Study of organic surfaces with UPS and Metastable Deexcitation Spectroscopy

Study of organic surfaces with UPS and Metastable Deexcitation Spectroscopy

133 Surface Science 126 (1982) 733-738 North-Holland Publishing Company STUDY OF ORGANIC SURFACES DEEXCITATION SPECTROSCOPY C. REYNAUD, C. JURET S...

330KB Sizes 2 Downloads 51 Views

133

Surface Science 126 (1982) 733-738 North-Holland Publishing Company

STUDY OF ORGANIC SURFACES DEEXCITATION SPECTROSCOPY C. REYNAUD,

C. JURET

Seruice France

des Atomes

Received

de Physique

24 August

WITH

UPS

AND

METASTABLE

and C. BOIZIAU et des Surfaces,

1982; accepted

for publication

CEN-

Saclay,

11 October

F- 91191 Gif-sur- Yvette

Cedex,

1982

We studied a thin polyacrylonitrile film by Metastable Deexcitation Spectroscopy, associated to UPS. Since it is a purely superficial and a non-destructive spectroscopy, MDS appears to be very well suited to the study of organic surfaces. The PAN sample is grafted on a Fe support by electropolymerization, a process which guarantees well-characterized and reproducible samples. The PAN reticulation produced by heat-treatment at T= 200°C appears claerly on the MDS spectra. The layer conductivity is better after the reticulation. The acrylonitrile molecular orbital energies seem to be weakly modified by the polymerization

Metastable deexcitation spectroscopy (MDS) is a well known tool for surface study [ 11. Metastable atoms have thermal energy and do not penetrate into the solid: MDS is a soft spectroscopy (i.e. non-destructive). So MDS is very well suited for organic surface study since organic layers are easily damaged by spectroscopies using electrons, ions of intense X-ray beams. On the other hand, organic layers have properties very interesting for a large range of technological applications, but the origin of these properties is still poorly understood [2]. Basic research on such a material is therefore necessary, including surface study. It has so far been impeded owing to the lack of well-characterized and reproducible samples. We dispose of such a sample. It is a thin polyacrylonitrile (PAN) film grafted on metallic surfaces by electropolymerization [3]. This technique requires drastic experimental conditions but gives reproducible and homogeneous layers having a pure chemical composition and an ordered molecular structure [4] (fig. la). PAN layers have interesting properties due to the presence of the nitrile group (-C = N). Their most interesting feature is a structural transformation produced by heat-treatments of several hours at temperatures above 200°C: the layer reticulation due to intrachain and interchain cyclization with opening of a CN bond (fig. lb). Our purpose is to study this transformation from a surface spectroscopy 0039-6028/83/0000-0000/$03.00

0 1983 North-Holland

C. Reynaud et al. / Organic surfaces

734

b) Fig. 1. Scheme of the polyacrylonitrile tion; (b) after reticulation.

grafted on a metallic surface: (a) after electropolymeriza-

point of view. The experimental set up has been described elsewhere [ 1,5]. The source simultaneously emits metastable helium atoms (23S, 19.8 eV) and photons (2 1.2 ev). Using a pulsed beam and a time-of-flight technique, one can record at the same time UPS and MDS spectra obtained in strictly identical experimental conditions. We report the evolution of MDS and UPS spectra with heat treatments at temperatures from 100 to lOOO*Cperformed on a 500 A thin Pan layer grafted on a Fe support. The comparison between UPS and MDS spectra provides interesting information; in particular it shows which mechanism occurs in MDS. If the Auger Deexcitation (AD) mechanism [I] takes place, MDS is identical to Penning

0

5

10

EK (eV)

15

spectrum of the PAN sample at T= 15O”C, before Fig. 2. MDS (- - -) and UPS ( -) reticulation. The kinetic energy scale (Ek) is valuable only for UPS spectrum. The MDS spectrum is shifted by I .4 eV. Thus the binding energy scale E, ( = 2 1.2 - E,) is available for both spectra. Arrows a, b, c, d and e indicate molecular levels of ref. [9].

C. Reynaud et

al. / Organicsurfaces

135

Ionization Electron Spectroscopy (PIES) and structures appearing in an UPS spectrum will appear in the MDS spectrum at a lower kinetic energy, shifted by the difference between the two incident energies (21.2 - 19.8 eV = 1.4eV) (neglecting the PIES mechanism shift). In the case of a Resonance Ionization + Auger Neutralization (RI + AN) mechanism [l], UPS and MDS spectrum shapes are quite different; their widthes are A(UPS) = hy - $I and A(MDS) = El - 2$, where + is the sample work function and E: the effective neutralization energy of the He+ ion. Sample charging phenomena occur with this sample. They are easily controllable through the surface electrostatic potential variations. Moreover, they permit us to study the sample conductivity. Since they disappear when the sample temperature is above lOO”C, we conclude that it is of phonon assisted hopping type [6]. We observe that the conductivity becomes higher after reticulation, a result interpreted by the existence of conduction pathes (bonds -C=N-C=N-) in this case. This result is very important since conductor organic layers have a lot of technological applications. All the spectra presented here have been recorded at temperatures above 100°C without any sample charging phenomena. We first studied our sample before any heat treatment. The spectra were the same when recorded at sample temperatures equal to 100, 110, 130 or 150°C. In this temperature range, no irreversible change occurred. A comparison

15

10

Es ieV1 .5

Fig. 3. Effect on MDS spectrum shape of heat-treatment at 200°C (- .-), 400°C (). Binding energy scale is such that Ea = 19.8 eV)- E,.

300°C (-

-

-)

and

C. Reynaud et al. / Organic surfaces

-15

-10

Fig. 4. MDS (- - -) and UPS (The Fe Fermi level is indicated.

-5

EF=O ) spectrum

of the sample after thermal

flash at 1000°C.

between UPS and MDS spectra (fig. 2) clearly indicates a PIES mechanism. This result was expected since resonance ionization is not possible owing to the molecular character of PAN layers. The same result has been found for condensed aromatics surfaces [7]. We can see that the MDS spectrum is more detailed than the UPS one. This is due to the escape depth difference. MDS is a pureIy superficial spectroscopy, which is very interesting owing to the complexity of the studied sample. On the other hand, UPS contributions from different depths are mixed, which explains the poor resolution. To assign structures appearing in these spectra, we dispose of an UPS measurement on acrylonitrile in gas phase [8] and of a computation of the molecular orbital energies of this molecule 191. These two works are in fair agreement for two levels, those labeled (a) and (b) in fig. 2: - level (a), corresponding to the C=C bond of the molecule does not appear clearly in our spectra; this bond is certainly the most altered by the polymerization; - level (b) corresponds to the nitrogen lone pair orbital. This level is, on the other hand, the more weakly perturbed. Moreover, it is remarkedly enhanced in PIES spectra recorded with some molecules containing the nitrile group [lo]. At the PAN surface, this orbital points outwards and its overlapping with the He(ls) orbital may be important. So we propose to assign the structure at E, = 13 eV to the nitrogen lone pair orbital. For the other levels, the two works cited above are in complete disagree-

C. Reynaud et al. / Organic surfaces

131

ment. In particular the CN r orbital is measured with a binding energy of 12.36 eV in the UPS experimental work and predicted at 14.28 eV in the theoretical work. This orbital is strongly perturbed by the reticulation process. So let us see the evolution of the MDS spectra shape with heat-treatment (fig. 3) (the UPS spectra are not sensitive enough to indicate anything). The structure centered around E, = 15 eV tends to disappear with heat-treatment up to 400°C whereas the structure at 13 eV remains and finally predominates. So we propose to assign the structure around 15 eV to the CN bond orbital, in better agreement with the theoretical work. After a heat-treatment at 450°C both MDS and UPS spectra change drastically. In the high kinetic energy part of the UPS spectrum, the Fe(3d) band appears [ 111. Since the UPS escape depth o,f a photoelectron with an energy between 10 and 20 eV is equal to about 10 A [ 121, we must affirm that the 500 A thick PAN layer has been destroyed. A well defined peak at 6 eV below the Fermi level (work function of Fe equal to 4.7 eV) may be attributed to carbon atoms remaining at the surface. After a thermal flash at 1000°C (fig. 4), this peak disappears in MDS but remains in UPS. The MDS spectral shape can no more be compared to the UPS one. We believe that the substrate has become an iron carbide and that the RI + AN mechanism takes place with this sample. We find El = 22.4 eV (see above), a value which is in good agreement with the usual ones [l]. In conclusion, we have learnt that heat-treatments at temperatures below 200°C do not produce any irreversible change. On the other hand, at T above 2OO’C and up to 400°C PAN layers undergo a reticulation process which is clearly visible on MDS spectra. The destruction of thin layers begins suddenly at T = 450°C. The assignment of the spectral structures seems to be possible with the computed molecular energy levels. This result, which must be confirmed, shows that acrylonitrile molecules retain their identity in spite of the polymerization. The results presented here show that MDS (associated to UPS) is an efficient tool for studying organic surfaces: MDS is very sensitive to the layer evolution; sample charging phenomena are usable and provide information on the bulk conductivity. Since organic surfaces are worth studying, this spectroscopy would be developed in the future. We want to thank G. Lecayon discussions.

for the PAN samples

supply

and for fruitful

References [l] C. Boiziau, in: Springer Series in Chemical Physics, Vol. 17 (Springer, Berlin, 1981) p. 48, and references therein.

738

C. Reynaud et al. / Organic surfaces

(21 C. Duke and L. Schein, Phys. Today (Feb. 1981) 42. [3] G. Lecayon, C. Reynaud, C. Boiziau, Y. Bouizem, C. Juret and C. Le Gressus, Chem. Phys. Letters, to be published. [4] G. Lecayon, Note CEA N 2181 (1981). [S] J. Roussel, C. Boiziau, R. Nuvolone and C. Reynaud, Surface Sci. 110 (1981) L634. [6] C. Reynaud, A. Richard, C. Juret, R. Nuvolone, C. Boiziau, G. Lecayon and C. Le Gressus, Thin Solid Films 92 (1982) 355. [7] T. Munakata, T. Hirrooka and K. Kuchitsu, J. Electron. Spectrosc. Related Phenomena 13 (1978) 219. [8] R.L. Lake and H. Thompson, Proc. Roy. Sot. (London) A317 (1970) 187. [9] J.B. Moffat, J. Phys. Chem. 81 (1977) 82. [lo] V. Cernak and A.J. Yencha, J. Electron. Spectrosc. Related Phenomena 8 (1976) 109. [ll] M. Pessa, P. Heimann and H. Neddermeyer, Phys. Rev. B14 (1976) 3488. [12] M.P. Seah and W.A. Dench, Surface Interface Anal. 1 (1979) 2.