Photovoltaic properties of polymer films

Solar Energy Materials 2 (1979/1980) 217-228 © North-Holland Publishing Company

PHOTOVOLTAIC PROPERTIES OF POLYMER FILMS P. J. REUCROFT and H. ULLAL Department of Metallurgical Engineering and Materials Science, University of Kentucky, Lexington, KY40506, USA Received 19 October 1979

The effect of metal electrode and film thickness on the photovoltaic energy conversion efficiency in (1:1) mole ratio films of poly (N-vinylcarbazole) (PVK) and 2,4,7-trinitrofluorenone (TNF) has been investigated. Low work function metals increase the Schottky barrier height which leads to increases in the photovoltaic energy conversion efficiency. A ten-fold decrease in film thickness produces a thousandq'old increase in photovoltaic energy conversion efficiency. A theoretical model which assumes that the photovoltaic current is limited by Child's law predicts photovoltaic efficiencies which are in good agreement with the measured efficiencies.

1. Introduction

Organic materials can usually be fabricated inexpensively in large area, thin film form and there has been considerable interest in the possibility of organic solar cells, in spite of their generally low measured efficiencies (1.0% or less). The photovoltaic effect has been measured for anthracene [1], tetracene [2, 3], various phthalocyanines [4-8], chlorophyll [9, 10], a squarylium dye [11, 12], a 1:1 complex of poly (N-vinylcarbazole) (PVK) and trinitrofluorenone (TNF) [13], and a merocyanine dye [14]. The observed phenomena in most of these systems have usually been attributed to the formation of a Schottky barrier photovoltaic cell. The low efficiencies have generally been attributed to low photocarrier generation efficiencies, low carrier mobilities and charge carrier trapping effects. Preliminary theoretical studies on PVK :TNF films indicated that in thin films (<0.1 #m), the photovoltaic energy conversion efficiency will be limited by the photocarrier generation efficiency [13, 15]. The model also predicted that the efficiency will be limited by space-charge-limited conduction in the case of films of thickness greater than 0.1 #m. Some experimental evidence was provided for this in that the experimental photovoltaic energy conversion efficiency showed an inverse cube dependence on the film thickness. Quantitative agreement between experimental and theoretically predicted efficiencies was not obtained, however, and the effect of varying the Schottky barrier height was not determined experimentally. To further test these models for the photovoltaic effect in organic films, additional studies have been carried out on PVK :TNF (1:1) films. Emphasis has been placed upon varying the Schottky barrier height by varying the metal electrode work function and directly 217

218

P. J. Reucro/L H. Ullal / Photoroltaic properties q/ polymer film,s

comparing measured photovoltaic energy conversion efficiencies with calculated efficiencies for each metal system. Polymer film thicknesses in the range 1-15 pm have been investigated.

2. Theoretical photovoltaic energy conversion efficiency The Schottky barrier model described previously, in which the photogenerated current density, Ip, becomes limited by space-charge considerations is assumed [13, 15]. In this model, photons are strongly absorbed by the polymer film in the region of the metal electrode-polymer interface. Excited states are produced which interact at the interface and promote charge separation at the barrier if the absorbed photon energy is equal to or greater than the barrier height. The photon flux is assumed high enough to provide a reservoir of charge carriers at the electrode. The I-V characteristics of a barrier cell can be expressed as: I=/o[exp (eV/kT)- 1]-Ip,

(1)

where Io is the saturation dark current density, V is the voltage, e is the electronic charge, k is Boltzmann's constant and T is the absolute temperature. Assuming negligible carrier trapping in the polymer film,/o and Ip can be estimated from the following expressions :

Ip =9eevpvz/8d 3,

(2)

where e is the film dielectric constant, e~ the permittivity of free space, # the carrier mobility and d the film thickness.

Io = A T E exp [-(~s/kr)],

(3)

where ~ is the electrode barrier and A the Richardson constant (1.2 x 106 A m -z K-2). Substitution of eq. (2) in eq. (1) and applying the condition O(IV)/OV=O, for maximum power [16], yields: (1 +eVmp/kT ) exp (eVmp/kT) = 1 +27e eo 12V2mp/8Iod3= 1 + 3Ip/lo,

(4)

where Vmp is the photovoltage at maximum power. Combining eqs. (4) and (1) yields the following expression for the maximum theoretical radiant energy conversion efficiency, when terms having a small contribution are neglected : ?/TH = ~-Q Vmplo(eVmp/kT) exp (eVmp/kT)P- 1,

(5)

where P is the radiant power density. Vmpcan be determined by solving eq. (4) by iteration employing literature mobility data [17] and using eq. (3) to determine Io. In the present calculations hole mobility data for PVK:TNF (1:1) were employed since experimental data indicated that photoemission of holes takes place at metal interfaces with PVK:TNF [13]. Substitution of Vmpand Io in eq. (5) allows determination of theoretical efficiency for PVK :TNF as a function Of~s and d.

P. J. Reucroft, H. Ullal / Photovoltaic properties of polymer films

219

Fig. 1 shows theoretical maximum photovoltaic conversion efficiencies calculated in this way for films having thickness in the range 1-10/~m and barriers in the range 1.5 to 3.5 eV. The radiant power intensity was 0.7 W m-2. This is the intensity of the radiation that was employed in the experimental measurements and which gave the maximum photovoltaic power output (at), = 5250 ,~). 10-I

10-2I 10-3~-

tO-4~

._o

10.5

E 10-6

/

P;~: 07 Wrn-2

10-7 /

Theorehcol Conversion Maximum Eceifnc Phlovortaic

10-8

tO~ 0

t~5

2JO

2~5 @~ (eV)

3LO

3~5

40

Fig. 1. Theoretical maximum photovoltaic conversion efficiency for P V K : T N F (1:1) as a function of barrier height at T = 2 9 8 K, input intensity=0.7 W m - 2 .

3. Experimental Polymer films (1-15 #m) were prepared by dipping the slide in a tetrahydrofuran solution of PVK :TNF (1:1). The polymer film was cast on the conduction side of ultrasonically clean Nesa glass. The back surface was cleaned off by wiping. Semitransparent (approximately 50~) metal electrodes were deposited at 10- a-10- 6 Torr, depending on the metal being evaporated, employing techniques described pre-

P. J. Reucro/t, H. Ullal / Photovoltaic properties o/polymer.#lm.s

220

viot.~ly [18-20]. The polymer film thickness was determined by capacitance measurements, employing a General Radio 1017-A capacitance bridge. The experimental set-up used for measuring the steadystate photovoltage has been described before [13]. A Keithley 610C electrometer in conjunction with a Keithley 370 recorder was used for the measurements. The light source monochrometer assembly has also been described in an earlier study [13].

4. Results and discussion

A typical current-voltage plot shown for a PVK :TNF (1:1) sample with a Cr electrode at a constant wavelength of 5000 ~ is illustrated in fig. 2. Similar results were obtained for the metal electrodes listed in table 1 and varying film thickness. In all the I - V plots, as the external load resistance decreases the current increases until some optimum value of R is reached. The current obtained under these conditions represents the maximum power output. This optimum value for the various sandwich cells using different metal electrodes and varying film thickness was generally found to be on the order of 10 9 ~'). The maximum power rectangle is also shown in fig. 2. Theoretically, this value is about 80~o of the product of I~ Voc,where Isc is the short-circuit current in amperes and Vo~is the open-circuit voltage in volts. of the cell. Experimental values were generally found to be much lower.

10 -9

I

I

I IIIII

I

I

t

i ittll

i

l

I

t lillt

I

I

t

i rittr

PVK :TNF (1:1) Cr Electrode llluminoted X : 5 0 0 0 ~, i0 -~C Isc

C~

Imp E

owe,

I

b,

I 0 -tt

Rec° I 1 i0 -12 I0 -4

,

,,,,,,,I

, I0 -3

,,,,,,,I

, i0 -2

,,l~,,,,l Vmp

li I0 "1

I IIIII

Voc

V (Volts)

Fig. 2. Plot o f l o g / - L o g V for C r - P V K : T N F - S n O 2 at a wavelength of 5000 ~.

iO°

P. J. Reucroft, H. Ullal / Photovoltaic properties of polymer films

221

Table 1 Electrode metal work functions [16] Element

Symbol

Work function (Ore) (eV)

Platinum Gold Chromium Silver Samarium

Pt Au Cr Ag Sm

5.35 4.75 4.60 4.30 3.30

It was observed that the PVK :TNF (1:1) complex responds mainly in the visible region of the electromagnetic spectrum (5000 to 6000 )~). Earlier studies have also indicated that PVK alone responds mainly in the ultraviolet region and the PVK: TNF complex in the visible region [21]. PVK :TNF absorbs photons in the visible region because of the formation of a charge transfer complex. This complex, in essence, needs less energy (hvct) to bring about the transition from the lower ground state to the higher available energy states (see ref. [22], fig. 11). A structural explanation for this is that one of the hydrogen atoms in the ethylene ( C 2 H 2 ) group has been replaced by a carbazole unit. The carbazole unit requires less ionization energy compared to the hydrogen atom to bring about an electronic transition. It has been observed that the response in the ultraviolet and infrared regions is significantly much lower in comparison to the visible region in all the samples tested. Current-voltage plots, such as the one shown in fig. 2, were used to determine the maximum power output and hence photovoltaic conversion efficiency for all the systems that were investigated. In order to compute the efficiency of the sandwich cell, it was necessary to measure the intensity of the xenon light source. The light intensity was measured at the surface of the photovoltaic specimens by replacing the specimens with an Eppley thermopile. The measured electromotive force was converted to intensity by means of a standard calibration chart supplied by Eppley Laboratories, Inc. The intensity of the lamp source deteriorates slowly with time. After using it for about 90 hours, however, it remained relatively constant. The measured values of the intensity_were in the range of 0.49 to 2.17 Wm -2 covering the spectrum from 4000 to 7500 ,~. Knowing the maximum power, and the intensity at specific wavelengths, which was determined at intervals of 250/~, the efficiency of the cell was computed. Fig. 3 illustrates the variation of the maximum photovoltage versus the metal work function. The maximum photovoltage, in the case of both 3 and 6 ~rn films, decreases as the metal work function increases. Alternatively, the maximum photovoltage increases as the metal electrode barrier heights increases. This is consistent with the model which predicts that the efficiency of the cell increases as the barrier height at the metal electrode-polymer film increases [13, 15]. Similar metal electrode effects have been reported in the investigation of amorphous silicon [23]. Fig. 3 also shows the effect of film thickness. The maximum photovoltage in the case of the 3/~m film is greater than that observed for the 6/~m film by a factor of 3-4. This is further

222

P. J. Reucro/i, H. Ullal / Photo~;oltaic properties oJ polvmer.films

I

~

I

PVK :TNF (1:1) Film Thickness

1.0

A ~3~. 0 ~6/~

09 = 0.8

i

07 Sm

0.6

~ 0.5 Ag

O.4 0.3 0.2

O I

0

I

5.0

4.0

1

I

J

I

5.0 6.0 Metol Work Function (eV)

I

70

Fig. 3. Variation of the maximum experimental photovoltage for PVK :TNF as a function of the metal work function for different film thickness.

evidence that the photovoltaic current shows a power law dependence on film thickness. A typical maximum efficiency versus wavelength plot is shown in fig. 4 for a Sm electrode cell of film thickness 2.96 #m. The maximum photovoltaic energy conversion efficiency varies significantly by changing the illuminated metal electrode and the film thickness. In almost all cases, the peak efficiency is in the range of 5250 ,~. In all the cases investigated, the illuminated metal electrode was connected to the input of the electrometer and the tin-oxide electrode to the ground. All the metal electrodes tested were approximately 50~o transparent and displayed a negative sign relative to the polymer film. This strongly indicates that the hole barrier is operative at the metal electrode-polymer film interface in most samples investigated. In two of the Pt samples a photovoltage of opposite sign was observed, however. Fig. 5 summarizes the maximum experimental efficiency as a function of the film thickness for the various metal electrode cells that have been investigated. In the case

P. J. Reucroft, H. Ullal / Photovoltaic properties of polymer films 10-3

I

I

223

I PVK " TNF(I:I) d = 2 . 9 6 MSm Electrode

Illuminated

,.~ iO-4

-

o

I0'

,

400O

I 500O

,

I 6O00

~

I 7OOO

xc~l Fig. 4. Maximum experimental efficiency versus wavelength for S m - P V K :TNF-SnO2.

of the Sm cells, the efficiency increases from 1.8 x 10-7~/o for a 15 /an film to 1.63 x 10-4Y/o for a 2.0 #m film. This data indicates that for a decrease in one order of magnitude in the film thickness, there is an increase of three orders of magnitude in the energy conversion efficiency. This marked increase in the efficiency can best be explained in terms of the photovoltaic current being limited by Child's Law. In this case, the current is dependent upon the inverse cube of the film thickness. The theoretical modelthen predicts that a tenfold decrease in film thickness will produce a thousandfold increase in photovoltaic conversion efficiency in agreement with the experimental findings. Fig. 5 also shows that there is a strong dependence on the illuminated metal electrode. The metal electrode with high work function and consequently lower metal electrod~polymer film barrier at the interface, as is the case for the Pt cells, have lower photovoltaic energy conversion efficiency in comparison to the metal electrode with low work function and higher metal electrode-polymer film barrier at the interface, i.e., the Sm electrode cells. The metal electrodes with intermediate values of work function, such as Ag, Cr and Au, fall in between these two. The highest experimental conversion efficiency measured was 1.63 x 10-4~o for a Sm

224

P ,I. ReucrolL H. Ullal Photovoltaic properties ~lpolvmerlilms

lO-31rl,l rO_4~_ ~ , ~ lr~,l e i~,--~-- iiiErlI i --i

>"10"5i

K F 10-7~ iO-e,LItl~ I00

" h,llll I I

tO ~-ilm Thickness C~I

It,ltlt I

IJ

OI

Fig.5.Maximumexperimentalefficiencyversusfilmthicknessfordifferentmetalelectrodes. electrode illuminated cell with a film thickness of 2.0 #m. In contrast, the efficiency is about three orders of magnitude less in films with Pt electrodes. Figs. 6-9 were plotted in order to make a more detailed comparison between the theoretical predictions of the model and the experimental data obtained for the various metal electrode cells. The theoretical values in each case were taken from fig. 1, corresponding to the particular electrode barrier (~s). Fig. 6 shows that the theoretical estimates are slightly greater than the experimental values obtained. The results also show that the efficiency is greatest in the case of the highest barrier height, as predicted by the theoretical model. The best agreement between the theoretical and experimental values was obtained in the case of the Ag electrode cell (fig. 7). The maximum conversion efficiency is lower for Ag compared to Sm, however, because of the lower barrier height (1.8 eV). In the case of the Cr and Au electrode cells (figs. 8 and 9), although the general trend is followed, the experimental values are again slightly higher than the theoretical values. These results suggest that the model predicts the general characteristics but still has room for refinement, especially in the case of high work function metal systems. The various barriers that can be operative, which include both electron and hole barriers, are illustrated in an earlier work [13]. The hole barrier at the metal electrode-polymer film interface can vary between 0.7 to 2.8 eV depending on the illuminated metal electrode.

P. J. Reucroft, H. Ullal / Photovoltaic properties of polymer films 10-2

' . . . . .

I' . . . . . . .

225

I ........

10 -3

A

~,

IO4

hi

E E ~ 10.5

PVK

o

iO-e

: TNF(I:I)

Sm electrode ~M =3.30 eV ~s = 2 8 0 eM mTheoretica I o Experimental

#

1 0 - 7 I,, . . . . . . I00

o

I ..... IO Film

, ,

,

Thickness

I ........ I

0 I

(H-)

Fig. 6. Theoretical and experimental maximum efficiency versus film thickness for S m - P V K :TNF-SnO 2.

5. Conclusions

The following conclusions can be drawn from the present work. The maximum photovoltaic energy conversion efficiency is controlled by two critical parameters, namely, the metal barrier height and the polymer film thickness. Low work function metals which establish high metal barrier heights such as in the case of Sm, display the maximum experimental energy conversion efficiency. In the case of a high work function metal such as Pt, the maximum energy conversion efficiency is several orders of magnitude lower than Sm for almost identical film thickness. This is due to the fact that the Pt electrode barrier height is comparatively much lower than the Sm electrode barrier height. Metals with work functions in between the high and low work function metals display intermediate values for the energy conversion efficiency. For the same metal electrode system, the maximum energy conversion efficiency is found to be much higher when thin films are used. Both these findings are in agreement with the predictions of the theoretical model. The model predicts that the photovoltaic energy conversion efficiency is dependent on the inverse cube of the film thickness for films greater than 0.1 #m in thickness. The model also predicts that the maximum energy conversion efficiency increases as the barrier height at the

I0~

"'''

'

'

["'''

' '

'

I"'''

' '

'

10-4 ~_

c e u_ LU

E •~: 10 -6 o

4

o / 10-7'

/

1 4

N F (11) Ag

I

electrode

~M : 4 3o ev

2

-=-Theoretical o Experimental

t

* s : ,8o ev

o/

! '

/

10-8

.......

I00

i0 Film

I Thickness

01 (H-)

Fig. 7. Theoretical and e x p e r i m e n t a l m a x i m u m efficiency versus film thickness for A g - P V K : T N F - S n O 2 104 . . . . . . . . .

t. . . . . . . .

~. . . . . . . .

--

i 0 -5

o

.~

go °°

I 0 "6

LIJ E

.E_ to-7

NF Cr ~M:

/ 10-8

lo

/

i. . . . . . . . IOO

ca:t)

electrode 460

eV

/

~s : i 5 o

I

--Theoretical

I

o Experimental

i. . . . . . . . lO Film

Thickness

ev

1. . . . . . . I

, Ol

(/~)

Fig. 8. Theoretical and e x p e r i m e n t a l m a x i m u m efficiency versus film thickness for C r - P V K : T N F - S n O z

226

P. J. Reucroft, H. Ullal / Photovoltaic properties of polymer films 104 -' . . . . . . .

I. . . . . . . .

227

r. . . . . . . .

10.5

O

~" 10-6 L~J

E E •~

i0 .7

° °°

10 - 8

:TNF(I:I) A u electrode (~M = 4 . 7 5 e V

[

/

I

~ s " 1.55 eV ~'FT~e°retiCa' o Experimental

/ 1° 'o6 ......

hlll,i I0

Film

i

i

I,,,,, I

, ,

,

o,

Thickness (,~)

Fig. 9. Theoretical and experimental maximum efficiency versus film thickness for A u - P V K :TNF-SnO 2.

metal electrode-polymer film interface increases. The maximum experimental photovoltaic energy conversion efficiency of 1.63 x 10-4~ has been obtained for a PVK :TNF (1:1) sample equipped with a Sm electrode and having a film thickness o f 2.0 #m.

Acknowledgements The research described was supported by the National Science Foundation (Grant No. GK-26154) and Ashland Oil Foundation.

References [1] [2] [3] [4] [5] [6] [7]

H. Killesreiter and H. Baessler, Chem. Phys. Lett. 11 (1971) 411. A. K. Ghosh and T. Feng, J. Appl. Phys. 44 (1973) 2781. L. E. Lyons and O. M. G. Newman, Australian J. Chem. 24 (1971) 13. N. N. Usov and V. A. Benderskii, Sov. Phys. Semiconductors 2 (1968) 580. M. I. Fedorov and V. A. Benderskii, Soy. Phys. Semiconductors 4 (1971) 1198. M. I. Fedorov and V. A. Benderskii, Sov. Phys. Semiconductors4 (1971) 1720. A. K. Ghosh, D. L. Morel, T. Feng, R. F. Shaw and C. A. Rowe, Jr., J. Appl. Phys. 45 (1974) 230.

228 [8] [91 [10] Ill] J12] J13] J14] [15] [161 [171 [t8] J19] [20] [21[ [22]

P. ,I. Reucrq/i, H. Ul/al : Photovo/taic properties q/ p(~lvmer /dm~

I-. R. Fan and L. R. Faulkner. J. Chem. Phys. 69 (1978) 3341. C. W. Tang and A. C, Albrecht, J. Chem. Phys. 62 (19751 2139. C. W. Tang and A. C, Albrecht. J. Chem. Phys. 63 119751 953. V. Y. Merritt, IBM J. Res. Develop. 22 (19781 353. V. Y. Merritt and H J. Hovel, Appl. Phys. Lelt. 29 (1976t 414. P. J. Reucroft, K. Takahashi and H. Ullal, J. Appl. Phys. 46 (19751 5218. A. K. Ghosh and T, Feng, J. Appl. Phys. 49 (1978) 5982. P. J. Reucroft, K. Takahashi and H. Ullal, Appl. Phys. Lett. 25 (19761 664. S. M, Sze, Physics of Semiconductor Devices iJohn Wiley, New York, 1969) p. 643. W. D. Gill, J. Appl. Phys. 43 11972) 5033. P. J. Reucroft and S. K. Ghosh, Phys. Rev. B8 (1973) 803. P. J. Reucroft, S. K. Ghosh and D. Keever. J, Polymer Sci. A-2 10 (1972) 2305. P. J. Reucroft and S. K. Ghosh, J. Non-Crystalline Solids 15 (1974) 399. M. Lardom E. Lell-Doller and J. W. Weigl, Mol. Cryst. Liquids Cryst. 2 (1967t 241, P. J. Reucroft. in: Electrical Properties of Polymers, eds. K. C. Frisch and A. Patsis tTechnomic Connecticut, 1972) p. 73. [23] D. E Carlson, IEEE Trans. Elect. Dev., ED-24 t19771 449.