Electronic structure and charge transport mechanisms in polyaniline

Electronic structure and charge transport mechanisms in polyaniline

Synthetic Metals, 29 (1989) E277 E284 E277 ELECTRONIC STRUCTURE AND CHARGE TRANSPORT MECHANISMS IN POLYANILINE. A.P. MONKMAN and D. BLOOR Dept. of ...

315KB Sizes 1 Downloads 111 Views

Synthetic Metals, 29 (1989) E277 E284

E277

ELECTRONIC STRUCTURE AND CHARGE TRANSPORT MECHANISMS IN POLYANILINE.

A.P. MONKMAN and D. BLOOR Dept. of Physics, Queen Mary College, University of London, Great Britain. G.C. STEVENS Central Electricity Generating Board, Leatherhead, Surrey, Great Britain. J.C.H. STEVENS Berkeley Nuclear Laboratories, Berkeley, Gloucestershire, Great Britain. P. WILSON Dept. of Chemistry, Imperial College, University of London, Great Britain.

ABSTRACT We present insitu ESR results along with optical and XPS valence band studies of thin polyaniline (PANi) films.

These results along with comparisons to model

PANi

systems have enabled us to shed new light on the proposed structures and conductive transport mechanisms of PANi. spectral

evidence

polyemeraldine

state

of

We also take this opportunity to present UV/VIS/NIR

semiquinone of

PANi.

radical

From

our

cation

(i.e.

results we

polaron) conclude

presence that

the

in

the

conduction

mechanism in PANi is that of variable range, inter and intra chain hopping between localised charge sites on the polymer backbone.

INTRODUCTION Over the last few years of our understanding of PANi has improved [1,8], however charge storage and transport mechanisms still remain unclear.

To address these problems,

we have undertaken a series of experiments in order to attempt to clarify the situation. From the literature it is clear that the way in which PANi is synthesised has a great bearing on the physical properties of the polymer.

Although we have adopted a single

synthesis route in this work [1] we feel that our results also yield some answers to these questions, which are of general applicability. We like other groups in the field have been using standard techniques, i.e. optical spectroscopy,

conductivity

measurements and

transport mechanisms in PANi.

ESR

studies,

to

characterise

the

charge

However, each of these techniques has short comings.

Certain electronic transitions may not be observed in optical spectroscopy due to the small oscillator strength of the transition [9] and electron correlation effects may distort the spectra.

Surface conductivities as measured by the four point method will be affected by

0379-6779/89/$3.50

© Elsevier Sequoia/Printed in The Netherlands

E278 20 04 N, N r D i p h e n y l

- 1,4-phen ylene-dia

mi n e

16

i 3oo

~\, ',

12

r

i , ~ 700 Wavelength, m/.~

4-

08 ~ \

Semiquinone

form

/// //

04

25o

400

6o0

800

looo

1150

WAVELENGTH (rim)

N,N'Diphe

n yl- 1,4-phe

n yle ne-diimine

Fig. 1. Insitu optical absorption spectra of PANi. Applied oxidation potentials were; A, -0.2 V. B, 0.4 V. C, 0.6 V. D, 0.8 V. E, 1.2 V and F, 1.4 V vs. SCE. Insert shows the optical absorption spectra of trimers I1 ( - - - ) and III ( ) [10].

surface condition and ionic conductivity contributions. could also give rise to erroneous spin signals.

Paramagnetic effects in the material

Thus, in this work we also report XPS

valence band studies and results from model PANi systems found in the literature to assist in the identification of the transport mechanism for PANi.

RESULTS Optical spectroscopy Fig. I shows the optical spectra for PANi during insitu electrochemical oxidation. The insert shows the optical absorption spectra of N , N ' - D i p h e n y l - 1 , 4 - phenylene-diimine and the corresponding semi reduced, semiquinone form [I0].

If one also compares the

insitu spectra with those of Wudl et al from various capped octomer model compounds [8], and taking into consideration the effect of delocalisation in going from a three ring to nine ring, and greater system respectively, identification of the absorption peaks in the PANi spectra becomes possible.

Thus, as stated previously [11 ] the absorption at ca. 340

nm is due to Iocalised exciton formation, as this absorption peak is observed even in the absorption spectrum of N , N ' - D i p h e n y l - l , 4 - p h e n y l e n e - d i a m i n e [10].

The PANi absorption,

which is seen to grow at ca. 840 n m and shift to high energy with oxidation, is due to imine (qninoid) moieties.

The behaviour of this peak is mirrored in the neutral capped

octomer compounds, where it can be seen that addition of one and two quinoids

E279

ii ~Y /I

Z

0 E

0I

SO0

," ~

1000

1500

2000

2500

500

3000

1000

~0

2000

250q

~0

iii

n

0

\

500

1000

\\

I

~00

2000

2500

3000

SO(]

WAVELENGTH

1000

1500

2000

2500

(nm)

Fig. 2. UV/VIS/NIR reflection spectra (in air) of PANi films originally electrochemically oxidised at; i, -0.2 V. ii and iv, 0.4 V. and iii, 0.8 V vs. SCE.

respectively onto the chain causes the observed absorption peak intensity to grow with only a modest bathochromic shift, however by the time four quiniods are present, one observes a large bathochromic shift.

Duke et al [12], have shown that this absorption is due to

the formation of self trapped excitons, centered on the quinoid moieties.

The observed

shoulder at ca. 425 nm in the PANi spectra could well be due to the presence of a small Iocalised

semiquinone population in the film.

Thus we deduce that there are neutral

chains or chain segments in the polymer film. Fig. 2 shows UVIVIS/NIR spectra of PANi films oxidised to various degrees. be clearly seen from the spectra a NIR absorption forms with oxidation.

As can

This peak is

seen to grow at ca. 1200 nm and shifts to ca. 1900 nm in the emeraldine state, on further oxidation this peak is seen to shift hack to high energy.

As it is not known

whether this peak is also observed in the model compounds, identification of the cause of this absorption is hindered.

It may be an intra band gap transition between levels that

converge, but do not close on oxidation.

Although two facts preclude this interpretation.

It should be noted that the wavelength scale is deceptive, and that this absorption peak is as sharp as the others seen in the spectra.

The sharpness is indicative of a transition

between energy levels and not energy bands.

Secondly, it is seen that the absorption has

a large intensity, indicating that it is more likely to be the transition from valence band (VB) to lowest polaron level, and not the transition between the two polaron levels.

3000

E280

t~ ,f"

2

z 0 0. nO (D m <

0 6

115

2,5

"1

0.5

eV

ENERGY Fig. 3. UV/VIS/NIR reflection spectra (in air) of a PANi film originally electrochemically oxidised at 0.2 V vs, SCE. Three distinct polaron absorptions are seen, at 2.5 eV, 1.3 eV and 0.57 eV.

Fig. 3 shows a UV/VIS/NIR spectrum of a thin PANi film which was electrochemically set to a potential of 0.2 V vs. SCE i.e. at the peak of the measured spin signal, see Fig. 4.

In this spectrum one observes the appearance of two new absorption

peaks, one at 2.5 eV (495 nm), and the second at 1.3 eV (950 nm), along with the NIR peaks at 0.57 eV (2170 rim).

We propose that this is the first true spectroscopic evidence

for the presence of semiquinone radical cations i.e. polarons, in a PANi film.

VOLTAGE

re.

Ag/AgCI

Ref.

OXIDATION -,2

Fig.

-.1

4.

O

.1

.2

.3

Insitu,

.4

(V)

REDUCTION .5

.E

.7

repetitive

.7

ESR

.6 .5

.4

scans

.3

.2

(5s

.1

0 -.1

-.2

/sweep)

of

electrochemically oxidised and reduced at a rate of 5 mV/s,

a

PANi

film which

was

E281

Insitu ESR. We have performed insltu ESR experiments (in aqueous electrolyte) in order to observe spin concentrations as a function of film growth and oxidation level.

As shown in

Fig. 4, as the polymer film is deposited an accompanying spin signal is also observed, as charge flows.

As can be seen, we observe a spin signal after the first oxidation peak in

the PANi cyclic-voltammogram, indicating the presence of polarons. It is seen that as the oxidation peak at 0.5 V grows (this oxidation peak is assumed to be caused by polymer degradation) the spin concentration that grows in between the

1.10

r

¢. 0.55

~

U) 14J

÷

0.00 0.50

1.00

1.50

2.00

(Microwave Power) 1/2

2.50 3.00 (mW)1/2

Fig. 5. Microwave power saturation curve for a PANi film which was electrochemically oxidised, and held at 0.2 V vs. Ag/AgCI.

main PANi oxidation peaks increases relative to the level observed at 0.2 V (maximum spin concentration).

Even though charge flows at this oxidation couple, no new spins are

seen to be injected during the oxidation process, thus we believe that spins are trapped at these defect sites, associated with polymer degradation.

Further it is seen that the spin

signal does not vanish when the films are fully reduced, again evidence for trapped spins. The ESR line shaped measured insitu is very asymmetric, whereas that of dry films, which are thicker than the microwave skin depth, are very symmetric. various polar,

Experiments with

nonpolar and ionic solvents have shown that the asymmetry is due to

dielectric shielding caused by the ionic aqueous electrolyte and is not due to Dysonian effects [13].

These results suggest that the spin carriers are not metallic like in nature.

For true metallic carriers, one would expect to observe a Dysonian line shape.

This has

been reported in the case of polythiophene [14]. The spin lattice relaxation time T I , for a PANi film, electrochemically oxidised to a potential of 0.2 V, i.e. at maximum spin concentration, was determined by a standard

E282

saturation technique [15]. lattice

From the straight line graph Fig. 5, we calculate that the spin

relaxation time of PANi in this state to be ca. 10 -6 s.

This indicates that the

spins are strongly coupled to the lattice.

300,~

~

- 1

lOOl o°

o ±

....

30

z~ SODi

~.~,~

_z

~

20

10

0

,' ~,

i

30

20

10

0

iii

~ / ' ~

200

"('~v~

250 -

\,

,'~

,

~

\

30

20

10

O

30

BINDING ENERGY

20

10

O

(eV)

Fig. 6. XPS valence band spectra of PANi films, originally electrochemically oxidised at; i, -0.2 V. ii and iv, 0.4 V. and iii, 0.8 V. vs. SCE.

XPS valence band studies. Fig. 6 shows the XPS valence band spectra of thin PANi films set to various oxidation potentials deposited on Pt. films are shown in Fig. 2. best possible resolution.

The corresponding UV/VIS/NIR spectra of these

The spectra were accumulated over long time periods to gain XPS was chosen over UPS as the

latter

is highly surface

sensitive, and surface pretreatment of the polymer films was deemed impractical. As can be seen from the spectra, above 10eV binding energy there only appears to be shifts in relative peak intensities, thus oxidation does not appear to greatly affect these deep lying states.

The lowest binding energy peak sharpens and shifts to lower binding

energy with oxidation to the emeraldine state, with further oxidation the peak is seen to shift back to higher binding energy.

As can be seen there is at no time an appreciable

density of states at the Fermi level.

However, in the emeraldine state the 2.7 eV binding

energy peak has a low binding energy tail which extends past the Fermi level, thus there is a small density of states at the Fermi level. a bulk metallic state is formed in the polymer.

This is not great enough to indicate that

E283

Conclusions. In this work we have attempted to obtain a consistent understanding of PANi from our results, and those of others. compounds,

we

deduce

the

Comparing PANi optical spectra and those of model

presence

of

lucalised

corresponding to neutral chains or chain segments.

excitons,

and

quinoid

moieties,

It is seen that increasing oxidation,

i.e. increasing the ratio of quinoid to benzenoid structures, causes the quinoid rings to become more localised, on the chain.

probably due to coulomb repulsion as more quiniods are formed

At an oxidation potential of 0.2 V vs. Ag/AgCI we observe the first clear

evidence of semiquinone (polaron) formation.

Three distinct peaks are observed at 2.50

eV (495 nm), 1.30 eV (955 nm) and 0.57 eV (2170 nm).

It is seen that the 0.57 eV

absorption has a far greater intensity than the other two, thus it would seem to be the absorption due to the transition from the VB to the lowest polaron level, and not that between polaron levels.

Upon oxidation it is seen that this absorption, and the other two,

shift to lower energy, and reaches a minimum when the measured spin concentration is at its maximum, i.e. 0.2 V.

This lowering of the polaron level closest to the VB implies

that the polarons become more localised. XPS valence bands studies show that the lowest binding energy peak lies ca. 2.7 eV below the Fermi level even in the emeraldine form of PANi, the tail of the peak does however reach to the surface.

Fermi level,

indicating a

small density of states at the

Fermi

This density is not large enough to be indicative of a bulk metallic state in the

emeraldine form of PANi. Insitu ESR studies show evidence for the formation of semiquinone radicals i.e. polarons.

The spin lattice relaxation time T1

strongly coupled to

the

lattice,

and

10 -6 s, indicating that the spins are

<

probably localised.

observed, suggesting that the spin carriers are not metallic.

No Dysonian line

shape

is

We propose that some spin

radicals become trapped at structural defect sites. Thus we have shown that polarons are formed on oxidation, but as the population grows they become more localised. addition, lattice,

of a to

the

Sum et al [16] have proposed a model whereby the

~r electron density dependent contribution to the elastic energy of the usual

Su-Schrieffer-Heeger

conjugation symmetry and supersymmetry.

model,

leads

to

the

breaking

of

charge

The physical consequences of this are that

oscillator strength is transferred to the VI3 to upper polaron level

transition (UPLT),

increasing the

that

observed absorption intensity of

this transition,

and

the

polaron levels are no longer symmetric about the Fermi level in the gap. shifted downward towards the VB.

localised They are

With greater localisation, more oscillator strength will

be transferred to the VB to UPLT, and the polaron levels will be shifted down further. Thus at 0.2 V we observe maximum spin concentration, maximum localisation of the polarons and consequently a new absorption at ca. 2.5 eV, due to the increased oscillator strength of the VB to UPLT.

This model fits our experimental results very well.

Hence, conduction in PANi will be via hopping from such lucalised sites, leading to the

E284

low maximum conductivity observed ca. 10 S cm-1, and temperature dependence of the conductivity [7].

Note, we have observed in the XPS Ci 2p core level spectra a low

binding energy peak, similar to that seen in metal chlorides, which is dominate in the emeraldine state films [17].

Thus, if the counter ion is of this form we believe that

counter ion pinning will play a major role in carrier localisation. Acknowledgements. We thank GEC Research Ltd., Marconi Research Centre and Berkeley Nuclear Laboratories for use of spectrometers.

I (APM) thank the SERC and CEGB for my

CASE studentship, Dr. D. Urch and Dr. I.A. Howard for such enlightening discussions and Prof. J. Albury for the use of his insitu ESR facilities. References. 1,

A.P.Monkman, D.Bloor, G.C.Stevens and J.C.H.Stevens, J. Phys. D:Appl.Phys.,

2.

R.J.Cushman, P.M.McManus and S.C.Yang, Makromoi.Chem., Rapid Comm.,8(1987)69.

3.

E.M.Geni~s and M.Lapkowski, Synth. Met., 21 (1987) 117.

4.

F.Zuo, M.Angelopoulos, A.G.MacDiarmid and A.J.Epstein, Phys. Rev. B.Rapid

20 (1987) 1337.

Comm., 36 (1987) 3475. A.J.Epstein et al., Synth. Met.,21 (1987) 63. 5.

A.J.Epstein, J.M.Ginder, M.O.Roe, T.L.Gustafson, M.Angelopoulos and A.G. MacDiarmid, Pros.Syrup. on Nonlinear Optical Properties of Polymers,Mat.Res. Soc.Mtg. Boston 11/30-12/5/8.

6.

S.Stafstr6m, J.L. Brddas,A.J. Epstein,H.S. Woo,D.B. Tanner,W.S. Huang and A.G.Macdiarmid, Phys.Rev.Lett., 59 (1987) 1464.

7.

F. Wudl, R.O.Angus, Jr.,F.L.Lu, P.M.Allemand, D.J.Vachon, M.Nowak, Z.X. Liu

8.

F.L.Lu, F. Wudl, M.Nowak and A.J.Heeger, J. Am.Chem.Soc., 108 (1986) 8311.

9.

K.Fesser, A.R.Bishop and D.K.Campbell, Phys. Rev. B, 30 (1984) 4804.

and A.J.Heeger, J. Am.Chem.Soc., 109 (1987) 3677.

10. H.Linschitz, J.Rennert and T.M.Korn, J. Am. Chem. Soc., 108 (1986) 83 11. D.Bloor and A.P.Monkman, Synth. Met., 21 (1987) 175. 12. C.B.Duke, E.M.Conwell and A.Paton, Chem. Phys. Lett., 131 (1986) 82. 13. H.Kodera, J. Phys. Soc. Jap., 28 (1970) 89. 14. H.Scharii, H.Kiess, G.Harbeke, W.Berlinger, K.W.Biazey and K.A.Muller, Synth. Met., 22 (1987) 317. 15. C.P.Poole, Electron Spin Resonance, Wiley, New York, 2nd edn., 1982. 16. U.Sum, K . F ~ r

& H.Buttner, Ber.Bunsenges. Phys.Chem., 91 (1987) 957. U.Sum

et al. J.Phys C:Solid State Phys., 20 (1987) L71.U.Sum et al., Solid State Comm., 61 (1987) 607. 17. A.P.Monkman, D.Bloor, G.C.Stevens and J.C.H.Stevens, in print.