (or phonon) absorption will give information about the atomic vibrations involved and thus the absorption of radiation normally occurs in the infrared region of the spectrum. The higher-energy parts of the spectrum, particularly those associated with the interband electronic transition, will provide further information about the electron states. Therefore the study of optical absorption and particularly the absorption edge is a useful method for the investigation of optically induced transitions and for getting information about the band structure and the energy gap of both crystalline and non-crystalline materials. There are two kinds of optical transition at the fundamental absorption edge of crystalline and noncrystalline semiconductors: direct and indirect transitions, both of which involve the interaction of an electromagnetic wave with an electron in the valence band. The electron is then raised across the fundamental gap to the conduction band. For a direct optical ~ransition from the valence band to the conduction !band, it is essential that the wave vector for the electiron be unchanged. In the case of an indirect transitioin, the interactions with lattice vibrations (phono~s) take place; thus the wave vector of the electron can change in the optical transition and the moment]um change will be taken or given up by phonon~. In other words, if the bottom of the conduction baOd lies at a different part of K-space from the 0957-4522

© 1995 Chapman & Hall

top of the valence band, a direct optical transition from the top of the valence band to the bottom of the conduction band is forbidden. The general formula for the optical absorption coefficient, a(c0), is given by a(o) = ~ In

(1)

where I0 and It are the intensities of the incident and transmitted light, respectively and L is the thickness of the sample. Hogarth and Hosseini [1] studied the optical absorption near the fundamental absorption edge in V2Os-PzOs and V2Os-PzOs-TeOz glasses. They found that the fundamental absorption of these glasses is dependent upon composition and arises from direct forbidden transitions and occurs at a photon energy in the range 1.9 to 2.6 eV, depending on composition. Infrared absorption spectra of glasses can give valuable information about atomic configurations in glasses even though quantitative analysis is rarely possible. For example, the modifier atoms enter the network of the basic glass or reside interstitially and the extent to which the double bond in phosphate networks, fo r example, is broadened, can be qualitatively assessed. Experimental and theoretical investigations of the vibrational spectra of vitreous solids have been undertaken by some investigators [2-4]. Several papers have dealt with the infrared absorption spectra of binary phosphate glasses, using the infrared spectroscopy technique. Attempts to apply this technique to the study of the structure of glassy materials were made by S u e t al. [5]. Quantitatively, the vibrational

409

TAB L E I Midband position (cm-1) of the absorption bands for the CuzO-Bi203 glass system Composition (mole %)

Bi-O-Bi (stretching)

Cu O (stretching)

O Bi-O (stretching)

O-B~O (bending)

Bi-O (stretching)

H20 (normal mode)

O H - ion (stretching

Cu-10% Cu-20% Cu-30% Cu~40% Cu-50%

441 477 445 438 436

632 610 599 610 605

698 704 703 704 715

863 876 810 877 866

975 989 1000 1000 989

1549 1568 1551 1555 1610

3224 3226 3404 3448 3511

spectra of several glass systems, including silicates [5], borates [6] and metaphosphates [7] were analysed, based on either the "localized model" or the "lattice model" as summarized by Borrelli and Su [8]. It is well known that oxide glasses containing transition metal ions (TMI) such as V, Fe, Cu, etc., are of great interest because of their technological applications [9-10]. Among the various oxide glasses with T M I such as copper, vanadium and iron, the physical properties of vanadium glasses have been the most extensively studied [11 14]. To the best of our knowledge, the preparation and characterization of the CuzO-Bi203 glass systems have riot yet been reported. The present paper reports the preparation of a new C u 2 0 - B i 2 0 3 oxide glass system and an investigation of its optical properties.

2. Experimental procedure 2.1. Glass preparation For the present .studies, binary Cu20-BizO3 glass samples were prepared with different ratios of C u 2 0 (see Table I). For each glass, the thermal history (i.e. melting temperature, melting time, annealing temperature and time, and all preparation conditions are as similar as possible. The binary C u 2 0 BizO3 glasses were prepared by melting the appropriate mixtures of bismuth oxide Bi203, of 99.99% purity, and copper oxide CuzO of 99.99% purity using open alumina crucibles heated in an electric furnace .open to the atmosphere. The mixture was heated f~st at 400 °C for an hour and then transferred to a second furnace and held at 900-950 °C for an hour. The glass melts were stirred occasionally with an alumina rod, to ensure homogeneous melts. The highly viscous melt was cast in a cylindrically shaped split-mou}d of mild steel. The glass produced was annealed at 350°C in a second furnace for an hour. Finally, the furnace was switched off and the glass allowed to cool in situ for 24 h. All the prepared samples were checked by X-ray diffraction; no diffraction lines were detected confirming the glassy state of these samples.

2.2. Infrared absorption spectra measurements (In) The infrared absorption spectra of the glass samples studied were recorded using a Perkin-Elmer doublebeam 598 spectrophotometer in conjunction with the

410

~2 <

L O

L i

4000

i

3000

i

i

i

i

r

2000

1500

1000

500

200

W a v e n u m b e r ( cm -1 )

Figure i Infrared absorption spectra for the Cu20-Bi203 glass system.

KBr disc technique, over the spectral range 4000-200 c m - 1 at room temperature. Glass powdered samples of 4 mg were thoroughly mixed and ground with 200rag KBr; after which the mixtures were pressed at 10 t c m -2 for 5 min under vacuum.

2.3. Ultraviolet and visible absorption measurements The optical absorption spectra in the visible and nearultraviolet region were recorded at room temperature. These curves were traced for highly polished glass samples of ~ 3 mm thickness using a PerkinElmer 402 double-beam spectrophotometer in the wavelength range 190-900 nm.

3. Results and discussion 3.1. Infrared absorption spectra studies Fig. 1 shows plots of the infrared absorption spectra for C u 2 0 - B i 2 0 3 glasses with different ratios of C u 2 0 in the spectral range 4000-200 c m - 1. The positions of the absorption bands of CuzO-Bi203 glasses are listed in Table I with specification of their attributed vibrational modes. For all CuzO-BizO3 glasses presented in this work, the spectra showed strong bands around 435-475, 600-632, 700-715, 810-875, 975-1000, 1550-1610 and 3225-3510 cm-1 which may be due to harmonics of the Bi-O-Bi stretching vibration, the C u - O stretching vibration, the O - B i - O stretching vibration, the

2000

4000

~I~Cu-O

Cu-O OOOOO Bi-O

00¢00 Bi-O ***** H20

*****

H20

1600 3000

A

T E 1200

tO

2000

813. "(3

800

"0

1000

0--

-0

O

0

0

[3

[3

El

400 D

O

llll[llllll

0

I111111111 I I I I I I I I I I I

10

20

II1111 I I I I I I I I I h l l l l l l l l

30

40

50

0

60

1:3 -

I I T I ' I I I S II11111111 I l l II II I I I I I I I l l I l l l

0

10

20

30

II11111111 I I I I I I I]I

40

50

60

Cu20 content ( mole % )

Figure 2 Compositional dependence of the wave number for each attributed band in the infrared spectra of the Cu20 Bi203 glass system.

70

70 Cu-O .OOO0~) Bi-O H20

60 60 50 50

¢-

40 <

40

>.

30 tin

30 20

Bi-O-Bi ~ ' ^ ^ ^ O - Bi - 0 OOO00 O - H

10

0

IIII,ITIIII,II,I,]IIIII,II,I,,,I,III

0

10

20

30

20

,I , , l l , , l l l h l l l l l l l l

40

50

10

60

l] l l l l l l l l 1 1 1 1 1 1 l l l

10

llllHllll

20

l111Hllllll~ll[llllllllllll

30

40

50

I

60

Cu20 content ( mole % )

Figure 3 Variation of the absorption intensity for absorption bands in the infrared absorption spectra of CuzO-BizO3 glasses with composition.

O - B i - O bending vibration, the Bi-O stretching vibration, the free H 2 0 normal-mode bending vibration and the free H 2 0 molecule or O H - ion, respectively. The presence of water molecules in our glasses may arise accidently during the preparation of pellets. Fig. 2 shows variations in the midband (peak position) wave number (cm -1) with C u 2 0 content (mole %) for each attributed band. It is clear that in general the compositional dependence of the midband positions is not pronounced. Fig. 3 shows variations of the absorption intensity for each band with C u 2 0 content; As the C u 2 0 content is progressively increased, the absorption intensity of all attributed bands d~splay their maximum at 30-40 mole % C u 2 0 contenti and beyond that the absorption intensity decreases with increasing C u 2 0 content.

As a result of the addition of up to 35 mole% of C u 2 0 to Bi203, the Cu ions reside in the glass inter-

stitially as network modifiers; by breaking some of the bridging bonds in the glass and replacing them by ionic bonds. The entrance of Cu ions interstitially into the glass network can also be seen in the optical spectra as a shift in the absorption band edge to lower energy. This alteration in structure will cause a decrease in the bond strength and consequently a shift of the various bands to lower energies. In the region 35 50 mole % of Cu20, if the Bi ions enter the network-forming positions, this probably weakens the structure and causes the band to shift to higher energy. If the Cu ions enter the glass network substitutionally, a strong C u - O bond stretching vibration would appear. 411

T A B L E I I Theoretical IR band positions compared with experimental wavenumbers for the stretching force constants for the glass system Cu20-Bi203 Bond

Reduced mass x 10 -z6 (kg U ~ 1)

Bi-O Cu-O O~Cu-O Cu O - C u O-Bi O Bi O - B i

Bond length (nm)

2.47 2_12 1.18 1.77 1.28 2.30

0.219 0.190 0.263 0.307 0_292 0.365

Force constant (N m - 1)

243 294 445 238 384 190

3. 1. 1. Quantitative justification of some absorption bands in Cu20-Biz03 glasses

Experimental wave n u m b e r

Theoretical wave n u m b e r

(cm 1)

(cm-1)

975 600-632 -

992 625 1215

700 715 435-475

790 482

615

30

In this part we attempt to make a quantitative justification for these suggestions. The wave numbers of the vibrational modes of the IR spectra are determined by the mass of the atoms and the interatomic force within the groups of atoms comprising the glass network. The wave number v is given by

q}ll ti f

~ Cu-lO% ocxx~ Cu-20% 0006(> Cu-30~ AAAAA Cu-40}~ ++-F++Cu-50%

20 O eO 4-,

v

=

~-1

=

(2)

f~c

o

where K is the bending or stretching force constant of the bond, c is the speed of light and g is the reduced mass of the molecule, which is given by

11

M1M2 -

M1 +

(3)

M2

0

where M1 and ME are the atomic masses of the two atoms. For a group involving two bonds with respective K1 and K2 force constants, and three atoms of masses M1, M2 and M 3 , w e write K I1-1

:

= M11

K

=

-t- M 2 1

-t-

M~ 1

(4)

~X~Xb] 3/4

1.67 U l -~T-- j

= 5.28 N k - - p - - j

+ 0.3 (rod A-1)

+ 30 (N m-

(5) (6)

where N is the bond order, Xa and X~ are the electronegativities and r is the bond length. If the experimental band-frequency does not fit Equation 2, then the band assignment is wrong and the band may be due to a more complicated form of vibration. The calculated values of v (cm-1) using Equations 2 and 5 as well as the experimental values of v are given in Table lI. Also, Table II summarizes the calculated reduced mass of the cation-anion stretching force constant. F r o m inspection of this table, we found

412

ill[

380 Figure 4 Optical [Cu20]x-[Bi203]

i i i i i

390

....

[

rllll

~

~

400 410 420 Wavelength ( nm )

absorption

spectra

I

18 ....

for

the

I

430

glass system

10o-x.

K 1 -t- K 2

Burger et al. [15] treated the interatomic force separately from the cation stereochemical arrangement where these two factors are independent. However, Higazy and Bridge [16] considered K as the bending or stretching force constant, which Can be calculated from one of the two empirical formulae [17] K

rlrll

that the theoretical values of v agree with the experimental values for all the recorded absorption bands except for the O - B i - O bond where the theoretical wavenumber was found to be higher than the experimental one, suggesting that this bond may be a band combining a stretching motion with a harmonic of the bending motion. Another suggestion is that this vibration might involve a mixed bending and stretching character.

3.2. Ultraviolet and visible spectroscopy studies The optical absorption spectra for binary C u 2 0 BizO3 glasses in the visible and UV range are shown in Fig. 4. This figure shows that there is a relatively sharp absorption edge. The reason for the sharpness of the absorption edge of these glassy samples is not clear now because this is a new type of oxide glass. The position of the fundamental absorption edge shifts to lower energy with increasing Cu20 content. Kordes and Nieder [18] showed a shift of the edge back and forth with increasing alkaline earth content in phosphate glasses. In the present work, the addition

T A B L E I I I Composition, optical energy gap Eo;t, and width of the localized states AE, of the glass system Cu20-Bi203

12.0

: []rn~Dcz Cu-10% - ,x,~zxz~xCu-20% cu-30 10.0 _

--H-II

No.

Cu-40% Cu-50%

A

>

o

CuO2

Bi203

10 20 30 40 50

90 80 70 60 50

Eopt (eV)

AE (eV)

3.10 3.03 3.02 3.01 3.00

0_107 0.100 0.095 0.092 0_080

8.0

1 2 3 4 5

Q

'7 E

Sample composition (%)

6.0

2.0

0.0

:1-/

E

[ .......

2.90

i Ii .......

2.95

It, l,l 3.00

~

.....

I .........

3.05

/

~i . . . . . . . . .

3.10

I I I ' .....

3.15

3.20

P h o t o n e n e r g y ( eV )

Figure 5 Dependence of (czhm)t/2 on photon energy (he)) for the glass s y s t e m [ C u 2 0 ] ~ - [ B i 2 0 3 ] lOO-x-

of C u 2 0 to Bi203 produces broken Bi-O-Bi bonds in the glass network which can be replaced by Cu-O, which is reflected in the absorption spectra by a significant shifting of the absorption edge to higher wavelengths. The shifts of the absorption edge are most likely related to the structural rearrangement of the glass and the relative concentrations of the various fundamental units. The fundamental absorption edge in such solids is less abrupt and well defined than in crystalline nonmetallic solids. In the high-absorption region, Tauc et al. [-19] and Davis and Mott [-20] gave an equation, derived independently, for the optical absorption coefficient a(e0) as a function of photon energy he0: ~(m) = ~B~(h0~- Eopt)"

(7)

where n is an exponent, co is the angular frequency of the incident radiation, B is a constant and Eopt is the optical energy gap of the material. Equation 7 with n = 2, as originally postulated by Tauc et al. [-19] and as predicted by Davis and Mott [20] for indirect transitions, has been found to represent the experimental results particularly at the higher values of absorption at the edge for the present glass system. For indirect forbidden transitions, the quantity (ahco) ~/2 is plotted against photon energy (hm) according to Equation 7. Fig. 5 shows a linear dependence of (ahm) 1/z on photon energy (he0) for the present glass system in the high photon-energy range. Deviation from linearity at low values of the photon energy was observed. Values of the optical energy gap, Eopt, are obtained by extrapolation of the linear region of the plots to (~hm) 1/2 = 0. These values for CuaO-Bi203 glasses are given in Table III. It is clearly evident that, for the present glass system, the values of Eop t slightly decrease with increasing C u 2 0 content. Stevels [-213 has suggested that the movement of the absorption band to lower energy corresponds to

transitions from the non-bridging oxygen, which has a less-tightly bound electron than bridging oxygen, F o r the present studies, the decrease in Eopt with increasing C u 2 0 content suggested that the nonbridging oxygen ion content increases with increasing C u 2 0 content, shifting the band edge to lower energies and leading to a decrease in the value of Eop t. The estimated values of Eop t for the C u 2 0 - B i 2 0 3 glass system are consistent with the reported values in many works [22]. The absorption coefficient ~(c0) near the band edge shows an exponential dependence upon the photon energy he0 and obeys the empirical Urbach relationship: 0t(oJ) =

~oexp(hos/AE)

(8)

where ~o is constant and AE is the width of the localized states in the band tails. The origin of the exponential dependence of the absorption coefficient on photon energy, he0, in the Urbach equation is not clearly known. Tauc and Zanini [23] suggested that it arises from electron transitions between localized states where the density of localized states is exponentially dependent upon energy. Davis and Mott [20] reported that this explanation is not valid for all disordered materials since the gradient of the observed exponential behaviour remains unchanged for many crystalline and non-crystalline materials. Dow and Redfield [-24] suggested that it may arise from random fluctuations of the internal fields associated with structural disorder in many amorphous solids. One possible reason suggested by them is that the gradients of the observed exponential edges obtained from the Urbach equation are very close to each other in many semiconductors. Fig. 6 shows the variation of In ~ with photon energy (ha)) for C u 2 0 B i 2 0 3 glasses. The values of AE calculated from the gradient of the linear part of these curves are listed in Table III. The value of AE for a range of amorphous semiconductors 1-25] lies typically between ~ 0.045 and 0.67 eV. For the glasses investigated in the present study, exponential behaviour is observed and the value of AE varies between 0.08 to 0.11 eV, depending on the glass composition. The dependence of the absorption coefficient a(e)) upon photon energy (he)) suggests that this glass system obeys the Urbach rule. The estimated values of AE and Eop t a r e seen to be small compared with other semiconducting oxide glasses. These small 413

4.0

2.

- ~DDD

Cu-10% : 0 0 0 0 0 Cu-2O%

3. 4.

- O 0 0 C ~ Cu - 3 0 % 3.0 ~-E--H-{- C u - 4 0 % - AAA/VX

0u-50%

5 2.0 6.

/

=

1.0

7.

7

~r//'// ~

0"0

-1.o

/

8.

f

9.

/

10.

?' X

I. S I M O N and H. O. MCHAHON, J. Chem. Phys: 21 (1953) 21 J. REITZEL, ibid. 23 (1955) 2407_ I. SIMON, in "Modern aspects of the vitreous state" (Butterworth, London, 1964). G.J. SU, N. U. BORRELLI and A. R. MILLER,Phys_ Chem. Glasses3(5) (1962) 167_6; N. U. BORRELLI and G. J. SU, ibid. 4 (1963) 206. H. HIRASHIMA, WATAMATE and T. YOSHIDA, J. NonCryst. Solids 95 & 96 (1987) 825. S.K. SHIN and G. J. SU, Proceedings of the 7th International Congress on Glass, Brussels, 1965, paper 48. N. U. BORRELLI and G. J. SU, MaSer. Res. Bull. 3 (1968) 181_ J.D. MACKENZIE, "Modern aspects of vitreous state" (Butterworth, London, 1964). T. YOSHEDA, H. HIRASHIMA and M. KATO, YogyoKyokai-Shi 93 (1985) 244.

11.

A. GHOSHandB. K. CHAUDHURI,Ind. J. Phys. 58A(1984)

12.

Idem, in "Metallic and semiconducting glasses-IF', edited by

62_ --2,0

IIIf~aaai==~J~alaaI

2.91

2.95

3.00

lllllll[ll,iI

IIIlIIIIIlllll

3.05

3.10

3.15

P h o t o n e n e r g y ( eV )

Figure 6 Variation of In ~ with photon energy h0) for the glass system [Cu20]~-[Bi203] 1oo-~.

values confirm the observed sharpness in the absorption edge of these glasses:

4. C o n c l u s i o n

Finally, we conclude that the addition of C u 2 0 to Bi203 improves both the electrical and optical properties of BizO~ glasses, where both the optical energy gap and the width of localized states decrease as a result of C u 2 0 addition.

A. K. Bhatnagar (Trans. Tech. Switzerland, 1986) p_ 515. 13. Idem, J. Non-Cryst. Solids 83 (1984) 151. 14. Idem, J. Mater. Sei. 22 (1987) 2369. 15. H. BURGER, W. VOGEL and K O Z H U K H A R O V , Infrared Phys. 395 (1985) 10; A. A. HIGAZY and B. BRIDGE, J. Mater. Sci. 20 (1985) 2345. 16. A. A. HIGAZY and B. BRIDGE, J. Mater Sei_20(1985) 2345. 17. B. BRIDGE and R. ROUND, J. Mater Sci_ Lett_ 7 (1988) 63. 18. E. KORDES and E. NIEDER, Glastechn_ Ber. 41 (1968) 41 19_ J. TAUC, R. G R I G O R O V I C I and A. VANCU, Phys_ Status Solidi 15 (1966) 627_ 20. E.A. DAVIS and N. F, MOTT, Phil. Mag. 22 (1970) 903_ 21. J . M . STEVELS, Proceedings of the l l t h International Congress on Pure and Applied Chemistry, 1953, Vol. 5, p. 519_ 22. s. K. A1-ANI, C. A. HOGARTH and R. A. E1-MALLAWANY, J. Mater Sei. 20 (1985) 661. 23_ J. TAUC and M. ZANINI, J. Non-Cryst. Solids 23 (1977) 349. 24_ J . D . DOW, and D. R E D F I E L D , Phys_ Rev. B5 (1972) 594. 25. A. ABDEL-KADER, A. A. HIGAZY and M. M. ELKHOLY, J_ Mater. Sei. Mater. Elec. 2 (1991) 204

References 1.

414

C . A . H O G A R T H and A. A. HOSSEINI, J_ Mater. Sci_ 18 (1983) 2697.

Received24 May 1994 and accepted 17 January 1995